You are on page 1of 511

Text-Critical and Hermeneutical Studies in the Septuagint


Vetus Testamentum

Editor in Chief
Christl M. Maier

Editorial Board
r.p. gordon – j. joosten – g.n. knoppers – a. van der kooij – a. lemaire –
s.l. mckenzie – c.a. newsom – h. spieckermann – j. trebolle barrera –
n. wazana – s.d. weeks – h.g.m. williamson


The titles published in this series are listed at

Text-Critical and Hermeneutical
Studies in the Septuagint

Edited by
Johann Cook
Hermann-Josef Stipp

Library of Congress Cataloging-in-Publication Data

Text-critical and hermeneutical studies in the Septuagint / edited by Johann Cook, Hermann-
Josef Stipp.
  p. cm. — (Supplements to Vetus Testamentum v. 157)
 Text-critical and Hermeneutical Studies in the Septuagint is the title of a bilateral research
project conducted from 2009 to 2011 by scholars from the universities of Munich (Germany)
and Stellenbosch (South Africa). The joint research enterprise was rounded off by a
conference that took place from 31st of August–2nd of September 2011 in Stellenbosch.
It was held in cooperation with the Association for the Study of the Septuagint in South
Africa (LXXSA). Scholars from Germany, the Netherlands, Ireland, Belgium, France, Canada
and the USA, as well as South Africa, delivered papers focusing on the history of the LXX;
translation technique and text history; textual criticism, and the reception of the Septuagint.
 Includes bibliographical references and index.
 ISBN 978-90-04-24078-0 (hardback : alk. paper)—ISBN 978-90-04-24173-2 (e-book : alk.
1. Bible. O.T. Greek—Versions—Septuagint—Congresses. 2. Bible. O.T.—Criticism,
interpretation, etc.—Congresses. 3. Bible. O.T.—Hermeneutics—Congresses. I. Cook, Johann.
II. Stipp, Hermann-Josef, 1954–

 BS744.T488 2013

This publication has been typeset in the multilingual “Brill” typeface. With over 5,100 characters
covering Latin, IPA, Greek, and Cyrillic, this typeface is especially suitable for use in the
humanities. For more information, please see

ISSN 0083–5889
ISBN 978-90-04-24078-0 (hardback)
ISBN 978-90-04-24173-2 (e-book)

Copyright 2012 by Koninklijke Brill NV, Leiden, The Netherlands.

Koninklijke Brill NV incorporates the imprints Brill, Global Oriental, Hotei Publishing,
IDC Publishers and Martinus Nijhoff Publishers.

All rights reserved. No part of this publication may be reproduced, translated, stored in
a retrieval system, or transmitted in any form or by any means, electronic, mechanical,
photocopying, recording or otherwise, without prior written permission from the publisher.

Authorization to photocopy items for internal or personal use is granted by Koninklijke Brill NV
provided that the appropriate fees are paid directly to The Copyright Clearance Center,
222 Rosewood Drive, Suite 910, Danvers, MA 01923, USA.
Fees are subject to change.

This book is printed on acid-free paper.


Preface ................................................................................................................. ix
Abbreviations .................................................................................................... xi
List of Contributors ......................................................................................... xvii

Part One

History of the Septuagint in General

The Pentateuch in Greek and the Authorities of the Jews ................. 3

Arie van der Kooij

The Biblical Canon and Beyond: Theological and Historical

Context of the Codices of Alexandria .................................................. 21
Heinz-Josef Fabry

Part Two

Translation Technique and Text History

Revisiting the Rock: Tsur as a Translation of Elohim in

Deuteronomy and Beyond ...................................................................... 37
Melvin K. H. Peters

Judges 3:12–30: An Analysis of the Greek Rendering of Hebrew

Wordplay ....................................................................................................... 53
Hans Ausloos

B or not B? The Place of Codex Vaticanus in Textual History

and in Septuagint Research .................................................................... 69
Siegfried Kreuzer

Übersetzungstechnik und Textkritik in den Esdrasbüchern:

Hendiadyoin, Doppelübersetzungen und
Wiederholungsvariationen in 1 Esdr ..................................................... 97
Dieter Böhler SJ
vi contents

In Search of the Old Greek Text of 4 Maccabees .................................. 127

Robert J. V. Hiebert

The Relationship between the LXX Versions of Proverbs and Job ..... 145
Johann Cook

An Analysis of the Use of Hebel as a Metaphorical and Symbolic

Device as Interpreted in LXX Ecclesiastes .......................................... 157
Lawrence Lincoln

The θεός and κύριος Terms in the Isaiah Text and their Impact on
the New Testament: Some Observations ............................................ 173
Peter Nagel

Revisiting the Original Greek of Ezekiel 18 .............................................. 193

Harry F. van Rooy

Theology after the Crisis: The Septuagint Version of Daniel 8–12 ... 207
Martin Rösel

Part Three

Textual Criticism

The Value of the Septuagint for Textual Criticism of the Hebrew

Bible as Illustrated by the Oxford Hebrew Bible Edition of
1 Kings ............................................................................................................. 223
Jan Joosten

Gottesbildfragen in den Lesartendifferenzen zwischen dem

masoretischen und dem alexandrinischen Text des
Jeremiabuches .............................................................................................. 237
Hermann-Josef Stipp

Two Difficult Passages in the Hebrew Texts of Lamentations 5:

Text-Critical Analyses of the Greek Translation ............................... 275
Gideon R. Kotzé

Amos 5:26—Überlegungen zur Textkritik, Textgeschichte und

Übersetzung eines schwierigen Bibelverses ....................................... 297
Eberhard Bons
contents vii

Part four


Die Schriftzitate im ersten Christentum und die Textgeschichte

der Septuaginta: Ein Wuppertaler Forschungsprojekt .................... 311
Martin Karrer und Johannes de Vries

Ιουδιθ und Iudith: Überlegungen zum Verhältnis der

Judit-Erzählung in der LXX und der Vulgata ..................................... 359
Barbara Schmitz

PsalmsLXX and the Christian Definition of Space: Examples

Based on Inscriptions from Central Asia Minor ............................... 381
Cilliers Breytenbach

The Testament of Job as an Adaptation of LXX Job ............................... 395

Jessie Rogers

On Kingship in Philo and the Wisdom of Solomon ............................. 409

Jonathan More

The Text Form of the Isaiah Quotations in the Sondergut

Matthäus Compared to the Dead Sea Scrolls, Masoretic Text
and Septuagint ............................................................................................. 427
Gert J. Steyn

Die Rezeption von Jer 38:31–34 (LXX) in Hebräer 8–10 und dessen
Funktion in der Argumentation des Hebräerbriefes ....................... 447
Wolfgang Kraus

Index of Ancient Sources .............................................................................. 463

Index of Subjects .............................................................................................. 489

The heading of this volume is owed to the fact that it arose from a bilat-
eral research programme titled “Text-critical and Hermeneutical Studies
in the Septuagint”, which was conducted from 2009 to 2011 by Johann
Cook, from the Department of Ancient Studies at the University of
Stellenbosch, and Hermann-Josef Stipp, from the Department of Old Tes-
tament at the Faculty of Catholic Theology of the University of Munich.
The project was funded jointly by the National Research Foundation of
South Africa and the Internationales Büro des Bundesministeriums für
Bildung und Forschung (International Bureau of the Federal Ministry of
Education and Research). The shared research enterprise was rounded off
by a conference that took place from 31st of August–2nd of September 2011
at the Wallenberg Research Centre, Stellenbosch Institute for Advanced
Studies (STIAS), and was held in cooperation with the Association for the
Study of the Septuagint in South Africa (LXXSA). Scholars from Germany,
the Netherlands, Ireland, Belgium, France, Canada and the USA, as well as
South Africa, kindly accepted the invitation to deliver papers. Most of the
lectures read on this occasion have been incorporated in the collection.
Moreover, Eberhard Bons contributed an essay although he was unable to
participate in the conference.
As editors of this volume we would like to express our gratitude
towards various persons and institutions. The National Research Founda-
tion of South Africa and the Internationales Büro des Bundesministeriums
für Bildung und Forschung contributed substantially towards the costs of
the research project and the conference. The Universities of Munich and
Stellenbosch offered an environment highly conducive to fruitful research.
The Division of Research Development at the University of Stellenbosch
has been extremely helpful in obtaining financial and other assistance.
A word of thanks to Brigitte Cyster and Tonya Hall from the Department
of Ancient Studies at the US is in order for their unstinting efforts in mak-
ing the arrangements for the congress.
Further, Gert Steyn and Pierre Jordaan in their capacity as executive
committee members of the LXXSA were very supportive throughout this
whole endeavour. We are especially indebted to the participants of the
conference and those who made their contributions available for publi-
cation. We also thank Prof. Christl M. Maier, the editor-in-chief of the
x preface

Supplements to Vetus Testamentum, for accepting the collection for publi-

cation in this series. The publishers Brill, Inc. should also receive a word
of gratitude for publishing this volume. A final word of appreciation goes
to Gideon Kotzé, research assistant of the Department of Ancient Stud-
ies at the US, who took upon himself all the hard work of preparing the
manuscript for publication. Without his professional dedication this vol-
ume would not have been possible.

The Editors, Johann Cook, Dept of Ancient Studies, University of

Stellenbosch, and Hermann-Josef Stipp, Faculty of Catholic Theology,
University of Munich, June 2012.

AASF Annales Academiae scientiarum fennicae

AB Anchor Bible
ABRL Anchor Bible Reference Library
AGBL Aus der Geschichte der lateinischen Bibel (= Vetus Latina: Die
Reste der altlateinischen Bibel: Aus der Geschichte der lateini-
schen Bibel)
AGJU Arbeiten zur Geschichte des antiken Judentums und des
AnBib Analecta biblica
ANRW Aufstieg und Niedergang der römischen Welt: Geschichte und
Kultur Roms im Spiegel der neueren Forschung (eds. H. Temporini
and W. Haase; Berlin: De Gruyter, 1972–)
AnSt Anatolian Studies
ANTF Arbeiten zur neutestamentlichen Textforschung
ASP American Studies in Papyrology
ATSAT Arbeiten zu Text und Sprache im Alten Testament
BA Biblical Archaeologist
BASOR Bulletin of the American Schools of Oriental Research
BBB Bonner biblische Beiträge
BCH Bulletin de correspondance hellénique
BdA La Bible d’Alexandrie
BDAG W. Bauer, F. W. Danker, W. F. Arndt and F. W. Gingrich. Greek-
English Lexicon of the New Testament and Other Early Christian
Literature. 3rd ed. Chicago: University of Chicago Press, 1999
BDF F. Blass, A. Debrunner and R. W. Funk. A Greek Grammar of the
New Testament and Other Early Christian Literature. Chicago:
University of Chicago Press, 1961
BEATAJ Beiträge zur Erforschung des Alten Testaments und des anti-
ken Judentum
BETL Bibliotheca ephemeridum theologicarum lovaniensium
BHK Biblia Hebraica (ed. R. Kittel; Stuttgart: Württembergische
Bibelgesellschaft, 31937)
BHQ Biblia Hebraica Quinta (eds. A. Schenker et al.; Stuttgart:
Deutsche Bibelgesellschaft, 2004–)
BHS Biblia Hebraica Stuttgartensia (eds. K. Elliger and W. Rudolph;
Stuttgart: Deutsche Bibelgesellschaft, 1977)
xii abbreviations

BHT Beiträge zur historischen Theologie

Bib Biblica
BiOr Bibliotheca Orientalis
BIOSCS Bulletin of the International Organization for Septuagint and
Cognate Studies
BiPa.Supp Biblia Patristica Supplément
BKAT Biblischer Kommentar, Altes Testament
BN Biblische Notizen
BRev Bible Review
BWANT Beiträge zur Wissenschaft vom Alten und Neuen Testament
ByzZ Byzantinische Zeitschrift
BZ Biblische Zeitschrift
BZAR Beihefte zur Zeitschrift für Altorientalische und Biblische
BZAW Beihefte zur Zeitschrift für die alttestamentliche Wissenschaft
CBET Contributions to Biblical Exegesis and Theology
CBQ Catholic Biblical Quarterly
CBQMS Catholic Biblical Quarterly Monograph Series
CCSL Corpus Christianorum: Series latina
CIG Corpus inscriptionum graecarum
CIIP Corpus Inscriptionum Iudaeae/Palastinae
CJZC Corpus jüdischer Zeugnisse aus der Cyrenaika
ConBOT Coniectanea biblica: Old Testament Series
COT Commentaar op het Oude Testament
CRAI Comptes rendus de l’Académie des inscriptions et belles-
CSCO Corpus scriptorum christianorum orientalium
CSEL Corpus scriptorum ecclesiasticorum latinorum
DCH Dictionary of Classical Hebrew (ed. D. J. A. Clines; Sheffield:
Sheffield Academic Press, 1993–)
DJD Discoveries in the Judaean Desert
DOP Dumbarton Oaks Papers
DSS Dead Sea Scrolls
EstBib Estudios bíblicos
ETAM Ergänzungsbände zu den Tituli Asiae Minoris
FAT Forschungen zum Alten Testament
FRLANT Forschungen zur Religion und Literatur des Alten und
Neuen Testaments
abbreviations xiii

Ges18 W. Gesenius and H. Donner. Hebräisches und Aramäisches

Handwörter-buch über das Alte Testament, 18th ed. Berlin /
Heidelberg / New York: Springer-Verlag, 2007
GKC W. Gesenius, E. Kautzsch and A. E. Cowley. Gesenius’ Hebrew
Grammar. 2nd ed. Oxford: Clarendon Press, 1910
HAL L. Koehler, W. Baumgartner and J. J. Stamm. Hebräisches
und aramäisches Lexicon zum Alten Testament. Leiden: Brill,
HBS Herders biblische Studien
HKAT Handkommentar zum Alten Testament
HNT Handbuch zum Neuen Testament
HR E. Hatch, H. A. Redpath, Concordance to the Septuagint and
Other Greek Versions of the Old Testament. 2 vols. Oxford:
Clarendon Press, 1897
HSM Harvard Semitic Monographs
HThKAT Herders theologisher Kommentar zum Alten Testament
HUCA Hebrew Union College Annual
IC Inscriptiones Creticae
ICC International Critical Commentary
IG Inscriptiones Graecae
IJO Inscriptiones Judaicae Orientis
IK Inschriften griechischer Städte aus Kleinasien
IOSCS International Organization for Septuagint and Cognate Studies
JBL Journal of Biblical Literature
JBS Jerusalem Biblical Studies
JIWE Jewish Inscriptions of Western Europe
JJS Journal of Jewish Studies
JNSL Journal of Northwest Semitic Languages
JQR Jewish Quarterly Review
JSem Journal of Semitics
JSHRZ Jüdische Schriften aus hellenistisch-römischer Zeit
JSJ Journal for the Study of Judaism in the Persian, Hellenistic, and
Roman Periods
JSJSup Supplements to the Journal for the Study of Judaism
JSOT Journal for the Study of the Old Testament
JSOTSup Journal for the Study of the Old Testament: Supplement
JSPSup Journal for the Study of the Pseudepigrapha: Supplement
xiv abbreviations

JSQ Jewish Studies Quarterly

JSS Journal of Semitic Studies
JTS Journal of Theological Studies
KAT Kommentar zum Alten Testament
KEK Kritisch-exegetischer Kommentar über das Neue Testament
KV Korte Verklaring van de Heilige Schrift
LA Letter of Aristeas
LBS Library of Hebrew Bible/Old Testament Studies
LEH J. Lust, E. Eynikel and K. Hauspie. Greek-English Lexicon of
the Septuagint. rev. ed. Stuttgart: Deutsche Bibelgesellschaft,
LSJ H. G. Liddell, R. Scott and H. S. Jones. A Greek-English Lexicon.
9th ed. with revised supplement. Oxford: Oxford University
Press, 1996
LXX Septuagint
LXX.D Septuaginta Deutsch
LUÅ Lunds universitets årsskrift
MAMA Monumenta Asiae Minoris Antiqua
MSU Mitteilungen des Septuaginta-Unternehmens
MT Masoretic Text
NA27 E. Nestle. Novum Testamentum Graece. 27th ed. (eds. B. Aland
et al.; Stuttgart: Deutsche Bibelgesellschaft, 1993)
NETS New English Translation of the Septuagint
NICOT New International Commentary on the Old Testament
NIGTC New International Greek Testament Commentary
NovTSup Novum Testamentum Supplements
NRSV New Revised Standard Version
NT New Testament
NTA Neutestamentliche Abhandlungen
NTS New Testament Studies
OBO Orbis biblicus et orientalis
OHB Oxford Hebrew Bible
OLSJ Online Liddell-Scott-Jones Greek-English Lexicon
OT Old Testament
ÖTK Ökumenischer Taschenbuch-Kommentar
OTE Old Testament Essays
OTL Old Testament Library
OTS Old Testament Studies
OtSt Oudtestamentische Studiën
PAAJR Proceedings of the American Academy of Jewish Research
abbreviations xv

PG Patrologia graeca (ed. J.-P. Migne; Paris, 1857–1886)

PL Patrologia latina (ed. J.-P. Migne; Paris, 1844–1864)
POT De Prediking van het Oude Testament
PVTG Pseudepigrapha Veteris Testamenti Graece
QFRG Quellen und Forschungen zur Reformationsgeschichte
RB Revue biblique
RECAM Regional Epigraphic Catalogues of Asia Minor
RevQ Revue de Qumran
RHR Revue de l’histoire des religions
RNT Regensburger Neues Testament
RQ Römische Quartalschrift für christliche Altertumskunde und
RSV Revised Standard Version
SBFA Studium Biblicum Franciscanum: Analecta
SBFLA Studii biblici Franciscani liber annus
SBL Society of Biblical Literature
SBLDS Society of Biblical Literature Dissertation Series
SBLSCS Society of Biblical Literature Septuagint and Cognate Studies
SBS Stuttgarter Bibelstudien
SC Sources chrétiennes
SCS Septuagint and Cognate Studies
SEG Supplementum epigraphicum graecum
SemClass Semitica et Classica
SNTSMS Society for New Testament Studies Monograph Series
SOTSMS Society for Old Testament Studies Monograph Series
SPap Studia papyrologica
STDJ Studies on the Texts of the Desert of Judah
StPB Studia post-biblica
SubBi Subsidia biblica
SUC Schriften des Urchristentums
SVF Stoicorum veterum fragmenta
TAM Tituli Asiae Minoris
TANZ Texte und Arbeiten zum neuentestamentlichen Zeitalter
TDOT Theological Dictionary of the Old Testament (eds. G. J. Botterweck
and H. Ringgren; Grand Rapids: Eerdmans, 1974–)
TECC Textos y estudios “Cardenal Cisneros”
TENT Texts and Editions for New Testament Study
ThHK Theologischer Handkommentar zum Neuen Testament
ThQ Theologische Quartalschrift
ThT Theologisch tijdschrift
xvi abbreviations

ThWQ Theologisches Wörterbuch zu den Qumrantexten

TLG Thesaurus linguae graecae
TLZ Theologische Literaturzeitung
TSAJ Texte und Studien zum antiken Judentum
TU Texte und Untersuchungen
TUGAL Texte und Untersuchungen zur Geschichte der altchristlichen
TWNT Theologische Wörterbuch zum Neuen Testament (eds. G. Kittel
and G. Friedrich; Stuttgart: Kohlhammer, 1932–1979)
VC Vigiliae christianae
VT Vetus Testamentum
VTSup Vetus Testamentum Supplements
WBC Word Biblical Commentary
WMANT Wissenschaftliche Monographien zum Alten und Neuen
WSt Wiener Studien
WTJ Westminster Theological Journal
WUNT Wissenschaftliche Untersuchungen zum Neuen Testament
ZAW Zeitschrift für die alttestamentliche Wissenschaft
ZDPV Zeitschrift des deutschen Palästina-Vereins
ZNW Zeitschrift für die neutestamentliche Wissenschaft und die
Kunde der älteren Kirche
list of contributors

Hans Ausloos Université catholique de Louvain, University

of the Free State
Eberhard Bons Université de Strasbourg
Dieter Böhler SJ Philosophisch-Theologische Hochschule
St. Georgen, Frankfurt am Main
Cilliers Breytenbach Humboldt University, Berlin
University of Stellenbosch
Johann Cook University of Stellenbosch
Johannes De Vries Kirchliche Hochschule, Wuppertal
Heinz-Josef Fabry University of Bonn
Robert J. V. Hiebert Trinity Western University, Langley
Jan Joosten University of Strasbourg
Martin Karrer Kirchliche Hochschule, Wuppertal
Gideon R. Kotzé University of Stellenbosch
Wolfgang Kraus Universität des Saarlandes, Saarbrücken
University of Pretoria
Siegfried Kreuzer Protestant University of Wuppertal
Lawrence Lincoln University of Stellenbosch
Jonathan More University of Stellenbosch
Peter Nagel University of Pretoria
Melvin K. H. Peters Duke University, North Carolina
Jessie Rogers Mary Immaculate College (University of
Limerick), Ireland
Martin Rösel University of Rostock
Barbara Schmitz Universität Würzburg
Gert J. Steyn University of Pretoria
Hermann-Josef Stipp Ludwig-Maximilians Universität, München
University of Stellenbosch
Harry F. Van Rooy North-West University, Potchefstroom
Arie Van Der Kooij Leiden University
Part One

History of the Septuagint in General

The Pentateuch in Greek and the Authorities of the Jews

Arie van der Kooij

It is generally held that the Septuagint version of the Pentateuch was

produced in Alexandria, in the first half of the third century B.C.E., but
questions such as its original setting and the reasons why this translation
was made are disputed. An intriguing question, hardly dealt with as far as
I know, concerns the issue on whose authority the Pentateuch was trans-
lated into Greek. In this contribution, I would like to address the question
of whether, as suggested by the Letter of Aristeas, the authorities of the
Jews in Jerusalem were involved in the translation project, or not.1
The Letter of Aristeas (hereafter LA) offers the most ancient answer to
the question about the original setting of the translation of “the Law of
the Jews”. According to this source,2 dating to the second half of the sec-
ond century B.C.E., the translation of the books of the Pentateuch was
part of the policy of the king Ptolemy II Philadelphus (282–246 B.C.E.) to
collect, if possible, all the books of the world. The royal librarian, Dem-
etrius of Phalerum, was commissioned to do so. Demetrius proposed to
include “the books of the Law of the Jews” (par. 30), for which a transla-
tion would be necessary. The king then sent a letter to the high priest of
the Jews, announcing his plan and requesting assistance. Eleazar, the high
priest, agreed to help, and sent a total of seventy-two translators, six men
of each tribe—men of good behaviour, experts in Hebrew and in Greek,

1 This contribution is based on the assumption that the Pentateuch was translated as
an entity, presumably so by a team of five (or six) translators which would account for a
variety of translation styles, and to some extent a lexical variety as well, on the one hand,
and for the fact that the five books are marked by a shared vocabulary as far as significant
terms are concerned, on the other. For another view, see M. Rösel, Übersetzung als Vollen-
dung der Auslegung: Studien zur Genesis-Septuaginta (BZAW 223; Berlin: De Gruyter, 1994),
257–258; C. G. den Hertog, “Erwägungen zur relativen Chronologie der Bücher Levitikus
und Deuteronomium innerhalb der Pentateuchübersetzung,” in Im Brennpunkt: Die Sep-
tuaginta. Studien zur Entstehung und Bedeutung der griechischen Bibel, Bd. 2 (BWANT 161;
eds. S. Kreuzer and J. P. Lesch; Suttgart: Kohlhammer, 2004), 216–228.
2 For an edition of the Letter of Aristeas, see e.g., A. Pelletier, Lettre d’Aristée à Philo­
crate: Introduction, texte critique, traduction et notes, index complet des mots grecs (SC 89;
Paris: Cerf, 1962).
4 arie van der kooij

and learned in the Law—to Alexandria to prepare the translation. The

work was done on the isle of Pharos, under the direction of Demetrius of
Phalerum. The new version was read to the leaders of the Jewish commu-
nity in Alexandria, as well as to the Ptolemaic king. It was received most
favourably by both parties.
Because LA clearly bears the marks of an apologetic document, the
question of its historical reliability has been disputed. For a long time,
modern scholarship has been sceptical about any historical clue that LA
might contain, except for the idea that the translation was produced in
Alexandria. In keeping with this tendency, Thackeray developed the view
that the translation originally was made “for synagogue use”, arguing that
the Jews in Alexandria were in need of a Greek version of the Pentateuch
that could be used for reading purposes in their synagogue.3 This the-
ory has found wide acceptance and remains popular up to the present
day. Other scholars, like Brock and Perrot, have suggested that, although
the liturgical matter may have been part of the picture, the translation
arose out of the educational needs of the Jewish community.4 A more
recent theory brought to the fore by Pietersma—the so-called “interlinear
paradigm”—fits in with this proposal.5
These and other suggestions6 made are based on the assumption that
the translation project of the Jewish Law should be seen an inner-Jewish
phenomenon. Interestingly, recent research is marked by a tendency to
regard the promulgation of the Greek Pentateuch as being due to a Ptole-
maic initiative or encouragement, reflecting in this way a less sceptical
attitude towards LA. A good example of this tendency is the theory pro-
posed by a number of scholars, like Bickerman and Barthélemy, that the
initiative of the whole project was taken by the Ptolemaic court.7 The idea

3 H. St. J. Thackeray, The Septuagint and Jewish Worship: A Study in Origins (London:
Milford, 1923), 9.
4 S. P. Brock, “The Phenomenon of the Septuagint,” OtSt 17 (1972): 16; C. Perrot, La lec-
ture de la Bible dans la synagogue: Les anciennes lectures palestiniennes du shabbat et des
fêtes (Hildesheim: Gerstenberg, 1973), 143.
5 A. Pietersma, “A New Paradigm for Addressing Old Questions: the Relevance of the
Interlinear Model for the Study of the Septuagint,” in Bible and Computer: The Stellenbosch
AIBI-6 Conference. Proceedings of the Association Internationale Bible et Informatique “From
Alpha to Byte”. University of Stellenbosch 17–21 July, 2000 (ed. J. Cook; Leiden: Brill, 2002),
6 See, e.g., J. Joosten, “Le milieu producteur du Pentateuque grec,” Revue des Études
Juives 165 (2006): 349–361.
7 E. Bickerman, Studies in Jewish and Christian History: Part 1 (AGJU 9; Leiden: Brill,
1976), 171–175; D. Barthélemy, “Pourquoi la Torah a-t-elle été traduite en Grec?” in Études
d’histoire du texte de l’Ancien Testament (OBO 21; Fribourg: Universitaires Fribourg, 1978),
the pentateuch in greek and the authorities of the jews 5

is that the king wanted to have a copy of the books containing the laws
and customs according to which the Jews in Alexandria and in Egypt were
allowed to live. Likewise, Modrzejewski has argued that “royal judges and
officials” needed access to the text of the Jewish Law in order to apply it
effectively.8 Other scholars are of the opinion that the translation proj-
ect rather was elicited by the cultural policy of the early Ptolemies. So
Rajak thinks of a translation made under Ptolemaic patronage,9 whereas
Fernández Marcos has argued that the scholarly milieu in Alexandria is
“the most fitting explanation for the origin of the translation”.10 These and
other proposals made are in line with LA insofar this document is marked
by the idea of a strong interaction between the cultural context in Alexan-
dria, on the one hand, and the religion of the Jews, on the other.11
A stimulating contribution in this regard represents the study by
Honigman, entitled The Septuagint and Homeric Scholarship in Alexandria
(2003). Her work offers a detailed and innovative analysis of LA, and it is
on this basis that she then develops her own thesis regarding the original
setting of the Greek version of the Pentateuch. She argues that the way
the origins of this translation are described in LA was strongly influenced
by the practice and ideology of Homeric scholarship in Alexandria. This
feature, she believes, opens up a window on the original setting of the
translation which was made for the sake of prestige, and not for reasons
of any religious need.
I am not going to discuss in this paper the theory advanced by Honig-
man—I have done so elsewhere12—but I agree with Honigman, and other
scholars referred to above, that in dealing with the complex issue of the

322–340; M. Harl, G. Dorival and O. Munnich, La Bible grecque des Septante: Du Judaïsme
hellénistique au Christianisme ancien (Paris: Cerf, 1987), 78 (“Il faut l’initiative officielle”
  8 J. Mélèze Modrzejewski, The Jews in Egypt: From Rameses II to Emperor Hadrian
(trans. R. Corman; Princeton: Princeton University Press, 1997), 104–111.
  9 T. Rajak, Translation and Survival: The Greek Bible of the Ancient Jewish Diaspora
(Oxford: Oxford University Press, 2009).
10 N. Fernández Marcos, “The Greek Pentateuch and the Scholarly Milieu of Alexan-
dria,” SemClass 2 (2009): 88.
11  Cf. also M. Rösel, Übersetzung als Vollendung der Auslegung; S. Kreuzer, “Entstehung
und Publikation der Septuaginta im Horizont frühptolemäischer Bildungs- und Kultur-
politik”, in Im Brennpunkt: Die Septuaginta. Studien zur Entstehung und Bedeutung der
Griechischen Bibel, Bd. 2 (BWANT 161; eds. S. Kreuzer and J. P. Lesch; Stuttgart: Kohlham-
mer, 2004), 61–75; A. Schenker, “Was führte zur Übersetzung der Tora ins Griechische?
Dtn 4,2–8 und Platon (Brief VII,326a-b),” in Die Septuaginta—Texte, Theologien, Einflüsse
(WUNT 252; eds. W. Kraus and M. Karrer; Tübingen: Mohr Siebeck, 2010), 23–35.
12 A. van der Kooij, “The Septuagint and Alexandrian Scholarship,” BiOr 68 (2011): 492–509.
6 arie van der kooij

origins of LXX Pentateuch the picture of LA in its main lines should not
be dismissed too easily. Of course, one could argue that the translation
itself is the only source available for the search after its origins. However,
the Greek version of the Law is only helpful if internal evidence can be
assessed in the light of external data regarding the period in which the
version supposedly was made (first half third century B.C.E.). As a mat-
ter of fact, LA is the only document from Antiquity that tells us a story
about the origins of the Greek Pentateuch. Honigman has shown that it is
marked by an antiquarian intent and that it also contains elements being
part of an oral tradition shared by members of the Jewish elite in Alex-
andria (“collective memory”). However, as she has also made clear, even
in the case of traditions used by the author, it remains difficult to show
whether they testify, in one way or another, to the original setting of the
In my view, the best way to proceed is to find out whether there are
specific data, external to LA, which may add to the plausibility of any
hypothesis based on LA.13 Any proposal made on the basis of LA can only
serve as a window on the original setting of the translation if there is some
evidence from sources other than LA which is in favour of the suggestion
made. In addition, it is important to see whether specific interpretive ele-
ments in the Greek version itself may fit in, or may even strengthen a
given proposal.
It is from this perspective that I would like to discuss the issue whether
the authorities of the Jews in Jerusalem, in particular the high priest,
played a role in the translation project. According to LA, the high priest
was the one who granted permission to translate the Law, who selected
the translators, and who is also presented as the prime interpreter of the
Law (par. 128–169). What to make of this? Is all this made up for propa-
gandistic reasons only, or does LA at the same time present a picture that
would make sense in the first half of the third century B.C.E.?
The Pentateuch, the Law of the Jews, was a most important part of
the literary heritage of the Jews, and also constituted, together with the
temple and the holy city, the identity of their religion.14 In view of its

13 Cf. Fernández Marcos, “The Greek Pentateuch,” 86.

14 There are good reasons to assume that the Pentateuch did already exist at that time
(ca. 300 B.C.E.). Chronicles, dating to the late fourth century B.C.E. (cf. G. N. Knoppers,
I Chronicles 1–9 [AB 12; New York: Doubleday, 2004], 116), clearly reflects a comprehensive
notion of the term Torah, encompassing not only Deuteronomy (as in Kings) but also
priestly laws of the Pentateuch. It likely presupposes the Pentateuch as an authoritative
writing (cf. e.g. M. Fishbane, Biblical Interpretation in Ancient Israel [Oxford: Clarendon
the pentateuch in greek and the authorities of the jews 7

significance the question of whether the Greek version of it was made on

the authority of the high priest and his colleagues in Jerusalem suggests


The high priest is depicted in LA as the political leader of Judea, and,

as far as we know, this is in line with the historical reality of the early
Hellenistic period, the time when the Law of Moses was translated. This
picture is attested by several sources, such as the work of Ben Sira (ch. 50)
as well as by a source dating to the first decades of Ptolemaic rule: the
description of the Jewish nation by Hecataeus of Abdera, a Greek scholar
who lived in Alexandria, around 300 B.C.E. It has come down to us in the
work of Diodorus Siculus (Bibliotheca Historica 40.3).15 Its focus is on the
laws and customs of the Jewish people. The reader is told that Moses,
“outstanding both for his wisdom and for his courage,” took possession of
the land and founded cities, such as Jerusalem. Hecateaus then continues,
“he established the temple that they hold in chief veneration, instituted
their forms of worship and ritual, drew up their laws, and ordered their
political institutions.” Thus, also in this statement the laws of Moses are
clearly related to the issue of the polity (πολιτεία) of the Jewish nation.
Interestingly, as to the form of government Hecataeus tells his readers
that the priests were selected (by Moses) “to head the entire nation.” They
should not only occupy themselves with the temple and the sacrificial
cult, but were also “appointed to be judges in all major disputes.” It was
to them that Moses entrusted “the guardianship of the laws and customs”.
Hecataeus then states:
For this reason the Jews never have a king, and authority over the people is
regularly vested in whichever priest is regarded as superior to his colleagues
in wisdom and virtue. They call this man the high priest.16

Press, 1985], 264). The evidence from Qumran also points to a date of its final stage before
the third century B.C.E.; see S. White Crawford, “The Use of the Pentateuch in the Temple
Scroll and the Damascus Document in the Second Century B.C.E.” in The Pentateuch as
Torah: New Models for Understanding Its Promulgation and Acceptance (eds. G. N. Knop-
pers and B. M. Levinson; Winona Lake: Eisenbrauns, 2007), 301–317.
15 For the text, see M. Stern, ed., Greek and Latin Authors on Jews and Judaism. Vol. 1:
From Herodotus to Plutarch ( Jerusalem: Israel Academy of Sciences and Humanities, 1974),
16 Ibid., 28. On this passage, see e.g. J. C. VanderKam, From Joshua to Caiaphas. High
Priests after the Exile (Minneapolis: Fortress Press, 2004), 118–122; and M. Brutti, The
8 arie van der kooij

The picture as presented by Hecataeus is that the Jewish nation was ruled
by “priests”, under the supreme direction of a high priest, a form of gov-
ernment which is regarded to be in accord with the laws of Moses.
The same view was expressed at another occasion of a much later date,
this time by leading Jews. I have in mind an incident that took place in
the year 63 B.C.E. when members of the Hasmonean house came before
Pompey, the Roman general, in Damascus with their dispute over the
kingship. As we are told by Diodorus Siculus:
Likewise the leading men, more than two hundred in number, gathered to
address the (Roman) general and explain that their forefathers [. . .] had sent
an embassy to the senate, and received from them the leadership of the Jews,
who were, moreover to be free and autonomous, their ruler being called
High Priest, not King. Now, however, these men were lording it over them,
having overthrown the ancient laws (τοὺς πατρίους νόμους) and enslaved the
citizens in defiance of all justice17
According to this passage, another party (“the leading men”) appeared
before Pompey, claiming that the ruler of the Jewish nation should not be
a king, but a high priest, because this type of leadership was considered
by them in agreement with the ancient, “ancestral” laws. Two forms of
government are at stake here—that of kingship, and that of aristocracy.
It is clear that the leading men were in favour of the latter, taken in the
sense of priestly rule, while rejecting the former as being a constitution
that brings about the enslavement of the citizens.
Thus, the sources just mentioned not only testify to the fact that the
Jewish nation of the time was ruled by a high priest, but also claim that
(high-)priestly rule was in line with the ancient laws, the Law of Moses. It
may be noted that whereas “the leading men” of Diodorus are focussing
on the figure of the high priest as leader—in contrast to the king—, the
passage of Hecataeus is more nuanced saying that it was the high priest
together with “the priests” who were ruling the nation. The latter ones
are obviously to be seen as leading priests, which is also in line with the
statement, by Hecataeus, that they were the judges “in all major disputes”.
This raises the question which priests might have been envisaged here.
I will come back to this question below.

­ evelopment of the High Priesthood during the pre-Hasmonean Period: History, Ideology,
Theology (Leiden: Brill, 2006), 138–141.
17 Diodorus Siculus, Bibliotheca Historica, 40.2. For this translation, see Stern, Greek and
Latin Authors 1, 185–186.
the pentateuch in greek and the authorities of the jews 9

As noted above, in LA the high priest is also presented as the (prime)

interpreter of the Law. This feature too is in line with the evidence we
have. In the description of the Jews by Hecataeus of Abdera, referred to
above, the following note on the high priest is found:
They call this man the high priest, and believe that he acts as a messenger
(ἄγγελον) to them of God’s commandments. It is he, we are told, who in
their assemblies and other gatherings announces what is ordained, and the
Jews are so docile in such matters that straight way they fall to the ground
and do reverence to the high priest when he expounds the commandments
to them.18
The high priest is depicted here as a “messenger” of God’s command-
ments who in assemblies and gatherings announces and expounds what
is ordained. As we know from other sources of the time, the priests were
considered experts of the law and hence the ones who should teach the
people (see e.g. Sir 45:17; Jub 31:15).19 This sheds light on the statement of
Hecataeus regarding the leading priests being appointed as “judges” which
of course presupposes their expertise of the laws. As these priests were
ruling the nation under the supreme authority of the high priest, the latter
was considered the prime authority as far as the law and its interpretation
is concerned.
This picture of the high priest is also implied in the Wisdom of Ben Sira.
As has been observed by scholars, the passage on Simon, the high priest,
in Sir 50 is strongly linked to the hymnic passage on wisdom and the law
in ch. 24. Both wisdom (Law) and the high priest are said to “minister”
before God (24:10; 50:14). The specific link between the two passages indi-
cates that the high priest was seen as the main authority as far as wisdom
and the Mosaic Law were concerned.20 All this reminds one of the figure
designated, in documents of Qumran, as “the Interpreter of the Law”, the
more so since it is very likely that he is to be regarded a high-priestly

18  Ibid., 28.

19  Cf. Deut 17:9; 33:10. For evidence from Qumran, see F. García Martínez, “Priestly
Functions in a Community without Temple,” in Gemeinde ohne Tempel/Community with-
out Temple. Zur Substituierung und Transformation des Jerusalemer Tempels und seines
Kults im Alten Testament, antiken Judentum und frühen Christentum (WUNT 118; eds. B. Ego,
A. Lange, and P. Pilhofer; Tübingen: Mohr Siebeck, 1999), 311–312 (“Priests as Judges”).
20 Cf. A. van der Kooij, “Authoritative Scriptures and Scribal Culture,” in Authoritative
Scriptures in Ancient Judaism ( JSJSup 141; ed. M. Popovic; Leiden: Brill, 2010), 69.
21  Cf. J. J. Collins, The Scepter and the Star. The Messiahs of the Dead Sea Scrolls and
Other Ancient Literature (New York: Doubleday, 1995), 144.
10 arie van der kooij

In the light of these (external) data it may be concluded that the pic-
ture of the high priest as leader and prime interpreter, as presented in
LA, concurs with the historical reality at the time when the translation of
the Law was made. Hence, in view of his role and position, one can imag-
ine that the Greek translation of the Law was made on the authority of
the high priest. This is even more plausible if one takes into account that
the Law was an official document and that, as has been argued convinc-
ingly by scholars, the translation should be seen as the result of an official


However, the question arises whether the translation itself contains any
piece of evidence that might shed light on the issue at stake. In this sec-
tion of my contribution, I will discuss a few passages in LXX Pentateuch
that seem to be of interest for our topic. The passages are Exod 19:6, and

Exodus 19:6, part of which reads thus,23

LXX: “You shall be to me a royal priesthood and a holy nation”
MT: “You shall be to me a kingdom of priests and a holy nation”
The Hebrew expression “a kingdom of priests” (‫ )ממלכת כהנים‬has been
rendered in Greek as βασίλειον ἱεράτευμα. This is not the only rendering
possible; an alternative, and more literal, translation is found in the ver-
sion of Aquila: βασιλεἰα ἱερέων. As to the phrase in LXX it has been sug-
gested to take both words as two substantives (“kingdom”, “priesthood”),24
but scholars like Wevers and others have convincingly argued that Greek
βασίλειον is best understood here as an adjective (“a royal priesthood”).25

22 See A. van der Kooij, “The Septuagint of the Pentateuch and Ptolemaic Rule,” in The
Pentateuch as Torah. New Models for Understanding Its Promulgation and Acceptance (eds.
G. N. Knoppers and B. M. Levinson; Winona Lake: Eisenbrauns, 2007), 292–293 (with lit.).
23 For a discussion of this passage, see also A. van der Kooij, “A kingdom of priests:
Comment on Exodus 19:6,” in The Interpretation of Exodus. Studies in Honour of Cornelis
Houtman (CBET 44; eds. R. Roukema et al.; Leuven: Peeters, 2006), 173–175.
24 So J. H. Elliot, The Elect and the Holy (NovTSup 12; Leiden: Brill, 1966), 70–73. For this
reading/interpretation, see also below.
25 J. W. Wevers, Notes on the Greek Text of Exodus (SCS 30; Atlanta: Scholars Press, 1990),
295. Cf. G. Schrenk, “ἱεράτευμα,” in TWNT III, 249; J. Lust, E. Eynikel and K. Hauspie, Greek-
English Lexicon of the Septuagint, rev. ed. (Stuttgart: Deutsche Bibelgesellschaft, 2003);
A. Le Boulluec and P. Sandevoir, L’Exode (BdA 2; Paris: Cerf, 1989), 200; C. G. den Hertog,
the pentateuch in greek and the authorities of the jews 11

The combination of adjective + noun, not that usual though, is attested at

other places in LXX Pentateuch as a type of rendering of a Hebrew expres-
sion consisting of stat. cstr. followed by a noun.26
The other term employed, ἱεράτευμα, occurs only three times in the
LXX—beside our text in LXX Exod 23:22 (in a large plus) and in 2 Macc
2:17. The word seems to be of late origin; it is only found in the LXX and
related Jewish or Christian writings (cf. Elliot). In distinction to the Greek
ἱερατεία, which occurs more often (e.g. Exod 29:9; 35:19) and denotes the
priesthood and the priestly office in general, the lexeme ἱεράτευμα is a
word which, just like στράτευμα and τεχνίτευμα, refers to a particular
group of people, a body, or college of priests.27
It is to be asked how the expression “a royal priesthood” is related to
the other phrase in the same verse, “a holy nation”. In the early Chris-
tian tradition, particularly so in 1 Peter 2:9, both expressions have been
taken as parallel phrases (1 Peter 2:9: “But you are a chosen race, a royal
priesthood, a holy nation”). This interpretation is also typical of the way
scholars have interpreted our text.28 For example, Wevers remarks that
“Exod’s ‘royal priesthood’ represents an ideal in which the priesthood is of
kingly stock, and in which all Israel constitutes such an ideal”.29 However,
from a Jewish perspective it is far from certain that “royal priesthood” in
Exod 19:6 should be understood as a designation of all Israel, parallel to

“Die griechische Übersetzung von Exodus 19:6 als Selbstzeugnis des frühhellenistischen
Judentums,” in The Interpretation of Exodus. Studies in Honour of Cornelis Houtman (CBET
44; eds. R. Roukema et al.; Leuven: Peeters, 2006), 184.
26 Cf. I. Soisalon-Soininen, Studien zur Septuaginta-Syntax (Helsinki: Suomalainen Tie-
deakatemia, 1987), 66.
27 Cf. Schrenk, “ἱεράτευμα,” 249 (“die Priesterschaft als Korporation”); T. Muraoka,
A Greek-English Lexicon of the Septuagint (Louvain / Paris / Walpole: Peeters, 2009), 338.
28 This also applies to the exegesis of MT Exod 19:6. This interpretation, however, has
been criticized in recent contributions; see in particular A. Graupner, “ ‘Ihr sollt mir ein
Königreich von Priestern und ein heiliges Volk sein’. Erwägungen zur Funktion von Ex
19,3b-8 innerhalb der Sinaiperikope,” in Moses in Biblical and Extra-Biblical Traditions
(BZAW 372; eds. A. Graupner and M. Wolter; Berlin: De Gruyter, 2007), 33–50 (“ein König-
reich, das von Priestern regiert wird” [43]).
29 Wevers, Notes on Exodus, 295. See also E. Schüssler Fiorenza, Priester für Gott: Stu-
dien zum Herrschafts- und Priestermotiv im der Apokalypse (Münster: Aschendorf, 1972), 83;
D. Munoz León, “Un reino de sacerdotes y una nación santa (Ex 19,6),” EstBib 37 (1978): 166
(he has the Jews of the diaspora in mind); O. Camponovo, Königtum, Königsherrschaft und
Reich Gottes in den Frühjüdischen Schriften (OBO 58; Freiburg / Göttingen: Vandenhoeck &
Ruprecht, 1984), 386; Le Boulluec and Sandevoir, L’Exode, 200; A. Schenker, “Drei Mosaik-
steinchen,” in Studies in the Book of Exodus: Redaction—Reception—Interpretation (BETL
126; ed. M. Vervenne; Leuven: Peeters, 1996), 374; W. Horbury, “Monarchy and Messianism
in the Greek Pentateuch,” in The Septuagint and Messianism (BETL 195; ed. M. A. Knibb;
Leuven / Paris / Dudley: Peeters, 2006), 91.
12 arie van der kooij

“a holy nation”.30 It seems to be more plausible to interpret our text from

the perspective the text was read in Jewish circles in Antiquity, such as
those behind the Targumim. The latter attest an interpretation according
to which the Hebrew phrase, “a kingdom of priests”, was understood as
referring to the leaders of the people, and not to the people as a whole:
Tg. Onq.: “kings, priests”
Tg. Neof. 1: “kings and priests”
Tg. Ps.-J.: “kings adorned with the crown, ministering priests”
These sources reflect a reading of the Hebrew expression by taking both
terms each on its own, as two substantives.31 It is to be noted that this type
of interpretation is also attested by a source dating to the late second cen-
tury B.C.E.—2 Maccabees. In 2 Macc 2:17 the Greek version of Exod 19:6
is quoted as follows: τὸ βασίλειον καὶ τὸ ἰεράτευμα. Here too both terms do
not refer to the people as a whole, but to the leadership, in this instance
to “the kingship” and “the priesthood” as two distinct institutions.32 Read-
ing from this perspective, the phrase “a royal priesthood” makes perfect
sense because a body, or college, of priests having a “royal” status strongly
evokes the notion of a body of leading priests.33 Interestingly, this is also
in line with other texts which testify to royal aspects of the priesthood,
such as 1Q21 (“kingdom of priesthood”) and the Aramaic Levi Document.34
Thus, our text seems to reflect the polity of the Jewish nation as ruled by a
particular body of priests, under the supreme direction of a high priest.35

30 The ancient Jewish literature does not offer an interpretation of Exod 19:6 in line
with that of the New Testament; see D. R. Schwartz, Studies in the Jewish Background of
Christianity (WUNT 60; Tübingen: Mohr Siebeck, 1992), 57–66.
31 Cf. Theodotion and Symmachus: regnum sacerdotes.
32 On this text, see A. van der Kooij, “The Use of the Greek Bible in II Maccabees,” JNSL
25 (1999): 129–130; T. Rajak, The Jewish Dialogue with Greece and Rome: Studies in Cultural
and Social Interaction (AGJU 48: Leiden: Brill, 2001), 52.
33 For a different interpretation of the phrase ‘royal priesthood’, see Den Hertog, “Die
griechische Übersetzung,” 186 (in his view, the adjective ‘royal’ indicates that Israel “als
eine einem nicht näher definierten externen βασιλεύς zugeordnete Priesterschaft bestimmt
34 See M. E. Stone and J. C. Greenfield, “Aramaic Levi Document,” in Qumran Cave
4 XVII: Parabiblical Texts, Part 3 (DJD XXII; G. J. Brooke et al.; Oxford: Clarendon Press,
1996), 17.
35 Horbury (“Monarchy and Messianism,” 92) subscribes to this view, but is also of the
opinion that the phrase ‘royal priesthood’ first of all fits “the view of Israel as priestly
the pentateuch in greek and the authorities of the jews 13

Exodus 23:21–23
The passage of LXX Exod 23:20–23 reads in translation,36
20 And look, I am sending my angel in front of you in order to guard you
on the way in order to bring you into the land that I prepared for you.
21  Mind yourself, and listen to him, and do not disobey him. For he shall
not hold you in undue awe, for my name is upon him.
22 If by paying attention you listen to my voice and do all that I tell you,
I will be an enemy to your enemies and will resist those who resist
23 For my angel will go, leading you, and will bring you in to the Amorrite
and Chettite and Pherezite and Chananite and Gergesite and Heuite
and Iebousite, and I will destroy them.37
This translation is based on the edition of Wevers, but, interestingly,
important MSS attest a text which includes a large plus prior to the first
word of v. 22:
If by paying attention you listen to my voice and do all what I command
you, and keep my covenant, you shall be for me a people special above all
nations. For all the earth is mine. And you shall be for me a royal priesthood
and a holy nation. These words you shall say to the sons of Israel.38
In LXX Exod 23:20–23 the figure of “my angel” dominates the scene. He is
designated in those terms both in v. 20 (MT without suffix, but see Sam-
Pent) and in v. 23 (cf. MT). God will send him in front of Israel, and he will
guard the people on the way to the land in order that they arrive in the
land safely. The underlying Hebrew text is about a heavenly messenger,
an angel,39 and it is commonly assumed that this also applies to the Greek
version. However, this version displays elements which seem to point in
another direction. V. 21 is of particular interest in this regard. It reads thus:
“Mind yourself, and listen to him, and do not disobey him. For he shall not
hold you in undue awe, for my name is upon him.”

36 The following discussion of this passage is a shortened version of A. van der Kooij,
“LXX Exodus 23 and the Figure of the High Priest,” in On Stone and Scroll. Essays in Honour
of Graham Ivor Davies (BZAW 420; eds. J. K. Aitken, K. J. Dell and B. A. Mastin; Berlin: De
Gruyter, 2011), 537–549.
37 NETS.
38 In the Septuagint edition of Rahlfs this plus is part of the Greek text. For a transla-
tion of the passage including the plus, see Le Boulluec and Sandevoir, L’Exode, 239. For a
discussion of the plus, see Van der Kooij, “LXX Exodus 23,” 539–541.
39 On the figure of the angel in the Hebrew text, see e.g., H. Ausloos, “The ‘Angel of
YHWH’ in Exod xxiii 20–33 and Judg. ii 1–5. A Clue to the Deuteronomistic Puzzle?”
VT 58 (2008): 1–12.
14 arie van der kooij

LXX, “Do not disobey him (μὴ ἀπείθει αὐτῶ)” for MT “do not show bit-
terness towards him”: The Greek version presupposes an interpretation
of the Hebrew different from that of the Masoretes. The latter reflects the
verb ‫( מרר‬cf. e.g. Sym), whereas the Greek phrase is based on the verb
‫( מרה‬cf. Deut 1:26; 9:7, 23, 24 [all cases Hiphil]; Isa 50:5; 63:10 [both cases
LXX “He shall not hold you in undue awe”, for MT “he will not pardon
your transgressions”: The Greek verb used here (ὑποστέλλομαι) means “to
draw, shrink back”. The idea is that “he will not refrain from judging you”.40
The Greek text alludes to the role of a judge, as is clear in the light of Deut
1:16–17 and of Job 13:8 as well.41

Deut 1:16–17 (LXX)

And I commanded your judges at that time, saying: “Give a full hearing
among your brothers, and judge rightly between a man and between his
brother and between his guest. You shall not recognize the person when
judging: like the small so you shall judge the great; you shall not shrink from
the face of a person, for the judgment is God’s. And the matter, if it be too
hard for you, you will bring it to me, and I will hear it.”
Job 13:8 (LXX)
Or will you shrink (from his face, i.e. God)? Nay, you should be judges
(MT: Will you lift up his face? Will you contend for God?)
As a judge, the messenger will not be overawed by you. The Greek render-
ing presupposes a reading of the underlying Hebrew text different from
MT, namely: “he will not lift up (sc. your face [cf. ‫ )]נשא פנים‬as to your
transgressions”. The issue at stake here is the sensitive matter of showing
partiality in court, pronouncing someone guilty who is not, or the other
way around. Or to put it with the words of Prov 18:5: “It is not good to lift
up the face of the guilty”.
The final clause of the verse reads “My name is upon him (ἐπ᾽ αὐτῶ)” for
MT “my name is in him (‫”)בקרבו‬. According to Wevers, the Greek text is
marked here by an interpretation that is meant to avoid an identification

40 Wevers, Notes on Exodus, 370.

41  Compare also Wis 6:7.
42 NETS: “Really? Will you prevaricate? Go ahead; become judges yourselves!” I prefer
a translation in line with Deut 1, because of the “judges” in the second part of the Greek
the pentateuch in greek and the authorities of the jews 15

of the angel with the Lord. His name is not within him because “he is not
himself the Lord,” and he cannot forgive sins because only God can do
so.43 Ausloos considers this difference being one of the elements in the
Greek version that point to a weakening of the role of the angel.44 How-
ever, both scholars do not address the question of what the phrase, “my
name is upon him,” may mean. It is said of the angel, or messenger, that
the name of the Lord is upon him. It is to be asked to which figure this
idea might apply. In the light of the available data the most likely answer
to this question is: the high priest.
As we know from Exod 28:36 and 39:30, the high priest is the one who
carries the name of the Lord as part of an inscription (“Holy to YHWH”)
being engraved on a plate of pure gold which was fastened on the turban.
In LA when describing the apparel of the high priest (par. 96–99), the
relevant passage reads thus:
Upon his head he has what is called the ‘tiara’, and upon this the inimitable
‘mitre’, the hallowed diadem having in relief on the front in the middle in
holy letters on a golden leaf the name of God (par. 98)45
In this description the focus is on the name of God as being inscribed, no
reference being made to the term “holiness” (ἁγίασμα) as the first word
is rendered in the LXX (Exod 28:32; 36:38). The same idea is attested by
Josephus, B.J. 5.235; Ant. 3.178, the latter of which reads as follows: “the
forehead . . . had a plate of gold, bearing graven in sacred characters the
name of God”.46 The name was inscribed in “holy” (LA), or “sacred”
( Josephus) letters, that is to say, it was not put in square characters, but
in the Old Hebrew script. At another place Josephus provides us with a
striking illustration of the significance attached to the fact that the name
of God is on the high priest. It is a story about Alexander the Great who,
having taken Tyre after a long siege, and also Gaza, went up to the city
of Jerusalem (Ant. 11.326). Alexander’s meeting with the Jewish leaders, at
Mount Scopus, is described as follows:

43 Wevers, Notes on Exodus, 370. For the idea that the (unusual) rendering in Greek is
to be seen as an interpretation by the translator, see also R. Sollamo, Renderings of Hebrew
Semiprepositions in the Septuagint (AASF 19; Helsinki: Suomalainen Tiedeakatemia, 1979),
44 H. Ausloos, “The Septuagint version of Exod 23:20–33. A ‘Deuteronomist’ at work?”
JSNL 22 (1996): 102.
45 Translation of R. J. H. Shutt, “Letter of Aristeas,” in The Old Testament Pseudepigrapha,
vol. 2 (ed. J. H. Charlesworth; New York: Doubleday, 1985), 19.
46 See also Origen, PG 12:1104; Tg. Ps.-J. Exod 29:6.
16 arie van der kooij

When Alexander while still far off saw the multitude in white garments
the priests at their head clothed in linen, and the high priest in a robe of
hyacinth-blue and gold, wearing on his head the mitre with the golden plate
on it on which was inscribed the name of God, he approached alone and
prostrated himself before the Name and first greeted the high priest [. . .] the
kings of Syria and the others were struck with amazement at his action and
supposed that the king’s mind was deranged. And Parmenion alone went up
to him and asked why indeed, when all men prostrated themselves before
him, he had prostrated himself before the high priest of the Jews, where-
upon he replied, “It was not before him that I prostrated myself but the God
of whom he has the honour to be high priest . . .” (Ant. 11.331–333).
According to this story, which is considered to be legendary, Alexander
prostrated before the high priest of the Jews because of the Name inscribed
on the head of the latter. The Macedonian king therefore greeted the high
priest first, before being greeted by him as one would expect. To be more
precise, Alexander did not prostrate himself before the high priest, but
before God, as he himself explains. This story which obviously reflects a
Jewish view, offers clear proof of the significance of the name of God on
the high priest, underlining the close relationship between him and God.
In the light of these data the phrase, “my name is upon him,” in the Greek
text of Exod 23:21 makes perfect sense if understood as referring to the
figure of the high priest.
It is to be asked however whether this also would apply to the term
ἄγγελος in v. 20 and v. 23. As such, this Greek word needs not to be taken
in the sense of a heavenly messenger, an angel, since it can also denote a
human messenger (see e.g. Gen 32:3). More importantly, there is evidence
that a priest could be designated that way. In Mal 2:7 the priest is called a
“messenger” of the Lord: “For the lips of a priest should guard knowledge,
and people should seek instruction from his mouth, for he is the messen-
ger of the Lord of Hosts”. Furthermore, as has been noted above, accord-
ing to Hecataeus of Abdera the Jews believed that the high priest acted “as
a messenger (ἄγγελον) to them of God’s commandments”.
All this fits the Greek text of Exod 23 well as it also sheds light on other
elements (modifications) in the passage. In v. 21, the people is urged to
listen and not to disobey the messenger of God of whom it is said that “he
shall not hold you in undue awe,” which, as we have seen, alludes to his
role as judge. It is true that in v. 22 it is not the messenger, but God and
his voice one should listen to, but, just as in the story about Alexander the
Great, our passage seems to be marked by the idea of a very close rela-
tionship between God and the high priest, particularly so because of the
Name being on the latter. It is interesting to note that this is fully in line
the pentateuch in greek and the authorities of the jews 17

with the following statement of Josephus: “Any who disobey him—i.e. the
high priest—will pay the penalty as for impiety towards God himself ”
(C. Ap. 2.194).
Finally, according to v. 20, the angel/messenger will go in front of Israel
“in order to guard you on the way in order to bring you into the land that
I prepared for you” (cf. v. 23). If read as referring to the high priest, this
verse too makes good sense because the priestly leader—Eleazar, the son
of Aaron—was the one who entered the land together with the people (cf.
Num 27:18–23; Josh 22:13; 24:33).


Both instances in LXX Exodus are very interesting as they testify to the
idea that the (leading) priests and the high priest are seen as governing
the Jewish people. They thus reflect a view of the constitution of the Jews
which is in line with the picture provided by Hecataeus of Abdera. For
the sake of clarity it is to be asked however which group of priests might
be envisaged as making up the body of leading priests within the Jewish
polity of the time.
A large number of priests, and Levites, were engaged in the temple ser-
vice, but given the hierarchy involved (see below) only a particular group
is to be regarded as representing the leading ones. It stands to reason to
think here of the priests who are often mentioned as accompanying the
high priest at official occasions. See e.g.:
(To our lord Bigvai, the governor of Judaea,) your servants Yedoniah and his
colleagues, the priests who are in Yeb the fortress (Cowley 30,1)47
(we sent a letter to your lordship and) to Johanan the high priest and his
colleagues, the priests who are in Jerusalem, and to Ostanes the brother of
Anani, and the nobles of the Jews (Cowley 30,18)48
(When he—the high priest—learned that Alexander was not far from the
city,) he went out with the priests and the body of citizens (Ant. 11.329)
With his colleagues [συνιερέων] he (i.e. the high priest, AvdK) will sacrifice
to God, safeguard the laws, adjudicate in cases of dispute, punish those con-
victed of crime (C. Ap. 2.194)

47 A. Cowley, Aramaic Papyri of the Fifth Century B.C. (Oxford: Clarendon Press, 1923
[repr. Osnabrück: Otto Zeller, 1967]), 113.
48 Ibid., 114.
18 arie van der kooij

From these examples we learn that the high priest together with “the
priests” is part of the leadership of the Jewish nation, the other part being
the nobles, the body of citizens. The latter college is comprised of the rep-
resentatives of the lay people, elsewhere referred to as “the elders”.49
Who were the priests that together with the high priest were head-
ing the nation? In my view, one has to think here of the priests who
are designated, both in the New Testament and by Josephus, ἀρχιερεῖς,
“chief priests.” They were the ones who constituted the highest level of
the priests acting in the temple, as we know from Josephus and rabbinic
sources.50 To quote Jeremias, the “chief priests permanently employed at
the Temple formed a definite body who had jurisdiction over the priest-
hood and whose members had seats and votes on the council (i.e. the San-
hedrin, AvdK)”.51 A writing from Qumran, 1QM, contains a passage which
is illuminating in this regard. 1QM 2:1–3 provides the following picture of
the priestly hierarchy in the temple:

– The chiefs of the priests behind the High Priest and of his second (in
rank), twelve priests to serve continually before God;
– The twenty-six chiefs of the divisions;
– The chiefs of the Levites to serve continually, twelve;
– The chiefs of their divisions.

The “chiefs of the priests”, representing the highest rank, are to be equated
with the chief priests.52 Therefore, it is likely that they were the ones who
together with the high priest constituted the priestly rule of the Jewish
nation. In the light of these data, it is reasonable to assume that the phrase
“a royal priesthood” refers to this body of leading priests.
These priests were also the ones who acted as judges because the priests
of the highest rank were members of the High Court in Jerusalem. This
is not only clear from the statement by Josephus quoted above, but also
from what Hecataeus tells us about the leading priests as being the ones
who were also appointed “judges in all major disputes”. The reference to

49 See, e.g., 1 Macc 7:33 and 1 Macc 14:28.

50 See J. Jeremias, Jerusalem in the Time of Jesus. An Investigation into Economic and
Social Conditions during the New Testament Period (London: SCM Press, 1976), 147–180.
51  Ibid., 180.
52 See A. van der Kooij, Die alten Textzeugen des Jesajabuches. Ein Beitrag zur Textge-
schichte des Alten Testaments (OBO 35; Freiburg / Göttingen: Universitätsverlag / Vanden-
hoeck & Ruprecht, 1981), 201.
the pentateuch in greek and the authorities of the jews 19

“the major disputes” points to the High Court as described in Deut 17:8–9
(MT; cf. LXX). Both verses read thus,
If any case arises requiring decision between one kind of homicide and
another, one kind of legal right and another, or one kind of assault and
another, any case within your towns which is too difficult for you, then you
shall arise and go up to the place which the Lord your God will choose, and
coming to the Levitical priests, and to the judge who is in office in those
days, you shall consult them, and they shall declare to you the decision.
According to this passage, “the priests” and “the judge”, the latter being
a lay person,53 are the ones who as members of the High Court will deal
with the “difficult” cases. All this also sheds light on the presentation of
the high priest as judge in LXX Exod 23:21.

LXX Exodus contains evidence testifying to a view of the constitution

of the Jews which is in line with the description given by Hecataeus of
Abdera—a body of chief priests, under the supreme direction of the high
priest, heading the nation. As this view obviously serves the interests of
the leading priestly authorities in Jerusalem, it seems likely that the inter-
pretation of the relevant passages actually reflects their own understand-
ing of these texts.
As noted above, in view of his role as leader and prime interpreter of
the Law it seems only natural that the Greek version was made on the
authority of the high priest. If not, that is to say, if the version was made
on the authority of leaders of the Jewish community in Alexandria, it
would have been considered, by the leadership in Jerusalem, a rival one.54
However, it is hardly likely that local leaders in Egypt were in a position to
authorise a translation of the Law, the more so since the Jewish commu-
nity in Alexandria, in its early days, was not significant at all.55 Moreover,
at that time Judea and Jerusalem were part of the Ptolemaic empire. So,
even if leaders of the Jewish community in Alexandria played any role, the

53 Compare Deut 19:17 (“the priests and the judges”).

54 As was the case with the SamPent.
55 Cf. P. M. Fraser, Ptolemaic Alexandria I, Text (Oxford: Clarendon Press, 1972), 55;
S. Honigman, Septuagint and Homeric Scholarship in Alexandria: A Study in the Narrative
of the Letter of Aristeas (London: Routledge, 2003), 98–101.
20 arie van der kooij

two passages in LXX Exodus do not support the idea of a rival version. On
the contrary, they rather indicate that the translation of the Pentateuch
was made on the authority of the leading priests in Jerusalem, and of the
high priest in particular.56

56 Compare the role played by king Jehoshaphat who according to 2 Chron 17 sent
nobles, Levites and priests to teach ‘the book of the law of the Lord’ in the cities of Judah
(vv. 7–9).
The Biblical Canon and Beyond:
Theological and Historical Context of
the Codices of Alexandria

Heinz-Josef Fabry

Presented to Emanuel Tov on his 70th Birthday

1. The Question

After the problematisation of the MT induced by the Dead Sea Scrolls

(DSS) and their early manuscripts, we cannot “longer posit MT at the cen-
ter of our textual thinking”;1 rather, we have to devote more attention to
the Hebrew “Vorlage” of the LXX. “We should not (any longer) presume
that the OG of any particular book was translated from the Masoretic
form of that book”.2 The fact is that the text form of the Vorlage is much
older than that of MT.
The well known “textual variety” in the biblical DSS testifies to the
extended textual variability of the text of the Hebrew Bible and the
Christian Old Testament during ancient times. The textual history shows
a gradual phase of stabilisation. This stabilisation may be caused both by
text-internal reasons (such as corrections of obvious mistakes) and, more
so, by text-external reasons (such as the need for a unique text for the
purpose of theological dialogue and defence against heresies). The stabi-
lisation of the Greek text is visible in the context of quotations of the Old
Testament (OT) in the New Testament (NT).3 The Hebrew text became
more solid much later for inner-Jewish theological reasons and because of

1  E. Tov, “Hebrew Biblical Manuscripts from the Judaean Desert: Their Contribution to
Textual Criticism,” JJS 39 (1988): 5–37, esp. 7.
2 E. Ulrich, “The Relevance of the Dead Sea Scrolls for Hexaplaric Studies,” in Origen’s
Hexapla and Fragments (TSAJ 58; ed. A. Salvesen; Tübingen: J. C. B. Mohr [Paul Siebeck],
1998), 401–407, esp. 403.
3 LXX in it earliest form as “Old Greek” has—undoubtedly—preserved a very early stage
of the text in the mode of translation. But E. Tov has instructed us about the splitting of text
traditions, as obviously extant in the Greek manuscripts in the DSS (OG: 7Q-papLXXEx,
4QLXXLeva, 4QpapLXXLevb and 4QLXXNum) and 8ḤevXIIgr (Proto-Theodotion; kaige).
22 heinz-josef fabry

its dissociation from the Christian Bible in the time of the Masoretes from
the 6th until the 9th centuries C.E.4
Against this background, the heuristic hypothesis of the present paper
is as follows: After the Greek text became fixed in the first centuries C.E.
and in the time the MT became fixed, the intentional interventions into
the text of the Bible came to an end. After that time the only opportunity
to introduce Bible-hermeneutical ideas into the text consisted of changing
the order of the biblical books (over-all composition). We will start the
survey by pointing out that the big codices of Alexandria in the 4th and
5th centuries C.E. are examples of interpretation by composition.
Since the LXX text-line shows “textual variety”, E. Ulrich rightly asks: “It
is instructive, if perhaps unfair, to ask the question again this way: which
witness presents an earlier or more original text, the second- and first-
century B.C.E. Qumran manuscripts, or the fourth- and fifth-century C.E.
Vaticanus and Alexandrinus, or the medieval minuscules?”5 The codices
A, B, and S are actually not congruent at all, so that an exhaustive com-
parison of the texts would be necessary.
This is in fact our actual issue. It is striking that in Alexandria in the
4th century C.E. three Greek Bibles (OT and NT) were prepared in a very
short timeframe. The texts of these Bibles are very close, since they origi-
nate possibly from the same Greek textual source. Scholars have certainly
worked off the variants of the codices from a text-critical perspective. This
is not our aim today. Our task, instead, is trying to understand whether
and to what extent these three Greek Bibles may differ in the arrange-
ment of the books which are included in the codices. It is also evident,
furthermore, that such compositional differences also extend to the NT. In
this regard, I intend to demonstrate that the arrangement of the biblical
books is not arbitrary; it rather reflects a specific system, which is what
I am going to focus on here.
To my surprise, this question seems to be new and very unusual, since
there is no hint of it in the history of research in this field. The three codi-
ces are normally not issues of exegetical or of patristic research. Therefore,
I think that the time has come to study the diverging arrangement of the
biblical books.

4 H.-J. Fabry, “Das ‘Alte Testament’,” in “What is Bible?” Part II.: Between Torah and Bible
(eds. K. Finsterbusch and A. Lange; Gütersloh; Gütersloher Verlagshaus, 2012) (in press).
5 Ulrich, “The Relevance,” 406.
the biblical canon and beyond 23

2. Stages of the Early History of the Christian Bible Canon

2.1. Guidelines from Early Judaism

It is difficult to get reliable information on the Jewish canon—apart from
the anachronistic terminology—in the 1st century C.E. We know a few
things about several discussions and disputations about canonical items,
such as the “Synod of Jamnia”. This “Synod”6 discussed purely Jewish
issues; there was not any controversy with or about early Christianity.
And also we know nothing about any discussions or decisions about the
arrangement of the biblical books.

2.1.1. The first available pieces of evidence on the arrangement of the

biblical books in the early Jewish tradition come from Flavius Josephus
(C. Ap. 1.7f.).7 His principle of arrangement follows the precise chrono-
logical sequence (ἀκριβὴς διαδοχή) from the creation until Artaxerxes,
which extends over all three parts of the text: 5 Books of Moses, 13 pro-
phetical books ( Job-Esther) and 4 poetical books (Ps/Prov/Qoh/Song). All
these books are about the revelation of God and were written in the time
between Moses and Artaxerxes. Later books are without prophetical suc-
cession. The hermeneutical principle of the list of Josephus and later Tal-
mud is historically orientated: the world and history are creations of God,
who leads mankind through history until the present time. The world of
today should be understood as being the result of God’s history and His
Magnalia Dei.

2.1.2. A Baraita in b.Baba Batra 14b (5th/6th century C.E.) provides an

arrangement of the 22 biblical books that goes back to older traditions.8
The three parts of the biblical books are the same as in Josephus. But
now we discover an obvious regrouping of the books, a reduction in the

6 To the problems of a “Synod of Jamnia” cf. A. Hahn, Canon Hebraeorum—Canon

Ecclesiae. Zur deuterokanonischen Frage im Rahmen der Begründung alttestamentlicher
Schriftkanonizizät in neuerer römisch-katholischer Dogmatik (Studien zu Theologie und
Bibel 2; Berlin: LIT-Verlag, 2009), esp. 182–193.
7 Cf. W. Fell, “Der Bibelkanon des Flavius Josephus,” BZ 7 (1909): 1–16; R. Beckwith, The
Old Testament Canon of the New Testament Church and its Background in Early Judaism
(Grand Rapids: Eerdmans, 1985), 78ff.
8 Beckwith, Old Testament Canon, 154–162. Beckwith suggest that this list has its origin
in Judas Makkabäus (164 B.C.E.), but see the refutation given by P. Brandt, Endgestalten
des Kanons. Das Arrangement der Schriften Israels in der jüdischen und christlichen Bibel
(BBB 131; Berlin: Philo Verlagsgesellschaft, 2001), 64.
24 heinz-josef fabry

quantity of prophetical books, because the books of Job, Dan, 1 & 2 Chron,
Ezra/Neh and Esther are now allocated to the Hagiographa. The historical
arrangement is dropped. But over all we find a strong stability in the basic
arrangement: the Moses books (not yet classified as torah), prophets and
writings. That corresponds to the Jewish canon.

2.2. The Early Christian Discussion

We would like to have pieces of evidence about the Jewish Bible in the
time of Jesus and even about Jesus’ conception of canonical books. His
vivid disputation against the Pharisees (Matt 23:35) may be of some rele­
vance, as Jesus speaks about the innocent blood of Abel (Gen 4) and of
Sacharja, son of Barachiasi (2 Chron 24:20–22); in fact, Jesus seems to
delineate a canonical outline from Genesis to 2 Chronicles. This single
piece of evidence, however, is not significant enough.9
It may be that there are a few lists of biblical books from early Christian
times (especially against the heretic position of Marcion), but the most
important list until now is the one of Melito, where we find the Chris-
tian idiosyncrasy of arranging the books in (a) historical, (b) wisdom and
(c) prophetical books. E. Zenger has told us about the main hermeneutical
differences in his “Introduction”.10 Later Church Fathers changed the pro-
phetical corpus in presenting first the Minor Prophets and then the Great
Prophets. In the following arrangements, one may detect some confusion,
but a closer investigation will bring the main arguments of the Fathers
to light.
First: The Hebrew Bible lost its primary role bit by bit, first in the West-
ern and later in the Eastern Church.11 The deuterocanonical books gradu-
ally became canonical.
Secondly: The middle position of the Prophets in the Jewish canons
was shifted very early to the final position by the Eastern Church Fathers
(Athanasius of Alexandria, Cyril of Jerusalem, Synod of Laodicea, Codex
B), with the aim to present the OT as the promise of Jesus Christ. Origen

 9 Cf. L. M. McDonald, The Formation of the Christian Biblical Canon (Peabody: Hen-
drickson, 1995), 46f.
10 E. Zenger, Einleitung in das Alte Testament (ed. C. Frevel; Stuttgart: Kohlhammer,
82012), 22–36.
11  R. Hennings, Der Briefwechsel zwischen Augustinus und Hieronymus und ihr Streit um
den Kanon des Alten Testaments und die Auslegung von Gal. 2,11–14 (Leiden: Brill, 1993), 189:
“Im Osten wird zwar die Autorität der Septuaginta nicht bestritten, aber zumindest in der
Kanontheorie tritt sie dort hinter der hebräischen Bibel zurück.”
the biblical canon and beyond 25

and Epiphanius steered a middle course. The Western Church (Synod of

Hippo, Synod of Carthago) held partly to the Jewish arrangement (S und A;
Hippo; Carthago against Codex B and Augustine).
Thirdly: The composition of the New Testament created a new situa-
tion, since now we have to respect mutual parallelisms and influences.

3. The Codices and Their Peculiarities

The Jewish arrangement of the biblical books remained nearly inflexible;

the Christian ones, on the contrary, remained modifiable until present
times. This is precisely the situation of the codices. An analysis of the
arrangements given by the Church Fathers clearly indicates that these
arrangements were not arbitrary.

3.1. H. B. Swete12 suggested that the arrangement of books according to

their chronology, literary nature, contents and authors reflects a typical
idiosyncrasy of the Alexandrian scholars. This explanation seems doubtful
to me. Therefore, we have to inquire about the reasons for the variations
in the arrangements. They are, as I argue, not arbitrary.

3.2. Our second issue is the quest for intentions. The canon lists until the
4th century were mostly homogeneous, but the arrangements in the codi-
ces departed from the established ways. Even if they did not reject their
Jewish roots, they revealed their peculiarity (a) in cancelling the tripartite
structure; (b) in regrouping the books of the prophets; (c) in accepting
the deuterocanonical writings; and (d) in compilating the books of the
Tanakh—now called the “Old Testament”—with the books of the early
Christianity—now called the “New Testament.” These compilations were
new Bibles with clear Christian intentions. They dismissed the historical
arrangement and depreciated the Pentateuch as Torah. It is amazing that
this new Christian hermeneutic was the starting point of a long develop-
ment; it had to grow and to try out several possibilities, and it needed
time to become consistent. Now we have to start our survey of the three
codices to find out how they disagree—despite their nearly concomitant
compilation of the books—and how they combine their Jewish and Chris-
tian hermeneutics.

12 H. B. Swete, An Introduction to the Old Testament in Greek (Cambridge: Cambridge

University Press, 1902 [repr. 1968]), Pt 2 Ch. 1.
26 heinz-josef fabry

4. Codex Vaticanus (B)

4.1. Arrangement of Books and Structure

Codex B is the oldest one; it reflects the tradition of the 39th Easter Letter
of Athanasius (367) and possibly served as the Bible for Athanasius, the
patriarch of Alexandria. B was written in Alexandria (or in Caesarea). The
Pentateuch, which is here understood as a collection of historical books,
is extended to an octateuch (Gen-Ruth), which is followed by a second
octateuch (1 & 2 Sam, 1 & 2 Kgs, 1 & 2 Chron, 1 Esd, Ezra/Neh). Subse-
quently the wisdom literature and the Trilogy of the Persian period follow
against the Jewish arrangement of the books. Tobit—whom Ambrosius
sees as a prophet13—prepares the transition to the corpus propheticum
(3rd Octateuch): first the Minor Prophets, after that the Great Prophets
with Baruch, Threni and EpJer. Daniel represents the apocalyptic conclu-
sion in the same way as in the lists of Cyril and Athanasius.
Athanasius designated Wisdom, Sirach, Esther, Judith and Tobit as
“ἕτερα βιβλία”, Cyril of Jerusalem as “ἀπόκρυφα”, but codex B now accepted
these books as fully canonical wisdom literature and gave them a place
between Job and the Minor Prophets. One may argue that this codex
reflects the biblical theologies of Athanasius and Cyril: moreover, the
place of wisdom literature before the prophetical books may reflect, on
the one hand, a conscious distance from the Jewish arrangement and, on
the other hand, a required dialogue with the philosophy in Alexandria.

4.2. The New Testament Books

The arrangement of the Old Testament books is to some extent parallel to
that of New Testament books, with Gospels, Acts (par. 2 historical octa-
teuchs), Letters of the Apostles (par. wisdom-octateuch) and the Book of
Revelation (par. prophetical octateuch). This is the arrangement given by
Cyril (sequence: Catholic letters-Pauline letters), which does not corre-
spond to the order of Athanasius. What is the ruling hermeneutic idea?

4.3. The Hermeneutical Concept

We have to be aware of the parallelism of both parts of the Bible. Appar-
ently, the book of Daniel stands in the final position (as in the list of the

13 Ambrosius, De Tobia (CSEL 32/II, Wien: Tempsky, 1897) 519.1 et seq.; 520.20 et seq.
the biblical canon and beyond 27

Greek Church Fathers Athanasius, Cyril of Jerusalem, Synod of Laodicea)

and forms a possible parallel to the book of Revelation.14 Here one may
possibly find the clue to the pan-biblical hermeneutic. This book deals
with the eschaton, the end of secular powers, the expectation of chang-
ing of all. The book of Daniel amounts to its final sentence: “But go your
way till the end. And you shall rest and shall stand in your allotted place
at the end of the days” (Dan 12:13). Such a final conclusion of the first part
of the Bible is an impressive bridge to the New Testament. The motifs of
the eschatological καιρός and the resurrection are central points of the
New Testament as well; moreover, the concept of the Christians as real
heirs of the Old Testament (see Letter of Barnabas and Irenaeus, Bishop
of Lugdunum in Gaul [Lyon]) lies behind this allocation!

4.4. The Church-Historical Background

The determination of the church-historical background depends on the
precise dating of the codex. If we accept the first part of the 4th century,
then we are close to the time of the Council of Nicaea (325), convened
by Constantine the Great. The main issue addressed by the Council was
the heresy of Arius and the struggle against him. His chief opponent
was Alexander, Bishop of Alexandria. The codex is probably written by
order of Constantine himself.15 Perhaps it was a gift of Athanasius for the
Roman Bishop,16 when he became bishop of Alexandria. The main issue
for the Council was the formula ὁμοούσιος, the consubstantiality of Jesus
Christ with his Father. In this regard, the Fathers of the Council attached
importance to the decisions of the Egyptian theologians, because of their
old pharaonic traditions, where the Pharaoh as Horus was consubstantial
with the Sun-God Re-Harachte.
The next issue at the Council was the dispute about the correct date
for the Easter feast. According to the calendar indications of the OT, the
date of the feast corresponded to the Sunday after the 14th of Nisan. This
issue may also be a reason to finalise the arrangement of OT books with
Daniel and the Easter hope of resurrection (Dan 12:13). And finally, we
can assume that the delight about the restoration of unity in Christian
belief under Alexandrian guidance could have been an excellent reason

14 The book of Revelation took long time getting canonical acceptance.

15 T. C. Skeat, “The Codex Sinaiticus, the Codex Vaticanus and Constantine,” JTS 50
(1999): 583–625.
16 As often suggested by A. Rahlfs.
28 heinz-josef fabry

for preparing the wonderful and precious copy of “the unique Bible”, the
Codex Alexandrinus.

5. Codex Sinaiticus (‫א‬/S)

5.1. Arrangement of Books and Structure

Codex S dates back to the first half of the 4th century. The text belongs
to the Alexandrian type, but displays a few Western readings. Since the
Codex is not complete, it is difficult to draw significant conclusions. Its
arrangement is close to the canon list of Athanasius, accepted later by the
Council of Laodicea. Here, however, one finds restrictions in accepting the
deuterocanonical writings included 1–4 Maccabees. Those texts follow, in
accurate historical sequence, the books Chron and Ezra; that means they
are historical books.
Since the main part of the historical books is lost, it is difficult to recon-
struct the initial structures of the Codex. It might be possible that the
historical section of the Codex (Gen-1 & 2 Ezra) was identical or similar
to that of Codex B and A. The trilogy from Persian times (Esther, Tobit,
Judith) and the (4?) Maccabees follow the historical part closely. After
that, however, the Codex unexpectedly goes its own way. Next in line is
not the wisdom literature—as in the other Christian codices—but, accord-
ing to the Jewish tradition, the prophets—first the Great Prophets, then
the Minor Prophets. Like the Jewish Bibles, the wisdom literature (not all
the ketûbîm though) forms the final books of the arrangement. The
Psalms—since the time of Ambrosius17 known as the Davidic prophecy
and the prophetic announcement of Jesus Christ—follow immediately
after the Minor Prophets.
Thus the main structure is close to that of the Talmud, where the wis-
dom literature also takes the final position. This alignment with the old
Jewish structure is surprising and, at the moment, enigmatic. It harks back
to the same tradition that is taken up again by Jerome, but now the book
of Job is at the end of the compilation. This is the unique characteristic of
the Codex Sinaiticus and how to interpret this unusual position is unclear.
Either Job is only a negligible annex—this is the assumption of Theodoret,
Bishop of Mopsuestia—or its position is of absolutely prime value. Since

17 Ambrosius, Explanatio Psalmorum I.8 (ed. M. Petschenig, Explanatio Psalmorum,

CSEL 64, Wien: Tempsky, 1999).
the biblical canon and beyond 29

Theodoret’s impeachment against Job had no consequences in the later

literature, I am convinced that the final position of Job corresponds to a
very deliberately intended theological concept.

5.2. The Books of the New Testament

The arrangement of the books of the NT is very peculiar in this Codex:
at the beginning are the Gospels, the Pauline letters, the Letter to the
Hebrews and the Pastoral letters, next(!) the Book of Acts, the Catholic
letters, and finally the book of Revelation. Perhaps the New Testament
canon was still under dispute, because we find in final position the Letter
of Barnabas and the “Shepherd of Hermas”.

5.3. The Hermeneutical Concept

If we look at the OT part of this Codex, we see the development of pro-
phetical ethics on top of the historical foundation, which acts as their
basis. The arrangement proceeds from history through the prophets to
wisdom. Wisdom is no longer placed before the prophets: rather, it is
understood as the result of learning from history and from prophets. Wis-
dom is pointed out as the goal of the OT, the perspective to the scriptures
about Jesus Christ. Mankind needs a sound knowledge of the Magnalia
Dei to organise life successfully, with wisdom on the base of prophetical
ethics. By placing the book of Job (including the notice of the resurrec-
tion in Job 42:17aLXX) in the final position, the Codex intensifies the idea
of resurrection and states more precisely that only human beings like Job,
who made a great effort to achieve wisdom and a successful life, will have
a share in resurrection after death. In any case, mankind will find substan-
tial wisdom in Jesus Christ, the first of the resurrected ones.

5.4. The Church-Historical Background

The church-historical background of this Codex may be not substantially
different from Codex Vaticanus. S was written possibly in the histori-
cal context of the struggle of the Arians against the highly controversial
Athanasius, Bishop of Alexandria. His opponents were successful in expel-
ling him from Alexandria to go and live with the monks of the Egyptian
desert. The dispute about the Nicaean formula ὁμοούσιος replaced by the
formula ἀνόμοιος, proposed by Aëtius of Antiochia, found many support-
ers in the Synod of Sirmium. Basileios of Ankyra proposed the alterna-
tive draft ὁμοιούσιος. The oppressive Synod of Seleukia/Rimini (359) gave
30 heinz-josef fabry

the formula ὁμοῖος τῷ πατρί. That was the end of the Nicaean formula for
the moment. “The world groaned and was astonished to find itself Arian.”18
After the proclamation of Julian the Apostate as emperor in 360, violent
pogroms broke out against the Christians as “godless Galilaeans,”19 espe-
cially in Alexandria. Christian doctrine was consequently excluded from
the programmes of education and Christian teachers were expelled. This
historical context is a good place for the Sinaiticus.

5.4.1. First: the structure of this Bible will demonstrate where to look for
true education. The Sinaiticus understood “re-Judaization” as the chosen
way against the persecution of the Christian as “godless Galilaeans.”

5.4.2. Secondly: the Sinaiticus essentially intended to give back the Tal-

mudic Bible to the Christians, in order to enable them to find again the
true roots of the doctrine of God in the time of christological turmoil. In
view of the old thesis of Hellenistic philosophers “presbyteron kreitton,”20
the Christians had to know more and more about the God-Father in order
to know the God-Son. The only way to achieve this knowledge is to read
the Bible preferably in its original form with the revelation of the Father.

5.4.3. Thirdly: the persecution of Christians in the time of Julian the Apos-

tate produced some martyrs, who were very soon greatly respected and
worshipped. Like in the veneration of saints, the martyrs were put side by
side with the great figures of the OT: Abraham, Moses and David; perse-
cuted prophets, Eleazar and the Maccabees were also added. Their vio-
lent death was an anticipated martyrdom and testimony for Jesus Christ.21
This is the reason for the acceptance of the books of the Maccabees
(1–4 Macc) in the Codex.

5.4.4. Fourthly: if we remember Didymus the Blind of Alexandria, we have

to acknowledge that he called Job a “visionary seer” of the incarnation of
God, on the one side, but, like the Sadducees in NT, he had great difficul-
ties in thinking about resurrection, on the other side. The ­idiosyncrasy

18 “Ingemiscens Orbis terrarum se arianum esse miratus est,” Jerome, Adv. Lucif. 29.
19  Julian, Ep. 84.
20 To this criterion cf. P. Pilhofer, Presbyteron Kreitton. Der Altersbeweis der jüdischen
und christlichen Apologeten und seine Vorgeschichte (WUNT II/39; Tübingen: J. C. B. Mohr
[Paul Siebeck], 1990), 17–75, 143–220.
21  K. Baus, Handbuch der Kirchengeschichte II/1 (Freiburg: Herder, 1973), 338.
the biblical canon and beyond 31

of the Sinaiticus is now to declare itself as the orthodox response to the

problematic of martyrs and to denounce Didymus in the course of the
struggle against the scholars of Origen.

6. The Codex Alexandrinus (A)

6.1. Arrangement of Books and Structure

Codex A was written in the first half of the 5th century C.E. after Athana-
sius, and the tractates of Eusebius and Athanasius are in fact included in
the text of the psalms. Codex A is the youngest of the three.22 The text is
divided in several octateuches: (1) Gen-Ruth; (2) 1 Kgs-2 Chron (Ezra/Neh
are absent); (3) XII-Dan; (4) Esther-3 & 4 Macc; (5) Ps-Sir (incl. Odes). The
books of the prophets begin with the Minor Prophets (as in B and in con-
trast to the MT) and close with Daniel. The trilogy of the Persian period
follows, enlarged by Ezra and Maccabees.23
It is noticeable that the Corpus Propheticum is completely framed by
historical books. Perhaps the prophets are intended to be seen through the
lenses of the historians. As in Sinaiticus, the wisdom literature is arranged
next. The sequence Ps (David)/Odes/Job/Prov (Solomon) is noteworthy,
because it claims Job as the contemporary of Solomon and of the Queen
of Sheba.24 The final position of Sirach is the most enigmatic peculiarity
of this Codex.

6.2. The Books of the NT

The NT books are arranged in a way which is very similar to that of Codex
B: Gospels, Acts, Catholic letters, Pauline letters, Letter to the Hebrews,
Pastoral letters and the book of Revelation. The NT is enlarged by 1 & 2
Clemens and Pss Sol. The Pss Sol seem to be in a marginal position.

22 The Patriarch Cyril noted that the book was written according to the transmission
of Thecla, a noble Egyptian woman, shortly after the Synod of Nicaea, and that her name
could be found at the end of the Codex (now fragmented and lost). Even if such a dating
is surely too early, the relevance of this tradition consists in the reference to Egypt.
23 The Trilogy Ezra-Dan-Esther occurs here for the first time. The trilogy is well attested
from the 3rd century C.E. onwards (Chester Beatty IX und X).
24 Cf. R. Nathan, 2nd century C.E.
32 heinz-josef fabry

6.3. The Hermeneutical Concept

The final emphatic position of the wisdom literature points to the same
theological pedagogics as in Codex Sinaiticus. The interpretation of the
books of the Prophets as historical books seems very strange and inexpli-
cable. The search for any hermeneutical concept has to look at the final
position of Sirach, where we read: “Do your work before the appointed
time and at the appointed time (ἐν καιρῷ) he will give you your reward”
(Sir 51:30). The “appointed time”, καιρός, is the very distinctive concept of
NT scriptures. The NT books are closely bound to the OT books through
the significant καιρός terminology. The eschatological perspective of
Sir 51 is a close bridge to the eschatologically imminent expectation in
the Gospels.

6.4. The Church-Historical Background

The determination of the church-historical background depends here also
on the precise dating of the Codex, which can be determined only approx-
imately. One may refer to three possible historical contexts.

6.4.1. The turn of the century was characterised by the argument of the

doctrine of the trinity. The emperor Theodosius promulgated the Edict
“Cunctos populos” (380) in order to oblige Christians to adhere to the
formula of Nicaea and to the belief in the godly trinity. The Trinitarian
formula of the Council of Constantinople (381), however, missed the
ὁμοούσιος quality of the Holy Spirit. That was the reason for a long con-
flict between Theodoret of Kyrrhos and Cyril of Alexandria (375–444); the
right interpretation of the ὁμοούσιος of the son was disputed again. When
Nestorius became bishop (428) in Constantinople, however, a new front
against Antiochia was built up. That may be the actual context of Codex

6.4.2. A second setting was shaped by Jerome and Augustine. Jerome was
Abbot of a monastery in Bethlehem and he was a zealous heir of Ori-
gen. But when Epiphanios of Salamis tried to condemn Origen, Jerome
changed his mind—to the astonishment of all. John, Bishop of Jerusa-
lem, and Rufin, translator of the works of Origen, became his opponents.
Jerome was encouraged by Theophil of Alexandria, who took up the
struggle against John. In Rome, Rufin wrote the voluminous work Apolo-
gia contra Hieronymum, to which Jerome responded with a multi-volume
the biblical canon and beyond 33

polemic paper against Rufin. Theophil succeeded in forbidding the read-

ing of the books of Origen.

6.4.3. In Hippo, Augustine became bishop in 395. Together with Aurelius

of Carthage, he suggested the annual Synod of North African bishops. The
Synod of Carthage (398) adopted a canonical decision: Wisdom, Pss Sol,
Tobit, Judith, and 1 & 2 Macc were decided as fully canonical. Beside this
decision, Augustine himself developed his own list of canonical books.
He subdivided the OT into historical (Gen-2 Chron) and prophetic books
(Ps-2 Ezra). He allocated the wisdom literature to the prophets and Job to
the historical books.
It is quite possible that the Codex Alexandrinus has its place in the
context of these controversies. On the one side, the Alexandrinus differs
distinctly from Origen in regrouping the canon list of Origen. On the other
side, the Codex does not accept the list of Jerome, because of its arrange-
ment, which was very similar to the Jewish one. Finally, the Codex does
not agree with Augustine and the Synod of Hippo. In this way, Alexandria
proved its neutral and impartial stance—for the last time. A few years
later the patriarch Dioscuros was banished by the Council of Chalcedon
(451). After that, the theology of Alexandria was declared monophysitic
and heretical.

7. Perspective

7.1. This survey has hopefully shown that the different arrangements of the
biblical books in the codices are not only well reasoned, but also reflect
clear theological intentions. Despite their imprecise dating, the codices
seem to point to specific purposes and intentions, which plausibly reflect
the prevailing historical controversies and originate from them. Writers
and compilers were actually theologians, who discussed the christological
and trinitarian questions asked by dogmatically informed theologians, in
a prudent and wise manner.

7.2. Writers and compilers did not actively modify these texts; they only
regrouped them in different ways. In this regard, my survey has extended
the possibilities of composition criticism.

7.3. Apart from these analyses of the early Christian Bibles, we have to

keep in view the parallel development(s) of the Hebrew text until its
34 heinz-josef fabry

final standardisation as MT at the end of the 1st millennium. This text

remained relatively unmodified until the time of the Alexandrian codices.
At least since the 6th century a small but important regrouping should
be noticed. The book of Lamentations was separated from Jeremiah, thus
forming an independent book. Ruth was separated from Judges and its
historical context, and regrouped to the ketûbîm. That was a crucial point
on hermeneutical thinking. The obvious early arrangement of the histori-
cal books Gen-Ruth as octateuch was accepted by all Christian Bibles and
understood as the basic document of Israel’s self-consciousness and iden-
tity. In Christian thinking, there was a close bridge between that octa-
teuch and the Tree of Jesse, the family-tree of Jesus in Matt 1, where Ruth
has a prominent place. In this sense, the OT was nothing other than a
salvation-historical introduction to Jesus Christ. The midrashic Jewish tra-
dition transferred the book of Ruth somewhere in the ketûbîm and broke
down this bridge. The octateuch was reduced to the Pentateuch.
In order to be able to understand those processes, we have to go back
to the interaction between the Jewish and Christian Bibles; we have to
analyse such interaction very attentively and understand it as a vivid and
fruitful dialogue between the religions. Once again: there is still a lot of
work to be done.
Part Two

Translation Technique and Text History

Revisiting the Rock: Tsur as a Translation of Elohim in
Deuteronomy and Beyond

Melvin K. H. Peters

The title of this paper telegraphs its objective. Whereas “revisiting”

acknowledges that the subject has been much discussed,1 the subtitle
signals my intention to demonstrate that the representation of ‫ אלהים‬by
the epithet ‫( צוּר‬or ‫ )הצוּר‬throughout the Hebrew Bible was the result
of deliberate, theologically motivated, editorial or redactional activity by
Hebrew tradents whose work was later adopted in the early centuries C.E.
as the received text of the Hebrew Bible.2 Therefore, such Hebrew read-
ings are subsequent to, or distinct from, those reflected in the Hebrew
parent text of the Septuagint, which consistently reflects Elohim in every
In a recent paper3 presented at a conference celebrating the publica-
tion of the New English Translation of the Septuagint,4 I suggested that in
Deut 32 the Septuagint Vorlage was prior to the received text, and that
modifications preserved in MT were in some instances due to theological
sensitivities. Discussion of ‫ צוּר‬in that paper was limited to some ­cryptic

1 See, for instance, the entries on ‫ צוּר‬by Thiel and Fabry in TDOT and the extensive
literature cited there.
2 This claim can obviously not be demonstrated empirically. However, it rests on care-
ful observation of the patterns of deviation between the Leningrad Codex and the Hebrew
behind the Septuagint throughout the biblical corpus. It is based, furthermore, on the
demonstrable premise that the Hebrew text of the MT was not the only one in circulation
in the pre-Christian centuries and therefore was not always or necessarily the basis for
Septuagint readings. Textual pluriformity has now been well established, especially since
the discovery of the Dead Sea scrolls. On this, see for example, the articles in A. Schenker,
ed., The Earliest Text of the Hebrew Bible: The Relationship between the Masoretic Text and
the Hebrew Base of the Septuagint Reconsidered (SBLSCS 52; Atlanta: Society of Biblical
Literature, 2003) especially E. Tov’s “The Nature of the Large-Scale Differences between the
LXX and MT S T V, Compared with Similar Evidence in Other Sources,” 125–143.
3 M. K. H. Peters, “Translating a Translation: Some Final Reflections on the Produc-
tion of the New English Translation of Greek Deuteronomy,” in Translation is Required:
The Septuagint in Retrospect and Prospect (ed. R. J. V. Hiebert; Atlanta: Society of Biblical
Literature, 2010), 119–134.
4 A. Pietersma and B. G. Wright, eds., A New English Translation of the Septuagint and
the Other Writings Traditionally Included under That Title (Oxford / New York: Oxford Uni-
versity Press, 2007).
38 melvin k. h. peters

comments, which are now amplified and supplemented with a more care-
ful examination of the entire phenomenon in Deuteronomy and beyond.
But first, some statistics.
The root ‫ר‬-‫ו‬-‫ צ‬appears 217 times in the Hebrew Bible in 19 different
forms. The largest group consists of the simple noun ‫“ צוּר‬rock,” occurring
73 times, followed by various verbal, adjectival or other nominal forms
such as place or personal names. Within the main cluster of nouns are
thirty-six5 where the basic meaning “rock” seems inappropriate, appear-
ing instead to be a substitute for the divine name. Exactly half of those
instances are in the Psalter; the others are scattered between Deut (7), 1&2
Sam (6), Isa (4) and Hab (1). All but one (Isa 30:29) are situated in “poetic”
passages. The first occurrence of ‫ צוּר‬6 is in Exod 17:6 where the god of
Israel is portrayed as standing ‫“ על הצור בחרב‬on the rock at Horeb.”
This may well be the basis for the identification of ‫ אלהים‬with ‫ צוּר‬in later
traditions. The first instance of the specialized use of the term occurs in
Deut 32:4, which in the LXX is rendered θεός; in the received tradition as
‫“ הצוּר‬the Rock.”
The main question of this paper then, is this: Did the translator of Greek
Deuteronomy have before him a text identical to the Leningrad Codex
and intentionally avoided transferring ‫ צוּר‬into Greek (or at least trans-
literating it) or does the translation θεός indicate the existence of a text
containing ‫אלהים‬, which the LXX translator faithfully rendered? Histori-
cally, the overwhelming response to that question in critical scholarship
has been that the Greek translator indeed had the received text before him
and was interpreting or even commenting upon it. The standard narrative
runs roughly as follows. Hellenistic Septuagint translators recoiled at the
presence of such “crass metaphors” for the divine, preferring instead to
use the name itself. Here for example is the critical commentary on this
verse offered by Carmel McCarthy in her recently published edition of
Biblia Hebraica Quinta of Deuteronomy.
“Rock.” When used figuratively of God, was always a problem for the early
translators. G and the other versions in varying typical ways, reinterpret this

5 Deut 32:4, 15, 18, 30, 31(2x), 37; 1 Sam 2:2; 2 Sam 22:3, 32, 47(2x); 23:3; Isa 17:10; 26:4;
30:29; 44:8; Hab 1:12; Ps 18(17):3, 32, 47; 19(18):15; 28(27):1; 31(30):3; 49(48):15; 62(61):3, 7, 8;
71(70):3; 73(72):26; 78(77):35; 89(88):27; 92(91):16; 94(93):22; 95(94):1; 144(143):1.
6 This presumes that the Pentateuch was the first unit collected in the Hebrew Bible,
and that the current order of the books reflects the order of their composition.
revisiting the rock 39

metaphor according to its sense here and at vv. 15, 18 and 30–31 below. See
also v. 37 regarding its use in relation to pagan gods.7
Quite apart from McCarthy’s unqualified use of the word “God,” at least
two striking things are implicit in her comments. First, is the clear notion
that the L text, the so-called MT,8 was the basis for the Septuagint transla-
tion, and second that the translators of “versions” presumably “consulted”
each other to determine their responses to textual problems within it.
Neither of those presuppositions is necessary or compelling.
The Septuagint translators being first had no other version with which
to consult. Translators of subsequent versions could have been aware of
the Greek and influenced by it, but the reverse was not possible. LXX
translators were flying solo; their only points of reference would have
been Hebrew. Indeed, the choice of the word “versions” in so much of
contemporary text critical discourse continuously reinforces the mislead-
ing idea of a single “original” from which copies were made. So, when
modern scholars group together readings found in disparate so-called ver-
sions and draw textual conclusions based on apparent shared patterns,
they risk missing important clues in specific forms of the text, especially
the Septuagint.
But McCarthy is not alone; several, perhaps most, scholars share such
views. Here for instance is John Wevers some years earlier in his Notes on
the Greek Text of Deuteronomy on this verse:
The term ‫ צוּר‬applying to God occurs not only in this verse, but also at
vv. 15, 18, 30, 31, and 37. The word means “rock,” but it is never translated
thus in LXX, which substitutes θεός throughout, except for v. 37 which see.
Presumably the term was used as a poetic name for God to designate the
deity as the one who was solid, firm, unmoveable, but the translator consis-
tently avoided a direct translation, thereby precluding any possible misun-
derstanding of the metaphor.9
Like McCarthy, Wevers uses the word God uncritically, adding theological
characteristics, which he presumes ancients held about divinity. He states

7 C. McCarthy, Biblia Hebraica Quinta editione cum apparatu critico novis curis elabo-
rato, Fascicle 5, Deuteronomy (Stuttgart: Deutsche Bibelgesellschaft, 2007), 139*–140*.
8 Hereafter MT is used only as a convenient equivalent for the Leningrad Codex, and is
not intended in any way to suggest either that there was a single Masoretic text or that the
text of L was (or should be) considered normative at the time the Septuagint Pentateuch
was translated.
9 J. W. Wevers, Notes on the Greek Text of Deuteronomy (SCS 39; Atlanta: Scholars Press,
1995), 510.
40 melvin k. h. peters

that ‫ צוּר‬is substituted by θεός throughout the LXX, a claim that we will
examine presently. Quite clearly Wevers had no doubt that it was the Sep-
tuagint translator who avoided a direct translation in order to preclude
any possible misunderstanding of the metaphor, and that a text identical
with the MT was before him.
The name most often identified with the subject of this paper is Staffan
Olofsson whose dissertation book,10 God is My Rock: A Study of Translation
Technique and Theological Exegesis in the Septuagint, was published in
1990. Olofsson was concerned with theology in the Psalter primarily and
surveyed in the course of his work the use of certain epithets including
‫ צוּר‬in other parts of the Hebrew Bible. His book, as its full title indicates,
is representative of a wide range of scholarship that makes a priori judg-
ments about theological tendencies in the Septuagint and assigns char-
acteristic patterns to its translators as a group.11 Embedded in that type
of scholarship are presuppositions about the activities and motivations
of the translators of the Septuagint. For instance, that there was theo-
logical exegesis within the Septuagint constitutes a starting point for this
kind of research. Conclusions are thus arrived at deductively, not induc-
tively. Septuagint translators are sometimes described as incompetent, at
other times as wildly radical in their responses to the text before them.
Of course that text is always the Hebrew of the Leningrad Codex. Only
rarely (and grudgingly) do such scholars grant the possibility of a differ-
ent Vorlage; every effort is made to conform the Greek translation to the
reading of the so-called MT.
Increasingly, I have grown suspicious of subjective explanations by
modern textual critics for the differences observed in ancient witnesses
to the text of the Hebrew Bible. These explanations are frequently too pre-
dictable and at times so speculative as to be of little value. I have suggested
in another place12 a covert reason for the traditional privileging of Hebrew
over Greek when they diverge. Especially after Qumran, where it has been
clear for some time that multiple textual traditions were in circulation
in the pre-Christian centuries, the persistence of “MT ­fundamentalism”

10 S. Olofsson, God is My Rock: A Study of Translation Technique and Theological Exegesis
in the Septuagint (ConBOT 31; Stockholm: Almqvist & Wiksell International, 1990). Also
frequently cited is A. Passoni Dell’Aqua “La metafora biblica di Dio come rocca e la sua
soppressione nelle antiche versioni,” Ephemerides Liturgicae 91 (1977): 417–453.
11  Cf. the literature cited in Olofsson, God is My Rock, 167–184 for an overview of the
state of scholarship at that time.
12 See Peters, “Translating a Translation,” 132–134.
revisiting the rock 41

seems particularly puzzling and out of step with our contemporary state
of knowledge. The standard argument in defense of current text critical
practice, as I have heard it, is simply this: unless and until we can unearth
a Hebrew text that corresponds precisely to the readings in the Septuagint
in every instance, we will suspect and blame the Septuagint translator
whenever there is variance between the Greek and the Hebrew of Codex
Now let me be clear. I am not a blind defender of Septuagint readings;
I am fully aware of recensional and revisionistic activity within the his-
tory of the transmission of the Septuagint—activity that invites caution in
regard to uncritical acceptance in every instance of the Greek text as we
now have it. I am also aware of Christian interpolations and interpreta-
tions in parts of the Old Greek text. It also has been shown that Hellenistic
influences are clearly present in parts of the Septuagint, particularly with
respect to personal and place names and the like. But those facts do not
place the unified reading of all existing Greek witnesses under suspicion
at all times in my view. I respect the work of the LXX translators and trust
that what they said is indeed what they meant. They did not create alter-
nate readings in order to be malicious, or to annoy and frustrate research-
ers two thousand years later; they were engaged in a serious enterprise.
The fiction of the “dishonest”13 translator flourishes only as long as we
continue to maintain the idea of the infallible and immutable MT.
In this first case, Deut 32:4, the noun ‫ הצוּר‬is not at all difficult; it did
not pose a problem in this or any other context for the translator of Deu-
teronomy, as McCarthy claims. Nor was the translator “consistently avoid-
ing a direct translation, thereby precluding any possible misunderstanding
of the metaphor,” as Wevers claims. The translator knew the noun and
had rendered it accurately in other places. One need look no further than
v. 13 to find a clear example of ‫ צוּר‬in parallel with ‫ סלע‬both appropriately
translated as πέτρα. Furthermore, a closer look reveals that while three
divine names are represented in the Greek—θεός twice and κύριος once—
only one, ‫ ֵאל‬, is in Hebrew. Furthermore, ὅσιος is not the most adequate
translation of ‫שׁר‬ ֖ ָ ָ‫ י‬either.14 So it is not at all certain that MT was the

13 By “dishonest” I state more bluntly and put in sharper relief what others have termed
“taking liberties,” “interpreting freely,” “harmonizing” and so on. All such modern judg-
ments presuppose the existence of MT as the LXX Vorlage. See also the comments of Orlin-
sky later in this paper.
14 The word ὅσιος usually renders ‫חסד‬, while ‫שׁר‬ ֖ ָ ָ‫ י‬is generally translated εὐθύς (εὐθής)
“right” “upright.”
42 melvin k. h. peters

­ orlage of the Septuagint in this verse. And even if it were, it would seem
counter intuitive for a translator to engage in such an irreverent practice
as explicitly using the divine name at a time and in a community in which
avoidance of calling the name had become normative.15 Translators were,
after all, Jews translating sacred scripture.
Rather, to state the obvious, it would make just as much (or better)
sense to presume that the Vorlage of the Septuagint indeed had ‫ אלהים‬at
the beginning of the verse and ‫ יהוה‬at the end—reflecting a stage in the
Hebrew textual tradition before the selective substitution of the metaphor
‫ צוּר‬had become common practice—as it would be to postulate, without
basis other than prejudice in favor of the received text, that a simple noun
like ‫ הצוּר‬presented a problem for the translator. The noun θεός is not
the usual rendering for ‫ צוּר‬as McCarthy claims16 but overwhelmingly it
is for ‫אלהים‬. Thus, on its face, θεός in the LXX should signal to an unbi-
ased observer the presence of ‫ אלהים‬in the Vorlage, just as κύριος should
signal ‫יהוה‬, as it does at the end of the verse. And Wevers, despite his
unquestioned erudition, exhaustive study of Septuagint materials and
monumental body of work, had no greater insight into the mind of an
ancient translator than any of us. His presumptions are as good as ours.
Why would representation of the divine as an inanimate object, or trans-
literating or translating the word ‫ צוּר‬into Greek be more objectionable
than writing the divine name itself anyway?
But how might one test this hypothesis that the LXX Vorlage had ‫אלהים‬
and not ‫ ?צוּר‬The answer: with careful attention to the treatment of this
expression in the remainder of this chapter and, in time, throughout the
whole Hebrew Bible. That then will be the focus of the remainder of this
paper. Prior treatments of the problem have been either “theological,”
that is to say, they were concerned with Hellenistic views of the divine,
or they have been “text critical,” that is, they were focused on matters of
approaches to translation. What I am proposing is that this be considered

15 Indeed, the observable pattern of movement in the Hebrew Bible is progressively

away from vocalizing the divine name, whether moving from the specific Yahweh to
the more generic Elohim, or from Elohim to any number of substitutes—The Holy One
Blessed be He, One Ancient of Days, etc.,—to ultimately avoiding calling the name alto-
gether in later periods. Analogously, the word choice in the LXX for instance of ἀλλόφυλοι
for Hebrew Philistine appears to be but one of many examples illustrating a tendency to
move away from specific forms to more generic and interpretative ones. Thus the move-
ment from ‫ צור‬to ‫אלהים‬, if the former were in the original, would appear to be out of
16 McCarthy, Deuteronomy, 151*.
revisiting the rock 43

a simple “literary-critical” project akin to other well-known examples of

variant forms of the text of the Hebrew Bible.
One question that could guide an inquiry of this kind would be this:
Were the adjustments based on theological/ideological grounds or were
they part of a mechanical process of “find and replace?” If they were made
for theological reasons, one would expect to find ‫ צוּר‬limited exclusively
to the god of Israel, who had come to be viewed in this way. Furthermore,
one might expect to find it replacing YHWH in places, since both forms of
the divine name were regularly interchanged. If on the other hand there
was a global process to selectively “find and replace” ‫ אלהים‬with ‫צוּר‬,
one might expect to discover evidence of inappropriate replacement, for
instance, in relation to ‫ אלהים‬as a plural form or in bound structures in
which the referent is a foreign divinity. That is precisely what one discov-
ers in the rest of Deut 32.
The next occurrence of ‫ צוּר‬in this specialized sense is in v. 15, which
provides insights into the translator’s method and his Vorlage. The Greek
makes explicit reference to Jacob in parallel with ֨‫ יְ ֻשׁרוּן‬Jeshurun. The for-
mer is transliterated as Ιακωβ, but the latter, ֨‫יְ ֻשׁרוּן‬, is interpreted as ὁ
ἠγαπημένος, “the beloved one.” This indicates minimally that the transla-
tor was not averse to rendering or interpreting metaphors, and certainly
points to a Vorlage different from MT, which here has no reference what-
soever to Jacob.
McCarthy explains the omission in MT as due to haplography,17 thus
granting the possibility that the LXX reading might be prior. Wevers by
contrast flatly states: “LXX made two lines out of the first hemistich, with
Ιακωβ as subject of line one and ὁ ἠγαπημένος, ‘the beloved’ freely translat-
ing ֨‫ יְ ֻשׁרוּן‬on the basis of its root ‫ ישׁר‬for the second line.”18 Both explana-
tions, insightful and creative as they seem, are totally subjective. What is
certain is that the Greek mentions Jacob by name; the Hebrew does not.
The texts are different and need not be reconciled.
The next lines ‫“ וַ יִּ טּ ֹ֙שׁ ֱא ֣ל ַוֹה ָע ָ֔שׂהוּ‬He abandoned Eloah who made him”:
‫“ ;וַ יְ נַ ֵ ֖בּל ֥צוּר יְ ֻשׁ ָע ֽתוֹ‬he spurned a rock that saved him” give the first clear
hint that ‫ ֥צוּר‬was likely a late insertion into an otherwise smooth line.
Here, unlike v. 4, ‫ ֥צוּר‬is anarthrous so it could theoretically point to any
rock. Literally translated, the stich would be rendered “he treated with
contempt a rock that saved him.” Whether the reviser forgot the article

17 Ibid., 144*.
18 Wevers, Notes, 518.
44 melvin k. h. peters

or simply presumed that subsequent readers would understand and sup-

ply it in context cannot of course be determined. But it seems significant
that of the 36 instances, only the first, 32:4, carries the definite article. This
might be a clue that the revision indeed began in Deut 32 and proceeded
Additionally, ‫ צוּר‬seems to disrupt the metrical flow of the line. The
hemistich ‫ וַ יִּ טּ ֹ֙שׁ ֱא ֣ל ַוֹה‬stands in parallel with ‫וַ יְ נַ ֵ ֖בּל אלהים‬, which would
be faithfully reflected if one reconstructs the Vorlage of the Septuagint as
it stands. The LXX reading here is again θεός, the standard gloss for ‫אלהים‬.
The translator does not distinguish between ‫ הצוּר‬and ‫צוּר‬, potential evi-
dence that the difference was not in his text.
In addition to its lack of reference to Jacob, the Hebrew of MT in v. 15
indicates its fractured character in one final way—it fluctuates between
second and third person pronouns: “You grew fat . . . He forsook the God
who made him . . .” The more expansive text reflected in the LXX is con-
sistent, beginning with the explicit naming of Jacob and appropriately
utilizing third person singular pronouns throughout. Indeed, without the
Septuagint, a reader might not have been able to determine the meaning
of the enigmatic ֨‫ יְ ֻשׁרוּן‬or for that matter the much-discussed ‫ ֥צוּר‬.
Verse 18 begins as noted with an anarthrous ‫ ֥צוּר‬. More revealing for
our purposes however are the next two instances, vv. 30 and 31, where a
confusion of ‫אלהי‬, the construct/bound form of the divine name, with
‫אלהים‬, the absolute/free form, seems to have taken place in the received
tradition. Verse 30 further shows that the LXX translator was not follow-
ing the MT tradition. Here is the context: “How could one have routed a
thousand/or two put ten thousand to flight/unless their Rock (‫ צורם‬MT),
God (θεός LXX) had sold them/the Lord had given them up?”
We begin with the obvious. If the LXX translator had ‫ צוּרם‬before
him and recognized ‫ ֥צוּר‬as a cipher for ‫אלהים‬, it would have been quite
straightforward to add αὐτῶν to ὁ θεός.19 But clearly that was not the case.
One of three things therefore seems to have occurred in the received
(MT) tradition. Either the Hebrew original from which modifications to
the divine name were made had ‫אלהיהם‬, or the final ‫ ם‬of an original
‫—אלהים‬the text reflected in the LXX—was read by the reviser as a third
person plural suffix with ‫ אלהי‬being glossed as ‫ ֥צוּר‬. Even more intriguing
and likely is the third prospect: that the original Hebrew, from which both
MT and LXX derived, contained the archaic form of the third plural ­suffix,

19 See comments on Ps 78:5 in n. 21 below.

revisiting the rock 45

‫אלהימו‬, as in v. 37. The LXX Vorlage would then be a variant upon, or mis-
reading of, this archaic form. Another way of framing the situation in the
parent text of the LXX would be to suggest haplography of (or parablepsis
on) the two ‫ה‬s in an original ‫אלהיהם‬, or the apocopation of a final waw
in an original ‫אלהימו‬.
Whatever the case, ‫ צוּרם‬creates ambiguity here. Their “rock,” ‫צוּרם‬, if
accepted as a reference to divinity, could now refer either to the god of
Israel through whose power the one or two persons were able to put thou-
sands to flight, or to the god of the myriads being put to flight, who for
whatever reason became angry and sold them to the one or two Israelite
pursuers. No such ambiguity is reflected in the Vorlage of the LXX. There,
it is the god of Israel who makes the improbable routs possible because
he had sold the many enemies into the hands of the few Israelites. The
equivalent of ‫ צוּרם‬is ‫( אלהים‬θεός) in parallel with Yahweh (κύριος). What
can be inferred here, then, is that the substitution of ‫ צוּר‬for ‫ אלהים‬or
‫ אלהי‬was a mechanical process that did not take into account contextual
or even grammatical concerns in some cases.
This situation is not limited to Deuteronomy or the Pentateuch. Paral-
lels can be seen in the so-called Elohistic Psalter (Pss 42–83) where ‫אלהים‬,
the more generic form of the divine name, was regularly substituted for
‫יהוה‬, the specific ineffable name of Israel’s god. Deuterographs like Ps
14 and 53 illustrate this quite clearly.20 Several other instances of appar-
ently uncritical or unthinking insertion of ‫ אלהים‬in the Elohistic Psalter
have been preserved, in MT. For instance, the ungrammatical expression
‫ יהוה אלהים צבאת‬in Ps 59:6; 80:5, 8, 15, 20 as well as 84:9, is probably
best explained as a substitution of ‫ אלהים‬for ‫אלהי‬. By 89:9, the normal
‫ יהוה אלהי צבאת‬resurfaces. Again, the presence of ‫ אלהים‬in bound
expressions like Elohim Elohai, Elohim Eloheka, Elohim Elohenu, Elohim
Elohe Israel, Elohim Elohe Teshuahti and so on, in the Elohistic Psalter
(43:4, 44:8, 48:15, 50:7, 51:16, 67:7, 68:9), almost surely indicates a substitu-
tion for an original Yahweh in every instance.
Even outside of the Psalter we find the expression Bet Elohim, Elohe
Israel “The house of God, god of Israel” (Ezra 6:22) and Elohim Elohe Avote-
hem “God, God of their fathers” (2 Chron 34:32). In this latter passage the
LXX reflects (the earlier) “Yahweh God of their fathers.” This ­pattern of

20 Note especially that Yahweh in Ps 14:2, 4 and 7 is represented/replaced by Elohim in

the virtual deuterograph Ps 53. One could argue that these are alternate forms of the same
Psalm but the case for 53 as subsequent to 14 in light of the attitude to the nomina sacra
seems convincing given its location in the collection.
46 melvin k. h. peters

deliberate substitution of the implicit for the explicit is thus not unusual
by any means. Something analogous seems to have occurred to produce
the confusion around ‫ צוּרם‬in Deut 32:30. Another parallel might be the
-boshet/-baal substitutions in Hebrew theophoric names involving Baal. I
could not imagine anyone seriously arguing that Ishboshet was an earlier
version of the name Ishbaal.
The following verse, 31, makes it abundantly clear that the substitution
of ‫ ֥צוּר‬for ‫ אלהים‬in the received tradition was a later mechanical process,
which could not and did not affect the Hebrew behind the Septuagint. It
has led to an undesirable theological situation in the text as it now stands.
The noun ‫ ֥צוּר‬could not have been intended as a metaphor for the god
of Israel exclusively because it is used here and in v. 37 in reference to a
foreign divinity. This indicates that the substitution was primarily at the
lexical level and not ideologically or theologically driven.
As in v. 30, ‫ צוּרם‬seems to represent an original ‫אלהים‬, ‫ אלהיהם‬or
‫( אלהימו‬mis)read, interpreted or (mis)understood in a similar way as
it was there. But the reviser seemed insensitive to the fact that in this
context ‫אלהי‬/‫ אלהים‬was plural since it refers to the gods of the nations,
and so he dutifully plugged in his default substitution for ‫אלהים‬, i.e., ‫ ֥צוּר‬.
Both the unpointed Vorlage of the LXX and the Hebrew version prior to
the revision later received by the Masoretes appear to have read: ‫כי לא‬
‫אלהימו‬/‫כאלהינו אלהיהם‬. The LXX accurately reflects that; the received
tradition retains a corruption that is both ideologically and lexically more
The final verse that makes any reference to ‫ צוּר‬in this specialized way—
v. 37—confirms and brings closure to the notion that the LXX Vorlage was
different from the received text and was uninfluenced by it. Note again,
the anarthrous ‫צוּר‬. Both the LXX and the received text recognize and
transmit the first part of the verse, though the Masoretes point the verb
‫ אמר‬as a waw consecutive with the perfect without a subject, i.e., ‫וְ ָא ַ ֖מר‬
“he will say,” and the LXX reflects a waw consecutive with the imperfect
with the subject being Yahweh, ‫“ ויאמר יהוה‬The Lord said.” I see no need
to harmonize or prioritize these readings. What is more interesting is that
the LXX reflects no evidence of ‫ צוּר‬in its parent text. Rather, the ἐφ᾿ οἷς
seems to suggest the presence of a relative particle, perhaps ‫אשׁר‬, instead
of ‫ צוּר‬in the Vorlage to which the final ἐπ᾿ αὐτοῖς points back.
So, what might we conclude from the evidence in Deut 32 regarding
‫ ?צוּר‬First, it cannot be shown definitively that the Vorlage of the Septua-
gint read the equivalent of MT in any place in this chapter. Indeed the
evidence seems just as, or even more, compelling that all the references
revisiting the rock 47

to θεός indicate that it is the normal representation of ‫אלהים‬. Second,

the Vorlage of the LXX demonstrates in some instances a more context
sensitive understanding of the text. Rather, it is the reading of the MT that
displays randomness, ambiguity and inelegance.
How would these findings hold in other sections of the Hebrew Bible
where ‫ צוּר‬seems to have been used metaphorically in MT? I am well
aware that individual translators might not have consulted each other,
and that patterns in one book may not be repeated in other books. But
it could be instructive to observe how the expression is treated in other
parts of the Septuagint. Such a survey might answer questions about the
Vorlage of each book and the habits of different translators. The books of
Samuel (1 & 2 Kingdoms) contain a few instances that might cast light on
the subject at hand. To those we turn now briefly.
In 1 Sam 2:2, the Septuagint and MT are not aligned completely. Only
the first clause, ‫יהו֖ה‬ ָ ‫ין־ק ֥דוֹשׁ ַכּ‬
ָ ‫“ ֵא‬there is no one holy like the Lord” is
shared; the LXX Vorlage also reflects an initial ‫ ִ ֣כּי‬. The second clause in
Hebrew, ‫“ ִ ֣כּי ֵ ֣אין ִבּ ְל ֶ ֑תָּך‬Truly there is no one beside you” finds no exact
parallel in Greek. If there is one, it is the final Greek clause, οὐκ ἔστιν
ἅγιος πλὴν σοῦ “There is no one holy besides you,” but that is a sim-
ple restatement of the first clause. The final clause that includes ‫צוּר‬,
ֽ ֵ ‫“ וְ ֵ ֥אן ֖צוּר ֵכּ‬there is no rock like our God” is not represented. Greek
has the equivalent of “there is none righteous like our God” as its second
clause. Some might wish to argue, and have suggested, that ‫ צוּר‬is repre-
sented by δίκαιος but that would be a stretch. What this example shows is
that the Hebrew and Greek of the verse, though similar in theme, are not
identical, and further that ‫צוּר‬, if it is accounted for at all, is not glossed
by ‫אלהים‬.
Five passages in 2 Sam (22:3, 22:32, 22:47 twice, 23:3) indicate that the
LXX Vorlage did contain ‫ צוּר‬which, in all but one instance, is glossed
with Greek φύλαξ “keeper/guard” in keeping with the verbal sense of
‫צוּר‬, “to enclose/besiege.” The variant case is 22:32 where ‫ צוּר‬is repre-
sented by κτίστης “creator,” obviously being read as related to the root ‫יצר‬
“to form or create.” But nowhere in the books of Samuel is there evidence
for ‫ אלהים‬as a translation of ‫צוּר‬. This further supports our conclusions
about the Vorlage in Deut 32 being indeed ‫אלהים‬.
The next cases are within the book of Isaiah. The first example (17:10)
is important because it is the only instance in which ‫ צוּר‬in MT corre-
sponds with ‫ יהוה‬in the Vorlage of the LXX. This confirms expectations
mentioned earlier, namely, that ‫ צוּר‬might replace Yahweh on occasion if
substitutions were ideologically driven. It also unfortunately ­disconfirms
48 melvin k. h. peters

Wevers’ unequivocal statement that the LXX substitutes θεός for ‫צוּר‬
throughout. But more important it shows that the LXX remains faithful
to its parent text, which obviously had Yahweh here and was thus not a
translation of ‫ צוּר‬in this case.
The end of Isa 26:4 ‫עוֹל ִ ֽמים‬
ָ ‫הוה ֖צוּר‬ ֔ ָ ְ‫ ִ ֚כּי ְבּ ָי֣הּ י‬is irrelevant to our discus-
sion because the Greek reflects a different Vorlage and makes no repre-
sentation of ‫צוּר‬. The example in 30:29, ‫ אל צוּר ישׂראל‬takes us back to
the pattern in Deuteronomy where ‫ צוּר‬apparently was not in the Vor-
lage of the Greek but rather ‫אלהים‬. The Greek, πρὸς τὸν θεὸν τοῦ Ισραηλ
seems to represent ‫אל אלהי ישׂראל‬. This would also confirm the earlier
observation that ‫ צוּר‬occasionally replaced the bound form ‫ אלהי‬of the
divine name. In 44:8 it would appear that ‫ צוּר‬was not in the parent text of
the Greek. Its presence in the received text suggests the independence of
the textual traditions.
Hab 1:12, like Isa 17:10, is a key passage in relation to this discussion.
Clearly, ‫ צוּר‬was in the received text and also in the Septuagint Vorlage
preceded in both instances by a ‫ ו‬i.e., ‫וצוּר‬. The LXX translator read his
text as a verb coming from ‫יצר‬, “to form or fashion,” and it appears that
verb was suffixed with the first person singular. Only the LXX recognizes
the waw because for ‫ וצוּר‬it reads: καὶ ἔπλασέν με = “and he has fashioned
me.” Its rendering of the rest of that verse ‫הוֹכ ַיח יְ ַס ְדתּוֹ‬ ֥ ִ ‫“ ְל‬to examine his
chastening” seems an accurate account of what is in the received text.
Modern translators of MT routinely ignore the waw, make ‫ צוּר‬a voca-
tive, and read the third singular suffixes as third plurals, something like:
“O Rock, you have made ‘them’ a cause for punishment/ complaint.”
So it is quite clear when ‫ צוּר‬was in the LXX parent text and how transla-
tors treated it. It is also clear when it was not. We only need to respect and
accept the information before us and not question the motives, actions or
integrity of the messenger.
We have observed thus far instances where ‫צוּר‬: (a) is translated rou-
tinely as πέτρα, (b) it is translated as a nominal relating to the verbal form,
hence φύλαξ “stronghold, keeper” etc., and (c) it is related to the verb ‫יצר‬
“to form/fashion/create.” There is thus no indisputable evidence that ‫צוּר‬
was ever translated as θεός or κύριος in the Septuagint. When those words
appear, they should be taken in their usual sense, namely to indicate the
presence of ‫ אלהים‬or ‫ יהוה‬in the parent text.
The remaining occurrences of ‫ צוּר‬as metaphor in MT occur in the Psal-
ter, and to these we now turn finally. The situation in the LXX Psalter
comports perfectly with our observations thus far. Two distinct patterns
emerge. First, ‫ צוּר‬is unquestionably present in the parent text of the LXX
revisiting the rock 49

in one third of the total cases, and in each case it is translated as related
to help or support: i.e., βοηθός—Ps 18:3, 19:15, 49:15, 94:22, and in 89:27,
ἀντιλήμπτωρ “supporter.” In 78:35, ‫ צוּרם‬is clearly present in the Vorlage,
and rendered as βοηθὸς αὐτῶν “their helper.” This supports our judgment
that the collocation ‫ צורם‬was not in the Vorlage of Deut 32: 30 and 31, for
here we see what its translation would look like.21
On the other hand, ‫ אלהים‬not ‫ צוּר‬appears to have been in the par-
ent text in all remaining instances—18:32, 18:47, 28:1, 31:3, 62:3, 62:7, 62:8,
92:16; 95:1; 144:1—in the Psalter. In 71:3 and 73:26 the LXX Vorlage probably
read ‫ לאלהי‬and ‫ אלהי‬respectively as the translations εἰς θεόν and ὁ θεός,
Those who would wish to argue otherwise bear the burden to show
credible reasons for the alternate translations of ‫ צוּר‬mentioned above.
Furthermore, they must also answer this basic question: “Why would a
translator follow his text so closely in the main and then in certain places
deviate dramatically from the text before him?” If his deviations are unre-
liable so also should be his agreements.
What then does all this mean? The most significant implication of these
findings is (something I have been emphasizing increasingly in my recent
work, namely,) that the well-established tradition of using the late text of
the Leningrad Codex (MT) as the starting point, the lemma, for textual
analysis is passé and needs to be rethought. Despite its wide acceptance
in religious circles and its rich history, it is just a manuscript like any other
and a rather late one at that. A corollary of this is that the suspicions
surrounding the Septuagint translators who worked centuries before the
Leningrad Codex was written, especially of their methods and motives,
must be lifted if we wish to understand the meaning of the evidence they
provide. Constant efforts to harmonize or explain away obvious differ-
ences in extant forms of the biblical text do violence to each of the forms
and do us no permanent good. The LXX is often the bearer of “inconve-
nient truths” but killing or doubting the messenger only serves to confirm
preexisting beliefs and does not assist in determining the true state of
affairs in the ancient world.
This is not just my view. Three other corroborating voices from dif-
ferent generations, among many other possible candidates, must suffice.
Already in 1974, Harry M. Orlinsky in his Grinfield lectures on the Septua-
gint at Oxford said the following:

21 Cf. n. 19 earlier.
50 melvin k. h. peters

I would put it more bluntly: the modern scholar has no right to assume a
priori that the Septuagint translator manipulated his Hebrew text; in point
of fact, it is the opposite assumption that would be valid viz., that unless
and until a case can be made for regarding the Septuagint translation of
any Book as an unreliable witness to the Hebrew text, the translator must
be regarded as a serious, forthright reproducer of that text.22
And later, speaking of variance between the received text and the Septua-
gint, in Deut 31 no less, Orlinsky says:
What is important is that scholars have not yet sufficiently recognized that
the translator did not edit or emend a text that contained our preserved
reading, for he was not an editor-harmonizer; he was a translator, and he
translated what his Hebrew text offered him. A pity indeed that so many
modern scholars have created anthropomorphic-theological-contextual-
textual problems for the Septuagint translators that they never had, and have
then proceeded to create “solutions” of the same ephemeral substance—
usually at the “expense” of the integrity and scholarship of the translators.23
Thirteen years later Anneli Aejmelaeus, in her frequently quoted 1987
ZAW article “What Can We Know about the Hebrew Vorlage of the Sep-
tuagint?” concludes:
In principle, the MT plus corresponding to the Septuagintal minus and the
Septuagintal plus ought to be treated on equal terms.
All in all, the scholar who wishes to attribute deliberate changes, harmo-
nizations, completion of details and new accents to the translator is under
obligation to prove his thesis with weighty arguments and also to show why
the divergences cannot have originated with the Vorlage. That the translator
may have manipulated his original does not mean that he necessarily did
so. All that is known of the translation techniques of the Septuagint points
firmly enough in the opposite direction.24
Finally, T. M. Law, a recent Oxford graduate, published an impressive and
well-argued defense of an independent pre-MT Vorlage for 3 Kingdoms,
Septuagint scholars must not go backwards to a pre-Qumran view of the
biblical text and its associated explanations of the Greek textual history, but
must instead approach the state of the question in the light of what is known
now about textual plurality and in the light of the most current research on

22 H. M. Orlinsky, “The Septuagint as Holy Writ and the Philosophy of the Translators,”
HUCA 46 (1975): 110.
23 Ibid., 112.
24 A. Aejmelaeus, “What Can We know about the Hebrew Vorlage of the Septuagint?”
reprinted in On the Trail of the Septuagint Translators (Kampen: Kok Pharos, 1993), 92.
revisiting the rock 51

the OG . . . Hebrew Bible scholars, however, must go backwards; for it was

in the 19th century that Hebrew Bible commentators on Kings . . . evaluated
3–4 Reigns as a different Hebrew Vorlage to MT.25
Instances abound where retroversion of the Hebrew behind the Septu-
agint is difficult and challenging to the modern scholar. We do well to
confine our intellectual energies to resolving those puzzles rather than
creating problems where they do not exist and need not be. There is noth-
ing problematic about the meaning of the divine names in either Hebrew
or Greek and we are well aware of the abundant evidence throughout
the LXX corpus that the Vorlage of the Septuagint, though similar in the
main with the received text, was often quite radically at variance with it.
I suggest that in this matter we have yet another such example. I hope to
have demonstrated in this brief survey, of a well-established but in my
view totally invalid notion, the risks of uncritically trusting late printed
Hebrew over early, easily retrovertable translation Greek.

25 T. M. Law, “How Not to Use 3 Reigns: A Plea to Scholars of the Books of Kings,”
VT 61 (2011): 294.
Judges 3:12–30: An Analysis of the Greek Rendering of
Hebrew Wordplay

Hans Ausloos

1. Wordplay and the translation technique of the Septuagint

If one would like to understand the full meaning of a biblical text, it is

not sufficient to have a good knowledge of Hebrew grammar and syntax.
From the first chapters of the Old Testament on, it becomes clear that
wordplay—whatever this may be—seems to be constitutive for bibli-
cal literature.1 If the reader does not succeed in tracing present forms of
wordplay within a pericope, its full meaning undoubtedly gets lost.
The present contribution does not, by any means, aim to clarify the
problem of Hebrew wordplay on a theoretical level.2 In this paper, I would
like to study if and how the Septuagint (LXX) translator has dealt with the
different types of wordplay with which he was confronted in his Hebrew
Vorlage. Evidently, it is one of the most difficult problems for a translator
to adequately render wordplay from a source language into a target lan-
guage, as becomes evident from the first pages of the Old Testament.
One of the first examples of wordplay in the Hebrew Bible can be
found in the creation narrative in Gen 2–3, where a clear link is made
between ‫ ָה ָא ָדם‬who is formed ‫ ִמן ָה ֲא ָד ָמה‬. Making use of ­similar-sounding
words, the Hebrew author clearly indicates that, in his view, there is a
close relationship between human beings and the earth (this does not
automatically imply that there is also an etymological link between

1 Despite the importance of wordplay for the understanding of the Hebrew Bible, one
had to wait until the end of the 19th century for the first systematic studies on this topic:
I. M. Casanowicz, Paronomasia in the Old Testament (Ph.D. diss., John Hopkins Univer-
sity, 1894); see also idem, “Paronomasia in the Old Testament,” JBL 12 (1893): 105–167.
Also in more recent times, systematic approaches to the field remain rather rare. See in
this respect, T. Cherry, Paronomasia and Proper Names in the Old Testament. Rhetorical
Function and Literary Effect (Ph.D. diss., Southern Baptist Theological Seminary, 1988);
S. B. Noegel, ed., Puns and Pundits. Wordplay in the Hebrew Bible and Ancient Near Eastern
Literature (Bethesda: CDL Press, 2000).
2 For some remarks with respect to the use of wordplay in biblical literature and an
attempt of classification of different categories, see V. Kabergs and H. Ausloos, “Parono-
masia or Wordplay? A Babel-Like Confusion. Towards A Definition of Hebrew Wordplay,”
Bib 93 (2012): 1–20.
54 hans ausloos

the two lexemes). In cases where one can hardly doubt that the Hebrew
author intended to play with Hebrew words, as is the case in Gen 2, a good
translator should at least try to render his Vorlage as adequately as pos-
sible. Taking for granted that the translator actually did notice the Hebrew
wordplay, he has three possible options. First, he could add a footnote
in order to clarify the wordplay that is present in the source language.
Doing so, he would indicate that he has noticed the Hebrew wordplay,
but at the same time he would admit that he was not able to find an
adequate translation equivalent, and therefore failed in his translation.
Second, the translator could transliterate those Hebrew words that are
constitutive for the wordplay. As such, he equally clearly indicates that
he has noticed the Hebrew wordplay and makes clear to his readers that
there is a link between the two words. At the same time, however, he
does not succeed as a translator in this case either, since he is not giving
a real translation. Third, he could translate the Hebrew words and search
for good alternatives in the target language. However, in practice, it is an
almost impossible task to find two or more terms in the target language
that do not only correspond in meaning, but have a similar connotation
or sound as well. Taking a look at Gen 2:7, one can conclude that the LXX
translator in fact translated ‫ ָה ָא ָדם‬and ‫ ִמן ָה ֲא ָד ָמה‬as καὶ ἔπλασεν ὁ θεὸς
τὸν ἄνθρωπον χοῦν ἀπὸ τῆς γῆς (NETS: “And God formed man, dust from
the earth”).3 In doing so, however, the link between both terms has been
completely lost for the reader of the LXX. Nevertheless, one should be
wary of drawing generalizing conclusions too easily. For example, within
the same context of the Eden narrative (Gen 3:20), the LXX translator suc-
ceeds very well in rendering the Hebrew wordplay into idiomatic Greek.
“Adam” calls his wife Eve (‫ ) ַחוָּ ה‬because she is “mother of all living ones”
(‫) ִכּי ִהיא ָהיְ ָתה ֵאם ׇכּל ָחי‬. The LXX translates it as καὶ ἐκάλεσεν Αδὰμ τὸ
ὄνομα τῆς γυναικὸς αὐτου Ζωη ὅτι αὕτη μήτηρ πάντων τῶν ζώντων (NETS:
“And Adam called the name of his wife Life, because she is the mother
of all the living”).4 In this case, the LXX translator did effectively translate
the proper name ‫ ַחוָּ ה‬with the Greek Ζωή and thus, he has chosen not to
transliterate it. Since ζωή in Greek means “life”, the implicit meaning of
the proper name ‫ ַחוָּ ה‬is preserved in Greek. Moreover, the sound similar-
ity with τῶν ζώντων is present as well.

3 R. J. V. Hiebert, “Genesis,” in A New English Translation of the Septuagint and the
Other Greek Translations Traditionally Included under That Title (eds. A. Pietersma and
B. G. Wright; New York / Oxford: Oxford University Press, 2007), 7.
4 Ibid., 8.
judges 3:12–30 55

Keeping these three major possibilities of rendering Hebrew wordplay

in mind, I am convinced that the study of the specific way in which Greek
translators render the Hebrew wordplay can be helpful in the attempt
to characterise the so-called translation technique of the LXX translators
as adequately as possible. The Greek rendering of Hebrew wordplay—
aetiologies can also be considered as such5—can therefore function as
a so-called content-related criterion6 in addition to the analysis of the
rendering of so-called Hebrew hapax legomena7 and Hebrew jargon-
defined vocabulary.8 These content-related criteria have to be considered
as complementing the “grammatically qualitative” research—as found in
the analyses of the Helsinki school, for example—from the specific angle
of analysing the procedure of the LXX translator when he was confronted
with difficult, highly specific and identifiable questions related to the con-
tent of his Vorlage.

5 H. Ausloos, “LXX’s Rendering of Hebrew Proper Names and the Characterization of
the Translation Technique of the Book of Judges,” in Essays on Septuagint, Hebrew Bible,
and Dead Sea Scrolls (FS R. Sollamo; JSJSup 126; eds. A. Voitila and J. Jokiranta; Leiden:
Brill, 2008), 53–71; H. Ausloos, “The Septuagint’s Rendering of Hebrew Toponyms as an
Indication for the Translation Technique of the Book of Numbers,” in Textual Criticism and
Dead Sea Scrolls Studies. Florilegium Complutense. (FS J. Trebolle Barrera; JSJSup 158; eds.
A. Piquer Otero and P. A. Torijano Morales; Leiden / Boston: Brill; 2012), 35–50.
6 A first methodological presentation of content-related criteria, named as such, has
been systematically presented and illustrated at the LXX.D-conference in Wuppertal, 2008.
See in this respect, H. Ausloos and B. Lemmelijn, “Content-Related Criteria in Characterising
the LXX Translation Technique,” in Die Septuaginta: Texte, Theologien und Einflüsse (WUNT
252; eds. W. Kraus, M. Karrer and M. Meiser; Tübingen: Mohr Siebeck, 2010), 357–376.
A refinement of the initial approach, and more particularly with respect to the Greek
rendering of Hebrew wordplay as a content-related criterion, has been presented at the
LXX.D-conference in Wuppertal, 2010 and has been published as H. Ausloos, B. Lemmelijn
and V. Kabergs, “The Study of Aetiological Wordplay as a Content-Related Criterion in the
Characterisation of LXX Translation Technique,” in Die Septuaginta: Entstehung, Sprache,
Geschichte (WUNT 286; eds. S. Kreuzer, M. Meiser and M. Sigismund; Tübingen: Mohr
Siebeck, 2012), 273–294.
7 H. Ausloos and B. Lemmelijn, “Rendering Love. Hapax Legomena and the Characteri-
sation of the Translation Technique of Song of Songs,” in Translating a Translation. The
LXX and its Modern Translations in the Context of Early Judaism (BETL 213; eds. H. Ausloos
et al.; Louvain / Paris / Dudley: Peeters, 2008), 43–61; H. Ausloos, “The Septuagint’s Render-
ing of Hebrew Hapax Legomena and the Characterization of its ‘Translation Technique’.
The Case of Exodus,” Acta Patristica et Byzantina 20 (2009): 360‑376. See also E. Verbeke,
“The Use of Hebrew Hapax Legomena in Septuagint Studies. Preliminary Remarks on
Methodology,” in Florilegium Lovaniense. Studies in Septuagint and Textual Criticism (FS
F. García Martínez; BETL 224; eds. H. Ausloos, B. Lemmelijn and M. Vervenne; Louvain /
Paris / Dudley: Peeters, 2008), 507–521.
8 B. Lemmelijn, “Flora in Cantico Canticorum. Towards a More Precise Characterisation
of Translation Technique in the LXX of Song of Songs,” in Scripture in Transition. Essays on
Septuagint, Hebrew Bible and Dead Sea Scrolls (FS R. Sollamo; JSJSup 126; eds. A. Voitila and
J. Jokiranta; Leiden, Brill, 2008), 27–51.
56 hans ausloos

The analysis of the Greek rendering of Hebrew wordplay therefore

functions in the context of a kind of artificial laboratory situation that
we create in order to observe the translator’s activity in a particular situ-
ation in which he has been forced to make a specific decision. Of course,
as is the case with hapax legomena (we cannot be entirely certain that
what we consider a hapax was equally a hapax for the translator), in some
cases there can be discussion on whether or not the Hebrew author has
intentionally played with words. As a result, it is theoretically possible
that what we consider as wordplay has not been perceived as such or even
remarked by the translator.9 Nevertheless, after receiving the results of
such a concrete, content-related analysis, the characterisation of a specific
LXX book can be evaluated more precisely. First, by comparing the differ-
ent LXX books ad intra on the basis of one and the same content-related
criterion (e.g. wordplay). Second, ad extra in a double way: comparing
the results on the basis of a particular content-related criterion to the
results of the study of another content-related criterion in different books
(e.g. wordplay compared to hapax legomena). Finally, by comparing the
content-related characterisation to the more traditional quantitative and
grammatically qualitative characterisation of the translation technique of
the respective LXX books.
Against this concise theoretical background, this paper will present a
concrete, elaborated example of an analysis of this kind in studying the
Greek rendering of Hebrew wordplay in Judg 3:12‑30. It will thereby demon-
strate the extent to which the combination and/or comparison of several
distinctive content-related criteria—in particular the analysis of the way
in which Hebrew hapax legomena are rendered and the way in which the
Greek rendering reflects the Hebrew wordplay—can give relevant results
in relation to a more sound characterisation of the LXX translation.

2. The Ehud and Eglon narrative in Judg 3:12–30

The stereotypical narrative pattern of the book of Judges is sufficiently

known. In reaction to Israel’s apostasy, God makes use of Israel’s enemies

9 Cf., however, the valuable remark of J. J. M. Roberts, “Double Entendre in First Isaiah,”
CBQ 54 (1992): 39‑48, esp. 40, with regard to the phenomenon of double entendre: “While
one must remain aware of the danger of overreading, however, it is far more likely that
our lack of familiarity with the wider connotations of classical Hebrew words and phrases
will result in underreading, of missing intentional double entendres”.
judges 3:12–30 57

in order to convert the Israelites, and, through the work of a “judge”,

brings them back on the right path. In the case of Judg 3:12‑30, it is the
corpulent Eglon, the Moabite king, to whose power God delivers the mis-
behaving Israelites. Together with the Ammonites and Amalekites, Eglon
fights Israel and conquers the “city of Palms”. After eighteen years of siege,
the Israelites beg God for help, after which the latter sends them a sav-
iour, Ehud the “Benjaminite”. Ehud and Eglon meet each other for the first
time at the moment that Ehud offers Eglon tribute from the Israelites. In
preparation of his visit, Ehud makes a “double-edged” sword one gomed
long, which he fixes at his right hip under his clothes. When tribute is
paid, Ehud gives permission to the carriers of the tribute to return home.
However, upon reaching the “sculptured stones” of Gilgal, Ehud returns to
Eglon’s palace and tells him that he still has a secret. Eglon immediately
orders all present persons to keep silent, which seems to be understood
as an order to leave the room.10 Ehud enters the throne room and, being
alone with Eglon, tells him that he has a divine message. In a gesture of
respect for the word of God which Ehud will speak out, Eglon rises from
his throne. With his left hand, Ehud takes the sword from his right thigh
and thrusts it in Eglon’s belly, so deep that even the hilt goes in after the
blade. Because Ehud does not withdraw the sword, the fat of Eglon’s belly
closes over the blade.
After the murder, v. 23 describes the way in which Ehud leaves the locum
delicti. He locks the doors of the roof chamber. Thereafter, i.e. after Ehud
has left, the servants come back and observe that the doors of the “roof
chamber” are locked. They speak cryptic words to each other: “Surely he
covers his feet in the cool chamber” (NRSV: “He is only relieving himself in
the closet of the cool chamber”). They wait a while and finally unlock the
doors to find their lord lying dead on the floor. In the meantime, Ehud has
been able to escape “beyond the sculptured stones” to Seira (as opposed
to v. 19, Gilgal is not mentioned in v. 26). When he reaches Ephraim, he
sounds the trumpet, gathers the Israelites and marches with them against

10 Judg 3:19 LXX (A) renders the Hebrew construction ‫אמר ָהס‬ ֶ ֺ ‫ וַ יּ‬as καὶ εἶπεν Εγλωμ
πᾶσιν ἐκ μέσου (NETS: “And Eglon said to everyone, ‘Away!’ ”). With respect to the double
text of Judges, see inter alia, W. R. Bodine, The Greek Text of Judges. Recensional Develop-
ments (HSM 23; Chico: Scholars Press, 1980); O. Munnich, “Le texte de la Septante,” in La
Bible grecque des Septante. Du judaïsme hellénistique au christianisme ancien (Initiations
au christianisme ancien; eds. M. Harl, G. Dorival and O. Munnich; Paris: Cerf, 1994), 175;
N. Fernández Marcos, The Septuagint in Context. Introduction to the Greek Versions of the
Bible (Boston / Leiden: Brill, 2001), 94–95.
58 hans ausloos

Moab. The pericope concludes with the mentioning of the killing of ten
thousand strong and fat Moabites, after which rest returns to Israel.
Judg 3:12‑30 represents a real pearl of ancient Hebrew narrative art that
has been passed down to us. Nevertheless, it is a difficult text, not least
because of several hapax legomena in it. In fact, there are three: the length
of Ehud’s sword as one “gomed” (‫ )גּ ֶֺמד‬long (v. 16), the noun ‫ַה ַפּ ְר ְשׁד ֺנָ ה‬
(v. 22) and the word ‫( ַה ִמּ ְס ְדּרוֹנָ ה‬v. 23). Having dealt with these three
hapax legomena in another contribution,11 the present paper will concen-
trate on the wordplay in this passage, where the usage of Hebrew lan-
guage is very much well-thought-out and often ambiguous. The author
constantly plays with words and their meanings, and it seems as if no
single element is accidental, although it is not always easy to define the
exact implications thereof.

3. Ehud as left-hander

The presentation of the Israelite protagonist of the narrative, Ehud, con-

fronts the reader of the Hebrew text with the first play on words. He
is introduced in Judg 3:15. When the Israelites cried unto God, YHWH
“raised up for them a deliverer, Ehud son of Gera, the Benjaminite, a left-
handed man” (NRSV) (‫מֹוׁש ַיע ֶאת ֵאהּוד ֶּבן ּגֵ ָרא ֶּבן ַהיְ ִמינִ י‬
ִ ‫וַ ּיָ ֶקם יְ הוָ ה ָל ֶהם‬
‫) ִאיׁש ִא ֵּטר יַ ד יְ ִמינֹו‬. Ehud is presented as son of Gera (‫) ֶּבן ּגֵ ָרא‬, who,
in turn, is said to be a “son of the Jemini” (‫) ֶּבן ַהיְ ִמינִ י‬.12 There can be
no doubt that the author is playing with words in his double use of the
forms ‫יְ ִמינֹו‬/‫ ַהיְ ִמינִ י‬. The text is not unambiguous. It does not become
clear whether the author of the Hebrew text links Ehud with the tribe
of Benjamin, or whether he would like to characterise Ehud as “son of
my (God’s?) right side”. Usually, when making reference to the person or
the tribe of Benjamin, the biblical authors make use of the proper name
‫ ִבנְ יָ ִמין‬.13 However, the fact that the author links Ehud with Gera seems

11 H. Ausloos and B. Lemmelijn, “Characterizing the LXX Translation of Judges on the
Basis of Content-Related Criteria. The Greek Rendering of Hebrew Absolute Hapax Legom-
ena in Judg 3,12–30,” in After Qumran: Old and Modern Editions of the Biblical Texts—The
Historical Books (BETL 246; eds. H. Ausloos, B. Lemmelijn and J. Trebolle Barrera; Louvain /
Paris / Dudley: Peeters; 2012), 171–192.
12 The combination ‫ ֶּבן ַהיְ ִמינִ י‬also occurs in 2 Sam 16:11; 19:17; 1 Kgs 2:8.
13 Cf., e.g., Gen 35:18; see also Judg 5:14; 10:9; 19:14; 20 (passim); 21 (passim). When Ben-
jamin is born after hard labour, his mother Rachel gives him the name Ben-Oni, which
means “son of my sorrow”. However, his father Jacob names his youngest son Ben-Jamin,
“son of (the) right hand”, which probably can be interpreted as “son of happiness”—cf.
judges 3:12–30 59

to be an argument in favour of Ehud’s identification as a Benjaminite,

because in Gen 46:21, Gera is considered to be a son of Benjamin. In
any case, the lexeme ‫ יָ ִמין‬plays a crucial role in the narrative. The same
v. 15, which presents Ehud as ‫ ֶּבן ַהיְ ִמינִ י‬, mentions that his right hand
(‫ )יְ ִמינֹו יַ ד‬is ‫ ִא ֵּטר‬. The precise meaning of this expression (‫) ִא ֵּטר יַ ד יְ ִמינֹו‬
is not clear. According to some exegetes, the author of Judg 3:15 aimed to
indicate that Ehud was handicapped at his right hand, which would imply
that he was left-handed.14 This is further used as the explanation for the
fact that Ehud fixes the sword at his right hip (v. 16). However, Judg 20:16,
the only other verse were the word ‫ ִא ֵּטר‬occurs, can be used against this
interpretation of a handicapped Ehud. In this verse, the same terminology
is used in characterising seven hundred soldiers of the Benjaminites—
here the term ‫ ְבנֵ י ִבנְ יָ ִמין‬is used ( Judg 20:15)—who are “picked men who
were left-handed; every one could sling a stone at a hair, and not miss”
(NRSV). In this respect, it is not plausible that all seven hundred warriors
are left-handed because of a handicap at their right hand. This leads schol-
ars to the supposition that, according to the biblical tradition, the Benja-
minites would have been specifically trained to fight with their left hands.15
So the adjective ‫ ִא ֵּטר‬has been related to a practice in which the right
hand was systematically bound in order to train soldiers in using their
left hand. In combat, in which attack and defence as well as the construc-
tion of armoury was mostly based on the use of the right hand, it could
be an advantage to be left-handed. It could have an effect of surprise to
the adversary, who was used to fighting against right-handed opponents.
In any case, when the author had the intention of saying that Ehud was
left-handed, it is remarkable that he uses the terminology ‫ִא ֵּטר יַ ד יְ ִמינֹו‬
and not ‫“( יַ ד ְשׂמֺאלוֹ‬his left hand”), as he does in Judg 3:21. Therefore, it
can be concluded that the term “left” is a keyword within the Hebrew
Ehud pericope. It is used again in v. 16 (Ehud fixes the dagger upon his
right thigh—‫ ) ַעל יֶ ֶרְך יְ ִמינֹו‬and finally in v. 21, where it is told that Ehud

C. Westermann, Genesis 12–36 (BKAT 1/2/2; Neukirchen-Vluyn: Neukirchener Verlag, 1981),

675. See also Y. Amit, The Book of Judges. The Art of Editing (Biblical Interpretation Series
38; Leiden / Boston / Köln: Brill, 1999), 171–198, esp. 179–180; U. Hübner, “Mord auf dem
Abort? Überlegungen zu Humor, Gewaltdarstellung und Realienkunde in Ri 3:12–30,”
BN 40 (1987): 130–140, esp. 133 n. 14.
14 Cf. NRSV: “Ehud son of Gera, the Benjaminite, a left-handed man”. Cf. F. Dexinger,
“Ein Plädoyer für die Linkshändler im Richterbuch,” ZAW 89 (1977): 268–269; E. Jugel and
H.‑D. Neef, “Ehud als Linkshänder. Exegetische und medizinische Anmerkungen zu Ri
3,15,” BN 97 (1999): 45–54, esp. 46–47.
15 See, e.g., B. Halpern, “The Assassination of Eglon—The First Locked-Room Murder
Mystery,” BRev 4/6 (1988): 32–41, esp. 35.
60 hans ausloos

takes his dagger with his left hand from his right thigh (‫) ֵמ ַעל יֶ ֶרְך יְ ִמינֹו‬
and thrusts it into Eglon’s belly.
Against this background, the question is raised as to which way the
LXX translator deals with this Leitmotiv of his Hebrew Vorlage. On the one
hand, his translation is rather disappointing. On the other hand, it will
become clear that he is a rather creative translator after all.
Let’s start with the disappointing element. The genealogical notion
‫( ֶאת ֵאהּוד ֶּבן ּגֵ ָרא ֶּבן ַהיְ ִמינִ י‬v. 15) is partially translated, partially very
literally transliterated: τὸν Αωδ υἱὸν Γηρα υἱοῦ τοῦ Ιεμενι ( Judg A and
B—NETS:16 “Aod son of Gera son of Iemeni”). The translator clearly dis-
tinguishes between the usual name of the tribe and the formula in Judg
3:15. In all instances in Judges where the usual Hebrew term ‫ ִבנְ יָ ִמין‬is used
to designate the tribe of Benjamin, the LXX reads Βενιαμιν. In Judg 3:15,
however, ‫ ֶּבן ַהיְ ִמינִ י‬is understood as a (singular) proper personal name.
The other instances in Judg 3, where it is said that Ehud uses his right or
left hand, are translated adequately. According to Judg 3:16, Ehud fastens
the dagger on his right thigh (ἐπὶ τὸν μηρὸν τὸν δεξιὸν αὐτοῦ), which is a lit-
eral translation of ‫ ַעל יֶ ֶרְך יְ ִמינֹו‬. When in v. 21 the murder is described, the
LXX likewise translates literally and correctly: with his left hand (A and
B: τὴν χεῖρα τὴν ἀριστερὰν αὐτοῦ), he takes the dagger from his right thigh
(A: ἀπὸ τοῦ μηροῦ τοῦ δεξιοῦ αὐτοῦ; B: ἐπάνωθεν τοῦ μηροῦ αὐτοῦ τοῦ
In sum, although the LXX renders the Hebrew Vorlage correctly, the
way in which the Hebrew author plays with the word pair “left” and “right”
in vv. 15, 16, 21, especially with regard to Ehud as son of Jemini, has been
completely lost for the Greek reader.
There is one final element, however, which has to be taken into con-
sideration too, namely Ehud’s characterisation as ‫ ִא ֵּטר יַ ד יְ ִמינֹו‬. Here, we
meet a translator who is looking for a good translation equivalent that
makes sense in the context of the pericope. He renders this enigmatic for-
mula with the term ἀμφοτεροδέξιον (NETS: “an ambidextrous man”). This
lexeme is composed of two common Greek words: the adjective ἀμφοτερος
(“both of two”) and the noun δέξιά (“right hand”) or the adjective δέξιος
(“on the right hand/side”). As a result, ‫ יַ ד יָ ִמין‬has an equivalent in the ele-
ment δέξιον.17 The term ἀμφοτεροδέξιον does not seem to be attested prior

16 P. E. Satterthwaite, “Judges,” in A New English Translation of the Septuagint and the
Other Greek Translations Traditionally Included under That Title (eds. A. Pietersma and
B. G. Wright; New York / Oxford: Oxford University Press, 2007), 204.
17 Compare, e.g., with Gen 48:17, where ‫ יַ ד יְ ִמינֹו‬is translated as τὴν δεξιὰν αὐτοῦ.
judges 3:12–30 61

to the LXX. Because of that, the lexicon LEH considers it a neologism,

meaning “ambidextrous”.18 Much later, the term can be found in the works
of Galenus Medicus (2nd century C.E.) and Aeristaenetus (5–6th century
C.E.).19 In the Greek papyri, the term does not seem to be ­present.20 In
any case, probably not knowing the precise meaning of the term ‫ ִא ֵּטר‬, the
translator seems to have been looking for an adequate translation equiva-
lent. Keeping very close to the Hebrew original (the element “right hand”
is conserved), he presents Ehud as able to fight with his left hand as well
as with his right hand. It seems that one can observe here a trace of the
translator’s contextual exegesis, taking care of the context of the narrative,
in which Ehud will use his left hand to kill the fat Eglon. It is with this
person that we will deal in the next part of this paper.

4. Eglon the calf

Besides the presentation of Ehud as left-hander, the Hebrew text seems

to give a specific appreciation of his adversary as well, who is called Eglon
(‫) ֶעגְ לוֹן‬. In the light of what happens in the story, this name could hardly
have been chosen accidentally. Whether the two last consonants (‫ון‬-)
function as a diminutive or not,21 the proper name ‫ ֶעגְ לוֹן‬is undoubtedly
related to the noun ‫ ֵעגֶ ל‬, which means “male calf”, “young bull”.22 The
proper name ‫ ֶעגְ לוֹן‬therefore seems to be an example of a double entendre.23
In Judg 3:12, 14, 17, 19, 20 and 24, ‫ ֶעגְ לוֹן‬is the name of the Moabite king.
However, due to its similar sound with the noun ‫ ֵעגֶ ל‬, the author seems
to give an appreciation of the king’s person.24 Even if the reader does not

18  J. Lust, E. Eynikel and K. Hauspie, Greek-English Lexicon of the Septuagint: Revised
Edition (Stuttgart: Deutsche Bibelgesellschaft, 2003), 34. T. Muraoka, A Greek-English Lexi-
con of the Septuagint (Louvain / Paris / Walpole: Peeters, 2009), 33 translates the term as
“able to use left and right hands equally well”. H. N. Rösel, “Zur Ehud-Erzählung,” ZAW 89
(1977): 270–272 interprets the Hebrew formula equally as indicating “Beidhändigkeit”.
19 H. G. Liddell and R. Scott, A Greek-English Lexicon. With a Revised Supplement
(Oxford: Clarendon Press, 1996), 95.
20 Cf. D. Hagedorn, Wörterlisten aus den Registern von Publikationen griechischer und
lateinischer dokumentarischer Papyri und Ostraka (s.l.: s.n., 132010; [cited 26 May 2011].
Online: ).
21  Cf. on this topic G. Mobley, The Empty Men. The Heroic Tradition of Ancient Israel
(ABRL; New York: Doubleday, 2005), 77–78; H.‑D. Neef, “Eglon als ‘Kälbermann’? Exeget-
ische Beobachtungen zu Jdc 3:12‑30,” VT 59 (2009): 284–294, esp. 288–289.
22 Cf. D. J. A. Clines, “‫ ֵעגֶ ל‬,” DCH 6, 248–249.
23 With respect to this terminus technicus, see Cherry, Paronomasia, 32–33.
24 Contrary to J. M. Sasson, “Ethically Cultured Interpretations. The Case of Eglon’s
Murder ( Judges 3),” in Homeland and Exile. Biblical and Ancient Near Eastern Studies
62 hans ausloos

immediately understand the wordplay—which seems rather implausible

because of the continuous hostility between Israel and Moab—it becomes
clear in the following verses. In a parenthesis in v. 17, the Moabite king
is presented as ‫ ִאיׁש ָּב ִריא ְמאֹד‬. The lexeme ‫ ָּב ִריא‬is often interpreted
as referring to Eglon as a very fat or corpulent man, not least because of
v. 22, in which it is said that Eglon’s fat closed upon the blade of the dag-
ger. However, human corpulence is normally referred to by making use of
the root ‫ ָשׁ ֵמן‬, as is the case in Judg 3:29.25 The term ‫ ָּב ִריא‬on the contrary,
is mostly applied to animals, meat or plants in order to indicate that they
are good or ripe for consumption.26 Using this term to characterise Eglon
seems therefore to have been done in order to present him as an object
that is ready to be slaughtered.27 In this respect, it is remarkable that the
author of the Ehud-pericope has used some more terms with a double
The term ‫ ִמנְ ָחה‬, that occurs four times in Judg 3:12‑30 (vv. 15, 17, 18[bis])
does not only mean “tribute” or “gift”,28 but often also has a cultic and
sacrificial meaning, as becomes clear from several instances in the book
of Leviticus and Numbers,29 where it is sometimes linked with the notion
of “fat”.30 Besides the noun ‫ ִמנְ ָחה‬, the verbal form ‫ וַ יַּ ְק ֵרב‬in Judg 3:17
(‫ ָק ַרב‬: “to present”) also seems to make a link with the offer cult (cf.
‫ וַ יַּ ְק ֵרב ֶאת ַה ִמּנְ ָחה‬in Lev 9:17: “he presented the offering”). The irony of
the Hebrew text is thus clear: while Ehud is bringing a tribute to Eglon
( Judg 3:15), it is in fact king Eglon who will be slaughtered as an offering
(to God?) by Ehud.31
It is beyond doubt that it is very difficult for a translator to transpose
these elements of double entendre or implicit polysemy, which, moreover,

(FS B. Oded; VTSup 130; eds. G. Galil, M. Geller and A. Millard; Leiden / Boston: Brill, 2009),
571–595, esp. 573.
25 Cf. also J. M. Sasson, “Ethically Cultured Interpretations,” 575–576 n. 13.
26 Cf. Gen 41:2, 4, 18, 20 (cows); 41:5, 7 (ears of grain); 1 Kgs 5:3 (oxen); Ezek 34:20
(sheep); Hab 1:16 (food), Zech 11:16 (flesh); see also as object in Ezek 34:3 (the fat animals).
In Ps 73:4, the term refers to the body of the wicked; in Dan 1:15 it refers to the flesh of
Daniel, Hananiah, Mishael, and Azariah.
27 Contrary to J. M. Sasson, “Ethically Cultured Interpretations,” 575, who considers the
term as indicating that Eglon was “imposing (. . .)—a notice that explains why he would
lack guards, a crucial element in the unfolding plot”.
28 See, e.g., Gen 32:14, 19, 21, 22; 33:10; 43:11, 15, 25, 26.
29 Cf., e.g., Lev 2:1, 4, 5, 6, 15; 6:13; Num 15:6.
30 See, e.g., Lev 2:5–6, 15; 14:10.
31 Cf. A. Wénin, “Le ‘point de vue raconté’, une catégorie utile pour étudier les récits
bibliques? L’exemple du meurtre d’Églôn par Éhud ( Jdc 3,15–26a),” ZAW 120 (2008): 14–27,
esp. 17–18.
judges 3:12–30 63

have a clear function in the narrative, to the target language. In what fol-
lows, we will have a look at the manner in which the LXX translator has
dealt with these elements of Hebrew wordplay.
First of all, the translator opts to transliterate Eglon’s name as Εγλωμ.
Doing so, the connotation of the proper name with an animal that will
be slaughtered completely disappears. However, he could have given the
king another name, such as Μόσχον, the noun which in Exod 32:4 is used
as the translation equivalent for the noun ‫ ֵעגֶ ל‬. Secondly, Eglon’s charac-
terisation as ‫ ָּב ִריא‬has been rendered in a very divergent way by the Greek
translator of Judges. In the LXX (A and B), this term is represented by the
Greek adjective ἀστεῖος as a translation equivalent: Εγλωμ ἀνὴρ ἀστεῖος
σφόδρα (NETS: “Eglom was a very handsome man”). This word is rather
rare in the LXX. As an equivalent of a Hebrew word, it occurs in Exod 2:2
(Moses’ mother saw that her baby was ‫—טוֹב‬ἰδόντες δὲ αὐτὸ ἀστεῖον) and
in Num 22:32 (the angel refutes Balaam with the words: οὐκ ἀστεία (‫ )יָ ַרט‬ἡ
ὁδός σου ἐναντίον μου; NETS: “your way was not pretty before me”).32 Only
in Judg 3:17, is ἀστεῖος the translation equivalent of the Hebrew ‫ ָּב ִריא‬.33
LEH translates ἀστεῖος in Judg 3:17 as “handsome”,34 a translation which
is followed by NETS. Muraoka renders it here as “charming”.35 In La Bible
d’Alexandrie, the Greek term is translated as “plantureux” (which in Eng-
lish can be translated as “ample”). The French translation therefore keeps
the middle ground between the meaning in classical Greek (“nice, ele-
gant”) and the meaning which is requested by the context.36 LXX.D argues
that ἀστεῖος was probably meant as a euphemistic or even ironic designa-
tion of Eglon, which would imply that the translator was well aware of the
tenure of Judg 3:12‑30.37 LXX.D translates: “Eglom war ein recht ansehnli-
cher Mann”.
Thirdly, it has been argued that in Judg 3 the term ‫ ִמנְ ָחה‬is used in a
double sense: as a tribute from Ehud to Eglon, but equally as the offering

32 See also Jdt 11:23; Sus (OG) 1:7; 2 Macc 6:23: ὁ δὲ λογισμὸν ἀστεῖον ἀναλαβὼν (NETS:
“making a high resolve”).
33 Elsewhere in the Old Testament, ‫ ָּב ִריא‬has ἐκλεκτός (Gen 41:2, 4, 5, 7, 18, 20; 1 Kgs 5:3;
Hab 1:16; Zech 11:16), ἰσχυρός (Ezek 34:20) or παχύς (Ezek 34:3) as translation equivalent.
34 Lust, Eynikel and Hauspie, Greek-English Lexicon, 90.
35 Muraoka, A Greek-English Lexicon, 98. In Num 22:32, Muraoka translates ἀστεῖος as
“proper, appropriate”.
36 P. Harlé, Les Juges (BdA 7; Paris: Cerf, 1999), 97: “Notre traduction tient la balance
entre le sens du grec classique, ‘de belle apparence, élégant, joli’, et le sens requis par le
37 W. Kraus and M. Karrer, eds., Septuaginta Deutsch: Das griechische Alte Testament in
deutscher Übersetzung (Stuttgart: Deutsche Bibelgesellschaft, 2009), 250.
64 hans ausloos

of Eglon by Ehud. Leaving aside the transliterations,38 ‫ ִמנְ ָחה‬has in the

LXX the nouns δῶρον, θυσία, θυσίασμα, ξένιον, ὁλοκαύτωμα and προσφορά
as translation equivalents. While some of these terms—such as θυσία,
θυσίασμα and ὁλοκαύτωμα—have an almost exclusively cultic or sacrifi-
cial usage, the translator of Judg 3 consistently translates ‫ ִמנְ ָחה‬as δῶρον.
Parallel to its Hebrew equivalent, this Greek word also has a double usage
in the LXX. As translation equivalent of the noun ‫ ָק ְר ָבּן‬, the term is used
in Leviticus and Numbers as indicating cultic and sacrificial offerings. In
Lev 1:2, for example, the expression ‫ ָק ְר ָבּן ַליהוָ ה‬is rendered as δῶρα τῷ
κυρίῳ. However, the term δῶρον can also be used in the more neutral way
of “gift”, as, for example, in Gen 32:14, 19, 21, 22 ( Jacob sends presents to
his brother Esau). Therefore, although it is difficult to determine whether
the translation equivalent in the LXX of Judg 3 with its double entendre has
been chosen deliberately or not, for the reader of the LXX, the term δῶρον
could have implied a similar double entendre as well.
Finally, the same conclusion can be drawn with respect to the transla-
tion of the verbal form ‫( וַ יַּ ְק ֵרב‬Judg 3:17). Here, LXX uses a form of the verb
προσφέρω (προσήνεγκεν). As in the Hebrew, this verb can have two con-
notations. Although προσφέρω, particularly in the books of Leviticus and
Numbers, is very often used in a cultic context as the translation equiva-
lent of the verb ‫“( ָק ַרב‬to approach, to bring near”), often in combination
with the noun θυσία (e.g. in Gen 43:26), the term has a profane connota-
tion as the translation equivalent of the Hiphil of ‫וַ יָּ בֺא‬: εἰσῆλθεν δὲ Ιωσηφ
εἰς τὴν οἰκίαν καὶ προσήνεγκαν αὐτῷ τὰ δῶρα (NETS: “And Ioseph came into
the dwelling, and they brought to him the presents . . .”).
In sum, whether or not the Greek double entendre was intentional
(but opinions diverge in this respect for the Hebrew as well), with the
exception of the rendering of Eglon’s name, the terminology in Greek can
be interpreted as referring to the profane offering of gifts by Ehud, but
equally to the cultic offering of the king by Ehud.39

38  Cf. the lexemes μαανα, μανα, μαναα(ν), μανααμ, μαναχ, μαννα, μαναειμ. See E. Hatch
and H. A. Redpath, A Concordance to the Septuagint and the Other Greek Versions of the Old
Testament (Including the Apocryphical Books) (2nd ed.; Grand Rapids: Baker, 1998), 292.
39 With regard to the question of how the reader of the LXX interpreted the meaning
of some Greek lexemes, see M. Harl, “La langue de la Septante,” in La bible grecque des
Septante. Du judaïsme hellénistique au christianisma ancien (Initiations au christianisme
ancien; eds. M. Harl, G. Dorival and O. Munnich; Paris: Cerf, 1994), 223–266, esp. 251–253.
judges 3:12–30 65

5. Further examples of wordplay

Besides the occurrences of wordplay within the characterisation of Ehud

as “left-hander” and Eglon as the young bull who will be sacrificed, there
are some minor elements in Judg 3:12‑20 in which the Hebrew author
seems to play with words. Again, the phenomenon of double entendre
seems to be prominent. The first one can be found in v. 16, where it is
said that Ehud made a dagger. The Hebrew text reads as follows: ‫וַ ּיַ ַעׂש לֹו‬
‫ ֵאהּוד ֶח ֶרב‬. According to the grammar of Joüon and Muraoka, this verse
functions as an example of the Hebrew use of the reflexive pronoun.40
Interpreting ‫ לֹו‬as such, the author would indicate that Ehud made a dag-
ger for himself. However, even if one accepts that Judg 3:16 is an example
of the reflexive usage of the personal pronoun, the particle preposition -‫ל‬
with the personal suffix could have another function as well in the pres-
ent context. In order to see so, let us compare it to a number of other
examples. Gen 19:3, for example, narrates how Lot prepares a feast “for
them”, i.e. the two messengers who visit him: ‫וַ יַּ ַעשׂ ָל ֶהם ִמ ְשׁ ֶתּה‬. In Exod
37:2, it is said that Bezaleel makes a crown of gold “for him”, i.e. the ark
(‫)וַ יַּ ַעשׂ לוֹ זֵ ר זָ ָהב‬. So it is at least possible that the author of Judg 3:16 uses
the personal pronoun with the particle preposition in a double sense: the
dagger that Ehud manufactures “for himself ” is actually made “for him”,
i.e. Eglon, who will be killed by the weapon. To a translator, it is not easy
to render this double entendre of the term. The LXX (A and B) has unam-
biguously translated one single meaning by rendering ‫ לוֹ‬as a reflexive pro-
noun (“to/for himself ”): καὶ ἐποίησεν ἑαυτῷ Αωδ μάχαιραν δίστομον (NETS:
“And Aod made for himself a double-bladed dagger”).
A final element of double entendre in the Hebrew text can be found
in the use of the term ‫ ָדּ ָבר‬in Judg 3:19–20. When Ehud returns to king
Eglon, he addresses him twice with parallelly structured words. In v. 19,
Ehud says: ‫“( ְּד ַבר ֵס ֶתר ִלי ֵא ֶליָך ַה ֶּמ ֶלְך‬I have a secret message for you, o
King”). When the servants have left the room, Ehud (v. 20) addresses the
king for the second time: ‫ֹלהים ִלי ֵא ֶליָך‬ ִ ‫“( ְּד ַבר ֱא‬I have a word of God for
you”). There can be no doubt that once more a term is interpreted differ-
ently by Eglon and the attentive reader of the text. For Eglon, the noun
‫ ָדּ ָבר‬refers to a secret message, which, moreover, is a message of God,41

40 P. Joüon and T. Muraoka, A Grammar of Biblical Hebrew (SubBi 14; Rome: Pontificio
Istituto Biblico, 1993), §146k.
41  It is interesting to note that within the Ehud pericope only in v. 20 the term ‫ֱאל ִֺהים‬
is used. Compare with the tetragram in Judg 3:12(ter), 15(bis), 28. J. M. Sasson, “Ethically
66 hans ausloos

and which lets him rise up from his throne (because of respect?). The
reader, on the contrary, can/will have interpreted ‫ ָדּ ָבר‬as a “thing”, refer-
ring to the dagger, which he interprets, moreover, as a divine weapon in
the hands of Ehud.42 In fact, it is not Ehud himself who will rescue Israel
by killing Eglon, but God who has sent Ehud as a deliverer (v. 15).43 It is
well possible that the author of Judg 3 has this double intention in mind,
although it is difficult to give a definitive answer in this regard.
The question at stake here pertains to the way in which the LXX trans-
lator has rendered this double entendre of the noun ‫( ָדּ ָבר‬which com-
monly occurs in Hebrew). In both verses (19 and 20), the LXX (A and B)
translates ‫ ָדּ ָבר‬by the Greek noun λόγος (θεοῦ). Next to the translation
equivalent λόγος, ‫ ָדּ ָבר‬has two other equivalents in the book of Judges:
besides πρᾶγμα (which clearly interprets ‫ ָדּ ָבר‬as “thing”), the term ῥῆμα
is also used. In Judg 11:10, for example, this term is an adequate ren-
dering of ‫ ָדּ ָבר‬in its meaning “word”. The elders of Gilead say to Jeph-
tah: “We will surely do conform to your word (‫”) ִכ ְד ָב ְרָך‬. However, in
Judg 8:1, where ‫ ָדּ ָבר‬clearly is used as “thing”, LXX equally reads ῥῆμα
(‫ית ָלּנוּ‬ ָ ‫— ָמה ַה ָדּ ָבר ַהזֶּ ה ָע ִשׂ‬τί τὸ ῥῆμα τοῦτο ἐποίησας ἡμῖν: “What is the
thing that you have done to us?)”. The same double meaning seems to
be true for the usage of λόγος, which, for example in Judg 11:28, is the
adequate rendering of ‫ ָדּ ָבר‬, meaning “word”: ‫וְ לֺא ָשׁ ַמע ֶמ ֶלְך ְבּנֵ י ַעמּוֹן‬
‫( ֶאל ִדּ ְב ֵרי‬NRSV: The king of the Ammonites did not heed the message
that Jephthah sent him”). The LXX here reads τῶν λόγων Ιεφθαε (A and
B). On the contrary, in Judg 21:11, where ‫ ָדּ ָבר‬seems to mean “thing”
(‫“—וְ זֶ ה ַה ָדּ ָבר ֲא ֶשׁר ַתּ ֲעשׂוּ‬This is the thing that you shall do”), LXX (A)
equally uses here the noun λόγος (καὶ οὗτος ὁ λόγος ὃν ποιήσετε).44 So, for
the LXX of Judg 3:19 and 20, it seems theoretically possible that the trans-
lator, as a sort of Hebraism, equally uses λόγος in a double sense: Ehud is

Cultured Interpretations,” 576 also mentions the possibility that the noun ‫ ֱאל ִֺהים‬func-
tions to denote a superlative: “the gravest message”.
42 In the book of Judges, the noun ‫ ָדּ ָבר‬occurs in Judg 2:4; 3:19, 20; 6:29(bis); 8:1, 3; 9:3,
30; 11:10, 11, 28, 37; 13:12, 17; 16:16; 18:7(bis), 10, 26; 19:19, 24; 20:7, 9; 21:11.
43 Moreover, texts such as 1 Sam 9:27–10:1 seem to mix both meanings of ‫ ָדּ ָבר‬. Samuel
announces to Saul that he has a ‫ ְדּ ַבר ֱאל ִֺהים‬for him. However, instead of saying some-
thing to him, Samuel takes a vial of oil and pours it on Saul’s head.
44 Cf., however, NETS’s translation: “And this is the word that you shall carry out” (καὶ
οὗτος ὁ λόγος ὃν ποιήσετε). The B text reads καὶ τοῦτο ποιήσετε (NETS: “And this is what you
will do”). Compare with LXX.D: “Das ist es, was ihr tun sollt” (A); “Dies sollt ihr tun (B)”.
judges 3:12–30 67

not only the messenger of a (divine) word, but the dagger in his hands is
a (divine) thing used to kill Eglon.45

6. Conclusion

There can be no doubt that the Hebrew text of Judg 3:12‑20 is character-
ised by several types of wordplay, of which the so-called double enten-
dre seems to be the most prominent one. As for the LXX translator, in
some instances—such as when he transliterates Eglon’s proper name—he
is not able to render the double meaning, although we cannot be sure
that he has noticed it. In other instances, however, the translator is look-
ing for some good solutions, such as, for example, with regard to Ehud’s
left-handedness. In translating the problematic Hebrew construction
‫ ִא ֵּטר יַ ד יְ ִמינֹו‬with the neologism ἀμφοτεροδέξιον, the translator clearly
indicates that he has understood the problem and the specificity of the
Hebrew text. Finally, it is sometimes not clear whether the translator him-
self also interpreted his Greek term as implying a double entendre, as is the
case with his rendering of ‫ ָדּ ָבר‬as λόγος. In any case, even if the translator
was using λόγος in a double sense, it is questionable whether the reader
of the LXX would have understood this usage.
When we compare these results with the creative way the translator
deals with the three Hebrew hapax legomena in this pericope,46 one has
to conclude that the LXX translator of Judges surely had the intention of
offering a qualitatively adequate rendering of his Hebrew Vorlage. In this
respect, and at least with respect to this pericope, his labelling as a weak
translator,47 needs to be nuanced. Indeed, on the one hand, the compari-
son of two different content-related criteria, i.c. the Greek rendering of

45 With regard to the possible theological connotation of λόγος as translation equiva-

lent of ‫ ָדּ ָבר‬, see C. Doignez and M. Harl, Le Deutéronome (BdA 5; Paris: Cerf, 2007), 40–43.
See, e.g., also ibid., 222 with regard to Deut 17:1, where ‫ ָדּ ָבר‬has been rendered by ῥῆμα:
“le mot rhema est un décalque de l’hébreu dabar, ‘parole’ ou ‘chose’ ”.
46 H. Ausloos and B. Lemmelijn, “Characterizing the LXX Translation of Judges,” 189:
“Summarizing the conclusions, one observes that the Greek translator of Judges has striven
to offer a meaningful rendering of the Hebrew hapax legomena in his Vorlage. Both with
regard to the mentioning of the weapon, which the translator deliberately calls a μάχαιρα
(dagger) with the length of a σπιθαμή in Judg 3:16, as well as with respect to the enigmatic
verses in Judg 3:22–23 in which he equally attempts to render an understandable Greek
text, the LXX translator demonstrates creativity and originality, each time taking serious
consideration of the particular context of the Ehud narrative”.
47 Cf. I. Soisalon-Soininen, Die Textformen der Septuaginta-Übersetzung des Richter-
buches (AASF 72; Helsinki: Suomalainen Tiedeakatemia, 1951), 60: “Man geht kaum fehl,
68 hans ausloos

Hebrew hapax legomena and the Greek rendering of wordplay, confirms

the rather creative attempts of the translator of Judges. On the other hand,
the confrontation ad extra with the more traditional evaluation of LXX
Judges as very literal seems to be questioned. Knowing that the charac-
terisation of a specific LXX translation must be based on as many differ-
ent criteria as possible, this preliminary evaluation is, of course, open to
debate and should surely be complemented by other research. Neverthe-
less, it opens up a surprising perspective.

wenn man Jdc. für die seinem Sprachgebrauch nach schwächste Übersetzung der ganzen
LXX hält”.
B or not B? The Place of Codex Vaticanus in textual
history and in Septuagint research

Siegfried Kreuzer

1. Introduction

One of the most basic facts in Septuagint studies is the primary text used
for analysis or comparison. This statement will hardly be challenged.
Yet it is a fact which needs to be reflected upon. For most studies on
the Septuagint, but also in studies on the biblical quotations in the New
Testament or even in comparisons with the daughter translations, the
Göttingen edition or Rahlfs’s so-called Handausgabe is often simply taken
as the starting point.1
Now, there is a reason for this approach: in his critical edition Rahlfs
wanted to reconstruct the Old Greek and the Göttingen edition maintains
that same goal. Yet it is and remains a goal only, and how this goal is pur-
sued and perhaps achieved depends on the methods and presuppositions
of the editor; and that’s exactly where Codex Vaticanus comes into play.
The critical editions use different manuscripts for their reconstructions, yet
Codex Vaticanus takes pride of place and dominates the editions. Rahlfs
states in his Handausgabe that he mostly relies on the Majuscules B, S and
A; among them—with a few exceptions as in the book of Judges—he basi-
cally followed Manuscript B, the Codex Vaticanus. The same holds true for
most of the volumes in the Göttingen edition.

1.1. The Importance of B for the Editions

Since the 16th century Codex Vaticanus has remained the most impor-
tant, i.e. the most highly esteemed, manuscript for Septuagint editions
and research. With the exception of the Aldina from 1518 and the Com-
plutensian Polyglot from 1514–1717, and starting with the Sixtina (1587),

1 Septuaginta: Vetus Testamentum Graecum. Auctoritate Academiae Scientiarum Gottin-

gensis editum (Göttingen: Vandenhoeck & Ruprecht, 1931ff.); A. Rahlfs and R. Hanhart, eds.,
Septuaginta: Id est Vetus Testamentum iuxta LXX interpretes (Stuttgart: Deutsche Bibelge-
sellschaft, 2006).
70 siegfried kreuzer

almost all editions of the Septuagint were more or less based on Codex
The first printed version of the Septuagint appeared in Venice in 1518
from the printer Aldine and is therefore called the Aldina. It used manu-
scripts which were available in the city at that time. At about the same
time the famous Complutensian Polyglot was prepared at the Alcala Uni-
versity in Complutum, near Madrid. For this edition manuscripts from
Rome and other cities were collected or borrowed.2 After about two gen-
erations the later Pope Sixtus V initiated and published a new edition;
manuscripts were sought and Codex Vaticanus became the basis of the
new edition, not only because it was available in Rome, but also because
of its excellent state of preservation and legibility.
Basically all Septuagint editions from the 17th through to the 19th
centuries,3 and even into the first half of the 20th century, have been based
on the Codex Vaticanus. More or less the only exception was Johannes
Ernestus Grabe’s edition of 1709–1720, which was based on Codex Alex-
andrinus. Practically all other editions are diplomatic editions of Codex
Vaticanus with an ever expanding critical apparatus, i.e. with the readings
of other manuscripts as they became available. This holds true also for the
editions of Holmes-Parsons,4 of Swete,5 and of Brooke-McLean.6 Even the
critical edition by Rahlfs and, to a large extent, the Göttingen edition rely
heavily on Codex “B.”

2 Cf. the basic study by F. Delitzsch, Studien zur Entstehungsgeschichte der Polyglot-
tenbibel des Cardinals Ximenes (Leipzig: Alexander Edelmann, 1871); and, more recently:
S. O’Connell, From Most Ancient Sources: The Nature and Text-Critical Use of the Greek Old
Testament Text of the Complutensian Polyglot Bible (OBO 215, Fribourg / Göttingen: Aca-
demic Press Fribourg / Vandenhoeck & Ruprecht, 2006).
3 H. B. Swete (An Introduction to the Old Testament in Greek [Cambridge: Cambridge
University Press, 1900], 182) counts at least 21 editions of that kind, deriving directly or
indirectly from MS B. Swete himself based his edition on a new collation of MS B.
4 R. Holmes and J. Parsons, eds., Vetus Testamentum Graecum Cum Variis Lectionibus
(Oxford: Clarendon Press, 1798–1820).
5 H. B. Swete, The Old Testament in Greek according to the Septuagint (3 vols.; Cam-
bridge: Cambridge University Press, 1887–1894).
6 A. E. Brooke and N. McLean, eds., The Old Testament in Greek according to the Text
of Codex Vaticanus supplemented from Other Uncial Manuscripts with a Critical Apparatus
Containing the Variants of the Chief Ancient Authorities for the Text of the LXX (Cambridge:
Cambridge University Press, 1906–1940).
b or not b? 71

1.2. Reasons for the Importance of Codex Vaticanus

There are good reasons for the importance of Codex Vaticanus. First, it
was and still is the oldest MS attesting almost all of the Old and the New
Testaments. There is a consensus that it was written in the 4th century,
but there is some debate about whether it was produced in Caesarea in
Palestine or in Egypt. The reason for Caesarea as a suggestion is its rela-
tion to Codex Sinaiticus; the reason for suggesting Egypt is the agreement
of the order of the books with the canon list of Athanasius and because of
some relation to the Egyptian text, especially in the Psalms. Yet both argu-
ments suppose that these traditions (the specific order of the canon list,
and the Greek Vorlage of the Egyptian text) had been available in Egypt
only. So what seems certain thus far is that Codex B originated in the
Eastern Mediterranean in the 4th century. It was written by at least two
scribes, and—as we will see—the Codex or its Vorlage was put together
from scrolls of different textual traditions.
As the ink had faded over the centuries, it was re-inked in the Middle
Ages, some time after the 10th century. This is evident because the so-
called enclitic nu has been re-inked according to the late Byzantine school
rule. The Codex probably came into the Vatican Library via southern Italy.
Unfortunately, the beginning of the Codex, i.e. most of the book of Gen-
esis, is missing.
There are only a few other large codices from late antiquity. There is
Codex Sinaiticus. The complicated history of its discovery, starting with
Tischendorf ’s journey to the St. Catherine Monastery in Sinai in 1844, does
not need to be repeated here.7 There are indications that this Codex origi-
nated in Caesarea, and one of its three or four scribes was possibly identi-
cal with a scribe of Vaticanus. The Codex has been dated between 340 and
360 C.E. It was probably bound only in the 6th century and brought to
St. Catherine’s soon afterwards. Most important is the fact that the Codex
underwent several corrections. Corrector A (Ca) seems to have worked in
the 6th century; he added readings from other traditions. Unfortunately
large parts of the Codex are missing: besides some newly published chap-
ters of Joshua and Judges, the older historical books and also large parts

7 It will be of interest that the famous debate about the legitimacy of its being taken
away from the St. Catherine Monastery has now been resolved, as the documents dealing
with the donation of the Codex to the Tsar by the monastery have been found in Moscow;
cf. C. Böttrich, “Neue Dokumente zur Geschichte des ‘Codex Sinaiticus’,” Early Christianity 1
(2010): 605–613.
72 siegfried kreuzer

of the Pentateuch are missing. With exceptions in a few books, the text,
at least in the Old Testament, is very close to that of Codex Vaticanus.
The remarks of Corrector A present an older textual tradition: some
of the corrections seem to presuppose a Hexaplaric text, others are close
to the Lucianic/Antiochene text. The fact that these corrections have been
added only shows that this tradition was important for the corrector, but
not how old the sources for these corrections are. The Codex with all of its
parts is now accessible in an excellent edition on the internet.8
The next codex is Codex Alexandrinus, generally dated to the 5th cen-
tury. There is practically no information about the origin of the Codex.
Around 1300 it came to Alexandria and in 1627 it was presented to King
Charles I of England. Its text is considered less consistent than Vaticanus,
but this is also a question of the standard for comparison, as Codex Vat-
icanus is also mixed, at least between kaige and non-kaige sections. In
Leviticus, Isaiah and Jeremiah, Alexandrinus seems to be close to Vati-
canus; in other parts it shows Hexaplaric influence; but it is also close
to the Antiochene text; and it has many idiosyncratic readings. As men-
tioned above, it was used for the edition by Grabe 1709–1720. Grabe also
wrote an essay about Alexandrinus attesting the oldest text in the book
of Judges, which was evidently the reason for its specific presentation of
the book of Judges in the editions of Brooke-McLean and Rahlfs. Different
from Vaticanus and Sinaiticus, the beginning of the codex, i.e. the whole
book of Genesis, is preserved in Codex Alexandrinus.
Regarding the older historical books, I would like to mention two more
codices. Codex M, which is Codex Coislianus from the 7th century, is
now in Paris. It comprises the Octateuch and the historical books. It has
many marginal readings with Hexaplaric material. In the historical books
its text is a mixture between the kaige text and the Antiochene text. Its
close counterpart is Codex N, also called Codex V like Codex Venetus.
The reason for the different designations is that N and V are two parts of
one codex. One part is in Rome, while the other part is in Venice. Because
of its close relation to M, the Roman part has been named N in Brooke-
McLean. It was later identified as the first part of a codex whose second
part had become known as Codex Venetus. Rahlfs therefore uses one and
the same siglum for both parts. Codex V dates from the 8th century and

b or not b? 73

in both parts it comprised the entire Old Testament.9 Both codices repre-
sent a text which is partly close to the Antiochene or Old Greek texts and
partly to kaige texts, just like Codex Vaticanus.
Looking at these codices with their later origin and their idiosyncrasies,
it is understandable that Codex Vaticanus became and still is considered
the most important single witness to the Septuagint. Yet there are other
and earlier documents as well.

1.3. Other Early Manuscripts and Witnesses of the Septuagint

The picture opens out if we take into account other manuscripts and the
witnesses of authors writing in Greek and the daughter translations. Besides
some manuscripts from pre-Christian times such as Papyrus Fouad 266,
several fragments from Qumran, and especially the Greek Minor Prophets
scroll from Naḥal Ḥever, today papyri from the 1st to the 4th century C.E.
from different parts of the Old Testament have been identified. For most
books of the Septuagint Codex Vaticanus is currently just one of many
witnesses that are relatively contemporary, but some witnesses are several
centuries older. I cannot go into the details here, but only refer to the
impressive and interesting list “Das handschriftliche Material für die ein-
zelnen Bücher des Alten Testaments” in the Verzeichnis der Griechischen
Handschriften edited by Detlef Fraenkel.10
Interestingly, amidst this wealth of manuscripts there is an excep-
tion for the older historical books, i.e. 1–4 Reigns, 1–2 Chronicles and 1–2
Esdras. There, with the exception of 2 Chron, Codex Vaticanus is still the
oldest witness of the Greek text in the form of a codex.
Yet also in these parts of the Septuagint there are other important
witnesses to the Greek text as well. First, there is the Jewish writer Jose-
phus, who in his Jewish Antiquities referred to the historical books and
quoted them; and even in his allusions it is sometimes possible to identify
the specific form of the text that he is referring to. Adam Mez made an
explicit comparison already in 1895.11 Later Henry St. J. Thackeray did the

  9 It probably was also originally a codex of the whole Bible, with the NT portion now
lost; cf. A. Rahlfs and D. Fraenkel, Die Überlieferung bis zum VIII. Jahrhundert (Vol. 1.1 of
Verzeichnis der griechischen Handschriften des Alten Testaments; Göttingen: Vandenhoeck
& Ruprecht, 2004).
10 Ibid., 472–497.
11  A. Mez, Die Bibel des Josephus, untersucht für die Bücher V–VII der Archäologie (Basel:
Jaeger & Kober, 1895). Mez’s results have been pushed aside by Rahlfs in A. Rahlfs, Lucians
Rezension der Königsbücher (Septuaginta-Studien III; Göttingen: Vandenhoeck & Ruprecht,
1911 [repr. 1965]), but they were confirmed by H. St. J. Thackeray, Josephus: The Man and the
74 siegfried kreuzer

same in connection with his studies on Josephus.12 The relevant passages

can now be found in the critical apparatus of the Madrid edition of the
Antiochene text.
But there are also Christian writers who quoted the Septuagint, and at
least some of them quoted not only Genesis, Psalms and the Prophets,
but also the historical books. Unfortunately, the older fathers like Justin,
Irenaeus or Clement of Alexandria do not offer much material from these
books. But in the 4th century there were Theodoret of Cyrus with his com-
mentaries, Chrysostom with his sermons, and other, although less prolific,
authors such as Asterius Sophista. With some caution their works can be
seen as witnesses of the texts they used. Theodoret in his commentaries
is an especially excellent witness to the biblical text he used, which is the
so-called Antiochene or Lucianic text.13
Considering this environment, one must say that the large old codices
have their special importance, but they are not the only textual witnesses.
Beyond that, looking at the whole field, one must admit that Vaticanus
(and Sinaiticus, where extant) represent a minority position within the
field, and that Codex Vaticanus too is of a mixed character as can be seen
especially in the change between kaige and non-kaige sections.
To appreciate the whole picture, a look at the daughter translations,
especially the Old Latin, is necessary. The Old Latin version originated
in the 2nd century, probably at more than one place, probably in more
than one attempt, and probably with even some Jewish origins. But this
does not further impact on our discussion. With its origin in the 2nd cen-
tury the Old Latin is older than Origen’s Hexapla and existed long before
Lucian. It therefore witnesses to a pre-Hexaplaric and a pre-Lucianic text.
As the Old Latin was produced for and among Latin-speaking people in
the western part of the Roman Empire, it must have been based on a
Greek Vorlage used in those western areas.14

Historian (New York: Jewish Institute of Religion, 1929), and E. Ulrich, The Qumran Text of
Samuel and Josephus (Chico: Scholars Press, 1978).
12 Thackeray, Josephus.
13 Cf. the editions: N. Fernández Marcos and A. Sáenz-Badillos, Theodoreti Cyrensis
quaestiones in octateuchum (TECC 17, Madrid: CSIC, 1979); N. Fernández Marcos and
A. Sáenz-Badillos, Theodoreti Cyrensis quaestiones in reges et paralipomena (Madrid, 1984);
Theodoret of Cyrus, The Questions on the Octateuch, Vol. 1. On Genesis and Exodus; Vol. 2.
On Leviticus, Numbers, Deuteronomy, Joshua, Judges, and Ruth (ed. J. F. Petruccione. Trans.
R. C. Hill, Washington: Catholic University of America Press, 2007).
14 Unfortunately the state and tradition of the Old Latin text is complicated. There are
only a few codices and fragments from palimpsests and, for some books, marginal notes in
biblical manuscripts. In spite of these problems and the fragmentary character of the Vetus
b or not b? 75

Something similar can be said about the earlier Egyptian translation,

the so-called Sahidic version from Upper Egypt. Different from the younger
Bohairic translation of Lower Egypt, the Sahidic version was produced in
the 3rd century for people who no longer spoke Greek or who never had
before. Through its Vorlage this Egyptian translation attests a Greek text
from the 3rd century, which was probably older than the Hexapla and also
certainly predates the floruit of Lucian.15
Taking these things together, there is an obvious conclusion: if there
is an agreement between the Antiochene text and the Old Latin and the
Sahidic translation, then we have a textual tradition which was wide-
spread from Syria to Egypt and to the West, and which most probably is
very old, if not identical with the original Septuagint or the so-called Old
Greek. From this wider picture we return to B, the Codex Vaticanus.

2. Codex Vaticanus and the other forms of the text

2.1. Some Peculiarities of Codex Vaticanus

In this section we will concentrate on the historical books. There is a spe-
cial problem in the book of Judges. Besides the textual tradition in Vati-
canus, there is definitely a different textual form in Codex Alexandrinus.
Brooke-McLean highlighted this text by using a different font.16 Rahlfs
went two steps further in his edition: (1) He juxtaposed both traditions;
and (2) he used the text of Alexandrinus as the basis for a critical recon-
struction. In doing so, Rahlfs presented what—at least in his opinion—

Latina, the Old Latin is an important witness to an old stage of the Septuagint. For the Old
Latin texts from Samuel, Kings and Chronicles cf. the apparatus in N. Fernandez Marcos
and J. R. Busto Saiz, El Texto Antioqueno (TECC 53/56/60; Madrid: CSIC, 1989/1992/1996).
15 For the Sahidic Version, cf. S. P. Brock, “Bibelübersetzungen I.5 Bibelüber-setzungen
ins Koptische 2. Altes Testament,” TRE 6: 199–200. Especially for the older Coptic Ver-
sions there is still much to be done. In a large project Karlheinz Schüssler has started
to collect and catalogue the manuscripts: K. Schüssler, ed., Das sahidische Alte und Neue
Testament: Biblia Coptica: Die koptischen Bibeltexte (Forschungsinstitut für Ägyptenkunde
und Koptologie der Universität Salzburg; Wiesbaden: Harrassowitz, 1995ff.).
For the books of Samuel there is the important edition by James Drescher: J. Drescher,
ed., The Coptic (Sahidic) Version of Kingdoms I, II (Samuel I, II) (CSCO 313–314; Scriptores
Coptici 35–36; Louvain: Secretariat du CSCO, 1970).
16 One reason for this decision was most probably the old treatise by Johannes Ernestus
Grabe on the priority of Codex Alexandrinus in the book of Judges: J. E. Grabe, Epistola Ad
Clarissimum Virum, Dn. Joannem Millium . . . Quâ ostenditur, Libri Judicum Genuinam LXX.
Interpretum Versionem eam esse, quam Codex Alexandrinum exhibet (Oxford, 1705).
76 siegfried kreuzer

was the Old Greek, which he called text A, and presented below that the
text of Codex B, as the evidently secondary text.17
Yet in spite of the evident secondary status of B in Judges, in the subse-
quent books Rahlfs basically followed the Codex Vaticanus. The basis for
this procedure was his investigation of the Lucianic text in Kings, pub-
lished in 1911.18 To put it briefly: in the 1860s some MSS had been iden-
tified as presenting the Lucianic or Antiochene text. Julius Wellhausen
in his book on the text of Samuel19 found that often those manuscripts
presented the oldest readings or even confirmed his conjectures. Accord-
ingly, Paul de Lagarde started his search for the Old Greek text by edit-
ing the Lucianic textual tradition. This line of research was supported by
Adam Mez, who—as mentioned above—had compared the Antiquities of
Josephus and had found that Josephus’ biblical quotations basically agree
with the Lucianic text. This meant that the Lucianic text is not the result
of a late reworking by Lucian around 300, but basically already existed in
the 1st century. Besides that, there was the Old Latin text from the 2nd
century, which also agreed with the Lucianic text, most of the time against
Contrary to this, Rahlfs basically defended the opposite view in his
investigation. For him Codex Vaticanus without a doubt represented the
oldest text and everything had to be compared with it. Whereas the Luci-
anic text was only a text type connected with the authority of that mar-
tyr (as the famous remark of Hieronymus can be understood as saying),20
Codex Vaticanus was equated with the Old Greek and all the differences
against Vaticanus were interpreted as the result of the comprehensive
redactional activity of Lucian.
For this purpose Rahlfs minimised all evidence that indicated the con-
trary: the agreements with Josephus were explained away; he accepted
only some forms of names as original. Also the evidence of the Old Latin

17  It should be noted that text “A” is not identical with Codex A, while text “B” is basi-
cally identical with Codex B (in the footnotes to B, Rahlfs only mentions corrections from
within the Codex).
18  Rahlfs, Lucians Rezension.
19  J. Wellhausen, Der Text der Bücher Samuelis (Göttingen: Vandenhoeck & Ruprecht,
20 “Alexandria et Aegyptus in Septuaginta suis Hesychium laudat auctorem, Constan-
tinopolis usque Antiochiam Luciani martyris exemplaria probat, mediae inter has pro-
vinciae palestinos codices legunt, quos ab Origene ela boratos Eusebius et Pamphilius
vulgaverunt, totusque orbis hac inter se trifaria varietate conpugnat.” Jerome, Preface to
Chronicles. Cf. R. Weber and R. Gryson, Biblia Sacra iuxta Vulgatam Versionem (Stuttgart:
Deutsche Bibelgesellschaft, 2007).
b or not b? 77

was explained away as being secondarily influenced by the Lucianic tra-

dition. In the same way, those quotations in the New Testament which
agreed with the Lucianic text were explained as a secondary influence
from the New Testament on the Lucianic manuscripts.
By this procedure—which he also had applied in his investigation on
the Psalms in 190721—Rahlfs came to his understanding of the Lucianic/
Antiochene text, which became most influential for the following decades
and in Septuagint research in general. It became one of the basic prin-
ciples for Rahlfs’s own edition and also for most editors in the Göttingen
edition, e.g. Ziegler and others.
In a description of the editorial work for the Septuagint volumes, Udo
Quast explained in 2000 that at the beginning of the editorial work one
knows little about the manuscripts and the recensional activities. “Lediglich
von dem Vorkommen der zwei großen christlichen Rezensionen des Ori-
genes und Lukian kann von vornherein—oder wenigstens in den meisten
Büchern—ausgegangen werden. Für sie stehen die Rezensionsmerkmale
außerdem weitestgehend fest.”22

2.2. The Problem of the Lucianic Text

What are these “well-known” characteristics of the Lucianic redaction?
Rahlfs mentioned three main traits: the first is the addition of articles;
the second is the addition of explanatory words such as the name of a
person speaking or acting. Yet those traits are irregular. Lucian not only
added words and articles, but he sometimes apparently deleted them.23
For Rahlfs this was not a question of his analysis; rather he made this
irregularity into a further characteristic of Lucian’s work and stated: “Der
Hauptcharakterzug dieser Rezension ist das Fehlen eines klaren Prinzips”.24

21  A. Rahlfs, Der Text des Septuagintapsalters (Septuaginta-Studien II; Göttingen: Van-
denhoeck & Ruprecht, 1907 [repr. 1965]).
22 “Only the occurrence of the two extensive Christian recensions, those of Origen and
of Lucian, can be assumed, at least for most of the books. Beyond that, the characteris-
tics of these recensions are well known and practically certain”. U. Quast, “Einführung in
die Editionsarbeit,” in Der Septuaginta-Psalter und seine Tochterübersetzungen (MSU 24,
eds. A. Aejmelaeus and U. Quast; Göttingen: Vandenhoeck & Ruprecht, 2000), 387–399
23 See in the synopsis below. Articles added in vv. 2, 6, 10; deleted in v. 10. Explanatory
words are added in: v. 2; deleted in v. 10.
24 “The main characteristic of this recension is that it has no clear principle”. Rahlfs,
Lucians Rezension, 293.
78 siegfried kreuzer

Yet one may say that the intention of Lucian was to improve the “Greek-
ness” of the text.25
This basic idea was taken over by Ziegler, among others. In his edition
of Jeremiah he mentioned the same characteristics as Rahlfs, and he also
wrote about Lucian’s irregularity. “Konsequenz war nicht seine Stärke”.26
Things changed with the discoveries from Qumran, especially the bib-
lical texts from Qumran and other places in the Judean desert. The first
scroll of the book of Samuel (4QSama) presented a text that was very close
to the Lucianic text, or rather its Hebrew Vorlage. This means that a sub-
stantial portion of the idiosyncrasies of the Lucianic text is not Lucianic,
but rather old, if not original. This new evidence from Qumran could not
be pushed aside as Rahlfs had done with the evidence from Josephus, from
the New Testament or from the Old Latin. The text from Qumran could
not have been influenced later by the Lucianic tradition. Qumran schol-
ars such as Frank Moore Cross, Eugene Ulrich or Emanuel Tov therefore
accepted the importance and the great age of the Lucianic text, or of the
Antiochene text as more accurate because it is more neutral. Yet they also

25 Rahlfs also mentions Lucian’s Atticising tendency. For Rahlfs this trait was less
important than it became in later scholarship: he mentions Atticising improvements (“atti-
sierende Verbesserungen”), but there are not many and they also are done irregularly.
The word Attic (“attisch”) is mentioned rather late in the study under the heading “other
changes” (“sonstige Änderungen”) (ibid., 176), where he first notices that the Aorist ειπα
was already in use in Attic, that it became common in the Hellenistic period, and that it
is found in both MSS A and B (ibid., 176–177); therefore it is not a sure sign. Indeclinable
δυο is used by the “Attiker”, and is found 4 times in L, while δυσιν, which was especially
favoured in Hellenistic times, is kept by Lucian in 2 Kgs 5:23 (“indeklinables δυο [kommt]
schon bei den Attikern vor (I 2,32; 22,31; II 21,5; 23,12)”), whereas δυσιν represents a “jüngere
Bildung, die in hellenistischer Zeit besonders beliebt ist, in II 5,23 [die] von Lukian beibe-
halten wurde”; (ibid., 259). For the change from ιερεις to ιερεας Rahlfs mentions 5 occur-
rences, but also 7 occurrences where Lucian did not make a change (ibid., 263). Altogether,
Rahlfs mentions seven forms or grammatical phenomena (ibid., 176, 204, 259, 260, 262, 263,
279) with 1 to 7 attestations (except ειπα, for which he does not mention any attestations),
and in almost every case also exceptions. Rahlfs concludes: “But Lucian is not a rigorous
Atticist because, in that case, he would have changed much more. And there are also cases
where Lucian uses a non-Attic form instead of the Attic form, e.g., τριτον instead of τρις,
also . . . I 6,7 πελυξ instead of πελεκυς and II 18,27 τον ουρον instead of το ουρον” (ibid., 281;
“Aber Lucian ist keineswegs strenger Attizist, er hätte sonst sehr viel mehr ändern müssen,
als er getan hat. Auch kommen Fälle vor, wo gerade L eine nichtattische statt der attischen
Form hat, wie τριτον statt τρις, ferner (in Abs. 1 nicht aufgeführt) I 6,7 πελυξ st. πελεκυς und
II 18,27 τον ουρον statt το ουρον”).
26 “Consistency was not his strength”. J. Ziegler, Beiträge zur Ieremias-Septuaginta
(MSU 6; Göttingen: Vandenhoeck & Ruprecht, 1958), 163.
b or not b? 79

embraced the old view about the Lucianic text and therefore developed
some compromise models.27

2.3. Naḥal Ḥever, Kaige and “Lucian Redivivus”

Another text has also become very important: the Greek Minor Prophets
scroll from Naḥal Ḥever. As is well known, this text has become the basis
for the identification of the kaige recension.28 This kaige recension, with
its close and formalistic adaptation of the Old Greek to the Hebrew text, is
not restricted to the Minor Prophets, but can be identified in other books
as well.29 What is most important for the current discussion is that the
kaige recension can also be found in Samuel and Kings. Barthélemy took
up an old observation by Thackeray, who had discerned four different
sections in Reigns, which he named with the Greek letters of the respec-
tive books.30 The sections βγ (2 Kgdms 10–3 Kgdms 2) and the section γδ
(3 Kgdms 22–4 Kgdms 25) are the kaige sections. These sections show the
same traits as the Naḥal Ḥever scroll. The other sections of Samuel and
Kings do not share these characteristics. Therefore they may be called the
non-kaige sections. The Greek text of these passages is of a different char-
acter and closer to the Old Greek.
It needs to be mentioned that this division refers to Codex Vaticanus
alone. While the kaige recension is a widespread phenomenon and can
be found in most books beyond the Pentateuch, the division within 1–4
Kgdms is a unique feature of Codex Vaticanus. It must have arisen from
the combination of different scrolls with different text types.
But what about the Antiochene text in these books? After identifying the
kaige sections, Barthélemy posed a logical question: if Codex Vaticanus is
secondary in the kaige sections, where is the Old Greek to be found? Can
it be identified in some other manuscripts or has it been lost? Barthélemy
compared the kaige text (in his book still called the Palestinian text) with
the Antiochene text and discovered that it must have been the base text
for the kaige recension. The questions and the answers can well be seen

27 E.g. E. Tov, “Lucian and Proto-Lucian,” RB 79 (1972): 101–113, for whom the Lucianic
text contains either the ancient Septuagint or an ancient Septuagint.
28 D. Barthélemy, Les Devanciers d’Aquila (VTSup 10; Leiden: Brill, 1963).
29 Ibid., 89: “Études sur quelques membres déjà connus du groupe καιγε.”
30 H. St. J. Thackeray, “The Greek translators in the four books of Kings,” JTS 8 (1907):
262–266, and idem, The Septuagint and Jewish Worship. A Study in Origins (Schweich Lec-
tures, London: Oxford University Press, 1921).
80 siegfried kreuzer

in the headings of the relevant chapters in his book. At first, by comparing

the two texts, he comes to the conclusion that there is a basic uniformity
between the two text forms: “identité de base entre la forme antiochienne
et la forme palestinienne du text grec”.31 This basic unity between the Pal-
estinian text and the Antiochene text matches clear differences that can
be explained in one way only: “la forme antiochienne ne peut être issue
de la forme palestinienne par abâtardissement”,32 i.e. the Antiochene text
cannot have originated from the kaige text. The dependence is the other
way round: the Antiochene text is the older text, probably very close to
the Old Greek or even identical to it (although not without corruptions in
the course of its transmission).33
This means that the discovery of the kaige recension and the new evalu-
ation of the Antiochene text are two sides of the same coin. This is not the
result of mere speculation or trepidation about the total absence of the
Old Greek for these texts; it is simply because the Antiochene text indeed
represents the text which was used and revised by the kaige revisers.
At about the same time as Barthélemy published his discoveries (1963),
a young man at Oxford had almost finished his dissertation. Sebastian
P. Brock worked on 1 Samuel in his dissertation, which was accepted in
1966.34 He did so in the rather traditional way and, just as many others
had done, followed Rahlfs. He did not use the few papers on the Qumran
scrolls that F. M. Cross had published by that time,35 but he took great
care with the Hexaplaric material and with his analysis of the Lucianic
recension. Evidently, he was not happy with Barthélemy’s book in which
the Lucianic recension had disappeared and the Antiochene text had
become the Old Greek, although with corruptions over time.
In 1965 Brock had the opportunity to deliver a paper on Barthélemy’s
book. Basically he defended the old ideas about Lucian. He gave it the fit-

31 Cf. Barthélemy, Devanciers, 92–102.

32 Cf. ibid., 110–113.
33 Cf. the famous statement: the Antiochene text is “la vielle septante, plus ou moins
corrompue et abatardie” (ibid., 127).
34 Printed much later: S. P. Brock, The Recensions of the Septuagint Version of 1 Samuel
(1966) (Torino: Silvio Zamorani, 1996).
35 F. M. Cross, “A New Qumran Biblical Fragment Related to the Original Hebrew
Underlying the LXX,” BASOR 132 (1953): 15–26; and idem, “The Oldest Manuscripts from
Qumran,” JBL 74 (1955): 147–72. See also W. F. Albright, “New Light on Early Recensions of
the Hebrew Bible,” BASOR 140 (1955): 27–33.
b or not b? 81

ting title “Lucian redivivus”.36 In this paper of only 6 pages he picked out
a few variant readings. His main argument was that the Lucianic text has
the same traits in the kaige and in the non-kaige section. He combined
this correct statement with the assumption that Codex Vaticanus in the
non-kaige sections represents the Old Greek and, therefore, the Lucianic
text is secondary. And he concluded: therefore, the Lucianic or Antiochene
text must also be secondary in the kaige sections.
At first this sounds convincing, but a closer look shows that Brock
jumps from a difference within Codex Vaticanus—i.e. the difference sig-
nalled by the terms kaige and non-kaige—to dating the Antiochene text.
This is simply false reasoning. The difference within the text of Codex
Vaticanus leads to a different relation to the Old Greek. But that is a prob-
lem within Vaticanus and does not say anything about the Lucianic text.
The variant characters and ages within the text of Vaticanus do not affect
the character of the Antiochene text. What changes is not the Antiochene
text, but its relation to the different parts of Vaticanus because of their
different characteristics.
Interestingly, Brock’s small paper was never seriously examined; at
least there is no publication.37 Yet this paper became most influential. Its
consequence was that Barthélemy’s discovery of the kaige recension was
widely accepted, while the other side of the coin, his new evaluation of
the Antiochene text, was pushed aside; the original combination of the
two facts remains practically unknown. Representative of this situation
is the statement in Fernandez Marcos’s “Introduction to the Septuagint”,
where he summarises the importance of Qumran/Naḥal Ḥever and Bar-
thélemy’s discovery only in regard to the kaige recension:
With the obligatory refinements in matters of detail, Barthélemy’s funda-
mental thesis, according to which these fragments belong to a consistent
revision of the LXX to bring it close to a Hebrew text very similar to, but not
identical with, the proto-Masoretic text, has been firmly accepted. Some of
the particular features of this revision which Barthélemy noted, and others
identified in later studies, can be debated . . . However, there is absolutely no

36 S. P. Brock, “Lucian redivivus. Some Reflections on Barthélemy’s Les Devanciers

d’Aquila,” in Studia Evangelica, Vol. V, Papers presented to the Third International Congress
on New ­Testament Studies held at Christ Church, Oxford 1965 (TUGAL 103; ed. F.L. [sic!]
Cross; Berlin: Akademie Verlag, 1968), 176–181.
37 For a first analysis and discussion see S. Kreuzer, “Lucian Redivivus or Barthélemy
and Beyond?” in IOSCS Congress Volume Helsinki 2010 (ed. M. K. H. Peters; Atlanta,
82 siegfried kreuzer

doubt that these fragments belong to the LXX, which we knew through more
reliable ancient witnesses, but it was revised to adapt it with greater literal-
ism to the current Hebrew text . . . The finds from Naḥal Ḥever, together with
its general interpretation within the framework of the early history of the
LXX provided by Barthélemy, became an obligatory reference point for all
later studies.38
Lucian as a reviser was indeed revived by Brock and many Septuagint
scholars still take it for granted that there was a general Lucianic recension39
with its specific characteristics.40 For many, Codex Vaticanus is still the
text most close to, or even more or less identical with, the Old Greek, at
least in the non-kaige sections. This problem is not only relevant in Sam-
uel and Kings and in the other historical books such as Judges, Chronicles
and 2 Esdras, but also in other books, i.e. in the prophetic and in some
poetic books.

3. Evaluation and a new approach

3.1. The Relation of Kaige to the Antiochene Text

The typical phenomena can be seen in practically any kaige text in the
historical books. As an example, I use a passage from where my discover-
ies began, i.e. 2 Sam 15.

The following synopsis41 of 2 Sam 15:2b, 5–6, 10 shows all the typical char-
acteristics as Rahlfs has proposed, and it also shows that Rahlfs even in
the kaige section almost exclusively understood the text of Codex Vati-
canus as the oldest text (except v. 2b with the introduction of ὁ ἀνήρ).

38 N. Fernández Marcos, The Septuagint in Context: Introduction to the Greek Version of
the Bible (Leiden: Brill, 2000 = Atlanta: Society of Biblical Literature, 2009), 72.
39 The question is not about Lucian as a person of the Syrian church, living around
300 C.E. (although his role sometimes has been questioned), nor that there are remarks in
ancient manuscripts which relate specific texts to (someone named) Lucian, but about the
assumed extensive recensional activity of Lucian (or a person of that time).
40 Representative again is Fernández Marcos, Septuagint in Context, 229: “However, no-
one has doubted the peculiar nature of the Lucianic or Antiochene text in the historical
books (Samuel‑Kings‑Chronicles).”
41  In the following tables, I use the vocalised MT as a matter of convenience and also
to indicate that the “text” never consisted of consonants only, but was always a “vocalised”
text with a reading tradition.
b or not b? 83

The typical addition of the article can be seen in v. 2b line 7 (2x!), v. 5

line 2; v. 6 line 8; v. 10 line 3 (2x). The addition of an explanatory word is
found in v. 2b line 6 (in this case even a whole sentence καὶ ἀπεκρίνατο ὁ
ἀνήρ) and v. 6 line 8 (παντῶν). There is also a change in words: σάλπιγγος
instead of κερατίνης (v. 10 line 7) is not exactly the same thing, but it better
represents the function.
Also the opposite can be identified: in v. 2b line 2 there is no πρóς; in
v. 10 line 7 both articles are deleted, and in line 9 the word βασιλεὺς is
missing. Lucian indeed worked in irregular and even contradictory ways.
Why would he add two articles in v. 10, when he deletes the articles a few
words later? Why would he add some words in v. 2b and delete a word
in v. 10?
On the other hand, taking up Barthélemy’s identification of the kaige
recension, the translation is indeed closer to the Hebrew, not only with
regard to πρóς in v. 2b line 2, the missing sentence in line 5, and the miss-
ing counterpart for παντῶν in v. 6 line 8, but also with the word κερατίνη
for ‫ ׁשפר‬in v. 10 line 7.
Yet the representation of the Hebrew in the kaige text is also question-
able. It follows the rules for the article only partially, as in v. 2b line 8,
where the determination of ‫ עבדך‬is expressed by ὁ δοῦλός σου, or in v. 10
line 7, where the determination is mirrored in τὴν φωνὴν τῆς κερατίνης.
But there are also other cases: the determination of ‫מאחד ׁשבטי־יׂשראל‬
has no article as a counterpart, nor does ‫ אנׁשי יׂשראל‬in v. 6 line 8 or the
‫ ׁשבטי יׂשראל‬in v. 10 line 3.
This seeming inconsistency can be explained by the underlying herme-
neutics: for the kaige revisers not only were the grammar and meaning
important but also, and even more so, the surface of the text. The result
was an isomorphic translation (or better, adaptation) of the text that mir-
rored its Vorlage. This explains the seeming irregularity: ‫ׁשבטי יׂשראל‬
(v. 2b line 7 and v. 10 line 3) and ‫( אנׁשי יׂשראל‬v. 6 line 8) are determined
genitive constructions, but there is no visible article. Also ‫ איׁש‬in v. 5 line 2
has no article; therefore there is no article with ἄνδρα either. On the other
hand, where there is an article or a similar visible element42 in Hebrew,
there is also an article in the kaige text.

42 This refers especially to the nota accusativi ‫את‬. As ‫ את‬is used before a determined
object, it has basically the same function as the article.
84 siegfried kreuzer

2 Sam/2 Kgdms 15:2b, 5–6, 10

MT KR (Rahlfs/B) Ant (Madrid Edition)
‫ וַ ּיִ ְק ָ ֙רא‬2b καὶ ἐβόησεν καὶ ἐκάλει
‫ַא ְב ָׁש ֤לֹום ֵא ָל ֙יו‬ πρὸς αὐτὸν Αβεσσαλωμ αὐτὸν Αβεσσαλωμ
‫אמר‬ ֶ ֹ ‫וַ ּ֗י‬ καὶ ἔλεγεν αὐτῷ καὶ ἔλεγεν αὐτῷ
‫יר ַ֔א ָּתה‬ ֙ ‫י־מ ֶּז֥ה ִע‬
ִ ‫ֵא‬ ἐκ ποίας πόλεως σὺ εἶ ἐκ ποίας πόλεως εἶ σύ
καὶ ἀπεκρίνατο ὁ ἀνήρ
‫אמר‬ ֶ ֹ ‫וַ ּ֕י‬ καὶ εἶπεν [ὁ ἀνήρ > B] καὶ ἔλεγεν
‫ֵמ ַא ַ ֥חד ִׁש ְב ֵ ֽטי־יִ ְׂש ָר ֵ ֖אל‬ ἐκ μιᾶς φυλῶν Ισραηλ ἐκ μιᾶς τῶν φυλῶν τοῦ
‫ַע ְב ֶ ּֽדָך׃‬ ὁ δοῦλός σου Ισραηλ
ὁ δοῦλός σου
‫ וְ ָהיָ ֙ה‬5 καὶ ἐγένετο καὶ ἐγίνετο
ִ֔ ‫ִּב ְק ָר‬ ἐν τῷ ἐγγίζειν ἄνδρα ἐπὶ τῶ προσάγειν τὸν
‫ְל ִה ְׁש ַּת ֲחֹו֖ ת ֑לֹו‬ τοῦ προσκυνῆσαι αὐτῷ τοῦ προσκυνεῖν αὐτῷ
‫וְ ָׁש ַל֧ח ֶאת־יָ ֛דֹו‬ καὶ ἐξέτεινεν τὴν χεῖρα καὶ ἐξέτεινε τὴν χεῖρα
‫וְ ֶה ֱח ִ ֥זיק ֖לֹו‬ αὐτοῦ καὶ ἐπελαμβάνετο αὐτοῦ
αὐτοῦ καὶ ἐπελαμβάνετο αὐτοῦ
‫וְ ָנ ַׁ֥שק ֽלֹו׃‬ καὶ κατεφίλησεν αὐτόν καὶ κατεφίλει αὐτόν
‫ וַ ּיַ֙ ַעׂש ַא ְב ָׁש ֜לֹום‬6 καὶ ἐποίησεν Αβεσσαλωμ καὶ ἐποίει Αβεσσαλωμ
‫ַּכ ָּד ָ ֤בר ַהּזֶ ֙ה‬ κατὰ τὸ ῥῆμα τοῦτο κατὰ τὸ ῥῆμα τοῦτο
‫ְל ָכל־יִ ְׂש ָר ֵ֔אל‬ παντὶ Ισραηλ παντὶ Ισραηλ
‫ֲא ֶׁשר־יָ ֥בֹאּו ַל ִּמ ְׁש ָ ּ֖פט‬ τοῖς παραγινομένοις εἰς τοῖς παραγινομένοις εἰς
‫ל־ה ֶ ּ֑מ ֶלְך‬
ַ ‫ֶא‬ κρίσιν πρὸς τὸν ὸν βασιλέα κρίσιν πρὸς τὸν βασιλέα
‫וַ יְ גַ ּנֵ ֙ב ַא ְב ָׁש ֔לֹום‬ καὶ ἰδιοποιεῖτο Αβεσσαλωμ καὶ ἰδιοποιεῖτο Αβεσσαλωμ
‫ת־ל֖ב‬ ֵ ‫ֶא‬ τὴν καρδίαν τὰς καρδίας
‫ַאנְ ֵ ׁ֥שי יִ ְׂש ָר ֵ ֽאל׃‬ ἀνδρῶν Ισραηλ παντῶν τῶν ἀνδρῶν τοῦ
֙ ‫ וַ ּיִ ְׁש ַל֤ח ַא ְב ָׁש‬10 καὶ ἀπέστειλεν καὶ ἀπέστειλεν
‫ְמ ַרּגְ ֔ ִלים‬ Αβεσσαλωμ κατασκόπους Αβεσσαλωμ
‫ל־ׁש ְב ֵ ֥טי יִ ְׂש ָר ֵ ֖אל‬
ִ ‫ְּב ָכ‬ ἐν πάσαις φυλαῖς Ισραηλ εἰς πάσας τὰς φυλὰς τοῦ
‫מר‬ ֹ ֑ ‫ֵלא‬ λέγων λέγων
‫ְּכ ָׁש ְמ ֲע ֶכ ֙ם‬ ἐν τῷ ἀκοῦσαι ὑμᾶς ἐν τῷ ἀκοῦσαι ὑμᾶς
‫ת־קֹול ַהּׁש ֔ ָֹפר‬ ֣ ‫ֶא‬ τὴν φωνὴν τῆς κερατίνης φωνὴν σάλπιγγος
‫וַ ֲא ַמ ְר ֶּ֕תם‬ καὶ ἐρεῖτε καὶ ἐρεῖτε
‫ָמ ַ ֥לְך‬ βεβασίλευκεν βασιλεὺς βεβασίλευκεν
‫ַא ְב ָׁש ֖לֹום‬ Αβεσσαλωμ Αβεσσαλωμ
‫ּב ֶח ְב ֽרֹון׃‬. ְ ἐν Χεβρων ἐν Χεβρων
b or not b? 85

This surprising observation fits the fact that in early Judaism not only
the meaning but also the surface of the text had become important. This
can be seen in the strange phenomenon of the kaige recension that the
short form of the Hebrew personal pronoun ‫ אני‬is rendered with ἐγώ and
the long form ‫ אנכי‬with ἐγώ εἰμί, even if a finite verb follows and this
combination in Greek is not only strange but simply wrong. This differ-
ence simply indicates the form of the Hebrew pronoun, although there
is no difference in meaning. The basic idea is that every detail and every
element in the holy text has some importance, even more so if it seems
superfluous, just because it is there.43
From this we can move on to the Antiochene text. If we—for a test—
leave aside the old assumptions about Lucianic redaction and try to see
things the other way around, we arrive at two surprising conclusions
which lead to a new solution. (1) Considering the Antiochene text as a
whole and not only looking at the supposed redactional differences, this
text is a faithful rendering of the Hebrew parent text and at the same
time a text demonstrating rather good Greek. (2) The seemingly irregular
and even contradictory changes by the supposed Lucianic redaction turn
into a consistent explanation as a redaction aiming at isomorphic equiva-
lence. The changes of the article can be explained consistently: as demon-
strated above, the articles in the Antiochene text are in accordance with
the grammar of the Hebrew text, while the kaige recension has added
or deleted the articles according to its isomorphic principle. The same
is the case with the so-called explanatory words. Evidently, the transla-
tor of the Antiochene text added explanatory words or—more probably,
in view of the Qumran texts—there was a Hebrew Vorlage which was
slightly different from the MT and which already in Hebrew contained
such explanatory words. The kaige recension again adapted the Greek to
its Hebrew reference text, in this case a text more or less identical with
MT. According to the principles of the kaige recension explained above,
this again led to additions and to omissions (see the sentence with ἀνήρ in
v. 2b line 6 and παντῶν in v. 6 line 8), according to the same clear principle
as for the article. This close adaptation also leads to changes in number
(e.g. v. 6 line 7: singular τὴν καρδίαν according to ‫ )את־לב‬and addition or

43 On Early Jewish hermeneutics see e.g. C. Dohmen and G. Stemberger, Hermeneutik
der Jüdischen Bibel und des Alten Testaments (Stuttgart: Kohlhammer, 1996); and D. Patte,
Early Jewish Hermeneutic in Palestine (Missoula: Scholars Press, 1975).
86 siegfried kreuzer

adaptation of prepositions (v. 2b line 2: πρός according to ‫ ;אליו‬v. 10 line 3:

ἐν according to ‫ ב‬instead of εἰς.
Also the semantic change from σάλπιγξ to κερατίνη can be explained in
terms of this close formal adaption. While the Antiochene text/Old Greek
aimed at functional equivalence (“trumpet”), the kaige recension wants
material equivalence (“horn”). This and other things like the substitution
of the historical past have been discussed by Thackeray and Barthélemy
and go beyond the redactional principle presented here.
The close adherence to the Hebrew reference text also allows a conclu-
sion regarding its Vorlage. In v. 2 line 3, the Hebrew text evidently had ‫לא‬
(cf. αὐτῷ), and in v. 10 line 9, there must have been a second ‫מלך‬, read
as ‫ ֶמ ֶלְך‬. The reference text of the kaige recension was therefore close to,
but not fully identical with, MT. The Hebrew reference text of the Anti-
ochene text/Old Greek was only slightly different. The plus in v. 2b lines
5–6 presupposes the typical Hebrew combination ‫ויען—ויאמר‬, which was
therefore most probably in the Hebrew Vorlage.
The emphasis (the heart of ) all (men of Israel) may be an emphasis
added by the translator, but may well go back to the Hebrew text.44 On
the other hand, v. 10 line 9 βασιλεύς is missing. This also may be a change
by the translator or—in view of the general closeness to the Hebrew
text—represent a minus in the Vorlage.

Basically the same situation can be seen in the kaige text of 2 Kgs
(4 Kgdms). Unfortunately, the Qumran evidence for this book is very
meagre. But this does not matter, because the argument developed above
relies on the inner relation of the texts. On the other hand, there is some
interesting evidence from the Old Latin, as the following brief synopsis

44 The closeness to the Hebrew reference text can be seen by the fact that in v. 2 the
assumed additional ‫ ויען האיש‬is confirmed by ‫ וענה האיש‬in both, 4QSama and 4QSamc
(DJD XVII, 260). The difference between ‫ ויען‬and ‫ וענה‬is regularly found in the Qumran
texts. Both forms express past tense. 4QSama, although in the reconstructed text, has ‫כול‬
as equivalent for παντῶν in v. 6 line 8. Unfortunately, 4QSama is not extant beyond the
beginning of v. 7.
b or not b? 87

2 Kgs/4 Kgdms 6:8–9

MT kaige/B (Rahlfs) Ant (Madrid Edition)
‫ּומ ֶלְך ֲא ָ ֔רם‬
֣ ֶ 8 καὶ βασιλεὺς Συρίας καὶ βασιλεὺς Συρίας
‫ָה ָי֥ה נִ ְל ָ ֖חם ְּביִ ְׂש ָר ֵ ֑אל‬ ἦν πολεμῶν ἐν Ισραηλ ἦν πολεμῶν τὸν Ἰσραήλ,
‫ל־ע ָב ָ ֣דיו‬
ֲ ‫וַ ּיִ ּוָ ַע ֙ץ ֶא‬ καὶ ἐβουλεύσατο πρὸς καὶ συνἐβουλεύσατο
‫מר‬ֹ ֗ ‫ֵלא‬ τοὺς παῖδας αὐτοῦ λέγων τοῖς παισὶν αὐτοῦ λέγων
֥ ִ ‫ל־מ ֛קֹום ְּפ‬ ְ ‫ֶא‬ εἰς τὸν τόπον τόνδε τινὰ Εἰς τὸν τόπον τὸν φελμουνεὶ
‫ַא ְלמ ִֹנ֖י ַּת ֲחנ ִ ֹֽתי׃‬ ελμωνι παρεμβαλῶ ποιήσωμεν ἔνεδρον,
και ἐποίησαν.
Et consilium habuit cum pueris suis: dicens: In
locum Phelminiim insidia faciamus L115 in locum
Phelmunim obsessionem faciamus L91–95
‫ וַ ּיִ ְׁש ֞ ַלח‬9 καὶ ἀπέστειλεν καὶ ἀπέστειλεν
‫ֹלהים‬ ִ֗ ‫ִ ֣איׁש ָה ֱא‬ Ελισαιε ὁ ἄνθρωπος τοῦ θεοῦ
֙‫ל־מ ֶלְך יִ ְׂש ָר ֵאל‬
֤ ֶ ‫ֶא‬ πρὸς [τὸν >B] βασιλέα πρὸς τὸν βασιλέα Ἰσραὴλ
‫מר ִה ָּׁ֕ש ֶמר ֵמ ֲע ֖בֹר‬ ֹ ֔ ‫ֵלא‬ Ισραηλ λέγων φύλαξαι μὴ λέγων Πρόσεχε τοῦ μὴ
‫ַה ָּמ ֣קֹום ַה ֶּז֑ה‬ παρελθεῖν διελθεῖν
‫י־ׁשם ֲא ָ ֥רם נְ ִח ִ ּֽתים׃‬ ֖ ָ ‫ִ ּֽכ‬ ἐν τῷ τόπῳ τούτῳ τὸν τόπον τούτον,
ὅτι ἐκεῖ Συρία κέκρυπται ὅτι ἐκεῖ Σύροι ἐνεδρεύουσιν.
Et mandavit homo dei L115

We find similar phenomena as above. In v. 8 line 2, the article gets replaced

by the preposition ἐν (Israel) according to the ‫ב‬. In v. 8 line 3, πρóς is
introduced according to the ‫ ֶאל‬in the Hebrew text. In v. 8 line 6, the
intentionally correct but free rendering ποιήσωμεν ἔνεδρον, “let us make an
ambush,” is replaced by the exact rendering παρεμβαλῶ, for ‫תחנתי‬.
In v. 9 line 3, the article is deleted because in ‫ מלך יׂשראל‬there is
no visible article. Verse 9 line 5 seems to presuppose ‫ במקום‬instead of
‫המקום‬, while the accusative and the article in the Antiochene text pre-
suppose the article of the MT. In v. 9 line 6, Σύροι is replaced by Συρία as
the exact rendering of ‫ארם‬.
In this passage the Old Latin is very interesting. The Latin expression
“com pueris suis” may represent either version. But “in locum Phelm-
iniim/Phelmunim” and “insidia/obsessionem faciamus” clearly represent
the Antiochene text. It is interesting that the expression ‫ פלני עלמני‬is
not translated but transcribed in both versions and represented with one
expression only, yet both “Phelminiim” and “insidia/ obsessionem” clearly
represent the Antiochene text. In v. 9 line 2 there is the change between
the name Ελισαιε and the title ἄνθρωπος τοῦ θεοῦ; both expressions can be
88 siegfried kreuzer

found in the context. Interestingly, the Antiochene text again agrees with
MT and it is confirmed by the “homo dei” in the Old Latin.
There are two conclusions:

(1)  The Old Latin confirms that the Antiochene text is pre-Hexaplaric and
pre-Lucianic. The comparison again shows that the text is deduced
from the Antiochene/Old Greek text;
(2) Differently from the situation in 2 Sam, in 2 Kgs the Antiochene text
seems closer to the MT, while the kaige text had a slightly different
reference text. Yet the only difference which presupposes a difference
in the Hebrew text is the change between the name Elisaias and the
title “man of God” in v. 9 line 2. With regard to the Hebrew text form,
this is a minor difference because of the repetition and the inter-
change of both elements in the context. But in regard of the age of
the text, the agreement with the Old Latin is proof of the antiquity of
the Antiochene text.

In sum, it can be said that, among other aspects, the two texts show that
in the kaige sections of both 2 Sam and 2 Kgs, the kaige text is a revision
and not the original text of the Septuagint. This proves that the text of
Codex Vaticanus in these sections is secondary.

3.2. The Relation of Codex Vaticanus to the Antiochene Text in

the Non-Kaige Sections
The situation in the non-kaige sections is more difficult. In these sections
of the historical books, especially 1–4 Kgdms, the text of Codex Vaticanus is
much closer to the Old Greek. Most authors hold it to be the witness closest
to the Old Greek, and many practically equate it with the Old Greek.
On the other hand, the Antiochene text has also proven to be very close
the Old Greek. The change between the kaige sections and the non-kaige
sections is a feature of Codex Vaticanus only. The Antiochene text shows
no such difference. Therefore, it can be assumed that also in the non-kaige
sections not only are the character but also the age and the relation to the
Old Greek about the same. So there are two textual traditions which are
seemingly very close to the Old Greek. If both B and the Antiochene text
are close to the Old Greek, B and the Antiochene text must also be closer
together. Yet there are many differences also in the non-kaige sections.
What is the relationship between these two text forms and how do they
relate to the Old Greek? We will consider some examples.
b or not b? 89

3.2.1. ἡ Βααλ—The Seemingly Female Baal

Starting with Judg 2:13 until 4 Kgdms 21:3, and 2 Chron 17:2 there is an
interesting feature: the name of the Canaanite god Baal is combined with
a female article: ἡ Βααλ, τῇ Βααλ or also ταῖς Βααλιμ. The meaning of this
strange feature is most probably a kind of Ketib-Qere in the Greek: the
female article indicates that the name of this God should not be pro-
nounced, but that it should be read as ἡ αἰσχύνη etc.45 This is confirmed by
the well-known change of the name of Ishbaal/Mephibaal to Ishboschet/
Mephiboshet and by the reading αἰσχύνη in 3 Kgdms 18:19, 25. This read-
ing practice can also be found in other books, especially in the book of
Jeremiah. Yet this reading practice was given up later. The kaige recension
in the historical books has returned to the male article with Baal, as can
be seen in Judg 2:13; 3:7; 10:6, 10.

τῇ Βααλ in 3 Kgdms 19:18 and Rom 11,4

Rom 11:4 1 Kgs/3 Kgdms 19:18 1 Kgs/3 Kgdms 19:18
Ant (Madrid) (Rahlfs)
4 ἀλλὰ τί λέγει αὐτῷ ὁ
χρηματισμός; κατέλιπον 18 καὶ καταλείψω ἐξ Ισραηλ 18 καὶ καταλείψεις ἐν
ἐμαυτῷ ἑπτακισχιλίους ἑπτὰ χιλιάδας ἀνδρῶν πάντα Ισραηλ ἑπτὰ χιλιάδας
ἄνδρας, οἵτινες οὐκ τὰ γόνατα ἃ οὐκ ἔκαμψαν ἀνδρῶν πάντα γόνατα
ἔκαμψαν γόνυ γόνυ τῇ Βααλ, καὶ πᾶν ἃ οὐκ ὤκλασαν γόνυ τῷ
τῇ Βάαλ. στόμα ὃ οὐ προσεκύνησεν Βααλ καὶ πᾶν στόμα ὃ οὐ
αὐτῷ προσεκύνησεν αὐτῷ

αὐτῷ] αὐτῇ 127

There is an interesting case in 3 Kgdms 19:18. This passage is quoted in

the New Testament in Rom 11:4. Rahlfs, in his analyses of the Lucianic
text, held the assumption that agreements between the Lucianic text
and the New Testament are not witnesses to an old text, but that the
New Testament has influenced the Lucianic tradition. In this way he also
explained the identical words in the quotation in Rom 11:4 and in the Luci-
anic text of 3 Kgdms 19:18. Yet Rom 11:4 has τῇ Βααλ. This word cannot be
explained as having influenced the Lucianic text, because Rom 11:4 is the

45 A. Dillman, Über Baal mit dem weiblichen Artikel (Monatsberichte der Königlichen
preussischen Akademie der Wissenschaften, Berlin, 1881), 601–620, had already proffered
this postulate and it remains the most probable explanation.
90 siegfried kreuzer

only ­occurrence of Baal in the New Testament and it would be impossible

to assume that this quotation would have produced all the occurrences
of ἡ Βααλ in the Septuagint. Rather, τῇ Βααλ in Rom 11:4 proves that not
only is this expression old, but the whole quotation uses the old textual
form. This form is the Antiochene text, which in this case again represents
the Old Greek while, on the other hand, the differences in the text of
Vaticanus must go back to a revision that not only changed the reading
of the name of Baal but other words as well. Unfortunately, while Rahlfs
discussed the other differences in this verse,46 he did not mention ἡ Βααλ,
and it is also missing at 3 Kgdms 19:18 in the apparatus of the Handaus-
gabe of 1935. Anyway, the expression τῇ Βααλ in the Antiochene text of 3
Kgdms 19:18 shows that at this place the Antiochene text represents the
old text, i.e. Old Greek, while B has the revised text.47

3.2.2. The Rendering of ‫הטוב בעיניך‬

As mentioned above (in 2.3.), soon after Barthélemy’s Les Devanciers
d’Aquila with its discovery of the kaige recension and its new evaluation
of the Antiochene text, Sebastin P. Brock delivered a paper with the tell-
ing title “Lucian redivivus”.48 In this paper he accepted the kaige recen-
sion, but defended the traditional view of an extensive Lucianic revision.
One of his (rather few) examples was the rendering of the expression
‫ הטוב בעיניך‬in the kaige and in the non-kaige sections. Brock observed
that in the Lucianic text the expression was translated with τὸ ἀρεστὸν
ἐνώπιόν σου, while in Codex Vaticanus there are two renderings of it: in
the kaige section (“Palestinian text”) it is τὸ ἀγαθὸν ἐν ὀφθαλμοῖς σου, and
in the non-kaige section one finds ἀγαθόν ἐνώπιόν σου. Brock rightly said
that the character of the Lucianic text is the same in the kaige and in
the non-kaige sections. But in a strange turn of the argument he did not
discuss the differences in the Vaticanus, but postulated that Vaticanus is
the Old Greek in the non-kaige section and that therefore—and because
of its identical character—the Antiochene text must be secondary, i.e.
the result of Lucian’s revision, in the kaige section as well. Expressed as a
table, his idea is the following:

46 Rahlfs, Lucians Rezension, 251.

47 Interestingly, there is also a case where the opposite is true. In 4 Kgdms 21:3, Codex
Vaticanus has kept ἡ Βααλ, while the Antiochene text has the masculine form. This shows
that Codex Vaticanus sometimes may have kept an older expression, even within a kaige
48 Brock, “Lucian redivivus”.
b or not b? 91

1 Kgdms 2 Kgdms (Pal.)

‫הטוב בעיניך‬ ‫הטוב בעיניך‬
B τὸ ἀγαθόν ἐνώπιόν σου τὸ ἀγαθὸν ἐν ὀφθαλμοῖς σου
Ant τὸ ἀρεστὸν ἐνώπιόν σου τὸ ἀρεστὸν ἐνώπιόν σου

Yet the real problem is the difference within Vaticanus. The difference
shows that there must be two levels. Barthélemy’s identification of the
kaige recension showed that the text in the kaige section belongs to
the revision and that it is the younger text, while the Antiochene text is
the older base text of that revision. If, as Brock had rightly maintained, the
character and the age of the Lucianic text is the same in both sections, it
is old in the non-kaige section as well. Barthélemy’s insights and analyses
lead to the following situation:

1 Kgdms 2 Kgdms 19:38(39)

and Pal. Throughout
Hebrew ‫הטוב בעיניך‬ ‫הטוב בעיניך‬
Ant τὸ ἀρεστὸν ἐνώπιόν σου τὸ ἀρεστὸν ἐνώπιόν σου
B (non-kaige) τὸ ἀγαθόν ἐνώπιόν . . .
B (Pal./kaige) τὸ ἀγαθὸν ἐν ὀφθαλμοῖς σου

The only question is: Is the text of Vaticanus in the non-kaige section
even older (and closer to the Old Greek) than the Antiochene text, or is it
younger, i.e. between the Antiochene tradition and the kaige recension?
As a look at the texts clearly shows, the expression τὸ ἀγαθόν ἐνώπιόν σου is
halfway between the Antiochene text and kaige. It has maintained ἐνώπιόν
σου but it has changed τὸ ἀρεστὸν into τὸ ἀγαθόν, which is formally closer
to the Hebrew. In other words, Brock’s example confirms Barthélemy’s
view: the Antiochene text preserves the oldest text and is close to the
Old Greek, while the kaige text is a later and formalistic revision of the
Greek text towards the Hebrew reference text. And, interestingly, even
the non-kaige sections in Codex Vaticanus show a—somewhat milder—
Hebraising revision.

3.2.3. Two Examples of Whole Verses

For further illustration we look at some verses in 2 Sam 4, comparing the text
of Codex Vaticanus, the Antiochene text, the MT and the text of 4QSama.
92 siegfried kreuzer

2 Sam 4:2
MT Line B Anted
‫ּוׁש ֵנ֣י ֲאנָ ִ ׁ֣שים‬ ְ 1 καὶ δύο ἄνδρες καὶ δύο ἄνδρες
‫דּודים‬֣ ִ ְ‫ָׂש ֵ ֽרי־ג‬ 2 ἡγούμενοι συστρεμμάτων ἡγούμενοι συστρεμμάτων
‫ָהי֪ ּו‬ 3 τῷ Μεμφιβόσθε τῷ Μεμφιβόσθε
‫ן־ׁש ֟אּול‬ ָ ‫ֶב‬ 4 υἱῷ Σαούλ· υἱῷ Σαούλ·
‫ֵׁש ֩ם ָה ֶא ָ֙חד ַ ּֽב ֲע ָ֜נה‬ 5 ὄνομα τῷ ἑνὶ Βαανά, ὄνομα τῷ ἑνὶ Βαναία,
‫וְ ֵ ׁ֧שם ַה ֵּׁש ִנ֣י‬ 6 καὶ ὄνομα τῷ δευτέρῳ καὶ τῷ δευτέρῳ ὄνομα
‫ֵר ָ֗כב‬ 7 Ῥηχάβ, Ῥηχάβ,
‫ְּב ֵנ֛י ִר ּ֥מֹון‬ 8 υἱοὶ Ῥεμμὼν υἱοὶ Ῥεμμὼν
‫ַה ְּב ֶ ֽאר ִ ֹ֖תי‬ 9 τοῦ Βηρωθαίου τοῦ Βηρωθαίου
‫ִמ ְּב ֵנ֣י ִבנְ יָ ִ ֑מן‬ 10 ἐκ τῶν υἱῶν Βενιαμείν, ἐκ τῶν υἱῶν Βενιαμίνd,
‫ם־ּב ֵא ֔רֹות ֵּת ָח ֵ ׁ֖שב‬ ְ ַ‫ִ ּ֚כי ּג‬ 11 ὅτι Βηρὼθ ὅτι καὶ Βηρὼθ
‫ל־ּבנְ יָ ִ ֽמן׃‬
ִ ‫ַע‬ 12 ἐλογίζετο τοῖς υἱοῖς ἐλογίζετο τοῖς υἱοῖς
13 Βενιαμείν. Βενιαμίν.

Qumran: 4QSama
‫ושני אנשים שרי גדודי]ם למפיבשת ֯בן֯ שאול ֗ש[ם הא]חד‬2[
‫בנימ[ין] ֯כי֯ גם‬
֯ ‫הבארתי מבני‬
֯ ‫[בענב ושם השני רכב] בני רמון‬
‫גת[י]ם [ויה]י֯ ו֯ שם‬
֯ ‫ו]י֯ ברחו הברתים‬3 ‫[בארות תחשב] ֯ע ֯ל בנימי[ן‬
As in all these chapters (of the non-kaige section), this verse clearly shows
the closeness of the texts of B and the Antiochene text. This demonstrates
that there were not two different translations (or “Septuagints”), but one
translation (i.e. the original Septuagint, the so-called Old Greek), which
was revised. Besides the general agreements, this is shown by the fact that
there are even agreements of B and the Antiochene text against the MT
such as Μεμφιβόσθε in line 3 and τοῖς υἱοῖς in line 12. Interestingly, the
first agreement Μεμφιβόσθε has its counterpart in the Qumran text, which
shows that it belongs to the Vorlage and not only to the translation. The
second agreement is harder to judge. Most probably it also goes back to
the Hebrew, because it can be explained more easily in Hebrew than in
Greek, as either an addition or an omission because of “homoioarkton”/
‫בני בנימין‬. Beyond that ‫ ֯ע ֯ל‬and the words before it in 4QSama are evi-
dently reconstructed according to MT.
Yet there are also small but significant differences. In line 5 the name
Βαανά is closer to MT than the Antiochene text with Βαναία, and in line 6
the word order in B agrees with MT. In both cases there is no real reason
that the Antiochene text should have changed the text represented by B.
In both cases the B text has no problem and the Antiochene text is not
really better Greek. On the other hand, both differences can be explained
b or not b? 93

as adaptations to the Hebrew text. Therefore, it is highly probable that

the text in B reflects an adaptation, which means that it has undergone
a—mild—revision towards a Hebrew reference text.

2 Kgdms 4:5
MT Line B Anted
‫ֽי־ר ּ֤מֹון‬
ִ ‫וַ ּיֵ֙ ְל ֜כּו ְּב ֵנ‬ 1 Καὶ ἐπορεύθησαν υἱοὶ Καὶ ἐπορεύθησαν οἱ υἱοὶ
֙ ִ ‫ַה ְּב ֵ ֽאר‬ 2 Ῥεμμὼν τοῦ Βηρωθαίου Ῥεμμὼν τοῦ Βηρωθαίου,
‫ּוב ֲע ָ֔נה‬ַ ‫ֵר ָכ֣ב‬ 3 Ῥεκχὰ καὶ Βαὰμ Ῥηχὰβ καὶ Βαναία
4 ὁ ἀδελφὸς αὐτοῦ,
‫אּו‬֙ ֹ ‫וַ ּיָ ֙ב‬ 5 καὶ εἰσῆλθον καὶ εἰσῆλθον
‫ְּכ ֣חֹם ַהּי֔ ֹום‬ 6 ἐν τῷ καύματι τῆς ἡμέρας ἐν τῷ καύματι τῆς ἡμέρας
‫ל־ּבית‬֖ ֵ ‫ֶא‬ 7 εἰς οἶκον εἰς τὸν οἶκον
‫ִ ֣איׁש ּ֑בֹ ֶׁשת‬ 8 Μεμφιβόσθε, Μεμφιβόσθε,
‫וְ ֣הּוא ׁש ֵֹ֔כב‬ 9 καὶ αὐτὸς ἐκάθευδεν καὶ αὐτὸς ἐκάθευδε
‫ֵ ֖את ִמ ְׁש ַ ּ֥כב‬ 10 ἐν τῇ κοίτῃ
‫ַ ֽה ָּצ ֳה ָ ֽריִ ם׃‬ 11 τῆς μεσημβρίας. τὸ μεσημβρινόν.

Qumran: Not extant.

In this verse, one can again observe that both Greek text forms gener-
ally agree and therefore go back to one single translation that was revised;
there are not enough differences to understand the two versions as two
independent translations.49 There are fewer differences than in the kaige
sections, yet they are clear and significant. There is the difference regard-
ing the article in lines 1 and 7, and in lines 4 and 11 there are differences
regarding addition and omission. According to the old theories about a
Lucianic redaction, Lucian would have added the articles and he would
have added the explanatory words in line 4. On the other hand, he would
have deleted the words in line 11. This was explained by Rahlfs and many
others as a trait of Lucian, i.e. that he worked irregularly and even con-
tradictorily. But this is only an assumption and does not explain why he
should have done this. Again, as in the kaige section, one should test the
other possibility. If one allows the Antiochene text to present the older
text, one arrives at a consistent explanation. Each of the differences can
be explained as an adaptation to the Hebrew reference text, which in
this case almost exactly matched the proto-MT text. The identification of
Benaia as “his brother” (line 4) is missing in the Hebrew text as well.

49 Cf. Tov, “Lucian and Proto-Lucian,” 101–113: The Antiochene text contains “the LXX”
or “a Septuagint.”
94 siegfried kreuzer

The place of Memphiboste’s rest (line 11) is an exact rendering of the

MT. The Hebrew Vorlage of the Antiochene text may have been shorter,
but it is probably only a freer rendering of the same Hebrew words by
naming the circumstance (“at noon”) of his sleeping, while B makes a
word-for-word translation. In any case, B makes an isomorphic adaptation
towards its reference text. The deletion of the articles in lines 1 and 7 fits
exactly into this picture and again demonstrates the isomorphic character
of the revision.50
The strange form of the names in line 3, both Ρεκχα and Βααμ, is a spe-
cific phenomenon of Codex Vaticanus, which in this case is only followed
by a few manuscripts (and for Βααμ also by the Ethiopian version, which
confirms a rather late date). The manuscripts show some variety and inse-
curity with these names.51 Ρεκχα could be explained by a Hebrew form
‫ רככ‬instead of ‫רכב‬. Basically, it looks like a scribal mistake, especially
since in v. 2 line 6 there is the correct form in Codex Vaticanus. More
amazing than the mistake is that it has not been corrected.
We could continue to analyse many more verses from the non-kaige
sections with similar results,52 but for the sake of space this must suffice.
The examples showed that with the text of Codex Vaticanus and the Anti-
ochene text there are two text forms which inseparably belong together
and are very close to the Old Greek. At the same time, there are many
differences, differences that are not scribal mistakes or other uninten-
tional corruptions, but which must be explained as intentional reworking
according to specific rules. The general trend of this reworking is a closer
adaptation of the Greek text towards the Hebrew reference text, mainly
in a formalistic, isomorphic way. This intention is the same as represented
in the kaige revision, i.e. bringing the text closer to its Hebrew reference
text and making it reflect the holy Hebrew text not only in its content,
but also in its form.

50 Interestingly, both text forms read the name Memphiboste and not Isboste, which
confirms that this is the Old Greek reading. The Hebrew equivalent is found in 4QSama,
cf. above v. 2, which confirms that the Vorlage of the Old Greek was not proto-MT but a
slightly different text form. Yet one should notice that in 4:4 MT also reads Mephiboshet/
Memphiboste (see also v. 7), while the Antiochene text has Memphibaal, and that in 4:8
MT again reads Ishboshet. But these are questions of the plurality of the Hebrew text
51  See the apparatus in Brooke-McLean.
52 This analysis is being undertaken in a research project at Wuppertal sponsored by
the Deutsche Forschungsgemeinschaft.
b or not b? 95

In comparing the text forms and testing the possible directions of

the change, it turns out that the text represented by Codex Vaticanus is
the one which has been revised, while the Antiochene text is closest
to the Old Greek.
As just stated, the formalistic Hebraising revision is not as strong as
in the kaige sections. One may ask if it represents an earlier stage in the
development. Considering the general development of the Septuagint
from a more liberal to a more literal translation, this seems logical. Yet
one must consider that we are dealing with a revision, not with the origi-
nal translation. The idea of greater closeness to the Hebrew in terms of
both content and structure may as well be the result of a cross-influence
from the kaige recension. Therefore, as long as we don’t have indications
of the chronology, it seems better to avoid a term such as pre-kaige and
refer to the revision by its characteristic: the text of Codex Vaticanus at
least in the non-kaige sections of the historical books represents a semi-
kaige text.

4. Conclusions: B or not B?

(1)  Codex Vaticanus (B) has become the most important single manu-
script of the Septuagint as a result of historical circumstances, as it
was the basic manuscript for the diplomatic editions from around
1600 until the 20th century, and as it remains the dominant manu-
script in the critical editions. It is probably still the most important
single manuscript, yet it has its worth no longer unto itself, but only in
the context of the other manuscripts and the other textual traditions
(including the Hebrew texts—especially from Qumran—on the one
hand, and the daughter translations—especially the Old Latin, but
also the Sahidic and the Syriac translation—on the other hand). Also
the quotations, both by Josephus and in the New Testament and by
the early Christian writers, have proven to be more important than
had been accepted in the first half of the 20th century.
(2) Most important is the fact that B consists of different text types. At
least the kaige sections and the non-kaige sections as they were iden-
tified by Thackeray and confirmed and interpreted by Barthélemy
give evidence of different text types. This clearly shows that B is a
mixed codex, i.e. a codex with different types of text. Since Barthé-
lemy’s study it is clear that the kaige sections represent a revised
text that is not the Old Greek. This result has been confirmed by the
96 siegfried kreuzer

a­ forementioned aspects, which demonstrate the early Jewish under-

standing of scripture and the related hermeneutics as the background
of the kaige recension. Yet the non-kaige sections of Codex Vaticanus
also show evidence of a Hebraising revision and therefore are not the
Old Greek, although much closer to it. The Hebraising recension of
the non-kaige sections—although milder—has a similar hermeneuti-
cal background to the kaige sections, i.e. it intends a formalistic adap-
tation towards the authoritative Hebrew reference text. Because it is a
more moderate revision, it can be called the semi-kaige recension. This
means that in both, the kaige sections and the non-kaige sections—at
least of the historical books, but probably also other books, especially
in the prophetic books—Codex Vaticanus represents revised texts,
although revised to different degrees.
(3) The kaige sections in B demonstrate that there had been a Hebraising
revision not only in the Minor Prophets (cf. Naḥal Ḥever scroll), but
also in large parts of the Septuagint—if not in all of it. Codex Vati-
canus is therefore important not only where it is relatively close to
the Old Greek, but also—and probably even more so—where its text
is secondary, as this shows an important and highly influential phase
of the transmission of the Septuagint.
(4) B as a manuscript from the 4th century is older than the other extant
codices and older than the manuscripts of the Antiochene texts, yet
one has to keep in mind that it is not older than the text of the Anti-
ochene fathers and that it is younger than Josephus, the Old Latin, the
Sahidic translation and (although they are in Hebrew) the Qumran
texts. This is the field where the place of Codex Vaticanus has to be
defined today.
(5) So, B or not B? The answer to this question is a clear “yes” and a clear
“no”: yes, in that Codex Vaticanus is still one of the most important
manuscripts of the Septuagint; no, in that it should not be understood
in the old way, as the most important witness of the original Sep-
tuagint, the so-called Old Greek, but rather in the way that it clearly
shows different stages of the transmission of the Septuagint and that
it allows a glimpse at the factors which accompanied this transmis-
sion and the forces which shaped it.
Übersetzungstechnik und Textkritik in den
Esdrasbüchern: Hendiadyoin, Doppelübersetzungen
und Wiederholungsvariationen in 1 Esdr

Dieter Böhler SJ

1. Einführung

In seinem Sendbrief vom Dolmetschen schreibt Martin Luther: „Denn wer

dolmetzschen wil, mus grosse vorrath von worten haben, das er die wol
koenne, wenn eins an allen orten nicht lauten will“1. Was Luther anstrebt,
ist eine durch und durch zielsprachenorientierte Übersetzung. Äquivalen-
tenkonstanz ist hier kein Ideal, vielmehr soll der je im deutschen Kon-
text passende Ausdruck gefunden werden, um einen flüssigen deutschen
Text zu erreichen, der jedes Hebraisieren vermeidet. Das Gegenmodell
zu dieser Übersetzungsphilosophie verwirklichen Buber und Rosenzweig,
die nicht mit Luther wie „die můtter ym haus und der gemeine man“2
reden wollen, sondern etwa durch Äquivalentenkonstanz ausdrücklich
das hebräische Kolorit übertragen und wiedergeben wollen. Beides hat je
nach Verwendungszweck der Übersetzung seine Berechtigung.
In der Septuaginta findet sich die ganze Bandbreite zwischen den
­beiden Extremen. Ja, es sind auch die Extreme selbst verwirklicht. So gibt
es bekanntermaßen von Daniel eine alte recht elegante griechische Über-
setzung, die meist als „Septuaginta“ oder „Old Greek“ bezeichnet wird. Ihr
wurde später eine Wort-für-Wort-Übersetzung an die Seite gestellt, die
in der Tradition Theodotion zugeschrieben wurde. Dasselbe gilt für die
beiden Esdrasbücher der LXX: Der älteren um 130 v.Chr. in bestem
Griechisch angefertigten Übersetzung 1 Esdr wurde später, im 1. nach-
christlichen Jahrhundert, die pedantische Interlinearübersetzung 2 Esdr
entgegengestellt. Obwohl im Fall der Esdrasbücher die literarischen Dif-
ferenzen zwischen den beiden Bucheditionen größer sind als im Falle der
Danielfassungen, handelt es sich doch weitgehend noch um zwei Über-
setzungen ein und desselben Buches. Ja es sieht geradezu so aus, als ob

1 M. Luther, „Sendbrief vom Dolmetschen,“ in D. Martin Luthers Werke (kritische

Gesamtausgabe 30/2; Weimar: Hermann Boehlaus Nachfolger, 1909), 627–646; 639.
2 Ibid., 637.
98 dieter böhler sj

jeweils ein ­Späterer die völlig zielsprachenorientierte Übersetzungsphilo-

sophie des alten Daniel und des 1 Esdr korrigieren oder ergänzen wollte
durch eine nun dezidiert ausgangssprachenorientierte Übertragung.
Die Durchführung zweier Extrempositionen von Übersetzungs-
philosophie am selben Text machen 1 und 2 Esdr zu einem sehr interes-
santen Studienobjekt. Die Übersetzungstechnik von 2 Esdr, die noch nicht
aquilagleich, aber dahin unterwegs ist, macht ihn für die Textkritik sehr
leicht handhabbar. Man muss nur das Griechische Wort für Wort ins Hebrä-
ische und Aramäische rückübertragen. Der späte Textzeuge zeigt sich dann
freilich als weitgehend MT-gleich und damit wenig interessant.
Das Gegenteil gilt für die ältere Version 1 Esdr. In den Textpartien,
wo diese Fassung sich mit Esr-Neh MT (und damit 2 Esdr) überlappt3,
bezeugt 1 Esdr zahllose größere und kleinere Varianten. Doch ist diese
zielsprachenorientierte Übersetzung textkritisch oft schwer auswertbar,
da eine mechanische Rückübersetzung unmöglich ist. Vielmehr muss in
jedem einzelnen Fall eingeschätzt werden, ob eine gut griechische Äqui-
valentenwahl einfach nur der Zielsprachenorientierung des Übersetzers
geschuldet ist, oder eben doch einen von MT abweichenden Vorlagetext
In vielen Fällen sind sich Textkritiker nicht darüber einig, welchen Text
1 Esdr bezeugt. Es ist bekannt, dass 1 Esdr, außer im Falle bestimmter
technischer Termini, keine Äquivalentenkonstanz pflegt. In der FS für
H.-J. Fabry habe ich etwa gezeigt, dass 1 Esdr ‫ ַהּכ ֵֹהן‬als „der Priester“ aber
je nach Kontext auch als „der Hohepriester“ übersetzen kann4. Das ist bei
ihm eine Frage der Übersetzungstechnik, nicht der Vorlagenvariante.
Es bedarf aber noch vieler Einzeluntersuchungen, um zu sehen, wie
sich der Übersetzer von 1 Esdr in bestimmten Situationen verhält, um ent-
scheiden zu können, ob eine bestimmte Äquivalentenwahl eine Variante
bezeugt oder nicht. Eine solche Einzeluntersuchung will der vorliegende
Versuch bieten.
Es ist in der Fachliteratur bekannt, dass 1 Esdr bisweilen einen einfa-
chen hebräischen Ausdruck durch ein griechisches Hendiadyoin wieder-
gibt, wenn er meint, damit die Bedeutungsbreite des hebräischen Wortes
besser abzudecken. Textkritische Folgerungen darf man daraus nicht

3 Esr 110 + Neh 8:1–13a.

4 D. Böhler, „Literarischer Machtkampf. Drei Ausgaben des Esrabuches im Streit um
das wahre Israel und um die Legitimation von Herrschaft,“ in Juda und Jerusalem in der
Seleukidenzeit (FS Fabry; BBB 159; Hg. U. Dahmen und J. Schnocks; Göttingen: Vanden-
hoeck & Ruprecht, 2010), 125–145; 130–132.
übersetzungstechnik und textkritik in den esdrasbüchern 99

z­ iehen. Wesentlich verschieden, aber doch benachbart, ist das Phänomen

der Doppelübersetzung. Beim Hendiadyoin geht es um die Kombina-
tion von zwei mehr oder weniger synonymen Ausdrücken. Mit Doppel-
übersetzung löst unser Übersetzer bisweilen das Problem, dass er dem
semitischen Vorlagenwort zwei semantisch völlig verschiedene griechi-
sche Äquivalente zuordnen kann. Manchmal wählt er beide. Eine solche
doppelte Wiedergabe bezeugt dennoch nur ein einfaches Vorlagewort.
Hendiadyoin und Doppelübersetzung werden wir uns gleich zuwenden.
Näher untersuchen möchte ich aber das Verhalten des Übersetzers von
1 Esdr bei Wortwiederholungen. Wenn im hebräischen Text zweimal hin-
tereinander oder mit geringem Abstand dasselbe Wort steht, wird 2 Esdr
das genau so wiederholen. 1 Esdr tut das meist nicht. Wann haben wir nur
Variation in der Äquivalentenwahl, und wann liegt die Bezeugung einer
Variante vor? Das ist die Frage, der diese Untersuchung nachgeht.

2. Die Übersetzungsphilosophien von 1 und 2 Esdr im Vergleich

Zu Beginn will ich ganz knapp an einem konkreten Beispielsatz in die

beiden Übersetzungsphilosophien von 1 und 2 Esdr einführen und die
jeweilige textkritische Auswertbarkeit darstellen.

1 Esdr 5:46 2 Esdr 3:1 Esr MT 3:1

ἐνστάντος δὲ τοῦ ἑβδόμου καὶ ἔφθασεν ὁ μὴν ὁ ‫יעי‬
ִ ‫וַ ּיִ ּגַ ע ַהח ֶֹדׁש ַה ְּׁש ִב‬
μηνὸς καὶ ὄντων τῶν ἕβδομος καὶ οἱ υἱοὶ Ισραηλ ‫ּובנֵ י יִ ְׂש ָר ֵאל ֶּב ָע ִרים ס‬
υἱῶν Ισραηλ ἑκάστου ἐν ἐν πόλεσιν αὐτῶν καὶ ‫וַ ּיֵ ָא ְספּו ָה ָעם ְּכ ִאיׁש ֶא ָחד‬
τοῖς ἰδίοις συνήχθησαν συνήχθη ὁ λαὸς ὡς ἀνὴρ
ὁμοθυμαδὸν εἷς

Der hebräische Text stellt drei Elemente in Parataxe zusammen: „und

genaht hatte sich,“ „und die Kinder Israel waren“ „und sie versammelten
sich“. 2 Esdr imitiert das genau so: 1. καὶ ἔφθασεν, 2. καὶ οἱ υἱοί, 3. καὶ συνήχθη.
1 Esdr dagegen löst die Dreierparataxe auf, macht aus den ersten beiden
Elementen je einen Genitivus absolutus und ordnet sie dem letzten Ele-
ment als dem Hauptsatz unter. Dadurch verschwindet das letzte Waw, das
mittlere Waw wird zu einem καί und das erste zu einem nachgestellten δέ.
Jedes Waw ist bezeugt, jedes auf eigene Weise. Das ist wichtig: 1 Esdr ist
eine sehr genaue Übersetzung, alles andere als frei oder paraphrastisch.
Aber sie ist nicht mechanisch wie 2 Esdr.
100 dieter böhler sj

Die im Hebräischen zwingende Anordnung Substantiv-Adjektiv ‫ַהח ֶֹדׁש‬

ִ ‫ ַה ְּׁש ִב‬wird von 2 Esdr imitiert: ὁ μὴν ὁ ἕβδομος. 1 Esdr hat öfter einmal
die im Griechischen wohlklingende Wortfolge Adjektiv-Substantiv5: τοῦ
ἑβδόμου μηνὸς.
Das hebräische ‫ ְּכ ִאיׁש ֶא ָחד‬gibt 2 Esdr, wenig überraschend, mit ὡς
ἀνὴρ εἷς wieder. Typisch für 1 Esdr ist die Wahl eines gut griechischen idio-
matischen Ausdrucks wie ὁμοθυμαδόν. Die Vorlage weicht von MT nicht
ab, aber wir hätten sie aus 1 Esdr nicht rekonstruieren können. Bei 2 Esdr
könnten wir das. Für das hebräische ‫ ֶּב ָע ִרים‬bietet 2 Esdr ἐν πόλεσιν αὐτῶν
und bezeugt damit die Variante ‫יהם‬ ֶ ‫ ְּב ָע ֵר‬, wie sie an der Parallelstelle
Neh 7:72 auch in MT steht. Diese Lesart ‫יהם‬ ֶ ‫ ְּב ָע ֵר‬liegt wohl auch dem
idiomatischen Ausdruck ἐν τοῖς ἰδίοις zu Grunde, der einfach „zu Hause“
bedeutet6 und die hebräische Vorlage höchst treffend wiedergibt.
Das ‫ ָה ָעם‬der letzten Zeile, das auch 2 Esdr mit ὁ λαός bezeugt, hat
1 Esdr aber nicht gelesen. Spurlos würde er es nicht verschwinden lassen.
Die „Kinder Israel“ bleiben Subjekt.
Das kurze Beispiel von 1 Esdr 5:46 zeigt, dass 1 Esdr eine sehr genaue
aber eben zielsprachenorientierte Übersetzung ist, deren Auswertung für
die Textkritik deutlich schwerer ist als die der Interlinearübersetzung
2 Esdr. Nach dieser allgemeinen Einführung nun zu der speziellen Frage von
Hendiadyoin, Doppelübersetzung und Variation bei Wortwiederholungen!

3. Hendiadyoin

Hendiadyoin ist die Wiedergabe eines einfachen Ausdrucks der Vorlage

durch zwei in etwa synonyme Äquivalente, die die Reichweite des Vorlage-
ausdrucks gemeinsam besser abdecken, als jeder für sich allein es könnte7.
Ein Beispiel findet sich in 1 Esdr 9:88. Esra fordert die in illegitime

5 Z. Talshir, 1 Esdras. From Origin to Translation (SCS 47; Atlanta: Scholars Press, 1999),
208; D. Böhler, „ ‚Treu und schön‛ oder nur ‚treu‛? Sprachästhetik in den Esrabüchern, in
Im Brennpunkt: Die Septuaginta, Bd. 3: Studien zur Theologie, Anthropologie, Ekklesiologie,
Eschatologie und Liturgie der Griechischen Bibel (BWANT 174; Hg. H.-J. Fabry und D. Böhler;
Stuttgart: Kohlhammer, 2007), 97–105; 101–102.
6 Vgl. Est 5:10; 6:12; Lk 18:28; Joh 19:27; Apg 21:6.
7 R. W. Klein, Studies in the Greek Texts of the Chronicler (unpubl. Diss., Harvard Uni-
versity, Cambridge, 1966), 273: „Numerous double translations and the use of hendiadys,
especially in 1 Esdras, show the strenuous efforts . . . by the translator to present the full
meaning of his Vorlage.“
8 Talshir, 1 Esdras. From Origin to Translation, 241; idem, I Esdras. A Text Critical Com-
mentary (SCS 50; Atlanta: Society of Biblical Literature, 2001), 464–465.
übersetzungstechnik und textkritik in den esdrasbüchern 101

­Mischehen verstrickten Judäer auf, Gott die Ehre zu geben und ein
Bekenntnis abzulegen. Im MT heißt das:
Esr 10:11: ‫יכם‬
ֶ ‫י־אב ֵֹת‬
ֲ ‫ֹלה‬
ֵ ‫תֹודה ַליהוָ ה ֱא‬
ָ ‫וְ ַע ָּתה ְּתנּו‬
ָ bedeutet als Substantiv zu ydh hif. soviel wie „loben“, „preisen“.
Daher übersetzt 2 Esdr:
καὶ νῦν δότε αἴνεσιν κυρίῳ τῷ θεῷ τῶν πατέρων ὑμῶν.
ָ fungiert aber auch als Substantiv zu ydh hitp. „bekennen“ wie in
Jos 7:19. Daher wählt 1 Esdr 9:8 ein Hendiadyoin als Äquivalent für den
einfachen hebräischen Begriff:
καὶ νῦν δότε ὁμολογίαν δόξαν τῷ κυρίῳ θεῷ τῶν πατέρων ἡμῶν
Mit „Bekenntnis“ und „Verherrlichung“ hat er beide Gebrauchsweisen
von ‫תֹודה‬
ָ eingefangen. Wenn wir nicht aus vielen anderen Beispielen
wüssten, dass Hendiadyoin eine Technik unseres Übersetzers ist, hätten
wir hier wegen der asyndetischen Zusammenstellung den Verdacht einer
Glosse von zweiter Hand9.

4. Doppelübersetzung

Während die Stilfigur des Hendiadyoin einen Begriff mit einem Doppel-
ausdruck aus zwei annähernd synonymen Ausdrücken wiedergibt, um ein
einziges semantisches Feld besser abzudecken, verstehe ich unter einer
Doppelübersetzung die Kombination von zwei Möglichkeiten, die seman-
tisch miteinander nichts zu tun haben. Wenn ein deutscher Übersetzer
sich beim englischen Wort „ear“ nicht entscheiden kann, ob es um das
Hörorgan geht oder eine Ähre („ear of corn“) und deswegen einfach „die
Ähre und das Ohr“ schreibt, dann ist das eine Doppelübersetzung, d.h.
zwei vollkommen verschiedene Wiedergaben eines einzigen Wortes.
In Esr 4:8, 9, 17 schreiben Rechum und Schimschai dem Perserkö-
nig einen Beschwerdebrief gegen die heimgekehrten Judäer. Rechum
trägt den Titel ‫ל־ט ֵעם‬
ְ ‫„ ְּב ֵע‬Berichterstatter,“ Schimschai den Titel ‫ָס ְפ ָרא‬
„Schreiber“ oder „Sekretär“.

9 Syndetisch und asyndetisch: ‫( ֵה ִביא‬2 Chr 36:7) λαβὼν . . . καὶ ἀπενέγκας (1 Esdr 1:39);
‫( ֵה ִביא‬2 Chr 36:18) ἀναλαβόντες ἀπήνεγκαν (1 Esdr 1:51); ‫( ְל ַה ֵּלל‬Esr 3:10) ὑμνοῦντες . . . καὶ
εὐλογοῦντες (1 Esdr 5:57); ‫( וְ ִׁש ְב ַעת יָ ֲעט ִֹהי‬Esr 7:14) καὶ τοῖς ἑπτὰ φίλοις συμβουλευταῖς (1 Esdr
102 dieter böhler sj

Esr 4:17: ‫ל־ט ֵעם וְ ִׁש ְמ ַׁשי ָס ְפ ָרא‬

ְ ‫ל־רחּום ְּב ֵע‬
ְ ‫ִּפ ְתגָ ָמא ְׁש ַלח ַמ ְל ָּכא ַע‬
2 Esdr scheint den Titel ‫ל־ט ֵעם‬
ְ ‫„ ְּב ֵע‬Berichterstatter“ an allen drei Stellen
als Eigennamen verstanden zu haben und transkribiert ihn daher einfach
nur. Dabei bleibt offen, ob Baaltam ein Zweitname Rechums ist oder eine
andere Person neben ihm.
καὶ ἀπέστειλεν ὁ βασιλεὺς πρὸς Ραουμ Bααλταμ καὶ Σαμσαι γραμματέα
1 Esdr dagegen bietet eine doppelte Wiedergabe ein und desselben ara-
mäischen Ausdrucks:
τότε ἀντέγραψεν ὁ βασιλεὺς Ραούμῳ τῷ γράφοντι τὰ προσπίπτοντα καὶ Βεελτεέμῳ
καὶ Σαμσαίῳ γραμματεῖ (1 Esdr 2:21)
Indem der Übersetzer zwischen den übersetzten Titel und den transkri-
bierten Eigennamen auch noch ein καί setzte, hat er aus ein und dem-
selben aramäischen Ausdruck zuerst einen Titel Rechums und dann
zusätzlich eine Person neben Rechum gemacht.
Der aramäische Ausdruck ‫ל־ט ֵעם‬ ְ ‫„ ְּב ֵע‬Berichterstatter“ fand sich bereits
in Esr 4:8 und 4:9. Der Übersetzer von 1 Esdr hatte sich im ersten Fall
für die Interpretation als Eigennamen entschieden und (καί) Βεελτέεμος
transkribiert (1 Esdr 2:15). Im zweiten Fall entschied er sich für die Über-
setzung „Berichterstatter“ und schrieb ὁ τὰ προσπίπτοντα (2:16). Im dritten
Fall Esr 4:17 kombiniert er die beiden Möglichkeiten zu einer Doppelüber-
setzung: τῷ γράφοντι τὰ προσπίπτοντα καὶ Βεελτεέμῳ (1 Esdr 2:21).10
Man würde die Doppelübersetzung an dieser Stelle einem Glossator
zuschreiben, wenn nicht der Übersetzer von 1 Esdr selbst zuvor schon
beide Möglichkeiten, freilich zunächst je für sich, eingesetzt hätte.
Noch deutlicher ist der folgende Fall. Auch hier setzt der Übersetzer
mal eine einfache Wiedergabe, mal eine Doppelübersetzung.
In Esr 5:5 werden die Tempelbauarbeiten der Juden durch einen per-
sischen Beamten auf ihre Rechtmäßigkeit hin überprüft. Der Erzähler
vermerkt, Gottes wohlmeinendes Auge habe auf den Ältesten der Juden
gelegen, so dass sie nicht behelligt wurden:
Esr 5:5: ‫הּודיֵ א‬
ָ ְ‫ל־ׂש ֵבי י‬
ָ ‫וְ ֵעין ֱא ָל ֲההֹם ֲהוָ ת ַע‬

10 Klein, Studies, 235f; Talshir, I Esdras. A Text Critical Commentary, 107 meint: „The
complex rendering of the title suits the nature of our translator.“ Sie lässt offen, ob ein
späterer Revisor in 2:16 die Transkription als Eigenname ersetzte und in 2:21 addierte oder
eine Doppelübersetzung des ursprünglichen Übersetzers vorliegt.
übersetzungstechnik und textkritik in den esdrasbüchern 103

2 Esdr versteht unter šby/sby hier nicht aramäisch „Älteste“ (‫) ָׂש ֵבי‬, sondern
‫„ ְׁש ִבי‬Gefangenschaft“, „Exulantenschaft“ und schreibt: τὴν αἰχμαλωσίαν
Kurz darauf schreibt der persische Beamte in Esr 5:9 dem König und
berichtet, er habe jene Ältesten befragt:
‫ֱא ַדיִ ן ְׁש ֵא ְלנָ א ְל ָׂש ַבּיָ א ִא ֵּלְך‬
In diesem Fall hat 2 Esdr sich nun für die andere Alternative, die Ältesten,
2 Esdr 5:9: τοὺς πρεσβυτέρους ἐκείνους
Der Übersetzer von 1 Esdr war in derselben Lage, was seine aramäische
Vorlage angeht. Im Fall von Esr 5:9 entscheidet auch er sich für die
1 Esdr 6:10: τότε ἐπυνθανόμεθα τῶν πρεσβυτέρων τούτων
Im Fall von Esr 5:5 allerdings wählt er die Doppelübersetzung:
1 Esdr 6:5: καὶ ἔσχοσαν χάριν ἐπισκοπῆς γενομένης ἐπὶ τὴν αἰχμαλωσίαν παρὰ τοῦ
κυρίου οἱ πρεσβύτεροι τῶν Ιουδαίων
Nur nebenbei sei bemerkt, dass für hebräisches „Auge“ ein Hendiadyoin
gewählt wird: zuerst χάριν, weil eben zielsprachenorientiert „Wohlwollen“
gemeint ist, dann aber auch noch ἐπισκοπῆς, um das hebräische Bild des
Auges nicht untergehen zu lassen.
Die Verteilung der beiden alternativen Äquivalente „Exulantenschaft“
und „Älteste“ auf den komplexen Satz 1 Esdr 6:5 schließt die Zweithand
eines bloß addierenden Glossators praktisch aus.11
Talshir kommentiert ganz zutreffend: „Unlike standard double transla-
tions, the two solutions do not simply adjoin each other; rather they are
interwoven to form one complex sentence“. Auch hier setzt der Übersetzer
wie in 1 Esdr 2:21 bei Beelteem die beiden Konzepte, die er dem doppel-
deutigen aramäischen Ausdruck entnimmt, als zwei verschiedene Größen
klar nebeneinander. Oder anders ausgedrückt: Die Doppelübersetzung
„Berichterstatter und Beelteem“ in 1 Esdr 2:21 konnte schon dort kaum
einem griechischen Glossator zugewiesen werden, weil die beiden

11 Die „Ältesten“ sind verwoben mit dem Plural des Verbs („sie hatten“) und dem
Genitiv „der Juden“. Die „Exulantenschaft“ hängt an dem „auf/über“ der Vorlage. Talshir,
I Esdras. A Text Critical Commentary, 330: „I Esd presents an ingeniously combined double
104 dieter böhler sj

einzelnen Möglichkeiten zuvor in 2:15 und 2:16 jeweils einmal schon

herangezogen worden waren. Die sicher dem Übersetzer von 1 Esdr 6:5
zuzuschreibende Doppelübersetzung „die Exulantenschaft“ und „die
Ältesten“ verstärkt die Sicherheit, dass ein solches Vorgehen, dem Über-
setzer selbst zuzuordnen ist.

5. Doppelausdrücke

Eine Gruppe von Ausdrücken steht zwischen einfachen Begriffen, die der
Übersetzer als Hendiadyoin wiedergibt, und Wortwiederholungen, die der
Übersetzer durch eine Variation von Äquivalenten überträgt. Ich meine
hebräische (oder aramäische) Doppelausdrücke, die doch nur einen einzi-
gen Begriff meinen, wie etwa „Tag für Tag“, ‫יֹום ְּביֹום‬, im Sinne von „jeden
Tag“, „täglich“. In allen vier (oder fünf) Fällen in denen MT und die Vorlage
von 1 Esdr einen solchen Doppelausdruck verwenden, gibt 2 Esdr (ebenso
wie 2 Par) ihn durch einen entsprechenden griechischen Doppelausdruck
Bei Joschijas Pascha in 2 Chr 35 (1 Esdr 1) waren alle Priester und Levi-
ten, die Sänger und Torwächter an ihren Plätzen. In 2 Chr 35:15 heißt es
von den Torwächtern:
2 Chr 35:15: ‫וְ ַהּׁש ֲֹע ִרים ְל ַׁש ַער וָ ָׁש ַער‬
Wort für Wort überträgt der Übersetzer von 2 Par nicht nur den Dop-
pelausdruck „an Tor und Tor“, sondern auch noch das wurzelgleiche
2 Par 35:15: καὶ οἱ πυλωροὶ πύλης καὶ πύλης
Ganz anders der Übersetzer von 1 Esdr! Charakteristisch ist für ihn die
Wiedergabe des hebräischen Doppelausdrucks mit einem echt griechi-
schen Äquivalent, das die hebräische Wortverdoppelung vermeidet:
1 Esdr 1:15: καὶ οἱ θυρωροὶ ἐφ᾽ ἑκάστου πυλῶνος
Die hebräischen ‫ ּׁש ֲֹע ִרים‬heißen bei 1 Esdr immer θυρωροί, in 1–2 Par und
2 Esdr ausnahmslos πυλωροί. Die Torwächter haben also als Fachtermi-
nus auch in 1 Esdr ein Standardäquivalent. Und doch zeigt sich schon an
diesem Beispiel, dass die Wahl der variierenden Äquivalente oft nicht aus-
tauschbar wäre. 1 Esdr hätte nicht umgekehrt formulieren können: „Und
die Torwächter an jeder Tür“, weil man zwar die Wächter an den Türen/
übersetzungstechnik und textkritik in den esdrasbüchern 105

Torflügeln eines Tors als Türwächter bezeichnen kann, nicht aber Tore
als Türen.
Ein ähnlicher Doppelausdruck („Tag für Tag“) findet sich in dem ara-
mäischen Text Esr 6:9:
Esr 6:9: ‫ֶל ֱהוֵ א ִמ ְתיְ ֵהב ְלהֹם יֹום ְּביֹום‬
Wenig überraschend lautet die Wort-für-Wort-Übersetzung von 2 Esdr:
ἔστω διδόμενον αὐτοῖς ἡμέραν ἐν ἡμέρᾳ
Gewählt und elegant dagegen drückt 1 Esdr denselben Gedanken aus:
1 Esdr 6:29: ἀναλίσκεσθαι καθ᾽ ἡμέραν („dass es täglich aufgewandt werde“).
Derselbe Doppelausdruck, nun aber hebräisch, steht in Esr 3:4—und zwar
gleich zweimal:
Esr 3:4: ‫ת־חג ַה ֻּסּכֹות ַּכ ָּכתּוב וְ ע ַֹלת יֹום ְּביֹום ְּב ִמ ְס ָּפר ְּכ ִמ ְׁש ַּפט‬ַ ‫וַ ּיַ ֲעׂשּו ֶא‬
‫ְּד ַבר־יֹום ְּביֹומֹו׃‬
Die griechische Übersetzung von 2 Esdr folgt wie gewohnt Wort für
2 Esdr 3:4: καὶ ἐποίησαν τὴν ἑορτὴν τῶν σκηνῶν κατὰ τὸ γεγραμμένον καὶ
ὁλοκαυτώσεις ἡμέραν ἐν ἡμέρᾳ ἐν ἀριθμῷ ὡς ἡ κρίσις λόγον ἡμέρας ἐν ἡμέρᾳ
Der Übersetzer von 1 Esdr wählt, wie im vorigen Beispiel die gut griechi-
sche Wendung καθ᾽ ἡμέραν.
1 Esdr 5:50: καὶ ἠγάγοσαν τὴν τῆς σκηνοπηγίας ἑορτήν ὡς ἐπιτέτακται ἐν τῷ
νόμῳ καὶ θυσίας καθ᾽ ἡμέραν ὡς προσῆκον ἦν.
1 Esdr bezeugt mit ὡς ἐπιτέτακται ἐν τῷ νόμῳ die hebräische Wendung
ָ ‫ ַּכ ָּכתּוב ַּב‬12.
Schwerer ist aber zunächst zu sagen, ob die Tatsache, dass 1 Esdr die
Wendung „Tag für Tag“ nur einmal hat und keine Spur ihrer Wiederho-
lung verrät, bedeutet, dass seine Vorlage in diesem Punkt abwich, oder
ob der Übersetzer die neuerliche Wiederholung einfach durch Weglassen
vermieden hat. Einfaches Weglassen ist im allgemeinen nicht die Art des

12 Die Wendung kommt in 2 Chr 25:4 und Neh 10:35, 37 vor. In Verbindung mit „Gesetz“
kann ‫ כתב‬mit ἐπιτάσσειν wiedergegeben werden, so wie bei königlichen Befehlen ‫אמר‬
mit ἐπιτάσσειν übersetzt werden kann (Esr 5:15 // 1 Esdr 6:18; ebenso DanLXX 1:18; 2:2, 46;
3:19, 20).
106 dieter böhler sj

Übersetzers von 1 Esdr. Und da er auch für ‫ ְּב ִמ ְס ָּפר‬kein Äquivalent bietet,
hat er es ziemlich sicher nicht gelesen. Denn ὡς προσῆκον ἦν gibt auf jeden
Fall (wie ὡς καθήκει in 1 Esdr 1:13 und Lev 5:10 und 9:16) ‫ ְּכ ִמ ְׁש ַּפט‬wieder.
Dann dürfte aber auch das Fehlen von ‫ ְּד ַבר־יֹום ְּביֹומֹו‬eher ein Minus in
der Vorlage widerspiegeln. Talshir kommentiert vermutlich richtig: „I Esd
reflects only ‫  ְּכ ִמ ְׁש ַּפט‬. . . ; the rest, it would seem, was not part of his Vor-
lage. It is pleonastic to a certain extent, especially ‫ ְּד ַבר־יֹום ְּביֹומֹו‬immedi-
ately following ‫“יֹום ְּביֹומֹו‬13.
Der letzte Fall eines derartigen Doppelausdrucks findet sich in Esr 10:14.
Die Mischehenangelegenheit soll durch eine Versammlung von Rich-
tern und Ältesten aus jeder Stadt beraten werden. Der hebräische Text
nennt sie:
Esr 10:14: ‫י־עיר וָ ִעיר‬
ִ ֵ‫זִ ְקנ‬
Die Nachahmung in 2 Esdr ist gewohnt sklavisch:
2 Esdr 10:14: πρεσβύτεροι πόλεως καὶ πόλεως
Einen besseren griechischen Ausdruck findet 1 Esdr 9:13:
1 Esdr 9:13: ἑκάστου δὲ τόπου τοὺς πρεσβυτέρους
Der spezielle Fall der eben behandelten hebräischen und aramäischen
Doppelausdrücke zeigt zunächst nur, dass 2 Esdr auch hier die sklavisch
stereotype Wiedergabe verfolgt, wo 1 Esdr keine Äquivalentenkonstanz
einhält. Und es deutet sich an, dass 1 Esdr die Wiederholung desselben
Worts, die die Vorlage ihm vorgibt, lieber vermeidet14. Manchmal tut er
dies nur aus stilistischen Gründen. Oft aber kann man auch sachliche, ja
theologische Überlegungen hinter seiner Wahl erkennen.

6. Wiederholungsvariationen

Mit der Zerstörung Jerusalems durch den babylonischen König Nebukad-

nezzar erfüllen sich nach 2 Chr 36:21 (1 Esdr 1:54f. Vorlage) die Drohungen
des Propheten Jeremia. Das verwüstete Land sollte Brache halten, bis 70
Jahre voll wären. Dieses doppelte Erfüllen der Prophezeiung Jeremias und

13 Talshir, I Esdras. A Text Critical Commentary, 300.

14 Er unterliegt hier aber keinem Zwang. Vgl. Esr 9:8f: ‫ת־לנּו ִמ ְחיָ ה‬
ָ ‫  ָל ֶת‬. . . ‫ּול ִת ֵּתנּו ִמ ְחיָ ה‬
ְ ;
δοῦναι ἡμῖν τροφὴν . . . δοῦναι ἡμῖν τροφὴν (1 Esdr 8:76, 78).
übersetzungstechnik und textkritik in den esdrasbüchern 107

der 70 Brachjahre heißt im hebräischen Vorlagetext von 1 Esdr ebenso wie

in 2 Chr 36:21 zweimal ‫ ְל ַמּלֹאות‬:
2 Chr 36:21: ‫ ְ ל ַמּלֹאות ִׁש ְב ִעים ָׁשנָ ה‬. . . ‫ְל ַמּלֹאות ְּד ַבר־יְ הוָ ה ְּב ִפי יִ ְר ְמיָ הוּ‬
Der griechische Übersetzer von 1 Esdr achtet darauf, dass er die Wortwie-
derholung durch Variation in der Äquivalentenwahl wiedergibt:
1 Esdr 1:54–55: εἰς ἀναπλήρωσιν τοῦ ῥήματος τοῦ κυρίου ἐν στόματι Ιερεμιου
55 . . . εἰς συμπλήρωσιν ἐτῶν ἑβδομήκοντα
In diesem Fall wären beide Äquivalente sowohl für übertragenes wie
numerisches Vollmachen geeignet gewesen. Wichtig war dem Übersetzer
die Variation.
Da die griechische Fassung von 2 Chr 36:21 (2 Par) dieselbe Variation
vornimmt wie 1 Esdr15, haben wir hier einen ersten Hinweis darauf, dass
der Übersetzer von 2 Par und der von 2 Esdr nicht derselbe sind, da sich
in diesem Punkt ihre Übersetzungstechnik merklich unterscheidet.
Unmittelbar nach ihrer Rückkehr aus dem Babylonischen Exil machen
sich die Heimkehrer unter Serubbabel daran, den Altar in Jerusalem wie-
derzuerrichten, um auf ihm Brandopfer darzubringen (Esr 3:2: ‫ ;עֹלֹות‬2 Esdr
3:2 und 1 Esdr 5:48: ὁλοκαυτώσεις). 1 Esdr kann hebräisches ‫ ע ָֹלה‬sowohl
mit ὁλοκαύτωσις (ὁλοκαύτωμα) als auch mit θυσία wiedergeben. In Esr 3:3,
wo zwei Arten von Brandopfern aufgezählt werden, wählt er nacheinan-
der die beiden Äquivalente. Er setzt sie aber wohlüberlegt:
Esr 3:3 ‫עֹלֹות ַליהוָ ה עֹלֹות ַלּב ֶֹקר וְ ָל ָע ֶרב׃‬
1 Esdr 5:49: θυσίας κατὰ τὸν καιρὸν καὶ ὁλοκαυτώματα τῷ κυρίῳ τὸ πρωινὸν καὶ
τὸ δειλινόν
MT spricht ebenso wie die Vorlage von 1 Esdr zweimal16 von ‫עֹלֹות‬, zuerst
generell von Brandopfern für den Herrn (MT) oder Brandopfern zu den
verschiedenen besonderen Terminen und Gelegenheiten (Sabbaten,

15 2 Par 36:21: τοῦ πληρωθῆναι λόγον κυρίου διὰ στόματος Ιερεμιου ἕως τοῦ προσδέξασθαι
τὴν γῆν τὰ σάββατα αὐτῆς σαββατίσαι πάσας τὰς ἡμέρας τῆς ἐρημώσεως αὐτῆς ἐσαββάτισεν
εἰς συμπλήρωσιν ἐτῶν ἑβδομήκοντα.
16 2 Esdr und die Vg bezeugen einen Text, in dem nur einmal vom Brandopfer die Rede
war und zwar im Singular: καὶ ἀνέβη ἐπ᾽ αὐτὸ ὁλοκαύτωσις τῷ κυρίῳ τὸ πρωὶ καὶ εἰς ἑσπέραν;
holocaustum Domino mane et vespere. Demgegenüber bezeugen 1 Esdr und Esr MT bei
allen kleineren Differenzen, die auch zwischen ihnen bestehen, doch gemeinsam einen
Text, der zweimal von Brandopfern sprach—und zwar je im Plural. Kleins Auffassung,
1 Esdr weise hier eine Doppelübersetzung für nur einmaliges „Brandopfer“ seiner Vor-
lage auf, scheitert am Text von MT (Klein, Studies, 36 und 233; dagegen zu Recht Talshir,
I Esdras. A Text Critical Commentary, 299).
108 dieter böhler sj

Neumonden, Festen, private Opfer)17 und dann von den täglichen Brand-
opfern, dem morgendlichen und dem abendlichen. MT nennt sie beide
‫עֹלֹות‬. 1 Esdr will seine beiden Äquivalente ὁλοκαύτωμα und θυσία zum
Einsatz bringen, verteilt sie aber so, dass er nur die täglichen Brandopfer
ὁλοκαυτώματα nennt, da sie wirklich immer Brandopfer sind, während er
die ersteren, die—etwa als private Opfer—auch ein Schlachtopfer oder
eine Mincha sein könnten, als θυσίας bezeichnet. Die beiden ihm sonst
gleich gültigen Äquivalente werden hier offenbar gezielt eingesetzt und
wären nicht einfach austauschbar.
Die Fortsetzung in Esr 3:4f. nennt die Opfer des Laubhüttenfestes, das
zweimal tägliche Tamid-Opfer und die anderen Festopfer unterschiedslos
‫( עֹלֹות‬2 Esdr: ὁλοκαυτώσεις). 1 Esdr differenziert in diesem Falle—offen-
bar mit leicht abweichender Vorlage („Sabbatopfer“)—so, dass die Opfer
des Laubhüttenfestes (nach Num 29:12–38 Brand- mit Speiseopfer, sowie
Sündopfer) mit dem allgemeinen Begriff θυσία, die täglichen Tamid-Opfer
diesmal προσφορά und die Festopfer wiederum mit θυσία wiedergegeben

Esr 3:4–5 1 Esdr 5:50–51 2 Esdr 3:4–5

ַ ‫ וַ ּיַ ֲעׂשּו ֶא‬4 50 καὶ ἠγάγοσαν τὴν τῆς καὶ ἐποίησαν τὴν ἑορτὴν τῶν
‫ַה ֻּסּכֹות ַּכ ָּכתּוב וְ ע ַֹלת‬ σκηνοπηγίας ἑορτήν ὡς σκηνῶν κατὰ τὸ γεγραμμένον
‫יֹום ְּביֹום ְּב ִמ ְס ָּפר‬ ἐπιτέτακται ἐν τῷ νόμῳ καὶ ὁλοκαυτώσεις ἡμέραν ἐν
‫ְּכ ִמ ְׁש ַּפט ְּד ַבר־יֹום‬ καὶ θυσίας καθ᾽ ἡμέραν ὡς ἡμέρᾳ ἐν ἀριθμῷ ὡς ἡ κρίσις
‫ְּביֹומֹו׃‬ προσῆκον ἦν λόγον ἡμέρας ἐν ἡμέρᾳ αὐτοῦ
‫יכן ע ַֹלת ָּת ִמיד‬ ֵ ‫ וְ ַא ֲח ֵר‬5 51 καὶ μετὰ ταῦτα προσφορὰς 5 καὶ μετὰ τοῦτο ὁλοκαυτώσεις
‫וְ ֶל ֳח ָד ִׁשים‬ ἐνδελεχισμοῦ ἐνδελεχισμοῦ
‫ל־מֹוע ֵדי יְ הוָ ה‬
ֲ ‫ּול ָכ‬
ְ καὶ θυσίας σαββάτων καὶ εἰς τὰς νουμηνίας καὶ εἰς
‫ַה ְמ ֻק ָּד ִׁשים‬ καὶ νουμηνιῶν πάσας ἑορτὰς τὰς ἡγιασμένας
καὶ ἑορτῶν πασῶν ἡγιασμένων

17 Ähnliche Aufzählungen sind in der Chronik nicht selten: 1 Chr 16:40: ‫ְל ַה ֲעלֹות עֹלֹות‬
‫ל־מזְ ַּבח ָהע ָֹלה ָּת ִמיד ַלּב ֶֹקר וְ ָל ָע ֶרב‬ ִ ‫ ; ַליהוָ ה ַע‬2 Chr 2:3: ‫וְ עֹלֹות ַלּב ֶֹקר וְ ָל ֶע ֶרב ַל ַּׁש ָּבתֹות‬
‫מֹוע ֵדי יְ הוָ ה‬
ֲ ‫ּול‬ְ ‫ ;וְ ֶל ֳח ָד ִׁשים‬1 Chr 23:31: ‫ ;עֹלֹות ַליהוָ ה ַל ַּׁש ָּבתֹות ֶל ֳח ָד ִׁשים וְ ַלּמ ֲֹע ִדים‬2 Chr 31:3:
‫ ָלעֹלֹות ְלעֹלֹות ַהּב ֶֹקר וְ ָה ֶע ֶרב וְ ָהעֹלֹות ַל ַּׁש ָּבתֹות וְ ֶל ֳח ָד ִׁשים וְ ַלּמ ֲֹע ִדים‬. Talshir, I Esdras. A
Text Critical Commentary, 299, vermutet mit Blick auf diese Reihungen der Chronik, 1 Esdr
5:49 habe für καιρός ‫ ַלּמ ֲֹע ִדים‬gelesen. In 1:17 steht καιρός für ‫( ֵעת‬2 Chr 35:17). In 1 Esdr 8:76
hat die Vorlage wohl ebenfalls so gelautet (‫ ;) ָּב ֵעת‬MT: ‫ ; ְמ ַעט‬vgl. Talshir, I Esdras. A Text
Critical Commentary, 449).
18 Die Vokabel προσφορά kommt in 1 Esdr nur hier vor. Häufiger steht sie in Sirach (14:11;
34:18, 19; 35:1, 5; 38:11; 46:16; 50:13, 14), da aber nie für eine spezifische Opferart wie Brand-,
Schlacht-, Speiseopfer.
übersetzungstechnik und textkritik in den esdrasbüchern 109

Nach der vom Statthalter Tattenai veranlassten Untersuchung über die

Legitimität der Jerusalemer Tempelbauarbeiten unter Serubbabel schreibt
König Darius in seiner Antwort, die Arbeiten seien zu fördern und wer
dem Reskript zuwiderhandle, dem „reiße man einen Balken aus seinem
Haus und pfähle ihn auf diesem Balken; sein Haus soll wegen seines Ver-
gehens zu einem Trümmerhaufen gemacht werden.“
Esr 6:11: ‫ּוביְ ֵתּה נְ וָ לּו יִ ְת ֲע ֵבד‬
ַ ‫ֹלהי‬
ִ ‫ן־ּביְ ֵתּה ּוזְ ִקיף יִ ְת ְמ ֵחא ֲע‬
ַ ‫יִ ְתנְ ַסח ָאע ִמ‬
‫ל־ּדנָ ה‬
ְ ‫ַע‬
1 Esdr 6:31 λημφθῆναι ξύλον ἐκ τῶν ἰδίων αὐτοῦ καὶ ἐπὶ τούτου κρεμασθῆναι καὶ
τὰ ὑπάρχοντα αὐτοῦ εἶναι βασιλικά
Für das zweimalige „sein Haus“ des aramäischen Textes schwingt sich
selbst 2 Esdr hier zu einer Variation auf, die für ihn an die Grenze des
Möglichen geht. Er schreibt zuerst οἰκία und dann οἶκος. Interessanter ist
1 Esdr. Er wählt zuerst, wo es buchstäblich um das Eigenheim des Täters
geht, τὰ ἴδια αὐτοῦ (so auch Est 5:10 und 6:12)19, während gleich darauf
„sein Haus“ den Besitz im allgemeinen meint und daher mit τὰ ὑπάρχοντα
αὐτοῦ wiedergegeben wird20. Beide Äquivalente sind exakt für ihren Kon-
text gewählt. Hätte man sich möglicherweise für das Eigentum im all-
gemeinen τὰ ἴδια vorstellen können, so doch nicht τὰ ὑπάρχοντα für das
In Esr 6:8–9 befiehlt Darius den Behörden, mit den Juden beim Jeru-
salemer Tempelbau zusammenzuarbeiten, sie bei den Kosten für die
Bauarbeiten und den anschließenden Unterhalt des Kultbetriebs aus der
königlichen Kasse zu unterstützen.

19  W. Bauer, Wörterbuch zum Neuen Testament (Hg. K. Aland und B. Aland; Berlin /
New York: De Gruyter 61988), 752.
20 Eine bemerkenswerte Berührung zwischen 1 Esdr (hier auch 2 Esdr) und DanLXX
zeigt sich bei der Übersetzung für das aramäische Wort ‫נְ וָ לּו‬/‫נְ וָ ִלי‬. 1 Esdr 6:31 und 2 Esdr
6:11 interpretieren die zweite Maßnahme gegen das „Haus“ des Übertreters als „Verstaatli-
chung“ (2 Esdr 6:11: καὶ ὁ οἶκος αὐτοῦ τὸ κατ᾽ ἐμὲ ποιηθήσεται). In den beiden Danielparalle-
len (Dan 2:5: ‫ּוב ֵּתיכֹון נְ וָ ִלי יִ ְּת ָׂשמּון‬
ָ und Dan 3:29: ‫ּוביְ ֵתּה נְ וָ ִלי יִ ְׁש ַּתּוֵ ה‬
ַ ) interpretiert DanLXX
ebenfalls „Verstaatlichen“ (2:5: καὶ ἀναληφθήσεται ὑμῶν τὰ ὑπάρχοντα εἰς τὸ βασιλικόν. 3:96
[MT: 3:29]: καὶ ἡ οἰκία αὐτοῦ δημευθήσεται), DanTh dagegen—wie die heutigen Überset-
zungen—„Zerstörung“ (2:5: καὶ οἱ οἶκοι ὑμῶν διαρπαγήσονται. 3:96: καὶ οἱ οἶκοι αὐτῶν εἰς
110 dieter böhler sj

Esr 6:8–9 1 Esdr 6:28–29 2 Esdr 6:8–9

‫ּומּנִ ְכ ֵסי ַמ ְל ָּכא ִּדי ִמ ַּדת‬
ִ 8 28 καὶ ἀπὸ τῆς φορολογίας 8 καὶ ἀπὸ ὑπαρχόντων
‫ֲע ַבר נַ ֲה ָרה ָא ְס ַּפ ְרנָ א‬ Κοίλης Συρίας καὶ βασιλέως τῶν φόρων πέραν
‫נִ ְפ ְק ָתא ֶּת ֱהוֵ א ִמ ְתיַ ֲה ָבא‬ Φοινίκης ἐπιμελῶς τοῦ ποταμοῦ ἐπιμελῶς
‫י־לא‬ָ ‫ְלגֻ ְב ַרּיָ א ִא ֵּלְך ִּד‬ σύνταξιν δίδοσθαι τούτοις δαπάνη ἔστω διδομένη τοῖς
‫ְל ַב ָּט ָלא׃‬ τοῖς ἀνθρώποις εἰς θυσίας ἀνδράσιν ἐκείνοις τὸ μὴ
τῷ κυρίῳ Ζοροβαβελ καταργηθῆναι
ִ ‫ּובנֵ י‬ ְ ‫ּומה ַח ְׁש ָחן‬ ָ 9 ἐπάρχῳ εἰς ταύρους καὶ
‫וְ ִד ְכ ִרין וְ ִא ְּמ ִרין ַל ֲע ָלוָ ן‬ κριοὺς καὶ ἄρνας 29 ὁμοίως 9 καὶ ὃ ἂν ὑστέρημα καὶ
‫ֶל ֱא ָלּה ְׁש ַמּיָ א ִחנְ ִטין‬ δὲ καὶ πυρὸν καὶ ἅλα καὶ υἱοὺς βοῶν καὶ κριῶν καὶ
‫ּומ ַׁשח‬
ְ ‫ְמ ַלח ֲח ַמר‬ οἶνον καὶ ἔλαιον ἐνδελεχῶς ἀμνοὺς εἰς ὁλοκαυτώσεις
‫אמר ָּכ ֲהנַ ּיָ א ִדי־‬ ַ ‫ְּכ ֵמ‬ κατ᾽ ἐνιαυτόν καθὼς ἂν τῷ θεῷ τοῦ οὐρανοῦ πυρούς
‫ירּוׁש ֶלם ֶל ֱהוֵ א ִמ ְתיְ ֵהב‬ ְ ‫ִב‬ οἱ ἱερεῖς οἱ ἐν Ιερουσαλημ ἅλας οἶνον ἔλαιον κατὰ τὸ
‫י־לא‬ ָ ‫ְלהֹם יֹום ְּביֹום ִּד‬ ὑπαγορεύσωσιν ῥῆμα τῶν ἱερέων τῶν ἐν
‫ָׁשלּו׃‬ ἀναλίσκεσθαι καθ᾽ ἡμέραν Ιερουσαλημ ἔστω διδόμενον
ἀναμφισβητήτως αὐτοῖς ἡμέραν ἐν ἡμέρᾳ ὃ
ἐὰν αἰτήσωσιν

Am Ende von 1 Esdr 6:28 findet sich mit εἰς θυσίας τῷ κυρίῳ Ζοροβαβελ
ἐπάρχῳ ein textkritisch, ja eigentlich literarkritisch relevanter Unterschied
zu Esr MT/2 Esdr, der mit dem Einschub der Pagenerzählung 1 Esdr 3–4,
einer Serubbabellegende, zusammenhängt. Diese echte Textdifferenz soll
uns hier nicht interessieren. Auch ἐνδελεχῶς κατ᾽ ἐνιαυτόν (1 Esdr 6:29)
spiegelt eine echte Vorlagedifferenz wider21. Uns beschäftigt hier nur
das Verhalten des Übersetzers von 1 Esdr. Das Partizip hitpe’el von yhb
in Esr 6:8 und 9 wird von 2 Esdr zweimal gleich mit ἔστω διδομένη/ἔστω
διδόμενον wiedergegeben. 1 Esdr wählt im ersten Falle δίδοσθαι, im zweiten
aber ἀναλίσκεσθαι22. In MT und 2 Esdr bezieht sich das erste Geben auf
einen Beitrag zum Tempelbau, das zweite auf eine Unterstützung zum
anschließenden Kultunterhalt. In der Vorlage von 1 Esdr bezieht sich das
erste Geben auf eine Unterstützung des Tempelbetriebs, das zweite aber
auf „Ausgaben“, die zu tätigen sind. Völlig passend und mit Überlegung
ausgewählt sind die beiden Äquivalente δίδοσθαι und ἀναλίσκεσθαι („auf-
wenden“). Sie könnten nicht ausgetauscht werden.
Diesem verwandt ist der Fall von Esr 7:20. Hier steht in einem ebenfalls
aramäischen Dekret des Perserkönigs zweimal ntn für „ausgeben“:

21 Talshir, I Esdras. A Text Critical Commentary, 361 und 363.

22 Ibid., 361 setzt in beiden Fällen ‫ יְ ַהב‬hitp. wie in MT voraus.
übersetzungstechnik und textkritik in den esdrasbüchern 111

Esr 7:20: ‫ן־ּבית ּגִ נְ זֵ י‬

ֵ ‫ל־לְך ְל ִמנְ ַּתן ִּתנְ ֵּתן ִמ‬
ָ ‫ּוׁש ָאר ַח ְׁשחּות ֵּבית ֱא ָל ָהְך ִּדי יִ ֶּפ‬ ְ
‫ַמ ְל ָּכא׃‬
2 Esdr schreibt dafür zweifaches δίδωμι:
2 Esdr 7:20: καὶ κατάλοιπον χρείας οἴκου θεοῦ σου ὃ ἂν φανῇ σοι δοῦναι δώσεις
ἀπὸ οἴκων γάζης βασιλέως
Hier nun scheint der Übersetzer von 1 Esdr den Infinitiv ‫ ְל ִמנְ ַּתן‬nicht gelesen
zu haben, denn er schreibt nur ὅσα ἂν ὑποπίπτῃ σοι (für ‫ל־לְך‬ ָ ‫) ִּדי יִ ֶּפ‬. Wahr-
scheinlicher aber ist, dass er das wiederholte „geben“ vermeidet, indem er
es in einem anderen Ausdruck aufgehen lässt23:
1 Esdr 8:18: καὶ τὰ λοιπά ὅσα ἂν ὑποπίπτῃ σοι εἰς τὴν χρείαν τοῦ ἱεροῦ τοῦ θεοῦ
σου δώσεις ἐκ τοῦ βασιλικοῦ γαζοφυλακίου
Talshir kommentiert ganz richtig: „The clause is reformulated in I Esd:
(1) the word order changes; (2) ‫ ְל ִמנְ ַּתן‬is not represented directly, probably
because of the following ‫ ִּתנְ ֵּתן‬.“ Wenn 1 Esdr hier einen MT-gleichen Vor-
lagetext hatte24, zeigt die Stelle, dass unser Übersetzer Wiederholungen
zwar nicht durch bloßes Weglassen, wohl aber durch Aufgehenlassen in
einem umfassenderen Ausdruck vermeiden kann. Hier hätte er dann ὅσα
ἂν ὑποπίπτῃ σοι deswegen vor εἰς τὴν χρείαν gestellt (anders als in MT),
weil er in letzterem den (Ausgabe-) Bedarf sah, womit in seinen Augen
das erste „ausgeben“ untergebracht wäre. Das erklärt dann auch die
In seinem aramäischen Firman verfügt Artaxerxes, Esra und die Juden
dürften etwaige Summen, die nach den Ausgaben für den Kultbetrieb
noch übrigblieben, nach dem Gutdünken ihres Gottes verwenden. In Esr
7:18 heißt es:
‫יטב ִּב ְׁש ָאר ַּכ ְס ָּפא וְ ַד ֲה ָבה ְל ֶמ ְע ַּבד ִּכ ְרעּות ֱא ָל ֲהכֹם‬
ַ ִ‫ל־א ָחיְך י‬
ֶ ‫ּומה ִדי ֲע ָליְך וְ ַע‬ ָ
‫ַּת ַע ְבדּון׃‬
Dem entspricht die buchstäbliche Übersetzung von 2 Esdr:
2 Esdr 7:18: καὶ εἴ τι ἐπὶ σὲ καὶ τοὺς ἀδελφούς σου ἀγαθυνθῇ ἐν καταλοίπῳ τοῦ
ἀργυρίου καὶ τοῦ χρυσίου ποιῆσαι ὡς ἀρεστὸν τῷ θεῷ ὑμῶν ποιήσατε

23 Ibid., 404.
24 Wenn 1 Esdr einen anderen Vorlagetext hatte, zeigt die Stelle in Esr 7:20 nur, dass
aramäisch ntn und yhb (Esr 6:8f ) gleichermaßen für (Geld) „ausgeben“ stehen können.
112 dieter böhler sj

1 Esdr vermeidet die Wiederholung desselben Äquivalents und wählt zwei

1 Esdr 8:16: καὶ πάντα ὅσα ἂν βούλῃ μετὰ τῶν ἀδελφῶν σου ποιῆσαι χρυσίῳ καὶ
ἀργυρίῳ ἐπιτέλει κατὰ τὸ θέλημα τοῦ θεοῦ σου
Der Übersetzer hat sich dabei genau überlegt, wann er das allgemeine
„machen“ und wann das spezifischere „ausführen“ setzt. Die Wahl der
beiden Äquivalente für aramäisches ‘bd wäre nicht austauschbar.
In Esr 7:6 wird mitgeteilt, dass der Priester und Schriftgelehrte Esra
einerseits kundig gewesen sei im Gesetz, das der Herr Israel gegeben hat,
dass andererseits der Perserkönig dem Priester alles gab, was er wünschte.
Beides, das Geben des Gesetzes durch Gott und die königlichen Gaben,
werden im hebräischen Text durch ntn ausgedrückt.
Esr 7:6: ‫ֹלהי יִ ְׂש ָר ֵאל וַ ּיִ ֶּתן־לֹו ַה ֶּמ ֶלְך ְּכיַ ד־‬
ֵ ‫תֹורת מ ֶֹׁשה ֲא ֶׁשר־נָ ַתן יְ הוָ ה ֱא‬
ַ ‫ְּב‬
‫ֹלהיו ָע ָליו ּכֹל ַּב ָּק ָׁשתֹו‬
ָ ‫יְ הוָ ה ֱא‬
2 Esdr schreibt zweimal δίδωμι und bildet den hebräischen Text Wort für
Wort ab.
2 Esdr 7:6: ἐν νόμῳ Μωυσῆ ὃν ἔδωκεν κύριος ὁ θεὸς Ισραηλ καὶ ἔδωκεν αὐτῷ ὁ
βασιλεύς ὅτι χεὶρ κυρίου θεοῦ αὐτοῦ ἐπ᾽ αὐτὸν ἐν πᾶσιν οἷς ἐζήτει αὐτός
1 Esdr muss einen leicht veränderten Vorlagetext gehabt haben, da bei
ihm die göttliche durch die königliche Gunst ersetzt ist25. Der Punkt aber,
auf den es uns hier ankommt, ist: Während 1 Esdr das königliche Geben,
das ja nur einem Manne gilt, ebenfalls mit schlichtem δίδωμι wiedergibt,
wählt er für das göttliche Geben der Tora, die ein Buch für die Öffent-
lichkeit ist, das ebenso simple wie ausnehmend passende Kompositum
ἐκδίδωμι („herausgeben“, „publizieren“):
1 Esdr 8:3–4: ἐν τῷ Μωυσέως νόμῳ τῷ ἐκδεδομένῳ ὑπὸ τοῦ θεοῦ τοῦ Ισραηλ
4καὶ ἔδωκεν αὐτῷ ὁ βασιλεὺς δόξαν εὑρόντος χάριν ἐναντίον αὐτοῦ ἐπὶ πάντα τὰ
ἀξιώματα αὐτοῦ
Der Firman des Perserkönigs Artaxerxes, mit dem Esra seinen Auftrag
erhält, endet in Esr 7:26 mit einer Strafandrohung an alle, die sich an das

25 Talshir (I Esdras. A Text Critical Commentary, 389f.) nimmt keine Vorlagedifferenz

an, sondern schreibt dem Übersetzer eine für ihn untypische Abweichung zu. Sie verweist
auf 1 Esdr 6:9 (Esr 5:8) für den richtigen Hinweis, dass δόξα für ‫ יָ ד‬zu stehen pflegt. Das
erklärt aber noch nicht das Fehlen von „des Herrn, seines Gottes“.
übersetzungstechnik und textkritik in den esdrasbüchern 113

Gesetz Gottes und das Gesetz des Königs nicht halten wollten. Der aramä-
ische Ausdruck für das Gesetz ist in beiden Fällen ‫ ָּדת‬.
Esr 7:26: ‫י־א ָל ָהְך וְ ָד ָתא ִּדי ַמ ְל ָּכא‬
ֱ ‫ָּד ָתא ִד‬
2 Esdr übersetzt mechanisch zweimal mit νόμος.
2 Esdr 7:26: νόμον τοῦ θεοῦ καὶ νόμον τοῦ βασιλέως
1 Esdr gibt das zweimalige aramäische ‫ ָּדת‬nicht mit zweimal demselben
griechischen Ausdruck wieder, sondern schreibt:
1 Esdr 8:24: τὸν νόμον τοῦ θεοῦ σου καὶ τὸν βασιλικόν
Klein notiert die Differenz unter „stylistic omissions“26. Talshir scheint
mir die Wahl des griechischen Übersetzers an dieser Stelle treffender zu
beschreiben, wenn sie anmerkt: „he makes a fine distinction between the
divine law . . . and the royal one for which he chooses to condense the
Aramaic phrase“27. Denn es ist völlig klar, dass er zweimal ‫ ָּדת‬gelesen
hat. Auch 1 Esdr redet nicht einfach von einem Gesetz Gottes und des
Königs (‫י־א ָל ָהְך וְ ִּדי ַמ ְל ָּכא‬
ֱ ‫) ָּד ָתא ִד‬, sondern von zweien, nennt aber nur
das erstere νόμος, das letztere τὸν βασιλικόν, ohne noch einmal dasselbe
Wort zu wiederholen, das das göttliche Gesetz bezeichnet hatte. Er inte-
griert das zweite ‫ ָּד ָתא‬in einen neu formulierten Ausdruck und lässt es
nicht einfach weg. Hätte 2 Esdr das zweite νόμος unterdrückt, müssten wir
uns fragen, ob seine aramäische Vorlage nur einmal ‫ ָּדת‬gelesen hat. Wir
hätten ein textkritisches Problem. Bei 1 Esdr aber zeigt sich kein textkriti-
sches Problem, sondern eine Übersetzungsphilosophie.
Ebenfalls im Firman des Artaxerxes findet sich die Vorschrift, dass die
Aufwendungen für den Jerusalemer Tempelgottesdienst aus der Kasse des
Königs beglichen werden. „Haus deines Gottes“ und „Haus der Schätze des
Königs“ stehen im aramäischen Text dicht bei einander:
Esr 7:20: ‫ן־ּבית ּגִ נְ זֵ י ַמ ְל ָּכא׃‬
ֵ ‫ל־לְך ְל ִמנְ ַּתן ִּתנְ ֵּתן ִמ‬
ָ ‫ֵּבית ֱא ָל ָהְך ִּדי יִ ֶּפ‬
2 Esdr setzt für ‫ ֵּבית‬zweimal unbeirrt οἶκος. In 2 Esdr 7:20 heißt es:
οἴκου θεοῦ σου ὃ ἂν φανῇ σοι δοῦναι δώσεις ἀπὸ οἴκων γάζης βασιλέως
Die Pluralform „Schatzhäuser“ muss textkritisch nichts bedeuten. Sie kann
ein Versuch sein, das aramäische „Schätzehaus“ wiederzugeben. 1 Esdr

26 Klein, Studies, 164.

27 Talshir, I Esdras. A Text Critical Commentary, 408f.
114 dieter böhler sj

wiederholt nicht οἶκος, ja wählt überhaupt zwei völlig verschiedene Äqui-

valente, „Heiligtum Gottes“ für „Gotteshaus“ und „königliche Schatzkam-
mer“ für „Haus der Schätze des Königs“:
1 Esdr 8:18: τοῦ ἱεροῦ τοῦ θεοῦ σου δώσεις ἐκ τοῦ βασιλικοῦ γαζοφυλακίου
1 Esdr hat auch zweimal ‫ ֵּבית‬gelesen, aber der Übersetzer vermeidet die
Wiederholung desselben Ausdrucks und wählt stattdessen zwei je an ihrer
Stelle exakt passende Äquivalente des Hebräischen.
In Neh 8:5–6 findet sich auf engstem Raum viermal hintereinander der
Ausdruck „das ganze Volk“.
‫ּוכ ִפ ְתחֹו ָע ְמדּו‬
ְ ‫ל־ה ָעם ָהיָ ה‬ ָ ‫י־מ ַעל ָּכ‬ ֵ ‫ל־ה ָעם ִּכ‬
ָ ‫וַ ּיִ ְפ ַּתח ֶעזְ ָרא ַה ֵּס ֶפר ְל ֵעינֵ י ָכ‬5
‫ל־ה ָעם ָא ֵמן ָא ֵמן‬
ָ ‫ֹלהים ַהּגָ דֹול וַ ּיַ ֲענּו ָכ‬
ִ ‫וַ יְ ָב ֶרְך ֶעזְ ָרא ֶאת־יְ הוָ ה ָה ֱא‬6 ‫ל־ה ָעם׃‬
ָ ‫ָכ‬
2 Esdr übersetzt ‫ל־ה ָעם‬
ָ ‫ ָכ‬erwartungsgemäß mit πᾶς ὁ λαός:
2 Esdr 18:5–6: καὶ ἤνοιξεν Εσδρας τὸ βιβλίον ἐνώπιον παντὸς τοῦ λαοῦ ὅτι αὐτὸς ἦν
ἐπάνω τοῦ λαοῦ καὶ ἐγένετο ἡνίκα ἤνοιξεν αὐτό ἔστη πᾶς ὁ λαός . . . καὶ ἀπεκρίθη
πᾶς ὁ λαὸς καὶ εἶπαν αμην
Die beiden Textdifferenzen zwischen MT und 2 Esdr, die diese Stelle auf-
weist, werfen textkritische Fragen auf. So hat die Vorlage von 2 Esdr offen-
bar nicht ‫ל־ה ָעם‬
ָ ‫ ֵמ ַעל ָּכ‬gelesen, sondern nur ‫ ֵמ ַעל ָה ָעם‬. Zwar bezeugt der
lukianische Text + πάντος, aber nach allen Regeln der Textkritik muss das
als spätere Angleichung an den Kontext oder den heutigen MT gelten.
Außerdem hat 2 Esdr nur einmal „Amen“ gelesen—ebenso wie 1 Esdr.
1 Esdr gibt das hebräische vierfache ‫ל־ה ָעם‬ ָ ‫ ָכ‬nicht mit dem viermal
gleichen griechischen Ausdruck wieder.
1 Esdr 9:45–47: καὶ ἀναλαβὼν Εσδρας τὸ βιβλίον τοῦ νόμου ἐνώπιον τοῦ πλήθους
προεκάθητο γὰρ ἐπιδόξως ἐνώπιον πάντων 46καὶ ἐν τῷ λῦσαι τὸν νόμον πάντες
ὀρθοὶ ἔστησαν . . . καὶ ἐπεφώνησεν πᾶν τὸ πλῆθος αμην.
Der Übersetzer splittet den Doppelausdruck „alles Volk“ in „alle“ und
„Volk“ und verteilt die beiden Ausdrücke πάντες und τὸ πλῆθος über die
vier Stellen: Menge—alle—alle—ganze Menge. Die Vorlage von 1 Esdr hat
vermutlich wie MT (2 Esdr L) ‫ל־ה ָעם‬ ָ ‫ ֵמ ַעל ָּכ‬gelesen und deswegen hier
πάντων für ein passendes Äquivalent gehalten. Wie 2 Esdr hat auch 1 Esdr
in seiner Vorlage offenbar nur einfaches „Amen“ gelesen. Hätte er nämlich
zweifaches hebräisches „Amen“ gelesen, aber die buchstäbliche Wieder-
holung in der septuagintaüblichen Formel γένοιτο γένοιτο (Num 5:22; Jdt
13:20; Pss 41:14; 72:19; 89:53; 106:48) vermeiden wollen, hätte er viele andere
Wege gehabt, die Doppelung ohne Wiederholung wiederzugeben.
übersetzungstechnik und textkritik in den esdrasbüchern 115

Da neben der Übersetzung γένοιτο (neben den genannten Stellen

z.B. auch durchweg in Dtn 27) auch die einfache Transkription αμην
vorkommt (1 Chr 16:36, Neh 5:13) aber auch die Übersetzung ἀληθῶς
( Jer 28[35]:6) hätte man sich auch irgendeine Kombination wie αμην
γένοιτο oder ἀληθῶς γένοιτο vorstellen können. Das einfache „Amen“ von
1 Esdr 9:47 ist keine Wiederholungsvermeidung, sondern wie in 2 Esdr
abweichende Vorlage. Die Transkription des hebräischen Wortes, wie sie
im NT normal ist (bei den Synoptikern über 50 Mal, immer einfach; bei
Joh immer doppelt: 25 mal „Amen, amen“), spricht dafür, dass der Über-
setzer den Ausruf als liturgischen Ruf kennt und in der Liturgie 1 Esdr 9/
Neh 8 entsprechend setzt.

7. Differenzierte Übersetzung, wo auch MT differenziert

In Esr 7:7–8 beginnt die Erzählung von Esras Heimkehr. Der Erzähler
datiert sehr genau, wann Esra und seine Karawane nach Jerusalem auf-
gebrochen sind:
7 Mit ihm zog im siebten Jahr des Königs Artaxerxes auch eine Anzahl von
Israeliten sowie von Priestern, Leviten, Sängern, Torwächtern und Tempel-
dienern nach Jerusalem. 8 Im fünften Monat dieses siebten Regierungsjahrs
des Königs kam Esra in Jerusalem an.
Der hebräische Text spricht einmal vom „Jahr sieben“ und einmal vom
„siebten Jahr“. 2 Esdr gibt alle Jahreszahlangaben wie die griechische Spra-
che das verlangt, mit Ordinalzahlen wieder28. Bei ihm heißt es also zwei-
mal ἔτος ἕβδομον. Dagegen differenziert 1 Esdr 8:5—freilich nicht beim
Zahlwort, sondern beim Äquivalent für „Jahr“: einmal schreibt er ἔτος ein-
mal ἐνιαυτὸς. Hier kann neben der Neigung von 1 Esdr bei zwei dicht auf-
einander folgenden gleichen hebräischen Ausdrücken in der Übersetzung
öfter einmal zu variieren29, der Wunsch mitgespielt haben, die leichte
Differenz im hebräischen Ausdruck, auch im Griechischen durchschei-
nen zu lassen. Freilich differenziert er gerade nicht bei der Zahlenangabe,
sondern beim Ausdruck für Jahr.

28 2 Esdr 1:1; 3:8; 4:24; 5:13; 6:3, 15; 7:7–8. Nur in 3:8 und 7:8 hat der hebräische Text Ordi-
nalzahlen, alle anderen Stellen sind hebräische und aramäische Kardinalzahlen.
29 Die Neigung ist kein Zwang, wie das zweimalige „fünfter Monat“ in 1 Esdr 8:5–6
116 dieter böhler sj

Esr 7:7–8 MT 2 Esdr 7:7–8 1 Esdr 8:5

‫ת־ׁש ַבע‬
ֶ ַ‫רּוׁש ָלםִ ִּב ְׁשנ‬ ָ ְ‫ ֶאל־י‬7 εἰς Ιερουσαλημ ἐν ἔτει ἑβδόμῳ εἰς Ιεροσόλυμα ἔτους
‫ְל ַא ְר ַּת ְח ַׁש ְס ְּתא ַה ֶּמ ֶלְך׃‬ τῷ Αρθασασθα τῷ βασιλεῖ ἑβδόμου βασιλεύοντος
‫רּוׁש ַלםִ ַּבח ֶֹדׁש‬ָ ְ‫וַ ּיָ בֹא י‬8 8καὶ ἤλθοσαν εἰς Ιερουσαλημ Ἀρταξέρξου ἐν τῷ πέμπτῳ
‫יׁשי ִהיא ְׁשנַ ת‬ ִ ‫ַה ֲח ִמ‬ τῷ μηνὶ τῷ πέμπτῳ τοῦτο μηνί οὗτος ἐνιαυτὸς ἕβδομος
‫יעית ַל ֶּמ ֶלְך׃‬ ִ ‫ַה ְּׁש ִב‬ ἔτος ἕβδομον τῷ βασιλεῖ τῷ βασιλεῖ

In Esr 6:14 vermeldet der Erzähler die Vollendung des Tempelbaus und
erwähnt, sie sei durch Dekret Gottes und Dekret der Perserkönige erfolgt.
Der aramäische Text spricht zweimal von ‫ ְט ֵעם‬, gebraucht aber die beiden
leicht verschiedenen (aber beide gebräuchlichen) Konstruktusformen:
zuerst ‫( ַט ַעם‬wie z. B. Esr 7:23), dann aber ‫( ְט ֵעם‬wie z. B. Dan 5:2). Auch
die Präposition ‫ ִמן‬wird einmal assimiliert und einmal nicht assimiliert
Esr 6:14: ‫ּכֹורׁש וְ ָד ְריָ וֶ ׁש‬
ֶ ‫ּומ ְּט ֵעם‬
ִ ‫ן־ט ַעם ֱא ָלּה יִ ְׂש ָר ֵאל‬
ַ ‫וְ ַׁש ְכ ִללּו ִמ‬
‫וְ ַא ְר ַּת ְח ַׁש ְׂש ְּתא‬
2 Esdr drückt die unterschiedliche Vokalisierung, die er vielleicht gar nicht
kennt, nicht aus. Bei ihm heißt es zweimal ununterschieden ἀπὸ γνώμης:
καὶ κατηρτίσαντο ἀπὸ γνώμης θεοῦ Ισραηλ καὶ ἀπὸ γνώμης Κύρου καὶ Δαρείου
καὶ Αρθασασθα.
1 Esdr differenziert hier, wie zu erwarten. Und es ist hier nicht allein der
übliche Wille zur Variation, sondern mit Sicherheit auch die Überzeu-
gung, göttliches und königliches Dekret seien nicht einfach von derselben
1 Esdr 7:4–5: καὶ συνετέλεσαν ταῦτα διὰ προστάγματος τοῦ κυρίου θεοῦ Ισραηλ
5καὶ μετὰ τῆς γνώμης Κύρου καὶ Δαρείου καὶ Ἀρταξέρξου.
Ganz richtig schreibt Talshir zu unserer Stelle: „The translator may have
chosen different equivalents in order to distinguish between God and the
kings“30. Den Willen, das menschliche Edikt vom göttlichen Ratschluss
zu unterscheiden zeigt ja später auch die masoretische Vokalisierung
(ebenso wie die verschiedenen Formen der Präposition mn). Es ist gut
möglich, dass eine solche Lesetradition schon zu Zeiten des Übersetzers
von 1 Esdr im Umlauf war, obwohl 2 Esdr davon keine Spur zeigt.

30 Talshir, I Esdras. A Text Critical Commentary, 373.

übersetzungstechnik und textkritik in den esdrasbüchern 117

8. Übersetzungsphilosophie und Textkritik

Die eingangs behandelte Praxis des Übersetzers von 1 Esdr, bisweilen mit
Doppelübersetzungen zu arbeiten, erlaubt an anderen Stellen ein siche-
reres textkritisches Urteil. Bei der Toraverlesung in Neh 8 heißt es in 8:6:
„Und Esra pries den Herrn, den großen Gott.“ 2 Esdr bezeugt exakt den-
selben Text. Dagegen heißt es in 1 Esdr 9:46: „Und Esdras pries den Herrn,
den höchsten Gott, den allherrschenden Gott Sabaoth.“

Neh 8:6 1 Esdr 9:46 2 Esdr 18:6

6 ‫ וַ יְ ָב ֶרְך ֶעזְ ָרא ֶאת־יְ הוָ ה‬καὶ εὐλόγησεν Εσδρας τῷ καὶ ηὐλόγησεν Εσδρας
‫ֹלהים ַהּגָ דֹול‬
ִ ‫ ָה ֱא‬κυρίῳ θεῷ ὑψίστῳ κύριον τὸν θεὸν τὸν μέγαν
θεῷ σαβαωθ παντοκράτορι

Hanhart schreibt die Erweiterung in 1 Esdr allen Ernstes dem Übersetzer

zu: „Es scheint, dass die vorgegebene Bezeichnung Gottes als ‫ ַהּגָ דֹול‬für
den Übersetzer der Anlaß war, die ihm in den Überlieferungen vorliegen-
den Gottesepitheta zu vereinigen“31. Die Übersetzungsphilosophie unse-
res griechischen Übersetzers schließt dergleichen kategorisch aus.
Richtiger liegt Talshir, die in σαβαωθ παντοκράτορι eine „double
representation“32, „double translation“33 von ‫ ְצ ָבאֹות‬erkennt. Einmal gibt
1 Esdr den wohl auch liturgisch gebräuchlichen hebräischen Ausdruck
in Transkription, dann aber auch noch zusätzlich in griechischer Über-
setzung wieder. Die beiden Wiedergaben σαβαωθ und παντοκράτωρ sind
gängig in der LXX und im NT34. Die hier vorliegende Doppelübersetzung
beweist, dass der Übersetzer in seiner hebräischen Vorlage ‫ֹלהי ְצ ָבאֹות‬
ֵ ‫ֱא‬
tatsächlich vor Augen hatte. Seine Vorlage weicht hier von MT ab. Daraus

31  R. Hanhart, Text und Textgeschichte des 1. Esrabuches (MSU XII, AAWG.PH 91;
Göttingen: Vandenhoeck & Ruprecht, 1974), 69.
32 Talshir, 1 Esdras. From Origin to Translation, 241.
33 Talshir, I Esdras. A Text Critical Commentary, 491. Talshir erwägt hier, anders als
idem, 1 Esdras. From Origin to Translation, 241, noch andere, weit weniger wahrscheinli-
che Möglichkeiten, etwa „der allherrschende Gott Sabaoth“ als Doppel-übersetzung von
„der große Gott.“
34 Das Gottesepitheton ‫ ְצ ָבאֹות‬wird in der LXX bei Jesaja transkribiert (σαβαωθ), im
restlichen AT übersetzt, und zwar in Sam-Kön, Pss und Zeph als Genitiv (κύριος τῶν
δυνάμεων, θεὸς τῶν δυνάμεων), in Ijob, Dodekapropheton (ohne Zeph), Jer mit der Apposi-
tion παντοκράτωρ. Im NT steht zweimal κύριος σαβαωθ (Röm 9:29; Jak 5:4) und einmal κύριος
παντοκράτωρ (2 Kor 6:18).
118 dieter böhler sj

folgt natürlich noch nicht, dass 1 Esdr hier den älteren Text bietet. Das ist
sogar eher unwahrscheinlich, denn ‫ ְצ ָבאֹות‬kommt zwar in den Chroniken
(1 Chr 11:9; 17:7, 24) als Gottesepitheton vor, nie aber in Esr–Neh. So ist
es in der hebräischen Vorlage von 1 Esdr wohl eine spätere Erweiterung,
die aber (gegen Hanhart) auf jeden Fall schon in der hebräischen Vorlage
stand und keinesfalls dem Übersetzer zuzutrauen ist.
In Esr 5:5—wir haben diese Stelle oben schon besprochen—fanden die
griechischen Übersetzer den aramäischen Ausdruck ‫הּודיֵ א‬ ָ ְ‫„ ָׂש ֵבי י‬die Älte-
sten der Juden“. 2 Esdr entscheidet sich für τὴν αἰχμαλωσίαν Ιουδα, also „die
Exulanten, die Ältesten der Juden“ (1 Esdr 6:5: ἐπὶ τὴν αἰχμαλωσίαν παρὰ τοῦ
κυρίου οἱ πρεσβύτεροι τῶν Ιουδαίων). Es war für den Übersetzer offenbar nicht
entscheidbar, ob er ‫„ ָׂש ֵבי‬Älteste“ oder ‫„ ְׁש ִבי‬Gefangenschaft“ lesen sollte.
Diese Technik der Doppelübersetzung verwendet er sogleich in v. 8
erneut. Hier nun hilft die Kenntnis seiner Übersetzungstechnik bei der
Entscheidung einer textkritischen Frage. 1 Esdr hat in 6:8 gegenüber Esr
MT 5:8 einen merklichen Textüberschuss. Er erwähnt die Ältesten der

Esr 5:8 MT 2 Esdr 5:8 1 Esdr 6:8

‫יְ ִד ַיע ֶל ֱהוֵ א ְל ַמ ְל ָּכא‬8 8 γνωστὸν ἔστω τῷ βασιλεῖ 8 πάντα γνωστὰ ἔστω τῷ κυρίῳ
‫י־אזַ ְלנָ א ִליהּוד‬ ֲ ‫ ִּד‬ὅτι ἐπορεύθημεν εἰς τὴν ἡμῶν τῷ βασιλεῖ ὅτι παραγενόμενοι
‫ ְמ ִדינְ ָּתא‬Ιουδαίαν χώραν εἰς τὴν χώραν τῆς Ιουδαίας
καὶ ἐλθόντες εἰς Ιερουσαλημ τὴν
πόλιν κατελάβομεν τῆς αἰχμαλωσίας
τοὺς πρεσβυτέρους τῶν Ιουδαίων ἐν
‫ְל ֵבית ֱא ָל ָהא ַר ָּבא‬ εἰς οἶκον τοῦ θεοῦ τοῦ Ιερουσαλημ τῇ πόλει
‫וְ הּוא ִמ ְת ְּבנֵ א ֶא ֶבן‬ μεγάλου καὶ αὐτὸς οἰκοδομοῦντας οἶκον τῷ κυρίῳ
‫ּגְ ָלל וְ ָאע ִמ ְּת ָׂשם‬ οἰκοδομεῖται λίθοις ἐκλεκτοῖς μέγαν καινὸν διὰ λίθων ξυστῶν
‫ְּב ֻכ ְת ַלּיָ א‬ καὶ ξύλα ἐντίθεται ἐν τοῖς πολυτελῶν ξύλων τιθεμένων ἐν τοῖς
τοίχοις τοίχοις

Es ist ziemlich klar und weitgehend anerkannt,35 dass der Rückverweis

auf „diese Ältesten“ im folgenden Vers Esr 5:9 beweist, dass sie auch in
5:8 einst erwähnt gewesen sein müssen, mithin das Plus von 1 Esdr 6:8

35 H. G. M. Williamson, Ezra, Nehemiah (WBC 16; Waco: Word Books, 1985), 70;
J. Blenkinsopp, Ezra-Nehemiah. A Commentary (OTL; London: SCM Press, 1989), 119;
D. Böhler, Die heilige Stadt in Esdra α und Esra-Nehemia. Zwei Konzeptionen der Wieder-
herstellung Israels (OBO 158; Freiburg / Göttingen: Universitätsverlag / Vandenhoeck &
Ruprecht, 1997), 154–158.
übersetzungstechnik und textkritik in den esdrasbüchern 119

den gegenüber Esr 5:8 ursprünglicheren Text bietet. Gunneweg jedoch

hält das Minus von MT für ursprünglich und das Plus von 1 Esdr für eine
erzählerische Auffüllung durch den Verfasser von 1 Esdr36. Da Gunneweg
1 Esdr für „ein Zeugnis des frühhellenistischen Judentums“37 hält, scheint
er an einen griechischen Verfasser oder Kompilator als Auffüller zu den-
ken. Die Kenntnis der Übersetzungspraxis des Übersetzers von 1 Esdr
erlaubt es uns jedoch, auszuschließen, dass das Plus in 1 Esdr 6:8 auf grie-
chischer Ebene entstanden ist.
Die Tatsache, dass er mit τῆς αἰχμαλωσίας τοὺς πρεσβυτέρους wieder eine
Doppelübersetzung für ‫ ָׂש ֵבי‬bietet, beweist, dass er einen aramäischen
Text vor Augen hatte und sich wie schon in 6:5 (Esr 5:5) nicht entscheiden
konnte, ob er ‫„ ָׂש ֵבי‬Älteste“ oder ‫„ ְׁש ִבי‬Gefangenschaft“ lesen sollte38. Das
Plus in 1 Esdr 6:8 ist sicher aramäischen Ursprungs und, wie die meisten
Ausleger mit dem textkritischen Apparat der BHS urteilen, obendrein der
gegenüber MT ursprünglichere Text.
Etwas anders gelagert und doch benachbart ist der Fall von 1 Esdr 1:4
(2 Chr 35:3). Nach MT weist König Joschija bei der Feier des Paschafestes
die Leviten an, Gott und seinem Volk zu dienen:
2 Chr 35:3: ‫יכם וְ ֵאת ַעּמֹו יִ ְׂש ָר ֵאל׃‬
ֶ ‫ֹלה‬
ֵ ‫ַע ָּתה ִע ְבדּו ֶאת־יְ הוָ ה ֱא‬
Der hebräische Text hat ein einziges Verb mit zwei Objekten. Auch 2 Par
35:3 bezeugt eben diesen Text: νῦν οὖν λειτουργήσατε τῷ κυρίῳ θεῷ ὑμῶν
καὶ τῷ λαῷ αὐτοῦ Ισραηλ.
Dagegen bietet 1 Esdr zwei verschiedene Verben mit je einem Objekt:
1 Esdr 1:4: 4καὶ νῦν λατρεύετε τῷ κυρίῳ θεῷ ὑμῶν καὶ θεραπεύετε τὸ ἔθνος αὐτοῦ
Die Frage ist nun, ob wir hier ein textkritisches Problem haben, wie
Talshir meint, oder nur eine Frage der Übersetzungstechnik, wie Klein
urteilt. Talshir39 geht davon aus, dass λατρεύετε für ‫ ִע ְבדּו‬steht und den
kultischen Dienst vor Gott meint. Für den Dienst am Volk und damit
das andere Verb θεραπεύετε nimmt sie, nicht ohne Zögern, an, dass ein
anderes Verb im hebräischen Text gestanden haben müsse, näherhin
‫ ָׁש ְרתּו‬. Dagegen glaubt Klein, die Wahl des ersten griechischen Äquiva-
lents habe den Übersetzer gezwungen für das zweite Objekt ein eigenes

36 A. H. J. Gunneweg, Esra (KAT XIX 1; Gütersloh: Gerd Mohn, 1985), 99f.
37 Ibid., 24.
38 Ebenso Klein, Studies, 221 und Talshir, I Esdras. A Text Critical Commentary, 334.
39 Talshir, I Esdras. A Text Critical Commentary, 10.
120 dieter böhler sj

­Verbenäquivalent zu suchen: „Since the translator chose λατρεύω (wor-

ship) to translate ‫ ִע ְבדּו‬, he had to use a more neutral word to indicate the
care for the people of Israel“40.
Die unterdessen wohlbekannte Bereitschaft unseres Übersetzers, einen
hebräischen Ausdruck durch eine griechische Doppelübersetzung oder
Hendiadyoin wiederzugeben oder aber ein und dasselbe wiederholte
hebräische Wort mit variierenden griechischen Vokabeln zu überset-
zen, macht es meines Erachtens völlig unnötig, hier ein zweites Wort für
„dienen“ in der hebräischen Vorlage zu postulieren. Allenfalls könnte man
doppeltes ‫ ִע ְבדּו‬vermuten, denn auch das würde der Übersetzer von 1 Esdr
mit λατρεύετε und θεραπεύετε wiedergeben. Es ist ihm aber nach allem,
was wir von ihm wissen, ohne weiteres zuzutrauen, dass er auch ein nur
einmal dastehendes hebräisches Wort mit einer Doppelübersetzung wie-
dergibt, wenn es ihm, wie hier, unpassend erscheint, dieselbe Art des Die-
nens auf Gott und Menschen anzuwenden.
Auf den ersten Blick etwas verwickelt liegt der Fall von Esr 9:12. Esra
richtet nach der Aufdeckung des Mischehenskandals ein langes Bußge-
bet an Gott (Esr 9:5–15). Dabei kommt er in 9:11–12 auf die Vorschriften
der Tora zu sprechen, wie die einwandernden Israeliten mit den im Land
befindlichen Nichtisraeliten umgehen sollen. Dabei zitiert er neben Lev
18 auch Dtn 23:7:
Ihr sollt nicht Friedens- noch Freundschaftsvertrag mit ihnen suchen in
Ewigkeit, damit ihr stark bleibt und das Gut des Landes essen und euren
Kindern vererben könnt in Ewigkeit.
Das hebräische Gebet sagt zweimal ‫ד־עֹולם‬
ָ ‫„ ַע‬in Ewigkeit“. 2 Esdr über-
setzt stereotyp zweimal mit ἕως αἰῶνος. 1 Esdr gibt aber das erste ‫ד־עֹולם‬
ָ ‫ַע‬
mit τὸν ἅπαντα χρόνον wieder, das zweite mit ἕως αἰῶνος.

Esr 9:12 MT 2 Esdr 9:12 1 Esdr 8:82

‫ֹלמם‬ ָ ‫א־ת ְד ְרׁשּו ְׁש‬ ִ ֹ ‫וְ ל‬ οὐκ ἐκζητήσετε εἰρήνην καὶ οὐ ζητήσετε εἰρηνεῦσαι
ָ ‫טֹוב ָתם ַע‬ָ ְ‫ו‬ αὐτῶν καὶ ἀγαθὸν αὐτῶν τὰ πρὸς αὐτοὺς τὸν ἅπαντα
‫ְל ַמ ַען ֶּת ֶחזְ קּו וַ ֲא ַכ ְל ֶּתם‬ ἕως αἰῶνος ὅπως ἐνισχύσητε χρόνον ἵνα ἰσχύσαντες
‫ֶאת־טּוב ָה ָא ֶרץ‬ καὶ φάγητε τὰ ἀγαθὰ τῆς φάγητε τὰ ἀγαθὰ τῆς γῆς
‫יכם‬ ֶ ֵ‫הֹור ְׁש ֶּתם ִל ְבנ‬ַ ְ‫ו‬ γῆς καὶ κληροδοτήσητε τοῖς καὶ κατακληρονομήσητε τοῖς
ָ ‫ַע‬ υἱοῖς ὑμῶν ἕως αἰῶνος υἱοῖς ὑμῶν ἕως αἰῶνος

40 Klein, Studies, 248.

übersetzungstechnik und textkritik in den esdrasbüchern 121

Klein41 erkennt das Zitat aus Dtn 23:7 in der ersten Vershälfte Esr 9:12a
und 1 Esdr 8:82a:
Dtn 23:7: ‫עֹולם‬
ָ ‫ֹלמם וְ ט ָֹב ָתם ָּכל־יָ ֶמיָך ְל‬
ָ ‫א־ת ְדר ֹׁש ְׁש‬
ִ ֹ‫ל‬
Er vergleicht daher nur die ersten Vershälften mit Dtn 23:7 und sieht,
dass der deuteronomischen doppelten Zeitbestimmung ‫עֹולם‬ ָ ‫ָּכל־יָ ֶמיָך ְל‬
in 2 Esdr das einfache ἕως αἰῶνος, in 1 Esdr 8:82 das ebenfalls einfache τὸν
ἅπαντα χρόνον entspricht. Er schließt nun daraus, dass das ‫ד־עֹולם‬ ָ ‫ ַע‬bzw.
ἕως αἰῶνος von MT bzw 2 Esdr einerseits und das τὸν ἅπαντα χρόνον von
1 Esdr andererseits im MT des Dtn zu einer Konflation verbunden worden
wären. Es scheint mir jedoch von vornherein sehr unwahrscheinlich, dass
der hebräische Text der Tora durch Varianten der eher späten Esrabücher
beeinflusst worden sein sollte. Klein berück-sichtigt vor allem überhaupt
nicht, dass in Esras Gebet gleich zweimal hintereinander ‫ד־עֹולם‬ ָ ‫ ַע‬steht,
wovon nur das erste mit τὸν ἅπαντα χρόνον, das zweite aber mit der für
Klein alternativen Variante ἕως αἰῶνος übersetzt wird.
Talshir42 neigt eher zu einer textkritischen Auswertung der Über-
setzung von 1 Esdr. Sie schließt zwar nicht aus, dass der Übersetzer von
1 Esdr, ebenso wie der von 2 Esdr, das masoretische ‫ד־עֹולם‬ ָ ‫ ַע‬gelesen
habe. Sie vermutet aber eher, 1 Esdr habe in seiner hebräischen Esravor-
lage das deuteronomische ‫ ָּכל־יָ ֶמיָך‬vorgefunden. Dann würde 1 Esdr einen
hebräischen Vorlagetext bezeugen, der das erste Mal ‫ ָּכל־יָ ֶמיָך‬das zweite
Mal aber ‫ד־עֹולם‬ָ ‫ ַע‬gelesen hätte. 2 Esdr bezeugte einfach den jetzigen MT
mit zweifachem ‫ד־עֹולם‬ ָ ‫ ַע‬.
Tatsächlich wäre τὸν ἅπαντα χρόνον eine für den Übersetzer von 1 Esdr
untypisch freie Wiedergabe von ‫ד־עֹולם‬ ָ ‫ ַע‬, ist doch griechisch χρόνος in
fünf von 10 Fällen die Wiedergabe von hebr. ‫ יָ ִמים‬43, dreimal aber auch
für ‫ זְ ָמן‬44. Insofern ist Talshirs Vermutung, er habe hier ‫ל־הּיָ ִמים‬ ַ ‫ ָּכ‬gele-
sen, durchaus begründet. Die Variante ist wohl von Dtn 23:7 MT her in
den hebräischen Vorlagetext von 1 Esdr gekommen. Sie hat sogar einige
Chancen, den ursprünglicheren Esratext wiederzugeben, da MT eine spä-
tere Assimilation darstellen könnte45. Hier spricht der Dtn-Text für eine

41  Ibid., 79.

42 Talshir, I Esdras. A Text Critical Commentary, 452f.
43 1 Esdr 1:18, 55; 2:15; 5:70; 8:73 // 2 Chr 35:18; 36:21; Esr 4:7; 4:5; 9:7.
44 In 1 Esdr 1:22 ist die Vorlage gänzlich verschieden, in 1:40 möglicherweise auch. In 6:3
gibt χρόνος aramäisch ‫ זְ ָמן‬wieder und in 9:12 hebräisch ‫זְ ָמן‬/‫ ֵעת‬.
45 Die LXX-Fassung der zitierten Dtn-Stelle zeigt, dass weder 1 Esdr noch 2 Esdr auf
diese griechische Dtn-Fassung zurückgegriffen haben. Dtn 23:7 LXX: οὐ προσαγορεύσεις
εἰρηνικὰ αὐτοῖς καὶ συμφέροντα αὐτοῖς πάσας τὰς ἡμέρας σου εἰς τὸν αἰῶνα.
122 dieter böhler sj

­Variation der Zeitbestimmung in der Vorlage von 1 Esdr. Der Variations-

wille des Übersetzers war hier vermutlich gar nicht gefordert.

9. Weitere Schlussfolgerungen aus den Beobachtungen

Eine Passage in 1 Esdr, die in Esr–Neh und damit auch in 2 Esdr keine
Entsprechung hat, ist die sogenannte Pagenerzählung 1 Esdr 3:1–5:6, eine
Serubbabellegende. Seit Torreys Ezra Studies aus dem Jahr 191046 ist weit-
hin anerkannt, dass sie im Original aramäisch war.
Pohlmann hatte in seinen Studien zu 1 Esdr die Auffassung vertre-
ten, die Pagenerzählung sei von einem anderen Übersetzer ins Griechi-
sche übertragen worden als der Rest von 1 Esdr47. Er verweist auf die
verschiedene Übersetzung vermutlich gleicher Vorlageausdrücke in den
beiden Teilen. Demnach wäre die Pagenerzählung erst auf griechischer
Sprachebene interpoliert worden. Talshir dagegen schreibt die griechi-
sche Übersetzung des ganzen Buches ein und derselben Hand zu48. Das
bedeutet, dass die Interpolation der aramäischen Erzählung schon in die
hebräisch-aramäische Tempelbauerzählung erfolgte.
Pohlmanns Argumente beweisen wegen der fehlenden Äquivalenten-
konstanz unseres Übersetzers wenig. Dagegen spricht die Technik der
doppelten Wiedergabe ein und desselben Vorlageausdrucks auch in 1 Esdr
3–4 für die Hand desselben Übersetzers wie im restlichen Buch.
In 1 Esdr 3:4 werden die drei Pagen des Königs zweimal als ­Leib-wächter
1 Esdr 3:4: οἱ τρεῖς νεανίσκοι οἱ σωματοφύλακες οἱ φυλάσσοντες τὸ σῶμα τοῦ
Das ist, wie Talshir wohl zu Recht vermutet, die doppelte Wiedergabe
eines aramäischen Ausdruckes: „Its first part offers a technical term,
σωματοφύλακες, well attested in the second century B.C.E. (papyri, Polybius).

46 C. Torrey, Ezra Studies (Chicago: University of Chicago Press, 1910 [repr. New York:
KTAV, 1970]), 23–25; K.-F. Pohlmann, Studien zum dritten Esra: Ein Beitrag zur Frage nach
dem ursprünglichen Schluß des chronistischen Geschichtswerks (FRLANT 104; Göttingen:
Vandenhoeck & Ruprecht, 1970), 48.
47 Pohlmann, Studien, 150f.
48 Talshir, 1 Esdras. From Origin to Translation, 103 und 106.
übersetzungstechnik und textkritik in den esdrasbüchern 123

The second part, οἱ φυλάσσοντες τὸ σῶμα τοῦ βασιλέως, presumably repre-

sents the wording of the Vorlage more closely“49.
Das spricht dafür, dass ein und derselbe Übersetzer das ganze Buch
übersetzt hat, mithin die aramäische Pagenerzählung schon in die hebrä-
isch-aramäische Vorlage integriert war.
Nach 1 Esdr 3:6 soll der Sieger des Pagenwettstreits „aus güldenen (Gefä-
ßen) trinken, auf einem Bett aus Gold schlafen und einen goldgezäumten
Wagen (fahren)“:
1 Esdr 3:6: ἐν χρυσώμασιν πίνειν καὶ ἐπὶ χρυσῷ καθεύδειν καὶ ἅρμα χρυσοχάλινον
Im Aramäischen stand hier zweifellos dreimal ‫ ְד ַהב‬. Der Übersetzer der
Pagenerzählung findet drei Variationen.
Nach Talshir50 liegt auch dem griechischen κυριεύει und δεσπόζει in 1
Esdr 4:3 doppeltes aramäisches šlt zu Grunde, ebenso wie dann erneut in
4:14. Der Übersetzer hätte variiert.
Hinter κατ᾽ ἐνιαυτὸν in 1 Esdr 4:51 und 4:52 steht aramäisch šnh bšnh51.
All diese Phänomene, die wir im übrigen 1 Esdr beobachtet haben,
zeigen sich anscheinend auch in der Pagenerzählung, für die wir keine
aramäische Vorlage besitzen. Das stützt die These ihrer Einfügung auf
semitischer Sprachebene52.

10. Ein Blick auf Daniel (MT, LXX, Th)

Der uns aus Esra bekannte Doppelausdruck ‫ יֹום ְּביֹומֹו‬findet sich auch in
Dan 1:5. DanTh schreibt dafür τῆς ἡμέρας καθ᾽ ἡμέραν. DanLXX dagegen καθ᾽
ἑκάστην ἡμέραν. Dieser geht vor wie 1 Esdr, DanTh wie 2 Esdr. Aber beide
hatten die gleiche Vorlage.
In Dan 2:47 spricht Nebukadnezzar von Gott als dem, der Geheimnisse
enthüllen kann, weil Daniel das Geheimnis des Königs enthüllen konnte.
Der aramäische Text spricht zweimal von glh.
Dan 2:47: ‫וְ גָ ֵלה ָרזִ ין ִּדי יְ ֵכ ְל ָּת ְל ִמגְ ֵלא ָרזָ ה ְדנָ ה‬. DanTh gibt das ohne jede
Variation wieder mit καὶ ἀποκαλύπτων μυστήρια ὅτι ἠδυνήθης ἀποκαλύψαι
τὸ μυστήριον τοῦτο. DanLXX variiert das Verb: ὁ ἐκφαίνων μυστήρια κρυπτὰ

49 Talshir, I Esdras. A Text Critical Commentary, 138. Sie denkt an die Vorlage dy ntryn
50 Ibid., 174 und 189f.
51  Vgl. ibid., 233f.
52 Wo der Übersetzer aramäische Wortwiederholungen nachahmt, wie in 4:7–9 und
4:37, ist die rhetorische Absicht der Wiederholungen offensichtlich.
124 dieter böhler sj

μόνος ὅτι ἐδυνάσθης δηλῶσαι τὸ μυστήριον τοῦτο. Neben dem allgemeinen

Variationswillen den der Übersetzer mit 1 Esdr teilt, mag auch die Überle-
gung mitgespielt haben, dass er göttliches und menschliches Offenbaren
nicht mit demselben Verbum beschreiben wollte. Eine textkitische Frage
stellt sich hier—entgegen dem ersten Anschein—nicht.
Nach Dan 1:8 war Daniel entschlossen, sich durch die königliche Ernäh-
rung nicht zu verunreinigen und bittet den Chefbeamten, sich nicht ver-
unreinigen zu müssen. Im hebräischen Text steht zweimal g’l hitp.
Dan 1:8: ‫ּוביֵ ין‬
ְ ‫ל־לּבֹו ֲא ֶׁשר לֹא־יִ ְתּגָ ַאל ְּב ַפ ְת ַּבג ַה ֶּמ ֶלְך‬
ִ ‫וַ ּיָ ֶׂשם ָּדנִ ּיֵ אל ַע‬
‫יסים ֲא ֶׁשר לֹא יִ ְתּגָ ָאל׃‬ ִ ‫ִמ ְׁש ָּתיו וַ יְ ַב ֵּקׁש ִמ ַּׂשר ַה ָּס ִר‬
DanTh übersetzt: καὶ ἔθετο Δανιηλ ἐπὶ τὴν καρδίαν αὐτοῦ ὡς οὐ μὴ ἀλισγηθῇ
ἐν τῇ τραπέζῃ τοῦ βασιλέως καὶ ἐν τῷ οἴνῳ τοῦ πότου αὐτοῦ καὶ ἠξίωσε τὸν
ἀρχιευνοῦχον ὡς οὐ μὴ ἀλισγηθῇ.
DanLXX variiert das Verb: καὶ ἐνεθυμήθη Δανιηλ ἐν τῇ καρδίᾳ ὅπως
μὴ ἀλισγηθῇ ἐν τῷ δείπνῳ τοῦ βασιλέως καὶ ἐν ᾧ πίνει οἴνῳ καὶ ἠξίωσε τὸν
ἀρχιευνοῦχον ἵνα μὴ συμμολυνθῇ. Die Kenntnis seiner Übersetzungsweise
erübrigt textkritische Bedenken.
In Dan 1:2 steht dreimal „Gotteshaus.“ Zunächst bezeichnet ‫ֹלהים‬ ִ ‫ית־ה ֱא‬
ָ ‫ֵב‬
den Jerusalemer Tempel. Dann ist ‫ֹלהיו‬ ָ ‫ ֵּבית ֱא‬ein babylonischer Götzen-
tempel. Und schließlich meint ‫ֹלהיו‬ ָ ‫אֹוצר ֱא‬
ַ ‫ ֵּבית‬die Schatzkammer in
diesem Götzentempel. DanTh schreibt dreimal οἶκος τοῦ θεοῦ und bezeugt
exakt den MT. DanLXX macht aus dem ersten „Gotteshaus“ ein ἱερόν, setzt
aber nicht einen Genitiv („Geräte des Tempels“), sondern ein Adjektiv
(„Tempelgeräte“) wie es 1 Esdr unzählige Male tut53. Birte Braasch ver-
kennt in ihrer Dissertation über DanLXX die Übersetzungstechnik, wenn
sie meint, die alte Danielübersetzung betone hier die „Heiligkeit“ der
Geräte, stelle „jedoch keinen expliziten Bezug zum Tempel her“54.
An der zweiten Stelle hat die Vorlage von DanLXX offensichtlich das
hebräische ‫ֹלהיו‬ ָ ‫ ֵּבית ֱא‬nicht gelesen. An der dritten Stelle aber hat er
ָ ‫ ֵּבית ֱא‬oder ‫ֹלהיו‬
ָ ‫אֹוצר ֱא‬
ַ ‫ ֵּבית‬gelesen, gibt es aber mit τῷ εἰδωλίῳ
αὐτοῦ wieder, weil ein Götzentempel für ihn nicht dieselbe Vokabel

53 Talshir, 1 Esdras. From Origin to Translation, 211; Böhler, „ ‚Treu und schön‛,“ 102.
54 B. Braasch, Die LXX-Übersetzung des Danielbuchs—eine Orientierungshilfe für das
religiöse und politisch-gesellschaftliche Leben in der ptolemäischen Diaspora (Diss. Ham-
burg, 2004. Veröffentlicht unter URN: urn:nbn:de:gbv:18–21588; URL: http://www.sub, 37f.
übersetzungstechnik und textkritik in den esdrasbüchern 125

verdient wie der Jerusalemer Tempel des wahren Gottes. Exakt dieselbe
Differenzierung hatte 1 Esdr in 2:7 auch vorgenommen55.

Dan 1:2 LXX Th

ִ ‫ית־ה ֱא‬
ָ ‫ּומ ְק ָצת ְּכ ֵלי ֵב‬ִ καὶ μέρος τι τῶν ἱερῶν καὶ ἀπὸ μέρους τῶν
‫ץ־ׁשנְ ָער ֵּבית‬
ִ ‫יאם ֶא ֶר‬ ֵ ‫וַ יְ ִב‬ σκευῶν τοῦ κυρίου καὶ σκευῶν οἴκου τοῦ θεοῦ
ָ ‫ֱא‬ ἀπήνεγκεν αὐτὰ εἰς καὶ ἤνεγκεν αὐτὰ εἰς
‫ת־ה ֵּכ ִלים ֵה ִביא ֵּבית‬
ַ ‫וְ ֶא‬ Βαβυλῶνα καὶ ἀπηρείσατο γῆν Σεννααρ οἶκον τοῦ
ָ ‫אֹוצר ֱא‬ַ αὐτὰ ἐν τῷ εἰδωλίῳ αὐτοῦ θεοῦ αὐτοῦ καὶ τὰ σκεύη
εἰσήνεγκεν εἰς τὸν οἶκον
θησαυροῦ τοῦ θεοῦ αὐτοῦ

11. Schluss

Wer dolmetschen will, hatte Luther gesagt, muss einen großen Vorrat an
Wörtern haben, damit er das je passende Äquivalent setzen kann. Das
gilt für einen dezidiert zielsprachenorientierten Übersetzer wie den von
1 Esdr (und DanLXX). Sie übersetzen exakt und keineswegs frei, variieren
aber in der Äquivalentenwahl—und oft sehr überlegt. Wer diese Technik
kennt, kann besser unterscheiden zwischen der Bezeugung textkritischer
Varianten und bloßer Variation in einer entschieden zielsprachenorien-
tierten Übersetzung. 2 Esdr und DanTh ersparen uns solche Überlegungen,
aber sie geben textkritisch ohnehin wenig her und „grosse vorrath von
worten“, wie Luther meinte, brauchten die auch nicht.

55 1 Esdr 2:7: καὶ ὁ βασιλεὺς Κῦρος ἐξήνεγκεν τὰ ἱερὰ σκεύη τοῦ κυρίου ἃ μετήγαγεν
Ναβουχοδονοσορ ἐξ Ιερουσαλημ καὶ ἀπηρείσατο αὐτὰ ἐν τῷ ἑαυτοῦ εἰδωλίῳ.
In Search of the Old Greek Text of 4 Maccabees

Robert J. V. Hiebert

1. Introduction

The story of the martyrdoms, during the reign of Antiochus IV Epipha-

nes, of an elderly Jewish scribe (γραμματεύς—2 Macc 6:18) or priest and
lawyer (ἱερεύς, νομικός—4 Macc 5:4) named Eleazar and of seven broth-
ers and their mother is told in 2 Macc 6:18–7:42, and in a considerably
expanded version, in 4 Macc 5–18. Antiochus and his officials are deter-
mined to make the Jews “forsake their ancestral laws” (2 Macc 6:1), to
force them “to renounce Judaism by tasting defiling foods” (4 Macc 4:26),
to compel them “to taste pork and food sacrificed to idols” (4 Macc 5:2),
and thereby to signal their willingness “to change over to Greek customs”
(2 Macc 6:9). Antiochus orders that the most barbarous forms of torture
should be inflicted upon the eight males, beginning with Eleazar, in his
attempt to make them yield. When they do not, but instead they defy him
boldly, he has them put to death. According to the 4 Maccabees version,
the mother of the seven brothers is the last one of the group still alive,
and she throws herself into the fire “so that no one might touch her body”
(4 Macc 17:1).1 In both versions, the martyrs’ courage and faithful obedi-
ence “to the holy God-given law” (2 Macc 6:23) are celebrated. In 4 Mac-
cabees, their story is recounted in detail to support the author’s thesis in
his philosophical treatise, i.e., that αὐτοδέσποτός ἐστιν τῶν παθῶν ὁ εὐσεβὴς
λογισμός “pious reason is absolute master of the passions” (4 Macc 1:1; cf.
vv. 7, 13, 30 passim).
Greek 4 Maccabees is the only one of the four books that traditionally
include the name Maccabees in the title for which to date there is no criti-
cal edition in the Göttingen Septuaginta series. I am currently at work to
remedy that lacuna, and a first draft of the critically-reconstructed text,

1 2 Macc 7:41 reports simply that “[l]ast of all, the mother died, after her sons,” without
offering any details as to the circumstances of her death. Throughout this paper, English
translations of Septuagint texts are, unless otherwise indicated, taken from A. Pietersma
and B. G. Wright, eds., A New English Translation of the Septuagint (New York / Oxford:
Oxford University Press, 2007), commonly referred to as NETS. My own translations of
texts are written in italics.
128 robert j. v. hiebert

based on the manuscript evidence of more than 70 extant Greek man-

uscripts along with the Syriac daughter version, is nearing completion.
This is not to say that there have been no preliminary efforts to recover
the wording of the original author. The first such undertaking was that of
Otto F. Fritzsche, whose 1871 edition is based mainly on the great codi-
ces Sinaiticus (S) and Alexandrinus (A)—dated to the fourth and fifth
centuries, respectively—though variant readings from close to a dozen
other manuscripts are recorded in the apparatus.2 A little more than two
decades later, the first edition of Henry B. Swete’s The Old Testament in
Greek according to the Septuagint appeared. Swete produced an essentially
diplomatic edition of 4 Maccabees in which he relied primarily on A, but
whose carelessly written text he corrected on the basis of S and Codex
Venetus (V), an eighth or ninth century uncial.3 In Alfred Rahlfs’ edition
of the Septuagint published in 1935, the text of 4 Maccabees is an eclectic
one that is based on S and A, with preference frequently being given to
S when the two diverge, and with substantive variants from V included
in the apparatus as well.4 The German translation of 4 Maccabees by
H.-J. Klauck, published in 1989, represents an advance in our understand-
ing of the text of this book as it left the hands of the original author.5
He renders a Greek text that has benefitted from text-critical proposals
communicated to him by Robert Hanhart, proposals that are based on
collations that have been carried out at the Septuaginta-Unternehmen in
The following list of Greek manuscripts and manuscript groups is based
upon the work of Klauck and Hanhart, but it has also been modified in
the light of my research and in consultation with Detlef Fraenkel of the
Uncials: A S V
 A 542 (11:5 – fin libri)
L: 236 491 534 728
q: 71 74 120 370 380 452 731 3002

2 O. F. Fritzsche, ed., ΜΑΚΚΑΒΑΙΩΝ ΤΕΤΑΡΤΟΣ, Libri apocryphi Veteris Testamenti

graece (Leipzig: Brockhaus, 1871), 351–386.
3 H. B. Swete, ed., The Old Testament in Greek according to the Septuagint, vol. 3:
Hosea—4 Maccabees, Psalms of Solomon, Enoch, The Odes (3rd ed.; Cambridge: Cambridge
University Press, 1894; 1905), vi.
4 A. Rahlfs, ed., Septuaginta: Id est Vetus Testamentum graece iuxta LXX interpretes
(Stuttgart: Württembergische Bibelanstalt, 1935).
5 H.-J. Klauck, 4. Makkabäerbuch ( JSHRZ 3.6; Gütersloh: Mohn, 1989), 641–763.
6 Ibid., 678–679.
in search of the old greek text of 4 maccabees 129

q1: 44 107 610

q2: 55 747
q´’ = q + q1 + q2 q´ = q + q1 q’ = q + q2 etc.
m: 316 317 322 325 391 397 446 457 467 472 473 586 591 592 594 595 596 597
607 617 639 640 641 656 677 682 683 686 695 699 713 714 774 778 782 789
m1: 455 585
m2: 587 738
m3 (init libri – 11:4): 62 542 747c/mg
m´’` = m + m1 + m2 + m3 m´ = m + m1 m’ = m + m2 m` = m + m3 etc.
11:5 – fin libri: m´’ 62 747c/mg
Codices mixti: 46 (52 332) 58 340 577 668 690 741 771 773 930
A few explanatory comments regarding these manuscript groupings are
in order.

1. Manuscript groups are delineated on the basis of regular patterns of

agreement among witnesses with regard to divergences from what the
editor of a Septuaginta edition reconstructs as the original text, based
on a careful assessment of the extant manuscript evidence.
2. The uncials, which are the most consistent witnesses to the original
text of 4 Maccabees, do not really constitute a manuscript group inas-
much as they often part company with one another when they do not
bear witness to the original text.
3. The four m groups (m´’`) consist of manuscripts that are either compo-
nents of Greek menologia (collections of lives of the saints arranged by
feast days) or that belong to generally the same textual tradition.
4. From the beginning of 4 Maccabees until 11:4, there exists a group that
I have called m3 and that consists of manuscripts 62 542 and 747c/mg.
Thereafter, 542 becomes aligned with A while 62 and 747c/mg are affili-
ated with m + m1 + m2 (m´’).
5. The original text of manuscript 747 belongs, with that of manuscript 55,
to the q2 group. However, 747 contains a host of corrections or alter-
native readings that are located either between the lines of the text
or in the margins. The majority of these readings are attributable to
a single scribe who did his work after the original scribe penned the
manuscript. The interlinear readings are designated 747c while mar-
ginal readings are designated 747mg, though, as indicated above, both
sets of readings are to be situated within the larger m tradition.
6. The manuscripts and groups that most frequently attest the original
text of 4 Maccabees are A S V L q q1 q2 46 (52 332) 58 340 577 668 690
741 771 773 930. The four m groups, however, are characterized by per-
vasive recensional activity.
130 robert j. v. hiebert

2. Daughter versions of Septuagint 4 Maccabees

Text-critically relevant translations of 4 Maccabees are few in number. In

fact, the Syriac version (Sy) is the only one that is fully extant. An eclec-
tic edition based on nine manuscripts was published by R. L. Bensly and
W. E. Barnes.7 A preliminary probe of the degree of affiliation between Sy
and the Greek manuscripts of the book conducted by means of my Web
Application for Textual and Exegetical Research (WATER) indicates that
the highest percentages of agreement are only in the range of about 7.5%.
Thus, for example, Sy agrees with manuscript 741 on 7.55% of the 93 read-
ings attested by one or both of Sy and 741, and Sy agrees with manuscript
668 on 7.54% of the 96 readings attested by one or both of Sy and 668.
Barnes reports that Sy generally agrees with S but seldom with A.8 The
statistics generated by WATER are as follows: S-Sy (89 readings—6.50%);
A-Sy (81 readings—6.01%); V-Sy (45 readings—5.46%). This indicates that
Sy is not very closely affiliated with any other witnesses to the text of this
book. While the Syriac version of 4 Maccabees is, to use Barnes’ terminol-
ogy, a “generally . . . faithful” translation of the Greek,9 it is not the sort of
hyper-literal version that Paul of Tella produced in his seventh century
Syriac translation of Origen’s Hexapla. Instead, it moves “back and forth
along a continuum from the reasonably literal to the freely interpretative.”10
What Hanhart says about the Syriac version of 3 Maccabees in his edition
of the Septuagint of that book is applicable to 4 Maccabees: “In Sy ist der
Text oft frei ausgestaltet, so daß die Textform der Vorlage nur noch schwer
erkennbar ist.”11 Relevant textual evidence from Syriac 4 Maccabees will,
however, be incorporated in the critical edition of this book.
As for other versions, Klauck says: “Nur Hinweise gibt es bislang auf
eine slavische und eine koptische Version.”12 E. Lucchesi reports that only

  7 R. L. Bensly and W. E. Barnes, eds., The Fourth Book of Maccabees and Kindred Docu-
ments in Syriac (Cambridge: Cambridge University Press, 1895).
  8 Ibid., xiv.
  9 Ibid.
10 R. J. V. Hiebert, “Preparing a Critical Edition of IV Maccabees: The Syriac Translation
and Passio Sanctorum Machabaeorum as Witnesses to the Original Greek,” in Interpreting
Translation: Studies on the LXX and Ezekiel in Honour of Johan Lust (BETL 192; eds. F. García
Martínez and M. Vervenne; Leuven: Peeters, 2005), 193–216 (202).
11  R. Hanhart, ed., Septuaginta: Vetus Testamentum graecum, IX, 3: Maccabaeorum liber
III (2nd ed.; Göttingen: Vandenhoeck & Ruprecht, 1980), 9.
12 Klauck, 4. Makkabäerbuch, 679–680.
in search of the old greek text of 4 maccabees 131

fragments of a Coptic version have survived,13 and so it will be of very

limited text-critical usefulness. The Slavic version will not be used at all
for the Göttingen edition.
Finally, there is also a fourth century Latin composition known as Pas-
sio Sanctorum Machabaeorum that is based on 4 Maccabees.14 Because it
is a free adaptation rather than a bona fide translation, however, it does
not have much text-critical value.

3. Selected examples involving text-critical decision-making

As is well known, the work of a textual critic involves judicious evalua-

tion of both external and internal evidence in extant textual witnesses. In
other words, in any given context one must make a decision with regard
to which reading is original and which one(s) is/are secondary, based both
on the age and reputation of the available manuscripts and on the assess-
ment of the probabilities with respect to the priority of one reading over
another/others. In the following discussion, I will focus on certain pas-
sages in which this process may be illustrated. I will begin in each case
by citing enough of the texts of Rahlfs’ edition and the NETS version to
provide context for the particular text-critical problem to be discussed.

1. ἀνδραγαθίας vs. καλοκαγαθιας (1:8)

Rahlfs 1:7–9
7 πολλαχόθεν μὲν οὖν καὶ ἀλλαχόθεν ἔχοιμ᾿ ἂν ὑμῖν ἐπιδεῖξαι ὅτι αὐτοκράτωρ ἐστὶν τῶν
παθῶν ὁ λογισμός, 8 πολὺ δὲ πλέον τοῦτο ἀποδείξαιμι ἀπὸ τῆς ἀνδραγαθίας τῶν ὑπὲρ
ἀρετῆς ἀποθανόντων, Ελεαζαρου τε καὶ τῶν ἑπτὰ ἀδελφῶν καὶ τῆς τούτων μητρός. 9
ἅπαντες γὰρ οὗτοι τοὺς ἕως θανάτου πόνους ὑπεριδόντες ἐπεδείξαντο ὅτι περικρατεῖ
τῶν παθῶν ὁ λογισμός.
NETS 1:7–9
7 On the basis of many and diverse considerations I could show you that reason
is absolute ruler of the passions, 8 but I can demonstrate it much better from the
bravery (ἀνδραγαθίας) of those who died for the sake of virtue (ἀρετή): Eleazaros,

13 E. Lucchesi, “Découverte d’une traduction copte du Quatrième livre des Maccabées
(BHG 1006),” Analecta Bollandiana 99 (1981): 302; idem, “Encore trois feuillets coptes du
Quatrième livre des Maccabées,” Écritures et traditions dans la littérature copte (Cahiers de
la Bibliothèque Copte 1; Leuven: Peeters, 1983), 21–22.
14 H. Dörrie, Passio SS. Machabaeorum (Abhandlungen der Gesellschaft der Wissen-
schaften zu Göttingen, Philologisch-historische Klasse 3/22; Göttingen: Vandenhoeck &
Ruprecht, 1938).
132 robert j. v. hiebert

the seven brothers and their mother. 9 All of these, in despising sufferings to the
point of death, showed that reason has full control over the passions.
1:8 ἀνδραγαθίας (-θειας V*mg Fritzsche Swete) A L q-q1-q2 m-m3 46 52 58 332 340
577 668 771 Sy Rahlfs Klauck] ανδρειας V* 690; καλοκαγαθιας S 74115

The term ἀνδραγαθία occurs nowhere else in 4 Maccabees, though it does

five times in 1 Maccabees—always in conjunction with πόλεμος (5:56; 8:2;
9:22; 10:15; 16:23)—and once in 2 Maccabees in association with ἀγών.16 In
4 Macc 1:10 the author does state that the men and their mother “died
for the sake of nobility of character” (καλοκαγαθία) and that it is there-
fore appropriate for him to praise them for their virtues (ἀρετῶν). The
term καλοκαγαθία occurs four other times in 4 Maccabees (3:18; 11:22; 13:25;
15:9) but nowhere else in the Septuagint. We therefore have a situation in
4 Macc 1:8 in which the great majority of witnesses attest ἀνδραγαθίας, a
term that does not occur elsewhere in the book but that suits the con-
text well—what with the talk of dying “for the sake of virtue” (v. 8) and
“despising sufferings to the point of death” (v. 9)—while just S and the
eleventh/twelfth century manuscript 741 attest καλοκαγαθιας, a term that
is found in the Septuagint only in 4 Maccabees and that might have come
to replace the original reading ἀνδραγαθίας because of its usage elsewhere
in the book and/or because of a tendency in the transmission history to
enhance the noble character of the martyrs even more. I would, therefore,
go with the majority text and the various listed editions (and against S) in
opting for ἀνδραγαθίας as original.

2. καὶ περὶ τὸ σῶμα vs. ø (1:20)

Rahlfs 1:20
παθῶν δὲ φύσεις εἰσὶν αἱ περιεκτικώταται δύο ἡδονή τε καὶ πόνος· τούτων δὲ ἑκάτερον
καὶ περὶ τὸ σῶμα καὶ περὶ τὴν ψυχὴν πέφυκεν.
NETS 1:20
Of the passions, the two most comprehensive types are pleasure and pain, and
each of these pertains by nature both to the body and to the soul.
1:20 καὶ περὶ τὸ σῶμα V L 370c m-m1-m2-m3-747c 577 690 Rahlfs Klauck] > A S
q(370*)-q1-q2(747*) 46 52 58 332 340 668 741 Fritzsche Swete Breitenstein

15 As is the case in the critical editions of the Göttingen Septuaginta series, throughout
this paper lemma readings include breathing marks and accents, variant readings do not.
16 Apart from the preceding references, ἀνδραγαθία is found in the Septuagint only in
Esth 10:2.
in search of the old greek text of 4 maccabees 133

Urs Breitenstein comes to the following conclusion regarding this pas-

sage: “Die von Rahlfs verworfene Lesart der wichtigsten Handschriften
Sinaiticus und Alexandrinus . . . würde ich als lectio difficilior vorziehen.
Sie wäre als Spitze gegen Epikur (wie manches in der Philosophie des
Ps-Ios [Pseudo-Iosephos]) zu verstehen, da nach Epikur alle ἡδοναί,
ἐπιθυμίαι, πόνοι (primär) körperbezogen sind.”17 David deSilva, who trans-
lates the text of S for his commentary in the Septuagint Commentary
Series published by Brill, renders this verse with the shorter reading: “The
two most comprehensive species of passions are pleasure and pain, and
both of these have grown up even around the soul” (τούτων δὲ ἑκάτερος*
[ἑκάτερονc] καὶ περὶ τὴν ψυχὴν πέφυκεν).18 Hanhart, in a communication
to Klauck, however, comments: “Aber es handelt sich doch eher um einen
Abschreibfehler aufgrund eines Homoioteleuton zwischen καί und καί
(Hanhart).”19 Hanhart is likely correct. It should be noted that in 4 Macc
1:28, pleasure (ἡδονή) and pain (πόνος) are, as in the longer reading of v. 20,
associated with both body (σῶμα) and soul (ψυχή): καθάπερ οὖν δυεῖν τοῦ
σώματος καὶ τῆς ψυχῆς φυτῶν ὄντων ἡδονῆς τε καὶ πόνου πολλαὶ τούτων τῶν
φυτῶν εἰσιν παραφυάδες “Just as pleasure and pain are two plants growing
from the body and the soul, so there are many offshoots of these plants.”
In that light, Breitenstein’s suggestion that the shorter reading in 4 Macc
1:20 is to be accepted as the original reading, and that it is to be regarded
as a challenge to Epicurus’s idea that all pleasures, desires, and suffer-
ing are primarily associated with the body, appears to be untenable. Thus
I would favour the longer reading attested by the majority of manuscripts
and included in Rahlfs’ and Klauck’s editions, and go against A S etc.
whose shorter reading is preferred by Fritzsche, Swete, and Breitenstein.

3. ἔργων vs. ετερων (2:9)

Rahlfs 2:8–9
8 αὐτίκα γοῦν τῷ νόμῳ πολιτευόμενος, κἂν φιλάργυρός τις ᾖ, βιάζεται τὸν αὑτοῦ
τρόπον τοῖς δεομένοις δανείζων χωρὶς τόκων καὶ τὸ δάνειον τῶν ἑβδομάδων ἐνστασῶν
χρεοκοπούμενος· 9 κἂν φειδωλός τις ᾖ, ὑπὸ τοῦ νόμου κρατεῖται διὰ τὸν λογισμὸν μήτε

17 U. Breitenstein, Beobachtungen zu Sprache, Stil und Gedankengut des Vierten Makka-
bäerbuchs (Basel / Stuttgart: Schwabe & Co. Verlag, 1976), 136 n. 1.
18 D. A. deSilva, 4 Maccabees: Introduction and Commentary on the Greek Text in Codex
Sinaiticus (Septuagint Commentary Series; eds. S. E. Porter, R. S. Hess, and J. Jarick; Leiden /
Boston: Brill, 2006), 4–5.
19 Klauck, 4. Makkabäerbuch, 692 n. 20c.
134 robert j. v. hiebert

ἐπικαρπολογούμενος τοὺς ἀμητοὺς μήτε ἐπιρρωγολογούμενος τοὺς ἀμπελῶνας. ¶ Καὶ

ἐπὶ τῶν ἑτέρων δὲ ἔστιν ἐπιγνῶναι τοῦτο, ὅτι τῶν παθῶν ἐστιν ὁ λογισμὸς κρατῶν·
NETS 2:8–9
8 As soon, indeed, as one adopts a way of life in accordance with the law, even
though a lover of money, one overpowers one’s own bent, lending without inter-
est to the needy and canceling the debt when the seventh year arrives. 9 Even if
one is stingy, one is overcome by the law through reason, neither gleaning the
harvest nor gathering the last of the vineyards’ grapes. ¶ And in all other matters
it is possible to recognize this principle, that reason overcomes the passions.
2:9 ἑτέρων A q-q1-610 46 52 58 340 668 741 771 Fritzsche Swete Rahlfs Klauck] εργων
S V L 610-q2 m-640*-m1-m2-m3 577 690 Sy; > 640*

David deSilva, who translates S, renders the relevant sentence in the lat-
ter part of v. 9: “And on the evidence of deeds, we recognize this—that
reason restrains the passions.”20 The external evidence of the textual wit-
nesses for both εργων and ἑτέρων is strong, so the original reading must
be reconstructed on the basis of internal evidence. It seems likely that
εργων represents a secondary development in the textual history in this
context, whereby the general reference to other (ἑτέρων) manifestations
of living “in accordance with the law” (v. 8) would have been replaced
by the more specific terminology of deeds (εργων) that embody that law.
Furthermore, given the orthographic similarity of the two words (espe-
cially their first and last parts), it is not hard to envision a scenario early
in the textual history in which textual corruption could have given rise
to the alternate reading. So in this case I would side with the editions of
Fritzsche, Swete, Rahlfs, and Klauck, which feature the minority reading
of A etc., and against the majority reading of S V etc.

4. εἴποιεν ἄν τινες vs. ειποι αν τις (1:5) / εἴποι τις ἄν vs. αν τις ειποι (2:24)

Rahlfs 1:5–6
5 πῶς οὖν, ἴσως εἴποιεν ἄν τινες, εἰ τῶν παθῶν ὁ λογισμὸς κρατεῖ, λήθης καὶ ἀγνοίας οὐ
δεσπόζει; γελοῖον ἐπιχειροῦντες λέγειν. 6 οὐ γὰρ τῶν αὑτοῦ παθῶν ὁ λογισμὸς κρατεῖ,
ἀλλὰ τῶν τῆς δικαιοσύνης καὶ ἀνδρείας καὶ σωφροσύνης ἐναντίων, καὶ τούτων οὐχ ὥστε
αὐτὰ καταλῦσαι, ἀλλ᾿ ὥστε αὐτοῖς μὴ εἶξαι.
NETS 1:5–6
5 Some might perhaps ask, “How then, if reason overcomes the passions, does it
not master forgetfulness and ignorance?” Their attempt at argument is ridiculous.

20 deSilva, 4 Maccabees, 6–7.

in search of the old greek text of 4 maccabees 135

6 For reason does not overcome its own passions but those opposed to justice,
courage and self-control, and it overcomes these not so that it destroys them but
so that one does not give way to them.
1:5 εἴποιεν ἄν τινες A q2 m-m1-m2-m3 58 332 340 577 690 Fritzsche Swete Rahlfs
Klauck] ειποι (ειπει 491; ειποιεν S* 534 340) αν τις S V L q-q1 46 52 340 668 741
771 Sy

Rahlfs 2:24–3:1
2:24 Πῶς οὖν, εἴποι τις ἄν, εἰ τῶν παθῶν δεσπότης ἐστὶν ὁ λογισμός, λήθης καὶ ἀγνοίας
οὐ κρατεῖ; 3:1 ἔστιν δὲ κομιδῇ γελοῖος ὁ λόγος· οὐ γὰρ τῶν ἑαυτοῦ παθῶν ὁ λογισμὸς
ἐπικρατεῖν φαίνεται, ἀλλὰ τῶν σωματικῶν.
NETS 2:24–3:1
2:24 How is it, then, someone may ask, if reason is master of the passions, that it
does not overcome forgetfulness and ignorance? 3:1 But this argument is entirely
ridiculous, for it is apparent that reason prevails not over its own passions but
over those of the body.
2:24 εἴποι (ειπη 370) τις ἄν A Sc V q-q1-q2 46 52 340 668 741 771 Fritzsche Swete
Rahlfs Klauck] ειποι αν τις L; ειποιτε S*; αν τις ειποι m-m1-m2-m3 58 577 690;
αν ειποι τις 316-473; αν τινες ειποιεν 397-467-686 577; ‫ܐܝܬ ܠܟܘܢ ܠܡܐܡܪ‬
Sy(c var)

This example involves two passages—4 Macc 1:5–6 and 2:24–3:1—that

express similar sentiments and wording. In 1:5, where the manuscript
support for both the singular and the plural readings is quite strong, the
concluding clause in 1:5—γελοῖον ἐπιχειροῦντες λέγειν “Their attempt at
argument is ridiculous”—with its plural participle is the determining fac-
tor in favour of the plural reading at the beginning of the verse attested by
A etc. and adopted in the editions of Fritzsche, Swete, Rahlfs, and Klauck.
In 2:24, however, a singular reading has overwhelming manuscript sup-
port, and the word sequence εἴποι τις ἄν attested by A Sc V etc. and fea-
tured in the same four editions is to be preferred.

5. πρός vs. ως (4:2)

Rahlfs 4:2
ὅθεν ἥκων πρὸς Ἀπολλώνιον τὸν Συρίας τε καὶ Φοινίκης καὶ Κιλικίας στρατηγὸν
ἔλεγεν . . .
NETS 4:2
So he came to Apollonius, governor of Syria, Phoenicia and Cilicia, and said  . . . 
4:2 πρός A L 610-q2 m-m1 46 52 58 340 577 668 690 741 773 Sy Swete Rahlfs Klauck]
ως S V q-q1-610 m3-747c Fritzsche
136 robert j. v. hiebert

David deSilva translates the S reading as follows: “For this reason, he went
when Apollonius [was] the governor of Syria, Phoenicia, and Cilica (sic),
and said. . . .”21 He comments that ως requires an implied verb “was,” as in
“when Apollonius [was] the governor.”22 But πρός would appear to make
better sense in this context. I wonder if a corruption developed early as
a result of attraction to the ending of the preceding word ἥκων that was
then mistakenly followed by ως instead of πρός.

6. ἱερέων vs. γεραιων (4:9)

Rahlfs 4:9–10
9 τῶν δὲ ἱερέων μετὰ γυναικῶν καὶ παιδίων ἐν τῷ ἱερῷ ἱκετευσάντων τὸν θεὸν
ὑπερασπίσαι τοῦ ἱεροῦ καταφρονουμένου τόπου 10 ἀνιόντος τε μετὰ καθωπλισμένης
τῆς στρατιᾶς τοῦ Ἀπολλωνίου πρὸς τὴν τῶν χρημάτων ἁρπαγὴν οὐρανόθεν ἔφιπποι
προυφάνησαν ἄγγελοι περιαστράπτοντες τοῖς ὅπλοις καὶ πολὺν αὐτοῖς φόβον τε καὶ
τρόμον ἐνιέντες.
NETS 4:9–10
9 While the priests, with women and children, were imploring God in the temple
to shield the holy place that was being treated with contempt 10 and while Apol-
lonius was going up with his armed forces to seize the funds, angels on horseback
appeared from heaven with lightning flashing from their weapons, instilling in
them great fear and trembling.
4:9 ἱερέων A q-q1-q2 m 46 52 58 340 577 668 771 Fritzsche Swete Rahlfs Klauck]
γεραιων (γερεων 491 728) S L 542-747mg 741; γηραιων (γιρ. V*; -ρεων m1) V m1-62
690; ‫ ̈ܣܒܐ‬Sy

This case presents an interesting situation. On the one hand, the textual
support for a reference to elders γεραιων/γηραιων/‫ ̈ܣܒܐ‬is quite strong. Fur-
thermore, it might be argued that such a reading would constitute a lectio
difficilior, given that priests might be expected to be in the temple “implor-
ing God” rather than elders. On the other hand, the difference between
γεραιων—or more specifically the itacized version of it (γερεων)—and
ἱερέων is one letter, and in uncial characters the two terms could appear to
be even more similar than in minuscules: ΙΕΡΕΩΝ/ΓΕΡΕΩΝ. So Fritzsche,
Swete, Rahlfs, and Klauck would presumably have concluded that an early
scribal error resulted in a change from priests in the temple to elders. That
palaeographical factor would, therefore, appear to be determinative. How
the reference to elders would have been interpreted once γεραιων came to

21 Ibid., 12–13.
22 Ibid., 13 n. 3.
in search of the old greek text of 4 maccabees 137

be part of the textual history, one can only speculate. Could it have been
that this reading was received favourably by some who had differences
with the priestly hierarchy, especially with those who, like Jason in the
time of Antiochus IV (4 Macc 4:15–20), were seen to be complicit with
Gentile authorities in exercising authority over the populace?

7. σφονδυλ- vs. σπονδυλ- (10:8 and 11:18)

Rahlfs 10:8
καὶ εὐθέως ἦγον ἐπὶ τὸν τροχόν, περὶ ὃν ἐκ σπονδύλων ἐκμελιζόμενος
NETS 10:8
Then they immediately brought him [the third brother] to the wheel. On it he
was disjointed at his backbone.
10:8 σφονδύλων A 728 q-452-q1-610-q2 317-542-682* 340 577 668 771 Swete] σφονδυλου
773; σπονδυλων S L-728 452-610 m-317 682*-m1-m2-62 46 52 58 690 741 Fritzsche

Rahlfs 11:18
ἐφ᾿ οὗ κατατεινόμενος ἐπιμελῶς καὶ ἐκσπονδυλιζόμενος
NETS 11:18
He [the sixth brother] was stretched tight upon it with great diligence; his back
was broken
11:18 ἐκσφονδυλιζόμενος (εκσφενδ. 457 592* 699; εκφονδ. 597*; -διλ. 71) A-542 491
q-452-q1-q2 m-316 397 467 473 586 640 686 789-m1 340 668c Fritzsche Swete] εκσπονδ.
(εκπονδ. 62) S 452 62-316-397-467-473-586-640-686-789-m2 46 52 58 668* 690
741 Rahlfs; σφονδ. L-491 577; μηεφονδυλιμενον (sic) 771

There is relatively strong manuscript support for both σφονδυλ- and

σπονδυλ- in 10:8 and for εκσφονδυλ- and εκσπονδυλ- in 11:18. A search in
Thesaurus linguae graecae reveals that the σφονδυλ- root occurs some
774 times in that database, while σπονδυλ- occurs 936 times.23 As Henry
St. J. Thackeray points out, Attic Greek attests forms in σφ-, whereas Ionic
and some κοινή writers use σπ-.24 In the only other place in the Septuagint

23 A Thesaurus linguae graecae (TLG) search yields two occurrences of the εκσφονδυλ-
root, and three of εκσπονδυλ-.
24 H. St. J. Thackeray, A Grammar of the Old Testament in Greek according to the Sep-
tuagint (Cambridge, Cambridge University Press, 1909), 106 (§7, 17). The LSJ entry for this
term is σφόνδυλος, but it is followed by the note “Ion. and later Greek σπόνδυλος” (The
Online Liddell-Scott-Jones Greek-English Lexicon [OLSJ], σφόνδυλος).
138 robert j. v. hiebert

in which any form of the σφονδυλ-/σπονδυλ- root is found (Lev 5:8), the
editions of Wevers, Swete, and Rahlfs all have σφονδυλ-:
σφονδύλου 802 Wevers Swete Rahlfs] σπονδυλου (-δηλου 528 125 646) 58-72-
707(mg)(vid) C´’ 118´-537 125 730 74-76 527 318 628´ 18 55c 59 424 646´
In 4 Maccabees, therefore, it would seem to be easier to account for the
more frequently occurring and later form σπονδυλ- as a secondary replace-
ment of the original σφονδυλ- than vice versa. In 4 Maccabees, it appears
that Rahlfs relied too much on the testimony of Sinaiticus when he opted
for the σπονδυλ- form.

8. βασανισμόν vs. + παραιτεισθαι (11:2)

Rahlfs 11:2
Οὐ μέλλω, τύραννε, πρὸς τὸν ὑπὲρ τῆς ἀρετῆς βασανισμὸν παραιτεῖσθαι
NETS 11:2
Tyrant, I am not about to beg to be excused from torture for the sake of virtue.
Klauck 11:2
“Ich will mich, Tyrann, zu den Foltern um der Tugund willen nicht erst auffordern
Scarpat 11:2
“Non mi rifiuterò, o tiranno, all tortura inflitta a causa della virtù.”26
deSilva 11:2
Οὐ μέλλω, τύραννε, πρὸς τὸν ὑπὲρ τῆς ἀρετῆς βασανισμόν
“I do not hold back, tyrant, in regard to torture for the sake of moral
11:2 βασανισμόν (βανισμον 457) S L q-q1 m-607-m2-m3-747c 46 52 58 773 Fritzsche]
θανατον 340 668 771; + παραιτεισθαι A 741 Swete Rahlfs Klauck; + δειλιαν (διλ.
340) q2(747*) 607 340 577 668 690 771; + ‫“( ܠܡܐܬܐ‬to come”) Sy

The presence of the verb παραιτεισθαι seems, on the basis of both external
and internal evidence, to be secondary. With regard to external evidence,
only A 741 attest it. As for internal evidence, it would appear that the
remaining manuscripts bear witness to the conviction that μέλλω requires
an infinitive complement. But that conviction seems to be predicated on

25 Klauck, 4. Makkabäerbuch, 731.

26 G. Scarpat, ed. and trans., Quarto libro dei Maccabei (Biblica testi e studi 9; Brescia:
Paideia, 2006), 305.
27 deSilva, 4 Maccabees, 36–37.
in search of the old greek text of 4 maccabees 139

the notion that in this context μέλλω means “be about to.” I would argue,
however, that here it denotes “delay, put off,”28 and that the verse means:
Tyrant, I do not put off torture for the sake of virtue.

9. ὀροφοφοιτοῦντα vs. οροφοιτουντα and οροφοκοιτουντα (14:15)

Rahlfs 14:13–16
13 θεωρεῖτε δὲ πῶς πολύπλοκός ἐστιν ἡ τῆς φιλοτεκνίας στοργὴ ἕλκουσα πάντα
πρὸς τὴν τῶν σπλάγχνων συμπάθειαν, 14 ὅπου γε καὶ τὰ ἄλογα ζῷα ὁμοίαν τὴν εἰς
τὰ ἐξ αὐτῶν γεννώμενα συμπάθειαν καὶ στοργὴν ἔχει τοῖς ἀνθρώποις. 15 καὶ γὰρ τῶν
πετεινῶν τὰ μὲν ἥμερα κατὰ τὰς οἰκίας ὀροφοιτοῦντα προασπίζει τῶν νεοττῶν, 16 τὰ
δὲ κατὰ κορυφὰς ὀρέων καὶ φαράγγων ἀπορρῶγας καὶ δένδρων ὀπὰς καὶ τὰς τούτων
ἄκρας ἐννοσσοποιησάμενα ἀποτίκτει καὶ τὸν προσιόντα κωλύει·
NETS 14:13–16
13 Consider how complex is the affection of a mother’s love for her children,
channeling all her feeling into a sympathy rooted deep within. 14 Even unreason-
ing animals show sympathy and affection for their offspring like that of human
beings. 15 For among birds, the tame ones that roam the mountains protect their
young on housetops, 16 and others, building their nests on mountain peaks, in
clefts of ravines, in holes of trees or on their tops, give birth to their young and
fend off the intruder.
14:15 ὀροφοφοιτοῦντα 741(|)] ‫ܥܐܠܐ‬ ̇ ‫ ܬܚܝܬ ܡܛܠܠܐ‬Sy; οροφοκοιτουντα (conjecture)
Bekker, Fritzsche, Deißmann, Klauck; οροφοιτουντα (-φητ. 71; -φυτ. S 457-472-586-
591-592-594-595-607-617-639-640-656-683-699-713-778-789 340 771; -φυτουντο
682) A S L q-q2 m-m1-m3 46 52 58 340 668 771 Swete Rahlfs; οροφοιτωντα V
577(-φυτ.); φοιτωντα m2

There are a number of problems in interpreting v. 15, specifically with

respect to the dependent clause and subsequent main clause τὰ μὲν ἥμερα
κατὰ τὰς οἰκίας ὀροφοιτοῦντα προασπίζει τῶν νεοττῶν (Rahlfs) “the tame ones
that roam the mountains protect their young on housetops” (NETS). The
association of tame birds with houses is logical enough, but what kind of a
connection is intended in the phrase that is introduced by the preposition
κατά? Furthermore, how does roaming the mountains (οροφοιτουντα) fit
with this picture? This brings us to v. 16, where other birds are described
building nests “on mountain peaks” (κατὰ κορυφὰς ὀρέων) as well as “in
clefts of ravines, in holes of trees or on their tops” (φαράγγων ἀπορρῶγας
καὶ δένδρων ὀπὰς καὶ τὰς τούτων ἄκρας). Although it is not stated explicitly,

28 Included in the OLSJ entry for μέλλω are the words “delay, put off . . . . inf. is freq. omit-
ted.” Among the cited examples are μακρὰ μέλλεται it is a long delay (Sophocles, Oed. col.
219), and μὴ μέλλωμεν ἔτι let us delay no longer (Plato, Leg. 712b).
140 robert j. v. hiebert

the birds described in v. 16 would appear to be wild ones in contrast to

the specifically-designated tame ones (ἥμερα) in v. 15. As such, their situ-
ation out in the wilds would also appropriately contrast with the location
among houses of the tame birds. In that light, again one asks, how do
mountains suit their situation?
It is undoubtedly such questions that have given rise to the variety of
alternatives for the participle in v. 15.
ὀροφοφοιτοῦντα 741(|)29
‫ܥܐܠܐ‬̇ ‫ ܬܚܝܬ ܡܛܠܠܐ‬Sy
οροφοκοιτουντα (conjecture) Bekker, Fritzsche, Deißmann, Klauck
οροφοιτουντα (-φητ. 71; -φυτ. S 457-472-586-591-592-594-595-607-617-639-640-656-
683-699-713-778-789 340 771; -φυτουντο 682) A S L q-q2 m-m1-m3 46 52 58 340
668 771 Swete Rahlfs
οροφοιτωντα V 577(-φυτ.)
φοιτωντα m2

Leaving aside for a moment the reading of 741 that I have chosen as my
lemma, let us consider first the conjecture, οροφοκοιτουντα, which has
been adopted by a number of textual critics over the years.30 This is a
compound consisting presumably of ὄροφος “roof ” and κοιτέω, which is
not attested but would likely be a verbal cognate of κοίτη “nest” (of a
bird).31 Klauck translates 14:15 with the conjecture (οροφοκοιτουντα) as fol-
lows: “Nehmen wir nur die Vögel. Die zahmen beschirmen ihre Jungen,
indem sie unter den Dächern der Haüser nisten.”32 Commenting on this
verse, Klauck suggests that this reading appears to be supported by the
wording of the Syriac version ‫ܥܐܠܐ‬ ̇ ‫ ܬܚܝܬ ܡܛܠܠܐ‬enter under the roof and
also by manscript 741 ὀροφοφοιτοῦντα roam the roofs. He does, however,
make the following acknowledgement: “Allerdings ist die Vokabel ander-
wärts nicht belegt. Rahlfs liest mit den Handschriften ὀροφοιτοῦντα (nach
Liddell-Scott s. v. ein Beleg bei Hesych.), was ‘die Berge durchstreifend’
bedeutet. Will man es beibehalten, müßte man so erklären: Die zahmen
Vögel in V. 15 benutzen die Haüser als ‘Berge’, wie die wilden Vögel in

29 Note that the word break occurs after iota in manuscript 741: οροφοφοι|τουντα.
30 I. Bekker, Flavii Josephi Opera Omnia, vol. 6 (Leipzig: Teubner, 1856), 295; Fritzsche,
ΜΑΚΚΑΒΑΙΩΝ ΤΕΤΑΡΤΟΣ, 378; A. Deißmann, “Das vierte Makkabäerbuch,” in Die Apokry-
phen und Pseudepigraphen des Alten Testaments, vol. 2 (ed. E. Kautzsch; Hildesheim: Georg
Olms, 1900), 169; Klauck, 4. Makkabäerbuch, 742.
31  OLSJ, ὄροφος, κοίτη.
32 Klauck, 4. Makkabäerbuch, 742.
in search of the old greek text of 4 maccabees 141

V. 16 die echten Berge (Hanhart).”33 As Klauck points out, the term attested
by Rahlfs and the majority of manuscript witnesses—ὀροφοιτέω—is not
found elsewhere in extant Greek literature. It is undoubtedly cognate to
ὀρειφοιτέω “roam the mountains,”34 which appears in a citation from the
first century B.C.E. grammarian Sostratus in the commentary on Homer’s
Odyssey by Eustathius, the twelfth century bishop of Thessalonica.35 The
form οροφοιτωντα is presumably based on the root ὀροφοιτάω which, in
LSJ/OLSJ, is listed as an equivalent to ὀρειφοιτέω. The form φοιτωντα is
based on the frequently attested simplex verb φοιτάω “go to and fro,” “on
the wing” (of birds), “roam wildly about.”36 The latter two readings in this
verse—οροφοιτωντα and φοιτωντα—can quite readily be accounted for as
variants of the majority reading, οροφοιτουντα.
This brings us back to my lemma reading, ὀροφοφοιτοῦντα, which is
attested in only the eleventh/twelfth century manuscript 741, one of my
codices mixti. This term occurs nowhere else in extant Greek literature, but
its occurrence in manuscript 741 in 4 Macc 14:15 means that there is more
textual evidence for it than there is for the conjecture, οροφοκοιτουντα,
that is favoured by Bekker, Fritzsche, Deißmann, and Klauck. Like that
conjecture, the first element in this compound word is ὄροφος “roof.” The
second element, however, is φοιτέω, an alternate spelling for φοιτάω.37 In
the entry in OLSJ for ὀροφοιτάω, one reads the following: “= ὀρειφοιτέω,
LXX 4 Ma. 14.15 (v.l. [varia lectio] -οῦντα, but perh. ὀροφοφοιτ- shd. be
read).”38 In his Italian edition of 4 Maccabees, Giuseppe Scarpat com-
ments that οροφοιτουντα “è risultato di un’aplologia *ὀροφο-φοιτέω ‘fre-
quento i tetti’.”39 The suggestion that the original reading ὀροφοφοιτοῦντα
could have given rise to the majority reading οροφοιτουντα is quite plausi-
ble.40 That possibility seems to me to represent a more logical progression
than the reverse one since, as noted above, tame birds are the subject of
the participle in question, and their association with the roofs of houses
makes more sense than the prospect of them roaming the mountains. At

33 Ibid., n. 14a.
34 OLSJ, ὀρειφοιτέω.
35 Eustathius, Commentarii ad Homeri Odysseam 1.390.9.
36 OLSJ, φοιτάω.
37 Ibid.
38 OLSJ, ὀροφοιτάω.
39 Scarpat, Quarto libro dei Maccabei, 353.
40 Oddly enough, despite this comment, Scarpat retains the majority reading
οροφοιτουντα “roam the mountains” in his Greek text of 14:15, and he translates it as
“s’annidano sotto il tetto” nest under the roof as though the Greek text were the conjecture
οροφοκοιτουντα nest in/under the roofs (Scarpat, Quarto libro dei Maccabei, 346–347).
142 robert j. v. hiebert

subsequent stages in the textual history, οροφοιτωντα and φοιτωντα could

have been derived from the majority reading. In this kind of a situation, it
would appear to be more prudent to settle on an attested reading than to
posit one that has no manuscript support, as is the case with the conjec-
̇ ‫ܬܚܝܬ ܡܛܠܠܐ‬
ture. Finally, in response to Klauck’s contention that ‫ܥܐܠܐ‬
in the Syriac version seems to support the conjecture οροφοκοιτουντα,
I would argue that one can make at least as good a case for associating
the Syriac reading with ὀροφοφοιτοῦντα of manuscript 741 as one can for
linking it with οροφοκοιτουντα.

10. ἐπιλογίσασθαι vs. επιλογισασθε (16:5)

Rahlfs 16:5
Καὶ γὰρ τοῦτο ἐπιλογίσασθε
NETS 16:5
Consider this also
16:5 ἐπιλογίσασθαι A S 534 q2(747*) 340 577 741 771 Swete] pr εστιν q-q1; επιλογεισθαι
542; επιλογισασθε L-534 46 52 668 Fritzsche Rahlfs; διαλογισασθαι 682 713; δει
λογισασθαι m1; διαλογισασθε m-682 713-m2-62-747c; + εστιν 577; + χρη 741

The complicating factor in evaluating various options in this case is the

fact that the development of alternative readings could well have been
the result of etacism (αι > ε as a termination or vice versa). I maintain,
however, that my lemma reading ἐπιλογίσασθαι is to be preferred as origi-
nal over Rahlfs’ lemma επιλογισασθε. The manuscript support for the for-
mer is stronger, and when one factors in the variant that involves the
addition of εστιν prior to my lemma (q-q1) and possibly also the readings
διαλογισασθαι (682 713) and δει λογισασθαι (m1), the case for an infinitive
rather than an imperative construction here is even more compelling.
One can more readily account for the progression from the infinitive
ἐπιλογίσασθαι and secondarily the longer alternative εστιν επιλογισασθαι
to the imperative επιλογισασθε than vice versa. My lemma could, in fact,
be regarded as the lectio difficilior. In the only other place in 4 Maccabees
where the verb ἐπιλογίζομαι occurs, the construction is Ἔστιν . . . τοῦτο . . . 
ἐπιλογίσασθαι “this can be explained” (3:6). So in 16:5, I would suggest that
the clause Καὶ γὰρ τοῦτο ἐπιλογίσασθαι is to be translated, For there is also
this to consider.
in search of the old greek text of 4 maccabees 143

11. ὡς οὐδέ vs. ουδ ως (16:12)

Rahlfs 16:12
Ἀλλὰ τούτῳ τῷ θρήνῳ οὐδένα ὠλοφύρετο ἡ ἱερὰ καὶ θεοσεβὴς μήτηρ οὐδ᾿ ἵνα μὴ
ἀποθάνωσιν ἀπέτρεπεν αὐτῶν τινα οὐδ᾿ ὡς ἀποθνῃσκόντων ἐλυπήθη
NETS 16:12
Yet the holy, God-fearing mother bewailed none of them with this lament and
neither attempted to dissuade any of them from dying, nor, as they died, did she
16:12 ὡς οὐδέ (ουδ q2 577; ουδεν 71) S V L q-q1-q2 m2 46 52 340 577 668 771] ως γαρ
ουδε m-m1-62; ουδ ως A-542 741 Fritzsche Swete Rahlfs

I think I understand why Fritzsche and Rahlfs chose ουδ ως as their lemma
(Swete typically follows A as a matter of course). I suspect that they opted
for ουδ ως because it maintains stylistic/syntactical symmetry with the
preceding clause that begins οὐδ᾽ ἵνα. Nevertheless, given the strength of
the support for ὡς οὐδέ, I question why the earliest extant witness (S) and
all the manuscript groups would choose to break that proposed original
symmetry by introducing an expression that is attested nowhere else in
the Septuagint (though there are plenty of examples of ὡς οὐδέ outside
the Septuagint). This is why I conclude that it was the other way round:
ὡς οὐδέ was original and ουδ ως secondary. My proposal for the original
text thus creates a sense of progression in the statement about the mother
by the author of 4 Maccabees, inasmuch as οὐδέ is construed to be more
adverbial than conjunctival: οὐδ᾿ ἵνα μὴ ἀποθάνωσιν ἀπέτρεπεν αὐτῶν τινα
ὡς οὐδὲ ἀποθνῃσκόντων ἐλυπήθη and she did not attempt to dissuade any of
them from dying, just as she did not even grieve while they were dying.41

4. Conclusion

The preceding discussion has dealt with examples of kinds of problems

that I encounter in assessing the manuscript evidence while attempting to
reconstruct the original Septuagint text of 4 Maccabees. Clearly each case
must be decided on its own merits. And, as indicated above, evaluating
both the available external and internal evidence is required.

41 My thanks to Albert Pietersma for his suggestion in a private communication (28 May
2011) that the sense of progression seems to be underscored when οὐδέ is understood to
be adverbial.
The Relationship between the LXX Versions
of Proverbs and Job

Johann Cook

1. Introduction

1.1. The Problem
On the issue of the relationship between the Septuagint versions of Prov-
erbs and Job scholars have deviating opinions. Gerleman expressed the
view that the same translator was responsible for both LXX Proverbs and
OG Job.1 Gammie2 tested Gerleman’s thesis and reached the opposite con-
clusion. Other scholars also dealt with aspects of the relationship between
these two translators. Heater3 studied the LXX of Job and found many
examples of intra-textual readings, what he called anaphoric translation
technique, in LXX Job taken from other parts of the Septuagint. However,
he did not directly address the issue at stake. In his series of articles,
Orlinsky 4 also analysed various characteristics of the Greek translation
of Job, without explicitly dealing with the relationship between the men-
tioned translators. Cox in passing refers to these translators and suggests
that they might come from the same circle of translators.5 In a paper deliv-
ered at IOSCS in New Orleans 2009 based on content analysis of these two
units, I came to the conclusion that these versions originate from different
translators and contexts.6 It is the intention of this article to contribute
towards this discussion.

1 G. Gerleman, Studies in the Septuagint I, Book of Job (LUÅ Bd 43. Nr 2; Lund: Gleerup,
1946), 14–17.
2 J. G. Gammie, “The Septuagint of Job: Its Poetic Style and Relationship to the Septu-
agint of Proverbs,” CBQ 49/1 (1987): 14–31.
3 H. Heater, A Septuagint Translation Technique in the Book of Job (CBQMS 11; Washing-
ton DC: The Catholic Biblical Association of America, 1982).
4 See especially H. M. Orlinsky, “The Character of the Septuagint Translation of the
Book of Job,” HUCA 39 (1958): 229–271.
5 C. E. Cox, “The Historical, Social, and Literary Context of Old Greek Job,” in XII Con-
gress of the International Organization for Septuagint and Cognate Studies, Leiden 2004
(SBLSCS 54; ed. M. K. H. Peters; Atlanta: Society of Biblical Literature, 2006), 116.
6 J. Cook, “Were the Septuagint Versions of Job and Proverbs Translated by the Same
Person?” Hebrew Studies 51 (2010): 129–156.
146 johann cook

1.2. Applicable Criteria
Appropriate criteria need to be formulated in order to address this issue.
As is well-known, there are two sets of criteria: firstly, linguistic ones and
more specifically on the micro level, lexically based criteria, as well as, on
the macro level, syntactical and stylistic issues; secondly, arguments from
content analysis. One would naturally expect the same translator to show
correspondence on these two levels. In this paper, I will deal with the first
criterion; since I have addressed the second criterion at the mentioned
meeting of IOSCS.
Some work has been done in this regard on the OG of Job by Cox.7 I
will take a cue from this research and in this paper I will concentrate on
a small number of linguistic criteria with a focus on lexical items. These
criteria will be chosen randomly. Firstly, I will research the way particles
have been applied in the two translated units; secondly, I will deal with
the manner in which certain lexical items are utilized in these two units.
Here I will endeavour to select those items that I deem typical of each
of the translators and try to determine how such items function in these

1.3. Textual Basis
Since the Old Greek of LXX Proverbs has not yet been determined sys-
tematically, the scholarly edition of Rahlfs8 must be utilised. Proverbs has
been allocated to Peter Gentry to prepare in the Göttingen edition. The
research into Job is based on the critical edition by Joseph Ziegler.9 In this
regard, see also Pietersma10 in his review of Ziegler and Gentry11 on the
asterisked materials.

7 C. E. Cox, “Tying it All Together: The Use of Particles in Old Greek Job,” BIOSCS 38
(2005): 41–54.
8 A. Rahlfs, Septuaginta. Id est Vetus Testamentum graeca iuxta LXX interpretes (Stutt-
gart: Deutsche Bibelgesellschaft, 1979).
9 J. Ziegler, Septuaginta: Vetus Testamentum Graecum. Auctoritate Scientiarum Gottin-
gensis editum XI,4: Job (Göttingen: Vandenhoeck & Ruprecht, 1982).
10 A. Pietersma, Review of Job. Septuaginta. Vetus Testamentum Graecum, II/4 ed. J. Zie-
gler, JBL 104 (1985): 305–311.
11  P. J. Gentry, The Asterisked Materials in the Greek Job (SCS 38; Atlanta: Scholars Press,
the relationship between the lxx versions 147

1.4. Conclusion
There is a consensus that both LXX Proverbs and Job are less faithfully
translated units.12 However, there are also differences between these trans-
lations. A pertinent one is that, whereas LXX Proverbs is fundamentally an
expansive text, the OG of Job, on the contrary, is a shortened, abbreviated
text. However, this does not mean that the former contains no minuses at
all, or that the OG of Job has no pluses. As a matter of fact, LXX Job con-
tains two (three?) important additions, one in ch. 2 vv. 9a–f, the diatribe
on the wife of Job and the second, ch. 42:17a–e. Another conspicuous dif-
ference is the change in the order of chapters13 in LXX Proverbs from ch.
24 onwards, which I have ascribed to the Greek translator.14 These units
are thus unique in the corpus of Septuagint translations.15
In one crucial respect these two units agree, namely in their size. The
number of words are comparable. LXX Job has 13561 words, which include
the Theodotionic text, and Proverbs totals 11164 words. Although there is
a subjective element in this counting it may be used as a basis for com-

2. The Micro Level (Lexical Items)

There can be no doubt that both translators of Proverbs and Job were
excellently educated in the Jewish and Greek cultures. They both were, as
Cox16 said about the OG Job, a work of good literary quality. As far as LXX
Proverbs is concerned, I identified the following significant pattern. This

12 E. Tov and B. G. Wright, “Computer-assisted Study of the Criteria for Assessing the
Literalness of Translation Units in the LXX,” Textus 12 (1985): 149–183. See also the mono-
graph by Tov, The Text-critical Use of the Septuagint in Biblical Research (JBS 3; Jerusalem:
Simor, 1981–1998).
13 See E. Tov, “Recensional Differences Between the Masoretic Text and the Septuagint
of Proverbs,” in Of Scribes and Scrolls, Studies on the Hebrew Bible, Intertestamental Juda-
ism, and Christian Origins presented to John Strugnell (eds. H. W. Attridge et al.; Lanham:
University Press of America, 1990), 43–56.
14 J. Cook, “The Greek of Proverbs—Evidence of a Recensionally Deviating Hebrew
Text?” in Emanuel: Studies in Hebrew Bible, Septuagint, and Dead Sea Scrolls in Honor of
Emanuel Tov (eds. S. M. Paul, R. A. Kraft, L. H. Schiffman and W. W. Fields; Leiden: Brill,
2003), 605–618.
15 C. E. Cox (“Job,” in A New English Translation of the Septuagint. A New Translation
of the Septuagint and the Other Greek Translations traditionally Included Under That Title
[eds. A. Pietersma and B. G. Wright; Oxford / London: Oxford University Press, 2007], 667)
thinks “OG Job is one of a kind in the Septuagint corpus.”
16 Ibid.
148 johann cook

translator clearly was a creative stylist with an exceptional knowledge of

Jewish and Greek culture which is observed on various niveaus. As far as
his creative approach is concerned, he made ample use of hapax legom-
ena which he borrowed from the Greek world. There are also a number
of neologisms;17 he moreover, applied a rather large number of lexemes
unique to Proverbs.18 As to be expected, he interprets extensively in some
instances. I have not made the same extensive analyses of LXX Job. But
the research by Cox points in the same general direction.

2.1. Individual Lexemes
A word on the databases that I used is in order. I utilized the Thesaurus Lin-
gua Graecae and combined it with Hatch & Redpath (HR). In connection
with LXX Job, I double checked these sources against Ziegler’s edition. I
also partially made use of Wordsmith which is a corpus concordance tool
developed by Mike Scott ( It compares word frequen-
cies from different corpora in order to determine which words are most
different between the two. It calculates a keyness score, which measures
the degree of difference in the two corpora, based on the assumption that
words with identical reference frequencies would have a keyness of 1. The
higher the keyness, the bigger the chance that an observed difference is
significant. The analyst selects the level of chance that a trivial finding is
mistakenly regarded as significant, which is by default one in a million in

2.1.1. Particles
Cox20 has dealt with this issue exhaustively. He deems these particles
important: “The use of such ‘little words’ provides nuance, continuity,
change of direction, qualification, color, and emotion, to what we say or
write.” However, the intention of this paper is different from Cox’s who
was interested in the way the translator used particles to tie the OG of

17 J. Cook, “The translator(s) of LXX Proverbs,” TC—a Journal of Biblical Textual Criti-
cism 7 (2002): 1–50.
18 J. Cook, The Septuagint of Proverbs Jewish and/or Hellenistic Proverbs. Concerning the
Hellenistic Colouring of LXX Proverbs (VTSup 69; Leiden: Brill, 1997), 335–342.
19 I was introduced to Wordsmith by a corpus linguist, Bertus van Rooy, from the Uni-
versity of North-West, Vaaltriangle campus. He provided this description and some of the
data utilised below.
20 Cox, “Tying it All Together,” 41–54. See also J. D. Denniston, The Greek Particles
(Oxford: Clarendon Press, 1934).
the relationship between the lxx versions 149

Job together. As stated above, the intention is to compare LXX Proverbs

and Job.
There are prominent correspondences as well as differences between
the way these items were applied by the individual translators. I will deal
with differences first of all. Differences
The first category is where a specific particle was used by either the trans-
lator of Proverbs and Job and not by both.
The particle ἄρα, as an inferential particle, namely “thus, accordingly,”21
is used 40 times in LXX, not at all in Proverbs and 11 times in Job. Another
inferential particle, μέντοι, “to be sure,” that is used 4 times in LXX Prov-
erbs, in Prov 5:4; 16:25; 22:9a and 26:12, does not appear in Job at all.
The coordinating conjunction ἀτάρ is used uniquely in the LXX. It
occurs only twice, namely in Job 6:21 (“But yet you too trod on me, without
mercy; so now that you have seen my hurt, be afraid!”)22 and 7:11 (“That
said, however, I will not be sparing with my mouth; I will speak, though I
am in anguish; I will open up the bitterness of my soul, though I am in dire
straits”). In both cases it occurs together with other particles in what Cox23
calls “the compounding of particles” (οὐ μὴν δὲ ἀλλά, “nonetheless,”—Job
2:5a; 5:8a; 13:3a; 17:10a and 33:1a). In Job 6:21, as “ἀτὰρ δὲ καί” and in 7:11
as “ἀτὰρ οὖν οὐδέ.” This compounding of particles, according to Cox, is
a characteristic of LXX Job. I made various searches in LXX Proverbs in
order to analyse this issue: There are 3 examples of δὲ καί, namely 3:16a;
6:3 and 14:22. However, this could hardly be seen as a compounding since
καί is used paratactically. I could find no examples of this compounding
feature in connection with the conjunction ἀλλά (Cox24 quotes ἀλλὰ μὴν
οὐδἕ—LXX Job 32:21b), nor with ὅτι (ὅτι μέν γάρ—LXX Job 9:19) in LXX
Proverbs. The closest is the combination of εἴ and γάρ as γὰρ εἴ in LXX
Proverbs 23:7 (however, this is a conditional clause); εἴ γάρ in 2:20 and εἰ δέ
in 23:3. But again this is not compounding in the order of those discussed
by Cox. Only one conclusion can be drawn on account of this data, that
the translator of LXX Proverbs did not compound particles, contrary to
LXX Job.

21  Cox, “Tying it All Together,” 42.

22 Translations from NETS.
23 Cox, “Tying it All Together,” 49.
24 Ibid.
150 johann cook

The subordinating conjunction ὁπόταν is a hapax legomenon in Job

29:22. According to HR, a related form, (ὁπότε), appears 10 times in the
LXX and in Job 26:14, but not in Proverbs.
Another category, adverbs and adverbials point in the same direction.
The adverb ἄλλως appears twice in Job, namely 11:12 and 40:8, but not in
LXX Proverbs. The adverb πάλιν is not used in Proverbs but appears 9
times in Job. The adverb καθώς occurs in Job 6:17; 10:4 (Theodotion) and
42:9 but not at all in Proverbs.
The emphatic particle δή, “indeed,” occurs 15 times in Job but not in
Proverbs. According to Van Rooy, it has a keyness of 19.3. The coordinat-
ing particle, διό, “therefore,” is used in Job 9:22; 32:6 and 10; 34:10 and 34;
37:24 and 42:6 but not in Proverbs. The subordinating conjunction διότι,
“because,” appears in Job 32:2 and 3; 33:9 and 36:12. It is not used in Prov-
erbs. According to HR, the coordinating conjunction τοίνυν, “for,” occurs
21 times in LXX and is used in Job 8:13 (together with οὕτως) and 36:14,
but not in Proverbs. The emphatic particle γε, “yet, even,” is not used in
Proverbs but it appears in Job 13:2, 9; 15:10 (2x); 21:25 and 30:2, 24. The
subordinating conjunction ἐπεί occurs in Job 13:15 and 35:7, and not in
Proverbs. According to HR, the negative combination μὴ οὐ/μὴ οὐχί occurs
18 times in LXX, not at all in Proverbs but once in Job 22:12.
It is rather difficult to interpret this data. For one thing many of these
particles are used in limited cases only. Nevertheless, at the least this cat-
egory is an indication that different translators were at work. The lack of
the compounded list of particles, οὐ μὴν δὲ ἀλλἄ, that appears in Job and
not in LXX Proverbs is the most significant piece of evidence in this regard.
One would surely expect the same translator to have used this list. Correspondences
According to Wordsmith, the particle δἕ is found 4852 times in the LXX.
Of these 735 occurrences are found in Job and 556 in Proverbs. Van Rooy
is of the opinion that this particle is the most conspicuous in LXX Job
compared to the rest of the LXX. It has a relative frequency of 4.3% of
all the words in LXX Job, compared to 0.7% in all the other books. This
unique characteristic applies to LXX Proverbs as well, which has a relative
frequency of 4.3%. This is an indication of large correspondence between
these corpora.
The adverb εἶτα occurs in Prov 6:11 and 7:13 and 12 times in Job. The
subordinating conjunction ἵνα occurs 44 times in Proverbs and 23 times in
Job. This is perhaps surprising since Job is probably more of an argumen-
the relationship between the lxx versions 151

tative text than Proverbs. However, the individual translators could have
made use of other particles.
The subordinating conjunction ἐπειδή occurs in Job 9:29 and 25:7 as
well as in Prov 1:24. The coordinating conjunction ἤ, “or,” occurs 29 times
in Proverbs and 76 times in Job. The negative particle μηδέ occurs in Job 7
times and 23 times in Proverbs. According to Van Rooy, this is a significant
difference with a keyness of 15. The inferential particle τοιγαροῦν occurs
12 times in the LXX, thrice in Job, namely in chs. 7:11; 22:10 and 24:22 and
twice in Proverbs, namely in 1:26 and 31. Of the conjunction καί there are
472 instances in LXX Proverbs. According to TLG, there are 594 cases in
LXX Job. However, this number must be relativized. A substantial por-
tion, 187 cases come from Theodotion, and moreover, some examples are
indicated to be καὶ γε in TLG, whereas Ziegler defined these as the particle
καίγε. Examples occur in Job 15:10 (2x), 30:2 from Theodotion. Another fac-
tor is the question as to what text should be deemed the OG. I recently
argued that the addition to ch. 2 v. 9 should be deemed the OG text.25
There are 5 instances of καί in these additions. On the contrary, the addi-
tions in ch. 42 v. 17 are most probably the result of at least two later hands.
One should distinguish between addition a and b–e in this regard.26 How-
ever, for the sake of this exercise, I will only discard the Theodotionic ver-
sions. Hence there are 407 instances of καί in LXX Job. This statistic is also
difficult to interpret, since each example could be understood differently
either as copulative, adversative, etc.27 Be that as it may, it is clear that
both translators used this conjunctive extensively. The coordinating con-
junction γάρ is used extensively in both units. There are 171 occurrences
in LXX Job of which only 4 come from Theodotion, thus 167 instances in
total. This particle is used 102 times in LXX Proverbs.
The combination of the interrogative τί and the coordinating conjunc-
tion γάρ occurs in Job, firstly the accusative neuter sg. form τί γάρ in 3:11
and 20; 4:17, 6:5 and 22; 7:17; 10:19; 15:7 and 9; 16:3; 18:4; 21:4; 22:3; 25:2; 31:14.
Also, the nominative m. sg. τίς γάρ occurs in Job 6:11; 13:19; 14:4; 15:14; 19:23;
27:8; 36:22; 41:2 and 42:3.
According to HR, the negative particle μήτε is used 18 times in the LXX
and not in Job and Proverbs. The particle καίτοι is a hapax legomenon in

25 See J. Cook, “Are the Additions in LXX Job 2,9a–e to be deemed as the Old Greek
Text?” Bib 19/2 (2010): 275–284.
26 See A. Y. Reed, “Job as Jobab: The Interpretation of Job in LXX Job 42:17b–e,” JBL 120/1
(2001): 31. See also Gentry, The Asterisked Materials in the Greek Job, 586.
27 Cox, “Tying it All Together,” 42.
152 johann cook

4 Macc 2:6. The coordinating conjunction οὖν is used 16 times in LXX Job
and 4 times in Proverbs.
The negative particle μηδέ is used in Job 3:4, 6, 7 and 9; 16:18 and 20:17
and 40:3. In Prov 1:10; 3:11 and 31; 4:5 and 6 (2x), 14 and 27; 5:20; 6:4, 7 (2x),
25 (2x); 9:18a; 23:6, 7 and 20; 24:1, 15, 19, 28 and 25:6.
According to TLG, the coordinating conjunction τε is used 18 times in
Job. There are, however, discrepancies between TLG and HR. In the lat-
ter the following instances do not occur in TLG, 1:5; 6:18; 12:7; 14:21; 26:13;
30:4; 37:13 and 39:11. These instances do not occur in Ziegler’s edition. This
brings the number of occurrences to 10. Τε is used 13 times in Proverbs. It is
striking that it is applied 5 times in the first six verses of ch. 1 in Proverbs.
This is done in order to tie together the introduction as to what is needed
to be wise, or to become wiser (v. 5).28 As can be seen there is some cor-
respondence between the way this particle is used in both units.
The interrogative ποτε is used 3 times in Job and 5 times in Proverbs.
The interrogative μήποτε occurs in Job 1:5 and in Prov 22:25; 23:9; 25:16
and 17 and 30:10. According to HR, μή τις/μή τι occurs 11 times in LXX and
in Job 6:22 and 25:3 (the coordinating conjunction γάρ is added in this
instance) and in Prov 3:30. The negative particle μή occurs frequently in
both units. The negative conjunction οὐδέ occurs approximately twice as
much in Job as in Proverbs.
The adverb ἔτι appears only once in Proverbs, namely in 31:7 and 28
times in Job. The adverb νῦν is used extensively in the LXX, but only 16
times in Job. In addition, 5 cases are of the form νυνί (6:28; 7:21; 30:1 and
9 and 42:5) but none appears in LXX Proverbs. There are three cases of
νῦν in Proverbs (5:7; 7:24 and 8:32). The adverb οὕτως is used frequently in
both Job and Proverbs. The interjection ἰδού occurs once in Proverbs (1:23)
and 26 times in Job. The subordinating conjunction ὅτε occurs twice in
Proverbs and 12 times in Job.
The data under this paragraph could be an indication that the same
translator was responsible for both units. However, there are no striking
characteristics in this regard.

28 J. Cook, “Inter-textual relations between the Septuagint versions of the Psalms and
Proverbs,” in The Old Greek Psalter. Studies in honour of Albert Pietersma (eds. R. J. V. Hie-
bert, C. E. Cox and P. J. Gentry; Sheffield: Sheffield University Press, 2001), 228.
the relationship between the lxx versions 153

2.2. Lexical Items Used Uniquely by the Translators

2.2.1. LXX Proverbs
Being wisdom literature one can naturally expect LXX Proverbs to have
religiously orientated lexemes. A prominent lexeme is found in ch. 8:22.
The translator intentionally opted for the verb κτίζω as equivalent for ‫ ָקנָ ה‬.
The Greek verb appears 63 times in the LXX, but only once in Proverbs
and not at all in Job. The Hebrew lexeme ‫ ָקנָ ה‬has various connotations
in its semantic fields. It is used in Prov 1:5; 4:5 and 7; 8:8; 15:32; 16:16; 17:16;
18:15; 19:8; 20:14 and 23:23. According to Lisowski,29 Prov 8:8 is the sole
example where the nuance of “to create” obtains. It therefore seems as
if the translator intentionally applied κτίζω, in order to avoid possible
LXX Proverbs also contains readings that appear in the political and
administrative contexts and that are not used in Job at all. The hapax
ἀβασίλευτόν (“without a king”) occurs in Prov 30:27. It appears in classi-
cal Greek literature in Thucydides Histories 2.80.5, 6; Xenophon Historia
Graeca 5.2.17 and Plutarch Theseus 24.2. It has a bearing on the issue of
kingship in Proverbs.31 The noun ἀντίδικος is used 8 times in the LXX and
once in Proverbs, namely in 18:17, but not in Job at all. Here it depicts the
adversary/opponent. The hapax δωρολήμπτης is found in 15:27 where it
describes “a receiver of bribes.” The verb παρεδρεύω occurs only in Prov
1:21 and 8:3 in the LXX in connection with the activity of the figure of
wisdom.32 The lexeme συκοφάντης occurs only twice in the LXX, in Ps 71
(72):4 and Prov 28:16. In the latter, it is used to describe a king who is a
swindler. Aitken33 relates it with Ptolemaic papyri.
Other significant lexemes have a bearing on political and religious insti-
tutions. The word συνέδριον appears 8 times in Proverbs with an interest-
ing distribution: 11:13 (‫ ;)סוד‬15:22 (‫ ;)סוד‬22:10 (-)(‫ ;)דין‬24:8 (-); 26:26 (‫;)כהל‬
27:22 (✝) and 31:23 (-). It is used 12 times in the LXX and not at all in Job.

29 G. Lisowsky, Konkordanz zum Hebräischen Alten Testament, Zweite Auflage

(Württembergische Bibelanstallt: Stuttgart, 1958).
30 Cook, The Septuagint of Proverbs Jewish and/or Hellenistic Proverbs, 21.
31 Cf. J. Aitken, “Poet and Critic: Royal ideology and the Greek translator of Proverbs,” in
Jewish Perspectives on Hellenistic rulers (eds. T. Rajak et al.; Berkely / Los Angeles / London:
University of California Press, 2007), 195 on the significance of the poetic lexical choices
this translator made.
32 Cf. ibid., 196. Aitken finds correspondences with the function of judges during the
Hellenistic period.
33 Cf. ibid.
154 johann cook

The following words appear only in LXX Proverbs and not in LXX Job:
The noun ἀγγελία Prov 12:25 (‫) ָדּ ַבר‬, 25:25 (‫מוּעה‬ָ ‫ ) ְשׁ‬and 26:16 (‫) ָט ַעם‬. This
noun is used 14 times in the LXX. Here it represents 3 Hebrew words in
LXX Proverbs. The lexeme αἴσθησις appears 25 times in LXX of which 21
times in Proverbs, consistently as equivalent for ‫ ַדּ ַעת‬. The noun ἀπαίδευτος
is found 16 times in LXX and 7 times in Proverbs. It represents 5 Hebrew
words in LXX Proverbs, namely ‫מוּסר‬ ָ , ‫ ְכּ ִסיל‬, ‫ ֵלץ‬, ‫ ֵאין נָ ָבל‬and ‫ ֱאוִ יל‬. The
noun γνῶσις is used frequently in the LXX and in Proverbs 15 times, mostly
for ‫ ַדּ ַעת‬. There are no equivalents in Job. The verb ἐγκωμιάζω appears
only in Proverbs, namely 5 times. Of these 4 times it represents ‫ הלל‬and
once ‫רבה‬.
As stated already, I have identified numerous hapax legomena in LXX

2.2.2. LXX Job
As far as OG Job is concerned, I did not make similar exhaustive lexical
analyses as in the case of LXX Proverbs. I identified a number of hapax
legomena in chs. 1, 2, 14, 19, 28 and 42. A few examples must suffice.
LXX Job ch. 2 contains 4 hapax legomena. The verb ἐκτίνω in v. 5 ren-
ders ‫יִ ֵ ּ֖תן‬. Three more examples come from the set of additions, namely the
verb διανυκτερεύω in 9c and πλανῆτις35 and λάτρις in 9d.
There are lexemes that can help to determine the provenance of OG
Job. Cox36 agrees with Gerleman that the Greek word φορολόγος, “tax
gatherer,” (3:18 and 39:7) is a term that reflects an Alexandrian (Egyptian)
context. This word appears only in 2 Esdr 4:7 and 18; 5:5; Job 3:18 and 39:7,
as well as 1 Macc 3:29 and not in Proverbs. Three other terms fall into the
same category. The noun πάπυρος (“papyrus”) appears only in Job 8:11 and
40:21 (in A) as well as in Isa 19:6. The noun κάλαμος (“reed”) is used 25
times in the LXX, but only once in Job, viz. 40:21. The third term, βούτομον
(“sedge”) also appears in Job 40:21 and in 8:11. Strikingly, none of these
terms occur in LXX Proverbs. This applies also to the term, τετράδραχμον,
which is a hapax legomenon, used as the standard denomination in the
monetary system of the Ptolemies.37 These lexemes seem to point to an

34 Cook, The Septuagint of Proverbs Jewish and/or Hellenistic Proverbs, 335–342.

35 This word occurs only in some MSS.
36 Cox, “The Historical, Social, and Literary context of Old Greek Job,” 108.
37 G. Gerleman, Studies in the Septuagint I, Book of Job (LUÅ Bd 43. Nr 2; Lund: Gleerup,
1946), 36.
the relationship between the lxx versions 155

Alexandrian provenance of LXX Job. I have argued that the Greek version
of Proverbs in fact came into being in Palestine.38
A number of lexemes occur only in LXX Job and not at all in LXX Prov-
erbs. The noun βροτός occurs only in Job namely 17 times as equivalents
for ‫ ֱאנוֹשׁ‬, ‫ ָא ָדם‬and ‫ ָבּ ָשׂר‬. The adjective γεννητός is used only in Job, 5
times, consistently as equivalent of ‫ילד‬. The verb δείδω is used 8 times in
Job and once in Isaiah. In LXX Job, it represents ‫ ָח ְפ ִשׁי‬, ‫ יגר‬and ‫שׁחח‬. HR
5 times uses the siglum ✝. Διάβολος is a word used 19 times in the LXX and
12 times in Job always for ‫ ָשׂ ָטן‬. Δίαιτα appears 11 times in Job and is used
for ‫נָ וֶ ה‬, ‫ א ֶֹהל‬and ‫ ַבּיִ ת‬. HR uses the siglum ✝ 3 times and the noun occurs
once in Judg 12:15. The lexeme ἐξαίσιος occurs only in Job (9x) for ‫ִמ ָקּרוֹב‬
‫ ֵשׁ ֶמץ‬, ‫פּלא‬, ‫ ֶצ ַלע‬, and ‫ ִפּ ְתאֹם‬respectively. The noun ἐπιστήμη is used 13
times in Job with different Vorlagen (‫ ִבּינָ ה‬, ‫ ְתּבוּנָ ה ֵדּ ַע‬, ‫ ַדּ ַעט‬and ‫ ) ָשׂכל‬and
not in Proverbs. The verb ἐφοράω occurs 4 times in Job and as equivalent
for 3 Hebrew words. In Job 21:16, for ‫ ;רחק‬22:12 for ‫ ;ראה‬28:24 for ‫נבט‬.
One example (34:24), according to HR, has no Semitic Vorlage. The noun
σαπρία (“refuse”) is used 9 times in the LXX and primarily in Job, namely
in 2:9; 7:5; 8:16; 17:14; 21:26 and 25:6, mostly in connection with ‫ ִר ָמּה‬.
As I stated earlier, the choice of lexemes has been made at random.
This exercise naturally needs to be extended, but can be deemed repre-
sentative. It is also evident that the Wordsmith option needs to be worked
out further.

3. Conclusion

I am forced by the evidence to reiterate the conclusion I reached in New

Orleans. I find it difficult to accept that two translators who exhibit such
divergent attitudes to their parent texts and to specific lexical items, could
have belonged to the same circle of translators, nor could be deemed as
one and the same translator.39

38 Cook, “Theological/ideological Tendenz in the Septuagint—LXX Proverbs a case

study,” in Interpreting Translation. Studies on the LXX and Ezekiel in honour of Johan Lust.
(eds. F. García Martínez and M. Vervenne; Leuven / Paris / Dudley: Leuven University
Press / Peeters, 2005), 79.
39 Cf. also Gammie, “The Septuagint of Job,” 29. Jan Joosten reached the same conclusion
in respect of the way another category, elaborate similes, is rendered differently in LXX
Proverbs and Job (“Elaborate Similes—Hebrew and Greek. A Study in Septuagint Transla-
tion Technique,” Bib 77 [1996]: 236).
An Analysis of the Use of Hebel as a Metaphorical and
Symbolic Device as Interpreted in LXX Ecclesiastes

Lawrence Lincoln

1. Introduction

This paper intends to examine how the Greek translation of Qohelet rep-
resented the lexeme ‫( הבל‬hebel) and how it incorporated the thematic
nature and concepts of the meanings of this lexeme into a Greek text
that would have been accessible and understandable to Jewish and Greek
reading audiences.
Qohelet is in many ways a controversial and unusual book when com-
pared to the rest of the Hebrew canon. As one of the five books com-
prising the group known as the Megillot (or Ketubim), its controversial
themes makes it stand more or less on its own in the collection of inspired
biblical works. Its enigmatic nature speaks not only of the world of the
third century B.C.E. when it was most probably composed,1 but also of a
universalistic humanism that is readily recognizable even today. Although
by tradition ascribed to the hand of King Solomon, most modern schol-
ars have rejected the book’s identification with Solomon as its author.2

1 The date of the composition of Qohelet has been extensively researched, not only
because of the controversial message of the book, but because of its attribution of Solomon
as its author. With few and somewhat vague historical allusions to go by, the extensive
scholarship has proposed a range of possible dates for its composition range from the
time of Solomon himself to the Persian period, the Greek age down to 100 C.E., a period
spanning six hundred years. Attempts to date the work by analysing the linguistic charac-
ter of the Hebrew have also fallen far short of expectations. Choon-Leong Seow probably
echoes the view of most modern scholars on Qohelet that it was most likely composed
shortly before the conquest of Palestine in 333 B.C.E. C. L. Seow, Ecclesiastes: A New Trans-
lation with Introduction and Commentary (AB 18; New York: Doubleday, 1997). In a PhD
dissertation L. R. Tyler analysed 66 individual lexical items to ascertain the lateness of
the book’s composition or, its possible earlier authorship attributed to King Solomon. In
all there were only six truly late lexical items that proved a later authorship, and as noted
by Tyler in his closing word: “The linguistic evidence, then, does not disprove Solomonic
authorship, nor does it prove that Solomon wrote Ecclesiastes.” L. R. Tyler, The Language
of Ecclesiastes as a Criterion for Dating (PhD diss., Texas, 1988). An earlier yet still relevant
study is that of E. Bickerman, Four Strange Books of The Bible (New York: Schocken Books,
1967), who also argues for a third century date of composition.
2 The prescript to the book states that it was written by the son of King David, although
the Hebrew version does not specifically refer to Solomon as such. Only the Aramaic
158 lawrence lincoln

Although Solomon is not mentioned specifically by name, the attribution

to Solomon by later rabbinical hands gave the text the cachet of legiti-
macy. Midrash Qohelet (known as Haggdat Qohelet in the Arukh) but only
called the Qohelet Rabbah since the Venice Edition of 1545,3 is a much later
work, probably eighth century, which fixes Solomon clearly as the spiri-
tual inspiration behind the creation of Qohelet: “The words of Koheleth,
the son of David, King in Jerusalem. That is what scripture declares by the
Holy Spirit through Solomon, King of Israel (1.1).”4
Hebrew Qohelet does not contain any stirring or dramatic narrative,
nor is it populated by sages or memorable characters (except obliquely
when referring to the putative authorship of King Solomon), nor does it
deal with sacred, legal or Hallachic matters. Singularly, as a Jewish crea-
tion, it contains no doctrinal imperatives; no messianic or end-of-days
grand plan; nor is mention made of the covenant and the term “Israel”
does not occur at all. The effect of all this is to give the work a philosophi-
cal tone rather than one of any direct religiosity.5 Compared to the other
books of the Hebrew Bible which deal with the historical experiences of
the Israelite nation and the development of a theological foundation for
the Jewish faith, the wisdom books, on the other hand, present a pan-Near
Eastern secular basis for Jewish wisdom thought.6 There is little there-
fore that identifies Qohelet, Job and Proverbs as particularly Israelite in
character. In fact, Job himself for instance, appeared to have dwelt in the
land of Uz, probably in southern Palestine in Edom or northern Arabia,7
but illustrates the wider context in which wisdom-type thought devel-

Targum adds “Solomon, the Son of King David in Jerusalem” as does the Midrash in
Qohelet Rabbah. The LXX also does not mention Solomon by name, although the para-
phrase by Gregory of Thaumaturgos goes further by stating in the opening sentence to the
book, “Solomon (the son of the king and prophet David) . . .” (Τάδε λέγει Σαλομών, ὁ τού
Δαβίδ βασιλέως καί προφήτου). J. Jarick Gregory Thaumaturgos’ Paraphrase of Ecclesiastes
(Atlanta: Scholars Press, 1990); likewise the Vulgate also only mentions the son of David:
filii David Regis Hierusalem.
3 See H. L. Strack and G. Stemberger, Introduction to the Talmud and Midrash (Edin-
burgh: T & T Clark, 1991), 342.
4 A. Cohen, Midrash Rabbah (vol. 8; London: Soncino Press, 1971).
5 Seow, Ecclesiastes, 54.
6 L. H. Schiffman, From Text to Tradition: a History of Second Temple & Rabbinic Juda-
ism (Hoboken: Ktav Publishing House, 1991), 32. As noted by James Crenshaw, “Wisdom
expresses itself with remarkable thematic coherence. Wise men and women address com-
mon problems, whether the dangers of adultery, the perils of the tongue, the hazards of
strong drink, the enigma of undeserved suffering, the inequities of life, or the finality of
death” (cf. J. L. Crenshaw, Old Testament Wisdom: An Introduction [London: SCM Press,
1981], 27–41).
7 See also Gen 36:28 and Lam 4:21.
an analysis of the use of hebel as a metaphorical 159

oped. This could partly explain why these three wisdom works were not
concerned with Israelite revelation and national redemption at all.8 In
Qohelet, God is mentioned 40 times (‫)אלוהים‬9 as such; Qohelet never
questions the existence of God, but is rather obsessed with the theme of
injustice and the capriciousness of human existence and a God who does
not appear to have a direct or determinative role in the fate of mankind.
Job too rails against the world that is filled with injustices in a universe
governed by a God who is supposed to be a God of justice. Qohelet’s treat-
ment of God (‫ )אלוהים‬is peculiar and confuses the reader; ‫ אלוהים‬is first
mentioned in 1:13 in a somewhat confusing context: “I gave my heart to
search and explore with wisdom all that is done under the heavens and
it is a hard thing that God has given men to preoccupy themselves with”
(1:13).10 God is therefore a problematic concept in Qohelet and, as H. Fisch
has observed, the God of Israel is “half-concealed and half-revealed” and is
only discerned through dialectic of presence and absence.”11
In Qohelet the cosmos and God are a mystery. God is undeniably in
control, but his actions and that of the cosmos are mysterious to men,
who have to deal with events over which they have little or no control,
while enduring a life that is unpredictable and harsh, and one in which
they are ultimately faced with death, like any of the animals of the earth.
Qohelet’s single most important feature therefore is its ontological
quest to finding out what being human is all about. Martin Heidegger, in
his monumental work, Being and Time, asked the same question but from
a different perspective: “if we are capable of dying, how can we possibly

8 See R. Alter, The Wisdom Books: Job, Proverbs and Ecclesiastes (New York: Norton,
2010), 3–10. The commentaries on wisdom literature are numerous and space does not
permit mentioning them in detail.
9 The term ‫אדוני‬, the commonly accepted personal appellation of God, does not occur
at all. Qohelet appears to speak to and of a “generic” God in keeping with the philosophic
tone of the book.
10 I have translated ‫ בו לענות‬as “preoccupy themselves with”; however, the Targum
has used the term “afflicted” ‫לאיסתגפא‬. The LXX uses περισπᾶσθαι. The Targum, however,
acknowledged the problematic nature of this passage and reinterpreted it entirely as fol-
lows: “And I set my mind to seek instruction from the Lord at the time when he revealed
himself to me at Gibeon to test me and to ask me what I wanted from him. And I asked
of Him only the wisdom to know the difference between good and evil understanding of
everything that happened under the sun in this world. I saw all the deeds of sinful people
were an evil matter which the Lord gave to the people so that they should be afflicted
by it”. I have used the translation here by P. Knobel, The Aramaic Bible (Collegeville: The
Liturgical Press, 1991) 24.
11  H. Fisch, Poetry with a Purpose: Biblical Poetics and Interpretation (Bloomington: Indi-
ana Press, 1988), 165.
160 lawrence lincoln

understand death?”12 In other words, if death is so fundamental to being

human, how then does one begin to understand what being human really
means? This kind of enquiry is by its very nature inherently melancholy,
a feature often ascribed to Qohelet. This is a universalistic theme and one
that has its roots in ancient philosophical enquiry, which is possibly why
it resonates so strongly and appears curiously familiar at first reading,
according to R. Whybray.13
It is therefore no surprise that this difficult book should have aroused
controversy before it was finally accepted as part of the Hebrew Bible
tradition. It has been traditionally accepted that the Council of Yavneh
played the determinative role in fixing the Hebrew collection of inspired
books at the end of the first century C.E. and thereby implying the time
when Qohelet’s reception was apparently finalised. This kind of linear
approach must be handled cautiously, as Schiffman has noted; there is no
evidence for supposing that the council at Yavneh was responsible for for-
mally approving and accepting the books of the Ketubim into the canon,
or that the canon as a whole had reached its final stage of acceptance.14
Qohelet continued to be a controversial book and generated internal
debate amongst the rabbis at least until as late as the fourth century C.E.,
according to Jerome.15 “The Hebrews say that . . . this book should be oblit-
erated, because it asserts that all the creatures of God are vain, and regards
the whole as nothing, and prefers eating and drinking and transient pleas-
ures before all things.” The Mishnah provides a succinct example of the
type of debate that Qohelet generated in m. Yad 3.5. This section names
the rabbis involved and includes the following observation:
All the Holy Scriptures renders the hands unclean. The Song of Songs and
Ecclesiastes render the hands unclean. R. Judah says: The Song of Songs
renders the hands unclean, but about Ecclesiastes there is dissension. R. Jose
says: Ecclesiastes does not render the hands unclean, and about the Song of
Songs there is dissension. R. Simeon says: Ecclesiastes is one of the things

12 M. Heidegger, Being and Time (trans. J. Macquarie and E. Robinson; Oxford: Black-
well, 1995), 41.
13 R. Whybray, “Qoheleth, Preacher of Joy,” JSOT 23 (1982): 87–98.
14 Schiffman, From Text to Tradition, 56–58.
15 Jerome, Commentarius Ecclessiasten in Patrologia Cursus Completus: Series Latina 23
(ed. J. P. Migne; Paris), 1172. Another point of reference is Josephus, who wrote in the last
decade of the first century: “For we have not an innumerable multitude of books among
us, disagreeing from and contradicting one another (as the Greeks have), but only twenty-
two books, which contain the records of all the past times” (Contra Apionem 1.38); cf.
W. Whiston, The Works of Josephus (Peabody: Hendrickson Publishers, 1987).
an analysis of the use of hebel as a metaphorical 161

about which the School of Shammai adopted the more lenient, and the
school of Hillel the more stringent ruling.16
Despite Talmudic references to the various debates around its acceptabil-
ity, the rabbinical pericopes are scholarly in nature, indicating that the
book had been in use already in some form or another, and that the decree
for its acceptance into the Hebrew canon was probably a fait accompli
but was still being debated and discussed.17 Arie van der Kooij explains
the ancient books and scriptures would not have been seen as carrying
authority if their teachings had not been brought into force and if they had
not been studied by the appropriate authorities—the scholar scribes. Inter-
pretation of books that were considered authoritative required authoritative
and authorised persons to bring the ideas into effect.18
A further clue as to the book’s earlier existence and influence before
the first century lies in the possibility that it was used by the author of
Ecclesiasticus, or the Wisdom of Ben Sirach, which was composed about
180–175 B.C.E. It is also likely that Qohelet was known to the author of the
Wisdom of Solomon, which was most likely compiled in the first century

2. Hebel

Qohelet begins the introductory passage with the following startling state-
ment: “Everything is futile, said Qohelet, it’s futile, it’s all completely futile”
(1:2) (my translation). The concept of hebel, most commonly rendered in
English as “vanity,” has been one of the most studied and disputed con-
cepts/lexemes in the Qohelet text and an extensive number of commen-
taries have been devoted to this one aspect of the Qohelet book alone.
The term hebel appeared in the King James Version as “vanity” and hebel
habalim as “vanity of vanities,” according to Robert Alter, most probably

16 English translation by H. Danby, The Mishnah (London: Oxford University Press,

1933), 781.
17 See Talmud b. Sab. 30b; Lev. Rab. 28.1; and Abot de R. Nat. 1.4.
18 A. van der Kooij, “Authoritative Scriptures and Scribal Culture,” in Authoritative
Scriptures in Ancient Judaism (JSJSup 131; ed. M. Popovic; Leiden: Brill, 2010), 55–71.
19 The author of the Wisdom of Solomon attempted to rectify certain of Qohelet’s say-
ings he deemed to be sinful. See G. Barton, A Critical and Exegetical Commentary on the
Book of Ecclesiastes (ICC; Edinburgh: T & T Clark, 1908), 53–58.
162 lawrence lincoln

as a result of the Vulgate influence from the Latin vanitas.20 In my reading

of the concept of hebel as used by Qohelet, I do not think that the modern
English equivalent of “vanity” has quite the correct sense of the meaning
of the Hebrew term hebel. In my opinion, Qohelet was principally trying
to convey a particular message, that in spite of all the advantages of fear-
ing God, the advantages to be gained in studying wisdom and pursuing
wealth and other material pleasures of life, in spite of all these, they are
all just pointless as nothing lasts in the cycles of life. In other words, the
lot of mankind is essentially pointless; as in the beasts of the earth, death
and rotting in a grave are the lot of mankind and it is futile to expect more
than that. Michael Fox has translated hebel as “absurd” and “absurdity of
absurdities.” He bases part of his argument on the work of Albert Camus
and his work The Myth of Sisyphus, and demonstrates that both Camus
and Qohelet are similar in that they excelled in portraying the irratio-
nalities of human existence, “The two thinkers are aligned also in their
unflinching determination to strip away illusions and to face life’s harsh
irrationality.”21 Fox, however, realizes that the “absurd” is not applicable
in all cases and advocates related synonyms, such as ephemeral, as in the
case of the comparison between animals and man when living beings are
referred to as ha-kol in 3:18–19. As Fox has noted, the death of an animal
cannot be regarded as “absurd” and equated with that of humans and
proposes in this case the term ephemeral when dealing with life forms
in general.22 The point is that life and the very notion of being human
and the way we live as it is desires rational outcomes and hope, but the
reality is in fact the opposite, which is that it is absurd and therefore futile
and pointless to be striving for any outcome (my emphasis). In a recent
book by Robert Alter, hebel and its variants have been translated as “mere
breath” and “merest breath,” which although it acknowledges the literal
and material nature of hebel, sounds contrived and also does not convey
the context, purpose and intent we can presume of the author or authors,
nor do these terms capture the extreme irony and pathos of human exis-
tence, according to Qohelet’s worldview.
The point is that hebel as a metaphorical and abstract concept is prob-
lematic from a translational point of view as well as when striving for

20 Alter, The Wisdom Books: Job, Proverbs and Ecclesiastes, 339.

21 M. V. Fox, A Time to Tear Down and a Time to Build Up: A Rereading of Ecclesiastes
(Grand Rapids: Eerdmans, 1999), 10.
22 Ibid., 39.
an analysis of the use of hebel as a metaphorical 163

a complete reading of the book. As Douglas Miller has noted: “Qohelet

employs hebel as a ‘symbol’, an image which holds together a set of mean-
ings, or ‘referents’, that can neither be exhausted nor adequately expressed
by any single meaning.”23 Significantly, Miller also states correctly that “no
single metaphor has been found adequate for the range of hebel’s diverse
usage” (my emphasis).24
On a practical level the lexeme hbl appears 38 times in the Hebrew ver-
sion of Qohelet, if one includes hebel habalim in 1:2.25 Hebel is a noun and
in its various uses in the MT it is used metaphorically to describe specific
conditions of human experience and in a few cases as a verb in the MT.26
It is also used more than one hundred times in the Babylonian Talmud
and in other biblical texts.27 The original Hebrew meaning is to describe
an exhalation of breath, a mist, light wind or vapour. In short, it describes
some form of breath/vapour/mist that can be seen, but has no physical sub-
stance beyond that and just disappears as quickly as it had been observed.28
Trying to take hold of it is thus physically impossible—thus it is futile.
The term thus acquired its metaphorical sense to describe the condi-
tions of living beings (humans and animals) that are observable, ironic,
and contradictory. Typical synonyms used are most notably “ephemeral,”
insubstantial, incomprehensible, enigmatic and inconsistent.29 In French
it is rendered as “fumeé,” in German “Vergänglichkeit,” “Nichtigkeit” and
“Eitelkeit,” in Dutch “ijdel” and in Afrikaans “tevergeefs.”
Fox has noted that in the absence of any clear history of the develop-
ment of the word in the Hebrew language, its metaphorical development

23 D. B. Miller, Symbol and Rhetoric in Ecclesiastes: The Place of Hebel in Qohelet’s Work
(Atlanta: Society of Biblical Literature, 2002), 15. Also, P. Wheelwright, Metaphor and Real-
ity (Bloomington: Indiana University Press, 1962), 92.
24 Miller, Symbol and Rhetoric, 14.
25 Some scholars believe that ‫ הכל‬in 9:2 should read as ‫ הבל‬and that at 9:9 the second
‫ הבל‬should be omitted because of scribal errors, as has been done in the LXX.
26 Job 27:12; Ps 62:11; Jer 23:16; 2 Kgs 17:15; Jer 2:5.
27 Miller, Symbol and Rhetoric, 187–194.
28 Seow, Ecclesiastes, 47; all the other studies on Qohelet are in broad agreement with
this explanation.
29 Also L. Koehler and W. Baumgartner, Lexicon in Veteris Testamenti Libros; (Leiden:
Brill, 1958); B. Davidson, The Analytical Hebrew and Chaldee Lexicon (London: Samuel Bag-
ster, 1959).
164 lawrence lincoln

is hinted at from its other uses in OT.30 Some of the other MT references
in which hebel and its variants appear are noted below.31

1.  Ephemeral connotations
Ps 39:6: “Surely every man walks in a vain show; surely they are disqui-
eted in vain (‫)אך הבל יהמיון‬.”
Prov 21:6: “The obtaining of treasures by a lying tongue is a vanity
(‫ )הבל‬tossed to and fro of them who seek death.”
Ps 144:4: “Man is like to vanity; his days are as a shadow (‫אדם להבל‬
‫ )דמה‬that passes away.”
Job 7:16: “I loathe it; I would not live forever—leave me alone, for my
days are vanity (‫)כי הבל ימי‬.”

2. Vain
Isa 49:4: “I have laboured in vain (‫)הבל‬.”
Isa 30:7: “Egypt will help only in vain (‫ )הבל‬and emptily (‫)רק‬.”
See also Lam. 4:17; Job 9:29; and Ps 94:11 for similar meanings of hebel
as meaning “vanity.”

3. Deceit
Hebel is also synonymous with “lies” and “deceit.” Whenever šeqer,
‘awen, ma‘al or kazab appears in the following verses, it includes hebel
within the same verse to highlight the vanity of deceitful deeds and
thoughts as in Prov 31:30; Job 21:34; Ps 62:10 and Zech 10:2.
The theme of deceit is taken further to also mean the uselessness of
pursuing false gods as in Jer 16:19; Deut 32:21; Jon 2:9; Zech 10:2; 2 Kgs
17:15; Jer 2:5, 8:19, 14:22.
But it is a moot point whether Qohelet was influenced in any way by
the passages noted above.32

30 Cf. Fox, A Time to Tear Down and a Time to Build Up, 27–30. See C. D. Ginsberg,
Coheleth, Commenly Called the Book of Ecclesiastes (London: Longman, Green, Longman
& Roberts, 1861), 13; and O. Loretz, Qohelet und der alte Orient: Untersuchungen zu Stil und
theologischer Thematik des Buches Qohelet (Freiburg: Herder, 1964), 223, who both describe
hebel as denoting “nothing,” but Fox (A Time to Tear Down and a Time to Build Up, 28)
counters with the argument that hebel is never used in Qohelet to mean “nothing.”
31 Adapted from a summary by Fox, A Time to Tear Down and a Time to Build Up,
32 G. S. Ogden, “The Meaning of the Term Hebel,” in Reflecting with Solomon: Selected
Studies in the Book of Ecclesiastes (ed. B. Zuck; Grand Rapids: Baker Books, 1994), 227–232.
an analysis of the use of hebel as a metaphorical 165

Hebel’s use is in fact quite extensive in numerous biblical and post-biblical

texts. A summary of these various usages and their loci is provided by
Douglas B. Miller (Appendix II), which shows how widespread the lexeme
was and the variety of metaphorical and non-metaphorical usages it had
in biblical and post-biblical literature. In short, it appears in hebel form
in Hebrew or as habla in its Aramaic form 96 times in noun forms and
9 times in verb forms in the MT. Hebel can also be found over 100 times
in the Babylonian Talmud.33
The term hebel and the metaphorical sense it connoted were there-
fore rooted in the Semitic consciousness and had a number of different
contextual meanings, all of which bear negative values and connotations
which would have been commonly understood as such by their intended
target audience. Different negative inferences and contexts included, diffi-
cult human conditions (Job 7:21), errors of speech (Job 35:16), futile labour
(Isa 49:4, Job 9:29), acquiring wealth dishonestly (Prov 21:6) and numerous
other contexts in which hebel was used symbolically and metaphorically.34
The debate on the intellectual influences on Qohelet has been widely
studied and the consensus amongst most scholars appears to be that it
is impossible to verify that there were any foreign, i.e. non-Jewish, influ-
ences on the author’s thought. Besides numerous Aramaisms, there are no
other convincing or discernible traces of language influences to be found
in his work.35

3. The Greek Version of Qohelet

There is no evidence of an Old Greek equivalent of Qohelet in the Alexan-

drian translation tradition as there are for the other biblical books of the
Hebrew canon. The Pentateuchal books were most likely translated in the
third century B.C.E. and the Ketubim after that, probably at various later
dates. The Greek texts are preserved in six uncial MSS with the fourth-
century Codex Vaticanus considered to be the most reliable.36

33 Miller, Symbol and Rhetoric, 187–194.

34 Ibid., 190–194.
35 Seow concludes that although there are some loan words from other languages,
these in themselves do not provide sufficient evidence to prove that the work was indeed
Phoenician, as proposed by M. Dahood, “The Language of Qoheleth,” CBQ 14 (1952): 227–
232; idem, “Canaanite Words in Qoheleth,” Bib 46 (1965): 210–212; idem, “The Phoenician
Background of Qoheleth,” Bib 47 (1966): 264–282.
36 Hamburg Papyrus Codex (possibly end of third century), written in both Greek and
Coptic and close to the Codex Vaticanus, confirms the existence of variant readings with
little or no other significance. A complete list of the Greek manuscripts is:
166 lawrence lincoln

Heinrich Graetz put forward the suggestion that LXX Qohelet was in
fact the work of Aquila, based on a number of similarities between the
LXX and the so-called Aquilan style.37 This generated vigorous and on-
going debate on whether the extant versions came from an Aquilan origi-
nal. Although this hypothesis has been widely accepted by some, some
scholars have argued against it, most notably Hyvärinen,38 John Jarick39
and Seow.40 The search for the correct answers probably lies somewhere
in between the opposing hypotheses, and Seow is most probably correct
when he concludes that the LXX is not the work of Aquila, but a version
that, like Aquila, is motivated by the desire to facilitate certain kinds of
exegesis promoted by the rabbis. Thus the style of the LXX may be bet-
ter explained as “proto-Aquilan” but not strictly Aquilan—“not Aquila
himself.”41 Thus the LXX version is “Aquilan” in nature, while the third
column of the Hexapla is in fact the work of Aquila as noted by Origen,
a view that is also supported by Fox.42 It is not known if there had been
an earlier version than Aquila’s first edition. According to Barton, “The
translation we have in the LXX was at all events made from a text which
differed a good deal from our present Hebrew, and was therefore made
from a text that Aquiba had not revised.”43 Barton goes further by add-
ing that if this is correct, then the LXX Qohelet version we have probably
dates from the second quarter of the second century.
The character of the translation of LXX Qohelet can be described as
being extremely literal and very close to the Hebrew original in most
respects. Fox calls it a mimetic approach and “consistent in word-
correspondences,” thus making it easier to identify and possibly explain

•  Hamburg Papyrus Codex—third century

•  Codex Vaticanus—fourth century
•  Codex Sinaiticus—fourth century
•  Codex Alexandrinus—fifth century
•  Codex Ephraemi—fifth century
•  Codex Venetus—eight/ninth century.
37 Graetz’s argument is based largely on the fact that the lexeme “et is translated as
σύν.” H. Graetz, Kohelet (Leipzig: Winter, 1871), 171–189; A. H. MacNeile, An Introduction to
Ecclesiastes (Cambridge: Cambridge University Press, 1904), 115–168.
38 D. K. Hyvärinen, Die Űbersetzung von Aquila (ConBOT 10; Lund: Gleerup, 1977),
39 J. Jarick, “Aquila’s Koheleth,” Textus 15 (1990): 131–139.
40 Seow, Ecclesiastes, 7–8.
41 Ibid., 8.
42 M. V. Fox, Qohelet and his Contradictions (JSOTSup 71; Sheffield: Almond, 1987), 164.
43 Barton, Ecclesiastes, 8–11.
an analysis of the use of hebel as a metaphorical 167

the variances.44 As a result of such a rigid translation formula, in which

each lexeme attempts to match the Hebrew equivalent, the LXX Qohe-
let ended up being probably somewhat stilted and alien-sounding to the
readers of Hellenistic Greek of the time—a point made by Peter Gentry
in the introduction to the NETS translation of Qohelet.45 The translator
tried to emulate some of the Semitic stylistic forms, thereby coming up
with non-conventional syntactical elements.
Several examples to illustrate this point and demonstrate the trans-
lator’s consistency and uniformity in finding rigid and consistent Greek
equivalents to match the Semitic equivalent.

 he Hebrew ‫ גם‬is consistently rendered as καίγε (1:17; 2:1, 7, 8, 14, 15bis,

•  T
19, 21, 23, 24, 26; 3:11; 4:4, 8tris, 11, 14, 16bis; 5:9, 16, 18; 6:5, 9; 7:14, 21, 22bis;
8:10, 12, 14, 16; 9:1bis, 6bis, 12, 13; 10:20; 12:5. The Hebrew lexeme ‫ וגם‬is
also consistently translated with καίγε, as in 1:11; 3:13; 5:15; 6:3, 7; 7:6, 18;
8:17; 9:3, 11bis; 10:3, 11:2.
•  To illustrate the pedantic exactness of the translator even further, the
Hebrew lexeme ‫את‬, where it is used as a marker denoting the accusa-
tive, were in all cases rendered consistently in the LXX as σύν (1:14; 2:12,
17, 18; 3:10, 11bis, 17bis; 4:1, 2, 3, 4bis, 15; 5:3, 6; 7:14, 15, 18, 26, 29; 8:8, 9, 15,
17bis; 9:1bis, 15; 10:19, 20; 11:5, 7; 12:9, 14).
•  There are a number of other less frequent translation idiosyncrasies. In
2:1 ‫ בתוב הראה‬is rendered in Greek as καί ἰδέ ἐν ἀγαθώ, thereby chang-
ing the Hebrew idiomatic phrase from “enjoy with pleasure,” to the lit-
eral “I see in good”, which contextually does not make any sense.

In terms of the Greek equivalent Qohelet hebel is translated by the lex-

eme ματαιότης (“vanity”) in all cases, except later in Origen’s third column,
where it is translated as ἀτμίς (“breath”).46 Ματαιότης therefore ensured
that in Greek there could be no ambivalence about the meaning of hebel in
its Greek forms and it may have been meant to consistently convey a clear
ethical message, with a range of abstract meanings—empty, frivolous, vain,

44 Fox, A Time to Tear Down and a Time to Build Up, 156.

46 The exceptions are Codex Vaticanus of the Septuagint and Hamburg Papyrus, where
alternatives are used at 9:9 where ἀτμος occurs. This can also be seen in Aquila, Symma-
chus and Theodosius at 9:9 and elsewhere occasionally. This has been noted by Miller,
Symbol and Rhetoric, 59.
168 lawrence lincoln

idle and trifling.47 One can conclude that the problems of translating the
concrete term hebel was not lost on Jerome, as he also used an abstract
term in Latin, vanitas and the corresponding adjective vanus.
Gregory Thaumaturgos recognised the problems inherent in the LXX
version based on the Hebrew and reinterpreted the work considerably,
especially when dealing with the contextual situations concerning hebel/
ματαιότης. He did so by using a range of alternative terms for hebel not
only as a means to clarify and interpret, but probably also to reduce the
repetition of ματαιότης.48
In 1:14, hebel is accompanied by the phrase, ‫רעות רוח‬, (“chasing the
wind”), which is unique to the MT and anyway would not have made
much sense if literally translated as such into Greek. This phrase appears
seven times with hebel in Qohelet: 1:14; 2:11, 17, 26; 4:4, 6, 16. It therefore
appears as προαίρεσις πνεύματος, which in Greek means “preference/choice
of spirit” (NETS), thus losing the metaphorical association of the wind, in
that ‫ רוח‬in the Hebrew also referred to spirit/ soul.49 ‫ רעות רוח‬therefore
referred to the useless/pointless chasing of something that one cannot
catch and thereby also extended the metaphorical sense in which hebel
was employed.50 In Greek, however, the intended meaning of προαίρεσις
πνεύματος was to point out the dangers and the harm that can come to a
man who seeks things that are unattainable, especially if left to chance.51
Qohelet deals with youth and with being young and how young people
should enjoy this early stage of their lives without forgetting that in the
end there will ultimately be God’s judgment awaiting them, presumably

47 H. G. Liddell and R. Scott, Greek-English Lexicon (Oxford: Clarendon Press, 1975), 489;
S. Holm-Nielsen (“On the Interpretation of Qoheleth in Early Christianity,” VT 24 [1974]:
168–177) also argues that ματαιότης gives hebel a more ethical slant. Robin Salters noted
that ματαιότης meaning “purposelessness” is preferable contextually because the alterna-
tive, ἀτμίς, did not possess any figurative connotations. R. B. Salters, The Book of Ecclesi-
astes: Studies in the Versions and the History of Exegesis (PhD diss., St. Andrews 1973).
48 Gregory used alternative words such as κενός, ἀνόνητος, ἀτοπία, ἄχρηστος, δολερός,
ἄθλιος, πονηρία, ἄνοια and πλάνη.
49 In Hebrew/Jewish eschatology man possesses a body as well as a life spirit (ruaḥ).
At the time of death the body and the spirit separate and God takes back the gift of life
he granted. See, for example, Ps 104:29: “you hide your face and they are troubled; you
take away their breath, they die, and return to their dust.” Refer also to Job 34:14–15 and
Gen 2:7.
50 The use of “wind” can also be found in other sources of wisdom literature, such as,
Job 15:2; Prov 27:16; Prov 11:29; Prov 30:4 and Job 16:3. See Miller, Symbol and Rhetoric, 94.
51 In the Vulgate this is translated as universa vanitas et adflicto spiritus (“all is vanity
and distress of purpose”). This is probably based on an assumption that ‫ רעות‬is derived
from the Aramaic root ‫רעע‬. See Miller, Symbol and Rhetoric, 93.
an analysis of the use of hebel as a metaphorical 169

with approaching old age (11:9–10). In Qohelet’s thought, the state of youth
is hebel, a mere transient phase (part of a cycle) of being human. The
Hebrew uses a metaphor for this description, ‫כי הילדות והשחרות הבל‬.
It is likely that the LXX translator may have struggled somewhat to find a
suitable metaphorical phrase to match ‫ ;ילדות והשחרות‬šahªrŭt, meaning
the prime of life (and most probably derived from ‫שחר‬, “dawn”), prob-
ably did not translate well into Greek and anyway would have sounded
repetitive.52 The Greek is therefore translated as ὅτι ἡ νεότης καὶ ἡ ἄνοια
ματαιότης—“Put away anger from your heart, and divert pain from your
body, for youth and lack of understanding are vanity.” In my view, this is
another example of how subtle changes were made to the Hebrew text
in order to translate Semitic metaphors and symbols, ambiguous mean-
ings, words and repetitiveness, and at the same time to include an ethi-
cal dimension without detracting from the overall literal approach in the
translation technique.
That the LXX translator took great care word-for-word to make sure
that the translation would make sense to a Greek-speaking person can be
seen, for example, in 6:7. The Hebrew Bible version reads: “All the labour
of man is for his mouth and yet the appetite (‫ )נפש‬is not filled.” In the
LXX, nephesh is replaced with ψυχή, thereby conveying a more theological
connotation than a literal form of nephesh would have and also avoid-
ing a Hebrew metaphor which would not have translated well into the
equivalent Greek.53
An additional example comes from 1:11 and, although it is not directly
connected with hebel, it is worth quoting in order to illustrate how care-
ful the translator was to preserve the original, but at the same time avoid
textual and contextual confusion. The verse in Hebrew has an apparent
repetition which may have been confusing in a translation setting, but
the interesting point is that the Greek replaces the last two words with
εἰς τὴν ἐσχάτην, “who will be born at the last.” This is a seemingly eschato-
logical viewpoint which may have been due to later rabbinical influences.

52 Koehler-Baumgartner has “prime of life” (Lexicon, 962), while Davidson in his Ana-
lytical Hebrew and Chaldee Lexicon attributes a meaning of “dawn of life” as a metaphor
for youth (709). There is also a semantic connection with “dark hair,” presumably as a
metaphor to distinguish between a young child and an adolescent. See also W. L. Holladay,
Hebrew and Aramaic Lexicon of the Old Testament (Grand Rapids: Eerdmans, 1988), 366;
similarly in F. Brown, S. R. Driver and C. A. Briggs, Hebrew and English Lexicon of the Old
Testament (Oxford: Clarendon Press, 1906), 1007.
53 According to Koehler and Baumgartner (Lexicon, 627), nephesh is translated in the
OT 755 times by ψυχή.
170 lawrence lincoln

However, the end of day’s connotation is clear in the Vulgate and may
have been what the LXX translator was also trying to imply.

4. Conclusion

The consistent use of some idiosyncratic words and morphological con-

structions such as the use of καίγε and σύν indicates that the translator, or
translators, went to great lengths to produce a relatively literal translation
based on the original Hebrew source text without compromising on the
overall theme or message of Qohelet. Where adaptations to the meaning
of a passage were required to make them understandable in Greek, the
translator was able to do so effectively. LXX Qohelet does not detract in
any way from the essential Jewishness of its content, but it does appear
to soften its tone slightly to convey a more ethical slant. This approach is
in fact even more evident in the Peshitta. Although the Syriac version was
apparently translated from a Hebrew original, there are some similarities
to the LXX and it goes even further to spiritualise the text.54 The Aramaic
version, a much later production, on the other hand, is wholly interpreta-
tive in nature and, while maintaining the major themes of Qohelet, it is a
combination of literal translation and midrashic paraphrases in line with
the Babylonian rabbinic world view of circa seventh century.55
The value of the LXX Qohelet is really to be found in its adoption by the
Christian church when it became part of the canon of Holy Scriptures for
Christians in the Hellenistic world. Jerome recorded that he had made two
translations of Ecclesiastes from the Hebrew and that he had made exten-
sive use of the LXX as well as other Greek translations from the second
century.56 Gregory Thaumaturgos (ca. 213–270) produced the first Chris-
tian version of Ecclesiastes apparently largely based on the LXX. This work
is known as a paraphrase because of the effort made by Gregory, inspired
by his beloved mentor, Origen, to follow in his tradition of reinterpreting
and clarifying original and obscure sacred texts.57

54 Cf. Holm-Nielsen “Interpretation of Qoheleth,” 168–177.

55 Knobel, The Aramaic Bible, 9–15; E. Levine, The Aramaic version of Qohelet (New York:
Sepher-Hermon Press, 1978), 5–13.
56 Jerome, Commentarius, 1009.
57 Cf. Jarick, Gregory Thaumaturgos’ Paraphrase of Ecclesiastes.
an analysis of the use of hebel as a metaphorical 171

James Barr said that words have meaning in specific contexts and in
relation to those who use them.58 Thus the translators were faced with
having to find suitable equivalents for hebel and other Semitic words
and phrases and select appropriate Greek equivalents without detracting
from their source texts. The translators of the LXX were aware of the com-
plexities associated with the lexeme hebel and were able to distinguish
its multiple uses.59 They recognised too that hebel was not restricted to
its elemental meaning of “vapour,” “wind” or “breath,” but was meant as
a symbolic and metaphorical device to produce a wider range of nuances
and meanings. This ultimately resulted in the necessity of making appro-
priate adaptations and additions that would have made sense to Greek-
speaking Jews.
The strength of the LXX is that it does not dilute the overall Judaic
nature of the book, albeit it lacking in direct historical references to Isra-
elite tradition. The LXX retains the integrity of Qohelet’s message as well
as, presumably, that of the Vorlage at its disposal, and it made the Greek
version accessible to “Jews who spoke Greek but did not read Hebrew
or Aramaic with a translation that would reflect the Hebrew as much as
possible.”60 The Greek translation recognised the central role played by
hebel throughout the book and used the Greek form consistently so as not
to detract from Qohelet’s central thesis. Hebel provided the means to link
the opening statement in 1:2 with the concluding one in 12:8 by using it
in appropriate contexts throughout the text. Hebel therefore became the
frame of reference by which Qohelet could apply his thought and belief
system to all of human experience.

58 J. Barr, The Semantics of Biblical Language (London: Oxford University Press, 1961),
59 For the multiple uses of hebel and a detailed exposition of its metaphorical and sym-
bolic place in Qohelet; see Miller, Symbol and Rhetoric.
60 Seow, Ecclesiastes, 7.
The Θεος and Κυριος Terms in the Isaiah Text and their
Impact on the New Testament: Some Observations

Peter Nagel

1. Introduction

According to the MT, the Hebrew deity is not only referred to using the
terms ‫אלוה‬, ‫ אלהים‬and ‫אל‬,1 but also named ‫ יהוה‬while being called ‫אדני‬.2
In addition, one should also consider the so-called “biblical” manuscripts
found in and around the Judean desert, in which the paleo-Hebrew terms,
such as ‫)אלהים( אלהימ‬, ‫)אל( אל‬, ‫ )יהוה( יהוה‬and ‫( אדוי‬Adonai),
among others, were used to represent the Hebrew deity.3 It goes without
saying that these terms had to be translated into the Greek frame of
conceptual reference; hence the Hebrew deity “became” θεός, was ‘named’
κύριος, while being ‘called’ δεσποτής. The latter is not to say that the
Hebrew deity was called “δεσποτής” as such, but according to the “rule
of thumb” accepted by scholars in general, these terms are considered to

1 Noteworthy is the philological and literary approach of A. Murtonen, A Philological

and literature treatise on the Old Testament divine names [El, Eloha, Elohim] and [Yahweh]
(Helsinki: Societas Orientalis Fennica, 1952); see also Gericke’s philosophical approach
towards El-ness in J. W. Gericke, “What is an ‫ ?אל‬A Philosophical Analysis of the Concept
of Generic Godhood in the Hebrew Bible,” OTE 22/1 (2008): 21–46.
2 The in-depth investigation of M. Rösel, Adonaj—warum Gott ‘Herr’ gennant wird (Göt-
tingen: Mohr Siebeck, 2000), is considered to be foundational regarding the “naming” and
“calling upon” the Hebrew deity using Adonaj; see also Rösel’s thorough yet condensed
contribution “‫‍אדו‌‌ן‬,” in Theologisches Wörterbuch zu den Qumrantexten ([ThWQ]; Bd. 1;
eds. H.-J. Fabry and U. Dahmen; Stuttgart: Kohlhammer, 2011); cf. M. Delcor, “Des diverses
manières d’écrire le tétragramma sacré dans les anciens documents hébraïques,” RHR 147
(1955): 145–173, as well as S. T. Byington, “‫אדני יהוה‬,” JBL 76/1 (1957): 58–59, who treats of
the attestation of ‫ יהוה‬and ‫ אדני‬in the Hebrew tradition.
3 Cf. J. P. Siegel, “The Employment of Paleo-Hebrew Characters for the Divine Names
at Qumran in the Light of Tannaitic Sources,” HUCA 42 (1971): 159–172; see also, what is
considered to be a “new” substitute for the Tetragrammaton, J. M. Baumgarten, “A New
Qumran Substitute for the Divine Name and Mishnah Sukkah 4.5,” JQR 83/1–2 (1992): 1–5.
Another valuable study in this regard is D. W. Parry, “4QSama and the Tetragrammaton,”
in Current Research and Technological Developments on the Dead Sea Scrolls: Conference on
the Texts from the Judean Desert, Jerusalem (eds. D. W. Parry and S. D. Ricks; Leiden: Brill,
1996), 106–125; cf. idem, “Notes on Divine Name avoidance in Scriptural Units of the Legal
texts of Qumran,” in Legal Texts and Legal Issues: Proceedings of the Second Meeting of the
International Organization for Qumran Studies, Published in Honour of Joseph M. Baumgar-
ten (eds. M. Bernstein, F. García Martinez and J. Kampen; Leiden: Brill, 1997), 437–439.
174 peter nagel

be the most suitable Greek equivalents for their corresponding Hebrew

counterparts. Hence, θεός = ‫אלהים‬, κύριος = ‫ יהוה‬and δεσποτής = ‫אדני‬.
The problem with such a “rule of thumb” presupposition is that one has a
tendency to overlook the discrepancies, even though they might seem to
occur infrequently. The question, is how valid and useful such a “one-to-
one” literary and conceptual transmission is? The intent with this study is
not to answer the latter question, but to make some preliminary remarks
based on observations on the use of the θεός, κύριος and δεσποτής terms
in the Isaiah text against its Hebrew backdrop.4 It would be beyond the
scope of this study to discuss the use of θεός, κύριος and δεσποτής in the
whole of the Isaiah corpus; the study would therefore limit itself to those
“out-of-the-ordinary” occurrences. Stated differently, those occurrences
that deviate from the so-called “rule of thumb” will be the focal point of
the investigation. The theory that the term θεός could have been the “first
suitable” Greek term used as equivalent for the Tetragram in particular
and for the Hebrew deity in general sprung from this; a theory introduced
by Traube;5 summarised, together with other theories, by Hurtado6; and
later re-interpreted by De Troyer.7
The first step in this investigation would be to briefly consider the
Vorgeschichte of these terms as per the Isaiah text; stated differently, one
is compelled to consider the state of a possible Hebrew Vorlage in this
regard.8 This section of the investigation will be followed by an overview

4 Numerous studies pertinent to the issue of the term κύριος as a reproduction of the
Tetragram have seen the light of day; the monumental work of W. W. Baudissin, Κυριος
als Gottesname im Judentum und seine Stelle in der Religionsgeschichte (ed. O. Eissfeldt;
Giessen: A. Töpelmann, 1926) as a standard reference work on the term κύριος deserves to
be named in particular; for references made to the Hebrew deity in the Greek Psalter, see
H. U. Steymans, “Die Gottesbezeichnung kyrios im Psalter der Septuaginta,” L’Ecrit et
l’Esprit: Etudes d’histoire du texte et de théologie biblique en hommage à Adrian Shenker
(OBO 214; eds. D. Böhler, I. Himbaza and P. Hugo; Fribourg / Gottingen: Academic Press /
Vandenhoeck & Ruprecht, 2005), 325–338.
5 L. Traube, Nomina Sacra: Versuch einer Geschichte der christlichen Kürzung (München:
C. H. Beck, 1907).
6 L. W. Hurtado, “The Origin of the Nomina Sacra: A Proposal,” JBL 117/14 (1998): 655–
673; see the summary on pp. 664–671.
7 K. de Troyer, “The Pronunciation of the Names of God,” in Gott Nennen (eds. I. U.
Dalferth and P. Stoellger; Tübingen: Mohr Siebeck, 2008), 145–172.
8 For a thorough and comprehensive investigation into the Textgeschichte of the Isaiah
text, see A. van der Kooij, Die Alten Textzeugen des Jesajabuches: Ein Beitrag zur Textge-
schichte des Alten Testaments (Göttingen: Vanderhoeck & Ruprecht, 1981); see also idem,
“Isaiah in the Septuaginta,” in Writing and Reading the Scroll of Isaiah (eds. C. C. Broyles
and C. E. Evans; Leiden / New York / Köln: Brill, 2008), 513–529 (517); cf. the work of Rösel,
Adonaj, esp. 78–124 with regard to Isaiah; J. Lust in turn investigated “The Divine Titles
‫ האדון‬and ‫ אדני‬in Proto-Isaiah and Ezekiel,” in Isaiah in Context: Studies in Honour of
the θεος and κυριος terms in the isaiah text 175

of possible tendencies where “disregarding” the “rule of thumb” would

demonstrate to be most prolific.9 The third and final step will reflect on
the impact these discrepancies might have had on the New Testament
authors. The comparison of the “Old” Greek (hereafter OG) Isaiah with
the MT10 as well as witnesses from the Judean desert11 will be done on a
textual level.

2. Vorgeschichte—The Hebrew Vorlage

There are principally three terms used in the Hebrew dialect when and
if reference is made to an omnipotent, transcendental, all-powerful being
(translated in virtually all English translations as “God” with a capital “G”);
they are ‫אל‬, ‫ אלהים‬and ‫אלוה‬.12 The Isaiah text is no exception; the term
‫ אלהים‬is used in 94 instances consisting of 84 phrase structures of which
most occur in Isa 21–66; the term ‫ אל‬is attested in more than 10 verses;13
while ‫ אלוה‬is deployed only once, in Isa 44:8. The term ‫ אלהים‬is used

Arie van der Kooij on the Occasion of his Sixty-Fifth Birthday (eds. M. N. van der Meer
et al.; Leiden: Brill, 2010), 131–149. For an investigation into the translation of the divine
name limited to the Pentateuch see M. Rösel, “The Reading and Translation of the Divine
Name in the Masoretic Tradition and the Greek Pentateuch,” JSOT 31/4 (2007): 411–428.
For “names” of deities in the Ancient Near East with specialized focus on Mesopotamia
as a region, with Sumerian and Akkadian as relevant dialects, see C. Uehlinger, “Arbeit
an altorientalischen Gottesnamen—Theonomastik im Spannungsfeld von Sprache, Schrift
und Textpragmatik,” in Gott Nennen (eds. I. U. Dalferth and P. Stoellger; Tübingen: Mohr
Siebeck, 2008), 23–72. In turn, F. Hartenstein, “Die Geschichte JHWHs im Spiegel seiner
Namen,” in Gott Nennen (eds. I. U. Dalferth and P. Stoellger; Tübingen: Mohr Siebeck, 2008),
73–96, focused his attention on the history of YHWH as “name” of the Hebrew deity (cf.
W. H. Brownlee, “The Ineffable Name of God,” BASOR 226 [1977]: 39–46) counterbalanced
by the treatment of Elohim as the so-called “name” for the Hebrew deity. He does this by
probing semantics and grammatical characteristics, among others. For “Gottesnamekri-
terium” see E. Blum, “Der vermeintliche Gottesname ‘Elohim’,” in Gott Nennen (eds. I. U.
Dalferth and P. Stoellger; Tübingen: Mohr Siebeck, 2008), 97–119. For an investigation of
the “substitutes” of the Tetragram see J. Z. Lauterbach, “Substitutes for the Tetragrammaton,”
PAAJR 2 (1930–1931): 39–46.
9 Cf. De Troyer, “Pronunciation,” 144–150, who offers a brief but thorough overview
if and to what extent the Tetragram was pronounced or not and how the oldest Hebrew
manuscripts assist in determining this.
10 The MT as represented by the Biblia Hebraica Stuttgartensia: SESB Version (Stuttgart:
German Bible Society, 2003).
11  As confirmed by A. Aejmelaeus, “What can we know about the Hebrew ‘Vorlage’
of the Septuagint,” ZAW 99/1 (1987): 58–89 (58), the manuscripts found in and around
the Judean desert assist us in getting closer to how the OG’s Hebrew Vorlage might have
12 Gericke, “What is an ‫ ”?אל‬21.
13 Cf. Isa 8:5; 10:21; 12:2; 40:18; 42:10; 44:10, 15; 45:14, 15, 20, 21; 46:6, 9.
176 peter nagel

Reference 1QIsaa MT Notes LXX

3:15a ‫יהוה‬ ‫אדני‬ Superscript 1 ‫ אדני‬QIsa a

3:15b ‫יהוה‬ –
3:16 ‫יהוה‬ ‫יהוה‬ κύριος
3:17a ‫אדני‬14 ‫אדני‬ Superscript ‫ יהוה‬in 1QIsaa ὁ θεός
3:17b ‫אדני‬ ‫יהוה‬ κύριος
3:18 ‫יהוה‬ ‫אדני‬ Superscript ‫ אדני‬in 1QIsaa κύριος

in correlation with ‫ יהוה‬in more than twenty instances,15 with the term
‫ אדני‬deployed in 48 verses in comparison to ‫ יהוה‬utilised in 450 verses,
occurring 394 times.16 In addition to the use of the term ‫ אדני‬in Isaiah,
the author(s) also made use of the term ‫אדון‬.17
If one compares the MT with 1QIsaa and other related manuscripts
found in and around the Judean desert, the Hebrew text tradition appears
for the most part intact. There are, however, some small variations which
require reflection. The variants found in Isa 3:15–18 are a classic example
of Hebrew variants used to represent the Hebrew deity. The table below
presents the Hebrew variants in comparison to their Greek equivalents.
The data suggest that the scribes responsible for 1QIsaa as well as the
Masoretes appear inconsistent in applying the terms used when referring
to the Hebrew deity. The underlying issue at hand is one definable as
the Ketib-Qere problem, which requires some clarification. The standard
explanation is represented by Rösel; the Masoretes vocalised ‫ יהוה‬with
the vowels assigned to ‫אדני‬. The latter “forced” the reader to pronounce
(Qere—what ought to be read) against what was written (Ketib—what
ought to be written).18 The exception would be that if and when ‫אדני‬
‫ יהוה‬is written in combination, the term ‫ יהוה‬would be vocalised to read
‫אלהים‬, this would counter the duplicate reading of Adonai Adonai.19 An
opposing stance on this matter, of which De Troyer is a representative, is
that the most “usual” form of the Tetragram in Codex Leningrad as well
as in Codex Aleppo testifies to ‫( יְ הוָ ה‬shema—what ought to be read) and

14 4QIsab (4Q56) reads ‫אדני‬.

15 Cf. Isa 7:11; 17:6; 21:10, 17; 24:15; 25:1; 26:13; 35:2; 36:7; 37:42, 16, 20, 21; 40:28; 41:13; 48:1;
17; 49:4, 5; 51:15, 22; 55:5; 60:9.
16 The use of ‫ אדון‬and ‫ אדני‬in relation to ‫ יהוה‬and ‫ צבאות‬in Isaiah has been covered
for the most part by Rösel, Adonaj, 78–124; see also Lust, “The Divine Titles”, 131–149.
17 Cf. Rösel, Adonaj, 79.
18 Cf. Rösel, Adonaj, 2.
19 Ibid., 3; cf. De Troyer, “Pronunciation,” 144–145.
the θεος and κυριος terms in the isaiah text 177

not ‫( יְ הו ֺה‬Αdonai), implying that the vowels adopted from the Aramaic
‫ ְׁש ָמא‬indicated what ought to be read;20 even though there are exceptions
to the rule.21
Returning to Isa 3:15–18, it is thus reasonable to assume that a redac-
tor of 1QIsaa wanted to make sure that the Tetragram in Isa 3:15a is pro-
nounced Adonai while the Masoretes “wrote” what they in all probability
heard, but wrote ‫ יהוה‬as an indication of what was implied by what was
read.22 Isa 3:17a in turn seems to indicate that the Masoretes copied what
ought to be read, while Isa 3:17b testifies to the fact that they interpreted
the term ‫( אדני‬1QIsaa) as an indication of what ought to be read, but wrote
what should be written. The reverse is again visible in Isa 3:18, with 1QIsaa
bearing witness to the Ketib form ‫יהוה‬, while the redactor of 1QIsaa indi-
cated what ought to be read, ‫אדני‬. In Isa 3:18 the Masoretes thus “inserted”
‫ אדני‬into the main body of the text either based on the superscript or
they wanted the Ketib form to be representative of the Qere form;23 the
latter explanation could be rejected as mere speculation. It could also be
argued that a Hebrew Vorlage was available to the Masoretes from where
they copied the text verbatim. A similar attested issue is found in Isa 28:16,
where the MT reads ‫ אדני יהוה‬while 1QIsaa attests to ‫ יהוה‬with a super-
script ‫אדני‬.24 The ‫ יהוה־אדני‬alternating readings continue in Isa 28:22
with 1QIsaa reading ‫ יהוה‬in comparison to the MT reading ‫אדני‬. Further-
more, the MT appears to have “ignored” ‫ יהוה‬in Isa 30:19 while 1QIsaa does
indeed read the Tetragram.25 These ‫ יהוה־אדני‬alternating variants, partic-
ularly attested to in Isa 3:15–18 (1QIsaa), confirm and reinforce Rösel’s posi-
tion that the vocalisation of Adonai testifies what ought to be read when

20 De Troyer, “Pronunciation,” 145.

21  E.g. Exod 3:2 (ΜΤ).
22 Cf. De Troyer, “Pronunciation,” 144.
23 Cf. D. Trobisch, Die Endredaktion des Neuen Testaments: Eine Untersuchung zur Ent-
stehung der christlichen Bibel (Göttingen: Vanderhoeck & Ruprecht, 1996), 21 n. 19. Note-
worthy is the fact that the Greek counterpart of Isa 3:18, as represented by the LXXGött, does
not attest to any equivalent of these terms, while some Greek manuscripts read κύριος
κύριος (cf. [※ 22-48-763-96] φησι[ν] κυριος κυριος [> oII 233]) others read πιπι πιπι (cf. ※
φησι κυριος κυριος [adn. πιπι πιπι]). Similar cases of Ketib-Qere “confusion” are detectable
in Isa 6:11; 7:14; 8:7 (‫ אדני‬as superscript in 1QIsaa with a probable reading of ‫ ;)יהוה‬Isa 9:7;
21:16; 28:2, 16; 30:15; 49:7; 61:1.
24 1QIsab however, appears to be closer to the MT with the second ‫ יהוה‬reading, with
an uncertainty of what term is to be read in the first instance. Isa 30:15 attests to a similar
issue; 1QIsaa again reads ‫ יהוה‬with ‫ אדני‬superscript, compared to 4QIsac most probably
reading ‫יהוה יהוה‬.
25 Additional discrepancies are found in Isa 9:7, where 1QIsaa reads ‫ יהוה‬as opposed to
‫ אדני‬attested in the MT. A similar case is found in Isa 28:2.
178 peter nagel

‫ יהוה‬was written.26 However, such alternating readings are very limited

and should thus not be taken as the “standard” practice of the time. The
Ketib-Qere problem surrounding the “naming” of the Hebrew deity might
not have been a case of reading aloud ‫ ְׁש ָמא‬, ‫ אלהים‬or ‫ אדני‬for ‫ ;יהוה‬but
it is indeed plausible that both practices could have been employed simul-
taneously by different scribes or scribal groups. It is nevertheless clear
that “naming” or making reference to the Hebrew deity was a complex
matter, at least from the 3rd century B.C.E. onwards.
Variant readings revolving around the term ‫ אלהים‬also occur; 1QIsaa (Isa
37:20) has ‫ יהוה אלהים‬while the MT only reads ‫יהוה‬.27 Furthermore, Isa
49:14 (1QIsaa) reads ‫ יהוה ואדוני‬with a superscript ‫ ואלוהי‬directly above
‫ ואדוני‬presumably implying that Elohim is to be read which would sup-
port the argument that if and when ‫ יהוה‬and ‫ אדני‬is consecutively read,
‫ יהוה‬should be pronounced ‫ אלהים‬to avoid the repetition of Adonai.28 A
slightly different but related issue is the MT reading in Isa 50:5 attesting
to both ‫ אדני‬and ‫ יהוה‬compared to 1QIsaa reading ‫ ;אדני אלהים‬the latter
seems to indicate that the Masoretes wrote what they considered to be
an indication of what ought to be read with the term ‫ אלהים‬in 1QIsaa.29
Finally, Isa 61:1 and Isa 61:11 testify to interesting variants, presented in the
table below.

Ref 1QIsaa 1QIsab 4QIsam MT

Isa 61:1 ‫יהוה אלהים‬ ‫יהוה אלהים‬ ‫אד‬ ‫אדני יהוה‬
Isa 61:11 ‫יהוה אלהים‬ – – ‫אדני יהוה‬

Thus, both 1QIsaa as well as the MT were consistent in applying the same
terms in Isa 61:1 and Isa 61:11. The only plausible assumption one could
draw from the data is that 1QIsaa and 1QIsab present a text tradition,
opposing the text tradition offered by 4QIsam and the MT, if a ‫אדני יהוה‬
reconstruction for 4QIsam is accepted. The Greek text traditions might
shed some light on the matter, and will consequently be considered in the
section to follow.

26 Cf. Rösel, Adonaj, 2ff.

27 The Greek and Syriac equivalents in turn only account for the term ‫אלהים‬.
28 Both 1QIsab and 4QIsab (4Q56) do not attest to any superscript.
29 Cf. Isa 54:6, which differs with regard to ‫( יהוה אלהים‬1QIsaa) and ‫( אלהים‬MT).
the θεος and κυριος terms in the isaiah text 179

3. The κύριος and θεός terms: an overview

The Greek manuscripts, dated to the 2nd century B.C.E. and 2nd century
C.E., testifies to the following terms as “suitable” Greek equivalents for rep-
resenting the Hebrew deity:30

– ΙΑΩ (4QpapLXXLevb)31—Tetragram representation

– θεος P.Fouad 266b—Elohim representation
– blank space P.Fouad 266b—most probably Tetragram representation
– ‫ יהוה‬θεος P.Fouad 266b—Tetragram and Elohim representation
– ‫ יהוה‬8ḤevXIIgr—Tetragram representation
– θεος κυριος P.Oxy 656—Elohim and Adonai representation

The data snippet above presents an array of Greek possibilities in repre-

senting the Hebrew deity; the variations and discrepancies found in the
Greek version(s) of Isaiah also confirm the complexity in deciding on a
“suitable” Greek equivalent for a Hebrew term used when representing
the Hebrew deity. Sensitivity for this complexity intensifies with the reali-
sation that even the Hebrew text tradition struggled with finding ways
and means to refer to or “name” the Hebrew deity. The data from OG
Isaiah, or rather, the data inferred from a reasonable critical construction
of the OG Isaiah,32 reveals that the term δεσποτής is used as Greek equiva-
lent for ‫ אדון‬in Isa 1:24; 3:1 and 10:33,33 with the term κύριος utilised as a
representation of ‫ אדני‬in a relatively large number of cases.34 Moreover,
the term θεός is used, not without exceptions, as an equivalent for ‫אלהים‬,35
as well as for ‫אל‬36 while the term κύριος is a rendition of ‫יהוה‬. Below is a
list of some supplementary characteristics and peculiarities:

30 Cf. De Troyer, “Pronunciation,” 160–161.

31  Cf. Rösel, Adonaj, 4–5 and De Troyer, “Pronunciation,” 153.
32 J. Ziegler, Septuaginta: Vetus Testamentum Graecum, XIV: Isaias (Göttingen: Vanden-
hoeck & Ruprecht, 1983).
33 The term δεσποτής is also used, given the presumptions, as equivalent for ‫ אדני‬in
Gen 15:2, 8; Jer 1:6; 4:10; as well as ‫ אלהים‬in Job 5:8b and for ‫ יהוה‬in Jer 15:11.
34 Cf. Isa 4:4, 6:1, 8, 11; 7:14, 20; 8:7; 9:8(7); 10:12; 11:11; 21:6, 8, 16; 24:2; 26:13b; 28:2; 29:13;
30:20; 36:12; 38:14, 16; 49:14.
35 Cf. Isa 2:3; 5:16 (‫ ;)אל‬7:11; 10:21 (‫ ;)אל‬12:2 (‫ ;)אל‬13:19; 14:13 (‫ ;)אל‬17:6, 10; 21:10, 17; 25:1, 9.
36 Cf. Isa 5:16; 8:8; 10:21.
180 peter nagel

a.) Isaiah OG (Isa 4:5; 5:13; 16:14; 25:9b; 27:3; 28:21; 30:32; 41:16; 45:24; 53:1;
55:5; 58:5; 59:13) ≠ The MT reads ‫;יהוה‬37
b.) Κύριος ὁ θεός in OG Isaiah = ‫ יהוה‬MT (Isa 17:6; 24:15; 25:1; 26:13; 30:18;
37:4, 20, 21; 38:5; 41:21; 42:5, 6, 8, 13, 21; 43:1, 10, 11, 12, 14, 15; 44:2; 45:1, 3,
5, 6, 7, 11);38
c.) Isaiah OG reads the term θεός ≠ MT (Isa 9:11, 17; 13:6; 23:16; 26:9, 12;
28:13b; 41:17c; 45:23, 25; 49:7;39 the opposite is in turn attested in Isa
21:9; 53:4;
d.) The θεούς = ‫( אליל‬MT);
e.) Κύριος ὁ θεός ≠ in MT (Isa 24:16; 27:4);40
f.) Κύριος = ‫( צור‬Isa 17:10);
g.) The term κύριος in Isa 40:18 = ‫ אל‬in the MT;41
h.) The 1st κύριος term in Isa 54:5 = ‫ בעל‬in the MT;
i.) The OG Isaiah ≠ The MT, the latter testifies to the term ‫ אלהים‬in Isa
55:7; 61:2.

3.1. The Term κύριος

The term κύριος dominates the OG Isaiah, with 422 occurrences in 365
verses, 5 times of which κύριος is used in the vocative case in relation to
the term θεός with its accompanying definite article;42 in 30 instances it
is used in the nominative case in combination with the term θεός.43 Cau-
tion is duly noted when concordance related statistics are given such as
being offered here. The statistics should not be considered as necessar-
ily “argumentatively sound data”, but merely introductory in nature. The
first noteworthy issue brought to the fore by the term κύριος is found in
Isa 3:15–18; it appears as if the Greek scribes “ignored” ‫אדני יהוה צבאות‬

37 In Isa 12:2; 37:14 the MT testifies to two ‫ יהוה‬terms, compared to one κύριος term each
in both the references. The opposite is true in Isa 37:6 offering two κύριος terms weighed
against the term ‫ יהוה‬by the MT; and in Isa 57:15 two κύριος terms are read in the LXX
compared to no reference made in the MT or any other Hebrew text tradition for that
matter; cf. Isa 58:6; 61:1 (2nd κύριος term); 63:7 (4th κύριος term); 64:2.
38 This “equivalent” was consistently used, at least in Isa 41–45.
39 In Isa 37:19 the OG Isaiah reads εἴδωλα as equivalent for ‫אלהים‬, while the Greek does
not contest the term ‫ אלהים‬in Isa 60:9.
40 Another variant is attested in Isa 41:21, where the MT reads ‫ יהוה‬with κύριος ὁ θεός as
the Greek equivalent; cf. Isa 42:5, 6, 8, 13, 21; 43:1, 10, 11, 12, 14, 15; 44:2; 45:1, 3, 5, 6, 7, 11.
41 Cf. Isa 7:13 representing the term κύριος as equivalent for ‫אלהים‬.
42 Cf. Isa 24:16; 25:1; 26:12, 13 and Isa 37:20.
43 Cf. Isa 8:10; 17:6; 27:4; 30:18; 37:4, 21; 38:5; 41:17, 21; 42:5, 6, 8, 13, 21; 43:1, 3, 10, 12, 14, 15;
44:2; 45:1, 3, 5, 6, 7, 11; 51:22; 52:12 and 57:21.
the θεος and κυριος terms in the isaiah text 181

(Isa 3:15).44 Greek Isaiah continues opposing the “rule of thumb”; in Isa
3:17a with ὁ θεός seemingly representing ‫( אדני‬MT, 4QIsab) or ‫ אדני‬with
superscript ‫( יהוה‬1QIsaa). One could infer from the evidence that, as the
term θεός is used to render ‫ יהוה‬in an overwhelming amount of cases, ὁ
θεός probably represents the 1QIsaa reading (‫ אדני‬with superscript ‫)יהוה‬
in Isa 3:17a. The term κυριός in Isa 3:17b appears to be the equivalent for
either ‫( אדני‬1QIsaa) or ‫( יהוה‬MT), with the term κύριος again utilised to
represent either ‫( אדני‬MT) or ‫ יהוה‬with superscript ‫( אדני‬1QIsaa) in Isa
3:18. The reason for this “unusual” and inconsistent representation could
be argued from both a theological as well as a literary stance.
There appears to be no obvious theological cue that supports a theo-
logically determined reason why ὁ θεός was chosen as an equivalent for
the terms in question (cf. Isa 3:17a). A literary induced reason, in combina-
tion with a theological-conceptual motivation is that the superscript ‫יהוה‬
is best rendered using ὁ θεός. The latter term would call the ‘‫אל‬-ness’ of
the Hebrew deity to mind, a term that would encapsulate an overarching
monotheistic deity, best represented by ‫אל‬, ‫ אלהים‬or ‫אדני יהוה צבאות‬,
none of which are present in the Hebrew of Isa 3:17a. The Hebrew text
does however attest to ‫ אדני יהוה צבאות‬in Isa 3:15, offering the transla-
tor ample opportunity to use ὁ θεός; instead, these terms are ironically
ignored. It is, also possible that the translator completely ignored his Vor-
lage and followed his own conceptual “creativity”.
The avoidance or ignorance of the term ‫ אדני‬introduced in Isa 3:15–18,
whereby a single κύριος term, or any other related term for that matter, is
used as a Greek equivalent for ‫אדני יהוה‬, is not confined to Isa 3:15–17.45
Another intriguing Hebrew text tradition that requires due consideration
in its Greek form is Isa 61:1 and Isa 61:11 (the table below summarises the
various readings).

44 Some Greek manuscripts read an additional κυριος κυριος (※ 22-48-763-96) φησι(ν)

κυριος κυριος (> oII 233) στρατιων V-oI’ L′’`-46–233 C 403′ 770 Chr.Tht. = 𝔐).
45 Cf. Isa 7:7; 22:5, 12 (B-Qmg-109 403′ 538 read κυριος κυριος which is considered to be
“closer” to the MT; minuscule 91 in turn reads κυριος ο θεος); 28:16, 22 (+※ κυριου QSyh
86); 40:10 (The hexaplaric tradition, together with Eusebius, accounts for two κυριος terms;
whereas Hieronymus “adds” deus equivalent for the term θεός); 49:22; 50:4, 5, 7, 9 (The BHS
text-critical apparatus suggests that the LXX, Ethiopian and Arabic versions do not read
an equivalent term for ‫אדני‬. The textual evidence from the Judean desert (1Q8 Isaiah and
1QIsaa) supports the MT text reading); 52:4; 56:8; 61:1, 11; 65:13 and Isa 65:15; Cf. Isa 36:9
where the Hebrew tradition reads ‫ אדני‬compared to the absolute silence of the Greek
182 peter nagel

Ref 1QIsaa 1QIsab 4QIsam MT LXXGött Qmg ʘ

Isa 61:1 ‫אד יהוה אלהים יהוה אלהים‬ ‫ אדני יהוה‬κύριος κύριος
Isa 61:11 ‫– יהוה אלהים‬ – ‫ אדני יהוה‬κύριος

If the “rule of thumb” presupposition is accepted as “reasonable”, then

one could assume that in this case the term κύριος represents the read-
ing of 1QIsaa best. When dealing with the Isaiah text, however, the more
argumentatively sound inference is that the only “rule of thumb” to be
adopted in general is that there are not any.46 The latter is confirmed
when one considers Isa 61:11 in relation to Isa 61:1a; the former testifies to a
single κύριος term as opposed to ‫( אדני יהוה‬MT), while 1QIsaa reads ‫יהוה‬
‫אלהים‬. Thus, in Isa 61:1a the term κύριος best represents either the MT
(used once to avoid repetition of the term ‫)אדני‬, or 1QIsaa. If the scribes
are considered to be consistent, the term κύριος best represents the MT in
Isa 61:11. One thing that could be deduced from the evidence is that LXXGött
seems to have simplified the matter by consistently reading a single term
κύριος, while manuscripts belonging to the hexaplaric tradition in general
follow the MT, as expected.
Another captivating issue is the “omission” of both the ‫ יהוה‬terms,
including ‫ צבאות‬in Isa 22:14a and ‫ צבאות‬. . . ‫ אדני‬in Isa 22:14b.47 The lat-
ter two occurrences do indeed strengthen the idea that the term ‫ אדני‬is to
a large extent “downplayed” by the translators of OG Isaiah. Isa 49:14 (‫יהוה‬
‫ )אדני‬offers another twist to the ‫—אדני יהוה‬κύριος dilemma; the LXXGött
renders κύριος ὁ κύριος as equivalent with an array of alternatives offered
by manuscripts, varying between the term θεός in the first instance,48 while
one is expected to read the term θεός in the second instance.49 The avoid-
ance of the term ‫אדני‬, especially when used in combination with ‫יהוה‬, is
wonderfully illustrated in the catena of occurrences found in Isa 50:1–10
(Isa 50:4, 5, 7 and 9); all the ‫( אדני יהוה‬MT) occurrences are rendered

46 Van der Kooij (“Isaiah in the Septuaginta,” 518) confirms that it is generally accepted
that the LXX Isaiah follows a more “free” approach towards the original, while mentioning
that Ziegler and Seeligmann are of the opinion, an opinion he supports, that the Vorlage
of the LXX Isaiah does not differ much from the MT. These statements might appear to
be confusing; why would a solid Hebrew text tradition cause such a “loose” or “free” Greek
equivalent, especially with theologically significant terms such as discussed here?
47 The “omission” of ‫ אדני‬in Isa 22:14b is also attested in the Syriac as well as in
48 κύριος1°] ο θεος 538.
49 κύριος2°] θεος A′ 88–oII L′’`-86c-233 564 403′ 534.
the θεος and κυριος terms in the isaiah text 183

with a single term κύριος. These instances, together with others, seem
to indicate a “discomfort” with the term ‫ אדני‬on the part of the Greek
translators, the reason of which is currently unknown and thus demands
further investigation.

2.2. The Term θεός

The term θεός occurs 195 times in 174 verses; in almost all cases it is accom-
panied with a definite article and is a term that implies the Hebrew deity
“proper”, except for its nominative plural use in Isa 37:19; 41:23; 42:17 and
the accusative plural use in Isa 44:15, the latter which refers to deities in
general, not the Hebrew deity in particular. The theory that the term θεός
was considered to be the most suitable Greek equivalent to represent ‫יהוה‬
appears to be a plausible one, at least in OG Isaiah. The data produced
from a “surface” comparison between the Hebrew and Greek text tradi-
tions of Isaiah, indicates that the term θεός is used, in a large number of
cases, as an equivalent for the Tetragram.50 Because the “general rule of
thumb” states that θεός is the Greek equivalent for ‫אלהים‬, the term is
conceptually devaluated as a reasonable candidate to represent ‫יהוה‬. A
channel of conceptual possibilities is opened up when θεός is seen as a
significant Greek equivalent for the Tetragram from a very early stage in
the Greek text tradition; a theory that would require a broader investiga-
tion into how the Tetragram was rendered elsewhere. For now the focus
will be confined to the Isaiah text.

Ref 1QIsaa MT LXXGött Sc 90–456 S* 36 O’-Qmg oII Rom

377–cII 9:28
393 403′
Isa ‫אדני‬ ‫ אדני‬ὁ θεος κύριος ὁ κύριος κύριος
10:23 ‫יהוה‬ ‫יהוה‬
Isa ‫אדני‬ ‫ אדני‬κύριος κύριος ὁ θεός
10:24 ‫יהוה‬ ‫יהוה‬ κύριος
Isa ‫יהוה‬ ‫ יהוה‬ὁ θεος

50 Cf. Isa 1:10; 2:2 4:2; 6:12; 7:17; 8:17, 18; 9:11(10); 10:20, 23, 26; 11:2, 3; 13:6; 14:2, 3, 5, 27;
24:21; 25:10; 26:4b; 27:1; 30:9, 18, 29; 31:1; 33:5, 22; 36:15, 18, 20b; 37:20b, 22; 38:7, 20b, 22; 40:7,
27, 28, 31; 41:4, 14; 42:12, 19, 24; 44:5, 6, 23; 48:17b; 51:13; 54:5, 13; 55:6; 58:8, 9, 11, 13; 61:2; 65:23.
The reverse is attested in Isa 7:13, where the term κύριος appears to be an equivalent for
‫ ;אלהים‬cf. Isa 61:10; 62:5.
184 peter nagel

In Isa 10:23 the LXXGött reads ὁ θεός as an equivalent for ‫ אדני יהוה‬supported
by the first hand of codices A and S, while codex B reads the term κύριος.
Other Greek manuscripts alternate between two possibilities; κύριος sup-
ported by some manuscripts belonging to the hexaplaric tradition,51 while
others include the definite article.52 Isa 10:24 in turn opted for the term
κύριος as an equivalent for ‫אדני יהוה‬. This reading is opposed by the first
hand of codex S which reads κυριος ο θεος, supported by some manuscripts
belonging to the hexaplaric tradition.53 A corrector of S offers another
possibility, reading κυριος κυριος.54 In Isa 10:26, the translators again opted
to use ὁ θεός to render ‫יהוה‬. In all these cases, the Hebrew text tradition
is supported by 1QIsaa. The “out-of-the-ordinary” cluster of attested terms
found in ch. 14 (Isa 14:2, 3, 5 and 27) all read the term θεός compared to the
MT which only attest to the term ‫יהוה‬. There seems to be no evidence to
suggest the contrary, showing that the Hebrew text tradition is intact. Isa
26:12 testifies to κύριος ὁ θεός as equivalent for ‫יהוה‬, the only occurrence
of its kind in the LXXGött.55 In 26:13, however, the MT reads ‫יהוה אלהים‬
with the expected Greek counterpart κύριος ὁ θεός. The reading in Isa 26:12
only makes sense if a “being consistent” argument is adopted to explain
the unusual Greek equivalent for the Tetragram. The reading κυρίου τοῦ
θεοῦ in Isa 28:13 is repeated only in Isa 51:20 in the whole of the LXXGött,56
while some manuscripts (Is 28:13) in the hexaplaric tradition together
with the Lucianic tradition calls for an “omission” of τοῦ θεοῦ,57 with the
catena manuscript group supporting the “omission” of κυρίου.58 A further
noteworthy case is the θεός reading in Isa 37:20b; the MT reads ‫ יהוה‬with
1QIsaa testifying to ‫יהוה אלהים‬, the latter which opens the possibility that
the 1QIsaa reading’s ‫ אלהים‬could present a possible alternative Hebrew
Vorlage for the LXXGött reading if the theory is not accepted that the term
θεός was considered to be the most “suitable” Greek equivalent for ‫יהוה‬.59

51 O’-Qmg, a reading that is confirmed by the Syro-Palestinian translation and the church
fathers Eusebius, Basilius and Tertullian.
52 (ο κυρ. oII).
53 + ο θεος S* 36 377–cII 393 403′; A 2nd hand of codex S reads κς ο κς while a 3rd hand
opted for κς κς. Some manuscripts belonging to the Lucianic tradition “omits” the definite
54 + (※) κυριος Sc 90–456 = 𝔐 ↓.
55 Cf. Isa 17:6; 24:15; 25:1; 26:13; 30:18; 37:4, 20, 21; 38:5; 42:6, 8, 13, 21.
56 Cf. Isa 7:11; 48:1.
57 om. τοῦ θεοῦ O′’ L′’`-233-456 301 403′ 449′ 534 Eus.Tht.Hi. = 𝔐; om. τοῦ 393.
58 om. κυρίου C′’-566.
59 Contra Van der Kooij, “Isaiah in the Septuaginta,” 518–518. In Isa 41:13 one finds a
similar case, where the MT reads ‫ יהוה אלהים‬while the LXX only reads θεος; cf. Isa 51:15.
the θεος and κυριος terms in the isaiah text 185

One would, however, have to justify why the translator would “discard”
‫יהוה‬. It is, probable that the translator did not consider it necessary to
reproduce an equivalent for both the Hebrew terms, as it was prohibited
to pronounce the former. The occurrence in Isa 38:11 seems to confirm the
theory that the term θεός is considered the most suitable equivalent for
‫ ;יהוה‬the MT reads ‫ יה יה‬with two Hebrew manuscripts testifying to ‫יהוה‬
supported by both Symmachus as well as the Syriac translation with the
LXXGött again reading θεός.
In Isa 41:17 κύριος ὁ θεός appears to represent ‫( יהוה‬MT) with no text
witnesses suggesting an alternative Hebrew reading,60 while κύριος ὁ θεός
reproduces ‫ האל יהוה‬in Isa 42:5; in the latter case an alternative Hebrew
text reading is found in 1QIsaa reading ‫אל אלהים‬.61 It is suggested that the
κύριος ὁ θεός reading in Isa 41:17 as well as Isa 42:5, among others, should
not be explained on account of its Hebrew Vorlage. One should rather
consider the κύριος ὁ θεός construction as theologically determined, espe-
cially in Isa 41–45. In these chapters, the κύριος ὁ θεός construction appears
to be governed by the first personal pronoun ἐγώ together with λέγει,62
which could imply that for the translators, if and when the Hebrew deity
declares or utters something in reference to the godly self, the existential
theological I, as logion, the theological maximus is required. If this is a
reasonable assumption, it fuels the proposal that neither κύριος nor θεός
are “suitable” Greek equivalents to render the Hebrew deity proper. This
theologically determined understanding of κύριος ὁ θεός is supported by
Isa 57:21 reading κύριος ὁ θεός in comparison to the MT which reads ‫אלהי‬
Another variant reading is attested in Isa 51:15; the term ὁ θεός with
a definite article appears to be the Greek equivalent for ‫יהוה אלהים‬64
where one would have expected a Greek equivalent κύριος ὁ θεός.65 In Isa
43:1 the Greek text reads κύριος ὁ θεός compared to the MT that only has

60 Some Greek manuscripts opted for ο αγιος (Q ʘ), while others “omitted” κυριος (198
538 Sa); cf. Isa 41:20.
61 Ms 51 “omits” κύριος while some manuscripts from the hexaplaric tradition, including
early Church fathers (oII 407 410 Ir. [hab.] Cyr.[lem]) in turn “omit” ὁ θεός.
62 A noteworthy exception is found in Isa 44:24, reading Οὕτως λέγει κύριος ὁ λυτρούμενός
σε καὶ πλάσσων σε ἐκ κοιλίας Ἐγὼ κύριος ὁ συντελῶν πάντα ἐξέτεινα τὸν οὐρανὸν μόνος καὶ
ἐστερέωσα τὴν γῆν; this phrase seems to suggest an opposing stance to what is offered as
an explanation for the frequent use of κύριος ὁ θεός in Isa 41–45.
63 1QIsaa, 1QIsab and 4QIsad confirm the Hebrew text reading.
64 Cf. Isa 41:13; 48:17; 55:5.
65 The Hebrew text tradition does not offer an alternative that would “justify” the cur-
rent Greek equivalent.
186 peter nagel

the term ‫יהוה‬. Why the inconsistency? In Isa 51:15 it is probably a case
of influence from the Vorlage; such a possibility is introduced by 4QIsac
which does not testify to the term ‫אלהים‬, but only to ‫יהוה‬, a term fre-
quently represented by the term θεός, as shown above. In the case of Isa
43:1, the reason behind the “out-of-the-ordinary” reading could be theo-
logical consideration, as was pointed out above. Finally, in Isa 60:9 “nam-
ing” the Hebrew deity is brought to the fore; according to OG Isaiah it is
the name κύριος that is deemed holy, while for the Hebrew text it is ‫יהוה‬
‫אלהים‬. Before moving on to how the variants, alternatives and peculiari-
ties could have impacted the New Testament authors, it might be worth-
while to examine how the OG Isaiah dealt with ‫ ִע ָ ּ֥מנּו ֵ ֽאל‬which came to
be known in the New Testament as Ἐμμανουήλ (Matt 1:23). The concept
‫ ִע ָ ּ֥מנּו ֵ ֽאל‬occurs only three times in the MT, all of which are confined to
Isaiah. In 7:14 the Hebrew terms are rendered with Ἐμμανουήλ. In Isa 8:8
however, the Hebrew concept is translated with μεθ᾿ ἡμῶν ὁ θεός. Finally,
in Isa 8:10 the phrase in turn is represented using μεθ᾿ ἡμῶν κύριος ὁ θεός;
three distinct Greek equivalents, one of which found its way into the New
Testament (Isa 7:14 cited in Matt 1:23a) and one cited in Matt 1:23b (Isa 8:8).66
The effect of this translation will be discussed in the next section of
this paper.

4. The New Testament impact

Reflecting upon the Isaiah text, its history within Christianity and its use
in the New Testament is by no means a new venture. A general overview
of the history of such influence was undertaken by John Sawyer in 1996,67
followed by a compilation of essays edited by Steve Moyise and Maarten
Menken on those New Testament books that contain Isaiah citations.68
Studies with a more focused and specialised approach are works such as

66 NA27 notes that the phrase μεθ᾿ ἡμῶν ὁ θεός is sourced from either Isa 8:8 or Isa 8:10,
which is technically incorrect; the reading in Isa 8:10 is clearly testifying to κύριος ὁ θεός
and not ὁ θεός.
67 J. F. A. Sawyer, The fifth gospel: Isaiah in the history of Christianity (Cambridge: Cam-
bridge University Press, 1996).
68 S. Moyise and M. Menken, eds., Isaiah in the New Testament: The New Testament and
Scriptures of Israel (London / New York: T & T Clark, 2005).
the θεος and κυριος terms in the isaiah text 187

that of Koch,69 Stanley,70 Wilk,71 Shiu-Lun Shum72 and Wagner;73 all of

whom investigated in some way or form the use of Isaiah as scripture in
the Pauline literature. Peter Mallen considered the reading and transfor-
mation of Isaiah in Luke–Acts.74 The objective with this section of the
investigation is to consider the impact the Isaianic use of the θεός and
κύριος terms might have had on the New Testament authors, using Matt
1:23 (Isa 7:14 and Isa 8:8) and Rom 9:27–29 as test cases; the latter which
testifies to two explicit Isaiah citations, Isa 10:22–23 (Rom 9:27–28) and
Isa 1:9 (Rom 9:29).

4.1. Isa 7:14; 8:8 (Matt 1:23)

The impact of Isa 7:14 and 8:8 in the text of Matt 1:23 and the significance
thereof should not be underestimated. The introductory formula launch-
ing the citation reads: ἵνα πληρωθῆ τὸ ῥηθὲν ὑπὸ κυρίου διὰ τοῦ προφήτου
λέγοντος (“for the fulfilment of what kyrios said through the prophet”).
Kyrios is thus the primary acting agent, speaking through the prophet Isa-
iah, in this case, saying: ἰδοὺ ἡ παρθένος ἐν γαστρὶ ἕξει καὶ τέξεται υἱόν, καὶ
καλέσουσιν τὸ ὄνομα αὐτοῦ Ἐμμανουήλ (“behold, the virgin will have in her
womb and she will give birth to a son, and they will call him Emman-
uel”). The author continues by offering a translation of the Aramaic term
with the words: ὅ ἐστιν μεθερμηνευόμενον μεθ᾿ ἡμῶν ὁ θεός (“which is trans-
lated: Theos with us”). Interpreting these citations within their immedi-
ate thought-structure (Matt 1:18–25), the import of the κύριος-θεός citation
becomes evident. Logically inferred, it is Ἰησοῦ Χριστοῦ (Matt 1:18) who
should be named Ἰησοῦν (Matt 1:21) and Ἐμμανουήλ (Matt 1:23a); the lat-
ter holding the semantic potential of μεθ᾿ ἡμῶν ὁ θεός (Matt 1:23b). The
citation in Matt 1:23a causes both author and reader to draw a direct con-
ceptual link between Ἰησοῦν and Ἐμμανουήλ. The latter would supposedly

69 D.-A. Koch, Die Schrift als Zeuge des Evangeliums (Tübingen: Mohr Siebeck, 1986).
70 C. D. Stanley, Paul and the language of Scripture: Citation technique in the Pauline
Epistles and contemporary literature (Cambridge / New York: Cambridge University Press,
71 F. Wilk, Die Bedeutung des Jesajabuches für Paulus (FRLANT 179; Göttingen: Vanden-
hoeck & Ruprecht, 1998).
72 Shiu-Lun Shum, Paul’s use of Isaiah in Romans: a comparative study of Paul’s Letter to
the Romans and the Sibylline and Qumran sectarian texts (Tübingen: Mohr Siebeck, 2002).
73 J. R. Wagner, Heralds of the Good News: Isaiah and Paul in Concert in the Letter to the
Romans (Leiden: Brill, 2003).
74 P. Mallen, Reading and Transformation of Isaiah in Luke-Acts (London / New York:
T & T Clark, 2007).
188 peter nagel

not have been an obvious link, if the Aramaic ‫ׁשּוע‬

ַ ֵ‫י‬, semantically related
to ‫ׁשּועה‬
ָ ְ‫י‬, was considered and understood as a literary-conceptual fore-
runner of Ἰησοῦς with the potential meaning “helper, saviour”. If such an
assumption is accepted as plausible, then the citation taken from Isaiah
expands the possible meanings of the term Ἰησοῦν from the semantic pros-
pect of “saviour” to Emmanuel to the semantic value of Theos with us. This
expansion of the semantic field results in the conclusion: Ἰησοῦς is Theos
among us. It is doubtful that the author of the Matthean gospel could have
“creatively” conceptualised that Ἰησοῦς is Ἐμμανουήλ and therefore μεθ᾿
ἡμῶν ὁ θεός without the support of Isa 7:14 and Isa 8:8. It remains an open-
ended question, if and to what extent the Matthean author considered the
theological implication of his citation.

4.2. Isa 10:22c–23 (Rom 9:27–28)

The author creates a Χριστός—θεός theological frame of reference in Rom
9:1–25 with its theological thrust fixated in and around Rom 9:5.75 The fol-
lowing theos-kyrios concept (Rom 9:26–33) is made possible by the three
cited texts: Hos 2:1b–c (Rom 9:26), Isa 10:22c–23 (Rom 9:27–28) and Isa
1:9 (Rom 9:29).76 Although the attention of this study is on the impact
the Isaiah citations might have had on Pauline thought, it is worth noting
that the textual integrity of both Hos 2:1b–c and Rom 9:26 in terms of θεός
and κύριος appears to be intact.77 Literarily and theological-conceptually
speaking the term θεός in Hos 2:1b–c could not have referred to any other
entity than ‫אל‬, the “wisdom” deity of ancient Israel. The same, though,
cannot be said for Isa 10:22c–23 in Rom 9:27–28. A variety of possibilities
are presented by the Greek witnesses for both Old Testament (hereafter
OT) and New Testament (hereafter NT) texts.78 All the NT text witnesses

75 Rom 9:5 sets the literary context in which the intriguing thrust of the inter-relatedness
of the Χριστός and θεός is put to the fore.
76 According to H. Schlier, Der Römerbrief: Kommentar (Freiburg: Herder, 1977), 303,
Paul wants to appeal that the Gentiles belongs to the people of God with the Hos 2:1
citation; in turn, the Isa 10:22–23 confirms the true Israel as the Ekklesia. These citations
also confirm the sovereign action of God (ibid., 304). For Schlier (ibid., 304), it is evident
that ἐκεῖ κληθήσονται υἱοὶ θεοῦ ζῶντος (Rom 9:26) indicates Paul’s apposing stance over and
against the Jewish-Apocalyptic and Rabbinic position. Koch (Schrift, 146), in turn suggests
that the redaction of the Israelites in Isa 10:22 is portrait as a legal act of Yahweh.
77 The Hebrew text tradition reads ‫ אל‬with the Greek text tradition offering θεός as
78 The Hebrew text tradition seems to be intact for the most part, except for a text
critical note in the BHS on Isa 10:23 that states that two Hebrew manuscripts “omits” ‫יהוה‬
the θεος and κυριος terms in the isaiah text 189

read κύριος, while the OT text witnesses vary between ὁ θεός, κύριος κύριος
and κύριος. From the textual evidence it can be deduced that the ‫ְצ ָבֹאות‬
term was either “ignored” from a very early stage of transmission or it had
been “omitted” by the Greek translators. The fact remains that ‫אדני יהוה‬,
when used in combination, often appears to have been contracted into
one single Greek equivalent. Paul’s “consistent” use of the term κύριος in
Rom 9:28 and Rom 9:29 as sourced from Isa 10:23 and Isa 1:9 respectively,
together with the fact that he had ample theologically determined reason
to “alter” the citation, points to the fact that he stringently followed his
Vorlage. If the term κύριος belongs to a Vorlage, the reading in Rom 9:28
would seem to be in line with some manuscripts from the hexapla. The
theological thrust of ‫“ אדוני ְצ ָבֹאות יהוה‬Lord God of Hosts” was already
watered down by the Greek translators from a very early stage because
they struggled to find a “suitable” Greek equivalent for these Hebrew
terms in combination. This struggle spilled over into the New Testament.
It is thus more than possible that “other” equivalents for these terms were
available, one of which might have primarily impacted Paul. This would
imply, taking Rom 9:26 which attests to the term θεός into consideration,
that in Rom 9:28 and Rom 9:29 Paul regarded the referent of both the
terms θεός and κύριος as one and the same entity: the Hebrew deity. Is
this really the case?
What then is the implication of these κύριος citations? Is Paul calling
Jesus to mind as the κύριος or as ‫ ?יהוה‬In Rom 9:8 Paul speaks of τοῦτʼ
ἔστιν, οὐ τὰ τέκνα τῆς σαρκὸς ταῦτα τέκνα τοῦ θεοῦ (“that is, not the children
of flesh are these children of Theos”),79 while he asserts in Rom 9:29 that
εἰ μὴ κύριος σαβαὼθ ἐγκατέλιπεν ἡμῖν σπέρμα (“if Kyrios, lord of hosts, did
not leave a remnant behind for us”). Both κύριος and θεός thus reserve the
right to decide which nation or clan to accept or deny, to include or to
exclude. Moreover, Paul also commands that ὁ λόγος τοῦ θεοῦ (Rom 9:6)
should not be considered invalid, for it will come to fulfilment once and
for all on earth (λόγον γὰρ συντελῶν καὶ συντέμνων ποιήσει κύριος ἐπὶ τῆς
γῆς [Rom 9:28]). The Hos 2:1b–c citation in Rom 9:26 reinforces the θεός
concept introduced in Rom 9:1–25; a concept that called the “supreme”
Hebrew deity to mind. The two κύριος citations (Isa 10:23 and Isa 1:9) seem
out of place. However, with these two citations, due to their overlapping
theme, Paul appears to regard the term θεός (Rom 9:6–8) and the term
κύριος (Rom 9:28) as having the same referent; the same could be said

79 Cf. Rom 9:7.

190 peter nagel

for Rom 9:29. It remains unclear if Paul adopted the “underlying” concept
of the term κύριος to be a rendering of the Tetragram, although it seems
quite plausible. Paul is thus not calling to mind Jesus as the κύριος, but the
authoritative nature of the Hebrew deity. These two Isaiah citations sug-
gest that the κύριος term within the explicit citations refers to a different
entity than the κύριος term which clearly refers to Jesus as the Christ. The
deployment of the two Isaiah citations would ultimately have an impact
on how one interprets the conceptual “relationship” between the terms
Χριστός and θεός in Rom 9:5. How do the quotations considered above and
the conclusions drawn about their impact reflect on the interpretation
and understanding of Rom 9:5?
The term Χριστός in Rom 9:5 should first and foremost be understood
in relation to the Χριστός terms in Rom 9:1 and Rom 9:3. In both these
instances the term Χριστός is presented within a prepositional clause:
Ἀλήθειαν λέγω ἐν Χριστῷ and ἐγὼ ἀπὸ τοῦ Χριστοῦ respectively. It should
further be noted that in both cases the first person singular pronoun in
the relevant clause, which refers to Paul, is used. In Rom 9:1 Paul’s truth is
considered to be justified ἐν Christ (Rom 9:1), while ἀπό refers to a second-
ary position over and against someone who is ἐν Christ (Rom 9:3).80 One
could thus, with a reasonable amount of certainty, conclude that the con-
cept underlying the term Χριστός in these verses are one of Christ being
a mediator. On the other hand, it would be difficult to deny that through
this mediating role, Χριστός, in the mind of Paul, should be praised as θεός,
who is θεός over all.81 This concept is strengthened when one considers
the idea that ἐγὼ ἀπὸ τοῦ Χριστοῦ ὑπὲρ τῶν ἀδελφῶν μου τῶν συγγενῶν μου
κατὰ σάρκα seamlessly fits into the concept of θεός’ free will to make or
appoint nations, clans or any group as “his sons” or “his children” as is evi-
dently assigned to both θεός and κύριος in ch. 9. Not only is θεός and κύριος
conceptually considered terms referring to the Hebrew deity, but in this
case the Χριστός term also belongs to this concept. The impact of the term
κύριος in both Isa 1:9 and Isa 10:23 leads one to deduce: (1) Paul stringently
follows his Vorlage. (2) He considered both the terms θεός and κύριος as
terms that “translate” or “represent” the Hebrew deity, even though this
is not always the case.

80 Cf. E. Käsemann, Commentary on Romans (London: SCM Press, 1980), 259.

81 To quote H.-C. Kammler, “Die Prädikation Jesu Christi als ‘Gott’ und die paulini-
sche Christologie: Erwägungen zur Exegese von Rom 9,5b,” ZNW 94/3–4 (2003): 171: “Der
Begriff θεός wäre dabei nicht im Sinne eines Nomen proprium verwendet, sondern als
the θεος and κυριος terms in the isaiah text 191

5. Conclusion

Rendering those Hebrew terms used when reference was made to the
Hebrew deity was as complex as the transmission of the Hebrew text
itself. The term ‫ אדני‬was frequently “misrepresented” if used in correla-
tion with the term ‫יהוה‬. Moreover, the Ketib-Qere perplexity had a major
influence on how the terms referring to the Hebrew deity was rendered;
this in turn spilled over and impacted the Greek translation of the Hebrew
text. The sensitivity, and hence, the prohibition in uttering the “name” of
the Hebrew deity became a major factor in the rendering of the terms
referring to the Hebrew deity in its Greek conceptual frame of reference.
The inference drawn from the investigated data of Isaiah is that the term
‫ אדון‬is consistently rendered using the term δεσποτής as a Greek equiva-
lent, while the term κύριος was primarily employed as the Greek equiva-
lent for the term ‫ אדני‬and ‫יהוה‬. However, the Greek translation of ‫יהוה‬
by κύριος is not without exception; in fact, the evidence appears to suggest
that the term θεός could be considered as the most suitable term not only
as a Greek equivalent for ‫יהוה‬, but as a term that represents the Hebrew
deity in general. The data suggests that the term θεός was considered, from
the 2nd century B.C.E. onwards, as a suitable Greek equivalent not only as
a rendering of the terms referring to the Hebrew deity proper, but also for
the personal Hebrew deity “named” ‫יהוה‬. Therefore, the “rule of thumb”
or any other systematised rendering of the Hebrew deity in OG Isaiah, at
least, should be discarded. Finally, due to the Greek variants available to
the New Testament authors, the theological and conceptual shift made by
the Greek translators had an impact not so much on the structure of the
New Testament text, but undoubtedly on the theo-logy and kyrio-logie of
the New Testament.
Revisiting the Original Greek of Ezekiel 18

Harry F. van Rooy

1. Introduction

When Ziegler1 prepared the critical text of Ezekiel for the Göttingen edi-
tion of the Septuagint (first published in 1952), he gave pride of place to
the B text, with Codex Vaticanus (B) and Papyrus 967 as the most impor-
tant witnesses. However, when he compiled his text, he had only a part of
this papyrus at his disposal. The publication of the parts of the papyrus in
Cologne2 and Madrid3 has made it possible to revisit the original Greek.
In some instances, Ziegler did not accept the reading of B as his origi-
nal Greek. The question is whether readings of Papyrus 967 may change
this choice in some instances. In many instances the papyrus supports
the choice of the reading of B as the original Greek. However, in some
instances where Ziegler did not accept the reading of B, this papyrus sup-
ports B, whereas in others it disagrees with B and supports the reading of
Ziegler. This paper will examine the readings of Ziegler in Ezekiel 18 in
the light of the readings of Papyrus 967. Many examples of Papyrus 967
agreeing with B and the critical text occur, but examples where Papyrus
967 and B agree with a reading not accepted by Ziegler (18:4, 9), or where
Papyrus 967 agrees with the critical text against B (18:10), are important.
These readings will be evaluated, with suggestions for revising the original
Greek of Ezekiel 18.

2. Papyrus 967 in Recent Discussions

As regards the value of Papyrus 967, two main views can be distinguished.
Some scholars regard the papyrus as very important for the history of the

1 J. Ziegler, Sepuaginta: Vetus Testamentum Graecum, XVI, 1: Ezechiel (Göttingen: Van-
denhoeck & Ruprecht, 1977).
2 L. G. Jahn, Der griechische Text des Buches Ezechiel nach dem Kölner Teil des Papyrus
967 (Papyrologische Texte und Abhandlungen 15; Bonn: Habelt, 1972.)
3 M. Fernández-Galliano. “Nuevas páginas del códice 967 del A.T. Grieco (Ez 28,19–
43,9),” SPap 10 (1971): 5–79.
194 harry f. van rooy

text of Ezekiel in the Hebrew transmission of the book, with the papyrus
representing an older version of the Hebrew. On the other hand, there are
scholars who restrict the importance of this papyrus to the transmission
of the Greek Ezekiel. It is impossible to discuss these questions in detail
in this paper. Important surveys of the problems and proposed conclu-
sions can be found in the works of Schwagmeier,4 Flanagan,5 Olley6 and
Schwagmeier, Flanagan and Olley support the idea that Papyrus 967 is
very important for the history of the Hebrew Ezekiel as well, while Flana-
gan argues in favour of its importance being restricted to the Greek trans-
mission of the book. In both these views, the fact that Ezekiel 36:23–30 is
omitted by the papyrus and the rearrangement of chapters 36–39 play an
important role. In the first view the insertion of 36:23–38 is regarded as
a late feature in the Hebrew tradition, while the rearrangement of 36–39
must be regarded as original. In the second view, these two factors are
ascribed to the transmission of Ezekiel in Greek. These issues will prob-
ably not find a consensus very easily. The aim of this paper is, however,
not related to that problem, but rather to the importance of Papyrus 967
for determining the original Greek. The manuscript frequently has read-
ings agreeing with the text as reconstructed by Ziegler, frequently agree-
ing with B. There are many instances, however, where the papyrus does
not agree with the text of Ziegler, agreeing or disagreeing with B. It will
be impossible to give attention to all the variants in Ezekiel 18, but a
re­presentative sample will be discussed to throw light on the use of Papy-
rus 967 in determining the reading of the original Greek in Ezekiel 18.
As Papyrus 967 is an important pre-hexaplaric witness to the Greek
text of Ezekiel, attention will first be given to passages that are important
for the hexaplaric tradition of Ezekiel, then to passages important for the
relationship between B and Papyrus 967, and finally to some other inter-
esting variants in Papyrus 967.

4 P. Schwagmeier, Untersuchungen zu Textgeschichte und Entstehung des Ezechielbuches

in masoretischer und griechischen Überlieferung (D.Th. diss., University of Zurich, 2004),
5 J. Flanagan, “Papyrus 967 and the Text of Ezekiel: Parablepsis or an Original Text?” in
Studies in Scripture in Early Judaism and Christianity (Library of Second Temple Studies 70;
ed. C. A. Evans; London: T & T Clark, 2009, 105–116).
6 J. W. Olley, Ezekiel. A Commentary Based on Iezekiel in Codex Vaticanus (Septuagint
Commentary Series; Leiden / Boston: Brill, 2009), 8–9.
7 D. M. O’Hare, “Have You Seen, Son of Man?” A Study in the Translation and Vorlage of
LXX Ezekiel 40–48 (SCS 57; Atlanta: Society of Biblical Literature, 2010), 7–19.
revisiting the original greek of ezekiel 18 195

3. Papyrus 967 and the Hexaplaric Tradition

3.1. Readings Marked with an Asterisk Not in Papyrus 967

Papyrus 967 was published over a period of time, with sections of the
papyrus in different localities. Ezekiel 18 is in that part of the papyrus in
Cologne, published by Jahn in 1972.8 In his discussion of the readings of
Papyrus 967, Jahn looks at all the instances where comparisons with hexa-
plaric readings can be made. In the first place he discusses the instances
where hexaplaric readings are indicated with an asterisk in hexaplaric
witnesses and where these readings do not occur in Papyrus 967, as is
to be expected. He discusses five such examples in Ezekiel 18. The first
of them is in Ezekiel 18:16. The MT has ‫ה־בגֶ ד‬ ָ ‫ וְ ֵעֹרום ִּכ ָּס‬at the end of
the verse. The last word is not translated in the original Greek accord-
ing to Ziegler. This omission is supported by Papyrus 967, A, B and some
other witnesses, while the hexaplaric reading adds ιματιον. Here Papyrus
967 and B have the reading of the original Greek, as is to be expected in
instances such as this. The same omission of the Hebrew occurs in 18:7,
supported by Papyrus 967, B and quite a number of witnesses. In this
instance an asterisk is not indicated in any hexaplaric witness, but the
reading of this word is clearly hexaplaric. Similar examples occur in 18:18,
22, 28 and 32.9 In all these instances Papyrus 967 agrees with the original
Greek as reconstructed by Ziegler.

3.2. Readings Marked with an Obelus Occurring in Papyrus 967

There are only two examples of this in Ezekiel 18, namely in vv. 2 and 8.
In 18:2, Ziegler has Υἱὲ ἀνθρώπου at the beginning of the verse. This reading
does not occur in the MT and is marked with an obelus in hexaplaric wit-
nesses. The fact that it occurs in Papyrus 967, B and many other witnesses
demonstrates that this reading is pre-hexaplaric and part of the original
Greek. It can probably be regarded as a pre-hexaplaric addition to bring
this address to the prophet in line with 12:22.10
In v. 8, Papyrus 967 omits the καί at the beginning of the verse, agree-
ing with the MT. It is interesting to note that this omission occurs in a
number of witnesses, including the Coptic in the Bohairic version.11 This

8 Jahn, Der griechische Text.

9 Ibid., 131–132.
10 Cf. W. Zimmerli, Ezekiel 1 (Hermeneia; Philadelphia: Fortress Press, 1979), 369.
11  The others are V-449 Armp Clem. Or.XIV 224 Lo. Tht. Tert.III 473.
196 harry f. van rooy

raises the question whether the reading without καί must be regarded as
the original Greek, although there are many examples in Ezekiel 18 of the
Septuagint reading καί at places where the MT does not have “and”.

3.3. Readings Marked with an Asterisk Occurring in Papyrus 967

There is only one example of this in Ezekiel 18, namely in v. 24. This is,
however, a very interesting example. Ziegler has the following reading:
ἐν δὲ τῷ ἀποστρέψαι δίκαιον ἐκ τῆς δικαιοσύνης αὐτοῦ καὶ ποιήσῃ ἀδικίαν
κατὰ πάσας τὰς ἀνομίας, ἃς ἐποίησεν ὁ ἄνομος, πᾶσαι αἱ δικαιοσύναι αὐτοῦ, ἃς
ἐποίησεν, οὐ μὴ μνησθῶσιν· ἐν τῷ παραπτώματι αὐτοῦ, ᾧ παρέπεσε, καὶ ἐν ταῖς
ἁμαρτίαις αὐτοῦ, αἷς ἥμαρτεν, ἐν αὐταῖς ἀποθανεῖται.
When this is compared to the MT, two differences can be observed. Zie-
gler does not have anything for the Hebrew ‫ יַ ֲע ֶׂשה וָ ָחי‬after ὁ ἄνομος. Zim-
merli regards these two words as a gloss in the MT and wants to delete it.12
Allen argues, however, that it should be regarded as original, as linking up
with the same verb earlier in the verse.13 Jahn has noted that some wit-
nesses have added something to correct this deficiency.14 The hexaplaric
witnesses read και ποιηση και ζησεται, or something similar, equivalent to
the MT. The Lucianic witnesses have εαν ποιηση ου ζησεται, or something
similar. The majority of witnesses do not have this addition, making it
probable that the omission of the phrase of the MT was already part of
the original Greek. Papyrus 967 has an addition, but only εαν ποιηση. Jahn
thinks that these different readings support Ziegler’s reading of the origi-
nal Greek, and that the difference between the Lucianic reading and the
reading of Papyrus 967 could be attributed to Papyrus 967 having omitted
the part appearing in the Lucianic reading, or that the Lucianic reading
has added it.15 As these witnesses all attest to attempts to bring the Greek
in line with the Hebrew, an omission by Papyrus 967 is more likely. In
this instance B has the original Greek and Papyrus 967 has an attempt,
pre-hexaplaric, to bring the Greek in line with the Hebrew.
The second difference in v. 24, the plural noun ἁμαρτίαις in the Greek
for the Hebrew singular, is one of the instances where all the Greek wit-
nesses agree against the Hebrew, with only two exceptions (Sa and Con-

12 Zimmerli, Ezekiel 1, 373.

13 L. C. Allen, Ezekiel 1–19 (WBC 28; Dallas: Word), 266. Cf. also D. I. Block, The Book of
Ezekiel 1–24 (NICOT; Grand Rapids: Eerdmans), 581.
14 Jahn, Der griechische Text, 55 and 145.
15 Ibid., 145.
revisiting the original greek of ezekiel 18 197

stit). However, in v. 21 the same Hebrew consonantal text appears, but

vocalised by the Masoretes as a plural (‫אתו‬
ָ ֹ ‫) ַחּט‬, while in v. 24 it is vocal-
ised as a singular. In both instances the Greek reads a plural, probably
indicating that the Greek used the same Vorlage, but interpreted the two
nouns as being both plurals, while the Masoretes took the one as plural
and the other as singular.

4. Readings with B and Papyrus 967 in Agreement

Many examples of Papyrus 967 agreeing with B and the critical text occur,
but important are those examples where Papyrus 967 and B agree with a
reading not accepted by Ziegler (18:4, 9) or where Papyrus 967 agrees with
the critical text against B (18:10). A typical example of Papyrus 967 agree-
ing with the critical text and B occurs in 18:2, as discussed above, where
the Greek has “son of man” as a plus at the beginning.
In 18:2 another very interesting example occurs. This example is noted
in his apparatus by Jahn,16 but not discussed in detail. The MT reads as
‫ל־א ְד ַמת יִ ְׂש ָר ֵאל ֵלאמֹר ָאבֹות‬
ַ ‫ת־ה ָּמ ָׁשל ַהּזֶ ה ַע‬
ַ ‫מ ְׁש ִלים ֶא‬ ֹ ֽ ‫ה־ּל ֶכם ַא ֶּתם‬ ָ ‫ַמ‬
‫אכלּו ב ֶֹסר וְ ִׁשּנֵ י ַה ָּבנִ ים ִת ְק ֶהינָ ה‬
ְ ֹ‫י‬
All the witnesses of the Septuagint do not have an equivalent for the
Hebrew phrase ‫מ ְׁש ִלים ֶאת־‬ ֹ ֽ ‫ ַא ֶּתם‬. Zimmerli regards the reading of the
Septuagint as a smoothing of the text in agreement with 12:22.17 Block
thinks it may be due to haplography.18 These proposals take the Hebrew
as the original text, with the Septuagint shortening it. It would not be so
easy to explain the reading of the MT as an expansion, disagreeing with
The more important variant, referred to by Jahn, is related to the phrase
‫ל־א ְד ַמת יִ ְׂש ָר ֵאל‬
ַ ‫ ַע‬. Ziegler has the verse as follows: Υἱὲ ἀνθρώπου, τί ὑμῖν ἡ
παραβολὴ αὕτη ἐν τοῖς υἱοῖς Ισραηλ λέγοντες Οἱ πατέρες ἔφαγον ὄμφακα, καὶ
οἱ ὀδόντες τῶν τέκνων ἐγομφίασαν. For the relevant phrase he reads ἐν τοῖς
υἱοῖς Ισραηλ. However, quite a number of variants appear in the different
manuscripts of the Septuagint. The MT says that the proverb is about the
land of Israel, while the Septuagint says that the proverb is current among

16 Ibid., 51.
17 Zimmerli, Ezekiel 1, 369. Cf. also Allen, Ezekiel 1–19, 265.
18 Block, Ezekiel 1–24, 557.
198 harry f. van rooy

the children of Israel. As the proverb deals with a saying in which the
children blame their fathers for the evil they are suffering, the phrase of
the Septuagint fits easier into the context. The proverb is not about the
land of Israel as such. Some of the variants in manuscripts of the Septu-
agint agree with the MT and can be regarded as editorial work to bring
the Greek closer to the Hebrew. Ziegler lists the following in his appara-
tus: ἐν τοῖς υἱοῖς] επι της γης Syhmg L´’ (449txt) Tht. Constit. (om. της) Aug.
Iul.op.imp. 3,38 = 𝔐; pr. επι της γης A´ ’-410; εν τω Cyr.IV 420 VI949: ex 3 |
Ισρ.] pr. του L´-V-449txt Tht.; > 46. The last two variants are not important
for this discussion. The other two are, however, important. If one looks at
the witnesses listed, B and related texts as well as hexaplaric texts support
Ziegler’s reading. A reading agreeing with the MT appears in the Lucianic
group, agreeing to some extent with the reference to Israel in the next
verse. The group of A has a combination of the two readings: επι της γης
εν τοις υιοις Ισραηλ (“about the land among the children of Israel”). It is
interesting to note that the MT has the same Hebrew expression in 12:22,
there rendered with επι της γης του Ισραηλ. This makes it quite possible
that the original Greek was translated from a different Vorlage in 18:2.19
Papyrus 967 has a unique reading: εν τῳ Ισραηλ επι της γης. It has the
reference to the land at the end of the phrase and refers to Israel, not
the children of Israel. This reference to Israel appears in v. 3 as well, and
it is also found in Cyrillus Alexandrinus. Papyrus 967 has a longer read-
ing than the MT and Ziegler in this instance, and this goes against the
general trend of the Septuagint having the shorter reading. The original
Greek probably had the reading as accepted by Ziegler and the reading of
Papyrus 967 can then be regarded as a pre-hexaplaric attempt to bring the
reading of the Greek closer to the Hebrew.
In Ezekiel 18:19 an interesting set of variants related to word order
occurs. Marquis made a very interesting study of word order variants in
Ezekiel.20 According to him, the Greek of Ezekiel 1–39 agree for 90% with
the word order of the MT, comparable to Jeremiah, and not much lower
than the figure for Psalm 1–78, 1 Samuel and 2 Kings.21 He regards Ezek-

19 Block, ibid., 557 n. 15, thinks that a different Vorlage may be possible, but also har-
monisation with other texts, such as 12:24. Zimmerli, Ezekiel 1–24, 369–370 regards the
reading of the Septuagint as secondary, pointing to other instances where the Septuagint
introduces the children of Israel.
20 G. Marquis, “Word Order as a Criterion for the Evaluation of Translation Technique
in the LXX and the Evaluation of Word-order Variants as Exemplified in LXX-Ezekiel,”
Textus 13 (1986): 50–84.
21 Ibid., 64–65.
revisiting the original greek of ezekiel 18 199

iel as quite literal in the rendering of word order. He discusses a large

number of examples from Ezekiel, but unfortunately not the example to
be discussed here. Ziegler has the following reading in 18:19: οὐκ ἔλαβε
τὴν ἀδικίαν ὁ υἱὸς τοῦ πατρὸς. The MT reads: ‫לֹא־נָ ָׂשא ַה ֵּבן ַּב ֲעֹון‮ ָה ָאב‬. The
reading of Ziegler does not follow the Hebrew word order, by putting the
subject of the verb between the two words of the Hebrew genitive con-
struction, and does not have an equivalent for the Hebrew preposition.
Ziegler lists the different variants: τὴν ἀδ. ὁ υἱὸς τοῦ π. B Q-62] την αδ.
του π. ο υιος 147´ Las Arm Didasc.; αδ. υιος π. αυτου A; ο υιος την αδ. του
π. αυτου 88 (om. αυτου) L´’-403´ (om. την) Tht.Constit. = 𝔐. Papyrus 967
agrees exactly with 88, a manuscript from the Alexandrian group, while
many manuscripts from the Lucianic tradition has the same word order,
with the addition of the genitive pronoun to father. The reading with the
pronoun occurs in two Alexandrinian manuscripts as well (403 and 416).
The reading of Papyrus 967 may attest to the original Greek, or it could
be an early attempt to bring the Greek closer to the Hebrew. The fact that
the reading of Papyrus 967 does not occur in hexaplaric witnesses, and
is close to readings of Alexandrian and Lucianic witnesses, may support
the possibility of Papyrus 967 having the reading of the original Greek,
with the other witnesses reflecting either a different Vorlage or stylistic
In vv. 10–20 the words for father and son occur a number of times in
the MT. In some instances the words occur with the article and in others
without the article, and in some instances it occurs with the pronomi-
nal suffix for the third person singular. The following list presents all the
instances in the order they appear in the MT, with the verse in the first
column, the Hebrew in the second, an indication of the context in the
third and Ziegler’s rendering in the fourth.

Verse Hebrew Context Ziegler

10 ‫ֵּבן‬ ‫ן־ּפ ִריץ‬
ָ ‫ֹהוליד ֵּב‬
ִ ְ‫ו‬ καὶ ἐὰν γεννήσῃ υἱὸν λοιμόν
14:1 ‫ֵּבן‬ ‫ֹהוליד ֵּבן‬
ִ ‫וְ ִהּנֵ ה‬ ἐὰν δὲ γεννήσῃ υἱόν
14:2 ‫ָא ִביו‬ ‫ל־חּטֹאת ָא ִביו‬
ַ ‫ת־ּכ‬
ָ ‫וַ ּיַ ְרא ֶא‬ καὶ ἴδῃ πάσας τὰς ἁμαρτίας
τοῦ πατρὸς αὐτοῦ
17 ‫ָא ִביו‬ ‫הּוא ל֥ א יָמּות ַּב ֲעֹון ָא ִביו‬ οὐ τελευτήσει ἐν ἀδικίαις
πατρὸς αὐτοῦ
18 ‫ָא ִביו‬ ‫י־ע ַׁשק ע ֶֹׁשק‬
ָ ‫ָא ִביו ִּכ‬ ὁ δὲ πατὴρ αὐτοῦ ἐὰν
θλίψει θλίψῃ
200 harry f. van rooy

Table (cont.)
Verse Hebrew Context Ziegler
19:1 ‫ַה ֵּבן‬ ‫ַמ ֻּד ַע לֹא־נָ ָׂשא ַה ֵּבן ַּב ֲעֹון ָה ָאב‬ Τί ὅτι οὐκ ἔλαβε τὴν
ἀδικίαν ὁ υἱὸς τοῦ πατρὸς
19:2 ‫ָה ָאב‬ ‫ַמ ֻּד ַע לֹא־נָ ָׂשא ַה ֵּבן ַּב ֲעֹון ָה ָאב‬ Τί ὅτι οὐκ ἔλαβε τὴν
ἀδικίαν ὁ υἱὸς τοῦ πατρὸς
19:3 ‫ַה ֵּבן‬ ‫ּוצ ָד ָקה ָע ָׂשה‬
ְ ‫וְ ַה ֵּבן ִמ ְׁש ָּפט‬ ὅτι ὁ υἱὸς δικαιοσύνην καὶ
ἔλεος ἐποίησε
20:1 ‫ֵּבן‬ ‫ֵּבן לֹא־יִ ָּׂשא׀ ַּב ֲעֹון ָה ָאב‬ ὁ δὲ υἱὸς οὐ λήμψεται τὴν
ἀδικίαν τοῦ πατρὸς
20:2 ‫ָה ָאב‬ ‫ֵּבן לֹא־יִ ָּׂשא׀ ַּב ֲעֹון ָה ָאב‬ ὁ δὲ υἱὸς οὐ λήμψεται τὴν
ἀδικίαν τοῦ πατρὸς
20:3 ‫ָאב‬ ‫וְ ָאב לֹא יִ ָּׂשא ַּב ֲעֹון ַה ֵּבן‬
‬ οὐδὲ ὁ πατὴρ λήμψεται τὴν
ἀδικίαν τοῦ υἱοῦ
20:4 ‫ַה ֵּבן‬ ‬
‫וְ ָאב לֹא יִ ָּׂשא ַּב ֲעֹון ַה ֵּבן‬ οὐδὲ ὁ πατὴρ λήμψεται τὴν
ἀδικίαν τοῦ υἱοῦ

The words for father and son appear in three forms each in the Hebrew:
without the article (10, 14 and 20), with the article (19 and 20) and with
the pronominal suffix for the third person masculine singular (14, 17 and
18). In vv. 14 and 18, the Greek uses the article before the noun as well, but
not in v. 17. The form without the article is rendered without the article in
Ziegler’s text of vv. 10 (a son) and 14 (a son), but with the article in v. 20
(the son, the father). The forms with the article in Hebrew are rendered
with the article in the Greek as well (vv. 19 and 20). For the forms with
the suffix the Greek uses the genitive of the personal pronoun in vv. 14, 17
and 18 (his father).
The two nouns appear twelve times in this section, with only two exam-
ples in v. 20 not following the Hebrew closely in Ziegler’s text. What is also
interesting to note is the rendering with the genitive pronoun but without
the article in vv. 17 and 18. In three instances no variants are listed by Zie-
gler (v. 10, 14:1 and 18) and in these instances Papyrus 967 has no variants
either. In another four instances the reading of Papyrus 967 agrees with
the text of Ziegler (17, 19:3, 20:1 and 20:2).
In v. 17, a number of witnesses insert the article (62’ L’’ 91 Tht). In 19:3
only one witness omits the article. The article is omitted by A and a few
other witnesses, while a number add the genitive pronoun to 20:2. The
fact that B and Papyrus 967 agree, supports Ziegler’s choice for the origi-
nal Greek in these instances. Wevers notes that the addition of the geni-
revisiting the original greek of ezekiel 18 201

tive pronoun appears in both hexaplaric and Lucianic witnesses, although

they are not asterisked.22
This leaves five of the instances to be considered: 14:2, 19:1 and 2 and
20:3 and 4. In 14:2 Papyrus 967 is the only witness to omit the genitive pro-
noun, and this is probably an error in transmission, as there is no reason
to suspect a different Vorlage here.
The two examples in 19:1 and 19:2 have been touched upon in the dis-
cussion of the word order of this verse above. Here the rendering of the
words for father and son is the issue under discussion. The Hebrew has:
‫ ַמ ֻּד ַע לֹא־נָ ָׂשא ַה ֵּבן ַּב ֲעֹון ָה ָאב‬. Ziegler reads: Τί ὅτι οὐκ ἔλαβε τὴν ἀδικίαν
ὁ υἱὸς τοῦ πατρός. The Greek follows the Hebrew by having both words
with the article and without the genitive form of the pronoun. The four
possible variants are the omission of the two articles and the addition of
the genitive pronoun. In this instance it is argued above that Papyrus 967
probably has the word order of the original Greek. It can further be noted
that Papyrus 967 agrees with the Hebrew with regard to having the article
before both nouns and not adding the genitive pronoun. This strengthens
the possibility that Papyrus 967 has the original Greek for the whole sen-
tence under discussion.
As far as 20:2 and 3 are concerned, Ziegler’s text follows the Hebrew
closely, with articles before both nouns and without the genitive pronoun.
However, Papyrus 967 omits the article before the second father and adds
the genitive pronoun after the second son. The omission of the article
before father occurs in a number of witnesses (A´ O C´’-86´). These wit-
nesses represent the Alexandrinian text (A and 26), the main hexaplaric
witnesses, the catenae group as well as 86 and 710. The spread of witnesses
can be regarded as supporting the reading of Papyrus 967 as representing
the original Greek. There can be no other reason for these witnesses to
omit the article, whereas the addition of the article to bring the reading in
line with the Hebrew and with the previous sentence makes sense.
The addition of the genitive pronoun to son also has widespread support.
It does not occur in B and in some hexaplaric witnesses.23 The reading of

22 J. W. Wevers, “The L Text of Ezekiel,” in Studies in the Text Histories of Deuteronomy
and Ezekiel (MSU XXVI; J. W. Wevers and D. Fraenkel; Göttingen: Vandenhoeck & Rupre-
cht, 2003), 78.
23 Ibid.
202 harry f. van rooy

Papyrus 967 can again be regarded as the original Greek, on account of

the same arguments as the previous example.

5. Other Interesting Readings in Papyrus 967

Jahn is of the opinion that Papyrus 967 is based in the original Greek, but
that it was corrected to bring it closer to the Hebrew in many instances.24
These are pre-hexaplaric corrections. In these instances the possibility of
Papyrus 967 representing the original Greek must be kept in mind.
A very interesting set of variants occur in Ezekiel 18, related to the
rendering of the Hebrew ‫ּוצ ָד ָקה‬ ְ ‫ ִמ ְׁש ָּפט‬that occurs in the MT in vv. 5,
19, 21 and 27. These two words occur as a pair in Ezekiel 33:14, 16 and
19 and 45:9 as well. Jahn discusses the examples in Ezekiel 18 and 45,25
but not the examples in Ezekiel 33 (probably because that chapter of the
book is not in the Cologne manuscript of Papyrus 967). These two words
were not treated in the same way in all the places where they occur. In
all the eight instances in the Hebrew of Ezekiel, the two words are the
object of the verb ‫ ָע ָׂשה‬. The verb in the perfect precedes the two words
in Ezekiel 18:5 and 21, 33:14 and 19, and follows them in 18:19 and 33:16. In
18:27 it precedes the verb in an imperfect with waw consecutive. In 45:9
the verb is an imperative and follows the two words. The Hebrew verb
is translated into Greek with forms of the verb ποιέω: in 18:5 a participle,
in 18, 19 and 27, 33:16 an aorist active indicative, in 18:21 and 33:14 and 19
an aorist active subjunctive and in 45:9 an aorist active imperative. In all
these instances the Septuagint follows the Hebrew word order. There are
only a few variants with regard to the form of the verb. The most impor-
tant one is in 18:19, where B and some hexaplaric (O-407) witnesses read a
perfect (πεποιηκε[ν]) for ἐποίησε. In this instance Papyrus 967 agrees with
the majority of the witnesses against B. This indicates that Papyrus 967
does not always agree with B.
However, there are many differences with regard to the rendering of
the two Hebrew words. According to Jahn, the normal rendering in the
Greek would be κρίμα καὶ δικαιοσύνη.26 This rendering occurs in most wit-
nesses in 18:5 and 33:14 and in all witnesses in 18:27, 33:16 and 19 and 45:9.
It is accepted by Ziegler as the reading of the original Greek in 18:5. How-

24 Jahn, Der griechische Text, 156.

25 Ibid., 162.
26 Ibid.
revisiting the original greek of ezekiel 18 203

ever, the fact that κρίμα καί is omitted by B, Papyrus 967 and LaCS in 18:5
makes it more probable that this was the reading of the original Greek.
This is supported by Jahn as well.27 Zimmerli regards the reading of B as
a summarising of the Hebrew.28
In 18:19 Ziegler has δικαιοσύνην καὶ ἔλεος ἐποίησε. In this instance there
is a variety of variants. Papyrus 967 is the only witness with the reading
δικαιοσύνην ἐποίησε καὶ ἔλεον. Lucianic witnesses have κρίμα καὶ δικαιοσύνην
καὶ ἔλεον ἐποίησε. Hexaplaric witnesses have ἔλεος καὶ δικαιοσύνην πεποίηκεν
and B δικαιοσύνην καὶ ἔλεος πεποίηκεν. It is interesting to note that manu-
script 26, a witness of the A group, omits καὶ ἔλεος.
In 18:21 Ziegler has ποιήσῃ δικαιοσύνην καὶ ἔλεος, with the verb at the
beginning, but the same rendering for the two Hebrew words as in 18:19.
The Lucianic witnesses have the same three words for the two Hebrew
words as in 18:19. Papyrus 967 (and 62) has δικαιοσύνην καὶ ἔλεον. A and
related witnesses have τα δικαιωματα μου.
In 33:14 Papyrus 967 is the only witness to omit καὶ δικαιοσύνην, while all
the others have the reading regarded by Jahn as the standard rendering:
κρίμα καὶ δικαιοσύνην. This standard rendering, however, occurs in 33:16
and 19 in Papyrus 967, as in all the other witnesses. If one keeps the prob-
ability in mind that in 18:5 the original Greek probably only had one word,
and not two, it may be possible that Papyrus 967 has the original reading
in 33:14.
Jahn discusses many examples where Papyrus 967 has unique readings.
Most of them are not really important, but some warrant discussion. He
refers to two variants in 18:10.29 The text of Ziegler is as follows: καὶ ἐὰν
γεννήσῃ υἱὸν λοιμόν ἐκχέοντα αἷμα καὶ ποιοῦντα ἁμαρτήματα. The variants
are related to the words αἷμα and ἁμαρτήματα. The MT reads:
‫ן־ּפ ִריץ ׁש ֵֹפְך ָּדם וְ ָע ָׂשה ָאח ֵמ ַא ַחד ֵמ ֵא ֶּלה‬
ָ ‫ֹהוליד ֵּב‬
ִ ְ‫ו‬
The ‫ ָאח‬is probably the result of dittography. The rendering of the Sep-
tuagint can be regarded as an attempt to make sense of the difficult
Hebrew at the end. The words for “blood” and “sins” have variants in the
Greek. The Hebrew word for blood in the singular is usually rendered as
a singular in the Greek as well. The word for “sins” is in the plural in the
Septuagint, with no direct Hebrew equivalent. Ziegler has the singular

27 Ibid.
28 Zimmerli, Ezekiel 1, 370.
29 Jahn, Der griechische Text, 52.
204 harry f. van rooy

αἷμα in the first instance and the plural ἁμαρτήματα in the second. In his
apparatus he lists 130 with the plural of the first word and B, 88, 106 and
LaCS with the singular in the second instance. Papyrus 967 has the plural
in both instances, in the first instance in agreement with 130 and in the
second instance in agreement with many witnesses, excluding B, 88, 106
and LaCS. In this instance Papyrus 967 supports the plural in the second
instance. However, the plural in the first instance in Papyrus 967 requires
more attention.
The combination of ‫ שפך דם‬occurs eight times in Ezekiel, in 16:38,
18:10 (as discussed above), 22:3, 6, 9, 12, 27 and 23:45.
In 16:38 Ziegler does not include the translation of this phrase in
his reconstructed text, but the phrase is added by many witnesses (και
εκχεουσων αιμα O-407 L-311-V-46 Tht.Hi. = 𝔐; pr. και εκχεουσης αιμα rel).
In all instances where it is added, the singular of the noun αιμα is used.
Unfortunately, this section is missing in Papyrus 967.
In 22:3 Ziegler has the plural in his text, with Papyrus 967 and a number
of witnesses having the singular (967 o L´’−36 Las* Co Aeth Arm Tht. = 𝔐).
In 22:6, 12 and 27 all the witnesses have the singular. In 22:9 all the wit-
nesses do not have the same word order, but they all have the noun in
the singular. In 23:45 the noun is also in the singular, but in the genitive,
in Ziegler’s text, without the verb for shedding (blood). In agreement with
the Hebrew, that verb is added in a number of witnesses, with only 233
having the noun in the plural. Although the singular is the normal render-
ing of the Hebrew noun, there are a number of instances where the plural
occurs in the majority of the manuscripts in 22:3 and some witnesses in
18:10 and 23:45. The plural would be the normal rendering of the Hebrew
noun in the plural, one would expect.
The Hebrew noun in the plural occurs in the MT in 9:9, 16:6 (twice), 9,
18:13, 22:2 and 24:6 and 9. In 9:9 the Hebrew says that the land will be full
of blood (plural). This phrase does not occur in the Septuagint. In 18:13 it
is stated that the blood (plural in MT) of the unrighteous person will be
on his head. The Hebrew plural is rendered by the singular in the Septu-
agint. In 16:6 (twice) and 9 the noun is used in the plural in the MT with
reference to blood at the birth of a baby. These are all rendered in the
singular in the Septuagint. The expression “city of blood” (plural) occurs
in the MT in 22:2 and 24:6 and 9. The phrase does not appear in 24:9. It
is rendered by a plural by the Septuagint in 22:2 and 24:6. Even in the
case of the Hebrew plural noun, the majority of instances are rendered
by the singular, except when it occurs in the expression “city of blood”.
Taking all these instances into account, it would be easier for a plural in
revisiting the original greek of ezekiel 18 205

the Septuagint to be changed to a singular than vice versa. This makes the
possibility of Papyrus 967 having the original Greek in 18:10 quite strong,
and this should probably be the reading.

6. Conclusion

In the light of the discussion above, some proposals are made for different
readings for the original Greek as proposed by Ziegler.

Verse Ziegler Proposal

5 κρίμα καὶ δικαιοσύνην δικαιοσύνην
10 αἷμα αἷματα
19 οὐκ ἔλαβε τὴν ἀδικίαν ὁ υἱὸς τοῦ πατρὸς οὐκ ἔλαβε ὁ υἱὸς τὴν
ἀδικίαν τοῦ πατρὸς
20 οὐδὲ ὁ πατὴρ λήμψεται τὴν ἀδικίαν τοῦ υἱοῦ οὐδὲ πατὴρ λήμψεται τὴν
ἀδικίαν τοῦ υἱοῦ αὐτοῦ

There is also one instance in 18:8 where a different reading has some sup-
port for replacing Ziegler’s reading. There is a related example in 33:14, but
these two examples warrant further study in the light of a discussion of
related variants elsewhere in Ezekiel.

Verse Ziegler Proposal

8 καὶ τὸ ἀργύριον τὸ ἀργύριον
33:14 καὶ δικαιοσύνην Omit

What this discussion demonstrates is that one must be careful to equate

Papyrus 967 too easily with the original Greek. This is often done by
referring to the important variants in Ezekiel 7, the omission of 36:23c–38
and the different order of chapters 36–39 in Papyrus 967. However, the
detailed analysis of the variants in Papyrus 967 in Ezekiel 18 indicates
that there are many variants in this manuscript that cannot be equated
with the original Greek. There are instances where the papyrus has the
original Greek, often in agreement with B, but this cannot be taken as a
general rule.
Theology After the Crisis:
The Septuagint Version of Daniel 8–12

Martin Rösel

The Book of Daniel is one of the most fascinating writings in the Bible—
but also one of the most difficult. I would like to remind of some of the
highly disputed questions concerning this book. First of all there is the
notorious problem of the two language switches from Hebrew to Aramaic
and back in Dan 2:4 and after chapter 7. It is generally agreed that the
Aramaic parts on the whole belong to the older stratum of the literary
history of the book, stemming from late Persian or early Hellenistic times.1
The Hebrew chapters 8–12 are obviously younger. Here the situation has
changed. The issue is no longer life in the Babylonian or Persian diaspora,
but the events in Israel itself, when the temple cult in Jerusalem and the
religious life of the Israelites were threatened by a wave of Hellenisation,
fuelled by the Seleucid king Antiochus IV Epiphanes.2
The events in the early 2nd century B.C.E. under the reigns of Anti-
ochus III and IV, and the violent Maccabean response to the desecration
of the temple led to the formulation of an apocalyptic concept of history.
The different contributors to the Book of Daniel held the conviction that
the obvious decline of the worldly kingdoms attests to God’s plans to end
this period of history and to introduce a new, just and eternal government.3
Thus the figure of Daniel, a diaspora Jew full of “enlightenment, under-
standing, and wisdom like the wisdom of the gods” (Dan 5:11), was trans-
formed into an apocalyptic seer who saw strange visions and received
insights and understanding from heavenly messengers.

1 Some fundamentalist circles still hold the traditional view that the book comes from
an exilic author; cf. as a recent example W. Vogel, The Cultic Motif in the Book of Daniel
(New York: Peter Lang, 2010).
2 This short characterisation is based on inter alia: K. Koch et al., Das Buch Daniel (EdF
144; Darmstadt: Wissenschaftliche Buchgesellschaft, 1980); J. J. Collins, Daniel. A Commen­
tary on the Book of Daniel (Minneapolis: Fortress Press, 1993), 24–64.
3 For a sound theological interpretation of the visions, cf. A. C. Merrill Willis, Disso­
nance and the Drama of Divine Sovereignty in the Book of Daniel (OTS 520; London-New
York: Continuum, 2010). For a detailed concept of the literary development of the visions,
cf. R. G. Kratz, “The Visions of Daniel,” in The Book of Daniel. Composition and Reception
(Vol. 1; VTSup 83/1; eds. J. J. Collins and P. W. Flint; Leiden: Brill, 2001), 91–113 (although the
argument presented there is not convincing).
208 martin rösel

In chapter 11 the historical events from the last Persian kings until the
beginning of the final days is revealed to Daniel by an angel. This difficult
chapter is largely a vaticinium ex eventu, retelling elements of the history
of the Seleucid and Ptolemaic wars in the 3rd and 2nd centuries. But since
the verifiable data end in the year 165 B.C.E. and the reconsecration of the
temple in Jerusalem is not mentioned, scholars generally agree that the
Hebrew/Aramaic version of the book was finished in exactly this year.4
But the literary development of the book was not finished. It is here
that the Septuagint comes into play,5 because the Greek versions attest to
several independent additions to the book, namely the Susanna story to
introduce the figure of the wise and just Daniel; the prayers in chapter 3;6
and the stories of Bel and the dragon, which highlight the strictly mono-
theistic theology of the book. Moreover, the Greek version also offers a
glimpse into the earlier history of the Daniel traditions, because in chap-
ters 4–6 the text of the OG differs to such an extent from the parent MT
that the assumption of a largely deviating Semitic Vorlage is unavoidable.7
To add another field of discussion, the textual tradition of the Greek
translation is in itself difficult, because very early a second translation
known under the name of Theodotion was made, which was much closer
to the Hebrew text than the Old Greek.8 Since there are some citations
in the NT from this proto-Theodotionic version, it must come from pre-
Christian times.9 Eventually the Old Greek version was no longer used;

4 Collins, Daniel, 38, 388–390.

5 O. Munnich, “Texte massorétique et Septante dans le livre de Daniel,” in The Earli­
est Text of the Hebrew Bible. The Relationship between the Masoretic Text and the Hebrew
Base of the Septuagint reconsidered (SBLSCS 52; ed. A. Schenker; Leiden-Boston: Brill, 2003),
6 Cf. J. Joosten, “The Prayer of Azariah (DanLXX 3). Sources and Origin,” in Septuagint
and Reception. Essays prepared for the Association for the Study of the Septuagint in South
Africa (VTSup 127; ed. J. Cook; Leiden: Brill, 2009), 5–16, who opts for an original Greek
composition, but see also K. Koch, Daniel 1–4 (BKAT 22/1, Neukirchen-Vluyn: Neukirchener
Verlag, 2005), 314–375 with arguments in favour of an Aramaic original.
7 R. Albertz, Der Gott des Daniel. Untersuchungen zu Daniel 4–6 in der Septuaginta­
fassung sowie zu Komposition und Theologie des aramäischen Danielbuches (SBS 131; Stutt-
gart: Katholisches Bibelwerk, 1988), who holds that the Vorlage of LXXDan 4–6 is older
than the Aramaic text of Dan 4–6.; see also Collins, Daniel, 5f.; Koch, Daniel, 377–380,
387–401 (on Dan 4 only, but with a theological exposition of the Greek text.)
8 A. A. di Lella, “The Textual History of Septuaginta-Daniel and Theodotion-Daniel,”
in The Book of Daniel. Composition and Reception (Vol. 2; VTSup 83/2; eds. J. J. Collins and
P. W. Flint; Leiden: Brill, 2001), 586–607.
9 A. Schmitt, “Stammt der sogenannte ‘θ’-Text bei Daniel wirklich von Theodotion?”
MSU 9 (1966): 279–392; idem, “Die griechischen Danieltexte (‘θ’ und o’) und das Theodo-
tionproblem,” BZ 36 (1992): 1–29.
theology after the crisis 209

we have the famous saying from the prologue of Jerome’s translation of

Daniel that he does not know why this has happened, but that he can
confirm that the text of the Seventy differs from the veritas (hebraica).10
The OG therefore survived in one pre-hexaplaric Ms (Pap. 967) and late
hexaplaric witnesses only (88, SyH).11 Because of the weak textual attesta-
tion, it is often difficult to decide about the oldest stratum of the Greek
Scholarly research on the Old Greek of Daniel has mainly concentrated
on chapters 1–7 and the problem of the deviating Vorlage in 4–6.12 Sev-
eral authors have focused on the translation of the “son of man” in 7:13,
because of its implications on the New Testament and the question of
the messianic titles of Jesus.13 Only a few studies are concerned with the
Hebrew chapters 8–12 and their translation into Greek.14 This is regret-
table, because the translation must have been carried out shortly after
the events of the Maccabean struggle. Most scholars opt for a dating in
the second half of the 2nd century B.C.E.15 Comparable to Ben Sira, the

10 Danihelem prophetam iuxta Septuaginta interpretes Domini Salvatoris ecclesiae non

legunt, utentes Theodotionis editione, et hoc cur accederit nescio . . . hoc unum adfirmare pos­
sum, quod multum a veritate discordet et recto iudico repudiatus sit.
11 Cf. the new “Einleitung” by O. Munnich in the second edition of the Göttingen edi-
tion of Susanna, Daniel, and Bel et Draco (J. Ziegler and O. Munnich, Septuaginta: Vetus
Testamentum Graecum, XVI, 2: Susanna, Daniel, Bel et Draco [Göttingen: Vandenhoeck &
Ruprecht, 1999], 9–121).
12 B. Braasch, Die LXX-Übersetzung des Danielbuches—eine Orientierungshilfe für das
religiöse und politisch-gesellschaftliche Leben in der ptolemäischen Diaspora. Eine rezeptions­
geschichtliche Untersuchung von Dan 1–7 (Diss., Hamburg 2003); Albertz, Gott; T. J. Mea-
dowcraft; Aramaic Daniel and Greek Daniel. A Literary Comparison (JSOTSup 198, Sheffield:
Sheffield Academic Press, 1995), with a short outlook on Dan 8–12, 245–261; T. McLay, The
OG and Th Versions of Daniel (SBLSCS 43; Atlanta: Scholars Press, 1996) with chapters on
Dan 8:1–10 and 12:1–13.
13 Cf. the recent discussion between O. Hofius and K. Koch: O. Hofius, “Der Septua-
ginta-Text von Daniel 7,13–14. Erwägungen zu seiner Gestalt und seiner Aussage,” ZAW
117 (2005): 73–90; K. Koch, “Der ‘Menschensohn’ in Daniel,” ZAW 119 (2007): 369–385. See
also M. Müller, The Expression ‘Son of Man’ and the Development of Christology. A History
of Interpretation (London: Equinox, 2008), esp. 326–374.
14 Worth mentioning is the dissertation of S. Pace Jeansonne, The Old Greek Translation
of Daniel 7–12 (CBQMS 19, Washington: Catholic Biblical Association, 1988); A. McCrystall,
Studies in the Old Greek Translation of Daniel (Diss., Oxford, 1980).
15 The main argument for this dating is the citation of βδέλυγμα ἐρημώσεως from Dan
11:31 in 1 Macc 1:54: Collins, Daniel, 8f. Cf. Meadowcraft, Aramaic Daniel and Greek Daniel,
275: “the stories were very soon translated into Greek”. See also the summary of Braasch,
Die LXX-Übersetzung, 291: “wenig wahr­scheinlich, dass der LXX-Übersetzer lange nach 164
v.Chr. die Übersetzungs­tätigkeit aufnahm”. McCrystall (Studies, 386) dates between 161
and 152 B.C.E.
210 martin rösel

book was translated not more than about 50 years after its final composi-
tion in Hebrew.
This temporal proximity raises the question of whether some reflection
of the events around the Maccabean crisis can be traced in the Greek
version. From 165 B.C.E. onwards the course of history was quite different
from what Daniel had predicted. Therefore we can ask: are there indica-
tions that the translator has slightly adjusted the message of the book
so that it fits the new historical situation? Exploring this question is not
mere academic guesswork, because we know that the Book of Daniel did
not lose its theological and political importance. In Josephus and other
sources such as the Revelation of John, the fourth kingdom of Dan 7 was
identified with the Romans16 after the end of the world and the beginning
of the eternal kingdom did not come about during the time of one of the
Hellenistic rulers. This identification with the Romans seemed to be plau-
sible, since the depiction of the “fourth beast, terrifying and dreadful and
exceedingly strong . . . (with) great iron teeth” (Dan 7:7) seemed to fit the
overwhelming power of the Romans much better than a regional Syrian
kingdom. So the question of this paper is whether there are elements in
the Greek translation of the book of Daniel which kept the promises alive
rather than presenting a closed chapter of Judaean history.
The Old Greek of Daniel has been described as a dynamic translation,17
which shows several attempts to explain the difficult text of its Vorlage
and to state things more precisely than a mere literal translation could do.

16 See M. Tilly, “Die Rezeption des Danielbuches im hellenistischen Judentum,” in Die

Geschichte der Daniel-Auslegung in Judentum, Christentum und Islam. Studien zur Kommen­
tierung des Danielbuches in Literatur und Kunst (BZAW 271; eds. K. Bracht and D. S. du Toit;
Berlin-New York: de Gruyter, 2007), 31–54; K. Koch, “Spätisraelitisch-jüdische und urchrist-
liche Danielrezeption vor und nach der Zerstörung des zweiten Tempels,” in Rezeption
und Auslegung im Alten Testament und in seinem Umfeld. Ein Symposion aus Anlass des 60.
Geburtstags von Odil Hannes Steck (OBO 153; ed. R. G. Kratz; Fribourg-Göttingen: Univer-
sitätsverlag-Vandenhoeck & Ruprecht, 1997), 93–123.
17 Cf. R. T. McLay “Daniel,” in A New English Translation of the Septuagint and the
Other Greek Translations Traditionally Included under That Title (eds. A. Pietersma and
B. G. Wright; New York-Oxford: Oxford University Press, 2007), 992; cf. H. D. Neef, “Daniel /
Das Buch Daniel,” in Septuaginta Deutsch. Erläuterungen und Kommentare zum griechi­
schen Alten Testament. Band II: Psalmen bis Daniel (eds. M. Karrer and W. Kraus; Stuttgart:
Deutsche Bibelgesellschaft, 2011), 3016; Tilly, “Rezeption des Danielbuches,” 34, with a list
of characteristics. Pace Jeansonne (Old Greek Translation, 131f.) has come to the result that
the Old Greek is “reasonably accurate and faithful”, but calls for caution when it comes to
assessing a theological Tendenz in the work of the translator. Cf. also her earlier article:
idea., “The Stratigraphy of the Text of Daniel and the Question of Theological Tendenz in
the Old Greek,” BIOSCS 17 (1984): 15–35 (on Dan 7:13). In the course of this paper it will
become clear that Pace Jeansonne’s approach is too narrow.
theology after the crisis 211

This becomes particularly obvious when comparing the instances in ques-

tion with the later, more literal version of Theodotion. In the first chapters
of the book there are several hints that the translation must originate from
Egypt. The figures of Daniel and the king are reshaped so that the Greek
story of Daniel enhances the idea that competent Jews can be promoted
by the king to get important positions in the hierarchy of the royal court.18
This fits nicely the situation of the Jewish community in Alexandria.
But it is not completely clear whether the LXX version of Daniel can
be seen as a unit; there are several differences between chapters 1–7, on
the one hand, and 8–12, on the other. Therefore caution is advised so that
one does not draw conclusions too quickly from one part of the book to
the other.
In chapters 8–12 one can see several attempts to clarify difficult histori-
cal or geographical details. Thus in every instance in 11:5–40 the Hebrew
‫ ֶמ ֶלְך ַהּנֶ גֶ ב‬has been clarified by βασιλεὺς Αἰγύπτου (9 times) while the
“King of the North” ‫ ֶמ ֶלְך ַה ָּצפֹון‬was translated literally as βασιλεὺς βορρᾶ
(7 times in 11:6–40); Theodotion has the more literal ὁ βασιλεὺς τοῦ νότου
throughout.19 For later readers the unspecific “king of the north” in oppo-
sition to “Egypt” must have opened the possibility to relate this wicked
kingdom to the empire of the Romans. This interpretation could then be
supported by 11:30, because here the “ships of the Kittim” (‫) ִצּיִ ים ִּכ ִּתים‬
were correctly identified with the Ῥωμαῖοι.20 In 11:43 the ‫ֻל ִבים וְ ֻכ ִׁשים‬
were identified with the Λίβυες καὶ Αἰθίοπες; in 10:4 we can read Τίγρης
where MT has ‫( ִח ָּד ֶקל‬cf. Gen 2:24; Theodotion gives the transcription
Εδδεκελ) and in 10:20 the “prince of Greece” (‫ ) ַׂשר־יָ וָ ן‬has been translated
as στρατηγὸς Ἑλλήνων.

18 This is one of the important findings of the dissertation by Braasch, Die LXX-Über­
setzung, 288–302.
19 Cf. McCrystall, Studies, 321f; F. F. Bruce, “The Earliest Old Testament Interpretation,”
OtSt 17 (1972): 37–52 (41f.). Unfortunately Pace Jeansonne (Old Greek Translation, 29) does
not comment on this striking actualisation, but only mentions it when presenting the
article by Bruce.
20 Neef, “Daniel,” 3046: “Die LXX ‘entziffert’ den MT ‫‘ )ציים( כתים‬kittäische (= zyprische
= mittelmeerische’ (Schiffe)”; J. A. Montgomery, A Critical and Exegetical Commentary on
the Book of Daniel (ICC; Edinburgh: T & T Clark, 1927 [repr. 1979]), 454: “a correct historical
midrash”. F. F. Bruce (“Prophetic Interpretation in the Septuagint,” BIOSCS 12 (1979): 17–26
[23]) hints at an intertextual connection with Num 24:24 (‫ )וְ ִצים ִמּיַ ד ִּכ ִּתים‬and assumes
that the translator has seen this prophecy fulfilled in the events of 168 B.C.E. G. Jahn (Das
Buch Daniel nach der Septuaginta hergestellt [Leipzig: Pfeiffer, 1904]) has seen the Vorlage
of the Septuagint as the older and better version than the MT, therefore he has offered a
large number of retranslations and conjectures, e.g. p. 111 on 11:30: The Vorlage had ‫ ָר ִמים‬,
which the translator read as ‫ר ִֹמים‬, MT later introduced the reference to Num 24:24.
212 martin rösel

Moreover, the actions of the Romans were characterised by καὶ

ἐξώσουσιν αὐτὸν καὶ ἐμβριμήσονται αὐτω “they will expel him and rebuke
him” (NETS), so that according to Robert Hanhart, “one believes to hear
the growling of the hard and curt man Gaius Pompilius”.21 Hanhart and
also Arie van der Kooij have convincingly shown that in 11:29f. and 11:1422
detailed knowledge about the wars between the Seleucid kings of the
North and the Ptolemaic kings of the south lies behind the translation.23
So one can conclude from these and other observations that the translator
was willing to interpret his text and to convey its meaning to his readers.
Although especially Sharon Pace Jeansonne and to a smaller extent also
Tim McLay24 have argued that those translations can all be explained as
technical solutions for the problems of the Vorlage, I have the impression
that it is possible to go—in the footsteps of Robert Hanhart and Arie van
der Kooij—one step further.25 I will concentrate on verses concerning the
events around the Maccabean crisis and the future.
In chapter 8 Daniel has a vision of a ram and a goat. Horns are growing
out of the goat; one of them “grew exceedingly great toward the south,
toward the east, and toward the glorious land” (8:9, MT ‫ל־ה ֶּצ ִבי‬
ַ ‫ ;וְ ֶא‬LXX
different: καὶ ἐπὶ βορρᾶν).26 The Greek version of vv. 10–12 is extremely
difficult to explain,27 but it is obvious that the cosmic dimension of the
threat against the temple is amplified: the horn was raised unto the stars

21 R. Hanhart, “Die Übersetzungstechnik der Septuaginta als Interpretation. Daniel 11,29
und die Ägyptenzüge des Antiochus Epiphanes,” in idem, Studien zur Septuaginta und zum
hellenistischen Judentum (FAT 24; ed. R. G. Kratz; Tübingen: Mohr Siebeck, 1999), 85, using
a characterisation by Mommsen; my translation into English.
22 A. van der Kooij, “A Case of Reinterpretation in the Old Greek of Daniel 11,” in Tradi­
tion and Re-interpretation in Jewish and Early Christian Literature (FS J.-C. H. Lebram; StPB
36; ed. J. van Henten et al.; Leiden: Brill, 1986), 72–80.
23 This is also the conclusion reached by McCrystall, Studies, 384.
24 See McLay, OG and Th Versions, 191 n. 9: “we have given considerable reason to
doubt McCrystall’s view that the OG translator actually intentionally introduced signifi-
cant changes to MT for theological purposes”. Although I am also not convinced that the
Greek translation has its own chronological system, as McCrystall argues, he has collected
several instances which show that the translator has had his own theological ideas.
25 See also Meadowcraft, Aramaic Daniel and Greek Daniel, 250, who states a “mixture
of interpretation and misreading” as being characteristic of Dan 8–12 and gives some
examples for theological differences between MT and LXX.
26 Obviously the translator did not understand the concept of the “glorious / beautiful
land” (‫)צבי‬, because it is never translated adequately; the allusion to the north seems to
be taken from 8:4 (Meadowcraft, Aramaic Daniel and Greek Daniel, 251). Pace Jeansonne
(Old Greek Translation, 108) rejects the possibility of an intentional change and assumes a
“faulty Vorlage” or a misreading: ‫ ָצפֹון‬.
27 Cf. e.g. Montgomery, Daniel, 335–340; Pace Jeansonne, Old Greek Translation, 87;
both were assuming that glosses were moved into the text of the Old Greek.
theology after the crisis 213

of heaven (ἀστέρων for ‫)צבא‬, some of the stars will fall (or: he will be cast
down from the stars; cf. NETS), the mountains which existed from eternity
were overthrown. It is interesting to notice that, while most verbs are in
the Aorist, the statements about the deliverance of the captives (8:11 ἕως
ὁ ἀρχιστράτηγος ῥύσεται τὴν αἰχμαλωσίαν for ‫ר־ה ָּצ ָבא ִהגְ ִּדיל‬ ַ ‫ )וְ ַעד ַׂש‬and
the desolation of the sanctuary (καὶ τὸ ἅγιον ἐρημωθήσεται 8:11; καὶ τὰ ἅγια
ἐρημωθήσεται εἰς καταπάτημα 8:13) are in the future tense, although the
Hebrew text has no yiqtol forms. In v. 14 the idea is introduced that the
sanctuary has to be purified after 2 300 days (only here καθαρισθήσεται for
‫וְ נִ ְצ ַּדק‬, is used). Although this translation is in itself not very significant, it
has to be seen together with 11:35 and 12:6. Here the idea of a final purifica-
tion has been introduced over against the Hebrew text; in 11:35 the intel-
ligent (‫ילים‬ ִ ‫ ) ַה ַּמ ְׂש ִּכ‬are preparing for their purification; in 12:6 two angels
are talking about the consummation of the wonders and the purification.
One may suspect that the translator held the position that the temple
still has to be cleaned—although the Maccabeans have reconquered and
reconsecrated it.28
Another interesting characteristic of the Greek Daniel is introduced a
little later in 8:16. Here the anthropos angel says that “the vision is for this
ordinance” (NETS; ἐπὶ τὸ πρόσταγμα ἐκεῖνο ἡ ὅρασις, without counterpart
in MT). From this verse on πρόσταγμα is used several times for prophe-
cies and visions,29 e.g. in 9:2, when it is about the “ordinance of the Lord
to Jeremiah” or in 9:23 when an ordinance went out from the Lord to
answer Daniel’s prayer and Daniel is asked to understand this ordinance.
In 9:25 he is then told that he will “discover ordinances to respond, and
you will build Jerusalem as a city for the Lord” (καὶ εὑρήσεις προστάγματα
ἀποκριθῆναι καὶ οἰκοδομήσεις Ιερουσαλημ πόλιν κυρίῳ; the Hebrew text has
‫ד־מ ִ ׁ֣ש ַיח נָ ֔ ִגיד‬
ָ ‫רּוׁש ַלםִ ַע‬ ָ ְ‫) ִמן־מ ָֹצא ָד ָבר ְל ָה ִׁשיב וְ ִל ְבנֹות י‬30. The role of Daniel
has been significantly changed, for now he is responsible for the future
restoration of Jerusalem.31 According to 12:4 and 12:9, Daniel is ordered to

28 Similarly McCrystall, Studies, 386: “The Temple needs the true line of Priesthood to
return”. He assumes that the translator of Daniel is opposed to pro-Seleucid members of
the Jewish community.
29 In the Aramaic parts πρόσταγμα is only used for orders of the kings (for ‫ ִמ ָּלה‬cf. 2:15;
3:22 [28]; 5:7; 6:13).
30 It is not clear whether the translator had exactly this text in his Vorlage; the transla-
tion of ‫ ָמ ִ ׁ֣ש ַיח נָ גיד‬by κύριος is unusual (Montgomery, Daniel, 378); a misreading of ‫ עד‬to
‫ > עיר‬πόλις is not impossible, cf. Pace Jeansonne, Old Greek Translation, 126f.
31 Thus already A. Bludau, Die alexandrinische Uebersetzung des Buches Daniel und ihr
Verhaeltniss zum massorethischen Text (Freiburg: Herder, 1897), 120.
214 martin rösel

seal and hide these προστάγματα until the time of the end. Again, one can
ask whether the translator had a different expectation of the final events
to the authors of the Hebrew book of Daniel. It is evident that the transla-
tion πρόσταγμα for ‫ ָּד ָבר‬has changed the meaning considerably, because
Daniel not only sees visions and hears the word of God, but he is now
receiving orders and is involved in the events of the last days.
Chapter 9 is the meditation about the 70 years of Jeremiah 29:10 / 25:11.
In v. 2 the problem is stated: ἐγὼ Δανιηλ διενοήθην ἐν ταῖς βίβλοις τὸν ἀριθμὸν
τῶν ἡμερῶν ὅτε ἐγένετο πρόσταγμα κυρίου32 ἐπὶ Ιερεμιαν τὸν προφήτην
ἐγεῖραι εἰς ἀναπλήρωσιν ὀνειδισμοῦ Ιερουσαλημ ἑβδομήκοντα ἔτη: “I, Daniel,
pondered in the books the numbers of days when the ordinance of the
Lord came to the prophet Jeremiah to be awake (or: to rise) until the
fulfilment of the disgrace of Jerusalem—seventy years.” We have already
noticed the use of πρόσταγμα for “word”. It is interesting to see that the
Greek text has introduced the contents of this order to Jeremiah: to be
awake to see when the disgrace of Jerusalem is fulfilled. So this chapter is
not about the destruction of the temple, but about its defilement; the text
is adapted to reflect the situation in the 2nd century. In v. 4 Daniel begins
his prayer to the Lord God who is maintaining his covenant to those who
keep his ordinances. Here πρόσταγμα is used in a usual way for ‫ ִמ ְצוָ ה‬.
Obviously the prophecies to Jeremiah and the visions of Daniel are seen
on the same level as God’s commandments (cf. also v. 12: καὶ ἔστησεν ἡμῖν
τὰ προστάγματα αὐτοῦ).
The translation goes even one step further when it states in v. 10 that
the Lord has given his law to Moses (not in MT) and to us through his
servants the prophets (τῷ νόμῳ σου ᾧ ἔδωκας ἐνώπιον Μωσῆ καὶ ἡμῶν διὰ
τῶν παίδων σου τῶν προφητῶν). Now Moses and his law are of secondary
importance when compared to the prophets.33 Daniel, on the other hand,
who has insights into the mysteries of the prophets, will receive ordi-
nances and rebuild the city of the Lord. For me it is obvious that another
concept of prophecy lies behind these translations. This assumption fits
the observation that, according to 9:6, the prophets were sent not only to

32 This verse is of special interest because in MS 88 obviously a previous Hebrew tet-

ragram ‫יהוה‬, which served to replace original κύριος (now attested by Pap. 967), was no
longer understood by a later scribe, who corrected it to the similar looking τῇ γῇ; cf. the
apparatus of the Göttingen edition. For the problem of the replacement of the divine name
see M. Rösel, “The Reading and Translation of the Divine Name in the Masoretic Tradition
and the Greek Pentateuch,” JSOT 31 (2007): 411–428.
33 Unfortunately, Pace Jeansonne did not comment on this interesting addition, which
is clearly intentional; cf. Neef, “Daniel,” 3034: “ergänzt präzisierend”.
theology after the crisis 215

“all the people of the land” (‫ל־עם ָה ָא ֶרץ‬ַ ‫) ֶאל ָּכ‬, but to “every nation on
earth” (παντὶ ἔθνει ἐπὶ τῆς γῆς).34
Chapter 9 shows some more interesting deviations. A minor point is
that here in this chapter we have the only instance in the LXX where ‫ּתֹורה‬ ָ
is translated by διαθήκη: 9:13: κατὰ τὰ γεγραμμένα ἐν διαθήκῃ Μωσῆ “what
is written in the covenant of Moses”. This is clearly a contextual transla-
tion, because vv. 13f. are not dealing with prescriptions of the torah, but
with the evils that have come to Israel because they have not obeyed to
the law. Moreover, Israel has not only transgressed the law, but has com-
pletely forsaken it; ἐγκαταλείπω in 9:11 is a highly unusual translation for
‫( עבר‬only Judg 2:20).
Special attention is obviously devoted to the problem of understanding
and comprehension. In chapters 8–12 the verb διανοέομαι and its cognates
are used for several Hebrew words like ‫ ִּבינָ ה‬, ‫בין‬, ‫חשב‬, ‫ׂשכל‬. This unifor-
mity is not in line with the translator’s usual habit of stylistic variation.
Some examples: the justice of the Lord has to be understood (9:13); Daniel
has to understand God’s prostagmata (9:23+25); the angels give insight /
understanding (9:22). Moreover, Israel’s sin against God can be described
as lack of understanding, cf. 9:15: ἡμάρτομεν ἠγνοήκαμεν for ‫ ָח ָטאנּו ָר ָׁש ְענּו‬.
Therefore the 70 weeks which were disclosed to Israel (9:24) are meant to
bring the people to end the sin and to understand the vision (διανοηθῆναι
τὸ ὅραμα, without parallel in MT).35 It is obvious that emphasis is laid on
the intellectual side of the Israelite religion. This is in line with the Hebrew
version of Daniel, because here ‫ ַה ַּמ ְׂש ִּכ ִלים‬are obviously the group from
which the final form of the text comes.
Daniel’s request to understand the 70 years of Jeremiah is answered
by Gabriel in vv. 24–27. We have already seen that the expectation of the
Greek version differs considerably from the Hebrew text when Daniel is
presented as the one who will rebuild Jerusalem (v. 25). The second part

34 Cf. F. F. Bruce, “The Oldest Greek Version of Daniel,” OtSt 20 (1977): 22–40 (24). Pace
Jeansonne (Old Greek Translation, 114–116) has tried to show that this interpretation of
greater prophetic universalism is not valid, because there are more instances where ἔθνος
has been used to translate ‫ ָעם‬and referring to Israel. But the references she gives are
not convincing. In 11:14 she misses the “decidedly pro-Ptolemaic interpretation” (Collins,
Daniel, 380) of this verse; ἔθνος here refers to one of the peoples in the Hellenistic empire.
Moreover, the verse refers back to Am 9:11; cf. the explanation of McCrystall, Studies, 144f.,
333–335. In 11:33 the context is negative, which may lay behind the choice of ἔθνος for ‫ ָעם‬.
Pace Jeansonne’s own conclusion is hesitant (Old Greek Translation, 118): “the OG does not
necessarily give us any consistent indication of greater universalism”.
35 Cf. also 10:1: τὸ πλῆθος τὸ ἰσχυρὸν διανοηθήσεται τὸ πρόσταγμα; “the forceful multitude
will understand the decree” (NETS).
216 martin rösel

of v. 25 on the Vorlage of the LXX must have differed from what we have
today in the MT, because there is another sequence of the events revealed
to Daniel. Moreover, there are important differences in meaning. In v. 26
the angel is not speaking about 62 weeks of Jerusalem being rebuilt but
of 77 + 62 (weeks), after which an anointed (χρῖσμα) will be removed.36
After that a king of nations will demolish the city and the sanctuary with
its anointed. Then the end of this king will come with wrath (v. 26b) and
in wars. The covenant will prevail, will return and will be rebuilt (v. 27a).
This passage reads like a summary of the events of the Maccabean wars.
But the text goes on; the interval until the final righteousness is much
longer in the Greek text. At the end of times sacrifice and libation will
cease and the abomination of desolations will be given until the end of
time (v. 27b).
Admittedly, the textual situation of this verse is extremely difficult, as
one can see in the Göttingen edition, where Olivier Munnich has given
some parts of the text in brackets only. But two important elements are
obvious: the Greek version abandons the scheme of 70 weeks or year-
weeks. Either it has no longer understood the way the Hebrew text was
calculating,37 or it wanted to enhance the time span until the events of the
last days.38 Moreover, as the text reads now, the events of the Maccabean
crisis are only a kind of prelude. The end of days has been postponed,
so that it was possible to actualise the text in the 1st century C.E.; cf. the
explicit reference to the βδέλυγμα τῆς ἐρημώσεως in Mark 13:14.39
Again, this interpretation of the end of chapter 9 is supported by obser-
vations from chapters 11 and 12, the long final vision of the end of days. I
have already mentioned that in some verses the translator has shown his
knowledge of historical matters. When it comes to questions of the final
days, differences between Vorlage and translation become obvious. In

36 For the purpose of this paper it is not necessary to engage in the discussion about
the textual condition of 9:24–25 and possible reconstructions of a deviating Vorlage; cf. e.g.
Montgomery, Daniel, 401f.; McCrystall, Studies, 237–260, although his theory of a consistent
chronology in the Greek Daniel is not convincing. The same is true for the calculation of
Bruce (“Prophetic Interpretation,” 25) according to whom the translator switches to year
139 of the Seleucid chronology, beginning in 311 B.C.E.
37 See already Bludau, Die alexandrinische Uebersetzung, 116f. and Neef, “Daniel,” 3036f.
for possible misreadings of the Hebrew text to explain the Greek numbers.
38 Thus Bruce, “Prophetic Interpretation,” 25f., but he favours an explanation that the
translator wanted to bring the dating of the events in line with the historical facts about
the Maccabean crisis; cf. already Bludau, Die alexandrinische Uebersetzung, 117.
39 C. Evans, “Daniel in the New Testament,” in The Book of Daniel. Composition and
Reception (Vol. 2, VTSup 83/2; eds. J. J. Collins and P. W. Flint; Leiden: Brill, 2001), 490–527
theology after the crisis 217

11:33 the Hebrew text introduces the ‫ילי ָעם‬ ֵ ‫ ַמ ְׂש ִּכ‬, a group which, accord-
ing to 12:3, has to its credit “leading many to righteousness”. In the Hebrew
text the maskilim are to be distinguished from the “little help” (11:34), who
support the wise. It is widely accepted that ‫ ֵעזֶ ר ְמ ָעט‬refers to the Mac-
cabees, while the maskilim are the apocalyptic group behind the book of
The Greek text displays a different idea of the sociology of the groups
of resistance. According to 11:33 LXX, there is a group of thoughtful ones
of the nation (NETS) or of people who care about the nation (LXX.D: “die
Nation im Sinn haben”), which will stumble by the sword and will be soiled
by pillaging. This group will then gather a little strength (συνάξουσιν ἰσχὺν
βραχεῖαν)41 and many will be gathered to that group. Then, according to
v. 35, some of this group of understanding people42 will be able to clean
themselves (cf. v. 33, they were soiled), and then they will be chosen and
purified. While the Hebrew text was speaking of two groups, the Greek
has only one. The Maccabees are expunged from history.
The Hebrew text then goes on to retell the last days of the king. The
final battle between the king of the north and the king of the south will
take place in the beautiful land—obviously Israel (11:41). In the LXX this
verse is entirely missing;43 according to v. 40, the king of the north will
come to Egypt to plunder the whole land (v. 43).
Only after the war against Egypt will the king move to a place “between
the seas and the mountain of the will of the holy one” (v. 45; NETS) where
he finds his end.44 Again, the LXX deviates in an important detail from

40 Cf. Montgomery, Daniel, 458f.; Collins, Daniel, 386, doubts that ‫ ֵעזֶ ר ְמ ָעט‬refers to a
group but is only stating that the maskilim received only little help from any group.
41 According to Collins (Daniel, 367), the translator had ‫ עז‬not ‫ עזר‬in his Vorlage; Jahn
(Das Buch Daniel, 113) assumes a reading ‫חיל‬.
42 It is interesting to notice that—perhaps for stylistic reasons—in 11:35 συνιέντων is
used for ‫ילים‬ ִ ‫ ; ַה ַּמ ְׂש ִּכ‬while in v. 33 ἐννοούμενοι was used. One could ask why the translator
did not use a cognate of διανοέομαι, which is one of the key words in Dan 8–12. Instead
the used διανοηθήσονται to render the following verb ‫( יכׁשלו‬perhaps reading ‫יׂשכלו‬, Neef,
“Daniel,” 3047). Anyway, it is obvious that he does not see ‫ ַמ ְׂש ִּכיל‬as a technical term, as
it is attested in texts from Qumran.
43 Rahlfs has: καὶ ἐπελεύσεται εἰς τὴν χώραν μου from MS 88 and the Syh; Collins (Daniel,
368) thinks of haplography in 967 and an attempt at reconstruction in 88 and Syh; already
Montgomery (Daniel, 468) was assuming a homoioteleuton. Also McCrystall (Studies, 381)
argues that a mistake must have occurred; A. Geissen (Der Septuaginta-Text des Buches
Daniel. Kap. 5–12, zus. mit Susanna, Bel et Draco sowie Esther Kap. 1, 1a-2,15 [Bonn: Rudolf
Habelt Verlag, 1968], 261) also thinks of a deviating Vorlage.
44 McCrystall (Studies, 383–385) suspects that v. 45 refers to the battles between Judas
Makkabaios and Antiochus IV at Emmaus and (because of the plural θαλασσῶν) near Bet
Zur, but this is not convincing.
218 martin rösel

its Vorlage because of the expectation of a final battle between northern

and Ptolemaic troops in Egypt. It is difficult to say whether or not the
translator had specific historical events in mind. After the campaign of
Antiochus IV against Egypt in 168 B.C.E., Demetrios II was in 129 B.C.E. the
next Seleucid king who tried to attack Egypt, but he did not get further
than the border at Pelusium.45 One can only assume that the battles of
Antiochus formed the matrix for the events of the final days which the
translator was expecting.
Chapter 12 then reveals what will happen in the last days. Again the
Greek version draws a considerably different picture. The first, minor
deviation is that the whole people of Israel will be exalted; LXX has πᾶς ὁ
λαός, while MT has ‫ ַע ְּמָך‬only (12:1). In v. 2 the Hebrew text speaks about
two groups who then will rise from the dust: some to everlasting life, and
some to shame and everlasting contempt: ‫עֹולם וְ ֵא ֶּלה ַל ֲח ָרפֹות‬ ָ ‫ֵא ֶּלה ְל ַחּיֵ י‬
ָ ‫ ְל ִד ְראֹון‬. In the Greek we have three groups: “some of everlasting life,
but others to shame and others to everlasting dispersion” (οἱ μὲν εἰς ζωὴν
αἰώνιον οἱ δὲ εἰς ὀνειδισμόν οἱ δὲ εἰς διασπορὰν αἰώνιον, translation: cf. NETS).46
One of the three groups will have a positive afterlife, v. 3, “those who are
intelligent will light up like the luminaries of heaven (καὶ οἱ συνιέντες
φανοῦσιν ὡς φωστῆρες τοῦ οὐρανοῦ), which is in accordance with the MT.
In the second part of this verse the credit due to this group is described
differently: they did not lead many to righteousness (‫יקי ָה ַר ִּבים‬ ֵ ‫ּומ ְצ ִּד‬
ַ ), but
they did strengthen God’s words (καὶ οἱ κατισχύοντες τοὺς λόγους μου); the
educational aspect of the work of the maskilim is no longer present.47 This
is in line with earlier observations concerning the ‫ ; ַמ ְׂש ִּכיל‬cf. footnote 42.
Then Daniel is ordered to hide these prostagmata and “to seal the book
until the time of consummation, until the many rage violently and the
earth is filled with injustice” (ἕως ἂν ἀπομανῶσιν οἱ πολλοὶ καὶ πλησθῇ ἡ γῆ
ἀδικίας, cf. NETS). The Hebrew text is quite different: many shall be running
back and forth, and knowledge48 shall increase (for ‫יְ ׁש ְֹטטּו ַר ִּבים וְ ִת ְר ֶּבה‬
‫) ַה ָּד ַעת‬. As observed at the end of chapter 9, the Greek version has a differ-

45 G. Hölbl, Geschichte des Ptolemäerreiches. Politik, Ideologie und religiöse Kultur von
Alexander dem Grossen bis zur römischen Eroberung (Darmstadt: Wissenschaftliche Buch-
gesellschaft, 2004), 178.
46 Bruce (“Oldest Greek Version,” 26) asks whether the third group is a result of a vari-
ant translation of one clause only.
47 Even if the Vorlage of the translator was different (cf. Neef, “Daniel,” 3049f.: perhaps
the translator has read ‫ )?ומחזיקי דברי‬it is interesting to notice that ‫ דבר‬was not trans-
lated with the usual prostagma.
48 NRSV and Collins, Daniel, 369: evil, reading ‫ הרעה‬for ‫הדעת‬, based on LXX: ἀδικίας.
theology after the crisis 219

ent expectation concerning the end of days: the worst is yet to come. And
as could be seen earlier, there is another important link between chapters
9 and 12 in the Greek version, because also 12:6 speaks—without reference
in the Hebrew text—about the need for a final purification: (πότε οὖν . . . ὁ
καθαρισμὸς τούτων).
Finally, in v. 9, another small but important deviation points in the
same direction. In the Hebrew text the angel orders Daniel: “Go your way,
Daniel, for the words are to remain secret and sealed until the time of the
end” (‫ד־עת ֵקץ‬ֵ ‫) ַע‬. In the Greek version only ‫ ַעד‬is translated and con-
nected with the following verse: ἕως ἂν πειρασθῶσι καὶ ἁγιασθῶσι πολλοί καὶ
ἁμάρτωσιν οἱ ἁμαρτωλοι . . . Again, the LXX gives the impression that in the
last days there will be a final threat. The same idea seems to lie behind
the expansion in v. 13 that there are still days and hours until the end. But
at least for Daniel the end will not be too bad, because here, in the last
sentence of the book, he receives the promise that he will rest and then
rise upon his glory (καὶ ἀναστήσῃ ἐπὶ τὴν δόξαν σου; the MT has ‫) ְלג ָֹר ְלָך‬.49
A short conclusion might be in order. It has become obvious that the
first Greek version differs from the Hebrew text, especially in respect to
the question of the events of the final days. In the eyes of the translator
the events of the Maccabean crises have only been a kind of prefigura-
tion of what will come in the future.50 God has revealed these coming
threats by way of his ordinances to the prophets and to the apocalyptic
seer Daniel. Those of the Israelites who think about these revelations can
gain eternal life and will light up like stars.
Admittedly, the textual situation of the Greek Daniel is so problematic
that not all of the collected observations are equally convincing and it is
not always clear whether a variant comes from the Vorlage or from the
translator. But seen in the larger framework of the history of reception,
it is understandable why the expectations of the book of Daniel could
be applied to the Roman empire, although the predictions of this book
about the Hellenistic rulers failed. The Greek book of Daniel can easily
be inscribed in the history of the apocalyptic movement, but this is a task
for another paper.

49 Only here δόξα is used for ‫ּגֹורל‬

ָ ; this translation is not mentioned in T. Muraoka,
Index, s.v.
50 Others like Bludau (Die alexandrinische Uebersetzung, 126f.) or McCrystall (Studies,
385f.) concluded that the Old Greek of Daniel wanted to comment on the Maccabean
events; they did not see that the translation opens the book to a new interpretation of
coming events.
Part Three

Textual Criticism
The Value of the Septuagint for Textual Criticism of the
Hebrew Bible as Illustrated by the Oxford Hebrew Bible
Edition of 1 Kings

Jan Joosten

1. Introduction

From the late nineteen eighties onward, Septuagint studies have been
going from strength to strength. Major projects aimed at producing trans-
lations into modern languages, with more or less extensive notes—La
Bible d’Alexandrie in France, NETS in North America, Septuaginta Deutsch
in Germany—have attracted lots of young scholars. There are schools of
Septuagint studies in Finland and Spain. The international meetings of
the International Organization for Septuagint and Cognate Studies (IOSCS)
have grown from intimate gatherings of a handful of experts to large-scale
conferences with parallel sessions—70 papers were programmed at the
latest IOSCS meeting in 2010 in Helsinki. And the boom is not coming to
an end: the Hexapla Project, the projected seven-volume Handbuch zur
Septuaginta, and the recently launched Historical and Theological Lexicon
of the Septuagint witness to the continuing vitality of the field. In South
Africa, too, the study of the Greek version of the Bible has grown to the
point where the creation of a special organization, the Association for the
Study of the Septuagint in South Africa became a necessity.1
The growth of Septuagint studies has gone hand in hand with the rise
of a relatively new approach to the old Greek version. In earlier times,
starting with Louis Cappel in the seventeenth century, the Septuagint was
mostly consulted as a tool, not as a scientific object in its own right.2 The
single most important context in which the Septuagint was exploited was
textual criticism of the Hebrew Bible. Old Testament scholars used the
Greek version as a quarry for readings apt to improve the Hebrew text
where it was felt to be faulty. Emanuel Tov’s celebrated introduction to
the Septuagint, first published in 1981, is called The Text-Critical Use of

1 A first volume of essays prepared for this new Association has recently been pub-
lished: Septuagint and Reception (VTSup 127; ed. J. Cook; Leiden: Brill, 2009).
2 L. Cappel, Critica sacra (Paris, 1650).
224 jan joosten

the Septuagint in Biblical Research.3 Over the last thirty years, however,
the angle of approach has shifted. More and more scholars declare their
interest in the Septuagint as a writing, or a collection of writings, in its
own right.4 Research has focused on questions such as how the Septua-
gint translators interpreted their Vorlage, and how the Greek text itself
became a source of meaning and interpretation. The historical origins of
the version, its original and subsequent status among Jews and non-Jews,
the inner coherence and continuity of the books making up the Greek
“canon” (if that is the right word) have been at the center of scholarly
interest. Questions of textual history are treated from the point of view
of Greek text forms such as the kaige recension or the Lucianic/ proto-
Lucianic amalgam.5
It would be an exaggeration to say that the new approaches have super-
seded the earlier focus on textual criticism of the Hebrew Bible. There
have been some vivacious exchanges, of course: I remember a vigorous
discussion between Martin Rösel and Ron Hendel on a number of Septua-
gint readings in Genesis—explained by Rösel as reflective of the transla-
tors’ exegesis and claimed by Hendel as witnesses to a different Hebrew
source text.6 But mostly what can be said is that the new spirit in Sep-
tuagint research is complementary to the earlier text-critical enterprise:
the Septuagint should be understood on its own terms before it can be
compared to Hebrew text forms; “retroverting” the Greek into Hebrew can
only be done when the version has thoroughly been analyzed as a transla-
tion. And in fact, the traditional text-critical exploitation of the Septuagint
has never gone away but has continued to be practiced by knowledge-

3 E. Tov, The Text-Critical Use of the Septuagint in Biblical Research (Jerusalem Biblical
Studies 8; Jerusalem: Simor, 1997).
4 See, e.g., M. Harl, “La ‘Bible d’Alexandrie’ et les études sur la Septante. Réflexions sur
une première expérience,” VC 47 (1993): 313–340; W. Kraus, “Hebräische Wahrheit und
Griechische Übersetzung. Überlegungen zum übersetzungsprojekt Septuaginta-deutsch
(LXX.D),” TLZ 129 (2004): 989–1007.
5 A representative cross-section of Septuagint studies is provided by the proceedings
of IOSCS conferences edited in the SBLSCS series (editor M. K. H. Peters). The proceed-
ings of Septuaginta-deutsch conferences held in Wuppertal also document the state of the
art well: Die Septuaginta—Texte, Kontexte, Lebenswelten (WUNT 219; eds. M. Karrer and
W. Kraus; Tübingen: Mohr Siebeck, 2008); Die Septuaginta—Texte, Theologien, Einflüsse
(WUNT 252; eds. W. Kraus, M. Karrer and M. Meiser; Tübingen: Mohr Siebeck, 2010); a
third volume is forthcoming.
6 See M. Rösel, “The Text-Critical Value of Septuagint-Genesis,” BIOSCS 31 (1998): 62–70;
R. S. Hendel, “On the Text-Critical Value of Septuagint Genesis: A Reply to Rösel,” BIOSCS
32 (1999): 31–34.
the value of the septuagint for textual criticism 225

able scholars in different places.7 An important niche for the text-critical

use of the Septuagint in our own time is the Oxford Hebrew Bible proj-
ect launched in the early 2000s by the aforementioned Ron Hendel. The
Oxford Hebrew Bible will be a new edition of the Hebrew Bible featuring a
critical text, an apparatus and a textual commentary.8

2. The OHB of 1 Kings

Producing a critical text of the Hebrew Bible is a daunting enterprise, for

many different reasons.9 On the theoretical level, an important question is
which textual form one should aim to reconstruct.10 Biblical literature was
shaped by tradition, and books may have known different Hebrew edi-
tions through time. For some books, such as Jeremiah and Joshua, there is
textual evidence of such divergent editions. Should textual critics aim for
the earliest form attainable, or for some later form, and if so, which one?
The answer to these questions varies a bit from book to book.11 In practice,
what will be done for most books is nothing more adventuresome than to
emend the more egregious errors of the received MT.12
Another thorny issue is what one should do with the Masoretic vocaliza-
tion and accents: one might argue they have no place in a reconstruction of

7 See, e.g., B. Albrektson, Text, Translation, Theology. Selected Essays on the Hebrew
Bible (SOTSMS; Farnham: Ashgate, 2010); O. Munnich, “Retouches rédactionnelles au
texte proto-massorétique: l’apport des versions grecques de Daniel,” in Congress Volume
Ljubljana 2007 (VTSup 133; ed. A. Lemaire; Leiden: Brill, 2010), 71–86; M. Richelle, Le Tes-
tament d’Élisée. Texte Massorétique et Septante en 2 Rois 13.10–14.16 (Cahiers de la Revue
Biblique 76; Paris: Gabalda, 2010).
8 See S. White Crawford, J. Joosten, E. Ulrich, “Sample Editions of the Oxford Hebrew
Bible: Deuteronomy 32:1–9, 1 Kings 11:1–8, and Jeremiah 27:1–10 (34 G),” VT 58 (2008): 352–
9 R. Hendel, “The Oxford Hebrew Bible: Prologue to a New Critical Edition,” VT 58
(2008): 324–351; H. G. M. Williamson, “Do We Need A New Bible? Reflections on the Pro-
posed Oxford Hebrew Bible,” Bib 90 (2009): 153–175; E. Tigchelaar. “Editing the Hebrew
Bible: An Overview of Some Problems,” in Editing the Bible (ed. J. Kloppenborg and J. New-
man; Toronto: University of Toronto Press, forthcoming).
10 See B. K. Waltke, “Aims of OT Textual Criticism,” WTJ 51 (1989): 93–108.
11  As can be seen in the sample published in the article referred to in n. 8, Jeremiah
will be edited in two columns, one representing the MT, the other the Vorlage of the Sep-
tuagint (see also the remarks on 1 Kgs 2:1 below).
12 Emanuel Tov has criticized the use of the term “emendation” in reference to the
adoption of preferred readings from other witnesses than the MT: the goal of textual criti-
cism is to restore the original text, not to emend the MT, see E. Tov, Textual Criticism of
the Hebrew Bible (Minneapolis / Assen: Fortress Press / Van Gorcum, 1992), 351; similarly
Hendel, “Oxford Hebrew Bible,” 331. Theoretically Tov is right, but it is also necessary to
recognize the limits of what is possible.
226 jan joosten

the “oldest text attainable”, because at the time when this text would have
circulated, no vowel pointing existed;13 but one might argue, too, that the
pointing encodes information going back to earlier times: when the text
was written in consonantal form only, people would still know how to
pronounce it.14 In the meantime, OHB policy is to keep the vocalization
and accents wherever the MT is adopted, and to vocalize—but not to
accentuate—emended passages.15
On the practical level, the main problem is the dearth—or rather, the
patchy attestation—of Hebrew textual material. A few passages in the
Pentateuch are attested in the MT, the Samaritan Pentateuch, and one
or more texts from Qumran. In Isaiah, or in parts of the books of Samuel,
the Qumran material is sufficiently extensive to start working. In many
other books, however, the only Hebrew witness available is the MT. The
MT is not entirely without variants, of course, but it is surprisingly unified
nonetheless. There would be scant justification for a critical edition of the
Hebrew Bible based solely on medieval Hebrew manuscripts. In passages
where no non-Masoretic Hebrew material is available, the lion’s share of
the text-critical enterprise will rest on a comparison of the MT with the
Greek version, and on conjecture. Without putting too fine a point on it,
this is the situation in regard to the first book of Kings.
Initially, the OHB of 1 Kings was attributed to Steve McKenzie. McKen-
zie did produce a critical edition of 1 Kgs 1:1–7—including an apparatus
but no textual commentary. But soon afterwards he abandoned the proj-
ect. In 2004, the book was reallocated to the present author. The proj-
ect of editing 1 Kings started in earnest in 2009, when funding became
available allowing close collaboration between Jan Joosten, of Strasbourg,
and Jean Koulagna, professor of OT at the Lutheran School of Theology
in Meiganga, Cameroon.16 Since 2009, two chapters of 1 Kings have been
analyzed and provisional versions of the critical text, the apparatus, and
the textual commentary have been created.
Although the critical text is based on a confrontation of all available
direct witnesses, and although the apparatus includes extensive references

13 See Williamson, “Do We Need A New Bible?”

14 For the antiquity of the oral tradition issuing in the Masoretic vocalization see, e.g.,
F. Buhl, Canon and Text of the Old Testament (trans. J. MacPherson; Edinburgh: T & T Clark,
1892), 236–239; J. Barr, Comparative Philology and the Text of the Old Testament. With Addi-
tions and Corrections (Winona Lake: Eisenbrauns, 1987), 194–207.
15 This will be done on the “copy-text principle,” see Hendel, “Oxford Hebrew Bible,”
16 The funding comes mainly from the Institut Universitaire de France.
the value of the septuagint for textual criticism 227

to readings in the Peshitta, the Targums, the Vulgate, and Flavius Josephus’
Antiquities, really meaningful divergences from the received Hebrew text
are almost wholly confined to the Greek versions. In practically all places
where a non-Masoretic reading has been adopted, our authority is the
Septuagint. While the aforementioned Aramaic and Latin versions hew
closely to the Masoretic tradition, diverging from it only rarely and usu-
ally in small details, the Greek tradition clearly reflects a different state of
the Hebrew text of 1 Kings.17 Unfortunately, the Greek textual tradition is
itself notoriously complicated in the books of Kingdoms/Samuel-Kings.

3. The Greek Text of 3 Kingdoms

3 Kingdoms is one of the books of the Septuagint for which the Göttingen
edition is still in preparation.18 Perusal of the old Brooke-McLean edition,
known also as the “Cambridge edition”, will quickly show the polymor-
phous nature of manuscript evidence.19 Successive stages in the history of
investigation have permitted, nevertheless, to perceive some measure of
order in the chaos. The most important discoveries were made by Thack-
eray, at the beginning of the twentieth century, and by Barthélemy in the
nineteen sixties.
On the basis of his grammatical analyses, Henry Thackeray realized that
the bulk of Greek manuscripts of 1–4 Kingdoms combine two distinct ver-
sions: a literal version in 1 Kgds 1–2 Kgdms 11:2 and in 3 Kgdms 2:12–21:43,
and an ultra-literal one, closely resembling Theodotion, in 2 Kgdms 11–3
Kgdms 2:11 and in 3 Kgdms 22–4 Kgdms 25.20 He correctly judged the less
literal parts to be older than the more literal ones:
Earlier portions: 1 Sam 1–2 Sam 11:1 1 Kgdms 1–2 Kgdms 11:1
1 Kgs 2:12–21:29 3 Kgdms 2:12–21:43
Later portions: 2 Sam 11:2–1 Kgs 2:11 2 Kgdms 11:2–3 Kgds 2:11
1 Kgs 22–2 Kgs 25 3 Kgdms 22–4 Kgdms 25

17 The textual nature of Josephus’ paraphrase in the Antiquities is a more complicated

matter. Josephus appears to know both the MT and the Old Greek, mixing elements as he
sees fit and adding others on his own account.
18 The Göttingen edition is being prepared in Madrid by Julio C. Trebolle Barrera and
his team, notably Pablo Torijano Morales and Andrés Piquer Otero.
19 The Old Testament in Greek. Vol. II, Part II, I and II Kings (eds. A. E. Brooke, N. McLean
and H. St. J. Thackeray; Cambridge: Cambridge University Press, 1930). 
20 See H. St. J. Thackeray, A Grammar of the Old Testament in Greek (Cambridge: Cam-
bridge University Press, 1909), 10–11.
228 jan joosten

Thackeray explained the presence of earlier and later portions in Kingdoms

as reflecting a translation in two stages: at first, only parts of Samuel–Kings
were translated, leaving aside the “more disastrous” bits (the gradual
downfall of David in 2 Sam and 1 Kgs 1–2, and the end of the monarchy in
2 Kgs).21 When, in a second stage, it was decided to add the missing parts,
this was done according to an ultra-literal technique that had become
fashionable in the meantime. Later research would disprove this expla-
nation. Moreover the exact demarcation of “earlier” and “later” portions
would be tweaked a bit.22 The identification and relative dating of the
different portions, however, has become an inalterable component of Sep-
tuagint research.
Half a century later, Dominique Barthélemy was able to show that
Thackeray’s “later portions” did not represent a translation, but rather the
recension of an older version.23 The basic hypothesis was suggested to
Barthélemy by the Greek Minor Prophets scroll discovered in Nahal Hever
a few years before. This scroll turned out to contain a recension of the
Septuagint text of the Dodecapropheton following the same principles of
literal translation as could be observed in parts of Kingdoms.
An important element in Barthélemy’s demonstration was the obser-
vation that an “Old Greek” version of Thackeray’s later portions was
preserved in the so-called Antiochene or Lucianic text, represented in
Kingdoms by five minuscules.24 While the majority text of 2 Kgdms 11:2–3
Kgdms 2:11 shows much similarity to Theodotion—and to the Minor
Prophets scroll of Nahal Hever—, the translation technique of the Anti-
ochene text coheres with that of Thackeray’s “earlier portions”. Thus it
turns out that the books of Samuel-Kings were from the start translated
completely. Later, however, the text was revised according to the prin-
ciples of a Palestinian school aiming at greater conformity between the
Greek version and the emerging MT. On the basis of one striking feature,
Barthélemy proposed to call this revision the kaige recension. Like Thack-
eray’s distinction of different portions, Barthélemy’s explanation of the
“later portions” as recensional has become a foundation stone of Septuagint

21 Ibid., 11.
22 See J. D. Shenkel, Chronology and Recensional Development in the Greek Text of Kings
(HSM 1; Cambridge: Harvard, 1968), 117–120.
23 D. Barthélemy, Les Devanciers d’Aquila (VTSup 10; Leiden: Brill, 1963), 47.
24 Referred to by the sigla b o c2 e2 in the Cambridge edition and by the numbers 19, 82,
93, 108 and 127 in the Göttingen edition (b = 19 + 108, the two manuscripts tending to go
together). For a description of the manuscripts, see N. Fernández Marcos and J. R. Busto
Saiz, El Texto Antioqueno de la Biblia Griega. I 1–2 Samuel (Madrid: CSIC, 1989), XV–XXIII.
the value of the septuagint for textual criticism 229

research. Thackeray’s “later portions” are nowadays usually referred to as

“kaige sections”.
In Barthélemy’s view, the mixture of Old Greek and kaige in the major-
ity text of Kingdoms reflected the partial nature of the revision in a first
stage. The manuscripts used by Origen for the Septuagint column in his
Hexapla were mixed because the kaige recension had been applied only
to 2 Kgdms 11:2–3 Kgdms 2:11 and 3 Kgdms 22–4 Kgdms 25.25 This remains
a possibility. It is difficult to explain, however, why a reviser would tar-
get only those two blocks, leaving alone the first 41 chapters as well as
some 20 chapters between 3 Kgdms 2:12 and 21:43.26 Later scholars have
tended, therefore, to ascribe the peculiar mixture of Old Greek and kaige
in the main manuscript tradition of 1–4 Kgdms to chance. In this view, the
recension was applied to the whole work. Later however, a scribe inadver-
tently copied parts of the work from a scroll containing the Old Greek and
parts from another scroll containing the kaige recension.27 By accident,
this mixed manuscript became the root of the later textual tradition—
excepting the Antiochene tradition, which preserved the original transla-
tion throughout.
In the post-Barthélemy period, two further adjustments were made.
Firstly, it was pointed out by several scholars that the Antiochene text
cannot simply be qualified as a representative of the Old Greek, not even
in the kaige sections (Thackeray’s “later portions”, where the main tradi-
tion reflects the kaige recension). Although the Antiochene text remained
largely untouched by the kaige recension, it underwent many other
changes, as can be proven easily from the non-kaige section (Thackeray’s
“earlier portions,” where the main tradition reflects the Old Greek). It is
more correct, therefore, to say that in the kaige sections the Old Greek can
be partially retrieved from the Lucianic tradition.
Secondly, detailed study has shown that all available manuscripts of 1–4
Kgdms are contaminated to some extent by different textual traditions.
While older studies habitually oppose the “B text” of Codex Vaticanus

25 Barthélemy, Devanciers, 140–141.

26 Barthélemy tried to justify the partial nature of the revision for the section 2 Kgdms–
3 Kgdms 2:11, but his explanations are not convincing, see ibid., 141–143.
27 The first scholar who proposed this hypothesis seems to be Emanuel Tov, see E. Tov,
“The Methodology of Textual Criticism in Jewish Greek Scriptures, with Special Attention
to the Problems in Samuel–Kings: The State of the Question: Problems and Proposed Solu-
tions,” in 1972 Proceedings: International Organization for Septuagint and Cognate Studies
and the SBL Pseudepigrapha Seminar (SCS 2; ed. R. Kraft; SBL, 1972), 3–15, reprinted in
E. Tov, The Greek and Hebrew Bible (Leiden: Brill, 1999), 489–499.
230 jan joosten

to the “Antiochene text” of minuscules 19, 82, 93, 108 and 127, and the
“Hexaplaric text” of Codex Alexandrinus, recent investigations indicate
that this terminology is at best approximative:

– In the non-kaige sections, Codex Vaticanus is expected to reflect the

Old Greek, but it also contains a number of recensional readings.
– Codex Alexandrinus and its congeners usually give a secondary text, but
here and there they alone appear to have preserved the most ancient
version available.
– The text of minuscules 19, 82, 93, 108, 127 is supposed to be imperme-
able to the kaige recension throughout 1–4 Kgdms, but it is not: several
kaige readings are found in it.

The textual situation of the Greek tradition in Kingdoms is very compli-

cated and the history of investigation has seen the emergence of many
hypotheses that later turned out to be unhelpful.28 The ideas of Thackeray
and Barthélemy, widely accepted as they are (although with slight modi-
fications), provide no easy protocol for deciding which Greek reading if
any reflects the Old Greek. They do however afford a certain perspective
when it comes to comparing the Greek data with the MT and other tex-
tual witnesses.

4. Case studies illustrating the Use of the LXX in Textual

Criticism of the Hebrew Text

As was stated above, the Greek text of 1 Kgs 1–2, in both the Old Greek and
the kaige recension, is rather faithful to its Hebrew Vorlage (or Vorlagen).
Nevertheless, there are a surprising number of divergences between the
received Hebrew text and the Greek tradition, particularly the Antiochene
text. No doubt some of these divergences go back to the Greek translator.
Some of them, however, appear to reflect a divergent Hebrew text. Among

28 To give only one notable example, Frank Cross has developed the idea of a “proto-
Lucianic” recension, postulating that the Lucianic/Antiochene tradition contains two dis-
tinct recensional strata: the “Lucianic” revision of the 4th Century, and an earlier revision
on the basis of a Hebrew manuscript textually akin to the 4QSama fragments. This theory
has proved almost impossible to root out, even although only Cross’ students ever really
supported it, against the unanimous agnosticism of other competent scholars. As has been
shown by Tuukka Kauhanen of Helsinki University in his recent doctoral dissertation, this
theory rests on no certain foundation. See T. Kauhanen, The Proto-Lucianic Problem in
1 Samuel (Ph.D. diss., Helsinki, 2011 [to be published by Vandenhoeck & Ruprecht, Göt-
the value of the septuagint for textual criticism 231

these are several variants that must be preferred to the reading of the
MT on internal grounds. We will explore two passages where the Greek
tradition may be held to offer a superior reading and one where, although
probably reflecting a Hebrew text, it is clearly secondary.

4.1. 1 Kgs 1:8

֣ ֶ ‫בּוֹרים ֲא‬
֖ ִ ִ‫יא וְ ִשׁ ְמ ִ ֣עי וְ ֵר ֔ ִעי וְ ַהגּ‬
֙ ‫וּבנָ יָ֙ הוּ ֶבן־יְ הוֹיָ ָ ֜דע וְ נָ ָ ֤תן ַהנָּ ִב‬
ְ ‫וְ ָצ ֣דוֹק ַהכּ ֵֺהן‬ BHS
‫ם־אד ֺנִ ָיּֽהוּ‬
ֲ ‫ְל ָדִו֑ד ל֥ ֺא ָהי֖ וּ ִע‬
“But Zadok the priest, and Benaiah the son of Jehoiada, and Nathan the prophet,
and Shimi and Rei, and the Heroes of David, were not with Adonijah.”

֣ ֶ ‫בּוֹרים ֲא‬
ִ ִ‫יא וְ ַשׁ ָמּא וְ ֵר ָעיו ַהגּ‬ ְ ‫ וְ ָצ ֣דוֹק ַהכּ ֵֺהן‬OHB
֙ ‫וּבנָ יָ֙ הוּ ֶבן־יְ הוֹיָ ָ ֜דע וְ נָ ָ ֤תן ַהנָּ ִב‬
‫ם־אד ֺנִ ָיּֽהוּ‬
ֲ ‫ְל ָדִו֑ד ל֥ ֺא ָהי֖ וּ ִע‬
“But Zadok the priest, and Benaiah the son of Jehoiada, and Nathan the prophet,
and Shammah and his comrades, the Heroes of David, were not with Adonijah.”

OHB Apparatus
‫ *ושמא ורעיו הגבורים‬GL (Σαμαιας καὶ οἱ ἑταῖροι αὐτοῦ οἱ ὄντες δυνατοί) ] ‫ושמעי ורעי‬
‫ והגבורים‬M
The lemma is taken from the eclectic text as edited in the OHB. The asterisk
shows that the Hebrew text is not attested as such, but retroverted from the Luci-
anic Greek (GL), quoted in parentheses. The non-preferred variant is quoted from
the MT (M). No other textual witnesses are indicated because they agree with
MT in the expected way: this is true for the Syriac, the Targum, and the Vulgate,
but also for the non-Lucianic Greek (καὶ Σεμεϊ καὶ Ρηι καὶ οἱ δυνατοί) which here
reflects the kaige recension.29

In MT, the proper names “Shimei and Rei” fall completely from the blue
sky. Occurring as they do between the well-known Zadok, Benaiah and
Nathan on the one hand and the famous group of “David’s heroes” (cf.
2 Sam 23:8) on the other, they are hard to interpret. The MT is hardly
correct, even although it is confirmed by the B text of the LXX (heavily
influenced by the kaige recension in this passage), by the Peshitta, the
Targum and the Vulgate.
The Antiochene text has several important variants. Where MT reads
the proper noun Rei, the Antiochene text has a common noun with a pos-
sessive pronoun: “his comrades.” The Greek reflects a different reading

29 Flavius Josephus reads differently from both the MT and the Antiochene Text: “And
Shimei, David’s friend, and all the bravest (warriors)” (Ant. 7.346).
232 jan joosten

of the consonants attested in MT: the waw was not attached to the word
‫ הגבורים‬as in MT, but to the letters ‫ ורעי‬to form the word ‫( ורעיו‬cf. rēϲāyw,
“his friends” Job 32:3).30 In addition, the first proper name seems to have
been read differently: in the LXX, Σαμαια(ς) corresponds most often to
‫ שמעיה‬and is found also for ‫( ישמעיה‬1 Chron 12:4) ‫( שמע‬Neh 8:4) and
‫( שמא‬2 Sam 23:11)—but nowhere, excepting this passage, for ‫שמעי‬.
Without being overly transparent, the Antiochene text makes good
sense in the context. The “heroes” (‫הגבורים‬, οἱ δυνατοί) of David are a
well-defined group of warriors accompanying David throughout his mili-
tary career. In 2 Sam 23:8–39, they are listed by name and divided into
two categories: the three, and the thirty.31 The third among “the three” is
‫—שמא‬Σαμαια(ς) in Greek (v. 11).32 Apparently, then, Shammah is named
in our verse as the leader of the famous “thirty” heroes of David at the
time of Adonijah’s coup. Why the third name in the list of 2 Sam 23 was
singled out and not the first (or the second) is impossible to say. The text
suggests that, between the glorious period described in 2 Sam 23:8–39
and the end of David’s life, something happened that made Shammah
the principal figure instead of Ishbaal.
In weighing the two alternative versions, no importance should be given
to the rule lectio difficilior potior. If MT is corrupt, as many critics admit,
its difficulty gives no indication of originality. The decision between the
two versions hinges exclusively on the possibility to explain one of them
as secondary to the other. Two scenarios seem possible:

– The Antiochene version is a facilitating interpretation of the corrupted

text of MT: the translator, or the Hebrew scribe producing the imme-
diate Vorlage of the Antiochene text, creatively rearranged the conso-
nants of proto-MT in order to arrive at a meaningful text. Casting about
for a likely identification of the proper nouns in his Vorlage, he hit upon
Shammah in the list of heroes in 2 Sam 23 and decided this was good
enough for his purposes.

30 Cf. BHS (where however there is no indication as to whether or not the Antiochene
reading is preferred).
31 The text of the list is much deteriorated in the MT and needs to be partly restored
on the basis of parallels in Chronicles and of the Greek versions. See S. R. Driver, Notes on
the Hebrew Text and the Topography of the Books of Samuel (2nd ed.; Oxford: Clarendon,
1913), 362–363.
32 In the Antiochene text of 2 Sam 23:11 it is clear that Samaias comes after Iesbaal
and Eleazar as the third of the three. The precise form of the proper name in the Hebrew
Vorlage of the Antiochene text can of course only be guessed.
the value of the septuagint for textual criticism 233

– Alternatively, the Hebrew reading reflected in the Antiochene is the

original from which the proto-MT corrupted. A possible reason for
the corruption would be the failure of an ancient scribe to identify the
Shammah to which his Vorlage referred. In view of the deplorable tex-
tual state of the list of heroes in 2 Sam 23, this seems to be a reasonable
hypothesis. Instead of Shammah, the scribe introduced the much more
frequent proper name Shimei. Once Shammah became Shimei, the link
between the proper name and the list of David’s heroes was broken,
leading to the further corruption of ‫ ורעיו‬into ‫ ורעי ו‬with ‫ רעי‬being
taken as a second proper name.

Since the whole principle of “tendentious palaeographical exegesis” is

moot,33 it seemed to us that the second scenario is more likely than the

4.2. 1 Kgs 2:3

֨‫מר ֻחקּ ָ ֺ֤תיו ִמ ְצו ָֺתיו‬

ֺ ֨ ‫הו֣ה ֱאל ֶֺ֗היָך ָל ֶל ֶ֤כת ִבּ ְד ָר ָכיו֨ ִל ְשׁ‬
ָ ְ‫ת־מ ְשׁ ֶ ֣מ ֶרת ׀ י‬
ִ ‫וְ ֽ ָשׁ ַמ ְר ָ֞תּ ֶא‬ BHS
‫וֹתיו‬ ָ֔ ‫וּמ ְשׁ ָפּ ָ ֣טיו וְ ֵע ְד‬
“And keep the charge of the LORD your God, to walk in his ways, to keep his
statutes, his commandments, and his judgments, and his testimonies.”

֨‫מר ֻחקּ ָ ֺ֤תיו ִמ ְצו ָֺתיו‬

ֺ ֨ ‫הו֣ה ֱאל ֶֺ֗היָך ָל ֶל ֶ֤כת ִבּ ְד ָר ָכיו֨ ִל ְשׁ‬ ִ ‫ וְ ֽ ָשׁ ַמ ְר ָ֞תּ ֶא‬OHB
ָ ְ‫ת־מ ְשׁ ֶ ֣מ ֶרת ׀ י‬
‫וּמ ְשׁ ָפּ ָ ֣טיו‬
OHB Apparatus
‫ ] משפטיו‬+ ‫ ועדותיו‬M GL, > G(-L)
The apparatus indicates that after the lemma there is a plus in MT (M) and in the
Antiochene text (GL). This element is lacking in the Greek tradition excepting the
Antiochene text (G(-L)).

What is striking in this example is that the non-Masoretic variant is

attested in the main witnesses of the Septuagint, but not in the minus-
cules representing the Antiochene text. This is contrary to expectation,

33 See Tov, Text-critical Use, 100–103.

34 The reading has been adopted by several other textual critics: See, e.g., H. Oort,
Textus hebraici emendationes, quibus in Vetere Testamento neerlandice vertendo usi sunt
(Leiden: Brill, 1900), 45. It is rejected in B. Stade and F. Schwally, The Sacred Books of the
Old Testament. A Critical Edition Of The Hebrew Text, Part 9: The Books of Kings (Leipzig:
Hinrichs, 1904), 62: “ingenious but merely conjectural.”
234 jan joosten

but it is not a unique constellation. The non-Antiochene Septuagint may

reflect the Old Greek here, or, perhaps more probably, it may reflect the
kaige recension made on the basis of a Hebrew Vorlage that didn’t yet
have the addition.
The secondary nature of the phrase ‫ ועדותיו‬in MT (followed by the
Antiochene [καὶ τὰ μαρτύρια αὐτοῦ], Peshitta, Targum, and Vulgate) is
shown by:

– the lack of a counterpart to this element in the non-Antiochene

– the unusual distribution of the connective waw in MT (“his statutes, his
commandments, and his judgments, and his testimonies); one would
expect either: A, B, C, and D, or: A, and B, and C, and D.
– the fact that ‫ עדות‬nowhere else co-occurs with the three other terms
of the list.

The phrase ‫ ועדותיו‬may have been added at a late stage through assimila-
tion to other texts where ‫ עדות‬figured in lists of this type (although not
with the exact same elements).
A common objection to text-critical operations on the Hebrew Bible is
that one should not take elements from one textual tradition to mend
another one.35 It is certainly true that MT and Septuagint represent differ-
ent streams of tradition in 1 Kings. As these two examples show, however,
textual emendations are not based solely on a divergence between wit-
nesses. Their foremost justification is a problem in the received Hebrew
text. In the two passages discussed, the MT is indeed anomalous to a cer-
tain extent. The comparison with other textual witnesses helps to identify
what is problematic in the MT, and it shows a way to a textual solution.
The emendation, however, is in a way called forth by the Masoretic tradi-
tion itself.36

35 See, e.g., D. Barthélemy, Études d’histoire du texte de l’Ancien Testament (OBO 21; Fri-
bourg / Göttingen: Editions Universitaires / Vandenhoeck & Ruprecht, 1978), 369.
36 I owe this insight to my former doctoral student, René Pfertzel.
the value of the septuagint for textual criticism 235

4.3. 1 Kgs 2:1

ֺ ֽ ‫מה ְבנ֖ וֹ ֵלא‬
ֺ ֥ ֺ ‫ת־שׁל‬
ְ ‫י־דִ ֖וד ָל ֑מוּת וַ יְ ַצ֛ו ֶא‬
ָ ‫וַ יִּ ְק ְר ֥בוּ יְ ֵ ֽמ‬ BHS
“Now the days of David drew nigh that he should die; and he charged Solomon
his son, saying”

‫י־דִ ֖וד ָל ֑מוּת‬ ָ ‫וַ יִּ ְק ְר ֥בוּ יְ ֵ ֽמ‬ ‫ויהי אחרי כן וימת דוד וישכב עם אבותיו‬
ֺ ֽ ‫מה ְבנ֖ וֹ ֵלא‬
ֺ ֥ ֺ ‫ת־שׁל‬
ְ ‫וַ יְ ַצ֛ו ֶא‬ ‫ויצו את בנו שלמה לפני מותו לאמר‬

OHB Apparatus
A = M, B = GL (καὶ ἐγένετο μετὰ ταῦτα καὶ ἀπέθανε Δαυιδ καὶ ἐκοιμήθη μετὰ τῶν
πατέρων αὐτοῦ καὶ ἐνετείλατο τῷ υἱῷ αὐτοῦ Σολομῶντι ἔμπροσθεν τοῦ θανάτου αὐτοῦ
Where textual variants line up in such a way as to indicate the presence of two
Hebrew recensions, the OHB format permits to print two parallel columns.38 We
have taken advantage of that possibility for the present verse. The A column, the
earliest text form according to us, contains the MT (M), the B column a Hebrew
text reconstructed on the basis of the Antiochene Greek (GL).

The narrative verse introducing David’s last will is transmitted in two dif-
ferent textual forms, the one represented by the MT, the main witnesses
of the Septuagint, Peshitta, Targum, and Vulgate, the other by the Antio-
chene text.
The Antiochene Greek, καὶ ἐγένετο μετὰ ταῦτα καὶ ἀπέθανε Δαυιδ καὶ
ἐκοιμήθη μετὰ τῶν πατέρων αὐτοῦ καὶ ἐνετείλατο τῷ υἱῷ αὐτοῦ Σολομῶντι
ἔμπροσθεν τοῦ θανάτου αὐτοῦ λέγων, can easily be translated into Hebrew
and probably goes back to a Hebrew Vorlage:

– καὶ ἐγένετο μετὰ ταῦτα = ‫( ויהי אחרי כן‬see Judg 16:4 etc.)
– καὶ ἐκοιμήθη μετὰ τῶν πατέρων αὐτοῦ = ‫( וישכב עם אבותיו‬see 1 Kgs 2:10
– ἔμπροσθεν τοῦ θανάτου αὐτοῦ = ‫( לפני מותו‬cf. 1 Chron 22:5 [LXX ἔμπροσθεν
τῆς τελευτῆς αὐτοῦ]).

37 “And it happened after this that David died and slept with his fathers; but before his
death, he commanded his son Solomon, saying . . .”
38 In 1 Kgs, this is a felicitous possibility. In many passages, the textual facts do indeed
indicate the presence of two (sometimes more) recensions, yet, it is not always possible to
indicate which recension is earlier. The two-column format permits the presentation of all
the textual evidence without needing to decide on the question of priority.
236 jan joosten

But this divergent Hebrew text is almost certainly secondary. In the books
of Kings, the expression ‫ וישכב עם אבותיו‬is always used as a euphemism
for “he died”.39 In Kings, it never combines with the verb ‫“ מות‬to die”.40
The combination of ‫ וישכב עם אבותיו‬with ‫ וימת‬cannot therefore repre-
sent the earliest stage of the text. The alternative version represented by
GL appears to be a later rewriting, drawing on 1 Kgs 2:10 (‫וַ יִּ ְשׁ ַ ֥כּב ָדִּ ֖וד ִעם־‬
‫ ) ֲאב ָ ֺ֑תיו‬and perhaps on a reminiscence of 1 Chron 22:5.

5. Conclusions

In 1 Kings (as in Samuel–Kings in general), the Old Greek is not readily

available, but needs to be recovered from the jungle of attested Greek
text forms. In kaige sections—such as 3 Kgdms 1–2:11—it must usually
be looked for in the Antiochene tradition, although this too represents a
mixture of primitive and secondary elements. Where the Old Greek can
be attained, it is seen to reflect a very good and old Hebrew text that is in
places superior to the MT. In the first two chapters of 1 Kings, we emended
the received Hebrew text twelve times on the basis of the Septuagint, of
which at least five times in a way that crucially alters the meaning of the
passage (as in 1 Kgs 1:8 discussed above).41 Some of our emendations will
perhaps not pass muster, but the large number gives an indication of the
important potential of the Greek tradition in textual studies on 1 Kings.
As the last example indicates, 3 Kgdms represents not only a differ-
ent Hebrew text, but also a different edition of the book. In the first two
chapters, this aspect of the relation between the MT and the Septuagint
is not prominent. But starting in chapter 3, it will become an important
issue. 1 Kings is one of the books (like Jeremiah and Daniel) where tex-
tual criticism bleeds into redaction-historical criticism. Fortunately, this
dimension needs not be dealt with in the present paper.

39 See Stade and Schwally, Books of Kings, 66.

40 It does so combine however in the late biblical Hebrew of Chronicles, see 2 Chron
41 The verses concerned are (an exclamation mark indicates important cases): 1 Kgs 1:2!,
8!, 9, 10 (2x), 11, 19, 35 ; 2:3, 4, 5 (2x)!, 22!
Gottesbildfragen in den Lesartendifferenzen zwischen
dem masoretischen und dem alexandrinischen
Text des Jeremiabuches

Hermann-Josef Stipp

1. Das Problem

Die ursprüngliche antike Übersetzung des Jeremiabuches ins Griechische

(JerG*) ist eine be­sonders interessante Zeugin für die literarische Genese
des Buches, weil sie eine Fassung be­wahrt hat, die etwa um ein Siebtel
kürzer ist als die masoretische Ausgabe und in der Makro­struk­tur sowie
bisweilen auch der Mikrostruktur von jener abweicht1. Dazu kommen
qualita­tive Va­rianten, die an Bedeutung jedoch weit hinter den quantitati-
ven zurückbleiben. Die textkri­ti­sche Auswertung von JerG* wird erleichtert
durch ihre typische Überset­zungssprache, die hoch­gradig standardisiert
ist und wenig Zugeständnisse an die Stilistik und Idiomatik des zeitge­
nössischen Griechischen macht. Diese Eigenart gestattet, die hebräische
Vorlage, wo sie von MT abweicht, zu­meist mit befriedigender Sicherheit
wie­derherzustellen. Seitdem Frag­mente des Jeremia­buches aus Qumran
veröffentlicht wurden, die der rekon­stru­ierten JerG*-Vor­lage nahestehen
(4Q71 bzw. 4QJerb und 4Q72a bzw. 4QJerd)2, setzt sich zuneh­mend die
Überzeu­gung durch, dass die durch JerG* repräsentierte Textform—sie
sei hier die alexandri­ni­sche genannt (AlT)—der masoreti­schen Ausgabe
(MT) an Alter vorausliegt. Ich habe mich diesem Urteil andernorts in dem
Sin­ne angeschlossen, dass JerAlT gegenüber JerMT eine globale Priorität
innehat; d. h. JerAlT ist kein Vorfahr von JerMT in direkter Linie, hat
sich aber seit der Gabelung der Textüberliefe­rungs­stränge in erheblich

1 Vgl. z. B. E. Tov, The Text-critical Use of the Septuagint in Biblical Research, Second
Edition, Re­vised and Enlarged (JBS 8; Jerusalem: Simor 1997), 243f.—Idem, Der Text der
Hebräischen Bibel. Handbuch der Textkritik (Stuttgart: Kohlhammer 1997), 265, bemisst die
Unterschüsse von JerG* sogar auf ein Sechstel des Buches.
2 Eine vorläufige Edition wurde erstellt von J. G. Janzen, Studies in the Text of Jere­miah
(HSM 6; Cambridge: Harvard University Press, 1973), 173–184; die editio princeps findet sich
bei E. Tov, „4QJera–e,“ in Qumran Cave 4. X. The Prophets (DJD XV; Hg. E. Ulrich; Oxford:
Clarendon Press, 1997), 145–207.
238 hermann-josef stipp

geringerem Maß weiter­ent­wickelt als JerMT 3. Hier sei lediglich nochmals

jene Datenklasse her­vorgehoben, der m. E. die größte Be­weiskraft zugun-
sten dieser Verhältnisbestimmung zukommt: Die Sonder­lesarten von
JerMT besitzen eine starke sprachliche Eigen­prägung durch den präma­
so­reti­schen Idiolekt, ein Repertoire von sprach­lichen Merk­malen, die im
maso­reti­schen Sonder­gut jeweils mindes­tens doppelt bezeugt sind, aber
in dem mit JerAlT geteilten Bestand des Bu­ches oder sogar im gesamten
Rest des AT fehlen. Derzeit (Mai 2012) zähle ich 82 solche Phäno­mene
mit zu­sammen über 250 Belegen; davon sind 38 Merkmale mit zu­sam­
men 92 Fällen aus­schließ­lich in den maso­retischen Sonderlesarten des
Jeremiabuches bezeugt4. Zu­sätzlich hat Jan Joosten auf Züge des nach-
klassischen Hebrä­ischen im masoreti­schen Sonder­gut hin­gewiesen5. Da
ein derartiges Kor­pus nicht aus Kür­zungen herrühren kann, ist der Befund
nur erklär­bar, wenn das ma­so­reti­sche Sondergut weit überwiegend aus
Nach­trägen besteht, die auf eine kleine Zahl von Er­gänzern zurückgehen.
Weil indes die globale Priorität von JerAlT definitionsgemäß ein geringes
Maß an Eigenentwicklung ein­schließt, bleibt die Einzel­prüfung der Vari-
anten unver­zichtbar.
Damit steht die generelle Verhältnisbestimmung der beiden wich­tigs­
ten antiken Text­formen des Jeremiabuches fest. Aller­dings haben in den
vergange­nen Jahren Georg Fischer und sein Schüler Andreas Vonach
in einer Serie von Publi­ka­tionen zäh die gegenteilige These ver­fochten.
Für Fischer, der sogar dem prämasore­ti­schen Idiolekt jede Beweiskraft
ab­spricht6, hat dem Schöpfer von JerG* nichts anderes als MT zur Vor­lage
gedient; entsprechend gingen sämtliche Differenzen auf den Übersetzer
zurück, der ein beträcht­liches Maß an Inter­preta­tion in sein Werk ein­
fließen ließ, wie zumal der Um­stand zeige, „dass Jer G dazu neigt, Gott

3 Vgl. H.-J. Stipp, Das masoretische und das alexandrinische Sondergut des Jeremia­
buches. Textgeschicht­licher Rang, Eigenarten Triebkräfte (OBO 136; Freiburg / Göttin­gen:
Universitätsverlag / Vanden­hoeck & Ruprecht, 1994), sowie die Aktualisierungen in idem,
„Zur aktuellen Diskussion um das Verhältnis der Textformen des Jeremiabuches,“ in Die
Septuaginta—Texte, Kontexte, Lebenswelten (WUNT 219; Hg. M. Karrer und W. Kraus;
Tübingen: Mohr Siebeck, 2008), 630–653.
4 Vgl. vorläufig H.-J. Stipp, „Linguistic Peculiarities of the Masoretic Edition of the Book
of Jeremiah. An Up­dated Index,“ JNSL 23/1 (1997): 181–202; sowie den in Anm. 3 genannten
Aufsatz. Ein vervollständigtes Inventar befindet sich in Vorbereitung.
5 J. Joosten, „L’excédent massorétique du livre de Jérémie et l’hébreu post-classique,“ in
Conservatism and Innovation in the Hebrew Language of the Hellenistic Period. Pro­ceedings
of a Fourth International Symposium on the Hebrew of the Dead Sea Scrolls & Ben Sira (STDJ
73; Hg. J. Joosten und J.-S. Rey; Leiden: Brill, 2008), 93–108.
6 G. Fischer, „Die Diskussion um den Jeremiatext,“ in idem, Der Prophet wie Mose. Stu-
dien zum Jeremiabuch (BZAR 15; Wiesbaden: Harrassowitz, 2011), 73–89 (87–89).
gottesbildfragen in den lesartendifferenzen 239

milder, weniger menschlich [d. h. we­ni­ger an­thro­po­morph, HJS] oder

gewalt­tätig zu schil­dern“7. Vonach er­klärt in der Einleitung zur Wieder­
gabe des griechischen Jeremiabuches in der „Septuaginta Deutsch“
(LXX.D): „Inhaltlich-theologisch bemüht sich die Septuaginta um ein mil­
deres Sprechen Got­tes selbst, aber auch über ihn . . . Sie formuliert generell
vorsichtiger und pietät­voller und be­tont mehr die Souveränität Gottes,
indem sie z. B. allzu anthropo­mor­phe Regun­gen vermei­det.“8 Fischer und
Vonach widersprechen damit dia­me­tral dem Urteil von Ber­nard Zloto-
witz, JerG* lasse keine Nei­gung zur Reduktion von Anthropo­morphismen
er­kennen9. Mehr noch: Die Modifikationen am Gottesbild bildeten sogar
bloß einen exem­pla­ri­schen Ausschnitt aus vielfältigen Freiheiten, die
sich der Übersetzer gestattete, um sei­nem Zielpub­li­kum den Zu­gang zum
Jere­mia­buch zu ebnen. In den Worten Fischers: „Jer-LXX be­zeugt einen
Über­setzer, der als from­mer Jude die Botschaft die­ses Buches seinen
Zeitge­nos­sen in ihrem helle­nistisch-ägypti­schen Kul­tur­kreis nahebrin­gen
wollte.“10 Vonach hat diese Überzeu­gung nun­mehr zur Grundlage seiner
Erläuterung des griechischen Jeremiabuchs im Kommen­tar zur LXX.D
Wie an anderer Stelle nochmals zu bekräftigen ist, liegt diese zusam­
men­fassen­de Cha­rakterisierung von JerG* fernab der Wirk­lichkeit12. Das
griechische Je­remiabuch repräsentiert vielmehr eine „imitative technique
of translation“13, wie sie typisch ist für eine „ausgangs­sprachliche“ Wieder-
gabe, die „wörtlich, aber grammatisch und/oder inhaltlich nicht korrekt“

7 Ibid., 85.
8 A. Vonach, „Jeremias—Das Buch Jeremia. Einleitung,“ in Septuaginta Deutsch. Das
griechische Alte Testament in deutscher Übersetzung (Hg. W. Kraus und M. Karrer; 2., ver-
besserte Aufl. Stutt­gart: Deutsche Bibelgesellschaft, 2010), 1288f. (1289).
9 B. M. Zlotowitz, The Septuagint Translation of the Hebrew Terms in Relation to God in
the Book of Jere­miah (New York: KTAV, 1981), mit dem Fazit S. 183: „The data throughout
this thesis have profusely demonstrated that the Septuagint translator(s ?) of the book of
Jeremiah reproduced literally and correctly the Hebrew terms related to God.“
10 G. Fischer, Jeremia. Der Stand der theologischen Diskussion (Darmstadt: Wiss. Buch-
gesellschaft, 2007), 52.
11 A. Vonach, „Jeremias,“ in Septuaginta Deutsch. Erläuterungen und Kommen­tare zum
griechischen Alten Testament, Bd. II: Psalmen bis Daniel (Hg. M. Karrer und W. Kraus;
Stuttgart: Deutsche Bibelgesellschaft, 2011), 2696–2814.
12 H.-J. Stipp, „Die Jeremia-Septuaginta als theologische Programmschrift. Zur Kommen-
tierung des griechischen Jeremiabuches im Rahmen der ‚Septuaginta Deutsch‘ (LXX.D),“
erscheint in: BZ.
13 J. Barr, The Typology of Literalism in Ancient Biblical Translations (MSU 15 = NAWG.
PH 11; Göttingen: Vandenhoeck & Ruprecht, 1979), 292.
240 hermann-josef stipp

verfährt14. „The most basic norm of the translator . . . was that of isomor-

phism. That is to say, typically each morpheme of the source text is rep-
resented in the target text.“15 Das Ergebnis muss seiner­zeit einen überaus
fremd­artigen, schwerfälligen Klang besessen haben, denn es strotzt von
Hebraismen und mutet seinen Benut­zern nicht selten kaum verständli-
che Passagen zu. Der Wille, den Text für die Leser­schaft aufzu­bereiten,
war in Wahrheit zwar nicht völ­lig in­existent, aber eng begrenzt. JerG*
enthält ein geringes Maß interpretierender Ele­mente, die eindeutig nicht
der Vorlage, sondern dem Übersetzer zuzurechnen sind. Um einige beson­
ders klare, da mehrfach bezeugte Beispie­le zu nennen: Der Klageruf ‫א ָההּ‬,ֲ
in Jer nur in Reden des Propheten an Jhwh belegt, wird auf die Wurzel
‫ היה‬bezogen und als Gottesanrede ὁ ὤν wieder­gegeben16, eine offenkundig
von ExG* 3:14 inspi­rierte Deutung17. Das Sub­­stantiv ‫ נָ ִביא‬wird in Fällen,
die von Falschpropheten reden, häufig mit ψευδο­προφή­της über­setzt18. Für
das Verb ‫ קרא‬tritt, wenn in Auf­forderun­gen Jhwhs an Jeremia gerichtet,
immer ἀναγινώσκω vorlesen ein, also auch dann, wenn der Kontext vom
Vortrag einer schriftlichen Fassung nichts weiß19.
Da ein beschränktes Ausmaß solcher Einflüsse nicht zu leugnen ist, soll
im Folgen­den ein Kernpunkt der Position Fischers und Vonachs geprüft
werden, nämlich die These, der Überset­zer habe bevorzugt die Gottes­
bilder seiner Vorlage entschärft20. Ein Ver­gleich der an­tiken Aus­gaben
des Jeremia­buches unter dieser Rücksicht erscheint lohnend, weil JerG*
zwar nicht entfernt einen solch ausgeprägten Hang zu inter­pre­tierenden
Wieder­gaben erkennen lässt, wie Fischer und Vonach be­haup­ten, aber
in der Tat mehrfach an solchen Stellen von MT abweicht, wo Aspekte
der Gottes­vorstellung zur Sprache kommen. Vorweg ist zu beto­nen, dass

14 F. Siegert, Zwischen Hebräischer Bibel und Altem Testament. Eine Einführung in die
Septuaginta (Münstera­ner Judaistische Studien 9, Münster: Lit, 2001), 133.
15 A. Pietersma und M. Saunders, „To the Reader of Ieremias,“ in A New English Trans-
lation of the Septuagint and the Other Greek Translations Traditionally Included under that
Title (Hg. A. Pietersma und B. G. Wright; New York: Oxford University Press, 2007), 876–881
16 1:6; 4:10; 14:13; 32:17.
17 Vgl. Siegert, Einführung (Anm. 14), 253–255; E. G. Dafni, „Οἱ οὐκ ὄντες θεοί in der Sep-
tuaginta des Jeremia­buches und in der Epistel Jeremias. Ein Beitrag zur Frage nach dem
Werdegang des sogenannten alexandrinischen Kanons,“ in The Biblical Canons (BETL 163;
Hg. J. M. Auwers und H. J. de Jonge; Leuven: Univer­sity Press, 2003), 235–245 (241f.).
18 6:13; 26:7, 8, 11, 16; 27:9/34:7; 28:1; 29:1, 8; ferner Sach 13:2.
19 3:12; 11:6; 19:2. Bei auch in MT vorausgesetzten Akten des Vorlesens: 29:29; 36:6, 8, 10,
13, 14, 15, ‌21, ‌23; 51:61, 63.
20 Vgl. schon die Bemerkungen bei Stipp, „Diskussion“ (Anm. 3), 651f., doch soll das Pro-
blem hier nochmals auf verbreiterter Beobachtungsbasis wiederaufgenommen werden.
gottesbildfragen in den lesartendifferenzen 241

kein objektivierbares oder gar automatisier­bares Ver­fah­ren zur Erschlie-

ßung des einschlägigen Materials existiert. Die Fall für Fall vor­ge­hende
Beurteilung, welche Devi­anzen die transportierten Gottesbilder beein-
flussen, ist nicht zu umgehen. Die unten besprochenen Stellen wur­den
über längere Zeit hinweg entlang der Arbeit am Text des Jeremia­buches
gesammelt. Damit wird zwar keine Vollstän­digkeit, aber hinrei­chende
Repräsentativi­tät bean­sprucht. Eingeschlossen sind die recht we­nigen
diskus­sionswürdigen Beispiele, die Fischer und Vonach zur Stüt­zung ihrer
These anführen21.
Selbstredend können hier unmöglich alle Argumente Fischers und
Vonachs zurück­ge­wiesen werden. Dies geht schon aus Raumgründen
nicht, weil hier ihr Verfahren, bei der Rechenschaft über die relevanten
Fak­ten recht sparsam vorzugehen, nicht fortgesetzt werden kann. Über­
gangen werden namentlich solche Fälle, die uner­hebliche Differenzen
aufbau­schen, offenkundige Fehler enthalten oder gravierende Tatsachen
ignorieren. So liest man etwa bei Vonach zu 2:20: „Die Ver­meidung har­ter
Aus­sagen über oder von Gott wird oft le­dig­lich durch eine einfache Ände-
rung der Kommunikationsstruktur oder des Subjekts erzielt. So spricht in
2,20 das Wort über die Hure­rei in der LXX durch Weiter­füh­rung der direk-
ten Rede das Volk selbst, in MT hingegen Gott.“22 Vonach scheint nicht
zu wissen, dass die Lesung διαχυθήσομαι auf den Verlust des Spatiums in
‫ ַא ְתּ צ ָֹעה‬zurückgeht, das grammatisch inkorrekt als ‫ ֶא ְת ָצ ֶעה‬gelesen wurde
(richtig: ‫)א ְצ ָט ֶעה‬.ֶ Die Verschiebung des Sprecher­wechsels resultiert daher
nicht aus einem interpretierenden Eingriff des Übersetzers, sondern aus
einer fehlerhaften Vorlage.—Ebenso deutet Vonach das Nebeneinander
von ‫ יָ ק ְֹשׁ ִתּי ָלְך‬und ἐπιθήσονταί σοι in 50:24 als „Subjektänderung“ des Über-
setzers zur Milderung des Gottesbildes23. Dabei deutet das Verb auf die
Vorlage ‫יָ ִשׁתוּ‬, die durch Ausfall des ‫ ק‬und Anpassung der Endung aus der
masoretischen Lesart hervorging.—Weiterhin wird erklärt: „Viele solche
‚Kor­rek­turen‘ des Gottesbildes werden in der LXX außer­dem da­durch
er­reicht, dass in 1. Sg. formulierte Aussa­gen Gottes oder solche in 3. Sg.
über ihn in ein unpersonales Passiv oder eine indiffe­rente Nominalkon-
struktion umformu­liert werden.“24 Das soll beispielsweise für 6:2 gelten.
Dabei fehlt der unerlässliche Verweis auf die Äquiva­lenz ‫יתי‬ ִ ‫ ≙ ָדּ ִמ‬τὸ ὕψος
σου, die zeigt, dass in der Vorlage das (offenbar defektiv geschriebene)

21 Vgl. unten Anm. 30, 34, 43, 46, 49, 63, 74, 83, 99.
22 Vonach, „Jeremias“ (Anm. 11), 2725.
23 Ibid.
24 Ibid.
242 hermann-josef stipp

Prädikat durch eine gängige ‫ר‬/‫ד‬-Ver­wechslung zu der nominalen Form

‫*ר ָמ ֵתְך‬
ָ verstümmelt und daher verlo­ren­gegangen war, was den Übersetzer
nötigte, aus dem Kontext ein neu­es Prädikat zu erra­ten. Dabei fiel seine
Wahl auf ‫וְ ַה ְמּ ֻענָּ גָ ה‬, eine Bildung, die seine Hebrä­ischkenntnisse ohnehin
vor eine ho­he Hürde gestellt haben dürfte. Das Vorurteil, die meisten
Devianzen gingen auf den Über­setzer zurück, hindert Vonach jedoch
daran, eine abweichende Vorlage überhaupt in Betracht zu ziehen. Die
gegebenen Um­stände bilden freilich eine enorme methodische Erschwer-
nis, will man aus JerG* konzep­tionelle Differenzen zu MT erheben.—Zu
49:37 (AlT 25:17) bemerkt Vonach: „In 25,17 richtet in MT Gott selbst Elam
aktiv zugrunde (‫‚ עד כלותי אותם‬bis ich sie zugrunde gerichtet habe‘),
während das zugrunde richtende Sub­jekt in der LXX durch die Wieder-
gabe im Infinitiv das Schwert des vorangehenden Satzglie­des ist“25. Dabei
wird übersehen, dass Jhwh in JerG* dieses Schwert aus­drücklich als τὴν
μάχαιράν μου mein Schwert reklamiert. Das auf ihn zurück­verweisende
Pro­no­men hat lediglich die Position gewechselt, und zwar derart, dass
an der göttlichen Urheber­schaft kein Zweifel bleibt.—Argumentationen,
wie sie diese Bei­spiele repräsentieren, sind bei Fischer und Vonach leider
keine Selten­heit.
Die Lesartenunterschiede sind zu befragen, ob sie dazu beitragen,
JerMT und/oder JerAlT mit distinkten konzep­tio­nellen Pro­filen zu ver-
sehen. Denn selbstverständlich können Differenzen nicht a priori dem
Überset­zer ange­lastet wer­den. Vielmehr kommen stets min­destens drei
verschiedene Ursachen in Betracht, zwi­schen de­nen zu ent­scheiden ist, was
indes nicht immer gelingt: (1) MT hat sich von AlT wegent­wickelt; (2) die
hebräi­sche Fassung von AlT hat sich von MT entfernt; (3) der Schöpfer
von JerG* ist von seiner Vorlage abge­wi­chen26. Es gilt je­weils zu klä­ren,
welche die­ser Quellen die größte Wahrschein­lichkeit besitzt. Nur wenn
Ort und Richtung einer Modifikation be­stimmt sind, lassen sich auch ihre
Motive ein­grenzen. Konvergenzen können Hinweise lie­fern auf Trends
oder gar spezi­fi­sche Theo­logien, die in dem betroffenen Texttyp bzw. in
der Übersetzung wirksa­m sind.
Die nachstehende Erörterung listet Lesartendifferenzen auf, die An­lass
zur Frage ge­ben, ob bei ihrer Entstehung die Reflexion von Gottes­

25 Ibid.
26 Die drei Alternativen vermeiden das methodische Manko der Arbei­ten Fischers, der
aufgrund sei­ner Vorentscheidung, Lesartdifferenzen seien immer den Übersetzern zuzu­
schreiben, nie­mals die Mög­lichkeit einer nichtmasoretischen Vorlage von JerG* in Be­tracht
zieht. In abgemilderter Form vertritt dieses Apriori auch Vonach, „Jeremias“ (Anm. 11).
gottesbildfragen in den lesartendifferenzen 243

bildproblemen eine Rolle gespielt hat. Die Analyse wird in einen Ver­such
münden, die Belege in Klassen zu ordnen, doch um nichts zu prä­judi­
zieren, wird das Material nach dem masoretischen Bucharrangement
dargeboten. Zitate aus dem Hebräischen folgen den Konventio­nen einer
textkriti­schen Syn­opse zum Jere­miabuch, die sich in Vorbereitung befin­
d­et27. Rückübersetzte alexandrinische Lesarten wer­den mit tiberischer
Vokalisation versehen, die als Symbolsystem zur Repräsentanz einer
be­stimmten Interpretation des Konsonantentextes dient. Eine Reihe
diakri­tischer Symbole hebt alexan­drinische Ab­weichungen hervor, die auf
die JerG*-Vorlage zurückgeführt werden:

[. . .] Eckige Klammern markieren masoretische Überschüsse.

<. . .> Spitze Klammern bezeichnen Überhänge der alexandrini­schen Aus-
. . . \ . . . Schrägstri­che geben qualitative Varianten an: zuerst in Leserichtung die
maso­retische Lesart, danach die alexandri­ni­sche.
* Drehachse: Die durch das Drehachsensymbol ge­trennten Passagen ste-
hen im alex­andrinischen Text in umgekehrter Reihenfolge.
#. . .§ Bei Bedarf stecken Grenzzeichen die Reichweite einer No­tation ab.
(. . .) Runde Klammern fügen erforder­lichenfalls weitere Infor­mationen
| Ein senkrechter Strich bezeichnet Satzgrenzen in AlT.
≙ entspricht
Gram! ungrammatisch
TT Tiberischer Text (d. h. MT in tiberi­scher Vokalisation).
JE12 Die Büchergruppe Jer, Ez und Dodekapropheton, die wahr­schein­lich
von dem­selben Übersetzer(kreis) ins Griechi­sche über­tragen wurde28.

Nötigenfalls werden die Verse in Sätze zerlegt (a, b, c, . . .). Wo es gebo­ten

erscheint, wird auch der griechische Text beigegeben. Dagegen wer­den
für den vorliegenden Zusammenhang unerheb­liche Varianten ausge­spart.
Die andersartige Kapitel- und Verszählung von AlT in Jer 26–51 wird nur
bei Bedarf nach Schrägstrich beigegeben.

27 H.-J. Stipp, Textkritische Synopse zum Jeremiabuch (Manu­skript, 7. korrigierte in­terne

Auflage, München, 2011).
28 Vgl. Stipp, Sondergut (Anm. 3), 17–19.
244 hermann-josef stipp

2. Das Material29

1:1 ‫ יִ ְר ְמיָ הוּ‬# ‫( | ֲא ֶשׁר ָהיָ ה ֶאל‬τοῦ θεοῦ) ‫§ ִדּ ְב ֵרי \ ְדּ ַבר־יְ הוָ ה‬
τὸ ῥῆμα τοῦ θεοῦ ὅ ἐγένετο ἐπὶ Ιερεμίαν
MT deklariert den Inhalt des Buches als Worte Jeremias, während AlT
das Folgende über­schreibt als Wort Jhwhs, das an Jeremia erging. JerG*
verwendet typisches Über­setzungs­griechisch (ῥῆμα . . . ἐγένετο ἐπί) und
unter­mauert so, dass die Variante eine hebräische Vorlage spiegelt. Diese
Fassung verdoppelt die Wort­ereignisformel V. 2 und un­terstreicht die
Kontrolle Jhwhs über das Offenbarungs­ge­schehen, was für den sekundä­
ren Einfluss theologi­scher Reflexion spricht. Somit wurde in der alexan­
dri­nischen Texttradition die Souveränität Jhwhs gesteigert, freilich schon
in der hebräi­schen Phase, nicht durch den grie­chischen Überset­zer30. Ein
analoger Eingriff ist in 51:59 AlT zu beobachten (s. u.).

1:17 ‫יהם‬
ֶ ֵ‫ ִל ְפנ‬# ‫ל־תּ ַחת‬ֵ ‫ן־א ִח ְתָּך \ וְ ַא‬ֲ ‫ § ֶפּ‬f ‫יהם‬
ֶ ֵ‫ל־תּ ַחת \ ִתּ ָירא ִמ ְפּנ‬
ֵ ‫ ַא‬e
<‫ילָך נְ ֻאם־יְ הוָ ה‬
ֶ ‫י־א ְתָּך ֲאנִ י ְל ַה ִצּ‬
ִ ‫> ִכּ‬
μὴ φοβηθῇς ἀπὸ προσώπου αὐτῶν μηδὲ πτοηθῇς ἐναντίον αὐτῶν ὅτι μετὰ
σοῦ ἐγώ εἰμι τοῦ ἐξαιρεῖσθαί σε λέγει κύριος
In MT trägt Jhwh eine scharfe Drohung vor: Wenn Jeremia sich von sei­
nen Geg­nern ein­schüchtern lässt, wird Jhwh diesen Effekt noch zusätz­lich
in die Höhe treiben. AlT dagegen bietet die mit ‫ ירא‬und ‫ חתת‬ge­bilde­te
er­weiterte Beruhigungsformel im üblichen Vetitiv: ‫ל־תּ ַחת‬ ֵ ‫ל־תּ ָירא וְ ַא‬
ִ ‫ ַא‬31.
Es folgt eine weitere Zu­sage des Mit­seins nebst Gottesspruchformel mit
wörtlichen Par­allelen in 8b—dort nach der einfachen Be­ruhigungsformel—
und 19c. Das Über­set­zungsgrie­chische (ἐγώ εἰμι für ‫ ; ֲאנִ י‬τοῦ + Infinitiv für
‫ ל‬+ Infinitivus constructus) bestätigt die Abkunft von einer hebräi­schen
Vor­lage. Der Über­hang widerspricht dem früher belieb­ten Standpunkt,
die alexandrini­sche Tradition habe Dub­letten getilgt; eher ist JerAlT
hier umge­kehrt konflationär angereichert worden. Laut dieser Textform

29 Der Text von JerG* ist entnommen aus der Edition von J. Ziegler, Ieremias, Baruch,
Threni, Epistula Ie­remiae (Septuaginta. Vetus Testamentum Graecum Auctori­tate Acade-
miae Scientiarum Gottingensis edi­tum 15, Göttingen: Vandenhoeck & Ruprecht ²1976).
Übersetzungen von JerG* sind angelehnt an die Wieder­ga­be von G. Fischer und A. Vonach
in LXX.D: Septuaginta Deutsch. Das griechi­sche Alte Testament in deutscher Übersetzung
(Hg. W. Kraus und M. Karrer; Stuttgart: Deutsche Bibelgesellschaft, 2009), 1288–1342.
30 So aber Vonach, „Jeremias“ (Anm. 11), 2712.
31 Jer 30:10 MT || 46:27; Dtn 1:21; Jos 8:1; 10:25; Jes 51:7; Ez 2:6; 3:9; 1 Chr 22:13; 28:20; 2 Chr
20:15, 17; 32:7; als Prohibitiv: Dtn 31:8; vgl. Narrativ 1 Sam 17:11.
gottesbildfragen in den lesartendifferenzen 245

gewährt Jhwh dem Pro­pheten un­eingeschränkten, in kli­schier­te Ter-

minologie gefassten Zuspruch. Der Über­schuss, der nur mit der alex-
andrinischen Fassung von 17ef harmoniert, und der kon­ventio­nel­lere
Sprach­ge­brauch deuten auf eine se­kundä­re Milderung des Gottes­bildes,
indem man die Drohung durch eine unbedingte Beistands­zusage ersetzte,
sodass der schroffe Umgang Jhwhs mit Jere­mia entfiel. Der Eingriff wurde
allerdings nicht vom Übersetzer vorgenommen, son­dern fand be­reits in
der vorausliegenden hebräischen Tradition statt.

6:11 ‫אתי‬
ִ ‫ ָמ ֵל‬# ‫ וְ ֵאת § ֲח ַמת יְ הוָ ה \ ֲח ָמ ִתי‬a
(συνετέλεσα √ ‫יתי ָה ִכיל )כלה‬ ִ ‫ נִ ְל ֵא‬b
‫חוּרים יַ ְח ָדּו‬
ִ ‫( וְ ַעל סוֹד ַבּ‬ἔξωθεν) ‫ל־עוֹלל ַבּחוּץ‬ ָ ‫ > ֶא< ְשׁפְֹך ַע‬c
 on der Zornesglut Jhwhs bin ich voll; ich bin überdrüssig, (sie) zurück-
Gieß (sie) aus über den Säugling auf der Gasse und über die Runde der
Burschen zugleich!

καὶ τὸν θυμόν μου ἔπλησα καὶ ἐπέσχον καὶ οὐ συνετέλεσα αὐ­τούς ἐκχεῶ
ἐπὶ νήπια ἔξωθεν καὶ ἐπὶ συναγωγὴν νεανίσκων ἅμα
 nd ich füllte meinen Grimm, aber ich hielt ihn zurück und bereitete
ihnen nicht das Ende. Ich werde (ihn) ausgießen über unmündige Kin-
der draußen und über eine Ver­sammlung von Jugendlichen zugleich.
Laut MT bekennt in 11ab ein menschlicher Redner—laut Buchkontext
Jeremia—, von nicht mehr be­herrschbarer Zornes­glut Jhwhs erfüllt zu
sein. In 11c rea­giert nach ei­nem un­markier­ten Spre­cherwechsel die frag-
los göttliche Stimme mit dem Be­fehl (Imperativ), die Zor­nesglut über die
Judäer zu ergießen. Diese Rederollen setzen einen weiteren un­mar­kierten
Sprecher­wechsel im Vortext voraus. Denn die Botenformel in V. 9 dekla-
riert das Fol­gende als Gottes­wort, aber wenn in 11a Jeremia zu vernehmen
ist, hat er ohne explizites sprachliches Signal das Wort ergrif­fen, was am
ehesten mit Beginn von V. 10 geschieht. In AlT hingegen bildet V. 9–15 eine
geschlossene Gottes­rede (die Selbst­referenz Jhwhs wie auf einen Dritten
nach Art von 10f ist gängig32 und daher kein Ein­wand). Folglich ist es dort
Jhwh, der laut 11ab mei­ne Zornes­glut (‫ ) ֲח ָמ ִתי‬nicht mehr zügeln will und

32 H.-J. Stipp, Elischa—Propheten—Gottesmänner. Die Kompositionsgeschichte des

Elischazyklus und ver­wandter Texte, rekonstruiert auf der Basis von Text- und Literar­kri­tik
zu 1 Kön 20.22 und 2 Kön 2–7 (ATSAT 24; St. Ottilien: Eos, 1987), 120–130.
246 hermann-josef stipp

in 11c ihre Entfesselung (‫ ) ֶא ְשׁפְֹך‬an­kün­digt33. Die unbefriedi­gend ange­

zeig­ten Spre­cherwechsel in MT könnten die ale­xan­dri­ni­schen Tradenten
veranlasst haben, den Passus zu vereinfachen, indem sie ihn durch­gehend
der Stim­me Jhwhs unter­stell­ten. Dann müss­ten sie zu diesem Zweck das
Gottes­bild aller­dings deut­lich getrübt ha­ben, und man fragt sich, warum
sie hierzu die Anrede an den Pro­pheten in 11c aufgaben, die ohnehin nur
aus dem Mund Jhwhs kommen kann. Da­her überzeugt es mehr, die Modi-
fikation in der ma­so­reti­schen Überlie­ferung zu suchen, wo man einen Teil
des Passus in den Mund Jeremias verlegte, um den Eindruck zu vermei-
den, Jhwh habe seine eige­nen Emo­tionen nicht im Griff. Der Impe­rativ
‫ ְשׁפְֹך‬mit (hier kontextgetilgtem) Objekt ‫ ֵח ָמה‬kann zu­sätz­lich durch die
Parallele in 10:25 angeregt worden sein.
Demnach erwuchsen die Devianzen zwar aus der theologischen Arbeit
am Gottes­bild, jedoch im maso­retischen Strang, und sie sollten auch
nicht Jhwh sanfter zeichnen, sondern seine Selbst­kontrolle dem Zweifel
entziehen. Sollte dagegen der Wandel tatsächlich auf alex­andri­ni­scher
Seite eingetreten sein, hätte man dort zur Behebung eines Pro­blems der
Text­kohärenz den Charakter Jhwhs erheblich verdunkelt. Dies gälte in
geringerem Maß für den Über­setzer, gin­gen die Abwei­chungen auf ihn
zurück. Indem er die Verb­form ‫ ָה ִכיל‬falsch analysierte, be­wahrte er Jhwh
vor dem Ver­dacht mangeln­der Herrschaft über sich selbst, doch hätte er
durch ἐκχεῶ den Vollzug des Unheils in Jhwhs eigene Hände verlegt.

9:9 ‫ל־ה ָה ִרים ֶא ָשּׂא \ ְשׂאוּ ְב ִכי ]וָ נֶ ִהי[ וְ ַעל־נְ אוֹת ִמ ְד ָבּר ִקינָ ה‬
ֶ ‫ַע‬
ἐπὶ τὰ ὄρη λάβετε κοπετὸν καὶ ἐπὶ τὰς τρίβους τῆς ἐρήμου θρῆνον
Der Satz ist Gottesrede, vorweg markiert durch die Botenformel V. 6, die
Gottesspruchformel V. 8 und ein Ich, das sich Handlungen zuschreibt,
die nur Gott offenstehen (V. 6, 8). Daher er­hebt Jhwh in MT ein Weinen
und ein Klagelied, während er in AlT eine sonst im näheren Kon­text nicht
genannte Hörerschaft zu diesen Riten aufruft. Dies kann den An­schein
erwe­cken, ein alexandrinischer Tradent oder der Übersetzer habe den
anthropomorphen Zug der offen zur Schau getragenen Trauer Gottes
beseitigen wollen34. Allerdings zitiert auch AlT Jhwh kurz zuvor in 8:23/

33 Die griechische Wiedergabe von 11b gibt kein Recht, eine abweichende Vorlage zu
erschließen, sondern erklärt sich aus dem irrigen Bezug von ‫ ָה ִכיל‬auf ‫כלה‬. Außerdem
bereitete dem Übersetzer das Verb ‫ לאה‬immer Schwierigkeiten; vgl. sonst 9:4; 12:5; 15:6;
34 So für den Schöpfer von JerG* Fischer, „Diskussion“ (Anm. 6), 84.
gottesbildfragen in den lesartendifferenzen 247

9:1 mit Worten, in denen er sich offen zu seinen Tränen bekennt, und
in 8:21 erweitert sogar ein alexandrinischer Überhang die Gottesklage um
die massiv anthropomorphe Vorstellung, ‫יּוֹל ָדה‬ ֵ ‫( ִחיל ַכּ‬ὠδῖνες ὡς τικτού­
σης) hätten Jhwh erfasst. Vom Weinen Gottes weiß AlT ferner in 48/
31:32.—Eine zu 9:9 analoge Varian­tenkon­stel­lation begegnet in 14:17 (s.
u.). Diese Befunde sind mit 13:17 zu vergleichen, wo laut MT Jeremia das
Schicksal der Judäer beweint, während AlT die Klage ebenfalls den Judä-
ern zuschreibt: ‫י\כם ִמ ְפּנֵ י‬ ֶ ‫י<כם> ִדּ ְמ ָעה ְבּ ִמ ְס ָתּ ִרים ִתּ ְב ֶכּה־נַ ְפ ִשׁ‬
ֶ ִ‫וְ ֵת ַרד ֵעינ‬
]. . .[ ‫גֵ וָ ה‬. Wäre es den ale­xandrini­schen Texttradenten darum gegangen,
An­thropomorphismen zu vermeiden, hätten sie in die Klage Jeremias
nicht einzugreifen brauchen. Nach 13:17 zu schließen, war AlT indes pri-
mär bestrebt, die Trauer über die Not der Judäer generell den Betroffenen
aufzubürden, während die Reduktion von Anthropomorphismen allen­
falls ein nachrangiges Motiv lieferte. Vor allem sollten die Judäer ihre aus
eigener Schuld erwachsene Misere selbst be­weinen. Die alexandri­ni­schen
Lesarten folgen somit einheitlich dem Muster von 9:17, wo der Appell
‫ וְ ִת ֶשּׂנָ ה ָע ֵלינוּ נֶ ִהי‬an Menschen (Klage­frauen V. 16) ergeht35 und auch den
Judäern die Tränen fließen: ‫וְ ֵת ַר ְדנָ ה ֵעינֵ ינוּ ִדּ ְמ ָעה‬. Im gegebe­nen Fall sind
keine Merkmale zu erken­nen, die ein Urteil erlauben, ob der Wandel im
Zuge der Übersetzung oder vielmehr schon in der he­bräi­schen Pha­se der
alexandri­ni­schen Text­über­lie­ferung eingetreten ist.

10:16 ‫ וְ יִ ְשׂ ָר ֵאל ֵשׁ ֶבט[ נַ ֲח ָלתוֹ‬c] ‫י־יוֹצר ַהכֹּל הוּא‬

ֵ ‫ ִכּ‬b ‫א־כ ֵא ֶלּה ֵח ֶלק יַ ֲעקֹב‬
ְ ֹ‫ל‬a
51:19 ‫ וְ ֵשׁ ֶבט[ נַ ֲח ָלתוֹ‬c] ‫י־יוֹצר ַהכֹּל הוּא‬ֵ ‫ ִכּ‬b ‫א־כ ֵא ֶלּה ֵח ֶלק יַ ֲעקֹב‬
ְ ֹ‫ל‬a
Die Dublette kontrastiert den „Anteil Jakobs“ mit den nichtigen Götzen­
bildern (10:14–16 || 51:17–19). Gegenüber den drei Sätzen der masoreti­schen
Ausgabe kennen die alexandrini­schen Fas­sungen nur zwei. Dabei fährt
der zweite Satz jeweils folgerichtig fort, weil ‫ נַ ֲח ָלתוֹ‬dort parallel mit ‫ֵח ֶלק‬
‫ יַ ֲעקֹב‬gefügt ist: ‫ נַ ֲח ָלה‬präzisiert ‫ ֵח ֶלק‬, und das Enklitikon vertritt ‫יַ ֲעקֹב‬,
sodass in beiden Sätzen Jhwh als Israels An- bzw. Erbteil proklamiert
wird. In 10:16c MT hingegen ist umgekehrt Israel der Stamm von Jhwhs
Erbteil, wozu 51:19c MT eine verstüm­melte Variante bietet. MT vertritt die
gängigere Konzeptionalisierung des Verhältnisses von Jhwh und Israel,
wie die näheren Parallelen in Jes 63:17 und Ps 74:2 sowie die entfernte-
ren in Dtn 32:9 und Ps 78:71 belegen. Zudem kommt ‫ ֵשׁ ֶבט‬in Jer nur hier
vor und hat damit idio­lek­talen Status. Folg­lich hat eine prämaso­retische

35 Von dort dürfte MT seinerseits ‫ וָ נֶ ִהי‬entlehnt haben.

248 hermann-josef stipp

Hand die beiden Verse einer Orthodoxie unter­worfen, der an der Domi-
nanz Jhwhs im Verhältnis zu seinem Volk gelegen war36.

10:18 ‫ת־יוֹשׁ ֵבי ָה ָא ֶרץ ] ַבּ ַפּ ַעם[ ַהזֹּאת‬

ְ ‫( ֶא‬Gram!) ‫קוֹל ַע‬
ֵ \ ‫צוֹל ַע‬
ֵ ‫ ִהנְ נִ י‬b
# ‫רוֹתי ָל ֶהם \ ַבּ ָצּ ָרה‬
ִ ‫ § וַ ֲה ֵצ‬c
ἰδοὺ ἐγὼ σκελίζω τοὺς κατοικοῦντας τὴν γῆν ταύτην ἐν θλίψει
Siehe ich, ich lasse die Bewohner dieses Landes stolpern37 in Drangsal.
Für 18c liest AlT nur in Drangsal, während MT Jhwh nochmals direkt
involviert mit und ich werde sie bedrängen. Vorweg in 18b kündigt Jhwh
ein massiv anthropomorphes Handeln an, das der Über­set­zer unver­
wässert wiedergab (wobei das Partizip in der Vorlage von ‫קוֹל ַע‬ֵ zu ‫צוֹל ַע‬
ver­schrieben war, dem er wider die Grammatik eine transitive Bedeutung
unterschob). Daher ist nicht ein­zusehen, weswegen allein 18c entschärft
worden sein sollte38. Mehr Wahr­scheinlichkeit hat für sich, dass in MT die
Beteiligung Jhwhs nachträglich gesteigert worden ist.

11:1; 18:1; 21:1; 32:1; 40:1 # ‫ַה ָדּ ָבר ֲא ֶשׁר ָהיָ ה § ֶאל־יִ ְר ְמיָ הוּ * ֵמ ֵאת יְ הוָ ה‬
Dies ist die konstante Abfolge der Relativsatz-Variante der Wortereig­
nisformel in MT 39, während AlT in den genannten Fällen die beiden Prä­
positionalverbindungen in umgekehrter Reihenfolge bietet. Der Befund
lässt zwei Deutungen zu, zwischen denen nicht leicht zu ent­scheiden ist:
Entweder hat man in AlT den Wortlaut mehrfach revidiert, um den Vor­
rang Jhwhs vor dem Propheten auch sprachlich zu repräsentieren40, oder
sämtliche Belege wurden in MT nach diesem Muster vereinheitlicht.

36 Vonach, „Jeremias“ (Anm. 11), 2724, 2756, vertritt den umgekehrten Entwick­lungsweg,
lässt aber alle ge­nannten Gesichtspunkte außer Acht. Der hebräische Wortlaut von 10:16
bietet überdies keinerlei „kom­plexe und nur schwer ver­ständliche Formulie­rung“ (2756),
die den Übersetzer zu Vereinfachungen hätte ver­anlas­sen können.
37 Fischer und Vonach (Anm. 29) übersetzen σκελίζω mit wegschleudern. Der Fehler ist
korrigiert im Kommentar z. St. bei Vonach, „Jeremias“ (Anm. 11), 2756, entsprechend H. G.
Liddell und R. Scott, A Greek-English Lexi­con, Revised and Augmented throughout by H. St.
Jones (Oxford: Clarendon Press, 1940), s. v.: trip up one’s heels, upset; J. Lust, E. Eyni­kel und
K. Hauspie, A Greek-English Lexicon of the Septuagint, Revised Edition (Stuttgart: Deutsche
Bibel­ge­sellschaft, 2003), s. v.: to overthrow, to upset.
38 Gegen Vonach, „Jeremias“ (Anm. 11), 2726.
39 Ferner 7:1 > AlT; 30:1; 34:1, 8; 35:1.
40 So noch entschieden Stipp, Sondergut (Anm. 33), 152f.
gottesbildfragen in den lesartendifferenzen 249

11:18 ‫יתי‬
ִ ‫יתנִ י \ ָר ִא‬
ַ ‫ ָאז ִה ְר ִא‬c ‫ וָ ֵא ָד ָעה \ וְ ֵא ָד ָעה‬b ‫י\ענִ י‬
ֵ ִ‫יענ‬
ַ ‫הוֹד‬
ִ ‫ [וַ ]יהוָ ה‬a
ֶ ‫ַמ ַע ְל ֵל‬
κύριε γνώρισόν μοι καὶ γνώσομαι τότε εἶδον τὰ ἐπιτηδεύματα αὐτῶν
Laut MT berichtet Jeremia in 18ab von einem Erkenntnisprozess, den
Jhwh bei ihm bewirkte (18a ‫יענִ י‬ ַ ‫הוֹד‬
ִ er ließ mich erkennen), wobei aus­
drücklich der Erfolg konstatiert wird (18b ‫ וָ ֵא ָד ָעה‬und ich erkannte: Nar­
rativ der 1. Person mit Endung -a41). Der Gegenstand der Ein­sicht wird erst
in 18c offengelegt, wo sich der Prophet direkt an Jhwh wendet: Da hast
du mir ihre Untaten ge­zeigt. Redet somit in MT der dritte Satz Jhwh an,
ist es in AlT der erste: 18a, der dort einen Appell an Jhwh richtet, indem
die Verbal­form als Im­perativ interpretiert wird (‫יענִ י‬ ֵ ‫הוֹד‬
ִ lass mich erken-
nen) und der asyndetische Gottesname als Anrede fungiert. 18b be­nennt
als Finalsatz mit syndetischem Kohortativ (wֺ=’iqtul‑a: ‫ וְ ֵא ָד ָעה‬da­mit ich
erkenne)42 die erhoff­te Folge der erbetenen Erkennt­nis, während 18c
die Erfüllung wie MT in die Vergan­genheit verlegt, aber keine Mitwir-
kung Jhwhs thematisiert: Da sah ich ihre Untaten. Die prä­teritale Dei­xis
steht im Einklang mit der Fortsetzung V. 19–20, der zufolge Jere­mia die
Ma­chen­schaf­ten seiner Geg­ner ebenfalls schon kennt. In bei­den Lesarten
ist das Verhältnis von 18ab zu c problematisch: in MT wegen des Wech­
sels der Rederichtung, in AlT wegen der temporalen Inkonsis­tenz. Einen
glatten Text ergibt nur eine Mischung beider Fassungen: ‫יענִ י‬ ַ ‫הוֹד‬
ִ ‫יְ הוָ ה‬
‫יהם‬ֶ ‫יתי ַמ ַע ְל ֵל‬
ִ ‫ וָ ֵא ָד ָעה ָאז ָר ִא‬Jhwh ließ mich erkennen und ich erkannte;
da sah ich ihre Untaten. Dies dürfte den originalen Wort­laut dar­stellen,
den die alexandrinische Tradition bzw. der Übersetzer in 18ab über eine
Re­vokalisierung umdeutete, wahrscheinlich angestoßen durch die relativ
seltene Forma­tion wa=’iqtul‑a in 18b und ermöglicht durch die Asyndese
des Gottes­namens in 18a. Die Modifikation auf alexandri­nischer Seite
war also an­scheinend vor allem grammatischer Natur, indem man mehr-
deutige Verbalformen nach einem geläufigeren Muster reinterpretierte.
Da­ge­gen griff die masoreti­sche Überlie­fe­rung tiefer in den Textbestand
ein: Sie hob in 18c die Urheber­schaft Jhwhs hervor, in­dem sie dem Verb
eine kausative Form ver­lieh, es dem Sub­jekt Jhwh zuordnete und so das

41 Vgl. GKC 134 (§ 49e); P. Joüon und T. Muraoka, A Grammar of Biblical Hebrew. Second
Reprint of the Second Edition, with Corrections (SubBi 27, Roma: Gregorian & Biblical Press,
2009), 129 (§ 47d).
42 Vgl. E. Kuhr, Die Ausdrucksmittel der konjunktionslosen Hypotaxe in der ältesten
hebräischen Prosa. Ein Beitrag zur historischen Syntax des Hebräischen (Beiträge zur semi-
tischen Philologie und Linguistik 7; Leip­zig: Hinrichs, 1929) 49–52; Joüon und Mura­oka,
Grammar (Anm. 41) 596 (§ 168b).
250 hermann-josef stipp

Sehen Jere­mias in das Zeigen Jhwhs ver­wandelte. Auch hier fand somit
Arbeit am Gottes­konzept statt, aller­dings im maso­reti­schen Strang, wo
man die Souveräni­tät Gottes über das Gesche­hen er­weiter­te. Ein gleich­
artiger Vorgang mit Wechsel von G- zu H-Stamm und von mensch­lichem
zu göttlichem Sub­jekt hat in 18:2 und vielleicht auch 19:9 stattgefunden.
An­ders zu bewerten ist wohl 15:17 (s. u.).

13:12 ‫ ָכּל־נֵ ֶבל יִ ָמּ ֵלא יָ יִ ן‬c [‫ֹלהי יִ ְשׂ ָר ֵאל‬

ֵ ‫ֹה־א ַמר יְ הוָ ה ֱא‬
ָ ‫ ]כּ‬b
Die Botenformel 12b fehlt in AlT. Die Meinung, sie sei dort wegen des
folgenden Trink­spruchs 12c gestrichen worden43, ist allerdings an der Tat-
sache zu messen, dass man die Phra­se häufig in MT ergänzt hat: 11:22;
18:11; 22:30; 27:21; 29:25; 31:37; 35:19, wobei in 27:21; 29:25 und 35:19 überdies
idiolektale Gottesepitheta den sekundären Charakter der Belege unter­
mauern44. Danach erfolgte der Eingriff eher auf masore­tischer Sei­te, wo er
jedoch keinen spezifischen konzeptionellen Inten­tio­nen ent­sprang, son­
dern zur routine­mäßigen Expansion geprägter Ele­mente gehörte, wie sie
für diese Text­form kennzeichnend ist.

14:17 ‫( ֵעינַ י> ֶכם< ִדּ ְמ ָעה‬2. Pl f, Gram!) ‫ ֵתּ ַר ְדנָ ה‬b

καταγάγετε ἐπ᾿ ὀφθαλμοὺς ὑμῶν δάκρυα
Der Passus eröffnet laut dem Redebefehl 17a ‫ת־ה ָדּ ָבר ַהזֶּ ה‬ ַ ‫יהם ֶא‬ ֶ ‫וְ ָא ַמ ְר ָתּ ֲא ֵל‬
und dem Enkli­tikon in ‫ת־ע ִמּי‬ ַ ‫ ַבּ‬17d eine Gottesrede. Folglich spricht MT
wie in 9:9 (s. o.) die Vorstellung aus, dass Jhwh weint: Meine Augen fließen
über von Tränen. G* da­gegen bildet das Verb hier wie dort als Imperativ
Plural. Die Form καταγάγετε deutet die Präfixkonjugation als 2. Ps Pl f in
Funktion Aufforderung, und zwar irrigerweise45, da 17a auf die Ad­ressaten
im maskulinen Plural verweist (‫יהם‬ ֶ ‫ ) ֲא ֵל‬und somit ‫יכם‬
ֶ ֵ‫ ֵעינ‬mit maskulinem
Pronomen zu rekonstruieren ist; 17d wiederum imagi­niert dieselbe Größe
singularisch als ‫ת־ע ִמּי‬
ַ ‫ ַבּ‬. Überdies bekräf­tigt bei der paralle­len Verbalform
‫ ִתּ ְד ֶמינָ ה‬17c auch G* die Interpretation als 3. Person. Dem­nach gebührt
dem Be­zug auf die 3. Ps Pl f die Priorität. Hier wie in 9:9 scheint sich der
Schluss nahe­zu­legen, man habe in AlT die Vor­stel­lung vom Weinen Got­
tes umgehen wollen46. Doch wie dort bereits dargelegt, gab wahr­schein­

43 So Fischer, „Diskussion“ (Anm. 6), 84.

44 Vgl. ferner 17:5 sowie als Bestand­teil größerer Ein­schübe 29:16, 17; 33:17, 20, 25.
45 Außerdem wird ‫‑ירד‬G ein kausativer Sinn beigelegt. Zur Äquivalenz ‫‑ירד‬G ≙ κατ­
άγω vgl. ferner 9:17c; 13:17d.
46 Vgl. Anm. 34; Vonach, „Jeremias“ (Anm. 11), 2762.
gottesbildfragen in den lesartendifferenzen 251

lich weniger die Aversion gegen An­thropomorphismen den Aus­schlag als

vielmehr der Wunsch, die Judäer ihre Notlage selbst beweinen zu lassen
(s. zu 9:9). Und wie dort bleibt unentscheidbar, ob die theologischen Kor-
rekturen schon in der he­bräi­schen Phase der alex­an­drinischen Text­tradi­
tion eintraten oder erst vom Übersetzer vorge­nom­men wur­den.

15:7 (ἠτεκνώθην) ‫ ִשׁ ַכּ ְל ִתּי \ ֻשׁ ַכּ ְל ִתּי‬b

Der Satz fällt innerhalb einer Gottesrede, in der Jhwh auf vergangenes
Strafhandeln an sei­nem Volk zurückschaut (6c–7c). Die Retrospektive
steht in TT durchgehend im Aktiv, wäh­rend G* in 7b eine passivische
Vo­kalisation voraussetzt. Diese Deutung des Konsonanten­textes ist in
einem einzelnen Punkt weniger hart, doch kann angesichts der Aussagen
im Kon­text von einer Milde­rung des Gottesbildes keine Rede sein. Zugleich
ist die G*-Lesart sogar anthro­pomorpher47. Wie immer sie sich zu TT ver-
halten mag, belegt sie jedenfalls keine Scheu von JerAlT oder JerG*, die
Gottesvorstellung mit anthropomorphen Metaphern zu artiku­lie­ren.

15:17 ‫אתי‬ִ ‫אתנִ י \ ָמ ֵל‬

ָ ‫ִכּי־זַ ַעם ִמ ֵלּ‬
Innerhalb der Konfession 15:10–21 schildert Jeremia mit 15:17d seinen
bedauernswerten Zustand. Dabei erhebt er in MT einen scharfen Vorwurf
an die Adresse Jhwhs, denn er macht ihn für seine Ge­mütsverfassung
verantwortlich, während AlT diese nur konstatiert. MT kann damit die
göttliche Lenkung des Ge­schehens herausstellen, aber AlT kann auch
die Anklage eliminiert haben, die Jhwh einen har­ten Umgang mit dem
Propheten vorhält. Angesichts der gering­fügi­gen Diffe­renz im Schrift­
bild kommt überdies ein bloßes Schreiberversehen—in beiden Rich­
tungen—in Be­tracht. Der analoge Befund in 1:17 spricht allerdings für eine
Modi­fikation in AlT. Vgl. auch den folgenden Fall 15:18.

15:18 ‫( ִלי ְכּמוֹ ַא ְכזָ ב ַמיִ ם‬Aor 3. f ) ‫  ָהיוֹ ִת ְהיֶ ה‬. . . ‫נוּשׁה‬
ָ ‫וּמ ָכּ ִתי ֲא‬
In dem Klagegebet Jeremias ist die Verbalgruppe ‫ ָהיוֹ ִת ְהיֶ ה‬auf der sprach-
lichen Oberfläche doppeldeutig, da sie einerseits für die 2. Ps Sg m eintre-
ten kann und somit Jhwh in einer besonders scharfen Anklage mit einem
Trugbach vergleicht. Andererseits kann die Verbal­gruppe die 3. Ps Sg f
bezeichnen; so würde sie allerdings auf ‫ ַמ ָכּ ִתי‬als ihr Subjekt verweisen

47 Die Variante ἠτεκνώθησαν hat Ziegler, Jeremias (Anm. 29), dagegen zurecht als Glät-
tung in den Apparat ver­wiesen; gegen Vonach, „Jeremias“ (Anm. 11), 2763.
252 hermann-josef stipp

und meine Wunde höchst implausibel mit einen Trugbach gleich­setzen48.

G* wählt trotzdem die zweite Möglichkeit, sodass der direkt an Jhwh
gerichtete Vorwurf entfällt (vgl. auch oben zu 1:17; 15:17)49. Unentscheid-
bar bleibt, ob der Über­setzer selbst da­mit den Text entschärfte oder bloß
eine Leseweise applizierte, die in jenen jüdischen Kreisen üblich war, in
denen er die Traditionen seines Glau­bens pflegte. Denn schließlich sind
die Väter der Septua­ginta ihrer Vorlage kaum erst im Zuge der Überset­
zung begegnet; weit mehr Wahrscheinlichkeit hat es für sich, dass sie ihre
heiligen Schriften schon vorweg in ihrer heimi­schen Gemeinde studiert
und herkömmliche Leseweisen des mehrdeuti­gen Konsonantentextes
ken­nen gelernt hatten50.

16:13 ‫יִתּנוּ‬
ְ \ ‫א־א ֵתּן‬
ֶ ֹ ‫[יוֹמם וָ ַל ָיְלה] ֲא ֶשׁר ל‬
ָ ‫ֹלהים ֲא ֵח ִרים‬
ִ ‫ת־א‬
ֱ ‫ם־שׁם ֶא‬
ָ ‫וַ ֲע ַב ְד ֶתּ‬
‫ָל ֶכם ֲחנִ ינָ ה‬
Weil dem ‫ ֲא ֶשׁר‬-Satz in MT ein Beziehungswort fehlt, gehört die Kon­
junktion dort zu den we­nigen Fällen, wo sie eine kausale Funktion aus­
übt51. AlT dagegen liest ein verbales Prädikat in der 3. Ps Pl und somit
ei­nen Relativsatz, der auf ‫ֹלהים ֲא ֵח ִרים‬
ִ ‫ ֱא‬bezogen ist. Der geringe Auf­
wand, mit dem sich der ‫ ֲא ֶשׁר‬-Satz der üblichen relativen Funktion zuord-
nen lässt, spricht ge­gen eine se­kun­däre Glättung. Vielmehr dürfte AlT
den originalen Stand repräsentieren, wäh­rend man in der masoretischen
Tradition der Vorstellung entgegentrat, dass Fremdgötter den judäischen
Exi­lanten Erbarmen gewähren oder verweigern könnten bzw. überhaupt
aktions­fä­hig wären52. Folglich erwuchs die masore­tische Lesart aus einer
theologischen Korrektur mit dem Ziel, die Souveränität Jhwhs zu sichern,
selbst um den Preis eines grausameren Gottes­bildes.

48 Der zitierte Passus enthält weitere Divergenzen, die sich aber nur auf Formulie­
rungsvarianten beziehen und daher hier ausgeklammert werden.
49 Vgl. Vonach, „Jeremias“ (Anm. 11), 2725, 2764.
50 Vgl. z. B. Siegert, Zwischen Hebräischer Bibel und Altem Testament (Anm. 17), 123–125;
H.-J. Stipp, „Be­mer­kungen zum griechischen Michabuch aus Anlass des deut­schen LXX-
Übersetzungspro­jekts,“ JNSL 29 (2003): 103–132 (120–122); G. D. Martin, Multiple Originals.
New Approaches to Hebrew Bible Textual Cri­ticism (SBL Text-Critic­al Studies 7; Leiden: Brill,
2011), 82f.
51 HAL 95b; vgl. Ges18 112a.
52 So z. B. W. L. Holladay, Jeremiah 1. A Commentary on the Book of the Prophet Jere­miah
Chapters 1–25 (Her­meneia; Philadelphia: Fortress Press, 1986), 474. Nach Vo­nach, „Jere-
mias“ (Anm. 11), 2765, hingegen „nimmt die LXX der Darstellung des Gottes Israels einiges
an Schärfe.“ Dann müsste der Übersetzer die polytheisti­schen Implika­tio­nen seiner Wie-
dergabe freiwillig in Kauf genommen haben.
gottesbildfragen in den lesartendifferenzen 253

16:15 ‫וּמכֹּל ָה ֲא ָרצוֹת‬

ִ ‫ת־בּנֵ י יִ ְשׂ ָר ֵאל ֵמ ֶא ֶרץ ָצפוֹן‬
ְ ‫ַחי־יְ הוָ ה ֲא ֶשׁר ֶה ֱע ָלה ֶא‬
‫( ָשׁ ָמּה‬ἐξώσθησαν) ‫יחם‬ ָ ‫ֲא ֶשׁר ִה ִדּ‬
Laut MT wird Jhwh die Israeliten aus aller Welt versammeln, wohin er
sie zuvor versprengt hat. Der kausativen Verbalform ‫יחם‬ ָ ‫ ִה ִדּ‬korres­pon­
diert die passivische Wieder­gabe ἐξώσθησαν, die den Akteur ver­schweigt.
Weil JerG* die Genera verbi großzügig behandelt, ist die Ent­scheidung
erschwert, ob das Passiv schon der Vorlage eignete (‫ )נִ ְדּחוּ‬oder auf den
Übersetzer zu­rückgeht, der Jhwh absichtlich oder bloß faktisch von der
Verantwortung für die Diaspora ent­band. Daher sind die Konse­quenzen
mehrerer Alternativen zu be­denken.
Wann immer das Passiv in den Text geriet, ist es jedenfalls kaum aus
einer theologisch moti­vierten Glät­tung erwachsen. 16:15 gehört in eine
Reihe mit 12 weiteren geprägten Wen­dungen in Jer, die ein topi­sches Por-
trät der Dia­sporasituation entwerfen, indem sie das Leben in der Fremde
durch die Verben ‫ נדח‬und ‫ פוץ‬auf Zerstreuung zurückführen und deren
Reich­weite mittels der Stichworte ‫ ֲא ָרצוֹת‬, ‫ גּוֹיִ ם‬und ‫ ְמקֹמוֹת‬weltweit ent-
grenzen53. Diese Passagen folgen weit über­wie­gend dem Muster von 16:15
MT, insofern sie ihre verbalen Prädi­kate als H-Stamm formen und Jhwh
die Subjektsrolle zuweisen. Dies gilt für 7 Fälle in dem Bestand, den MT
mit AlT gemeinsam hat; dort bekennt sich auch JerG* zur Urheberschaft
Jhwhs und gibt somit keinen Anlass zur Frage nach ab­weichen­den Vorla-
gen oder Eingriffen des Überset­zers54. Hinzu kommen 3 weitere Belege in
masoretischen Über­schüs­sen, die, wie idiolektale Komponen­ten erweisen,
nicht in AlT ausgeschieden wurden, sondern erst nach der Gabelung der
Texttra­di­tionen in MT eingedrungen sind55. Daneben existie­ren nur zwei
passi­vische Fälle, die ebenfalls dem ma­sore­tischen Sondergut angehö­ren,
eine Art der Bezeugung, die sie selbst zu einem Merk­mal des prämaso-
retischen Idiolekts erhebt (40:12; 43:5)56. Akteurs­blinde Varian­ten der
fraglichen geprägten Wendung treten also erst in den jüngsten Zutaten
zur masoreti­schen Ausgabe des Buches auf, während die alexandrinische
Edition gerade die pas­sivischen For­mu­lie­rungen vermissen lässt, die theo-
logisch die geringsten Probleme berei­tet haben müssten.

53 Vgl. H.-J. Stipp, Deuterojeremianische Konkordanz (ATSAT 63; St. Ottilien: Eos, 1998),
54 8:3; 9:15; 23:3, 8; 24:9; 32:37; 46:28 (|| 30:11 MT). In 23:8 MT wird Jhwh sekun­där in 1.
Ps repräsentiert.
55 29:14, 18; 30:11 (|| 46:28). Die idiolektalen Züge sind verzeichnet bei Stipp, „Up­dated
Index“ (Anm. 4), 192f.
56 Vgl. Stipp, „Updated Index“ (Anm. 4), 193f. Sollte in 16:15 das Passiv in der Vor­lage
wurzeln, heben sich 40:12; 43:5 immerhin durch die Form ‫ ָשׁם‬statt ‫ ָשׁ ָמּה‬ab.
254 hermann-josef stipp

Das Bild wird bekräftigt durch nicht-formelhaftes Vergleichsmaterial,

das ge­eigne­te Verben mit dem Subjekt Jhwh und dem Objekt Israel bzw.
Juda verbindet. G* gibt auch solche Passagen unge­schönt wie­der: Vgl. mit
‫‑פוץ‬H 13:24; 18:17; mit ‫ זרה‬15:7; 31:10; mit ‫‑גלה‬H 22:12 (wo sogar allein AlT
die Deportation in die Hände Jhwhs verlegt, wes­wegen der Passus weiter
unten zu erörtern ist); 29:4, 7. Der Beleg mit ‫‑גלה‬H in 29:14 fällt in einen
ma­soretischen Überschuss, den idiolektale Elemente als späte Interpola-
tion markieren57. Nä­herer Analyse bedarf lediglich die Sachlage bei ‫‑נדח‬H
in 27:15, wo AlT einzig den pronomi­nalen Bezug auf Jhwh übergeht; siehe
dazu unten.
So ergibt sich: Bei 16:15 sind keine Indizien zu erkennen, die ein Ur­teil
gestatten, ob das Passiv bereits in der JerG*-Vorlage stand oder erst vom
Übersetzer geschaffen wurde. Jedenfalls zeigt sich die alexan­drini­sche
Tradition weder in ihrer hebräischen noch in ihrer grie­chischen Pha­
se geneigt, Jhwhs Regie hinter der Ent­wurzelung der Isra­eliten zu ver­
schlei­ern. Der masoretische Strang hat erst auf seiner jüngsten Stufe zwei
akteursblinde For­mu­lierungen hervorgebracht, allerdings neben weiteren
Belegen herkömmlichen Typs, die wie gehabt Jhwh als Verursacher iden-
tifizieren. Sofern sich die diskutierte Lesartendifferenz in 16:15 über­haupt
mit in den Text­formen wirksamen Trends verbinden lässt, ergeben sich
zwei Möglich­keiten, je nach dem, wo die passive Variante erstmals auftrat.
War sie schon in der JerG*-Vorlage enthalten, bildete sie dort kaum einen
sekundären Zug, sondern darf eher die Priori­tät beanspruchen, während
der maso­retische Kausativ nachträglich aus der Adaption an einen domi-
nanten Sprachgebrauch erwuchs und/oder dem wieder­holt beobachtba-
ren Bedürf­nis nachgab, die Steuerung der Ge­schichte ausdrücklich in die
Hände Jhwhs zu ver­legen. Hat erst der Übersetzer das Passiv eingeführt,
wollte er ebensowenig Jhwh ein freundliche­res Gesicht verleihen, son­
dern er pendelte spontan zwi­schen den Genera verbi, wie er es häufig tat,
ohne dass er dazu spezieller Gründe bedurfte. Welche Alternative auch
immer den Vorzug ver­dient, gibt der Befund jedenfalls kein Recht, der
alexandrinischen Texttradi­tion bzw. dem Übersetzer einen Hang zur Mil-
derung der Gottesbilder zu be­scheinigen. Dafür gibt es zu viele Gegenbei-
spiele58. Für den Übersetzer ließe sich dieses Fazit durch weitere Beispiele
aus JE12 erhärten59.

57 S. Anm. 55.

58 Gegen Vonach, „Jeremias“ (Anm. 11), 2765.
59 Vgl. namentlich die bei Stipp, Konkordanz (Anm. 53), 88f., gesammelten Belege aus
gottesbildfragen in den lesartendifferenzen 255

18:2 ‫ת־דּ ָב ָרי‬
ְ ‫יעָך \ ִת ְשׁ ַמע ֶא‬
ֲ ‫יּוֹצר וְ ָשׁ ָמּה ַא ְשׁ ִמ‬
ֵ ‫קוּם וְ יָ ַר ְד ָתּ ֵבּית ַה‬
Anstelle des Kausativstamms in MT setzt AlT den Grundstamm voraus
(ἀκούσῃ). Plau­sibel ist allein der Schritt von der alexandrinischen zur
masoretischen Lesart, die das Bestreben verrät, die Kontrolle Jhwhs über
das Geschehen herauszustreichen. Nach gleichem Muster und mit iden­
tischer Absicht hat man in MT 11:18 sowie vielleicht auch 19:9 Stämme und
Subjekte ausgetauscht.

19:9 ‫יהם‬ ֶ ‫יהם וְ ֵאת ְבּ ַשׂר ְבּנ ֵֹת‬ ֶ ֵ‫ת־בּ ַשׂר ְבּנ‬
ְ ‫ וְ ַה ֲא ַכ ְל ִתּים \ וְ ָא ְכלוּ ֶא‬a
ֶ ‫ ֲא ֶשׁר ִיָציקוּ ָל ֶהם אֹיְ ֵב‬c ‫וּב ָמצוֹק‬ְ ‫אכלוּ ְבּ ָמצוֹר‬ֵ ֹ ‫ר־ר ֵעהוּ י‬ ֵ ‫ וְ ִאישׁ ְבּ ַשׂ‬b
Dies ist eine typische Konstellation, die die Frage nach sekundärer Auf-
hellung des Gottes­bil­des in der alexandrinischen Überlieferung anregt:
Hat man dort die grauenvolle Vorstellung ausgelöscht, dass Jhwh wün-
schen könne, die Angehörigen seines Volkes zum Kanniba­lismus an ihren
eigenen Kindern zu nötigen (‫אכל‬-H)60? Allerdings bezeu­gen in 9b beide
Textformen über­ein­stim­mend den Grundstamm von ‫אכל‬, und die alex­
andrinische Variante von 9a ent­spricht der Parallele Dtn 28:53 ‫וְ ָא ַכ ְל ָּת‬
‫ּוב ָמֹצוק‬ְ ‫ֹלהיָך ְּב ָמֹצור‬ ֶ ‫ן־לָך יְ הוָ ה ֱא‬
ְ ‫ּובנ ֶֹתיָך ֲא ֶׁשר נָ ַת‬
ְ ‫י־ב ְטנְ ָך ְּב ַׂשר ָּבנֶ יָך‬
ִ ‫ְפ ִר‬
‫ ; ֲא ֶׁשר־יָ ִציק ְלָך אֹיְ ֶבָך‬vgl. ferner die ebenfalls nicht-kausativen Belege Lev
26:29 und Bar* (hebräisch) 2:361. Der Befund lässt daher entgegengesetzte
Deutun­gen zu: Entweder hat man 19a in AlT theologisch geglättet und/
oder an 19b samt der Parallelen angeglichen, oder der deu­te­rono­mistische
Autor von 19:1–1362 benutzte von vornherein die ge­prägte Wen­dung mit
‫‑אכל‬G, die in der masoretischen Tradition im Interesse eines Macht­er­
weises Jhwhs kau­sa­tiv abgewandelt wurde. Weil MT gehäuft im Interes­se
der Sou­veränität Jhwhs modifiziert und dazu wiederholt der G-Stamm mit
menschlichem Subjekt gegen den H-Stamm mit göttlichem Verursa­cher
ausgewechselt wurde (11:18; 18:2), hat der Verdacht der ale­xandri­nischen
Priorität gute Gründe auf seiner Seite, ohne dass die wünschens­werte
Sicher­heit zu gewinnen wäre. Jeden­falls ist der Befund nicht ohne die
genannten Paral­le­len sachge­recht zu beurteilen63.

60 So G. Fischer und A. Vonach, „Tendencies in the LXX Version of Jeremiah,“ in

G. Fischer, Der Prophet wie Mose (Anm. 6), 64–72 (69).
61 Diese und entferntere Parallelen sind zusammengestellt bei Stipp, Konkordanz
(Anm. 53), 16.
62 Vgl. zur Verfasserfrage H.-J. Stipp, „Jeremia und der Priester Paschhur ben Immer.
Eine redak­tions­ge­schichtliche Studie,“ in Kulte, Priester, Rituale—Beiträge zu Kult und
Kult­kritik im Alten Testament und Alten Orient (FS  Th. Seidl; Hg. St. Ernst und M. Häusl;
ATSAT 89; St. Ottilien: Eos, 2010), 375–401.
63 Gegen Fischer und Vonach (Anm. 60).
256 hermann-josef stipp

20:16 <‫ר־ה ַפְך יְ הוָ ה > ְבּ ַאף‬

ָ ‫וְ ָהיָ ה \ יְ ִהי ָה ִאישׁ ַההוּא ֶכּ ָע ִרים ֲא ֶשׁ‬
Die wohl von Dtn 29:22 (weniger wahrscheinlich: Ijob 9:5) inspirierte
Glosse belegt, dass man auch in der alex­andrinischen Tradition bereit
war, Jhwh zusätzlich mit negativen Lei­denschaften zu asso­ziie­ren.

21:7 ‫ת־ע ָב ָדיו וְ ֶאת־‬

ֲ ‫הוּדה וְ ֶא‬
ָ ְ‫ת־צ ְד ִקיָּ הוּ ֶמ ֶלְך־י‬
ִ ‫י־כן נְ ֻאם־יְ הוָ ה ֶא ֵתּן ֶא‬
ֵ ‫ וְ ַא ֲח ֵר‬a
. . . ‫ָה ָעם‬
‫וּביַ ד[ ְמ ַב ְק ֵשׁי נַ ְפ ָשׁם‬
ְ ] ‫יהם‬ֶ ‫ְך־בּ ֶבל וּ[ ְביַ ד אֹיְ ֵב‬
ָ ‫אצּר ֶמ ֶל‬ ַ ‫בוּכ ְד ֶר‬
ַ ְ‫] ְבּיַ ד נ‬
‫יהם‬ ֶ ‫ לֹא־יָ חוּס \ ָאחוּס ֲע ֵל‬c ‫י־ח ֶרב‬ ֶ ‫ וְ ִה ָכּם \ וְ ִה ֻכּם ְל ִפ‬b
‫ וְ לֹא יְ ַר ֵחם \ ֲא ַר ֵחם‬e [‫ ]וְ לֹא יַ ְחמֹל‬d
In AlT 7a kün­digt Jhwh an, König Zidkija und seine Untertanen ihren
Feinden auszuliefern, die zuvor in V. 2, 4 als die babylonischen Belage­rer
Jerusalems identifiziert wurden. Anschlie­ßend sollen die Sieger laut 7b ein
Blutbad anrichten, das Jhwh in 7ce als Aus­druck seiner Gnaden­losigkeit
deutet (1. Ps Sg). In MT sind die Gegner in einer Reihe von Nomi­nal­
gruppen aus­differenziert, angeführt vom babylonischen König, der zwar
in beiden Fassungen der Ein­heit 21:1–10 erwähnt wird (V. 2), aber nur in
der masoretischen Aus­gabe einen Namen trägt (ferner V. 2 MT)64. Das
Mas­saker gilt in 7b TT als Tat Nebukad­nezzars und in 7c–e MT als Folge
von dessen Mangel an Erbar­men (3. Ps Sg). Dabei bezeugt MT mit 7d die
zusätzliche Aussage ‫וְ לֹא יַ ְחמֹל‬, im Einklang mit der Paral­lele 13:14, wo die
Trias kein Mitleid haben—nicht schonen—sich nicht erbarmen allerdings
wie in 21:7 AlT Eigen­schaften Jhwhs aufreiht. Der Wandel von MT zu AlT,
mithin die Umwidmung der Passage von Nebu­kadnezzar auf die Baby-
lonier (7b) und Jhwh (7ce) ist unwahr­scheinlich, denn dann wäre die
Strei­chung von Namen und Titel aus 7a kaum zu motivieren (die nicht
jünger sein können als die 3. Person in 7b–e, der sonst die Bezugsgröße
fehlt). Folglich wurde 21:7b–d umgekehrt auf Nebukadnezzar umdirigiert,
möglicherweise gefördert durch die Defek­tiv­schreibung von ‫וְ ִה ֻכּם‬65. Als
Konse­quenz ging die Mitleidlosigkeit von Jhwh auf das menschliche
Strafwerk­zeug über. Diese Annahme passt zu 13:14 und dem Umstand,
dass die maso­reti­schen Überschüsse an Namen und Titeln den baby­lo­ni­
schen Großkönig ohnehin besonders freizügig be­denken66, darunter nicht
selten an Stellen, die sich aus anderen Grün­den als se­kundär erwei­sen.

64 In V. 4 erwähnt ihn MT zusätzlich mit seinem Titel.

65 Theoretisch kann der Vorgang mehrstufig verlaufen sein: Erst wurde 7b singula­risch
gelesen, dann hat man 7c(d)e auf die 3. Ps umgestellt.
66 Vgl. die Zusammenstellung bei Janzen, Studies in the Text of Jeremiah (Anm. 2),
gottesbildfragen in den lesartendifferenzen 257

Demnach wurde hier zwar das Gottes­konzept abgemildert, aber erstens

nicht im alexan­drinischen, sondern im masore­tischen Strang. Zweitens
hat der Ein­griff allem Anschein nach gar nicht die Auf­hellung des Gottes­
bilds bezweckt, denn dafür hätte die Um­wid­mung auf die babylo­nischen
Widersacher ausgereicht, indem man 7c–e in die 3. Person Plural ver-
setzte. Vielmehr ging es dem Revisor vor allem um die Einführung Nebu-
kadnezzars, mit dem sanfte­ren Profil Jhwhs als Nebeneffekt. Wer dagegen
MT die Priorität zuspricht, muss statt­dessen eine Ver­dunkelung des Got-
tesbildes in AlT annehmen67.

22:12 ‫ר־הגְ לוּ \ ִהגְ ֵל ִתי אֹתוֹ ָשׁם יָ מוּת‬

ִ ‫<אם> ִבּ ְמקוֹם ֲא ֶשׁ‬
ִ ‫ִכּי‬
Laut MT wurde Joahas bzw. Schallum von Menschen unwiderruflich
deportiert, während dies nach AlT ein Werk Jhwhs gewesen ist. Die
Priori­tätsfrage ist schwer entscheidbar, doch in Frage kom­men nur eine
Auf­hellung des Gottesbildes in der masoretischen Tradition oder eine
Ver­schärfung im alexandrinischen Strang, um hinter dem Vorgang die
Fügung Gottes aufzu­weisen. Auch ein schlichter Schreibfehler ist nicht
auszu­schließen, der in beiden Richtungen eingetreten sein kann: entwe­
der durch Ausfall des ‫ ת‬mit anschließender Korrektur ‫ ו ⇒ י‬in MT oder
umgekehrt durch Verschreibung ‫ י ⇒ ו‬und Ergänzung des ‫ת‬. Die Hypo­
these von Glät­tungstendenzen in JerG* findet jedenfalls erneut keine

23:23 ‫ֹלהי ֵמ ָרחֹק‬
ֵ ‫ֹלהי ִמ ָקּר ֹב ָאנִ י נְ ֻאם־יְ הוָ ה וְ לֹא ֱא‬
ֵ ‫[א‬ֱ ‫]ה‬ ַ
In MT ist V. 23 eine rhetorische Frage, in AlT hingegen eine Feststel­lung;
infolgedessen dekla­riert MT Jhwh als einen „Gott von ferne,“ während
AlT ihn als „Gott aus der Nähe“ be­schreibt. Die Konsequenz ist laut V.
24 in beiden Texttypen gleich: ‫א־א ְר ֶאנּוּ‬ ֶ ֹ ‫ ִאם־יִ ָסּ ֵתר ִאישׁ ַבּ ִמּ ְס ָתּ ִרים וַ ֲאנִ י ל‬.
24b ist aufgrund der negierten Apodosis nur als rhetorische Frage sinn-
voll, die einschärft: Kein Versteck vermag vor Jhwhs Blick zu schützen.
Der fragende Charakter von 24b spricht dafür, dass diese Eigenart auch
vorweg formal markiert wurde, und damit zugunsten der masoretischen
Fragepartikel, die eine disjunktive Alternativfrage . . . ‫  ִאם‬. . . ‫ ֲה‬erzeugt68.
Zudem ist die Streichung der Fragepartikel in AlT gut begründbar. In MT
verbürgt gerade die Ferne und damit der geweitete Blickwinkel Gottes seine

67 Diese größere Härte des Gottesbildes in JerG* wird von Vonach, „Jeremias“ (Anm.
11), 2771, folgenlos notiert.
68 Vgl. HAL 59a; Ges18 70b.
258 hermann-josef stipp

Allwissenheit. Diese Logik hat man in der alexandrinischen Texttradition

offenbar nicht mehr ver­standen; deshalb wandelte man die Frage in eine
Feststellung um, sodass nun umgekehrt die Nähe seine Un­entrinnbarkeit
garantiert. Damit ändert sich nicht der maßgebliche Zug des Gottesbildes,
son­dern nur der gedankliche Anweg dort­hin. Mittels einer anderen Logik
hat AlT den Sinn des Passus gewahrt69.

25:6 ‫ ָל ֶכם‬# ‫ § וְ לֹא ָא ַרע \ ְל ָה ַרע‬c ‫יכם‬

ֶ ‫אוֹתי ִבּ ַמ ֲע ֵשׂה יְ ֵד‬
ִ ‫א־ת ְכ ִעיסוּ‬
ַ ֹ ‫ וְ ל‬b
MT bietet zwei Sätze: In 6b warnt Jhwh davor, ihn zu kränken; für den
Fall, dass sein Ruf Gehör findet, verspricht er den Judäern in 6c, ihnen
nichts Böses zuzufügen. Anstelle von 6c kennt AlT nur einen Infinitiv­
ausdruck, der die schlimmen Konsequenzen als selbst verursach­te Frucht
der Beleidigung Jhwhs stilisiert. Keine Fassung schreibt Jhwh schmerz­
hafte Taten an Juda zu; der masoretische Satz 6c mit dem Subjekt Jhwh
dient ja dem Zweck, für den Fall des Ge­horsams seinen Verzicht auf Straf-
handlungen zu erklären. Die alexandrinische Lesart kann also nicht dem
Wunsch entspringen, das Gottesbild aufzuhellen. Für die Bewertung ist
zu be­ach­ten: Die masoretische Variante ‫ ָא ַרע‬ist in Jer singulär, während
der alexandrinische In­fini­tiv ‫ ְל ָה ַרע‬mehrere Parallelen besitzt (25:29; 4:22;
31:28), ein ambivalenter Tatbestand, da er sowohl einen äl­teren Wortlaut
als auch Adaption an eine gebräuchlichere Formulierung anzeigen kann.
Sollte die masoreti­sche Version sekundär sein, dürfte sie aus dem Motiv
her­rühren, das Geschehen der Lenkung durch Jhwh zu unterstellen.

25:30 ‫ ָ שׁאֹג יִ ְשׁ ַאג ַעל־נָ וֵ הוּ ֵה ָידד ְכּד ְֹר ִכים יַ ֲענֶ ה‬. . . ‫יְ הוָ ה ִמ ָמּרוֹם יִ ְשׁ ָאג‬
JerG* 32:16 κύριος ἀφ᾿ ὑψηλοῦ χρηματιεῖ . . . λόγον χρηματιεῖ ἐπὶ τοῦ
τόπου αὐτοῦ καὶ αιδεδ ὥσπερ τρυγῶντες ἀποκριθήσονται
Der Vers fällt auf durch die ungewöhnliche, doppelte Gleichung ‫≙ שׁאג‬
χρηματίζω. Das Stan­dardäquivalent zu ‫ שׁאג‬brüllen lautet ἐρεύγομαι brül-
len; JE12 kennt darüber hinaus noch ὠρύομαι mit derselben Bedeutung70.
Alle diese Verben bezeichnen vor­wiegend Rufe von Raubtie­ren. Einzig
im zitierten Passus wird ‫ שׁאג‬mit χρηματίζω wiedergegeben. G benutzt

69 Nach der Erklärung von Vonach, „Jeremias“ (Anm. 11), 2775, wird in den beiden
Fassungen „inhaltlich das direkte Gegenteil ausgesagt . . . Drückt der Satz im Hebräischen
nämlich v.a. die Ferne Gottes aus, so wird in der LXX gerade seine Nähe betont.“
70 Vgl. E. Hatch und H. A. Redpath, A Concordance to the Septuagint and the Other Greek
Versions of the Old Testament, 2. Aufl. (Grand Rapids: Baker, 1998), s. v.; E. C. dos Santos, An
Expanded Hebrew Index for the Hatch-Redpath Concordance to the Sep­tuagint (Jerusa­lem:
Dugith, o. J. [1973]), s. v.
gottesbildfragen in den lesartendifferenzen 259

das Verb in seiner Grundbedeu­tung handeln, beschäftigt sein in 1 Kön/

3 Bas 18:27 und Ijob 40:8; dane­ben kommt es nur und gehäuft in JerG*
32–43 (MT 25–36) vor. Es bildet also ein Spezifikum der zweiten Hälfte
des griechischen Je­remiabuches, die sich durch eine Anzahl von Über­set­
zungskonventionen von der ersten Hälfte abhebt71. Dabei reprä­sentiert
χρηματίζω, von 25:30/32:16 abge­sehen, sonst immer ‫דבר‬-D, und zwar stets
bezo­gen auf Reden Jhwhs (30:2; 36:2, ‌4) oder eines Pro­pheten (26:2cd),
der auch ein Falsch­prophet sein kann (29:23). In JerG* hat χρηματίζω
daher die Spezialbedeutung (prophe­tisch) reden ange­nommen72, die im
NT fort­wirkt73.
In 25:30/32:16 charakterisiert das Verb ebenfalls Kundgaben Jhwhs.
Der Vers besitzt zwei Parallelen mit ‫ שׁאג‬und Subjekt Jhwh, wo­bei G*
gleichermaßen die Standardäquivalente vermeidet: In Joël 4:16 wird ‫שׁאג‬
mit ἀνακράζω aufschreien und in Am 1:2 mit φθέγγομαι äu­ßern, reden
wieder­gege­ben. Die Fälle ste­hen in JE12, deren grie­chische Ausgaben ‫שׁאג‬
in andersartigen Kon­tex­ten durchaus wörtlich reprä­sentie­ren (ἐρεύγομαι
Hos 11:10; Am 3:4, 8; ὠρύομαι Jer 2:15; Ez 22:25; Hos 11:10; Zef 3:3). Diese Pra-
xis legt den Schluss nahe, dass der JE12–Über­setzer ‫ שׁאג‬in unverfängli-
chen Zusam­men­hängen wörtlich wiedergab, aber das Subjekt Jhwh nicht
mit Verben verbinden wollte, die sonst für das Ge­brüll von Tieren eintre­
ten. Aus ähnlichen Pietätsgründen dürfte er für ‫ יַ ֲענֶ ה‬den Plural gewählt
und so den ‫ ֵה ָידר‬-Ruf auf (unge­nannte) menschliche Akteure ver­schoben

25:31 ‫ ָה ְר ָשׁ ִעים נְ ָתנָ ם \ נִ ְתּנוּ‬c ‫ל־בּ ָשׂר‬

ָ ‫ נִ ְשׁ ָפּט הוּא ְל ָכ‬b ‫ ִכּי ִריב ַליהוָ ה ַבּגּוֹיִם‬a
‫ַל ֶח ֶרב‬
Laut MT liefert Jhwh eigenhändig die Frevler dem Schwert aus; AlT 32:17
dagegen gebraucht das Passiv, wobei indes der Vortext keinen Zweifel
zulässt, wer die Fäden in der Hand hält. Da Jhwhs Gerichts­handeln den

71 Vgl. E. Tov, The Septuagint Translation of Jeremiah and Baruch. A Discussion of an

Early Revision of the LXX of Jeremiah 29–52 and Baruch 1:1–3:8 (HSM 8; Missoula: Scholars
Press, 1976); speziell zu χρηματίζω S. 71. Zu den Konsequenzen der Zweitei­lung vgl. zuletzt
T. S. L. Michael, „Bisectioning of Greek Jeremiah: A Problem to Be Revisited?“ BIOSCS 39
(2006): 93–104; A. Pietersma, “An Excursus on Bisectioning Iere­mi­as” (2007): http://ccat
72 Lust, Eynikel und Hauspie, Lexicon (Anm. 37), s. v.
73 W. Bauer, Griechisch-deutsches Wörterbuch zu den Schriften des Neuen Testa­ments
und der frühchrist­lichen Literatur, 6., völlig neu bearb. Aufl. hg. von K. Aland und B. Aland
(Berlin: De Gruyter, 1988), s. v.
74 In diesem Fall ist daher der Meinung von Fischer, „Diskussion“ (Anm. 6), 85, und
Vonach, „Jeremias“ (Anm. 11), 2791, beizupflichten.
260 hermann-josef stipp

Übeltätern gilt und obendrein primär die Fremd­völker heimsucht, lud es

nicht dazu ein, dem Porträt Jhwhs freundlichere Züge zu verlei­hen. MT
weist mit der Pendenskonstruktion . . . ‫ ָה ְר ָשׁ ִעים נְ ָתנָ ם‬die syntaktisch kom­
plexere Struktur auf, doch zeigt die alexandrini­sche Tra­di­tion generell
keine Neigung, dieses Bildemuster durch einfa­chere zu ersetzen. Daher
hat wahrscheinlich AlT den originalen Wortlaut bewahrt, während MT
wieder die gött­liche Regie ins Licht rückt.

26:3 und öfter: ‫נחם‬-N mit göttlichem Subjekt ≙ (ἀνα)παύομαι

In der ersten Hälfte von JerG* 75 wird ‫נחם‬-N bereuen überwiegend mit
μετανοέω bereuen wie­der­gegeben, gleichgültig ob Menschen (8:6) oder
Jhwh (4:28; 18:8, 10) als Subjekt fungieren. Daneben wird in 20:16 bei
göttlichem Subjekt μεταμελέομαι bereuen, bedauern eingesetzt; ferner
bietet 15:6 in Gottesrede die Äquivalenz ‫יתי ִהנָּ ֵחם‬ ִ ‫לא‬
ֵ ִ‫ נ‬ich bin es müde,
Mitleid zu haben ≙ καὶ οὐκέτι ἀνήσω αὐτούς und ich werde sie nicht mehr
freilassen, wo das Schrift­bild des Infinitivs anscheinend von ‫נוח‬-H abge-
leitet wird (‫) ַהנִּ ָחם‬. Diese Praxis bewegt sich in den Bahnen der übri­gen
Bücher aus JE12, die für ‫נחם‬-N mit Subjekt Jhwh konstant μετα­νοέω wäh-
len76. Die zweite Hälfte von JerG* hingegegen gebraucht bei göttlichem
Subjekt (ἀνα)παύομαι aufhören, was ebenfalls an eine Ab­leitung von ‫נוח‬
denken lässt (26/33:3, 13, 19; 42/49:10), während für die Reue Efraims in
31/38:19 weiter­hin μετανοέω eintritt. Die Glei­chung ‫נחם‬-N ≙ παύομαι ist
nochmals in Jes 1:24 belegt, wiederum bei göttlichem Akteur. Allerdings
findet sie sich auch bei Rahel in 31/38:15 ‫נחם‬-N getröstet werden, sich
trösten lassen. Oben­drein wird παύομαι für ‫דכא‬-D pass (Subjekt: Judäer)
44/51:10 herangezogen. Die Basis ‫ דכא‬nebst ihren Vari­anten ‫דכה‬, ‫דכך‬
und ‫ דוך‬ist sonst in JE12 nicht belegt und könnte dem Übersetzer unbe-
kannt gewesen sein. Der Befund legt den Schluss nahe, dass man in JerG*
29–52 eine theologisch motivierte Wie­dergabe vorge­zogen hat, die den
Anthropo­morphismus vermeiden sollte, Gott könne Reue empfinden. Das
Äquiva­lent (ἀνα)παύομαι dürfte von der graphischen Ähnlichkeit ‫נחם‬/‫נוח‬
angeregt sein und wurde noch in weiteren Fällen eingesetzt, wo es dem
Übersetzer wünschenswert er­schien.

75 Vgl. oben die Ausführungen zu 25:30 und speziell Anm. 71.

76 Jhwh/Gott: μετανοέω: Joël 2:13, 14; Am 7:3, 6; Jon 3:9, 10; 4:2; Sach 8:14; Men­schen:
μεταμελέομαι: Ez 14:22.
gottesbildfragen in den lesartendifferenzen 261

27:8 . . . # ‫יהם‬ ֶ ‫]וּב ֶדּ ֶבר[ ֶא ְפקֹד § ַעל־הגּוֹי ַההוּא \ ֲע ֵל‬

ַ ‫וּב ָר ָעב‬
ָ ‫ַבּ ֶח ֶרב‬
‫( \ ֻתּ ָמּם ]א ָֹתם[ ְבּיָ דוֹ‬Gram!) ‫ד־תּ ִמּי‬
ֻ ‫ַע‬
Der Satz kündigt Völkern, die sich der göttlich verfügten Weltherrschaft
Nebukadnezzars (V. 6) verweigern, den Untergang von der Hand des
ba­bylonischen Großkönigs an. MT be­tont eigens die Urheberschaft Jhwhs
in diesem Geschehen: ‫ד־תּ ִמּי א ָֹתם ְבּיָ דוֹ‬ ֻ ‫ ַע‬mit un­gram­mati­scher kausativer
Funktion des Grundstamms von ‫תמם‬. AlT bietet eine regelhafte Va­rian­te,
allerdings ohne Verweis auf das göttliche Zutun: ‫ד־תּ ָמּם ְבּיָ דוֹ‬ ֻ ‫ ַע‬. Weil die
Dro­hung nicht nur Juda, sondern allen Völkern gilt, ist so wenig wie bei
25:31 ersichtlich, warum der Passus den Drang zur Aufhellung des Got­
tesbildes geweckt haben sollte. Allenfalls könnte man in AlT den Wort­
laut an 24:10 ‫ד־תּ ָמּם‬ ֻ ‫ ַע‬angegli­chen haben. Die mangelnde Grammatizität
spricht indes für eine späte Korrektur auf masoretischer Seite, als die
Si­cherheit im Umgang mit der hebräi­schen Sprache bereits im Schwin-
den begriffen war. Offenbar hat man erneut dem Be­dürfnis genügt, die
Autonomie und Majestät Jhwhs zu steigern77. In ähnlicher Weise wurde
auch in AlT 42:12 und MT 43:10, ‌12 eine Initiative vom König von Ba­bel auf
Jhwh verlagert. Wenn der Vorgang in 21:7 umgekehrt verlief, so deshalb,
weil dort Nebukadnezzar erstmals in den Vers eingeführt wurde.

27:10 ‫ ִכּי ֶשׁ ֶקר ֵהם נִ ְבּ ִאים ָל ֶכם ְל ַמ ַען ַה ְר ִחיק ֶא ְת ֶכם ֵמ ַעל ַא ְד ַמ ְת ֶכם‬a
[‫ וַ ֲא ַב ְד ֶתּם‬c ‫ ]וְ ִה ַדּ ְח ִתּי ֶא ְת ֶכם‬b
Laut MT wird die Zerstreuung der von den Falschpropheten verleiteten
Fremdvölker von Jhwh selbst vollstreckt (10b), während AlT 10bc über­
springt. Wie schon im vorigen Beispiel ist kein Grund zu erkennen,
wa­rum ein alexandrinischer Tra­dent gewünscht haben sollte, die­sen Zug
aus dem Charakterprofil Jhwhs zu tilgen. Ohnehin hätte es dafür gereicht,
10b auszu­scheiden oder in einen Infinitiv umzuformen, der wie ‫ַה ְר ִחיק‬
10a der Präposition ‫ ְל ַמ ַען‬unter­geordnet wäre und die Zerstreuung als Ziel
bzw. Effekt der Falschprophetie hinstellen würde, wie es auch V. 15 AlT
tut (dazu sogleich). Da dies nicht geschehen ist, hat wahrscheinlich nicht
AlT das Gottesbild gemil­dert, sondern MT hat es zugespitzt, offenbar im
Interesse der Ge­schichts­souveränität Jhwhs.

27:15 ‫תכם‬ֶ ‫יח]י[ ֶא‬

ִ ‫וְ ֵהם נִ ְבּ ִאים ִבּ ְשׁ ִמי ַל ָשּׁ ֵקר ְל ַמ ַען ַה ִדּ‬
Nach MT hat die judäische Falschprophetie zur Folge, Jhwh zur Zer­
streuung Judas zu treiben, während AlT diese Beteiligung Jhwhs nicht

77 Vgl. Anm. 102.

262 hermann-josef stipp

kennt. Für sich betrachtet, lässt die Variante die Erklärung zu, die alex­
andrinische Tradition habe das Gottesbild um ein unerwünschtes Detail
bereinigt, doch im Lichte des Parallelfalls in V. 10 (s. o.) hat die Stei­
gerung der gött­lichen Verfü­gungs­macht in der masoretischen Tradition
die höhere Plausibilität für sich.

28:13 ‫יהן מֹטוֹת ַבּ ְרזֶ ל‬

ֶ ‫\תי ַת ְח ֵתּ‬
ִ ‫ית‬
ָ ‫מוֹטֹת ֵעץ ָשׁ ָב ְר ָתּ וְ ָע ִשׂ‬
Während laut MT Hananja jene Verendgültigung der babylonischen Welt-
herrschaft vollzieht, die metaphorisch als die Ersetzung des hölzer­nen
Jochs durch ein eisernes vergegenwärtigt wird, ist dies für AlT ein Werk
Jhwhs. Plausibel ist nur die masoretische Priorität: Offen­kun­dig gab AlT
dem Drang nach, eine Maßnahme von solcher Reich­weite in die Verfü-
gung Got­tes zu verlegen, wie dies auch V. 14 tut, wo es Jhwh ist, der den
Nationen das eiserne Joch aufbürdet: ‫ל־הגּוֹיִ ם‬ ַ ‫ל־צוַּ אר ָכּ‬
ַ ‫עֹל ַבּ ְרזֶ ל נָ ַת ִתּי ַע‬
]‫[ה ֵא ֶלּה‬
ָ . In Dtn 28:48 han­delt Jhwh ebenso an Israel. Das härtere Auftre-
ten Gottes hat man dafür in Kauf genom­men.

29:14 ‫( ָל ֶכם‬καὶ ἐπιφανοῦμαι = ‫יתי‬ ִ ‫אתי \ וְ נִ ְר ֵא ִתי )? וְ נִ גְ ֵל‬

ִ ‫וְ נִ ְמ ֵצ‬
Die Rückübersetzung von G* καὶ ἐπιφανοῦμαι ist unsicher; die sonsti­gen
verwertbaren Bele­ge von ἐπιφαίνομαι in JE12 stehen für ‫‑גלה‬N (Ez 39:28
AlT) und ‫‑ראה‬N (Zef 2:11)78. Ez 39:28 AlT pro­pa­giert ähnlich wie hier die
Idee, dass Jhwh sich Judäern in der Fremde offen­bare (‫יהם‬ ֶ ‫לוֹתי ֲא ֵל‬
ִ ָ‫ְבּ ִהגּ‬
‫ ָבּגּוֹיִ ם‬anstelle von MT ‫ל־הגּוֹיִ ם‬ ַ ‫לוֹתי א ָֹתם ֶא‬ ִ ְ‫ ְבּ ַהג‬.), weswegen damit zu
rechnen ist, dass ‫יתי‬ ִ ‫ וְ נִ גְ ֵל‬eine über­greifende Tendenz von AlT wider­
spiegelt. Andererseits kön­nen ‫אתי‬ ִ ‫ וְ נִ ְמ ֵצ‬und ‫ וְ נִ ְר ֵא ִתי‬durch Ver­schreibung
auseinander hervorgegangen sein. Ferner kann ‫אתי‬ ִ ‫ וְ נִ ְמ ֵצ‬dem Bestreben
ent­springen, das sinnlichere Konzept der Gottesnähe, wie von ‫‑ראה‬N
verkörpert, über den Austausch gegen ‫‑מצא‬N zu spiritualisieren bzw.
zu transzenden­tali­sieren. Zugleich lässt sich MT als Ver­such deuten, den
Kontext V. 13–14 geläu­fi­geren Aus­drucksweisen anzunähern: Die engste
Parallele zu V. 13–14 MT ist Jes 65:1 mit drei gemein­samen Verben (‫בקשׁ‬-D,
‫דרשׁ‬, ‫מצא‬-N); vgl. ferner die Beschreibungen von Gotteser­fah­rungen mit
den Wort­paaren ‫דרשׁ‬/‫מצא‬-N Jes 55:6; 1 Chr 28:9; 2 Chr 15:2; ‫בקשׁ‬/‫מצא‬-N
2 Chr 15:4, 15. Bei die­ser unklaren Argumen­ta­tions­lage wird man besser
auf ein Prioritätsurteil ver­zichten. Jedenfalls vertritt G* mit ἐπιφανοῦμαι
eine Lesart, die eher anthropomorpher ist als MT.

78 Vgl. ferner Ez 17:6 für ‫פנה‬.

gottesbildfragen in den lesartendifferenzen 263

30:12 [‫ ָאנוּשׁ ] ְל[ ִשׁ ְב ֵר]ְך‬b ≙ ἀνέστησα σύντριμμα

Das Adjektiv ‫ ָאנוּשׁ‬unheilbar war dem JerG*-Übersetzer unbekannt (vgl.
15:18; 17:9, 16)79. In 30:12 leitete er das Schriftbild behelfsmäßig von der
Wurzel ‫ נשׂא‬ab mit dem Ergebnis Ich (Jhwh) richtete Zusammen­bruch
auf. Infolgedessen “the Lord was presented . . . as the one who brought to
Israel the grief mentioned in the unit Jer. 30:12–17.”80 Das Missverständnis
bedingte, dass sich Jhwh in JerG* über die Vorlage hin­aus eine grausa­me
Tat zuschrieb.

30:23 ‫ ≙ ַס ֲע ַרת יְ הוָ ה‬ὀργὴ κυρίου; ‫גּוֹרר‬

ֵ ‫ ≙ ַס ַער ִמ ְת‬ὀργὴ στρεφομένη81
‫ ַס ֲע ָרה‬wird in JerG* sonst in der Parallele 23:19 durch σεισμός reprä­sen­tiert;
für ‫ ַס ַער‬steht dort συσσεισμός und in 25:32 λαῖλαψ. Die interpre­tierende
Wiedergabe von ‫ ַס ֲע ַרת יְ הוָ ה‬und ‫ ַס ַער‬durch ὀργή (κυρίου) zeigt, dass der
Über­setzer keineswegs zöger­te, Gott sogar über die Vorlage hinaus mit
Unmutsausbrüchen zu assoziieren. Die Bedeu­tung der bei­den Substantive
wird regelrecht auf intensive negative Emotio­nen Jhwhs umgelenkt, pas-
send zur Fortsetzung V. 24. Vielleicht hat sich der Über­setzer dabei von
Anklängen an das Substantiv ‫ זַ ַעם‬leiten lassen, das in JE12, wenn es einen
Affekt Jhwhs beschreibt, überwiegend durch ὀργή repräsentiert wird (Jer
50:25; Ez 21:36; 22:24; Nah 1:6; vgl. Ez 22:31; Hab 3:12). Sollte hinge­gen schon
die Vorlage ‫ זַ ַעם‬gelesen haben, ist die Verschärfung des Gottesbildes für
die hebräische Phase der alexan­drini­schen Tradition anzunehmen. In die-
selbe Kategorie gehört die Äqui­va­lenz ‫( ְמזִ ָמּה‬bezo­gen auf Jhwh) ≙ ὀργή
51:11; vgl. ferner 36:7 (s. u.).

31:20 ‫ לוֹ‬# (ἔσπευσα) ‫ל־כּן § ָהמוּ ֵמ ַעי‬

ֵ ‫ַע‬
In Gottesrede wird ‫ ָהמוּ ֵמ ַעי‬meine Eingeweide gerieten in Wallung durch
ἔσπευσα ich eilte re­präsentiert. Dies ist ein glaubwürdiges Beispiel für
interpretierende Wiedergabe in JerG*. Of­fenbar wollte der Übersetzer die
Vorstellung vom Bauch (κοιλία; so JE12 für ‫ ֵמ ֶעה‬82) Jhwhs vermei­den83.

79 Die funktionslose Präposition in [‫ ] ְל[ ִשׁ ְב ֵר]ְך‬ist aus einer Konflation mit Nah 3:19
‫  ְל ִשׁ ְב ֶרָך נַ ְח ָלה ַמ ָכּ ֶתָך‬. . . ent­standen. Das enklitische Personalpronomen hingegen ist kaum
zu entbehren und dürfte daher in der noch paläohebräisch geschriebenen alexandrini-
schen Tradition durch Haplo­gra­phie mit dem folgen­den ‫ נַ ְח ָלה‬ent­fallen sein: ‫ן → ך‬.
80 B. Becking, Between Fear and Freedom. Essays on the Interpretation of Jeremiah 30–31
(OTS 51; Leiden: Brill, 2004), 27.
81 Anstelle von ‫גּוֹרר‬ ֵ ‫ ִמ ְת‬las die hebräische Fassung von AlT wahrscheinlich ‫חוֹלל‬ֵ ‫ִמ ְת‬
(vgl. Jer 23:19).
82 Jer 4:19; Ez 3:3; 7:19; Jon 2:1, 2.
83 Dies ist weithin anerkannt. Vgl. z. B. Becking, Fear (Anm. 80), 37; Fischer und Vonach,
„Tendencies“ (Anm. 60), 69; Vonach, „Jeremias“ (Anm. 11), 2799.
264 hermann-josef stipp

Er setzte daher an die Stelle emotionaler Auf­gewühltheit ein ge­steuertes,

wenn­gleich vieldeutiges Verhalten. Hinreichender Anlass zur Frage nach
einer devianten Vorlage besteht nicht84.

32:28 ‫וּל ָכ ָדהּ‬

ְ ‫ְך־בּ ֶבל‬
ָ ‫[ ֶמ ֶל‬. . .] ‫ת־ה ִעיר ַהזֹּאת ְבּיַ ד‬
ָ ‫ ֶא‬# ‫§ ִהנְ נִ י נ ֵֹתן \ ִהנָּ תֹן ִתּנָּ ֵתן‬
34:2 <‫וּל ָכ ָדהּ‬ ְ > ‫ְך־בּ ֶבל‬
ָ ‫ת־ה ִעיר ַהזֹּאת ְבּיַ ד ֶמ ֶל‬
ָ ‫ ֶא‬# ‫§ ִהנְ נִ י נ ֵֹתן \ ִהנָּ תֹן ִתּנָּ ֵתן‬
39:17 ‫יהם‬ ֶ ֵ‫ר־א ָתּה יָ גוֹר ִמ ְפּנ‬
ַ ‫וְ לֹא ִתנָּ ֵתן \ ֶא ֶתּנְ ָך ְבּיַ ד ָה ֲאנָ ִשׁים ֲא ֶשׁ‬
Die drei Übereignungsformeln sind gemeinsam zu behandeln. Die ersten
beiden stimmen weitge­hend wörtlich überein und teilen überdies die­selbe
Lesartendifferenz: MT bietet eine aktivi­sche Fassung im Futurum instans,
während AlT eine passivische Figura etymologica bezeugt. In­folgedessen
macht jeweils nur MT explizit Jhwh als Akteur der Ausliefe­rung Jeru­sa­lems
an die Babylonier kenntlich. Zu vergleichen sind hierzu die Parallelen in
32:3, wo MT und AlT übereinstimmend die aktivische Fassung vertreten,
und 38:3, wo beide Text­formen passi­visch formulie­ren. Auf Quer­einflüsse
zwischen diesen Passagen weist die konfla­tionäre Entlehnung von ‫וּל ָכ ָדהּ‬ ְ
aus 32:28 bzw. 38:3 in 34:2 AlT. Bei der ne­gierten Übereig­nungs­for­mel
in 39:17, die Ebed-Melech die Bewahrung vor ano­nymen Gegnern ver­
heißt, sind die Ge­nera verbi umgekehrt verteilt: MT gebraucht das Pas­
siv, während AlT durch das Aktiv Jhwh als Handeln­den identifiziert. Die
Zusammenstellung zeigt, dass kein Texttyp bei den betroffe­nen Übereig-
nungsformeln konsequent ein bestimmtes Ge­nus verbi bevorzugt.
Wie sind die Abweichungen zu bewerten? Die Varianten in 32:28 und
34:2 können zunächst einfach auf die spontane Adaption an 32:3 oder
38:2 zurückgehen. Sollten die For­meln in der alexan­drinischen Tra­dition
entschärft worden sein85, ist dies, wie 32:3 demons­triert, jedenfalls nicht
konsequent geschehen. Ferner ist hervorzuheben, dass JerMT sich von AlT
durch zusätzliche Belege des Deiktikons ‫ ִהנֵּ ה‬unter­scheidet, die sich oben-
drein gerade in Kap. 32 häu­fen86. Daher ist eher den alexan­drinischen
Lesarten von 32:28 und 34:2 die Prio­rität zuzuerkennen, während der
masoretische Wortlaut sich der Vorliebe für ‫ ִהנֵּ ה‬verdanken dürf­te, der
ein Tradent unter konflationärer Anlehnung an Vorbilder nachgab. Die
These der Auf­hel­lung des Gottes­bildes findet hier jeden­falls keine brauch-
bare Stütze.

84 σπεύδω repräsentiert in JE12 sonst ‫עוז‬-H (Jer 4:6), I ‫( נהר‬Mi 4:1, in der G*-Vorlage
verschrieben zu ‫)?מהר‬, ‫מהר‬-D (Nah 2:6); vgl. Ez 30:9.
85 So Vonach, „Jeremias“ (Anm. 11), 2801, 2803, für den Übersetzer.
86 Vgl. 24:1; 32:24, 27, 28; 49:12, 35.
gottesbildfragen in den lesartendifferenzen 265

39:17 bot, da Heilszusage, keinen Anreiz zur theologischen Glättung. Die

sekundäre Wahl des Passivs (MT) ist daher kaum zu motivieren. Dagegen
kann die alexandrinische Fas­sung an die Parallele 38:16 ‫ם־א ֶתּנְ ָך‬ ֶ ‫ַחי־יְ הוָ ה ִא‬
‫  ְבּיַ ד ָה ֲאנָ ִשׁים ָה ֵא ֶלּה‬. . . angepasst sein; viel­leicht hat auch der Wunsch mit-
gespielt, den Vorgang deutlicher der Regie Jhwhs zu unter­stellen. Folglich
hat man allenfalls in 39:17 AlT die göttliche Kontrolle über den Vor­gang
zu erhöhen gesucht, doch eine Milderung des Gottes­bildes lässt sich dem
Material nicht ent­nehmen.

36:7 ‫ל־ה ָעם ַהזֶּ ה‬

ָ ‫ר־דּ ֶבּר יְ הוָ ה ֶא‬
ִ ‫ִכּי־גָ דוֹל ָה ַאף וְ ַה ֵח ָמה ֲא ֶשׁ‬
ὅτι μέγας ὁ θυμὸς καὶ ἡ ὀργὴ κυρίου ἥν ἐλάλησεν ἐπὶ τὸν λαόν τοῦτον
Der Übersetzer hat die Wiedergabe des Gottesnamens aus dem Relativ­satz
in den Matrixsatz verlagert. Die Abfolge kann nicht in der Vorlage gestan-
den haben, denn sie ergäbe eine Kon­struktuskette mit mehreren Nomina
regentia, eine in hebräischer Prosa über­aus seltene Wort­verbindung87.
Demzufolge ist es der Übersetzer gewesen, der mit dem Ausdruck der
Grimm und der Zorn des Herrn die Assoziation Jhwhs mit heftigen nega-
tiven Emotionen nach­gera­de verstärkt hat. Eine regelrechte Neigung, den
Zorn Jhwhs hervorzuheben, zeigt die alexandrinische Tra­dition auch in
30:23 und 51:11.

36:26 ‫] § וַ יַּ ְס ִתּ ֵרם‬. . .[ ‫] וְ ֵאת יִ ְר ְמיָ הוּ‬. . .[ ‫ת־בּרוְּך‬

ָ ‫ ָ ל ַק ַחת ֶא‬. . . ‫וַ יְ ַצוֶּ ה ַה ֶמּ ֶלְך‬
# ‫יְ הוָ ה \ וַ יִ ָסּ ֵתרוּ‬
Baruch und Jeremia entgehen dem Zugriff von Jojakims Häschern, indem
sie ein Ver­steck aufsuchen. Laut MT ist dies ein Werk Jhwhs, wo­von AlT
nichts weiß. Glaubhaft ist nur der Weg von AlT zu MT, mithin der Nach-
druck auf der aktiven Intervention Jhwhs in MT88.

40:3 ‫ל־ה ָמּקוֹם ַהזֶּ ה׃‬

ַ ‫ת־ה ָר ָעה ַהזֹּאת ֶא‬ָ ‫ֹלהיָך ִדּ ֶבּר ֶא‬ ֶ ‫ יְ הוָ ה ֱא‬2
‫אתם ַליהוָ ה‬
ֶ ‫י־ח ָט‬
ֲ ‫ ִכּ‬d [‫ ] ַכּ ֲא ֶשׁר ִדּ ֵבּר‬c ‫ וַ יַּ ַעשׂ יְ הוָ ה‬b [‫ ]וַ יָּ ֵבא‬a 3
Die Rede Nebusaradans an Jeremia in 40:2–5 beginnt mit einer kurz­ge­
fassten theologischen Deutung der Katastro­phe Judas. Über AlT hinaus
betont MT, dass Jhwh das angekündigte Un­heil (V. 2) auch herbei­ge­führt

87 Vgl. B. K. Waltke und M. O’Connor, An Introduction to Biblical Hebrew Syntax (Winona

Lake: Eisen­brauns, 1990), 139 (§ 9.3b).
88 Vgl. Anm. 102.
266 hermann-josef stipp

habe (‫ וַ יָּ ֵבא‬3a)89. Für eine Tilgung ist kein Grund er­sicht­lich. Ferner sind
Aussagen, die ‫בוא‬-H mit dem Subjekt Jhwh verbinden, in Jer derart häu­
fig90, dass der über­wiegende Teil der Belege von ‫בוא‬-H dieses gött­liche
Subjekt trägt. Eine Ergän­zung wie 40:3a entspricht daher den typi­schen
Mustern des ma­so­reti­schen Sonderguts. Die Position des Gottes­namens
beim zweiten, von AlT bestätigten verbalen Prädikat unterstützt die
Annahme einer Zutat, die er­neut dem Ziel gedient haben dürfte, die gött-
liche Geschichts­len­kung zu exponieren.

42:12 ‫ וְ ֶא ֵתּן ָל ֶכם ַר ֲח ִמים‬a

‫ל־א ְד ַמ ְת ֶכם‬
ַ ‫ וְ ֵה ִשׁיב \ וַ ֲה ִשׁב ִֹתי ֶא ְת ֶכם ֶא‬c ‫ וְ ִר ַחם \ וְ ִר ַח ְמ ִתּי ֶא ְת ֶכם‬b
Der Vers aus dem Orakel Jeremias an die angehenden Ägyptenaus­wan­
derer ist seinem Kon­text nach dazu bestimmt, den Judäern jene Heils­
gaben zu verheißen, die sie erwarten dürfen, wenn sie auf die gewünsch­te
Emigration verzichten. Faktisch wendet sich der Passus jedoch in einem
Fik­tionsbruch an die Adressaten des Erzählers im Exil, um ihnen die
Heimkehr in Aussicht zu stellen (12c)91. Die Rückführung ist nach MT
eine Tat des Königs von Babylon, laut AlT hingegen wird sie von Jhwh
bewirkt. Die letztere Variante wird aus doppeltem Grund als se­kundär
er­wiesen: Erstens lässt sich 12c AlT aus dem historischen Wissen ablei­ten,
dass es nicht die Babylonier waren, die die Erlaubnis zum Aufbruch nach
Juda erteilten; schon deshalb hätte man nicht nachträg­lich die­sen Akt in
ihre Hände verlegt. Zweitens hat die Revision in 12b AlT eine Du­blette zu
12a hervorgebracht. Infolgedessen hat hier die alexan­drinische Tra­dition
bestimmte Hand­lungen explizit aus menschlicher in göttliche Regie über­
tragen, doch nicht aus inner­theologi­schen Gründen, sondern um retro-
spektiv den Einklang mit dem Geschichtsverlauf herzu­stel­len.

89 Überdies ist ‫ ַכּ ֲא ֶשׁר ִדּ ֵבּר‬3c mit Sicherheit prämasoretischer Zusatz, wie das idiolek­
tale Gepräge zeigt; vgl. Stipp, „Updated Index“ (Anm. 4), 194; idem, Konkordanz (Anm. 53),
35. Zu objektlosem ‫ עשׂה‬im Sinne von han­deln, es ausführen vgl. von Jhwh 1 Sam 14:6; Ps
119:126; mit anderen Akteuren z.B. Gen 41:34; Jer 12:5; Hag 2:4; Spr 13:16; 21:25; 31:13; Esr 10:4;
1 Chr 28:10; 2 Chr 19:7.
90 2:7; 3:14; 4:6; 5:15; 6:19; 11:8MT, 11, 23; 15:8 u. v. a.
91 Diese Interpretation wird näher gerechtfertigt bei H.-J. Stipp, Jeremia im Parteien­
streit. Studien zur Textent­wicklung von Jer 26, 36–43 und 45 als Beitrag zur Geschichte Jere-
mias, seines Buches und judäi­scher Par­teien im 6. Jahrhundert (BBB 82; Frankfurt am Main:
Hain, 1992), 188f., 254–256; ferner idem, „Das judäische und das babylonische Jeremiabuch,“
in Congress Volume Ljubljana 2007 (VTSup 133; Hg. J. Krašovec; Lei­den: Brill, 2010), 239–264
(262 Anm. 106). Dort werden auch weitere Details der Textbezeu­gung von 42:12 erörtert.
gottesbildfragen in den lesartendifferenzen 267

43:10 [‫ְך־בּ ֶבל ] ַע ְב ִדּי‬

ָ ‫אצּר ֶמ ֶל‬ ַ ‫בוּכ ְד ֶר‬
ַ ְ‫ וְ ָל ַק ְח ִתּי ֶאת־נ‬d ‫ ִהנְ נִ י שׁ ֵֹל ַח‬c
‫ וְ ַשׂ ְמ] ִתּי[ ִכ ְסאוֹ ִמ ַמּ ַעל ָל ֲא ָבנִ ים ָה ֵא ֶלּה‬e
Laut 10e MT kündigt Jhwh an, für Nebu­kad­nezzar in Ägypten einen Thron
zu errichten, wo­hingegen nach AlT der Babylonier dies selbst tun wird.
Hier hat anerkanntermaßen die maso­reti­sche Tradition Souveräni­tät
vom König von Babel auf Jhwh verlagert92. Eine gleichartige Inter­vention
bezeugt sogleich das nächste Beispiel.

43:12 ‫ֹלהי ִמ ְצ ַריִ ם‬
ֵ ‫וְ ִה ַצּ ִתּי \ וְ ִה ִצּית ֵאשׁ ְבּ ָב ֵתּי ֱא‬
In MT legt Jhwh selbst Feuer an die ägyptischen Tempel, was nach AlT
Nebukadnezzar ob­liegt. In der Fortsetzung 12b–df, 13a weisen dann beide
Textformen dem Großkönig die Sub­jektsrolle zu. Dies und der Be­fund in
V. 10 (s. o.) bezeugen, dass erneut die masoretische Tra­dition die Macht-
fülle Jhwhs auf Kosten Nebukadnezzars gesteigert hat93.

46:10 <‫וְ ָא ְכ ָלה ֶח ֶרב > ַליהוָ ה‬

Der Präpositionalausdruck wurde in AlT konflationär aus dem paralle-
len Satz in 12:12 inter­poliert, mit der Konse­quenz, dass die Grausamkeit
Jhwhs gegenüber Ägypten deutlicher her­vortrat.

49:35 ‫ילם‬ ָ ‫־ק ֶשׁת ֵע‬

ֶ [‫] ִהנְ נִ י[ שׁ ֵֹבר \ יִ ָּשׁ ֵבר ] ֶאת‬
Das Deiktikon ‫ ִהנֵּ ה‬ist in MT mehrfach nachgetragen worden94, was dafür
spricht, dass dies auch hier geschah. In­dem man zudem das Prädikat ins
Aktiv versetzte95 und dem göttlichen Subjekt un­terstellte, be­tonte man
Jhwhs weltweites Walten. Weil ein Fremdvolk betroffen ist, bestand
wenig Anreiz zu abmilderndem Eingreifen.

50:20 ‫ִכּי ֶא ְס ַלח ַל ֲא ֶשׁר ַא ְשׁ ִאיר \ נִ ְשׁ ֲאר‬

Die in das Babylonorakel eingeflochtenen Heilsworte für Israel und Juda
V. 17–20 sagen an ihrem Ende dem überlebenden Rest Vergebung zu. MT
rechnet das Überdauern durch ‫שׁאר‬-H aus­drücklich dem Eingreifen Jhwhs
an, während AlT durch das Passiv (‫שׁאר‬-N) kei­nen Urheber the­matisiert.

92 Vgl. z. B. W. L. Holladay, Jeremiah 2. A Commentary on the Book of the Prophet Jer-

emiah Chapters 26–52 (Her­meneia; Philadelphia: Fortress Press, 1989), 277; anders Vonach
(Anm. 102).
93 Anders wieder Vonach (Anm. 102).
94 Ferner 24:1; 32:1, 24, 27, 28; 34:2; 49:12; 50:12; vgl. 49:15.
95 Dabei ist wohl ebenfalls ‫ ֵאת‬ergänzt worden, das in passivischen Sätzen indes auch
das Subjekt markieren kann.
268 hermann-josef stipp

Weil sich nur der Übergang von AlT zu MT einsichtig machen lässt, ist
aber­mals eine Steigerung der göttlichen Geschichtssouveränität in MT zu

51:11 ‫יתהּ‬
ָ ‫( ְל ַה ְשׁ ִח‬ἡ ὀργὴ αὐτοῦ) ‫ל־בּ ֶבל ְמזִ ָמּתוֹ‬
ָ ‫י־ע‬
ַ ‫ת־רוּח ַמ ְל ֵכי ָמ ַדי ִכּ‬
ַ ‫ֵה ִעיר יְהוָ ה ֶא‬
Die Äquivalenz ‫ ≙ ְמזִ ָמּה‬ὀργή ist in G singulär. JerG* wählt für das Sub­
stantiv sonst βδέλυγμα Abscheulichkeit (11:15) und ἐγχείρημα Vorha­ben, Ver-
such (23:20 || 30:24). Für das Verb ‫ זמם‬tritt ἐγχειρέω vorhaben, trachten ein
(so im folgenden Vers 51:12)97, für ‫ זִ ָמּה‬ἀπαλλοτρίωσις Entfremdung (13:27).
Wie der mehrmalige Gebrauch von ἐγχειρ- er­weist, war dem Überset­zer
die Bedeutung der Wurzel ‫ זמם‬vertraut, doch gab er sie uneinheitlich wie-
der. War seine Vorlage mit MT identisch, hat er hier mit ὀργή ein beson-
ders scharfes Äquivalent eingesetzt. Ähnlich wie in 30:23 stellt sich die
Frage, ob ihm der Anklang an die Wur­zel ‫ זעם‬als Inspiration diente, und
wie dort ist nicht völlig auszu­schlie­ßen, dass bereits die Vorlage entspre-
chend zugespitzt worden war. Der Vergleichs­fall 36:7 lässt den Ursprung
der Variante indes eher beim Übersetzer suchen, der folglich keine Scheu
kannte, von Jhwhs Zorn zu reden und da­bei so­gar über seine Vorlage hin-

51:59 ‫ר־צוָּ ה >יְ הוָ ה ֵאת< יִ ְר ְמיָ הוּ ַהנָּ ִביא > ֵלאמֹר< ֶאת־ \ ֶאל ְשׂ ָריָ ה‬
ִ ‫ַה ָדּ ָבר ֲא ֶשׁ‬
MT berichtet von dem Wort, das der Prophet Jeremia dem Seraja auftrug;
AlT hingegen spricht von dem Wort, das Jhwh dem Propheten Jeremia auf-
trug, (es) dem Seraja zu sagen. Der Gebrauch von εἰπεῖν für ‫ ֵלאמֹר‬98 deu-
tet auf eine unge­schickt redigierte hebräische Vor­la­ge. Dem­nach hat AlT
ähnlich wie in 1:1 (s. o.) die Kontrolle über einen Offen­ba­rungs­vorgang
ausdrücklich von Jeremia auf Jhwh übertragen99.

3. Auswertung

Ein sachgerechtes Urteil über das vorgelegte Material hat zunächst zu

berücksichtigen, dass die Belegsammlung nur Unterschiede zwischen dem
tiberischen Text und JerG* auf­listet. Die be­handelten Fälle sind je­doch

96 Laut Vonach, „Jeremias“ (Anm. 11), 2781, wird sogar hier „Gott in der LXX eher milder
97 Vgl. auch 4:28: ὁρμάω?
98 Sonst 27:4/34:3; 30/37:1; 37/44:17; vgl. 44/51:26 (MT ‫)א ֵֹמר‬.
99 Vgl. Vonach, „Jeremias“ (Anm. 11), 2785, für den Übersetzer.
gottesbildfragen in den lesartendifferenzen 269

eingebettet in ei­ne Fülle von Passagen, in denen die beiden Zeu­gen

überein­stim­mend Jhwh eine Härte im Denken und Handeln zu­schreiben,
die bei der Vorliebe für mild getönte Gottesbilder, wie sie in unserem
Kul­turkreis verbreitet ist, erfah­rungsgemäß erheb­lichen Anstoß erre­gen
kann. Schon dies nötigt zur—immerhin banalen—Einsicht, dass antike
jahwistische Textproduzen­ten und -rezipienten andere Maßstäbe an
ihre autorita­tive Literatur anlegten als wir heute; was Leser des 21. Jahr­
hunderts verstört, braucht ein antikes Publikum keineswegs befremdet
zu haben. Wie die Lektüre des ge­sam­ten Buches be­stätigt, hat JerG* die
schroffen Züge Jhwhs norma­ler­weise unge­filtert an ihre Leser weiter­ge­
reicht. Dies gilt uneingeschränkt auch für Jhwhs Umgang mit Israel und
Juda100. Dasselbe lässt sich—wie hier nicht mehr näher ausgeführt werden
kann—für die zahlreichen Anthropo­morphismen und -pathismen feststel-
len101. Eine zusammen­fassende Würdigung muss also in Rech­nung stellen,
dass die unter­suchten Lesartendifferenzen keine Regelfälle reprä­sentie­ren,
son­dern ein recht kleines Kor­pus von Ausnahmen bilden.
Die Analyse hat nicht immer eindeutige Priori­täts­urteile geliefert. Sie
hat freilich erge­ben, dass in der Tat vereinzelt textliche Modi­fika­tio­nen
plausibel auf das Anliegen rückführ­bar sind, uner­wünsch­te Details aus
dem Porträt Jhwhs zu til­gen, also Gottesbilder an vorgän­gige Erwartun­
gen anzugleichen. Diese Ein­griffe sind aber gering an Zahl und haben
den Sta­tus begrenzter Anomalien, zeigen also kei­neswegs vorherrschen­de
Ten­den­zen an, schon gar nicht in JerG*. Insgesamt fördern die hinter den
Devianzen aufscheinenden Moti­ve nicht den Ein­druck, dass die Traden­ten
der masore­tischen und der alexan­drini­schen Ausgabe des Jere­mia­buches
mitsamt dem griechi­schen Überset­zer von der Sorge über grau­same
Im­plika­te der Gotteskonzepte des Werkes umgetrie­ben wur­den. Entge­gen
anders lautenden Thesen war ihr Drang, solche Eigen­arten zu dämp­fen,
wenig ausgeprägt.
Sofern sich ein Trend aus­machen lässt, der eine nennens­werte An­zahl
von Retu­schen hervorgebracht hat, betrifft dieser die Neigung des maso­

100 Vgl. z. B. 2:30; 4:4, 8–10, 26; 5:3; 5:9 (|| 5:29; 9:8); 5:14–17; 6 passim; 7:14–16, 20, 33f.;
8:1–3, 10, 17f.; 9:10/11, ‌14f./15f., 20f./21f.; 11:11, 14, 17, 22f.; 12:7–13; 13:13f., 17, ‌24–27; 14:11f., 15–18;
15:1–9, 13f.; 16:3–5, 9, 13, 18; 17:27; 18:17; 19:3, 7–13, 15; 20:4–6; 21:4–14; 22:5–7, 22; 23:19f., 39f.;
24:8–10; 25:9–11 u. v. a.
101 So ist in JerG* die Rede vom Grimm (ὀργή, θυμός, παροξυσμός) Jhwhs ebenso gängig
wie in MT und wird, wie oben gezeigt, in 30:23; 36:7; 51:11 sogar noch aus­ge­weitet. Vgl.
ferner z. B. 1:9; 2:21, 30; 4:19; 8:19–9:1/2; 9:14/15; 10:12f. (|| 51/28:15f.), 17f.; 11:17; 12:2, 7f.; 13:11, 13;
16:17; 17:27; 19:7, 11; 21:4–7, 14; 22:24–26; 23:15, 19f., 39; 24:6; 25:3; 46/26:15; 48/31:36; 49:10/29:11;
49:38/25:18; 50/27:25; 51/28:25, 39, 57 u. ö.
270 hermann-josef stipp

reti­schen Strangs, die souve­räne gött­liche Herrschaft über alles Gesche­hen

zu betonen. Diese Tendenz charakterisiert also gerade nicht JerG*, son-
dern im Gegenteil die masoretische Textüberlieferung. Mehrfach wurden
Handlun­gen aus mensch­lichen Händen in die Initiative Gottes ver­legt.
Während Jere­mia nach dem älteren Wortlaut in 11:18 die Un­taten seiner
Wider­sacher sieht (‫ראה‬-G) und in 18:2 das Gotteswort hört (‫שׁמע‬-G), ver-
schiebt MT die Ver­ben in den H-Stamm und stellt so die Wahr­nehmungen
des Propheten als von Gott verursacht hin. Diese Fälle stützen besonders
nach­drücklich die alexandrinische Priorität, weil dort keinerlei Anlass
be­stand, die Ur­heberschaft Jhwhs zwecks Aufhellung seines Porträts oder
Vermeidung von An­thropo­mor­phismen herun­terzu­spielen. Dies bekräf-
tigt das Urteil über die folgenden Beispiele: In 10:18 hat man Jhwhs Regie
durch den Austausch von ‫ ַבּ ָצּ ָרה‬gegen ‫רוֹתי ָל ֶהם‬ ִ ‫ וַ ֲה ֵצ‬zusätzlich unter­
strichen. Dasselbe geschah in 49:35 MT, wo man das akteurs­blin­de Passiv
‫ יִ ָּשׁ ֵבר‬gegen das auf Jhwh bezogene Aktiv ‫ ִהנְ נִ י שׁ ֵֹבר‬vertauschte; ebenso
hat man in 50:20 MT ‫ נִ ְשׁ ֲארוּ‬durch den Kausativ ‫ ַא ְשׁ ִאיר‬ersetzt. Wenn
ferner in 36:26 AlT Baruch und Jeremia sich vor den Nach­stel­lun­gen Joja-
kims ver­bergen (‫)וַ יִ ָסּ ֵתרוּ‬, macht MT eine Tat Jhwhs daraus (‫)וַ יַּ ְס ִתּ ֵרם יְ הוָ ה‬.
Nach 27:8 AlT gehen die rebellischen Völ­ker von der Hand Nebu­kad­nez­
zars zugrunde (‫ד־תּ ָמּם ְבּיָ דוֹ‬ ֻ ‫) ַע‬, doch laut MT vollzieht dies Jhwh, wäh-
rend der Groß­kö­nig zum ausführen­den Organ herab­sinkt (‫ד־תּ ִמּי א ָֹתם‬ ֻ ‫ַע‬
‫) ְבּיָ דוֹ‬. Wo AlT zufolge Nebu­kadnezzar in Ägyp­ten sei­nen Thron errich­tet
(43:10) und die Tempel einäschert (43:12), er­kennt MT gött­liche Akte102.
In 25:6 dürfte älteres ‫ ְל ָה ַרע‬dem negierten Kau­sativ ‫ וְ לֹא ָא ַרע‬ge­wichen
sein. Außerdem hat man in 6:11 den Verdacht man­gelnder Selbstkontrolle
Jhwhs unterbunden. Und während Jhwh in 10:16 || 51:19 AlT als Erbteil
Jakobs deklariert wird, kehrte man in MT mit Rücksicht auf die göttliche
Majestät die Beschreibungs­rich­tung um, sodass nun Israel als der Stamm
von Jhwhs Erbteil erscheint.
Wiederholt hat die masoretische Tradition im Inter­esse der Souverä­
nität Jhwhs sogar zu­sätzliche Kanten in seinem Profil in Kauf ge­nom­
men. Um in 16:13 dem Eindruck der Aktionsfähigkeit von Fremd­göttern
einen Riegel vorzuschie­ben, zog man es vor, deren Erbar­mungs­lo­sigkeit

102 Die Meinung von Vonach, „Jeremias“ (Anm. 11), 2725, in 27:8; 36:26; 43:10, 12 habe
umgekehrt der Überset­zer zwecks Vermeidung von Anthropomorphismen die gött­lichen
Akte in menschliche Hände verlegt, ist an den in Anm. 101 genannten Passagen zu messen.
Nach Vonach müsste sich der Übersetzer auf wenige Bei­spiele gängigen Redens von Gott
beschränkt und gerade die massiven Anthropomorphis­men und ‑pathismen unan­getastet
gelassen haben.
gottesbildfragen in den lesartendifferenzen 271

auf Jhwh zu übertragen (‫א־א ֵתּן ָל ֶכם ֲחנִ ינָ ה‬

ֶ ֹ ‫)ל‬. In 27:15 wurde die Zerstreu­
ung Judas ausdrück­lich als sein Werk deklariert (‫יח]י[ ֶא ְת ֶכם‬ ִ ‫) ְל ַמ ַען ַה ִדּ‬
und in 40:3 seine Ur­he­berschaft der Kata­strophe deutlicher hervorge­kehrt
([‫)]וַ יָּ ֵבא‬. Daher würde es zum Gesamt­bild passen, wenn—was allerdings
nicht mit Si­cherheit zu klären ist—in 19:9 der Kanni­balis­mus der Judäer
an ihren Kindern (AlT: ‫אכל‬-G) nach­träglich ex­plizit auf das Wirken Jhwhs
zu­rückge­führt wor­den wäre (MT: ‫אכל‬-H mit gött­lichem Subjekt).
Dass diese Modifikationen tatsächlich im masoretischen Strang ein­
traten mit dem Ziel, die göttliche Geschichtslenkung selbst um den Preis
zusätzlicher Härten zu exponieren, und nicht etwa umgekehrt auf alex­
an­drinischer Seite die Züge Jhwhs freundlicher gestaltet wur­den, belegen
insbesondere jene masoretischen Sonderlesarten, die Jhwh Grausam­
keiten an Frevlern und Fremdvölkern zuschreiben und daher kaum Anlass
zur Abschwächung uner­wünschter Charakteristika bieten konnten (25:31;
27:8, 10; 49:35).
Geringfügig hat der Nachdruck auf Jhwhs Regie hinter dem Gesche­hen
auch in AlT seine Spuren hinterlassen. So wandel­ten sich in 1:1 die Worte
Jeremias in das Wort Jhwhs, das an Jeremia erging, und in 51:59 wurde der
Auf­trag Jeremias an Seraja ersetzt durch das Geheiß Jhwhs an Jeremia
zur Beauftragung Se­rajas. In 28:13 ging der weltgeschichtliche Akt, der
metaphorisch aufscheint als Herstel­lung der eisernen Joch­stangen, von
Hananja auf Jhwh über. Ferner ist damit zu rechnen, dass man aus Respekt
vor der Hoheit Jhwhs bei meh­reren Belegen der Wort­ereignisformel die
Nennung des Gottesnamens sekundär vor das Glied mit dem Adressaten
Jeremia gerückt hat (11:1; 18:1; 21:1; 32:1; 40:1). Dagegen erscheint Zurück-
haltung geboten bei der Frage, ob die alexan­drini­sche Tradition in 22:12
die Deporta­tion des Joa­has nach­träg­lich in gött­liche Hände gelegt hat;
gleiches gilt für die aktivische Fas­sung der Zusage an Ebed-Melech in
39:17, ihn nicht seinen Feinden auszuliefern. Auszuklam­mern ist jeden­
falls 42:12, wo AlT aus der Rück­schau die Erlaubnis zur Heimkehr aus
dem Exil vom König von Babylon auf Jhwh verlagert hat, denn hier gaben
keine theologischen Motive (im engeren Sinne) den Ausschlag, son­dern
die Kenntnis des tatsächlichen Ge­schichtsverlaufs. Ist somit eine für MT
cha­rakteristische Ten­denz stark vermindert auch in AlT anzutreffen, ent-
spricht dies dem ge­nerellen Pan­orama, wonach in beiden Überlieferungs­
strän­gen ähnliche Praktiken der Textpfle­ge wirksam waren, ob­gleich in
sehr unterschiedlichem Aus­maß103.

103 Vgl. Stipp, Sondergut (Anm. 3), 165.

272 hermann-josef stipp

Leicht an Grausamkeit gewann das Gottesbild in AlT 20:16 durch den

Zusatz, wohl aus Dtn 29:22 (vgl. Ijob 9:5) entlehnt, Jhwh habe sein Zer­
störungswerk <‫ > ְבּ ַאף‬verrichtet; fer­ner wurde 46:10 an 12:12 adaptiert mit
dem Ergebnis <‫וְ ָא ְכ ָלה ֶח ֶרב > ַליהוָ ה‬. Der Nach­druck auf Jhwhs Zorn in 36:7
ist dagegen mit Sicherheit dem Übersetzer anzu­rechnen, der sich folglich
mitnich­ten scheute, den Gott Israels mit negati­ven Gefühlsaus­brü­chen
zu verbinden. Dasselbe gilt mit hoher Wahrscheinlichkeit auch für 30:23
und 51:11; andernfalls wurzeln diese heftigen Zuspitzungen schon in der
hebräischen Phase der alexandrinischen Texttradition. Be­dingt durch ein
Missverständnis der Vorlage, hat der Übersetzer in 30:12 mit ἀνέστησα
σύντριμμα Jhwh noch eine weitere Unheilstat an Israel zu­geschrieben.
Anthropomor­phismen, die sogar über MT hinaus­gehen, bietet AlT bzw.
G* in 15:7 ‫ ֻשׁ ַכּ ְל ִתּי‬/ἠτεκνώ­θην und 29:14 ‫וְ נִ ְר ֵא ִתי‬/καὶ ἐπιφανοῦμαι.
Von einer Entschärfung des Gottesbildes kann dagegen nur in weni­
gen Ausnahme­fäl­len die Rede sein. Einigermaßen sicher ist lediglich,
dass AlT in 1:17 und wohl auch 15:17, 18 den barschen Um­gang Jhwhs
mit seinem Propheten abschwächte104. Der Schöp­fer von JerG* be­schnitt
Anthropomor­phis­men, als er die Vor­stel­lungen vom Brüllen (25:30) und
vom Bauch Jhwhs (31:20) seinem Publikum vorenthielt; weiterhin scheint
er aus dem­selben Grund in der zweiten Hälfte des Buches die Rede von
der Reue Gottes vermieden zu haben, indem er für ‫נחם‬-N mit göttlichem
Subjekt (ἀνα)παύομαι aufhören wählte. Wenn jedoch ein alexandrini­scher
Tra­dent oder der Übersetzer das Weinen Jhwhs in 9:9; 14:17 auf das Volk
umleitete, so zeigt die gleichartige Umlenkung des Weinens Jeremias 13:17,
dass nicht Vor­behalte gegen Anthropo­pathis­men die Feder führten, son-
dern das Aussageziel, die Schuldigen sollten ihre Not­lage selbst betrauern.
Eine Milde­rung des Got­tesbildes könnte die masoreti­sche Tra­dition in
22:12 beabsichtigt haben, wenn, was offen bleiben muss, die Deportation
des Joahas nicht in AlT auf Jhwh über­tra­gen (‫) ִהגְ ֵל ִתי‬, sondern umgekehrt
in MT auf Men­schen verscho­ben worden wäre (‫) ִהגְ לוּ‬. AlT hätte das Got-
tesbild in 19:9 auf­gehellt (‫‑אכל‬G), sofern nicht im Gegenteil die maso-
retische Überlieferung Jhwh se­kundär zum An­stifter des Kanni­balis­mus
erhoben hat (‫‑אכל‬H MT), was insgesamt plausibler ist.
Keinen signifikanten Einfluss auf das Gottesbild hat dagegen die ale­
xandrinische Transformation der Frage 23:23 in eine Feststellung, sodass

104 Vielleicht meldet sich hier dieselbe Sorge um die Reputation der authentischen Pro-
pheten zu Wort, die in 26:21 zur Streichung der Notiz von der Furcht und Flucht des Urija
ben Schemaja geführt hat; vgl. Stipp, Sondergut (Anm. 3), 153.
gottesbildfragen in den lesartendifferenzen 273

Jhwh von einem ‫ֹלהי ֵמ ָרחֹק‬ֵ ‫ ֱא‬in einen ‫ֹלהי ִמ ָקּר ֹב‬
ֵ ‫ ֱא‬ver­wandelt wird, denn
die Neufassung leitet dasselbe Theologumenon—die Unentrinnbarkeit
Gottes—mit ei­ner anderen Logik her, nachdem das Verständnis für die
ältere Begründung geschwun­den war. Kaum zum Thema gehören die prä-
masoretischen Eingriffe in 21:7, die nicht die Ent­lastung Jhwhs bezweck­
ten, sondern Nebu­kadnez­zar in den Kon­text einführen sollten. Einem
Urteil entzieht sich die Pas­siv­konstruktion ἐξώσθησαν beim Rück­blick auf
die Zerstreu­ung Israels in JerG* 16:15, denn ihr text­geschichtlicher Ort
und die Triebkräfte hinter ihrer Entste­hung sind nicht be­friedigend zu
erhellen. Ein negatives Ergeb­nis lässt sich immerhin konstatie­ren: Aus
sekundä­rem Bemühen um das Gottesbild ist das Passiv wahr­scheinlich
nicht erwach­sen.
Die verbleibenden Diffe­renzen verdanken sich wohl ganz anders­arti­
gen Beweggrün­den. Die nur von MT bezeugte prophetische Boten­formel
in 13:12 rührt aus der rou­tine­mäßi­gen Vermehrung klischierter Phra­sen in
diesem Texttyp her, und die ehemals pas­si­ven Über­eignungs­formeln in
32:28 || 34:2 nahmen aktive Form an, weil ein Rezensor in der masore­ti­
schen Tradition das Deiktikon ‫ ִהנֵּ ה‬besonders schätzte.
Die Behauptung, die alexan­drinische Tradition bzw. der Übersetzer
hätten anstößige Züge der Gottesvorstellung abgeschliffen, trifft nach all-
dem nur vereinzelt zu. Die weni­gen gültigen Belege werden jedoch durch
gegenteilig orientierte Beispiele aufge­wogen, und ohne­hin ändern sie
nichts am extremen Übergewicht der Unheilsansagen, die eine Unmenge
massiver Härten in den Gottesbildern mit sich bringen, sodass selbst dann,
wenn sich wei­tere Mil­derungen nachweisen ließen, das Por­trät Jhwhs nur
marginal nuanciert würde.
So lässt sich resümieren: Zu den Motiven, die nach der Gabelung der
Überlieferungs­stränge des Jeremiabuches die Textentwicklung antrieben,
gehörten auch Impulse, die man als Arbeit am Gottesbild klassifizieren
kann. Sie waren jedoch erstens insgesamt nur recht schwach ent­wickelt.
Zweitens wirkten sie bloß minimal in der alexandrinischen Tradition—
vom griechi­schen Über­setzer ganz zu schweigen—, während die meisten
einschlägigen Retu­schen aus prämasoreti­schen Händen hervorgingen,
wie es das generelle Verhältnis der beiden Arme der Textüberlieferung
ohnehin erwarten lässt. Drittens ging es den Tradenten nur in sel­tenen
Aus­nahmefällen darum, Jhwh mildere Züge zu verleihen, sondern sie
wollten vor allem seine Lenkung allen Geschehens propagieren, und sei
es sogar um den Preis gestei­gerter Här­te sei­nes Profils. Allem Anschein
nach galt ihr Interesse angesichts ihrer eigenen politischen Macht­losig­
keit viel mehr der Geschichtssouveränität Jhwhs als seiner Gnädigkeit.
274 hermann-josef stipp

Was theo­logische Pro­gramme der Septuaginta-Übersetzer angeht, bleibt

der Er­trag bescheiden: Der Schöpfer des griechischen Jeremiabuches
setzte alles daran, seinem Publikum durch getreuliche Wie­der­gabe des
Mut­tertextes ein zuverlässiges Bild vom verehrten hebräi­schen Original
zu ver­mitteln—so, wie er es verstand.
Two Difficult Passages in the Hebrew Texts of
Lamentations 5: Text-Critical Analyses of the
Greek Translation

Gideon R. Kotzé

1. Introduction

The Greek translation of Lamentations counts amongst the more interest-

ing translation units in the corpus of Greek Jewish scriptures traditionally
known as the Septuagint (LXX). LXX Lamentations is generally charac-
terised as a “literal”1 or “quantitative formal equivalent”2 translation of a
Hebrew Vorlage that was close to the consonantal base of the MT.3 It is
also considered as a member of the kaige group of translations and revi-
sions.4 Concerning the “literal” nature of LXX Lamentations, Albrektson5
is of the opinion that the translator rendered his Vorlage in a mechanical
way and often word for word. By disregarding the context of passages, he

1 According to H. Ausloos and B. Lemmelijn (“Content-Related Criteria in Characteris-

ing the LXX Translation Technique,” in Die Septuaginta—Texte, Theologien, Einflüsse. 2.
Internationale Fachtagung veranstaltet von Septuaginta Deutsch [LXX.D], Wuppertal 23.—
27.7.2008 [eds. W. Kraus and M. Karrer; Tübingen: Mohr Siebeck, 2010], 358), “a transla-
tion is labelled as ‘literal’ when it renders its Vorlage very accurately, translating it in an
almost mechanical way and often word for word”. In contrast to such literal translations,
“a translation is characterized as ‘free’ when it does not proceed in this manner, in other
words when it handles its Vorlage, both linguistically and exegetically, more freely”. They
point out, however, that, in the past, the “distinctive features of ‘literalness’ have not been
precisely defined and there has been very little interest in the ‘freedom’ of a translation”.
Ausloos and Lemmelijn (“Content-Related Criteria,” 359) insist that one must determine
in which ways a translation unit is either literal or free. In this regard, see J. Barr, The
Typology of Literalism in Ancient Biblical Translations (MSU XV; Göttingen: Vandenhoeck
& Ruprecht, 1979).
2 K. J. Youngblood, “The Character and Significance of LXX Lamentations,” in Great
Is Thy Faithfulness? Reading Lamentations as Sacred Scripture (eds. R. A. Parry and H. A.
Thomas; Eugene: Pickwick, 2011), 65.
3 See G. R. Kotzé, “The Greek Translation of Lamentations: Towards a More Nuanced
View of its ‘Literal’ Character,” in Septuagint and Reception (VTSup 127; ed. J. Cook; Leiden:
Brill, 2009), 78.
4 See D. Barthélemy, Les Devanciers d’Aquila (VTSup 10; Leiden: Brill, 1963), 158–160;
and I. Assan-Dhôte and J. Moatti-Fine, Baruch, Lamentations, Lettre de Jérémie (BdA 25.2;
Paris: Cerf, 2005), 155–157.
5 B. Albrektson, Studies in the Text and Theology of the Book of Lamentations with a Criti-
cal Edition of the Peshitta Text (Lund: CWK Gleerup, 1963), 208–210.
276 gideon r. kotzé

fails, at times, to capture the meanings of the Hebrew sentences. Simi-

larly, Hirsch-Luipold and Maier6 depict the approach of the translator as
Der Septuaginta-Übersetzer versucht, die hebräische Vorlage möglichst dec-
kungsgleich nicht primär im Blick auf den Sinn, sondern auch im Blick auf
die Satzstruktur, die Grammatik sowie die Bedeutung und Form der Einzel-
wörter ins Griechische zu übernehmen. Dabei entstehen äußerst gewagte
Satzkonstruktionen und eine fremdartige, bisweilen poetische Sprache,
aber auch Sätze, die im Griechischen rätselhaft bleiben. Der Sinn der grie-
chischen Übersetzung weicht gerade wegen dieser Übersetzungsmaxime an
manchen Stellen erheblich von dem des hebräischen Textes ab.
With regard to the membership of LXX Lamentation to the kaige group,
only five of the nine key features of the group discussed by Barthélemy are
present in this translation unit.7 First, all the occurrences of ‫ גם‬in Lam-
entations are rendered by καί γε (Lam 1:8, 2:9, 3:8, 4:3, 4:15 and 4:21). Sec-
ondly, in LXX Lam 3:53, ἀνήρ serves as the translation equivalent of ‫איש‬,
but the latter is not used in a distributive sense in this case. Thirdly, forms
of ‫ נצב‬appear in Lam 2:4 and Lam 3:12. In the first passage, ‫ נצב ימינו‬is
translated as ἐστερέωσε δεξιὰν αὐτοῦ. The rendering does not reflect the
kaige group’s characteristic translation equivalent for ‫נצב‬, στηλόω. In Lam
3:12, however, ‫ ויציבני‬is rendered by a form of στηλόω. Fourthly, the Greek
present is only found once in LXX Lam 4:19 (= MT Lam 4:18). Gentry there-
fore remarks that “it can safely be said that the translator was not in the
habit of using historical presents”.8 Finally, ‫ אין‬is translated six times as
οὐκ ἔστιν (Lam 1:9, 1:17, 1:21, 2:9, 4:4 and 5:8), three times as οὐχ ὑπάρχει/
ὑπάρχουσιν (Lam 1:2, 5:3 and 5:7) and once as οὐκ ἦν (Lam 1:7). Gentry9
concludes from this that “only three of the five patterns provide strong
support that the Greek Lamentations belongs to the καίγε tradition . . . It

6 R. Hirsch-Luipold and C. M. Maier, “Threnoi/Die Klagelieder,” in Septuaginta Deutsch.

Das griechische Alte Testament in deutscher Übersetzung (eds. W. Kraus and M. Karrer;
Stuttgart: Deutsche Bibelgesellschaft, 2009), 1349. Cf. also idem, “Threnoi/ Threni Seu
Lamentationes/Die Klagelieder,” in Septuaginta Deutsch. Erläuterungen und Kommentare
zum griechischen Alten Testament. Band II: Psalmen bis Daniel (eds. M. Karrer and W. Kraus;
Stuttgart: Deutsche Bibelgesellschaft, 2011), 2829–2830.
7 Cf. Barthélemy, Les Devanciers d’Aquila, 48–80. In the wake of Barthélemy’s initial
study, scholars have suggested a large number of other features that supposedly also char-
acterise the kaige group. R. T. McLay (“Kaige and Septuagint Research,” Textus 19 [1998]:
127–139) shows that these features are not consistently shared by the members of the
group nor are the differences among these texts sufficiently recognised.
8 P. J. Gentry, “Lamentations,” in A New English Translation of the Septuagint (eds.
A. Pietersma and B. G. Wright; New York / Oxford: Oxford University Press, 2007), 934.
9 Ibid.
two difficult passages in the hebrew texts 277

is clear that some relationship exists between Greek Lamentations and

other texts in the καίγε tradition, but it is not systematic.”
Another interesting issue, relevant to the discussion on the kaige-
membership of LXX Lamentations, is the possibility that it is the work of
Theodotion. This possibility is raised, on the one hand, by the fact that
Origen indicated that the versions of Aquila and Theodotion did not exist
for Lamentations.10 On the other hand, a few readings are attributed to
Aquila in the critical apparatus of Ziegler’s Göttingen edition of Lamen-
tations.11 Gentry12 refers to a study of one of his students, Kevin Young-
blood, who investigated this possibility by comparing and contrasting LXX
Lamentations with the Greek Minor Prophets scroll from Naḥal Ḥever, the
Greek Psalter, Aquila and the materials in the book of Job that are attrib-
uted to Theodotion. He reports that the “comparison between Lamenta-
tions and Theodotion Job was interesting. The evidence showed not only
strong similarities, but also enough differences to make it fairly certain
that the same person did not produce both translations”.13 As a result,
Gentry concludes: “Theodotion Job and OG Lamentations have the same
type of approach to translation, but the Greek Lamentations is probably
not Theodotion”.14
Although the “literal” character of LXX Lamentations seems well estab-
lished, it is necessary and possible to produce a more nuanced profile of
this translation unit.15 One way to achieve this is to study how the Greek
translator rendered passages in the Hebrew text that can be considered

10 Cf. P. J. Gentry, “Old Greek and Later Revisors: Can We Always Distinguish Them,” in
Scripture in Transition. Essays on Septuagint, Hebrew Bible, and Dead Sea Scrolls in Honour
of Raija Sollamo (eds. A. Voitila and J. Jokiranta; Leiden / Boston: Brill, 2008), 326; and
Youngblood, “LXX Lamentations,” 65–66.
11 J. Ziegler, Septuaginta. Vetus Testamentum Graecum Auctoritate Academiae Scien-
tiarum Gottingensis editum XV: Jeremias, Baruch, Threni, Epistula Jeremiae (Göttingen:
Vandenhoeck & Ruprecht, 1976).
12 Gentry, “Old Greek and Later Revisors,” 326.
13 Ibid., 327.
14 Ibid.; L. J. Greenspoon (“The Kaige Recension: The Life, Death, and Postmortem
Existence of a Modern- and Ancient Phenomenon,” in XII Congress of the International
Organization for Septuagint and Cognate Studies, Leiden 2004 [ed. M. K. H. Peters; Atlanta:
Society of Biblical Literature, 2006], 5–16) also argues that the members of the kaige group
cannot always be linked to the version of Theodotion.
15 Kotzé, “The Greek Translation of Lamentations,” 93.
16 This suggestion is based on the conclusions of a previous study: G. R. Kotzé, “Lamen-
tations 4:7 and 4:14. Reflections on the Greek Renderings of the Difficult Hebrew Wordings
of These Verses,” JSem 20/1 (2011): 250–270.
278 gideon r. kotzé

2. Difficult Readings and Textual Criticism

James Barr defines a “difficult” reading as follows:17 A reader

feels that [the reading] ‘does not make sense’. The grammar is “wrong”, i.e.
does not fit with usual patterns of usage. The use of words is anomalous.
Or perhaps the text contradicts what is said elsewhere in the same literary
work, so that it seems to “spoil the effect” of the whole; or it may contradict
something well known from altogether other sources.
Thus, a textual difficulty can pertain to the semantics of words, the gram-
mar and syntax of a passage and/or the content of the passage in relation
to its larger literary contexts.
From the perspective of Old Testament textual criticism, a reader can
deal with such textual difficulties in different ways. Traditionally, the dis-
cipline of Old Testament textual criticism studies the biblical texts, their
development during the processes of transmission (copying and transla-
tion) and evaluates different readings in order to determine which read-
ing is the original one.18 With regard to a textual difficulty, the reader who
follows this approach to textual criticism will, in the first instance, want
to determine whether there are readings in the available textual repre-
sentatives that differ from the difficult reading. If a collation of the extant
Hebrew manuscripts and/or an analysis of the ancient translations, which
were based on Hebrew Vorlagen, reveal the existence of diverging Hebrew
readings, the reader must account for these different readings and how the
difficult reading came into being. An examination of the different readings
may lead the reader to assume that the difficult reading was the result of
some scribal error. Alternatively, the reader can come to the conclusion
that the other readings developed from the difficult reading and therefore,
that this reading, despite its difficulties, represents the original wording of

17 J. Barr, Comparative Philology and the Text of the Old Testament (Oxford: Clarendon
Press, 1968), 3.
18 See E. Tov, Textual Criticism of the Hebrew Bible. Second Revised Edition (Assen / Min-
neapolis: Van Gorcum / Fortress Press, 2001), 289–290. Text-critics formulate their findings
in connection with the evaluation of diverging readings in different ways (Tov, Textual
Criticism of the Hebrew Bible, 310): “Some speak in terms of preferable readings, others refer
to better or (more) original readings, and again others try to identify the reading from
which the other ones presumably derived”. See, for example, A. van der Kooij, “Textual
Criticism of the Hebrew Bible. Its Aim and Method,” in Emanuel. Studies in the Hebrew
Bible, Septuagint, and Dead Sea Scrolls in Honor of Emanuel Tov (eds. S. M. Paul, R. A. Kraft,
L. H. Schiffman and W. W. Fields; Leiden / Boston: Brill, 2003), 729–739; and B. Lemmelijn,
“What Are We Looking For in Doing Old Testament Text-Critical Research?” JNSL 23/2
(1997): 69–80.
two difficult passages in the hebrew texts 279

the passage.19 If the reader is satisfied that none of the readings preserved
in the textual representatives constitute the original text, the wording can
be emended so as to restore its presumed original form. This is referred
to as conjectural emendation, given that the suggested original reading is
conjectured from the actual readings in the textual representatives.
In cases where the Hebrew manuscripts and the Vorlagen of the ancient
translations (as far as their details can be restored with certainty) present
a uniform text, a textual difficulty in the Hebrew wording can be seen as
a copyist error. The reader can then solve the difficulty by means of con-
jectural emendation.20 However, if the reader does not regard the textual
difficulty as an error, attempts can be made to make sense of the difficult
reading with the help of comparative philology.
In the approach to textual criticism, where the available textual rep-
resentatives are treated as witnesses to the original Hebrew texts, the
ancient translations are relevant to the discussion of textual difficul-
ties, insofar as they were based on Hebrew wordings that differ from
the Hebrew text containing the difficulty. The text-critical “significance”
of the variant readings in the ancient translations will, in this approach,
depend on whether they are considered to be more original than the dif-
ficult readings in the Hebrew text. In the philological treatment of tex-
tual difficulties, the ancient translations are used primarily to confirm an
understanding of the Hebrew text that is based on evidence culled from
the cognate languages.21
Methodologically speaking,22 these ways of dealing with the ancient
translations are appropriate to text-critical efforts at editing or elucidating

19 One of the (internal) criteria for evaluating diverging readings is that the more dif-
ficult reading is the preferable one (lectio difficilior praeferanda est). According to R. W.
Klein (Textual Criticism of the Old Testament. From the Septuagint to Qumran [Philadelphia:
Fortress Press, 1974], 75), “Grammatical, historical, theological, and lexical difficulties often
were eliminated or modified by the scribes as they copied the manuscripts. The scribes
would not knowingly insert a more difficult form for a common one or an archaic or rare
word instead of one in everyday usage.” P. K. McCarter (Textual Criticism. Recovering the
Text of the Hebrew Bible [Philadelphia: Fortress Press, 1986], 21) notes that this text-critical
rule of thumb is another way of posing the question: Which reading would have been
more likely to give rise to the others (utrum in alterum abiturum erat)?
20 Barr (Comparative Philology, 3) correctly indicates that conjectural emendations pro-
ceed from the supposition that textual difficulties constitute errors in the wording of a
21 See Barr, Comparative Philology, 238–272, concerning the use of the ancient versions,
especially the LXX, in philological treatments of textual difficulties.
22 On the appropriate methodologies for using the ancient translations in text-critical
research, see, for example, E. Tov, The Text-Critical Use of the Septuagint in Biblical
280 gideon r. kotzé

the Hebrew text of Old Testament writings, especially in cases of prob-

lematic passages. However, text-critical research can also be relevant to
the interpretation of Old Testament writings in other ways than solving
difficulties in the Hebrew wordings and providing exegetes with prob-
lem-free Hebrew texts to base their exegesis on. If the putative original
Hebrew text and the textus receptus are not regarded as the most impor-
tant wordings of an Old Testament writing and therefore, as (practically)
the only legitimate representatives of the writing’s content, the wordings
of the ancient translations gain importance as witnesses to the content
of a biblical book.23 When the ancient translations are studied as repre-
sentatives of an Old Testament writing’s content, the reader must have a
good understanding of why the wordings of the translations look the way
they do.24 Seeing as text-critics investigate the ways in which readings in
the textual representatives were created during the processes of copying
and translation, textual criticism can contribute to such a better under-
standing of the wording of ancient translations and how they present the
content of passages in an Old Testament writing.
Accordingly, this study, which focuses on LXX Lamentations, takes as
its point of the departure the assumption that text-critical analyses of the
Greek renderings of difficult passages in the Hebrew wordings of Lam-
entations can open new vistas on the way in which the Greek translator
translated his Vorlage. This, in turn, can make a small but significant con-
tribution to a better understanding of the wording of LXX Lamentations,
as well as a more nuanced profile of this translation unit.

Research. Revised and Enlarged Second Edition (Jerusalem: Simor, 1997); and J. A. Adair, “A
Methodology for Using the Versions in the Textual Criticism of the Old Testament,” JNSL
20/2 (1994): 111–142.
23 The content of an Old Testament writing refers to what a reader takes the wording
of the writing to convey. In other words, it is the reader’s perception of what the wording
of the writing means.
24 This is one of the lessons learned from rhetorical criticism as practiced by, for instance,
Phyllis Trible. Trible (God and the Rhetoric of Sexuality [Philadelphia: Fortress Press, 1978])
stresses the organic unity of the form or wording of a text and its content: “Form and
content are inseparable. On the one hand, the text is not a container from which ideas or
substance can be abstracted to live an independent life. On the other hand, the text is not
a subject matter from which stylistic and structural wrappings can be removed to exist
autonomously. How the text speaks and what it says belong together in the discovery of
what it is. To convey content is to employ form; to convey form is to employ content”. See
also P. K. Tull, “Rhetorical Criticism and Intertextuality,” in To Each Its Own Meaning. An
Introduction to Biblical Criticisms and Their Applications. Revised and Expanded (eds. S. L.
McKenzie and S. R. Haynes; Westminster: John Knox Press, 1999), 159.
two difficult passages in the hebrew texts 281

3. Method of Analysis

Bearing this assumption in mind, two verses of Lamentations 5, which

contain textual difficulties will be subjected to text-critical analyses. The
problematic passages are Lam 5:10 and 5:13. In each case, the problematic
aspect of the Hebrew wording(s)25 is first identified and discussed before
the Greek text26 is subjected to analysis. The analyses will treat the Greek
texts of Lam 5:10 and 5:13 as both translations and as witnesses to the con-
tent of the verses in their own right.27 The analyses therefore entail exami-
nations of the readings in the Greek text28 that attempt to determine the
most plausible explanations of how these readings were created during
the processes of translation. The analyses also include comparisons of
the Greek translation with other ancient translations. Such comparisons
serve three purposes. First, they highlight unique renderings of difficult
Hebrew words, phrases or clauses in the Greek translation. Secondly, they
reveal instances where ancient translators arrived, independently from
each other, at the same interpretation of the Hebrew text, as well as cases
where the Greek text might have influenced later translations. Thirdly,
similar readings in the Greek text and another ancient translation raise
the possibility that these translations were based on Hebrew Vorlagen
that differ from the extant Hebrew text(s) used in the text-critical analy-
sis. For the purposes of this study, comparisons with the Peshitta29 and

25 For the wording of the MT, I primarily make use of the BHQ fascicle edition:
R. Schäfer, “Lamentations,” in Biblia Hebraica quinta editione cum apparatu critico novis curis
elaborato. General Introduction and Megilloth (eds. A. Schenker et al.; Stuttgart: Deutsche
Bibelgesellschaft, 2004), 54–72, 113*–136*. I also consult the critical apparatuses of BHK and
BHS. Codex Leningradensis is used as the base text in all these editions. With regard to the
Qumran manuscripts, the wordings of some verses of Lamentations 5 were preserved on a
fragmentary manuscript found in cave 5 (5QLama). I make use of the DJD edition prepared
by J. T. Milik (“Lamentations [Premier Exemplaire],” in Les “Petites Grottes” de Qumran:
Exploration de la Falaise, Les Grottes 2Q, 3Q, 5Q, 6Q, 7Q à 10Q, Le Rouleau de Cuivre [DJD III;
M. Baillet, J. T. Milik and R. de Vaux; Oxford: Clarendon Press, 1962], 174–177).
26 The Göttingen edition of Lamentations was prepared by J. Ziegler. I make critical use
of this edition, as well as of A. Rahlfs’s edition of the LXX, which was edited by R. Hanhart:
Septuaginta. Id est Vetus Testamentum graece iuxta LXX interpretes (Stuttgart: Deutsche
Bibelgesellschaft, 2006).
27 Cf. J. Cook, “Exegesis in the Septuagint,” JNSL 3/1 (2004): 2.
28 Given that this study has the work of the original Greek translator of Lamentations in
view, the Old Greek text of LXX Lamentations must be used for the analysis. The Old Greek
text refers to the earliest form of the translation’s wording that can be (re)constructed on
the basis of the available manuscripts and conjectural emendation.
29 B. Albrektson’s critical edition of the Peshitta text of Lamentations serves as the
source for readings of the Syriac translation.
282 gideon r. kotzé

the Vulgate30 will suffice. After the most plausible explanations for the
(Old) Greek readings are identified, the analyses subsequently establish
the effect that the particular readings in the wording of the Greek text
have on the content of the passages in question.

4. Analysis of the Hebrew and Greek Wordings of

Lamentations 5:10 and Lamentations 5:13

4.1. Lamentations 5:10
4.1.1. The Hebrew Wordings

MT ‫ֹעור ֙נּו ְּכ ַתּנ֣ ּור נִ ְכ ָ֔מרּו ִמ ְּפ ֵנ֖י זַ ְל ֲע ֹ֥פות ָר ָ ֽעב׃‬
Our skin is wrinkled like an oven, because of the rages of hunger.

5QLama ‫‍עורי֯ [נ]ו֯ ֯כתנור נכמרו ֯מ[פ]ני ז֯ לפות ֯רעב‬

[Our] skins are wrinkled like an oven, be[cau]se of the rages31 of hunger.

The wordings of Lam 5:10 are almost identical in 5QLama and the MT.
In the Masoretic manuscript, there is incongruence between the verb
‫נִ ְכ ָמרּו‬, which is plural, and its subject, ‫עֹורנּו‬
ֵ , which is singular. This dis-
crepancy in number was corrected by the scribe who copied 5QLama. He
initially copied the opening word of the verse as ‫עורנו‬, but then (presum-
ably after copying ‫ )נכמרו‬inserted a yôd in the interlinear space after the
rêš of ‫עורנו‬. This scribal correction changes the suffix from one denoting
a singular noun to one indicating a plural noun;32 therefore, the word
can be restored as ‫עורינו‬. The form ‫ עורינו‬is also found in a number of

30 For the Vulgate, I use the fifth edition of R. Weber edited by R. Gryson (Biblia Sacra
iuxta Vulgatam Versionem [Stuttgart: Deutsche Bibelgesellschaft, 2007]).
31  E. Y. Kutcher (A History of the Hebrew Language [Jerusalem: Magness Press, 1982],
96) notes that gutturals weakened in Qumran Hebrew. These consonants were sometimes
confused with one another or elided altogether. See also M. G. Abegg, “The Linguistic
Analysis of the Dead Sea Scrolls: More Than (Initially) Meets the Eye,” in Rediscovering
the Dead Sea Scrolls. An Assessment of Old and New Approaches and Methods (ed. M. L.
Grossman; Grand Rapids: Eerdmans, 2010), 59; and E. Qimron, The Hebrew of the Dead Sea
Scrolls (Atlanta: Scholars Press, 1986), 25–26. The reading ‫ ז֯ לפות‬in 5QLama is an example
of the elision of the pharyngeal ‘áyĭn.
32 See Schäfer,