John Lee, Collected Essays On The Greek Bible and Greek Lexicography

BIBLICAL EXEGESIS & THEOLOGY 112
C O N T R I BU T I O N S TO
John A. L. Lee
Collected Essays on
the Greek Bible
and Greek
Lexicography
PEETERS
COLLECTED ESSAYS ON THE GREEK BIBLE AND GREEK
LEXICOGRAPHY
CONTRIBUTIONS TO BIBLICAL EXEGESIS AND THEOLOGY
SERIES EDITORS
K. De Troyer (Salzburg)
G. Van Oyen (Louvain-la-Neuve)
ADVISORY BOARD
Reimund Bieringer (Leuven)

Lutz Doering (Münster)
Mark Goodacre (Duke)
Bas ter Haar Romeny (Amsterdam)
Annette Merz (Groningen)
Madhavi Nevader (St Andrews)
Thomas Römer (Lausanne)
Jack Sasson (Nashville)
Tammi Schneider (Claremont)
John A. L. LEE
COLLECTED ESSAYS ON THE GREEK

BIBLE AND GREEK LEXICOGRAPHY
PEETERS
LEUVEN – PARIS – BRISTOL, CT
2022
A catalogue record for this book is available from the Library of Congress.
© 2022 — Peeters, Bondgenotenlaan 153, B-3000 Leuven
ISBN 978-90-429-4806-8
eISBN 978-90-429-4807-5
D/2022/0602/33
All rights reserved. No part of this publication may be reproduced, stored in a

retrieval system, or transmitted, in any form or by any means, electronic, mechan-
ical, photocopying, recording or otherwise, without the prior permission of the
publisher.
PREFACE
Whether one ought to publish a collection of one’s essays is a question

not always to be answered in the affirmative. The author may not be the
best person to judge. Everything seems ‘important.’ The old standby, ‘at the
urging of many friends,’ would not serve in this case (if it ever did): only
Jim Aitken urged me to do it; Trevor Evans was supportive. I thank them
both. Later, I was greatly encouraged by the welcoming response of Kristin
De Troyer as editor of the CBET series, and of Peeters of Leuven.
I have had to disappoint some of my darlings. Not everything could be
included anyway, for space and copyright reasons, and others were out of
date or too slight to be worth reviving. But others were worth including
and I am sorry to lose them; perhaps a second collection might be pos-
sible. I see that one essay, number 11, was almost a monograph in itself;
yet it is only the beginning of what awaits in that inviting field of study.
The papers may seem ποικίλα, but through them runs a single thread:
Greek language. Not lexicography alone, not the Septuagint and the New
Testament as such, nor Renaissance Greek, but the study of Greek itself
and what it brings to other areas of scholarship.
The papers are published unrevised, except for formating and style
changes and occasional improvements in wording. I have added small
updates here and there in square brackets, and longer postscripts at the end
of some essays.
My best thanks are due to Anneli Aejmelaeus for agreeing to write an
introduction, both an honour to me and a token of friendship; to Elisabeth
Hernitscheck of Peeters for her meticulous help with preparation and sub-
mission of copy; and to my wife Diane, who has always understood and
encouraged my need to keep at work.
John Lee
University of Sydney
October, 2021
TABLE OF CONTENTS
PREFACE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . V
ABBREVIATIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . IX
INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . XV
1 A Note on Septuagint Material in the Supplement to Liddell and

Scott . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
2 Aposkeuê in the Septuagint . . . . . . . . . . . . . . . . . . . 11
3 Equivocal and Stereotyped Renderings in the LXX . . . . . . 21
4 The Future of zên in Late Greek . . . . . . . . . . . . . . . . 37
5 Some Features of the Speech of Jesus in Mark’s Gospel. . . . 47
6 A Non-Aramaism in Luke 6:7 . . . . . . . . . . . . . . . . . 71
7 The United Bible Societies’ Lexicon and its Analysis of Meanings 79
8 A Lexical Study Thirty Years On, with Observations on ‘Order’
Words in the LXX Pentateuch . . . . . . . . . . . . . . . . . 105
9 The Present State of Lexicography of Ancient Greek . . . . . 117
10 Review of Muraoka, A Greek-English Lexicon of the Septuagint:
Chiefly of the Pentateuch and the Twelve Prophets (2002) . . 129
11 Dimitrios Doukas and the Accentuation of the New Testament
Text of the Complutensian Polyglot. . . . . . . . . . . . . . . 143
12 Exapostellô . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183
13 A Lexicographical Database for Greek: Can It Be Far Off? The
Case of ἄμφοδον . . . . . . . . . . . . . . . . . . . . . . . . 199
14 The Complutensian Polyglot, the Text of Sirach, and a Lost
Greek Word . . . . . . . . . . . . . . . . . . . . . . . . . . . 207
15 Review of Muraoka, A Greek-English Lexicon of the Septuagint
(2009) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223
16 Auxiliary thelô . . . . . . . . . . . . . . . . . . . . . . . . . . 237
17 Releasing LSJ From Its Past. . . . . . . . . . . . . . . . . . . 257
18 Why Didn’t St Basil Write in New Testament Greek? . . . . . 279
19 The Atticist Grammarians . . . . . . . . . . . . . . . . . . . . 293
20 Etymological Follies: Three Recent Lexicons of the New
Testament . . . . . . . . . . . . . . . . . . . . . . . . . . . . 319
21 The onoma Rule . . . . . . . . . . . . . . . . . . . . . . . . . 339
VIII TABLE OF CONTENTS
22 The Literary Greek of Septuagint Isaiah . . . . . . . . . . . . 351

23 The Sayings of the Desert Fathers: Their Evidence for Late
Koine Greek . . . . . . . . . . . . . . . . . . . . . . . . . . 373
24 The Vocabulary of the LXX and Documentary Evidence . . . 389
25 The Puzzle of John 21:15–17: A Formality Solution . . . . . 403
26 The Brill Dictionary of Ancient Greek: New Clothes for an Old
Emperor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 407
27 Greek Idiom in the LXX-Pentateuch: The Preposition para . . 427
28 Back to the Question of Greek Idiom . . . . . . . . . . . . . . 441
BIBLIOGRAPHY . . . . . . . . . . . . . . . . . . . . . . . . . . . . 455
NOTICE OF ORIGINAL PUBLICATION . . . . . . . . . . . . . . . . . . 479
INDEX OF GREEK WORDS . . . . . . . . . . . . . . . . . . . . . . . 483
INDEX OF HEBREW WORDS . . . . . . . . . . . . . . . . . . . . . . 489
INDEX OF BIBLICAL REFERENCES . . . . . . . . . . . . . . . . . . . 491
INDEX OF MODERN AUTHORS . . . . . . . . . . . . . . . . . . . . . 497
ABBREVIATIONS
Abbreviations for Greek authors and works are those used by LSJ or an
expanded version thereof. Papyrological publications follow Oates et al.,
Checklist, and epigraphical publications follow Horsley and Lee, ‘A Pre-
liminary Checklist.’ References to text numbers in PCairZen follow LSJ’s
practice and omit the first 59(00). All works not listed below are referred
to by author and short title (see bibliography). Abbreviations of biblical
books, standard works, and series titles follow SBLHBS.
BAG = W. Bauer, trans. W. F. Arndt and F. W. Gingrich, Greek-English Lexi-

con of the New Testament (Chicago and Cambridge, 1957).
BAGD = W. Bauer, W. F. Arndt, F. W. Gingrich, and F. W. Danker, A Greek-
English Lexicon of the New Testament and Other Early Christian Literature
(2nd ed.; Chicago, 1979).
Bauer2 (1928) = W. Bauer, Griechisch-Deutsches Wörterbuch zu den Schriften
des Neuen Testaments (2nd ed.; Gießen, 1928).
BDAG = W. Bauer, F. W. Danker, W. F. Arndt, and F. W. Gingrich, A Greek-
English Lexicon of the New Testament and Other Early Christian Literature
(3rd ed.; Chicago, 2000).
BDB = F. Brown, S. R. Driver, and C. A. Briggs (eds.), A Hebrew and English
Lexicon of the Old Testament (Oxford, 1907).
BDF = F. Blass and A. Debrunner, A Greek Grammar of the New Testament and
Other Early Christian Literature (trans. and rev. R. W. Funk; Cambridge,
1961).
BDR = F. Blass and A. Debrunner, Grammatik des neutestamentlichen Griechisch
(rev. F. Rehkopf; Göttingen, 1975).
Beekes, EDG = R. Beekes, with L. van Beek, Etymological Dictionary of Greek
(2 vols.; Leiden, 2010).
BHS = Biblia Hebraica Stuttgartensia (Stuttgart, 1983).
Biblia Sacra = Biblia Sacra: Iuxta Vulgatam Versionem, ed. R. Weber et al. (3rd ed.;
Stuttgart, 1983).
BM = A. E. Brooke and N. McLean, The Old Testament in Greek (3 vols.;
Cambridge, 1906–1940).
BRAA = W. Bauer, Griechisch-Deutsches Wörterbuch zu den Schriften des Neuen
Testaments (6th ed.; rev. V. Reichmann, K. Aland, and B. Aland, Berlin,
1988).
Brenton = [L. C. L. Brenton], The Septuagint Version of the Old Testament, with
an English Translation (London, [1844]).
CGL = J. Diggle et al. (eds.), The Cambridge Greek Lexicon (2 vols.; Cambridge,
2021).
X ABBREVIATIONS
CHB = S. L. Greenslade (ed.), The Cambridge History of the Bible, Vol. 3: The
West from the Reformation to the Present Day (Cambridge, 1963).
Chantraine, DELG = P. Chantraine, Dictionnaire étymologique de la langue grecque:
histoire des mots (Paris, 1968–1980; new ed. with supplement 2009).
Class. = Classical.
DDBDP = Duke Databank of Documentary Papyri.
DGE = F. R. Adrados et al. (eds.), Diccionario Griego-Español (7 vols. to date;
Madrid, 1980–2009).
DGENT = J. Mateos et al. (eds.), Diccionario Griego-Español del Nuevo Testa-
mento (5 fascs. to date; Córdoba, 2000–2012).
Dizionario Biografico = A. M. Ghisalberti (ed.), Dizionario Biografico degli
Italiani (Rome, 1960–2020).
DMLBS = R. E. Latham, then D. R. Howlett et al. (eds.), Dictionary of Medieval
Latin from British Sources (London, 1975–2013).
Engl. = English.
EDNT = H. Balz and G. Schneider (eds.), Exegetical Dictionary of the New
Testament (3 vols.; Grand Rapids, 1990–1993).
Fraser, LGPN, IIIA = P. M. Fraser, E. Matthews et al. (eds.), A Lexicon of Greek
Personal Names (vol. IIIA; Oxford, 1997).
Frisk, GEW = H. Frisk, Griechisches Etymologisches Wörterbuch (3 vols.;
Heidelberg, 1960–1972).
GE = F. Montanari, et al. (eds.), The Brill Dictionary of Ancient Greek (Leiden,
2015).
GI = F. Montanari, Vocabolario della Lingua Greca (GI) (1st ed.; Torino, 1995;
2nd ed. 2004; 3rd ed. 2013).
Göttingen ed. / Gött. = Septuaginta Vetus Testamentum Graecum: Auctoritate
Academiae Scientiarum Gottingensis editum.
HALOT = L. Koehler and W. Baumgartner, The Hebrew and Aramaic Lexicon of the
Old Testament (rev. W. Baumgartner, J. J. Stamm et al., trans. M. E. J. Richard-
son; 2 vols.; Leiden, 2001).
HDM, La Bible grecque = M. Harl, G. Dorival, and O. Munnich, La Bible grecque
des Septante: Du judaïsme hellénistique au christianisme ancien (Paris, 1988).1
Heb. = Hebrew
HR = E. Hatch and H. A. Redpath, A Concordance to the Septuagint and the
Other Greek Versions of the Old Testament (Oxford, 1897).
HR Suppl. = HR Supplement, by H. A. Redpath (Oxford, 1906), Part II: Concord-
ance to Portions of Ecclesiasticus with Hebrew Equivalents.
IE = Indo-European.
KB, Lexicon = L. Koehler and W. Baumgartner, Lexicon in veteris testamenti
libros (Leiden, 1958).
KB, Grammatik, I = R. Kühner and F. Blass, Ausführliche Grammatik der griechi-
schen Sprache, I.1, 2 Elementar- und Formenlehre (Hannover, 1890, 1892).
KG, Grammatik, II = R. Kühner and B. Gerth, Ausführliche Grammatik der grie-
chischen Sprache, II.1, 2 Satzlehre (Hannover and Leipzig, 1898, 1904).
1
The title page has the names in the order Dorival, Harl, Munnich, but the front cover
has Harl, Dorival, Munnich. I have chosen the latter τιμῆς ἕνεκα Marguerite Harl.
ABBREVIATIONS XI
KJV = King James Version (1611).

Kriaras, Λεξικό = E. Kriaras, Λεξικὸ τῆς μεσαιωνικῆς Ἑλληνικῆς δημώδους γραμ-
ματείας 1100–1669 (14 vols. to date; Thessaloniki, 1968–1997).
Lampe, PGL = G. W. H. Lampe, (ed.), A Patristic Greek Lexicon (Oxford, [1968]).
Lat. = Latin.
LBG = E. Trapp et al. (eds.), Lexikon zur byzantinischen Gräzität: Besonders
des 9.–12. Jahrhunderts (8 fascs.; Vienna, 2017).
LEH, Lexicon = J. Lust, E. Eynikel, and K. Hauspie, Greek-English Lexicon of
the Septuagint: Revised Edition (2 vols.; Stuttgart, 2003).
Lewis and Short = C. T. Lewis and C. Short, A Latin Dictionary (Oxford, 1879).
Lex. Acad. = Λεξικὸν τῆς Ἑλληνικῆς Γλώσσης. Αʹ. Ἱστορικὸν Λεξικὸν τῆς Νέας
Ἑλληνικῆς, τῆς τε κοινῶς ὁμιλουμένης καὶ τῶν ἰδιωμάτων (5 vols.; Athens,
1933–1989).
Lex. Proia = Λεξικόν της Ελληνικής Γλώσσης (Athens, 1933).
LN = J. P. Louw and E. A. Nida et al., Greek-English Lexicon of the New Testa-
ment: Based on Semantic Domains (2 vols.; New York, 1988). [In paper 7]
LS1 (1843) = H. G. Liddell and R. Scott, A Greek-English Lexicon: Based on
the German work of Francis Passow (Oxford, 1843).
LS1 Abridged = [G. Marshall], A Lexicon Chiefly for the Use of Schools Abridged
from the Greek-English Lexicon of H. G. Liddell, M.A., and R. Scott, M.A.
(Oxford, 1843; 28th ed. 1903).
LSJ = H. G. Liddell and R. Scott, A Greek-English Lexicon (rev. and augm.
H. Stuart Jones; Oxford, 1940).
LSJ Suppl. (1968) = H. G. Liddell, R. Scott, and H. Stuart Jones, Greek-English
Lexicon: A Supplement, ed. E. A. Barber et al. (Oxford, 1968).
LSJ Suppl. (1996) = H. G. Liddell and R. Scott, A Greek-English Lexicon: Revised
Supplement, ed. P. G. W. Glare, with assist. of A. A. Thompson (Oxford,
1996).
LXX.D = W. Kraus and M. Karrer (eds.), Septuaginta Deutsch: Das griechische
Alte Testament in deutscher Übersetzung (Stuttgart, 2009).
LXX.D Komm. = M. Karrer and W. Kraus (eds.), Septuaginta Deutsch: Erläuterun-
gen und Kommentare zum griechischen Alten Testament (2 vols.; Stuttgart,
2011).
LXX.H, 3 = E. Bons and J. Joosten (eds.), Handbuch zur Septuaginta: Handbook
of the Septuagint, vol. 3 Die Sprache der Septuaginta: The Language of the
Septuagint (Gütersloh, 2016).
Mega Lexikon = I. S. Zervos (ed.), Μέγα Λεξικόν της Ελληνικής Γλώσσης (9 vols.;
Athens, 1953).
Moulton, Grammar, I = J. H. Moulton, A Grammar of New Testament Greek,
vol. I Prolegomena (Edinburgh, 1908).
MH, Grammar, II = J. H. Moulton and W. F. Howard, A Grammar of New Tes-
tament Greek, vol. II Accidence and Word-Formation (Edinburgh, 1919–
1929).
MHT, Grammar, III = J. H. Moulton, A Grammar of New Testament Greek, vol. III
Syntax, by N. Turner (Edinburgh, 1963).
MHT, Style = J. H. Moulton, A Grammar of New Testament Greek, vol. IV Style,
by N. Turner (Edinburgh, 1976).
XII ABBREVIATIONS
Migne, PG = J.-P. Migne, (ed.), Patrologiae cursus completus: Series Graeca

(162 vols.; Paris, 1857–1866).
Migne, PL = J.-P. Migne, (ed.), Patrologiae cursus completus: Series Latina
(217 vols.; Paris, 1844–1864).
MM = J. H. Moulton and G. Milligan, The Vocabulary of the Greek Testament
Illustrated from the Papyri and Other Non-Literary Sources (5 vols. 1914–
1929; one-vol. edition London, 1930).
MT = Masoretic Text.
NA28 = B. and K. Aland et al. (eds.), Novum Testamentum Graece (28th ed.;
Stuttgart, 2012).
NETS = A. Pietersma and B. G. Wright (eds.), A New English Translation of the
Septuagint and other Greek Translations Traditionally Included under that
Title (New York/Oxford, 2007).
New Documents = G. H. R. Horsley et al. (eds.), New Documents Illustrating
Early Christianity (vols. 1–6; Macquarie University, 1981–1992; vol. 7 ed.
S. R. Llewelyn, 1994).
OED = Oxford English Dictionary (2nd ed.; 20 vols.; Oxford, 1989).
OGELD = D. N. Stavropoulos (ed.), Oxford Greek-English Learner’s Dictionary
(Oxford, 1988).
OLD = P. G. W. Glare (ed.), Oxford Latin Dictionary (Oxford, 1968–1982).
Pauly-Wissowa, RE = Paulys Real-Encyclopädie der classischen Altertumswis-
senschaft (50 vols.; Stuttgart, 1894–1980).
Pent. = Pentateuch.
Preisigke, Wörterbuch = F. Preisigke and E. Kiessling, Wörterbuch der griechi-
schen Papyrusurkunden (vols. 1–3 by Preisigke, Berlin, 1925–31; vol. 4 by
Kiessling et al., Berlin/ Marburg/ Wiesbaden, 1944–1993).
Rahlfs / Ra. = A. Rahlfs, Septuaginta: Id est Vetus Testamentum graece iuxta
LXX interpretes (rev. R. Hanhart; Stuttgart, 2006).
RBLG = P. Boned Colera et al. (eds.), Repertorio Bibliográfico de la Lexicografía
Griega (Madrid, 1998).
SBLHS = The SBL Handbook of Style (2nd ed.; Atlanta, 2014).
Schwyzer, Grammatik, I = E. Schwyzer, Griechische Grammatik, I. Allgemeiner
Teil. Lautlehre. Wortbildung; Flexion (Munich, 1938).
Schwyzer, Grammatik, II = E. Schwyzer, and A. Debrunner, Griechische Gram-
matik, II. Syntax und syntaktische Stilistik (Munich, 1950).
Shipp, MGE = G. P. Shipp, Modern Greek Evidence for the Ancient Greek
Vocabulary (Sydney, 1979).
Th. = Theodotion.
Thomson-Muses = C. Thomson, The Septuagint Bible (rev. C. A. Muses; 2nd ed.;
Indian Hills, 1960).
TLG = Thesaurus Linguae Graecae (online at <stephanus.tlg.uci.edu>).
TLG Canon = L. Berkowitz, and K. A. Squitier, Thesaurus Linguae Graecae:
Canon of Greek Authors and Works (3rd ed.; New York, 1990).
TLL = Thesaurus Linguae Latinae (Leipzig, 1900–).
Vetus Latina: Sirach = Vetus Latina: Die Reste der Altlateinischen Bibel, Band 11/2
Sirach (Ecclesiasticus), 9. Lieferung Sir 23,7–24,47; Register, ed. Walter
Thiele (Freiburg, 2005).
ABBREVIATIONS XIII
Note abbreviations of the following biblical books:

1 Sam = 1 Samuel (MT) = 1 Kgdms = 1 Kingdoms (LXX)
2 Sam = 2 Samuel (MT) = 2 Kgdms = 2 Kingdoms (LXX)
1 Kgs = 1 Kings (MT) = 3 Kgdms = 3 Kingdoms (LXX)
2 Kgs = 2 Kings (MT) = 4 Kgdms = 4 Kingdoms (LXX)
Siglum
~ = corresponds to
INTRODUCTION
By Anneli AEJMELAEUS
Reading through the essays of this collection has been for me like an
adventure through well-familiar terrain but taking new paths, going through
thickets and rocky ground, and seeing everything from a fresh angle. Ever
since I got my hands on the freshly printed book titled A Lexical Study
of the Septuagint Version of the Pentateuch (1983), by a scholar named
J.A.L. Lee, I have been a great fan of this scholar (whose first name I
learned only decades later), who elucidated the language used by the Sep-
tuagint translators so impressively from contemporary sources, either newly
discovered or at long last thoroughly examined. It was a major opening in
Septuagint scholarship, and for me, it was like a revelation that gave a dif-
ferent perspective on the Septuagint translators and their milieu in Hellenistic
Alexandria, very different from my own translation-technical approach that
concentrated on the Hebrew-Greek correspondence of syntactical features.
Reading these essays gave me a vivid recollection of that early eye-opening
experience.
This collection contains essays written over a period of five decades,
organized in their order of appearance. It displays a kind of cross-section of
a long and successful scholarly career and finely complements the book-
form production of the author. The essays also reflect the great changes
that have taken place in the conditions and practices of scholarly work
during this long career. The themes discussed in the collection range from
classical Greek to koine and biblical Greek, from lexicography to the cre-
ation of databases, from the analysis and history of single words to assess-
ments of stylistic levels of expressions, and from Atticism to the mysteries
of the Complutensian polyglot. The strength of John Lee is that nothing
Greek is without interest for him. Whatever the starting point, his unparal-
leled knowledge of the Greek language, his thorough familiarity with texts
from different periods, as well as his perseverance in going through all the
relevant material allow him to draw conclusions on the development of
the language through centuries, to place linguistic phenomena on that con-
tinuum, and to elucidate the evolving meanings and uses of expressions
against that background.
XVI INTRODUCTION
There is no need to give a summary of the contents of the essays at

hand in this introduction. Each essay begins with an abstract introducing
its aims and results. At the end of each essay, the results and further con-
clusions are once more collected in a summary. To some essays, the author
has added actualizing comments concerning discussions that were sparked
by the essay or interesting details about the process of writing. At this point,
it is befitting to highlight the major strands in the collection and in John
Lee’s scholarly work.
The collection opens and closes with essays on Septuagint Greek, which
seems to have been a central object of John Lee’s enthusiasm and scholar-
ship from early on and throughout his career. His specialty is to look at the
language produced by the Septuagint translators as Greek representative
of contemporary koine Greek usage. His ambition is to match words and
expressions that have been labelled as special Septuagint language or trans-
lationese with parallels from genuine Greek texts, contemporary documen-
tary sources as well as literature. At the same time, unlike many others who
interest themselves in the Septuagint per se, he always pays due attention
to the Hebrew source text in order to be able to differentiate between inter-
ference from the underlying Hebrew and expressions of the translators’
Greek usage and competence. He works like a true detective, and the results
are both convincing and amazing. The availability of contemporary koine
texts is naturally a decisive prerequisite for this kind of research, but it
takes a scholar like John Lee not to be satisfied with mere sampling but to
pursue the study as far towards perfection as possible. Of course, the com-
puters that arrived on the scene midway through his career and the ever-
growing number of texts in electronic searchable form have had a tremen-
dous effect on this project.
Moreover, the results of this kind of thorough investigation can be made
fruitful in different ways: in the linguistic and lexical study of the Sep-
tuagint as well as for using the Septuagint as evidence for contemporary
koine Greek, in the relative dating of the various parts of the Septuagint
based on the development of the language, in the assessment of the influ-
ence of the Septuagint on New Testament Greek, in describing the devel-
opment of the meaning of vocabulary in focus throughout the history of
the Greek language, not to speak of the benefits to numerous other areas
of biblical study, not least of all textual criticism. The essays at hand can be
warmly recommended to scholars and students of the Greek Bible what-
ever their area of study.
Mutatis mutandis, the principles that govern research on Septuagint Greek
apply to the linguistic study of the New Testament as well. In each case,
the relationship between biblical Greek and contemporary Greek usage is
INTRODUCTION XVII
in principle the same, just with a difference in time period. Corresponding

to the possible interference from the underlying Hebrew in the Septuagint,
the New Testament carries the possibility of influence from the Septua-
gint, which was a sacred text for New Testament writers, a text that was
frequently studied as well as quoted, alluded to, and imitated. In both cases,
John Lee underlines the priority of the criterion of whether similar usage can
be detected in contemporary sources. Old inherited explanations of features
in New Testament Greek, and corresponding definitions of meaning, that
claim Semitic/Aramaic or Septuagint influence are repeatedly found to be in
need of correction in the light of contemporary parallels.
An enormous amount of time and energy has been invested by John Lee
in examining various lexica of New Testament and Septuagint Greek. In
the case of the Septuagint, the history of lexicography is much shorter
than in the case of the New Testament, to which John Lee has dedicated
a whole book. Several essays in this collection deal with various dictionar-
ies of the Septuagint, of the New Testament, as well as of ancient Greek
in general. Writing a review article on a lexicon is not a task for anybody,
but John Lee is matchless in this regard. He is thorough and critical in
his scrutiny of each lexicon and strict in his argumentation, illustrating his
points by rich examples. He brings to light all the shortcomings and incon-
sistences in the presentation of the material as well as in adherence to the
principles of definition and glossing. He can be very sharp in his criticism,
but he can also express it in a gentle tone, reminding the reader again and
again that he does not wish so much to criticize or blame as to help improve
the lexicon for a new edition and to enhance the level of ancient Greek
lexicography.
Every student of languages has a high regard for dictionaries, and they
are no doubt an important tool for working especially on texts in classi-
cal languages. But to what extent can we rely on them? Discussions on
the various dictionaries reveal the complexities of producing a dictionary
as well as the very human nature of most of the available products with
which we need to get along as students and scholars. It is striking to see
how most dictionaries are dependent on their predecessors, repeating the
glosses and quotations given in them and taking for granted questionable
explanations of rare cases and even downright erroneous definitions of
meaning. It is important for a dictionary-user to be aware of such problems,
to know what to expect from a dictionary and where to suspect deficiencies.
While reviewing existing lexica, John Lee proceeds in a most pedagogical
way, giving ample evidence and clear arguments for each of his points,
offering sample entries and explaining the principles to be followed, so
that the reader actually receives a complete introduction into lexicography.
XVIII INTRODUCTION
At the same time, a wide range of language phenomena are investigated.

Theory goes hand in hand with praxis.
The essays at hand reveal a deep concern for the future of ancient Greek
lexicography. Dictionaries in book-form have obviously come to the end
of their time, at least as comprehensive treasuries of the Greek language.
When envisioning the future of Greek lexicography, John Lee suggests
that a database be founded for the collection of all relevant material, pieces
of research on Greek vocabulary scattered in different publications and
information on the occurrence and usage in the various texts of different
periods of the Greek language. In fact, he already proposed this a quarter
of a century ago and has repeated it at regular intervals. Such a database
would be the only sensible solution for the advancement of ancient Greek
lexicography, but it might take some time to find an agent powerful enough
to organize such an immense endeavour. In the meantime, the work could
be launched on a smaller scale limited to certain periods or literatures,
and the various databases that come about this way could ultimately add up
to a real “Thesaurus Linguae Graece.” As for the area of Septuagint Greek,
John Lee emphasizes that he believes this to be the new desideratum of
Septuagint studies.
Even a database that only compiles all the existing material for the Greek
language of the Septuagint is a huge task. Nevertheless, this is no doubt the
sole correct way to proceed. One major obstacle to the realization of this
proposal might be the present-day science policy that does not favour long-
term projects. This is really a paradox. Now that we have the technology
to handle all that huge amount of data much faster and more effectively
than the previous generations that have worked from one century to another
on book-form dictionaries almost beyond the size that can be handled by
human hands, modern scholars are rather expected to concentrate on short-
term projects, and such fundamental research is no longer in fashion. And
yet, the need for a new approach to ancient Greek lexicography, or any
part of it, is obvious, as demonstrated by John Lee. Let us hope that this
collection of essays finds its way into the hands of young scholars who get
inspired by it and are bold enough to begin to realize this proposal.
It remains for me to congratulate John Lee on his long and successful
career – half a century so far – as well as on his forthcoming milestone
birthday. I wish him many happy returns and hope that we shall all soon
see his dream come true.
Anneli Aejmelaeus
1
A NOTE ON SEPTUAGINT MATERIAL IN
THE SUPPLEMENT TO LIDDELL AND
SCOTT
1969
Abstract
In 1968 a Supplement to LSJ was published, containing 153 pages of ‘correc-

tions and additions’ to LSJ (1940) that had accumulated in the hands of Oxford
University Press up to that time. Among them were many insertions and changes
related to the LXX. This review examines (some of) the new LXX material. A
number of methodological faults are detected and demonstrated by examples. The
most notable is that of attributing to a Greek word, without adequate justification,
a new meaning derived from the meaning of the Hebrew word it translates.
In the recently published Supplement to Liddell and Scott’s Greek-

English Lexicon there are a large number of entries introducing new mate-
rial from the LXX. Below are discussed certain of these which are open to
criticism, less in points of detail than as regards the method and assump-
tions underlying the treatment of the words concerned.
It is a basic principle of lexicography that in order to establish the exist-
ence of a new sense of a given word incontrovertible examples of that
sense must be found. So long as the word can be understood in one of
its established senses without undue strain, it ought to be classified under
that sense. Only if the new sense is clearly demanded by the context can
it be regarded as definitely established. The authors of the Supplement
have, for particular reasons, introduced unjustified new senses for a num-
ber of words in neglect of this principle.
Thus under ψυχή an additional sense is given which arises from the
error of including as part of the indicator value of a word a notion expressed
by the context in which it is used.
2 COLLECTED ESSAYS ON THE GREEK BIBLE
Add ‘VIII. slave, πᾶσαν ψυχήν LXX Ge. 12.5, 1Ma. 10.33; dub. sens.,
ψυχὰς ἀνθρώπων Apoc. 18.13.’
The two LXX passages are as follows:

Gen 12:5 καὶ ἔλαβεν Ἀβρὰμ τὴν Σάραν γυναῖκα αὐτοῦ καὶ τὸν Λὼτ
υἱὸν τοῦ ἀδελφοῦ αὐτοῦ καὶ πάντα τὰ ὑπάρχοντα αὐτῶν, ὅσα ἐκτή-
σαντο, καὶ πᾶσαν ψυχήν, ἣν ἐκτήσαντο ἐν Χαρράν, καὶ ἐξήλθοσαν.
1 Macc 10:33 (Ra.) καὶ πᾶσαν ψυχὴν Ιουδαίων τὴν αἰχμαλωτισθεῖσαν
ἀπὸ γῆς Ιουδα εἰς πᾶσαν βασιλείαν μου ἀφίημι ἐλευθέραν δωρεάν.
‘Slave’ is a possible translation of ψυχήν in these contexts, but it does

not follow that ψυχή had itself developed a new sense ‘slave.’ Such a
development would not be impossible (cf. σῶμα in late Greek), and it
could have originated in transitional contexts such as these, but there is
no evidence whatever for it. What ψυχή refers to [emphasis added] in
these two places (household slaves; captives in war) is indicated by other
words in the sentence, not by ψυχή itself, which can be readily trans-
lated ‘soul’ or ‘person.’ Unless an example of ψυχή meaning ‘slave’ quite
apart from the indications given by other words in the context can be
found, the listing of these instances under a new heading is unjustified;
they belong with the other examples of ψυχή referring to persons, in
section IV.1.
Moreover, in Gen 12:5 ψυχή is, as very frequently throughout the LXX,
simply a literal translation of the Hebrew word ‫נפשׁ‬, ‘soul,’ ‘person.’ The
same is probably true of the example in 1 Macc, the Hebrew original of
which is now lost. The inclusion of the difficult Rev 18:13 is misleading.
ψυχάς cannot have the sense ‘slaves’ there. It is the whole phrase ψυχὰς
ἀνθρώπων that may mean ‘slaves.’
The same underlying error is seen in the following entries. S.v. ἐκκαίω:
Add ‘IV. ἐ. ὀπίσω τινός pursue with fierce enmity, LXX 3 Ki. 20(21).21.’
The passage is:

3 Kgdms 20(21):21 (Ra.) ἰδοὺ ἐγὼ ἐπάγω ἐπὶ σὲ κακὰ καὶ ἐκκαύσω
ὀπίσω σου (‫ )ובערתי אחריך‬καὶ ἐξολεθρεύσω.
The rendering ‘pursue with fierce enmity,’ though no doubt a possible
translation of the Greek, obscures the actual value of ἐκκαύσω, which
is clearly intended to be a literal rendering of the Hebrew word, meaning
‘burn,’ etc.1 If there is any idea of pursuit it is conveyed by the combining
of ἐκκαύσω with ὀπίσω σου, not by ἐκκαύσω itself; and the idea of
1
‫ בער‬piel, ‘kindle, burn, consume, devastate,’ BDB.
A NOTE ON SEPTUAGINT MATERIAL 3
enmity is suggested by the general tone of the passage. Admittedly the

construction in which ἐκκαίω is here used is unusual for Greek, but that
is a result of the word-for-word rendering of the original. There is no new
sense of ἐκκαίω here.2
S.v. ἐπίπεμπτος. For the present section II we are to substitute:
ἐπίπεμπτον, τό, one-fifth of the votes in a trial, Ar. Fr. 201, cf. Eup. 65.
2. a fifth in addition, LXX Le. 5.16; τὸ ἐ. a fifth as a fine, IG 7.3073.1
(Lebadea, ii B.C.).
Lev 5:16 reads as follows:

καὶ ὃ ἥμαρτεν ἀπὸ τῶν ἁγίων, ἀποτείσαι, καὶ τὸ ἐπίπεμπτον προσθή-
σει ἐπ’ αὐτὸ καὶ δώσει αὐτὸ τῷ ἱερεῖ.
That the fifth is additional is indicated by προσθήσει. The use of ἐπί-

πεμπτον here is no different from that in 1. and ought to be listed there.
So also the IG example, which, as the Supplement’s own translation shows,
does not differ from sense 1.
II
A somewhat different error of method, that of supposing without ade-

quate reason that the Greek word has the same meaning as the Hebrew
word it renders, has produced unwarranted new meanings in other entries
in the Supplement. A typical example is found under κόλπος:
III, after ‘6.’ insert ‘bowl or dish, LXX Pr. 19.21(24), 26.15.’
The passages are:

Prov 19:21(24) (Ra.) ὁ ἐγκρύπτων εἰς τὸν κόλπον αὐτοῦ χεῖρας ἀδί-
κως, οὐδὲ τῷ στόματι οὐ μὴ προσαγάγῃ αὐτάς.
Prov 26:15 (Ra.) κρύψας ὀκνηρὸς τὴν χεῖρα ἐν τῷ κόλπῳ αὐτοῦ οὐ
δυνήσεται ἐπενεγκεῖν ἐπὶ τὸ στόμα.
The proposed meaning is based solely on the meaning of the Hebrew

word to which κόλπος corresponds in these two places, i.e., ‫‘( צלחת‬dish,’
BDB; ‘pan,’ KB, Lexicon). There is no support for such a sense in the
known uses of κόλπος; if it had developed this sense it is hardly to be
doubted that we should have other examples of it. Nor are there any grounds
2
Cf. the uses of ἐκκαίω in its fifty or so other occurrences in the LXX: (1) act. ‘kindle,’ ‘set
alight,’ pass. ‘be kindled,’ lit. and fig. (ὀργὴ κυρίου etc.); pass. ‘be stoked up’ (Dan 3:22
κάμινος); (2) act. ‘burn’ (trans.); ‘scorch’; pass. ‘be burnt.’
for assuming that κόλπος here is an exact equivalent of the word it trans-
lates. There is no point of connexion between the root ‫ צלח‬and κόλπος
which would suggest that the translation is an etymologising one of the
kind quite often found in the LXX, and the context of the Greek does not
demand that κόλπος should have the same meaning as the Hebrew. Trans-
lating κόλπος in these two passages as ‘bosom,’ ‘pocket,’ gives just as
good a sense as translating it ‘bowl,’ since precisely what the sayings
as a whole are intended to convey in the Greek is something we can only
guess at. There is no difficulty in supposing the translator did not know
the meaning of the uncommon word ‫צלחת‬.3
There is moreover a certain improbability in the suggestion that the
Proverbs translator, whose rendering Thackeray classified as ‘literary,’
under ‘paraphrases and free renderings,’4 would use the word κόλπος in
a very unusual sense, unlike his use of it elsewhere,5 in a context where
its normal sense would seem most natural. It is not as if no other words
for ‘bowl,’ ‘dish,’ were available.
That it must not be assumed that a word in the LXX will automatically
have the same sense as the Hebrew word it renders is a principle which
is not likely to be doubted by anyone.6 But this is not to say that the LXX
can always be understood purely as Greek. Words are in fact very fre-
quently used in senses as far as we know abnormal for Greek, and their
meaning can be established only from the Hebrew original rendered in
each instance. The cases, however, where a word can be assumed to have
a sense unlike its known senses and the same as that of its Hebrew orig-
inal are quite special ones to which certain conditions apply.
These conditions are, I suggest, somewhat as follows: (1) if a Greek
word whose main sense coincides more or less with what is felt to be the
main sense of a certain Hebrew word is used to render that Hebrew word
3
Elsewhere only 2 Kgs (= LXX 4 Kgdms) 21:13 (rendered ἀλάβαστρος), and perhaps
2 Chr 35:13 ‫צ לחות‬,ֵ taken as pl. of ‫צלחת‬, KB, Lexicon (LXX rendering not clear). That
the LXX translators at times display a limited knowledge of Heb. is generally accepted:
cf. Swete, Introduction, 319, 330; Jellicoe, Septuagint, 329. It is interesting to note,
however, that according to Toy, Proverbs, on 19:24, ‫ צלחת‬was taken as ‘bosom’ by
some medieval Jewish commentators; as ‘armpit’ by Aq., Sym., Targ., Lat., and ‘Rashi
reports a rendering slit in a garment (= bosom).’ It would seem possible then that the
LXX rendering is not just a mistake, but represents a traditional understanding of the word
in these passages.
4
Thackeray, Grammar, 13.
5
See Prov 6:27; 16:33; 17:23; 30:4. There and elsewhere in the LXX, about 35 times,
κόλπος is used in its normal senses, except perhaps in 3 Kgdms 22:35 bis ὁ κόλπος τοῦ
ἅρματος (κόλπος 2º ~ ‫קיח‬, ‘bosom,’ etc.). There is similarly no sign of a meaning ‘bowl’
in any of its NT occurrences.
6
See, e.g., Deissmann, Bible Studies, 73–4.
in an instance where the Hebrew word has a sense unparalleled in the

Greek word (‘Hebraism,’ e.g. δίδωμι rendering ‫ נתן‬both when it means
‘give’ and when it means ‘put,’ ‘make’);7 or, what is virtually the same, if
a Hebrew word is rendered by its literal Greek equivalent in a context (e.g.,
an idiomatic expression) which is unusual for the Greek word (‘literalism,’
e.g., υἱός rendering ‫ בן‬in Gen 11:10 Σὴμ υἱὸς ἑκατὸν ἐτῶν, ‘Sem was a
hundred years old’); (2) if the Greek word appears to copy the etymologi-
cal connexions, real or fancied, of the Hebrew word (‘etymologising,’ e.g.,
Cant 4:13 ἀποστολαί σου for ‫שׁלחיך‬, cf. Suppl. s.v.); (3) if the context of
the Greek, preferably in more than one place, is such that it is impossible
to understand the word in any other sense than that of its Hebrew original
(e.g., κῶλον rendering ‫פגר‬, ‘corpse,’ Num 14:29, etc.).8 This last type is
rare and it may be that further evidence of the Greek word would show
the use to be normal for Greek.
Without an indication like one of these there can be no certainty that the
LXX rendering in a given instance is an exact equivalent of the Hebrew
original, and it is precarious to suggest for it on the basis of the Hebrew
a meaning which differs considerably from the known uses.9
Another entry affected by the same faulty procedure is that under
βάσανος:
III. 2, add ‘; … shame, disgrace, LXX Ez. 16.52, al.’
Ezek 16:52 (Ra.) καὶ σὺ κόμισαι βάσανόν σου, ἐν ᾗ ἔφθειρας τὰς
ἀδελφάς σου ἐν ταῖς ἁμαρτίαις σου, αἷς ἠνόμησας ὑπὲρ αὐτὰς καὶ
ἐδικαίωσας αὐτὰς ὑπὲρ σεαυτήν.
Here and in Ezek 16:54; 32:24, 30 βάσανος corresponds to MT ‫כלמה‬,

(‘insult,’ ‘reproach,’ ‘ignominy,’ BDB). It is difficult to explain why the
Hebrew was rendered in this way, since it is clear that the translators of
Ezekiel were familiar with the root ‫כלם‬: cf., e.g., in this same verse 16:52
the rendering of ‫ כלמה‬by ἀτιμία, and in 34:29 etc. by ὀνειδισμός. It is
nevertheless dubious to assume that βάσανος was intended to be an exact
7
Cf. Weinreich, Languages in Contact, 48 on one of the major types of lexical interference
possible in the speech of bilingual speakers: ‘If two languages have semantemes, or units
of context, which are partly similar, the interference consists in the identification and
adjustment of the semantemes to fuller congruence.’
8
[In ‘Context is King,’ 172–4 Pietersma has challenged the meaning ‘corpse’ for κῶλον,
arguing that it has its usual Greek sense of ‘limb,’ ‘member’ (of the body). I think he is
right. In Gen 15:11 the translator renders ‫ פגרים‬by σώματα then explicates it by adding
τὰ διχοτομήματα αὐτῶν. The meaning of ‫ פגר‬itself could be looked into and refined.]
9
I leave aside of course the difficulty that it is impossible to be certain that the translator
had the same Hebrew as found now in MT. It is usual to assume that it was the same
unless there are strong indications to the contrary.
equivalent of ‫כלמה‬. The evidence we have for the meaning of βάσανος in

Greek is quite clear: ‘touch-stone,’ etc.; ‘torture,’ ‘torment,’ are its usual
senses.10 In the four places where βάσανος renders ‫ כלמה‬it can be read-
ily translated ‘torment.’ Cf. also its three other occurrences in Ezekiel, at
3:20; 7:19 ~ MT ‫מכשׁול‬, ‘stumbling-block’; 12:18 ~ MT ‫רגזה‬, ‘quivering,’
‘quaking.’ Are we to suppose that βάσανος has the same meaning as the
Hebrew in those places also?
The meanings proposed in the following cases are I believe open to
suspicion on the same grounds:
ἀδόκιμος ‘ἀ. ἀργύριον dross,’ Prov 25:4; Isa 1:22.11
ἀθετέω ‘treat despotically,’ Deut 21:14.
ἀντιπίπτω ‘dispute,’ ‘strive,’ Num 27:14.
ἐγκολλάω ‘reach as far as,’ Zech 14:5.
ἐκκρούω Pass., ‘to be thrust out,’ Deut 19:5.
ἐμπυρισμός ‘rust’ (on plants), 3 Kgdms 8:37.
πνοή ‘gap,’ ‘vent,’ Neh 6:1.
πύρωσις ‘rust’ (on plants), Amos 4:9.
σωτήριος ‘peace-offering,’ Lev 3:6; 7:11, al.12
III
In certain entries we find not only that the Greek word has been wrongly
assumed to have the same sense as the word it translates but also that the
meaning of the Hebrew word is itself not generally agreed upon. Thus the
meaning given for the Greek word derives from only one of the possible
meanings of the Hebrew word, selected by the authors of the Supplement
as the right one. For instance, s.v. κάνθαρος:
Add ‘VII. = Lat. cantherius, spar, rafter, LXX Hab. 2.11.’
Hab 2:11 διότι λίθος ἐκ τοίχου βοήσεται, καὶ κάνθαρος (~ ‫ )כפיס‬ἐκ
ξύλου φθέγξεται αὐτά.
The meaning of ‫( כפיס‬found only here in MT) upon which the proposed
meaning of κάνθαρος is clearly based is given by BDB with some hesi-
tation, ‘appar. (si vera l.) term. tech. for some beam in a house, perhaps
10
See LSJ; BDAG; and for examples contemporary with the LXX, Preisigke, Wörterbuch.
11
Apart from the main objection to which this entry is open there is the additional point
that the phrase ἀδόκιμον ἀργύριον does not in fact occur in Isa 1:22. The text is: τὸ
ἀργύριον ὑμῶν ἀδόκιμον· οἱ κάπηλοί σου μίσγουσι τὸν οἶνον ὕδατι (‘your money is
not genuine …’).
12
For a full discussion of this term in the LXX, giving no support to the Supplement’s
meaning, see Daniel, Recherches, 272–81. [Cf. now Lee, Greek of the Pentateuch, 197.]
rafter, or girder,’ and does not appear at all in KB, Lexicon, who give the
meaning ‘stucco-work (laid on panelling).’
What κάνθαρος here was in fact intended to mean cannot be known
with certainty, at least with our present evidence. It is usually ‘beetle’ and
various objects (a type of cup, boat, fish, ornament, and the mark under
the tongue of the Apis-bull, Hdt. 3.28) apparently so called from similar-
ity to a beetle. The sense ‘beetle’ may be possible in Hab 2:11, but seems
unlikely. Perhaps the meaning is ‘knot,’ suggested by Schleusner.13 This,
though unsupported by a parallel, would give a satisfactory sense and be
more in keeping with the known uses than ‘rafter.’ (κάνθαρος occurs only
here in the LXX.)
The identification of κάνθαρος with cantherius is mistaken and the Latin
word therefore irrelevant. κάνθαρος is to be identified with cantharus,14
and cantherius probably with κανθήλιος (-ον).15
S.v. σκελίζω:
Delete ‘LXX Je. 10.18’, [i.e. from its present place in 1.] and add ‘2. used
for translating Hebr. qāla (uprooted), LXX Je. 10.18.’
Jer 10:18 ἰδοὺ ἐγὼ σκελίζω τοὺς κατοικοῦντας τὴν γῆν ταύτην ἐν
θλίψει, ὅπως εὑρεθῇ ἡ πληγή σου.
We note first of all the minor point that the Hebrew word translated
by σκελίζω is in fact ‫( קולע‬participle of ‫)קלע‬. The usual meaning given
for this word, in accordance with its use elsewhere (twice of slinging
stones), is ‘sling forth’ (so BDB; KB, Lexicon). The meaning ‘uproot’
for it in this passage derives from a proposal by G. R. Driver.16 I would
not presume to criticize this proposal;17 the point is merely that unless
the meaning of the Hebrew word can be regarded as definitely established
it is at least doubtful to use the word as evidence for the meaning of the
Greek word.
There is no reason to suppose that this Greek word has the same mean-
ing as ‫ קלע‬here, as the classification of the example under a new heading
would suggest and as is stated explicitly in the entry under ὑποσκελίζω
13
Schleusner, Lexicon in LXX (‘nodus’).
14
So TLL, s.v.
15
‘Cant(h)erius a gr. κανθήλιος tractum esse videtur; cf. κανθήλια, κάνθων,’ TLL. For
κανθήλιον, τό, ‘rafter’ see LSJ, s.v. κανθήλια III.
16
Driver, ‘Linguistic and Textual Problems,’ 107. Driver does not suggest that σκελίζω
means ‘uproot,’ though his footnote 42 (‘The LXX’s σκελίζω “I trip up, supplant” is not
far from this, the true meaning’ [i.e., ‘uproot’]) seems to be a hankering after using the
Greek in support of the meaning ‘uproot.’ In fact ‘trip up’ is not really much closer to
‘uproot’ than to ‘sling forth.’
17
For a comment on it see Barr, Comparative Philology, 108.
below. The evidence at present available to us indicates clearly that σκε-

λίζω usually meant ‘trip up,’ ‘upset’: cf. the examples in LSJ, and also
the much better attested ὑποσκελίζω, which has the same meaning. Note
especially:
Jer 23:12 γενέσθω ἡ ὁδὸς αὐτῶν αὐτοῖς εἰς ὀλίσθημα ἐν γνόφῳ, καὶ
ὑποσκελισθήσονται καὶ πεσοῦνται ἐν αὐτῇ.
A similar entry, relating to the same passage (Jer 10:18), is found under
ὑποσκελίζω:
2. add ‘b. uproot, used in translating Heb. qāla (uprooted) LXX Je. 10.18.’
I am however unable to discover any sign of a reading ὑποσκελίζω for
this place.
The entry under παρατείνω, although it does not propose a new category
for the LXX example, is of a similar kind, involving the use of a Hebrew
word of uncertain meaning as the basis for the meaning of the Greek:
II. 3, add ‘; ὅλην τὴν παρατείνουσαν (sc. ἡμέραν) all the lengthening
day, i.e. throughout the morning, LXX 2 Ki. 2.29.’
2 Kgdms 2:29 (Ra.) καὶ Αβεννηρ καὶ οἱ ἄνδρες αὐτοῦ ἀπῆλθον εἰς
δυσμὰς ὅλην τὴν νύκτα ἐκείνην καὶ διέβαιναν τὸν Ιορδάνην καὶ ἐπο-
ρεύθησαν ὅλην τὴν παρατείνουσαν καὶ ἔρχονται εἰς τὴν παρεμβολήν.
(ὅλην τὴν παρατείνουσαν ~ MT ‫כל־הבתרון‬.)
The meaning of ‫בתרון‬, a word found only here, must be regarded as
still under discussion. BDB take it as the proper name of a place, while KB
give ‘commonly: gully,’ noting forenoon proposed by Arnold, and also (in
Suppl.) Batruna the name of a place mentioned in the El-Amarna tablets.
The Supplement’s entry gives the impression of being an attempt to force
the Greek to agree with the meaning ‘morning’ for ‫ בתרון‬at any price.18
To understand παρατείνουσαν in this way is very far-fetched. Granted
παρατείνω could have the sense ‘become longer,’ but how ‘the lengthen-
ing day’ can be understood to mean ‘the morning’ is not clear, and whether
it is a likely sense of the Greek here seems not to have been considered.
A satisfactory and more natural sense is obtained by taking παρατείνω
in the sense of ‘stretch, lie, beside’ (LSJ II.l): ‘they travelled across all the
adjacent country (sc. χώραν or γῆν).19
18
Neither W. R. Arnold, who first proposed this meaning for ‫‘( בתרון‬The Meaning of ‫בתרון‬,’
274–83) nor Eissfeldt, supporting it (‘Ein gescheiterter Versuch,’ 117 n. 9), makes any
attempt to use the Greek translation as evidence for that sense. Arnold thought the LXX
rendering to be ‘obviously based on Aramaic ‫בתר‬, after, or one of its derivatives’ (275).
19
So Schleusner, Lexicon in LXX, s.v. παρατείνω. Cf. Ezek 27:13 ἡ Ἑλλὰς καὶ ἡ σύμπασα
καὶ τὰ παρατείνοντα, οὗτοι ἐνεπορεύοντό σοι. … The acc. after ἐπορεύθησαν is a
IV
In another type of entry the Supplement classifies certain occurrences

of words in the LXX under new headings because the Greek is a ‘mis-
translation’ of the Hebrew original. Thus s.v. ὀπωροφυλάκιον:
Add ‘2. mistranslation in LXX Mi. 3.12 (cf. Ps. 78[79].1, Je. 33[26].18)
of Hebr. ‘îyîn ‘heap of ruins’ (perh. from resemblance to Syr. ‘ay(y)en
‘guarded’).’
The separation of the example under a new heading, suggesting that

because it is a mistranslation it somehow differs from other examples, is
unjustified. In fact the information that the translator understood the
Hebrew wrongly and the reason why he did so have no relevance at all
to establishing the meaning of the word here. We should not know there
had been mistranslation if we did not already know what the Greek word
could be expected to mean (i.e., ‘garden-watcher’s hut’) and feel it to be
most improbable that it meant ‘heap of ruins.’
Similar remarks apply to the entry under πλατύνω:
Add … ‘6. in LXX Pr. 24.43(28) a mistranslation of Hebr. pittāh
‘deceived’ in the sense of the Syr. pattî ‘widened’.’
In conclusion we may notice the Supplement’s entry s.v. ὑφοράω:

Add ‘II. overlook, disregard, LXX De. 3.26, Ps. 77.59, 62.’
In the three places cited the text has ὑπερεῖδεν (< ὑπεροράω).
Postscript
The opening observations about the need for ‘incontrovertible’ examples were
written with input from John Chadwick. The term ‘indicator value’ was derived
from Barr but is not one I would use now.
Chadwick wrote in 1994 (‘Replacing Liddell and Scott,’ 7–8) that his ‘suspi-
cions [of the 1968 Supplement] were first aroused by J. A. L. Lee, then a research
student at Cambridge, who was working on the Greek of the Septuagint.’ He
went on to say, ‘I decided that the most probable explanation of these faults was
that the editor had received a series of notes from some eminent Hebrew scholar,
whose advice he had accepted without bothering to check it.’ The LXX mate-
rial in the first Supplement was said to have been contributed by G. R. Driver
(preface, vi).
consequence of literal representation of the Heb., but in any case can be paralleled else-
where in Greek, e.g., Arr., An. 6.23.1.
All the entries commented on in this paper were dropped or rewritten in the
Revised Supplement (1996), though there is no mention in the preface. Further
examination of the LXX entries in both LSJ Supplements (1968 and 1996) may
be found in Hauspie, ‘The LXX Quotations’ (2004).
In 2015 Al Pietersma, in ‘Context is King,’ 165, wrote (to my surprise): ‘I like
to date the beginning of a new era [in LXX lexicography] with the publication of
John Lee’s ‘A Note on Septuagint Material.’
2
ΑΠΟΣΚΕΥΗ IN THE SEPTUAGINT
1972
Abstract
This paper examines the word ἀποσκευή, a third century BC formation that
shows a surprising semantic development, leading to the meaning ‘military depend-
ants, family.’ The non-LXX evidence establishes this and other meanings, and
the LXX occurrences are shown to have the same meanings, including ‘depend-
ants, family.’ Contemporary usage is seen to be indispensible to understanding the
word in the LXX. This evidence is a corrective to the false attribution of the mean-
ing ‘children’ to ἀποσκευή on the basis of Hebrew ‫טף‬, which it often renders.
Further, ‫ טף‬itself sometimes has a meaning close to ‘dependants, family.’
In early Koine Greek ἀποσκευή, a new formation attested first in the

third century BC, undergoes a remarkable semantic development. The evi-
dence for this in papyri and Koine authors has previously been observed,
but the LXX examples of the word, which are quite numerous, have received
little attention. My purpose here is to examine the word in the LXX in the
light of the contemporary usage. It will be seen that a knowledge of the
latter helps considerably in establishing the meaning in the LXX. This word
provides a good illustration of the importance of investigating the LXX
vocabulary in conjunction with the vocabulary of its time.
The evidence of the word in non-biblical Greek is extensive. It occurs
frequently in Koine authors, especially Polybius and Diodorus Siculus,
and a number of times in the papyri. These examples have been dis-
cussed at some length by others.1 Here I merely summarise what has been
established.
1
Holleaux, Études, 3:15–26; Kiessling, ‘Die Aposkeuai’; Wilcken, ‘III. Referate,’ 88.
See also Preisigke, Wörterbuch, 4, s.v. ἀποσκευή, and discussion by the editors of
PHal 1.
The primary sense is, as the etymology leads us to expect, ‘movable

property,’ ‘baggage’ (both in the sing. as a collective, and in the pl.).2 This
is seen, for example, in D.S. 19.32.1:
τὰ βαρύτατα τῆς ἀποσκευῆς προαπέστειλεν εἴς τινα τῶν πλησίον
πόλεων.
The heaviest of the baggage he sent on ahead to one of the nearby cities.
Similarly Plb. 1.68.3; 2.70.5; 2.3.8, and probably in PCairZen 1.93.9

(257 BC).
The interesting step in meaning is that the word came to include per-
sons as well as inanimate objects. The context in which this development
took place was, as Holleaux shows (18 ff.), a military one. Along with
each army went its ἀποσκευή, its ‘baggage-train,’ which included not
only baggage proper, but also the soldiers’ wives and children and all the
other persons attached to the army.3 This is clearly seen, for example, in
Polyaen. 4.6.13:
Ἀντίγονος δὲ τῶν Εὐμενείων στρατιωτῶν εἰδὼς ἑπομένην τὴν
ἀποσκευήν, ἐν ᾗ γυναῖκες ἦσαν αὐτῶν καὶ τέκνα καὶ παλλακαὶ καὶ
οἰκέται καὶ χρυσὸς καὶ ἄργυρος καὶ ὅσα ἄλλα ἐκτήσαντο. …
Antigonos, knowing that the baggage-train of the Eumenian soldiers was
following, in which were wives, children, concubines, household slaves,
gold, silver, and everything else they had acquired. …
Other examples of ἀποσκευή in the sense of ‘baggage-train’ are found

in Plb. 2.3.7; D.S. 19.42.2; 43.7.
Moreover, each soldier had his own individual ἀποσκευή, comprising
his family and all his worldly possessions. A good example of this use is
found in the following passage in Polybius 1.66.7–9. The Carthaginians
request their mercenary troops to leave Carthage and withdraw to another
town until they can be paid off. The mercenaries agree, but wish to leave
behind τὰς ἀποσκευάς (‘their families’). At this the Carthaginians demur,
fearing
… μήποτε διὰ χρόνου παραγεγονότες, καί τινες μὲν τέκνων, ἔνιοι
δὲ καὶ γυναικῶν ἱμείροντες, οἱ μὲν οὐκ ἐκπορευθῶσι τὸ παράπαν, οἱ
δ᾽ ἐκπορευθέντες αὖθις ἀνακάμπτωσι πρὸς ταῦτα.
2
Cf. ἀποσκευάζομαι, ‘pack up and carry off’ (see LSJ; Holleaux, Études, 3:23 n. 6).
ἀποσκευή basically = ‘all the personal property that one can pack up and take away.’
3
See also Parke, Soldiers, 207, on the contents of the ἀποσκευή and its importance to a
Hellenistic army. The ἀποσκευή naturally included animals: see D.S. 19.84.7; Holleaux,
Études, 3:20 n. 1.
ΑΠΟΣΚΕΥΗ IN THE SEPTUAGINT 13
… lest, longing to be with their wives or children after their recent pro-
tracted absence, they might in many cases refuse to leave, or, if they did,
would come back again to their families. (Paton, Loeb ed.)
Consequently the Carthaginians compel the mercenaries to take τὰς

ἀποσκευάς with them. The ἀποσκευαί clearly include wives and children.4
Other similar examples are found in D.S. 18.30.2, 20; 47.4.
In a number of examples in Egyptian papyri we find a further develop-
ment in meaning, though a slight one. Here ἀποσκευή is used exclusively
of persons, namely, the family left behind by a soldier on active service.
Most of the occurrences are in legal contexts: in Ptolemaic Egypt the fami-
lies of soldiers in the field were accorded special privileges in legal pro-
ceedings.5 Thus in PHal 1.128–44 (III BC) the word is found six times in
regulations concerning the treatment of cases involving soldiers’ families,
e.g., 134–6:
ἐὰν δέ τινες φάσκωσιν εἶναι τῆς ἀποσκευῆς, οἱ δικασταὶ περὶ τούτου
διαγινωσκέτωσαν καὶ ἐὰν γνωσθῶσιν ὄντες τῆς ἀποσκευῆς. …
If any claim to be (in the category) of military family, the dikastai are
to make a decision on the matter and if they judged to be of military
family. …
Compare PBaden 4.48.9 (126 BC), a letter in which a wife writes to her
husband describing the action of an adversary in a law-suit:
ἔλεγε γὰρ μήτε σε στρατεύεσθαι μήτ᾽ ἐμὲ εἶναι ἀποσκευήν.
For he said that neither were you on campaign nor was I military family.
Similarly in UPZ 1.110.90; 2.199, 206 (164 BC); SB 5.8009.3 (I BC).

The senses of the word, then, in non-biblical Greek may be classified
as follows:
1. movable property
2. the baggage-train of an army
3. a soldier’s personal property and family
4. a soldier’s family
It is to be noticed that in senses 3 and 4 ἀποσκευή refers specifi-

cally to a soldier’s (property and) family. Whether it could be used more
generally of any man’s (property and) family is not indicated by our
evidence.
4
On this passage cf. Walbank, Commentary, ad loc.
5
See esp. Kiessling, ‘Die Aposkeuai.’
II
In the LXX ἀποσκευή occurs over thirty times.6 We have, first of all,
the use in the original sense of ‘movable property.’7 This is found com-
monly, especially outside the Pentateuch. Examples are:
Gen 15:14 μετὰ δὲ ταῦτα ἐξελεύσονται ὧδε μετὰ ἀποσκευῆς (‫)רכושׁ‬
πολλῆς.
After that they will come out to here with much property.
2 Chr 20:25 (Ra.) εὗρον κτήνη πολλὰ καὶ ἀποσκευὴν (‫ )רכושׁ‬καὶ
σκῦλα καὶ σκεύη ἐπιθυμητά.
They found many cattle and (much) property and booty and desirable
objects.
In other passages, however, it is clear that we must recognise a sense

similar to sense 4 above, for example:
Exod 12:37 ἀπῆραν δὲ οἱ υἱοὶ Ἰσραὴλ ἐκ Ῥαμασσὴ εἰς Σοκχώθα,
εἰς ἑξακοσίας χιλιάδας πεζῶν οἱ ἄνδρες πλὴν τῆς ἀποσκευῆς (‫)טף‬,
(38) καὶ ἐπίμικτος πολὺς συνανέβη αὐτοῖς, καὶ πρόβατα καὶ βόες καὶ
κτήνη πολλὰ σφόδρα.
The sons of Israel set out from Ramasse to Sokkhotha, up to 600,000 foot-
soldiers, (that is) the men apart from their families, and a large mixed
crowd went up with them, both sheep and beef animals and a great many
cattle.
Here the word plainly refers to all persons apart from the full-grown men;
that is, the men’s wives and children and all the other persons attached
to them. (The singular is used collectively.) So too in Gen 43:8 the refer-
ence is to living persons, the families of Judah and his brothers. Judah
appeals to his father to let him take Benjamin and go to Egypt in order to
buy food:
Gen 43:8 ἵνα ζῶμεν καὶ μὴ ἀποθάνωμεν καὶ ἡμεῖς καὶ σὺ καὶ ἡ ἀπο-
σκευὴ ἡμῶν.
So that we might and live and not die, both we and you and our families.
Two further examples may be seen in Exod 10:8–11. Pharaoh agrees

to let the Israelites go and offer worship, and asks who are to go (vs. 8).
Moses replies that they wish to take young and old, sons and daughters,
sheep and cattle (vs. 9). Pharaoh objects, saying:
6
See below for the full classified list. The word is not in the NT.
7
This is a definition to cover all cases; the translation may vary from case to case depend-
ing on the context (‘baggage,’ ‘property,’ ‘goods,’ etc.).
Exod 10:10 καθότι ἀποστέλλω ὑμᾶς, μὴ καὶ τὴν ἀποσκευὴν ὑμῶν; …

(11) μὴ οὕτως· πορευέσθωσαν δὲ οἱ ἄνδρες, καὶ λατρεύσατε τῷ θεῷ.
As I let you go, am I also to let your families go? … (11) Oh no. Let the
men go and worship your God.
The plague of locusts follows, then the three days of darkness. Pharaoh
relents, and says:
Exod 10:24 βαδίζετε λατρεύσατε κυρίῳ τῷ θεῷ ὑμῶν· πλὴν τῶν προ-
βάτων καὶ τῶν βοῶν ὑπολείπεσθε· καὶ ἡ ἀποσκευὴ ὑμῶν ἀποτρεχέτω
μεθ᾽ ὑμῶν.
Go, perform your cultic service to the Lord your God, except you must
leave behind the sheep and beef animals; but your families may leave
with you.
In the first instance ἀποσκευή may well include the cattle besides the
men’s dependants – the vagueness of the word is apparent – but in the
second it clearly means the men’s families alone.
In the four examples just quoted ἀποσκευή renders Hebrew ‫טף‬, on which
some comment is necessary. ‘Children,’ ‘little ones,’ the meaning given
by BDB, is one of the senses of ‫ טף‬but not the only one. BDB themselves
note in their Addenda et Corrigenda that ‘the word includes (or implies)
women as well as children,’ citing Gen 47:12; Exod 12:37; 10:10, 24;
Num. 32:16, 17.8 Even this does not go far enough. It seems clear that
in fact ‫ טף‬had, in addition to ‘children,’ a wider sense similar to that of
ἀποσκευή, namely, ‘a man’s family, including wife, children, and other
dependants.’9 Moreover, this sense of ‫ טף‬is probably to be recognised not
only in the places cited by BDB but elsewhere as well, for example:
2 Sam 15:22 ‫ויעבר אתי הגתי וכל־אנשׁיו וכל־הטף אשׁר אתו‬
And Ittai the Gittite went past and all his men and all his dependants
who were his.10
The LXX translators’ renderings of ‫ טף‬also indicate that they recog-

nised this general sense of ‫טף‬. Apart from ἀποσκευή, and words meaning
‘children,’11 we find the following renderings: οἰκίαι (Gen 50:21), πανοι-
κία (Judg 18:21A), συγγένεια (Gen 50:8), λαός, (Judg 21:10A), ὄχλος
8
Similarly Skinner, Genesis on Gen 47:12.
9
Cf. NEB, in which ‫ טף‬is almost invariably rendered as ‘dependants.’
10
Cf. Smith, Samuel, note ad loc.: ‘he marched by with a train which embraced his men
and their families.’ (‫ ~ טף‬LXX ὄχλος.) I would see this sense of ‫ טף‬also in Gen 43:8;
50:21; Num 32:24; Judg 18:21; Ezra 8:21.
11
Renderings of ‫ טף‬meaning ‘children’: παιδία (Gen 45:19; Num 14:3, 31; Deut 1:39;
3:6; Josh 1:14; 9:2f (MT 8:35); 2 Chr 20:13); τέκνα (Deut 2:34; 3:19; Judg 18:21B;
(2 Sam 15:22), κατὰ σῶμα (Gen 47:12 ~ ‫)לפי הטף‬, τὰ λοιπά (Jer 48(41):16);
also ἀπαρτία (Num 31:17, 18).12 There is therefore no reason to suppose
that in the four passages just considered (Gen 43:8; Exod 10:10, 24; 12:37)
ἀποσκευή is intended to have the meaning ‘children.’ Other passages where
the same sense of ‘family’ appears certain are Num 32:16, 17, 24 (‫ טף‬in all),
and Judith 7:2.
The widest sense of ἀποσκευή is probably to be discerned in:
Gen 14:12 ἔλαβον δὲ καὶ τὸν Λὼτ υἱὸν τοῦ ἀδελφοῦ Ἀβρὰμ καὶ τὴν
ἀποσκευὴν (‫ )רכשׁ‬αὐτοῦ καὶ ἀπῴχοντο.
And they also took Lot, the son of Abram’s brother, and his property and
went away.
To judge from the context, the word seems to cover both goods and
persons. Later, when Lot is rescued (vs. 16), mention is made of the recov-
ery also of τὰ ὑπάρχοντα αὐτοῦ καὶ τὰς γυναῖκας καὶ τὸν λαόν. The
last-mentioned would presumably be the miscellaneous crowd of relatives
and slaves which formed part of Lot’s ἀποσκευή. On the other hand, the
fact that ἀποσκευή here renders ‫ רכשׁ‬means that we cannot entirely rule
out the sense of ‘movable property.’ ἀποσκευή may have been intended
simply as a literal rendering of the Hebrew word, without regard to what
follows in the narrative.13
III
A number of LXX passages raise a special problem. They are those in

which ἀποσκευή, translating ‫טף‬, occurs side by side with γυναῖκες, for
example:
Deut 20:13–14 καὶ πατάξεις πᾶν ἀρσενικὸν αὐτῆς ἐν φόνῳ μαχαίρας,
(14) πλὴν τῶν γυναικῶν καὶ τῆς ἀποσκευῆς καὶ πάντα τὰ κτήνη.
And you shall strike every male in it (the city) with slaughter of the
knife, except the women and the families and all the cattle.
What is the meaning of our word here? ‘Baggage,’ ‘possessions’ will

not fit the context and can be eliminated. Is it ‘children,’ as the immediate
context and the Hebrew original seem to indicate? Clearly this is a possible
2 Esd 8:21); ἔκγονα (Deut 29:10; 31:12); νήπια (Jer 50(43):6; Ezek 9:6); ἐπιγονή
(2 Chr 31:18).
12
[I omit a lengthy note on ἀπαρτία included in the original article.]
13
Similarly 2 Macc 12:21 προεξαπέστειλεν ὁ Τιμόθεος τὰς γυναῖκας καὶ τὰ τέκνα καὶ
τὴν ἄλλην ἀποσκευήν (καὶ τὰ τέκνα om. A).
sense. In my view, however, there is no compelling reason for seeing a

new sense of ἀποσκευή here.
The Hebrew original is an unsure guide to the meaning because:
(a) assuming ‫ טף‬means ‘children’ here, there is no certainty that the trans-
lators intended their rendering to have the same sense;14 and (2) it has not
been established beyond doubt that ‫ טף‬itself has the specific sense of ‘chil-
dren’ here. A less precise meaning is possible.
I suggest that it is possible to take ἀποσκευή tautologically, in the same
wide sense that has been established for other passages, namely ‘family’
(including wife, children, and other dependants). The word has a vague
meaning, and we ought not to give it such a precise one as ‘children’ unless
the context overwhelmingly demands it. This suggestion is strongly sup-
ported by a further example where ἀποσκευή is plainly tautological:
Num 16:27 καὶ Δαθὰν καὶ Ἀβιρὼν ἐξῆλθον, καὶ εἱστήκεισαν παρὰ
τὰς θύρας τῶν σκηνῶν αὐτῶν, καὶ αἱ γυναῖκες αὐτῶν καὶ τὰ τέκνα
αὐτῶν καὶ ἡ ἀποσκευὴ (‫ )טף‬αὐτῶν.
And Dathan and Abiron came out and stood beside the doors of their
tents, and their wives and their children and their dependants.
Here the word must be vague and general, and include the women and
children just mentioned and other persons in these families. (‘Baggage’ is
ruled out, as is ‘children.’)
Several other passages in the Pentateuch are similar to Deut 20:14
above in having ἀποσκευή and γυναῖκες together, with the added compli-
cation that the immediate context will permit ἀποσκευή to be taken in
the sense of ‘baggage,’ ‘property,’ as well as ‘family.’ Although one can-
not say for certain, the latter seems the more probable in all these places.15
I take two examples, in the first of which the mention of property in the
next verse (6) strongly suggests that ἀποσκευή refers to persons:
Gen 46:5 καὶ ἀνέλαβον οἱ υἱοὶ Ἰακὼβ Ἰσραὴλ τὸν πατέρα αὐτῶν
καὶ τὴν ἀποσκευὴν (‫ )טף‬καὶ τὰς γυναῖκας αὐτῶν ἐπὶ τὰς ἁμάξας …
(6) καὶ ἀναλαβόντες τὰ ὑπάρχοντα αὐτῶν καὶ πᾶσαν τὴν κτῆσιν, ἣν
ἐκτήσαντο ἐν γῇ Χανάαν. …
And the sons of Jacob took up their father Israel and their families and
their wives on the wagons … (6) and taking up their property and eve-
rything they had acquired in the land of Canaan. …16
14
Cf. my remarks on this subject in ‘Note on Septuagint Material,’ 237–8 [= Essay 1].
15
They are: Gen 34:29; 46:5; Num 31:9; 32:26, 30. ἀποσκευή ~ ‫ טף‬in all except Num 32:30,
where there is no equivalent in MT.
16
Cf. LSJ Suppl. (1968), s.v. ἀποσκευή: ‘children, little ones, LXX Ge. 46. 5, al.’ This is
plainly based on the supposed meaning of the Heb. word (note ‘little ones’).
The second is:

Num 32:26 ἡ ἀποσκευὴ (‫ )טף‬ἡμῶν καὶ αἱ γυναῖκες ἡμῶν καὶ πάντα
τὰ κτήνη ἡμῶν ἔσονται ἐν ταῖς πόλεσιν Γαλαάδ. …
Our households and our wives and all our cattle shall be in the cities
of Galaad. …
Here the sense is likely to be the same as in the nearby verses 32:16,
17, 24, all of which concern the Reubenites’ and Gadites’ intention to
leave their families and cattle secure in Gilead while they themselves cross
the Jordan to aid the other tribes in their entry into Canaan. The meaning
‘dependants,’ ‘family’ seems certain in these.17
A textual question arises in one instance, where the reading adopted by
Rahlfs is questionable:
Exod 39:22 (Ra.) ὅσα συνέταξεν κύριος τῷ Μωυσῇ, οὕτως ἐποίησαν
οἱ υἱοὶ Ισραηλ πᾶσαν τὴν ἀποσκευήν [of the tabernacle].
αποσκευην Bahqu: κατασκευην FbMegjnsvwz: παρασκευην AF*ia?
rell. [BM]
Ἀποσκευή is hardly possible in this context. If read it will have to bear an

unparalleled sense: none of the established senses is suitable. On the other
hand, either of the alternative readings will give a satisfactory sense (κατα-
σκευή ‘fittings,’ ‘furniture’; παρασκευή ‘preparation,’ or ‘equipment’).18
IV
To sum up, I give a full list of the LXX examples, classified according
to the senses I propose (omitting Exod 39:22):
1. movable property (including animals): Gen 15:14; 31:18; 1 Chr 5:21;
2 Chr 20:25; 32:29bis; 2 Esd 1:4, 6; 1 Macc 9:35, 39: probably
2 Chr 21:14, 17; 1 Macc 5:13, 45. (‫ רכושׁ‬in all except 1 Chr 5:21;
2 Chr 32:29[1°] ‫מקנה‬.)
2. movable property and persons comprising a man’s family:
2 Macc 12:21; probably Gen 14:12 (‫)רכושׁ‬.
3. the group of persons comprising a man’s family: Gen 34:29, 43:8;
46:5; Exod 10:10, 24; 12:37; Num 16:27; 31:9; 32:16, 17, 24, 26,
30; Deut 20:14; Judith 7:2. (‫ טף‬in all except Num 32:30: no Heb.
equivalent.)
17
Outside Pent., there is a similar uncertainty about deciding between ‘property’ and
‘family’ in: 2 Chr 21:14, 17; 1 Macc 5:13, 45: I take all as probably ‘property.’
18
Support for κατασκευήν may be seen in Exod 35:24, where it is used of the tabernacle
fittings and translates ‫עבדה‬. [Göttingen ed. now reads παρασκευήν.]
We have seen, then, that the usage of ἀποσκευή in the LXX is closely
linked with that in the Greek of the time. In particular, the use of the word
in reference to persons has now been established in the LXX.
There is, however, one point of difference between LXX and contem-
porary usage. As was noticed earlier, according to our present evidence
the word is used outside the LXX only in reference to soldiers’ families.
In one part of the LXX, the Pentateuch, however, it is frequently used in
a more general way of any man’s family. It is difficult to tell whether this
was an innovation in the translators’ own Greek. The extension is a slight
one and could easily have occurred already in the Greek of the time. It
can hardly be due to Hebraism. Nevertheless, there is a possibility that the
translators themselves extended the usage of this convenient term. The
extension would have been helped by the fact that many of the contexts
in which such a word is needed are quasi-military ones. In particular, the
Israelite host in its journey from Egypt into Canaan is very like an army
on the march, and is readily described in military language.19 Outside the
Pentateuch ἀποσκευή in reference to persons is confined to military con-
texts (Judith 7:2; 2 Macc 12:21).
Another point of difference between books of the LXX is also to be
noticed, though it is difficult to tell whether it is significant. The use of
ἀποσκευή in reference to persons is practically restricted to the Pentateuch.
Outside the Pentateuch it is found only in the two instances just mentioned;
otherwise the word is used of inanimate property and animals, mostly trans-
lating ‫רכושׁ‬. Linked with this is the fact that ἀποσκευή as a rendering of ‫טף‬
is found only in the Pentateuch; in the other books ‫ טף‬is most often rendered
by words for ‘children.’ In this the later translators do not depart from the
example of the Pentateuch: ‫ טף‬is several times so taken there [and ren-
dered παιδία etc. (see n. 11)]. But in four instances where they take ‫ טף‬in
a wider sense the later translators use words of a quite different kind from
ἀποσκευή.20 It seems possible therefore that they consciously avoided
ἀποσκευή in reference to persons. If so, we might see in this a deliberate
rejection of the use because of the pejorative implication it might seem to
carry. On the other hand, the passages where ἀποσκευή might have been
expected are few, and the choice of another word may have been accidental.
19
See esp. Num 31:9, where ἀποσκευή occurs in the context of a battle against the Midi-
anites, and Exod 12:37, where in the same verse as ἀποσκευή the men are called πεζοί.
20
πανοικία (Judg 18:21A); λαός (Judg 21:10A); ὄχλος (2 Kgdms 15:22); τὰ λοιπά
(Jer 48:16).
Postscript
This paper supersedes what was first presented in my 1970 dissertation, and
was subsequently published without change in Lexical Study, 101–7 (1983). The
suggestion made in the final paragraph of the above paper, that ἀποσκευή in ref-
erence to persons may have been consciously avoided in later books of the LXX,
was included at the behest of the JTS reader; I was doubtful then and still am.
There is nothing to show that ἀποσκευή was felt to be pejorative. The Pentateuch
examples seem quite neutral (e.g., Exod 12:37; Gen 43:8 quoted above), and the
documentary examples (e.g., PHal and PBaden above) are clearly not pejorative.
There may be contextual reasons why ἀποσκευή was not chosen in the later
LXX passages.
In O’Connor and Lee, ‘The Case of Hebrew tap and Greek aposkeuê’ (2007),
the semantic developments in ‫ טף‬and ἀποσκευή are set out and compared. The
authors note the approaches to the words in recent commentaries and elsewhere,
remarking that ‘for these two words confusion and lack of acquaintance with
recent developments have lasted too long in certain quarters’ (409). Attachment
to the meaning ‘children’ for ‫ טף‬to the exclusion of others persists (see, e.g.,
DCH), as does reluctance to accept the new meaning of ἀποσκευή (see Wevers,
Numbers, 512).
3
EQUIVOCAL AND STEREOTYPED
RENDERINGS IN THE LXX
1980
Abstract
This paper takes up some methodological issues raised by Tov’s article ‘Three
Dimensions’ (1976). While agreeing with the suggested two dimensions of mean-
ing of a LXX rendering, that is, (a) if read simply as Greek, and (b) if read with
reference to the original Hebrew whereby the translator’s intention becomes
clear, I put forward some instances in which the translator may have been play-
ing with the two levels intentionally, creating an ‘equivocal’ rendering. The paper
then turns to stereotyped renderings and the question whether they can be said to
have a lexical meaning identical to that of the word they translate, as Tov proposes
in some cases. I argue against this, particularly in the case of Tov’s example παντο-
κράτωρ, and in the case of παραπικραίνω, which Walters regarded as having the
meaning of ‫מרה‬, ‘rebel,’ the word it translates in the Psalms.
In a recent article Emanuel Tov has tackled some basic questions of

method in LXX lexicography, and made several valuable suggestions.1 In
concluding he remarks that ‘it will be to the benefit of LXX lexicography
if attention be directed not only to single words and semantic fields, hut also
to methodological questions, even if the answers provided are partial and
tentative, as in the present paper.’ It is in the same spirit that I offer here
some further observations on the subject, prompted by Tov’s discussion.
Tov points out that what was intended by the LXX translator in one
way may subsequently be understood in another. Actual instances of
this in the later versions and the Church Fathers are given by Tov, and of
course in our own reading of the LXX today we often meet with the same
dilemma between reading a passage simply as Greek and interpreting it
1
Tov, ‘Three Dimensions.’ [A revised version has since appeared in Tov, Collected Essays
(1999), 85–94.]
with reference to the original Hebrew, by means of which the translator’s

real intention becomes clear.2 There are, then, potentially two dimensions,
or levels, of understanding of any LXX rendering.3 Tov stresses that
the meaning intended by the translator is the dimension that has priority
in LXX lexicography. This is clearly right, and the establishing of this
principle is, I believe, an important step forward. It is what the translator
intended his rendering to mean at the time of translation that the lexicog-
rapher must try to recover, not what a subsequent reader might take it to
mean.4
II
I want to suggest, however, that there are instances in which the trans-
lator himself may have been playing a double game. That is to say, in some
cases it looks as if the translator himself, at the time of translation, was
aware of, and was making use of, the possibility of these same two levels
of understanding, when he gave a rendering that read one way as Greek and
another way in the light of the Hebrew. Though this seems to be uncom-
mon, it is a phenomenon to be reckoned with. I put forward the following
three examples for consideration.
(1) Gen 39:10 ἡνίκα δὲ ἐλάλει τῷ Ἰωσὴφ ἡμέραν ἐξ ἡμέρας, καὶ
οὐχ ὑπήκουσεν αὐτῇ καθεύδειν μετ᾽ αὐτῆς τοῦ συγγενέσθαι αὐτῇ.
:‫ויהי כדברה אל־יוסף יום יום ולא־שׁמע אליה לשׁכב אצלה להיות עמהּ‬
If we read the Greek on its own, the meaning of συγγενέσθαι is unmis-

takable. We have been plainly told in the earlier part of the narrative what
Potiphar’s wife had in mind, so it is natural to take it that Joseph refused
‘to go to bed with her in order to have sexual intercourse with her.’ This is
a well-established sense of συγγί(γ)νομαι, attested in Classical and later
Greek.5 Of the several possible senses of the word, this is clearly the one
2
Cf. Ottley, Handbook, 167–9 for an earlier appreciation of the problem, leaving it unre-
solved.
3
[Tov’s ‘third dimension’ refers to (new) meanings attached to LXX words in the ‘post-
Septuagintal period.’]
4
Tov, ‘Three Dimensions,’ 529–32. Though the argument is not affected, Tov’s third illus-
tration, κύριος τῶν δυνάμεων, is open to doubt: it is hardly certain that the translators
intended no more than the sense of ‘army’ for δύναμις here. This principle, that what the
author intended his/her words to mean is the meaning, has always been followed in read-
ing original Greek texts, but there it has not been necessary to formulate it.
5
Examples form Xenophon to Plutarch in LSJ, s.v. II.3. The sexual sense is plain also in
Jdt 12:16 and Sus (Th.) 11. Though this use originates in euphemism, the reference, in
EQUIVOCAL AND STEREOTYPED RENDERINGS IN THE LXX 23
that suits the context best. But when we look at the Hebrew a new facet of
the rendering shows itself. We now see that the words τοῦ συγγενέσθαι
αὐτῇ are open to being interpreted as a rendering of the literal meaning
of the original, ‘to be with her,’ with no further connotations.6 Is this the
meaning the translator intended?7 It should be noticed that συγγίνομαι
is not normally used in this basic etymological sense but rather in senses
that are developments of the primary meaning (‘associate with,’ ‘meet,’
‘converse with,’ etc.),8 and that a really literal rendering would have been
(τοῦ) εἶναι μετ᾽ αὐτῆς, as found elsewhere.9 Nevertheless it is clear enough
that the translator has availed himself of the possibility that συγγενέσθαι
can be understood in its etymological sense, and that he intended to rep-
resent the Hebrew words literally.10
But is that all? I do not think it can be. I suggest that it is highly likely
that the translator was also well aware of the final effect of the rendering,
read as Greek and in the context of the story so far. He has managed both
to be faithful to the original and to produce a meaning natural to the con-
text. The existence of a Greek word with the appropriate ambiguity made
this possible.. Whether he deliberately sought out such a word can hardly
be known for certain, but it seems significant that the strictly literal ren-
dering that was available was not used. It is safe to say, I think, that the
translator was at least conscious of the double significance. The transla-
tor’s intention, then, in itself had two dimensions.11
the right context, is quite definite. For parallel semantic development cf., e.g., σύνειμι,
συνουσία, ὁμιλῶ, -ία, and Engl. intercourse. For τοῦ + infin. in a final sense see, e.g.,
MHT, Grammar, III, 141.
6
The exact significance of the Heb. is debatable. The confused statements in BDB and
KB, Lexicon (s.v. ‫ היה‬III.4.d(a) and 7) are difficult to interpret, but most translations and
commentaries take it simply in its literal meaning. To me it seems more likely that this
expression meant ‘have sexual intercourse with’ (cf. esp. 2 Sam 13:20); simil. Weinfeld,
Genesis, ad loc. (I am grateful to Mr Brian Parker for this reference.) If this was the
meaning of the Heb., it might complicate but not vitiate my argument.
7
Here and throughout I use ‘translator’ in the sing. for practical convenience, without
meaning to exclude the possibility of more than one translator.
8
See LSJ, s.v. The first sense in LSJ is in fact to be born with, not ‘be with,’ which does
not appear.
9
E.g., Gen 26:3; 31:3; 48:21. συγγίνομαι is used to render ‫ היה עם‬only here in
Gen 39:10.
10
Cf. Tov’s classification (‘Compound Words,’ 211) of this example along with other
compounds described as ‘more or less literal renderings.’
11
My argument is nicely illustrated by the divergence in the renderings given by the two
main Engl. translations of the Greek: ‘… to sleep with her, so as to be with her’ (Brenton);
‘… to lie down with her to have commerce with her’ (Thomson-Muses). [NETS: ‘… to
sleep with her in order to have relations with her.’]
My next example is:

(2) Exod 36:2 καὶ ἐκάλεσεν Μωυσῆς τὸν Βεσελεὴλ καὶ Ἐλιὰβ …
καὶ πάντας τοὺς βουλομένους προσπορεύεσθαι πρὸς τὰ ἔργα ὥστε
συντελεῖν αὐτά.
‫ויקרא משׁה אל־בצלאל ואל־אהליאב … כל אשׁר נשׂאו לבו לקרבה אל־המלאכה‬
:‫לעשׂת אתה‬
Of course προσπορεύεσθαι is a more or less literal equivalent of ‫לקרבה‬.

The sense of ‘go to,’ ‘draw near to’ (physically, in space) is attested for
the Greek word since Aristotle, and occurs quite often elsewhere in the
LXX, as, e.g., Exod 30:20 ὅταν προσπορεύωνται πρὸς τὸ θυσιαστήριον
λειτουργεῖν … (MT ‫)בגשׁתם‬. But if we look more closely into the usage
of the word we find reason to think that the translator has again succeeded
in killing two birds with one stone.
Just as English words such as go to, approach, show a tendency to move
away from the purely spatial sense and to develop additional, figurative,
uses,12 so too προσπορεύομαι shows a number of developed senses, attested
especially in papyri contemporary with the Pentateuch translators. One of
these is ‘apply oneself to,’ ‘proceed with’ (an activity). This, and not the
merely spatial sense, is clearly required, e.g., in PCairZen 1.60.6 (257 BC)
προσπορεύεται δὲ καὶ πρὸς [ταῦτα] καὶ πρὸς τὰ λοιπὰ μαθήματα, ‘he is
proceeding with this and his other studies.’13 If we look again at Exod 26:2,
it is evident that such a sense would be most appropriate there. Moreover,
there is enough evidence to suggest that the very phrase found in Exodus
was a stock expression in the Greek of the time: προσπορεύομαι πρὸς
τὰ ἔργα (something like ‘get on with the job’) is attested more than once
in documents of the third century BC.14 Could our translator have been
unaware of all this? It seems hardly likely. On the contrary, I think it
probable the translator welcomed, if he did not actually seek, the oppor-
tunity afforded by this word to be both literal and idiomatic. A final sup-
porting detail is that if a merely literal rendering had been wanted, ἐγγίζω
was available: it is used 8 times in the Pentateuch as a rendering of ‫קרב‬
qal.
12
Similarly, e.g., Lat. adeo, accedo.
13
Similarly PCairZen 1.132.4 (256 BC). Note also that προσέρχομαι develops the same
meaning: see LSJ, s.v. I.6. Cf. also προσχωρῶ in Pl., Rep. 539a.1.
14
PSI 4.352.8; PCairZen 3.531.2, 16. Though both the last two are partly restored, one
fills the gap in the other in such a way that the restorations are hard to doubt. Helbing
(Kasussyntax, 300) noted the parallel between the PSI ex. and Exod 36:2, translating
the latter as ‘sich heranmachen an.’ It seems significant that the plural ἔργα in Exod is
not required by the original. [I.e., it is influenced by contemporary usage.]
A third example is:

(3) Num 14:16 παρὰ τὸ μὴ δύνασθαι κύριον εἰσαγαγεῖν τὸν λαὸν
τοῦτον εἰς τὴν γῆν, ἣν ὤμοσεν αὐτοῖς, κατέστρωσεν αὐτοὺς ἐν τῇ
ἐρήμῳ.
In MT the word corresponding to κατέστρωσεν αὐτοὺς is ‫וישׁחטם‬

(i.e., ‫‘ שׁחט‬slaughter’). Seeing that the basic sense of στρώννυμι is ‘spread,’
it is reasonably certain that the translator, whether because of misreading
or because he actually had such a form in front of him, has rendered a
form of ‫‘ שׁטח‬spread.’ But what exactly did he intend his rendering to
mean?
Two senses of καταστρώννυμι present themselves. The primary sense
of ‘spread out,’ ‘strew’ obviously corresponds literally to the Hebrew
word the translator was trying to translate. Taking into account the origi-
nal as the translator saw it, then, we should be led to give καταστρώννυμι
some such meaning as ‘scatter’ here. On the face of it this would seem an
acceptable understanding of what the translator intended.
But there is also a well-established sense of the word that takes a
quite different direction: καταστρώννυμι is attested in both Classical
and later Greek in the sense of ‘lay low,’ with more than a hint of the idea
of killing.15 There are examples of this sense in the LXX in Judith and
2 Maccabees and a very similar use turns up in a papyrus of the second
century BC.16 So there is no reason to doubt that it was known to our
translator. This latter meaning makes equally good sense; in fact, much
better sense, if we look at the whole context. The sentence occurs in words
spoken by Moses as he tries to persuade God to change his mind. Moses
imagines aloud what other nations will say if God exterminates the Israel-
ites, as he has just threatened to do. That the threat is to wipe them out,
and not something milder, is clear from what precedes: in v. 12 God’s
words are πατάξω αὐτοὺς θανάτῳ καὶ ἀπολῶ αὐτούς, and in v. 15 Moses,
referring to this threat, says καὶ ἐκτρίψεις (‫ )והמתה‬τὸν λαὸν τοῦτον. A
word meaning ‘slaughter’ or something similar is therefore expected here
in v. 16.17
15
See BAG, s.v. 1; LSJ, s.v. III and s.v. καταστορέννυμι III. See also LSJ, s.v. στόρνυμι
I.2.b.
16
Jdt 8:14, 25; 14:4; 2 Macc 5:26; 11:11; 12:28; 15:27; UPZ 77.II.28 of throwing a
cow down on the ground (but without any suggestion of killing).
17
It may be wondered if we have not come full circle and proved that the translator was,
after all, trying to render the same text as MT, without any confusion with ‫שׁטח‬. This
seems unlikely: the standard rendering of ‫ שׁחט‬in Pent. is σφάζω (qal × 41; niph × 3);
and if σφάζω was felt unsuitable to apply to human slaughter, ἀποκτείνω (e.g.) was
Moreover, a NT passage provides evidence that κατέστρωσεν was under-

stood in just this way by at least some subsequent readers. In 1 Cor 10:5
we find κατεστρώθησαν ἐν τῇ ἐρήμῳ in reference to the destruction of
some of the Israelites, as the context as a whole makes clear, and the word-
ing plainly derives from Num 14:16. The same understanding of the word
is shown by the Old Latin rendering contriuit.18
It will already be clear that I am suggesting that both meanings were
in the translator’s mind. His rendering κατέστρωσεν was faithful to the
Hebrew as he saw it, but also open to an interpretation better suited to
the context. In effect such a rendering puts the responsibility for the final
decision elsewhere. The reader must decide, ideally with reference to the
original Hebrew.19 For his part the translator can honestly claim that he
has not misrepresented his original.
III
The three examples considered so far follow the same pattern: the Greek
rendering is both a literal representation of the Hebrew and at the same
time capable of a different meaning equally suited to the context. My next
(and last) example, while still an illustration of the phenomenon we are
considering, shows a somewhat different pattern.
(4) Gen 19:4–5 καὶ οἱ ἄνδρες τῆς πόλεως οἱ Σοδομῖται περιεκύκλω-
σαν τὴν οἰκίαν … (5) καὶ ἐξεκαλοῦντο τὸν Λώτ, καὶ ἔλεγον πρὸς
αὐτόν Ποῦ εἰσιν οἱ ἄνδρες οἱ εἰσελθόντες πρὸς σὲ τὴν νύκτα; ἐξά-
γαγε αὐτοὺς πρὸς ἡμᾶς, ἵνα συγγενώμεθα αὐτοῖς.
The word of interest is once again συγγίνομαι. But here we find a

neat reversal of what we saw before. This time the sexual meaning is not
clear in the Greek but explicit in the Hebrew.
Let us begin with the Greek on its own. I submit that if the story is read
through without knowledge of the underlying Hebrew, there is nothing
available. καταστρώννυμι is unusual (only here in Pent.) and seems an unmistakable

clue that ‫ שׁטח‬was involved. This is taken for granted by BHS and Gray, Numbers, ad
loc. Note also that ‫( שׁטח‬an uncommon word: × 6 in MT) is rendered once elsewhere
by καταστρώννυμι, in Job 12:23. I refrain from defending the Num translator at length
against the imputation in Gray’s note: ‘G foolishly ‫וישׁטחם‬.’ I would say only that far
from folly, we may be looking at an intelligent attempt to overcome a difficulty created
by a faulty text.
18
As reported by Brooke-McLean.
19
One is reminded here of Brock’s observation that ‘the free translation aims at bring-
ing the original to the reader, while the literal one the reader to the original’: Brock,
‘Phenomenon,’ 17; cf. 28.
to lead one to understand συγγενώμεθα in anything but a vague sense

of ‘meet,’ or ‘converse with.’20 This is difficult to prove, depending as it
does on personal impression, but it must at least be agreed that the vaguer
meaning is possible in the context. The Sodomites could well be repre-
sented as merely asking to meet the strangers, without stating plainly their
real objective. The fact that their real objective is known to Lot, and may
be guessed from his subsequent offer (v. 8), does not alter this. Nor does
the likelihood that most readers would already know the Sodomites’
reputation.21
Turning to the original, we find that συγγενώμεθα αὐτοῖς corresponds
to ‫ונדעה אתם‬. At once it is clear that a quite different implication of the
Greek word is possible. But is it the full story to say that since the original
has the meaning ‘know sexually,’ that or something close to it was the
meaning the translator intended? The translator’s purpose appears to me
to have been more subtle. Avoiding the literal rendering with γινώσκω,
employed regularly elsewhere,22 he has chosen a word capable of an innoc-
uous interpretation in the context. It is hardly fanciful to suppose that the
motive behind this was a sense of delicacy. At the same time, however,
the meaning of the Hebrew is there in his rendering for anyone who can
recognise it, whether by reference to the original or in some other way.23
IV
Enough has been said to show the existence in the LXX of the phe-
nomenon of equivocal rendering.24 If my analysis is right, in such cases
we are dealing with a double intention on the part of the translator. The
lexicographer has therefore to deal with a double meaning. It must be
20
To the ample Class. attestation of these senses noted by LSJ we can add from the
papyri SB 3.6300.10 (I BC) and probably PGrenf 1.35.1 (II BC).
21
Not, however, from reading other parts of the OT: this incident is the only indication
that the Sodomites’ offence was the one now associated with the name. Elsewhere their
crime is vague or, indeed, different: see Speiser, Genesis, 142.
22
Six times in Gen alone: 4:1, 17, 25; 24:16; 19:8, the last just three verses away from
our example. This too is the rendering in the closely parallel incident in Judg 19:22AB.
συγγίνομαι as a rendering of ‫ ידע‬is unique in Gen 19:5 and clearly calls for explanation.
23
Boling (Judges, 276) suggests that ‫( ידע‬in the parallel Judg 19:22) is ‘deliberately
ambiguous.’ This does not seem necessary, but if it was accepted, and we assumed also
that the Gen translator recognised it, then the ambiguity of his rendering would be a
deliberate reflection of the Heb.
24
Two further examples I have noted are προσπορευέσθωσαν in Exod 24:14 and
ὕπανδρος/ ὑπ᾽ ἀνδρός in Num 5:20, 29. [See Lee, ‘Accuracy and Idiom,’ 96–7 on the
text in Num.]
stressed that his task is not simply to choose between the two; he must
come to terms with the fact that both are present and both have equal claim
on his attention. In practice this means that a lexicon of the LXX will need
to record in one way or another that the word exhibits a combination of
senses in the place concerned.
As to the frequency of this type of rendering, it can be said only that
it does not appear to be common. Nevertheless the lexicographer will be
wise to keep the possibility of it in mind. I have not so far noticed any
example outside the Pentateuch, but that is probably accidental. Further
study may however show that some translators were more prone to this
habit than others, and perhaps that it does not occur at all in some books.
It may be useful to put forward some terminology. I suggest that, where a
double significance is possible, we apply the terms exterior and interior meaning
to the two interpretations involved. The exterior meaning is the sense the word
appears to have when read in the context of the Greek without reference to the
original. The interior meaning is the probable sense when the word is considered
as a representation of the corresponding Hebrew. When there is reason to suppose
that the translator consciously intended both, we have an equivocal rendering. Most
of the time, of course, the question of equivocal rendering does not arise, because
the exterior and interior meanings are one and the same.
One final point must be made to complete the picture. Of course not
all the renderings that we find to be open to two interpretations need have
been produced deliberately. There are at least some, and probably many
cases where we may reasonably deduce that the translator did not intend
the double meaning. I give one example where this seems to be so:
(5) Gen 44:32 ὁ γὰρ παῖς σου ἐκδέδεκται τὸ παιδίον παρὰ τοῦ πατρὸς
λέγων Ἐὰν μὴ ἀγάγω αὐτὸν πρὸς σὲ καὶ στήσω αὐτὸν ἐναντίον σου,
ἡμαρτηκὼς ἔσομαι πρὸς τὸν πατέρα πάσας τὰς ἡμέρας.
... ‫כי עבדך ערב את־הנער מעם אבי לאמר‬
The construction in the Greek is misleading: παρὰ τοῦ πατρός seems to

require us to take ἐκδέδεκται as ‘received,’ a possible sense of the word.
But reference to the original makes it clear that here, as in the earlier occur-
rence in the same story,25 ἐκδέχομαι has the particular sense of ‘stand
surety for,’ which is well attested in third century BC papyri.26 The mis-
leading construction in the Greek has come about through word-for-word
adherence to the Hebrew. The translator apparently did not feel able, or
did not take the trouble, to abandon the Hebrew construction with ‫מעם‬
25
Gen 43:9 (MT ‫)ערב‬. The speaker in 44:32 is recounting this.
26
See Lee, Lexical Study, 59–60.
in favour of one more natural to Greek. There is, however, nothing here
to suggest that he was trying for an equivocal rendering. The exterior
meaning ‘received’ has no advantage in the context, nor can we see any
other motive for wanting the word to be open to that interpretation. We
may conclude therefore that the interior meaning, and no other, was what
the translator intended, and consequently that it alone is to be reported by
the lexicon.
I turn now to a different aspect of the problem of determining the

translators’ intentions, namely stereotyped renderings. On this subject,
too, Tov has offered a valuable discussion, which is my starting-point.27
While in agreement with him at many points, I wish to raise objections in
one area involving a fundamental question of principle.
Though it is difficult to prescribe exact conditions for recognising
one, stereotyped renderings are a well-known feature of LXX translation
methods. The problem for the lexicographer is to arrive at satisfactory
principles for determining their meaning. In particular, difficulty arises
when stereotyped representation leads a translator into using a word in
contexts where its Hebrew counterpart has a sense outside the normal
semantic range of the Greek word. How are we to define the meaning of
the Greek word in these cases? In accordance with the principle enunci-
ated by Tov, the answer lies in trying to discover the meaning intended
by the translator.
Tov’s discussion of εἰρήνη in this connexion (535–6) seems to me to
establish a sound basis for approaching this difficult question. The use of
εἰρήνη as a standard equivalent of ‫ שׁלום‬leads to a number of occurrences
of the Hebraism ἐρωτῆσαί τινα εἰς εἰρήνην (and the like). Tov’s con-
clusions regarding this use of εἰρήνη are (in part) as follows:
[I]t appears that several translators did not use εἰρήνη in accordance
with ordinary Greek usage, and that they did not have a definable mean-
ing of εἰρήνη in mind. They simply equated ‫ שׁלום‬with εἰρήνη on a
practical level. Consequently, one might say that for many of the trans-
lators εἰρήνη was merely a symbol representing ‫שׁלום‬. If this is correct,
a Greek lexicon should not create a new meaning of εἰρήνη on the basis
of the LXX alone, claiming, as it were, that the translators enlarged the
semantic range of εἰρήνη.
27
Tov, ‘Three Dimensions,’ 532–42.
Whether or not our final view of this particular example will need to be
modified, these conclusions appear to me to be right as far as they go, and
to offer a valuable starting-point for dealing with all stereotyped render-
ings. There are, I think, three essential points here: (a) The Greek word
is equated in a mechanical fashion, for the sake of practical convenience,
with the Hebrew word, and may be no more than a symbol for it.28 (b) It
is not certain that the translator has a definable meaning in mind for the
Greek word. (c) The Greek word, as such, does not automatically acquire
a new sense from being used in this way.
Tov follows the same lines in analysing the example of προσήλυτος
(537–8), and, further, some Hebraistic uses of δικαίωμα, which, whether
or not they are classed as stereotyped renderings, raise the same questions
(539–40). In both cases Tov concludes that the translators used the words
mechanically, and recognition of new senses is not called for.
Tov’s view of προσήλυτος seems to me unsoundly based, however. First
the record must be put straight regarding the mistaken reference to Apollonius
Rhodius I, 834 in BAG and BDAG. This derives from the correct observation
of Hort (see MM) that the word occurs in the Scholia on that line (see edition of
A. Wellauer, 1828). We have there two examples, both clearly in the sense of
‘stranger,’ not ‘proselyte,’ in a non-Jewish source of uncertain, but not necessar-
ily late, date (see Fränkel, Einleitung, 105f.). There are also Patristic examples
of the same sense cited by Lampe, PGL. ‘Stranger,’ ‘newcomer,’ is, then, a pos-
sible sense; but even more, as etymology and the parallel formations ἔπηλυς, etc.,
suggest, it is likely to be the original one (so Chantraine, DELG, s.v. ἐλεύσομαι).
The difficulty in Exod 22:20, etc. then disappears: ‘sojourner’ was the meaning
intended there (and often elsewhere). J. A. Loader’s thesis, that προσήλυτος is
derived from προσέρχομαι in the technical sense of ‘be accepted as a convert to
Judaism’ (‘An Explanation,’ 270–7), would be convincing if he could cite clear
LXX examples of the verb in that sense, but there are none. (I am grateful to
Professor G.P. Shipp and Dr A.W. James for assistance with this note.)
VI
But now Tov takes a further step, tentatively at first, but with increasing
confidence. Having remarked (538) that ‘if a certain Greek word represents
a given Hebrew word in most of its occurrences, almost by implication
it has become a mere symbol for that Hebrew word in the translation,’
he continues: ‘In other words, if a lexeme as πατήρ represents ‫ אב‬in most
of its occurrences, its lexicographical description could be identical with
28
John Strugnell has orally suggested: YHWH read Adonaï or melek read Shah in Persian.
[Note added by RB editor.]
that of the Hebrew word, because it follows its Hebrew equivalent in

all its meanings and usages.’ When we reach the discussion of παντο-
κράτωρ (541), this possibility has become a certainty: ‘the lexical mean-
ing of παντοκράτωρ in Jer and the Minor Prophets is ‫ )ה(צבאות‬because
it reflects only this Hebrew phrase …,’ and further: ‘The second level or
dimension refers to the stage when the word came to be used as a stereo-
typed equivalent for ‫ )ה(צבאות‬in Jer and the Minor Prophets. At this stage
the meaning of the Greek has to be expressed as ‫ )ה(צבאות‬since it repre-
sented that word in all its usages.’29
This step seems to me hazardous. We have moved from regarding the
Greek word as a symbol, whose meaning in the translator’s mind may be
hard to define, to the quite different position that its lexicographical descrip-
tion is, beyond a doubt, exactly that of the Hebrew word. To do this is to
lose sight of our guiding principle that the translator’s intention determines
the meaning.
What in fact can be said about the translator’s intention in a case of
stereotyped rendering? At the outset we must recognise the limitations
to our knowledge. What was in the translator’s mind cannot be known for
certain. Care must he taken therefore not to assume anything without great
caution, and to do so only if there is some indication on which to build.
Now if a translator uses a stereotyped rendering, even with complete
consistency, does it automatically follow that he had full knowledge of,
and consciously intended to reproduce, the usage of the Hebrew word
as it would have been known to the original speakers and writers of the
language? It does not. It must surely be recognised that this method of
rendering is consistent with ignorance as much as knowledge. The trans-
lator’s comprehension of the true meaning of the Hebrew word may be
anywhere between completely accurate and completely inaccurate. It
cannot be assumed that it was the former simply on the basis that the
rendering is stereotyped. Therefore the lexicographical description of the
Greek word must not automatically be equated with that of the Hebrew
word.
The impetus towards making this equation is strong, for reasons not
difficult to see. The distribution of the Greek word in the text is identical
with that of the Hebrew word; it is hardly surprising, then, that in each
context the former appears, to one familiar with Hebrew, to have the same
meaning as the latter. But this is a matter of what one might call physical
29
Similarly on διαθήκη (‘Three Dimensions,’ 542): ‘the second dimension of διαθήκη is
fully identical with ‫ברית‬.’
location; it does not of itself tell us what meaning was in the translator’s
mind, and this, we have agreed, is to be the basis of the lexicographical
description.
These remarks are of course aimed at trying to establish the general
principles that must underlie our approach. I am not suggesting that a Greek
word used as a stereotyped rendering will never be described lexicographi-
cally in terms that coincide more or less fully with the description of its
Hebrew counterpart. There are likely to be a good many such examples
(e.g., νύξ ~ ‫)לילה‬. But the description will be arrived at primarily on the
basis of other observations than that the rendering is stereotyped. Each
word will be examined on its own merits and every indication utilised, but
the fact that there is stereotyped rendering will not be taken necessarily to
imply full knowledge of the sense of the Hebrew word.
VII
Let us look again at παντοκράτωρ. Tov, as we have seen, asserts that

‘the lexical meaning of παντοκράτωρ in Jer and the Minor Prophets is
‫)ה(צבאות‬.’ The point I am making is that this is true only if we can assume
that the translator(s) of those books had in mind the true meaning of ‫צבאות‬
and intended παντοκράτωρ to convey it. If there is any doubt about that
assumption we cannot say ‘the meaning of παντοκράτωρ is ‫צבאות‬.’ I think
most would agree that there is doubt in this case. The meaning of ‫צבאות‬
continues to be a problem today.30 It is very likely that the translator’s
conception of it did not correspond exactly with the true meaning. This is
to give the best estimate of his knowledge. At worst, it is possible that he
had no clear idea at all of it, but having adopted παντοκράτωρ for what-
ever reason as his standard equivalent, used it mechanically throughout.
What meaning of παντοκράτωρ he had in his thoughts in that case is of
course hard to say, especially as it need not have been clearly defined at
all. But I would suggest that if he was conscious of any it is most likely
to have been something close to the one current in Greek, since after all,
as Tov says (541), ‘the Greek word had a meaning of its own before it
was used in the LXX.’ In short, we cannot simply define the meaning of
παντοκράτωρ as ‫צבאות‬, but must try, however hard it is, to define the
word in terms of what was in the translator’s mind as he used it.
30
[The entry in HALOT, s.v. ‫ צבא‬B, presents two columns of data, lists seven proposed
meanings, and makes no decision; DCH s.v. ‫ צבא‬11: ‘the meaning of ‫ צבאות‬may be
(1) military … (2) celestial … (3) … intensive abstract plural.’]
The same argument applies to a well-known problematic case occur-

ring in the Psalms (and occasionally elsewhere), namely, παραπικραίνω,
which would be expected to mean ‘embitter,’ provoke.’31 No non-biblical
examples of this word are known apart from Philo, but the indications
of its meaning are clear: the simplex πικραίνω is well attested (Class.
Greek and later, including LXX) in the meaning ‘embitter,’ ‘irritate.’ ἐκπι-
κραίνω in Deut 32:16 (~ ‫ כעס‬hiph, parallel to παροξύνω) has a similar
meaning. In some LXX examples παραπικραίνω itself renders ‫ כעס‬hiph
(Jer 39:29 +): ‘provoke’ is clearly the sense intended. Such is the meaning
also in the two Philo examples (see BDAG). Finally, Hesychius glosses
it παροργίζω. Whether or not this compound is a coinage of the LXX, I
suggest it is formed on the analogy of the old word παροξύνω. Compare
the compounds παραζηλῶ, παροργίζω, παρενοχλῶ.
παραπικραίνω is the Psalms translator’s stock equivalent of ‫‘ מרה‬rebel.’
It is possible, but it does not at once follow, that παραπικραίνω has the
sense of ‘rebel.’ The step of supposing that the translator intended the
word in that sense should be taken with care, and only after all available
clues have been weighed. Though it is difficult to feel confident of any
conclusion, the indications seem to me to point away from the meaning
‘rebel.’
It is generally accepted that this rendering has arisen from associating
‫ מרה‬with the root ‫ מרר‬and especially the adjective ‫מר‬, ‘bitter.’ It follows
that its originator supposed the notion of bitterness to be present in the
Hebrew word and intended this notion to be conveyed by the Greek. In
other words παραπικραίνω was intended in something like its normal
meaning ‘embitter’ rather than in the sense of ‘rebel,’ which the translator
failed to recognise in the Hebrew word. If this line of argument is denied,
it becomes hard to explain why a word containing πικρ- was used at all.
παραπικραίνω can only have been used because it has the idea of ‘embit-
ter,’ which the translator saw in ‫מרה‬. After all, if a translator understood
‫ מרה‬to mean ‘embitter,’ ‘provoke,’ how else would he indicate it than by
using a Greek word that would normally mean that?
Of course it may be, in fact it is likely, that the Psalms translator’s ideas
were less clearly thought out than this, and that παραπικραίνω was a make-
shift rendering, with an ill-defined meaning in his mind. But the result is
the same: παραπικραίνω cannot be equated with ‫מרה‬. This is still true
if, following Flashar and Walters, we see this equivalent as having been
simply lifted by the Psalms translator from Deut 31:27. Walters himself
31
For details, with references to earlier discussions, see Walters, Text, 150–3.
says ‘the characteristic feature of these borrowed equivalents … is that

they betray a lack of understanding’ (152). It is precisely this lack of under-
standing that would make it impossible to equate παραπικραίνω with ‫מרה‬.
If the translator does not understand the true meaning of ‫ מרה‬then it can-
not be said that he intends his rendering to convey it.32
Walters reached the opposite conclusion to mine, namely, that the trans-
lation of παραπικραίνω ought to be ‘rebel’ in all cases (153). I have tried
to show that his own premises point the other way. Walters himself is seen
to go back on his conclusion when later on the same page he discusses
renderings of this type generally and says that if the lexicon records such
blunders it must do so ‘without ascribing the meaning of the Hebrew to
the Greek word, which in fact reflects its homonym.’
It should also be noticed that there are a few instances in which the root
‫ מרה‬is rendered by other, better known, words for ‘provoke,’ ‘irritate,’
namely παροξύνω and ἐρεθίζω with its derivatives.33 If παραπικραίνω
is to be taken as ‘rebel’ on the basis of the Hebrew, then why not these
too? But this cannot be seriously maintained. Rather, these examples show
that some of the translators did actually believe ‫ מרה‬means ‘provoke.’
It is fair to suppose that among them was the translator of Deut 31:27,
who in the one verse renders ‫ מרי‬by ἐρεθισμός and ‫ מרה‬hiph by παρα-
πικραίνω.34
VIII
Finally I return to the expression ἐρωτῆσαί τινα εἰς εἰρήνην and vari-
ations of it.35 As I have indicated above, Tov’s conclusions appear to me
sound as far as they go, and perhaps we should go no further. But it may
be that here is a case in which the further step of giving the Greek the
same meaning as the Hebrew is justified. What can be said about the prob-
able state of the translators’ understanding here? This is a fairly common
32
Note that ‘rebel’ could have been expressed by ἀπειθῶ or ἀφίσταμαι, both of which
are found in the LXX, sometimes even as renderings of ‫מרה‬.
33
Num 20:24 παροξύνω, Deut 21:18 ἐρεθιστής, 20 ἐρεθίζω, (all these ~ ‫)מרה‬, 31:27
ἐρεθισμός (~ ‫)מרי‬. Walters notes the last three (Text, 150).
34
Cf. Tov’s discussion of the rendering of ‫ שׁדי‬by ἱκανός (‘Three Dimensions,’ 540). ‘We
must … ascribe to ἱκανός that meaning of the Hebrew word which the translator had in
mind and not the one which we ascribe to the Hebrew word.’ The same principle applies
to these unexpected renderings of ‫מרה‬.
35
List of examples in Thackeray, Grammar, 40–1. In some instances τά is added before
εἰς, which looks like an attempt to make the expression more intelligible as Greek.
Hebrew idiom, of a kind that even a translator with a poor knowledge of

Hebrew might be expected to know. Some of the translators certainly had
a good grasp of it, as idiomatic renderings elsewhere show.36 In any case
it is not difficult to interpret: its meaning is fairly obvious in context, and
it is readily linked with the well-known use of ‫ שׁלום‬in greetings and the
like. In the light of these considerations it might be argued that the transla-
tors who used this rendering were well aware of the meaning of the Hebrew
expression and did intend the Greek in the same sense. If so, the conse-
quences for the LXX lexicographer are clear: the Greek expression is to
be given the meaning of the Hebrew one.
IX
These few remarks on stereotyped renderings do not go very far into

a highly complex subject. The enormous range of individual examples, no
two exactly alike, makes generalisation difficult and uncertain. Principles
will always need to be open to modification in the light of practical expe-
rience. But it is clear, I think, that in cases of stereotyped rendering we
cannot proceed at once to give the Greek word exactly the meaning of the
Hebrew word, but must take the more difficult course of trying to establish
what meaning the translator actually intended. If the result is that the lexi-
cographer cannot give an exact definition of the word, this may often be an
accurate reflection of the fact that the translator himself had no definable
meaning in mind.
Postscript
The first topic discussed, ‘equivocal’ renderings, which gives its name to the
paper, can be seen to connect with an emerging perspective on LXX translation
technique that sees the potential for ‘multicausality’ in producing renderings. This
approach has recently been introduced from translation studies by James Aitken
and elaborated further by Marieke Dhont.37
Extensive discussion of how to determine the meaning of a Greek rendering
in the LXX has taken place since 1980, much of it generated by the NETS project.
Our understanding of the translators’ methods, bilingual interference, the ‘text as
produced,’ the role of context and the Hebrew original, and more, has advanced
36
E.g., ἠσπάσαντο ἀλλήλους/ αὐτόν Exod 18:7; Judg 18:15A; ἠρώτησεν αὐτούς Πῶς
ἔχετε; Gen 43:27.
37
Aitken, ‘Origins of καί γε,’ 37; Dhont, ‘Multicausality.’
greatly. Even so, the above observations on stereotypes retain their validity and
usefulness. The examples παντοκράτωρ and παραπικραίνω are prototypical
cases.
I continue to hold to the principle enunciated by Tov, that the text means what
the translator intended it to mean. It is important as our ultimate guide, even
if determining intention is problematic and the translator’s intention can never
be fully known. This too has been the subject of continued debate. See further
Essay 15.
4
THE FUTURE OF ZHN IN LATE GREEK
1980
Abstract
The paper takes as its starting-point the interpretation by G. D. Kilpatrick of

a statement in the Antiatticista about the forms of the future of ζῆν. I argue that
he has misunderstood the lemma and taken it in the opposite way to what was
intended, in view of the usual purpose of the Antiatticist’s remarks. The evidence
for the forms ζήσω, ζήσομαι, and βιώσομαι is then collected from surviving texts
and an interpretation of the historical development is offered, taking account of the
significant role of Atticism. Conclusions about NT usage are offered but not deci-
sions on the NT text.
The form of the future of ζῆν in the NT has been the subject of some
discussion since G. D. Kilpatrick’s important paper ‘Atticism and the Text
of the Greek New Testament.’ Kilpatrick used it to illustrate his thesis that
where the NT textual tradition shows a Koine and an Atticising variant the
former is to be regarded as original, since a change from Koine to Atti-
cistic under the influence of Atticism is more likely than a change in the
opposite direction.1 He argued in favour of reading the middle ζήσομαι
in John and elsewhere on the premise that ζήσομαι was the Koine form
and ζήσω the Atticistic. There is however a fundamental error in the argu-
ment by which Kilpatrick established this premise. This error remains
undetected in the most recent discussion of the example, by Fee.2 The
necessary correction has been suggested by Martini, but tentatively and
without much supporting detail.3 It seems worthwhile therefore to discuss
1
Kilpatrick, ‘Atticism and the Text,’ 132–3.
2
Fee, ‘Reasoned Eclecticism,’ 188–9. Fee opposes Kilpatrick’s view on other grounds;
similarly in his earlier brief discussion in Fee, Papyrus Bodmer II, 49. Kilpatrick’s premise
is accepted at face value also by Kieffer, Au dela des recensions?, 160–1 and Elliott,
‘The UBS Greek NT,’ 299.
3
Martini, ‘Eclecticism,’ 154–5. [Martini correctly explains the import of the Antiatticist’s
lemma.]
the point fully and establish that Martini is correct and that Kilpatrick’s
view of this example rests on a misunderstanding. In addition there is need
for a fuller examination of the attestation of the two forms outside the NT.
All discussions so far have relied on the material in the standard lexicons
and grammars. As we shall see, this hardly gives an adequate picture. Until
these points have been cleared up, the debate over the NT readings cannot
proceed on a sound footing.
Kilpatrick quotes the following observation of the ancient author known
as the Antiatticista:
ζήσει· Πλάτων Πολιτείας ἕκτῳ, οὐ ζήσεται.4
He continues:
ζήσει is the correct form supported by a reference to Plato’s Republic
and ζήσεται is condemned. This is confirmed by the evidence of the
authors. ζήσω is used by Attic writers and the Atticisers, ζήσομαι
(except for [Dem.] xxv.82) is the form current among non-Attic and
later writers. On this evidence we would expect the New Testament
writers to use ζήσομαι.
What needs to be recognised is that the Antiatticista’s remark implies

the opposite of what Kilpatrick supposes.
The Antiatticista was a rebel against the Atticising fashion of his time,
which looked back to Classical Attic as the model of correct Greek and
condemned anything in the living language that did not conform to it.
He did not, however, simply reject all such appeal to the past as mis-
taken, but sought to refute the Atticists by applying their own methods
against them. Where they stigmatised a feature as unacceptable, he
produced an occurrence in a Classical author, thus proving them wrong
by their own criterion.
His purpose and method can be clearly seen from an entry such as
the following:
γέμειν τὴν ναῦν μόνον φασὶ δεῖν λέγειν, τὰ δ᾽ ἄλλα μεστὰ λέγειν.
ἐλέγχει δ᾽ αὐτοὺς Εὔβουλος Εἰρήνῃ.5
4
Bekker, Anecdota Graeca, 1:97 [Valente, Antiatticist, ζ 2]. Πλάτων Πολιτείας ἕκτῳ =
Plato, Rep. 9.591c.7. [Additional notes by NovT ed.: The text of the entry has not been
commented on by Guil. F. A. van Dam, Observationes in Lexica Segueriana (thesis
Leiden; Roterodami, 1873), nor by L. J. Sicking, Annotationes ad Antiatticistam (thesis
Leiden; Amstelodami, 1883). Photius’ Lexicon (ed. S. A. Naber, Leiden, 1864, 247)
contains the entry ζήσει· ἀντὶ τοῦ ζήσεται, which seems to mean: where the unusual
ζήσει occurs, it stands for the more normal ζήσεται.]
5
Bekker, Anecdota Graeca, 1:86. For similar entries see, e.g., γοργός (86), εὐειδής (92),
κομψόν (102).
THE FUTURE OF ZHN IN LATE GREEK 39
They say one ought to say only ‘the ship is full’ (γέμειν), but with
reference to other things (one ought) to say (they are) ‘filled’ (μεστά).
Eubulus in Peace refutes them.
The Atticists (‘they’) condemned the current use of γέμειν with

reference to objects other than ships, and advocated instead the use of the
adjective μεστός. An occurrence of the condemned feature in Eubulus
proves its respectability.6
The Antiatticista expresses the point in a variety of ways. Many entries
(at least in their surviving form) are much more abbreviated than the one
just quoted. The extreme is reached with those like διάφορος· ἐχθρός (88),
and βόλβιτον· Ἱππῶναξ (86), where nearly everything is taken for granted.
A little less cryptic are those on the pattern of ἀποκριθῆναι· οὐκ ἀποκρί-
νασθαι.7 This remark is to be understood as follows: ‘ἀποκριθῆναι, the
Koine form condemned by the Atticists, is in fact found in Attic, instead
of ἀποκρίνασθαι, the standard Attic form approved by the Atticists.’8 This
type may be expanded by the addition of a reference to an Attic occur-
rence, as for example ἀρτύματα· οὐχ ἡδύσματα. Σοφοκλῆς Φαίαξιν.9 The
meaning is: ‘Sophocles in Phaiakes has ἀρτύματα, not the recommended
ἡδύσματα.’ With this we reach the closest available parallel to ζήσει·
Πλάτων Πολιτείας ἕκτῳ, οὐ ζήσεται.10
In any case it is quite plain that however it is expressed, the Antiat-
ticista’s point is always the same: a Classical example is cited (or it is
implied that one is available) to prove the acceptability of a Koine feature
rejected by the Atticists.
The correct interpretation of the remark on ζήσει, then, is clear: the
active was in use in the Koine of the writer’s time, the second century AD;11
the Atticists objected to it and advocated the middle. The fact that the
Antiatticista cites Plato in support of the active makes this certain.
This interpretation, moreover, agrees with what we should expect on
general grounds. It is well known that the tendency of the Koine was to
6
That is, Eubulus Comicus (IV BC). The play is not otherwise extant. For other Class.
exx. see LSJ, s.v. γέμω 2.
7
Bekker, Anecdota Graeca, 1:77. Similarly ἀλήθειν (78), γαγγαλίζειν, γραία (87).
8
By ‘Attic’ is meant strictly the dialect of Classical Athens. This should be carefully
distinguished from what the Atticists thought was Attic (= ‘Atticistic’).
9
Bekker, Anecdota Graeca, 1:82. Cf. αὐτοσχεδιάζειν (83), ἐκδημία (93).
10
It is worth noting in addition that the censured form or word is almost invariably put
first. Exceptions are at once made clear by the words that follow: e.g., γενέθλιον
ἡμέραν ἀξιοῦσιν ἀεὶ λέγειν, οὐ γενέθλια (86). [Thus ζήσει is likely to be the cen-
sured form.]
11
For the date see Debrunner–Scherer, Geschichte, 2:17.
replace middle futures of active verbs by the active.12 This was a slow
process and not all verbs were affected at the same time, but there can be
no doubt of the existence of the trend, and that it was always in the same
direction. It is most unlikely that any Atticist would have wished to, or
needed to, advocate ζήσω, when the tendency in the spoken language was
to introduce this form anyway. Rather, we can expect the Atticists to have
tried to defend the threatened middle, which they naturally assumed to be
good Attic.13
II
Let us now consider the evidence of the surviving texts. I begin by

setting out such attestation as I have been able to find, arranged chrono-
logically. Completeness cannot be claimed, but the list goes well beyond
existing collections, and is sufficient to allow us to reconstruct with some
confidence the history of ζήσω and ζήσομαι. Besides these two forms,
there is another to be taken into account, namely βιώσομαι (-σω).14
2.1. ζήσω
Attic of V–IV BC: Ar., Plut. 263 -ειν, Fr. dub. 899a -εις. Pl., Rep. 5.465d.2
-ουσι, 9.591c.7 -ει (var. causa: βιώσεται precedes), Leg. 7.792e.7
διαζήσει.
Of uncertain date and origin: [Epich.] 267 -ων.15 ‘Menander,’ Γνῶμαι
μονόστιχοι, 186 -εις.16
12
Moulton, Grammar, I, 154–5; BDF, §77; Thackeray, Grammar, 231–3; Bauer in BAGD,
xiv; Mandilaras, Verb, §367.
13
Cf. Browning, Greek, 52: ‘Atticising writers continuously fail in their purpose. Either
they admit Koine forms censured by the grammarians … or they overcompensate and
produce false Atticisms. … The literature of this period is full of middle voices where
Attic uses in fact the active, of wrongly used datives. …’ Martini’s interpretation (‘Eclec-
ticism,’ 154–5) agrees in essentials with mine.
14
Sources of information were LSJ; Mandilaras, Verb; Schmid, Atticismus; Stephanus,
Thesaurus; Thackeray, Grammar; Veitch, Greek Verbs; and esp. the standard author
indexes. The following authors are fully covered, thanks to the indexes noted: Aristoph.
(Todd); Demosth. (Preuss); Epict. (Schenkl); Eur. (Allen–Italie); Joseph. (Rengstorf);
Philo (Mayer); Plato (Brandwood); Plb. (Mauersberger: no exx.). Coverage is extensive
if not complete for: Arist. (Bonitz); Dio Cass. (Boissevain); Lucian (Jacobitz); Plut.
(Wyttenbach); Xenoph. (Sturz).
15
That is, attributed to the Doric poet Epicharmus Comicus (V BC). Perhaps to be read
ζησῶν (= Doric fut.): so KB, Grammatik, I.2, 436; but ζήσων Kaibel.
16
The collection known as the ‘Maxims of Menander’ continued to be augmented right
up until Byz. times, and contains much that Men. ‘could not have written’: Edmonds,
Fragments, IIIB, 901–3.
LXX MS tradition: Prov 9:11 ζήσεις B etc. (-η, -ει AV); Sir 37:26
ζήσει A (-εται B etc.); Amos 5:6 ζήσετε A (-εσθε L, ζῆτε B etc., ζήσατε
Rahlfs). All three of these are likely to be secondary.17 It is hard to know
when they entered the tradition.
A fourth example is of special interest: at 4 Kgdms 18:32 the majority
reading is ζήσετε, while bdhjoprvxyc2e2 have ζήσεσθε. ζήσετε is thus
the reading of the Kaige recension, and ζήσεσθε in all probability the
original LXX.18 We may therefore roughly date this example of the active
to the first half of I AD.19 (The two occurrences of the active at Ps 137:7
and 142:11 (ζήσεις με) are not relevant, as ζήσεις is causative, a quite
different use which requires the active.)
I AD: Philo, Quis Heres 258, Congr. 176 both ζήσεις, in quotations
of Gen 20:7 and 27:40 respectively (but LXX ζήσῃ).20 Joseph., AJ 12.315
ζήσουσι, 16.119 -ομεν, BJ 7.394 ἐπιζήσουσι.
I–II AD: Epict. 3.26.24 -εις, 4.1.104 -οντα, Plut., Mor. 194a κατα-
ζήσεις, in a quotation of a saying of Epaminondas, reported also in
Aelian, VH 9.9, in almost identical words. Both authors, then are quoting
a source.
II AD: Aelian, NA 7.2 καταζήσουσι, VH 11.9 καταζήσεις (see above).
Artem., Onir. 200.7 -εις, 200.18 -εις, 201.1 -εις, -ει, 201.4 -εις.21 Barn.,
Ep. 6.17 -ομεν. Lucian, Alex. 34 -εις, in a quotation of an oracle of
Alexander the false prophet.
II–III AD: Dio Cass. 53.9.2 -ειν, 58.27.3 -ων, 58.28.2 -ων. Here may
be added the example in Cramer, Anecdota Graeca Oxon. II, 372: ζῶ· τὸ
ἐπὶ τῆς ζωῆς, ὁ μέλλων ζήσω ἐστίν.22
17
Rahlfs accepts act. in Prov 9:11, without compelling reason. Arguments in favour of
mid. are: (a) this is the usual LXX form; (b) Prov itself has mid. in 9:6; 28:16; (c) Prov
is in ‘literary’ Greek. There are other, weakly attested exx. of the act. to be found, as,
e.g., 4 Kgdms 20:1 ζήσεις v, ζήσῃ rell.
18
Barthélemy’s discovery (Les devanciers) concerning Kgdms βγ and γδ (our example
is in the latter) is well known. The Kaige recension of Kgdms γδ has the mid. other-
wise (4 Kgdms 1:2; 4:7; 7:4, etc.; likewise 2 Kgdm 12:22 (βγ), and Hab 2:4 in Barthé-
lemy’s Dodecaproph. fragments (Les devanciers, 175). The act. in 4 Kgdms 18:32 is,
then, an accidental change, not one of the deliberate changes that characterise this
recension.
19
See Barthélemy, Les devanciers, 144, 148; Cross and Talmon, Qumran, 313.
20
Were these readings already in Philo’s LXX, or did they enter later in the MS tradition
of Philo? (No v.ll. are reported by Cohn-Wendland.) They appear to have no connexion
with the recension identified by Katz in Philo’s Bible.
21
Refs. to p. and l. number in Pack (Teubner ed., 1963).
22
It is hard to know what to make of this, esp. without a date; it is presumably not earlier
than II AD, but could be much later. The fact that the writer states the form of the fut.
(there is no mention of other tenses) suggests there was difficulty about it.
2.2. ζήσομαι
Ionic, probably V–IV BC: Hippocr., Nat. Puer. 30.8 -εται.23

Late Attic-Hellenistic, IV BC: Arist., Pol. 7.1327b.5 -εται. Probably
to be included here is [Demosth.] 25.82 -εται.
LXX: the middle is usual throughout (c.102 times).
I AD: Philo, Leg. All. 2.81 -εται, Somn. 2.31 -εσθαι, Ioseph. 264
-εται (+ 6 exx. in LXX quotations). Joseph., AJ 1.193 -εται, 4.302 -ομέ-
νους, 9.75 -όμενοι, 12.281 -εσθε. Antiphilus in Anth. Gr. VII 175 -εσθε.
I–II AD: Plut., Mor. 79e -εται, 1082b -εσθαι, c -εσθαι, -εται.24
II AD: Apostolic Fathers 43 times (+ 4 exx. in quotations).25 Artem.,
Onir. 199.8 -εται, 201.8 -εται. Lucian, Paras. 112 -εται × 3, [Lucian]
Asin. 27 -εται.
II–III AD: Dio Cass. 53.34.2 -ῃ. Philostr., VA 1.7 -ομαι, 37 -εται.
III AD: Diog. L. 6.44 -οιτο. POxy 6.907.23 (276 AD) -εται.
2.3. βιώσομαι (-σω)
Attic of V–IV BC: e.g., Ar., Equit. 699. Eur. Alc. 784. Isoc., Areop.
78, Pax 35. Xenoph., Ages. 9.4, Cyn. 7.4, Cyr. 5.4.33, Mem. 1.7.2; 4.8.8;
2.1.23 (δια-). Plato 18 times (incl. 2 in cmpds., δια-, συμ-).26
Late Attic-Hellenistic, IV(–III) BC: Arist., Eth. Nic. 1101a.17, 1127b.27,
Magn. Mor. 1212b.31 (συμ-), 1213a.28 (συμ-), Pol. 1267b.36. Menand.,
Epitr. 694, Peric. 977 (Sandbach), Fr. 581.6 (Körte). ‘Menand.,’ Γνῶμαι 270
(act.).
LXX: Sir 13:5 (συμ-); Prov 7:2; Job 29.18; 4 Macc 6:20 (ἐπι-) (last
three act.).
I BC: Philodem., Oec. 13.27.
I AD: Philo, Leg. ad G. 85, Virt. 47, 127, Spec. Leg. 3.154, Dec. 49,
Vita M. 2.48, Somn. 2.31, Quod Det. Pot. 62, 154, Opif. M. 172 (total 10).
Joseph. 11 times (twice act.: AJ 1.46, 8.337).27
23
Ed. Joly (Budé ed., vol. XI, 1970) = Littré VII, 536. Joly (p. 23) accepts the traditional
dating at the end of V BC for this; Pauly-Wissowa, RE, VIII.2, col. 1828, say 375 BC at
latest.
24
In Plut., Mor. 1082b, c ζήσομαι, instead of Plut.’s usual βιώσομαι, is more or less
demanded by the context: … εἰ … ὁ βαδίζων ἐβάδισε καὶ βαδιεῖται; τὸ δεινῶν,
φασί, δεινότατον, εἰ τῷ ζῶντι τὸ ἐζηκέναι καὶ ζήσεσθαι συμβέβηκεν …, and so on
throughout.
25
For the full list of references see Goodspeed, Index Patristicus.
26
See Brandwood, Index to Plato.
27
See Rengstorf, Concordance.
I–II AD: Epict. 1.4.27, Ench. 24.1, Fr. 8.12 (δια-). Plut., Mor. 100c,
101d, 141d (συμ-), 160b, c, 187e, 563d, 1043c x 2 (one συμ-), 1119d,
Demetr. 1.3, 47.4 (κατα-), Publ. 3.1 (δια-), Titus Flam. 20.3 (κατα-)
(total 14 in Plut.).
II AD: Lucian, Charon 20, Dial. Meretr. 15.1, Herm. 6.77, Nav. 26,
Par. 12, Tox. 37, VH 1.12 (κατα-) (total 8). Appian, BC 4.119 (act.).
II–III AD: Dio Cass. 52.18.6, 27.5, 39.2; 58.23.2, 27.3, 28.1; 69.17.1
(total 7).
III AD: Diog. L. 2.68.
III
This evidence, though it appears unwieldy at first sight, in fact yields

a clear picture of the history of these forms. I suggest the following
interpretation.
The usual Attic future was neither ζήσω nor ζήσομαι, but, as everyone
who has learnt their ‘principal parts’ knows, βιώσομαι. The occurrences
noted are sufficient to confirm the traditional statement.28 Plato’s 18 exam-
ples, compared with only three of ζήσω, one of which is introduced for
a special reason, are conclusive on this point. It is also significant that
βιώσομαι is preferred by Xenophon, the ‘forerunner of the Koine,’ and
is still to be found in Aristotle and Menander, on the threshold of late
Greek. Thereafter it declines in the spoken language.
Alongside βιώσομαι, Attic occasionally used ζήσω, but it did not last
beyond Plato’s time. The examples attributed to Epicharmus and Menander
are too untrustworthy to affect this conclusion.
ζήσομαι cannot be traced back to Classical Attic. Its attestation, first
in Ionic, next in late Attic, then in the Koine, follows a familiar pattern.
It is an innovation of the Koine, originating in Ionic.29 Its appearance
on the scene is linked with the obsolescence of βιώσομαι, on which it is
modelled.30 This explains why the middle takes over so completely from
the already existing active, a development that is puzzling otherwise.
28
See, e.g., Cobet, Variae lectiones, 610; Schwyzer, Grammatik, I, 675; Smyth, Grammar,
698.
29
This interpretation is strongly supported by the fact that the aor. ἔζησα, the late Greek
equivalent of Attic ἐβίων, also originates in Ion.: see BDF, §75; Bauer in BAGD, xiv;
LSJ, s.v. ζῶ. On the Ionic element in Koine Greek generally see, e.g., Thumb, Hellenis-
mus, 68, 209.
30
Schwyzer, Grammatik, I, 782 n. 5: ‘Hippokr. ζήσομαι (: ἔζησα) für älteres βιώσομαι.’
A parallel is φάγομαι on the analogy of ἔδομαι, the old form that φάγομαι replaces.
Aristotle represents the transitional stage when ζήσομαι and βιώσομαι

were in use side by side.
ζήσομαι continued as the reigning form throughout the early Koine,
as the LXX attests. Then early in the first century AD, perhaps some-
what earlier, the form was affected by the current which we have already
noticed, the trend towards replacement of future middles by the active.31
ζήσω now begins to appear in our evidence, alongside the middle. The
καίγε recension provides the earliest datable post-Classical example; then
we find it sporadically in writers from the first century onwards. We may
be sure that it was the usual form in the spoken language. This was a new
development, not a revival of the old Attic ζήσω.
Close on the heels of the appearance of ζήσω came the Atticist reaction
against it. This was part of their general stand against the change that
was affecting future middles. The active was condemned and ζήσομαι
promoted as the correct form. The Antiatticista testifies to this. His evi-
dence relates to the second century, but the reaction probably began ear-
lier.32 It does not matter that the middle was not the true Attic form; the
important thing is that it was believed to be. The case of ζήσω/-ομαι was
thought to be just like all the others (ἀκούσω/-ομαι, etc.). Besides ζήσο-
μαι, genuine Attic βιώσομαι was revived and encouraged by the Atticists,
as we can deduce from its frequency in Atticising authors. This was quite
artificial, and could have had little impact on ordinary language.
All this is likely to have led to considerable uncertainty, if not confusion,
among writers of the first century and later. (The efforts of the Antiatticista
in support of ζήσω would only have added to it.) That this was indeed the
case seems clear from our evidence. All three forms are liberally attested,
and hardly any author keeps consistently to one or the other. A general
pattern can nevertheless be seen: ζήσω tends to be avoided in favour of
ζήσομαι or βιώσομαι. This is what we expect, given that the bulk of the
evidence is more or less ‘literary.’ This is true even of the POxy example
(see below). Among the authors noticed, Epictetus and Artemidorus are
generally thought to be the nearest to the ordinary Koine.
Let us look more closely at the practice of individual writers. Philo,
Plutarch, and Lucian have ζήσομαι occasionally (3 or 4 times each), but
βιώσομαι predominates and is evidently their usual word (10, 14, and
31
The fact that some verbs are affected much earlier than this (e.g., ἀκούσω appears at
least as early as III BC) makes no difference. Each verb has a separate history.
32
As is well known, Atticism shows itself as early as Dionysius of Halicarnassus, in I BC,
and its roots can be traced even earlier: see, e.g., Thumb, Hellenismus, 252; Schwyzer,
Grammatik, I, 130.
8 times respectively). ζήσω occurs, but only in quotations: this is some-

one else’s usage, not theirs. Josephus is much the same (ζήσομαι 4 times,
βιώσομαι, -σω, 11), but does not succeed in avoiding ζήσω altogether
(3 times). The Apostolic Fathers represent as it were the intermediate level.
They keep to ζήσομαι with one exception, but do not aspire to βιώσο-
μαι. Epictetus and Artemidorus accord with expectation in having more
examples of ζήσω than the middle, but Epictetus surprises us by also using
βιώσομαι (3 times; the evidence is incomplete for Artemidorus). Dio
Cassius, a little later, also seems uncertain of himself. βιώσομαι is his
usual word (7 times), but he has ζήσω more often than the middle (3:1).
For the remaining authors the evidence is too fragmentary to make much
of. As to the example of ζήσομαι in POxy 6.907.23, this is not, as one
might be tempted to suppose, evidence of popular Greek. The document
is a will, likely to be cast in formal language.33 Moreover its literary level
is clearly marked by τε and βούλομαι, to look no further than the imme-
diate context.
IV
Rather a lot of ground has had to be covered in order to reach our goal,
but we are now in a position to draw some conclusions for the NT. One
thing is abundantly clear. The usage of the first century AD was unsettled.
We cannot say with certainty which out of ζήσω and ζήσομαι a given
writer would have used, or whether he would have been consistent. The
most that can be said, in my opinion, is that the closer the writer’s lan-
guage to everyday Koine the more likely it is that he would have used
ζήσω. On this basis we should expect ζήσω in John, unless he was quoting
or wished to achieve a special effect. (The same would apply to Mark and
Revelation.) The remaining books might be expected to show a tendency
to prefer ζήσομαι.
Subsequent influences on the MS tradition are equally difficult to reduce
to absolutes. It is quite likely that there was a tendency to alter the active
to the ‘correct’ middle in accordance with Atticist precepts. The reverse
change, though less likely, is also possible (cf. the LXX). Where this hap-
pened it would be a case of an occasional, unintentional alteration in the
direction of the scribe’s own speech.
33
καταλείπω δοθῆναί τε βούλομαι κατ᾽ ἔτος ἐφ᾽ ὃν ζήσεται χρόνον. …
The text of the NT has been purposely left out of this enquiry, so that
the linguistic background could be established independently. I must leave
it to others to examine the MS readings in the light of the results. Apart
from its importance for the NT, the case we have studied also has some
interest as a specimen of what happened when the Atticists tried to turn
the clock back.
Postscript
Kilpatrick responded in ‘Atticism and the Future of ζῆν’ (1983). He did not
concede that he had taken the Antiatticist’s statement the wrong way, and mar-
shalled many arguments in support of the correctness of his interpretation. Among
them he put forward the suggestion that ‘abbreviation or corruption [may] both
have caused the inconsistency’ (147), thus admitting that his interpretation was
inconsistent with the usual purpose of the Antiatticist. It is hard to see what cor-
ruption has occurred or what emendation would be plausible to reverse the import
of the lemma, when the text of Plato confirms the occurrence of the active ζήσει
claimed by the Antiatticist.
Other discussions are found in: Kilpatrick, ‘Eclecticism and Atticism’ (1977),
109–11; De Jonge, ‘De nieuwe Nestle’ (1980), 312–3. There is further description
of the Antiatticist’s methods in Lee, ‘Atticist Grammarians’ [= Essay 19].
5
SOME FEATURES OF THE SPEECH OF
JESUS IN MARK’S GOSPEL
1985
Abstract
All five examples of the obsolescent particle μέν in Mark’s Gospel (leaving
aside 16:19) occur in Sayings of Jesus. It is argued in this paper that the location
is significant: Jesus uses a prestige feature appropriate to his status. The con-
nexion with insights from sociolinguistics and the phenomenon of ‘formality’
are elaborated. Then four further features of a similar kind in Mark are presented
in detail (εὖ, φάγοι, ὦ, οὐ μή), and several others are noticed briefly. Parallels
in the LXX and elsewhere in the NT are noted. Brief remarks on the implications
for the history of the Sayings are offered.
In the Gospel of Mark the particle μέν seldom occurs. ‘Surprisingly

rare,’ Taylor calls it, but does not pursue the matter.1 There is, however,
something noticeable about the five occurrences in Mark (leaving aside
one in the Longer Ending): they are all in words spoken by Jesus. This fact
must have been noticed before, but as far as I am aware no significance
has been attached to it. I wish to suggest, however, that this distribution is
not accidental, but is in fact the key to understanding the use of the word
in Mark. Furthermore, I shall argue in this paper that we have here only
one instance of an important but overlooked phenomenon in New Testa-
ment Greek. I begin with μέν, the feature first observed, which also pro-
vides a good starting-point here.
It is well known that in post-Classical Greek there is a general decline
in the particles.2 That is to say, the wide and subtle range of Classical par-
ticles is sharply curtailed, leaving a few to do most of the work; at the
1
Taylor, Mark, on 4:4.
2
See Blomqvist, Greek Particles, 132, with references to the main discussions; also Thrall,
Greek Particles, 3–6, showing the decrease in combinations of particles.
same time there is a long-term tendency for all post-positives to be elimi-

nated in favour of words that come first in the clause. Though there is
room for further study of the details, the general trend is incontestable.
It is clear that μέν (…δέ) shares in this decline. In the Ptolemaic period
there is already a perceptible falling-off in its frequency.3 But this is not
all. There is now a new factor, one which makes all the difference for our
purpose. When μέν (…δέ) does occur, it tends to be found, Mayser tells
us, in educated, official writing rather than private letters, which are our
best evidence for everyday language. This indicates that we are dealing
with one more instance of a familiar phenomenon in post-Classical Greek,
the tendency to look back to the earlier language as the model of good
Greek and to maintain its idioms after they had become obsolete in the
spoken language. This reaches its peak in the second century AD, but
the trend is discernible long before. It seems clear, then, that quite early
in the Koine μέν (…δέ) was no longer the living idiom that it undoubt-
edly had been in Classical Greek, but had become, on the whole, a sign
of an attempt at more educated Greek, a prestige feature consciously
used.
The evidence of the LXX points in the same direction. To begin with,
the total number of occurrences, 215, is not high.4 Then if we distinguish
between the literary books and the rest, following Thackeray’s classifi-
cation,5 the former are found to have 181 examples (of which 2–4 Macc
account for no less than 117), while all the rest together have only 34. In
the Pentateuch, which in level of language is a fair specimen of unpreten-
tious Greek of the third century BC, μέν is used very sparingly: there are
only 19 examples.6 The contrast with 2–4 Macc is striking, especially when
one bears in mind that the Pentateuch is more than four times the length
of 2–4 Macc.7
To what extent is the rarity of μέν in the translated books due to the
absence of any Hebrew equivalent? There is no way of knowing for cer-
tain, but I would argue that this is not the main factor. If it was, we might
3
Mayser, Grammatik, II.3, 128; sample figures 125. An indication of frequency of μέν
in Class. Greek is given by figures for these authors (all approx. except Hdt.): Ar. 630;
Hdt. 2567 (Powell); Pl. 9000; D. 4070; Th. 1750; X. 1420 (last two from Jannaris,
Grammar, 400).
4
Figures based on HR and Rahlfs’s text. [The ‘Accordance’ tally is close: 218 (Ra.).]
5
6
Seven of these are μέν solitarium. A curious detail is that a high proportion of the
exx. (12) are alike in that μέν follows εἰ (ἐάν), sometimes with, sometimes without εἰ
(ἐάν) δέ in the next clause. Perhaps we should think of the obsolescence of μέν in the
living language as proceeding at a different rate for different kinds of use, the use in
εἰ-clauses being one of the last to go.
7
Pages of Rahlfs’s text: Pent. 354; 2–4 Macc 85.
SOME FEATURES OF THE SPEECH OF JESUS IN MARK’S GOSPEL 49
expect μέν not to occur at all. And the comparison with δέ is suggestive:
δέ is used with great frequency in some of these books, despite the availa-
bility of the more literal equivalent καί.8 I conclude that the translators were
not necessarily constrained by the original in their use of particles, just
as in general they show readiness, especially in the Pentateuch, to employ
idiomatic Greek were appropriate. It seems probable, therefore, that the
LXX reflects fairly well the status of μέν in the language generally.
The history of μέν in the post-Ptolemaic Koine has not yet been fully
studied, but there is no reason to doubt that the earlier trend continues,
and the NT itself confirms this. Nevertheless our conclusions will be more
secure if we can obtain further evidence of how μέν was being used in this
period. I have therefore made a survey of 18 volumes of The Oxyrhynchus
Papyri, covering a large sample of texts of the first to fifth centuries AD.9
The result agrees convincingly with what we have seen already.
The total number of texts surveyed was 659, of which 90 are private
letters, and 30 or so domestic ephemera (lists, private accounts, and the
like). The rest are official and private documents, all of a more or less for-
mal nature. These include such texts as a letter of the Emperor, records of
court proceedings, leases, and contracts. The occurrences of μέν total 100,
of which 81 are in the formal documents, and 19 in the private letters.
The figure for the former group causes no surprise. Examination of the
latter brings to light some significant points. (a) It is at once noticeable that
μέν tends to occur in stilted, formulaic language, especially at the beginning
of a letter. One particular phrase, πρὸ μὲν πάντων (παντός) recurs 8 times
as part of an opening formula, usually the well-known πρὸ μὲν πάντων
εὔχομαί σε ὑγιαίνειν.10 (b) In nearly half the examples (9 to be exact), μέν
is not followed by δέ or other correlative. (c) Some of the letters in which
μέν occurs show definite signs of a higher level of education. An unmis-
takable instance is POxy 17.2109 (late I AD), with two examples of μέν
(lines 17, 40), both followed by correct δέ. Other features include γε (50),
the learned-sounding word ἕξις (16), and a potential optative with ἄν (47).
It is no accident that the writer is a student in Alexandria writing home
to his father – no doubt in his best Greek.11 (d) This is the only letter using
8
[‘Accordance’ figures for δέ in Pent. (Gött. ed.): Gen 853; Exod 413; Lev 128; Num 74;
Deut 100. Μέν 19 times in Pent.]
9
POxy vols. 17–46 except those containing only literary texts, and vol. 22 with unreliable
index. The few texts of VI AD and later are omitted from consideration.
10
For a full treatment of this formula, and of others in which πρὸ μὲν πάντων appears,
see Steen, ‘Les clichés,’ 155–6; Exler, Ancient Greek Letter, 107–13. This one phrase
must account for a large number of the occurrences of μέν in the papyri.
11
The only illiterate letter with μὲν ... δέ, 42.3087.11–14 (III–IV AD), interestingly has
εἰ μὲν ... εἰ δέ (cf. note 6 above).
μέν that dates from before the second century. (The sample contains 10 pri-
vate letters of I AD.) We may suspect that in the second century greater
efforts were made to revive μέν. All this points to the conclusion that in
the later as in the earlier Koine μέν (…δέ) was used with considerable
artificiality, and that both user and reader (or hearer) felt it to be indicative
of formality and correctness.
The increased use of μέν without a following δέ (or its equivalent) is
in keeping with this.12 In Classical Greek this use of μέν was idiomatic
within certain limits;13 it was hardly common compared with the normal
use. It is plain that later writers who indulge in it have little concern for
the semantic value of μέν, but simply insert it, or a fixed phrase contain-
ing it, to add a touch of style.14
Turning now to the NT, we find first that μέν (…δέ), as has long been
remarked, is not very common.15 Moreover, the distribution is signifi-
cant. The books closest to the level of everyday language, Mark, John,
and Revelation, have only 14 examples out of a total of 180: Mark has 6
(including 16:19), John 8, and Revelation 0. In contrast Acts, for example,
has 48, and 1 Corinthians and Hebrews have 20 each.16 Μέν thus occurs
predominantly in the group of books in which it would be expected. But
even here it is noticeable that some books have it rarely or not at all, and
that Luke, whose concern for good style is well known, uses it little in his
Gospel (10 examples). In frequency and overall distribution, then, μέν in
the NT reflects the usage in the language generally.
So we come to Mark, where a more subtle reflection of current usage
is to be seen. The examples are as follows (16:19 aside):17
4:4 καὶ ἐγένετο ἐν τῷ σπείρειν ὃ μὲν ἔπεσεν παρὰ τὴν ὁδόν …
(5) καὶ ἄλλο ἔπεσεν ἐπὶ τὸ πετρῶδες … (7) καὶ ἄλλο ἔπεσεν εἰς
τὰς ἀκάνθας.
12
For further indications of its incidence see Mayser, Grammatik, II.3, 129–30; MHT,
Grammar, III, 332 (but Turner’s NT figures do not separate μέν without any correlative
from μέν correlated with καί, ἀλλά, or πλήν).
13
See KG, Grammatik, II.2, 271–2; Denniston, Greek Particles, 380–4.
14
According to Mr George Harris, a member of one of my classes in 1976, it is still the
practice in some Mod. Greek writing to insert a μέν, preferably early in the piece, ‘to
show that you know how to use particles.’ Cadbury (Style of Luke, 145–6) seems to me
rather to miss the point when he plays down μέν (οὖν) solit. as a test of style because
it is ‘of doubtful literary excellence.’ μέν is still a sign of a conscious attempt at ‘better’
Greek even if the user cannot handle it in the Classical manner.
15
See BDF, §447; MHT, Grammar, III, 331; BAGD, s.v. μέν.
16
Figures from Aland, Konkordanz, II ‘Wortstatistik.’ There is of course considerable
MS variation.
17
NT text is NA28 (2012). Variant readings in Mark depend on Legg, NTG: Marcum.
9:12 Ἠλίας μὲν ἐλθὼν πρῶτον ἀποκαθιστάνει πάντα· καὶ πῶς …;

(13) ἀλλὰ λέγω ὑμῖν ὅτι. …
12:5 κἀκεῖνον ἀπέκτειναν, καὶ πολλοὺς ἄλλους, οὓς μὲν δέροντες,
οὓς δὲ ἀποκτέννοντες.
14:21 ὅτι ὁ μὲν υἱὸς τοῦ ἀνθρώπου ὑπάγει καθὼς γέγραπται περὶ
αὐτοῦ, οὐαὶ δὲ τῷ ἀνθρώπῳ ἐκείνῳ.
14:38 τὸ μὲν πνεῦμα πρόθυμον ἡ δὲ σὰρξ ἀσθενής.
The noticeable point, as I have already indicated, is that all five exam-
ples are found in words of Jesus. We may also note that μέν is three times
correlated with δέ in the best manner (12:5; 14:21, 38), and is not without
correlative in the other two places.18 As to the text, there is uncertainty
only in 9:12, where a minority of MSS omit μέν.19 One less example,
however, would make no difference to my argument. Moreover, if μέν in
9:12 is secondary, it is of interest, and may be significant, that it has been
introduced into the words of Jesus.20
The conclusion to which my argument is leading will already be clear.
Μέν, we have seen, is a prestige feature associated with formal and edu-
cated Greek. In Mark it appears only in the speech of Jesus. The natu-
ral conclusion is that as regards this feature at least, Jesus speaks ‘better,’
more formal Greek than is spoken by those around him and used by Mark
in the narrative. It may be difficult to establish how this came about, but
the essential reason for it is readily seen: Jesus speaks in a manner appro-
priate to his status.
What of 16:19? It is interesting, to say the least, that the use of μέν here
differs markedly from that elsewhere in the Gospel. Not only is it used in
the combination μὲν οὖν, not found in the rest of Mark, but it occurs in
narrative.21 This reinforces the commonly-held view of the Longer Ending,
and highlights the special character of the use of μέν elsewhere in Mark.
18
The correlation with μέν in 4:4–7 is described by BAGD (s.v. μέν 2.d) as ‘an inexact
usage.’ In 9:12 the correlative is ἀλλά, a stronger equivalent of δέ (BDF, §447.6); this
would be quite acceptable Class. usage: see Denniston, Greek Particles, 5–6.
19
DLWΨ fam.1 (exc. 118) 28.565.892. If μέν is not original here, the parallel Matt 17:11
could be the source (so Torrey, Gospels, 57–8). But Matt’s change of ἀλλά to δέ (vs. 12)
is perhaps a clue that μέν was already present in Mark, otherwise ἀλλά could have been
retained: Matt is always careful to correlate μέν with δέ. In the other places in Mark μέν
is either unanimously attested (4:4; 12:5) or omitted only in some versions (14:21, 38).
20
A search through Legg’s apparatus brings to light five more examples of μέν in the MS
tradition of Mark: 1:8; 4:28; 8:28; 10:27, 39. Of these, three are in sayings of Jesus
(4:28; 10:27, 39).
21
The first point is noted (e.g.) by Farmer, Last Twelve Verses, 99–100, and Elliott, ‘Text
and Language,’ 261, but the second seems to have been missed. Μὲν οὖν, a favourite of
Luke in Acts, has a pretentious sound to it. This and a number of other features suggest
that the Longer Ending is on a different level of language from the rest of Mark.
It is natural to ask if Mark’s Gospel is alone in this use of μέν or whether

anything similar is to be found in the other Gospels. A brief answer must
suffice. In Matthew there is a clear link between μέν and the speech of
Jesus: out of 20 examples only two appear in other contexts (one is used
by John the Baptist, the other by the disciples). Luke and John, however,
do not show anything comparable.
II
At this point we may consider more fully the nature of what has been
found. It seems to me that the use of μέν in Mark is best understood as an
instance of a general linguistic phenomenon that is well known to all of
us, though we are often not conscious of it. In recent years it has received
increased attention from linguists within the relatively new discipline of
sociolinguistics.22
Language varies with the occasion. The way something is said (or written)
is adjusted according to who is speaking to whom, the circumstances, the
subject-matter, and so on (the ‘social context’). An important ingredient
in this adjustment is the degree of ‘formality.’ The different varieties (‘reg-
isters,’ ‘styles’) appropriate in different situations show a gradation from
informal to formal, from ‘slang to ceremony,’ in G. W. Turner’s phrase.23
This appears in the level of formality of the individual features used. Every
feature or ‘variable,’ in all spheres of the language (pronunciation, mor-
phology, syntax, and vocabulary), is related in some way to the scale of
formality.
Formality is a complex thing, difficult to define. It involves many ele-
ments, such as propriety, technicality, politeness, seriousness, and correct-
ness, each of which has as it were a scale of its own.24 The last is espe-
cially to be noticed. Correctness seems always to play a part in formality.
The use of a feature felt to be ‘correct’ brings a formal tone into any con-
text, while formal speech (or writing) to be formal must be on the whole
22
I cannot claim any expertise in this subject, or more than slight acquaintance with its
extensive literature. I have made use especially of: Bolinger, Aspects; Ellis and Ure,
‘Varieties,’ 251–9; Gregory and Carroll, Language and Situation; Trudgill, Sociolinguis-
tics; G. W. Turner, Stylistics. Stylistic levels of language in connexion with translation
of the Bible are usefully discussed in Nida, ‘Varieties of Language,’ 316–22.
23
Turner, Stylistics, 185. Terminology is still under discussion. For a full system of clas-
sification see Gregory, ‘Aspects.’ In Gregory’s system formality is comprised under
‘tenor of discourse.’
24
Cf. Trudgill, Sociolinguistics, 110.
‘correct.’ Whatever the difficulties of definition, in practice we recognise

at once the level of formality of any feature in our own language, and
can learn to do so in others, though hardly with the full appreciation of
the native speaker.
It would not be easy to describe in full the ‘social context’ of Jesus’s
utterances. There is no need to do so, however, in order to see the most
important element in the situation, the high status of Jesus as the Messiah
and Son of God, a being acting with authority and dignity and deserving
of respect. It would be consistent with this if features of a higher level of
formality occurred in his speech. Μέν is just such a feature. It is another
matter to decide whether such features could have originated with Jesus
himself or whether they were incorporated into his sayings by others who
felt them to be appropriate. Whatever the answer to that question, it seems
to me that the rationale behind the phenomenon remains the same.
We have, then, what appears to be a secure example of greater formality
in the speech of Jesus as recorded by Mark. The next step is obvious: are
there any other features of the same kind as μέν? I have found that there
are, and the clearest of them will be set out in the remainder of this paper.
Attention will be concentrated on Mark, in order to establish the case as
securely as possible for that Gospel. It is apparent, however, that the same
thing is at work in the other Gospels, and their evidence will be noticed
for some features. Before we proceed, my general approach needs to be
more fully explained.
III
Greek in the first century AD undoubtedly allowed for variation in

degree of formality. The kind of study undertaken here depends on our
being able to form an idea of the place of a given feature on the formal-
ity scale. What means are available for doing so? We have already seen
most of them employed in dealing with μέν, but it will be useful to gather
them together and state more explicitly what they are.
A pervading element in post-Classical Greek is the scale of ‘literariness,’
or gradation in quality of Greek. Something of the kind is found in most
languages, but Greek in this period is especially affected, for this was the
time when there first emerged the diglossia that has been a characteristic
of the language ever since.25 We are still a long way from a satisfactory
25
See, e.g., Browning, Greek, 49.
analysis and description of this scale, as of the post-Classical language

generally. Recent works have reopened the discussion, without clear result.26
Our terms remain ill-defined, and exact criteria for categorising the level
of a given text or feature are lacking. Nevertheless it is certain that vary-
ing degrees of literary refinement can be recognised in the texts. Lucian
writes ‘better,’ more literary Greek than Theon the schoolboy (POxy 1.119
[II–III AD]). Formality in post-Classical Greek is bound to be intimately
connected with this scale: formality and correctness go hand in hand.
Therefore what we can learn about the literary level of a feature will also
give an indication of the degree of formality it was felt to have.
The use of the papyri for this purpose requires care. As is now coming
to be recognised, they do not provide a homogeneous body of evidence for
popular Greek.27 The occurrence of a feature in the papyri does not auto-
matically make it colloquial or ‘vulgar.’ There are not only differences
of level between documents; we have also to remember that stereotyped
expressions, which often contain obsolete features, occur frequently, and
that even a poorly-educated writer may have his literary pretensions. Μέν
has provided an illustration of these points.
In dealing with the NT, it is important to recognise that the books dif-
fer in literary level. This is fundamental to the present enquiry. Though we
still await an adequate treatment of this subject,28 I take it as generally
accepted that on the scale from vernacular to literary Mark, John, and
Revelation belong to the lower levels, Revelation being lowest of the
three;29 Matthew, Luke-Acts, and the Epistles are decidedly higher, though
with differences between them, the high point being marked by Hebrews.30
None of the books reaches the level of the Atticists proper; but we should
26
Rydbeck, Fachprosa; Frösén, Prolegomena. For criticisms of Rydbeck’s proposed
‘Zwischenschichtsprosa’ see Frösén, 25–6, 50–1, 92–3, and reviews by Kilpatrick and
Thesleff. In my view the language is best regarded as a continuum. Any divisions we
make are artificial and do not indicate discrete varieties. Atticistic Greek, when it becomes
the fashion, is not isolated but shows mutual interference with lower levels. Cf. the Mod.
Greek situation: though Katharevousa and Demotic have separate identities, there is in
practice much interference between the two: see Browning, Greek, 112–3; Sotiropoulos,
‘Diglossia,’ 7.
27
Cf. esp. Rydbeck, Fachprosa, 26, 127, 176, 195.
28
N. Turner’s recent continuation of Moulton’s Grammar (MHT, Style), is disappointing in
this respect, thanks to his preoccupation with the question of Semitic influence. The older
discussions in MH, Grammar, II, 5–11 and BDF, §3 leave a good deal to be desired.
Moule’s remarks (Idiom Book, 2–3), though brief, are, as always, valuable.
29
Kilpatrick’s attempt to add the Pastorals to this group seems to me to fail (‘What John
Tells Us,’ 79–81). Some of Kilpatrick’s tests, especially παρά + acc., and σύν, are uncer-
tain, and there are indications contrary to his thesis, such as μέν (2 Tim 1:10; 2:20; 4:4),
οὖν (1 Tim 2:1, 8; 3:2; 5:14; 2 Tim 1:8; 2:1, 21), βούλομαι (1 Tim 2:8; 5:14; 6:9;
Titus 3:8), ἡμέτερος (2 Tim 4:15; Titus 3:14), and ὦ (1 Tim 6:11, 20).
30
Bauer in BAGD, xii.
be misled if we therefore placed Matthew, Luke-Acts, and the Epistles

low on the scale. It is not true, as is sometimes believed, that the NT was
written in popular Greek pure and simple.31 We are of course speaking of
the general character of the books. There is undoubtedly unevenness within
books (some of it is the subject of this paper). It should also be noticed that
level of language, which we are talking of here, is quite distinct from indi-
vidual style. Mark and John, for example, are similar in level of language
but distinguishable in style.32 Semitic influence is yet another matter. There
is no simple correlation between incidence of Semitic idiom and literary
level.
The books of the LXX likewise vary in level of language. We have
already seen in the case of μέν how this may help to establish the tone
of a feature. We must also keep in mind that there is a gap of two to three
hundred years between the LXX and the NT, during which changes are
certain to have taken place in the language, among them changes that affect
literary level. So we may find, for example, that a word that was in ordi-
nary use in the time of the LXX has become obsolete and belongs to more
elevated language in the NT. Similarly, when the LXX itself is quoted in
the NT its level of language will tend to differ from that of the context.
What has been said so far concerns distribution: by observing where a
feature tends to occur we may learn something about its literary level and
hence degree of formality. What other indications of level of formality
are available to us?
An appreciation of the long-term trends of the language, and their rate
of progress, is essential. This is because it seems to be a general axiom of
language that archaism and formality go together: the use of an obsolete
or obsolescent feature instead of its current equivalent suggest more for-
mal speech. (Contrast is an essential element here. An old feature is notice-
able, or ‘marked,’ as regards formality only if it contrasts with some other
feature which has replaced it in normal use; otherwise inherited features,
which in fact make up the bulk of any language, are unmarked.)33 This
principle is especially applicable to post-Classical Greek, where concern
with past forms of the language was a dominant influence. We can be
sure that when a writer of this period wished to be formal he or she would
tend to employ obsolete features.
We also depend on more elusive indications of formality. Some kinds of
expression ‘sound’ formal in any language. Such reliance on impression
31
Similarly N. Turner, ‘Literary Character,’ 107–14; Wifstrand, ‘Stylistic Problems,’ 174.
32
Kilpatrick, ‘Some Thoughts,’ 290. Cf. Frösén, Prolegomena, 134.
33
In the case of μέν, the contrast is between use and non-use of the particle. In speech
the function of μέν may well have been taken over by intonation.
is clearly hazardous, but we can define to some extent the features con-
cerned. Elaboration and periphrasis, in particular, seem likely always to sug-
gest formality. The preface to Luke’s Gospel, for example, clearly owes much
of its highly formal tone to these, as well as to other characteristics less easy
to name. That it was felt to be formal can hardly be doubted, even though
detailed demonstration is not easy. In the same way, most would agree that
poetry usually has a high degree of formality, despite the difficulty of proof.
Finally, there are the Atticist grammarians.34 Their remarks are a valu-
able guide to what was felt to be ‘correct’ in their time, and must always
be taken into account. But their usefulness for the NT and our present
purpose is limited. The reasons for this are worth noticing. (a) Their obser-
vations date from the second century, when the Atticising tendency was at
its height, and do not necessarily apply to the first century. (b) The object
of their endeavours is the highest literary level; approval is reserved for the
very ‘best,’ that is, Attic, Greek. Information about this does not always
tell us what standards applied at less exalted levels. (c) Their statements
are frequently crude and oversimplified; they need much interpretation in
the light of other evidence. (d) Their material is incomplete, both because
much of what they wrote has not survived and because some features were
not noticed. In sum, they give us only a fragment of information about
the complex picture of language varieties in Greek of the first century AD.
By combining these approaches we may reach a fair understanding of
the degree of formality of a given feature in the time of the NT. It must be
accepted, however, that our knowledge is and will remain imperfect. Only
the native speaker of a language can fully appreciate the variations in for-
mality it employs. When the language is no longer spoken, we depend on
making the best deductions we can from the available clues; certainty and
exact knowledge are unattainable.
IV
4.1. εὖ
It is generally accepted that in later Greek εὖ was gradually superseded

by καλῶς, a near synonym from early in Classical Greek.35 Some such
34
A useful outline is found in Kilpatrick, ‘Atticism,’ 128. [See now Lee, ‘Atticist Gram-
marians’ = Essay 19.]
35
BDF, §102.3; Chantraine, DELG, s.v. ἐύς; Mayser, Grammatik, I.3, 126; MM, s.v. εὖ;
Schmid, Review of LSJ Part 4, 705.
development was to be expected, given the shortness and irregularity of εὖ.

It is in keeping with the tendencies of the Koine that the replacement is
both ‘fuller-sounding’ and a regular formation related to a familiar word.36
Some specimen figures are enough to indicate the trend. Whereas καλῶς
is hardly known to Homer (καλῶς 1, εὖ 215), and is still well outnum-
bered in Herodotus (καλῶς 32, εὖ 125) and Aristophanes (καλῶς 80,
εὖ 130), in Plato the figures are more nearly even (καλῶς 528, εὖ 575),
and by the time of Polybius the older word is definitely on the wane
(καλῶς 87, εὖ 31). The papyri show the same development, as Mayser and
Moulton-Milligan observe, and as may be illustrated by PTebt I–IV, con-
taining texts of III BC to III AD: occurrences of καλῶς number 45, of
εὖ 9 (not counting restorations). In Vita Aesopi G, a popular Greek text
of I–II AD, καλῶς occurs 10 times, εὖ not at all.37
The LXX appears to go against the trend, with καλῶς 33 times and
εὖ 56 times (also εὖγε 14 times, often doubled). But the higher figure for
εὖ is largely due to the Pentateuch (καλῶς 3, εὖ 25, of which 16 are in
Deut); its early date is at least part of the explanation.
By the time of the NT it is clear that καλῶς has become the usual word
and εὖ is obsolete. καλῶς occurs 37 times and is preferred even in the
more literary books (Matthew twice, Luke 4 times, Acts 3, Epistles 18);
εὖ is not only rare (6 examples, including one of εὖγε) but there is also
something special about its use in every case.
One example, in Eph 6:3, is in a quotation from the LXX. Another is
found at the conclusion of the letter from the Jerusalem Council:
Acts 15:29 ἐξ ὧν διατηροῦντες ἑαυτοὺς εὖ πράξετε. Ἔρρωσθε.
This is unequivocally labelled ‘literary language’ by BDF (§102.3). Its

formulaic, and hence fossilised, character is also noticeable, whatever the
sense of πράξετε.38
There remain four examples in the Gospels. Mark has one, in words
spoken by Jesus:
Mark 14:7 πάντοτε γὰρ τοὺς πτωχοὺς ἔχετε μεθ᾽ ἑαυτῶν καὶ ὅταν
θέλητε δύνασθε αὐτοῖς εὖ ποιῆσαι.39
36
Cf. BDF, §126.1.
37
Data from my own index.
38
Cf. BAGD, s.v. εὖ; Haenchen, Acts, ad loc.
39
There is no variant to εὖ. In place of αὐτοῖς a number of witnesses (ΑΧΨΘΠΣΦ etc.)
read αὐτούς, thus making the Greek more ‘correct’ (cf. MHT, Grammar, III, 245). It is
hard not to suppose the identity of the speaker helped to produce this reading (assuming
it is secondary).
It is important that the more usual equivalent of this expression can be

cited often from the NT: καλῶς ποιεῖν occurs 11 times, including once
in Mark (7:37). There can be no doubt therefore that καλῶς could have
been used in 14:7.40
The other examples in the Gospels, found in Matt 25:21, 23 (εὖ) and
the parallel Luke 19:17 (εὖγε, v.l. εὖ), might be satisfactorily explained
as due to the more literary tendency of these books. Moreover they are
exceptional in that εὖ(γε) is here an interjection (‘well done!’).41 But it is
a remarkable fact that they too occur in words spoken by Jesus (Parable of
the Talents/Pounds).42
The conclusion seem inescapable. The location of εὖ in the speech of
Jesus in Mark has not come about by accident. The old-fashioned word,
which we may be sure had a formal, elevated tone, has been deliberately
chosen for its appropriateness to the speaker.
4.2. The optative
The post-Classical decline of the optative is well known and needs no

demonstration here.43 The position reached by the first century AD may be
summed up as follows. While not rare, the optative is no longer an integral
part of the verb system; except for stereotyped expressions such as καλῶς
ἂν ποιήσαις and μὴ γένοιτο, which occur at all levels, it is associated
with the literary language, and even there some of its uses are greatly cur-
tailed. Subsequently (from II AD onwards) it mcreases in frequency under
the influence of Atticism but always remains an artificial feature con-
sciously used.
The NT unmistakably reflects current usage.44 There are some 67 exam-
ples, a modest total, and of these 15 are the set phrase μὴ γένοιτο and
11 the form εἴη. Some uses are barely represented, and others absent alto-
gether. Most important of all is the distribution: all the examples except
40
A further equivalent is ἀγαθοποιῶ (10 times NT including v.l. Mark 3:4, but not in Mark
otherwise.
41
BDF (§102.3) imply that this use had a longer life. But note that καλῶς is used in a
similar way: see Mark 12:32; Rom 11:20; LSJ, s.v. καλός C. 6, and Dio Cass. 67.18.1
καλῶς Στέφανε, εὖγε Στέφανε.
42
The words are those of a speaker within the parable, not of Jesus in propria persona.
It is interesting, however, that in both forms of the parable the speaker represented
allegorically is Jesus.
43
See esp. MHT, Grammar, III, 118–33; BDF, §§65.2, 384–6; Moulton, Grammar, I, 194–
9; Robertson, Grammar, 936, 1408; Debrunner–Scherer, Geschichte, 123-8; Browning,
Greek, 37, 41–2; Mandilaras, Verb, §§603–58. [Evans, Verbal Syntax, 175–8.]
44
Data from NT reference works already cited.
one occur in Luke–Acts and the Epistles. Even here there are notable
restrictions: the literary Greek of Hebrews, for example, has it only once
(13:21 καταρτίσαι, in a prayer). Moreover, the link between this anti-
quated feature and its social setting is often plain to see. For example,
the rare optative in a conditional clause is used by Paul in his speech of
defence before governor Felix (Acts 24:19 εἴ τι ἔχοιεν); Turner aptly
comments: ‘the use of slightly antique language in the presence of Felix …
is understandable.’45 Similarly in Acts 26:29 we find a potential optative
‘where in the royal presence of Agrippa, Paul employs the stilted εὐξαίμην
ἄν.’46
In view of all this it is not surprising that Mark, John, and Revelation
make hardly any use of the optative. That there is even one example is
cause for comment. When that example is found in words spoken by Jesus
it is hard to doubt that the location is significant. In Mark in the Cursing
of the Fig Tree, we find:
Mark 11:14 μηκέτι εἰς τὸν αἰῶνα ἐκ σοῦ μηδεὶς καρπὸν φάγοι.
It is noteworthy that this example is in no way formulaic but has been
created for this context, and that it is one of only two instances of the opta-
tive in imprecations in the NT.47 On the other hand, the volitive use was
the most enduring of the uses of the optative and is the commonest in
the NT, so there need be no suspicion that it was beyond the range of
an author such as Mark.48 To complete the picture we need to notice that
equivalents were definitely available; here the imperative (μηδεὶς …
φαγέτω) is the probable alternative.49
There is, however, the question of the text in Mark 11:14. The majority
reading φάγοι has been generally accepted without question, but Kilpatrick
has argued that φάγῃ (DUWΨ fam.1 fam.13, etc., Or.) is original, on the
principle that readings that are arguably stylistic improvements are to be
regarded as secondary.50 Apart from the general objection to any rigid
45
MHT, Grammar, III, 126.
46
MHT, Grammar, III, 123. Similarly BDF, §385.1: ‘The few examples [of the potential
opt. in main clauses] are literary language befitting the occasion.’ This useful observation
finds no place in BDR, §385.
47
BDF, §384; MHT, Grammar, III, 122. The other is at Acts 8:20: τὸ ἀργύριόν σου εἴη
εἰς ἀπώλειαν.
48
As is certainly the case with Mark 12:2 ἵνα λάβοι ‫ *א‬238 579. This quite exceptional
final opt. (‘the hall-mark of a pretty Attic style,’ Moulton, Grammar, I, 197) is generally
agreed to be secondary. Other opts. noted from Legg’s apparatus are at Mark 12:19
(λάβοι 579); 9:34 (τίς μείζων: add. εἴη fam.13, etc.).
49
See, e.g., 10:9; 13:15, 16, and Acts 1:20 λαβέτω for LXX λάβοι. For this and other
equivalents see BDF, §384; MHT, Grammar, III, 120.
50
Kilpatrick, ‘Atticism,’ 135; ‘What John Tells Us,’ 86.
application of this principle, there seem to me to be a number of difficul-

ties that would need to be resolved before rejecting φάγοι. To what extent
is the reading φαγη (also φαγει 472 etc.) a product of confusion of the
sounds of οι, η, and ει?51 What is to indicate that the obsolescent optative
has not been altered in the direction of more everyday Greek? Finally, is
φάγῃ an exact equivalent of φάγοι, or does it alter the sense slightly, from
a wish for evil to something like ‘no one must eat,’ ‘don’t anybody eat’?52
This seems to me a definite possibility, and if correct would be decisive
against Kilpatrick’s proposal, since a clear motive for the change to φάγῃ
could then be seen, that is, to tone down the curse.53 The same motive no
doubt lies behind the alteration to μηκέτι ἐκ σοῦ καρπὸς γένηται in the
parallel Matt 21:19.54
In my view, then, φάγοι, with its slightly archaic and elevated tone, is
to be explained not as a correction of later scribes but as an original feature
in keeping with the status of the speaker, and of the same character as μέν
and εὖ.
4.3. ὦ
The history of the incidence of ὦ is clear enough.55 In the earliest Greek

it is rare, but it increases in frequency until in Classical Attic the vocative
with ὦ is commonplace, while the vocative without it is less usual and in
some way ‘marked.’ The details of Classical usage need not be gone into,
for with Koine Greek comes a quite new picture. The bare vocative is now
the norm, and the insertion of ὦ attracts notice. Very few examples of ὦ
can be found in the papyri,56 and this evidence is supported, for example,
51
See Gignac, Grammar, 1:265–7. Mussies (Apocalypse, 246) notes Ruth 1:9 εὕροιτε B,
εὕρητε A as an instance.
52
So the other NT exx. of 3rd pers. subj. in prohibitions suggest: 1 Cor 16:11; 2 Cor. 11:16;
2 Thess 2:3 (add Mark 13:36 μὴ εὕρῃ?). Turner (MHT, Grammar, III, 94) describes
these as ‘hortatory.’ Cf. Mussies, Apocalypse, 245–6, on the opposition between opt.
and subj.
53
It is evident from the commentaries that this has long been felt a difficulty: see, e.g.,
Lane, Mark, 402; Taylor, Mark, 460.
54
The reading μηκέτι γένηται is that of the majority of MSS and read by NA28. Matthew’s
γένηται may have helped to produce φάγῃ in Mark.
55
Schwyzer, Grammatik, II, 60–1; Humbert, Syntaxe, §§499–502; Scott, ‘Vocative in
Homer and Hesiod’; Vocative in Aeschylus and Sophocles’; ‘Additional Notes’;
Moulton, Grammar, I, 71, and NT reference works cited below.
56
Mayser, Grammatik, II.1, 55 notes one, ‘in dem feierlich gehobenen Artemisiafluch,’
UPZ 1.1 (IV BC) ὦ δέσποτ᾽ Ὀσερᾶπι. MM found BGU 2.665.III.8 (I AD) ἐρρῶσθαί σε
εὔχομαι ὦ πάτερ (cf. πάτερ alone in II.11, 20). To these I can add only PFouad 25v.I.2
(II AD, document about a lawsuit) ὦ πονηρότατον πάντων τῶν ἀνθρώπων μειράκιον,
by Vit. Aes. G, where vocatives are extremely frequent but ὦ occurs only
three times, in fables inserted into the narrative.57 Epictetus similarly shows
restricted use of ὦ.58
On the other hand, literary authors make frequent use of it, as a glance
at Lucian or Plutarch and the figures for Philo (148) and Achilles Tatius
(ὦ c.50, ὤ 17) are sufficient to confirm. In the same way, many of the LXX
examples are found in the more literary books (4 Maccabees alone accounts
for 36 out of the LXX total of 80).59 Imitation of the Classical language is
obviously at work here. In sum, ὦ in late Greek shows the signs of being
a prestige feature consciously used.
In the NT the vocative without ὦ is usual. There is general agreement
that ὦ is exceptional and calls for explanation whenever it occurs. To put
it another way, ὦ is ‘marked’ and we need to try and appreciate what it
conveys that the bare vocative does not. According to current opinion,60
ὦ is to be explained in three ways: (a) In Acts 1:1; 18:14, and 27:21 it is
a stylistic refinement in accordance with Attic usage. (b) In Rom 11:33 it
is exclamative ὦ (= ὤ), also in accordance with Attic usage. (This group
is considerably expanded by BDF, following Katz, but none have followed
suit.) (c) In the majority of instances it is said to express emotion, namely
Matt 15:28; 17:17; Mark 9:19; Luke 9:41; 24:25; Acts 13:10; Rom 2:1,
3; 9:20; Gal 3:1, 1; Tim 6:20; Jas 2:20, to which we should presumably
add 1 Tim 6:11 (this example seems to have been accidentally overlooked
by all except Zerwick). We may note that Turner is compelled to observe
that ‘there is no great emotion’ in some of these instances (Rom 2:1, 3;
9:20; Jas 2:20).
This analysis is difficult to accept. Explanations (a) and (b) are clearly
right, but (c) raises doubts. Emotion may well be present in many of
these contexts, but it is puzzling that ὦ is not used in the many other cases
POxy 6.937.9 (III AD) ὦ κυρία μου ἀδελφή, and 31.2603.17 (IV AD, elegant letter
of commendation) ὦ φίλτατε. Lest it be thought that vocatives (without ὦ) hardly ever
occur in the papyri, I note that there are over 20 in POxy vol. 1 alone; see also exx. in
Mayser, ibid.
57
Vit. Aes. G 129.14, 131.12, 140.8. The usage of the fables is consistent with this: in
Fab. Aes. 1–60 (ed. Halm) I find 16 vocs. with ὦ, 2 without.
58
Total occurrences c.28. ἄνθρωπε illustrates the point well: it occurs 63 times, but only
once with ὦ (Schenkl’s index). I have not seen Johannessohn, Präpositionen; Turner
(MHT, Grammar, III, 33) reports J.’s finding that Epict. has ὦ with proper names ‘nur
da, wo es sich um Grössen der Vergangenheit handelt.’
59
See Walters, Text, 229, 231. Walters comments (232) that the author of 4 Macc is
‘anxious to equal the best models.’ Breitenstein, Beobachtungen, 106–9, makes it clear
that ὦ in 4 Macc is a stylistic device, inserted or omitted at will.
60
BDF, §146; BDR, §146; MHT, Grammar, III, 33; Zerwick, Biblical Greek, §35; BAGD,
s.v.
where emotion is equally discernible.61 The fact is, the vocative without
ὦ seems to be the norm, whether emotion is expressed or not. There is
also the inconsistency between (a) and (c): we are asked to accept that
ὦ is a literary refinement in some places but not in others, and some-
times expresses emotion and sometimes does not. Further, it is hard to
see any difference at all between some of the examples in (c) (especially
Rom 2:1, 3; 1 Tim 6:11; Jas 2:20) and those in (a).
I suggest that much better sense can be made of the NT evidence if,
in accordance with contemporary usage, and taking up the unmistakable
clue afforded by examples like Acts 1:1, we regard ὦ as primarily a fea-
ture of higher style. Its chief effect in all places is to give a formal and
elevated tone.62 In emotional contexts this may well enhance the expres-
sion of emotion, but emotion is not what ὦ itself expresses.
On this basis the NT distribution becomes fully intelligible. The major-
ity of examples occur in the literary Greek of Acts and the Epistles, whose
writers use it as an occasional embellishment; the rare occurrences in
the Gospels share a significant characteristic: they are all found in words
spoken by Jesus (Mark 9:19, Matt 17:17, Luke 9:41 ὦ γενεὰ ἄπιστος,
Matt 15:28 ὦ γύναι, μεγάλη σου ἡ πίστις, Luke 24:25 ὦ ἀνόητοι καὶ
βραδεῖς τῇ καρδίᾳ). Why it should occur in these particular places and
nowhere else in the words of Jesus may be hard to discover, but this does
not alter the point that when it does occur its appropriateness to the speaker
can be felt.
What we have in Mark, then, is a single example of this elevated fea-
ture, conspicuous against the background of Mark’s generally low-level
Greek, but located where its effect is most telling.63
I have left aside so far the question of the differentiation between ὦ and
ὤ, which has been the subject of a detailed examination by Walters.64 His
results for the LXX seem clear, but are less persuasive for the NT. Here
he finds exclamative ὤ in more than half the instances; in the Gospels
he would read ὢ ἀνόητοι and ὢ γενεὰ ἄπιστος, leaving the vocative ὦ
61
E.g., Mark 8:33 || Matt 16:23 σατανᾶ. Mark 10:47, 48 || Matt 20:30, 31; Luke 18:38, 39
υἱὲ Δαυίδ (Ἰησοῦ). Matt 3:7 || Luke 3:7 γεννήματα ἐχιδνῶν. Matt 26:39 || Luke 22:42
πάτερ. Luke 22:33 κύριε. 34 Πέτρε, John 13:7, 9 κύριε, 19:26 γύναι, 21:17 κύριε.
62
I have here the support of Jannaris (Grammar, 327): ‘its presence in postchristian
compositions usually adds a certain solemnity or emphasis.’ The idea that ὦ in the NT
expresses emotion seems to originate with Blass (Grammar, 1st ed. 1896, p. 85). No
supporting evidence from outside the NT has yet been offered.
63
The view that Jesus is here presented as a deity temporarily on earth complements my
argument very well: see esp. Tabachovitz, Die Septuaginta, 113–5, where it is argued
that several LXX reminiscences, including ὦ, aid the portrayal.
64
Walters, Text, 228–36.
only in ὦ γύναι. One may doubt the reality of the distinction here; if it is
real it is a fine one and does not as far as I can see affect the conclusion
I have drawn about the tone of this feature in the NT.
4.4. οὐ μή
The use of οὐ μή with the subjunctive or future for a prohibition or

strong denial has already been much discussed, and most of the data have
been collected. I simply propose a reinterpretation, linking this feature
with the phenomenon we are considering.
It will be useful to summarise my argument at the outset. Οὐ μή is com-
mon in Classical Greek, but rather less so in post-Classical. In the LXX,
however, it is noticeably frequent. The NT follows suit, but with the impor-
tant restriction that apart from LXX quotations and reminiscences most
of the examples are in sayings of Jesus. My conclusion is that the NT
shows LXX influence in the use of this feature, and that it conveyed not
only emphasis but more important a solemn, biblical tone, especially suited
to prophetic utterances and the speech of Jesus.
These points may now be taken up in more detail. The origin of the
idiom has been debated, but has no importance here, nor is it necessary
to enter into discussion of the differing constructions (aor., pres. subj.;
fut. indic.). The wide spread of the idiom in Classical Attic is clear.65
Examples are found both in prose (Thuc., Plato, etc.) and poetry (Aesch.,
Soph., Eur.), in elevated language (Demosth., the dramatists) just as much
as popular (Aristoph., Xen.). No particular trend is evident, except that its
rarity in Aristotle and Menander suggests there was some decline in late
Classical.66
The post-Classical evidence is not easy to assess, but overall there does
seem to be a drop in frequency, as others have observed. Turner found no
instances among 1370 negatives in his sample of six literary authors,67 and
only sporadic examples are observable elsewhere in literature (e.g. one in
Ach. Tat., three in Epict.).68 In the papyri, however, it is not exactly rare,
as is often said. Besides the 17 Ptolemaic occurrences listed by Mayser,69
65
LSJ, s.v. οὐ μή; KG, Grammatik, II.2, 221–3; Schwyzer, Grammatik, II, 317.
66
Aristotle only Top. 108a.28; Mir. 157b.28; Menander barely once, Fr. 893 (Körte)
ap. Photius: οὐ μηκέτι· ἀντὶ τοῦ οὐκ ἔτι. Μένανδρος.
67
Aratus, Callimachus, Lycophron, Aelian, Philostratus, and Alciphron: MHT, Grammar,
III, 96 n. 2.
68
Ach. Tat. 2.19.1; Epict. 1.25.27; 3.7.8, 22.33. Jacobitz records one ex. in Lucian:
Pisc. 18.
69
Mayser, Grammatik, II.1, 233; II.2, 564.
I have been able to find more than 20 examples.70 A further five in Vit.
Aes. G should also be noticed.71 All this implies that while it was not com-
mon it was still very much alive, and more at home in popular Greek than
literary.72
Turning to the LXX, one is struck by the great frequency of this feature.
More than 800 occurrences are recorded by Hatch-Redpath.73 These are
spread throughout, though with notable concentrations in the Pentateuch,
Job, Psalms, Sirach, and the prophetical books. The exact figures for the
Pentateuch and Isaiah, for example, are 82 and 103 respectively (Rahlfs’s
text). Why the translators made such frequent use of it we can only guess.
Hebraism can be ruled out at once: there is no equivalent Hebrew (or Ara-
maic) idiom. Another approach may be more fruitful. In the Pentateuch,
from the first example onwards, an association with divine pronouncements
is noticeable.74 Thus at its first appearence in Genesis we find:
Gen 3:1 Τί ὅτι εἶπεν ὁ θεός Οὐ μὴ φάγητε ἀπὸ παντὸς ξύλου τοῦ
ἐν τῷ παραδείσῳ;
and typical of later examples is:

Exod 20:5 οὐ προσκυνήσεις αὐτοῖς οὐδὲ μὴ λατρεύσῃς αὐτοῖς.
The connexion between this and the generous use of the idiom in
the prophetical books is hard to miss. While this is not the full story,75
it suggests that the Pentateuch translators, and then others following their
lead, found in the οὐ μή construction a powerful and somewhat for-
mal expression suited to biblical style generally and divine utterances in
particular.
70
Already noted in MM and BAGD: POxy 1.119.4, 6, 14, 15 (II–III AD); 6.903.16
(IV AD); 12.1483.10 (II–III AD); PGM 5.279 (= PLond 1.46); 13.321 (= PLeidW);
Wessely, Specimina, 26.4 (I AD). Add: PHaun 10.28 (III BC); POxy 42.3070.6 (I AD),
and, with some dozen exx., SB 6.9641 recto I.17–, II.6– (II AD), a series of prom-
ises made on entering one of the Egyptian priesthoods. From the inscriptions there is
SIG 2.1042.16 (II–III AD) [= IG II 1366.16 (I AD); also IAbydosMemn B.1 (n.d.);
IPessinous 6.4 (II BC)].
71
Vit. Aes. G 28.5; 50.10; 61.11; 67.10; 80.11.
72
Suzanne MacAlister has drawn my attention to a much later example in Malalas,
Chron. 7.172.6 (VI AD) οὐ μὴ σταθῇ ἡ πόλις σου Ῥώμη. It seems significant that the
context is a Pythian prophecy. [οὐ μή ‘more at home in popular Greek’: I am doubtful
now of this statement.]
73
HR, s.vv. οὐ μή, οὐδὲ (οὐ) μή, οὐκέτι (οὐ) μή, οὔτε (οὐ) μή. This figure includes
instances where a v.l. is involved.
74
In Gen 10 out of 14, in Exod 12 out of 12 exx. are in divine speech.
75
Among other questions to be considered is the extent to which οὐ μή is used for variety
(see, e.g., Exod 20:5 above). [Fuller discussion in Lee, Greek of the Pentateuch, 56.]
Whatever the reason for the frequency of οὐ μή in the LXX, it must

have been a very noticeable characteristic of the version for those who
heard and read it. In view of this and its unexpected frequency in the NT
as compared with contemporary Greek, we may reasonably conclude that
it came into NT usage largely from the LXX.76 It follows that its tone
is likely to have been formal and dignified, as is true generally of LXX
features in the NT.77 Moreover in this case there is the association with
prophecy and the like to add further to its formal character.
We are now in a position to appreciate NT usage more fully. Its most
striking aspect has long been observed: the numerous examples in the
Gospels are almost entirely confined to sayings of Jesus.78 The conclu-
sion I would draw is that οὐ μή is not distributed in this way by accident,
but because it lends appropriate solemnity and biblical flavour to Jesus’s
words, recalling in particular the LXX use of the idiom in prophecy. It is
worth adding that sayings containing οὐ μή are frequently introduced by
ἀμὴν λέγω ὑμῖν (and variations), which so obviously ‘adds emphasis and
solemnity to that which follows.’79
The examples outside sayings need not be examined in detail here.
In brief they are as follows.80 In the Gospels there are 7 (Mark 14:31 ||
Matt 26:35; 16:22; Luke 1:15; John 13:8; 20:25; 11:56), of which only
the last (in a question) is in any way inconsistent with the view that οὐ μή
is formal in tone. In Acts, the Epistles, and Revelation the great majority
are in LXX quotations or reminiscences. In addition there is an example
that is possibly part of an agraphon in 1 Thess 4:15.81
This last example leads us to a point of considerable interest. Moulton
noted, besides 1 Thess 4:15 and Mark 16:18, the appearance of this fea-
ture twice in the Oxyrhynchus Logia (POxy 1.1.2; 4.654.1). To these can
be added occurrences in Gos. Thom. A 3:2 and B 2:2 (ed. Tischendorf):
76
So Lagrange, Saint Marc, XCIX; Tabachovitz, Die Septuaginta, 106 n. 3; MHT, Style,
33, 69; more tentatively Moule, Idiom Book, 22.
77
Cf. BDF, §4.2: ‘the language of the LXX appeared to be very appropriate to a solemn
and dignified style.’ This aspect of Lucan style is often remarked on by Haenchen
(Acts, e.g., 80, 159, 254, 351), and cf. Sparks, ‘Semitisms of Acts,’ 27.
78
Moulton, Grammar, I, 191; BDF, §365. Robertson (Grammar, 1405) gives a useful table
of occurrences, classified according to speaker. His figures are: total 66 in the Gospels,
58 in Sayings. But John 8:52 (quoted words of Jesus) should be added to the latter.
79
Taylor, Mark, on 3:28; similarly BAGD, s.v. ἀμήν 2. In Mark, 5 out of 7 sayings with
οὐ μή begin with this formula. Ἀμήν is of course another example of the phenomenon
under consideration: it is used only by Jesus, and its solemn character is unquestioned.
80
Cf. Moulton, Grammar, I, 191–2.
81
So Moulton, ibid.; cf. Frame, Thessalonians, 171–2. But according to Jeremias (Unknown
Sayings, 80 n. 3) it is ‘now generally agreed’ that the agraphon consists of 16–17a, and
15 is an ‘introductory summary’ by Paul.
both times Jesus is the speaker, and the idiom occurs only here in these
texts. From this evidence it appears that οὐ μή became at some stage a
conventional feature for sayings of Jesus. In this connexion it is noteworthy
that Luke in his rephrasing of the saying in Mark 10:29–30 introduces
οὐ μή (Mark 10:30 ἐὰν μὴ λάβῃ ἑκατονταπλασίονα || Luke 18:30 ὃς
οὐχὶ μὴ [ἀπο]λάβῃ πολλαπλασίονα).82
Finally, further mention needs to be made of Moulton’s discussion,
which has influenced, and I think confused, many subsequent treatments.83
Moulton reached the conclusion that the frequency of οὐ μή in OT quota-
tions and sayings of Jesus may be put down to ‘a feeling that inspired
language was fitly rendered by words of a peculiarly decisive tone.’ This
conclusion is not far from the mark, but on the way to it Moulton argued
that excessive use of οὐ μή is linked with translation from Semitic origi-
nals, which includes sayings of Jesus; that it has weaker emphasis in ‘trans-
lation Greek’ than elsewhere; and (less clearly) that this over-use of unem-
phatic οὐ μή is connected with ‘elementary Greek culture.’
Moulton here seems to me to be following a quite false trail. Without
a corresponding Hebrew or Aramaic idiom, there can be no causal con-
nexion between the use of οὐ μή and translation from Semitic originals.
Translation therefore does not explain the concentration of οὐ μή in say-
ings any more than it explains its frequency in the LXX. In any case the
sayings cannot be lumped together with the LXX as if the translation pro-
cess was the same in both. As to the degree of emphasis, the question seems
largely irrelevant. Examination fails to confirm a correlation between ‘trans-
lation Greek,’ that is, the LXX and the Sayings, and unemphatic οὐ μή.
There is little emphasis, e.g., in Epict. 1.25.27 ἄνθρωπε, καὶ μὴ θεώρει καὶ
οὐ μὴ θλιβῇς, ‘do not become a spectator and you will not be crowded’ (cf.
Luke 6:37); POxy 42.3070.6 οὐκέτι οὐ μὴ δείρομέν σε ἐὰν δώσῃς ἡμεῖν τὸ
ποιγίσαι, ‘we won’t hit you any more if …’; PCairZen 2.225.6; Vit. Aes. G 61.11.
This is true even of some Classical examples, such as Aristoph., Pax 1226, Nub. 296.
On the other hand some ‘translation Greek’ examples seem quite emphatic,
e.g., Mark 10:15 ἀμὴν λέγω ὑμῖν, ὃς ἂν μὴ δέξηται τὴν βασιλείαν τοῦ θεοῦ
ὡς παιδίον, οὐ μὴ εἰσέλθῃ εἰς αὐτήν, Gen 21:16 οὐ μὴ ἴδω τὸν θάνατον τοῦ
παιδίου μου, Matt 5:26 ἀμὴν λέγω σοι, οὐ μὴ ἐξέλθῃς ἐκεῖθεν, ἕως ἂν ἀπο-
δῷς. … Furthermore the same expressions can be cited from both kinds of Greek:
οὐ μὴ φάγω is found in Gen 24:33; Tob 7:12S; 1 Cor 8:13; POxy 1.119.14
82
[Another instance: Mark 4:22 οὐδὲ ἐγένετο ἀπόκρυφον ἀλλ᾽ ἵνα ἔλθῃ εἰς φανερόν ||
Luke 8:17 οὐδὲ ἀπόκρυφον ὃ οὐ μὴ γνωσθῇ καὶ εἰς φανερὸν ἔλθῃ.]
83
Moulton, Grammar, I, 39, 187–92; MM, s.v. οὐ. Cf., e.g., Lagrange, Saint Marc, XCIX;
Taylor, Mark, on 9:1; Zerwick, Biblical Greek, §444. [Also Decker, ‘Markan Idiolect,’
51 n. 31.]
and SB 6.9641 recto 1.17; οὐδ᾽ οὐ μὴ γένηται is found in Matt 24:21 and
ChrWilck 122.4. Clearly it would be fanciful to detect any difference of emphasis
or ‘decisiveness’ between the biblical and non-biblical occurrences.
So far as it can be judged at all, the emphasis varies, both in ‘translation

Greek’ and elsewhere. If οὐ μή shows loss of emphasis, it is more likely
to be a case of the universal tendency for strong expressions to lose their
force than a phenomenon peculiar to translators of Hebrew and Aramaic,
who were no more lacking in ‘Greek culture’ than many others who used
Koine Greek.
4.5. Other probable features

Other features besides those examined above have been found. A full
treatment is not possible here, however, and I shall simply indicate them
briefly. While not all can be regarded as certain, I have examined them
far enough to be confident that they are probable.
(1) ἄν with indicative in the apodosis of an ‘unreal’ condition is found in
Mark only in 13:20 (and very doubtfully 7:11),84 in contrast with indicative
alone (Mark 9:42; 14:5, 21, and sporadically elsewhere in the NT).
(2) ἔξωθεν in its original sense of ‘from outside,’ with the force of -θεν
felt, occurs only in Mark 7:15, 18 in the NT; elsewhere ἔξωθεν is equiv-
alent to ἔξω.
(3) ἔσωθεν similarly has the obsolescent sense of ‘from inside’ only in
Mark 7:21, 23, Luke 11:7.
(4) θύραι of one door is rare in the NT; in Mark it is found only in 13:29
ἐπὶ θύραις. The appearance of equivalents using the singular (Mark πρὸς
τὴν θύραν 1:33; 2:2; 11:4; Acts 3:2; ἐπὶ τὴν θύραν Rev 3:20; πρὸς
τῇ θύρᾳ John 18:16) makes it likely the old set phrase was obsolescent
and formal-sounding.
(5) ὁράω pres. and impf. is obsolescent in the NT, being replaced by
βλέπω (and θεωρῶ). Out of 11 occurrences in the Gospels (not including
John 6:2 ἑώρων l. ἐθεώρουν), 9 are in words of Jesus (Mark 1:44; 8:15;
Matt 8:4; 9:30; 16:6; 18:10; 24:6; Luke 12:15; 16:23.).
(6) οὐαί, like οὐ μή, is a feature with OT associations and a solemn bibli-
cal flavour: all examples in the Gospels are in sayings.
(7) οὖν is greatly affected by variation in the MSS, but it is hardly acci-
dental that the five Marcan examples accepted in NA28 occur as follows:
84
It is to be understood that the noted feature occurs in words of Jesus.
in words of Jesus (10:9; 13:35), in supposed words of Jesus (11:31 [οὖν]),

in words of Pilate (15:12), and in narrative only in the Longer Ending
(16:19). A tendency towards localisation in sayings is also evident in Mat-
thew (though not in Luke or John).
(8) The substantival neuter participle, except for stock phrases, shows signs
of being a feature of more elevated style. The majority of NT examples are
in the more literary books (63 in Luke-Acts, 29 in Matt, 58 in the Epistles).
In Mark there are only seven, as follows: in narrative τὸ γεγονός (5:14) and
τὰ μέλλοντα (10:32); in words of Jesus τὰ κοινοῦντα (7:15), τὰ ἐκπο-
ρευόμενα (7:15), τὸ εἰσπορευόμενον (7:18), τὸ ἐκπορευόμενον (7:20),
and ἐκ τοῦ περισσεύοντος (12:44). The two in narrative are set phrases
whereas the others are clearly more special to their contexts.
This paper has not attempted to cover more than a part of a large sub-
ject. Attention has been concentrated on the most telling examples and a
full explanation of the methods used. Nevertheless the evidence examined
gives good grounds for thinking that there is some tendency in Mark’s
Gospel, and probably elsewhere, for features having a formal, dignified
tone to be localised in the words of Jesus. It also suggests that further
investigation along the same lines would be fruitful, and lead us ultimately
to a more complete appreciation of NT usage.
It must be stressed that this phenomenon is no more than a tendency.
The features concerned are slight and occur only sporadically. There is no
suggestion that a more elevated, formal tone is consistently maintained in
the sayings of Jesus or that a sharp difference between them and the sur-
rounding contexts can be detected. The phenomenon is in fact such as to
escape the notice of all but the native speaker. But this is just what might
be expected: as modern sociolinguistic study has amply demonstrated, lin-
guistic variables are often unobtrusive, and speakers show remarkable sen-
sitivity in using and recognising them.
It may be asked what proportion of the Gospels is made up of sayings.
Clearly, the greater the extent of the sayings, the less meaningful any con-
trast between their usage and that of the remainder. The answer is reassur-
ing. A calculation based on lines of Nestle-Aland’s text shows that sayings
comprise 37% of Mark, 42% of John, 48% of Luke, and 57% of Matthew.85
85
Line-counts were: Mark 543 out of 1482; John 812 out of 1943; Luke 1228 out of 2562;
Matt 1361 out of 2395. I do not know if a similar calculation has been made before.
An important parallel to the suggested phenomenon is to be noticed.

In the LXX Pentateuch, as I have argued in detail elsewhere,86 there is
at least one case of the use of more elevated vocabulary when God is
the speaker: the everyday word βρέχω, usual in narrative, is replaced by
the antiquated, more dignified equivalent ὕω in the speech of God in
Exod 9:18; 16:4.87 Parallels from further afield are no doubt abundant.
To take only one example, it has been observed that in Virgil’s Aeneid
the speech of the gods is coloured by archaisms and other features add-
ing impressiveness.88
My argument finds further support in a related phenomenon observ-
able in the NT, the tendency for respectful, dignified terms to be used in
describing Jesus and his actions. Some examples are as follows: (a) In
the description of the death of Jesus in Mark 15:37, 39; Luke 23:46, dig-
nity and solemnity are imparted by the use of ἐκπνέω, a rare euphemism
with poetic antecedents.89 (b) The obsolescent word βοῶ, largely replaced
by κράζω, appears in Mark, apart from a LXX quotation, only when
Jesus is being described (15:34). (c) Δακρύω, a milder and more digni-
fied term than the usual κλαίω, is confined to a single occurrence in the
NT, in John 11:35 ἐδάκρυσεν ὁ Ἰησοῦς.90 (d) A more complex example,
involving also the characterization of the speaker, is seen in Luke 23:41
οὗτος οὐδὲν ἄτοπον ἔπραξεν, where ἄτοπον is a carefully chosen word,
avoiding the possible offence of the more direct πονηρόν or κακόν (con-
trast 23:22 where Pilate says τί γὰρ κακὸν ἐποίησεν οὗτος;).91 In a simi-
lar vein are the numerous changes, especially omissions, made by both
Matthew and Luke because of ‘an increasing feeling of reverence for
the person of Christ’ (Allen).92
Both the phenomenon just mentioned and the one described in this
paper, though extensive in themselves, are only facets of the whole sub-
ject of adjustment of language to situation and speaker in the NT. Many
Black (Aramaic Approach, 52) gives some estimates evidently based on impression
alone.
86
Lee, Lexical Study, 123–4.
87
It is not clear how extensive this phenomenon is. Other probable features known to me
are: ἡμέτερος (Gen 1:26), ὑμέτερος (Gen 9:5), μέσαι νύκτες (Exod 11:4), and cf. οὐ
μή above. [The first two are developed in Lee, Greek of the Pentateuch, 49–53.]
88
Palmer, Latin, 112–3.
89
Simpson, Words, 14. Cf. Matt 27:50 ἀφῆκεν τὸ πνεῦμα, John 19:30 παρέδωκεν τὸ
πνεῦμα.
90
I owe exx. (b) and (c) to Shipp, ‘Distribution,’ 134–5.
91
A useful discussion of the semantic history of ἄτοπος is found in Arnott, ‘The Con-
frontation,’ 119–22.
92
Allen, Matthew, xxxi–iii; on Luke see Cadbury, Style of Luke, 90–6.
observations on this subject have been made, but in a piecemeal fash-

ion; more systematic study would be likely to find much of interest, and
enhance our appreciation of NT language and the impact it had in its
own time.
The purpose of this paper has been to observe a phenomenon, not to
advance any theory about how it came into being. Nevertheless the obser-
vation, if correct, clearly has implications for the history of the sayings
of Jesus. I must leave it to others with the necessary expertise to explore
the subject properly; I mention only the most immediate consequences as
I see them. None of the features examined has any counterpart in Aramaic
(or Hebrew). This means that they cannot go back to an original Aramaic
form of the sayings. It follows that, unless the sayings were originally
spoken in Greek, these features do not derive from Jesus himself but were
introduced into the sayings by others at some stage in their history. Who
introduced them is a matter for conjecture. It does not seem likely that
Mark himself was responsible, at any rate for all of them; if he was we
might expect to see more use of such features in his own writing. (It would
be wrong, however, to suppose that he was not fully aware of them). It is
more likely that Mark simply used what came to him, and that the sayings
had already undergone a certain amount of development of this kind. That
the sayings were not felt to be unalterable is demonstrated in the NT itself,
where Matthew and more especially Luke are seen to make changes, many
of them stylistic, in sayings just as readily as elsewhere. Finally there is the
question of whether anything in the nature of a convention operated to pro-
duce our phenomenon. While this could apply to some features (especially
οὐ μή), it seems more likely that most of them were introduced indepen-
dently and haphazardly, out of a natural desire for the dignity of the speaker
to be reflected in his language.
Postscript
Studies that draw on or make significant mention of this paper are: Aitken,
‘Characterization of Speech’; Bentein, ‘Particle-usage,’ 724 and n. 12; idem,
‘Dimensions of Social Meaning,’ 157 and n. 68. The subject of οὐ μή in the LXX
is further developed in Lee, Greek of the Pentateuch, 54–8.
6
A NON-ARAMAISM IN LUKE 6:7
1991
Abstract
The use of εὑρίσκω in Luke 6:7 with the meaning ‘find a way (to),’ ‘be able
(to)’ has been suspected of being an Aramaism. While the Aramaic evidence has
been established, the evidence for this use in Greek has been ignored or not fol-
lowed up, both by advocates of the proposal and in the NT lexicons. Evidence
is here presented to show that it is a thoroughly Greek use in the Koine period,
with roots in earlier Greek. There is no reason to see it as due to the influence of
Aramaic on Luke’s language.
In Luke 6:7 εὑρίσκω is found in a context requiring the sense of ‘find

a way (to),’ ‘be able (to)’:
Luke 6:7 παρετηροῦντο δὲ αὐτὸν οἱ γραμματεῖς καὶ οἱ Φαρισαῖοι
εἰ ἐν τῷ σαββάτῳ θεραπεύει, ἵνα εὕρωσιν κατηγορεῖν αὐτοῦ.1
This use, also found in the text of D at 11:54 and 13:24, has been the
subject of remark since the time of Wellhausen, who suggested that an
Aramaism was involved.2 In more recent discussion, notably by Black,
Fitzmyer, and Wilcox, attention has been focused on establishing the nec-
essary evidence on the Aramaic side, namely the existence of a verb with
the two senses of ‘find’ and ‘be able.’3 It now seems beyond doubt that
in Palestinian Aramaic of the NT period ‫ שׁכח‬was in use in just this way.
With this settled, Fitzmyer and Wilcox are happy to conclude that the use
of εὑρίσκω in the sense of ‘be able’ reflects Aramaic interference.
1
This reading continues to hold the field in NA28, with the support of P4vid ‫ *א‬B Θ f1 1241 sa.
The majority reading has κατηγορίαν for κατηγορεῖν (often combined with κατ᾽ before
αὐτοῦ).
2
Wellhausen, Einleitung, 17.
3
Black, Aramaic Approach, 133–4; Fitzmyer, ‘Aramaic Background,’ 12–3; Wilcox,
‘Semitisms,’ 1011–2. Fitzmyer’s material reappears without substantial change in his
commentary, Luke I–IX, at 117, 610–1.
But what of Greek usage? Is it safe to assume that in the Koine period
εὑρίσκω was used only in the sense of ‘find,’ and that ‘be able’ was not
part of its semantic range? For Black and Wilcox the question simply does
not arise. Fitzmyer is aware of its importance, but his investigation into
Greek usage goes no further than a consideration of the one papyrus occur-
rence noted by BAG (1957).4 Having (rightly) rejected this as not being a
true parallel to Luke 6:7, he treats the matter as closed.
In fact there is good evidence for the use of εὑρίσκω in the sense of
‘be able’ in post-Classical Greek, much of it readily available in standard
reference works. The examples I have been able to find, without attempting
to be exhaustive, are set out below. For the first of them we need look no
further than LSJ:5
(1) Epictetus (I–II AD) 2.12.2 δὸς γοῦν ᾧ θέλεις ἡμῶν ἰδιώτην τινὰ
τὸν προσδιαλεγόμενον· καὶ οὐχ εὑρίσκει χρήσασθαι αὐτῷ, ἀλλὰ
μικρὰ κινήσας τὸν ἄνθρωπον, ἂν παρὰ σκέλος ἀπαντᾷ ἐκεῖνος,
οὐκέτι δύναται μεταχειρίσασθαι, ἀλλ᾽ ἢ λοιδορεῖ ἢ καταγελᾷ καὶ
λέγει· ἰδιώτης ἐστίν· οὐκ ἔστιν αὐτῷ χρήσασθαι.
At all events, give to anyone of us you please some layman with whom
to carry on an argument; he will find no way of dealing with him, but
after moving the man a little, in case the latter thwarts him, our man
gives up trying to handle him, and thereafter either reviles him, or
laughs him to scorn, and remarks, ‘He is a mere layman; it is impossible
to do anything with him.’ (Oldfather, Loeb ed.)
The parallel with our Lucan example is clear: οὐχ εὑρίσκει χρήσασθαι
αὐτῷ = ‘he does not find a way/is unable to deal with him.’ We notice,
too, that Epictetus, varying his word, uses δύναται to express much the
same thought further on in the sentence, when he says οὐκέτι δύναται
μεταχειρίσασθαι = ‘he is no longer able to handle him.’
II
There are further examples in literary writers of the same period:

(2) Plutarch (I–II AD), Pyrrhus 6.1 ἐκ δὲ τούτου πολλὰ μὲν περι-
νοῶν καὶ μεγάλα τῇ γνώμῃ, ταῖς δὲ ἐλπίσι μάλιστα καὶ πρῶτον
ἀντιλαμβανόμενος τῶν πλησίον, εὗρεν ἐμφῦναι τοῖς Μακεδόνων
πράγμασιν ἐκ τοιᾶσδέ τινος προφάσεως.
4
Namely, PPar 45.7 (153 BC) προσέχων μὴ εὕρῃ τι κατὰ σοῦ ἰπῖν. This was evidently
taken over by BAG from MM, who unjustifiably cited it in this connexion and failed
to notice any better parallel. It continues to appear in BAGD (1979), and BRAA (1988),
still with the outmoded reference to PPar 45 (= UPZ 69). [Also still in BDAG (2000).]
5
LSJ, s.v. εὑρίσκω I.3. ‘Act., also, find means, be able.’
A NON-ARAMAISM IN LUKE 6:7 73
After this, he began to revolve many large projects in his mind; but his
hopes were fixed first and more especially on undertakings close at
hand, and he found a way to take direct part in Macedonian affairs, on
grounds something like the following. (Perrin, Loeb ed.)
Here again we find εὑρίσκω without an object and linked directly with
an infinitive (ἐμφῦναι intrans. aor. 2, ‘grow in,’ ‘plant oneself in,’ etc.).
(3) Achilles Tatius (II AD) 1.10.2 ἐὰν γὰρ ἡ ὠδὶς παραγένηται καὶ
ἐνστῇ τῆς ἀνάγκης ἡ προθεσμία, μηδὲν πλανηθείς, κἂν πρωτοκύ-
μων ᾖς, εὑρήσεις τεκεῖν ὑπ᾽ αὐτοῦ μαιωθεὶς τοῦ θεοῦ.
For when you begin to feel the pangs and it is clear that the destined day
is at hand, you cannot go wrong, even though it be your first labour,
but you will find the way to bring forth and the god himself will deliver
you. (Gaselee, Loeb ed.)6
(4) Dio Chrysostom (I–II AD) 33.44 καὶ τινὲς μὲν ὑμῶν ὀργίζονται
καί φασί με ὑβρίζειν τὴν πόλιν, τοὺς δὲ ταῦτα ποιοῦντας οὐκ αἰτιῶ-
νται· τινὲς δὲ ἴσως καταγελῶσιν, εἰ περὶ μηδενὸς κρείττονος εὗρον
εἰπεῖν.
Some of you are angry and say that I am insulting your city, but they do
not find fault with those who do these things; some on the other hand are
perhaps laughing at the fact that I was not able to talk about anything
more important.
Next these three examples in Vita Aesopi G, a more popular text of I AD:7
(5) Vita Aesopi G 35.16–7 ὁ Ξάνθος ἀκούσας φιλοσόφου μὲν ζήτημα,
μὴ εὑρίσκων αὐτὸ δὲ ταχέως ἀναλῦσαι, [τῷ Αἰσώπῳ ἑστῶτι] λέγει·
πάντα τῇ θείᾳ προνοίᾳ διοικεῖται.
Xanthos, hearing a problem for a philosopher, but not being able to solve
it readily, said, All things are under the control of divine providence.
(6) Vita Aesopi G 48.27–9 εἷς τῶν σχολαστικῶν πρὸς τοὺς ἑτέρους
λέγει· διὰ τί τὸ μὲν πρόβατον ἐπὶ θυσίαν ἀγόμενον οὐ κέκραγεν,
τὸ δὲ χοιρίδιον μεγάλα τονθρύζει; μηδενὸς οὖν εὑρίσκοντος λῦσαι
τὴν ἐρώτησιν, Αἴσωπος λέγει. …
6
The meaning is defined as ‘discover how to’ in O’Sullivan, Lexicon to Achilles Tatius,
s.v. 3a. Rather different, I think, is 2.35.5 ἀπῆλθεν ὁ ἐραστὴς οὐχ εὑρὼν πιεῖν, where
πιεῖν is virtually a substantive and εὑρίσκω has the sense of ‘get’; cf. John 4:7, 10 δός
μοι πιεῖν. O’Sullivan however defines it as ‘get a chance to’ (3b), which makes it very
close to Luke’s use. The same uncertainty appears in Epict. 3.19.5 ἂν μὴ εὕρωμεν φαγεῖν
ἐκ βαλανείου, where again I feel φαγεῖν is quasi-substantival, though it might be adduced as
a further parallel to Luke. I wish to thank Dr Suzanne MacAlister for bringing the Achilles
Tatius examples to my notice.
7
Refs. to section and page line in Perry, Aesopica. This version of the Life, dated by Perry
to I AD, is a rare example of a work of literature written (for the most part) in popular
Koine Greek.
One of the students said to the others, Why is it that a sheep being led
to slaughter does not cry out, while a pig makes a great deal of noise?
When no one was able to answer the question, Aesop said. …
(7) Vita Aesopi G 102.4–7 ἐπ᾽ ἐκείνοις δὲ τοῖς καιροῖς ἔθος εἶχον
οἱ βασιλεῖς παρ᾽ ἀλλήλων φόρους λαμβάνειν διὰ τῆς ἐναρέτου
μάχης· οὔτε γὰρ ἐν πολέμοις συνίσταντο οὔτε μάχαις· ἔγραφον γὰρ
προβλήματα φιλοσοφίας δι᾽ ἐπιστολῶν, καὶ ὁ μὴ εὑρίσκων διαλύ-
σασθαι φόρους ἐτέλει τῷ πέμψαντι.
In those times the kings had a custom of receiving tribute from one
another by an excellent form of contest: they did not engage in wars or
battles, but used to send each other philosophical problems by letter, and
the one who was not able to solve (the problem) paid tribute to the sender.
III
In the papyri Preisigke lists some dozen examples, adding ‘usw.’ for
good measure.8 Of these the clearest are:
(8) BGU 3.822.28 (105 AD)9 καὶ [ἐά]ν σοι φανῇ, πέμψον μοι ἄγρα-
φον χάρτην, ἵνα εὕρο[με]ν ἐπιστολ[ὴν] | γράψαι.
And if it seems good to you, send me some clean writing paper, so that
I can write a letter.
(9) BGU 1.38.12 (II-III AD) ἐκομισάμην παρὰ Πτο|λεμίνου μῆλα
ξγ καὶ | παρὰ τῆς γυναικὸς Σερή|νου ἄλλα ια, καὶ γινώ|σκειν σε
θέλω, ὅτι πάντα | ἠφάνισται καὶ οὐκ εὕ|ρηκα δῶναι οἷς ἔγρα|ψές
μοι, ἀλλὰ ἠγόρακα, | εἵνα πέν[ψ]ω αὐτοῖς.
I received from Ptoleminus 63 apples and from the wife of Serenus an
additional 11, and I want you to know that they have all gone bad and
I have not been able to give them to the persons to whom you wrote
to me (to give them), but I have bought (some) to send to them.10
(10) PCairMasp 2.III.23 (VI AD) ... ὅπως εὕρωμεν ἡσύχως βιῶναι.
… so that we may be able to live quietly.
In addition to the evidence given so far, a new source of examples of this

use has been brought to light in BAGD (1979). Here we find three examples
8
Preisigke, Wörterbuch, s.v. Several of these are found in LSJ, s.v. II.2. [PCairPreis 2.12
(362 AD) in the original article has been omitted here: it is mostly restored. See Post-
script for new examples.]
9
[New date from DDBDP; previously ‘II–III AD.’]
10
ἠφάνισται ‘spoiled’: see LSJ, s.v. I.6; Shipp, MGE, 117; Lietzmann, Griechische
Papyri, no. 9 (‘sind verdorben’). This fits better than ‘have disappeared/ been lost.’ The
writer has received the apples, but there is something wrong with them.
cited from Astrampsychus (ed. Hercher [1863]), with the remark that ‘it is
perhaps better to conclude that εὑρίσκω with inf. = be able.’ While this is
a welcome addition to the evidence, BAGD simply adds it to the PPar exam-
ple mentioned above, and adduces none of the other material available.
BAGD’s references here can also be updated in the light of the valuable
work done by Browne on this text, now known as the Sortes Astrampsy-
chi, and edited in the Teubner series (1983). In the text now published by
Browne, which he has convincingly argued was first compiled in III AD
in Egypt, there are some dozen examples of the use we are looking for.11
They are all very similar, one in a question, others in responses. A selec-
tion will suffice:
(11) Sort. Astr. Quaest. 83 εἰ εὑρίσκω πωλῆσαι.
Sort. Astr. Resp. 3.8 οὐχ εὑρίσκεις ἄρτι δανείσασθαι.
Sort. Astr. Resp. 58.1 οὐχ εὑρήσεις κερδῆσαι ἀπὸ τοῦ πράγματος.
Sort. Astr. Resp. 66.3 εὑρίσκεις πωλῆσαι μετὰ κέρδους.
We may round off this collection with three examples from the Patristic
literature that happen to have come to light but might be no more than the
tip of the iceberg.12
(12) Apophth. Patrum PG 65.253B καὶ οὐχ εὗρον πρὸς ταῦτα ἀπο-
κρίνασθαι αὐτῷ.
(13) Apophth. Patrum PG 65.333C τίς ἐκ τῶν δύο εὑρίσκει ζῆσαι;
(14) Ephraem Syrus (IV AD) 2.105C ἰδοὺ περισπῶμαι πάντοθεν, καὶ
οὐδὲ τὴν μικράν μου σύναξιν εὑρίσκω βαλεῖν.
I am distracted on all sides, and am not even able to recite my Small
Office.13
The above evidence attests this use of εὑρίσκω from the first to the
sixth centuries, in sources ranging from everyday documents to highly
conscious works of literature. The conclusion seems warranted that it was
a normal, though possibly infrequent, feature of Koine Greek idiom. [More-
over, the dating of many of the examples in I–II AD (including the new
ones noted below) makes them contemporary with Luke.]
11
For the date see Browne, Sortes, p. v, and esp. Browne, ‘Origin and Date.’ BAGD’s note
in its list of ‘Writers and Writings in Antiquity,’ xxxiii, also needs revision in the light
of Browne’s researches. Remarks here and above about BAGD apply equally to BRAA
(1988), which shows no change from BAGD. [Similarly BDAG.]
12
I owe the first two to Sophocles, Greek Lexicon. There is no mention of this use in
Lampe, PGL, s.v. εὑρίσκω.
13
Quoted from Lampe, PGL, s.v. σύναξις.
IV
Where did such a use come from? Can anything be said about its ante-
cedents within Greek? In fact the origins of this use are not difficult to
detect. To begin with, there is actually a Classical occurrence very like the
examples we have been considering:
(15) Euripides, Medea 195–8
στυγίους δὲ βροτῶν οὐδεὶς λύπας
ηὕρετο μούσῃ καὶ πολυχόρδοις
ᾠδαῖς παύειν, ἐξ ὧν θάνατοι
δειναί τε τύχαι σφάλλουσι δόμους.
But the hateful griefs of mortals, no one has found a way to stop, by
music and many-stringed songs, because of which deaths and dread for-
tunes overthrow houses.14
Semantically this seems indistinguishable from the later use, though it

differs in showing the middle and not the active.15 There is no other example
exactly like it attested for the Classical period, so it might be regarded as
an isolated poetic innovation. Even at that, it shows the potential for such
an extension of the meaning of εὑρίσκω in ordinary language. But there
are other signs that it was no very striking development.
An example of εὑρίσκομαι + infinitive in Herodotus is cited by Page
and Smyth in this connexion:
(16) Hdt. 9.28.3 μετὰ δὲ τούτους ἵσταντο Κορινθίων πεντακισχίλιοι,
παρὰ δὲ σφίσι εὕροντο παρὰ Παυσανίεω ἑστάναι Ποτειδαιητέων τῶν
ἐκ Παλλήνης τοὺς παρεόντας τριηκοσίους.
Next to them stood 5000 Corinthians, and they had secured from Pau-
sanias that beside them should stand the 300 Potidaeans from Pallene
who were present.
Here we have εὑρίσκομαι in the sense of ‘find/get for oneself,’ ‘obtain,’

a frequent use in Classical Greek (LSJ IV), to which is added an infinitive
with subject in the accusative. While semantically the end result is some-
what different from the Euripides example (and our one in Luke), syn-
tactically it shows the same development.16
14
LSJ, s.v. I.4: find out or discover how to, citing this example; Page, Medea, ad loc.: ‘has
discovered how to stop’; Smyth, Grammar, §2135: ‘find how to.’
15
Cf. Humbert, Syntaxe, 105: ‘Il arrive souvent que l’interêt personnel que l’on porte à
l’action donne à celle-ci une nuance « déterminée »: en face de εὑρίσκειν « trouver »,
εὑρίσκεσθαι signifie fréquemment « arriver à trouver ».
16
Further examples just like the one in Hdt. are found in Dio Cassius: 42.20.3 συγκαθέ-
ζεσθαί τε γὰρ ἐπὶ τῶν αὐτῶν βάθρων καὶ ἐς τἆλλα συνεξετάζεσθαί σφισιν … εὕρετο
In much the same way we find Thucydides amplifying εὑρίσκω with

a ὅπως clause in:
(17) Th. 7.67.2 … πολλοὶ δὲ καὶ ἀκοντισταὶ χερσαῖοι ὡς εἰπεῖν
Ἀκαρνᾶνές τε καὶ ἄλλοι ἐπὶ ναῦς ἀναβάντες, οἳ οὐδ᾽ ὅπως καθεζο-
μένους χρὴ τὸ βέλος ἀφεῖναι εὑρήσουσι.
… and many javelin-men also, land animals so to speak, Acarnanians
and others, going aboard ships, who will not even have the ability to
throw their missiles while sitting down.17
Finally it may be observed that the same development was established

early in the case of a verb of similar semantic force, namely ἔχω. From
Homer onwards, and continuing into later Greek, this verb is occasionally
found constructed with an infinitive, giving the sense of ‘have means (to),’
‘be able (to)’ (see LSJ A.III., BAGD 1.6.a.). A NT example affords the
most apt illustration:
(18) [John] 8:6 τοῦτο δὲ ἔλεγον πειράζοντες αὐτόν, ἵνα ἔχωσιν
κατηγορεῖν αὐτοῦ.18
We may reasonably conclude, then, that our use of εὑρίσκω is not only
amply established for Koine Greek, but also has its roots in the earlier
language. There is no reason to regard it as in any way unGreek or as una-
vailable to a writer such as Luke.
Postscript
Electronic searching provides further documentary examples of εὑρίσκω, find

a way (to), be able (to), some in newly-published texts. I quote one and simply
note three others:
(‘he obtained for himself that he should sit on the same benches with them and in other
respects be reckoned among them’); 47.31.5 ἥ τε Κλεοπάτρα … εὕρετο τὸν υἱὸν …
βασιλέα τῆς Αἰγύπτου κληθῆναι.
17
The construction crops up again later: Dio Chrys. 11.108 οὐδὲ ὅπως ἀποκτείνῃ τὸν
Ἕκτορα εὑρίσκων.
18
[NA28 text] Interestingly, ευρωσιν appears as a variant to ἔχωσιν. Mention should also be
made of εὐπορῶ (‘have plenty,’ etc.), which shows a parallel semantic development:
as early as Aristotle (Top. 102a.13: LSJ I.2) it is found constructed with an infin. to
give the meaning ‘have the means (to),’ and ends up as the usual Mod. Greek word for
‘be able,’ μπορῶ.
SB 26.16608.34 (c.100 AD) ἐρεῖς | δὲ Ζοιλᾶτι ὅτι | ὡς ἐδήλωσά

σοι | οὐχ εὗρ[ο]ν δοῦναι | Ἑρμοδώρῳ.
You are to tell Zoilas that, as I informed you, I didn’t find a way to give/
couldn’t give (it) to Hermodoros.
PMich 8.477.35 (II AD); 488.15 (II AD); SB 3.6264.14 (II AD?).
7
THE UNITED BIBLE SOCIETIES’ LEXICON
AND ITS ANALYSIS OF MEANINGS
1992
Abstract
The paper focuses on one aspect of the Louw-Nida lexicon of the NT arranged
by semantic domains, namely, its approach to the lexical analysis of individual NT
words. The general characteristics of Louw and Nida’s method are first described
and illustrated, then a number of problem areas are identified. These mainly
concern questions of ‘lumping’ and ‘splitting,’ the lack of coverage of some mean-
ings, problems with definitions, and the exclusion of syntactical elements. The
potential of this lexicon to compete with the Bauer/ BDAG lexicon is considered
as well, despite its very different character as a domain lexicon.
1. INTRODUCTION
The United Bible Societies’ lexicon by Louw and Nida forms a major
new contribution to the field of NT lexicography.1 Not only does it adopt
a new approach to the organisation of the vocabulary, by arranging it
according to areas of meaning or semantic domains; it also offers a fresh
lexical analysis of each word in the NT. While taking existing dictionaries
as their starting-point,2 the editors have analysed afresh the meaning or
meanings of every word, re-thought the criteria of classification, and for-
mulated new definitions for all meanings, rejecting the long-established
but unsatisfactory method of defining by glosses.
Although the editors have not presented their material in the conven-
tional alphabetical form, the new lexical analysis that the dictionary offers
deserves to be taken very seriously, since it forms an alternative to that
1
Joannes P. Louw and Eugene A. Nida, Greek-English Lexicon of the New Testament Based
on Semantic Domains, Vol. 1 Introduction and Domains, Vol. 2 Indices (United Bible
Societies: New York, 1988) [= LN]. Parts of this paper were presented in a seminar
in the Division of Religious Studies at La Trobe University in 1990, and in a panel at
the Annual Meeting of SBL in New Orleans in November 1990.
2
See LN, 1:iv.
found in existing NT lexicons. Given the lexicographical expertise and

experience of its authors, and their determination to analyse the data anew
on the best principles, there is good reason to expect their treatment to be
an advance on all other dictionaries, including the current leader in the
field, Bauer/BAGD. That the authors intend it to be seen in this light is clear
from one of their introductory remarks: ‘The principal reason for a new
type of Greek NT lexicon is the inadequacy of most existing dictionaries,
which for the most part are limited in indicating meanings, since they
depend principally upon a series of glosses’ (1:viii).
The present paper focuses on this aspect of the Louw-Nida dictionary,
that is, its handling of the lexical analysis of each NT word. It is intended
as a contribution towards evaluating this part of the work fully and decid-
ing its potential as a rival to Bauer/BAGD. To be a serious competitor it
would presumably need to be recast in the familiar alphabetical format,
with the definitions and examples belonging to a word gathered under that
word, instead of being distributed through the various semantic domains.
Whether an edition in this form might be published is of course a matter
for Professors Louw and Nida to decide, but it is an option that could be
considered at a later stage when assessment of the work has progressed
further.
Most of what I have to say is directly concerned with, or arises out of,
the question of ‘the fineness of the semantic grid.’ I begin, then, with some
preliminary remarks on this topic.
Having collected a corpus of occurrences and assigned a provisional
meaning in each place, the lexicographer has the task of deciding how
many senses (or ‘meanings’) to recognise, where to place the divisions, and
how to define the senses recognised. In doing so, it is necessary to steer
a course between setting up too many meanings and setting up too few.
At the one extreme, it is possible to frame a separate definition for each
occurrence; at the other, one can adopt a single meaning broad enough
to embrace all the examples. How best to make the divisions cannot be
reduced to rules or predicted in any way, but will be found to vary from
word to word. It is also true that each lexicographer will make somewhat
different decisions, and there will always be room for disagreement.
All this of course is well known to Louw and Nida, and they have
offered valuable comment on the process in their own writings.3 For
example:
3
Besides the Introduction in the Lexicon itself, I have drawn on Nida and Louw, Lexical
Semantics, which the authors have kindly shown me in a 1990 draft and given me
permission to quote from in advance of publication. [The work was published in 1992.]
THE UNITED BIBLE SOCIETIES’ LEXICON 81
How many different meanings there are for a particular term depends on
the fineness of the semantic grid, that is to say, the extent to which one
tends to lump or to split differences. The ultimate criteria for such lump-
ing or splitting depend upon the entire semantic system.4
Again:
[O]ne must recognize the fundamental problem in all scientific descrip-
tions, namely, the opposing tendencies in lumping and splitting. Those who
favor lumping often insist that this simplifies the description and high-
lights the relations, but those who favor splitting insist that lumping only
moves the difficulties to another level of abstraction or analysis. There is
generally no final answer to the issue of lumping versus splitting. …
The second principle of organizing related meanings is to adopt no more
meanings than are completely necessary to account for the evidence.
This is, of course, a fundamental principle of all scientific work, since
the fewer the necessary categories or principles needed to explicate the
data the more likely is the analysis to be correct.5
These remarks, and the lexicographical principles embodied in them,
are undoubedly sound, and there can be no question that in general Louw
and Nida have carried out their task well, and arrived at satisfactory deci-
sions about the number of meanings to recognise for each word and how
to define them. Nevertheless I find that there are problems. The most seri-
ous of these, and one which seems to go beyond mere difference of per-
sonal opinion, is that there are cases in which a NT use is not covered by
the definitions given for the word in question. This can mostly be traced
to excessive ‘lumping’: the grid is simply not fine enough to catch all the
examples. There are times, however, when it seems to be due to oversight.
Besides this, one can identify a number of other problems, concerning
definition and the like.
In what follows, these points will be set out under headings, with rep-
resentative examples treated in detail. Before proceeding to this critique,
however, it will be useful to describe the general features of LN’s approach
to analysis of meanings.
2. GENERAL CHARACTERISTICS OF LOUW AND NIDA’S ANALYSIS
As already mentioned, LN uses definition as the primary means of indi-

cating meaning, an undoubtedly better method than that found in most
4
LN, 1:xvi.
5
Lexical Semantics, 75 and 42; cf. also 61. These are taken from Chapter 3, a full discus-
sion of the process of analysing different meanings of the same lexeme.
NT dictionaries.6 Glosses are not rejected altogether, however, but are

added to each definition to indicate possible ways of translating. The two
are distinguished by typographical means. Then the sense in question is
illustrated by at least one NT example, quoted in context, with translation
into English. A typical entry looks like this:
57.212 τόκος, ου m: the interest on money that has been loaned – ‘inter-
est.’ κἀγὼ ἐλθὼν σὺν τόκῳ ἂν αὐτὸ ἔπραξα ‘then I would have received
it back with interest when I returned’ Lk 19.23.
In framing definitions LN have used common sense in deciding the
appropriate amount of definition required.7 It would be possible, for exam-
ple, to define ἀστήρ in full scientific terms; LN sensibly opt for the equiva-
lent English word as the definition, thus (1.30): a star or a planet – ‘star,
planet.’ In the same way they frequently leave the definition out altogether
and allow a gloss to stand in its place. For example: 1.28 ἥλιος, ου m
– ‘the sun.’ 3.5 συκῆ, ῆς f – ‘fig tree.’ All this is unobjectionable, since
the English terms are sufficiently meaningful to the English speaker and
there is no ambiguity. On the other hand, some meanings are given the full
treatment, as for example: 19.53 πνίγωa: to apply pressure around the neck
in order to harm or kill – ‘to choke.’ In between are perhaps the majority
of LN’s definitions, such as 15.227 περιπατέωa: to walk along or around –
‘to walk, to go.’, in which it is rightly assumed that a term such as walk is
adequate in place of a more elaborate definition of the activity.8
It does not take long to discover that LN’s definitions are generally an
improvement on the indications of meaning given in BAGD. One exam-
ple will serve for the very numerous instances in which this is true. In
BAGD the word δεῖπνον is divided into two senses, described thus:
1. dinner, supper, the main meal (toward) evening
2. (formal) dinner, banquet
It is not easy to grasp from these poor descriptions what the two mean-
ings are and how they differ. LN’s analysis is not fundamentally different,
but the definitions make clear what is unclear in BAGD:
23.25 δεῖπνονb, ου n: the principal meal of the day, usually in the
evening – ‘supper, main meal.’
23.22 ἄριστονa, ου n; δεῖπνονa, ου n; βρῶσιςb, εως f: a meal whether
simple or elaborate (this generic meaning of ἄριστον and δεῖπνον is
6
Cf. LN, 1:vii.
7
Cf. LN, 1:xiii.
8
The small superscript letters attached to words mark distinct meanings; the full list of
meanings for each word is given in the Greek-English Index in Vol. 2.
in contrast with the more specific meanings of ἄριστονb, 23.23, and

δεῖπνονb, 23.25) – ‘a meal, a banquet, a feast.’
Another feature of LN’s approach to definition is that different parts of

speech, especially verb and noun, are sometimes covered by the one defini-
tion. So for example:
25.59 ἐλπίζωa; ἐλπίςa, ίδος f: to look forward with confidence to that
which is good and beneficial – ‘to hope, to hope for, hope.’
The definition given here (and in most instances, it seems) is in fact that
of the verb; the user must extract the noun’s definition from it (though a
noun gloss is provided). This method is considered by the authors to be
one of the advantages of the semantic domain approach (see 1:x), though
in practice it is not resorted to very often. One wonders whether it really
is an advantage, since it means that some parts of speech are not given a
definition in their own right. There is some awkwardness for example in
a case like the following, where it is not really clear what the definition
of εὐλογία would be:
33.470 εὐλογέωb; εὐλογίαc, ας f; κατευλογέω: to ask God to bestow
divine favor on, with the implication that the verbal act itself constitutes
a significant benefit – ‘to bless, blessing.’
To arrive at the meaning of the noun, it is not enough simply to turn

the definition into noun form: this gives ‘asking God to bestow divine
favor on. …’ In effect one needs to know the meaning already in order
to get the right result. Nevertheless one can accept the principle involved,
that ‘certain verbs and corresponding nouns have essentially the same
meaning. … Whether a writer employs a verb or a noun depends largely
upon the syntactic structure and the stylistic features of a passage’ (1:xii).
LN’s treatment of the middle and passive of verbs is also rather dif-
ferent from what we are accustomed to see in our dictionaries. The general
pattern is this: only the active of a verb is defined, and no mention is made
of the middle or passive unless it is distinct semantically, when it is given
its own definition and listed quite separately in the Greek–English Index.
So for example ἄρχομαι ‘begin’ and ἄρχω ‘rule’ are treated as separate
words, and their meanings fully defined. αἰτέω on the other hand is listed
in the active with one definition, and no mention is made of the middle;
likewise ἀποστέλλω is defined only in the active, and it is up to the user
to know whether the passive occurs and to apply the definition in the
appropriate way. This, whether or not we agree with it, is in accordance
with LN’s general policy of excluding syntactical information (on which
more will be said in a moment). Even so, some inconsistencies are to be

found. αὐξάνω ‘increase’ and αὐξάνομαι ‘increase (degree)’ get separate
entries, but one definition has to serve for both λούω and λούομαι (47.12
λούω: to wash the body). The passive of ἀποστέλλω, as we have seen,
is not mentioned, yet λυπέω ‘make sad’ and λυπέομαι ‘be sad’ are treated
as semantically distinct.
An important feature of LN’s analysis is that the criteria of classifica-
tion are strictly semantic. This has far-reaching consequences. It means
that a large amount of the material that we are used to finding in BAGD
is simply excluded. In BAGD we are given a classification not only by
lexical meaning but also by various other criteria, such as reference, the-
ological content, collocation, and above all syntax – and the mixture varies
from one word to another. None of this finds a place in LN unless it affects
the semantic content. There is a useful discussion of this point, with illus-
trations, in LN’s introduction,9 but let us take our own example, ἐσθίω.
In BAGD two senses of ἐσθίω are recognised, 1. eat, 2. consume,
devour, but the first is broken down into five sections, with further sub-
divisions within these. Sections 1. a, b, c, and d have to do with the con-
structions found with this verb – whether it is used with a direct object,
a preposition, or absolutely. Then e takes a different tack and gathers the
examples of ἐσθίω in combination with πίνω, subdividing these according
to the kind of eating referred to. Besides the numbered subdivisions, we
find along the way further breakdown of the examples, depending on the
kind of food being eaten and so on, and we are told the various expres-
sions or phrases that crop up in the NT, such as τὸ ϕαγεῖν, διδόναι τινὶ
ϕαγεῖν.
All this is familiar to users of BAGD, and we have come to take it for
granted. We certainly feel that it is useful information to be given, and
we may also admire Bauer’s ingenuity in getting it all in. Nevertheless this
type of classification, incorporating so much that has nothing to do with
lexical meaning, is, as LN say, ‘not only unsystematic but misleading.’10
LN’s treatment of this word, by contrast, has a very lean appearance. The
same two meanings are recognised, though of course better defined (see
23.1 and 20.44), but there is no syntactical information given, nor are any
phrases or collocations noticed except for one ‘unit’ or idiom, τὸν ἑαυ-
τοῦ ἄρτον ἐσθίω, which has its own entry and definition (57.190).
9
LN, 1:viii–ix, xvi–xvii. See also Louw’s discussion of πνεῦμα in Lexicography and
Translation, 70–7, and Silva, Biblical Words, 172–4. Silva brings out well the incon-
sistency of the criteria used in BAGD.
10
LN, 1:viii.
The absence of syntactical data is perhaps the most striking aspect here,
and the one which may cause us most doubts. It is easy to recognise that
omission of this material is correct according to the principles LN have
set themselves; we may feel that we can live with not knowing the con-
structions used with ἐσθίω and a multiplicity of other verbs so carefully
set out in BAGD; but we feel uneasy that ἵνα and ὅπως appear without
any mention of the mood of the accompanying verb, and quite surprised
that even prepositions are treated as much as possible without reference
to case. More will be said on this below, where we shall consider instances
in which the strain of separating syntax and semantics seems to be too
great.
As far as phrases or collocations are concerned, these have no place
in LN unless they form a semantically distinct unit. If so, they are listed
as ‘units’ under their head-words in the Greek–English Index, and then
defined separately in the appropriate domain. Nearly all LN’s ‘units,’ as
far as I can tell, are idioms, and are described as such in the place where
they are defined.11 Thus, to continue with our example ἐσθίω, the expres-
sion διδόναι τινὶ ϕαγεῖν gets no special mention because the meaning
of the whole expression is no different from the meaning of its parts; τὸν
ἑαυτοῦ ἄρτον ἐσθίω, on the other hand, has a meaning all its own, and
so appears in a domain far away from that of ἐσθίω, with its own defini-
tion, as follows:
57.190 τὸν ἑαυτοῦ ἄρτον ἐσθίω: (an idiom, literally ‘to eat one’s own
bread’) to earn a living by one’s own efforts – ‘to earn a living, to earn
a livelihood.’12
It can be seen, then, that LN’s analysis of any word will generally be
characterised by fewer divisions than BAGD’s, if only because many of
the criteria used by BAGD are simply not relevant: the only criterion is the
semantic one. But within that, what can be said of LN’s general approach
to making divisions? Do they like many fine distinctions, or do they incline
towards a small number of broad definitions? It is of course impossible
to assess this with precision, but the impression gained from a fairly wide
range of examples is that LN prefer to lump rather than split.
Let us take ἄλλος as an illustration. BAGD manage to analyse this
into about a dozen divisions. Of course not all of these are really separate
11
The exceptions are mainly ‘titles’ (e.g., 12.15 υἱὸς τοῦ θεοῦ). I have noticed one ‘fixed
phrase’ (12.38 πνεῦμα πονηρόν). It is not clear why πνεῦμα ἀκάθαρτον achieves the
status of a unit: at 12.39 it is not given a description but simply defined.
12
On the treatment of idioms see LN, 1:ix; Nida and Louw, Lexical Semantics, 106–7.
meanings, since as usual BAGD are dividing by other criteria as well, but
there still appear to be about 5 or 6 different meanings – it is hard to count
them exactly – including other, one … another, different, more, the other
(of two). LN cover all this with two definitions of a quite broad character
(also embracing ἕτερος):
58.36 ἄλλοςa, η, ο; ἄλλως; ἕτεροςa, α, ον; ἑτέρως: pertaining to that
which is different in kind or class from all other entities – ‘different,
differently, other than.’
58.37 ἄλλοςb, η, ο; ἕτεροςb, α, ον: pertaining to that which is other
than some other item implied or identified in a context – ‘other, another.’
Another example, briefly: παρακαλέω. This is a difficult word to handle,

because the senses merge imperceptibly into one another, and it is hard
to know where to draw the lines. In BAGD we find a division into five
senses, with the first further subdivided into three (or is it two?), making
seven or eight different meanings. LN settle for only four.
As I have said, this appears to be the general trend of LN’s analysis.
At the same time, instances arise in which LN split more than BAGD,
usually with good reason. So for example ἥκω: only one meaning is given
in BAGD, have come, be present, whereas in LN there are three, glossed
as: a arrive, b be here, c happen. It should be noted that the appearance
of BAGD’s entry can be deceptive in a case like this. There we find ἥκω
divided into several headings, as if there is more than one sense, but
inspection shows that all this classification is based on other criteria, and
there is in fact only one meaning recognised (given at the beginning).13
The corpus of material analysed by LN is the NT as a whole, but no
attempt is made to list occurrences exhaustively. Under each definition
we are given selected examples, usually one for each word, sometimes two,
rarely more. These have the virtue of being full quotations with translation
into English. On the other hand, there is no listing of examples under each
meaning, of the kind found in BAGD. This means that users are not told
where the editors would have placed any occurrence that is not quoted,
and they must make their own deductions. There is real uncertainty about
this at times, as will be seen further below. It also means that there is no
information about frequency or distribution.
13
For an even more deceptive case see γαμέω, in which BAGD manage to give the impression
that they are dividing into some five different senses, whereas even their own definitions
show that there is only one sense; the divisions are made on the basis of reference and
grammar. They are thus fundamentally in agreement with LN, who naturally recognise
but one sense.
Besides the restricted presentation of NT data, there is limitation of

another kind: parallels outside the NT and historical data of any kind
simply do not come within LN’s scope.14
On another front, LN go beyond what is usually provided in our NT
lexicons. In addition to the definitions and illustrative examples, many
sections have short notes which are ‘designed primarily to help translators
deal with some of the frequently recurring difficulties’ (1:xiii). The lexicon
thus complements the United Bible Societies’ well-known series of trans-
lators’ handbooks on the NT.
In concluding this section, it is worth reiterating that here and through-
out the paper I have chosen to focus on LN’s approach to the lexical analy-
sis of individual words. The classification into semantic domains and sub-
domains is however an equally important part of the lexicon, and equally
deserving of serious assessment. We should also be aware that the two are
not really independent, but interconnect with each other, as LN themselves
well explain (esp. 1:xvi).
Let us turn now to the areas in which I see problems. As already men-
tioned, representative examples will be treated. These are not in any way
exhaustive, or the outcome of a systematic search; they serve only to show
that the problems exist.
3. PROBLEMS WITH LUMPING
In LN’s analysis of λέγω (including its suppletives) a division is made

into five meanings (plus two idioms) defined as follows:
33.69 λέγωa; ϕημίa: to speak or talk, with apparent focus upon the
content of what is said – ‘to say, to talk, to tell, to speak.’
33.129 καλέωa; λέγωb; ἐπιλέγω: to speak of a person or object by
means of a proper name – ‘to call, to name.’
33.131 καλέωb; λέγωc; ϕωνέωc; ἐπικαλέω: to use an attribution in
speaking of a person – ‘to call, to name.’
33.138 λέγωd: to mark the correspondence in the meaning of foreign
expressions – ‘to mean.’
33.140 ϕημίb; λέγωe: to say something in order to explain more
fully the implications or intent of what has been said – ‘to mean, to
imply.’
14
There are occasional exceptions. At 67.183 it is mentioned that ἐπιούσιος ‘may very
well have been coined by the Gospel writers.’
It does not take long to realise that the first of these senses has to cover
a great deal. BAGD’s analysis, with its profusion of categories and senses,
is not easy to use for comparison, but it still alerts us to various areas of
meaning that LN evidently subsume under meaning a. We can probably
agree that BAGD’s separate senses ask, answer, assure, maintain, s.v. II.1.a,
etc.), are not really different from ‘say, speak’ and therefore do not need
to be split off. But what of order? This seems to me clearly a different sense
which stretches the definition in 33.69 to the limit, and certainly does not fit
under any of the other four.
As illustrations of this sense let us take:
Matt 5:34 ἐγὼ δὲ λέγω ὑμῖν μὴ ὀμόσαι ὅλως.
But I tell you not to swear at all.
Matt 4:3 Εἰ υἱὸς εἶ τοῦ θεοῦ, εἰπὲ ἵνα οἱ λίθοι οὗτοι ἄρτοι γένωνται.
If you are the son of God, order these stones to become loaves of
bread.15
The meaning here seems to go beyond merely speaking or saying some-

thing: there is an additional notion of directing that something must be
done, the semantic development being similar to that in English tell. LN’s
own definition in 33.323, in the Subdomain 33F’ Command, Order, would
have covered it very nicely: to state with force and/or authority what
others must do. But this use of λέγω has not been recognised, and does
not appear alongside κελεύω and the rest, where we might have expected
to find it.
This – to my mind excessive – lumping under meaning a contrasts with
the careful divisions made in the case of the other four meanings. One
might have thought that if λέγωd ‘mean,’ for example, deserved to be split
off into a meaning of its own then this needed to be done for ‘order’ too.
Consider LN’s example quoted in 33.138: Acts 9:36 μαθήτρια ὀνόματι
Ταβιθά, ἣ διερμηνευομένη λέγεται Δορκάς. I think it is right to recog-
nise a separate sense ‘mean’ here, but if ‘order’ is caught by meaning a,
then this use could be too (‘a disciple named Tabitha, which interpreted
says Dorcas’).16
A further difficulty is worth noticing before we leave this word. LN’s
meaning a must also serve to cover a use such as Luke 9:31 οἳ ὀϕθέντες
ἐν δόξῃ ἔλεγον τὴν ἔξοδον αὐτοῦ ‘who appeared in glory and spoke of
15
For further examples see BAGD, s.vv. λέγω II.1.c, and εἶπον 3.c, though not all those
listed are sound.
16
BAGD absurdly lump Acts 9:36 and similar exx. under II.3 call, name, at the same
time as giving ‘mean’ as the translation.
his departure.’ Presumably this has not been separately recognised because
of LN’s general policy of excluding syntax from consideration. Never-
theless one has doubts that the coverage is adequate. There does seem
to be a semantic difference between λέγω τὴν ἔξοδον and, e.g., λέγω τὴν
παραβολήν, or λέγω τὴν ἀλήθειαν.
Another example is παρακαλέω. As was mentioned above, LN keep
the divisions down to four, and this seems a fairly sensible analysis in
general. Its weakness is that once again one meaning has to cover a great
deal of ground. LN’s meanings are:
33.168 παρακαλέωa; παράκλησιςb, εως f : to ask for something ear-
nestly and with propriety – ‘to ask for (earnestly), to request, to plead
for, to appeal to, earnest request, appeal.’
33.315 καλέωe; ϕωνέωd; παρακαλέωb: to ask a person to accept offered
hospitality – ‘to invite.’
33.310 παρακαλέωc: to call to come to where the speaker is – ‘to call
together to.’
25.150 παρακαλέωd; παράκλησιςa, εως f : to cause someone to be
encouraged or consoled, either by verbal or non-verbal means – ‘to
encourage, to console, encouragement.’
Meanings b and c deal with two fairly narrow senses, d takes care of
‘comfort, encourage,’ and meaning a is left to cover the rest of the field.
This works well enough for most examples, but there is definitely some
strain at the edges. What do we do, for example, with these two?:
1 Tim 6:2 ταῦτα δίδασκε καὶ παρακάλει.
Titus 2:15 ταῦτα λάλει καὶ παρακάλει καὶ ἔλεγχε μετὰ πάσης ἐπι-
ταγῆς.
BAGD (s.v. 2) gloss the meaning as impress upon someone, urge, exhort,
which seems to be on the right track. In both places ταῦτα refers back
to teaching and instructions set out in the preceding verses. The idea
of recommending a course of action is what is involved, rather than sim-
ply asking for something. Yet in LN’s analysis, meaning a, ‘to ask for
something earnestly and with propriety,’ is the only one of LN’s mean-
ings that comes anywhere near, and is presumably the one intended to
cover them.
We also wonder if LN were right to discard ‘call upon for help’ as a
separate sense, in favour of lumping into meaning a. BAGD (s.v. 1.c) find
two examples of this use. One (2 Cor 12:8) cannot be relied on, but the
other looks as if it ought to be taken seriously, especially in view of the
long history of this sense in Greek (see BAGD and cf. LSJ, s.v. II.1):
Matt 26:53 ἢ δοκεῖς ὅτι οὐ δύναμαι παρακαλέσαι τὸν πατέρα μου,

καὶ παραστήσει μοι ἄρτι πλείω δώδεκα λεγιῶνας ἀγγέλων·
Cf., e.g., RSV: ‘Do you think that I cannot appeal to my Father, and
he will at once send me more than twelve legions of angels?’
Again, meaning a is the only one available, and must be what LN intend
here.
4. THE PLACING OF A USE IS UNCERTAIN
In the examples looked at so far, though we may not agree with their
decisions about lumping, LN still appear to be in control of the NT data.
That is, their definitions can, perhaps with some stretching, be seen to
cover all the occurrences, and the placing intended is not in serious doubt.
We move on now to instances in which it is not clear where a use is to be
placed, and we wonder whether LN have in fact addressed the question.
In Luke 24:38 we find ἀναβαίνω used as follows:
Τί τεταραγμένοι ἐστὲ καὶ διὰ τί διαλογισμοὶ ἀναβαίνουσιν ἐν τῇ
καρδίᾳ ὑμῶν·
BAGD (s.v. 2) describe it as ‘fig(urative)’ and gloss as arise. The ren-

dering of the RSV, for example, is: ‘Why are you troubled, and why do
questionings rise in your hearts?’
LN analyse ἀναβαίνω into four meanings and one idiom, as follows:
15.101 ἀναβαίνωa; ἀνέρχομαι: to move up – ‘to come up, to go up,
to ascend.’ The upward movement may be of almost any gradient, for
example, in going up a road to Jerusalem … or in going up into a tree
… or in ascending into heaven. …
15.99 ἀναβαίνωb: to move up onto an object, with specialization of
meaning in reference to boats – ‘to go aboard, to embark.’
23.196 ἀναβαίνωc: to grow, as of plants, from the time of sprouting to
mature size – ‘to sprout and grow.’
23.189 ἀναβαίνωd: to grow taller (restricted in the NT to the growth
of plants) – ‘to grow up.’
30.17 ἀναβαίνω ἐπὶ καρδίαν: (an idiom, literally, ‘to arise in the heart’)
to begin to think about something – ‘to begin to think, to think, to have
a thought occur to someone.’ ἐπὶ καρδίαν ἀνθρώπου οὐκ ἀνέβη ‘what
no one ever thought could happen’ 1 Cor 2.9.
Where is our example intended to find its home? Meanings b and d

are definitely out; c tempts us with ‘as of plants,’ as if our use might be
a metaphorical one based on the growth of plants, but this is not very
appealing. Meaning a looks as if it is intended to describe physical move-
ment of persons in space. That leaves the idiom in 30.17. Our example
could likewise be regarded as an idiom, and the meaning given there is
suitable; the difficulty is that the wording of the two expressions is not
exactly the same. In the end we are left in some doubt about what LN
intended to do with Luke 24:38.
A second instance may be mentioned without going into full detail.
In one of its uses ζητέω goes beyond simply ‘seek’ and has the idea of
enquiry by discussion (BAGD s.v. 1.c, and cf. parallel development in
ζήτησις). John 16:19 seems a pretty clear example of this use, which we
might gloss as ‘discuss, debate’:
καὶ εἶπεν αὐτοῖς, Περὶ τούτου ζητεῖτε μετ᾽ ἀλλήλων ὅτι εἶπον,
Μικρὸν καὶ οὐ θεωρεῖτέ με, καὶ πάλιν μικρὸν καὶ ὄψεσθέ με·
In LN’s analysis there are 7 meanings and one idiom. None of these,
as far as I can see, offers the meaning mentioned. We are thus uncertain
whether LN have simply missed this use, or whether they decline to recog-
nise it and intend John 16:19 to come under a definition such as 27.34: to
attempt to learn something by careful investigation or searching – ‘to try
to learn, to search, to try to find out, to seek information.’
5. A USE IS NOT COVERED17
We come now to the area of more serious weakness in LN’s analysis.

It can be shown, I believe, that there are times when LN quite fail to pro-
vide for a NT use. Whatever the reason for it, the spread of meanings
recognised for some words simply does not cover all the data.
Let us begin with θέλω. LN recognise four meanings and one idiom,
as follows:
30.58 θέλωa: to purpose, generally based upon a preference and desire –
‘to purpose.’ … Col. 1.27.
31.4 θέλωb: to have a particular view or opinion about something –
‘to be of an opinion, to think something is so.’ … 2 Pe 3.5 … θέλω
in 2 Pe 3.5 is the only NT occurrence with this meaning. …
25.1 θέλωc; θέλησις, εως f: to desire to have or experience something –
‘to desire, to want, to wish.’ … Mt 20.21 … Ga 4.9 … 1 Cor 12.18.
17
[For ‘use’ it would have been more accurate to say ‘lexical meaning.’ Cf. Lee, History,
173 n. 10.]
25.102 θέλωd: to take pleasure in something in view of its being desir-

able – ‘to like, to enjoy.’ … Mk 12.38.
33.136 θέλω εἶναι: (an idiom, literally ‘it wishes to be’) something
which is to mean or to be understood in a particular manner – ‘it means,
this is to be understood as.’ … Ac 2.12.
It is well known that from early on θέλω includes in its range of mean-
ing a sense usually glossed as ‘be willing.’ This is distinct from ‘wish,’
‘desire,’ which is also old, though not usual in Attic. In Koine Greek the
word is used freely in both senses. In the NT a clear example of ‘be will-
ing’ is:
Matt 26:14 Τότε πορευθεὶς εἷς τῶν δώδεκα, ὁ λεγόμενος ᾽Ιούδας
᾽Ισκαριώτης, πρὸς τοὺς ἀρχιερεῖς (15) εἶπεν, Τί θέλετέ μοι δοῦναι,
κἀγὼ ὑμῖν παραδώσω αὐτόν·
It is not satisfactory to see θέλω here as expressing purpose or desire;

it has merely the idea of consent or acquiescence: ‘what are you willing to
give me if I betray him to you?’18 As can be seen, none of LN’s definitions
comes near this. Nor is θέλω mentioned in Subdomain 25E Be Willing,
where ἑκούσιος and the like are gathered (25.65–67). It appears, then, that
LN have simply not noticed that this sense occurs in the NT.
There is another use of θέλω that has failed to achieve recognition in
LN’s analysis. Of several probable NT examples, a good one is:
Mark 6:48 καὶ ἰδὼν αὐτοὺς βασανιζομένους ἐν τῷ ἐλαύνειν, ἦν γὰρ
ὁ ἄνεμος ἐναντίος αὐτοῖς, περὶ τετάρτην ϕυλακὴν τῆς νυκτὸς ἔρχε-
ται πρὸς αὐτοὺς περιπατῶν ἐπὶ τῆς θαλάσσης· καὶ ἤθελεν παρελθεῖν
αὐτούς.
Although this use is poorly tabulated by BAGD, and more work needs
to be done on collecting the examples and parallels, it has long been clear
that the sense that fits the context best is something like ‘be about to.’
To quote the note on this passage from another work co-authored by
Nida:
kai ethelen parelthein autous ‘and he was going to pass them by’: it is
generally agreed that the verb thelo here does not mean ‘will’ or ‘wish,’
but is used in a weakened sense, being the equivalent of an auxiliary
(cf. Turner … Taylor … Manson).19
18
BAGD treat this and similar examples under sense 2., which begins with wish, will,
but goes on to embrace be ready, be prepared, be willing, and more. Other NT exx.:
Matt 11:14; John 5:40; 2 Thess 3:10. [In Essay 16 I take this to be the auxiliary use of
θέλω: ‘what will you give me?’]
19
Bratcher and Nida, Translator’s Handbook, 213.
Nevertheless LN do not give this sense a separate existence. In their

present analysis θέλωa is the only one available that comes near fitting
the context, but this is clearly intended to be different from ‘be about to.’
Next let us take εἰ, in the treatment of which there is a glaring omission.
LN give three meanings:
89.65 εἰa: a marker of a condition, real or hypothetical, actual or con-
trary to fact – ‘if.’
89.30 εἰb: a marker of cause or reason on the basis that an actual case
is regarded formally as a supposition – ‘since, because.’
90.26 εἰc: a marker of an indirect question as content – ‘whether, if,
that.’
In addition there are three units, εἰ τύχοι, εἰ μή, and εἰ μήν. We see
that the use of εἰ in indirect questions is adequately dealt with, but where
is εἰ in direct questions? It is clearly not covered by any of the three
meanings, and its absence from 69.11–16, where other words introduc-
ing questions are collected, confirms that LN have failed to take account
of it.
It is well known that this use occurs in the NT, but lest there be any
doubt, let us see an example:
Acts 19:2 εἶπέν τε πρὸς αὐτούς, Εἰ πνεῦμα ἅγιον ἐλάβετε πιστεύ-
σαντες·20
Another, but less well-known use of εἰ has also slipped through LN’s
net. In the NT there are two occurrences of the Hebraistic use of εἰ in
oaths, in which a negative is implied and εἰ amounts to ‘certainly not.’ So
in Mark 8:12:
καὶ ἀναστενάξας τῷ πνεύματι αὐτοῦ λέγει, Τί ἡ γενεὰ αὕτη ζητεῖ
σημεῖον; ἀμὴν λέγω ὑμῖν, εἰ δοθήσεται τῇ γενεᾷ ταύτῃ σημεῖον.21
… Truly, I say to you, no sign shall be given to this generation. (RSV)
It would hardly be satisfactory to counter that, because this use derives

(in Hebrew!) from an ordinary ‘if’ clause with aposiopesis of the main
clause, it is adequately covered by εἰa, ‘if.’ No one would imagine, with-
out being told, that a negative is to be supplied, and therefore it deserves
to be described as a separate use.
20
Other examples: Matt 12:10; 19:3; Luke 13:23; 22:49; Acts 1:6; 7:1; 21:37; 22:25.
Cf. BAGD, s.v. εἰ V.1.
21
The other is in a repeated quotation from the LXX at Heb 3:11; 4:3, 5.
One further example. One of the senses of βλέπω, from Classical

times onwards, is ‘look,’ i.e., ‘cast the gaze in a specified direction.’ We
would expect this to occur in the NT at some time or other, and so it does.
John 13:22 seems a convincing enough instance:
ἔβλεπον εἰς ἀλλήλους οἱ μαθηταὶ ἀπορούμενοι περὶ τίνος λέγει.
The disciples looked at one another, uncertain of whom he spoke. (RSV)22
LN’s coverage of βλέπω appears thorough, with 7 meanings and two
units. One of these seems at first sight to provide the required sense, but
inspection shows that it is dealing with something different:
82.10 βλέπωg: (a figurative extension of meaning of βλέπωa ‘to see,’
24.7) to be oriented in a particular direction – ‘facing.’ λιμένα τῆς Κρή-
της βλέποντα κατὰ λίβα καὶ κατὰ χῶρον ‘a harbor in Crete facing
southwest and northwest’ Ac 27.12.
This, as the example helps to make clear, describes only orientation, and
lacks the idea of using the eyes to look in a certain direction. So it appears
that LN have simply failed to observe one of the standard uses of this com-
mon verb. This is all the more surprising in view of their good handling
of ἐμβλέπωa, defined at 24.9 as: to direct one’s vision and attention to
a particular object – ‘to look straight at, to look directly at,’ a definition
which could also have embraced the missing use of βλέπω.23
There are further examples that can be added to those noticed so far in
this section. A detailed treatment would be tedious, however, and I simply
indicate the missing uses briefly, without reproducing LN’s analysis in
full.
(1) εὑρίσκω + infinitive, with the sense of ‘find a way (to), be able (to),’
found in Luke 6:7. LN’s 27.1 is the nearest, but hardly good enough. This
use deserves a place in Domain 74 Able, Capable, alongside the similar
use of ἔχω (74.12).
(2) ζῶ. The phrase ὕδωρ ζῶν (John 4:10 etc.) has a double meaning, ‘life-
giving water’/‘fresh water.’ LN deal with the former (23.88) but fail to
provide for the latter.
(3) ἡγέομαι. In 1 Thess 5:13 this verb has an extra dimension not
catered for by LN’s 31.1. The sense required is ‘hold a high opinion of’
(cf. BAGD’s esteem, respect, s.v. 2).
22
Others in BAGD, s.v. βλέπω 3.
23
LN’s bracketed note in 82.10 above also deserves comment. It is surely clear that βλέπωg
‘face’ is not a ‘figurative extension’ of βλέπω ‘see,’ but of the unrecognised βλέπω
‘look.’
(4) καταβαίνω. LN’s one definition (15.107 to move down, irrespective

of the gradient) is inapplicable to Acts 8:26, where the subject is a road.
This use would have gone nicely with κατάβασις in 15.109: the slope or
declivity of an object which one may descend.
(5) μή and μήτι in questions. The subject deserves fuller study, but there
is enough to show that these two are sometimes used to introduce neu-
tral questions, not only those expecting a negative response. See, e.g.,
John 4:33; 7:35; 18:17, 25; 4:29 (μήτι).24 LN give no recognition to
this at all: μή and μήτι appear in Subdomain 69E Markers for a Negative
Response to Questions, and there is no Subdomain for neutral questions
(not even those introduced by ἆρα).
(6) ὅστις appears only in 92.18, defined as an indefinite relative, and ὅς
is listed alone in 92.27 as the only definite relative. LN thus overlook the
well-known fact that ὅστις in the NT (and the Koine generally) is very
frequently not indefinite at all, but simply complements the declension
of ὅς.25
(7) παρά of location, ‘among, at, by,’ is covered in 83.9 and 83.25, but
παρά with reference to movement, ‘along, beside’ (e.g., Mark 1:16) is not
catered for. The word is not included, except in the meaning ‘from’ (84.5),
in Domain 84 Spacial Extensions, where we would have expected to see
it alongside διά and κατά (84.29–32).
(8) Words for ‘why?’ These appear in two places: πόθεν, ἱνατί, and λεμα
in 89.38; τί in 92.15. Apart from the unexpected division – the defini-
tions are almost identical – where is διὰ τί? And is the (admittedly rare)
use of ὅτι totally undeserving of mention?26 I can find no sign of these
two anywhere.
6. A USE IS COVERED AFTER ALL
πολλά is worth a section of its own, for its notably misleading treat-
ment. Suppose we are looking for the familiar adverbial use of πολλά as
an intensive. BAGD handle this well enough with ‘the acc. is used as an
24
Cf. BDF, §427.2; MHT, Grammar, III, 283; Barrett, John, 439, 482. The English trans-
lations of John 18:17, 25 diverge entertainingly: NIV, NRSV dutifully render as
expecting the answer ‘no’; ARV, NEB make it neutral; and KJV, RSV translate as if
the question were introduced by οὐ!
25
A good recent discussion is: Spottorno, ‘Relative Pronoun.’ LN compound the error by
quoting in 92.18 an example of ὅστις which is certainly definite (Acts 5:16).
26
Five possible examples in the NT. Mark 9:11, 28 seem secure. Cf. BAGD, s.v. ὅστις
4.b.
adv. greatly, earnestly, strictly, loudly, often, etc.’ (s.v. πολύς I.2.b.β). A
typical example is Mark 5:10 καὶ παρεκάλει αὐτὸν πολλὰ ἵνα μὴ αὐτὰ
ἀποστείλῃ ἔξω τῆς χώρας ‘and he begged him eagerly not to send them
out of the country.’
In LN we find in the Greek–English Index ‘πολλά (adv) often,’ and a
reference to 67.11. This duly provides the definition: a number of related
points of time – ‘often, many times,’ with Matt 9:14 as the example. But
that is all; no other entry for πολλά is to be seen. LN appear to be telling
us that πολλά as an adverb is used in the NT only in reference to time.
But if we happen to look under πολύς, something more emerges. Three
meanings of this word are recognised, the third of which is:
78.3 πολύςc the upper range of a scale of extent, but probably somewhat
less than for μέγαςb, μεγάλως, and μέγεθος (78.2) – ‘great, greatly,
much, a great deal.’ ἠγάπησεν πολύ ‘the great love she has shown’
Lk 7.47; τὸ πολὺ αὐτοῦ ἔλεος ‘his great mercy’ 1 Pe 1.3; πολλὰ σπα-
ράξας ‘caused him to convulse severely’ Mk 9.26; καὶ πολλὰ ἐπετίμα
αὐτοῖς ‘and he sternly warned them’ Mk 3.12. …
So here is the rest of the story of adverbial πολλά, but subsumed under
the adjective, and in a quite separate Domain (Degree). It is fair to say that
this is a bad arrangement, for at least two reasons. It is misleading for the
user of the dictionary, who may well think that πολλά is used only of time;
and it is inconsistent, in that this use is picked out and given a standing of
its own, while other uses (referring to size, number, degree) are lumped
together and placed out of sight, as it were, under one use of the adjective.
In fact πολλά as an adverb is hard to split into different senses, and
would be better regarded simply as an intensifier. LN’s own example of
the supposed sense of ‘often’ illustrates the point: Matt 9:14 διὰ τί ἡμεῖς
καὶ οἱ Φαρισαῖοι νηστεύομεν πολλά, οἱ δὲ μαθηταί σου οὐ νηστεύου-
σιν· Here πολλά is not unambiguously ‘often,’ but might just as well be
translated ‘much,’ or ‘severely.’
7. PROBLEMS WITH DEFINITIONS
While it is true, as has been said above, that LN’s policy of using defi-
nitions to indicate meaning is welcome, and the results are generally an
improvement on existing treatments, there is much in the way they have
framed their definitions that causes uneasiness. Even after due allowance
has been made for individual taste, and for the demands imposed by the
domain structure, many of the definitions seem unsatisfactory.
One problem that stands out may be illustrated by the following entry:
13.3 γίνομαιb : to possess certain characteristics, with the implication
of their having been acquired – ‘to be.’ γίνεσθε οὖν ϕρόνιμοι ὡς οἱ
ὄϕεις ‘therefore be wise as serpents’ Mt 10.16; διότι ἀγαπητοὶ ἡμῖν
ἐγενήθητε ‘because you are dear to us’ 1 Th 2.8.
What are we to make of such a definition? It may first be remarked that

if we did not already know what word was being defined, it would be dif-
ficult to guess – ‘to possess certain characteristics’ sounds like something
to do with ἔχω. But more important, it is only by looking at the examples
(and the gloss!) that we can see that the meaning being defined is: ‘be
(something specified by an adjective).’ The definition does not convey
this. We cannot see how ‘to possess certain characteristics’ fits in with
ϕρόνιμοι or ἀγαπητοί in the examples. Similar difficulty is encountered
with γίνομαιc at 13.48, defined as: to come to acquire or experience a state
(LN’s example: ὁ λόγος σὰρξ ἐγένετο Jn 1.14). How can this work in
John 1:14? Is σάρξ a ‘state’?
The problem here is that the definition is not a substitution equivalent.
That is to say, it is not a form of words that can take the same place in the
context as the word being defined and convey the same meaning. Yet the
framing of definitions in this form is usually taken to be a basic procedure
of lexicography. LN’s practice is inconsistent: many definitions are cast in
this form, but many are not. This is clearly a deliberate choice by the edi-
tors, not an oversight, but for the user it is a difficulty running through the
whole dictionary, making the definitions hard to rely on. At any time one
may encounter a definition that does not ‘fit the context.’27
Further illustrations are:
71.22 χρή: that which should be or happen, with the implication of
propriety – ‘should, ought.’ οὐ χρὴ . . . ταῦτα οὕτως γίνεσθαι ‘this
should not happen’ Jas 3.10.
71.28 ὄϕελον: that which ought to be if one only had one’s wish –
‘would that.’ καὶ ὄϕελόν γε ἐβασιλεύσατε ‘and would that you really
were kings’ … 1 Cor 4.8. …
13.68 ἐκβάλλωf: to cause a significant change of state by decisive
action – ‘to cause to be, to make become.’ ἕως ἂν ἐκβάλῃ εἰς νῖκος
27
I of course exclude from this criticism the many definitions that are not intended to be
any sort of equivalent, but employ a different mode of definition, namely description
of the function of the word. See many examples in Domains 83–92, e.g., ‘91.3 μένa: a
marker of linkage in discourse.’ Even some verbs come into this category, e.g., ‘90.45
ποιέωa: a marker of an agent relation with a numerable event.’ As this example shows,
such definitions invite another kind of criticism, on the grounds of obscurity.
τὴν κρίσιν ‘until he makes justice to triumph’ or ‘until he causes justice

to prevail’ Mt 12.20.
13.54 ὑπάγωe: to undergo a significant change – ‘to undergo, go to.’ καὶ
εἰς ἀπώλειαν ὑπάγει ‘and he goes to destruction’ or ‘he is destroyed’
Re 17.8.
A more striking manifestation of this problem is the occasional defining

of a noun as a verb. This is consistent with LN’s principle, noted earlier,
that a verb and its related noun may have essentially the same meaning
and one definition can cover both, but the result is distinctly odd when the
definition of the verb is given for the noun alone. For example:
88.131 ὕβριςa, εως f: to be insolently mistreated – ‘maltreatment, inso-
lence, and mistreatment.’
88.139 ἀνομία, ας f: to behave with complete disregard for the laws or
regulations of a society – ‘to live lawlessly, lawlessness, lawless living.’
Similarly, in at least one case an adjective is defined as if it were an

abstract noun:
41.31 βάρβαροςc, ον: a pattern of behavior associated with a low cul-
tural level – ‘uncivilized, barbarian.’28
Other problems with LN’s definitions vary considerably from case to

case and are difficult to classify. One can only say by way of summary
that at any point the reader may encounter a definition that is uninform-
ative or inaccurate or puzzling in some way. I simply offer a few samples
to illustrate.
36.1 ἡγέομαιb; προΐσταμαιa; κατευθύνω; ϕέρωd; ἄγωd: to so influence
others as to cause them to follow a recommended course of action –
‘to guide, to direct, to lead.’
The definition seems very broad, even vague. This is not a consequence
of having to cover five different words: all share the same meaning here.
Whereas the glosses are clear, the definition, rather than helping to clarify
exactly what meaning is under consideration, makes it indistinct. It sounds
just as much like a definition of πείθω as of any of these verbs. How then
is πείθω distinguished? At 33.301 we find πείθωa defined as: to convince
someone to believe something and to act on the basis of what is recom-
mended. This does not seem very different, and the impression is sup-
ported by the fact that the NT examples given by LN fit happily enough
28
Cf. the definition of πολύς quoted above, and other adjectives in Domain 78A, most of
which will not work as substitution equivalents.
under either definition. In short, we are looking at definitions that are

simply not distinctive enough to be useful.
13.85 προϕέρω: to cause to exist in an evident manner – ‘to cause
to exist clearly, to bring forth, to produce.’ ὁ ἀγαθὸς ἄνθρωπος ἐκ τοῦ
ἀγαθοῦ θησαυροῦ τῆς καρδίας προϕέρει τὸ ἀγαθόν ‘the good person
from the good treasure of the heart brings forth that which is good’ or
‘. . . causes to exist that which is good’ Lk 6.45.
The definition on its own is quite unhelpful. What does ‘to cause to
exist in an evident manner’ mean? It is only with the help of the glosses
and by studying the example that we can appreciate what is going on. Part
of the difficulty is LN’s decision to treat this use as an ‘established figura-
tive meaning,’ i.e., one which has lost its figurative impact.29 This is very
unexpected in a context where προϕέρω is combined with ἐκ τοῦ θησαυ-
ροῦ. And sure enough, in their treatment of θησαυρός in the parallel pas-
sage Matt 12:35, LN go the other way and take it as figurative (see 7.32).
In sum, one feels a lack of sureness of touch here.
13.6 περίκειμαιd: to be in a state involving various aspects – ‘to be in
various ways, to be in many ways.’ ἐπεὶ καὶ αὐτὸς περίκειται ἀσθέ-
νειαν ‘since he himself was weak in many ways’ He 5.2.
This definition suffers from all the weaknesses we have just been see-
ing: its import is hard to grasp, it is vague, and it cannot be fitted into the
context of the quoted example. We are better off with BAGD’s ‘be sub-
ject to,’ or Newman’s ‘be subject to, be beset with.’
8. SYNTAX IS THE HIDDEN FACTOR
As we have seen, the policy of the new lexicon is to exclude syntax

from consideration. This is followed with great rigor: apart from indi-
cations of case with some prepositions, virtually no syntactical data are
given. While this may seem sound as a principle, the consequences in
practice give us pause. In terms of the focus of this paper, i.e., the value
of LN’s lexical analysis of each word, doubts must arise whether a lexical
treatment that omits syntax in this thoroughgoing way remains useful.
Of course much of the syntactical material included in a dictionary such
as BAGD is not essential and may be omitted without serious conse-
quences. I am thinking rather of instances where keeping syntax out of
29
See LN, 1:xviii.
the picture seems artificial, and causes strain or is even misleading. Some
specimens follow.
In Domain 59 Quantity, two meanings of πολύς appear. At 59.1 πολύςa
is defined as: a relatively large quantity of objects or events (example: τῷ
σῷ ὀνόματι δυνάμεις πολλὰς ἐποιήσαμεν ‘in your name we did many
mighty deeds’ Matt 7:22). Then at 59.11 we find πολύςb: a relatively large
quantity (example: ἦν δὲ χόρτος πολὺς ἐν τῷ τόπῳ ‘there was much grass
there’ John 6:10). Why two separate yet so similar meanings? The hidden
factor, as we learn from the examples, is that πολύςa is plural (‘many’),
and πολύςb is singular (‘much’). The difference of meaning depends on a
syntactic feature which underlies LN’s analysis but is not mentioned. The
user who does not already know the facts is left in the dark.
At 89.49 we have ἵνα defined as: a marker of result, though in some
cases implying an underlying or indirect purpose. Then at 89.52 an almost
identical definition appears for ὡς and ὥστε: markers of result, often in
contexts implying an intended or indirect purpose. Both entries are in the
same Subdomain (89H Result). Why are they separated? Is there any real
difference between them? The reason for the separation must surely be the
unmentioned and unmentionable fact that their constructions are different.
LN’s treatment of ἄν in ‘unreal’ conditions seems the reductio ad
absurdum of the method.
71.14 ἄνa: (in combination with a past tense, as part of the apodosis of
a conditional sentence) pertaining to being certain, in view of particu-
lar circumstances of a condition contrary to fact – ‘would.’4 οὗτος εἰ
ἦν προϕήτης, ἐγίνωσκεν ἄν ‘if this person were a prophet, he would
know’ Lk 7.39.
4
The meaning of ‘certainty’ in this type of conditional sentence can-
not be attributed solely to ἄν but to the combination of ἄν with certain
tense forms. This is a typical case of a combined lexical and syntactic
meaning.
This word really has no meaning without the syntactic combination.

LN themselves are telling us this and giving us syntactical information, but
cannot bring themselves to tell us the really crucial piece, that the com-
bination with the indicative mood is what gives this meaning.
Perhaps the most difficult to understand and the most likely to mislead is
LN’s handling of prepositions. Prepositions occur in a variety of Domains,
usually without any reference to case, though occasional mentions occur.
In Domain 84 Spacial Extensions, for example, some 18 prepositions
appear. Cases are not given, except in three instances (84.5, 84.29, 84.32).
The reason for the exceptions is not immediately apparent. Then we find
the strange Domain 90, called Case, in which, along with other words,
many prepositions appear, now with cases uniformly indicated as far as
Subdomain M Experiencer, where they suddenly stop. So in 90.41 one
meaning of εἰς has the case added, but in 90.59 another does not.30 The
introduction to this Domain (pp. 796–7), lengthy compared with that given
to other Domains, does nothing to dispel the mystery. Case, we are told,
‘is defined in terms of the relation of participants to events or states,’ and
‘there are a number of different prepositions which are employed to mark
quite diverse case relations.’ Whatever all this means, it seems clear that
it is very difficult to analyse the semantics of prepositions and present them
in a useful way without taking the accompanying case into account.
9. CONCLUSIONS
First let us recognise that the semantic domain approach adopted by LN

is of great value, offering a new way of working with the NT vocabulary.
The stated objective of providing a tool for translators seems also to have
been successfully achieved. The value of the work in these ways is not to
be underestimated.
But what of the lexical analysis as such? Again, let it be clearly stated
that the greater part of it is of very high quality and an advance on pre-
vious work. Time and again LN are seen to have improved the rather
muddled material often found in BAGD, by re-analysing the meanings,
using definitions instead of glosses, and applying consistent criteria of
classification.
Nevertheless, one has to say that the work is flawed. The gaps in cover-
age of the NT data are the major worry. At first it appeared possible that
these were few and not very serious, in that they were matters of differ-
ence of opinion about placing of NT uses. As this investigation has pro-
gressed, however, it has become clear that there are real and serious gaps,
and that they are unpredictable in extent: at any time one may be looking
at an analysis that fails to cover NT usage fully. The problem is made
worse by the fact that it is next to impossible to tell if anything is missing
without an extensive investigation. The unsatisfactory character of many
of the definitions is also a serious concern.
What could be done in a revised edition? I suggest that another edition
in alphabetical format would be of value. In it all the senses of each word
30
σὺν χειρί (90.2) is also given a case indication, but not χράομαι, καταχράομαι (90.13)
or ἐπιρίπτω ἐπί (90.18).
could be gathered under the word, and the NT examples listed more fully.
But at the same time it would be necessary to correct the deficiencies in
coverage and wording of definitions. To do this, nothing less than a com-
plete check and reworking of the NT data would be required. This is no
small job; yet only if it were done could the dictionary be reliable and
offer the really sound lexical analysis of the NT that we still lack. Whether
it is possible to carry out such a revision is a matter for the authors.
10. LN AND THE NEW MM
The question of the value of LN’s lexical analysis of the NT has particular
relevance to the task of producing a new edition of Moulton and Milligan’s
Vocabulary of the Greek Testament, a project in which G. H. R. Horsley
and this writer are currently engaged.31 The object of the work is, in essence,
to present the documentary evidence available for each NT word and use.
The collecting of data from the inscriptions and papyri is of course the main
task, to which an essential adjunct is a clear statement of the NT mean-
ings to be illustrated. Some such description is obviously required both as a
guide to what is being sought in the documents and, at the editing stage, as
a framework for the presentation of parallels. Though revisions will almost
certainly need to be made in the light of the evidence found, a starting-point
of some kind must be adopted.
A simple but reliable lexical analysis of the NT vocabulary is therefore
a necessity. Surprisingly, such a thing is not easily found. At the outset
of the project it was taken for granted that BAGD’s would be the analysis
on which we depended. The problems with this have become increasingly
evident. The most notable of them – the inadequate handling of definition
of meaning, and the classification by diverse criteria – are the same as led
LN to produce a new lexicon. The effect of these and other difficulties is
that a lexical analysis cannot be extracted easily from BAGD. Except in the
case of the most straightforward words, anything up to a full re-analysis
may be required.
31
The project was first proposed at Macquarie University in 1979 and has been pursued
continuously since then, first at Macquarie, and then (since 1988) as a joint project
of Sydney, La Trobe, and Macquarie Universities in Australia. Detailed testing of the
proposal has appeared in G. H. R. Horsley, New Documents Illustrating Early Christianity
(5 vols., North Ryde, NSW, 1981–89). [Since 2005 no further collaborative work on this
project has been done. Nevertheless it is worth keeping this section as showing the issues
on our minds at the time and still relevant to such a project.]
At this point, clearly, the new lexicon of Louw and Nida presents itself
as a possible alternative. The investigation undertaken in this paper, how-
ever, leads to the conclusion that our problem remains unsolved. LN’s
analysis, though it is certainly an advance on BAGD’s, is simply not reli-
able enough in its descriptions of meaning and, more particularly, in its
coverage of the NT data.
Postscript
Louw responded to this paper with a vigorous defence in ‘Analysis of Meaning’

(1993). A fuller assessment of LN appears in Lee, History, 155–65. In considering
its place in the history of NT lexicography, I called it a ‘breakthrough’ because
of its thorough application of the definition method, while also giving due appre-
ciation to its adoption of a domain structure (155–6, 180). The issue of coverage
of meanings in the NT remains, however. I returned to it in History, 162–4, pre-
senting new examples. A more recent description and critique of LN is found in
Peláez, ‘NT Lexicography,’ 37–54, and for comment on my assessment of LN,
see Porter, ‘Lexical Semantics,’ 124–5. The question of an alphabetic arrangement
of LN’s data has not gone away: see Peláez, ibid., 53–4; cf. Lee, History, 173
n. 6. The temptation to use the ‘Greek–English Index’ in vol. 2 with its handy
glosses as a de facto small dictionary is hard to resist.
8
A LEXICAL STUDY THIRTY YEARS ON1
2003
Abstract
The author’s Cambridge dissertation, A Lexical Study, was completed thirty

years ago. The aims of this paper are (a) to set that work in the context of LXX
studies at the time, the late 1960s; (b) to consider what investigation has been done
along the same lines since then (to 2003); and (c) to present a specimen of simi-
lar work waiting to be done. Words for ‘to order’ in the LXX-Pentateuch are
selected for examination. The three main words used in the Pentateuch, ἐντέλλομαι,
συντάσσω, and προστάσσω are first illustrated, then the documentary evidence for
the third century BC is presented, illustrated by examples; a fourth word, κελεύω,
not actually used in the Pentateuch, is also noticed. Conclusions are drawn about
the usage of the words and differences between them in third-century BC Greek.
It is argued that any differences are found not in lexical meaning but in connota-
tion or ‘tone.’
1. A LEXICAL STUDY
My doctoral dissertation entitled A Lexical Study of the Septuagint Ver-

sion of the Pentateuch was published in 1983. This was an unrevised ver-
sion of a work that had been completed and submitted in 1970, from research
begun in 1966. Its main objective was to demonstrate the links between
the vocabulary of the LXX Pentateuch and that of contemporary Koine
Greek, especially as evidenced in the Egyptian papyri. This contributed
to the debate about the nature of LXX Greek, in support of the view that
the Greek of the LXX is essentially the Greek of its time. It was not a new
idea to use the documentary evidence for this purpose: the initial discovery
had been made by Deissmann. But it had not been followed up for the
1
An earlier form of this paper was presented at a Language Colloquium at Macquarie
University, 14 September, 2001. I thank Trevor Evans for the invitation to participate
and for running an observant eye over this final version. [The essay found an appropri-
ate place in Paul, et al. (eds.), Emanuel, in honour of my long-standing friend Emanuel
Tov.]
LXX (unlike the NT) after the first enthusiasm. On the contrary, a reac-
tion had set in and the phantom of ‘Jewish-Greek’ had made a comeback.
Moreover, the evidence had not been thoroughly exploited by Deissmann
and had continued to grow in the meantime. There was clearly a need to
pursue the question further and make the case stronger.
The work was well received by most reviewers and is regularly referred
to in the literature. There seems to be general acceptance that the case has
been proved.2 This is gratifying, but it is not the end of the story. There is
much still to be done to make full use of this evidence, not only in order
to strengthen the case further but, perhaps more importantly, to throw light
in all sorts of ways on the meaning and usage of LXX words. In this paper
I want to look back and set that study in the context of its time, consider
developments since, and present a specimen of the kind of work that is
still waiting to be done.
My training at the University of Sydney, where I did my first degree,
was primarily in Classical Greek, and as part of it I was introduced to Koine
Greek in a course conducted by G. P. Shipp.3 I had also spent some years
learning Hebrew, and did some work on the LXX in my honours year.
Even before I left Australia to go to Cambridge, it struck me that there
was a topic of interest in some combination of the LXX and the language
of the papyri, though I did not have a clear idea of how it could be
focused.
Early in the century Cambridge had been a centre of LXX studies,
but by 1966 the great names of the past, Swete, Thackeray, Brooke, and
McLean, were a distant memory.4 Sebastian Brock, who had known Peter
Walters, and had recently arrived in Cambridge, was, I think, the only
LXX specialist there at the time. My supervisor, Barnabas Lindars, SSF, a
fine scholar who was subsequently elected to the Rylands Chair of Bibli-
cal Criticism and Exegesis at Manchester, was primarily a NT expert and
a little mystified by my approach. Important help came to me from John
Chadwick, who taught me the basics of lexicography.
2
See, e.g., LEH, Lexicon, VIII; Muraoka, Lexicon, IX n. 14; HDM, La Bible grecque,
234; Jobes and Silva, Invitation, 263.
3
G. P. Shipp (1900–1980), Professor of Greek 1954–1965, best known for his work on
Homer, but a philologist whose range covered all of Greek and Latin. The Sydney BA
provided (as it still does) the opportunity to mix subjects from different areas. I had
done enough Latin as a schoolboy to satisfy me.
4
For a helpful summary of the life and work of these scholars, see Jobes and Silva, Invita-
tion, 247–52. On one occasion, in 1970, I dined as a guest in Christ’s College. When I
mentioned my work on the LXX to one of the older Fellows, he recalled McLean, who
had been a Fellow and Master of the College (he died in 1947).
A LEXICAL STUDY THIRTY YEARS ON 107
Although there was contemporary debate about the nature of LXX Greek,
the issue was not a priority in current LXX research, and the idea of har-
vesting the rich documentary evidence did not seem to be much on any-
one’s mind.5 This despite the fact that Deissmann’s discovery dated from
1895, and the subsequent work of Moulton and Milligan was focused on
the NT (and covered that only partially) and was itself well out of date.
I felt isolated in choosing this topic; it even seemed that I was venturing
into strange and dubious territory. As to the lexicography of the LXX,
though everyone said that a new lexicon was needed, there was no project
in place to make it a reality.
LXX studies at that time were quiet compared with the remarkable bur-
geoning that has occurred since. Much of the work was on the text – as
it had been in the previous generation also – and the important Göttingen
edition was a focus of effort. Sidney Jellicoe’s 1968 book, which marks a
sort of turning-point, being the first general book since Swete’s 1902 classic,
appeared during the time I was working. The IOSCS only began at the same
time, holding its inaugural meeting in December, 1968.6 Contact between
scholars in those days was not what it is today. Email did not exist, and fly-
ing to and fro between countries was yet to become common (my trip from
Australia to England in 1966 was actually a sea voyage of four weeks).
My first visit to the USA did not take place until 1985.
My work in Cambridge of course predated the computer. Those who
have not lived through that revolution may not fully appreciate how far-
reaching are the changes it has wrought. Electronic tools and resources
did not exist. Not just hours or weeks, but months were spent searching
for words in the indexes of documentary volumes and confirming occur-
rences. Every text had to be laboriously copied by hand in the library, then
recopied into the manuscript when written, before finally being handed
over for typing. Typing of Greek could just be achieved by the latest
advance in technology, the IBM ‘golf ball’; even so, accents and other
diacritics had to be inserted by hand. Photocopying was only just begin-
ning to be possible. As to other tools, even such a basic thing as a list of
editions of documentary papyri was not to hand when I began.7
5
The papyrologist Orsolina Montevecchi was a notable exception. She had already drawn
attention to the links and continued to do so. See HDM, La Bible grecque, 233, 243,
248; Lee, Lexical Study, 6.
6
For an account of the IOSCS, see Jobes and Silva, Invitation, 311–3. I still possess a copy
of the original first Bulletin, of June 1968, consisting of four sides of mimeographed
foolscap. This was reprinted in BIOSCS 2.
7
I remember the kindness of E. G. Turner, Professor of Papyrology at University Col-
lege, London, in showing me an advance copy of the list in his Greek Papyri (1968)
and marking for me the volumes that contained Ptolemaic papyri.
These remarks indicate the practical constraints under which work of

this kind proceeded. Today [in 2003] searches for words in both inscrip-
tions and papyri can be done in seconds via the CD ROM PHI 7, texts
can be inspected on screen and portions copied, and the whole LXX is
available electronically. There are also two commentaries on the Pen-
tateuch, a lexicon (or two) of the LXX, an English translation project in
progress, and for good measure a CD ROM containing most of Greek
literature to 1453.
Obviously the book has limitations. In general it lacks the compre-
hensiveness that could be achieved today (even within the length limit
imposed). Moreover, it was restricted to a certain kind of parallel, namely
new developments in the language, so as to establish the point as persua-
sively as possible; and I was at all times very cautious to include only the
most convincing examples. A less restrained collection of evidence (even
if limited to the Pentateuch) could be twice as long, and even then would
not exhaust the potential of this line of enquiry.8
What has happened since? What sort of follow-up work has there been?
It is safe to say not very much. There have been minor studies in articles,
but there has been no book or major work that pursues the subject thorough-
ly.9 There is no ‘Moulton and Milligan’ for the LXX. The two lexicons that
have appeared since I completed my study (Lust, Eynikel, and Hauspie for
the whole LXX and Muraoka for the Twelve Prophets)10 have performed
important service in providing tools after such a long time when the only
lexicon was Schleusner (1820–1821), but, as their authors admit, they have
not attempted fresh searching of documentary sources: they have relied on
existing collections. For myself, one discovery in Lexical Study (on ὕω and
βρέχω) was the inspiration for a line of investigation on ‘formality’ in the
NT; but my interests generally since 1970 have gone in other directions.11
8
I think the decision to publish the work as it stood was right, though the delay was
regrettable. Even so, nothing occurred up to 1983 to change the picture significantly;
it is not as if the book has been replaced even now. Various minor weaknesses were
pointed out by reviewers, mostly rightly. But Ina Willi-Plein’s wholly negative review
(in ThLZ: see bibliog.) occasioned me as much perplexity (‘Ratlosigkeit’) as she said
she felt in trying to review the book.
9
A noteworthy article is that of Hélène Cadell, ‘Vocabulaire de la législation.’ For further
bibliography, see HDM, La Bible grecque, 243. Documentary evidence is used spro-
radically in the papers of Tov, Swinn, and Lee in Muraoka (ed.), Melbourne Symposium.
That memorable occasion was one of many valued contacts with the honorand of the
present volume (Emanuel Tov).
10
LEH, Lexicon (2003); Muraoka, Lexicon (Twelve Prophets) (1993). [Now Muraoka,
Lexicon for the whole LXX.]
11
The NT study was published as ‘Some Features’ (1985). This could still be followed
further in the LXX. Chapter 8 of Lexical Study on linguistic evidence for dating the
2. ‘ORDER’ WORDS IN THE LXX-PENTATEUCH
I turn now to a specimen of what might still be done to illuminate

the LXX vocabulary using contemporary documentary evidence. I have
selected the group of words for ‘to order.’ One naturally expects to begin
with the ordinary word for ‘to order’ in ancient Greek, κελεύω. It is a
surprising fact that there is not a single example in the LXX Pentateuch.
Opportunities to express the idea were certainly not lacking: the usual
Hebrew word (‫ )צוה‬occurs many times. The translators must have been
using something other than κελεύω. What they were using is soon dis-
covered. The following three verbs account for almost all expressions of
the idea in the Pentateuch: ἐντέλλομαι (156 times), συντάσσω (82), and
προστάσσω (12).12 Not so obvious is why this should be so. Why not
κελεύω? What has happened to it? I think the documentary evidence
helps considerably in finding an answer.
Let us begin with a summary of the usage of the two commonest words
in the LXX Pentateuch and a selection of examples. A look through the
occurrences soon shows that ἐντέλλομαι and συντάσσω are semantically
close; it is difficult to find any grounds for giving them separate lexical
meanings. Both are used for orders given by an authority figure (God,
Moses, Pharaoh, a parent, a priest). The manner of ordering is oral, rather
than written, in both. Both are used for instructions to be immediately
acted on as well as longer-term directions, and both are used to order any
kind of action. ἐντέλλομαι with the law (or provisions of the law) as
object is noticeably frequent. Finally, it is easy to find places in which the
translators use the two interchangeably in near proximity, referring to the
same event and rendering the same Hebrew word.13
LXX Pentateuch was very experimental and limited (as I knew full well). Gathering
the data for this purpose could be done much more thoroughly and with less effort now;
finding valid tests is the tricky part. [Ross, in ‘Lexical Possibilities (2016), reinvestigated
ὁράω and βλέπω and concluded that ‘even in light of new data, Lee’s conclusions
prove remarkably accurate, suggesting the potential of his methodology for further
application.’ (341).]
12
Figures from LEH, Lexicon, drawn in turn from the CATSS files. Other words used in
Pent.: ἐπιτάσσω (x 1); ἐξηγοῦμαι (x 1); also λέγω/εἶπον, of which possible examples
are: Exod 35:1 (cf. Le Boulluec and Sandevoir, L’Exode, ad loc.), Lev 9:6; Num 32:27
(cf. LEH). Not in Pent.: διατάσσω, παραγγέλλω. Not in this sense in Pent.: διαστέλλω.
Outside the Pent. κελεύω occurs 29 times, but only in books not in the Hebrew Bible;
ἐντέλλομαι is by far the favourite word.
13
E.g., Exod 31:6 (συντ.), 11 (ἐντ.); Lev 8:4 (συντ.), 5 (ἐντ.), 9 (συντ.); Num 2:33
(ἐντ.), 34 (συντ.); Deut 5:12 (ἐντ.), 15 (συντ.), 16 (ἐντ.). All ~ ‫ צוה‬piel. Cf. Dorival,
Les Nombres, 54, 56. Wevers (Numbers, 31, 494) says he can find no distinction
between ἐντέλλομαι and συντάσσω. A full translation-technical study is not undertaken
A selection of typical examples first of ἐντέλλομαι (with one of προ-

στάσσω), and then of συντάσσω:
(1) Gen 3:11 καὶ εἶπεν αὐτῷ Τίς ἀνήγγειλέν σοι ὅτι γυμνὸς εἶ, εἰ
μὴ ἀπὸ τοῦ ξύλου, οὗ ἐνετειλάμην σοι τούτου μόνου μὴ ϕαγεῖν ἀπ᾽
αὐτοῦ, ἔϕαγες;
And he said to him: Who told you that you are naked, unless you have
eaten of the tree of which I commanded you of this one alone not to
eat of it?
(2) Gen 12:20 καὶ ἐνετείλατο Φαραὼ ἀνδράσιν περὶ Ἀβρὰμ συμπρο-
πέμψαι αὐτὸν καὶ τὴν γυναῖκα αὐτοῦ καὶ πάντα, ὅσα ἦν αὐτῷ, καὶ
Λὼτ μετ᾽ αὐτοῦ.
And Pharao gave orders to some men regarding Abram to conduct
him on his way and his wife and all his property, and Lot with him.
(3) Gen 28:1 προσκαλεσάμενος δὲ ᾽Ισαὰκ τὸν ᾽Ιακὼβ εὐλόγησεν
αὐτόν, καὶ ἐνετείλατο αὐτῷ λέγων Οὐ λήμψῃ γυναῖκα ἐκ τῶν θυγα-
τέρων Χανάαν·
Isaak summoned Iakob and blessed him and ordered him saying, You
shall not take a wife from the daughters of Khanaan.
(4) Exod 7:6 ἐποίησεν δὲ Μωυσῆς καὶ Ἀαρών, καθάπερ ἐνετείλατο
αὐτοῖς κύριος, οὕτως ἐποίησαν.
Moyses and Aaron did as the Lord instructed them, thus they did.
(5) Deut 27:1 καὶ προσέταξεν Μωυσῆς καὶ ἡ γερουσία ᾽Ισραὴλ
λέγων Φυλάσσεσθε πάσας τὰς ἐντολὰς ταύτας, ὅσας ἐγὼ ἐντέλλομαι
ὑμῖν σήμερον.
Moyses and the council of elders of Israel ordered (them) saying,
Keep all these commandments which I command you today.
(6) Exod 1:17 ἐϕοβήθησαν δὲ αἱ μαῖαι τὸν θεόν, καὶ οὐκ ἐποίησαν
καθότι συνέταξεν αὐταῖς ὁ βασιλεὺς Αἰγύπτου, καὶ ἐζωογόνουν τὰ
ἄρσενα.
The midwives feared God and did not do as the king of Egypt directed
them, and they kept the male children alive.
(7) Exod 12:35 οἱ δὲ υἱοὶ ᾽Ισραὴλ ἐποίησαν καθὰ συνέταξεν αὐτοῖς
Μωυσῆς, καὶ ᾔτησαν παρὰ τῶν Αἰγυπτίων σκεύη ἀργυρᾶ καὶ χρυσᾶ
καὶ ἱματισμόν·
The sons of Israel did as Moyses ordered them and asked for silver
and gold objects and clothing from the Egyptians.
here, but would complete the picture. In two of my examples (Exod 1:17; 12:35)
συντάσσω ~ ‫ דבר‬piel/subst.
(8) Lev 13:54 καὶ συντάξει ὁ ἱερεύς, καὶ πλυνεῖ ἐϕ᾽ οὗ ἂν ᾖ ἐπ᾽ αὐτοῦ
ἡ ἁϕή, καὶ ἀϕοριεῖ ὁ ἱερεὺς τὴν ἁϕὴν ἑπτὰ ἡμέρας τὸ δεύτερον·
The priest shall give directions and shall wash [i.e., get someone to
wash] the article on which the mark of infection is, and the priest shall
set apart the mark of infection for seven days a second time.
(9) Num 2:34 καὶ ἐποίησαν οἱ υἱοὶ ᾽Ισραὴλ πάντα, ὅσα συνέταξεν
κύριος τῷ Μωυσῇ, οὕτως παρενέβαλον κατὰ τάγμα αὐτῶν καὶ
οὕτως ἐξῆρον.
The sons of Israel did everything that the Lord instructed Moyses,
thus they encamped according to their units and thus they set out on
the march.14
3. DOCUMENTARY EVIDENCE
Now for the documentary evidence. A search (via PHI 7) in nineteen

volumes of papyri produced these numbers of occurrences in third cen-
tury BC texts: ἐντέλλομαι 22; κελεύω 25; προστάσσω 32; συντάσσω
285.15 Let me summarise what a perusal of this evidence indicates, and
then give some representative examples.
It is obvious at once that συντάσσω was a favourite word in Ptolemaic
officialese in the time of the LXX Pentateuch. In meaning it is definitely in
the ‘order’ field with the other verbs. This is a somewhat unexpected flour-
ishing: the word is old, but in Classical usage it mostly means ‘arrange,’
‘organise,’ and the like, though the beginnings of the development to ‘order’
can just be seen in some occurrences.16
ἐντέλλομαι goes back to earlier Greek, though it was not very com-
mon, it seems, outside Herodotus. It is still running along in Ptolemaic
Greek with its meaning as before (‘command,’ etc.) but less favoured than
συντάσσω in the kind of language found in papyrus documents.
προστάσσω was well established in Classical usage as one of the
‘order’ words and maintains its ground.
14
ἐξῆρον sic Rahlfs, and Wevers, Göttingen ed.; but a 2nd aor. of αἴρω and cmpds. is
unusual if not unknown. Read ἐξῆραν?
15
PHib 1, 2; PLille 1, 2; PSI 4, 5, 6; PPetr 1–3; PMich 1; PCol 3, 4; PCorn; PCairZen 1–5.
The counts cannot be absolutely precise. I have dropped grossly restored examples, but
opinions would differ. Only III BC texts are counted. I have not made any attempt
on the inscriptions for the usual reasons: III BC texts are scattered through numerous
volumes; they come from a variety of ancient localities, not just Egypt; and their genres
vary markedly.
16
See LSJ, s.v. II.4.a and b (but Aeschin. 2.22 is quite different). More examples and
comment in Anz, Subsidia, 324–5.
κελεύω is the oldest member of the group, being the standard word from
Homer onwards. It is well known, however, that it may show a weaker
sense translatable as ‘request,’ ‘urge,’ ‘tell.’ It loses some of its vigour
as a result of long and frequent use. In the Ptolemaic documents it has
clearly suffered something of an eclipse, when compared with the over-
whelming popularity of its new rival συντάσσω. But this does not mean
it had dropped out: it was always still available in both its weaker and its
full sense. Later, from the first century AD onwards, it makes a comeback
alongside other words, and συντάσσω in its turn fades away.17
All of these verbs are close semantically. The difference between them
is not one of lexical meaning so much as of connotation, of ‘flavour’ or
‘tone.’ To determine this exactly is a difficult exercise; only a speaker of
the time could appreciate it fully. We can choose, then, to give up the
attempt altogether, or do what we can, relying on the hints gleaned from
context and situation.
I suggest the key is this: ἐντέλλομαι, προστάσσω, and συντάσσω are
more formal and official-sounding than κελεύω. They are in use, or have
come into use, to meet the need for words with just that flavour as com-
pared to the plain, slightly worn-out feel of κελεύω. συντάσσω is the new-
comer and still fresh; it is businesslike, with some hint of its origin in the
meaning ‘arrange.’ προστάσσω is not markedly different in flavour from
συντάσσω. How ἐντέλλομαι differs from these two is not obvious; perhaps
it is a little elevated. At the same time as all this, we have to allow for the
use of κελεύω in its original sense with full force, especially in literary writ-
ing. This is why most of the LXX examples are in the more literary books
(2–4 Maccabees account for two-thirds), and a writer like Polybius uses it
124 times.18
A turn-over in ‘order’ words is in fact a general phenomenon: they tend
to lose their punch and become weak, or to lack the authoritative sound
felt desirable in certain contexts. In English we have a kaleidoscope of
17
A full study of the history of these words, from Class. to Mod., is a desideratum. It
would include the curious later development of κελεύω to ‘please.’
18
Mauersberger, Polybios-Lexikon. Figures for the other words in Polybius: ἐντέλλο-
μαι 17, προστάσσω 60, συντάσσω 84 (the last two from my own search via TLG).
Other attempts to distinguish ‘order’ words have generally looked for differences in the
manner or substance of the order. For Pelletier, κελεύω ‘tend à se réserver pour les cas
où l’ordre est donné de vive voix,’ and ἐντέλλομαι ‘a été choisi pour exprimer le ton
d’autorité bienveillante’ (‘Vocabulaire,’ 239, 240). Pelletier proceeds entirely without
reference to the papyrus evidence, and uses one inscription from Teos. The idea that
κελεύω refers to verbal orders is an old one, found already in Schmidt’s treatment
(Synonymik, 1:199–214). Harl (La Genèse, 54, 103) follows Pelletier on ἐντέλλομαι.
Cf. another treatment in Louw and Nida, Lexicon, 33.323, 325, 329.
words and expressions of different ages and flavours at our disposal: bid,
enjoin, charge, command, prescribe, direct, order, instruct, issue instruc-
tions, issue orders, issue a directive, tell.19
Next a selection of examples from third century BC papyri, the first a
nice parallel to (6) Exod 1:17 above:
(10) PCairZen 2.155.1–3 (Philadelpheia, 256 BC, letter)
Ἀπολλώνιος Ζήνωνι χαίρειν. ὁ βασιλεὺς συνέτασσεν ἡμῖν | δισπορῆ-
σαι τὴν γῆν. ὡς ἂν οὖν ἐχθερίσηις τὸν πρώιον σῖτον, | εὐθέως πότι-
σον τὴν γῆν ἀπὸ χερός.
Apollonios to Zenon greetings. The king has issued instructions to us
to do the second sowing of the land. When therefore you have finished
harvesting the early grain, at once irrigate the land by hand.

Ἀρτεμίδωρος Πανακέστορι χαίρειν. παραγινομένου μου ἐγ Βουβάστου
εἰς Μέμ[ϕιν] | ἐνετέλλετο Ἀπολλώνιος μάλιστα μὲν αὐτὸν διελθεῖν
πρὸς σέ, εἰ δὲ μή, ἀποστεῖλα[ί] | τινα παρ᾽ ἐμοῦ ὃς ἀναγγελεῖ σοι
τὰ παρ᾽ αὐτοῦ. ἠκηκόει γὰρ ὅτι οὐ πᾶσα κατασπείρετα[ι] | ἡ γῆ αἱ
μύριαι ἄρουραι. συνέτασσεν οὖν ἀναγγέλλειν σοι ἵνα ξυλοκοπηθῆι
πᾶσα καὶ ποτισθῆι.
Artemidoros to Panakestor greetings. When I arrived at Memphis from
Boubastos, Apollonios instructed that preferably I myself should come
to you, but failing that to send someone from me to convey to you his
wishes. For he heard that not all the land, the ten thousand arourai, are
sown. He gave instructions therefore to convey to you that all of it is
to be cleared and irrigated.
This passage illustrates the interchangeability of the two verbs as well

as their general character. Apollonios, also met in (10), was a high official
in the administration.
Ζωίλος Ἀλεξάνδρωι χαίρειν. καλῶς ἂν ἔχοι εἰ ἔρρωσαι· ὑγιαίνο-
μεν δὲ καὶ αὐτοί. | ἐκομισάμην τὰς ἐπιστολὰς καὶ οὐκ ἀμελῶ περὶ
ὧν ἂν ἡμῖν γράϕηις. σὺ δὲ καλῶς | ποιήσεις ἐκπονήσας ἵνα ἡμῖν
γίνηται ἅ σοι ἐνετειλάμεθα κατὰ τὸ ὑπόμνημα, | τούτου δὲ γενομέ-
νου ἐπί[στ]ασο ὅτι ὀϕειλήσω σοι χάριν ἱκανήν.
Zoilos to Alexandros greetings. It would be good if you are well; we too
are in good health. I have received your letters and I am not neglecting
the matters about which you wrote to us. You will do well to ensure
that what we instructed you according to the memorandum is done for
us, and if this happens, know that I will owe you considerable gratitude.
19
Cf. Buck, Dictionary, 19.45. Note esp. developments from Latin to Romance.
Here in (12) we have an ordinary example of ἐντέλλομαι in a friendly

though businesslike letter.
(13) PCairZen 5.852.7–10 (Philadelpheia, III BC, polite letter to Zenon
asking for placement)
εἰ δ᾽ ἐπὶ τοῦ παρόντος ἡσυχίαν ἄγειν κελεύεις, τὸμ μὲν καιρὸν αὐτὸς |
εἰδήσεις, ἐμοὶ δὲ καλῶς ἂν ποιήσαις συντάξας ὅπως ἂν παρα|δειχθῆι
μεθ᾽ ὧν οἰκίαν τε ἕξω καὶ τὰ λοιπὰ δέοντα, ἵνα μὴ πρεσ|βύτερος ὢν
ῥέμβωμαι.
But if for the present you direct me to do nothing, you will know when
it is the right time, but as for me please be kind enough to give instruc-
tions that it be indicated with whom I am to live and the other necessary
matters, so that, being an old man, I may not be left unattached.
κελεύω seems to function the same way as συντάσσω here in (13).

The writer is an older man, and the style shows touches of ‘better’ Greek
(μὲν … δέ, τε, ὅπως ἄν).
(14) PSI 4.420.1–19 (Philadelpheia, III BC, letter)
Ζήνωνι χαίρειν | Σεμθεύς. συνέτα|ξάς μοι κεραμεῦ|σαι ἅπαν‹τα›
τὸν κέ|ραμον ἕως τοῦ | ἰσιόντος ‹μηνὸς› ε τὴν | ἡμέραν ν. ποιῶ |
οὖν κατὰ ταῦτα. | ἐκελεύοσαν δέ με | καταβαίνοντα | συγχωνεύειν.
ἐγὼ | οὖν οὐκ ὠιχόμην, | ἕως ἄν μοι σὺ συντά|ξηις. νῦν οὖν ἄλλοι |
πάρισιν κεραμεῖς· καὶ | γὰρ ὁ χωνεύων με|μαλάκισται, ὁ ξένος. |
τί οὖν μοι συντάσσεις; | ὅπως ἂν οὕτω ποιῶ.
Semtheus to Zenon greetings. You ordered me to make pots and nothing
else until the 5th of the ensuing month, 50 per day. So I am doing that.
But they told me to go down and join in glazing. I have not gone until
you order me. Now other potters have come; and the glazer has fallen
ill, the foreigner. What do you order me, then? so that I may do that.
συντάσσω is what Zenon does, κελεύω describes the others’ action.

I detect a difference: συντάσσω is formal and polite in reference to the
authority figure, κελεύω is informal.
(15) PCol 4.66.1–9 (Philadelpheia, 256/5 BC, letter)
… Ζήνωνι χαίρειν. καλῶς ποιεῖς εἰ ἔρρωσαι. ἔρρω|μαι δὲ καὶ αὐτός.
ἐπίστασαι ὡς κατέλιπές με ἐν Συρίαι μετὰ | Κρότου καὶ ἐποίουν
πάντα τὰ προστασσόμενα τὰ κα|τὰ ‹τὰ›ς καμήλους καὶ ἤμην σο[ι]
ἀνέγκλητ[ο]ς. σοῦ δὲ προστά|ξαντός μοι ὀψώνιον διδόναι ἃ σὺ
συνέταξας οὐκ ἐδίδου | μοι οὐθέν. ἐπεὶ δὴ πολλάκ[ι]ς μου δεομένου
διδόναι μοι | ἃ σὺ συνέταξας οὐκ ἐδίδου μοι οὐθὲν Κρότος, ἀλλ᾽
ἐκέλευ|έμ με ἀπαλλάσσεσθαι, χρόνον μὲν οὖν πολὺν ἐκαρτέ|ρουν
σε προσδεχόμενος.
… to Zenon greetings. You do well if you are in good health. I also am
well. You know that you left me in Syria with Krotos and I did everything
that was ordered in regard to the camels and you found no fault with
me. But though you gave orders to give me wages/provisions, what you
ordered he did not give me any [of it]. When then, though I asked him
often to give me what you ordered, Krotos did not give me anything
but told me to leave off, I held on for a long time waiting for you.
This text (15) shows examples of προστάσσω as well as συντάσσω,

all for Zenon’s action without much distinction between them. κελεύω,
as in (14), is for the person cast as the bad guy.
4. CONCLUSIONS
Let us draw our conclusions for the LXX. If nothing else, we have a
clear explanation why the somewhat unexpected word συντάσσω makes
such a strong showing in the Pentateuch: it was a standard word in Ptole-
maic Egypt in the time of the translators.
προστάσσω needs little comment: it is one of the staple formal words,
but less frequently used than συντάσσω in the Pentateuch just as in con-
temporary sources.20
As to ἐντέλλομαι, while the contemporary evidence confirms that it
is alive and well, the high frequency in the Pentateuch (156, as against
22 in papyri) calls for explanation. One reason can certainly be discerned.
This verb is often (26 times) joined with ἐντολή, giving a neat etymologi-
cal combination that matches the Hebrew (‫)צוה מצוה‬. Most of these occur
in one book (Deuteronomy), which also has a very large proportion of all
the examples (85). Personal taste may well come into it: this translator, like
the translator of Genesis, strongly prefers ἐντέλλομαι to συντάσσω.21
Finally, κελεύω. The total absence from the Pentateuch remains a lit-
tle surprising, but my conclusion is that it was just not suitable in tone.
Formal-sounding words were felt more appropriate in the dignified con-
texts the translators were dealing with throughout the Pentateuch. They
availed themselves of those in use, notably συντάσσω, the characteristic
word of Ptolemaic officialdom and business.
This is merely a specimen of what the documentary evidence still has
to offer for illuminating the LXX. Further follow-up could certainly be
done even on these words. There is a need for such work to continue in
20
[προστάσσω is formulaic in petitions, for the King’s action in reponse to the petition;
see esp. PEnteux 2–112 (all III BC).]
21
Distribution in Pent.: ἐντέλλομαι: Gen 20, Exod 19, Lev 16, Num 19, Deut 85: total 156.
συντάσσω: Gen 2, Exod 39, Lev 12, Num 26, Deut 2: total 82.
order to place LXX lexicography on as sound a footing as possible, espe-

cially as translation and commentary work on the LXX gathers pace and
calls for more refined understanding of the vocabulary. Finding the best
way to collect (and store) the material systematically is the challenge for
the coming generation of LXX scholars.22
22
[The discussion of ‘order’ words here formed the basis of their treatment as specimens
of ‘officialese’ in Lee, Greek of the Pentateuch, 64–6.]
9
THE PRESENT STATE OF LEXICOGRAPHY
OF ANCIENT GREEK
20041
Abstract
The paper surveys and critiques the existing lexicographical resources for
Ancient Greek, considering the questions of coverage, dependence on predeces-
sors, and quality of the analysis of meaning. There is no single lexicon that covers
everything. The lexicon that comes nearest is the renowned LSJ, but it is primarily
a lexicon of Classical Greek. Others that attempt to cover the same ground are either
incomplete or dependent on LSJ. Lexicons of individual authors or bodies of text,
of which the NT is one, are numerous but variable in quality and dependent on
predecessors. The pattern of indicating meaning by the ‘gloss’ method pervades
the tradition, and has only been broken by Louw and Nida’s NT lexicon, employ-
ing definitions. In this it has been followed partially by the latest edition of Bauer’s
NT lexicon, BDAG (2000). Some limited suggestions are made for improving the
situation, directed mainly at the creation of a database to collect data on individual
words.
As long as Ancient Greek texts are read in the original, there will be a
need for lexicons, that is to say, tools which provide, first and foremost,
meanings of words, and also a certain amount of other helpful information,
on occurrences, collocation, and the like. The form that they take – whether
it is to be electronic or some other – is not an issue for the present pur-
pose: we can safely leave this to common sense and market forces. But what-
ever form they might have, there will be a need for tools which tell the
reader of a Greek text the meaning of words not already known. Lexicons
will continue to be produced to meet this need.
1
This paper was originally presented at the SBL Annual Meeting in Philadelphia in
November 1995, in the Biblical Lexicography Consultation. Updating and minor expan-
sions have been incorporated; the general thrust remains the same. Some characteristics
of oral presentation have been retained.
Not only will lexicography be in demand, but it will continue to carry

a weighty responsibility. This is because of the special character of lexi-
cons. Lexicons are regarded by their users as authoritative, and they put
their trust in them. Lexicons are reference books presenting a compressed,
seemingly final statement of fact, with an almost legal weight. The mere
fact that something is printed in a book gives it authority, as far as most
people are concerned. And understandably: if one does not know the mean-
ing of a word one is predisposed to trust the only means of rescue from
ignorance.
Yet this trust is misplaced. The concise, seeemingly authoritative state-
ment of meaning can, and often does, conceal many sins – indecision,
compromise, imperfect knowledge, guesswork, and, above all, dependence
on predecessors. Lexicographers have to make a decision and put down a
definite statement, and they are fallible like everyone else. But the ordi-
nary user has no means of knowing where the mistakes have been made,
where the ignorance has been covered up, what has been lifted from some-
where else without checking, and so on.
All this means that we will not only go on producing lexicons, but we
have a special responsibility to ensure that the lexicons we produce get it
right.
The main part of my paper will be a survey and critique of the existing
lexicographical resources for Ancient Greek. Ancient Greek covers Myce-
naean, followed by Classical, then Koine Greek (300 BC to 600 AD),
within which of course the LXX and NT find their place.2 The language
continues on into Byzantine and eventually Modern Greek. All this forms
a continuum, of a gradually evolving language, not a series of disparate
entities. This fact is important for the lexical study of Greek words: all the
evidence, from all these periods, is potentially relevant. For the purpose of
this survey, however, I am going only as far as Byzantine Greek.
The question I want to address is this: what lexicons or dictionaries are
available to provide guidance on meanings of words in all this vast body
of Greek texts? What sort of coverage is there? The ideal would be one
lexicon that covered everything. This we don’t have. The coverage is in
fact partial, unsystematic, and uneven in quality. To put it more bluntly:
there are gaps everywhere, and even those things that seem to have been
done have not been done as well as they could, and are in need of reas-
sessment.
2
The terms are simply labels, without a commitment to any theory or judgement about the
nature of these phases.
THE PRESENT STATE OF LEXICOGRAPHY OF ANCIENT GREEK 119
In saying this, my intention is not to denigrate everything that exists,

but to honestly assess the present situation, so that we can go forward. I
am very conscious of the centuries of work behind the resources that we
now have, and that it is only because of what our predecessors have done
that we can go on to build something better.
Before I turn to the lexicons themselves, there is one other general
point to be made, a most important one. As we all know, in the last few
years Ancient Greek texts have become available in electronic form. We
all know of the magnificent work done to produce the Thesaurus Linguae
Graecae (currently disk ‘E’) and the databank of documentary texts (cur-
rently found on the disk ‘PHI7’). And it is already common practice to
draw on these in word studies. But I don’t know if it has been generally
realised that this development brings about a major shift in Greek lexicog-
raphy. This resource is not just a useful extra for word studies; it means
that all available evidence not only can be, but must be used in any work
at an advanced level. This was not possible for earlier dictionary makers;
consequently all the existing lexical entries in all our dictionaries are now
obsolete and await reassessment in the light of the full evidence, or at least
checking to see if there is further evidence to be added.
II
Let us turn now to a survey of the present lexical resources. Given time
constraints it is not possible to include illustrations of the statements I will
make, but it can be taken for granted that supporting evidence is available.
We begin of course with LSJ. We all know and love LSJ. It has served
us well since 1843. We know it has some faults, but where would we be
without it? Actually its faults are much worse than most would suppose.
To sum them up briefly under three heads: it has no coherent method of
defining meaning, but relies on glosses; its basic material is derived from
predecessors, in some cases descending from the ancient lexicographers;
and the organisation is chaotic as a result of piecemeal revisions especially
in the ninth edition. All this is before taking into account new evidence
afforded by electronic searching. The LSJ Supplements (1968 and 1996)
could do no more than make further piecemeal changes.
All this has been well known to those who have had occasion to work
closely with LSJ and scrutinise its entries, but it is not generally recog-
nised, and most scholars continue to trust LSJ. Fortunately, criticisms have
now been expressed in print, particularly by John Chadwick. I refer you
especially to his 1994 paper, where abundant illustrations can be found;

he says bluntly: ‘It is about time that Greek scholars recognised the need
for a thorough overhaul of this indispensible tool.’3 Since then his book
Lexicographica Graeca has appeared (1996), offering many word-studies
that show how LSJ’s treatment needs improvement.
That is how things stand with what is our only general lexicon of Ancient
Greek. Moreover LSJ is basically a lexicon of Classical Greek and does
not pretend to cover the post-Classical material in a thorough way. And yet
this is the work on which we not only still rely heavily but which has been,
for generations, the resource from which everyone, including the authors
of other lexicons, has derived information. One can see its influence every-
where.
The only competitor to LSJ in attempting to cover the same ground is
the Diccionario Griego-Español (DGE), currently in progress.4 This shows
every sign of being an improvement on LSJ and of not simply copying
their material uncritically. It also makes a proper attempt to cover the post-
Classical period. But at the present moment it is not finished and will not
be in the lifetime of many of us. And there is a further point, which raises
a question that ought to be pursued at some time: given the size of the pro-
jects we are faced with, how many languages should we operate in in our
major reference works? Spanish speakers and advanced scholars will be
well served; how many others will use it?
Another work, with the title Μέγα Λεξικὸν τῆς Ἑλληνικῆς Γλώσσης,
which appeared some fifty years ago, attempts to cover Greek from Ancient
to Modern, but does little to remedy the situation. It is basically LSJ’s
material, with additions from various sources.
III
Everything else that remains deals with a special area. That is, all other
lexicons are focused on a particular text or set of texts, and draw on out-
side material, if at all, only to illustrate the words that occur in that text.
Let us take first the NT. It has a rich lexicographical tradition of its own
going back to the sixteenth century. Yet there are disturbing features to
be seen when one looks closely. It is not possible to go over this ground
in detail here, but the salient points are these: first, there is the legacy of
the long tradition of indicating meaning by glosses rather than definitions,
3
Chadwick, ‘Replacing Liddell and Scott,’ 10.
4
[Vol. 7, 2009 has reached ἔξαυος.]
which leads to many problems (as Louw and others have shown). Secondly,
there is the fact that even the latest lexicons derive their material from their
predecessors, and a great deal of it has been passed on uncritically over the
course of centuries. Thirdly, there is an aspect that I think is not well known:
meanings given in the NT lexicons are contaminated by glosses from the
standard translations, going back as far as the Vulgate. There is a fourth
tendency which has become evident to me lately: NT lexicons are unsys-
tematic in their control of other discussions, and may or may not take up
useful contributions to the understanding of the meaning. All this mainly
concerns the major lexicon series of our time, Bauer (1928, 6th ed. 1988)
and its offshoots in English.5
Louw and Nida’s 1988 work6 presented two major innovations: it
adopted a semantic domains approach, and it put into practice the method
of using definitions to indicate meaning. But the understanding of the
words is derived primarily from the existing tradition (mainly via Newman7),
and the other tendencies that I mentioned are just as evident, even if they
are under the surface.
BDAG (2000), whose reviser and editor is justly honoured in this vol-
ume8 for his indefatigable lexicographical labours over nearly half a century,
incorporates not only a host of minor revisions but a significant advance:
for the first time in the Bauer series, definitions in the manner of Louw and
Nida have been added for a majority of the words.
Nevertheless, the NT lexicographical tradition, although it is in an
advanced state compared with other areas, would benefit from a thorough
rethinking.
We come now to the LXX. Until now we have had only Schleusner dat-
ing from the 1820s,9 and we have been desperately in need of a new work
for more than a century. Well, at last the time has come. We now have not
one but two lexicons of the LXX. I refer of course to LEH, covering all
the LXX, and Muraoka’s lexicon to the Twelve Prophets.10 I am sure we
all feel grateful to these authors for having at last provided us with some-
thing. Obviously LEH, which is complete, will be the standard tool for
some time to come.
5
These observations are argued in Lee, ‘Hebrews 5:14.’ A fuller treatment is planned in
my book on the history of NT lexicography, currently in preparation. [Published 2003.]
6
Louw and Nida, Lexicon.
7
Newman, Concise Dictionary.
8
[Taylor et al. (eds.), Language and Lexicography.]
9
Schleusner, Lexicon in LXX.
10
LEH, Lexicon (2003); Muraoka, Lexicon (Twelve Prophets) (1993). [See postscript for
an update on Muraoka, Lexicon (2009).]
But we must look at it a little more closely and honestly. The most
obvious strength of LEH is in making full use of previous discussions
of LXX words outside lexicons. So it gathers up everything that has been
done in the past by way of preparation for such a lexicon. But there are
also significant weaknesses, which the authors themselves would readily
admit. Two major points stand out: (a) most of the meanings are taken
wholesale from LSJ; and (b) there has been no systematic gathering of
non-LXX parallels that might throw new light on the meanings. In other
words, it is based primarily on existing lexicons; and so we continue to
move around in this circle in which the faults of one lexicon are passed
on to the next.
IV
As regards the Greek documentary texts, i.e., the inscriptions and

papyri, that vast body of data that is still being added to and assimilated,
the attempt to provide any sort of lexical coverage is a struggle. There
is Preisigke’s lexicon, followed by various supplements by Kiessling and
others,11 all out of date or incomplete, and there is a sense of never being
able to get on top of it. The only sure way to tap into this material is to
work directly on the texts either in printed or in electronic form. But this
is daunting for the non-specialist.
The study of these documentary texts is one of the major challenges
of today. Their importance for illuminating the biblical texts was of course
recognised long ago, and my colleague Greg Horsley and I continue with
the attempt to produce a new Moulton and Milligan, that is, an up-to-date
exploitation of this evidence for the light it sheds on the NT vocabulary.12
The same task needs to be carried out for the LXX (as Lust and his col-
leagues well recognise).
But there is more to this documentary evidence than simply quarrying it
for parallels. To produce a proper lexicon of Ancient Greek, all of it needs
to be brought into conjunction with the evidence from the literary texts.
We are only at the beginning of doing this. The existing lexicographical
tradition is based on the literary texts, which have been known all along.
Documentary examples have been added piecemeal to the lexicons in recent
times, but that is not the same as combining all the evidence to arrive at a
complete lexical analysis of a word.
11
Preisigke, Wörterbuch.
12
See recently Horsley and Lee, ‘Interim Entries, 1’; Lee and Horsley, ‘Interim Entries, 2.’
So much for the documentary texts. What other lexical resources are
there? We must mention, however briefly, the many dictionaries focused
on individual ancient authors (Polybius, Josephus, etc.). They are all very
well, and of course are very useful in relation to the particular author,
but they obviously approach the language piecemeal rather than as a
whole. In addition one knows that LSJ will have been the main guide
to meaning.
What else? There is one other major work, Lampe’s Patristic Greek
Lexicon (1968). This rivals LSJ in size and is packed with mind-boggling
detail related to the Patristic texts. It is obviously an indispensible and
valuable tool. But its drawback is that it is intended to be used in com-
bination with LSJ. That is, all the material in LSJ is taken for granted;
Lampe aims only to add new material where necessary. In theory this
might sound all right, but in practice it is a recipe for endless uncertainty.
One simply cannot know the full range of meaning or usage, or even
words, actually occurring in the Patristic authors. Lampe’s lexicon, then,
must be recognised for what it is: a partial lexicon to a particular group
of texts. Nevertheless it is the only glimmer of light in a vast territory
that would otherwise be completely dark.
Lampe takes us some way into the Byzantine period, but for coverage
in time after that, what do we have? Not very much. A ‘lexicon of Byzan-
tine Greek’ has just begun to appear. The first part, published in 1994,
reached a point somewhere in alpha.13 Inspection immediately reveals
that this is not a full dictionary at all, but takes the form of a supplement
to . . . LSJ, and also Lampe. So it is rather like Lampe, a partial coverage
of a certain body of authors. Actually it seems to be mainly a list of addi-
tional words.14
Finally we have the lexicon of Kriaras, covering late Byzantine popu-
lar literature. The size of this valuable work gives some idea of what we
are dealing with here: there have been 14 volumes since 1968, reaching
παραθήκη in 1997. But again it is a question of trying to cover a certain
portion of the data, a preliminary to a full-scale lexical treatment.15
13
Lexikon zur Byzantinischen Gräzität (LBG) [completed 2017].
14
[The older lexicon of Sophocles, Greek Lexicon of the Roman and Byzantine Periods
(1887), ought to be mentioned here. It offers useful, but again partial, coverage of its
period.]
15
Kriaras, Λεξικό.
VI
This is as far as our survey needs to go, and it is time to bring things
together. The general picture, as I see it, is this. First of all, in terms simply
of coverage, there is only one general lexicon of the period, namely LSJ
– and it is primarily a lexicon of Classical Greek. Classical Greek is
covered well; post-Classical Greek has only been dealt with by focusing
on patches of the total area, and there are large areas still untouched.
Further, this process of working through the texts piecemeal, even if it
were to be completed, would still give us a series of lexicons, each dealing
with its own segment and, as it were, cut off from the rest. The essential
overview of all the evidence would still be lacking. I find it hard to judge
at this stage how much DGE, the Greek–Spanish lexicon, will alter this
situation.
Secondly, there is the problem of the quality of the coverage. It is sim-
ply a fact that what has been done so far cannot be relied on. This does not
mean that it is all badly wrong; it does mean that until a thorough check
has been done, both to eliminate the mistakes of the past and to use the
full resources now available, we cannot know for certain that what we find
in front of us when we look up a word is sound.
Even without doing any investigation, one knows that all is not well
from the fact that if you are reading almost any text of the post-Classical
period – sometimes even one from the Classical period – you will sooner
or later come across something poorly dealt with, or not covered at all,
by LSJ or any other available tool.
At this point one might well object that it is a question of practicalities.
The resources we have, and are likely to have, are governed by what is
possible, and lexicography is very slow and difficult work. I am, of course,
all too aware of this. Nevertheless, it seems to me important methodo-
logically that we recognise, and not lose sight of, the ideal, the true goal
to be aimed at, if we are to get things really right. And that is a complete
lexical treatment of Greek from its beginnings right through to the end of
Byzantine Greek (or better still into Modern Greek), making use of all the
occurrences, to ensure that all the evidence is brought to bear at once. The
whole history of a word hangs together, and it is quite misleading to think
in segments consisting of particular authors or groups of texts.
The practical difficulties of producing an actual lexicon done in this
way are of course very great, but we must recognise this as the only sound
method if we are to achieve a full lexical treatment of any word.
Let me add that I’m well aware that even if such an ideal lexicon is
achieved, there will still be a need for smaller tools that will provide con-
cise information – using glosses if need be – on particular authors and
periods of Greek. But behind the handy tools must stand the ideal complete
lexical treatment that has done the work thoroughly; the smaller works
should be derived from it, not stand in its stead.16
VII
In conclusion, let me turn to immediate goals. My suggestions may

seem modest after the far-reaching critique you have just heard. I am
not about to propose that we immediately begin work on a ‘Thesaurus of
Ancient Greek,’ as was suggested and fairly hastily abandoned early in
the twentieth century. I make only some limited suggestions, and hope
that others will be stimulated to think about these matters and come up
with proposals.
As regards the NT and the LXX, I think it should be recognised that the
work is not over. Now that LEH is complete, a thorough revision ought
to be undertaken. The NT is more difficult, because existing lexicons
are generally regarded as the last word. Nevertheless all is not well with
the NT lexical tradition, and long-term plans for a complete overhaul are
needed.
As to LSJ, we all shrink from suggesting a major revision, knowing
how huge the task will be. Nevertheless, sooner or later something must
be done. The availability now of LSJ in electronic form (in the Perseus
digital library) is a useful step, but it is what we do next that will be impor-
tant. Are there any plans for re-analysis of the entries? Will new examples
be sought via the TLG and added? The idea of having an LSJ database on
line for anyone to add to or comment on is worth exploring. But unless
there was some ongoing sifting, not to mention reorganisation of the accumu-
lating material under each word, we might be no better off. Everyone would
enjoy adding to such a database, but they would not enjoy using it.
16
An instance of the reverse pattern is seen in Greek etymological studies. We have mas-
sive, exhaustive, and authoritative reference books on Greek etymology – Frisk, GEW
and Chantraine, DELG – but for most people they are virtually unusable. The practi-
cal tool derived from them, a ‘Concise Etymological Dictionary of Ancient Greek’ is
yet to be produced. [There is now also Beekes, EDG, similar to the others in size and
scope. The need for a concise etymological dictionary remains: cf. concluding remarks
in Essay 20, ‘Etymological Follies.’]
I have two smaller and more immediate proposals. First, I suggest that
we need to establish a world database to collect the results of all published
work on individual NT words. This would be not just a bibliography, but
a concise summary of the result of each study. It would have a lexical
focus, and be linked to words, not passages. I am suggesting that this would
be initially for the NT alone, but it could eventually cast its net wider.
Secondly, a matter which is in itself trivial but causes major difficulty
and frustration, the lack of a consistent system of reference, in particu-
lar the abbreviations, for all the texts that we deal with. There is no one
method of abbreviation even for the books of the NT. I doubt if anything
can be done about them, but for the rest a plea for common sense may
be worth expressing. We need to set to work to establish standard abbre-
viations for all authors and texts in Greek. In the case of literary texts the
excellent TLG Canon provides an authoritative list of authors and their
works, but does not offer any abbreviations; here I think a valuable oppor-
tunity has been missed.
As far as documentary texts are concerned, I can report that Horsley and
Lee have produced a list of abbreviations for volumes of Greek inscrip-
tions, which at the moment have no agreed system. This has appeared in
Epigraphica for 1994, and we hope it is the first step towards a consensus,
matching the standard list of Oates (et al.) for the papyri.17
To conclude. Clearly there is much to do; perhaps it seems to be too
much. On the other hand, it is good to know that there remains interesting
work to be done, if scholars are willing to take it on. And I for one am
optimistic that in the long run even the most daunting objectives will be
achieved. For the present, if we do nothing else, we can at least recognise
the true state of affairs in Ancient Greek lexicography, and be cautious.
Postscript
Though there have been some useful developments, the situation remains funda-
mentally unchanged since this paper was written in 2004. A new lexicon purporting
to cover Classical and later Greek (i.e., GE, translated from GI) appeared in 2015.
It represents little advance on LSJ, from which it is ultimately derived (see Essay 26).
The very recent Cambridge Greek Lexicon (2021) ‘is designed primarily to meet
the needs of modern students’ of Classical Greek (p. vii). While it offers a new
presentation of meanings, it is confined to a mainly Classical selection of authors
and cites no references beyond authors’ names. Its usefulness remains to be tested.
17
Horsley and Lee, ‘Preliminary Checklist.’ Since the presentation in 1995, BDAG has
adopted this list. [Also the SBLHBS, which describes it as an ‘indispensable checklist’:
see 99–100.]
In the NT field, the Diccionario Griego-Español del Nuevo Testamento (DGENT),

employing a carefully-designed definition method, began in 2000 and has reached
the end of beta (in 2012). (Cf. remarks in Lee, History, 165–6 with n. 13.) García
Santos, Diccionario del Griego Bíblico (2011), covering both the LXX and NT,
is useful and fairly comprehensive. It employs a mixed method of stating meaning
but is discerning in its sense-divisions. The major NT lexicons (BDAG, BRAA,
LN) continue as they were.
In the LXX field, a significant advance has been achieved in the publication
of Muraoka’s A Greek-English Lexicon of the Septuagint (2009), with complete
coverage of the LXX. This lexicon applies the definition method as fully as pos-
sible, with some mixing of methods. All the LXX data have been analysed afresh.
While there are some flaws, and further areas to be covered (see Essay 15), this
lexicon is the best guide to the LXX at present and the sound foundation for all
future development of LXX lexicography. For further commentary and update on
LXX lexicons see Aitken, No Stone Unturned, 9–15.
Documentary texts, both papyri and inscriptions, have become available in elec-
tronic, searchable form online (via <papyri.info> and <epigraphy.packum.org>)
and the databases are constantly being updated. Access to the texts and fast search-
ing have become easy, but the lexicographical tools to understand the texts have not
kept pace, in fact have not changed at all. The researcher has to rely on the same
printed works (LSJ; Preisigke, Wörterbuch) as before. Translations (for papyri,
not inscriptions) are intermittently provided in the electronic databases, but are not
entirely reliable.
My suggestions for a database have been developed further in Lee, History, 182–
3 and ‘Releasing LSJ’ (Essay 17); see also ‘Lexicographic Database’ (Essay 13).
There is no sign yet of action to make such a thing a reality.
10
REVIEW OF MURAOKA, A GREEK–ENGLISH
LEXICON OF THE SEPTUAGINT:
CHIEFLY OF THE PENTATEUCH AND
THE TWELVE PROPHETS (2002)
2004
Abstract
This extended review of the second instalment of Muraoka’s Lexicon first sets
it in its context among other lexicons, then describes its general features, illus-
trated by sampling of entries. The method of indicating meaning is next exam-
ined more closely, with extensive examples. The use of the ‘definition’ method,
as opposed to that of the ‘gloss,’ is an important and welcome aspect of the
Lexicon, and is generally applied with success. The point is made here that a
‘one-word definition’ is an acceptable means of defining, provided the distinc-
tion between it and a gloss is maintained. Some flaws and inconsistencies in
Muraoka’s treatment are noticed and suggestions made for improvement. The
term ‘mixed method’ is introduced, to characterise a method by which defini-
tions and glosses are used together to state meaning, without clear distinction of
one from the other. Muraoka is not alone in falling into this pattern: it is found
in some well-known major lexicons. Muraoka’s use of explanatory additions and
his approach to collocation are also examined. In the final assessment this lexi-
con is considered to be a good lexicon, which offers the prospect of development
into the standard lexicon of the LXX. (Essay 15 is a review of the complete
lexicon. It is intended to be read in the light of the present one.)
It is given to few scholars to compile a lexicon, to even fewer to com-

pile one to the LXX. Compared with the lexicography of the NT or of
Ancient Greek generally, this has been the most neglected of fields. Until
the late twentieth century, lexicons devoted entirely to the LXX, from
the elementary effort of Rosenbach in 1634 to Schleusner’s classic of
1820–21, could be counted on the fingers of one hand. The reasons for
the long gap after Schleusner in modern times are not difficult to discern,
but need not detain us here. The LXX was certainly read, or at least con-
sulted, and there was great interest in it as an adjunct to other disciplines.
But those who ventured into it had to rely on the guidance of the general
lexicons of Greek, notably Liddell and Scott, with some help from the
NT lexicons and their partial coverage of LXX material. That this was the
cause of much inadequate if not mistaken understanding of the Greek of
the LXX cannot be doubted.
By the mid-1980s the pressure to address the problem had built up
greatly and more than one project was begun or mooted. A kind of scrum
developed, from which two players emerged. The competition can only
have been beneficial, however much one might have preferred to see, in
an ideal world, a single major project combining all resources. As so often
in the history of lexicography, individuals simply set to work. Around
the same time as Lust, Eynikel, and Hauspie began their LXX lexicon
project in Leuven, Muraoka began work in Melbourne on a lexicon to one
portion of the LXX, the Twelve Prophets. This was no small objective in
itself, but was achieved by 1993.1 It was, as it turned out, only the prelude
to a much larger effort culminating in the present impressive lexicon. In
the new lexicon the earlier material on the Twelve Prophets is expanded
to cover the Pentateuch, together with, for a significant proportion of
the words, coverage of the whole LXX. With the Pentateuch included,
the tipping-point has been reached and the lexicon is now a major tool
for work on the LXX, even though of course much still remains to be
covered.
It is known that Muraoka intends to continue work along the same
lines and add treatment of further LXX books, perhaps all of them, until
a full lexicon of the LXX is achieved.2 This step-by-step approach is
both acceptable and commendable, given the size of the task of producing
a major lexicon of the LXX providing full data on every word. The all-
too-familiar alternative, of a project that sets out to cover everything and
makes a start with a fascicle reaching a point somewhere in alpha, then
takes another forty or fifty years to complete, is not inviting.
1
Muraoka, Lexicon (Twelve Prophets). The present reviewer’s contribution to that lexicon,
generously acknowledged in the introduction, was made at an early stage, and ended well
before the lexicon reached completion; the subsequent work on the Pentateuch has been
entirely Muraoka’s own.
2
Cf. Muraoka’s entry in ‘Record of Work Published or in Progress’ in BIOSCS 35 (2002),
33: ‘Next phase started. Finished Isaiah, now working on Jeremiah.’ Professor Muraoka
has kindly informed me (by email 11 Jan, 2005) that he has now covered about 70% of
the whole LXX.
REVIEW OF MURAOKA, A GREEK–ENGLISH LEXICON OF THE SEPTUAGINT 131
II
Muraoka’s lexicon uses the traditional method of presentation, namely

a list of headwords in alphabetical order, each accompanied by an entry
that gives the meaning(s) of the word along with occurrences and a selec-
tion of other information. There are 4,478 headwords, of which 1,553 are
treated for the whole LXX. The layout is pleasing and easy to read, with
large print and conspicuous divisions. In each entry the following infor-
mation appears: headword and its morphology; the sense(s) of the word,
accompanied by full (though not necessarily exhaustive) listing of occur-
rences (many quoted and translated); range of application and syntactic
breakdown; a list of semantically related words; brief references to sec-
ondary literature; a summary of the corresponding words in the Hebrew
original. A set of special symbols is used (eleven in all) to indicate various
things, for instance, that data are complete for the whole LXX. It must be
admitted that the symbols are not easy to remember and most users will
tend to ignore them, but they are a useful system for supplying information
that may be wanted at some point.
The great virtue of this lexicon is its offering of a full lexical analysis of
each word. All the occurrences in the books covered have been examined
and classified semantically. Splits into different senses are made when
required, each of the senses is defined, and under each a thorough if not
complete list of occurrences is given. Along the way one gets, as already
noted, an indication of the various applications of the word and the syn-
tactic relations it enters into.
The framework thus established is very valuable not only for the books
covered but also, in general, as a guide to the meaning of the word in any
LXX occurrence not yet treated. It will also be the basis for the expansion
of the lexicon to cover the whole LXX, when that is undertaken. There
are inevitably gaps at present where a word happens to be used in a differ-
ent sense in some book(s) not yet dealt with, and of course the lexicon can
give no assistance with words that are not yet entered at all. But as far as
it goes, and that is a long way, it provides an extremely good guide to the
meaning of the LXX vocabulary.
How good are the lexical analyses? My impression is that they are
generally sound; I can detect few faults. Muraoka has adopted, at least in
principle, a definition method of indicating meaning, avoiding the problems
created by glosses, and ensuring a reliable breakdown and description
of the senses. (More will be said on this topic below.) That Muraoka has
tackled this task in a language not his first is highly commendable; the
reader cannot fail to be impressed by the results. There are times when
one might quibble over details of English expression, but this is not a
serious problem: the intended meaning is always sufficiently clear.
On the question of taking into account evidence from Greek outside the
LXX, that essential adjunct to deciding meanings, Muraoka’s approach
is as follows. First, he has relied primarily on what has so far been gath-
ered into the major reference works or noted in special studies of the
LXX vocabulary: he has not attempted any fresh searching, as he himself
explains (p. ix). This is an acceptable approach, given that further search-
ing and assessment, even of a limited kind, would have slowed progress
immeasurably. It must not be forgotten, however, that the existing col-
lection of evidence on each word is not necessarily complete, and further
collecting needs to go on, especially from the documentary sources con-
temporary with the LXX. But this is work for the future. Secondly, none of
this evidence is reported in the entries (with some exceptions). This too I
think is the right way to proceed: it is simply impossible to give an adequate
summary within the constraints of the printed page. It is also better for the
reason that what is reported in lexicons tends to become authoritative and
immutable (like the meanings). Such data will be better assembled system-
atically, over a longer time, in an electronic database, as I have argued
elsewhere in relation to NT lexicography.3 Some mention of parallels does,
however, slip in, in a very sporadic way.4 I am not sure that this is a good
thing, though one can understand the temptation to mention some particu-
larly apt parallel. The risk is that it will be taken to be definitive.
In regard to secondary literature, Muraoka has made use of existing
studies as much as possible, but in this case the material does gain a men-
tion in the entries, though somewhat sparingly. Muraoka explains (p. xvi)
that mention in the entry implies that the study has made a ‘substantive
contribution’ to determining the semantics of the word concerned. This
makes good sense, and avoids the problems created by simply listing works
that may or may not have anything useful to offer. But it can be unclear
whether a particular article or discussion has been omitted because it has
nothing to offer, or because it has been missed. Again, a collection in elec-
tronic form that was systematic and ongoing could keep track of everything,
and solve the problem of keeping up to date. But this too is for the future.
3
See Lee, History, 182–6.
4
So, e.g., s.v. ἀγαπητός an example from Homer is quoted; s.v. ἐκλύω 2 one from
Aristotle; s.v. ἵστημι I.3 one from Demosthenes; s.v. ἐλπίς 2 we are told that ‘this
sense is also attested by Polybius’; s.v. ἐγχάσκω a parallel collocation of εἰς with χάσκω
in Philostratus is noted; and s.v. ἔλεος Aristotle’s definition of the word is quoted fully
(in translation).
A comparison with the other current lexicon of the LXX, LEH, is inevi-
table.5 First it needs to be said that LEH is compact, easy to use for a
quick indication of meaning, and above all complete for the whole LXX
(as things stand at present). It is also, paradoxically, quite thorough in
its collection of references to secondary literature. But what LEH gains
in completeness it loses in brevity of analysis. Though full statistics
are given, the selection of occurrences for each word is small, and most
important, there is no systematic analysis and breakdown of the senses.
A series of glosses is offered, mostly derived from LSJ. These cannot
adequately represent the range of meanings found in the LXX. In short,
LEH is ideal as a smaller lexicon, with a different aim from Muraoka’s.
It was a welcome holding effort produced at a time when there was noth-
ing but LSJ and Schleusner, and such it will remain. But the future major
lexicon of the LXX will need to be built on a different and better founda-
tion, and Muraoka’s work is that foundation.
III
I offer a sample of words to show the character of the work and illus-
trate some of the points mentioned.
(1) λόγος. Muraoka makes seven divisions in his analysis, as follows:
1.[a.]6 word spoken or uttered.
b. of divine message communicated to a human or humans [sic
plain text].
2. report, news, rumour.
3. esp. pl., a chain of connected events, ‘a story, an account.’
4. what is or may be a good reason.
*5. course of action, step to be taken.
6. matter under discussion.
Under each of these senses, references and brief quotations are given,
for example under 6 we have ‘ἐν τῷ ~ῳ τούτῳ “in this matter” De 1.32;
a legal case, 22.20.’ The number of references ranges up to 24 in the case
of sense 1. It is a little disconcerting to realise, however, that all of the
references under sense 1 are to the Twelve Prophets and none to the Pen-
tateuch, though of course the meaning occurs there. The references for this
sense are simply continued unchanged (with one exception, the correction
of Hos 1:3 to 1:2) from the 1993 lexicon, and no Pentateuch references
5
LEH, Lexicon (1992, 1996); LEH, Lexicon (2003).
6
[a.] added by me for clarity; Muraoka mostly omits this ‘a.’
have been integrated into them. There is no great harm in this; but it could
mislead the user into thinking there are no occurrences in the Pentateuch.
The use of the asterisk, seen here in sense 5, ‘signifies that the word is
not attested earlier than the LXX Pentateuch, i.e. the third century B.C.E.’
(p. xiv). The purpose is to alert us to a possible innovation in the Greek
of the Pentateuch. Muraoka is careful to add that ‘words so marked do not
have to be neologisms created by Septuagint translators.’ Besides new
words, Muraoka marks previously unattested meanings in the same way,
as in the case of λόγος 5. It seems likely that this sense is indeed a neolo-
gism and has arisen from stereotyped rendering of the Hebrew word (‫)דבר‬,
but Muraoka does not foreclose on the question: by the use of the asterisk
he simply points out the absence of earlier evidence for this sense and
allows the user to draw a conclusion (or not).
In the analysis of λόγος, the divisions seem well taken and show sub-
tlety in detecting differences not obvious at first sight. The division at
1.b however gives one pause. Is this a separate lexical meaning, or simply
a difference of reference under the one lexical meaning word spoken or
uttered? Comparison with other entries suggests that these divisions usu-
ally indicate differences of reference, syntax, or collocation, not a new
lexical meaning, and that would seem to be what is intended here; but some-
times they do offer a separate lexical meaning, as, e.g., in the case of δίδωμι
2.d ‘to entrust for temporary safe-keeping.’7 This is a matter that needs
clarification and a consistent policy.
As to completeness of coverage of the meanings of λόγος, one might
wonder where Exod 18:19 (καὶ ἀνοίσεις τοὺς λόγους αὐτῶν πρὸς τὸν
θεόν) is intended to fit. The meaning required in the context seems to be
‘matter in dispute,’ something different from sense 6 and its cited occur-
rences (even Deut 22:20, which seems to me to belong in sense 1). This
example highlights the problem of bulk that the lexicographer constantly
faces. Muraoka has dealt with the 110 or so occurrences of λόγος in the
Twelve Prophets and the Pentateuch as best he can, perhaps missing
even some of these, but to cover the whole LXX there are still over 1000
to go!
(2) κάτεργον. The meaning given, operating costs (for Exod 30:16;
35:21), has evidently benefited from the evidence of the word in documen-
tary sources: LSJ give examples only in III BC papyri and the Pentateuch,
7
Other examples like δίδωμι s.vv. ἄρχω 3.b; αὐξάνω I.b; ἐκτός 3.b; ἐπικαλέω B.1.a, b,
c, d; κινέω A.b, c. I do not find this explained in the relevant section of the Introduction
(xiv–xv).
offering the glosses wages … labour-costs … service (sic for Exod 30:16;
35:21). But we also see the difficulty of keeping track of sources and
coverage of the evidence. Muraoka refers to Wevers on Exod 30:16 as
his source (where there is no mention of papyri); the similar discussion in
Le Boulluec and Sandevoir is not noted (where there is a brief mention);
none of these writers refers to LSJ, which must be their ultimate source
for the meaning. A full examination of the evidence is presumably still to
be done: LSJ cite only five papyri; there are in fact up to 120 occurrences,
50 of them in III BC texts. Such an investigation is likely to lead to a more
refined understanding of the word’s meaning in the time of the Pentateuch
(e.g., does it refer to wages strictly, or more broadly to expenses including
wages?)8
(3) παιδίον. Connotation, a feature difficult to handle in the tradi-
tional method of lexical description, comes into the picture for this word.
Muraoka’s solution here is first to define (young child) and give the usual
breakdown of occurrences, then add a separate note: ‘The feature of endear-
ment and affection appears to be present in some cases: e.g. … [examples].’
Such notes do not appear often. In the case of τέκνον there is a different
treatment. Four senses are given first with normal definitions, then comes
‘5. an endearing address to a youth: … Ge 43.29 (Joseph to Benjamin).’
Strictly speaking this implies that there is no semantic content, only con-
notation, which may be right. But what of the two examples in 1.b? These
fall under the definition 1. immediate offspring, with only the following
description: ‘b. voc. without a posses. pron.: τί ἐστιν, τέκνον; “What is the
matter, child?” Ge 22.7; 22.8.’ They miss out, somewhat unexpectedly, on
any special connotation. The whole subject is clearly one well worth future
debate and experiment.9
(4) λαμπάδιον. Muraoka has changed his mind on the form of the head-
word: it was λαμπαδεῖον in the 1993 lexicon and is now λαμπάδιον,
for what reason one cannot deduce. The only secondary literature men-
tioned in both editions is Walters, Text (1973). There is no reference to
8
Wevers, Exodus, 496. Le Boulluec and Sandevoir, L’Exode, 309. LSJ shows the impact
of the documentary discoveries early in the twentieth century: in LS 8th ed. (1897)
κάτεργον was cited only from the Pentateuch, with the meaning work. Preisigke’s some-
what baffling contribution in Wörterbuch (1925), s.v. will also need to be assessed:
‘1) Arbeitslohn … 2) Leistungsmaß, Arbeitsnachweis … 3) Lehenbeartungsdarlehen. …’
9
The term ‘connotation’ is used here as defined in Zgusta, Lexicography, 38–41, not as
Caird uses it in ‘Towards a Lexicon. I,’ following John Stuart Mill. [Zgusta (38): ‘We
can describe connotation as consisting of all components of the lexical meaning that add
some contrastive value to the basic, usually designative function.’]
Boyd-Taylor’s lengthy and penetrating discussion, bearing on the mean-

ing as well as the form of the word.10 This may have appeared too late to
be taken into account, but will be relevant to future treatment. The exam-
ination of the LXX vocabulary is young and ongoing: almost every word
is potentially on the brink of further elucidation by a new study.
(5) διαστέλλω, διαστολή. Muraoka has a good analysis of the tricky
verb διαστέλλω (but aren’t they all?), rightly recognising a sense 4. to
state precisely in Gen 30:28; Lev 5:4 bis, along with five other senses.
But things seem to have gone wrong with the occurrence of the noun in
Num 30:7, where a parallel nominal sense such as precise statement seems
required by the context (and where the Hebrew has the noun correspond-
ing to the verb in Lev 5:4 bis). Muraoka has placed this example of δια-
στολή under 2. express orders or instruction, along with Num 19:2 (where
it is right), but this does not make proper sense in Num 30:7. Muraoka
refers to Caird’s old but decisive discussion, and to the more recent one
of Dorival. The latter seems to have put him off course. Dorival (and his
colleagues) saw three possible senses, but could not come to a decision
between them and opted for the translation disposition, ‘qui peut avoir ces
trois sens.’ This stratagem is hardly helpful to the lexicographer trying
to decide the lexical meaning. Fortunately Muraoka has ignored another
discussion of Num 30:7, that of Wevers, which could only have created
deeper confusion.11
This instance could serve as a model of the peculiar demands of pre-
paring an entry in a LXX lexicon. To reach a result, one must assess the
entries in the major lexicons (the splendid confusion of LSJ, partially cor-
rected in later works; the entry in the very recent DGE, which in the case
of διαστολή is six or seven times the size of LSJ’s); make sense of the
now generous but still somewhat raw commentary tradition; seek out other
studies and try to absorb their conclusions; search for and assess, as far
as one can, the evidence of the word in texts both literary and documentary
outside the LXX, not trusting older collections that may be out of date;
take into account the Hebrew equivalent in the source text; notice paral-
lel formations and passages; compare the meaning of semantically related
words; and check for equivalent words used in the context (in this case
ὁρισμός is relevant). All this can be a confusing – and time-consuming –
10
Boyd-Taylor, ‘Evidentiary Value’ (2001).
11
Caird, ‘Towards a Lexicon. I,’ 467–8; Dorival, Les Nombres, 513; Wevers, Numbers,
497–8. Yet another meaning appears in DGE, vol. 5 s.v., where Num 30:7 slips in at
the end of διαστολή C II 2 glossed as compromiso (= ‘commitment’?).
mixture to absorb. In the end, the only guide through the maze is one’s
own sense of the likely meaning in the context. Muraoka is as aware as
anyone of these demands, and has coped with them to great effect. But no
one should imagine that there is no work left to be done.12
IV
I turn now to a closer look at the definition method employed in

Muraoka’s lexicon. My remarks are not intended to amount to a negative
assessment of the work as a whole but made with a view to possible
improvement in future editions, and as further reflections on the current
debate over method in stating meaning in Greek lexicons.
Muraoka describes at the outset (pp. xii–xiii) the method he intends to
apply. Instead of giving a ‘translation equivalent,’ the problems of which
are succinctly summarised, he has ‘chosen to go for definition, to describe
senses of a given word in sentence form or as fragments of a sentence.’
A definition is formated in italics; occasionally a translation equivalent
will be added, in plain text within single quotes. So far all is clear. The
definition method will be used, and a definition is understood to have a
form involving the use of more than one word. Anything else will be a
translation equivalent and formated differently. But then Muraoka adds
this statement: ‘Where there is no or little likelihood of misunderstand-
ing, we have given translation equivalents: e.g., τροϕός wet-nurse instead
of woman employed to suckle another’s child.’ This step causes confu-
sion. It means that a translation equivalent (or ‘gloss’) will sometimes
take the place of a definition proper, but will be formated in italics like
a definition, even though it is still a gloss. It seems that the only way we
will know it is a gloss and not a definition is that it will be a single word,
since a definition is a ‘sentence’ or ‘fragment of a sentence’ and a gloss
is by implication one word. But the user will have difficulty maintaining,
or seeing any point in, the distinction. In effect what Muraoka is saying is
that a single word can serve as a definition in certain cases. It would have
been better to formulate it that way. To me this is a perfectly acceptable
12
I do not enter here into the issue of the theoretical basis of LXX lexicography, the focus
of Boyd-Taylor’s stimulating paper (‘Evidentiary Value’). I would say only that in my
opinion the ultimate criterion for deciding the meaning is the intention of the translator,
who works with Greek as he knows it and as he expects readers and hearers of his own
time to understand it. We cannot, of course, know the intention of the translator: we
can only make the best deduction from the indications available.
method of defining in a Greek–English dictionary for English speakers,

provided there is no ambiguity or lack of clarity in the chosen word, as in
the case of Muraoka’s example, wet-nurse, which needs no further unpack-
ing. When such a statement of meaning is used, it is to be regarded, not as
a translation equivalent or gloss, but as a one-word definition.13
In practice however Muraoka does not keep consistently to his stated
method. From what has been said in the introduction, we expect to find
each word or sense defined by means of a word or phrase – one of either –
formated in italics; a translation equivalent in plain text and quotation
marks might sometimes be added. This would be the strict definition
method. Such definitions do occur, as in:
δίκαιος 2. conforming with set and agreed standards:
δικαιόω *1. to pronounce innocent, ‘to acquit’:
ἐϕοδιάζω to furnish with supplies for a journey:
ἔτος year:
θηλυκός female:
πεῖρα 1. attempt: … 2. testing: … 3. experience:14
But what are we to make of the following?:

καύχημα that which makes one proud, object of pride:
μακρόθεν 2. at a distance, at a remove:
παράϕρων wandering from reason, out of one’s mind:
πορεύομαι II.3. to conduct oneself, follow a certain moral life-style:
In these we have two definitions, each one, presumably, helping to

explain the other. In the same way, we often find the meaning stated by
means of not one, but two or even three single words, as in:
θέλημα desire, interest:
λογίζομαι 4. to reason, consider:
λογίζομαι 5. to deem, consider:
παίγνιον plaything, toy:
περιστερά dove, pigeon:
ἐπιθυμέω to desire, long for, covet:
καλός 1. advantageous, beneficial, desirable:
καταϕρονέω to despise, belittle, treat with contempt:
ὀπή opening, hole, crevice:
13
[Italics added, 2021.] Cf. the discussion in Lee, History, 22. Anne Thompson, in her
review (2003) attributes to me (114) the term ‘definition gloss’ for such a one-word
definition. I did not use it in the book, and would be reluctant to accept it: I think it is
liable to be misleading.
14
Colons are retained from the original, where they mark the end of the statement of mean-
ing, before other material begins.
πεδίον plain, open land or field:

ποιέω 4. to fashion, construct, manufacture:
πόνος 1. toil, suffering, hardship:
πορεία 2. journey, going, marching:
Are these really one-word definitions? That is, in Muraoka’s terms,

translation equivalents that have replaced a definition because there was
‘no likelihood of misunderstanding.’ It seems unlikely that they are.
Rather, as the use of multiple equivalents suggests, they are in fact old-
fashioned glosses that do not have the character of definitions.
In addition to these, we find cases in which there is a combination of
phrase and single word; that is, apparently, definition plus gloss, but with
both in italics, not just the definition:
δορά skin of body when taken off, hide:
ἐκεῖ 1. there, in that place already mentioned:
καταλαμβάνω 1. to lay hold of, seize:
πλῆθος 1. large quantity, multitude:
σιωπή silence, abstinence from speech:
Let us take stock at this point. What we have seen shows that glosses,
the traditional method of indicating meaning in lexicons, have maintained
their hold in Muraoka’s lexicon and are being used along with definitions,
in a variety of combinations. This is a mixture of methods, and Muraoka
is not alone in using it. The mixed method, as I think it deserves to be
called, turns up often elsewhere. It appears at times in a predominantly
definition lexicon like the OLD, as in:
ingens 1 Of very great size or dimensions, huge, vast.
insanus 1 Of unsound mind, demented, frenzied, mad.
Conversely the DMLBS, which is primarily a gloss lexicon, at times

mixes the methods, as in:
gratanter 1 with joy or pleasure, thankfully.
improvisus 1 appearing or occurring without warning, unexpected.
The mixed method also appears occasionally even in the OED, the
premier definition lexicon and pioneer of the method. For example:
glad 1. Bright, shining, beautiful.
subside 1. intr. To sink down, fall to the bottom, precipitate.15
15
The mixed method may be seen in Chadwick, Lexicographica Graeca, e.g., 87, 100, 207,
261. It is to be noticed that the mixed method is not used in BDAG (2000): definitions
are always typographically distinguished from glosses, even though the two may occur
together. The strict definition method is also followed in Louw and Nida, Lexicon. My
Can the mixed method work? Many would say yes, and one would
have to concede that no great harm is done, e.g., by defining περιστερά
as dove, pigeon. The user gets, as it were, two hits of information which
together, presumably, are clearer than one – though one could say, even
in this simple case, that we are just being given an older and a newer word
for the same thing and there is no need for both. Even so, no great harm
is done, nor perhaps in πόνος 1. toil, suffering, hardship, or σιωπή silence,
abstinence from speech. But we should be aware of the hazards of this
method when close attention to the meaning is needed in exegesis or for
other purposes in reading a text. Take the case of ἐπιθυμέω to desire,
long for, covet, quoted above. It matters very much what exactly this word
means in Exod 20:17 and Deut 5:21 (οὐκ ἐπιθυμήσεις τὴν γυναῖκα τοῦ
πλησίον σου. οὐκ ἐπιθυμήσεις τὴν οἰκίαν. …). Simply glossing it in this
way (the glosses, incidentally, are the same as LSJ’s) fails to resolve the
question of what it actually means, in particular whether it implies sexual
desire. The English words could have more than one meaning as well as
differing from one another, so all sorts of possibilities are left open by
stating the meaning in the form of these three glosses. The exegete may
or may not be aware of this, and may proceed to decide the meaning on
the basis of the English words as he or she understands them.16
But this is not by any means the end of the variations that Muraoka
allows himself in the statement of meaning. A frequent and significant ele-
ment, though one on which no information is offered, are words in plain text
joined with the definition or gloss in italic. Examples:
ἀϕαιρέω 1. to remove and cause to disappear:
βρῶμα 1. food, not cooked:
γερουσία assembly of elders as a decision-making body in ancient Israel:
διαβιβάζω to cause to cross from one side to the other:
ἐπέχω 1. to wait without proceeding to next action:
ζωή 1. life, vitality, inclusive of non-physical dimension:
ζωογονέω 2. to preserve alive without killing:
θῦμα slaughtered animal:
κατάλοιπος left remaining untouched, unharmed or not mentioned:
κατανοέω to observe closely to find out about:
κῆτος huge sea-fish:
own thinking and practice have undergone development: the analysis of συνίστημι in
Lee, ‘Συνίστημι,’ 3–4, shows the same faults I am criticising in Muraoka.
16
The meaning of ἐπιθυμῶ is a crucial question for Loader’s discussion of the Decalogue
in Sexuality, 5–25, but he seems unaware of the inadequacy of a gloss like ‘desire.’ He
would get no help from Muraoka. Notice further that ‘desire’ also appears in Muraoka’s
meanings for θέλω 1. to desire, wish: and βούλομαι to wish, desire: What does ‘desire’
mean there and is it the same as in ἐπιθυμῶ?
κόσυμβος decorative fringe of a garment:

κράσπεδον tassel on the four corners of an ancient Israelite’s outer
garment:
λανθάνω to escape notice of, mostly of inadvertent omission on the part
of the beholder:
λοιπός left and remaining out of a given number or quantity:
πειρασμός putting of character to the test:
πένθος grief, esp. over the dead:
σαλπίζω to blow a trumpet:
The formating suggests that these are explanatory additions not intended
to be part of the definition, but amplifying it in some way. Yet many of them
do seem to be part of, indeed an essential part of, the definition proper.
So, for example, in θῦμα slaughtered animal, where it is surely clear that
animal alone is not sufficient to define the word. Similarly κῆτος is prob-
ably not just any sea-fish, but one that is of a specially large size, or huge.
On the other hand, some of these examples may have to do with indicating
the collocation, that is, the word(s) with which the defined word is or may
be combined in context. Such may be the case in πειρασμός putting of
character to the test, though it is hardly clear.
Sometimes the explanatory addition is bracketed, as in: τηρέω to watch
(to attack at an opportune moment). There are also instances in which the
whole definition is in plain text. At first one might think these are simply
the result of a formating slip, but I am not so sure: some of them look like
explanatory material that has invaded the whole definition. Examples:
ἁϕή 2. part of body infected by leprosy:
γραμματεύς 2. a person in leadership position of one sort or another:
[contrast 1. one who records]
καρδία 3. seat where human thoughts, intentions and attitudes are
generated and take shape:
λόγιον 3. woven piece of an ephod placed on the breast worn by the
high priest and used to arrive at legal decisions by means of
δήλωσις καὶ ἀλήθεια, ‘oracle box’:
ὄνυξ 3. a precious stone, ‘onyx’:
I note also instances in which definitions include bracketed explanatory

words that are also italicised, for example:
ἐπιθύω to offer (heathen, cultic) animal sacrifice:
κοιμάομαι 1. to lie (down):
οἰκογενής 2. belonging to the (royal) household:
Finally we come to collocation and how it is treated by Muraoka. The

best and clearest method, as it seems to me, is the one seen in examples
like these:
ἡγέομαι 1. to direct (organisation, group of individuals):

μανθάνω 1. to learn (a skill, art, habit):
ὀρθρίζω 2. to seek and turn in eager anticipation (to sbd, πρός τινα):
The collocation is in plain text enclosed in brackets, and not part of

the definition itself. But we find numerous variations, with the brackets
sometimes omitted, the collocation sometimes in italics, and elements of
definition sometimes treated as if part of the collocation (as in μολύνω):
ἀϕαιρέω 2. to shift from place X to Y:
εἰσακούω 1. to give ear to sbd or sth:
ἐξαποστέλλω 5. to part with, give a send-off to guest or visitor:
εὐλαβέομαι 2. not to venture to do, to refrain from doing out of a sense
of awe:
καλύπτω 1. to place sth over an entire surface:
λατομέω 1. to dig a cavity in the ground:
λογίζομαι 7. to put down to sbd’s account:
μολύνω 2. to violate (a woman sexually):
περιαιρέω II. mid. to remove from oneself or sbd close to oneself:
περιζώννυμι I. act. to put (garment) round sbd’s body:
In the last two features discussed, I do not see any fundamental problem.
In the case of collocation it is only a matter of applying consistent for-
mating. As to the explanatory elements of definition currently formated in
plain text, further thought is needed. The simplest solution would be to
change all of them into italic format as being part of the definition, if they
can be so regarded. If not, at least an explanation of their function needs
to be formulated and a consistent policy maintained in their use. The ques-
tion of the ‘mixed method’ is more difficult. It would help at least if the
distinction between a one-word definition and a translation equivalent or
gloss were maintained. Further revision along the lines of conversion of
glosses into definitions (multi-word or one-word) and removal of super-
fluous glosses would be a larger enterprise.
It has been truly said that it is not possible to make a good lexicon
out of a bad one. In the case of Muraoka’s lexicon the possibility exists of
making a good lexicon even better. These remarks are offered as a con-
tribution to that end.
11
DIMITRIOS DOUKAS AND
THE ACCENTUATION OF
THE NEW TESTAMENT TEXT OF
THE COMPLUTENSIAN POLYGLOT
2005
Abstract
This paper first notices the puzzled descriptions of the accentuation of the
Greek NT text of the Complutensian Polyglot (1514), then goes on to point out the
unobserved fact that the accentuation follows a monotonic system almost exactly
the same as that now in use in Modern Greek. Next is considered the informa-
tion on the matter in the preface to the volume. The Greek text of the preface is
presented with English translation and notes. A number of current misconceptions
are dealt with. The question of the identity of the inventor of the accentuation is
then explored in full. The evidence in favour of Dimitrios Doukas as editor of
the text and author of the preface is summarised and augmented. The paper then
argues that it was he who conceived and applied the system of accentuation.
Possible other sources of the idea are considered and eliminated. Finally the ques-
tion of who might have been behind the initial intention to print an unaccented
text is discussed.
1. DESCRIBING THE ACCENTUATION
The fifth volume of the great Complutensian Polyglot, dated 1514,

contains the Greek text of the NT, the first printed (though not the first
published) edition. As such it has been of enormous importance, and the
subject of scholarly scrutiny, ever since. The authorship, the nature of the
text, the manuscripts on which it was based, and the typography have all
received close attention, and the questions associated with them have
been to a great extent unravelled. But the accentuation used in the NT text
has been observed with some puzzlement, even disapproval, and not pur-
sued further.
Metzger’s standard work on the text of the NT offers this statement:

[The Polyglot New Testament] is printed without rough or smooth
breathing marks and is accented according to a system never heard of
before or since: monosyllables have no accent, while the tone syllable
in other words is marked with a simple apex, resembling the Greek
acute accent mark.1
The equivalent work of the nineteenth century, that of Scrivener, though

fuller, is in similar vein (and also clearly Metzger’s main source2):
The Greek type in the other volumes is of the common character, with
the usual breathings and accents; in the fifth, or New Testament volume,
it is quite different, being modelled after the fashion of manuscripts of
about the thirteenth century, very bold and elegant … without breath-
ings, and accentuated according to a system defended and explained
in a bilingual preface πρὸς τοὺς ἐντευξομένους, but never heard of
before or since: monosyllables have no accent, while in other words
the tone syllable receives the acute, the grave and circumflex being
discarded.
In a subsequent note, on the laudatory verses included in volume 5,

Scrivener quotes approvingly those of ‘the native Greek editor, Demetrius
Ducas,’ in which full accentuation and breathings are displayed, and adds
this comment: ‘the fantastic mode of accentuation described above was
clearly not his work.’3
Another comment, itself somewhat fantastic, is found in a 1953 paper
by Tasker:
The regular system of breathings and accents, the bane of all printers
of Greek, is not used; but acute accents are inserted in an unparalleled
manner on the tone syllable in words of more than one syllable, as
though Spanish and not Greek was being printed.4
1
Metzger, Text of the NT, 97. An earlier, shorter version of this paper was presented at a
Language Colloquium in Memory of G. P. Shipp at Macquarie University, 16 May, 2003,
organised by Trevor Evans. I thank Vrasidas Karalis for his reactions and helpful infor-
mation at an early stage. I am especially grateful to Michael Curran, who read a late draft
of the whole article and put my arguments to the test by many acute observations. I am
indebted to a number of others for assistance, as mentioned in the appropriate places.
2
Along with Tregelles, Printed Text: cf. his statement (p. 10) that the editors say ‘that
they have marked the tone-syllable of each word with a simple apex.’ The ‘simple apex’
in turn derives from simplex tantum apex in the Polyglot NT preface (Latin version).
3
Scrivener, Plain Introduction, 2:177, 178 n. 1. Scrivener was quite wrong about the
authorship of the ‘fantastic mode of accentuation,’ as we shall see.
4
Tasker, ‘Complutensian Polyglot,’ 204. The fact that this was originally ‘a lecture given
to the Spanish Department of King’s College London and others’ may go some way to
explain Tasker’s facile remark.
DIMITRIOS DOUKAS AND THE ACCENTUATION OF THE NEW TESTAMENT 145
Even Geanakoplos, a much more sensitive observer, had only this to

say in 1962:
Anticipating the reader’s questions as to the differences in the Greek
type utilized in the printing of the Old and the New Testament volumes,
the editor carefully explains why it was chosen to employ Greek char-
acters without accents. (Actually it would be more accurate to say that
a peculiar system of accentuation was employed, the acute accent alone
being used and only on polysyllabic words, while aspirates were entirely
omitted.)5
Basil Hall, a year later, uses the same key word ‘peculiar’ to character-
ise the system:
The accentuation of the Greek was peculiar: it was justified by the
editors on the ground that it formed no part of the genuine text and was
absent from older manuscripts; and no ‘breathings’ were provided (yet
the editors gave the normal accentuation to their text of the Septuagint).6
Most recently (1990) M. A. Screech offers, in passing, this puzzling

remark:
When we turn to the beautiful text of the Complutensian New Testa-
ment (with its absence of breathings and its accents designed to show
tonic stress not modern Greek usage) we read. …7
In these discussions the writers have attempted to describe the accen-

tuation in the NT text of the Polyglot without an appreciation of its true
nature. It is in fact a monotonic system, one that is almost exactly the same
as that now in use in Modern Greek, which was officially introduced in
1982.8
At this point it will be helpful to see a specimen of the Polyglot text.
An illustration showing the beginning of the Gospel of John (1:1–14)
is given below. This short extract is sufficient to demonstrate the main
5
Geanakoplos, Byzantium, 240. Whether peculiar is intended to mean ‘strange, odd’ or
simply ‘all its own’ is not clear; either way it is inappropriate.
6
CHB, 58–59. For Bentley in 1983 (Humanists, 92) ‘the Greek type is a bit odd’ in its
omission of ‘all breathing marks and most accents’; he further speaks of ‘the inconve-
nience caused’ by the omission (to whom?). The following report the bare facts: Gregory,
Textkritik, 2:925; Nestle, Introduction, 2; Darlow and Moule, Historical Catalogue, 2:4.
There is no remark in Aland and Aland, Text of the NT, 3–4. Tregelles in 1854 had also
used the word peculiar: ‘In their preface, the editors refer to the peculiar manner in which
they had printed the Greek’ (Printed Text, 10). The views of Proctor and others regarding
the type are discussed below.
7
Reeve and Screech, Erasmus’ Annotations, xv (Screech).
8
For a full statement of the system, see, e.g., Triantaphyllides, Γραμματική: Αναπροσαρμογή,
22–26; Holton, et al., Greek, 38–40.
characteristics of the system and to confirm that it is indeed monotonic,

as the term is applied today. That is, a single accent mark, the acute, is used,
and it is employed to mark the stress in words of more than one syllable.
Further comment on features of the system as seen in the Polyglot will then
be made.
The text of John 1:1–14 in vol. 5 of the Complutensian Polyglot

(Alcalá, 1514). Image from the holdings of the Biblioteca Nacional de España.
2. FEATURES OF THE ACCENTUATION
The mark of accentuation is clearly the traditional acute, not some-

thing resembling it but in some way different, as implied by Metzger. It
is indistinguishable from the acute used, for instance, in the NT lexicon
at the end of volume 5, with full accentuation.9
9
Longer and shorter varieties of the acute accent mark are detectable. In the preface
it is nearly always the longer, rarely the shorter; in the NT lexicon it is the other way
round; in the NT text I have not detected any of the longer form (in a very small
sample).
Monosyllables are generally not accented, but sometimes they are. In the
specimen text we have ο λόγος σάρξ εγένετο, with an accent on σάρξ.
I have investigated this phenomenon by extensive sampling. My finding
is that while the great majority of monosyllables are not accented, a sub-
stantial minority are. Those that are accented are mostly nouns and verbs,
occasionally other parts of speech. There is no evident system. I interpret
this phenomenon as reflecting the actual pronunciation of the phrases con-
cerned, in which the accented word, though monosyllabic, would in fact
have a stress on it and not be pronounced like an unemphatic monosyllable
(mostly the old proclitics and enclitics such as εις, το, μου). The person
responsible for marking the accents intuitively placed a mark on a stressed
monosyllable from time to time, contrary to what the system provided,
because oral realisation of the text was too strong to ignore; but it was
done inconsistently and no regular system for dealing with such cases
was developed. This flexibility in accentuation of monosyllables, though
unsystematic, parallels to some extent the Modern Greek system, where
in specified cases monosyllabic words are written with an accent.10
Some examples of accented monosyllables in the Polyglot text:
όπου θέλει πνεί (John 3:8)
μέλλοντες πλείν (Acts 27:2)
οίνω ολίγω χρώ (1 Tim 5:23)
συ τις εί (John 1:19, 8:25)
είπον ουν αυτώ τις εί (John 1:22)
συ εί ο διδάσκαλος (John 3:10)
ηλίας ει σύ (John 1:21), but ο προϕήτης ει συ (1:21)
τον βούν αυτού (Luke 13:15)
ου ϕιμώσεις βούν αλοώντα (1 Cor 9:9), but βουν αλοώντα ου ϕιμώσεις
(1 Tim 5:18)
υμείς εστέ το ϕώς του κόσμου (Matt 5:14), but το ϕως των ανθρώπων
(John 1:4, in the specimen passage)
είδε ϕώς μέγα (Matt 4:16)
λευκά ως το ϕώς (Matt 17:2)
και χείρ κυρίου ην μετ αυτού (Luke 1:66)
αυτού ϕλόξ πυρός (Rev 19:12)
πρίν η συνελθείν (Matt 1:18)
πού έθηκαν αυτόν (John 20:2), but που έθηκαν/ας αυτόν (20:13, 15)
πού σοϕός (1 Cor 1:20)
εκ της ρούθ (Matt 1:5)
10
See Holton, et al., Greek, 38–9. Modern Greek in fact accents πού the interrogative, as
in the Polyglot samples below; but not σαρξ or the others.
The handling of enclitics also calls for comment. The practice of the
editor(s) is to ignore the special rules of accentuation of enclitics, and treat
them as ordinary words. Disyllabic enclitics, then, invariably have an acute
on the second syllable, regardless of what precedes, and words before
monosyllabic enclitics never receive an additional accent on the final syl-
lable. In the former case the resulting markings, though mostly contrary
to the enclitic rules, may well coincide with pronunciation in practice.
So εν εισίν (1 Cor 3:8), σαρκικοί εστέ (3:3), άγιος εστίν (3:17), υμείς
εστέ (1:30), οποίον εστί (3:13), των ανθρώπων εστί (1:25), εγώ μεν
ειμί (1:12). But the total absence in the latter case is more unexpected. It
means that in instances like απέστειλε με (1:17), το κήρυγμα μου (2:4),
πνεύματι τε (4:21) the accent marks failed to help the reader place the
second stress where it occurred in pronunciation, assuming it did so then
as it does today, i.e., απέστειλέ με, etc. (There can be no question that
the editor(s) knew the relevant rules: see, in the preface below, γνωμώνιόν
τι (l. 35), ἔδοξέ τι (l. 39), and the accentuation throughout the LXX text
of the Polyglot.) One must suppose that the editor(s) regarded this as a
refinement that was not worth attempting in a reduced system of accent
marking.11
Besides the acute accent, one other diacritic is used, the diaeresis. This
is exactly as in the modern system. A few examples: ησαΐας (John 1:23),
βηθσαϊδά (1:44), νοΐ (1 Cor 1:10), γάϊον (1:14). The iota subscript is
omitted altogether, as can be readily seen from the opening words of
the sample, εν αρχή. This is not a new step, but was already standard
practice at the time, even in texts written or printed with full polytonic
accentuation.12
What we have in the NT text of the Polyglot is a fully thought-out and
effective system of monotonic accentuation, consistently applied apart
from occasional variations in the treatment of monosyllables. It is not
some unheard of, as if to say outlandish, idea without a place in the his-
tory of the language and unworthy of serious notice.
11
For modern survival and writing of the accent pattern in απέστειλέ με, etc., see Holton,
et al., Greek, 30–1.
12
As, e.g., the preface to the NT and the LXX text in vols. 1–4. The letter υ often appears
in the NT text with a small dot between the prongs. This is clearly not a breathing. I take
it to be an accidental feature of one of the sorts of upsilon in the Polyglot font. It occurs
not only initially but in all positions; it is even found with an accent added above (see,
e.g., Παύλος in 1 Cor 1:13). I notice the same sporadic dot in eta and omega. This then
will account for Tregelles’ assertion that breathings are omitted ‘except in the case of Υ’
(Printed Text, 10).
3. THE ACCENTUATION IN HISTORICAL PERSPECTIVE
Let us now try to put this remarkable feature of the Polyglot into his-
torical perspective. The official introduction of the monotonic system of
accentuation in 1982 was preceded by a long period of experimentation
and debate in the nineteenth and twentieth centuries. Its competitor was
the traditional polytonic system inherited from late antiquity, which had
been preserved all through into the modern period and was the prevailing
system, with some modification, until that time; it can still be seen today
in formal writing. Debate over the accentuation (and use of breathings,
iota subscript) formed part of the larger language question, that is, what
the modern language actually was – or ought to be – and what its orthog-
raphy ought to be. The first attempt at reform of the traditional method
of writing the accents appeared in 1814 with the work of Y. Vilaras, who
published texts without any accents and breathings. By 1900 a monotonic
system essentially the same as that in use today had been proposed by
D. Melandinos, which gathered supporters as the century progressed, cul-
minating in the result we have indicated. Such a system is practical, eco-
nomical, and adequate, removing superfluous distinctions while not aban-
doning accent marking entirely.13
The movement towards a revised system of accent marking begins in
modern times no earlier than the nineteenth century. The Polyglot NT thus
anticipates by 300 years the earliest modern proposal for change, and by
over 450 years the establishment of a full monotonic system. This is not
to suggest a connexion, but to observe the surprising but apparently unno-
ticed fact that someone in the early sixteenth century arrived at exactly the
same solution as modern thinkers on the subject.
In its own time the accentuation of the Polyglot NT was, as far as I can
discover, quite without precedent. Other early printed Greek texts show
the traditional system, or, less frequently, complete omission of markings.14
13
For a historical summary of the question, on which I rely here, see Argyriades, Νεοελλη-
νική Γλώσσα, 154–9. A specimen of Vilaras’s orthography may be seen in Triantaphyl-
lides, Γραμματική 1: Ιστορική Εισαγωγή, 443–4. I am grateful to Elizabeth Kefallinos
of Macquarie University for drawing my attention to the former work.
14
Numerous illustrations may be seen in: Proctor, Printing of Greek; Scholderer, Greek
Printing Types; Morison, Politics and Script; and Barker, Aldus Manutius. Barker’s col-
lection of 46 illustrations includes script as well as type of the period: none show any-
thing but the polytonic system. See also, e.g., the facsimile edition of Lascaris’s grammar
(Amsterdam, 1966); Geanakoplos, Byzantium, 206; Staikos and Sklavenites, Πεντακόσια
Χρόνια, 11–17. For texts without any accent or breathing marks, see esp. Proctor, fig-
ures 2–4, 6, 9, 10, 13, 20, 26, 28, 31, 35, 38. Some of these are the earliest attempts at
printing Greek, with a few words of Greek embedded in Latin text. On fig. 31, which is
The specimen page of a polyglot projected by Aldus in 1504 uses the

traditional system, and is therefore no forerunner of the Polyglot in this
respect.15 Besides this evidence, the fact that the accentuation needed to be
explained and defended confirms that it was a novelty. In the Compluten-
sian Polyglot itself, the full polytonic system is used in the Greek text of
the Septuagint in volumes 1–4, in the NT lexicon at the end of volume 5,
and in all the other places where Greek is printed, including some where
the NT font is re-used.16 As for MSS of the NT, the general picture is
well known: older MSS, such as the great uncials, are unaccented (unless
markings are added in a later hand, as in Codex Vaticanus), but by the
ninth century both uncials and minuscules are regularly adorned with full
polytonic accentuation and two breathings. A similar situation applies in
the case of MSS of other ancient literary texts.17
The Polyglot accentuation, then, was an innovation in its time. We are,
however, not without information on the thinking behind it. As mentioned
in the quotations above, there is a preface on the subject in the work itself.
I propose now to reproduce this preface in full, with a translation, since
not all of it has been adequately or correctly appreciated – in fact a number
of gross misinterpretations persist – and no one has yet, to my knowledge,
made a translation of it in its entirety. I shall examine its contents closely
to see what it reveals about the accentuation, and then go on to consider
the question of the identity of its inventor.
of special interest, see below. Some may also be seen in Scholderer, facsimiles 2, 3,
6, 19 (= Proctor, fig. 31).
15
A facsimile can be seen in Renouard, Annales, 390–1 (I am grateful to Erik Hamer for
checking this for me). Cf. Geanakoplos, Byzantium, 246 n. 89.
16
See below for a list of the last-mentioned. Doukas’s editions of the Erotemata, etc. and
of Musaeus, Hero and Leander (both Alcalá, 1514), published in the same year but after
vol. 5 of the Polyglot, use the Polyglot NT font but with standard accentuation. Illus-
trations of the Erotemeta, etc. may be seen in: Geanakoplos, Byzantium, 241; Legrand,
Bibliographie, 1:118; Proctor, Printing of Greek, pl. 24; of the Musaeus in: Thomas,
Sixteenth-Century Printing, pl. 12; Scholderer, Greek Printing Types, fig. 24; Morison,
Politics and Script, pl. 171. (Scholderer, followed by Morison, gives the date of the
Musaeus as 1510: this seems to be a mistake.)
17
For illustrations of NT MSS, see, e.g., Metzger, Manuscripts. On accentuation, p. 12.
A great variety of texts and MSS from III AD to XVI AD can be seen in Wilson, Mediae-
val Greek Bookhands, showing accentuation at various stages of development, from
none through incomplete to full polytonic. Instances of partial accentuation do not show
anything approaching a monotonic system. Further illustrations in: Cavallo and Maehler,
Greek Bookhands, and Turner, Greek Manuscripts. Cf. Turner’s summary (11) of accen-
tuation in his sample of MSS (late IV BC to VI/VII AD): ‘Accents are rarely written in
prose literature, still more rarely in private letters. … But they are likely to be used fairly
frequently in texts of lyric verse, especially in verse in difficult dialects. …’ The earliest
occurrence of accents so far known is from II BC. The establishment of the polytonic
system was a lengthy and somewhat complicated process reaching finality in IX/X AD;
for a summary, see Schwyzer, Grammatik, I, 373–5.
4. THE PREFACE
4.1 Introduction
The NT volume of the Polyglot, that is, volume 5 (1514), opens with a
lengthy preface in Greek, entitled Πρὸς τοὺς ἐντευξομένους, filling one
folio page and four lines of a second. It is followed at once by a transla-
tion into Latin. The heading of the latter makes clear that the Latin is a
translation, not the original: Precedens Greca prefatiuncula in latinum
versa. Ad Lectorem. We are thus spared the need to establish this point.
The Latin version (but not the Greek) is printed also in volume 1 (1517),
with the heading Prologus in nouum testamentum: et de causis quare in
eo apices graeci sunt praetermissi. Ad lectorem. This text matches the
one in volume 5 very closely, with some minor differences in orthography
and very slight differences in substance.18 Differences between the original
Greek and the Latin translation are another matter, to be noticed in pass-
ing as we proceed.
4.2 Text of the Preface

Πρὸς τοὺς ἐντευξομένους.
῞Ινα μὴ θαυμάσῃς ὦ σπουδαῖε ϕιλολόγε μηδὲ σχῇς πρὸς ἡμᾶς
ἀηδῶς εἰ ἐν παρούσῃ τῆς νέας διαθήκης γραικοτυπώσει ἑτερο-
τρόπως ἢ ἐν τῆς παλαιᾶς τὰ στοιχεῖα μόνα ἄνευ τῶν πνευμάτων
5 καὶ τῶν τόνων ἐντετυπωμένα ἐξεδόθη, προύργου νενομίκαμεν
τὴν τοῦ πράγματος τούτου αἰτίαν πᾶσιν κατὰ τὰ πρῶτα ἐμπεϕα-
νίσθαι. ἔστι δὲ καὶ τοιαύτη.
τοὺς ἀρχαιοτάτους τῶν ἑλλήνων χωρὶς τούτων ἐν τοῖς χαρα-
κτῆρσι κορυϕῶν γράϕειν εἰθισμένους, σαϕέστερόν ἐστιν ἢ πολ-
10 λῶν δεῖσθαι μαρτυριῶν. καὶ γὰρ δηλοῖ τοῦτο ϕανερῶς παλαιά τινα
κοὐκ ὀλίγα τῶν ἀντιγράϕων, οἷον καλλιμάχου ποιήματα καὶ τὰ
σιβύλλης ἔπη, καὶ πεπαλαιωμέναι ἐν τῇ πόλει λίθων γλυϕαὶ
μόνοις ἁπλῶς γράμμασι ἐγκεχαραγμέναι. ὥστε πάντως εἶναι πρό-
δηλον, ῥαβδίσκων τούτων καὶ κεραιῶν ὑπερθέσεις ἐν τῇ πρώτῃ
15 ἐκείνῃ τῆς γλώττης ἑλληνικῆς γενεσιουργίᾳ μὴ ἐπινενοημένας,
μήτε πρὸς ὁλόκληρον τῆς αὐτῆς ϕωνῆς τελειότητα πανταχόθεν
συντετακυίας.
18
Differences in substance: heading (above); the word hac is omitted from in hac noui
testamenti greca editione to suit relocation to vol. 1; rursus is replaced by rursum (once).
Differences of orthography: mainly more frequent use of ae for e; th in authoritas is
changed to t (twice). There are differences of punctuation, esp. more frequent use of the
colon (:) in vol. 1.
20 ἐπειδὴ δὲ καὶ πᾶσαν τὴν νέαν διαθήκην τοῦ κατὰ ματθαῖον

εὐαγγελίου καὶ τῆς πρὸς ἑβραίους ἐπιστολῆς δεόντων, ἑλληνικῇ
ἀπ᾽ ἀρχῆς διαλέκτῳ ὥσπερ καὶ ὑπὸ τοῦ ἁγίου πνεύματος ἐχρη-
ματίσθη, συγγεγράϕθαι πάντες ὁμολογοῦσιν, ἔδοξε καὶ ἡμῖν
ἀρχαίαν ἐν αὐτῇ τῆς αὐτῆς γλώττης παλαιότητα καὶ μεγαλο-
25 πρέπειαν ὁσίως διατηρεῖν, καὶ τὴν βίβλον ἄνευ ὁπωσοῦν ἐλα-
χίστων προσθηκῶν, ἀρχαίων γραϕῶν δίκην ἐκδιδόναι, ὡς μὴ
ϕανῶμεν πρᾶγμα οὕτως δή τοι ἱερὸν καὶ αἰδεσίμης μεγαλοϕρο-
σύνης πλῆρες κατ᾽ ἀλλοτρίας αὐτῷ προσεπιβεβλημένας καὶ
νέας ἐργασίας ἀνακαινοτομῆσαι. πρὸς δὲ τούτοις εἰ τἀληθὲς
30 λεκτέον ἐστί, ἡ τῶν πνευμάτων καὶ τόνων ἔνδεια τοῖς ὁσονδή-
ποτε ἐν τοῖς ἑλληνικοῖς λόγοις γεγυμνασμένοις οὐδὲν πρόσκομμα
ἐπιϕέροιεν ἄν. λέγω δὲ πρὸς τὴν εἰλικρινῆ τῶν λεγομένων
ἔννοιαν ἀναϕέρων.
οὐ μὴν ἀλλὰ ἵνα μή τις ἀπορήσῃ ἐν ποίᾳ τῶν συλλαβῶν προ-
35 σήκει τὸν τόνον ἐϕαρμόζειν, ἁπλῆ μόνον ἐν ταῖς πολυσυλλάβοις
λέξεσι κεραία προσηρτήθη. οὐ μέντοι τόνος αὐτὴ ἑλληνικὸς
ὑπολαβέσθω, ἀλλὰ γνωμώνιόν τι καὶ σημεῖον ὑϕ᾽ οὗ ἀπευθύνοιτ᾽
ἂν ὁ ϕιλομαθής, ὥστε μὴ διαμαρτάνειν ποτὲ ἐν τῇ ἐκϕορᾷ καὶ
εὐρυθμίᾳ τῶν λέξεων. ἐν δὲ τῆς ἀρχαίας διαθήκης ἑλληνικῇ
40 ἐκδόσει ὅτι αὐτὴ μετάϕρασις παντῇ που ἀλλ᾽ οὐ πρωτοποίητος
σύνταξις οὐκ ἔδοξέ τι ἐκ τῆς κοινῆς ἐν τῷ γράϕειν συνηθείας ἢ
ἀϕαιρεῖν ἢ ἐναλλάξαι.
ἐπειδὴ δὲ οὐ τοῖς πολυμαθέσι μόνον ἀλλ᾽ ἅπασι καθόλου τοῖς
περὶ τὴν ἁγίαν γραϕὴν σχολάζουσι συντελεῖν ἡ βίβλος αὕτη
45 πεϕιλοτίμηται, ἐϕ᾽ ἑκάστῃ τῶν λέξεων γραμματίδια ῥωμαϊκὰ
κατὰ στοιχεῖον ἐπετέθη, ἃ ἐπιδεικνύει τὴν ἀμϕοτέρων πρὸς τὰς
ἑτέρας καταντικρὺ κειμένας ἐπάλληλον σύγκρισιν, ὅπως οὐδὲν ᾖ
πρὸς τὸ σϕάλλεσθαι ἐνδόσιμον τοῖς μαθητιῶσι καὶ μήπω εἰς
ἄκρον τῶν ἑλληνικῶν ἀϕιγμένοις. ἔτι δὲ ἐπείπερ ἐγχωροῦσιν
50 ἐνίοτε λέξεις ἑλληνικαὶ πολυσχιδεῖς δοκοῦσαι ἔχειν καὶ ἀμϕι-
βόλους καὶ ἄλλας τινὰς διχογνώμους σημασίας, διὰ σπουδῆς
ἡμῖν ἐγένετο σημειοῦν καὶ τοῦτο, ὑποκειμένῃ στιγμῇ ῥωμαϊκῷ
γραμματιδίῳ ὑπὲρ τὴν ἑλληνικὴν καθεστηκότι λέξιν.
καὶ ἵνα παύσωμεν προοιμιάζοντες, κακεῖνο τὸν ϕιλομαθῆ μὴ
55 λανθανέτω, οὐ ϕαῦλα ἡμᾶς οὐδὲ τυχόντα ἐπὶ τῇ ἡμετέρᾳ ἐντυ-
πώσει ἐσχηκέναι ἀντίγραϕα, ἀλλ᾽ ἀρχαιότατα καὶ καθόσον
οἷόν τε ἦν ἐπηνορθωμένα, καὶ δὴ καὶ κατὰ τὴν παλαιότητα
οὕτως ἀξιόπιστα, ὥστε μὴ πείθεσθαι αὐτοῖς, πρὸς δυσκόλου εἶναι
τὸ παράπαν καὶ βεβήλου. ἃ δὴ καὶ αὐτὰ ὁ ἁγιώτατος ἐν χριστῷ
60 πατὴρ καὶ κύριος ἡμῶν ὁ μέγιστος ἀρχιερεὺς λέων δέκατος τῇ

ὁρμῇ ταύτῃ συλλαμβάνειν προθυμούμενος, ἐκ τῆς ἀποστολικῆς
βιβλιοθήκης ἀγόμενα, ἔπεμψε πρὸς αἰδεσιμώτατον κύριον τῆς
ἱσπανίας καρδηνάλιον, οὗ χορηγοῦντος καὶ κελεύσαντος τὸ παρὸν
βιβλίον ἐτυπώσαμεν.
65 ὑμεῖς δὲ οἱ τῆς παιδείας ἐρῶντες θεῖον τουτὶ καὶ ἱεροπρεπὲς
τοὖργον νεωστὶ ἐντετυπωμένον μετὰ πάσης δέχεσθε προθυ-
μίας. κἂν τοῦ χριστοῦ τοῦ κυρίου καὶ θεοῦ ἡμῶν μιμηταὶ ϕανῆ-
ναι καὶ τοῖς ἔργοις γενέσθαι σπουδάζετε, ἴστε ὅτι οὐδεμία ὑμῖν
ἔτι πρόϕασις λοιπὴ τὸ μὴ τῇ θείᾳ γραϕῇ συνομιλεῖν. οὐκέτι
70 ἀντίγραϕα διεϕθαρμένα, οὐ μεταϕράσεις ὕποπται, οὐκ ἀπορίαν
τῆς ἀπογράϕου ϕράσεως ἀπορεῖτε. τὸ θέλημα μόνον τὸ ὑμέτε-
ρον καὶ προθυμίαν προσδεχόμεθα. ἧς μὴ ὑπολειπούσης, ἀνα-
ντιρρήτως ἔσται οὕτως. τῆς γὰρ γλυκύτητος τῶν θείων λόγων
γευσάμενοι, τὰς λοιπὰς τῶν ἐπιστημῶν μακρὰν χαίρειν ἐάσετε.
75 εὐτυχεῖτε, καὶ τὸν καινουργισμὸν τῆς ἡμετέρας ἐργολαβίας μὴ
ἐκϕαυλίσητε.
4.3 Notes on the Greek text

The original is in general very accurate in regard to spelling, accentuation,
punctuation, and word-spacing. Some errors in these are clearly the typeset-
ter’s, not the author’s. I have attempted to keep changes to a minimum consist-
ent with present-day practice and comprehensibility. The punctuation is that
of the original, with a change in two places only, as noted below (lines 26, 47;
Legrand made the same changes in his printing of the text in Bibliographie,
1:115–7). There are no capital letters except pi, used once: I have retained this
feature. The original is printed in a solid block without paragraphs: I have intro-
duced them.
There appears to be only one place where the spelling is in error: line 37
γνωμώνιον, of which the attested form is γνωμόνιον. This is likely to be a mor-
phological as much as phonological slip: the writer could have been misled by the
association with γνώμων as well as the phonetic identity of ο and ω. Cf. εὐγνομό-
νως in the extract from Doukas’s preface to Plutarch, Moralia (1509; see below),
but εὐγνωμόνως correctly in his preface to the Erotemata, etc. (1514; Legrand,
Bibliographie, 1:119).
Throughout the text I have made two general changes, not individually noted.
Iota subscript, not used at all in the original, is here introduced where required.
Second, in initial diphthongs the original regularly has the breathing, or combi-
nation of breathing and accent, placed over the first vowel, in accordance with a
common (though not universal) practice at that time and still sometimes found
long afterwards: I have moved these to the second vowel. Similarly, in the (rare)
instances where an accent is placed over the first vowel of an internal diphthong
I have moved it to the second. All other changes are noted.
As far as I know, the text has been reproduced only once before, by Legrand in
1885 (Bibliographie, 1:115–7). Legrand’s text differs from mine in a few places,
the most important as follows: he inserts an article twice (lines 2, 62), alters to
τ[οι]ούτων in line 8 and corrects the spelling of ἐπηνορθωμένα in line 57. These
I regard as improvements beyond necessity. Legrand also omits the words τοῦ
χριστοῦ in line 67, presumably by accident.
In the following, the text I have printed above is given first, the original (or
Legrand’s emendation) second.
2 ῞Ινα] Ινα, first letter drop cap 2 μηδὲ σχῇς] μὴ δὲσχῆς 3 ἐν παρούσῃ]
ἐν [τῇ] παρούσῃ Legrand 8 τούτων] τ[οι]ούτων Legrand 13 ἁπλῶς] ἀπλῶς
22 ἀπ᾽ ἀρχῆς] and 28 κατ᾽ ἀλλοτρίας] elision mark and breathing printed
together over α 26 ἐκδιδόναι, ὡς] ἐκδιδόναι. ὡς 27 δή τοι] δήτοι 30 λεκτέον
ἐστί,] λεκτέον ἐστὶ, 31 οὐδὲν πρόσκομμα] οὐδέν πρόσκομμα 32 ἐπιϕέ-
ροιεν sic 32 ἄν.] ἂν. 34 μή τις] μήτις 35 ἁπλῆ] ἀπλὴ 35 ἐν ταῖς] ἔν
ταῖς 37 γνωμώνιόν τι sic. γνωμόνιόν τι Legrand 37 καὶ σημεῖον] καί
σημεῖον 37 ὑϕ᾽ οὗ] ὑϕοῦ 37–8 ἀπευθύνοιτ᾽ ἂν] ἀπευθύνοιτἂν 40 παντῇ
που] παντῆπου 40 ἀλλ᾽ οὐ] ἀλλ οὐ 43 ἀλλ᾽ ἅπασι] ἀλλ ἅπασι 45 ἐϕ᾽
ἑκάστῃ] ἐϕ ἑκάστη 47 σύγκρισιν, ὅπως] σύγκρισιν. ὅπως 50 ἐνίοτε]
ἑνίοτε 54 κακεῖνο sic. κἀκεῖνο Legrand 56 ἀλλ᾽ ἀρχαιότατα] ἀλλ ἀρχαι-
ότατα 56 καὶ καθόσον] και καθόσον 57 ἐπηνορθωμένα] ἐπηνωρθωμένα
Legrand 59 τὸ παράπαν] τοπαράπαν 61 ἀποστολικῆς] similarly original.
ἀποιολικῆς Legrand 62 πρὸς αἰδεσιμώτατον] πρὸς [τὸν] αἰδεσιμώτατον
Legrand 65 ἐρῶντες] ἑρῶντες 65 καὶ ἱεροπρεπὲς] και ἱεροπρεπὲς 66 τοὖρ-
γον sic. Cf. Smyth, Greek Grammar, §173.a. 67 κἂν τοῦ χριστοῦ τοῦ κυρίου]
κἂν τοῦ κυρίου Legrand 69 λοιπὴ τὸ] λοιπή τὸ
4.4 Translation
To future readers.
So that you may not be surprised, diligent scholar, or be displeased with
us that in the present Greek printing of the NT, in a way different from that
of the Old, only the letters, without the breathings and accents, have been
set in print and published, we have thought it important that the reason for
this be made clear to all at the outset. It is as follows.
That the most ancient of the Greeks were accustomed to write without
these points (κορυϕαί) on the letters is too clear to need many testimonies.
For certain old copies (ἀντίγραϕα), not a few in number, clearly show this,
such as poems of Kallimakhos and the verses of the Sibyll, and carvings
of great age on stone in the city, engraved simply with letters alone. So it
is quite evident that, in that first bringing into being of the Greek language,
the placing on of these small strokes and marks was not devised, nor con-
tributed to the full completeness of the said language in any way.
Since also all acknowledge that the whole NT, apart from the Gospel
according to Matthew and the Epistle to the Hebrews, was written down
in the Greek language from the beginning just as it was imparted by the
Holy Spirit, we too decided piously to preserve the archaic antiquity and
majesty therein of the same language, and to publish the book without the
least addition whatever, in the manner of the ancient writings, so that we
may not seem to have introduced novelty into something so holy and full
of revered lofty thought, by means of alien and new operations imposed
on it. Moreover, if the truth be told, the lack of breathings and accents
could cause no obstacle to those with any training at all in Greek letters.
I mean this with reference to the pure thought of what is said.
Nevertheless, lest anyone be in doubt on which syllable it is proper to
apply the accent, a simple mark (κεραία) has been attached only to words
of more than one syllable. Let it not be supposed, however, that this is
the Greek accent [mark], but a small pointer and sign by which the studi-
ous person might be guided, so as never to go astray in the pronunciation
(ἐκϕορά) and modulation (εὐρυθμία) of the words. But in the Greek edi-
tion of the OT, because it is more or less entirely a translation and not an
original composition, it was decided not to remove or change anything
from what is common practice in writing.
Since this book earnestly aspires to be of service not only to those
advanced in learning but to all universally who study the holy scriptures,
small roman letters in alphabetical order have been placed on each of the
words, which indicate in succession the match between each word and the
other lying opposite, so that there may be no cause for those who are learners
and have not yet reached the peak of Greek studies to go wrong. Further,
since Greek words are capable at times of seeming to have multiple and
ambiguous meanings and others that give rise to difference of opinion, we
have been diligent to indicate this also, by a dot underneath the small roman
letter standing above the Greek word.
And so that we may make an end of our preface, this also should be
brought to the notice of anyone eager for learning, that we did not have
inferior or just any copies for our printing, but the most ancient and correct
as possible, and in particular so reliable in regard to (κατὰ) their age that
not to trust them is altogether the act of an obstinate and impious (βέβη-
λος) person. Which very copies our most holy father and lord in Christ, the
most high chief priest Leo the tenth, being eager to assist this enterprise,
sent, when they were brought out of the apostolic library, to the most rev-
erend lord cardinal of Spain, at whose provision of funds (χορηγοῦντος)
and command we have printed the present book.
You who love learning, receive with all zeal this holy and reverent work
newly printed. And if you are eager to be seen as imitators of Christ our
Lord and God, and in your deeds to be such, know that there is no excuse
remaining to you for not engaging with (συνομιλεῖν) holy scripture. No
more corrupted copies, no suspect translations, you are not at a loss from
the copied manner of expression (τῆς ἀπογράϕου ϕράσεως). We await
only your wish and readiness. If this is not lacking, it will undoubtedly be
so. For when you have tasted the sweetness of the sacred words, you will
put far from your mind the remaining branches of learning. Farewell, and
do not disparage the novelty of our undertaking.
4.5 Notes on the Greek and the translation

The language is a learned, almost timeless Byzantine Greek, unalloyed by
vernacularisms and ultimately based on Classical models, yet idiomatic within its
own parameters. The style is dense, verbose, and showy. There are no false forms,
even in the case of such difficulties as adjectives in -ης, and no mistakes in con-
cord. The syntax is sometimes idiosyncratic, but hard to fault. Confident handling
of particles, the relative pronoun, and most noticeably the participle, is on dis-
play. Hyperbaton is used, but not to excess. Vocabulary from all eras is deployed
as necessary, including neologisms called for by the recent advance in technology,
the invention of printing.
Besides LSJ (with Supplement) and Lampe, PGL, I have consulted seven
lexicons of Ancient, Byzantine, and Modern Greek: Contopoulos, Neon Lexikon;
BDAG; DGE; Lex. Proia; Mega Lexikon; Sophocles, Lexicon; and OGELD.
Even these do not completely cover the vocabulary available to the author. TLG
CD ROM E (1999) has been checked for the hapax legomena ῥαβδίσκος, ἀνακαι-
νοτομῶ, πρωτοποίητος.
There are some unexpected omissions of the article, notably: 3 ἐν παρούσῃ τῆς
νέας διαθήκης γραικοτυπώσει, 4 ἐν τῆς παλαιᾶς [γραικοτυπώσει], 14 ῥαβδίσκων
τούτων καὶ κεραιῶν ὑπερθέσεις (for, e.g., τὰς τῶν ῥαβδίσκων...), 24 ἀρχαίαν...
παλαιότητα καὶ μεγαλοπρέπειαν, 39 ἐν δὲ τῆς ἀρχαίας διαθήκης ἑλληνικῇ ἐκδό-
σει. Rather than signs of inadequate command of the language, these are to be seen
as a deliberate mannerism of the style: the writer boldly plays against expectation for
effect.
There is a noticeable preponderance of perfect participles. I take this to be a
sign of conscious display.
2 σπουδαῖε. Perhaps as I have translated, in the original sense and as the Latin
version has it (studiose); but it might be no more than an empty cliché, ‘excellent.’
12 πεπαλαιωμέναι. Not just equivalent to ‘old,’ but with a certain flavour,
like ‘aged,’ ‘of long standing.’
14 ῥαβδίσκων. ῥαβδίσκος is not found in LSJ; Lampe, PGL; Mega Lexikon;
Contopoulos, Neon Lexikon; Lex. Proia; OGELD; Sophocles, Lexicon. The
neologism (if such it is) provides a neat descriptor of the accent mark (literally
‘little rod/stick’).
14 κεραιῶν. κεραία is old as a term for any mark or serif above or forming
part of a letter (see LSJ, s.v. II.3, BDAG, s.v.). Here it evidently refers to the
breathing marks, perhaps alluding to the horn shape. But in 36 it refers to the
acute accent mark used to indicate the stress.
15, 24 γλῶττα. The Atticistic form signals literary language. There are no
other occurrences of σσ/ττ here.
16 πανταχόθεν. It makes the best sense if taken closely with the negative, i.e.,
‘not in every way’ = ‘not in any way,’ ‘not at all.’ I do not know of a parallel with
πανταχόθεν, but cf. οὐ᾿ πανταχοῦ, οὐ πάντῃ, οὐ πάντως (LSJ, s.vv.). Otherwise
it is very tautological if taken with ὁλόκληρον τελειότητα (‘the full complete-
ness throughout’).
17 συντετακυίας. For συντείνω intr., ‘contribute,’ ‘conduce’ (to, πρός), see
LSJ, s.v. συντείνω II.2, Lex. Proia, s.v. The thought is that the marks were not
necessary to make the writing of the language complete. One might translate
‘nor was important for …,’ ‘nor had any bearing on. …’
26 δίκην + gen., ‘as, like.’ An ancient usage (LSJ, s.v. I.2; DGE, s.v. A.II)
still sufficiently alive in modern times to be recorded in Lex. Proia.
29 ἐργασίας. An exactly suitable sense is not evident in the lexicons. A trans-
lation such as ‘operations’ or ‘procedures’ fits the context.
29 ἀνακαινοτομῆσαι. Not attested, only καινοτομῶ (with καινοτομία, etc.)
from Classical onwards. A model for the compound in ἀνα- can be seen in ἀνα-
καινοποιῶ (recorded in Lampe, DGE). For the meaning here, cf. LSJ and Lampe,
PGL, s.v. καινοτομῶ. Kαινοτομῶ is used in Doukas’s preface to Plutarch’s
Moralia (1509; Legrand, Bibliographie, 1:93) and Mousouros’s preface to Epis-
tolae philosophorum, etc. (1499; Legrand 1:54).
32 ἐπιϕέροιεν. Strict grammar would require ἐπιϕέροι, with singular sub-
ject ἡ ἔνδεια. But it is plural ad sensum, the subject being in effect ‘the lack of
breathings and the lack of accents.’
36 αὐτή. I take this to be the post-Classical (and Modern) demonstrative use
(subject of ὑπολαβέσθω, complement τόνος ἑλληνικός). For earlier examples of
the development, see Horrocks, Greek (1st ed.), 74, 93 (papyri), 181–3 (Malalas),
226 (in general); BDAG, s.v. 2.a (NT).
37 γνωμώνιον. See above on the spelling. LSJ’s meaning for γνωμόνιον is
‘pointer or dial-hand.’ Cf. γνώμων, ‘pointer’ of a sundial. Here we have a new
application, to describe the accent mark. The Latin version renders notula (‘a little
mark’).
39 εὐρυθμία, ‘modulation.’ I have used a rather general equivalent; I am
not sure if the word is meant to have a more precise reference, such as ‘correct
accentuation.’
40 πρωτοποίητος. Not attested. There are 37 other compounds in -ποίητος
(-τός) in Buck-Petersen, Reverse Index, 482. The meaning is evidently ‘made for
the first time,’ hence ‘original,’ ‘first-hand.’
44 συντελεῖν. For the meaning ‘help,’ be of service’ (to, dat.), see LSJ, s.v. II.2;
Lex. Proia.
45 ἐϕ᾽ ἑκάστῃ τῶν λέξεων, ‘on each of the words.’ 53 ὑπὲρ τὴν ἑλληνικὴν
καθεστηκότι λέξιν, ‘standing above the Greek word.’ The small roman letters
are superscripts standing at the beginning of the word to which they relate.
46–7 ἃ ἐπιδεικνύει … σύγκρισιν. The intended meaning is clear, but the
expression appears somewhat awkward. σύγκρισις = ‘comparison,’ ‘parallel,’
according to the lexicons. τὴν ἀμϕοτέρων πρὸς τὰς ἑτέρας καταντικρὺ κειμένας
ἐπάλληλον σύγκρισιν literally = ‘the of both [words] in relation to the others
lying opposite successive comparison.’ I have paraphrased slightly, in the interests
of natural English.
48 ἐνδόσιμον. A post-Classical meaning, ‘incentive, cause, occasion,’ for which

see Mega Lexikon, s.v. 3; Contopoulos, Neon Lexikon; Lex. Proia. The earliest
attestation is in Arrian, Anabasis 1.7.8 (Mega Lexikon).
50–1 πολυσχιδεῖς … καὶ ἀμϕιβόλους καὶ ἄλλας τινὰς διχογνώμους. All
three adjectives qualify σημασίας.
51 διχογνώμους. A rare word (one citation in LSJ; DGE), but there is sufficient
evidence to determine its meaning: see it and its group in LSJ; Mega Lexikon;
Lex. Proia.
63 ἱσπανίας (hispanie in the Latin version). Likewise (e.g.) in Doukas’s pref-
aces to the Erotemata, etc. and Musaeus, Hero et Leander (both 1514). The use
of the rough breathing implies that its original value was known and it was still
potentially realisable as [h] in pronunciation at this time.
63–4 καρδηνάλιον, οὗ χορηγοῦντος καὶ κελεύσαντος τὸ παρὸν βιβλίον
ἐτυπώσαμεν. The device of the relative clause with participle and hanging finite
verb, in the best Classical manner. Cf., e.g., Demosthenes 19.1 τὸν ὅρκον ὃν εἰσε-
λήλυθεν ὑμῶν ἕκαστος ο᾿μωμοκώς. For a liturgical example, see Αἶνοι at Ὄρθρος,
ἦχος πλ. ά, 1: τὸ μυστήριον, ὃ ἀνυμνοῦσι ἀπόδος ἡμῖν ἀγαλλίασιν καὶ τὸ
μέγα ἔλεος. The form with gen. abs., as in the preface, is a clever variation. For
a use of χορηγῶ like that here, cf. Doukas’s preface to Plutarch, Moralia (1509):
τοιαύτης μὲν οὖν ἠξιώθητε δωρεᾶς, Ἄλδου χορηγοῦντος, ἡμῶν τε διορθού-
ντων.
67–8 κἂν … σπουδάζετε. Is this a slip for subj. σπουδάζητε (after κἄν =
καὶ ἐάν)? Probably not. Rather, it reflects the advance of the merging of indic.
and subj. endings, and the loss of distinction esp. between εἰ and ἐάν/ἄν, ὅτε
and ὅταν, which had begun much earlier. Cf. Mandilaras, Verb, 241–8; Hor-
rocks, Greek (1st ed.), 181, 246–7; BDAG, s.vv. ἐάν, ὅταν for NT examples
and references to the standard treatments. LSJ’s example of κἄν with indic.
(s.v. κἄν II.1) in Apollonius Dyscolus, Synt. 70.22 (II AD) reads: κἂν γὰρ οὕτω
ϕαμέν. …
69 συνομιλεῖν. An interesting choice of word. It appears to be a metaphori-
cal use of the standard meaning ‘converse with.’
71 τῆς ἀπογράϕου ϕράσεως, ‘the copied manner of expression.’ ἀπόγραϕος
is often neut. as subst., ‘copy,’ but here the adj. ‘copied’ (exx. in Sophocles,
Lexicon; none in LSJ; DGE), hence ‘at one remove,’ etc. (A modern meaning
‘crossed out, cancelled’ leads nowhere.) ϕράσις can also mean ‘an (individual)
expression, idiom, phrase’ (cf. LSJ), but this does not fit as well. I take this to
be a succinct description of one of the problems of reading a translation, i.e.,
trying to understand translated idiomatic language without access to the original.
The Latin version gives up and paraphrases all of οὐκ ἀπορίαν τῆς ἀπογράϕου
ϕράσεως ἀπορεῖτε as: non inopia textus originalis, ‘no lack of the original text.’
75 ἐργολαβία. A more general meaning than the older ones in LSJ (‘contract
for the execution of work’; ‘profit-making’). See Contopoulos, Neon Lexikon.
The new vocabulary of printing is seen in 5, 66 ἐντυπῶ, 55–6 ἐντύπωσις,
64 τυπῶ, all old words now put to work in a new way. Other prefaces of the time
use the same terminology. 3 γραικοτύπωσις seems to be a novelty, but γραικό-
τυπος/ον occurs in Mousouros’s preface to Etymologicum Magnum (1499:
Legrand, Bibliographie, 1:61).
5. MISTAKEN VIEWS
It will be as well to begin by clearing away some of the misconcep-

tions that have gathered about this preface. First, it is sometimes said that
it was written by the printer, Brocar. There is no reason whatever to sup-
pose this or evidence to support it. On the contrary, its content and the
character of the Greek clearly suggest, to put it no stronger, that it was
composed by the editor(s) of the volume, as would be one’s first expec-
tation in any case. This error originates with Robert Proctor, in 1900, who
simply refers to ‘the preface by the printer, Arnaldo Guillen de Brocar,’
without explanation or argument.19 I surmise that Proctor was so focused
on the type that he just assumed that anything said about it came from the
printer. He was certainly not well acquainted with the content of the pref-
ace: he misunderstood even the Latin version at a crucial point, as we are
about to see.
A second, more serious mistake is the assertion that the preface states
that the font was cut on the model of a single MS sent from the Vatican
library. This has spread through a number of writers, not all of whom
acknowledge a source, but it can be clearly traced to Proctor. Proctor
says that ‘it can be affirmed with certainty that [the Greek type] is based
on the writing of a particular manuscript,’ and goes on to say ‘it seems …
to be beyond dispute that the type was cut on the model of the writing in
the ‘archetypa tantae uetustatis, ut fidem eis abrogare nefas uideatur,’ sent
to Cardinal Ximenez by Leo X from the Vatican Library.’20 Even without
returning to the original Latin text (let alone the Greek), and relying just
on Proctor’s elliptical quotation, one can see that something is wrong:
eis is plural; it follows that archetypa probably is also. In fact the context
runs as follows (Proctor’s quoted words underlined):
illud lectore[m] no[n] lateat: no[n] queuis exemplaria i[m]pressioni
huic archetypa fuisse: s[ed] antiquissima eme[n]datissimaq[ue]: ac
tante preterea vetustatis: vt fidem eis abrogare nefas videatur. Que
19
Proctor, Printing of Greek, 144. Repeated in Lyell, Cardinal Ximenes, 48; Tasker, ‘Com-
plutensian Polyglot,’ 201; Hall, ‘Trilingual College,’ 5:143; cf. Scholderer, Greek Print-
ing Types, 9.
20
Proctor, Printing of Greek, 144. Similarly Scholderer, Greek Printing Types, 9–10;
Turner, Printed Editions, 12; Tasker, ‘Complutensian Polyglot,’ 204; Hall, ‘Trilingual
College,’ 143; Morison, Politics and Script, 298–9; Barker, Aldus Manutius, 6. In
Bentley, Humanists, 93 there seems to be a remnant of Proctor in: ‘The preface to the
New Testament claimed great antiquity and accuracy for them [the manuscripts used],
but troubled to mention only a single one, lent to Ximénez by Pope Leo X from the papal
library.’
s[an]ctissimus in [Christo] p[ate]r [et] d[omi]n[u]s n[oste]r Leo deci-

mus pontifex maximus huic i[n]stituto fauere cupie[n]s ex ap[osto]lica
bibliotheca educta misit ad Reuerendissimu[m] d[omi]n[u]m Cardinale[m]
hispanie:
Let the reader be aware, that not just any copies were the originals for
this printing, but very ancient and correct ones, and moreover of such
antiquity that to refuse to trust them is wrong. These our most holy father
and lord in Christ Leo the tenth, supreme pontiff, desiring to favour this
plan, sent, after they were brought out of the apostolic library, to the most
reverend lord cardinal of Spain.21
It is quite clear that the preface here (and in the Greek original, lines 55–
62) speaks of manuscripts, plural: no[n] queuis exemplaria i[m]pressioni
huic archetypa fuisse … (‘not just any copies were the originals for this
printing …’). It is clear moreover, as pointed out by Woody in 1971 in
a valuable but neglected article, that the remarks are not about the font, but
about the text and the authorities lying behind it. The ‘manuscript origin’
of the font ‘belongs to a well-populated world of historical myth,’ as
Woody puts it.22 In the perpetuation of this myth, what is most striking
is that writers not only take it on trust from a previous source but often
repeat Proctor’s Latin quotation word for word, oblivious of what it actu-
ally says.
Proctor was responsible for another false assertion, which bears on the
question of the accentuation. He thought the reason given in the preface
for the lack of accents and breathings in the NT text was an ‘ingenious
excuse,’ and the true reason was the ‘defective state of the type’ at this
time. This is instantly disproved by the fact that the same font appears with
full accentuation elsewhere in the volume, namely: in the preface itself;
in the letter of Eusebios to Karpianos; in the miscellaneous introductory
matter to the Epistles (12 pages); in the verses of Doukas and Fausto; and
in the lexicon at the end of the volume (75 pages). All of this amounts to
a substantial body of text.23 The fact that some accentuation, albeit in the
21
I quote the Latin text in the form found in vol. 5, not that in vol. 1. Square brackets
indicate resolution of abbreviations. This passage had been quoted and (correctly) trans-
lated in Tregelles, Printed Text, 4.
22
Woody, ‘Greek Fonts,’ 143. Geanakoplos already in 1962 ignored the myth and stated
the facts correctly (Byzantium, 240–2). Woody (145) makes much out of the preface to
the OT as a means of elucidating the meaning of the NT preface: the OT preface, in a
similar discussion, is clearly talking of the text not the font. This is true and helps to
clinch the case, but the NT preface is already clear. Even Woody manages to misquote
the Latin, printing archetype for archetypa (144). The Cardinal’s prologue in vol. 1 also
clearly speaks of codices of both the Greek OT and the NT sent from the Vatican library.
23
The same point is made by Woody (‘Greek Fonts,’ 146 and n. 7), but he instances only
the poem by Doukas. (The OT volumes, though fully accented, are not relevant here,
reduced form of the acute alone (and the diaeresis), was in fact employed in
the NT text further undermines the theory of a defective font: a full set of
letters with the acute (and the diaeresis) was available. We may doubt, too,
that the editors would have bothered to concoct such an elaborate and plau-
sible justification just to conceal the true reason. A final argument against
Proctor’s assertion is seen in Nicolaus de Lyra’s Liber differentiarum veteris
testamenti, which was printed before the Polyglot, in about 1512. In it a
smaller version of Brocar’s Polyglot NT font is used for occasional patches
of Greek: it has full polytonic accentuation.24
6. INFORMATION IN THE PREFACE
Let us return now to the monotonic accentuation and what can be learnt
about its origin from the preface. The question of the accentuation of the
text was clearly of deep concern to the editors. It is the first topic on their
minds, and occupies more than half the preface. The tone is defensive,
expecting adverse criticism and attempting to forestall it. The question of
the text and the manuscripts is deferred till well on in the preface, and on
this subject, to the regret of later scholarship, it is tantalisingly brief.
The argument presented is as follows. Ancient Greek writing did not
originally use accents and breathing marks. They had not yet been devised,
and were in fact unnecessary. Two kinds of evidence are available as
proof: old ‘copies’ of some authors, such as Kallimakhos and the Sibylline
oracles, and ancient inscriptions still visible in ‘the city.’ In both cases
texts without any accent or breathing marks are observable. The NT was
compiled and written in the same ancient period, and may be assumed to
have been originally written in the same manner. It is therefore historically
accurate to print it without them; and indeed they are not essential. More-
over, it is an act of piety to present the sacred text imparted by the Holy
because a different, Aldine font is used.) On the technical difficulties of printing Greek
with accents and breathings, cf. Proctor, Printing of Greek, 17–21. I do not pretend
to expertise in this area, and am willing to be corrected if I have missed something
that supports Proctor’s view. Proctor also put a false slant on the preface’s statement
about the Septuagint. Where the editors are clearly speaking of the accentuation and
giving their reason for not departing from the usual practice (lines 39–42), Proctor
sees a statement about the choice of font: ‘According to the printer … the Greek of
the Old Testament is merely a translation, and therefore not worthy of his fine special
type’ (144).
24
I first learnt of this work from Norton, Printing in Spain, 40. I rely on Norton for the
attribution to Brocar and the dating (‘not later than 1512’): there is no statement of either
in the volume. Cf. Norton, Descriptive Catalogue, 9: ‘c. 1512?’, and 10: ‘dating based
on ornamental material.’ Norton makes no remark on the accents.
Spirit in its original form without any innovations. It is clear that authen-
ticity is the aim. The NT will be presented in its ancient, original form, that
is, the Greek text without the accentual markings that historically came later.
To the same end, the text is based, as we are told further on in the preface,
on the ‘most ancient and correct’ manuscripts.25
The two pieces of evidence mentioned by the editors invite further
attention at this point. First Kallimakhos and the Sibylline oracles. Woody
astutely perceived that the singling out of these authors for mention ought
not to be accidental, and he found out the reason for it. In the Miscellanea
of Angelo Poliziano, published in Florence in 1489, portions of these very
authors are printed without accentuation or breathing marks of any kind.
As Woody pointed out, Poliziano states or implies that he was follow-
ing his manuscripts in this respect. Clearly, the Polyglot editors selected
Kallimakhos and the Sibylline oracles for mention because they knew
this book and it provided visible testimony to ancient practice. By ‘cop-
ies’ (ἀντίγραϕα) they mean manuscripts (as they do elsewhere), i.e., the
manuscripts lying behind Poliziano’s text.26
There is something more. In the same work, the Kallimakhos verses are
introduced by Poliziano in these words: Sed aures ad Callimachi iam ver-
siculos subscriptos veteri more sine ullis accentiunculis arrigamus (‘Let
us now pay attention to the verses of Callimachus written below in the
ancient manner without any accent marks’). The words veteri more, ‘in
the ancient manner’ are significant: Poliziano already uses the same jus-
tification for omission as the Polyglot editor(s). They could hardly have
been unaware of his remark. It does not follow that they knew of ancient
practice only from this source, but Poliziano’s work was a precedent for
printing Greek in this way.27
25
Woody (‘Greek Fonts,’ 145) makes an interesting observation about the fonts used in
printing the Latin texts. In the OT volumes the font is the more modern roman, in the
NT the archaic gothic. The Latin fonts thus parallel the Greek, i.e., the more modern
Aldine cursive in the OT, the traditional ‘Greco–Latin’ in the NT.
26
Woody, ‘Greek Fonts,’ 146. Illustrations of Poliziano’s Kallimakhos text may be seen in
Proctor, Printing of Greek, fig. 31 and Scholderer, Greek Printing Types, facsim. 19.
The Greek in the rest of the volume has normal accents and breathings (Woody, 146 n. 8).
I have had access to the Poliziano volume only in a later edition (Antverpiae, 1567),
which seems to be a fairly faithful copy. At any rate, it prints the Kallimakhos and Sibyl-
line oracles without accents or breathings, even though the font used is a later Aldine
cursive. Poliziano’s reliance on MSS is stated in the case of the Sibylline oracles (576)
but only implied in that of Kallimakhos (604–5), it seems to me.
27
Scholderer (Greek Printing Types, 5) quotes the words veteri more sine ullis accentiun-
culis in his discussion of the font, obviously aware of their significance; Woody missed
them. Proctor (Printing of Greek, 133) again saw only typographical motives: ‘this
The other evidence mentioned is ancient inscriptions that can be seen

in ‘the city.’ To anyone reading or writing in Greek, this could mean only
one thing: Constantinople. Personal acquaintance with that city, either
by the author or by others who had visited it and could report what they
had seen, is implied. The Latin version, somewhat surprisingly, translates
this as ‘Rome.’ Two possible explanations present themselves. Either the
translator simply misunderstood ἡ πόλις, or, as seems more likely, it was
a deliberate shift by someone who could not allow Rome to appear less
well endowed with ancient monuments than the old capital of the Eastern
Empire. At any rate the change throws into relief the Greek orientation
of the preface, a matter that is important for the question of authorship,
as we shall see.
The argument of the preface now takes a new turn. The text of the
NT, it has just told us, will be printed without accents or breathings, in
fact ‘without the least addition whatever.’ Nevertheless, there will be some
accentuation after all, a simple mark placed on words of more than one
syllable. In practice this means that the inessentials, the three varying
accent marks and the breathings, are dispensed with, and a single accent,
the acute, is used to indicate where the stress is to be placed in all cases
where there might be any doubt. It is a system that supplies the minimum
necessary for practical purposes. Though the diaeresis is not mentioned,
the author of the system knows that it too is necessary, and uses it where
required.
We are told that this system is to provide guidance in the correct ‘pro-
nunciation and modulation’ of the words. The purpose, then, is practical.
It is also pedagogical: the learner needs this assistance. We note that
reading aloud is taken for granted. The pedagogical concern is visible, and
explicitly stated to be the motive, in another feature of the text, the keying
of each Greek word to the corresponding Latin word. Elsewhere in the
volume, too, the needs of beginners in Greek are served, namely in the
Introductio quam brevissima ad grecas litteras and the NT lexicon that
appear at the end of the work. It might be said that volume 5 of the Poly-
glot is both an authoritative presentation of the Greek and Latin texts and
a tool for learning Greek in order to compare them.28
rejection of accents [in the Poliziano vol.] must be held to be due to a deliberate decision
in favour of greater simplicity. This is certainly attained; for the entire fount consists
of twenty-seven Greek sorts.’
28
The keying of the Greek and Latin texts is at the expense of the general appearance of
the page. Cf. Scholderer, Greek Printing Types, 9. Besides the small roman letters, the
text also indicates, by a small forward slash, every Greek word that has no equivalent
There is more than a hint of compromise in the fact that some accen-
tuation is provided after all. It is a step back from the original intention.
These, it seems, are the accents you have when you’re not having any
accents. Why the compromise? Was it simply to answer the stated need?
Or was it from fear that the break with tradition was too radical, despite
the earlier example of Poliziano’s book? We can only speculate: it could
have been either or both. But the latter reason is not very satisfying: critics
would be just as offended by the novelty of a reduced accentuation as
by none at all.
There is a third possibility. The compromise could be the outcome of
conflicting views among the editors. While one side wanted to go all the
way and print nothing but the bare letters, the other insisted that there must
be some accentuation, and if it was not to be the full polytonic system,
at least some minimal indication must be given as a practical guide. We
shall return to this later and see if anything can be deduced about the par-
ties involved. But even if we can find out nothing about them, a conflict
of opinion would provide an explanation for something rather puzzling
otherwise.
The accent system that resulted was clearly not proposed for general
use. The editors are careful to warn the reader that it is not the standard
system of Greek accentuation (‘let it not be supposed …’). The standard
remains the full polytonic system, as used in the printing of the Septua-
gint and elsewhere. The novel one has been created only for use here in
the special circumstances of the NT text. Again we see a concern for the
student, in this case not to allow any misconception to arise. We also see
a conservative outlook on Greek accentuation.
7. DOUKAS THE EDITOR OF THE TEXT AND AUTHOR OF THE PREFACE
I turn now to the question of authorship. Who was the inventor of the
monotonic accentuation? It is not possible to say with certainty, but there
must be a strong suspicion that it was Dimitrios Doukas the Cretan. Before
we approach that conclusion, let us set out what can be said about his
editorship of the volume and authorship of the preface.
in the Latin. Most are occasioned by the Greek definite article. One might have thought
they could be dispensed with, but they are inserted with great diligence, further detracting
from the general appearance. The OT volumes likewise provide the tools for learning
Hebrew and studying the original text. On the dotted letters (ll. 49–53) see below. On
the significance of the NT lexicon, see Lee, History, 47–51.
Bataillon, in 1937, conjectured that Doukas played a leading role in

the preparation of the NT text of the Polyglot. Then Geanakoplos, in his
important work Byzantium and the Renaissance (1962), attempted to bring
this role into clearer focus. Geanakoplos’s cautious conclusion from the
evidence he could find was that ‘it would seem justifiable to affirm that
Ducas was not only the author of [the] preface but primarily responsible
for editing the Greek text of the NT volume.’ This view was accepted and
developed further by Bentley in 1983. The case depends on a series of
data, some no more than clues but others good evidence, which it will be
useful to summarise.29
(1) Doukas is known to have held the chair of Greek at Alcalá. The
date of his arrival cannot be determined, but there is definite evidence
of his presence by 1513. He came to Spain at the invitation of Cardinal
Ximénes, and was already experienced in editing Greek texts: he had
been the principal editor of two major Greek works of the Aldine press in
Venice. In Spain, there is evidence of his work as a teacher and publisher
of textbooks on the rudiments of Greek for students.
(2) In the series of dedicatory verses addressed to the Cardinal in
the NT volume (five sets in all, two in Greek, three in Latin, with their
authors’ names attached), Doukas’s head the list, followed by those of
Nikitas Faustos, Juan de Vergara, Hernán Núñez de Guzmán, and Barto-
lomeo de Castro. The placing of Doukas’s name and verses appears sig-
nificant. None of the others named, or of the other Spanish scholars known
to have been involved, Antonio de Nebrija and Diego López de Zúñiga,
can be shown to have a better claim as leading editor.
29
Bataillon, Érasme et l’Espagne, 21–2, 42; Geanakoplos, Byzantium, 223–55, esp. 239–
43; Bentley, Humanists, 76–9. Cf. also Sáenz-Badillos, La filología bíblica, 399: he
sees Doukas as the only person in the Polyglot team with the capacity to edit the
Greek columns in the NT and the OT, and concurs with Geanakoplos that he was the
author of the Greek preface. Legrand (Bibliographie, 1:117; similarly p. cv) already
asserted that the Greek text of the NT was due ‘aux soins réunis de Démétrius Ducas et
de Nicétas Faustos.’ See also De Jonge, Opera Omnia Erasmi, IX.2, 14–7 for an analy-
sis of Zúñiga’s share in the Polyglot: there is no sufficient ground to speak of him as
‘the principal editor,’ as has often been done (17 with n. 72). Lowry (Aldus Manutius,
286 with n. 121) dissents from Geanakoplos, finding the attribution of the preface to
Doukas ‘on stylistic grounds’ to be ‘ingenious rather than convincing,’ and doubting
that Doukas could have been the editor of ‘a text which appeared in January 1514 if
he had been in Alcala [sic] only since the previous year.’ Lowry’s first argument has
no substance; and as far as I can see there is no certainty that Doukas did not arrive
in Spain before 1513. His last known location before Spain is Venice with Aldus in
March 1509. Aldus shut down his business and left Venice later the same year, when
a French invasion threatened (see Lowry, 159–60); Doukas is likely to have been out
of work then.
(3) A strong link to Doukas is provided by the textual note attached

to Matt 6:13 in the NT, concerning the conclusion to the Lord’s Prayer.
‘Its author exhibits a precise knowledge of the Greek Church’s liturgy’
(Bentley, 79). Doukas, who was both a Greek and subsequently the edi-
tor of the editio princeps of the Greek liturgies (Rome, 1526), certainly
had this knowledge.
The only other person in the team who might conceivably have been as well
informed was Fausto, concerning whom it will be useful to say a little more at
this point. Vittore/Vittorio/Vettor Fausto (c.1480–c.1540) was certainly an Ital-
ian, a Venetian, not a Greek, as has sometimes been said: Νικήτας Φαῦστος is
simply the Greek equivalent of his name. Apart from noting that he was elected
to the chair of Greek at Venice in 1518, and was the editor of a Greek liturgical
book, writers on the Polyglot have offered little information on him. He seemed
a shadowy figure who might or might not have played an important role in the
Polyglot. But a lot more is known about him, and, most surprisingly, that his
‘chief claim to fame is his work as a naval architect’ (Wilson, 92). This began
in Venice in 1525–1526 when he proposed the building of a quinquereme, sup-
posedly based on ancient models. The vessel was in fact built, with some suc-
cess, and Fausto went on to further experiments in ship-building and to become
‘a celebrity in Venice and elsewhere’ (95). Wilson seems sceptical that Fausto
even took part in the Polyglot project, and says (89) that his collaboration ‘does
not appear to have been the original purpose of his travels [in Spain and else-
where], and he cannot have stayed very long with Ducas.’ Fausto’s experience
of editing up to this point was in Latin texts (Cicero and Terence); later, in Paris,
he edited a Latin translation of Aristotle’s Mechanica. The liturgical book, the
Paraklitiki, was published in Venice in 1522 by the de Sabio press at the expense
of Andreas Kounades. There is a Greek preface by Fausto, from which it appears
that he was called in at a late stage as a metrical expert to correct the text from
that point of view. He clearly writes as a non-Greek, addressing his preface τοῖς
καθ᾽ Ἑλλάδα χριστιανοῖς, and employing ‘you’ throughout. All this seems to
justify regarding Fausto as a junior player in the editing of the Polyglot NT text
and not specially expert in liturgical matters.30
30
On the subject of this paragraph, see esp. N. G. Wilson, ‘Vettor Fausto,’ and Diziona-
rio Biografico, 45:398–401. Cf. Legrand, Bibliographie, 1:cv; Bataillon, Érasme et
l’Espagne, 42; Geanakoplos, Byzantium, 166 n. 179, 240. For details of the Paraklitiki,
see Legrand, 1:173 and, with much additional material including Fausto’s preface, Ladas
and Chatzidimos, Προσθῆκες, 7–11. Fausto’s explanation of his part in editing this book
begins: ἔπειτα δὲ [ὁ Κουνάδης] πολλῶν ἀντιγράϕων συνειλεγμένων τὸ ἐλλεῖπον ἐϕ᾽
ἑκάστου ἐξ ἀλλήλων ἀναπληρώσας τὸ σύνολον, ἅτε ποιητικοῦ τινὸς [sic] ἀνδρὸς
δεόμενον διὰ τὸ ἔμμετρον αὐτοῦ, ἐμοὶ παραδέδωκεν, ἵνα δὴ τὸ ϕαῦλον ἐκκρίνας
τὸ ὀρθὸν δοκιμάσαιμι. Incidentally, the Paraklitiki (also called the Oktoïkhos) is not, as
Geanakoplos has it (Byzantium, 240 n. 72), a book of ‘hymns to the Virgin,’ but the book
containing the hymns of the eight-tone cycle. The Dizionario Biografico (‘una raccolta
di preghiere alla Vergine’) betrays its source. Geanakoplos (240) says Fausto was ‘prob-
ably a pupil of Ducas.’ This is a guess, but a plausible one. Bataillon’s guess (Érasme et
l’Espagne, 42) that he was ‘peut-être le typographe qui a composé le texte hellénique sous
la direction de Doucas’ seems improbable: surely this was Brocar’s task.
(4) While the older Spanish sources all play down the role of Doukas,
Nebrija once ‘let slip,’ as Bentley puts it (78), ‘the information that Hebrews
(i.e., conversos) and Greeks were charged with editing, respectively, the
Hebrew and Greek scriptures.’ Doukas was the only native Greek working
on the project.
(5) Geanakoplos observed (243) that the language of the preface ‘seems
to be that of one intimate with the nuances of Greek style.’ This is right,
but can be put a good deal more strongly. In my opinion the character of
the Greek is such that it could not have been written by anyone but an
advanced Hellenist. It is a tour de force in the Byzantine style. While there
were other Hellenists among the Polyglot scholars, they were all non-
Greeks, and all Spaniards apart from Fausto. Doukas stands out among
them all as the person with the necessary expertise to write it. I offer an
additional observation that reinforces this. The Latin translation is inex-
pertly done, with simplifications, gaps at difficult points and some outright
mistakes in translation. This suggests that none of the Polyglot scholars
who could write Latin – and that would have been all of them apart from
Doukas – had sufficient skill in Greek to translate such a text well. Still
less could they have composed it.31
(6) The reference to ancient inscriptions in ‘the city,’ i.e., Constan-
tinople, implies acquaintance with the place, something unlikely to be
available to anyone but Doukas. Geanakoplos was inclined to see this as
evidence that Doukas himself had visited Constantinople. This of course
cannot be ruled out, but it is not certain: word of mouth could have con-
veyed this information to the Greeks in the West; in fact it may have been
common knowledge among them. To me the strength of this point lies more
in the Greek perspective evidenced here and highlighted by the change to
‘Rome’ in the Latin version. The writer writes as a Greek. Doukas seems
the best, if not the only, candidate.32
(7) The preface uses the first person plural frequently, but the singular
‘I’ occurs at one point (l. 30 λέγω …, ‘I mean this with reference to the
pure thought of what is said’). Geanakoplos was rightly cautious about
31
Geanakoplos (Byzantium, 242 n. 78) pointed out that the Latin version is ‘simpler in its
wording.’ But he went on to say, ‘which may mean it was translated from the Greek,’
showing he had missed the statement in the heading.
32
Geanakoplos, Byzantium, 225 with n. 7, 242–3. There seems to be good evidence that a
journey to Constantinople was possible after 1453. But the ‘vividness’ of the phrasing in
the preface, appealed to by Geanakoplos as a sign that Doukas had seen the inscriptions
in situ, does not strike me. In fact the Latin version suggests autopsy in a way the Greek
does not: monumenta vetustissima que rome adhuc visuntur, ‘monuments of great age
which are still on view in Rome.’ The Greek simply says (l. 12) πεπαλαιωμέναι ἐν τῇ
πόλει λίθων γλυϕαί, ‘carvings of great age on stone in the city.’
using this, but thought that it ‘may well indicate that one person was
mainly responsible for editing the Greek text.’ I am not sure that it does
that; but it certainly suggests that a single person was responsible for
writing the preface. The ‘we’ elsewhere, though it could mean ‘I,’ is more
probably inclusive, reflecting the collaborative character of the project.
The ‘I’ emerges in explanatory λέγω, ‘I mean,’ because the singular is
usual in this expression unless one really means ‘we.’ The Latin trans-
lation does not, as Geanakoplos asserted, change ‘I’ to ‘we’ here, but
drops the whole sentence! In fact the Latin version eliminates all but two
of the thirteen occurrences of ‘we/us/our.’ I am not sure what significance
is to be seen in this. Perhaps it is simply stylistic. It has a depersonalising
effect, at any rate.33
8. FURTHER INDICATIONS
To the above can be added further evidence that I have not seen adduced
before. The first is in lines 67–68, where we read κἂν τοῦ χριστοῦ τοῦ
κυρίου καὶ θεοῦ ἡμῶν μιμηταὶ ϕανῆναι καὶ τοῖς ἔργοις γενέσθαι σπου-
δάζετε, ‘and if you are eager to be seen as imitators of Christ our Lord and
God, and in your deeds to be such.’ The phrase τοῦ χριστοῦ τοῦ κυρίου
καὶ θεοῦ ἡμῶν has the sound of Greek liturgical language. Though I have
not found this exact phrase, combinations of a similar kind are common-
place throughout liturgical texts:
τοῦ Κυρίου καὶ Θεοῦ καὶ Σωτῆρος ἡμῶν Ἰησοῦ Χριστοῦ
Κύριε Ἰησοῦ Χριστὲ ὁ Θεὸς ἡμῶν
Χριστὸς / Χριστὲ ὁ Θεὸς ἡμῶν
Χριστῷ τῷ Θεῷ
Χριστῷ τῷ Βασιλεῖ καὶ Θεῷ ἡμῶν
τοῦ ἀγαπητοῦ Υἱοῦ αὐτοῦ, Κυρίου δὲ ἡμῶν Ἰησοῦ Χριστοῦ
33
The two that survive are 60 ὁ … κύριος ἡμῶν … λέων ~ d[omi]n[u]s n[oste]r Leo;
64 ἐτυπώσαμεν ~ imprimi fecimus. The others are at lines 2, 5, 23, 27, 53, 54, 55 bis,
65, 70, 73. That first person plurals do not necessarily mean a plurality of persons is
demonstrated with clarity by another preface of Doukas’s, that found in the Erotemata,
etc. (Alcalá, April 1514). Here we have χαρακτῆρσιν οἷς ἐνετύχοδεν [sic] in a series of
first person singulars where Doukas is definitely speaking, or rather complaining, of his
own labours in preparing the book. The text of this is reproduced in Legrand, Bibliogra-
phie, 1:119. Geanakoplos (Byzantium, 235) translated: ‘using letters [characters in Greek
type] which I found here.’ His translation implies a reading ἐνέτυχον ὧδε. But I suggest
the true reading is ἐνετύχομεν, the δ being simply a typesetter’s error. Cf. in Doukas’s
prefaces to Rhetores (1508; Legrand, 1:88) and to Plutarch (1509; Legrand, 1:93): τοῖς
ἀντιγράϕοις οἷς ἐνετύχομεν. Bataillon translated χαρακτῆρσιν οἷς ἐνετύχοδεν as
‘avec les caractères que j’avais sous la main’ (Érasme et l’Espagne, 22).
The similarity is evident and the point reinforced by the strikingly dif-
ferent formulation in the Latin version: et si christi Opt[imi] Max[imi]
sectatores videri vultis [et] esse. Just as the textual note on Matt 6:13
links Doukas to the editing of the NT text, this – possibly unconscious –
reminiscence of Orthodox liturgical language links Doukas to the writing
of the preface.34
The application of the title Optimus Maximus to Christ in the Latin text is note-
worthy. It was of course originally a title of Jupiter in Roman antiquity. It was then
transferred at some point to the Christian God, as seen in the well known D.O.M.
(Deo Optimo Maximo) inscribed on churches and elsewhere, and modelled on
I.O.M. (Iovi Optimo Maximo). The application to Christ rather than God the
Father is unexpected, possibly even of questionable orthodoxy. Who at Alcalá
could have written such a thing? And is it an isolated instance? I have no answer
to the first, but in answer to the second can point to some parallels, though far
from Spain of 1514. Optimus Maximus Jesus occurs in a Scottish source of the mid
sixteenth century.35 The application of the title to Christ is also found in Rabelais:
in a passage in the Quart Livre (1552), in which he interprets the death of Pan in
Plutarch as the Crucifixion, he adds ‘… car cestuy tresbon, tresgrand Pan, nostre
unicque Servateur, mourut lez Hierusalem, regnant en Rome Tibere Cesar.’36 The
whole topic is clearly one worth further study.
In the same sentence there is another difference of interest. In μιμηταὶ …
γενέσθαι we have a distinct NT allusion (cf. 1 Cor 4:16; 11:1 μιμηταί μου
γίνεσθε, Eph 5:1 γίνεσθε οὖν μιμηταὶ τοῦ θεοῦ ὡς τέκνα ἀγαπητά, 1 Thess 1:6;
2:14). The Latin translator misses the allusion by translating μιμηταί as sectatores
(‘followers,’ ‘adherents,’ ‘enthusiastic supporters’), whereas the Vulgate ren-
ders μιμηταί as imitatores in all places; the Complutensian NT lexicon likewise
gives imitator as its only gloss on μιμητής. The reason for the rendering is not
quite clear, but it may have arisen, as Harm W. Hollander points out to me, from
the association of μίμησις with ζῆλος/ζήλωσις, and of μιμητής with ζηλωτής.
Cf. Moeris p. 168 (ed. Pierson, 1759) ζήλωσις, Ἀττικῶς. μίμησις, Ἑλληνικῶς
34
In the NT the only comparable passages are 1 Thess 1:12 κατὰ τὴν χάριν τοῦ θεοῦ ἡμῶν
καὶ κυρίου Ἰησοῦ Χριστοῦ, which however is usually taken as ‘according to the grace
of our God and the Lord Jesus Christ’ (RSV); and John 20:28 ἀπεκρίθη Θωμᾶς καὶ
εἶπεν αὐτῷ· ὁ κύριός μου καὶ ὁ θεός μου. There is a parallel use of liturgical language
in Doukas’s preface to Αἱ Θεῖαι Λειτουργίαι (Rome, 1526; Legrand, Bibliographie,
1:193–4), where the memorable phrase τὸ ἀρχαῖον κάλλος, ‘[their] ancient beauty,’
occurs twice. This comes from the Orthodox funeral service: … εἰς τὸ καθ᾽ ὁμοίωσιν
ἐπανάγαγε, τὸ ἀρχαῖον κάλλος ἀναμορϕώσασθαι (Νεκρώσιμα Εὐλογητάρια 2).
35
Ferr[erius], Kinloss 81, recorded in DMLBS, s.v. optimus. I thank Jim Adams, David
Howlett and Theodor Christchev for their assistance, and especially for making the
dictionary entry available to me in advance of publication. I also thank Edwin Judge
for valuable help along the way.
36
Marichal, Quart Livre, 138. See also Screech, Rabelais, 328. I am grateful to Michael
Screech for supplying me with this information. He also kindly informs me, in answer
to my query, that he has never consciously come across such a use of the title in Eras-
mus’s works and would expect to recall it if he had.
[ζ 4 Hansen]; Ammonius 209 (ed. Nickau, 1966) ζῆλος καὶ ζηλοτυπία διαϕέ-
ρει … ζῆλος δὲ μίμησις καλοῦ …, cf. 210–211; Hesychius p. 677 (ed. Schmidt,
1867) ζηλωτής· ἐρεθιστής. μιμητής. In the NT text, 1 Peter 3:13 ζηλωταί has
a v.l. μιμηταί, the majority reading. Finally, the rendering of ζηλωτής as secta-
tor occurs in the Vulgate at Titus 2:14. At any rate it seems that the writer of
the Greek preface was familiar with the Greek NT in a way the Latin translator
was not: the latter neither recognises the Greek allusion nor recalls the Vulgate
equivalent.
The preface also contains a not so veiled criticism of the Vulgate and
its manuscript witnesses. When it says (ll. 69–71) ‘no more corrupted
copies, no suspect translations, you are not at a loss from the copied man-
ner of expression,’ the target can only be the Vulgate. What this means,
put bluntly, is: the Vulgate text is corrupt, its translations faulty, and its
meaning difficult to grasp. Instead of grappling with these problems, the
scholar can now engage directly with τῇ θείᾳ γραϕῇ, the original Greek
NT. The sentiments expressed earlier about ‘piously’ preserving the text
in its ancient form and avoiding the introduction of ‘novelty into some-
thing so holy’ are further evidence of the same high regard for the origi-
nal. While this may represent the views of some of the Polyglot scholars,
they were in general deeply conservative in their approach to the Vulgate;
at the extreme end of the spectrum was one like Zúñiga, who later vehe-
mently defended the traditional Latin text against the perceived attacks
of Erasmus. The Cardinal himself, though he could assert the need for
recourse to the originals when the Latin manuscripts differed or appeared
corrupt, revered Jerome’s translation and took the conservative side in the
dispute between Nebrija and the other editors over the Latin version,
restraining Nebrija from presenting a new translation based on the Greek
in place of the Vulgate. The conclusion all this tends to is that the views
espoused in the preface are those of one who takes for granted the superi-
ority of the original Greek over the Latin translation, that is, someone like
Doukas or Nebrija.37
37
Cf. Bentley, Humanists ch. 3, esp. 89–91 (dispute with Nebrija), 97–111 (respect for
the Vulgate; evidence of unpublished annotations); also Homza, Religious Authority,
87–8. For a summary of Zúñiga’s outlook, see De Jonge, Opera Omnia Erasmi, 18–20.
The Cardinal is said to have exclaimed to Nebrija, ‘God forbid that I should alter a word
of the Blessed Jerome’s!’: so Fernández-Armesto, ‘Cardinal Cisneros,’ 157. While the
Cardinal’s prologue expresses much in common with the views of Nebrija, it also dis-
plays what Fernández-Armesto characterises as ‘pre-humanist techniques,’ ‘echoes of
cabbalistic lore,’ and ‘the world of mystical devotion’ (157–8). For the well known pas-
sage Accedit … examinetur, in the Cardinal’s prologue, see, e.g., Darlow and Moule,
Historical Catalogue, 2:4; Homza, Religious Authority, 248–9. An English translation of
the whole prologue can be found in Olin, Catholic Reform, 61–4. The Latin version of
The fact that there is a Greek preface at all is something to reflect on.
Only in this volume of the Polyglot is there a Greek preface; all the other
prefaces are in Latin. A connexion is thereby made to all the other publica-
tions of Greek texts up to this time, each usually equipped with a learned
preface in Greek, written by a Greek scholar involved in the editing of
the volume. These range from the 1476 epitome of Laskaris’s grammar,
with a preface by Dimitrios the Cretan (a different Dimitrios), to the 1513
edition of Plato’s works, introduced in Latin by Aldus and in Greek by a
200-line poem (soon to be famous) by Markos Mousouros, the joint editor.
And the practice continued long after the time of the Polyglot. The Greek
preface in the NT volume of the Polyglot signals the presence of a Greek
scholar involved in the work, and the quality of his preface is a sign of his
competence in Greek.38
In other Greek prefaces both before and after the Polyglot the indica-
tion of the author’s name at the top is standard practice. There is no name
at the head of our preface. This absence, however, seems to me readily
explicable. The fact that the volume was a team effort would have played
some part, but a more compelling reason can be suggested. The great pro-
ject was the Cardinal’s, and all honour for it belonged to him. As Doukas’s
own verses said, ἔργα αὐτοῦ ἥδε βίβλος, ‘this book is his work.’ Of
course this was exaggerated, but it was the right tone to adopt in deal-
ing with a powerful hierarch and patron. It would have been unseemly
for someone else’s name to appear at the top of the first page. The col-
laborators on the volume could put their names to their verses later on
in the book, but the opening preface remained tactfully anonymous. In
due course the Cardinal’s own prologue, addressed to Pope Leo and set-
ting out his titles in their full glory, appeared on the opening page of
volume 1.
the NT preface does not tone down the above passage much, No[n] mendosa exemplaria:
non suspecte translationes: non inopia textus originalis, though mendosa (‘faulty’) is
less harsh than διεϕθαρμένα, and the last clause is a weak paraphrase (cf. notes above
on l. 69): perhaps the translator missed what was going on. A further point that seems to
have drawn no comment before: the preface to vol. 5 has nothing to say about the Latin
text printed in the volume. It is only from the title page that one learns that it is the Vul-
gate (cum latina beati hieronymi translatione); there is no other information.
38
A large body of Greek prefaces from 1476 to 1563 can be seen in Legrand, Bibliographie,
vol. 1. Five other Greek prefaces or notes by Doukas, three of them quite lengthy, are
found in: (1) Rhetores Graeci (vol. 1. Venice, 1508; Legrand, 1:85–88); (2) Plutarch,
Moralia (Venice, 1509; Legrand, 1:92–93); (3) Erotemata, etc. (Alcalá, April 1514;
Legrand, 1:119); (4) Musaeus, Hero et Leander (Alcalá, [1514]; Legrand, 1:120);
(5) Αἱ Θεῖαι Λειτουργίαι (Rome, 1526; Legrand, 1:193–4). There does not appear to
be one in Doukas’s final publication, Alexandri Aphrodisiei Commentaria (Rome, 1527;
Legrand, 3:316–7).
There is then a strong case for concluding that Doukas was the editor
of the NT Greek text and the writer of the preface to the volume. This of
course is not to rule out or diminish the contribution of others: it is cer-
tain that the book as a whole was a collaborative effort, and that is likely
to have included some of the labour on the Greek text. The preface also,
though written by one person, could have received input from others in
one way or another. It certainly had to reflect the general policy decided
on by the team under the direction of its head, Cardinal Ximénes de
Cisneros.
9. DOUKAS THE INVENTOR OF THE ACCENTUATION
Was Doukas the inventor of the monotonic accentuation? There is no

way to prove decisively that he was; we can only gather the available clues
and draw the most probable conclusion. The clues seem to me to be as
follows.
Doukas had general responsibility for editing the Greek text, and wrote
the preface. He must at least have been closely involved with the decisions
about the accentuation of the Greek and the execution of them.
We know that he was a native Greek, a Greek scholar and teacher, and
an experienced editor of Greek texts. He was the only person in the team
with all those qualifications. Concern over the accentuation of the Greek
would seem to belong to such a person, rather than the Latinists and
experts in the Vulgate; they are more likely to have felt it was a matter
of small consequence.
That Doukas himself did indeed worry about accents is shown by
another preface of his, in the edition of Plutarch’s Moralia (Venice, 1509).
At one point the following memorable passage occurs:
Ὑμέτερον δ᾽ ἂν εἴη τὸν ἡμέτερον κάματον εὐγνομόνως ἀποδέξα-
σθαι, μὴ δ᾽ ἐκϕαυλίζειν, εἴ τι που γράμμα διεστραμμένον ἢ ἄλλο
τι τοιοῦτον ἡμᾶς παρέδραμεν, οὐκ ὂν ἐμποδὼν οὐδὲ τοῖς μήπω
ϕωνῆς γευσαμένοις ἑλληνικῆς. Πολλοὶ μὲν ἀπείρως ἔχοντες τῆς
περὶ τὸ διορθοῦν ϕιλοπονίας ἤ, βέλτιον εἰπεῖν, ταλαιπωρίας,
καὶ τοῦ γνῶθι σαυτὸν ἀμνημονοῦντες, ὅλως τε σοϕοὶ δοκεῖν ἐϕιέ-
μενοι, μεγάλῃ τῇ ϕωνῇ λαρυγγίζουσιν· «ὀξεῖα ἀντὶ βαρείας
ἐτέθη· παῖε, παῖε τὸν κατάρατον, διέϕθαρται γὰρ ἡ βίβλος ἣν
ἐκεῖνος ἐπετέτραπτο κατασκευάσαι τοῖς ϕιλολόγοις ὁλοσχερῆ τε
καὶ ἀνενδεᾶ.»
Your role would be to receive the product of our toil kindly and not
disparage it if some misshapen letter or other such thing has got past us,
when it is no hindrance even to those who have not yet tasted the Greek
language. Many who have not experienced the labour, better to say
pain (ταλαιπωρία), of editing, and forget the maxim ‘Know yourself,’
and are anxious to seem very clever, shout out with a great voice: ‘An
acute has been put instead of a grave! Hit him, hit him, the cursed
fellow! For the book that he was entrusted with preparing for scholars
complete and without deficiency is ruined.’39
We can see that Doukas had suffered over Greek accents. This outburst
is a cry from the heart. He is unlikely to have given the matter anything
less than his best attention when it came to the printing of the NT text.
The extensive treatment of the topic in the preface is another sign of that
preoccupation.40
That Doukas was responsible for the execution of the system in the
Polyglot text seems extremely likely, not only from the points made so far,
but from the nature of the task. Let us consider what was involved. The
manuscripts exhibited the polytonic accentuation. To make the changeover
to monotonic would have been too complicated for a typesetter to carry
out alone, even with detailed instructions. Someone needed to go through
the entire text and mark the new accents and diaereses on the copy that
went into the print-room. That person would have been able to follow the
polytonic accents of the manuscripts as a guide to the position of the stress,
but expertise in NT Greek would be required to ensure that the appropri-
ate changes into the new system were made; some acquaintance with the
pronunciation tradition of the Greek NT, if not with spoken Greek, would
have been an advantage.
There are some indications that the latter was indeed brought to bear.
The occasional accentuation of monosyllables, discussed above (§2),
39
Greek text from Legrand, Bibliographie, 1:92–3; my translation. See Lowry, World of
Aldus, 240 for a lively picture of the methods – or lack of them – in Aldus’s workshop
when this edition was produced. In Doukas’s ed. of the Erotemata, etc. (April 1514),
there is another heartfelt preface on the difficulties of the editor’s lot, including almost
the same phrase as above: ἐν ταῖς ταλαιπωρίαις τῆς διορθώσεως.
40
Sáenz-Badillos (La filología bíblica, 392), after noting what the preface says about omis-
sion of accents and breathings and the use of ‘pequeñas señales’ as an aid to pronuncia-
tion, attributes the decisive voice in both to Doukas: ‘Es muy probable, que sea Demetrio
Ducas, – el más familiarizado con los antiguos textos, y el que mayor experiencia edito-
rial tiene – quien haya hecho prevalecer su opinión en este punto concreto.’ Cf. also 399:
‘… puede ser suya la decisión de editarlo sin acentos ni espíritus.’ Sáenz-Badillos does
not discuss the accentuation further. It may be no coincidence that a work published
long afterwards by one of Doukas’s pupils was on accentuation: Diego Sigeo, De rati-
one accentuum (Lisbon, 1560). Sigeo refers explicitly to his old teacher (… praecep-
torem meum in schola Complutensi Demetrium Ducam. …). I owe the information here
to Legrand (Bibliographie, 1:194 n. 2); cf. Geanakoplos, Byzantium, 233. I have not
(yet) been able to see this ‘livre rarissime,’ or to check whether it contains any remnant
of Doukas’s views on accentuation.
indicates more than a mechanical approach to placing the stresses. Though

it created inconsistency, this phenomenon arose from familiarity with the
actual pronunciation of Greek. The placing of the diaereses was also not
a simple matter. The manuscripts often have superfluous diaereses, and
the retention of only those that are necessary requires judgement. Again,
actual pronunciation is the guide. Someone who knew what he was doing
carried out these tasks. Whether it was Doukas himself, or whether he
delegated some or all of the work, he at least supervised it.41
It is harder to say for certain whether Doukas himself conceived the
basic idea of the monotonic accentuation, that is, the idea of a system using
an acute accent mark to indicate the stress to the extent necessary and no
more. It could have come from some other person or source, either in
Spain during the work on the Polyglot, or at an earlier time when he was
in Venice or even Crete. So far everything points to Doukas in response
to a certain need at Alcalá, but we need to consider other possibilities.
10. PARALLELS IN THE OLD TESTAMENT TEXT
First there is something of direct relevance to be seen in the other part

of the Polyglot, the OT. The OT volumes (1–4, 1517) are introduced by
their own preface, which runs in part:42
Illud est etia[m] co[n]sidera[n]du[m]: q[uod] in hebraicis characterib[us]
scie[n]ter omisim[us] apices illos: quib[us] nu[n]c vtu[n]tur Hebraei p[ro]
acce[n]tibus. Na[m] hi cu[m] ad nulla[m] vel significati vel p[ro]nu[n]-
ciationis differentia[m] pertinea[n]t: sed ad sola[m] ca[n]tus ipsoru[m]
modulatione[m]: merito a veteribus Hebraeis reiecti sunt: quos in hoc
imitari maluimus. Veru[m] ne locus acce[n]tus cuiusq[ue] dictionis igno-
raretur: hoc modo p[ro]uidimus: vt quonia[m] dictiones hebraicae vt
plurimu[m] in vltima habe[n]t accentu[m]: o[mn]es huiusmodi dictiones
nullo prorsus apice notare[n]tur: reliquae vero no[n] habentes accentu[m]
41
One might speculate that Fausto was Doukas’s main assistant (cf. above). An illustra-
tion of what might be involved in the diaeresis: in John 1:19 the Polyglot text prints
λευίτας, where λευΐτας is found in some MSS, as in a gospel lectionary of XII AD
(Metzger, Manuscripts, no. 38). The lack of a second accent in απέστειλε με, etc. (see
above) remains puzzling, however. Why did they leave these out, if pronunciation was
the guide?
42
This preface is headed Prologus. Ad lectorem. De his que ad lectionem veteris tes-
tamenti diuersis linguis nunc primum impressi sunt praenotanda and comes next
after the Cardinal’s prologue addressed to Pope Leo X. Both are printed in all four
OT volumes.
in vltima (quae rarissime occurru[n]t) Super syllaba[m] vbi praedomi-

natur acce[n]tus: apice signare[n]tur: hoc modo ‫א ֵר ץ‬.
ֶ֨ [sic for ‫]א ֶר ץ‬
ֶ
This also is to be observed, that in the Hebrew characters we have
knowingly omitted those marks which the Jews now use for accents. For
since these do not relate to any difference of meaning or pronunciation,
but only to the melody in chanting them, they were with good cause left
out by the Jews of old, whom we have chosen to imitate in this matter.
Nevertheless, so that the place of the accent of any word may not be
unknown, we have made provision in this way, namely, since most
Hebrew words have their accent on the final syllable, all the words of
this kind are distinguished by no mark at all, but the remainder that do
not have the accent on the final syllable (which occur very rarely) are
indicated by a mark over the syllable where the accent predominates,
thus ‫א ֵר ץ‬.
ֶ֨
We learn that the cantillation marks, which also function as indicators

of the accent, will be omitted from the Hebrew text as not being part of
the ancient way of writing. An indication of the position of the stress will
however be given by a simple system which marks the stress when neces-
sary, namely, when there is an exception to the rule that the final syllable
receives the stress. The mark (apex) illustrated is not unlike a Greek grave
accent and is placed above the syllable. Thus the student or anyone else
not completely familiar with Hebrew receives the necessary guidance in
reading aloud. Inspection readily confirms that this system has been applied,
and with great consistency, throughout the Hebrew text. Moreover, although
nothing is said about it, the system has already been put into practice in
volume 6 containing the Hebrew lexicon, printed in 1515.
The whole procedure here runs parallel to that seen in the editing of the
Greek text of the NT. In both, an unnecessary part of the accentuation is
abandoned for the sake of a return to ancient practice, and instead a simpli-
fied accent is introduced as a practical guide to pronunciation.
There is a further notable parallel between the OT Hebrew and the NT
Greek texts. Not only do they both show the same method of keying each
word to the corresponding word in the Vulgate by a small roman letter, but
both use the method of marking selected words by a dot under the letter.
This phenomenon is easy to miss, and I am not aware that it has been com-
mented on before. It is introduced in the NT preface in the sentence ἔτι δὲ
ἐπείπερ … (see above, lines 49–53) and in the Old by an explanation begin-
ning Ex his praeterea litterulis nonnullas reperies: quibus tenue punctu[m]
subscribitur hoc modo a Quod quidem punctum designat: … (‘Moreover
among these small letters you will find some under which is written a slight
point, thus a, which point indicates. …’). Examples can readily be seen on
inspection of the texts.43 This phenomenon deserves a study of its own,
which will not be undertaken here. For the moment we note it only as a
further significant parallel between the texts of the two Testaments.44
It is evident that the editors of the two parts conferred, and decided
on similar procedures in printing the accents and other features of the
texts. But can we deduce which came first? The order of printing of the
volumes is not much of a guide. Decisions about how the Hebrew text was
to be printed would have been made long before 1517, given the huge size
of the task. We know the system of accentuation was already in place
in 1515 in the Hebrew lexicon, and work on it must have begun well
before then. Further, the writer of the NT preface already knows by 1514
how the LXX will be printed in the OT volumes, that is, with full accen-
tuation (see lines 2–5 above). So priority is difficult to determine by the
dates.45
Perhaps it is futile to attempt to determine priority, but the decision
about the Hebrew accentuation appears to me, at least, to be based on what
was done in the Greek rather than independently evolved for the Hebrew.
If authenticity was the aim, we might have expected the editors of the
Hebrew to decide to omit all vowel signs and print just the bare letters.
This would indeed have been a return to the most ancient practice. Instead
they decide to omit the so-called ‘accents,’ which they know are primar-
ily musical notation marks. There is no mention of precedents, as there is
in the case of the Greek, no allusion to old manuscripts without accents.
Further, the accent system introduced is, as far as I am aware, an extreme
novelty for Hebrew, unlike the Greek one, which arises out of the existing
43
[An instance can be seen in the image of John 1:1–14 above: εγεννήθησαν in the
second-last line (1:13) is preceded by a dotted superscript ‘c.’ The dot indicates that there
is a discrepancy between the Greek meaning and the Vulgate Latin nati sunt.]
44
Woody (‘Greek Fonts,’ 145) already noticed the statement about omission of accents,
but not the rest. Homza (Religious Authority, 86) mistakenly speaks of the Polyglot as
printing Hebrew without vowel markings.
45
The OT preface echoes the NT (Latin) preface in the discussion of the Greek MSS
employed, even using the same words at some points: see in the NT preface illud
lectore[m] no[n] lateat … hispanie (quoted above) and cf. in the OT preface: Quod
aut[em] ad Graeca[m] scriptura[m] attinet: illud te no[n] latere volumus: no[n] vulgaria
seu temere oblata exemplaria fuisse huic nostrae impressioni archetypa: sed vetustissima
simul et eme[n]datissima: quae sanctissimus D[omi]n[u]s noster Leo Decimus Po[n]tifex
maximus caeptis nostris aspira[n]s ex ipsa apostolica Bibliotheca ad nos misit: ta[n]
tae integritatis: ut nisi eis plena fides adhibeatur: nulli reliqui esse videa[n]tur: quibus
merito sit adhibe[n]da. It is natural to suppose that the OT preface of 1517 was written
later and drew on the earlier one.
system. And the accent symbol chosen is like the Greek one but turns the
opposite way. Why so? Because it mirrors the Greek. Moreover, in the
Hebrew lexicon we find that an accented Hebrew word and its trans-
literation into roman letters show accents facing opposite ways, thus:
áleph. ‫ ָ֨א ֶלף‬and: ‫ ָ֨מיִ ם‬Máim. The suggestion of a mirror image is further
supported by the way the small roman letters are placed. In both the Greek
and the Hebrew texts they come before the word: in the Greek that means
to the left, since one reads from left to right, but in the Hebrew it means
the opposite side, that is, to the right, because one reads from right to
left.
I conclude from this that the idea of using an acute accent mark to indi-
cate the stress where necessary originated in the editing of the NT Greek
text and was adapted for use in the Hebrew text. The Hebrew editors’
primary reason for giving up the cantillation marks may well have been
convenience: it was challenging enough to print the intricate vocalisation
in full, without attempting those as well. The justification that they were
not ancient could then, with some truth, be brought in.46
11. LATIN, SPANISH, AND ITALIAN
The basic notion of the monotonic accentuation, then, does not seem
to have come from the Hebrew of the Polyglot. But there are other pos-
sible sources to be considered and eliminated. A system of marking Latin
accentuation was long used, and is still found, in Latin service books, as
also in various related works. In its basic form it indicates the stress by an
acute in words of three or more syllables, this being the minimum neces-
sary. Its purpose was obviously to guide those whose grasp of Latin accen-
tuation needed some assistance when reading aloud. How old is this sys-
tem? Might it already have been established in Spain and Italy in the time
of the Polyglot? I see no evidence that it was.
I have checked a number of mapping points, as follows. There are no marks
in (1) the editio princeps of the Missale Romanum (Milan, 1474), in (2) Missale
ad Usum Sarum (Basle, c.1489), and in (3) Missale Romanum (Venice, 1549).
There is frequent but still incomplete marking of the stress (the circumflex is also
46
On the accents/ cantillation marks, see Bauer and Leander, Historische Grammatik, 136–
54; Joüon-Muraoka, Grammar, 61–9. There is a cantillation mark ‫’( ֨א‬azla), that is simi-
lar to the grave used. But the choice of it from among the thirty or so marks available
can only have been motivated by its similarity to the Greek accent.
used) in (4) Missale Romanum (Salamanca, 1587), and very limited marking, some-
times none at all in (5) Missale Romanum (Munich, 1613). Finally, full application
of acute marks where necessary, and also occasional use of grave and circumflex
(for a different purpose), can be seen in (6) Missale Parisiense (Paris, 1738), and
(7) Uffizio della B. Vergine Maria (Rome, 1782).47
Latin of the early sixteenth century in general is written without marks,

as in the prefaces and Latin texts of the Polyglot. At the same time, how-
ever, there begins the occasional use of the grave, not to mark stress but
to indicate vowel length, as, e.g., in adverbs (longè, verò), a phenomenon
still seen long afterwards. But there is no sign of a system of marking the
stress.48
Nevertheless the occasional use of an acute to indicate stress can be
found, when it is needed for a special purpose. Thus Melancthon’s Greek
grammar (1527), in the discussion of accents, illustrates some points by
writing Latin words with an acute; Erasmus does the same in his work
on Greek and Latin pronunciation (1528); and Nebrija marks the stress
of some Latin words in his Latin-Spanish lexicon (1492).49 Similarly, as we
have just seen, volume 6 of the Polyglot itself uses an acute to mark the
stress in transliterations of Hebrew words. The idea as such was certainly
about at the time. But there is no sign of the application to Latin of a fully
developed system; moreover the inspiration for the sporadic use of the
acute mark seems to come from Greek.
The same question needs to be asked regarding Spanish and Italian.
Both employ accentual markings today. How far back do these systems
go? Could Doukas or someone else associated with the Polyglot have
seen in the contemporary writing of the vernacular languages a model
for Greek? The Spanish system was clearly not in place at this early
date. In Nebrija’s grammar of Spanish (Castilian) (1492) and his Latin/
Spanish lexicons (1492, 1495), accent marks are not used.50 Spanish
47
A facsimile of (2) can be seen in Johnson, First Century of Printing, pl. 4, and of (4) in
Thomas, Sixteenth-Century Printing, pl. 46. John Sheldon kindly made available his
copy of (7). Other copies or facsimiles were seen in the Veech Library, Catholic Insti-
tute of Sydney (I thank Charles Hill, Margaret Watt and Michael Cullen), and Fisher
Library, University of Sydney. I have been unable to discover a discussion of this
phenomenon.
48
See the plates in Johnson, First Century of Printing; Johnson, Italian Sixteenth Century;
Thomas, Sixteenth-Century Printing.
49
Melancthon [sic], Grammatica graeca. I have not seen the earlier ed. (1518), entitled
Institutiones graecae grammaticae. Erasmus, De recta latini graecisque sermonis pro-
nuntiatione, e.g., 200–1. For Nebrija, see next note.
50
Antonio de Nebrija, Gramatica castellana (1492); Diccionario Latino-Español
(1492); Vocabulario español-latino (1495?). Nebrija does use the acute as a means
documents up to and well after the time of the Polyglot are also seen to
be unaccented.51
In Italian there is still ‘chaos’ in the writing of accents today (or at least
up to 1966), in the words of a leading authority, but the beginnings of accent
marking are seen further back than in Spanish. The situation in the early
sixteenth century is summed up by Migliorini (trans. Griffith) as follows:
Accents, too, [like the apostrophe] were modelled on Greek, as can be
seen from the preference for an acute accent internally in a word (in the
rare cases when an accent was used) and the grave accent on the final
syllable. After sporadic appearances in the Quattrocento, they were used
by Bembo and Manuzio in the Asolani (1505), which sometimes had the
grave on the final vowel (menò, altresì; but also amista, castita, etc.)
and very rarely an accent within the word (restío).52
It is clear that at this time there was no regular system based on the acute
alone as far as necessary, as in the Greek monotonic. Moreover, Greek
itself was the source of the sporadic and unsystematic accentuation that
there was.
None of these, then, provides a model for what was done in the accen-
tuation of the NT Greek text of the Polyglot. Nor do we have any indica-
tion that any other person had thought of such a system for Greek before.
It seems most probable that the idea was conceived by Doukas, who devel-
oped it into a workable system and saw to its application in the printing
of the text. He had the knowledge and experience – one might add confi-
dence – to invent a new system in response to the need that had arisen, and
to come up with one that was simple, practical, and elegant.
of indicating the stress in Spanish when discussing points of accentuation, but only
then: see Gramatica castellana, lib. 2, cap. ii–iiij. In his Latin-Spanish lexicon, some
Latin headwords have the stress marked (e.g., Dídymus, Tolétum), but the Spanish is
unaccented.
51
I have checked the 129 plates in Arribas Arranz, Paleografía documental hispánica,
showing documents ranging from 850 to 1641, including several in the period 1490–
1516: there are no marks of accentuation (unlike ç and ñ, which appear sometimes). In
Thomas, Sixteenth-Century Printing, I see no marks in the Spanish texts until, and only
in, pl. 48, Vida del P. Ignacio de Loyola (Madrid, 1594), which has very occasional use
of final grave or acute.
52
The description as ‘chaos’ is in Migliorini, The Italian Language, 474; quotation 237
(= Migliorini, Storia della lingua italiana, 383–4). See further Migliorini, The Italian
Language, 181, 231, 278, 322–3, and the plates in Migliorini, Storia, esp. the Aldine
Petrarch of 1501. A description of the present-day system may be seen in Dardano and
Trifone, La lingua italiana, 395–6. It is clearly not comparable to the monotonic accen-
tuation of Greek: both acute and grave (also, rarely, circumflex) are used only in cer-
tain words and conditions, and not solely to indicate stress. (I am grateful to Gianluca
Alimeni of Macquarie University for assistance here.)
12. THE PROPONENT OF AN UNACCENTED TEXT
Who was responsible for the initial intention to print the Greek text
without accents? This is an interesting question that does not have to be
answered, but is too tantalising to leave untouched. One could of course
argue that it was Doukas himself. Persuaded by the ancient model of the
inscriptions in Constantinople and aware of the precedent in Poliziano’s
Miscellanea, he as editor of the Greek text could have initiated the idea
and led the rest of the team to agree with him. On the other hand, the final
outcome smacks so much of compromise that it is tempting to see behind
it a difference of opinion among the editors, as I have argued above. If
there was such a conflict, on which side is Doukas likely to have been?
In my view Doukas is likely to have been conservative in this matter. As
a native Greek educated from an early age in the writing of Greek, and
later a teacher of the language, he would have been totally at home with
the polytonic system and, in the general nature of teachers, committed
to its retention despite the difficulty of learning it. The suggestion to print
without accents on the ancient model I think came from someone else, and
it was Doukas who clawed back some accentuation to serve as a practical
guide. Who was that someone? We need to look for a radical.
Woody came up with a name behind the proposal to omit accents and
breathings, though he makes no mention of the other ingredient in the
situation, the decision to insert a simplified accentuation:
The fact that the omission of accents was borrowed from Poliziano
may enable us to guess who among the Complutensian editors was
responsible for the idea. It was probably Elio Antonio de Nebrija, who
had studied for ten years in Italy and whom we find, on at least one occa-
sion, placing Poliziano at the head of a list of Italian humanists (even
before Pico della Mirandola).53
This is an inviting suggestion, but there is a problem about dates. Nebrija

did not actually join the Polyglot team until the autumn of 1513, when prepa-
ration of the NT text must have been well advanced (printing was complete
in January 1514). Nevertheless the idea is not to be abandoned entirely.
The link to Poliziano is stronger than Woody supposed. Sáenz-Badillos has
shown that some of Nebrija’s material in his Tertia Quinquagena (Alcalá,
53
Woody, ‘Greek Fonts,’ 147; cf. 144, 146. Woody voiced his reservations: ‘[t]here
is, however, no proof that the idea was Nebrija’s, and the point is not one I will insist
upon.’
1516) depends directly on Poliziano’s Miscellanea, the very source of the

notion of printing Greek without accents, as we saw.54
Nebrija (1441 or 1444–1522) was among the most famous and distin-
guished scholars of his day. He was a renowned linguist, the author of the
first grammar of Spanish and a compiler of lexicons. His approach to the
NT was along humanist lines and his work was of the calibre of Valla’s.
In Spain ‘il devance Erasme lui-même’ (Bataillon, 27). He held the chair
of grammar at Salamanca until mid 1513; it was not until then that Car-
dinal Cisneros was able to attract him to Alcalá, ‘at the cost of a hand-
some salary’ (Bentley, 88). We also know that Nebrija held radical views
on some other matters bearing on the content of the Polyglot, so much
so that he withdrew rather than go on associating himself with it. From a
surviving letter we know that he felt thwarted on two counts, both issues
in which he was on the side of innovation and the rest of the team were
conservative, namely his desire to revise the Vulgate, and his strong objec-
tion to the planned inclusion of a ninth-century list of interpretations of
proper names by Remigius. Nebrija was radical enough to envisage a NT
text without accents, and senior enough to be listened to. If he was not
in Alcalá at the relevant time, his views could still have had an impact at
a distance, especially through those whom he had taught and who were
members of the project team, Núñez de Guzmán and López de Zúñiga.
In the end there is of course no necessity for Nebrija alone to have been the
source: the contents of Poliziano’s book could have been widely known
among Spanish Hellenists, including those at work on the Polyglot.55
54
Sáenz-Badillos, La filología bíblica, 71–3. Bentley (Humanists, 84–5) argues that Nebrija
began work on the Tertia Quinquagena well before 1505.
55
On Nebrija’s NT work, date of arrival, contribution to the Polyglot, and letter to Cardinal
Cisneros: Bataillon, Érasme et l’Espagne, 26–41; Sáenz-Badillos, La filología bíblica,
162–7; Bentley, Humanists, 88–91; Homza, Religious Authority, 248 n. 34 (letter).
Sáenz-Badillos (166–7) finds some reflection of Nebrija’s views in the interpretations
of proper names provided in vol. 5. If this is right, it is an important clue that Nebrija
did have some input even into vol. 5, whether in person or through his writings. On
Núñez and Zúñiga as pupils of Nebrija: Sáenz-Badillos, 175, 199. But Zúñiga’s teacher
of Greek was Barbosa: see De Jonge, Opera Omnia Erasmi, 14. Barbosa, too, had stud-
ied in Italy under Poliziano (De Jonge, ibid.; Sáenz-Badillos, 199). Juan de Vergara (born
1492) was still a youngster, though full of promise. According to Sáenz-Badillos (322),
he studied Greek under Doukas and Núñez. Homza (4) says an inventory of his books
after his death included a work or works of Poliziano (his later brush with the Inquisition
is described in entertaining detail in Homza’s ch. 1, pp. 1–48). Full polytonic accentua-
tion was used in the LXX, of which Doukas was probably editor also (Bataillon, Érasme
et l’Espagne, 42; Geanakoplos, Byzantium, 239 n. 70; Sáenz-Badillos, 399). If my recon-
struction is correct, Doukas got what he wanted there, and it had nothing to do with the
13. CONCLUSION
Whatever the truth may be on the question of who proposed the print-
ing of an unaccented text, the outcome was the invention of the monotonic
system, and it occurred in Spain nearly four centuries before the system
emerged again in Greece in modern times. Its inventor was almost cer-
tainly Dimitrios Doukas of Crete, principal editor of the NT volume of the
Polyglot.
reason given in the NT preface. This would explain its specious logic. The justification
given is that the LXX was a translation and not the original. But it was just as ancient as
the NT, as they well knew: if printing without accents was historically authentic for the
NT, it was also for the LXX.
12
ΕΞΑΠΟΣΤΕΛΛΩ
2007
Abstract
This paper was contributed to a collection of essays with the theme ‘Septuagint
Greek and its significance for the New Testament.’ Each contributor chose a word
or words found in both texts and explored the relationship between them, con-
sidering also the possible input of other sources. In this study of ἐξαποστέλλω it
is shown that the word was adopted by the Pentateuch translators when it was in
vogue in the third century BC, but that over the next two centuries it gradually
dropped out of use in ordinary Greek. It reappears in the NT almost exclusively
in Luke-Acts. I argue that Luke used it as a ‘Septuagintalism,’ rather than taking
it from contemporary Greek. Later reappearances of the word suggest that it had
entered the ‘biblical’ style from the NT (and LXX). The study illustrates the neces-
sity of taking into account contemporary Greek in tracing the influence of the LXX
on NT Greek. The other ‘send’ words, πέμπω and ἀποστέλλω, are included in the
discussion, as being in the same semantic field as ἐξαποστέλλω and part of its
history.
INTRODUCTION
The word ἐξαποστέλλω is used in both the LXX and the NT. It occurs
commonly in the former and occasionally in the latter, almost exclu-
sively in Luke–Acts. It would be possible to proceed at once to assess the
LXX and NT data and begin to draw conclusions about the links between
the two. But there is a third ingredient to be taken into account, namely
the usage in Greek outside the biblical texts and contemporary with them,
that is, in the period from the third century BC to the first century AD.
As this paper will show, a proper appreciation of the LXX and NT can-
not be achieved without knowing what was going on in the background.
A further point is that ἐξαποστέλλω was part of a semantic field, which
it did not occupy alone but shared with other words, notably πέμπω
and ἀποστέλλω. These also will need to be taken into acccount in our
investigation.
ΕΞΑΠΟΣΤΕΛΛΩ OUTSIDE THE BIBLICAL TEXTS
The reference books tell the story up to a point. LSJ offer some 19 instances,
nearly all in post-Classical Greek, including some from documents and
from the LXX and NT. A few more documentary references are to be found
in MM. BDAG provides the usual brief summary and occasional citations,
including some not in LSJ (e.g., Josephus), without figures. Lampe, PGL
does include the word (not always the case) and lists a few occurrences in
patristic authors. The Mega Lexikon covers much the same ground as LSJ.1
Lee in 1983 drew attention to some further examples in III BC papyri and its
frequency in Polybius, and noted that the one apparently Classical example
(in a purported letter of Philip in Demosthenes 18.77) is likely to be a later
insertion.2 In TDNT there is a brief survey, which sums up with ‘this word
became common in Gk. from the time of Polybius with essentially the same
meaning as ἀποστέλλειν.’3 All this is more or less what we expect from
our present reference works, which have necessarily relied on the old methods
of gathering data. The picture so far is of a word that came on the scene in
III BC as ‘a more vigorous form of the older word [ἀποστέλλω]’ (Lee), and
continued in use in Koine authors right through to the patristic period, being
equally at home, as far as one could tell, in both vernacular and literary Greek.
But let us now consider what can be learnt from more comprehensive
data obtainable by electronic searching. We begin with documentary texts,
accessible via the disk PHI7, which contains most of the published inscrip-
tions and papyri.4 The number of occurrences of ἐξαποστέλλω in these
texts turns out to be around 270. The majority of these by far are in inscrip-
tions (over 200). The figures for ἀποστέλλω and πέμπω in the same cor-
pus, however, are somewhat over 3,000 each.5 We can immediately deduce
that ἐξαποστέλλω was never a strong contender against these two older,
better established words.
The documentary evidence is also interesting for another, more signifi-
cant reason. Inspection of the search results reveals that nearly all the 270
or so occurrences of ἐξαποστέλλω are dated to the first three centuries BC;
texts from I AD and later number fewer than ten. Within the latter group,
1
[DGE has since reached this word and has an extensive selection of examples. I do not
see anything that contradicts my findings. See n. 9 below for one inscr.]
2
Lee, Lexical Study, 93–4. References to the useful older studies of Anz (1894) and Glaser
(1894) are found there. Cf. also Cadbury, Style of Luke, 32.
3
TDNT, 1:406; cf. 398-400 on ἀποστέλλω and πέμπω. Tagliaferro, in ‘Nota linguistica
filoniana,’ explores Philo’s attempted distinction between ἐξαποστέλλω and ἐκβάλλω
in Gen 3:23, 24, together with a review of their attestation in non-biblical Greek.
4
Packard Humanities Institute CD ROM 7.
5
It is important to note that figures given are approximate; it is very difficult to make them
exact.
ΕΞΑΠΟΣΤΕΛΛΩ 185
the papyrus texts show something interesting too. There are six, and only
two date from I AD, the rest from IV and V AD. The two I AD examples
are in petitions, PRyl 2.127.22 (29 AD) and SB 6.9150.42 (5 AD).6 The
relevant wording in PRyl 2.127 runs … ἀναζητῆσαι ὑπὲρ τοῦ μέρους καὶ
τοὺς αἰτίους ἐξαποστεῖλαι ἐπὶ σὲ πρὸς τὴν ἐσομένην ἐπέξοδ(ον), ‘… to
enquire into the matter and despatch the culprits to you for the punishment
they will receive.’ This is a formulaic section of the petition, the conclusion
requesting action of the official, and something of the kind occurs with
variations in countless documents of this type. The SB 6.9150 example is
similarly located: … ὅπως ἀσϕαλισάμεν[ος] τὸν ἐνκαλούμενον ἐξαπο-
[σ]τείλῃς ἐϕ᾽ οὓς καθήκει. From earlier centuries the same use of ἐξαπο-
στέλλω in petitions can be readily exemplified, as in PDion 10.13 (109 BC)
ὅπως οἱ αἴτιοι ἀναζητηθέντες ἐξαποσταλῶσι ἐπὶ τὸν στρατηγόν.7
It is clear that these two texts of 29 and 5 AD are employing traditional
language continued from petitions of previous centuries. This not only
explains the two occurrences in documents of I AD but offers a clue to the
‘tone’ that ἐξαποστέλλω might now have, given that it was becoming rare.
Its tone is likely to have been formal and official.
The later occurrences in papyri, i.e., in IV and V AD as mentioned,
are rather isolated and not inconsistent with what has been suggested. They
are mainly in official letters in formal style; only one is in a private letter.8
I interpret these as coming from a time when ἐξαποστέλλω had under-
gone a revival – but not in popular speech: it remained a word of formal-
ity. In any case these much later examples are an insignificant number in
the overall picture.
As to the occurrences in inscriptions, there are, as we have already
noted, over 200. Among these I can find only one of certainly AD date,
SEG 18.143.45 (Corinth, c.43 AD, copy of decrees and letter of the con-
federation of the Lycians).9 The inscriptional texts, like this one, tend to
employ formulaic, traditional, and formal language.
6
SB 6.9150 is entered in PHI7 with the date ‘Vspc’ but in the edition it is ‘Jahr 5 n. Chr.’
(not ‘5. Jahrh. n. Chr.’).
7
Others are: SB 14.11274.13 (4 BC), PRainCent 51.17 (I BC), PGrenf 1.38.18 (170 BC).
The editor of PRainCent 51, Bastianini, notes that ἐξαποστέλλω is frequent at this point
in petitions; similarly Tagliaferro, ‘Nota linguistica,’ 417 n. 11, with further references.
8
The official letters are: PPanopBeatty 2.115, 154 (300 AD); POxy 50.3577.6 (342 AD);
PPrinc 3.183.5, 11, 19, 21 (345 AD); The private letter is SB 6.9158.[11], 21 (V AD)
θέλησον δὲ ἐν τάχι ἐξαποστῖλαι τὸ σημῖον μὴ ὀκνουμένη, ἀμμή. Notice also
PapGraecMag 15.16 (III AD) Βαρβαρε Λαιλαμψ …, ἐν τῷ οὐρανῷ θεὸς ὁ μονογενής,
ὁ ἐκσαλεύων τὸν βυθόν, ἐξαποστέλλων ὕδατα καὶ ἀνέμους. In BDAG there is a reference
to ‘PFouad 203, 2ff (I-IIAD)’ s.v. ἐξαποστέλλω 1.b, but no such text can be located.
9
Editor’s statement of date: ‘paullo post a. 43p.’ Another text of possibly AD date is
IG VI 1.14.7 (Sparta), given as ‘primo a. Chr. n. saeculo vel paulo post.’ There remains
The documentary evidence, then, gives a clear indication that ἐξαπο-

στέλλω was a word that enjoyed a vogue in III–I BC, but thereafter became
obsolete in ordinary language. Its use continued then sporadically, but with
some noticeable ‘affect’ or additional connotation.
Let us now turn to the evidence of Greek literature.10 The overall figure
for occurrences of ἐξαποστέλλω is high, around 2,400, but dwarfed by
those for ἀποστέλλω (13,000+) and πέμπω (17,000+). The data are more
diverse and complicated than in the case of the documents, involving, as
they do, a great range of authors of all dates. I offer the following repre-
sentative sample:
ἐξαποστέλλω ἀποστέλλω πέμπω

Polybius (III–II BC) 208 66 89
Diodorus Siculus (I BC) 102 331 93
Dionysius Halic. (I BC) 4 104 60
Philodemus (I BC) 0 0 0
Philo (I BC–I AD) 15 7 14
Josephus (I AD) 12 55 297
Vita Aesopi G (I AD) 0 3 13
Arrian (I–II AD) 0 9 62
Dio Chrysostom (I–II AD) 0 3 12
Epictetus (I–II AD) 0 5 9
Plutarch (I–II AD) 6 126 332
Achilles Tatius (II AD) 1 2 12
Lucian (II AD) 0 5 34
Pausanias (II AD) 0 43 18
Cassius Dio (II–III AD) 1 61 376
Xenophon Ephes. (II/III AD) 0 1 1
Basilius (IV AD) 33 108 23
Eusebius (IV AD) 106 228 28
John Chrysostom (IV–V AD) 103 497 500
We see the predominance of occurrences in the centuries BC, followed by

a decline and then a revival in IV AD. It turns out that only Polybius, often
thought of as a writer of ‘literary Koine,’ strongly preferred ἐξαποστέλλω
a small residue of texts undatable for me, thanks mainly to the reluctance of epigraphers
to indicate any kind of date, even within wide parameters (the refusal of some to add ‘AD’
and ‘BC’ (or ‘CE/BCE’) to dates is a further source of uncertainty and frustration).
But even if all the uncertainly dated inscriptions turned out to have an AD date they
would be too few to affect the general picture. [DGE record one additional AD inscr.,
IAdramytteion 16.25 (‘II d.C.’), but this turns out to be a mistake: in PHI Greek Inscrr.
the date is ‘106 v.Chr.’]
10
Accessed via TLG CD ROM E (1999). Again, figures are not exact.
to its rivals in all its history.11 Other pointers are the following. Philo’s
examples are all either in LXX quotations or in re-uses of LXX wording
following a quotation. In Josephus, ἐξαποστέλλω, though not in quota-
tions as such, often occurs in contexts with obvious biblical associations.
The single example in Achilles Tatius is a curious one: it looks as if it
was deliberately used to secure a succession of ἐξ- compounds.12 When
we come to the sample patristic authors, Basil, Eusebius, and John Chrys-
ostom, the majority of their examples by far are in biblical quotations or
allusions, or in discussions using the words from a previous quotation. In
the rather popular Greek of the Apophthegmata Patrum (V–VI AD), ἐξα-
ποστέλλω is not one of the standard ‘send’ words: there is but one occur-
rence, in a LXX quotation.13
All the data support the interpretation that the heyday of ἐξαποστέλλω
was in the third to first centuries BC and it declined thereafter until it
entered the vocabulary of ecclesiastical writers under the influence of the
biblical texts. It was not an established vernacular word in the first and
later centuries AD.
I have said nothing so far about the semantics of ἐξαποστέλλω and its
competitors ἀποστέλλω and πέμπω, i.e., about their lexical meanings and
range of application; I have simply compared the figures. It is of course
unlikely that the three words were totally interchangeable. Nevertheless
existing analyses show that the overlap was great, especially in the case
of ἐξαποστέλλω and ἀποστέλλω, and little can be done to separate them
without an exhaustive examination of the examples (together with the minor
players ἐκπέμπω and ἀποπέμπω).14 For the present study this would yield
little of interest; what is of interest here is the frequency of one word vis-
à-vis its main competitors over time. The result is that we can track its rise
and fall, and gain an insight into its currrency and its probable ‘feel,’ or
connotation, at any given time. The first of those times is when the LXX
came into being.
11
[For useful remarks on ‘literary Koine’ and on Polybius see Horrocks, Greek, 96–8.]
12
Ach. Tat. 6.9.1 ἐβόα πάλιν, Τὸν μοιχὸν ἐξέκλεψας σύ, τῶν δεσμῶν ἐξέλυσας, καὶ τῆς
οἰκίας ἐξαπέστειλας·
13
Abba Petros Pionites 2. Figures for ἀποστέλλω and πέμπω in the Apophthegmata: 47 and
41. I have left many loose ends because they do not affect the picture. For example, the
TLG search produces an occurrence of ἐξαποστέλλω in Aeschylus (Fragmenta 314c.11),
but it turns out not to be in any play of Aeschylus but in a later summary by Diod. Siculus.
Dates of authors are taken from the TLG Canon.
14
See Muraoka, Lexicon, s.vv.; Lee, Lexical Study, 94; Dorival, Les Nombres, 308;
BDAG, s.vv.; Louw and Nida, Lexicon, 15.66, 67, 68. An analysis of πέμπω, based
mainly on Classical data, is found in Chadwick, Lexicographica Graeca, 233–40.
ΕΞΑΠΟΣΤΕΛΛΩ IN THE SEPTUAGINT
The overall figures are: ἐξαποστέλλω 287, ἀποστέλλω 691, πέμπω 22.15
When we look within these at the Pentateuch, we find that ἐξαποστέλλω
and ἀποστέλλω are used almost equally, the former 80 times, the latter 94.
The Pentateuch translators clearly had a liking for ἐξαποστέλλω, despite
the availability and semantic equivalence of the older word ἀποστέλλω.
The big loser is πέμπω, which is used only once (Gen 27:42): for some
reason it was not thought to be up to the job. Outside the Pentateuch, ἐξα-
ποστέλλω is spread through the different books without any obvious con-
centrations, but ἀποστέλλω, as the massive figure suggests, overtakes it in
frequency in most books after the Pentateuch. πέμπω, however, is notice-
ably concentrated in books not in the Hebrew Bible (17 times), and among
these in books known for their literary Greek (there are 3 in Wis, 12 in
2–3 Macc). I interpret this as due to a feeling on the part of those writers
that πέμπω, as the old-fashioned, unadorned word, was appropriate to the
classicising style.16
It was with the Pentateuch translators, then, that the fondness for ἐξα-
ποστέλλω began, and we naturally assume this to be a reflection of the
currency of the word in their time. This is true up to a point. But let us
dig a little deeper. A search of third-century BC Egyptian papyri reveals
556 examples of ἀποστέλλω, 21 of πέμπω, but only 5 of ἐξαποστέλλω.17
These figures are quite surprising. Such a marked disparity between ἐξα-
ποστέλλω and ἀποστέλλω in particular is at odds with the Pentateuch
frequencies. It begins to look as if the Pentateuch translators favoured
ἐξαποστέλλω even beyond what would have been usual at the time. Why
would that be? We have no means of knowing for certain, but the explana-
tion could be that ἐξαποστέλλω was marked as the slightly more formal
and impressive word in comparison with ἀποστέλλω, and might therefore
have been regarded as especially suited to the biblical style. It was not a
15
Figures from LEH, Lexicon, in turn derived from the CATSS files. The Hebrew equiva-
lent in nearly all instances is ‫ שׁלח‬qal or piel.
16
The figures for πέμπω in authors of literary Greek (above) tell a similar story. There
is a parallel in the distribution of κελεύω in the LXX: see Lee, ‘Thirty Years On,’ 520
[= Essay 8].
17
In twenty vols. containing the majority of III BC texts, viz. PHib 1, 2; PLille 1, 2;
PSI 4–6; PPetr 1–3; PMich 1; PCol 3, 4; PCorn; PCairZen 1–5; PYale 1 (part). The
larger figures are not precise. The five III BC examples are: PSI 4.384.4; PPetr 2.37.3;
PCairZen 1.93.7; 4.578.2; PYale 1.39.12. [A fresh search via Papyri.info locates five
more: PHal 7.2; PKoeln 6.258.2; SB 18.14042.9; 28.16923.1; PSorb 1.50.4. The previ-
ously noted PYale 1.39.12 is now dated c.190 BC, not quite in III BC, but near enough.
The tally, then, is 10.]
total equivalent of ἀποστέλλω, but because of the doubled-up prefix car-

ried some additional connotation, however slight. Of course it is obvious
that often the translators are simply ringing the changes on the two words,
but this need not invalidate the point.18
The result we have for the LXX is this. The LXX translators adopted a
word that was current in the early Koine of their time and made it a charac-
teristic LXX word by frequent use; there is, in addition, a possibility that
it was slightly more formal in tone than its main equivalent. With the pas-
sage of time, however, changes would occur in these words. Writers of the
next period would make different choices, in accordance with the vocabu-
lary of their own day; their choices might also reflect acquaintance with an
older text that had significance for them.
ΕΞΑΠΟΣΤΕΛΛΩ IN THE NEW TESTAMENT
In the first century AD ἐξαποστέλλω was no longer a standard part of

the Greek vocabulary. That it might still appear sporadically is not to be
denied – no word ever disappears completely from Greek! – but the evi-
dence has shown that both vernacular texts and literary authors made little
use of it now. We would not expect to see it in the Greek of the NT writ-
ers, unless there were special factors at work.
The statistics for the three main ‘send’ words in the NT are:
ἐξαποστέλλω ἀποστέλλω πέμπω

NT as a whole 12 132 79
Matt 0 22 4
Mark 0 20 1
Luke–Acts 10 49 21
John 0 28 32
Gal 2 0 0
All other Epistles 0 9 16
Rev 0 3 5
18
E.g., Gen 8:7 ἀπέστειλεν, 8 ἀπέστειλεν, 10 ἐξαπέστειλεν, 12 ἐξαπέστειλεν.
Exod 3:10 ἀποστείλω, 12 ἐξαποστέλλω, 13 ἀπέσταλκεν, 14 ἀπέσταλκεν, 15 ἀπέ-
σταλκεν, 20 ἐξαποστελεῖ. 1 Kgdms 5:8 ἀποστέλλουσιν, 10 ἐξαποστέλλουσιν, 11 ἐξα-
ποστέλλουσιν, ἐξαποστείλατε. In these passages ἐξαποστέλλω has the role of ‘varier’
and ἀποστέλλω that of ‘varied.’ I thank Michael Curran for this observation and his
comments on this paper generally.
We see that there are some examples of ἐξαποστέλλω after all, but they
are quite few in number, and concentrated in the writings of one author,
Luke. The predominant word throughout is ἀποστέλλω, even in Luke.
The old plain word πέμπω makes a strong showing in John, who has a
special fondness for it in ὁ πέμψας με (22 times) and similar expressions,
but otherwise it shows up most in the more literary Greek of Luke and the
Epistles.19
It is a commonplace of NT commentary that Luke was a ‘Septua-
gintaliser.’ That is, he not only incorporated many direct quotations from
the LXX into his work, but consciously imitated its language and style by
frequent use of LXX expressions and vocabulary. Moreover, some degree
of adaptation to context is observable: that is, LXX features may be delib-
erately placed to suit the character of the context. Though details remain
debatable, this seems to be the consensus view; it is certainly the best
explanation of the phenomena.20
The conclusion is easily reached. Luke displays a fondness for LXX
features in his writing; our word was a characteristic LXX word; none
of the occurrences in Luke are simply a matter of direct quotation; and
we know that ἐξαποστέλλω was obsolete in contemporary Greek. We
conclude that Luke drew it from the vocabulary of the LXX.21 We can
also conclude that the two occurrences in Galatians must be explained as
either older Greek or Greek in the Septuagintal style, not contemporary
usage.
Let us take a closer look at the NT occurrences in the light of this
conclusion and see what else can be observed. The first example is in the
Infancy Narrative, in Mary’s Song of Praise:
19
Mark’s one instance of πέμπω is odd: 5:12 πέμψον ἡμᾶς εἰς τοὺς χοίρους, ἵνα εἰς
αὐτοὺς εἰσέλθωμεν. Other NT words: ἐκπέμπω twice (Acts 13:4; 17:10), ἀποπέμπω
not at all; perhaps ἐκβάλλω, included with ἐκπέμπω and ἐξαποστέλλω by Louw and
Nida, Lexicon in 15.68, instanced by Matt 9:38.
20
From the extensive literature: Fitzmyer, Luke I–IX, 113–25; Ó Fearghail, ‘Imitation of the
LXX’; Haenchen, Acts, 74–5, 80; Barrett, Acts, 2:xlv–xlviii; Witherington, Acts, 123–5;
Arnold, ‘Luke’s Characterizing Use.’ Sparks’s articles ‘Semitisms of Luke’ and ‘Semi-
tisms of Acts’ retain their value (ἐξαποστέλλω is not mentioned). ‘Context’ embraces
situation, speaker, subject-matter. Some Marcan instances of adaptation to social context
are presented in Lee, ‘Some Features.’
21
Commentators often point out that ἐξαποστέλλω is restricted to Luke–Acts (and Gala-
tians) in the NT, and may add that it occurs in the LXX, but the conclusion does not
seem to be drawn. Fitzmyer lists ἐξαποστέλλω among characteristic Lucan vocabulary
(Luke I–IX, 110), but not as a ‘Septuagintism.’ Tagliaferro, ‘Nota linguistica,’ 417 n. 13
drew the opposite conclusion to mine, namely that ἐξαποστέλλω in the NT ‘appare
conforme al contemporaneo greco profano piuttosto che a quello dei LXX.’ But this was
based on an incomplete survey of the evidence.
Luke 1:53 πεινῶντας ἐνέπλησεν ἀγαθῶν καὶ πλουτοῦντας ἐξαπέ-

στειλεν κενούς.
It is well known that the Magnificat (vss. 46–55) is a canticle composed

in OT style, with similarities to the Song of Hannah in 1 Kgdms 2:1–10.
It is, in Fitzmyer’s words, ‘a mosaic of OT expressions drawn from the
LXX.’ The actual combination seen here, ἐξαποστέλλω + κενόν, is
attested in the LXX (Gen 31:42; Deut 15:13; 1 Kgdms 6:3; Job 22:9).
Many commentators argue moreover that Luke has re-used a pre-existing
hymn, with his own adaptation only in vs. 48, if that. However that may
be, the Septuagintal character of the whole context and of this instance of
ἐξαποστέλλω is very distinct, and our investigation does no more than
confirm what is already clear.22
Next are the two examples in Luke 20:10, 11. Here it is useful to see
the context as a whole and to take into account the Synoptic parallels.
Luke 20:10 καὶ καιρῷ ἀπέστειλεν πρὸς τοὺς γεωργοὺς δοῦλον ἵνα
ἀπὸ τοῦ καρποῦ τοῦ ἀμπελῶνος δώσουσιν αὐτῷ· οἱ δὲ γεωργοὶ
ἐξαπέστειλαν αὐτὸν δείραντες κενόν.
11 καὶ προσέθετο ἕτερον πέμψαι δοῦλον· οἱ δὲ κἀκεῖνον δείραντες
καὶ ἀτιμάσαντες ἐξαπέστειλαν κενόν.
12 καὶ προσέθετο τρίτον πέμψαι· …
13 εἶπεν δὲ ὁ κύριος τοῦ ἀμπελῶνος· τί ποιήσω; πέμψω τὸν υἱόν
μου τὸν ἀγαπητόν·
|| Mark 12:2 καὶ ἀπέστειλεν πρὸς τοὺς γεωργοὺς τῷ καιρῷ δοῦλον …
3 καὶ λαβόντες αὐτὸν ἔδειραν καὶ ἀπέστειλαν κενόν.
4 καὶ πάλιν ἀπέστειλεν πρὸς αὐτοὺς ἄλλον δοῦλον· κἀκεῖνον ἐκεϕα-
λίωσαν καὶ ἠτίμασαν.
5 καὶ ἄλλον ἀπέστειλεν· …
6 ἔτι ἕνα εἶχεν υἱὸν ἀγαπητόν· ἀπέστειλεν αὐτὸν ἔσχατον …
|| Matt 21:34 ὅτε δὲ ἤγγισεν ὁ καιρὸς τῶν καρπῶν, ἀπέστειλεν τοὺς
δούλους αὐτοῦ …
35 καὶ λαβόντες οἱ γεωργοὶ τοὺς δούλους αὐτοῦ ὃν μὲν ἔδειραν, ὃν
δὲ ἀπέκτειναν, ὃν δὲ ἐλιθοβόλησαν.
36 πάλιν ἀπέστειλεν ἄλλους δούλους …
37 ὕστερον δὲ ἀπέστειλεν πρὸς αὐτοὺς τὸν υἱὸν αὐτοῦ …
It can be seen that the other Synoptics use only one word and repeat it
as often as necessary, but Luke introduces variation. His resources include
all three ‘send’ words: their word ἀποστέλλω, the old standby πέμπω,
and his LXX-flavoured ἐξαποστέλλω. Where Mark had simply ἀπέστει-
λαν κενόν, he changes the verb to ἐξαποστέλλω, thus securing a match
22
Fitzmyer, Luke I–IX, esp. 358–62; quotation 359. Cf. Marshall, Gospel of Luke, 77–9,
82–5.
with the LXX (and an echo of Luke 1:53), and then repeats the expres-
sion in the next verse in place of what Mark had.23 It appears that Mark’s
ἀπέστειλαν κενόν provided the trigger for Luke to introduce a Septua-
gintalism. We notice also Luke’s introduction into the context of the obvi-
ous Septuagintalism προσέθετο + inf., ‘do x again’ (11, 12), which is not
present in the parallels.24
In Acts there are seven examples. In some, though not all, there is
discernible adaptation to context in the choice of this word. Three of the
examples are in speeches, which are usually significant contexts in Acts.
For one thing, they may contain concentrations of OT quotations, which
are part of the characterisation of the speaker as a figure in ‘biblical history’
speaking in a biblical style. This is especially noticeable in the speeches
in 1–4, 7 and 13.25 Besides quotations proper, OT allusions and linguistic
features may be introduced for a similar purpose. All this is one aspect of
the generally higher register of the speeches in Acts.
Thus we have in the speech of Stephen before the Sanhedrin:
Acts 7:12 ἀκούσας δὲ Ἰακὼβ ὄντα σιτία εἰς Αἴγυπτον ἐξαπέστειλεν
τοὺς πατέρας ἡμῶν πρῶτον.
Here the content of Gen 42:2 is summarised. But ἐξαποστέλλω is not
actually used in that place in the LXX. Luke introduces it, so producing
a Septuagintal turn of phrase. We can almost watch Luke at work.
Similarly in Paul’s speech in the synagogue of Antioch, ἐξαποστέλλω
plays a part in the ‘biblical’ (and elevated) effect:
Acts 13:26 Ἄνδρες ἀδελϕοί, υἱοὶ γένους Ἀβραὰμ καὶ οἱ ἐν ὑμῖν
ϕοβούμενοι τὸν θεόν, ἡμῖν ὁ λόγος τῆς σωτηρίας ταύτης ἐξαπε-
στάλη.
Another example occurs in the speech of Paul from the steps of the
temple in Jerusalem:
Acts 22:21 καὶ εἶπεν πρός με· πορεύου, ὅτι ἐγὼ εἰς ἔθνη μακρὰν
ἐξαποστελῶ σε.
The Septuagintal style can be seen as appropriate to God as the speaker
of the words Paul is reporting, as well as the right style for the audience.
23
The distinctiveness of ἐξαποστέλλω κενόν is hightlighted by the fact that ἀποστέλλω
is never used in this combination in the LXX.
24
Cadbury (Style of Luke, 167) sees the change to ἐξαποστέλλω in vs. 10 as due to
Luke’s preference for compound verbs; this is an undeniable feature of Luke’s style,
but not the main factor here. Marshall (Gospel of Luke, 729) aptly comments on vs. 11:
‘πέμπω is a literary variant for ἀποστέλλω, diff. Mk.’
25
Cf. esp. Arnold, ‘Luke’s Characterizing Use,’ with references there.
A similar use of ἐξαποστέλλω, in that God is again the one doing the
sending, is in:
Acts 12:11 καὶ ὁ Πέτρος ἐν ἑαυτῷ γενόμενος εἶπεν· νῦν οἶδα ἀλη-
θῶς ὅτι ἐξαπέστειλεν [ὁ] κύριος τὸν ἄγγελον αὐτοῦ καὶ ἐξείλατό με
ἐκ χειρὸς Ἡρῴδου.
The other examples in Acts are in narrative, as follows:

Acts 9:30 ἐξαπέστειλαν αὐτὸν εἰς Ταρσόν.
Acts 11:22 ἐξαπέστειλαν Βαρναβᾶν [διελθεῖν] ἕως Ἀντιοχείας.
Acts 17:14 εὐθέως δὲ τότε τὸν Παῦλον ἐξαπέστειλαν οἱ ἀδελϕοὶ
πορεύεσθαι ἕως ἐπὶ τὴν θάλασσαν.
The lexical meaning is not quite the same in all of these. In 11:22 we
have simply ‘despatch’ a person to perform some action or mission, as
in 7:12; 12:11; 22:21 above; but in 9:30 and 17:14 it is something like
‘see (a person) off on his/her way,’ a LXX usage seen, e.g., in Gen 26:31;
31:27; 1 Kgdms 9:19, 26. Whether or not there is anything significant
here in terms of location, the use of the word in this sense confirms Luke’s
intimacy with LXX usage.26
Outside Luke–Acts there are only the two examples, in Galatians, which
occur in the same context and are close in thought and wording:
Gal 4:4 ὅτε δὲ ἦλθεν τὸ πλήρωμα τοῦ χρόνου, ἐξαπέστειλεν ὁ θεὸς
τὸν υἱὸν αὐτοῦ, γενόμενον ἐκ γυναικός. … 6 ῞Οτι δέ ἐστε υἱοί, ἐξα-
πέστειλεν ὁ θεὸς τὸ πνεῦμα τοῦ υἱοῦ αὐτοῦ εἰς τὰς καρδίας ἡμῶν
κρᾶζον·
There is no reason for Paul to introduce this word instead of his usual
ἀποστέλλω and πέμπω (note Rom 8:3 ὁ θεὸς τὸν ἑαυτοῦ υἱὸν πέμψας),
unless one wishes to argue that he chose it specially to suit this weighty
pronouncement. But it would remain a little odd that Paul used the word
only here. The explanation probably lies elsewhere. Other lines of enquiry
by NT scholars have led to the conclusion that Paul incorporates a pre-
Pauline statement into vs. 4b, and possibly another in vs. 6, or his own
adaptation of the preceding one.27 ἐξαποστέλλω may not be Paul’s usage,
then, but someone else’s. On the basis of what we have established about
the character of the word, it would seem to be a case of deliberate choice
26
In 17:14 there is an element of variatio, since ἐκπέμπω is used in the same way shortly
before (vs. 10 οἱ δὲ ἀδελϕοὶ εὐθέως διὰ νυκτὸς ἐξέπεμψαν τόν τε Παῦλον καὶ τὸν
Σιλᾶν εἰς Βέροιαν).
27
See Betz, Galatians, 205–10; Bruce, Galatians, 194–5; Longenecker, Galatians, 166–
74; Lambrecht in EDNT, 2:2–3; Tarazi, Galatians, 201–8.
of ‘biblical,’ i.e., Septuagintal, vocabulary by the composer of the state-

ment (a parallel with Wis 9:10 has been noted).
Finally, two other occurrences not counted above may be briefly noticed:
Luke 24:49 καὶ [ἰδοὺ] ἐγὼ ἀποστέλλω τὴν ἐπαγγελίαν τοῦ πατρός
μου ἐϕ᾽ ὑμᾶς·28
εξαποστελλω ‫ א‬2 Β Δ 33. … ‫ آ‬εξαποστελω L 892 ‫ آ‬txt ो75 ‫ *א‬Α C
DΚΝWΓΘΨ…ै
The reference is to the sending of the Spirit; hence there is a connexion
to Gal 4:6 just discussed. But given the uncertainty of the text, it is hard
to see what might follow. As to the choice of reading, one could argue
that ἐξαποστέλλω was introduced because it was felt to be appropriate
to Lucan style in such a context.
Mark 16:8 [[Shorter Ending: μετὰ δὲ ταῦτα καὶ αὐτὸς ὁ Ἰησοῦς
ἀπὸ ἀνατολῆς καὶ ἄχρι δύσεως ἐξαπέστειλεν δι᾽ αὐτῶν τὸ ἱερὸν καὶ
ἄϕθαρτον κήρυγμα τῆς αἰωνίου σωτηρίας. ἀμήν.]]
This occurrence, though in a passage unlikely to be a genuine part of
Mark’s Gospel, leads us into the next stage in the history of our word
in Christian circles. I have suggested above that ἐξαποστέλλω ‘entered
the vocabulary of ecclesiastical writers under the influence of the biblical
texts.’ The Shorter Ending, whatever the date of writing, clearly shows
how ἐξαποστέλλω might be used in compositions in the biblical style in
the early Church.29
Another sign of this is to be seen in the occurrences in the ‘Apostolic
Fathers.’ A couple of samples will give an idea of the flavour of the pas-
sages. The model now is not only the LXX but the NT. Thus 2 Clement 20:5
(II AD):
Τῷ μόνῳ θεῷ ἀοράτῳ, πατρὶ τῆς ἀληθείας, τῷ ἐξαποστείλαντι ἡμῖν
τὸν σωτῆρα καὶ ἀρχηγὸν τῆς ἀϕθαρσίας, . . . αὐτῷ ἡ δόξα εἰς τοὺς
αἰῶνας τῶν αἰώνων. ἀμήν.30
Hermas, Vis. 4.2.6 (II AD):
πιστεύσατε τῷ κυρίῳ, οἱ δίψυχοι, ὅτι πάντα δύναται καὶ ἀποστρέϕει
τὴν ὀργὴν αὐτοῦ ἀϕ᾽ ὑμῶν καὶ ἐξαποστέλλει μάστιγας ὑμῖν τοῖς
διψύχοις.31
28
[Text and app. crit. NA28.]
29
Note this characterisation of the Shorter Ending in Metzger, Textual Commentary, 125–
6: ‘Besides containing a high percentage of non-Markan words, its rhetorical tone differs
totally from the simple style of Mark’s Gospel.’
30
Ed. K. Bihlmeyer (Tübingen, 1956).
31
Ed. M. Whittaker (Berlin, 1967). Others at Sim. 8.6.3; 9.14.3. It is notable that God is
the sender in all of these.
Subsequently, ἐξαποστέλλω enters into liturgical texts, and not just in

quotations, so clearly it had become established in the vocabulary of the
Church. So in the Divine Liturgy of St John Chrysostom:
καὶ ἐπὶ πάντας ἡμᾶς τὰ ἐλέη σου ἐξαπόστειλον.
And in this poetic composition of VII AD:

Λελυμένους σοὺς δεσμίους, Σιών, ἐξαπόστειλον καὶ ἐκ λάκκου
ἀγνωσίας ἀνύδρου ἐξάγαγε·32
CONCLUSION
The word ἐξαποστέλλω was a new formation that entered Koine Greek
in the third century BC. Unlike numerous other newcomers of the period,
it was not destined to be the vernacular replacement for some older obso-
lescent word, but remained less common than its near synonyms ἀπο-
στέλλω and πέμπω. While it was close to both in lexical meaning, there
are signs that its connotation differed, even from the beginning, in the
direction of greater formality. The Pentateuch translators of the LXX took
it up and made extensive use of it, followed by other translators, thereby
making it a characteristic LXX word. It continued in use until the end of
the first century BC, particularly in the formal language of inscriptions and
in the ‘literary Koine’ of writers like Polybius, but thereafter it declined,
so that by the first century AD its occurrence in ordinary Koine Greek was
unusual. The NT author Luke took it up not from the language of his day
but from the LXX, whose language he extensively imitated. It imparted a
‘biblical’ flavour, whether by being part of a LXX expression or by being
in itself an old-fashioned word associated with the LXX. In Luke’s writ-
ings it is to some extent significantly located, in that it appears in contexts
where its flavour plays a role in the general effect. Two examples outside
Luke, in Galatians, are also likely to derive from Septuagintal language.
After the NT the word quickly entered the language of Christian writers as
a NT as well as LXX word, and continued into ecclesiastical Greek of the
32
The first is in the Holy Anaphora in the section beginning Μνήσθητι, Κύριε, τῆς πόλεως
ταύτης, the second in the Kanon (Ode 6) for Orthros of Palm Sunday, attributed to
Kosmas the Monk (or, according to some, St Andrew of Crete). ἐξαποστέλλω continued
into medieval Greek and survives today in modern Greek (demotic form ξαποστέλνω).
It is unclear what role may have been played by ecclesiastical language or puristic Greek.
It seems significant that in the modern Greek versions of the NT by Kolitsaras (Athens,
1964) and the UBS (Athens, 1989), ἐξαποστέλλω is always replaced, mostly by στέλνω.
For medieval Greek, see Kriaras, Λεξικό, s.v. εξαποστέλλω. For modern, see, e.g.,
Lex. Proia; Mega Lexikon; Babiniotes, Λεξικό; all s.v. ξαποστέλνω.
patristic period. It also survived to be used sporadically in late Koine,

especially in official contexts.
Some reflections on method may be offered. In NT language study it is
usual to regard the usage of the LXX as having a major, sometimes cru-
cial, role to play in understanding the NT, and rightly so. But it is never a
simple matter of comparing the one with the other. The ‘third ingredi-
ent,’ the usage of the language contemporary with both, must be taken into
account. Before any feature of NT language can be categorised as a Septua-
gintalism, or as anything else for that matter, it is necessary to know what
non-biblical Greek usage was at the time of the NT. A moment’s reflection
will show that this knowledge is a prerequisite in all such attempts at cate-
gorisation, otherwise we might simply be looking at a feature common to
the NT and Koine Greek. The frequent NT expression ἀποκριθεὶς εἶπεν,
for example, is not recognised as a Septuagintalism simply because it is
common in the LXX, but also because it is unlike anything that normally
occurs in Greek at any time; conversely no one would suggest that λέγει,
‘he says’ is a Septuagintalism, even though it is common in the LXX as
well as the NT, because we know it is ordinary Greek.
This methodological underpinning, though obvious once pointed out,
is easy to neglect and has often been inadequately followed in practice,
partly because of the difficulty of gathering the data. Today the situation
has changed. Electronic resources make it possible to assemble fuller data
than ever before. A truer picture of the usage of any word can be seen, as
ἐξαποστέλλω has demonstrated. This potential for improved information
on the state of the language cannot be neglected in future study.
A further important point is evident. Koine Greek was not a static entity.
Over the long span of 900 years change was inevitable. This too is obvi-
ous, but it is easy to forget to look for change when studying some LXX
or NT feature. In the past there has been a tendency to rest content with
establishing that a feature was found in Koine Greek, as though it would
have been spread evenly through all periods. Attention to the progress of
the language is also needed. There are many signs that in the 300 years
between the LXX and the NT there were important shifts in usage. In fact
there are grounds for dividing Koine Greek at the point where I BC ends
and I AD begins, as Thumb long ago suggested.33 Our word ἐξαποστέλλω
has proved a good instance of the changes that could occur from one
period to the next.
33
Thumb, Hellenismus, 9–10. Cf. Moulton, Grammar, I, 41. It is time to go further, I think,
and recognise a division into Early (III–I BC), Middle (I–III AD) and Late Koine (IV–
VI AD).
Postscript
The proposed division of Koine Greek into three periods, put forward in passing
in footnote 33, has been taken up by several scholars in the field. It is proving a
useful adjunct to the study of post-Classical Greek.34
34
See Aitken, ‘Phonological Phenomena,’ 262, 277; Bentein, ‘Perfect Periphrases,’ 207;
La Roi, ‘Diachronic Analysis,’ 230; ‘Insubordination,’ 21; Stolk, ‘Dative and Genitive,’
188 n. 10.
13
A LEXICOGRAPHICAL DATABASE
FOR GREEK: CAN IT BE FAR OFF?
THE CASE OF AMPHODON
2008
Abstract
This paper was presented in Wuppertal in 2006, in a session on the possible

contribution of databases to LXX studies. I chose the case of ἄμϕοδον as an
illustration. Information on this word in lexicons and scholarly discussions and
the data on its occurrences have never been systematically collected or analysed.
The meaning of ἄμϕοδον consequently remains unsettled and false meanings have
been perpetuated. There are two occurrences in the LXX in Jeremiah, which I
discuss. An interim lexical analysis is offered. The overall conclusion is that the
only solution to the problem of scattered information is to store data on each word
in an ongoing electronic database.
A partial database of the Septuagint vocabulary already exists. It was

created by James Aitken, initially as part of the AHRB Greek Bible Project.
From that starting-point it has continued to be developed and will shortly
be available on line as the ‘Demetrios database.’1 The benefits of the data-
base approach are already evident from the samples available. But this
paper is designed to take matters a step further. A database for the LXX,
valuable though it is, inevitably constitutes the first step towards a data-
base that covers all of Greek, which is also an important desideratum.
Once electronic data are collected on any word in the LXX, data have also
been collected for the study of that word wherever else it may occur in
Greek. Conversely, any collection of data related to the occurrence of
a word in some other text outside the LXX offers data of value to the
Septuagintalist if the word also occurs in the LXX. Instead, therefore, of
1
[Archived at: https://ota.bodleian.ox.ac.uk/repository/xmlui/handle/20.500.12024/2528
(Jim Aitken).]
creating different databases for different Greek texts and keeping things
in separate compartments, there is everything to be gained by aggregating
the data and serving the needs of everyone at once. Given the nature of
the electronic medium, there is nothing to stop us doing so.
As to the importance of such a tool, let me briefly describe the situa-
tion in Greek lexicography generally and mention previous contributions I
have made on this topic. In a short paper presented in 1995 (but not pub-
lished until 2004), I surveyed the state of lexicography of ancient Greek
and pointed out that ‘the coverage is partial, unsystematic, and uneven
in quality.’ Ideally, what is needed is not a series of updated lexicons but
‘a complete lexical treatment of Greek from its beginnings right through
to the end of Byzantine Greek … making use of all the occurrences, to
ensure that all the evidence is brought to bear at once.’ I made only ten-
tative suggestions about an electronic database. One of them was that we
could establish ‘a world database to collect the results of all published
work on individual New Testament words,’ this being not just a bibliog-
raphy but a summary of results. I also considered the possibility of using
the recently created electronic LSJ as the basis of ‘an LSJ database on line
for anyone to add to or comment on.’2 I was, however, wary of this because
of the problem of sifting and organising the contributions.3
Much more developed ideas are to be found in my History, where I put
forward proposals for the direction that NT lexicography might take in
the future. Fundamental to my view is the creation of an electronic resource
for gathering the material. The goal would remain the same, that is, ‘the
practical one of delivering an accurate description of the meaning of each
Greek word in the New Testament,’ but the database would be an ongoing,
long-term means of assembling and assessing all the data relevant to the
task.4
Similar ideas, applied to the lexicography of ancient Greek as a whole,
were developed in my paper ‘Releasing LSJ from its Past,’ designed to
address the question: what is the future of LSJ?5 This represents my full-
est views so far. There are serious flaws in LSJ, despite its great value.
2
Lee, ‘Present State,’ 67, 72–3, 74, 73 [= Essay 9]. That paper was presented at the SBL
Annual Meeting 1995 in Philadelphia.
3
My remarks misled at least one person into thinking I was advocating ‘a kind of lexico-
graphical chat-room’ – which is far from what I intended: see Büchner, Review of LEH,
Lexicon, 147.
4
Lee, History, ch. 11; quotation 182.
5
Delivered in 2002 at Cambridge Colloquium on Ancient Greek Lexicography; MS for
publication 27.10.03; a volume of colloquium papers is projected. [Published 2010 =
Essay 17.]
A LEXICOGRAPHICAL DATABASE FOR GREEK 201
They are the result of the constraints, both practical and methodological,
within which earlier lexicographers worked. An old-style revision would
not solve them but perpetuate them. The way forward is by the creation of
an electronic database to assemble and reassess all the data on every word
in ancient Greek.
II
I turn now to an illustration, taking a word in the LXX as my starting-

point. From it we see how inadequate is the information currently con-
tained in our lexicons, how extensive the data available if all sources are
tapped, and how useful a full collection would be for deciding the mean-
ing of the word, in the LXX or anywhere else.
The word is ἄμϕοδον, which occurs twice in the LXX, translating
Hebrew ’armôn:
Jer 17:27 καὶ ἀνάψω πῦρ ἐν ταῖς πύλαις αὐτῆς, καὶ καταϕάγεται
ἄμϕοδα Ιερουσαλημ καὶ οὐ σβεσθήσεται.
Jer 30:16 (MT 49:27) καὶ καύσω πῦρ ἐν τείχει Δαμασκοῦ, καὶ
καταϕάγεται ἄμϕοδα υἱοῦ Αδερ.
It is not entirely clear what the word ἄμϕοδον means. Let us suppose
we want to find out. The first step would be to consult a lexicon, let us
say LEH, which gives the meaning as: block of houses surrounded by
streets.6 You would most probably turn then to LSJ, where you would find:
‘I. street … II. block of houses surrounded by streets …: hence, ward,
quarter of a town, LXX Je. 17.27. …’7 You might think this was LSJ’s final
word on the subject and not think to check the 1996 supplement, which
has further adjustments, no small ones.8 We are now told to delete all of
‘hence … town’ and after the Jeremiah reference to add ‘as administrative
unit.’ The upshot is rather puzzling: we lose ward, quarter, generally under-
stood to be the commonest meaning, and gain a clarification of Jer 17:27
that seems irrelevant to it. But nothing has been done about block of houses
surrounded by streets, which is a questionable meaning, though the ordi-
nary user would have no reason to suspect this, not knowing that it goes
back to the 5th ed. of Liddell and Scott (1861) and takes no account of
new evidence or discussion since that time. Its perpetuation in LEH is an
unfortunate effect of relying on LSJ for meanings.
6
LEH, Lexicon, s.v.
7
Part 1 of LSJ containing ἄμϕοδον appeared in 1925.
8
LSJ Suppl. (1996).
Let us suppose we investigate further in the lexicons. There are several

to sample. In BDAG the focus is naturally on the one NT occurrence
(Mark 11:4), given as ‘street,’ but we are also told that this comes from the
meaning ‘a city quarter.’ Sophocles’s old lexicon offers a few references,
in all of which it is taken to mean ‘street, in a city,’ including Jer 17:27.9
DGE is also to be consulted. Its meanings are 1 calle … 2 plazuela …
3 barrio … with Jer 17:27 and some documentary examples placed under
sense 3, i.e., ‘quarter.’ Montanari’s lexicon (GI) similarly assigns Jer 17:27
to isolato, quartiere, rione.10 There is also the Mega Lexikon, which has
its own take: (1) ‘road running around a neighbourhood’; (2) ‘group of
houses separated by roads from other habitation, separate neighbourhood,
quarter.’11
III
Vistas are opening up, and we seem to be finding more questions than
answers. Yet we have hardly begun. Next must be considered what is in
the literature, that is, scholarly discussions of the word in articles and
books. The lexicons give almost no clue to this: LEH (2003) lists Milli-
gan12 and one article by Tov;13 BDAG (2000) refers to MM and Preisigke,
Wörterbuch, in all a total of four items. LSJ, DGE and others attempt no
coverage of this kind. How would one know that there are two very recent
articles, a short but useful discussion by Lolos (2003)14 and a much fuller
treatment by Du Bouchet (2004)?15 Or that there is a long history of dis-
cussion of this interesting word? Among the older items is one by Buck
(1922), a master philologist whose opinion ought not to be neglected.16
Buck declared that the meaning commonly adopted in the lexicons is sim-
ply wrong: the primary notion is likely to be ‘that which is on both sides
of the street,’ so the entity is ‘the complex of houses facing a street and
forming a court which may readily be closed at each end.’
9
Sophocles, Lexicon.
10
[GE translates GI as block, quarter, district.]
11
My translation. Orlandos and Travlos (Λεξικὸν Ἀρχαίων Ὅρων) work with two similar
senses, adding a long list of references including Mark 11:4 and Jer 17:27, but without
indicating which meaning is to be assigned to them. Lampe, PGL does not cover this
word. [Muraoka, Lexicon (2009): quarter of a town.]
12
Milligan, Selections, 81.
13
Tov, ‘Three Dimensions,’ 530–1.
14
Lolos, ‘Greek Roads,’ 165–6.
15
Du Bouchet, ‘Quartiers.’
16
Buck, ‘Greek ἄμϕοδον,’ 114.
By following other trails and sources I have brought to light over two
dozen items.17 Even so, the list does not include translations and commen-
taries, or anything published before 1900. Though not all these discussions
may be useful (and some are indeed useless), they all have some relevance
and need to be assessed for their contribution. And it is not enough simply
to read the last of them: what is noticeable is that there has been no sys-
tematic accumulation of data, such as to allow one to find out where the
discussion is up to by consulting one item. The lexicons, whose task one
supposes is to do just that, are the worst of all. For reasons beyond their
control, lexicons simply cannot cover the data in a thorough and up-to-date
way. The authoritative and reliable statement that everyone supposes they
give may be anything but.
But this is the end of the story, isn’t it? By no means. We have not yet
dealt with evidence of the word in use, i.e., occurrences in ancient Greek
texts. The collection of citations in the lexicons may seem extensive, but
electronic searching reveals that there is much more available. In Greek
literature, on the basis of TLG CD-ROM E, there are over 200 occurrences;
in documents, if we use PHI CD-ROM 7, there are over 1,700 examples.
The majority of the latter are in papyri, where it is a word of common
occurrence, mostly in connexion with named ἄμϕοδα.18 This evidence has
not yet been systematically studied. Ideally, all of it needs to be gathered
and assessed to ensure that the usage of our word is correctly appreciated.
Du Bouchet’s study, the most recent in the field, offers a good summary,
but it is not a systematic presentation with lexical meanings, nor is it clear
how complete his collection is.
The nature of the problem in front of us is clear, and its solution ready
to hand. An ever-growing body of data has never been systematically col-
lected and analysed; it could all be gathered electronically in one place,
where it could be organised, discussed, and kept up to date. Furthermore,
whatever is collected in this way will be relevant to the study of ἄμϕοδον
wherever it occurs.
17
Additional to those noted already or below are the following (11 items): Garlan, ‘Cités,
armées,’ 21–2; id., Recherches, 382; Hennig, ‘Staatliche Ansprüche,’ 250–1; Hohlwein,
L’Égypte Romaine, 95–6, 334; Jouguet, La vie municipale, 136–7, 282–91; Klamp,
‘Testament des Taharpaesis,’ 113; Llewelyn, New Documents, 7:41–2; Preisigke,
Fachwörter, s.v.; Rink, Strassen- und Viertelnamen, 7–17; Robert, À travers l’Asie
Mineure, 152–4; Thackeray, Septuagint and Jewish Worship, 34.
18
Daris (‘I Quartieri’) has recently compiled a list of the known ἄμϕοδα of Oxyrhynchus,
which number over 40.
IV
Let us return to the two examples of ἄμϕοδον in the LXX. What can
be said about their meaning? Obviously it would be advisable to approach
them with a prior knowledge of the meaning in Greek generally, includ-
ing any progression over time. A final statement of this cannot be offered
here; it must await the full study of the evidence that is still required. Nev-
ertheless I put forward the following as an interim analysis and a corrective
to the misleading information in the lexicons:
1 a complex of houses facing a street and forming a court closable at
either end
2 a street in such a complex
3 one such complex or a group of them forming an administrative dis-
trict of a town, often bearing a name; a ‘quarter’19
The Hebrew original in LXX Jer 17:27 and 30:16 (MT 49:27) must be
considered next, before any attempt to decide what meaning was intended
by the translator. I note the following meanings of ’armôn from a few stand-
ard sources:
HALOT: (fortified) palace
BDB: citadel, then: citadel, castle, palace
KB, Lexicon: Wohnturm … dwelling-tower (fortified building of small
square base a[nd] several stories)
DCH: fortress … fortress, fortification
Zorell: domus elegantior et magis munita divitum, palatium [‘a more
elegant and fortified house of the rich, a palace’]20
I do not myself see clearly what the entity is, beyond a certain vague
impression. Each of the equivalents given by the Hebrew lexicons means
something slightly different. We should note that the Hebrew word in
Jeremiah is plural (like the Greek), so that there are multiples of them
in one city – which hardly suits ‘citadel’ or ‘palace’ (if the royal palace is
meant). If the translator was trying to represent one of these senses, ἄμϕο-
δον is nothing like what we would expect. More likely equivalents would
be βασίλειον, βᾶρις, πυργόβαρις, or πύργος, the first three of which do
occur elsewhere, along with other more far-fetched renderings.21 For Tov,
19
Cf. Du Bouchet, ‘Quartiers,’ 50: ‘Quelles que soient les incertitudes de détail, il faut
donc distinguer pour ἄμϕοδον trois référents distincts, … la parcelle urbaine, le quartier
et la rue.’
20
Zorell, Lexicon Hebraicum. Caird (‘Towards a Lexicon. I,’ 460–1, discussing ἄντρον)
confidently states the meaning of ’armôn as ‘the inner keep or redoubt of a fortified
enclosure or palace.’ He mentions no source or authority.
21
βασίλειον Prov 18:19; βᾶρις Ps 47(48):4 + 4; πυργόβαρις Ps 121(122):7. Other ren-
derings: ἄντρον, γῆ, ἐναντίον, θεμέλιον, ναός, οἶκος, πόλις, χώρα.
the variety of renderings indicates that the Jeremiah translator and oth-
ers did not know the meaning of ’armôn and guessed from context. Tov
nevertheless regards ἄμϕοδον (taken to be a ‘block of houses surrounded
by streets’) as close to the meaning of the Hebrew word (understood as
‘palace’), and intended to be so by the Jeremiah translator. I do not see
this: there is no similarity between an ἄμϕοδον and a palace; and if the
translator does not know the meaning of the Hebrew word, it cannot be
used as an indicator of the intended meaning of ἄμϕοδον.22 All in all, it
seems to me fruitless, in the present case, to attempt to draw out the mean-
ing of ἄμϕοδον from the Hebrew. The best guide is what the word would
normally mean in Greek, combined with the indications of the (Greek)
context. On that basis, sense 1 or 3 above would suit well, and sense 1,
the less technical, would be preferable. The picture is of fire consuming
the dwelling complexes that make up the large towns of Jerusalem and
Damascus. The translator chooses a term that would be an everyday one
and immediately meaningful to his readers, many of whom would have
lived in an ἄμϕοδον (unlike us, who have to work to imagine one).
I come now to a conclusion. One could say that all scholarly work on
Greek words up till now has been a preparation for the day when it could
be systematically collected and organised. We publish articles and books
on the assumption that everything will somehow be read and absorbed
into the scholarly consciousness, or at least taken into the lexicons at the
next updating. But we know it doesn’t happen. The process is sluggish,
haphazard, and subject to all kinds of omission and duplication, simply
because of the practical constraints. Who has time to read everything, let
alone remember it? But the day has come when a systematic collection
can become a reality. An electronic gathering-point is now possible; it is
also, as I have tried to show, necessary. Good decisions about the meaning
22
Tov has discussed this word four times, each with slight variations, but his understand-
ing of ἄμϕοδον, dependent on LSJ, does not advance (Jeremiah and Baruch, 33; ‘Three
Dimensions,’ 530–1; ‘Did the Septuagint Translators?,’ 56–8; ‘The Septuagint,’ 164).
Du Bouchet’s discussion of the LXX examples (‘Quartiers,’ 47–8) is hampered by the
attempt to get the meaning to match the Hebrew, taken to be ‘palais.’ Case’s discussion
(‘The Use of ἄμϕοδον’) achieves little. Like Tov, Caird (‘Towards a Lexicon. I,’ 461)
noted the variety of renderings as an indication that the word ’armôn ‘gave the LXX
translators a great deal of trouble.’ The variety of definitions in the Hebrew lexicons
suggests its meaning is still uncertain today.
of words cannot be made without taking full account of the available dis-
cussions and evidence, and this is not happening with current methods.
Even if one were simply to take the bibliography noted in this paper, enter
it into a database and keep it up to date, it would be a big step forward.23
At least the next person to work on this word would not have to do it all
again. If a list of the occurrences of ἄμϕοδον, merely that, were added as
well, we would be well on the way to a major resource for lexical study.
My paper title is cast in the form of a question. Let me end with another.
If, as seems certain, the creation of a database for Greek lexicography
would be of great value, is there any reason why it can’t be done?
Postscript
There has been no answer to my final question, and (as of 2021) no news of
the creation of a database along the lines suggested.
23
An existing print resource attempting this is the RBLG (1998). The problem is that it
cannot be readily updated or augmented. In the case of ἄμϕοδον it records only three
items. [RBLG Suppl. 3 (2008) (online) adds two more, Lolos and Du Bouchet.]
14
THE COMPLUTENSIAN POLYGLOT,
THE TEXT OF SIRACH,
AND A LOST GREEK WORD
2009
Abstract
An otherwise unknown word, πάλαθος, recorded in the Complutensian lexi-

con of the NT, Wisdom, and Sirach (1514), is tracked down and explained in
this paper. It turns out to occur in Sirach at 24:15 in one part of the MS tradi-
tion known as ‘Gk II,’ found in MS 248, which the Polyglot editors used for
their text of Sirach. Though πάλαθος is probably a corruption of a similar word in
Sir 24:15, its status as a ‘real word’ is worth considering. Comparison of the Old
Latin leads to a proposal for an improved reading of another word accompany-
ing πάλαθος in the text of Sir 24:15 in the Göttingen edition based on ‘Gk I.’
Some other instances of weakly attested words occurring in the Gk II variants
are noted as useful additions to the lexicographical record.
This paper tells a tale of detection. It all started with an entry for an
unknown Greek word in a forgotten lexicon. The quest for an explanation
led to the differing recensions of the book of Sirach, to a textual prob-
lem in the Greek text, a suggestion for amendment in the light of the OL,
the discovery of a nest of unrecorded words, and some lessons for Greek
lexicography.1
The ‘forgotten lexicon’ is not really forgotten, just old and not well
known. It is found in the great Complutensian Polyglot, printed at Alcalá
(Latin Complutum) in Spain in 1514–1517. This six-volume work presents
the biblical texts in their original languages, together with the ancient
versions. In volume 5 (1514) the Greek NT is printed for the first time;
the volume also contains a lexicon of the NT, another first. This lexicon,
1
This is an expanded version of a paper delivered at the SBL International Meeting
in Vienna, in July 2007. I am grateful to the participants for their comments, especially
Anneli Aejmelaeus for drawing my attention to O’Connell’s book (see n. 8), and to
Michael Curran for reading the final version. The BIOSCS reviewers’ comments have
also contributed to its improvement.
somewhat surprisingly, sets out to cover not only the NT but also the
two wisdom books in the Greek Apocrypha, namely, Wisdom and Sirach
(Ecclesiasticus). Each entry in the lexicon gives the Greek word, its forms,
and a Latin equivalent, but not references. The number of entries totals over
9,000. My interest in this lexicon originated in the investigation under-
taken for my history of NT lexicons.2 I am now engaged in preparing a
new edition of the lexicon with a full study of its content.3
A MYSTERY WORD
In the Complutensian lexicon the following entry appears:

πάλαθος. ου. ὁ. Massa. et πάλαθος ἀρωμάτων. vulgo poma.
We have first the headword, πάλαθος, with an indication of the genitive

(-ου) and gender (masc.), then the meaning, ‘lump/cake,’ followed by a
phrase in which it occurs, namely, πάλαθος ἀρωμάτων, ‘cake of spices,’
and finally a comment, ‘commonly [applied to] fruits.’ The quotation of
a context is unique in the lexicon, and additional comments are almost as
rare. Apart from that, there is nothing unusual about the entry. But the
problem is, the word πάλαθος does not exist. If we look in LSJ at the point
where we would expect it, we find this:
παλάθ-η... ἡ, cake of preserved fruit, Hdt. 4.23, Thphr. HP 4.2,10, LXX
1 Ki. 25.18, al., Amynt. ap. Ath. 11.500d, Luc. Pisc. 41, Vit. Auct. 19.
-ιον, τό, Dim. of foreg., Polem. Hist. 88; cf. παλάσιον. -ίς, ίδος,
ἡ, = foreg., Ph. Bel. 89.28, Str. 2.3.4. -ώδης, ες, like a παλάθη,
Dsc. 1.67.
This is a nice little word-group, and clearly the one to which πάλαθος
belongs – or would belong if it existed – but πάλαθος is not there and is
apparently unknown. A search of other lexicons, old and new, failed to
find any entry for it.4 So why is πάλαθος entered in the Complutensian
lexicon?
2
Lee, History, esp. 45–51; for Polyglot title details see 329–30.
3
[Collecting the data was in progress at the time of this paper but was diverted by work
for the Grinfield Lectures on the LXX and the ensuing book, The Greek of the Pentateuch
(2018). Some fruits of that research may still be hoped for.]
4
Such as Hesychius (ed. Hansen); Etymologicum Magnum (ed. Gaisford, 1848); Suda
(ed. Adler); Συναγωγὴ Λέξεων (ed. Cunningham); Kriaras, Λεξικό; Sophocles, Lexicon.
A TLG search also did not produce any examples (though many of παλάθη). It was only
at the final stage of preparing this paper that I checked Schleusner, Lexicon in LXX (1822)
and found that he includes it, from the Complutensian text: ‘πάλαθος, pila vel massa,
THE COMPLUTENSIAN POLYGLOT, THE TEXT OF SIRACH 209
The NT can quickly be ruled out as the source. A glance in a concord-

ance, if such were needed, establishes its absence. This leaves Wisdom
and Sirach. But the usual tools fail to help: πάλαθος does not appear in
those books or any others, as far as the concordances know.5 It was at this
point that I began to learn more about the Greek text of Sirach, in particu-
lar the form of it printed in the Complutensian Polyglot (in vol. 3, 1516),
and this led to the solution. The text of Sirach in the Polyglot is based on
a MS that represents a markedly different version from that in the stand-
ard editions of Rahlfs and Ziegler, namely, MS 248, on which more will
be said shortly. It is in that other version that πάλαθος occurs, at Sir 24:15,
and accordingly in the Polyglot printed text, which reads as follows:
Ὡς κιννάμωμον καὶ ὡς πάλαθος ἀρωμάτων,
καὶ ὡς σμύρνα ἐκλεκτὴ ἔδωκα εὐωδίαν.
Ὡς χαλβάνη καὶ ὄνυξ καὶ στακτή,
καὶ ὡς λιβάνου ἀτμὸς ἐν σκηνῇ.
The text in Ziegler, on the other hand, based on a majority of MSS includ-
ing the major uncials, is rather different:6
ὡς κιννάμωμον καὶ ἀσπάλαθος ἀρωμάτων
καὶ ὡς σμύρνα ἐκλεκτὴ διέδωκα εὐωδίαν,
ὡς χαλβάνη καὶ ὄνυξ καὶ στακτὴ
καὶ ὡς λιβάνου ἀτμὶς ἐν σκηνῇ.
The older editions of Rahlfs (1935) and Swete (1891) likewise read
ἀσπάλαθος ἀρωμάτων (+ δέδωκα ὀσμήν). Since most concordances and
lexicons depend on this form of the text, the reading of the Polyglot text
is not covered; hence the absence of πάλαθος.7
THE LEXICON ENTRY
The lexicon editor included πάλαθος in the lexicon because it was in

the Polyglot text, and he quoted the phrase in which he found it. He also
i.q. παλάθη. Sir. XXIV.15 sec. Compl. ὡς πάλαθος.’ He goes on to suggest that ἀσπάλα-
θος should be read.
5
That is, HR and HR Suppl.; Auwers, Concordance. There is no mention of πάλαθος in:
Ziegler, ‘Zum Wortschatz’; idem, ‘Ursprüngliche Lesarten’; Smend, Griechisch-Syrisch-
Hebräischer Index; Wagner, Septuaginta-Hapaxlegomena. For further discussion of the
indexing problem in Sirach, see Ziegler, ‘Vokabel-Varianten,’ 188–90.
6
Ziegler (ed.), Sapientia Iesu Filii Sirach (Göttingen ed.).
7
Even Auwers’s Concordance to Gk II does not fully cover the Polyglot text of Sirach:
it covers Gk II as printed in Ziegler’s text, with some (most?) of the variants of 248, but
not all of them. I eventually found πάλαθος when I thought to look for ἄρωμα in Sirach.
gave it a meaning. How did he know what it meant? Before answering

that, we must go back a step and ask where he found the text that he
worked from. The lexicon was printed in the fifth volume dated Janu-
ary 1514; the text of Sirach is in the third volume, printed later, certainly
after May 1515 and most likely at the end of 1516.8 It is improbable that
the sheets of the third volume were already printed and accessible three or
more years earlier. The next possibility is that the editor worked from the
fair copy, that is, a final handwritten copy prepared by the text editor for
the printer to work from. This too seems unlikely: a fair copy was prob-
ably made, but not as early as 1513.9 In fact the answer is that the lexicon
editor compiled his lexicon direct from the MS of Sirach (and Wisdom).
This MS is 248, which has long been known to be the basis of the text
of Sirach and Wisdom printed in the Polyglot.10 A comparison of entries
in the lexicon with 248 and the printed text proves that the lexicon editor
worked from the MS: in certain instances the editor who prepared the text
for printing in the Polyglot made changes to what was in 248, or a type-
setting error occurred, but these deviations are unknown to the lexicon
editor and he enters only the original reading of 248.11
The lexicon editor, then, compiled his word-list for Sirach from MS 248.
Deciding the meanings of the words came next. For help with this he
had very limited resources. In the Introductio to the lexicon some are
mentioned, among them the lexicon of Cyril, Suidas (or the Suda), and
the Etymologicum Magnum; in addition it can be shown that the editor
made extensive use of another current work, the Greek–Latin lexicon of
Crastonus.12 Apart from these he had Jerome’s Vulgate (= OL in Sirach),
that was all. For a word resembling πάλαθος the information at his dis-
posal was:
8
Cf. O’Connell, Ancient Sources, 6–7.
9
See O’Connell, Ancient Sources, 144, on the likelihood of an editor’s fair copy of Sirach.
The task of type-setting the Greek text and Latin interlinear matching word-for-word
would, I think, make a fair copy essential.
10
MS 248 (Holmes and Parsons) is a minuscule of the thirteenth century in the Vatican
Library (Vat. gr. 346), lent to Cardinal Ximénes for the editing of the Polyglot. It was
the primary source for all the LXX books in vol. 3. See Ziegler, Sirach, 42; O’Connell,
Ancient Sources, 127–8. An edition of 248 in Sirach is available in Hart, Ecclesiasticus
(but not without errors: Ziegler, Sirach, 53).
11
One example will suffice from many: Sir 37:11 μεταβουλίας 248; μεταβολίας Compl.,
Ziegler; μεταβουλία Lexicon (μεταβολία not in). Deviations of the Complutensian
text from 248 are quite numerous: see Ziegler, Sirach, 42, with a select list. O’Connell
(Ancient Sources, 144–5) argues that all are explicable as editorial changes to 248.
12
First published in 1478, followed by many editions; the one available in Alcalá was
probably the Aldine, i.e., [Crastonus,] Dictionarium graecum (1497).
Crastonus, Dictionarium (1497): παλάθη. ης. ἡ. massa.

Suda (ed. Adler): παλάθαι· μᾶζαι σύκων.
Etymologicum Magnum (ed. Gaisford): not in.13
Lexicon Cyrilli (MS): ?14
Vulgate (as in Polyglot, vol. 3): Sicut cinamomu[m] [et] balsamu[m]
aromatizans odore[m] dedi.
The real help obviously came from Crastonus. Though πάλαθος was
not showing up, the lexicon editor made an intelligent guess on the basis
of παλάθη in Crastonus. He had no reason to doubt it was a real word, and
took it as a variation on παλάθη. If παλάθη meant massa, πάλαθος would
probably mean the same; and it made good sense in his Greek text: ‘Like
cinnamon and like a cake of spices, and like choice myrrh I gave a pleas-
ant odour.’ His text of the Vulgate offered balsamum (‘basalm’) as the
meaning of πάλαθος, but he evidently didn’t trust it: he preferred to rely
on Crastonus. He had no reason to think of the word ἀσπάλαθος, or to
suspect that πάλαθος in his text might be a corruption. He added the remark
that πάλαθος is commonly applied to fruits (vulgo poma) on the basis of
the statement in the Suda, that παλάθαι (pl. of παλάθη) is used with refer-
ence to ‘cakes of figs.’15 He quoted the phrase in which he found πάλαθος
because it showed a rather different, though similar use. So we arrive at
the entry with which we began.
The volume containing the text of Sirach was printed subsequently, and
it presented not only the Greek text and the Latin Vulgate but an inter-
linear Latin rendering of the Greek, as in all the Old Testament volumes
of the Polyglot. It is known that in Sirach this translation was the work of
Juan de Vergara.16 It retains the Vulgate where possible, but changes the
wording to match the Greek, which is often very different. It is a fair guess
that this rendering was prepared later than the lexicon, and that Vergara
was able to make use of the lexicon, already printed in 1514. From it he
13
The first editions of both the Suda and the Et. Mag. had appeared in 1499. Hesychius
(1514) was not yet to hand.
14
The copy of Lexicon Cyrilli now in Madrid, a MS of X/XI AD, and almost certainly
the one used by the Complutensian editors, does not contain the relevant page, as far
as I can ascertain from the scanned images online. If it did, the entry was probably very
similar to that in the Suda (above): the same lemma, thought to derive from Lexicon
Cyrilli, appears in the Συναγωγὴ Λέξεων and Hesychius. [A return visit to the MS online
in the Biblioteca Histórica, Universidad Complutense de Madrid revealed the relevant
entry in the lexicon of Cyril. It is indeed the same as that in the Suda, except for the
accents: πάλαθαι· μάζαι συκῶν. (22.9.21)]
15
This itself is based on LXX examples, cf. 4 Kgdms 20:7 παλάθην σύκων, Isa 38:21
παλάθην (ἐκ) σύκων.
16
See Sáenz-Badillos, La filología bíblica, 327. He was also responsible for the transla-
tion of Wisdom and several other books.
took massa as the meaning of πάλαθος.17 The Polyglot text of Sir 24:15,
with Vergara’s interlinear rendering, is as follows:
Sicut ci[n]namomu[m] [et] sicut massa aromatum:
Ὡς κιννάμωμον καὶ ὡς πάλαθος ἀρωμάτων,
[et] q[ua]si myrrha electa dedi suaue[m] odore[m].
καὶ ὡς σμύρνα ἐκλεκτὴ ἔδωκα εὐωδίαν.
Quasi galbanu[m] [et] onyx [et] stacte:
Ὡς χαλβάνη καὶ ὄνυξ καὶ στακτή,
[et] q[ua]si thuris vapor in tabernac[u]lo.
καὶ ὡς λιβάνου ἀτμὸς ἐν σκηνῇ.
THE TEXT OF SIR 24:15
Having found πάλαθος, we could proceed at once to the question of

what status it has as a word. But the text of Sir 24:15 invites attention, and
will prove to be an interesting trail to follow. The text history of Sirach
is one of the most difficult and complex in the Greek Bible.18 It is not the
aim of this paper to make a contribution to this subject; I simply report the
current consensus, as background to a closer look at the text. The origi-
nal Hebrew version of Sirach (Hb I) was the basis of a Greek translation
(Gk I); then came an expanded version of the Hebrew (Hb II), which was
in turn the basis of a revised and expanded Greek version (Gk II). Gk I is
transmitted in the major uncials and dependent minuscules; Gk II is rep-
resented by a number of other witnesses, including 248, but neither it nor
any other MS preserves a pure text of Gk II.19
Ziegler’s text and app. crit. present the data on the MS readings in
Sir 24:15, as follows:
ὡς κιννάμωμον καὶ ἀσπάλαθος ἀρωμάτων
B S A V O L l a b c min. La verss.
ἀσπάλαθος] παλ. 248-672 46 336 534ʹ; απαλ. V 705; σπαλ. 543;
fort. κάλαμος Sm.; pr. ως O-Sc-V 248-672 46 336 534ʹ 543 Aeth ArmII
17
It is of course theoretically possible that Vergara’s translation was made first and the
lexicon editor used it, but practically very unlikely, both because of the time-frame and
the fact that the lexicon editor worked direct from MS 248.
18
Ziegler, ‘Ursprüngliche Lesarten,’ 461; Skehan and Di Lella, Ben Sira, 59. Cf. Jellicoe,
Septuagint, 306–10.
19
Ziegler, Sirach, 74; on witnesses to Gk II, 58–69.
The reading ἀσπάλαθος ἀρωμάτων adopted by Ziegler is that of BSAV

and various minuscules. The Polyglot text matches 248 and others that
share ὡς πάλαθος ἀρωμάτων. Clearly Ziegler regarded ἀσπάλαθος as
original and variants such as ὡς πάλαθος as secondary, and there is no
reason to argue with him. At some point in the tradition ὡς was introduced
before ἀσπάλαθος, leading to various corruptions including ὡς πάλαθος.
But there is still a problem in Ziegler’s text: ἀρωμάτων does not make
proper sense.
The word ἄρωμα itself (aromatic herb or spice, LSJ) is not the prob-
lem: it is well attested in Greek from early on, and occurs in the LXX and
NT. Likewise ἀσπάλαθος is a well known word, even if its meaning is
somewhat hazy: it is the term for some sort of thorny aromatic plant, and,
though found only here in the LXX, is attested from the fifth century BC
to Modern Greek.20 But how does the gen. pl. ἀρωμάτων fit syntactically?
The best one can think of is ‘among (the) spices,’ but this is awkward. At
this point it may be of interest to see how the translations have dealt with
it. The older ones reflect a text that is different from Ziegler’s and similar
to Rahlfs’s, i.e., with the additional words δέδωκα ὀσμήν, ‘I gave a sweet
smell,’ after ἀρωμάτων, but this does not affect their rendering of ἀρω-
μάτων, except in the case of the NEB (see further below):
KJV (1611): I gave a sweet smell like cinnamon and aspalathus, and
I yielded a pleasant odour like the best myrrh, as galbanum, and onyx,
and sweet storax, and as the fume of frankincense in the tabernacle.
[London: Bagster, n.d.]: I gave a sweet smell like cinnamon and
aspalathus, … = KJV
Smend (1906): Wie Zimmt und wohlriechender [Kalmus und Kassia],
und wie Myrrhenfluss duftete ich süss. Wie Galbanum und Räucher-
klaue und Stakte, und wie Weihrauch war mein Duft in der Hütte.
Giannakopoulos (1955–68): Ὡς ἡ εὐώδης κανέλλα καὶ ὁ ἀρωματικὸς
σπάλαθος δίδω καὶ ἐγὼ τὴν εὐωδίαν μου. …
NEB (1970): Like cassia or camel-thorn I was redolent of spices; I
spread my fragrance like choice myrrh, like galban, aromatic shell, and
gum resin; I was like the smoke of incense in the sacred tent.
Kolitsaras (1981): Ὅπως ἡ εὔοσμος κανέλλα καὶ ὁ ἀρωματικὸς
ἀσπάλαθος ἔδωσα καὶ δίδω ἐγὼ τὴν εὐωδίαν. …
NJB (1985): Like cinnamon and acanthus, I have yielded a perfume,
like choice myrrh, have breathed out a scent, like galbanum, onycha,
labdanum, like the smoke of incense in the tent.
20
See LSJ, s.v.; Chantraine, DELG, s.v.; Kriaras, Λεξικό, s.v.; Lex. Acad., s.v.
Skehan-Di Lella (1987): Like cinnamon, or fragrant cane, or precious

myrrh, I give forth perfume; Like galbanum and onycha and mastic,
like the odor of incense in the holy Tent.
NRSV (1989): Like cassia and camel’s thorn I gave forth perfume,
and like choice myrrh I spread my fragrance, like galbanum, onycha,
and stacte, and like the odor of incense in the tent.
NETS (2007): Like cinnamon and camel’s thorn for spices,* and like
choice myrrh I gave forth a fragrance, like galbanum and onycha and
stacte and like the vapor of frankincense in a tent.
[Fn. *Possibly of spices; + I gave off a fragrant smell = Ra.]
LXX.D (2009): Wie Zimt und Gewürzstrauch und wie ausgewählte
Myrrhe habe ich den Wohlgeruch verbreitet; wie Galbanum und
Onyx und Myrrhen-Öl und wie Duft von Weihrauch im (Heiligen)
Zelt.
Some of these, namely, KJV, NJB, NRSV, LXX.D, do not appear to ren-
der ἀρωμάτων at all: they simply pass over it.21 NETS makes the attempt,
but ‘for spices’ is not a possible meaning of the genitive and does not make
much sense. NEB, working with Swete’s text, joins ἀρωμάτων with ὀσμήν,
which is possible with that text but not Ziegler’s. Skehan and Di Lella’s
‘fragrant cane’ is not a rendering of ἀσπάλαθος (+ ἀρωμάτων?) but, as
far as I can make out, a rendering of κάλαμος εὐώδης, the phrase found in
Exod 30:23. They take it for granted that Sir 24:15 is based on Exod 30:23,
34, in the passage describing the perfumes and incense used in the service
of the Tent, and allow that passage to influence the interpretation here,
where Wisdom likens herself to a similar list of perfumes and incense.22
Smend works from the same premise, but at least his alterations are overt.23
Giannakopoulos and Kolitsaras render ἀρωμάτων as equivalent to ἀρω-
ματικός, ‘aromatic,’ which makes good sense and is what we would like
it to say, but is not an accurate rendering of what we actually have in the
text.24
21
I do not know if LXX.D’s Gewürzstrauch (‘spice-bush’) could include representation
of ἀρωμάτων.
22
Skehan and Di Lella, Ben Sira, 328, 334–5. Cf. Wright, No Small Difference, 248, with
n. 26, expressing some caution. ἀσπάλαθος is not in the Exodus passage (or anywhere
in the LXX outside Sir 24:15).
23
Translation in Smend (ed.), Weisheit des Jesus Sirach, 41; discussion 219–20. Smend
emends to match Exodus: ‘Vielleicht ist (ἀσ)πάλαθος Fehler für (ὡς) κάλαμος’ (219);
hence ‘fort. κάλαμος Sm.’ in Ziegler’s app. crit. (above). Smend’s wohlriechender
(‘fragrant’) implies ἀρωματίζων not ἀρωμάτων. He notes the OL reading without
comment.
24
Giannakopoulos, Ἡ Παλαιὰ Διαθήκη, vol. 26; Kolitsaras, Ἡ Παλαιὰ Διαθήκη, vol. 4.
All this demonstrates the difficulty of ἀρωμάτων: none of the transla-

tions has been able to make sense of it.25 There are no MS variants to the
word, so we get no help from that direction. But there is another avenue
to follow.
THE OLD LATIN
The oldest witness to Sirach, apart from the original Hebrew (not extant
in Sir 24:15), is the OL version. This predates the earliest Greek MSS,
the IV AD uncials, and is generally regarded as a witness of high value.26
What does it have as the rendering of ἀρωμάτων? It reads as follows in
a modern critical edition:27
sicut cinnamomum et aspaltum aromatizans odorem dedi
quasi myrra electa dedi suavitatem odoris
et quasi storax et galbanus et ungula et gutta
et quasi libanus non incisus vaporavi habitationem meam
The OL corresponding to ἀρωμάτων is aromatizans, a pres. part.,

agreeing with aspaltum.28 This clearly implies a Greek original ἀρωματί-
ζων, masc. pres. part. agreeing with ἀσπάλαθος, with the straightforward
meaning ‘aspalathus/camel’s thorn giving off an aroma.’29 It is hard to
see how the Latin rendering could have arisen from any other form of the
Greek, when Latin aromatizo is a rarity that appears to have been cre-
ated for this place, on the model of ἀρωματίζω (which is a normal Greek
word).
Lewis and Short cite aromatizo only here; TLL adds two examples, one in a
glossary and one in Oribasius (VI AD). The interpretation offered in Lewis and
Short, aromatizans odorem dedi (‘giving off an aroma I gave a sweet smell’),
25
There is of course no such difficulty with the Complutensian text πάλαθος ἀρωμάτων,
as understood by the lexicon editor and Vergara, i.e., ‘cake of spices.’ Interestingly, the
Geneva Bible (1560) reflects the same text: ‘I smelled as the cinnamom, and as a bagge
of spices.’
26
See Ziegler, Sirach, 14, 75; Skehan and Di Lella, Ben Sira, 56–7; Wright, No Small
Difference, 5–6; Reiterer, ‘Review of Recent Research,’ 26. The displacement of chap-
ters seen in the Greek MSS is not present in the OL, which therefore precedes all the
extant Greek MSS.
27
Vetus Latina: Sirach, at 24:20. The text of the Vulgate in Weber’s edition (Biblia Sacra)
is identical, except in having murra for myrra (at Sir 24:20).
28
For the form aspaltum (v.l. balsamum), see TLL, s.v. aspalathus. It is obviously a
(neuter) variation on aspalathus, itself derived from ἀσπάλαθος.
29
The gender of ἀσπάλαθος is commonly fem., but masc. is also found (LSJ); by Mod.
Greek the masc. is standard, as already earlier: see Lex. Acad., s.v.; Kriaras, Λεξικό, s.v.
though a possible reading of the Latin, would not be possible in the Greek original,
which would require ἀρωματίζουσα (Wisdom/Σοφία is subject). The attestation
of ἀρωματίζω includes occurrences in Aquila (see HR, s.v.). The additional
words δέδωκα ὀσμήν in most MSS (~ OL odorem dedi) are regarded by Ziegler
as secondary (‘ex 15b’); the implications for the OL reading are not clear. In
Herkenne, De Veteris Latinae Ecclesiastici capitibus I–XLIII, 191–2 the OL of
Sir 24:15 is noted without remark on aromatizans or ἀρωμάτων. Thiele (Vetus
Latina: Sirach, 24:20, app. crit. to aromatizans) had the solution almost within
his grasp: ‘habuitne interpres latinus (αρωμα)των pro participio?; non negle-
genda est lectio sequens (murra) [sic] electa (substantivum + adiectivum).’
My proposal, then, is that the original form of the text was ἀσπάλαθος
ἀρωματίζων, which was corrupted early to ἀσπάλαθος ἀρωμάτων, by
miscopying of ἀρωματίζων. The original reading of the Greek was the
basis of the OL translation, but the corruption occurred soon after, early
enough to enter our oldest Greek witnesses and to be transmitted in all
subsequent extant MSS.
For another instance where the OL preserves a better reading than all
the Greek witnesses and has been used to restore the Greek original, there
is Sir 35(32):9 ὅπου γέροντες μὴ πολλὰ ἀδολέσχει (Ziegler’s text). Here
γέροντες is not found in any of the Greek MSS, which all read λέγοντος/
ες, but it matches the OL senes. The correction was first made by the
Complutensian editor, probably by retranslation from the Latin, as Ziegler
says, and is accepted by Ziegler, with support now from the Hebrew and
the Syriac (and the Sahidic?).30
Before leaving the question of the text, it will be useful to consider what the
original Hebrew of ἀσπάλαθος ἀρωματίζων might have been. It seems likely to
have been the same phrase as in Exod 30:23, ‫( קנה־בשׂם‬lit. ‘reed/cane of per-
fume’: LXX κάλαμος εὐώδης, ‘sweet-smelling cane’), but translated differently.
A match of ἀρωματίζων with ‫ בשׂם‬is very probable: ἄρωμα often renders ‫בשׂם‬
elsewhere, in fact more frequently than εὐώδης. The rendering of ‫ קנה‬by ἀσπάλα-
θος, however, would be a one-off: the common renderings are κάλαμος and καλαμί-
σκος. Even if this means that κάλαμος was the original rendering (in Gk I?), later
changed to ἀσπάλαθος (in Gk II?), the change is too far back to justify restoring
κάλαμος to the surviving Greek text.
30
Ziegler, Sirach, 42–3. Ziegler records the support inconsistently: app. crit. at 35:9 (p. 275)
has ‘= La (ubi sunt senes) Sa: cf. H,’ but at p. 43 he says ‘… diese Lesart, die auch
dem hebr. und syr. Text entspricht. …’ An anonymous reviewer of this paper has con-
firmed that the Syriac does have ‘elders.’ Further comment on this example is found in
O’Connell, Ancient Sources, 144. At Sir 3:17 there is what seems to me a missed oppor-
tunity for improvement of the text on the same basis: the reading of the MSS is ὑπὸ
ἀνθρώπου δεκτοῦ, for which Ziegler adopts Smend’s ὑπὲρ ἄνθρωπον δότην, ‘cf. super
hominum gloriam La.’ But the OL implies ὑπὲρ ἀνθρώπων δόξαν, which gives good
sense: ‘you will be loved beyond the glory of men.’ Smend’s δότην agrees with the
Hebrew ‫ )מ(נותן‬not the OL: see Beentjes, Book of Ben Sira, 23.
THE STATUS OF ΠΑΛΑΘΟΣ
Πάλαθος, ‘lump/cake’ is found in six MSS of the eleventh to four-

teenth centuries AD, then in the Complutensian lexicon dependent on
MS 248, in Vergara’s translation dependent on the lexicon, in the Com-
plutensian text of Sirach based on MS 248, in Schleusner’s lexicon depend-
ent on the Complutensian text, and nowhere else in any lexicon or text.
Moreover it appears to be a corruption of another word. Nevertheless
it makes sense as a word in context, as read by the lexicon editor and
the text editor. As far as formation goes, it could be a word: πάλαθος
belongs to a type of formation in -θος (-θον), -θη.31 Mostly only one or
the other suffix appears, as, e.g., in the case of κάλαθος and σπάθη, but
if one exists there is potential for the other to be formed, as in: κολο-
κύνθη and κολόκυνθος; λαπάθη, λάπαθος and λάπαθον; ὄχθη and
ὄχθος; Πλαθάνη and πλάθανον. The potential for a masc. form along-
side παλάθη is therefore clear, and πάλαθος, though not attested, could
have existed.
Is πάλαθος then a real word, deserving of a place in our lexicons?
This of course raises the question of what a ‘real word’ is. Obviously,
not every misspelling and corruption in the surviving Greek MSS can be
regarded as a real word. Something else is needed. We need to establish
in some way that the word existed at some time in the language of Greek
speakers. We might appeal to the knowledge of the copyists, who copied
πάλαθος as if it was a word they knew and understood. But that is not very
reliable; they might equally well not have understood it and thought noth-
ing as they copied it. The case is different with another variant in Sir 24:15,
namely, σπάλαθος, the reading of MS 543, dated 1186 AD (see Ziegler’s
app. crit. above). The form σπάλαθος is one of the large number of vari-
ant forms of ἀσπάλαθος recorded in medieval and dialectal Greek and is
still alive today, as evidenced by Giannakopoulos’s translation (above).32
Thus the variant spelling has the support of evidence of the living lan-
guage outside the MS, and is a ‘real word.’ We do not have such evidence
for πάλαθος – yet.
For these reasons I think πάλαθος cannot count as a real word; but at
the same time I think it ought to be recorded in some way in the lexicons
– with a suitable indication of uncertainty – because if one day another
31
See Chantraine, Formation des noms, 366–8; Buck-Petersen, Reverse Index, 444–9;
Schwyzer, Grammatik, I, 510–1.
32
See esp. Lex. Acad., s.v.; cf. Kriaras, Λεξικό, s.v. There is no sign of πάλαθος among
the variant forms of ἀσπάλαθος.
example is found, the link can be made and its status upgraded.33 If one
were inclined to think that all the evidence of Greek is in, here is a
fact to ponder: the Oxyrhynchus papyri that have been deciphered and
published to date are only 1% of the total held in Oxford.34 From this
source alone, new data on the Greek language will certainly be brought
to light.
OTHER OVERLOOKED DATA
In πάλαθος we have discovered a gap in the lexicographical record, but

it is not the only one of its kind. As pointed out above, the lexicons do
not fully cover the text of Sirach preserved in the MS tradition. Data from
Gk II are not noted in LSJ (and others) because they depend on editions
based on Gk I, covering only part of the tradition. The problem affects
not only the additional verses of Gk II (printed by Ziegler in smaller type),
but Gk II variants within verses common to Gk I and Gk II. The latter type
may make their appearance only in the app. crit. of Ziegler. Sometimes we
may discover an unknown word like πάλαθος, but this is rare; more often
we gain a useful attestation of a word that is weakly attested so far. Here
are some examples.
(1) ἔκπρακτος in Sir 10:8:

φιλαργύρου μὲν γὰρ οὐδὲν ἀνομώτερον·
οὗτος γὰρ καὶ τὴν ἑαυτοῦ ψυχὴν ἔκπρακτον ποιεῖ.
This is a verse found in the representatives of Gk II and printed by

Ziegler in smaller type. ἔκπρακτος is unknown to LSJ and Suppl. and
other lexicons.35 But it is a plausible member of the group ἐκπράκτης,
ἔκπραξις, ἐκπράσσω, ‘exact payment,’ etc., i.e., an adjective meaning
‘payment-exacting.’ So we might translate ‘there is nothing more law-
less than a person who loves money, for he makes his own soul a debt-
collector.’36
33
The place for this is of course not in a printed book but an electronic database. See Lee,
‘Lexicographical Database,’ 214–20.
34
Michael Theophilos, at SSEC conference, Macquarie University, Sydney, 5 May, 2007.
The POxy series has reached vol. 72 (2008).
35
LEH, Lexicon; Lampe, PGL; Mega Lexikon; Sophocles, Lexicon, s.vv.
36
The NETS rendering ‘… makes his own soul a commodity’ is evidently influenced by
the OL (10:10) … animam suam venalem habet (venalis = ‘for sale, open to bribes’),
(2) λαλητός in Sir 18:33:

ἔσῃ γὰρ ἐπίβουλος τῆς ἰδίας ζωῆς.
+ λαλητός 248-672
The additional word found in two MSS is recorded by LSJ only in

Job 38:14 (‘endowed with speech’) and the Etymologicum Magnum (‘talked
of’); two later examples in the Fathers are noted in Lampe (‘endowed
with speech’; ‘argumentative’). An additional occurrence is not without
value. Whether or not the text is better with λαλητός added at the end,
it yields the meaning ‘for you will be talked of as a schemer against your
own life.’37
(3) πολύλαλος in Sir 21:25:

χείλη ἀλλοτρίων ἐν τούτοις διηγήσονται,
ἀλλοτρίων] πολυλάλων 248 Anton. p. 993 Mal.
This example of πολύλαλος can be added to five out-of-the-way

occurrences in LSJ, one of them in Symmachus Job 11:2. The variant
makes good sense in the text of 248, which as a whole reads: χείλη
πολυλάλων τὰ οὐκ αὐτῶν διηγήσεται, ‘the lips of the talkative will nar-
rate things not their own.’38
(4) περιψήχω in Sir 30:7:

περιψύχων υἱὸν καταδεσμεύσει τραύματα αὐτοῦ,
καὶ ἐπὶ πάσῃ βοῇ ταραχθήσεται σπλάγχνα αὐτοῦ.
περιψύχων] -ψήχων 248
LSJ’s entry for περιψήχω reads (in full): ‘περιψήχω, sine interpr., Gloss.’
Further data of any kind would obviously be welcome. LSJ’s ‘Gloss.’ refers
to a seventh century glossary in Cod. Harl. 5792.39 Sir 30:7 could be the
which was probably arrived at by (wrongly) connecting ἔκπρακτον with πρᾶσις, ‘sale,’
πιπράσκω, ‘sell,’ and related forms.
37
Et. Mag. 588.54 (ed. Gaisford) notes the word without meaning; ‘talked of’ is LSJ’s.
The KJV rendering, ‘For thou shalt lie in wait for thine own life, and be talked on,’ shows
that they, or rather a predecessor, worked from a text with the additional word. NRSV
and NETS: ‘For you will be plotting against your own life.’
38
KJV again reflects this text: ‘The lips of talkers will be telling such things as pertain not
unto them.’ Similarly NRSV: ‘The lips of babblers speak of what is not their concern.’
It appears that the Complutensian text was the basis of the KJV or an earlier English
version; cf. n. 25 on the Geneva Bible.
39
Goetz and Gundermann, Glossae, 406.
source of the lemma in that glossary. The reading περιψήχων is not to

be dismissed as a mere misspelling. Though this is the only occurrence
we know of, it is not an improbable compound and its meaning is readily
deducible as ‘wipe clean’: cf. περιψῶ (άω), ‘wipe all around, wipe clean,’
and the simplex ψήχω, ‘rub down,’ etc.40 The resulting meaning is at least
satisfactory: ‘Wiping clean his son he will tie up his wounds, and at every
cry his insides will be agitated.’ The majority reading περιψύχων, how-
ever, remains preferable, even though the meaning ‘cherish’ etc. rests on
a slender foundation.41 Another variant, περὶ ψυχῶν, found in several MSS
and reflected in the OL, is difficult to make sense of. But whatever the
merits of the reading, the variant text of MS 248 yields an occurrence of
περιψήχω worth noting in the lexicons.
(5) συγγηρῶ (άω) in Sir 11:16:

τοῖς δὲ γαυριῶσιν ἐπὶ κακίᾳ συγγηρᾷ κακία.
Another verse included by Ziegler from Gk II. The point of interest is

the attestation, for only the second time, of a present stem συγγηρῶ, as
opposed to συγγηράσκω: the latter is attested in Classical texts since
Herodotus, the former known from one occurrence in Aretaeus (II AD).42
Clearly there are data here that would be valuable in the lexicographi-
cal record but have been missed. These five words are only a sample: the
text of Sirach, with MS traditions that vary markedly, is certain to have
more to offer of the same kind; but it is likely that variant texts in other
parts of the LXX will have similar useful material. The lesson for Greek
lexicography is that standard critical texts are not the only potential source
of vocabulary items. The variant readings of those texts, the ones that end
up on the editor’s cutting-room floor, are equally worthy of attention.
40
Mega Lexikon, rightly: περιψήχω γλωσσ. ἄνευ ἑρμ. δ[ιάφορος] τ[ύπος] τοῦ περιψῶ.
The Complutensian lexicon editor made a good fist of it: circunfrico. consumo. rado.
emundo. abstergo (‘I rub around; I consume; I scrape; I clean, I wipe off’). Vergara’s
interlinear follows the lead of the lexicon with abstergens.
41
LSJ, s.v. II: ‘metaph. refresh, revive, cherish,’ citing Sir 30:7 and ‘D.H. 7.46 (cj. Reiske),
Alciphr. 1.39.’ The conjecture carries no weight; the two examples in Alciphron (Rhet.
et Soph. II/III AD), Epistulae 4.14.3, 8 (ed. Schepers, 1905) are much later and describe
a courtesan’s behaviour towards her lover. Cf. NETS: ‘When one cherishes a son, one
will bind up his wounds.’ Other versions add their own spin: KJV: ‘He that maketh too
much of his son shall bind up his wounds.’ NRSV: ‘Whoever spoils his son will bind up
his wounds.’ Skehan and Di Lella, Ben Sira: ‘Whoever spoils his son will have wounds
to bandage.’ All these renderings surely require ὁ before περιψύχων, as in vss. 1, 2, 3.
42
See LSJ. The NETS rendering ‘and evil things grow old along with those who take pride
in evil’ appears to render the v.l. κακά (to κακία).
CONCLUSION
A previously unknown word, πάλαθος, recorded in the Complutensian

lexicon of the NT, Wisdom, and Sirach turns out to exist in the MS tradi-
tion of Sirach, in one of its text-types, Gk II, of which a leading repre-
sentative, MS 248, was the basis of the text of Sirach in the Complutensian
edition. The lexicon editor worked direct from MS 248, where he found
the word at Sir 24:15, and entered it in his lexicon. He assigned a plausible
meaning with the aid of the tools of the time and some guesswork. The
word arose from a corruption and is not yet secure as a real word; never-
theless it should be placed in the lexicographical record. The same verse
contains an undetected corruption of a different word, ἀρωμάτων, which
should be restored to ἀρωματίζων on the basis of the OL. There are other
instances of unrecorded words that occur in the MS tradition of Sirach,
of which five examples have been given: ἔκπρακτος (Sir 10:8), λαλη-
τός (Sir 18:33), πολύλαλος (Sir 21:25), περιψήχω (Sir 30:7), συγγηρῶ
(Sir 11:16). Greek lexicography could record these and be more aware
generally of variant texts as a potential source of new attestations of words.
Postscript
The paper is admittedly something of a showpiece. The results are less sig-
nificant than the demonstration of what comes to light when a simple question is
asked about a word, and what a tour of manuscripts, texts, and obscure vocabulary
is required to find the answer. For the history of Greek lexicography, the paper
illustrates the challenges the Renaissance pioneers faced and the limited resources
they had at their disposal. The most useful result is not so much the explanation
of the unknown word πάλαθος (‘lump,’ ‘cake’), as the discovery of a case for
emendation of the problematic ἀσπάλαθος ἀρωμάτων in Sir 24:15 to ἀσπάλαθος
ἀρωματίζων on the basis of the OL.
15
REVIEW OF MURAOKA, A GREEK-ENGLISH
LEXICON OF THE SEPTUAGINT (2009)
2010
Abstract
This review assesses the final instalment of Muraoka’s lexicon, completing the
coverage of the whole LXX. It builds on the review (2007) of the preceding lexi-
con that covered not all, but a substantial part of the LXX. The features criticised
in the previous review are briefly revisited and changes or reactions noticed. The
review then turns to the question of Muraoka’s approach to determining mean-
ing in the LXX and how it compares with that of NETS, the new translation of
the LXX into English. NETS was preceded by lengthy discussion of method, and
this evolved over time. The aim is to reach a clear understanding of how the two
approaches differ and the issues involved, in particular the widely misunderstood
‘interlinear model’ introduced by NETS. The discussion is illustrated by examples.
The review concludes with suggestions for future development of LXX lexicogra-
phy on the basis of Muraoka’s lexicon.
In a mere twenty-five years, a short span in lexicographical time,

T. Muraoka has singlehandedly brought his lexicon of the Septuagint to
completion. After two earlier instalments, the full work now appears, cov-
ering both the translated books of the Hebrew Bible and the original Greek
works. It is not dependent on any other Septuagint lexicon but is based
on a fresh lexical analysis of the material; this too is unusual in the world
of lexicon production. Thus the desideratum of Septuagint scholars for
over a century has been fulfilled. Muraoka is to be congratulated and his
achievement celebrated.
In an earlier review in this journal I examined the previous instalment
of the lexicon in some depth, describing its general characteristics as well
as offering a critique of Muraoka’s method of stating meaning.1 That
1
Muraoka, Lexicon (Pentateuch and Twelve Prophets), reviewed in BIOSCS 37 (2004),
127–39 [= Essay 10].
material does not need to be repeated here. Readers are encouraged to

read the present review in the light of the earlier one. Here I consider
(a) what may have changed in the present lexicon, especially in response
to my earlier comments on definition; (b) the general issue of how lexical
meaning is determined in the LXX, comparing Muraoka’s approach with
that of the NETS project; and (c) future work in LXX lexicography.
The general contours of the lexicon remain unchanged from the
previous instalments. That is, it presents a list of headwords for the LXX
vocabulary in alphabetical order, with information on morphology, a break-
down into senses with full or selective listing of occurrences, information on
collocation, and finally a list of semantically related words, and references
to selected literature. One change from its predecessors is the omission of
the summary of corresponding Hebrew words, which had appeared at the
end of each entry. This is an appropriate step, and Muraoka reports that he
intends to publish the data in a separate work.2
The Introduction, which is short and to the point, will repay reading by
all who use the lexicon. It not only explains the layout of the entries but
sets out – rather too briefly perhaps – Muraoka’s approach to LXX lexi-
cography and his method of indicating meaning. The choice of the ‘defini-
tion’ method as the basis of the latter is the most important feature of this
lexicon. Whatever faults may be found with the execution, the application
of this method offers the best prospect of stating meaning unambiguously
and separating different lexical meanings, as opposed to the ‘gloss’ method
with its great propensity for imprecision.
In the earlier review I looked in some detail at Muraoka’s practice in
framing definitions, with lists of examples. I pointed out that although
the definition method was generally applied, there were many meanings
that had the appearance of glosses, either in combination with a definition,
or standing alone, or in groups of two or three. I called Muraoka’s approach
the ‘mixed method’ and noted that this method is also encountered in some
leading ‘definition’ lexicons (OED, OLD). It seemed to me that it would be
helpful to clarify the difference between a gloss and a one-word definition,
and to recognise the latter as an acceptable form of definition. Further, I
questioned Muraoka’s highly developed practice of adding explanatory
additions (in plain text) to the definition or gloss (in italic), and occasional
inconsistency in the formating of collocations.
In the present edition, Muraoka has taken the point about one-word
definitions and in the Introduction has spelt out his position (p. xii); on this
2
See ‘Introduction,’ xv. The data are retained in a few instances ‘for which HR provide
no information.’
REVIEW OF MURAOKA, A GREEK-ENGLISH LEXICON OF THE SEPTUAGINT 225
we are now in agreement, at least in principle. As regards the examples

that I thought required adjustment, changes have been made, but only to
about half. Moreover, a great many other instances of the same features
remain throughout the lexicon. Further, there is no new information on the
purpose or role of the explanatory additions, nor improved consistency in
their use. It is still not quite clear if these are part of the definition proper,
or an explanation of the definition, or an indicator of collocation (or a
combination of these). Collocations also remain inconsistently formated.
It will be useful to note a few examples to illustrate these points. One or
two are from my earlier selection, but the rest are new; all are taken from
the entries as they now stand in the 2009 edition.
Two or three glosses, not definitions:
καλός 1. advantageous, beneficial, desirable:
νεανίσκος young man, lad:
νεῖκος quarrel, contention:
πόνος 1. toil, suffering, hardship:
πρᾶγμα 1. deed, action:
ῥαντός spotted, speckled:
τάχος swiftness, speed:
φοβερός awe-inspiring, formidable, frightful:
φόβος 1. fear, dread:
φόβος 2. religious fear, awe, reverence:
Explanatory additions:
ἀκούω 1. of sense perception: to hear, + acc.
δεῖπνον meal, dinner, usually sumptuous or formal:
ἐπακούω 5. to react to oral message:
ἐπανέρχομαι to return to the point of origin:
ἐπανήκω to move back to the point of origin:
καταλήγω to leave off speaking:
κῦμα wave of the sea:
λουτήρ washing-tub for ritual use:
παιδεία 1. education, instruction (in religion and morality):
ῥάβδος 2. shepherd’s staff, crook:
σθένος 1. bodily strength:
τέρας portentous, extraordinary event with some symbolic meaning per-
formed by God, or by man (though ultimately by God):
Collocation mostly without brackets, sometimes with:
γυμνόω 3. to remove one’s upper garment:
εἰσδύω to move oneself into a space:
κύριος 1. one who owns and controls (possessions):
τίθημι I. 1. to place, lay (+ acc.):
τίθημι I. 5. to direct sbd to do sth:
τίθημι II. 2. to institute (law etc.):
It can be seen that Muraoka is inclined to variation. These instances are

of course only a selection: many more could be added. At the same time,
there are just as many other entries that maintain a consistent approach.
The reason for some of the variations can be surmised. For instance, it is
certainly difficult to compose a definition of καλός sense 1, though pre-
sumably possible (cf. sense 2. morally good and acceptable). Hence three
glosses, each helping to focus the intended meaning: the time-honoured
gloss method comes to the rescue. But I confess to finding the system
behind the explanatory additions (if there is one) elusive. I admit that they
are not necessarily unhelpful, if one suspends one’s wish to know exactly
what they are doing; some may even like this idiosyncratic feature of the
lexicon.3
A general tidying up of these matters could be hoped for at a later
stage in the life of the lexicon. In the meantime, users are not likely to
be seriously misled by the inconsistencies in definition method discussed
here. There are of course other aspects that invite comment; some are
taken up in the fourth part of this review, which considers future steps in
LXX lexicography.
II
The production of Muraoka’s lexicon coincided with another major ini-

tiative in LXX studies, a new translation into English (NETS), which was
set in train under the leadership of Al Pietersma in the mid-1990s. Trans-
lating involves deciding what the text means, and very soon the editors of
NETS found themselves facing questions very like those faced by a lexi-
cographer. The difficulty of the questions is greatly magnified in dealing
with a work that is itself a translation. To their credit, Pietersma and his
colleagues began addressing theoretical issues early on, and the work has
been accompanied by a wealth of discussion and explanation. It has also
been accompanied by controversy and at times incomprehension.4 My aim
is not to wade into this difficult and sometimes bewildering debate, but
3
Danker in his review of Muraoka, Lexicon (2009) draws attention to the ‘different for-
mulae and formulaic combinations used in entries,’ concluding in characteristically mild
fashion that ‘some of them invite the reader to be on special alert.’
4
See, e.g., the polemics between Muraoka and Pietersma in Muraoka, ‘Recent Discus-
sions,’ and Pietersma, ‘A Response’; and the corrective to Dines, Septuagint in Wright,
‘The Septuagint,’ esp. 107–10. It is evident from Muraoka’s ‘Introduction’ (p. ix) that
he misunderstands the approach of NETS (and LEH); likewise Rajak in Translation and
Survival, 143–5.
to try to reach a clear understanding – for myself as much as the interested

reader – of where the difference lies between the approaches of Muraoka
and NETS. My own view necessarily emerges.
The NETS approach. It was easy to get the impression in the initial stages
of the project that the basic principle of NETS was to treat the Hebrew
original as determining the meaning of the Greek translation. The NRSV,
i.e., a translation of the Hebrew text, was to be the base text, with revisions
as required to match the Greek. Further, the ‘interlinear’ model was intro-
duced as a means of describing the translators’ approach to their original.
This seemed to imply that the translators’ method was a mechanical one
in which each Greek word would have the meaning of its Hebrew counter-
part. Statements of the kind still found in the introduction to NETS added
to the impression, e.g., ‘the Greek had a dependent and subservient lin-
guistic relationship to its Semitic parent.’5 Extended discussion of transla-
tion theory, though valuable in itself, tended not to clarify the issue for the
non-expert.6 In all this I speak of impression.
The outcome. I think it fair to say that the NETS approach has evolved
over time, gaining focus and consistency. The relationship to the NRSV
can be seen to have moved from greater dependence to less, in the face
of dealing with Greek that departs in its own peculiar ways from the
Hebrew, even in the ‘literal’ translations.7 ‘Interlinearity’ is clearly stated
to be a metaphor, a way of conceptualising the translators’ approach, not
a description of an actual written form.8 At any rate, the true nature of the
NETS approach is now the opposite of what is commonly supposed: NETS
attempts to translate the Greek according to the meaning it has as Greek,
not to transfer the Hebrew meaning to the Greek and thence to the English
translation. One might think a lexicographer’s aim to be the same, but this
is not necessarily so. To understand how Muraoka’s approach differs, we
need to look deeper.
The focal point. The focus of NETS is on the meaning at the point of
translation, that is, ‘what the original translator thought his text to mean.’9
Muraoka on the other hand looks to the subsequent meaning, that is, ‘what
sense a reader in a period roughly 250 B.C. – 100 A.D. who was ignorant
5
‘To the Reader of NETS,’ xiv. Similarly Pietersma, The Psalms, ix.
6
See, e.g., papers from the Panel Discussion ‘LXX and Descriptive Translation Studies’
in BIOSCS 39 (2006).
7
See, e.g., Silva, ‘Esaias,’ 824 on the shift in dependence on the NRSV in Isaiah. In van
der Meer, Review of NETS (2007), ‘interference’ from the NRSV is assessed as slight
(118, 119).
8
‘To the Reader,’ xiv.
9
‘To the Reader,’ xv.
of Hebrew or Aramaic might have made of the translation.’10 This is a

significant difference, and it can lead to fundamentally different results
between the translation in NETS and the lexical meaning in the lexicon.
(This is not always the case, of course; in the majority of instances NETS
and Muraoka would agree.) A simple example: in Gen 13:14 θάλασσα,
in a context where it is one of the four points of the compass, is translated
‘sea’ in NETS, but given the meaning *2. west in Muraoka. Thus NETS
treats θάλασσα as having its normal Greek meaning ‘sea,’ while Muraoka
takes it as having the sense of the word it translates, Hebrew ‫‘( ים‬west’ as
well as ‘sea’).11 The difference of focal point is the cause, but how exactly
did they arrive at these different meanings?
Context. As θάλασσα shows, the role given to context is crucial. In a
non-translated Greek text, context is the standard tool of the lexicogra-
pher for determining the meaning: the meaning required by the context
is what the text means. Muraoka uses it in that way. This is natural if one’s
aim is to discover what the LXX would mean to a later reader. The context
in Gen 13:14, as it stands in the Greek, requires θάλασσα to be taken as
meaning ‘west,’ with confirmation from the meaning of the original. For
NETS, on the other hand, the Greek context does not have this determina-
tive role.12 When context requires a Greek word to have a meaning outside
its normal range, NETS does not accept it, but gives the word its normal
Greek meaning. This is because of its view of how the translator operates
within the ‘interlinear’ paradigm: he may simply write down a standard
equivalent (θάλασσα = ‫)ים‬, disregarding how it fits in the Greek context
as a whole. So NETS translates ‘sea’ because the translator himself disre-
gards context in his choice of rendering. The Hebrew of the source text
is obviously relevant to the debate here and we must consider it next.
The Hebrew original. Muraoka does take into account the original, even
though his target is the meaning as understood by a later reader who knows
no Hebrew. How it is done is not clearly spelt out.13 Examination of entries
suggests that Muraoka’s practice is to allow the meaning of the Hebrew
to be transferred to the Greek where the Greek context seems to require
it. The meaning is then usually marked with an asterisk, indicating non-
standard Greek. Our example θάλασσα is one such case. The application
10
‘Introduction,’ viii. The approach of La Bible d’Alexandrie is in essence the same as
Muraoka’s: see Harl, La Genèse, 10.
11
Discussions in, e.g., Boyd-Taylor, ‘Lexicography and Interlanguage,’ 58–60; Muraoka,
‘Recent Discussions,’ 235; and Pietersma, ‘A Response,’ 334–6.
12
See esp. Pietersma, ‘Context is King.’
13
‘Introduction,’ viii is rather vague.
is largely ad hoc, each instance being decided as it arises; this is not

unexpected, given the lexicographer’s experience that every case is dif-
ferent. The NETS project also attaches importance to the original Hebrew:
‘what this Septuagint says, and how it says it, can only be understood in its
entirety with the help of the Hebrew.’ But apart from ‘arbitrating between
competing meanings of the Greek,’14 the Hebrew meaning is not permitted
to override the normal Greek meaning. If the Hebrew and the normal Greek
meaning match well enough, we notice nothing; but if the Hebrew meaning
is outside the normal range of the Greek word, NETS keeps the Greek mean-
ing and does not allow a transfer from the Hebrew. So the NETS introduc-
tion can truthfully say, ‘perhaps paradoxically, the interlinear paradigm
safeguards the Greekness of the Septuagint.’15
Intention. Another issue lurks in the discussion so far, namely transla-
tor’s intention. Though nothing is stated in the introduction, NETS is wary
of resort to the ‘intention’ of the translator: we simply cannot read the mind
of an ancient translator.16 This is true scientifically, but in practice unwork-
able. A translator has some intention when he translates, however difficult
it may be to determine it. The postulated ‘interlinear’ model of the transla-
tor’s method itself involves an attempt to capture to some degree the inten-
tion of the translator. The editors of NETS say as much in using the words
‘what the original translator thought this text to mean.’17 Muraoka alludes
to the subject, but it does not loom large because his focus is on the subse-
quent meaning.18
In my view, if we seek to understand the text ‘as produced’ (NETS’s
words), we have no choice but to attempt to deduce the translator’s inten-
tion, and I believe it can be done, using certain clues. These are: (a) context
(of the Greek as created by the translator); (b) meaning of the Hebrew origi-
nal (as understood by the translator); (c) Greek usage (in the translator’s
time). From these we deduce, however imperfectly, the meaning intended
by the translator, and this in turn is the meaning, at the point of produc-
tion. In this I adhere to the insight of Emanuel Tov, enunciated in 1976.19
14
15
16
Compare Boyd-Taylor, ‘Evidentiary Value,’ 51–55; idem, Reading Between the Lines,
4–5.
17
‘To the Reader,’ xv. Boyd-Taylor’s older paper, ‘Place in the Sun,’ includes subtle (and
convincing) attempts to reconstruct the translator’s thought processes.
18
Muraoka, ‘Introduction,’ viii: ‘The translator’s intention is something rather elusive and
not easy to comprehend with confidence.’
19
See Tov, ‘Three Dimensions,’ 529–30, 532. Compare Lee, ‘Equivocal Renderings,’ 104
[= Essay 3].
I am, then, in basic agreement with the NETS approach, but with trans-
lator’s intention as the path to establishing the meaning of the text as
produced.20
Bilingual interference. This is a large topic, extensively discussed in the
context of language generally. In a translated text, all sorts of potential for
interference arises; there are also many problems of definition and under-
standing. It is not possible to deal with the topic adequately here. For the
present purpose I focus on one issue, raised by our example of θάλασσα.
Is it possible that the translator himself intended θάλασσα to mean ‘west’
in the context of Gen 13:14 (and elsewhere)? That would be a ‘loan-shift’
or ‘semantic loan.’21 It is very difficult, if not impossible, to know. We
have no means of showing that the shift had occurred in the mind of the
translator. The alternative, that θάλασσα was simply the default rendering
of the Hebrew word, not intended to mean something different from its
usual Greek sense, remains the safer assumption. That is the position of
NETS. Are there, then, any cases of interference of this type? That is, are
there any genuine loan-shifts? I used to think there were, but now I am not
so sure.22 Pietersma has argued persuasively that their existence remains
to be demonstrated.23 The interesting consequence is that renderings like
θάλασσα, ‘sea,’ instead of a normal Greek word for ‘west,’ reflect a choice
to translate in that way and thus to retain some of the un-Greek character
of the original in the translation.24 To what extent the choice was free and
conscious is a topic for further thought.
Conclusion. The approaches of NETS and Muraoka to determining the
meaning of a word in the LXX are fundamentally different, and the main
reason is the difference of focus, between the meaning ‘as produced’ on
20
Aejmelaeus’s thoughtful exploration of the topic in ‘Translation Technique’ (1989) led
her to a position like Muraoka’s and to rule out translator’s intention as a means of inter-
preting the text. For her, intention could even be absent, as in the case of standard render-
ings. I see intention as involved in these choices like all others; it is simply very variable
or flexible.
21
See Adams, Bilingualism, 461–3; Silva, Biblical Words, 87. Older Septuagintalists spoke
simply of ‘Hebraisms,’ without much analysis: see, e.g., Thackeray, Grammar, 39–
55.
22
See Lee, ‘Note on Septuagint Material,’ 237–8; ‘Equivocal Renderings,’ 116–7. [I have
changed my mind about κῶλον and now agree it is not a ‘loan-shift’: see Essay 1, n. 8.
But in Greek of the Pentateuch, 195, 198 I identify what I think are definite cases of
‘loan-shifts’ (certain exx. of ἁμαρτία, πλημμέλεια).]
23
Pietersma, ‘Context is King,’ 172–5. We have come a long way since the days when
Gehman could claim that ‘the translator had in mind a certain kind of tree’ when he used
ἄρουρα (‘field’) to translate ‫אשׁל‬, ‘tamarisk,’ and even more startlingly that ἀπό meant
‘toward’ in ἀπὸ λιβός (‫)נגבה‬: see Gehman, ‘Adventures,’ 130, 126.
24
The point is well made by Satterthwaite, at ‘Judges,’ 199.
the one hand, and what a later reader ‘might have made of the translation’
on the other. The two approaches are not reconcilable and are bound to
lead to different results when the conditions require them. Yet they are not
unable to co-exist, and the LXX student or scholar can accept and make
use of each on its own terms. As well as a lexicon on Muraoka’s princi-
ples, it is valuable for us to have a rendering of what the LXX appears to
have meant as Greek at the point of production.
In practice, of course, things are not always so simple. There are diffi-
culties for both in attaining their intended targets. Muraoka faces the pros-
pect of not one but multiple later readers spread over centuries, and uncer-
tainty about how any of them would have understood the LXX, especially
where it presents extremes of un-Greek usage.25 The NETS approach at least
has only one person to focus on (at a time) and has a clear theoretical basis
for assigning a meaning to the Greek, but it is not easy to do, and NETS
is seen to compromise its principles at times.26 Part of the difficulty is the
ever-present problem of determining what Greek usage actually was in
the translator’s time. A change in the evidence can require a change in the
rendering, from a Greek meaning differing from the Hebrew to one match-
ing it, and vice versa.
The NETS project and the debates generated by it have been beneficial
to LXX studies. The insistence of the NETS editors on establishing a theo-
retical basis for their work has brought out and clarified issues that were
unclear to most of us. It seems to me, indeed, that we have seen the most
significant advance in a century in our understanding of what used to be
called ‘the Semitic element’ in the LXX.
III
I offer now just a few examples to illustrate the issues discussed in the
preceding section.
Num 5:12–13 ἀνδρὸς ἀνδρὸς ἐὰν παραβῇ ἡ γυνὴ αὐτοῦ καὶ παρίδῃ
αὐτὸν ὑπεριδοῦσα, (13) καὶ κοιμηθῇ τις μετ᾽ αὐτῆς.
Of a man, of a man, if his wife transgresses and disregards him disdain-
ing (him) and someone sleeps with her.
‫ ושׁכב אישׁ אתה‬:‫אישׁ אישׁ כי־תשׂטה אשׁתו ומעלה בו מעל‬
25
For example, ἐν at Hos 12:12, where Muraoka has it under *4. in return for, for the
price of. Would a reader without Hebrew guess this?
26
See examples below. One understands, of course, the difficulty of maintaining consistency
across the work of over thirty contributors.
The rendering is isomorphic (‘interlinear’), i.e., the Greek and Hebrew

words match in order and meaning (more or less). But the interesting thing
is the way the translator, using the syntactic markers natural to Greek, cre-
ates a context that makes the sentence hang together and convey a mean-
ing, despite its strangeness as Greek. Especially significant is his choice of
genitive case in ἀνδρὸς ἀνδρός: he looks ahead and links them to αὐτοῦ,
instead of using the totally ‘literal’ rendering ἀνὴρ ἀνήρ. In this I see clear
proof of ‘intention.’ But there is nothing to allow us to take the step of
transferring the meaning of the Hebrew idiom to the Greek (‘if any man’s
wife ...’). The translator certainly knows what the Hebrew idiom means,
as renderings elsewhere show,27 but chooses not to represent it in normal
Greek but to keep the un-Greek flavour of the original. The right trans-
lation into English, then, is in accordance with NETS principles, i.e., as
above, ‘of a man, of a man.’28 The rendering of ‫ אישׁ‬by τις in the next
clause shows the translator choosing to go the other way and use natural
Greek instead of the stereotypical equivalent ἀνήρ.29
The same phenomena could be illustrated at length. For example, at
Gen 8:10–12, the translator first renders the Hebrew idiom ‫ יסף‬hiph +
infin. by a natural Greek equivalent (πάλιν ἐξαπέστειλεν) and then two
verses later by a ‘literal’ equivalent, unnatural for Greek (οὐ προσέθετο
τοῦ ἐπιστρέψαι). To represent what the translator intended when he chose
this rendering we must translate ‘she (the dove) did not add to return.’30
Determining a later reader’s understanding of it is harder.31
Deut 19:5 καὶ ἐκκρουσθῇ ἡ χεὶρ αὐτοῦ τῇ ἀξίνῃ κόπτοντος τὸ ξύλον,
καὶ ἐκπεσὸν τὸ σιδήριον ἀπὸ τοῦ ξύλου τύχῃ τοῦ πλησίον.
and (if) his hand is knocked aside as he is cutting the tree with the axe,
and the iron axe-head falling out of the wooden handle hits his neighbour.
‫… ונשׁל הברזל מן־העץ ומצא את־רעהו‬
To render ‫ מצא‬the translator declines the stereotypical equivalent εὑρί-
σκω and draws on the deep resources of the Greek language for the appro-
priate word. His choice is τυγχάνω + gen. in its ancient, indeed original,
27
For example, Num 1:4 ἕκαστος (κατὰ …).
28
NETS itself compromises: ‘Man by man – if his wife goes astray …’ with note: ‘I.e. Any
man.’ Both are concessions to natural English and not what the Greek says.
29
The Leviticus translator adopts the same strategy for ‫ אישׁ אישׁ‬in Lev 15:2 ἀνδρὶ ἀνδρὶ
ᾧ ἐὰν γένηται ῥύσις. He too can render idiomatically if he chooses, as in Lev 20:2
ἐάν τις. …
30
NETS again compromises, with ‘did not continue to turn back’: LSJ, s.v. προστίθημι
B.III is the culprit.
31
Muraoka, s.v. προστίθημι *2., takes it to be that of the Hebrew idiom: ‘still to do sth
as formerly, do sth again.’
meaning ‘hit’ (the mark/target), attested since Homer. The total context
he creates tells us exactly what he intends. This use of τυγχάνω is not an
everyday one and possibly came to the translator from an education in
the Classics.32
Isa 48:22 (simil. 57:21) οὐκ ἔστι χαίρειν τοῖς ἀσεβέσι, λέγει κύριος.
There is no ‘greetings’ for the ungodly, says the Lord.
:‫אין שׁלום אמר יהוה לרשׁעים‬
This example shows how the original Hebrew can be decisive in dis-
cerning the translator’s intention; acquaintance with Greek usage is also
essential. If the Greek is read on its own it appears to mean ‘it is not pos-
sible for the ungodly to rejoice’ (compare Muraoka, s.v. εἰμί 1.d) or ‘there
is no rejoicing, says the Lord, for the impious’ (NETS). But χαίρειν trans-
lates ‫ שׁלום‬in the original. How does that work? Why not εἰρήνη? The
answer strikes us when we bring together the twin facts that ‫ שׁלום‬is a
standard greeting in Hebrew, and χαίρειν has the same function in the
greeting formula in Greek letters (‘x to y greetings’). The Isaiah transla-
tor skilfully captures in the Greek of his day the meaning of the Hebrew
as he understands it.
1 Kgdms 1:26 (+ 4) ἐν ἐμοί, κύριε.
‫בי אדני‬
This classic has been discussed so often there is not much left to say.
But I include it as a final test of the approaches of Muraoka and NETS.
The basic facts are: Hebrew ‫ בי‬is an idiom meaning ‘I pray, excuse me’
(BDB), or ‘by your leave’ (HALOT). The translation ἐν ἐμοί simply rep-
licates the (perceived) components of the Hebrew (‘in/on me’) and con-
veys nothing more than ‘in me’: there is no comparable idiom in Greek.
We do not know if the translator understood the Hebrew idiom; we cannot
say that he intended ἐν ἐμοί to have its meaning; we only know that he
chose to represent the Hebrew ‘literally.’ The translation of the Greek as
produced is therefore ‘in me.’33
In the introduction to NETS, the editors offer ἐν ἐμοί as a specimen
of what they call an ‘isolate,’ giving the meaning as ‘in/with me.’34 But
32
NETS is not quite on the mark: ‘happens to strike his neighbour.’ Dogniez and Harl, Le
Deutéronome have got it: ‘atteint son prochain.’ τυγχάνω is only here in the Pentateuch.
Homer, Iliad 23.857, with τυγχάνω and its opposite, is too good to miss: ὃς δέ κε μηρίν-
θοιο τύχηι, ὄρνιθος ἁμαρτών (‘he who hits the string [with his arrow], missing the
bird. …’).
33
Compare Barr’s unerring analysis in Typology, 393.
34
‘To the Reader,’ xvii.
the actual translations are: ‘by me …?’ (Judg 6:13AB, 15AB), ‘by me’
(Judg 13:8AB, 1 Kgdms 1:26) and ‘with regard to me’ (3 Kgdms 3:17,
26). There have been some compromises, apparently in the direction of
more intelligible English. I do not think ‘with regard to’ is possible, and
‘with’ and ‘by’ (instrumental?) would be hard to pick up without a con-
text to suggest them; the treatment as introducing a question is simply
puzzling. So the NETS outcome is only approximately in line with what
one expected.
Muraoka (s.v. ἐν *18) rightly labels ἐν ἐμοί a calque on the Hebrew.35
But he gives no meaning. He thus does not commit himself to a decision on
what it would mean to a Hebrewless later reader, no doubt because it is dif-
ficult to say, just as that reader would be hard put to make sense of it.
IV
Muraoka’s lexicon is a major step forward in LXX lexicography. It

will now be both the standard tool and the foundation for future devel-
opment. As to the latter, the lexicography of the LXX is young and will
certainly not stand still in coming years. Let us consider what might hap-
pen next.
It goes without saying that refinements to the lexicon will need to be
made. I am not suggesting a major overhaul, but minor improvements that
will ensue from further observation and discussion. These are of the kind
already pointed out by other reviewers; my own list can wait.36 They can
easily be done in future editions – which are certain to come, if the history
of NT lexicography is any guide. At the same time there are other needs
in LXX lexicography that go beyond small adjustments to a printed work,
though the results will end up there.
First and most important is the further pursuit and evaluation of evi-
dence of Greek outside the LXX. This evidence obviously plays a vital
role in determining meaning in the LXX. The easy assumption that the
evidence can all be found in works like LSJ is mistaken. Muraoka, as
he explains, drew on whatever was available in the reference books and
special studies, but did not undertake fresh searching of his own; he also
35
NETS has its own use of the term calque that seems at variance with the usual under-
standing. For a recent definition see Adams, Bilingualism, 459. NETS’s stock example,
διαθήκη ~ ‫ברית‬, cannot be a calque in that sense.
36
See, e.g., the reviews by Danker; Shaw; and Silva. The long review of Muraoka, Lexicon
(Pentateuch and Twelve Prophets) in Andersen, ‘Advances’ raises many questions but
is short on answers.
does not attempt to report the evidence in his entries.37 This approach
was the right one. But the task of completing the evidence for all the
vocabulary of the LXX now needs to be undertaken in earnest. This will
be ongoing and involve many contributors, not one lone lexicographer.
I also do not believe that the lexicon should attempt to assemble this evi-
dence; it should do no more than draw on the results. The place for the
assembly of the evidence is in an electronic database. A database is also
the place for the material that meets the second need, which we will come
to in a moment.
I give just one example of the importance of non-LXX evidence and
how it can change. The word προσήλυτος has been supposed, on the basis
of lack of attestation, to be a creation of the LXX translators and even,
despite the inherent improbability, to have had the sense ‘proselyte’ when it
was coined.38 But at the 2009 SBL meeting David M. Moffitt presented a
papyrus document containing an occurrence of the word: the date is III BC,
there is no suggestion of a Jewish connection, and the meaning can hardly
be other than the expected one, i.e., ‘newcomer, temporary resident.’39 Of
course, proper searching of the evidence is not just for the purpose of find-
ing parallels outside the LXX, but also to establish when a word or use is
unattested and so might be a creation of the translators.
The other need is for an ongoing, up-to-date, complete record of existing
discussions of words in the LXX, and indeed in all of Greek.40 Muraoka
includes a selection in each entry, following his sound principle that he
names only works that have made a contribution to his decision.41 LEH
include much fuller lists. But we have no way of knowing if the lists are
complete, and they are impossible to keep up to date; further, mere lists
tell us nothing about what the authors have to say. The only way this can
be remedied efficiently is in an electronic database.
Again just one example, the ‘famished bear’ of Hos 13:8 (ἄρκος ἀπο-
ρουμένη). The discussions of this disputable case seem to be proliferat-
ing, almost before one has had time to take it in. Muraoka mentions only
Bons 2001. Besides that there is his own discussion in the Introduction
37
‘Introduction,’ ix. Compare comments in my review of Muraoka, Lexicon (Pentateuch
and Twelve Prophets), 129 [= Essay 10].
38
Labelled ‘neol.’ by LEH, marked * by Muraoka. See both for literature as far back as
Allen in 1894.
39
The papyrus is P.Duk.inv.727R. Publication is expected in the near future (email from
David Moffitt, 6.5.10). [See Lee, Greek of the Pentateuch, 15 for an update.]
40
See Lee, ‘Lexicographical Database.’
41
Muraoka, ‘Introduction,’ xv. Compare remarks in Lee, Review of Lexicon (Pentateuch
and Twelve Prophets), 129–30.
(p. viii), where the views of Harl and Joosten are mentioned; LEH add
no more; but I have come across one by Boyd-Taylor, with a further refer-
ence to Jobes and Silva.42 Are there more? Who can say? And what do
they each contribute? To see where the discussion is up to, it would be
helpful, one might say essential, to have this information. The same applies
to countless other cases. The place for all this information is in an electronic
database, and I believe this to be the new desideratum of LXX studies.
Reviewers tend to focus on faults, often [self-]indulgently, showing off
their skill at spotting mistakes and raising questions. In the final impres-
sion the flaws become magnified out of proportion. The criticisms voiced
in this review cannot detract from Muraoka’s outstanding achievement.
The work as a whole is reliable, helpful, and thorough. In it students of the
LXX have a tool that we can use with confidence and respect long into
the future.
42
Boyd-Taylor, ‘Evidentiary Value,’ 54–5.
16
AUXILIARY THELO
20101
Abstract
Over the history of Greek, the verb θέλω, ‘wish,’ ‘be willing,’ evolves into an
auxiliary expressing futurity. Instances as early as Classical Greek are known;
partial collections of examples have been made. My aim in this paper is to gather
as many instances as possible up to 600 AD. Existing sources as well as my own
observations are used, and 70 + instances are found. Even so, this is not an exhaus-
tive collection. The conclusion is reached that θέλω was available as a future
auxiliary over the course of 1000 years, but was never more than sporadic until
after the end of Koine Greek. All the examples are of θέλω + infinitive, none of
θέλω ἵνα + subjunctive, a detail of significance and contrary to what is some-
times assumed. In the course of the investigation another, previously unnoticed
use of θέλω, namely, imperative of θέλω + infinitive with the meaning ‘(please)
do/don’t,’ was discovered; instances of this use are presented.
1. THE GRAMMATICALISATION OF ΘΕΛΩ
The future tense in Modern Greek is formed with θα + subjunctive, as

for example, θα γράϕω, θα γράψω. This form of expression has its origin
in a periphrasis with θέλω. The ultimate base is θέλω + infinitive, with
θέλω in its original meaning ‘wish to,’ which evolves into an expression
of simple futurity. The development is parallel to that in many languages,
among them of course English, in which futurity is expressed by an aux-
iliary that originally meant ‘wish/want’; or to put it in terms of grammati-
calisation, the lexical item ‘wish/want’ has evolved along the cline of
grammaticality to a grammatical function, namely, to express futurity.2
The detailed history of the development in the Byzantine period is not
the concern of this paper and will be touched on only briefly. It is more
1
This paper was originally presented at the conference ‘Buried Linguistic Treasure: The
Potential of Papyri and Related Sources for the Study of Greek and Latin,’ at Christ
Church, Oxford in 2006, convened by Trevor Evans and Dirk Obbink.
2
Cf. Hopper and Traugott, Grammaticalization, 6–7.
complicated than one might have expected, and debate continues on the
details. It is not simply a matter of a single line of development θέλω +
infinitive > θέλω ἵνα + subjunctive > θέλω να + subjunctive > θα + sub-
junctive; there are more steps and variants involved, as shown especially
by Brian Joseph’s study, which is a warning against over-simplification.3
For our purposes, let us simply note that when the ancestor of the Modern
Greek particle first appears in the twelfth century, in the form θε να, it
does so alongside θέλω + infinitive as an expression of futurity: the latter
was still in use and continued to be for some time before its final displace-
ment by θα + subjunctive.
By the Byzantine period the periphrasis with θέλω had clearly prevailed
over the other, earlier contenders as the means of expressing the future.
The other main contenders, at the end of the Koine period, were: the old
monolectic form; the present with future sense; μέλλω + infinitive; ἔχω +
infinitive; the aorist subjunctive.4 We know with hindsight that θέλω was
to prevail, but the issue had not yet been decided.
2. A NEW ASSESSMENT OF THE EVIDENCE BEFORE 600 AD
Where are the beginnings of this development of θέλω? When, in the

period before 600 AD, does θέλω start to show signs of being a future
auxiliary? Where are the examples, and how many are there? That is the
question that I want to (and will) address in this paper.5
It is commonly said that θέλω + infinitive appears as a future-equivalent
in the Koine period. Browning is the most authoritative voice on the sub-
ject. He simply states it as a fact, adding the rider that it is not common
till after 600, but gives no examples.6 Browning’s book was intended as
a general survey, without detailed references; but if we look elsewhere it
3
Joseph, Morphology and Universals, 114–59. Cf. Horrocks, Greek, 130, 228–9. See also
Pappas, ‘Microcosm’; Joseph and Pappas, ‘On Some Recent Views’; Holton, ‘Formation
of the Future,’ 119–20, 127–8; Hock and Joseph, Language History, 402–5 (Balkan
developments); earlier Jannaris, Grammar, 552–9; Thumb, Handbook, §226; Hatzidakis,
Μεσαιωνικά, 1:197.
4
See Browning, Greek, 33–5. Other less frequent future-equivalents are also noted there.
5
I am well aware of the older usage of shall and will maintained by some, but as this is
not my own practice and is artificial for me, it is not followed in this paper.
6
Browning, Greek, 34. He goes on (35) to list the numerous ways of expressing futurity
in John Moschos, again without citing examples except one (not of θέλω). The unnamed
source from which these data are derived, Mihevc-Gabrovec, Études, 64–5, noted only
one instance of θέλω as a future auxiliary in Moschos (see under no. 1 in my list of
examples below).
AUXILIARY THELO 239
is much the same. Gignac, in his generally thorough grammar of the papyri
of the Roman and Byzantine periods, speaks of the ‘increasingly frequent
replacement of the future tense by periphrastic constructions in the later
Koine, mainly by θέλω ἵνα and the subjunctive,’ but offers only two
examples – in the same text, and in fact of θέλω + infinitive.7 Mandilaras
likewise asserts it, but gives no examples.8 Joseph simply refers to Brown-
ing.9 Horrocks takes it for granted and does not amplify.10 Back in 1898,
Karl Dieterich did much better: he noted some instances in late funerary
inscriptions, a source which proves to be a rich one when modern search-
ing techniques are applied; but his observations slipped out of sight.11
Besides these there are a number of specialised studies, notably those
of Joüon, Riesenfeld, Rödiger, Schrenk, and Wifstrand, that offer useful
collections of examples.12 But they all focus on their own area of interest;
they do not connect with one another nor study the phenomenon across
time.
In the lexicons there is a certain amount of material, very partial, but
useful as far as it goes. LSJ offer a sense II.1 ‘to express a future event, like
our will or shall,’ with eight instances cited (plus an ‘etc.’). The examples
are all Classical, and only half seem to me to be right, but even so, this is
a beginning.13 Lampe also recognises this sense, but has only two exam-
ples, the same two that had been noted (with three others) by Sophocles
7
Gignac, Grammar, 2:290, with n. 3. He adds a reference to Burguière, Histoire de
l’infinitif, but this work yields no Koine Greek examples of auxiliary θέλω. Gignac’s
examples are at no. 6 below.
8
Mandilaras, Verb, 180.
9
Joseph, Morphology and Universals, 114, 116, with nn. p. 150.
10
Horrocks, Greek, 76, cf. 229–32.
11
Dieterich, Untersuchungen, 245–6. See nos. 7 to 10 in the list of examples. The lengthy
discussion in Mirambel, ‘Essai,’ 179–88, yields one example, the same one as noted by
Mihevc-Gabrovec in Moschos (see n. 6 above).
12
Joüon, ‘Θέλειν’; idem, ‘Les verbes’; Riesenfeld, Zum Gebrauch von θέλω; Rödiger,
‘βούλομαι und ἐθέλω’; Schrenk, ‘θέλω, θέλημα, θέλησις’; Wifstrand, ‘Verba für wol-
len.’ Other works consulted: Adrados, ‘Sobre los orígenes’; Braun, ‘Nota sui verbi’;
Fleischman, The Future; Fox, ‘βούλεσθαι und (ἐ)θέλειν’; Pinkster, ‘Development of
Future Tense’; Preisigke, Wörterbuch. I have not been able to see Psichari, Quelques
travaux.
13
LSJ, s.v. ἐθέλω. The whole section II is headed ‘of inanimate things’ and examples of
that kind are cited first under II.1; then LSJ add ‘very rarely of living things’ and proceed
to cite an equal number. The distinction has no effect on the lexical meaning, but, as
Willy Clarysse pointed out to me after my paper at the ‘Buried Linguistic Treasure’
conference, examples applied to inanimates are strong proof of the development. Andreas
Willi made the somewhat similar point that the clearest examples will be those where
the verb is in the third person, and not in an if-clause (as no. 35 below). The ‘etc.’ in
LSJ covers some good Plato examples that had been in the 7th and 8th editions but
were dropped in the 9th, leaving only Rep. 370b, an unconvincing case.
back in 1887.14 The NT lexicons, on the other hand, are not aware of the
question at all; even the probable NT examples escape notice, let alone
others.15 Most surprising is DGE, which has no instances of this sense and
apparently does not recognise its existence.16
My purpose has been to gather as many examples of θέλω as a future
auxiliary as I can from all previous sources, as well as those I have found
myself. It must be said at once that the collection is not exhaustive. While
most of the documentary evidence has been checked (via PHI7), I have
not done the full examination of Greek literature that would be possible
– though forbidding – by means of the TLG and would be likely to yield
further material. But what I have goes some way towards answering the
question. My list of examples is presented below, in reverse chronological
order. Needless to say, all the items in the list have been thoroughly vet-
ted; I have rejected any proposals that are doubtful. The list is therefore
not a list of all the proposals but only of those that have a good chance
of being what we are looking for.
What are we looking for? This needs to be clear at the outset. We are
looking for cases where the usual or established senses of θέλω do not
seem to work, where any such sense has faded away to the point where
there is not much left but futurity. By the usual senses I mean ‘wish/want’
and ‘be willing’; I do not include among them a meaning ‘intend,’ as I am
not sure that it is clearly established for this word (as it is for μέλλω). This
fading does not rule out the possibility, even likelihood, that θέλω retained
some nuance that distinguished it from the monolectic future expressing
simple futurity and from other future expressions. But it is difficult if not
impossible for us at this distance to appreciate such a nuance; even to
define the usual senses of θέλω is notoriously difficult.
In a quest to find any new semantic development, one needs to be able
to produce examples that are better than just possible, but highly probable
(or as John Chadwick would have put it, ‘incontrovertible’). It is a severe
test in this case, because it is in the nature of the phenomenon that there is
gradual shading from one meaning into another, and it is hard to know in
a particular instance whether the meaning really has shifted from the lexi-
cal area into the grammatical.17 I cannot claim that my examples all pass
this test, but there are certainly some.
14
Lampe, PGL, s.v. θέλω IV; Sophocles, Lexicon, s.v. θέλω 5.
15
See BDAG, s.v. θέλω; Louw and Nida, Lexicon, Subdomains 25.1,102; 30.58; 31.4.
Louw (‘Analysis of Meaning,’ 142) specifically rejects Mark 6:48 (no. 23 below) in
reply to me (Lee, ‘UBS Lexicon,’ 179).
16
DGE, vol. 6, s.v. ἐθέλω tb. [también] θέλω.
17
Cf. Hopper and Traugott, Grammaticalization, 6–7, 9.
AUXILIARY THELO 241
3. SAMPLES
Let us take some samples from the list to illustrate these points. In the
case of no. 38 Hdt. 1.109.4 εἰ δ᾽ ἐθέλει … ἀναβῆναι ἡ τυραννίς, it is hard
to see how, with the inanimate subject ‘sovereignty,’ the verb can continue
to have its sense of ‘wish’ or any other distinct semantic content; we are
left with futurity. The same can be said of no. 35 Pl., Rep. 423b μέχρι οὗ
ἂν ἐθέλῃ … εἶναι μία, where the subject is ‘the city.’ These are just two
items from the surprisingly extensive evidence in the Classical period,
notably in Herodotus and Plato.
For good examples from much later, consider no. 21 Aesopi Fab. 142
πῶς πάλιν ἐξ ὄνου ἵππον θέλεις ἔχειν; and no. 12 POxy 14.1763.10
λέγουσι δὲ ὅτι ǀ μέχρι ιε θέλομεν ǀ ἐξελθεῖν. The latter in particular
seems to be a periphrasis for the future. The context, with its reported
speech and time expression, makes it clear that the volition of the parties
described as ‘we’ is not in the picture.
In no. 7 MAMA 1.160.4 ἄν τις θεǀλήσι ἀνῦξεν, it is not a question of
someone merely wanting to open the tomb; the text envisages someone
actually doing so in the future and suffering the penalty. All these funerary
texts (nos. 7–10, with list) are similar and provide a strong bloc of exam-
ples. We note that the dates are not from the end of Koine Greek but from
the fourth, third and even second centuries AD (many are of course not
precisely datable). Among them no. 10 IPrusaOlymp 1.83.9 is an interest-
ing variation. The words ἐὰν δέ τις θ[ε]ǀλήσει stand without an infinitive
expressed: it is to be supplied by extrapolation from ἀνεξοδίαστον, ‘inal-
ienable,’ ‘not to be taken over,’ to give the meaning ‘if anyone shall/does
(alienate it, take it over).’ The simple future-auxiliary function of θέλω
seems inescapable.
Included in the list are one or two examples where there is some doubt.
Example no. 32 PHib 1.65.25 was proposed by Mayser.18 At first sight
one would be inclined to take θέλομεν as ‘I want to,’ but a reading of the
whole letter suggests that Mayser was right. The writer is explaining his
plans and simply states what he will do to make up the deficit if he can
get some help with the rest. It is not, then, an expression of a wish but a
description of future action. If accepted, this would be contemporary with
no. 31 LXX Exod 2:14 μὴ ἀνελεῖν με σὺ θέλεις, which I think is sound.19
18
Mayser, Grammatik, II.1, 226.
19
Cf. Evans, Verbal Syntax, 229, where θέλω is taken to be ‘mean/intend’: ‘surely you
don’t mean to kill me?’ My understanding of μή here as introducing a neutral question
makes a slight difference.
In an example like no. 16 Hermas, Vis. 3.1.9 θέλοντος οὖν μου καθίσαι
εἰς τὰ δεξιὰ μέρη, we seem to catch θέλω at the point of transition. Is
it ‘wishing to’ or ‘being about to’? Either is possible. But the two clearer
examples of ‘be about to’ in the same author (no. 17, and 3.3.1) tip the
balance in favour of the latter.
Two general observations may be made at this point. Quite a number of
the examples are of a past tense (ἤθελον, ἠθέλησα), where I have trans-
lated ‘was going to,’ ‘was about to.’ Some overlap or competition with
μέλλω, the standard word for this from early Greek onwards, seems obvi-
ous. A full study is needed before any firm conclusions can be reached
about μέλλω in Koine Greek and its relation to θέλω, but one suggestion
may be put forward here. If we take the NT, where μέλλω is common (109
occurrences), as a sample, we see that while μέλλω is sometimes used like
θέλω (as Acts 16:27 σπασάμενος τὴν μάχαιραν ἤμελλεν ἑαυτὸν ἀναι-
ρεῖν), it most often refers to the more distant future (as John 6:71 οὗτος
γὰρ ἔμελλεν παραδιδόναι αὐτόν). This loss of immediacy could be the
reason for another contender to appear, to supply the meaning ‘be on the
point of.’
Secondly, I draw attention to a noteworthy fact: none of the examples
in my collection shows θέλω ἵνα + subjunctive; all are of θέλω + infini-
tive. This is significant in the light of what came later. It is consistent with
the evidence of the continuing use of θέλω + infinitive as a future expres-
sion in Byzantine Greek. A shift from infinitive to ἵνα + subjunctive in
this expression appears not to have been a feature of the Koine period
at all.
4. AN UNOBSERVED AUXILIARY FUNCTION OF ΘΕΛΩ
Before I come to my conclusion on θέλω as a future auxiliary, there

is another to be noticed. In the course of this investigation I came across
what amounts to another auxiliary function of θέλω, one that as far as I
know has not been observed before. It was from the search of the papyri
for θέλω that this discovery emerged; a connection could then be made to
some literary examples not considered in this light before. A selection of
examples is given below (§§6.2, 3).
What we see is θέλω in positive and negative commands + infinitive
(rarely imperative: POxy 14.1776.6 below) in which the full semantic con-
tent of θέλω has faded and the verb is simply a means of introducing or
AUXILIARY THELO 243
in some way nuancing the instruction contained in the infinitive. The com-
bination appears to form a polite request, but the semantic value of θέλω
itself is hard to pin down. ‘Please’ or ‘be so kind as to’ are makeshifts and
not true equivalents, though they give the general effect. The development
would seem to come from θέλω in its sense of ‘be willing.’ The upshot
is that we have an auxiliary reminiscent of the English auxiliary do in ‘do
say,’ ‘do not say,’ etc.
5. CONCLUSION
Finally, a conclusion on θέλω as a future auxiliary. We have a large

number of examples, over seventy, spread across a time span of more than
1000 years. How do we assess the significance of this evidence? What
does it mean?
On the one hand, it seems clear that θέλω as a future auxiliary was more
common than has been supposed. If we apply the argument that people
write more carefully than they speak, and assuming that θέλω as future
auxiliary was a vernacular feature, it could be concluded that it was in fact
in frequent use in speech throughout the whole time, and has simply not
surfaced much in our evidence. On that basis it might already have been
the front-runner well before the end of Koine Greek. On the other hand,
we find as late as the end of the Koine period, in reasonably vernacular
texts, various other future expressions still competing with θέλω. More-
over, some of these are not old but new contenders that had appeared in the
middle and later Koine (present with future sense; ἔχω + infinitive; aorist
subjunctive), thus showing that no one form had yet established itself. So
one might conclude that θέλω as future auxiliary was never very common,
and even by the end of the Koine period was still some way from estab-
lishing the dominance that it was to gain later. I incline to the latter view,
that θέλω was available as a future auxiliary for 1000 years, but was never
more than sporadic until after Koine Greek. Though the time span of θέλω
as future auxiliary seems remarkably long, and one might be surprised that
different, nearly synonymous ways of expressing the future could co-exist
for centuries, grammaticalisation studies have shown that just such charac-
teristics are part of the phenomenon.20
20
Cf. Hopper and Traugott, Grammaticalization, 97 on the persistence of alternative future
markers in English since the time of Beowulf.
Obviously what would be helpful next is a statistical study, in as large

a corpus as possible, of all the ways of expressing futurity up to the end
of Koine Greek. But that is a task for the future.21
6. EXAMPLES
6.1 θέλω as future auxiliary

1. John Moschos 19 (M 87:2865C–D) (c.600 AD) [Sophocles; Jannaris; Dieterich;
Lampe]22
διὰ τοιαύτην οὖν ἡδονὴν βλέπε πόσους κόπους θέλεις ἀπολέσαι, ἴδε διὰ
ποίαν ἁμαρτίαν θέλετε ἑαυτοὺς ἀποστερῆσαι τῆς βασιλείας τῶν οὐρανῶν.
ἀβάλε τῇ ἀνθρωπότητι. διὰ μίαν ὥραν ὅλον ἐκεῖνον τὸν κάματον θέλεις
ζημιωθῆναι;
For the sake of such pleasure look how much work you are going to destroy, see
for what kind of sin you are going to deprive yourselves of the kingdom of heaven.
Oh human nature! Are you willing to/going to forfeit all that toil for one hour
[of pleasure]?
See also 184 (M 87:3057A) [Mihevc-Gabrovec]
2–4. Apophthegmata Patrum (M 65:76–440) (c.500 AD)
2. Arsenios 29
ἦλθέ ποτε μαγιστριανός, ϕέρων αὐτῷ διαθήκην τινὸς συγκλητικοῦ συγγενοῦς
αὐτοῦ, ὃς κατέλιπεν αὐτῷ κληρονομίαν πολλὴν σϕόδρα· καὶ λαβὼν αὐτήν,
ἤθελε σχίσαι. καὶ ἔπεσεν ὁ μαγιστριανὸς εἰς τοὺς πόδας αὐτοῦ, λέγων, δέομαί
σου, μὴ σχίσῃς αὐτήν·
Once a magistrianus came to him bringing the will of a certain senator his kins-
man, who had left him a very large inheritance. He took it and was going to tear
it up. The magistrianus fell at his feet saying, ‘I beg you, don’t tear it up.’
3. Paphnoutios 1
[ὁ ἀρχιλῃστὴς] … ἐγέμισε ποτήριον οἴνου, καὶ τὸ ξίϕος ἐν τῇ χειρὶ αὐτοῦ,
καὶ λέγει τῷ γέροντι· ἐὰν μὴ πίῃς, ϕονεύω σε. γνοὺς δὲ ὁ γέρων ὅτι ἐντολὴν
Θεοῦ θέλει ποιῆσαι, βουλόμενος αὐτὸν κερδῆσαι, ἔλαβε καὶ ἔπιεν.
21
A recent Cambridge PhD dissertation by Theodore Markopoulos, ‘The Category “Future”
in Greek: A Diachronic Investigation of Three Future-referring Periphrastic Forms,’
deals with μέλλω, ἔχω and θέλω from the Hellenistic to the late Medieval period. I
thank the author and others for bringing this to my notice after my paper. My examples
have been collected independently. [Published in 2008; see Postscript.]
22
A name in square brackets after a reference indicates the scholar who proposed this exam-
ple; if there is no name, it is my own proposal. See bibliography for works referred to.
AUXILIARY THELO 245
[The robber chief] … filled a cup of wine and with his sword in his hand said to
the elder, ‘If you don’t drink, I’ll kill you.’ The elder, knowing that he was about
to perform a command from God, and wishing to win over the robber, took it
and drank.
4. Silouanos 1
καὶ ἐξελθόντων αὐτῶν, εὗρεν ὁ μαθητὴς αὐτοῦ ὕδωρ ἐν τῇ ὁδῷ καὶ ἤθελε
πιεῖν· καὶ λέγει αὐτῷ ὁ γέρων· Ζαχαρία, νηστεία σήμερον·
After they had set out, his disciple found water on the way and was going to take
a drink. The elder said to him, ‘Zacharias, fast day today!’
See also Makarios 1; 11.
5. Acta Conc. Ephes. 1.1.2, p. 40 l. 19 (ed. Schwartz) (431 AD)

ἢ πῶς Χριστιανοὶ θέλουσιν ὀνομάζεσθαι οἱ λέγοντες εἰς ἄνθρωπον ἅγιον
ὡς ἐπὶ ἕνα τῶν προϕητῶν ἐληλυθέναι τὸν λόγον καὶ μὴ αὐτὸν ἄνθρωπον
γεγονέναι λαβόντα ἐκ Μαρίας τὸ σῶμα, ἀλλ᾽ ἕτερον εἶναι τὸν Χριστὸν καὶ
ἕτερον τὸν τοῦ θεοῦ λόγον τὸν πρὸ Μαρίας καὶ πρὸ αἰώνων υἱὸν ὄντα τοῦ
πατρός;
Or how will they be called Christians who say that the Word came into a holy man
as upon a prophet and that he did not become a man by taking his body from
Maria, but that Christ is one thing and the Word of God, who was the Son of the
Father before Maria and before the ages, is another?
6. PMichael 39.10, 14 (V AD?) [Gignac]

ει (= ἡ) μητέρα σου ǀ ἀσθενῖ, ἀποθανῖν θέλι. ǀ ἐὰν δυνασθῇς σοὺ βάλλε ǀἠ
ινοψεσουσωπου. ει μιǀτέρας σου εἶπι ἀποθανῖν θέλιν. [Punctuation and some
accents added.]
Your mother is sick, she is going to die. If you can, put. … Your mother said she
is going to die.
7. MAMA 1.160.4 (IV AD?)

ας ᾽Ωαλεντίνη ǀ ζῶν ϕρονῶν ἀ[ν]ǀήστησα μνήμης ǀ χάριν. ἄν τις θεǀλήσι ἀνῦξεν
(= ἀνοῖξαι) ἔǀξ[ωθε]ν τοῦ γένǀο[υς, ἕξ]η πρὸς τὴǀν [Τριάδ]αν.
I …-as Valentini while alive and in my right mind set up [this tomb] as a memorial.
If anyone outside my family shall open it, he will have to face the Trinity.
8. IG XIV 1563.2 (Rome) [Dieterich]

Εἰρῆνα ἔζησεν ἔτη ζ. ἐὰν οὖǀν τις αὐτὴν θελήσῃ ἀνορύξǀαι, τὸν μέλλοντα
αἰῶνα μὴ κλǀηρονομήσῃ.
Eirena lived seven years. If anyone shall dig her up, may he not inherit the age
to come.
9. TAM 5.1.213.7 (261/2 AD)

εἴ τις θελήσει ǀ σκυβαλλίσαι τὸ μνῆμα τοῦτο, ἕǀξει τὸν Ἀπόλλωνα κεχολωμέǀνον
καὶ τὴν κυρίαν Ἀναεῖτιν διὰ ǀ τέκνα τέκνων, ἔγονα (= ἔκγ-) ἐγόνων.
If anyone shall foul this tomb, he will incur the wrath of Apollo and the lady
Anais for his children’s children, his descendants’ descendants.
10. IPrusaOlymp 1.83.9 (II AD)

… τὸ μνημ[εῖ]ǀον κατασκεουασάντων ǀ Μαρκέλλας τῆς γυναικὸς ǀ αὐτοῦ καὶ τῶν
υἱῶν αὐτοῦ ǀ [- - -] ἀνεξοδίαστον σ[ὺν] τ[ῷ] ǀ περικήπῳ. ἐὰν δέ τις θ[ε]ǀλήσει,
δώσει τῇ πόλει προσǀ[τ]είμου (δην.) ͵ βϕʹ. χαίρετε.
… the tomb, prepared by Marcella his wife and his sons, … not to be taken over
by someone else, together with the garden around it. If anyone does [take it], he
will give to the city a penalty worth 2500 dinaria. Farewell.
Similarly (total 12): IG XIV 238.10 [Dieterich]; 625.6 [Dieterich]; MAMA
6.234a.6 (c.275 AD); IGBulg 1.218.7; 3.1.996.2; IKios 100.8; TAM 2.3.1086.5;
TAM 5.1.741.8 (244/5 AD); 5.1.776.11 (305/6 AD); 5.2.1077.10; 5.2.1083.7;
5.2.1107.6.
11. Acta Xanthippae et Polyxenae 7.25 (ed. James) (III AD) [Jannaris]
εἰ ἦν μοι δυνατόν, ἤθελον ἅψασθαι τοῦ κρασπέδου τῶν ἱματίων αὐτοῦ, ἵνα
ἴδω [1. εἰδῶ?] τὴν εὐμένειαν καὶ τὴν πρόσδεξιν αὐτοῦ καὶ εὐωδίαν.
[Xanthippe has caught sight of Paul walking in the street outside her house and
says:] If it were possible for me, I would take hold of the hem of his clothes, so
that I might see/know his goodwill and acceptance and fragrance.
12. POxy 14.1763.10 (III AD, after 222) [Costas; Joüon, ‘Les verbes’]
οὔπω μέχρι σήμεǀρον τὰ πλοῖα τῆς ἀνǀνώνας ἐξῆλθεν ǀ ἵνα δυνηθῶμεν ἐξελǀθεῖν,
καίτοι ἐμοῦ μηǀδὲν ἔχοντος πρᾶξαι ǀ ἐνθάδε. λέγουσι δὲ ὅτι ǀ μέχρι ιε θέλομεν ǀ
ἐξελθεῖν σὺν θεῷ.
Up till today the grain-supply ships have not left so that we could leave, yet
I have nothing to do here. They say that we will leave by the 15th with god’s
help.
13. PMichael 17.3 (II–III AD)

εἰ ǀ οἶδας ὅτι θέλεις μετενέγκαι τὸν ǀ σῖτον εἰς Πέψα, γράψον πῶς ǀ μέλλομεν
αὐτὸν μετενέγκαι.
If you know that you will transfer the grain to Pepsa, write [and tell me] how we
are going to transfer it.
14. X. Eph. 2.13.3 (II–III AD)

ἔδει δὲ τὴν Ἀνθίαν οὕτως ἱερουργηθῆναι. ὡς δὲ πάντα ἕτοιμα ἦν καὶ κρεμνᾶν
τὴν κόρην ἤθελον, ψόϕος τῆς ὕλης ἠκούετο καὶ ἀνθρώπων κτύπος.
AUXILIARY THELO 247
It was necessary for Anthia to be sacrificed in that manner. When all was ready
and they were about to hang the girl up, a rustling was heard in the bushes and
the sound of men moving.
15. POxy 10.1293.18 (117–138 AD)

τοῦτο οὐχ ἕνεκα ἡμῶν ǀ ποιῶ ἀλλὰ ἕνεκα τῶν καμηλειτῶν, ǀ μὴ θελήσῃ τις
ἀϕεῖναι μέρος ǀ μὴ ἐνένκας.
I am not doing this [i.e., asking for confirmation of receipt of a previous load of
oil] on our account, but because of the camel drivers, in case any of them leaves
part [of the load] and does not bring it.
16–17. Hermas, Vis. (II AD)

16. 3.1.9
λέγει μοι· κάθισον ὧδε. λέγω αὐτῇ· Κυρία, ἄϕες τοὺς πρεσβυτέρους πρῶτον
καθίσαι. ὅ σοι λέγω, ϕησίν, κάθισον. θέλοντος οὖν μου καθίσαι εἰς τὰ δεξιὰ
μέρη οὐκ εἴασέ με, ἀλλ᾽ ἐννεύει μοι τῇ χειρὶ ἵνα εἰς τὰ ἀριστερὰ μέρη καθίσω.
She says to me, ‘Sit here.’ I say to her, ‘Lady, let the elders sit first.’ ‘Do as I say,’
she says, ‘Sit.’ Then when I was about to sit on her right she stopped me and sig-
nalled to me with her hand to sit on her left.
17. 3.2.3
ταῦτα εἴπασα ἤθελεν ἀπελθεῖν· πεσὼν δὲ αὐτῆς πρὸς τοὺς πόδας ἠρώτησα
αὐτὴν κατὰ τοῦ κυρίου ἵνα μοι ἐπιδείξῃ ὃ ἐπηγγείλατο ὅραμα.
After she said this she was going to leave; but I fell at her feet and asked her by
the Lord to show me the vision which she had promised.
See also 3.3.1
18. Hypothesis of Euripides, Alexandros. POxy 52.3650.29 (early II AD)23

… οἵτινες ἡττῆσθαι διαλαβ[ό]ντες ὑπὸ δούλου κατηξίωσαν τὴν Ἑκάβην ὅπως
ἂν αὐτὸν ἀποκτείνῃ. παραγενηθέντα δὲ τὸν Ἀλέξανδρον Κασ[σάν]δρ[α] μὲν
ἐμμανὴς ἐπέγνω καὶ π[ερὶ τῶ]ν μελλόντων ἐθέσπισεν, Ἑκάβη [δὲ ἀπο]κτεῖναι
θέλουσα διεκωλύθη. π[α]ρα[γενό]μενος δ᾽ ὁ θρέψας αὐτὸν διὰ τὸν κίνδυνον
ἠναγκάσθη λέγειν τὴν ἀλήθειαν.
… who, supposing they had been defeated by a slave, urged Hekabe to kill him.
When Alexandros arrived, Kassandra in a raving state recognised him and prophe-
sied what would happen; but Hekabe as she was about to kill him was prevented:
the man who reared him arrived and because of the danger was compelled to tell
the truth.
23
I first noticed this example some years ago at a seminar on this text by the late Kevin
Lee in Sydney. [My suggestion that θέλουσα might mean ‘being about to’ was met with
little enthusiasm at the time.] Coles’s translation in the ed. princ. in POxy was ‘Hecabe
who wished to kill him’; in Collard et al., Euripides, 2:50–1, it is ‘Hecuba, who was
ready to kill him.’ There is no comment on θέλουσα in either edition.
19–20. Vita Aesopi G (ed. Perry; Ferrari) (I AD)
19. 99.6
καθ᾽ ὃν καιρὸν ἦν ὁμόϕωνα τὰ ζῷα τοῖς ἀνθρώποις, πένητα ἀπορούμενον
τροϕῆς ἐπιλαβέσθαι [δὲ] ἀκρίδας τὰς λεγομένας ‹τερετιστρίας› καὶ ταύτας
ταριχεύειν καὶ πωλεῖν ϕανερῆς τιμῆς. πιάσας δέ τινα ἀκρίδα ἠθέλησεν
αὐτὴν ἀποκτεῖναι. ἡ δὲ ἰδοῦσα τὸ μέλλον πρὸς τὸν ἄνθρωπον εἶπεν. …
At the time when animals had the same speech as human beings, a poor man lack-
ing sustenance caught grasshoppers called hummers and pickled them and offered
them for sale at a certain price. He caught a certain grasshopper and was about to
kill her, but she, seeing what was going to happen, said to the man. …
20. 91.9
ἡ μὲν οὖν τοῦ σημείου λύσις ἐστὶν αὕτη· πάντως τις τῶν βασιλευόντων θελή-
σει ὑμῶν τὴν ἐλευθερίαν καταδουλῶσαι καὶ τοὺς νόμους ἀκυρῶσαι καὶ ἐπι-
σϕραγίσαι τῇ ἰδίᾳ δυνάμει.
The interpretation of the sign is this: one of the reigning kings will for certain
reduce you from freedom to slavery, nullify your laws and put the stamp of his
power on you.
21. Aesopi Fab. (ed. Chambry) 142 ῞Ιππος καὶ στρατιώτης.

ὅτε δὲ ὁ πόλεμος κατέπαυσεν, εἰς δουλείας τινὰς καὶ ϕόρτους βαρεῖς ὁ ἵππος
ὑπούργει. … ὡς δὲ πάλιν πόλεμος ἠκούσθη καὶ ἡ σάλπιγξ ἐϕώνει, τὸν ἵππον
χαλινώσας ὁ δεσπότης καὶ αὐτὸς καθοπλισθεὶς ἐπέβη. ὁ δὲ συνεχῶς κατέπι-
πτε μηδὲν ἰσχύων· ἔϕη δὲ τῷ δεσπότῃ· ἄπελθε μετὰ τῶν πεζῶν [τῶν] ὁπλιτῶν
ἄρτι· σὺ γὰρ ἀϕ᾽ ἵππου εἰς ὄνον με μετεποίησας, καὶ πῶς πάλιν ἐξ ὄνου ἵππον
θέλεις ἔχειν;
When the war ended, the horse served at various tasks and carried heavy loads.
… When war was declared again and the trumpet called, the master put the bridle on
the horse, put his armour on and mounted. But the horse continually fell down because
he had no strength. He said to his master: ‘Go with the foot soldiers now. You turned
me from a horse into a donkey; how will you get a horse again from a donkey?
22–26. New Testament (I AD)
22. Matt 26:15

Τότε πορευθεὶς εἷς τῶν δώδεκα ὁ λεγόμενος Ἰούδας Ἰσκαριώτης πρὸς τοὺς
ἀρχιερεῖς εἶπεν, Τί θέλετέ μοι δοῦναι, κἀγὼ ὑμῖν παραδώσω αὐτόν;
Then one of the Twelve called Ioudas Iskariotes went to the chief priests and
said, ‘What will you give me and I will hand him over to you?’
23. Mark 6:48 [Turner; Schrenk; Joüon, ‘Θέλειν’; Taylor; Bratcher and Nida; al.]
καὶ ἰδὼν αὐτοὺς βασανιζομένους ἐν τῷ ἐλαύνειν, ἦν γὰρ ὁ ἄνεμος ἐναντίος
αὐτοῖς, περὶ τετάρτην ϕυλακὴν τῆς νυκτὸς ἔρχεται πρὸς αὐτοὺς περιπατῶν
AUXILIARY THELO 249
ἐπὶ τῆς θαλάσσης· καὶ ἤθελεν παρελθεῖν αὐτούς. οἱ δὲ ἰδόντες αὐτὸν ἐπὶ τῆς
θαλάσσης περιπατοῦντα ἔδοξαν ὅτι ϕάντασμά ἐστιν, καὶ ἀνέκραξαν·
Seeing them struggling in their rowing, since the wind was against them, about the
fourth watch of the night he came to them walking on the sea; and he was going
to go past them, and when they saw him walking on the sea they thought it was a
ghost, and cried out.
See also John 6:21 [Riesenfeld; Schrenk]
24. John 1:43 [Riesenfeld; Schrenk; Joüon, ‘Θέλειν’]

Τῇ ἐπαύριον ἠθέλησεν ἐξελθεῖν εἰς τὴν Γαλιλαίαν καὶ εὑρίσκει Φίλιππον.
καὶ λέγει αὐτῷ ὁ Ἰησοῦς, Ἀκολούθει μοι.
The next day he was about to go out into Galilee and he finds Philip. And Jesus
says to him, ‘Follow me.’
25. Acts 14:13 [Riesenfeld; Schrenk]

ἐκάλουν τε τὸν Βαρναβᾶν Δία, τὸν δὲ Παῦλον Ἑρμῆν, ἐπειδὴ αὐτὸς ἦν ὁ
ἡγούμενος τοῦ λόγου. ὅ τε ἱερεὺς τοῦ Διὸς τοῦ ὄντος πρὸ τῆς πόλεως ταύρους
καὶ στέμματα ἐπὶ τοὺς πυλῶνας ἐνέγκας σὺν τοῖς ὄχλοις ἤθελεν θύειν. ἀκού-
σαντες δὲ οἱ ἀπόστολοι Βαρναβᾶς καὶ Παῦλος, διαρρήξαντες τὰ ἱμάτια αὐτῶν
ἐξεπήδησαν εἰς τὸν ὄχλον κράζοντες καὶ λέγοντες, Ἄνδρες, τί ταῦτα ποιεῖτε;
… καὶ ταῦτα λέγοντες μόλις κατέπαυσαν τοὺς ὄχλους τοῦ μὴ θύειν αὐτοῖς.
They called Barnabas Zeus and Paul Hermes, since he was the leader in the speeches.
The priest of Zeus Before the City brought bulls and garlands to the gateway and
together with the crowd was about to offer sacrifice. The apostles Barnabas and
Paul, when they heard of it, tore their clothes and leapt into the crowd crying out
and saying, ‘Men, what is this you are doing? …’ With their words they barely
stopped the crowd from offering sacrifice to them.
26. Acts 19:33

ἐκ δὲ τοῦ ὄχλου συνεβίβασαν Ἀλέξανδρον, προβαλόντων αὐτὸν τῶν Ἰουδαίων·
ὁ δὲ Ἀλέξανδρος κατασείσας τὴν χεῖρα ἤθελεν ἀπολογεῖσθαι τῷ δήμῳ. ἐπι-
γνόντες δὲ ὅτι Ἰουδαῖός ἐστιν, ϕωνὴ ἐγένετο μία ἐκ πάντων ὡς ἐπὶ ὥρας δύο
κραζόντων, Μεγάλη ἡ Ἄρτεμις Ἐϕεσίων.
Some of the crowd instructed(?) Alexandros, whom the Jews put forward. Alexan-
dros motioning with his hand was about to make a defence to the popular assembly.
But when they realised he was a Jew, with one voice they all cried out for about two
hours, ‘Great is Artemis of the Ephesians!’
27. Test XII Patr, TReub 1.7 (ed. De Jonge) (c.50 AD?) [Sophocles; Jannaris;
Lampe]
λέγω γὰρ ὑμῖν, ὅτι ἐνέπληξέ με πληγὴν μεγάλην ἐν ταῖς λαγῶσί μου ἐπὶ
μῆνας ἑπτά· καὶ εἰ μὴ Ἰακὼβ ὁ πατὴρ ἡμῶν προσηύξατο περὶ ἐμοῦ πρὸς
Κύριον, ὅτι ἤθελε Κύριος ἀνελεῖν με.
I tell you that he struck me with a great affliction in my flanks over seven months,
and if Jacob our father had not prayed for me to the Lord, [I tell you] that the Lord
would have destroyed me.24
28. POxy 55.3806.7 (15 AD)

[ἠ ἠ ]ἠ ωνίου ἀναπλέοντος ἀναγκαῖον ἔγνων ἀ[σπ]άǀ[σα]σθαί σε διὰ γραπτοῦ
καὶ παρακαλέσσαι σε γράϕειν ǀ μοι περὶ ὧν ἐὰν θέλῃς. ἥδιστα γὰρ ποιήσωι.
τῶι ǀ ἀδελ(ϕῷ) σου κατὰ ἑκάστην ἡμέραν παρεδρεύωι, ǀ μὴ θέλει ἐπιστολάς
σοι πέμψαι. τὸ δῖγμα τοῦ ǀ [ἐ]ριδίου δῖξον Φιλοῦτι καὶ γράψον μοι ἠ (= εἰ)
ἀρέσǀκει αὐτῆι ἢ οὔ.
Since ...-onios is sailing up I decided I must greet you in writing and ask you to
write to me about what you want, for I will gladly do it. I apply to/attend on your
brother every day in case he will send you letters. Show Philous the specimen of
the wool and write and tell me if she likes it or not.
29–31. LXX
29. Tobit 3:10 GII (II BC?) [Schrenk]

ἐν τῇ ἡμέρᾳ ἐκείνῃ ἐλυπήθη ἐν τῇ ψυχῇ καὶ ἔκλαυσεν καὶ ἀναβᾶσα εἰς τὸ
ὑπερῷον τοῦ πατρὸς αὐτῆς ἠθέλησεν ἀπάγξασθαι. καὶ πάλιν ἐλογίσατο καὶ
λέγει Μήποτε ὀνειδίσωσιν τὸν πατέρα μου. . . . χρησιμώτερόν μοί ἐστιν μὴ
ἀπάγξασθαι, ἀλλὰ δεηθῆναι τοῦ κυρίου ὅπως ἀποθάνω.
On that day she was grieved in her soul and wept, and she went up to her father’s
upper room and was going to hang herself. And she considered again and said,
‘Never let them reproach my father. … Better for me not to hang myself but to
pray to the Lord that I might die.’
30. Tobit 6:15 GII

ἤκουσα ὅτι ἑπτὰ ἤδη ἐδόθη ἀνδράσιν, καὶ ἀπέθανον ἐν τοῖς νυμϕῶσιν αὐτῶν
τὴν νύκτα, ὁπότε εἰσεπορεύοντο πρὸς αὐτήν, καὶ ἀπέθνῃσκον. καὶ ἤκουσα
λεγόντων αὐτῶν ὅτι δαιμόνιον ἀποκτέννει αὐτούς. καὶ νῦν ϕοβοῦμαι ἐγώ –
ὅτι αὐτὴν οὐκ ἀδικεῖ, ἀλλ᾽ ὃς ἂν θελήσῃ ἐγγίσαι αὐτῆς, ἀποκτέννει αὐτόν.
I have heard that she has already been given in marriage to seven men and they
died in their bridal chamber, on the night when they went in to her they would
die. And I heard it said that a demon kills them. And now I am afraid, because
[the demon] does not harm her, but whoever comes near/wants to come near her
it kills.
Cf. GI ὅτι δαιμόνιον ϕιλεῖ αὐτήν, ὃ οὐκ ἀδικεῖ οὐδένα πλὴν τῶν προσα-
γόντων αὐτῇ.
24
In their translation, Hollander and De Jonge, Testaments of the Twelve Patriarchs, supply
‘I would have died’ before the ὅτι clause, rendered ‘because the Lord wanted to kill me.’
Such a supplement is unnecessary if θέλω has the auxiliary use (which they may not have
considered).
AUXILIARY THELO 251
31. Exod 2:14 (III BC)

περιβλεψάμενος δὲ ὧδε καὶ ὧδε οὐχ ὁρᾷ οὐδένα, καὶ πατάξας τὸν Αἰγύπτιον
ἔκρυψεν αὐτὸν ἐν τῇ ἄμμῳ. ἐξελθὼν δὲ τῇ ἡμέρᾳ τῇ δευτέρᾳ ὁρᾷ δύο ἄνδρας
Ἐβραίους διαπληκτιζομένους καὶ λέγει τῷ ἀδικοῦντι Διὰ τί τύπτεις τὸν πλη-
σίον; ὁ δὲ εἶπεν Τίς σε κατέστησεν ἄρχοντα καὶ δικαστὴν ἐϕ᾽ ἡμῶν; μὴ
ἀνελεῖν με σὺ θέλεις, ὃν τρόπον ἀνεῖλες ἐχθὲς τὸν Αἰγύπτιον;
He looked this way and that, and saw no one, so he struck the Egyptian and buried
him in the sand. Going out the next day he saw two men, Hebrews, fighting. He
said to the one hurting the other, ‘Why do you strike your neighbour?’ He said,
‘Who appointed you ruler and judge over us? Are you going to kill me the way
you killed the Egyptian yesterday?’
32. PHib 1.65.25 (c.265 BC) [Mayser]

[θέλ]ομεν οὖν ἐγ δηǀ[μο]σίου τὸν λοιπὸν ǀ [συν]αγοράσαι σῖǀ[το]ν ἵνα μη[θὲ]ν ǀ
[εἰς ἐ]μὲ ὑστερήσηι.
[The writer has asked the recipient for help in supplying part of the grain he owes
and explains how he will obtain the rest:] I want to/am going to purchase the
remainder of the grain from the state, so that there might be no arrears against me.
33–35. Plato (IV BC)
33. [Plato], Alkib. I 122d [Wifstrand]

τοῦτο μὲν γὰρ εἰ ἐθέλεις τοὺς Λακεδαιμονίων πλούτους ἰδεῖν, γνώσῃ ὅτι πολὺ
τἀνθάδε τῶν ἐκεῖ ἐλλείπει.
For in this matter, if you consider the wealth of the Lacedaemonians, you will
recognise that things here are very much inferior to those there.
34. Protag. 334b [Wifstrand]

… οἷον καὶ ἡ κόπρος, πάντων τῶν ϕυτῶν ταῖς μὲν ῥίζαις ἀγαθὸν παραβαλλο-
μένη, εἰ δ᾽ ἐθέλοις ἐπὶ τοὺς πτόρθους καὶ τοὺς νέους κλῶνας ἐπιβάλλειν, πάντα
ἀπόλλυσιν.
… as for example dung, which when applied to the roots of any plants is a good
thing, whereas if you were to put it on the young shoots and twigs it destroys them
all.
35. Rep. 423b [LS7, LS8]

Τίς, ἔϕη, ὅρος; Οἶμαι μέν, ἦν δ᾽ ἐγώ, τόνδε· μέχρι οὗ ἂν ἐθέλῃ αὐξομένη εἶναι
μία, μέχρι τούτου αὔξειν, πέρα δὲ μή.
‘What measure?’ he said. ‘In my opinion this one,’ I said, ‘To the point where [the
city], while increasing in size, will still be one, to that point will [the guardians]
let it grow, and no further.’
See also Pl., Theaet. 162e; Parm. 158c; Alkib. I 122b, c; Charm. 174c; Lysis 217a,
c; Meno 71a; Hipp. Mi. 373a; Rep. 581c [all Wifstrand]; 436b [LS7, LS8].
36. Aristophanes, Wasps 536 (422 BC) [LS7, LS8; Jannaris; LSJ]25
ὁρᾷς γὰρ ὥς σοι μέγας ἐστὶν ἁγὼν
καὶ περὶ τῶν πάντων,
εἴπερ – ὃ μὴ γένοιτο –
νῦν ἐθέλει κρατῆσαι.
[Chorus to Bdelykleon:] You see how great the contest is for you and how every-
thing is at stake, if he wins now – which I hope won’t happen.
37. Antiphon 4.2.7 (V BC) [Rödiger]

ὡς μὲν οὖν οὐ δικαίως κατηγοροῦμαι, ἐπιδέδεικταί μοι· ἐθέλω δὲ τοὺς κατη-
γοροῦντάς μου πᾶσιν οἷς ἐγκαλοῦσιν ἐνόχους αὐτοὺς ὄντας ἀποδεῖξαι.
That I am unjustly accused I have demonstrated; but I will show that my accusers
are themselves liable to all the accusations they bring against me.
38–40. Herodotus (V BC)
38. 1.109.4 [LSJ]

εἰ δ᾽ ἐθέλει τούτου τελευτήσαντος ἐς τὴν θυγατέρα ταύτην ἀναβῆναι ἡ τυραν-
νίς, τῆς νῦν τὸν υἱὸν κτείνει δι᾽ ἐμεῦ, ἄλλο τι ἢ λείπεται τὸ ἐνθεῦτεν ἐμοὶ
κινδύνων ὁ μέγιστος;
If on his death the sovereignty passes to this daughter, whose son he is now killing
by my hand, am I not hereafter in extreme danger?
39. 2.14.1 [Waddell]

εἴ σϕι θέλοι, ὡς καὶ πρότερον εἶπον, ἡ χώρη ἡ ἔνερθε Μέμϕιος (αὕτη γάρ
ἐστι ἡ αὐξανομένη) κατὰ λόγον τοῦ παροιχομένου χρόνου ἐς ὕψος αὐξάνε-
σθαι, ἄλλο τι ἢ οἱ ταύτῃ οἰκέοντες Αἰγυπτίων πεινήσουσι . . . ;
If, as I said before, the land below Memphis (this is the part that is increasing)
should increase in height at the same rate as in the past, isn’t it inevitable that
the Egyptians who live there will go hungry . . . ?
40. 9.89.2
ὁ δὲ ᾽Αρτάβαζος γνοὺς ὅτι, εἰ ἐθέλει σϕι πᾶσαν τὴν ἀληθείην τῶν ἀγώνων
εἰπεῖν, αὐτός τε κινδυνεύσει ἀπολέσθαι καὶ ὁ μετ᾽ αὐτοῦ στρατός. …
Artabazos, realising that if he were to tell them the whole truth about the battles,
he and his army would be in danger of destruction, …
See also Hdt. 1.32.3; 2.11.4 [LSJ]; 2.99.3; 3.12.1; 7.10d.2 [Rödiger]; 7.49.4
[LSJ].
25
The source lies earlier: this example is noted in the Paris ed. (1831–65) of Stephanus’s
Thesaurus.
AUXILIARY THELO 253
41. Sophocles, Ant. 1040 (V BC) [Rödiger]

τάϕῳ δ᾽ ἐκεῖνον οὐχὶ κρύψετε,
οὐδ᾽ εἰ θέλουσ᾽ οἱ Ζηνὸς αἰετοὶ βορὰν
ϕέρειν νιν ἁρπάζοντες ἐς Διὸς θρόνους
[Creon to Teiresias:] You shall not bury that man, not even if Zeus’s eagles seize
him as food and carry him to Zeus’s throne.
42. Aeschylus, Cho. 851 (458 BC) [Rödiger]

Χο. ἠκούσαμεν μέν, πυνθάνου δὲ τῶν ξένων
ἔσω παρελθών. οὐδὲν ἀγγέλων σθένος
ὡς αὐτὸν αὐτῶν ἄνδρα πεύθεσθαι πάρα.
Αἰ. ἰδεῖν ἐλέγξαι τ᾽ εὖ θέλω τὸν ἄγγελον,
εἴτ᾽ αὐτὸς ἦν θνήισκοντος ἐγγύθεν παρών,
εἴτ᾽ ἐξ ἀμαυρᾶς κληδόνος λέγει μαθών·
Chorus: We heard the story; but you pass inside and learn [it] from the strangers.
There is no strength in messengers’ reports like learning in person from them.
Aegisthus: I will see and examine well the messenger, whether he himself was
present close by when [Orestes] died, or tells by learning from an uncertain report.
See also Sophocles, OC 1291 (end of V BC); Euripides, Or. 770 (408 BC);
Aeschylus, Cho. 904 [all Rödiger].
6.2 θέλησον, θελήσατε, ‘(please) do’

PAnt 3.197.2 (VII AD) θελήσατε στρῶσαι τῷ παρόντ(ι) ἀνθ(ρώπῳ) ǀ ἀνερ-
χο(μένῳ) ἱππάρ(ια) τρία.
Please saddle three ponies for this man who comes up.
PCairMasp 1.67061.1 (VI AD) θελήσατε οὖν παραχρῆμ[α τοῦτο] ǀ ποιῆσαι.
POxy 10.1300.5, 8, 9 (V AD) θέλησον οὖν, κύρα μήτηρ, ἀποστῖλε … θέλησον
οὖν, κύρα μήτηρ, ἀγοǀράσε … καὶ θέλ[η]σον λαβῖν τὸ κοῦκλιν.
POxy 16.1941.5 (V AD) θέληǀσον ἀποστῆναι τῆς γεωρǀγίας μηχανῆς Στύμονος.
SB 6.9158.20 (V AD) θέλησον δὲ ἐν τάǀχι ἐξαποστῖλαι ǀ τὸ σημῖον ǀ μὴ ὀκνου-
μένη, ǀ ἀμμή, οὐκ ἔχοǀμεν εἰς ἡπερεǀσίαν (= ὑπηρ-).
Do send the signet quickly without delay, mama: we don’t have one to use.
BGU 3.948.11, 13, 18 (IV–V AD) θέλησον οὖν πέμψεν μοι … θέλησον οὖν ǀ
ποιῆσεν μοι … θέλησον … ἀγοράσιν (= -ειν) ǀ μοι.
POxy 14.1776.6 (late IV AD) θέλησον ὅσον ǀ χρῄζουσι ἔν τε σίτῳ ἢ ǀ ὄξους
παράσχου αὐτοῖς.
Aesopi Fab. 40 (ed. Chambry) θέλησον οὖν τοὺς ἐμπροσθίους πόδας ἐρεῖσαι
τῷ τοίχῳ, ὀρθῶσαι δὲ τὰ κέρατα, ἀναδραμοῦσα δὲ ἐγὼ καὶ σὲ ἀνασπάσω.
Just rest your front feet against the wall and hold your horns up, and I will run up
[out of the well] and pull you up too.
6.3 μὴ θελήσῃς, μὴ θελήσητε, ‘(please) don’t’

POxy 59.4005.1 (VI AD) μὴ θελήσῃς μηδεμίαν ϕροντίδαν ἔχειν περὶ τῆς
ἡμῶν ἀδελϕῆς Μαρίας, σοῦ δὲ ǀ συμβίου. οὐκ ἐοῦμεν (= ἐῶμεν) γάρ, τοῦ θεοῦ
θέλοντος, αὐτὴν δεηθῆναί τινος.
Don’t have any concern about our sister Maria, your wife. For we will not, God
willing, allow her to want for anything.
Acta Conc. Ephes. 1.1.2, p. 68 l. 14 (ed. Schwartz) (431 AD) Εἰ θέλετε ἀκοῦσαι,
ἡσυχάσατε καὶ μανθάνετε· μὴ θελήσητε ἐμποδίσαι τοῖς λεγομένοις, ἀλλὰ
μακροθυμήσατε, ἵνα τὰ ῥήματα ἀκριβῶς ἀκούσητε.
If you are willing to hear, keep quiet and learn; do not obstruct what is being said,
but be patient. …
PAmh 2.143.16 (IV AD) μὴ θελήǀσῃς οὖν, κύριε, μῖνε ἐκǀτὸς ἡμῶν αὔριον διὰ
τὴν ǀ ἀϕορμὴν τοῦ ὕδατος ǀ εἵνα δηνηθῶμεν ǀ ποτίσαι τ[ὸ]ν μέγαν κλῆǀρον.
Don’t stay away from us tomorrow, sir, for the release of the water, so that we
can irrigate the large allotment.
PWisc 2.74.5, 20 (III–IV AD) καὶ πάνοι (= πάνυ) χρίαν ἔχω{.ι}μεν τῇ παρου-
σίᾳ σου. μὴ θεǀλήσῃς οὖν παραμῖναι παρά σοι τοῦ σαι ἀπαντῆǀσαι πρὸς ὑμᾶς
καὶ ἀπολέσωμεν τὴν ǀ οἰκίαν ἡμῶν. … μὴ θέλῃς ἀμαιλῆσαι καὶ ὕστερα μετα-
μελή‹σ›ῃς.
PFlor 2.210.9 (255 AD) ἀλλὰ μὴ θελήσῃς μοι πέμǀψε ὄξος· καὶ γὰρ ἐντολὰς
ἔσχον ǀ διὰ τοῦ ἐπιστολιδίου παρὰ τοῦ ǀ ε[ὐσ]χήμονος.
Tobit 4:5 GI, simil. GII πάσας τὰς ἡμέρας, παιδίον, κυρίου τοῦ θεοῦ ἡμῶν
μνημόνευε καὶ μὴ θελήσῃς ἁμαρτάνειν καὶ παραβῆναι τὰς ἐντολὰς αὐτοῦ·
Postscript
In Markopoulos, The Future in Greek, the development of θέλω is discussed

several times (40–2, 76–7, 80–1, 104–12, 164–77) and examples of auxiliary θέλω
are noted. Our collections overlap: some of his examples are missing from mine,
and some of mine from his. A consolidated collection would be useful. Markopou-
los discusses θέλω in requests at 81–2, with PWisc 2.74.20 and POxy 14.1776.6
(quoted above) as examples, and at 110, with PPrag 2.193.2 (V AD) θέλησον,
εἴπερ βούλει, ἐκμισθῶσαι αὐτὸ Καναίῳ τῷ ἁλιεῖ.
Another relevant study is Tronci, ‘Future Forms in Postclassical Greek,’ with
a discussion of ‘future-oriented’ periphrases using μέλλω and θέλω in the LXX
and NT (133–7), including nos. 22, 29–31 above. Tronci takes the LXX exx.
to be ‘ingressive,’ as allegedly implied in my translations, concluding that ‘the
AUXILIARY THELO 255
periphrases with θέλω … can never be translated as plain futures in the LXX’
(134). Her conclusion is similar for the NT (137). La Roi, however, disagrees
about Exod 2:14, arguing that ‘this example is best interpreted as a future use
in this context’ (‘Diachronic Analysis,’ 222).
In observing the use of θέλω in polite requests (§4, examples at §6.2, 3),
I recorded only aorist imperatives and did not notice that the present imperative
can be used in the same way (‘please do/don’t). Aitken’s discussion in ‘Literary
Attainment,’ 116–7 alerted me to this. He records two cases in Sirach (Sir 6:35
θέλε; 7:13 μὴ θέλε) and two in non-biblical sources (Vita Aesopi [‘P, 247’]26 μὴ
θέλε; Life of Adam and Eve 31 μὴ θέλε), with the remark that these are ‘among
many koine examples’ (117). A search in <papyri.info> brings to light this nice
example:
BGU 4.1205.7 (28 BC) κομ|[ψῶ]ς ἐμοὶ θέλετε τὴν αἰτίαν | γράψαι το[ῦ]
φακοῦ καὶ πίσον | [. .] οὐδὲ γὰρ σύνστοιχοι ἑατῶν | [γ]ίνεσθε γεγρα-
φηκότες Πα|[νίσ]κωι, ὅτι πεπράκαμεν, ὅτε | [. . .] οὐ πεπράκαμεν.
Do kindly write to me the reason for the lentils and peas … for you are
not even consistent with yourselves, having written to Paniskos that we
sold (them), when [you know?] we have not sold (them).
26
This is a ref. to ‘Vita Planud.’ (diff. || Vita G 32.7). I have doubts about this one.
17
RELEASING LSJ FROM ITS PAST
20101
Abstract
What is the future of Liddell and Scott’s lexicon to be? That was the main
question posed for the Cambridge Colloquium in 2002. This paper addresses the
topic by examining the history of the entry for one word, ἀγαπητός, from the first
edition of the lexicon in 1843 to LSJ. The treatment of the word in NT lexicons
and commentators is part of the story. Six characteristics of the tradition emerge
from the analysis, most of them inimical to satisfactory lexicographical practice.
The paper moves on to propose a new approach, utilising the potential of an elec-
tronic database. The possible content of such a database is outlined and a revised
analysis of ἀγαπητός is offered.
INTRODUCTION
‘Liddell and Scott’ is an institution, the great lexicon of Greek not

only familiar to every scholar and student of Greek and Latin, but known
to the world at large, if only by name. For more than a century and a half
it has been the chief lexical resource for ancient Greek in the English-
speaking world, and far outside it. The first edition, prepared by the two
young Oxford scholars on the basis of the lexicon of Franz Passow,2 was
published in 1843. After a succession of editions, each incorporating new
material and revisions, it reached an eighth in 1897. A more extensive
revision was undertaken by Stuart-Jones and McKenzie to produce the
current ‘new edition’ or ninth, issued in parts from 1925 to 1940. Numer-
ous ‘Addenda et Corrigenda’ had already accumulated at the end of the last
part. There followed a supplement in 1968, and another in 1996, gather-
ing, and sometimes correcting, all the earlier additional material and adding
more. Offspring of the larger lexicon were also produced, the medium-sized
1
This is an expanded version of a paper delivered at the Colloquium on Ancient Greek
Lexicography held in Cambridge, 7th–9th July, 2002. I am grateful to Anne Thompson
for reading it and suggesting a number of improvements.
2
Passow, Handwörterbuch (1831).
‘Intermediate’ in 1889, based on the 7th edition, and the smaller ‘Abridged’
in 1843, followed by a 2nd edition in 1846. There was also a little-known
anonymous derivative in 1850, the ‘Copious Greek–English Vocabulary.’3
Now at the beginning of the twenty-first century the great lexicon has
reached a decisive stage in its history. It is evident that the work cannot
stand still and remain as it is, but must move forward. It already requires
further revision, not only because of the build-up of new material, but
because of the pressing need for correction of many of the existing entries.
That the latter is true is not readily conceded by all, especially those who
have not moved beyond implicit trust in what a lexicon tells them. But the
fact is well known to those who have had occasion to study LSJ’s entries
closely, and has often been demonstrated, in recent years especially by
John Chadwick.4 What is LSJ’s future to be? This paper is designed to
address that question and consider what part an electronic database might
play in the process.
In the first half of this paper I focus on LSJ’s entry for one word, not
primarily to show that LSJ have got something wrong – that is easy to
do for any number of entries – but to examine the history of the entry. This
will enable us to draw out the reasons why the entry is now as it is, namely,
in a state of disarray. Its present state will be seen to be as much a result
of its history, and the methods and constraints imposed on the lexicogra-
phers as they worked on it, as of poor judgement and scholarship on their
part. I will go on to suggest what will be required to cure the problems
in the treatment of this word, and by extension all the words in the lexicon,
and to argue that these processes will be best carried out in an entirely new
medium. By this means LSJ will be enabled to escape the mistakes and
pressures of the past and go on to a new (and better) stage in its life.
THE HISTORY OF AN ENTRY
The word I have chosen as an illustration is ἀγαπητός. I have recently

had occasion to examine this word in a study of the history of lexicons
of the NT, a history lasting five hundred years.5 Its treatment there con-
nects directly with the way it has been treated in the main lexicons of
3
Details of the editions of the larger LS(J) may be seen in Lee, History, 355–6. The com-
piler of the ‘Copious’ was George Marshall, as also of the ‘Abridged.’ I thank Chris Stray
for this information.
4
Chadwick, ‘Semantic History’; idem, ‘Replacing Liddell and Scott’; idem, Lexicographica
Graeca.
5
See Lee, History.
RELEASING LSJ FROM ITS PAST 259
ancient Greek, and in a special way with LSJ itself, as we shall see. Let
me begin with some remarks on the semantic history of this word as I
understand it.
The point of most interest is the supposed meaning ‘only,’ whether
alone or in combination with some other notion, to give meanings like
‘only-beloved, only-begotten.’ Current NT lexicons put forward such a
meaning in a number of NT occurrences and adduce parallels in other texts
from Homer onwards. The Septuagint is also supposed to offer support.
I have set out in detail elsewhere the case for rejecting ‘only’ as a false
meaning.6 It can be traced back through a long line of commentators and
lexicon entries, with a few doubters along the way. Its leading proponent
in the twentieth century was the Oxford NT scholar C. H. Turner, who
argued the case fully in a 1926 article.7 The ultimate source turns out to
be a Homeric gloss (on Il. 6.401), preserved in Hesychius and elsewhere.
Scrutiny shows that neither this nor any of the other alleged examples pro-
vides a sound basis for recognising such a meaning. I am not alone in
my scepticism: Chadwick also rejected this meaning, though he did not
pursue its history in LS and elsewhere.8 In what follows I trace the pro-
gress of this meaning into LSJ and its effect on the shape of the entry
for ἀγαπητός.
The line of semantic development in ἀγαπητός seems clear. The pri-
mary meaning is ‘beloved.’ From ‘beloved’ come weakenings to ‘likable,’
and ‘to be acquiesced in’; the first sense often has an intensified nuance like
‘especially loved, precious.’ The pattern is parallel to that in the verb ἀγα-
πάω and in its equivalent στέργω, i.e., ‘love’ > ‘feel affection for’ > ‘be
content with.’9 There is no meaning ‘only.’ Still less can it be the primary
meaning, as Turner would have us believe.
Let us turn to the entries at four points in the history of Liddell and
Scott’s lexicon. We begin with LS1 (1843):
ἀγαπητός, ή, όν, verb. Adj. from ἀγαπάω, beloved, Hom.;
μοῦνος ἐὼν ἀγαπητός (the only, and so) doubly beloved son, Od.
2. 365: longed for, desired, Pind. N. 8. 6: worthy of love,
6
Lee, History, 193–211.
7
Turner, ‘Ὁ υἱός μου.’
8
Chadwick, Lexicographica Graeca, 32–4.
9
The early semantic history of ἀγαπάω is complicated by the presence of ἀγαπάζω along-
side it: both are found in Homer, ἀγαπάζω only in the meaning ‘treat/receive with affec-
tion,’ ἀγαπάω once in that sense and once as ‘be content’ (with/that). The problem of how
these denominatives (as they would normally be) could be formed before the appearance
of ἀγάπη (earliest attested c.500 BC, as a name) remains unsolved. Cf. Chadwick, Lexi-
cographica Graeca, 32.
lovable, Plat. Alc. 1. 131 E. II. neut. ἀγαπητόν (ἐστι),

one must be content, εἰ–, ἐάν–, Plat. and Arist. Adv. -τῶς,
lovingly, readily, contentedly, Plat.; ἀγαπητῶς ἔχειν, to be con-
tented, like ἀγαπᾶν : – but in Att. prose, so as only just to
content one, i. e. only just, barely, scarcely, = μόλις, Heind.
Plat. Lys. 218 C; ἀγαπητῶς σωθείς Dem., cf. Meineke Menand.
p. 108.
There is little to disagree with here; at any rate, the general structure is
right. Beloved is primary and the other senses are developments of that.
‘Only’ is far from being put forward as a meaning: it appears as the equiva-
lent of μοῦνος in Od. 2.365, where ἀγαπητός is rightly treated as beloved.
The suggestion that it might in fact mean doubly beloved in this sort of
context seems to me to have merit. (This was derived from Passow.)10
What ἀγαπητός is most unlikely to mean in Od. 2.365 is ‘only.’ Yet this
is exactly the meaning Turner wishes to give it, apparently on the basis
that it refers to an only child. But as Chadwick says, ‘the epithet μοῦνος
would be otiose, if ἀγαπητός meant the same. It is evident that [the phrase]
means “being greatly loved because you are an only son”.’11
Next let us go to LS4 (1855), where things begin to go wrong:
ἀγαπητός, ή, όν, verb. Adj., beloved, μοῦνος ἐὼν ἀγαπητός the
only (and so doubly) beloved son, Od. 2. 365; hence without
μοῦνος, of an only son, Ἑκτορίδην ἀγαπητόν Il. 6. 401, cf. Od. 4.
817; so in Att., Νικήρατος . . ὁ τοῦ Νικίου ἀγ. υἱός Dem. 567. 24;
cf. Arist. Rhet. 1. 7, 41. 2. desirable, delightful, ἀγαπατά (ἐστι),
c. inf., Pind. N. 8. 6: – freq. in Att. Prose, worthy of love, loveable,
Plat. Alc. 1. 131 E, etc. II. to be acquiesced in (as the least
in a choice of evils), Andoc. 26. 15: – hence, ἀγαπητόν (ἐστι) one
must be content, εἰ . . , ἐάν . . , Plat. Prot. 328 A, Xen. Oec. 8. 16,
Dem. 302. 1, Superl., -ότατος, Plat. Phil. 61 E. So Adv. -τῶς,
readily, contentedly, Plat. Legg. 735 D, etc.; ἀγαπητῶς ἔχειν, to
be contented, like ἀγαπᾶν: – but also in Att. Prose, so as only just
to content one, i.e. only just, barely, scarcely, = μόλις, Heind. Plat.
Lys. 218 C: ἀγαπητῶς σωθῆναι Lys. 107. 16, so, ἀγαπητόν,
Menand. μέθη 1.
The general structure is still much the same, with beloved as the starting-
point. There are additional examples here and there, and an improvement
in the definition of sense II. But in ‘hence without μοῦνος, of an only son’
LS not only make a false deduction, but confuse reference and lexical
meaning. ‘Of an only son’ would be a correct statement, i.e., that the word
10
[Passow4 (1831): ‘der einzige, darum doppelt geliebte Sohn.’]
11
Chadwick, Lexicographica Graeca, 33.
refers to an only son; but by italicising ‘only’ they turn only into the lexical
meaning. Now they have set running, almost by accident, the meaning
only, in addition to beloved. Do they really mean that the examples that
follow mean nothing more than only? Possibly they do, but it is not quite
clear, and one wonders if they have thought through the consequences of
writing ‘of an only son.’
Now let us see what shape the entry had reached in LS8 (1897), before
the major revision for the ninth edition was undertaken:
ἀγαπητός, ή, όν, Dor. -ατός, ά, όν, verb. Adj., beloved, μοῦνος ἐὼν
ἀγαπητός the only dearly beloved son, Od. 2. 365; more commonly
without μοῦνος, of an only son, Ἑκτορίδην ἀγαπητόν Il. 6. 401, cf. Od.
4. 817; so in Att., Ar. Thesm. 761; Νικήρατος . . ὁ τοῦ Νικίου ἀγ. παῖς
Dem. 567. 24, cf. Arist. Rhet. 1. 7, 41, al.; Comically, δαπίδιον ἓν ἀγ.
Hipparch. Ἀνασ. 1. II. of things, worthy of love, loveable,
desirable, dear, Plat. Alc. 1. 131 E, etc.; Sup. -ότατος Id. Phil. 61 E;
τὸ ἀγαπητόν an object of desire, Arist. Rhet. 1. 7, 41, al. 2. to be
acquiesced in (as the least in a choice of evils), Andoc. 26. 15: – hence,
ἀγαπητόν [ἐστι] one must be content, εἰ . . , ἐάν . . , Plat. Prot. 328 A,
Xen. Oec. 8. 16, Dem. 302. 1, Arist., etc.; c. inf. Eth. N. 9. 10,
6. III. Adv. -τῶς, readily, gladly, contentedly, Plat. Legg.
735 D, Dem. 409. 7, etc. 2. to one’s heart’s content, Diphil. Incert.
4. 3. just enough to content one, only just, barely, scarcely, = μόλις,
Plat. Lys. 218 C; ἀγαπητῶς σωθῆναι Lys. 107. 16; so also, ἀγαπητόν
Menand. Μέθη 1.
There has been more tinkering since the fourth edition, most of it not
beneficial. Od. 2.365 is now clearly (and rightly enough) stated to mean
dearly beloved. But ‘of an only son’ for the next group of examples sur-
vives unscathed, with the minor change of ‘hence without μοῦνος’ to
‘more commonly without μοῦνος,’ presumably because two more exam-
ples have been added. Other changes are serious missteps. The new posi-
tion of ‘II’ separates what was sense [I.]2 desirable . . . worthy of love, etc.
into a new major category away from where it belongs with [I.] beloved,
and to be acquiesced in is demoted to a subcategory of that, surely not an
improvement. Worse, ‘of things’ has been inexplicably introduced ahead
of Pl., Alc. 1.131e in II.1, which it will be useful to see at once:
οὔτ᾽ ἐγένεθ᾽, ὡς ἔοικεν, Ἀλκιβιάδῃ τῷ Κλεινίου ἐραστὴς οὔτ᾽ ἔστιν
ἀλλ᾽ ἢ εἷς μόνος, καὶ οὗτος ἀγαπητός, Σωκράτης ὁ Σωφρονίσκου καὶ
Φαιναρέτης.
Alkibiades son of Kleinias, it seems, neither had nor has any lover
except one only, and that very much loved, namely Socrates, son of
Sophroniskos and Phainarete.
Socrates, it seems, is a thing. What caused this lapse we cannot say;

what we can say is that making any change to a sense without checking the
examples cited under it is fraught with risk.
Finally we come to the result of the revisions made to produce the ninth
edition, the one we rely on today. LSJ (1925–40) reads as follows:
ἀγαπ- … … … … … … … … ητός, ή, όν, Dor. -ατός, ά, όν, that
wherewith one must be content (cf. ἀγαπάω III), hence of only children,
μοῦνος ἐὼν ἀ. Od.2.365; Ἑκτορίδην ἀ. Il.6.401, cf. Od.4.817, Sapph.
85, Ar. Th.761, Pl.Alc.1.131e; Νικήρατος . . ὁ τοῦ Νικίου ἀγ. παῖς D.21.
165, cf. Arist.Pol.1262b23, EE 1233b2; αὕτη μονογενὴς αὐτῷ ἀγαπητή
LXX Jd.11.34, cf. To.3.10, Ev.Marc.12.6 (but cf. also II.2), etc.; ἀγα-
πητός· μονογενής, Hsch.: – so of things, Arist.Rh.1365b16; δαπίδιον
ἓν ἀ. Hipparch.Com.1; προβάτιον Men.319.3. 2. to be acquiesced
in (as the least in a choice of evils), And.3.22, J.BJ 5.10.3: – hence,
ἀγαπητόν [ἐστι] one must be content, εἰ . . , ἐάν . . Pl.Prt.328b, X.Oec.
8.16, D.18.220, Arist.Metaph.1076a15, etc.; c. inf., EN 1171a20. II.
of things, desirable, ἤθη X.Mem.3.10.5; βίος Pl.Phlb.61e (Sup.). 2.
of persons, beloved, ἀδελφὲ ἀγαπητέ LXX To.3.10: in letters, as a
term of address, Ep.Rom.12.19, cf. PGrenf.2.73, etc. III. Adv.
-τῶς, gladly, contentedly, Pl.Lg.735d, D.19.219, etc. 2. just
enough to content one, barely, scarcely, Pl.Lys.218c; ἀ. σωθῆναι Lys.
6.45, cf. Diph.89.2, etc.
The semantic structure of the entry, and with it the presumed semantic
history of the word, has been reversed. The meaning previously taken to
be primary, beloved, is relegated to sense-category II. Here, apparently,
it begins as a weaker sense desirable applied to things, before developing
to beloved, which is allegedly not found before the book of Tobit in the
LXX. As the primary meaning we now have the impressive that where-
with one must be content. This seems a carefully crafted and insightful
definition. That was certainly how it struck me as an undergraduate when
I acquired (with some excitement) my own copy of the big LSJ and saw
this entry. But mature reflection suggests otherwise. And I can now trace
where it originated – in a source outside LSJ.
To begin with, that wherewith one must be content is the definition
of a noun, not an adjective. A definition in this form is unjustified and
unhelpful. The more serious question is, is this meaning really intended to
apply to all the examples that are listed in [I.1]? Since no other meaning
is stated, and ‘hence of only children’ indicates reference only, not lexi-
cal meaning, we must suppose it is. Let us take Od. 2.365. Do LSJ really
mean that Eurykleia says to Telemakhos, ‘you being an only son, one with
whom his parents must be content’? Or again that Pl., Alc. 1.131e (above)
. . . εἷς μόνος, καὶ οὗτος ἀγαπητός means ‘Alkibiades had one lover only,
one that he had to be content with’? Or that Il. 6.401 παῖδ᾽ . . . Ἑκτορίδην
ἀγαπητόν can be tortured into meaning ‘the boy, Hektor’s son, the one he
had to be satisfied with’? It seems unlikely.12
The fact is that it is difficult or impossible to make the meaning that
wherewith one must be content work in the examples cited (with one pos-
sible exception, Men. 319.3 = 224.3 Kock-Austin: cf. the lexical analysis
below). It appears rather that LSJ intend this as the base meaning from
which sense [I.1] is supposed to derive. But what is that sense? They have
failed to say. Chadwick thought they did not mean to say it was ‘only.’13
I strongly suspect they did. The Hesychius gloss quoted is a good clue;
and we know they were working from LS8, with its ‘of an only son,’ fol-
lowed by many of the same examples. Perhaps they just forgot to italicise
‘only.’ But it is very unsatisfactory that the user of the lexicon is reduced
to guessing what meaning is being stated.14
Some further details. The reference to ‘To. 3.10’ appears twice, in [I.1]
and II.2, the first preceded by ‘cf.,’ an unhelpful and irritating device
which leaves the reader guessing, and may be suspected to cover up uncer-
tainty or indecision on the part of the editor. Are we meant to suppose that
ἀδελφὲ ἀγαπητέ might possibly mean ‘brother I have to accept,’ or even
‘only brother,’ as well as ‘beloved brother’? It gets worse. The reference
in II.2 to To. 3.10 ἀδελφὲ ἀγαπητέ, the alleged first instance of the mean-
ing beloved, is wrong: Tob 3:10 (GII text) reads μία σοι ὑπῆρχεν θυγάτηρ
ἀγαπητή. It is in Tob 10:12 (GI text) where ἄδελφε ἀγαπητέ is found.
Clearly the reference in II.2 should be changed to 10:12; but what of the
other in [I.1]? Should it stay as it is or be changed also?
The separation of meaning [I.]2 to be acquiesced in . . . hence . . . one
must be content from the first meaning was justified, indeed required,
in the structure that prevailed in preceding editions. But to retain it as a
separate category after introducing as sense [1.1] that wherewith one must
be content is asking for trouble: the two overlap in meaning and therefore
a reassessment of the second definition and its examples is called for.
The point in all this is that not only are LSJ’s meanings unsatisfactory,
but the placing of the examples is in disarray. We can see that this has
12
LSJ Suppl. (1996) adds two more to this group, ‘Ge. 22.2, cf. Am. 8.10.’ Again, it is
impossible to see how the definition can apply. Rather, the editor seems to be thinking
of the meaning ‘only.’ On these two examples see further Lee, History, 203–6.
13
Chadwick, Lexicographica Graeca, 33.
14
It is curious that in Lampe, PGL, s.v. ἀγαπητός, LSJ is quoted thus: ‘(cf. LS: with which
one must be content, hence of only children),’ i.e., with only italicised and the definition
adjusted to adjectival form.
come about through the process of revision, with its attendant difficulties.
A reversal of the semantic structure along with the introduction of a new
meaning is a particularly hazardous operation to carry out. All the cita-
tions in the entry need to be checked if it is to be done without introducing
mistakes. One cannot simply strike out one definition and insert another
and assume that all will be well.
I said I could trace the origin of that wherewith one must be content.
C. H. Turner, as already mentioned, published his influential article in Jan-
uary 1926, and in it he refers to part one of LSJ, which had just been pub-
lished in 1925. Turner in fact uses the new shape of LSJ’s entry to bolster
his case for taking υἱὸς ἀγαπητός to mean ‘only son’:
It is most instructive here to compare the new edition of Liddell and
Scott – the first part is happily available for the subject of this lecture –
with the previous one, and to note the change in the treatment of the
word.
He goes on to extract some of LSJ’s Classical examples and discuss
them, before drawing the conclusion that
the assertion may safely be hazarded that when ἀγαπητός is used in
connexion with υἱός, θυγατήρ [accent sic], παῖς, or similar words, no
Greek of pre-Christian times would have hesitated in understanding
it of an ‘only child,’ or would for a moment have thought of any other
meaning as possible.15
Apart from offering the latter highly questionable statement, Turner has
been a little less than candid with us. Before the 1926 paper, there appeared
in 1919 an anonymous article containing the same argument and much of
the same material as in 1926. There is no clue to the author’s identity in
the 1919 paper; but a footnote in the 1926 paper contains a clear indication
that it was Turner.16 In the 1919 paper we find the following statement:
With regard to pre-Christian usage, it may be noted that in classical
Greek the word ἀγάπη is unknown . . . and that ἀγαπάω means mainly
‘to be contented with’ . . . : ἀγαπητός would therefore properly be
‘what one has to be contented with,’ and so ‘all that one has,’ and then
finally ‘the exclusive object of interest or affection.’17
15
Turner, ‘Ὁ υἱός μου,’ 115, 117.
16
Turner, ‘Ὁ υἱός μου,’ 125 n. 1: ‘Most of the patristic passages have already been col-
lected in an article on the word ἀγαπητός which appeared in J.T.S. xx 339–344 (1919) as
a preliminary draft for the Lexicon of Patristic Greek. I should like to acknowledge here
the help given me on the classical side on that occasion by my colleague Prof. J. A. Smith
and by the late Charles Cannan.’
17
Turner, ‘Ἀγαπητός,’ 339 n. 2. The statement that ‘in classical Greek . . . ἀγαπάω means
mainly “to be contented with”’ is wide of the mark. See the survey of usage of Classical
So Turner had already proposed ‘what one has to be contented with.’ It

is extremely likely, then, that Turner was the source, by whatever channel,
of LSJ’s new definition for sense I.18 Thus Turner himself had a signifi-
cant influence on the entry which he then used to support his case. The
interest of all this, apart from the tale itself, is in being able to rediscover,
and thus assess, the origin of this major shift in the LSJ entry. The revisers
of LS8 appear to have relied, with insufficient caution, on a NT scholar’s
enthusiastic advocacy of a meaning that he had set his sights on for the sake
of its exegetical value in the NT, but that needed more objective assess-
ment by someone less involved.19
CHARACTERISTICS OF THE TRADITION
One could go on unpacking the detail, but that is quite enough. LSJ’s entry
for ἀγαπητός is a mess. Why? One might simply answer that the editors
of successive editions have not done their job properly; but that only touches
the surface. There are deeper reasons, most of them beyond the control of
the editors. They lie in the characteristics of the tradition, both methodo-
logical and practical. These have helped to bring about, indeed ensured, the
result we now have. Let me summarise these features as I see them.
(1) Control of the evidence. By ‘evidence’ I mean the raw data for the
lexicographer’s task, the occurrences of the word being treated; by ‘con-
trol’ I mean gathering and assessing them adequately. How many occur-
rences of ἀγαπητός are there? LS1 cites 3 (not counting broad categories
like ‘Hom.’); by LSJ there are 27 (not counting ‘etc.’). In DGE there are
about 70. Electronic searching via the TLG and PHI7 reveals a tally of more
than 4,000 examples of this word in literature up to 1453 and in documen-
tary texts.20 No lexicographer of the past could hope to find all these and
authors in Joly, Vocabulaire de l’amour, 14–9. In Isocrates, for example, there are 50 occur-
rences, of which 38 show the sense ‘love’; in Plato there are 10, of which 8 have that
sense; in Demosthenes 27, of which 12 (all figures from Joly).
18
The likely channel, as Chris Stray suggests to me, was Cannan, who was Secretary to the
Delegates of OUP (1898–1919) and whose help is acknowledged by Turner (see n. 16).
Patrick James suggests that Alexander Souter may have had a hand in it too. Cf. mention
of him in the LSJ preface (p. x) and his short note (Souter, ‘ Ἀγαπητός’) supporting
Turner: cf. Lee, History, 207 n. 1.
19
Further critique of Turner’s methods and alleged examples of the meaning ‘only’ can be
seen in Lee, History, 197–202 and Chadwick, Lexicographica Graeca, 32–4. Chadwick
(ibid. 32) already implied, somewhat cryptically, that Turner might have been behind
the change in LSJ, but did not elaborate.
20
PHI7 = The Packard Humanities Institute, CD ROM 7 Greek Documentary Texts (1991–
1996).
assess them. Even those they had were not assessed each time they were
re-used in LS, as is evident from the case of Pl., Alc. 1.131e. I am not
suggesting that Liddell and Scott or their successors ought to have tried
to find all this material, and then to have checked every example before
writing their entry, which would somehow have encompassed it all. Given
the physical difficulty of finding such a vast body of data by the methods
available to them and of examining it in a reasonable amount of time,
it was impossible. My point is that we now know there is much more
abundant evidence than the small amount LSJ had to rely on, and that that
evidence not only exists but is readily accessible and could be brought to
bear on the lexical analysis. There is a further reason, besides the practi-
cal and time constraints, why LSJ based their entry on a small sample, and
that relates to the next point.
(2) Dependence on predecessors. Lexicographers in general (with some
rare exceptions) rely for most of their material on preceding lexicons. This
is not only because it is easier to reuse what someone has already assem-
bled and analysed (or not), but because it is impossible to do otherwise
unless one is willing to spend, say, 50 years gathering and assessing all the
data afresh. In the case of a long-running series like LS(J), it was natural
each time to trust the material in the previous edition and simply make
revisions to it. But there is a risk that dubious lexical decisions remain
unquestioned and so are perpetuated. A case in point is the perpetuation
of ‘of an only son’ through several editions, without anyone noticing the
confusion of reference and meaning; another is the questionable survival
of the long-standing to be acquiesced in as a separate sense in LSJ after
something similar was newly introduced to replace the existing sense I.
(3) Piecemeal revision. At the same time as reusing the bulk of the
material of a previous lexicon or edition, editors of lexicons make piece-
meal changes and additions. This spasmodic intervention may cause more
trouble than leaving the entry alone. This is because any change may impact
on the rest of the entry but tends not to be accompanied by a reexamination
of the entry as a whole. A glaring case in the LSJ tradition is the intro-
duction of a new definition for sense I, as we have seen, without adequate
reflection on how it is meant to work in the examples cited for that sense.
(4) Control of secondary literature. Lexicons do not proceed in a
vacuum. A vast body of discussion in books and journals accompanies their
progress. In that literature all sorts of additions, corrections, reactions,
and comments can be found. Editors of lexicons attempt to take account
of this material. No doubt the revisers who produced LSJ did so to the
best of their capacity. But what means do we have of knowing? And what
reporting of it might usefully have been provided, both to show the sources
of information and as an aid to further research? Such questions have not
entered LSJ’s horizon, nor at any stage in the history of LS. In the world
of NT lexicons some coverage of the secondary literature is almost de
rigueur.21 It would of course have been impossible to do this in LSJ with-
out doubling the size of the work. But it is a gap we are entitled to feel,
and that we may look for a way of remedying. In the case of ἀγαπητός,
it would have been helpful indeed to find in LSJ a reference to the source
that inspired the change to sense I of ἀγαπητός.
(5) Physical constraints. This is a factor of major importance. I refer
to the limitations of presenting lexicon entries on the printed page. All
the way from the simple ancient glossary made up of lemmata in the form
‘word x = word y’ through to LSJ and modern printed lexicons, there is
a single line of development. From one point of view all that has happened
over time is the accumulation of further information up to the limits of the
printed book. The ultimate size of the book has always constrained the
amount of material that could be presented in each entry and determined
the format of presentation, including the need for abbreviation. For the lexi-
cographer at work on compiling the entry, there is always space pressure.
A selection has to be made from a larger body of data and whatever is
finally included has to be compressed as much as possible. The actual tasks
of lexical analysis and description, difficult enough, are made more diffi-
cult: the lexicographer is worrying about how to keep the size of the entry
down, and material that would have been helpful or illuminating has often
to be omitted to save space.
To this of course is linked the matter of ease of reading and compre-
hension by the user of the lexicon. The user tends to be on the losing side
of the contest. In the case of LSJ, besides the general compression, we
have had to accept the practice of shortening headwords by the use of
hyphens and combining groups of headwords into large paragraphs. The
motive was obviously to save space; the result is very inconvenient to the
reader.22 None of this would have been necessary if the content of entries
had not been governed by what can be presented in a printed work.
21
Cf. esp. BDAG, building on earlier material gathered by Bauer from 1928 onwards.
Older lexicons, such as Thayer’s at the end of the nineteenth century, provided similar
bibliographical matter.
22
See Chadwick, ‘Replacing Liddell and Scott,’ 2 for caustic comments on the innovation
and the name of the person who had the ‘effrontery’ to claim credit for ‘this astonishing
lack of perceptivity.’ Another device is the use of cross-references to other words, intro-
duced by ‘=’. This too is presumably to save space, but may end up not doing so, and
simply wastes the reader’s time. See, e.g., s.v. σκήνωμα: first we are told ‘= σκήνημα,’
Under this heading is also to be noted the sheer physical difficulty of

preparing copy. This impacts on the work of the editors and their deci-
sions about how to arrange, or re-arrange, the entries. LSJ provides a telling
illustration. The preface to LSJ gives a description of how the revisions to
LS8 were done in the 1920s. Corrections and insertions were written in the
margins of a paste-up of the existing entries.23 This procedure would have
been intricate and prone to error, and a discouragement to major changes.
For Chadwick it was a ‘basic fault’ in preparing the new edition; it ‘effec-
tively prevented any attempt to modernise the style or to recast the struc-
ture of the articles.’24 Anyone who uses a computer for word-processing
knows how radically different things are today.
(6) Method of indicating meaning. This is a deeper methodological
problem, one that will not be fixed in a day. The lexicon of Liddell and
Scott, from LS1 through to LSJ, like all its predecessors from time imme-
morial, has relied primarily on the unsatisfactory gloss method. The alter-
native is the ‘definition’ method, which was pioneered by Murray in the
Oxford English Dictionary and handed on to the Oxford Latin Dictionary.
This development came too late to have an impact on LS(J). In the defi-
nition method of indicating lexical meaning, rather than simply using
one or more translation equivalents, an attempt is made to define the area
of meaning in an exact and unambiguous way, and to distinguish it from
other areas of meaning. In the NT sphere, there has been lengthy discus-
sion of the problems associated with the use of glosses and the advan-
tages of a shift to definitions. In the case of ἀγαπητός there is, as it hap-
pens, considerable use by LSJ of what might pass for definition; but the
usual glosses are also seen, as e.g., in III.2 just enough to content one,
barely, scarcely, earlier only just, barely, scarcely. These last three are
simply possible translations in context. Similarly in II.[1] desirable. This
gloss shows the potential for the method to be misleading, even with the
qualifier ‘of things.’ One might be tempted to think, wrongly, of ‘wanted,
sought after,’ or even ‘attractive.’ (Cf. below for an attempted definition).25
then under the latter we find ‘= σκηνή.’ The meaning tent nevertheless appears further
on under σκήνωμα and might just as well have been given in the first place. More seri-
ously, this device fails to tell which meaning of the cross-referenced word is the relevant
one. Thus in the case of ‘σκήνημα = σκηνή,’ when we look up σκηνή we are left to take
our pick from no less than nine senses of this word.
23
LSJ, ‘Preface 1925,’ x. [The marked-up galley proofs of the sixth edition (1869), repro-
duced in Imholtz, ‘Liddell and Scott,’ 130, vividly illustrate the horrors of such a process.]
24
Chadwick, Lexicographica Graeca, 7–8.
25
For discussion of definition versus glossing, see Lee, History, 15–29, with references
esp. to the work of Louw and Nida on the NT; see also BDAG, viii. The gloss method
A NEW APPROACH
These six points indicate the constraints under which the editors of LSJ
worked. No lexicographer of the past could escape them. What ought to
be done next? I do not believe an old-style revision of LSJ is worth doing
now. That is, the sort in which piecemeal changes and additions are made
without a complete overhaul, and the result is another printed book of the
same size and scope as LSJ. My reasons for saying this are two. First, to
carry out such a revision would be to continue the methods of the past
and accept the same constraints; it would not cure the problems of LSJ,
but would make them worse. Secondly, there is now an opportunity to
do things in an entirely different and better way, that is, by using the elec-
tronic medium, and transferring the whole operation to electronic form.26
The rest of my paper is concerned with exploring what this would mean.
I should make it clear that I am concerned with the conception rather than
the technical detail. The technical expertise to carry out the task is already
available; lexicographers need to consider how they would like to see it
harnessed to their needs.
What I have in mind is not simply to make LSJ available on screen,
but to reconceptualise our approach to the tasks of lexicography and the
production of lexicons.27 The text of LSJ has already been entered elec-
tronically, and this could be used as a first step, but it would be no more
than that. What needs to be developed is a database in which everything
can be stored: not only the meanings and citations that we currently con-
sult the lexicon for, but all the data that lie behind them and are in any way
relevant to them. Moreover, as well as being a storehouse of information,
it would be a record of work carried out to produce it. There is of course
no limit on the size of an electronic storage system, so one of the problems
of the past would simply disappear.
Obviously this is a big and long-term enterprise. It may seem to some
to be an unrealistic ideal. The practical difficulty of actually getting started
and then continuing with the work over many decades certainly looks
is of course the staple of smaller lexicons such as Morwood and Taylor, Pocket Oxford,
in which the meanings given for ἀγαπητός are: ‘beloved; amiable; welcome; sufficient.’
These are derived unchanged from Feyerabend, Langenscheidt’s Pocket Greek Dictionary
(1910). The potential for misunderstanding is evident.
26
[John Chadwick had already arrived at a similar view in 1996. In Lexicographica Graeca,
28 he wrote: ‘There is no case at all to be made for an indefinite expansion of the book
[i.e., LSJ] in printed form. … Thus we have arrived at the parting of the ways: the total
lexicon must henceforth be kept in electronic form. …]
27
The proposals here for a database are similar to those made for NT lexicography in Lee,
History, 182–8.
forbidding; it may seem better not to start. If this is so, it is because we

are still in transition from the old way of doing things, that is, working in
the medium of print. Sooner or later, in my opinion, an electronic database
for Greek lexicography will be seen to be inevitable and will be created,
and then maintained indefinitely. To take a simple illustration. Many small
revisions and additions that might be made to LSJ have steadily accumu-
lated since the publication of the last Supplement in 1996. Would anyone
wish to see another printed supplement to LSJ to take account of them?
Rather, it will seem natural to gather such material electronically and make
it available in that form.
In Stuart-Jones’s 1925 preface to LSJ, the tale is told of how the idea
of a ‘Thesaurus Graecus’ emerged in 1904 but was quickly abandoned as
unachievable, not least because of the likely bulk of the resulting work.28
It was indeed impossible – in printed form, which was the only one that
could be envisaged at the time. The thinking then naturally ran along the
lines of a vast collection of handwritten slips, like those for the TLL and
OED, which would eventually lead to a printed work. Diels’s guess that
it might run to 120 volumes was enough to frighten anyone. In some ways
the preservation of this story in our leading lexicon has had an unfortunate
effect. It has cemented the idea that ancient Greek is too immense to do
more than dip into as far as lexicography is concerned, and a complete
collection of examples is neither possible nor necessary. But by using the
electronic medium it can be done. The texts of Greek literature are now
available in the TLG, and the majority of published documentary texts
in PHI7; it is only a matter of re-sorting the data to arrive at a complete
list of occurrences of every word, the first step towards a ‘thesaurus’ of
Ancient Greek. Individual researchers are already doing this every day on
a small scale when they carry out electronic searches for words they are
studying.29
What then might the ideal database contain for a word such as ἀγαπη-
τός, and by extension every word in LSJ? What information would one
like to be able to find and see presented on screen, if one were to look up
this word? It is not difficult to say, at least as regards the primary informa-
tion. As to the rest, I do not see why our horizons should be limited in any
way. There is nothing to stop us including all the data related to a given
28
LSJ, ‘Preface 1925,’ v–vi. A fuller account and discussion of the ‘Thesaurus’ proposal
may be found in Cohn, ‘Griechische Lexikographie,’ 724–30.
29
It should be understood that the TLG is not a lexicon: it provides raw data, without mean-
ings or lexical analysis of any kind.
word, much of which is currently consigned to separate reference works

and has never been assembled in one place. A suggested list of categories
of information follows, with discussion and, in the case of the ‘main entry,’
a sample layout. The question of the electronic structure that would organ-
ise the data is one I leave to others. The categories have of necessity been
placed in linear order here, in accordance with logic and convention, but
of course in the electronic form the order is immaterial.
CATEGORIES IN THE DATABASE
MAIN ENTRY
This would be like the lexicon entry for any word in a larger lexicon
such as LSJ, with lexical meanings defined and classified, each illustrated
by selected examples. Ample quotation of the context would be included
with each citation. Translations would also be given. I regard the latter as
not just an aid to the user, but a valuable discipline for the lexicographer.
It is remarkable how often one’s understanding of the meaning of a word
is sharpened, even completely revised, by the attempt to set down a trans-
lation of the passage. Since there is no limit on space, all this could be
presented in tabular form with the utmost clarity as the objective. The
setting-out in the electronic OED is a model of how well this can be done.
Besides tabulation and different font sizes, colour could be used to catch
the eye.
There would of course be links from this view to every other part of the
database. The word itself could be the lead-in to most links; in addition,
some other items, such as the citations, could be linked to further data on
them. All these other data would be material of potential interest, but not
essential to grasping the semantic contours of the word. What one saw in
the main entry would be complete in that respect; all the rest would be
amplification or support. I indicate possible links by the standard conven-
tion of underlining. The format is rudimentary; the final product would be
visually much more attractive as well as space-efficient.
The lexical analysis and selection of examples are illustrative and not
intended to be final. The selection is small and confined to the Classical
period; in the final product it would be larger and include texts of later
dates. I have expanded LSJ’s abbreviations a little. With the space con-
straint lifted, there is no need for extreme abbreviation of authors’ names
and works; in fact there may be no need for any. Nevertheless, a standard
system of abbreviation for all Greek authors and works, not forgetting
documentary texts, remains a desideratum, and could be supplied some-
where in this database.30
It is important to note the difference between the definitions and the
translations given below. The definition is an attempt to state succinctly
and unambiguously the lexical meaning seen in the examples listed under
it; the translation of a given example is simply a translation in context,
and does not necessarily use the exact words of the definition, though its
general import is the same. The test of the definition is that it fits the con-
text in place of the defined word without difference in meaning in all the
cited occurrences.
ἀγαπητός
etymology, formation, orthography, morphology
statistics, full evidence
lexical analysis, semantic history, lexical structure, syntax, bilingual data
literature
mini entry
1 regarded with special love and affection

Homer, Iliad 6.401
ἥ οἱ ἔπειτ᾽ ἤντησ᾽, ἅμα δ᾽ ἀμφίπολος κίεν αὐτῇ
παῖδ᾽ ἐπὶ κόλπῳ ἔχουσ᾽ ἀταλάφρονα, νήπιον αὔτως,
Ἑκτορίδην ἀγαπητόν.
She met him then, and with her went the maidservant holding the
tender boy in her arms, just a child, the beloved son of Hektor.
author work date more context text discussion
Sappho 132.2
ἔστι μοι κάλα πάις χρυσίοισιν ἀνθέμοισιν
ἐμφέρην ἔχοισα μόρφαν Κλέις ἀγαπάτα,
ἀντὶ τᾶς ἔγωὐδὲ Λυδίαν παῖσαν οὐδ᾽ ἐράνναν. . . .
I have a beautiful daughter, who looks like golden flowers, Cleis
my darling, whom I would not exchange for all Lydia or lovely. . . .
(Page)
30
Documentary texts, i.e., inscriptions, papyri, ostraca, etc., currently suffer from another
problem, the difficulty of identifying the same text published in different editions and
collections over the past century and more. This could be easily remedied by assigning
to each text a unique number, which would thereafter be quoted whenever the text was
re-edited or used in any way. A database could then collect all information on each
(numbered) text. Since this was written, the ‘Trismegistos’ database constructed by Mark
Depauw (www.trismegistos.org) has adopted just such a procedure.
Plato, Alc. 1.131e

οὔτ᾽ ἐγένεθ᾽, ὡς ἔοικεν, Ἀλκιβιάδῃ τῷ Κλεινίου ἐραστὴς
οὔτ᾽ ἔστιν ἀλλ᾽ ἢ εἷς μόνος, καὶ οὗτος ἀγαπητός, Σωκράτης
ὁ Σωφρονίσκου καὶ Φαιναρέτης.
Alkibiades son of Kleinias, it seems, neither had nor has any
lover except one only, and that much loved, namely Socrates,
son of Sophroniskos and Phainarete.
author work date more context text discussion more examples
of sense 1
2 regarded with pleasure and satisfaction

Xenophon, Mem. 3.10.5
πότερον οὖν, ἔφη, νομίζεις ἥδιον ὁρᾶν τοὺς ἀνθρώπους δι’ ὧν
τὰ καλά τε καὶ ἀγαθὰ καὶ ἀγαπητὰ ἤθη φαίνεται ἢ δι’ ὧν τὰ
αἰσρχά τε καὶ πονηρὰ καὶ μισητά;
Which then, he said, do you consider more pleasing to see,
men in whom fine and good and likable characteristics are
displayed, or those in whom shameful and bad and detestable
ones?
Plato, Protag. 328a

ἀλλὰ κἂν εἰ ὀλίγον ἔστιν τις ὅστις διαφέρει ἡμῶν προβιβάσαι
εἰς ἀρετήν, ἀγαπητόν.
But if there is someone who excels us even a little in leading [us]
on towards virtue, it is something to be glad of.
author work date more context text discussion31
Plato, Philebus 61e

οὐκοῦν εἰ τἀληθέστατα τμήματα ἑκατέρας ἴδοιμεν πρῶτον συμ-
μείξαντες, ἆρα ἱκανὰ ταῦτα συγκεκραμένα τὸν ἀγαπητότατον
βίον ἀπεργασάμενα παρέχειν ἡμῖν, ἤ τινος ἔτι προσδεόμεθα
καὶ τῶν μὴ τοιούτων;
So then, let us see, when we have first mixed the truest segments
of each, are these mixed together sufficient to effect and provide us
with the most agreeable life, or do we still need something more
that is of a different kind?
author work date more context text discussion more examples
of sense 2
31
This example has been understood as ‘one must be content’ since LS2 (1845), but some
rethinking may be appropriate.
3 to be tolerated as the best attainable

Andocides 3.22
μετὰ δὲ τοῦτο ὅρκους ὀμόσαντες αὐτοῖς καὶ τὴν στήλην
εὑρόμενοι ‹παρ᾽› αὐτῶν στῆσαι, κακὸν ἀγαπητὸν ἐν ἐκείνῳ τῷ
χρόνῳ, σπονδὰς ἤγομεν ἐπὶ ῥητοῖς.
After this, when we had sworn oaths with them and had secured
their permission to set up the stele, a misfortune to be accepted
at that time, we made a truce on specified terms.
Xenophon, Oec. 8.16

ἀπειλεῖ γὰρ ὁ θεὸς καὶ κολάζει τοὺς βλᾶκας. ἐὰν δὲ μόνον μὴ
ἀπολέσῃ τοὺς μὴ ἁμαρτάνοντες, πάνυ ἀγαπητόν·
For God threatens and punishes stupid persons; and if he only
refrains from destroying those who are innocent of any wrong, it
is something to be glad of.
Menander 224.3 (Kassel-Austin)

εἶτ᾽ οὐχ ὅμοια πράττομεν καὶ θύομεν;
ὅπου γε τοῖς θεοῖς μὲν ἠγορασμένον
δραχμῶν ἄγω προβάτιον ἀγαπητὸν δέκα,
αὐλητρίδας δὲ καὶ μύρον καὶ ψαλτρίας, . . .
μικροῦ ταλάντου. . . .
Then don’t we fare in proportion as we sacrifice? When I bring to
the gods a poor sheep just good enough, bought for ten drachmas,
but flute-girls and perfumed oil and harp-players, . . . worth nearly
a talent, . . .
author work date more context text discussion32 more examples
of sense 3
MINI ENTRY
The purpose of this view is of course to supply essential information
quickly. The many small (‘manual,’ ‘pocket,’ ‘concise’) dictionaries are
32
This example illustrates the need for discussion; in fact it is a good candidate for the
‘pool of unplaced or problematic occurrences’ mentioned below. The meaning is not
obvious, and opinions will differ. LSJ place it in [I.1], where, as we have seen, it is hard
to be sure what meaning they intend. DGE places it in I.2, muy amado, apreciadísimo,
estimadísimo. Gomme and Sandbach (Menander, 698) remark that προβάτιον is ‘depre-
ciatory’; cf. the translations of Edmonds (‘a lean lamb’) and Allinson (‘a scrawny
sheep’). If this is right, sense 3 seems appropriate. Another example in a comic fragment,
Hipparch. Com. 1.3 (Kock-Austin) δαπίδιον ἓν ἀγαπητὸν ποικίλον, cited since LS7,
is also hard to place.
the model. Meanings would be given, but not citations (unless in special
cases). More information about forms, to help the beginner, could be
included here. Glosses might be offered as well as the definitions estab-
lished in the main entry.
ETYMOLOGY; WORD FORMATION; ORTHOGRAPHY; MORPHOLOGY
These might possibly be treated together, since there is much overlap.

Some words would call for considerable detail in each category, others
only routine information or possibly none at all. In the case of ἀγαπητός,
one would like to see displayed the etymology (unknown); how the word
is formed (verbal adj. in -τός from ἀγαπάω); a full list of related words;
dialectal forms; declensional forms (regular).33
FULL EVIDENCE
Here would be stored and accessible on screen all occurrences of the

word in surviving texts up to whatever date was chosen as the cut-off.
Every citation would be provided with the same kinds of information as
indicated above for the citations in the main entry, and all would be clas-
sified semantically on the basis of the analysis there. A pool of unplaced or
problematic occurrences, linked to ongoing discussion, would be a natural
adjunct. The data would of course be searchable in a variety of ways, such
as by author, date, genre of work, meaning, textual variants. Summaries
and statistics would be available.
WORK AND AUTHOR DATA
Here would be found a list like the TLG Canon, which could be con-
sulted for itself, but would also be linked to author names and citations
occurring anywhere in the database.
LEXICAL ANALYSIS
It would be helpful if the database offered an explanation of the think-
ing behind the definitions and allocation of occurrences, and provided an
opportunity for ongoing discussion and fine-tuning of them. Definitions
are never arrived at easily, and those given at first might call for improve-
ment; similarly the decision about the meaning of an individual occurrence
33
Much of the information is of the kind found in Chantraine, DELG. In the case of verbs,
the database could record not only ‘principal parts’ but a complete list of verb forms,
as provided by a number of handbooks today.
is not always easy to make and could sometimes change. As assessment

of the data on the word went on, it would be useful to have a record of
such decisions, especially on contested questions (such as the existence of
the meaning ‘only’). Unsolved problems would also be noted and offered
for consideration. The earlier history of the lexical analysis, i.e., its treat-
ment in previous lexicons of all ages, even as far back as antiquity, could
find a place here.
DIACHRONIC ANALYSIS
An outline of the semantic history of the word (and its field) from its
origin onwards, ideally as far as Modern Greek. This goes hand in hand
with the lexical analysis and the ordering of senses: implicit in them is a
reconstruction of the word’s semantic evolution.34
LEXICAL–STRUCTURAL DATA
Here would be an opportunity to take account of the insights of modern
linguistics and attempt to locate the word in its place in the lexical struc-
ture of the language. The obvious first requirement would be to list the near
synonyms and antonyms, and attempt to discern the differences, if any,
between them. In the case of ἀγαπητός, one would want to learn how it
related semantically to φίλος, φίλιος, προσφιλής, στερκτός, μισητός,
ἠγαπημένος, and so on. Data on connotation and register would form part
of the picture.
SYNTAGMATICS
Data on collocation. Our lexicons are already strong on this, partly,
one suspects, to provide an aid to prose composition. Comprehensive data
could now be given. For instance, one could find a breakdown of all occur-
rences of ἀγαπητός by application (to persons, animals, things, situations).
BILINGUAL DATA
I am thinking of words where bilingual interference or interaction
between Greek and another language is relevant to the semantics of the
Greek word, as when φίλος functions in the same way as Lat. amicus to
indicate a courtier of the Roman emperor. In the case of ἀγαπητός, the
questions raised by its use as a translation of Hebrew yahid in the Sep-
tuagint would be pursued.35
34
For examples, see Chadwick, ‘Semantic History,’ and Lexicographica Graeca, passim.
35
On this see Lee, History, 203–6.
SECONDARY LITERATURE
Some remarks on this topic have been made above, under (4), among
the constraints affecting LSJ. It is safe to say there exists some discussion,
whether in special articles or monographs or in commentaries, of almost
every word in the lexicon. It would be most helpful to have all this mate-
rial gathered together in one place. Moreover, this would not only be a
collection of what had been done in the past, but an ongoing record of
discussion as it proceeds in the future. A mere list of titles is not enough;
summary of the content, if not reproduction in full, is essential. In the case
of ἀγαπητός, there are nine items listed in RBLG, and more can be found
in the NT lexicons. That is only the recent material, of the last century or so:
there is a lot more in older works, all of it potentially relevant to improving
our understanding of the word.36 Everyone who works on this word has
to chase them up; then the next person does it all again. This duplication
of effort can be brought to an end.
CONCLUSION
This then is an outline of the possible content of the ideal electronic

database for Greek lexicography. The question of the printed form is in
my opinion a non-issue. Whatever printed book is desirable can be produced
from the database. Any such book can select the right material from the
larger store, and thus be properly targeted to the intended user, unlike most
older works, which try to meet the needs of many at once. One printed
book that will remain a desideratum is an equivalent to LSJ, on a some-
what reduced scale, the book for general use by more advanced under-
graduates and scholars. This will offer a certain degree of detail, but not
attempt to supply everything. A lot of what is in LSJ now is superfluous
to the needs of the general user. As John Chadwick said, ‘much of the
information . . . ought to be consigned to a special lexicon of record.’37
The database I am envisaging is that ‘special lexicon of record.’ In it all
the information on each word will be stored and readily accessible to those
who really want it. If we go along this path, the attempt to put a sort of
digest of all the data on the printed page can be given up and be recognised
to be a thing of the past. Most of the problems I have listed will be over-
come, and LSJ will be released from the confines of its history.
36
See Lee, History, 193–211 passim.
37
Chadwick, Lexicographica Graeca, 19. Cf. also this remark (ibid.): ‘Part of the trouble
with LSJ, and even more with its Supplements, is that no attempt has been made to
define the user for whom the work is created.’
Postscript
Recent work on LSJ includes a notable volume of essays, Liddell and Scott,
edited by Stray, Clarke, and Katz, published in 2019. This book offers many
insights into the history of Liddell and Scott’s Lexicon, its methods, and possible
future directions. Among the contributions, Anne Thompson’s paper ‘Βάπτω: An
Illustration’ forms a strong complement to my essay above. Thompson presents a
close analysis of the entries for βάπτω in all the major lexicons and in Chadwick’s
Lexicographica Graeca, and then offers a new lexical analysis of her own based
on their source material.38 The need for a complete overhaul of LSJ is again evi-
dent. Further, a lucid explanation of the definition method as opposed to the gloss
method is offered in Thompson’s ‘Response’ to the remarks of a reviewer of Lid-
dell and Scott (in BMCR).
The chapter by John Henderson, ‘A1–Zythum,’ in Stray, Classical Dictionaries,
though on the history of the OLD, is both entertaining and pertinent. Much of the
history of the OLD is a ‘re-run’ of that of the OED. The familiar themes appear:
underestimation of time, cost, and staffing, overestimation of abilities,39 production
quandaries, and intrusion by non-experts. Future planners of major lexicon projects
would be well advised to read before acting.
My enthusiastic embrace of a lexicographical database that would include
everything was mainly a ‘scoping’ exercise to show what could be done. I did not
and do not imagine it can be done easily or quickly, or in its entirety in the near
future. The main point is that the move into some kind of database cannot be
avoided and a start ought to be made as soon as possible. As far as I know, there
has been no progress in that direction. The future of LSJ, and with it Ancient Greek
lexicography, remains tied to the past. Meanwhile new lexicons have appeared in
print or electronic equivalent, following traditional methods: GE, its predecessor
GI, and the recent CGL (2021); DGE has reached ἔξαυος in Vol. 7 (2007).40
38
To this analysis (with 16 senses) is added a ‘concise outline’ listing seven senses of βάπτω
(388–9). This would form the ideal basis of an entry in a shorter lexicon, with minimal
illustrations added. I do not disagree with Anne’s remarks at 393–4: ‘There is … no need
initially for exhaustive collection of source material and bibliography. A relatively small
sample of texts can be used to set up reasonably accurate preliminary models that can be
added to, corrected, and modified.’
39
Souter (in 1921) asserted that ‘Such a scheme [his plan to revise Lewis and Short]
requires no genius or group of geniuses to carry it out’ (Henderson, ‘A1–Zythum,’ 148).
Yet his own attempt at a specimen entry was so poor that his unsuitability as editor of
the OLD was plain to see.
40
The proposal for a database appears also in Essays 9 (delivered 1995) and 13 (delivered
2006).
18
WHY DIDN’T ST BASIL WRITE IN
NEW TESTAMENT GREEK?
20101
Abstract
St Basil’s Greek is noticeably different from that of the NT, though both
fall within the Koine period of ancient Greek. This paper sets out to explain the
reason. St Basil’s Greek and the Greek of the NT are briefly characterised, then
the discussion surveys the rise of Atticism, the ancient education system, and
St Basil’s education. The conclusion is reached that St Basil’s choice of the higher
language and style of the pagan Classics was the natural one for him, though
paradoxical in that he, as a Christian, rated the content of the biblical books as
superior.
PREAMBLE
Saint Basil the Great was a prolific author, and there is a large body
of his works surviving. There are over three hundred letters alone; other
works number at least a hundred. Besides the count of his works, the tally
of words in some of the longer works gives an idea of how prolific he was:
the Homiliae in Hexaemeron numbers 35,834 words, and the De Spiritu
Sancto 26,309, according to the TLG electronic count.2 All of this was
written in Greek, of course. But what sort of Greek? To say that it was Greek
as written in the fourth century AD provides only a partial answer. True
enough, that locates it within the Koine Greek period (300 BC to 600 AD)
of ancient Greek, but the Greek of that period is a complex phenomenon,
affected by currents and influences not obvious without investigation.
1
This paper was presented at the first Patristic Symposium of St Andrew’s Greek Orthodox
Theological College in 2009, dedicated to St Basil the Great. I am grateful to His
Eminence Archbishop †Stylianos and St Andrew’s College for the invitation to present
in the Symposium, and to Fr Doru Costache and Edwin Judge for valuable suggestions.
[The paper has been reprinted in Costache and Kariatlis (eds.), Cappadocian Legacy,
63–77.]
2
TLG Canon, 81–4.
ST BASIL’S GREEK
As my starting-point I want to quote just the first sentence of one of

St Basil’s most famous works, To Young Men, Πρὸς τοὺς Νέους:3
Πολλά με τὰ παρακαλοῦντά ἐστι ξυμβουλεῦσαι ὑμῖν, ὦ παῖδες, ἃ
βέλτιστα εἶναι κρίνω, καὶ ἃ ξυνοίσειν ὑμῖν ἑλομένοις πεπίστευκα.
Many are the things encouraging me to advise you, children, what I
judge to be best and what I am convinced will be of benefit to you when
you make your choice.
In this one short sentence, we already have striking indications of the

kind of Greek we are dealing with in St Basil’s works; and this is not an
isolated instance, but typical of his style everywhere. To begin with, the
opening phrase imitates the first words of a speech attributed to the famous
Athenian orator Demosthenes, who flourished in the fourth century BC.
His sentence begins Πολλά με τὰ παρακαλοῦντα ἦν, ὦ ἄνδρες Ἀθη-
ναῖοι, γράψασθαι Νέαιραν τὴν γραφὴν ταυτηνί (‘many were the things
encouraging me, men of Athens, to bring this suit against Neaira’).4 The
resemblance is unlikely to be accidental. St Basil knew the Classics (as
we shall see), and the sentence includes an idiomatic feature of Classi-
cal Greek, the placing of the enclitic pronoun (in this case με) as second
word.5
Other significant features in the sentence are: the choice of the Attic, or
rather old Attic, form ξυν- instead of συν- in two of the verb compounds;
the idiomatic Classical use of the perfect πεπίστευκα describing a state;
and the skilful use of the aorist participle ἑλομένοις attached to ὑμῖν,
meaning literally ‘to you (at the point of) choosing.’6 These are not natural
features of the everyday Greek of St Basil’s time but the result of con-
scious effort by the writer.
There is something else, a subtlety that one would hardly notice unless
alerted to it, the use of clausulae. This involves choosing or contriving a
certain pattern of syllables in the closing phrase of a sentence or clause:
3
In full Πρὸς τοὺς νέους, ὅπως ἂν ἐξ Ἑλληνικῶν ὠφέλοιντο λόγων, ‘To young men,
on how they might benefit from Greek literature.’ The exact title is in fact uncertain:
see Wilson, Saint Basil, 8–9. The TLG Latin title is De Legendis Gentilium Libris; others,
such as Rousseau (Basil of Caesarea, e.g., 49), use Ad adulescentes.
4
[D.,] Against Neaira 59.1; Wilson, Saint Basil, 37. Modern and ancient scholarship has
doubted the attribution to Demosthenes, but the speech dates from his time and was just
as much a model for later generations as the genuine speeches.
5
See Wilson, Saint Basil, 37. Stasinopoulos (Μορφές, 277) detects further Demosthenic
phrasing in ἃ βέλτιστα … and ἃ ξυνοίσειν … in St Basil’s opening sentence.
6
On πεπίστευκα see Wilson, Saint Basil, 37.
WHY DIDN’T ST BASIL WRITE IN NEW TESTAMENT GREEK? 281
the last two stressed syllables are separated by an even number of unac-
cented syllables, as here in ἑλομένοις πεπίστευκα. Another illustration,
this time with four syllables between the accented syllables, may be seen
further on in the same work, at 5.9: καὶ τελευτήσαντι παραμένουσα.7
The effect of this device is to give a pleasing balance and smoothness as
the sentence comes to a conclusion.
Numerous other linguistic features found generally in St Basil’s works
could be mentioned, all tending in the same direction, that is, towards
old-fashioned, Classical Attic, of the kind that was no longer spoken in
St Basil’s day. Examples ready to hand are forms in -ττ-, as in φυλάττω,
instead of -σσ-; the frequent use of obsolete particles such as μήν, γε, δή;
and recourse to the full range of optative forms and uses in the verb.8 These
are all connected with the phenomenon of ‘Atticism,’ on which more will
be said shortly.
Besides using individual features adopted from the Classical language,
St Basil refers frequently to the works of Classical authors, both by quoting
them and alluding to them. In To Young Men can be found quotations and
reminiscences from a list of authors including Homer, Hesiod, Theognis,
Plato, Aristotle, and Solon. Some are direct quotations, as when three lines
of Solon are quoted (5.9), or an expression is borrowed verbatim from Plato
(ὅτι μὴ πᾶσα ἀνάγκη, 9.2).9 Others are more subtle, as when a Homeric
phrase, ἔνθεν ἑλών (‘taking it up from this point’) is slipped into a sen-
tence (1.5).10 There is even an allusion (9.6) to the famous maxim on the
temple of Apollo at Delphi, γνῶθι σεαυτόν, but this is done without actu-
ally quoting the words.11 In 5.6–8 St Basil uses the episode of the meeting
of Odysseus and Nausicaa in Odyssey 6 to draw a lesson about virtue, intro-
ducing it without even mentioning the names: he is able to assume that his
educated hearers will be as familiar with the story as he is. In his works as
a whole, the same Classical authors and many others are cited or alluded to
with great frequency.12
7
The rule is set out in Maas, Greek Metre, 17; cf. Wilson, Saint Basil, 13. St Basil was
not alone in following this rule; his practice ‘corresponds closely to that of other well
known writers of the same date’ (Wilson).
8
As in the title of To Young Men: ὠφέλοιντο (see n. 3). Cf. the list of features in Fabricius,
‘Der sprachliche Klassizismus,’ esp. 192.
9
Wilson, Saint Basil, 64.
10
This was ‘a favourite phrase with Atticists’ (Wilson, Saint Basil, 40). The sentence as a
whole reads: τίνα οὖν ἐστι ταῦτα, καὶ ὅπως διακρινοῦμεν, τοῦτο δὴ καὶ διδάξω ἔνθεν
ἑλών.
11
See Wilson, Saint Basil, 65.
12
See the discussion and index of references in Deferrari, Basil: Letters 249–368, pp. 366–
8, 458–61.
St Basil’s writing style in general has attracted high compliments. Of

the Hexaemeron, Quasten expressed the view that ‘there is no work in late
Greek literature which could be compared with these homilies in rhetori-
cal beauty.’13 More tellingly, his friend and contemporary St Gregory of
Nazianzus regarded St Basil’s letter-writing skills as far superior to his
own.14 The playful exchange between St Basil and the renowned rhetori-
cian Libanius in Letters 338, 339 and 340 indicates as surely as anything
that St Basil was on an equal footing with Libanius in mastery of refined
writing. Each tries to outdo the other in εἰρωνεία, Libanius claiming that
on reading St Basil’s letter he declared he had been defeated in beauty
of letter-writing (ἐν κάλλει μέν, ἔφην, ἐπιστολῶν ἥττημαι), to which
St Basil responds by describing his letter as ‘that shabby letter of mine’
(τὴν ἐπιστολὴν ἐκείνην τὴν ῥυπῶσαν), and so on.15
We have then, from this brief sketch, an indication of the style and lan-
guage of St Basil, and the impression can be quickly confirmed by reading
any of his works. St Basil writes a high level, literary Atticistic Greek, densely
constructed with all the devices and subtleties of language and rhetoric avail-
able to the trained writer.
THE NEW TESTAMENT
Three centuries before St Basil, the NT authors also wrote in Greek,

within the same time-period of ancient Greek as St Basil, that is, the Koine
Greek period, mentioned before. But the contrast with the language and
style of St Basil is marked. The NT is written, in general, in ordinary Koine
Greek, and the style of writing, though varying from book to book, in no
case reaches the same high plane as St Basil’s. This is the general picture,
within which there is a considerable range; this range we must now briefly
explore.
To focus first on language. Simple, low-level Koine Greek is charac-
teristic of the Gospels of Mark and John; the Gospel of Matthew is simi-
lar but on a slightly higher level. The two books by St Luke, the Gospel
13
Quasten, Patrology, 3:216–7. In the ninth century St Photius offers a glowing tribute to
St Basil’s style and rhetorical skill in the Hexaemeron (Bibliotheca, 141).
14
See Gallay (ed.), Saint-Grégoire, 53. Cf. Bardy, ‘Basilius von Caesarea.’
15
On the question of the authenticity of the correspondence, see Rousseau, Basil of
Caesarea, 57 n. 130; Deferrari, Basil: Letters 249–368, xiii–xv. There seems to be
no doubt that some are genuine, among them the three mentioned. For further remarks
on St Basil’s rhetorical skills see Kinzig, ‘Greek Christian Writers,’ esp. 646, with bib-
liography.
of Luke and the Acts of the Apostles, show a distinct effort to write more
literary Greek than the other Gospel authors. Many features indicative of
‘better’ Greek can be noticed, as, for example, certain choices in vocabu-
lary, in which popular words are avoided in favour of more Attic ones.
This is noticeable when St Luke rewrites material derived from the Gos-
pel of Mark. St Luke can also write a showpiece of high level Greek in the
preface to his Gospel (1:1–4). But he does not keep it up. He quickly turns
to a different, ‘biblical’ sort of language (and style), which owes a great
deal to the Greek of the LXX. This is his model for a ‘classic’ style, not
the authors of Classical literature; and features of Atticistic Greek, though
present, are not maintained with the consistency seen in St Basil.
The other books of the NT, apart from Revelation, which is below even
Mark and John, are rather different and show middle-range educated Greek.
The Epistles of St Paul are a definite step above Mark and John. This is
seen in a variety of features – forms, syntactical structures, choice of vocabu-
lary – that contrast, sometimes sharply, with the simpler Greek of Mark and
John. At the top among the books of the NT, in terms of language and
style, is the Epistle to the Hebrews, long recognised as the work of a highly-
educated writer.
We are already touching on style, which is a somewhat different matter
from level of language, though linked. The NT writers have their indi-
vidual styles. In trying to characterise these different styles, an important
question is the influence of rhetorical techniques. A good deal of modern
study has turned attention to this topic, which has much relevance to the
comparison with St Basil, since St Basil’s works are the product of advanced
training in rhetoric. Among the NT books, the Gospels and Acts reveal very
little evidence of rhetorical techniques, and in the case of Mark none at
all.16 The case is different with the Pauline Epistles. The question of St Paul’s
training in rhetoric and the extent of his use of rhetoric in his letters has
perplexed ancients and moderns alike. He appears not to have been formally
trained in rhetoric, but it is clear that he was not ignorant of the art; he
claims to be a ‘layman in speech’ (ἰδιώτης τῷ λόγῳ, 2 Cor 11:6), but this
cannot be taken at face value; later exponents saw that he was no expert
by Classical standards, yet the power of his eloquence was unmistakable;
his writings can be analysed to a considerable extent according to formal
rhetorical categories, but the results have been diverse and inconclusive.17
16
For a thorough exploration see Burridge, ‘The Gospels and Acts.’
17
On Paul see Judge, ‘Paul’s Boasting’; Porter, ‘Paul of Tarsus.’ Judge’s paper, delivered
as long ago as 1967, was full of insights and set out the parameters of the debate. Porter
covers all recent discussion without reaching any firmer conclusions.
The situation is clear, however, with regard to the Epistle to the Hebrews:
there is general agreement that it reveals a high level of rhetorical training
and competence.18
As regards reference to Classical authors, the strong characteristic of
St Basil’s writing that we noted above, the NT authors have little to offer,
naturally enough. It is true that Paul quotes Menander (1 Cor 15:33 φθείρου-
σιν ἤθη χρηστὰ ὁμιλίαι κακαί, ‘bad company corrupts good morals’),
but even this was a popular epigram and unlikely to derive from direct
acquaintance with Menander.19 Paul’s speech in Athens, as recorded by Luke
(Acts 17:22–31), contains, appropriately in that learned setting, a quotation
from Aratus (Acts 17:28 τοῦ γὰρ καὶ γένος ἐσμέν, ‘for we too are his
offspring’), but there is little more of that nature in the NT. What the authors
do quote extensively is the Christians’ own ‘classical’ text, the OT in Greek,
the LXX.
In sum, the books of the NT show a range of language, from higher
to lower, but the majority remain ordinary Koine Greek, and are not in
the same league as St Basil and most of the Fathers. Those who looked
back to them from the high point of the fourth century saw writing that
struck them as inelegant and uneducated – ‘als stilistische Monstra,’ as
Norden put it – and from their point of view it was true.20 St Basil him-
self speaks of the material communicated by the biblical authors as ‘in
sense true, but in style unlearned’ (νοῦν μὲν ἀληθῆ, λέξιν δὲ ἀμαθῆ).21
Several other Fathers observe and discuss this problem, sometimes at
length.22
One feature alone is sufficient to sum up the gulf between the Hoch-
sprache of St Basil and the language of the NT. The form ξύν, which we
observed in St Basil’s sentence at the beginning of this paper, is not used
by a single NT writer.23
18
See Thurén, ‘New Testament Writings,’ esp. 589–92. There is a brief but useful
survey of NT rhetoric in BDF, §§485–96 (‘The Arrangement of Words: Figures of
Speech’).
19
Fee, First Epistle to the Corinthians, 773 n. 61.
20
Norden, Die antike Kunstprosa, 2:517; see further 516–25; Kinzig, ‘Greek Christian
Writers,’ 634–6.
21
Letter 339. St Basil means this playfully (cf. Norden, Die antike Kunstprosa, 2:529,
n. 1), so that he can pretend that his own style has been infected by theirs and it
shows.
22
The passages are set out in full in Triantaphyllides, Γραμματική 1: Ιστορική Εισαγωγή,
418–22. Cf. Kinzig, ‘Greek Christian Writers,’ 635. While noting the poor quality of
NT writing, the Fathers offer good defences and explanations.
23
See BDF, §34.4.
KOINE GREEK AND ATTICISM
Why is St Basil’s Greek so different from what we read in the NT? The
short answer is: St Basil was an educated man. But what did that mean in
the fourth century AD? This is what I want to explore next. To understand
what it meant we need to go back, well beyond St Basil’s time, all the way
back to Classical Athens. As is well known, there was a great flowering
of Greek culture in Athens of the Classical age. Art, science, drama, his-
toriography, philosophy, oratory – everything – reached a pinnacle in that
one small πόλις in the fifth and fourth centuries BC. When one looked
back from a later time in the ancient world, the achievements seemed
unsurpassed, and indeed unsurpassable. Even today it seems so in many
respects.
The language of the texts of that period was the dialect of Athens, that
is, primarily Attic. Other dialects and styles were on the scene, especially
poetic genres; for example, Homer, the foundation text for all study of
literature, was in Ionic dialect combined with other elements. But Attic
was the primary vehicle of the literature of Classical Athens. When Alex-
ander’s conquests carried Greek and Greek culture all over the Eastern
Mediterranean at the end of the fourth century BC, a new phase of the
language had already begun to emerge. This was Koine Greek, based on
Attic, but with numerous differences that were characteristic of the vernacu-
lar. Koine Greek became the spoken language wherever Greek was used,
all over the area of Alexander’s empire. At this time, works of literature
could be written in a literary Koine, without adhering to strict Attic or
attempting to follow Classical models.
But as time went on, a new outlook arose.24 This was the tendency to
view the early post-Classical period as one of decline in Greek language
and culture, and to look back to the golden age for inspiration and for
models to imitate, not only in literature but in everything else – art, archi-
tecture, and so on. It became a matter of importance not only for Greeks
but for Romans, the new masters of the ancient world, who looked to the
Greeks for education and culture. This tendency emerged in the first cen-
tury BC and led to the phenomenon of ‘Atticism’ in language, that is, the
attempt to revive Attic of the fourth century BC as the instrument of good
writing. The great classics of the literature of Athens in the fourth and fifth
centuries BC became the models, and Greek with Attic features came to
24
On the subject of this paragraph, see Browning, Greek, 44–50; Kazazis, ‘Atticism,’ 1200–
17, and esp. 1200-6; Lee, ‘Atticist Grammarians,’ 283–5, with further references.
be taught and learnt in the schools, as opposed to Koine Greek, the spoken
language. The two of course are not completely separate, but overlap. What
Atticism involved, however, was trying to revive and maintain Attic usage
wherever the spoken language had moved on and changed. It was essen-
tially a purifying process, and it must be admitted, artificial. To take a sim-
ple example: a popular word for ‘bed,’ κράβαττος, appeared in early Koine
Greek, but Atticists recommended avoiding it because it had no Classical
precedent. In the NT it is found several times, but Matthew and Luke, influ-
enced by the puristic tendency that they had been introduced to in school,
remove Mark’s examples or replace them (with κλίνη or κλινίδιον).25
This is only a brief glimpse of the history of the period, but it gives us
the background we need to appreciate the character of ancient education,
our next topic.
AN ANCIENT EDUCATION
Let us turn now to the education system in antiquity, especially in the

Greco-Roman world. What was it like? It is easy for moderns to be una-
ware that there was any such thing, still less that it was highly developed
and systematic. This is partly because we have never had reason to give
the matter any thought; but it is also because everything before a cou-
ple of centuries ago is often imagined as primitive, enveloped in a sort of
dark age. This of course is entirely mistaken. Education already existed
in the Classical period and developed further in the Greco-Roman period;
by the time we reach St Basil it had been in place for centuries, and was
an organised and confident system, sure of its goals and methods and its
role in training elites. (It was, it may be added, not a public, free edu-
cation system such as we know today, but was available only to those
who could afford to pay for it; even so, most people learnt at least the
rudiments.)
The full program involved three stages or levels.26 Stage one, naturally,
aimed at teaching the pupil to read and write. To learn the rudiments,
recitation, not just of the alphabet, but of syllables combining all the con-
sonants and vowels, was a staple technique. Fragments of Classical authors,
25
Cf. the anecdote in Browning, Greek, 47 n. 34. [Bp Triphyllios of Ledra is said to have
been rebuked by St Spyridon for replacing κράβαττος with σκίμπους in a Gospel reading
(Mark 2:11 / John 5:8). For σκίμπους see Phryn. etc. in Essay 19, §8.]
26
See Cribiore, Gymnastics of the Mind, 2–3, 160–244; Whitmarsh, Second Sophistic, 13–
5. For a full study see Morgan, Literate Education.
especially Homer, were introduced even at this stage; they were copied out
as exercises and learnt by heart. So a start was made on Classical Greek
literature here.
At stage two the curriculum covered γραμματική (τέχνη), loosely
equivalent to our ‘grammar,’ though embracing more than what we mean
by that. It involved analysis of texts, with close study of details of language
– vocabulary, grammar, etymology, and so on. Letter-writing, regarded
as an important skill, was introduced and practised here. There was fur-
ther study of the poets and the prose authors, the ‘Classics,’ as models of
expression and sources of themes. Most students would not advance even
to this stage, but there was one more beyond it.
Stage three was the equivalent of our tertiary level today, but based on
different premises. The training offered here brought the student to the
pinnacle.27 To get into stage three there was a sort of entrance exam, as
Libanius tells us.28 As well as learning to write in the best style, the student
learnt the art of rhetorical performance, which was just as important. The
aim was to train the student in Classical models so well that he could offer
a complete oration (μελέτη) in a similar style. This skill was regarded as
‘the crowning achievement of the student of rhetoric.’29 To get to that point,
the student advanced through a series of increasingly demanding and com-
plex exercises, called προγυμνάσματα. In these he was expected to read,
copy, and memorise large portions of text, and to immerse himself in the
works of orators, poets, and historians of the Classical age. Finally, after
years of intense practice, he would be able to create his own display, weav-
ing together elements he had learnt and adding his own. The full course
lasted five or six years, though not necessarily all stayed the distance.
Many would not have felt the need to: it is said that after two years the
student could have acquired enough skill to plead in court and take up
legal advocacy as a profession.30
We know that there were public displays of oratory before an audience,
sometimes before the Roman Emperor himself. The top performers, the
‘Sophists,’ as they were known then, attracted large audiences. We have
descriptions of these performances.31 They were clearly highly-charged
occasions, demanding the utmost of the speaker’s skill and training. The
speech was often delivered on a theme proposed by the audience on the
27
Whitmarsh, Second Sophistic, esp. ch. 2; Cribiore, Gymnastics of the Mind, 220–44.
28
Cribiore, Gymnastics of the Mind, 224.
29
Cribiore, Gymnastics of the Mind, 223; cf. Whitmarsh, Second Sophistic, 20–1.
30
Cribiore, Gymnastics of the Mind, 224.
31
E.g., in Philostratus, Lives of the Sophists: see Whitmarsh, Second Sophistic, 27–9.
day.32 Thus the speaker at this level had not only to compose, on the
spot, an oration displaying his familiarity with Classical texts and themes,
but also to be careful not to make any slips in purity of Attic in his own
speech. The audience would be keeping a close watch on every word, and
failure to impress was a real possibility. But success brought renown. It
was towards this ultimate prize that the third stage of education led, for
those who could reach it.
ST BASIL’S EDUCATION
What we have described is indeed the course of training that St Basil

had been through. A good deal is known about his education and career.
He was born in Caesarea in Cappadocia, and received his elementary
training from his own father, himself a distinguished rhetorician. He then
attended schools in Caesarea, Constantinople, and finally the great ‘Uni-
versity’ of Athens. It was not possible to go higher than that. Athens the
city was still at this time well preserved and the centre of the ancient inher-
itance of literature, philosophy, and culture. It was, as Deferrari puts it,
‘the pattern of excellence to a world that elected to see in Atticism the
cultural ideal.’33
It seems certain St Basil was acquainted with the greatest pagan rhet-
orician of the fourth century, Libanius, whom we have mentioned before,
though whether he actually studied under him (in Constantinople) is less
certain.34 We do know that in Athens one of his teachers was Prohaeresius
(276–366 AD), who held the chair of rhetoric and was another celebrated
rhetorician.35 All this is important, because the school you attended and the
quality of your teachers mattered – just as it does today. Finally, St Basil
himself came to hold a ‘chair of rhetoric,’ that is, a chair in pagan litera-
ture and learning, in his home town of Caesarea.
It was only after all this rigorous training, and the career culminating in
the achievement of a chair of rhetoric, that St Basil turned to the Church
and was ordained. Then began a different life, his celebrated career in the
32
It was rather like a great organist improvising on a theme given to him or her just before
the performance.
33
Deferrari, Basil: Letters 1–58, xviii.
34
Wilson, Saint Basil, 52; Basil: Letters 1–58, xvii–xviii. A large quantity of Libanius’s
writings survives: 64 orations, 51 declamations, 1544 letters, a Progymnasmata, plus
lesser works (TLG Canon, 245).
35
Joyal, et al., Greek and Roman Education, 257–8. Stasinopoulos (Μορφές, 99) says he
was also a Christian.
Church, during which he established his famous monastic Rule, became

Archbishop of Caesarea, and so on. In doing that, he left behind the world
of pagan learning – but he did not forget what he had learnt.
A fitting conclusion to this section is the statement that can be found
in the Menaion on St Basil’s feastday (1 January), in the short biography
recorded there:
Ἐν δὲ λόγοις, οὐ μόνον τοὺς καθ᾽ αὑτόν, ἀλλὰ καὶ τοὺς παλαιοὺς
ὑπερέβαλε· διὰ γὰρ πάσης ἐλθὼν ἰδέας παιδεύσεως, ἐν ἑκάστῃ τὸ
κράτος ἐκτήσατο.
In Classical learning he surpassed not only his contemporaries but even
the ancients; for passing through every kind of training, he acquired
mastery in each.36
This sentence sums up the first phase of his life succinctly and perfectly,
and ‘every kind of training’ fits exactly the educational curriculum of the
time. The author of this summary knew his subject – and also how to
write the sort of Greek St Basil himself would have been comfortable
with.
ST BASIL’S CHOICE
We can see now ‘where St Basil was coming from’ – to use an expres-
sion of today – and we have come to the answer to our question: St Basil
writes as he does because that was the way an educated man naturally
wrote in his time. The choice was as it were predetermined. That was
especially the case for one who had been educated right up to the top level.
He used, to the full extent of his ability, the Atticising Greek and style that
he had learnt and that was the status form of his time. It was the style
expected of him by his contemporaries, and it would have been difficult
for him to abandon it. This does not mean he could not speak everyday
Greek: Attic Greek was learnt in the schoolroom, not the nursery. It also
does not mean that he was incapable of preaching and writing in everyday
Greek if he chose.
36
Μηναῖον τοῦ Ἰανουαρίου (Ἐν Ἀθήναις· Ἀποστολικὴ Διακονία, 1991), 24. The meaning
of λόγοι here can be demonstrated from other examples, some in St Basil himself: see
Deferrari, Basil: Letters 337, 338; To Young Men, title (see n. 3); 4.8 (quoted below);
5.1; see also Lampe, PGL, s.v. λόγος I. A.13. It is clear that (οἱ) λόγοι (pl.) refers to
learning in Classical pagan literature, sometimes specifically to the literature itself, as
in 4.8, 5.1. It is easy, as Rousseau does (Basil of Caesarea, 41, 49), to miss the full
significance of the term and translate simply as ‘words’ or ‘rhetoric.’ In To Young Men,
10.4 St Basil refers by way of contrast to οἱ ἡμέτεροι λόγοι, ‘our (Christian) literature,’
and to ἱεροὶ λόγοι (2.5, quoted below).
The authenticity of certain of St Basil’s works has been called into

question on this very issue. The two homilies On the Origin of Humanity 1
and 2 (Περὶ τῆς τοῦ ἀνθρώπου κατασκευῆς) are written in a style that
differs noticeably from St Basil’s other writing, and specifically from the
other nine homilies of the Hexaemeron, with which these two seem to
belong.37 The style ‘lacks the refinement and clear, balanced syntax of his
usual rhetorical prose’ (Harrison). Some take this as an indication that they
were not St Basil’s work; others that he did compose them but did not
have time to polish them for publication. While it is not intended to resolve
the issue here, it is worth noticing the possibility, as Fr Doru Costache
suggests to me, that St Basil intentionally composed them in that style for
delivery to a less educated audience; the question of polishing was sepa-
rate, and he may not have had the time or inclination to attend to it. We
may thus have a glimpse in these works of the simpler Greek of which
St Basil was capable in the appropriate setting.38
So St Basil writes in the language and style of the great pagan Classics
and rejects the simple Koine Greek of the NT as the vehicle for his pub-
lished teaching and thoughts. But there is a paradox at the heart of this
choice, one already hinted at in St Basil’s remark on the biblical texts
quoted above (‘in sense true, but in style unlearned’). The basis of faith
and moral guidance for St Basil, as for all Christians, is the NT and the
Scriptures (which means the OT). The pagan Classics are not without
value in this respect, as the work To Young Men is written to demonstrate,
but only where they have good to offer: a higher truth governs what may
be selected from them as valuable.39 As St Basil himself puts it (2.5):
‘To this [other, that is, eternal, life] the Holy Scriptures lead the way, teach-
ing us through secrets’ (εἰς δὴ τοῦτον ἄγουσι μὲν ἱεροὶ λόγοι, δι᾽ ἀπορ-
ρήτων ἡμᾶς ἐκπαιδεύοντες). Yet despite this, St Basil’s chosen linguistic
vehicle for his writing is not biblical Greek but the elevated language of
the pagan Classics. He cannot envisage a writing style based on the NT
(still less on the LXX), or an education system having only the Christian
texts as its models.40
37
Harrison, Saint Basil the Great, 14–5; Rousseau, Basil of Caesarea, 318 n. 1. I am grate-
ful to Fr Doru Costache for drawing this issue to my attention and for assistance with
bibliography.
38
While the style is certainly simpler, it seems to me the language remains Atticising and
rather demanding for the uneducated listener. If the story in Browning, Greek, 50 can be
trusted, St John Chrysostom once demonstrated his capacity to preach in everyday Greek
when called upon to do so.
39
See To Young Men 1.5, 4.6–7; Quasten, Patrology, 3:214; Jaeger, Early Christianity, 81.
40
On this last point cf. Wilson, Saint Basil, 9; Jaeger, Early Christianity, 83–4; Louth in
Young, et al., Cambridge History, 294: St Basil ‘clearly envisages no alternative way
CONCLUSION
I end with a paragraph from the work with which we began. In this pas-
sage we see the paradox played out, as it were, as St Basil makes his point
about selective use of Classical literature while at the same time demon-
strating his mastery of both the content of that literature and its linguistic
form.
Κατὰ πᾶσαν δὴ οὖν τῶν μελιττῶν τὴν εἰκόνα τῶν λόγων ἡμῖν
μεθεκτέον. ἐκεῖναί τε γὰρ οὔτε ἅπασι τοῖς ἄνθεσι παραπλησίως
ἐπέρχονται, οὔτε μὴν οἷς ἐπιπτῶσιν ὅλα φέρειν ἐπιχειροῦσιν,
ἀλλ᾽ ὅσον αὐτῶν ἐπιτήδειον πρὸς τὴν ἐργασίαν λαβοῦσαι, τὸ λοι-
πὸν χαίρειν ἀφῆκαν· ἡμεῖς τε, ἢν σωφρονῶμεν, ὅσον οἰκεῖον ἡμῖν
καὶ συγγενὲς τῇ ἀληθείᾳ παρ᾽ αὐτῶν κομισάμενοι, ὑπερβησόμεθα
τὸ λειπόμενον.41
We should, then, partake of pagan literature (λόγοι) exactly in the man-
ner of the bees: for they do not approach all the flowers equally, nor do
they try to carry off the whole of those on which they land, but taking
as much of them as is suitable for their work, they bid the rest goodbye;
so we, if we are wise, receiving from them (i.e., pagan literature) as
much as is suitable (οἰκεῖον) to us and related (συγγενές) to the truth,
will pass over the remainder.
The simile of the bee is itself not new but a commonplace going back
to Isocrates, and the passage ‘bears a close resemblance’ to one in Plu-
tarch.42 In vocabulary we see at one point a choice, surely intentional, of
words with suggestive dual meanings: οἰκεῖον can mean ‘suitable, proper’
but also ‘of one’s own family, household’; similarly συγγενές has both a
figurative sense (‘related’), as well as a literal one (‘of the same kin, fam-
ily’). Thus is implied that the truth by which the pagan Classics are to be
measured is the truth found within the Christian family. The passage is
replete with the expected linguistic features marking it as high style based
on Attic models: the Attic -ττ- in μελιττῶν; a ‘gnomic aorist’ in ἀφῆκαν;
controlled use of particles, δὴ οὖν, τε, γάρ, μήν; Ionic/Old Attic ἤν
(= ἐάν); and the verbal adjective μεθεκτέον, a rarity in everyday language.
Finally, the sentences exhibit a flow and balance (every sentence-end and
of learning Greek than using pagan literature.’ This topic is connected with the issue
of tension between ‘two sets of standards’ in St Basil’s thought over time, explored at
length by Rousseau (Basil of Caesarea, 40–57). Rousseau goes so far as to conclude
that ‘Basil consistently rejected, probably from the earliest period following his departure
from Athens, the claims of the traditional curriculum’ (56).
41
To Young Men 4.8. Text from Wilson, Saint Basil. My translation, based on Deferrari
and McGuire, Basil: To Young Men, pp. 391–3.
42
Wilson, Saint Basil, 48; Deferrari and McGuire, Basil: To Young Men, 391 n. 2; Stasi-
nopoulos, Μορφές, 280.
most clause-ends obey the rule of clausulae), that achieves a pleasing effect
without showing quite how it is done.
Thus St Basil puts on display his omnivorous training in Classical lite-
rature at the same time as he cautions his young readers not to partake of
all of its flowers.
19
THE ATTICIST GRAMMARIANS
2013
Abstract
The first aim of this essay is to introduce the Atticist grammarians and explain
their purpose; how they present their material; and how they can be used as
evidence for the history of Greek. The point is made that their assertions are not
statements of fact but require interpretation. The Atticists (or works) covered are:
Phrynichus, Moeris, the Antiatticista, Aelius Dionysius, the Philetaerus, Thomas
Magister, Ammonius, and the Onomasticon of Pollux. The discussion then turns
to the relevance of the Atticist grammarians to the NT. The grammarians’ evi-
dence helps to prove that the language of the NT was by and large the Koine
Greek of their day. There is, however, some influence of Atticism on the NT,
mainly in Luke–Acts and the Epistles. Specimens are presented. Finally, the phe-
nomenon of Atticising variants in the MS tradition and the questions raised for
NT textual criticism are briefly outlined.
1. ATTICISM
Atticism was a complex phenomenon that has been described and

explained in numerous ways, but its essential element was the tendency to
look back to the language and literature of a former era as the model to
follow, from a later time when the spoken language had changed and origi-
nal composition of that literature was in the past. The former era was the
Classical period, and access to it was through its texts, which were studied
and imitated in the education system. The later time was almost the whole
of the post-Classical period: the phenomenon appeared in the first cen-
tury BC, reached a peak in the second century AD, and continued its influ-
ence through the Byzantine era, with effects still today. The results were
seen not only in writing but in the shape of the language as a whole, that
is, in a differentiation of the spoken and written varieties, or a ‘diglossy/
diglossia,’ that affected Greek for the rest of its history.1
1
For an accurate definition of the term (too long to repeat here), see Ferguson’s in Hudson,
Sociolinguistics, 54. I am grateful to Michael Curran for reading this paper and offering
many useful comments.
Atticism was not simply a linguistic phenomenon but part of a larger

enterprise to recover the Classical past. The cultural world of fifth and
fourth century BC Athens was to be recreated, in literature, rhetoric, the
arts, and philosophy. Powerful cultural and social forces fostered the enter-
prise, and even emperors supported it. The whole was an exercise in imi-
tation, but it had a deep and enduring – some would say disastrous –
impact.2
The attempt to turn the clock back created complexity for users of the
language, especially when they wished to commit anything to writing, or,
at a more advanced level, engage in rhetorical display. Anyone who had
been to school had been introduced to the notion that some forms, words,
and uses were ‘better’ or more ‘correct’ than others. They were the ones
that matched the Classical models, contrasting with new features that had
arisen as a result of change in the spoken language. The contemporary spo-
ken language was disparaged, the older, obsolete features were approved
and accorded prestige. The need for guidance in finding one’s way through
these artificial shoals soon led to the creation of works offering instruc-
tion in how to do it. The authors of these works were the Atticist gram-
marians.3
There is a long list of names of Atticist grammarians, though most of
their works survive today only in fragments. The effects of their efforts,
however, and of the whole movement to treat the Greek of the past as the
model of good Greek, are to be seen in almost all the written remains of
post-Classical Greek. Most works of literature from that period exhibit
artificial Atticising features to varying degrees, and even the lower levels
of everyday writing display some influence.4
Although the Atticising movement had marked success, and writers at
the top of the scale could write a whole work that appeared to reproduce
the Attic of fifth-century Athens, this Atticising Greek was not a spoken
language or dialect separate from Koine Greek; it was a stylistic variety,
or rather group of varieties, added on to the living language, Koine Greek.
2
The literature is extensive. Recommended are: Anderson, Second Sophistic, esp. 86–
100; Bowie, ‘Greeks and Their Past’; Browning, Greek, 44–50, 104–13; Horrocks,
Greek, 78–86, 151–3; Kazazis, ‘Atticism’; Swain, Hellenism and Empire, esp. 17–41;
Whitmarsh, Second Sophistic; Cribiore, Gymnastics of the Mind, esp. ch. 8. It is worth
noting that ‘Asianism,’ against which Atticism was supposedly a reaction, is largely a
fiction: see Whitmarsh, Second Sophistic, 7–8, 50–1.
3
The epithet ‘grammarian’ is not quite appropriate, but it is hard to find a better alternative.
4
This last topic has hardly been touched. There are a limited observations in Mandilaras,
Verb (mostly on the optative); Schmid, Atticismus, 3:19–33 passim; Kotsevalov, ‘Koine
Syntax.’ The only extensive study known to me is an unpublished work: Connolly,
‘Atticism in Non-literary Papyri’ (see bibliography).
THE ATTICIST GRAMMARIANS 295
The latter was the mainstream that went on being spoken and evolving
over the course of nine centuries, from 300 BC to 600 AD, before passing
into the next phase termed Byzantine Greek. Atticism is to be viewed
as one constituent of the wide range of variation of which Koine Greek
was capable.5
2. THE ATTICIST GRAMMARIANS
This essay focuses on a selection of three Atticist grammarians whose

works survive in substantial remnants, Phrynichus, Moeris, and the Anti-
atticista. The aim is to introduce them, with specimens of their work, and
to show how they can be used as evidence for the progress of the lan-
guage and for the linguistic character of a text such as the NT. We begin
with some general observations about them.
The aim of the Atticist grammarians was to identify what was accept-
able ‘Attic.’ Such an aim always had a corollary, that something else was
not acceptable, whether this was stated openly or merely implied. One
feature was approved, another condemned. Since their purpose was to
revive features that were no longer alive and did not come naturally to
speakers of the language, it follows that the condemned feature was in
fact the live one used in the language of the day; the recommended one
was the feature that was obsolete. Evidence from the texts of the time,
when brought to bear, confirms this. We can, then, as a general principle,
take the Atticists’ statements as good information about what was actually
spoken Koine Greek, though this was not their intention.6 Their opinions
about what was good Attic are another story. For various reasons they vary
in reliability: they always need to be checked against the evidence of the
texts. The features favoured by the Atticists are what they thought was
good Attic of the fifth century BC, rather than what actually was.
Furthermore, the Atticists had their own ideas about which authors
provided models of good Attic. Not all the authors of the Classical period
were accepted by any means. Controversies developed between Atticists
over the choice of authors. The strictest Atticists worked with a limited
canon of approved writers, regarding the rest as below standard and not
to be imitated. They even went so far as to correct or rebuke an ancient
5
Cf. Kazazis, ‘Atticism,’ 1208–9; Horrocks, Greek, 81; Frösén, Prolegomena, 99–100,
177–8, 229.
6
In Hatzidakis, Einleitung, 285–303 numerous features condemned by the Atticists are
shown to descend into the modern language.
author for lapsing into a use that they did not accept. They also mixed
the evidence of prose and poetry, and of authors of widely varying dates
and styles, so that their picture of Classical Attic prose usage was, over-
all, a distorted amalgam.7
Many of the features that they focus on seem trivial in the extreme, but
this is not surprising in a movement requiring great attention to detail if
error was to be avoided; everything written and spoken could be subjected
to pedantic scrutiny by others. On the other hand, we find no remark in the
Atticists on many of the changes between Classical and Koine Greek. This
may be partly because the remains of their works are not complete; but
some features, especially broader trends, simply slipped under their radar.8
A key point about use of the Atticists’ material is that their assertions
are not to be taken at face value as objective statements of fact; they
require interpretation. They were made with a purpose, and the purpose
was prescriptive rather than descriptive. Interpretation needs to take into
account two things: the purpose of the Atticist grammarian and the evi-
dence of usage. The discussion that follows is intended to illustrate these
in practice. Since the NT is an important specimen of Koine Greek, the
illustrations often involve it in some way. The history of the features over
the course of time, from Classical to Koine Greek and beyond, is always
part of the picture also.
3. A FIRST SPECIMEN
Let us consider an example now, as a general illustration of what has

been said so far. In one of the statements by the Atticist Phrynichus (of
whom more shortly), the concern is over so small a matter as the form
and gender of the word for ‘flea.’ Phrynichus says:
ψύλλος βάρβαρον, ἡ δὲ ψύλλα δόκιμον ὅτι καὶ ἀρχαῖον.
[The form] ψύλλος (masc.) is barbarous; but ψύλλα (fem.) is approved
because it is also ancient. (Phryn. 306)
The assertion that the masculine form is ‘barbarous’ is of course not a

statement of fact, but an opinion, a derogatory description of something
7
Cf. Anderson, Second Sophistic, 88; Kazazis, ‘Atticism,’ 1206. To be fair to Phrynichus,
he was well aware of the ancient dialects and that Herodotus, for instance, has Ionic
features that are not Attic: see, e.g., Phryn. 235.
8
For example, the long-term change from a three-voice to a two-voice system in the
verb, with absorption of the middle functions into the active; see, e.g., Browning,
Greek, 30.
that aroused the Atticist’s distaste. The approved form is the feminine,
which is described as ‘ancient,’ a not very exact term, but evidently refer-
ring to Attic Greek of the Classical period. The Atticist instructs us to avoid
the masculine form and use the feminine. The masculine must of course
have been in use in his time, otherwise there would have been no reason
to condemn it. This much we can deduce from the statement itself. We
also have a parallel statement in Moeris (ψ 2), saying the same thing more
baldly: ψύλλα θηλυκῶς Ἀττικῶς. ἀρσενικῶς Ἕλληνες.
The evidence of actual usage in surviving texts, incomplete though of
course it is, fills out the picture and gives us food for thought. It is true
that ψύλλα is the original formation and attested early in Aristophanes
and Xenophon, then in Aristotle.9 But according to the Suda, ψύλλος was
used by Epicharmus, a Sicilian comic poet older than Aristophanes.10
There is also an example in Aristotle (HA 537a.6, referring to a sea-flea).
The next appearance of ψύλλος is in the LXX, at 1 Kgdms 24:15, a use-
ful piece of evidence; further, we have evidence of Ψύλλος as a name
in Herodotus and Menander and in documents from the third century BC
onwards.11 The form approved by the Atticists, ψύλλα, is common in liter-
ary sources from the second century AD onwards.12 Finally the modern
language comes in with unequivocal evidence of ψύλλος as the normal
spoken form.13
What does all this tell us? We can confidently say that the original form
ψύλλα had competition as early as the Classical period from the masculine
ψύλλος, which may have been in Attic already in the time of Aristotle;
it was well established by the early Koine (cf. the LXX) and continued to
be used in the spoken language from then until modern times; the Atticists
revived ψύλλα and caused it to be used artificially by educated writers
of their time; it then persisted in puristic writing for centuries. The actual
attestation of the words in this case is meagre, but we can do a lot with
it. What we are doing is reconstructing the full history of the forms from
whatever clues and evidence we have. The statements of the Atticists are,
or may be, useful clues to the general direction, even if they are not fully
accurate as to the facts.
9
See Chantraine, DELG, s.v.; LSJ, s.v.
10
Suda ψ 152 (Adler); ‘καὶ ἕτεροι,’ says the Suda.
11
Hdt. 4.173; Men. 272 (Körte); Fraser, LGPN, IIIA, s.v. The fem. Ψύλλα is known from
Lycophron (IV BC) onwards.
12
A TLG search produces over 200 occurrences, from Lucian and Galen in II AD till well
into the Byzantine period.
13
See, e.g., Lex. Proia, s.v. The fem. (no surprise!) is retained in puristic Greek.
4. PHRYNICHUS
Phrynichus was a famous strict Atticist who lived in the second cen-
tury AD. His major work, the Ecloga, can be dated to 178 AD. He was
born, not in some part of old Greece such as Athens, as one might have
thought, but in Bithynia, in NW Asia Minor. He is said to have been a
competitor for the chair of rhetoric at Athens, against his successful rival
Pollux. His birthplace and career are a good reflection of the cultural land-
scape of the Greco-Roman world.14
The standard up-to-date edition of the Ecloga is that of Fischer,15 but
the 1820 edition of the fine scholar Lobeck, which also incorporates
his predecessors’ observations, retains much value.16 The later edition of
Rutherford is still useful at times, though he adopts the role of a sort of
latter-day Atticist.17
Another work of Phrynichus, called the Praeparatio sophistica (Σοφι-
στικὴ Προπαρασκευή), survives in a sizeable epitome. The current edition
is that of de Borries;18 the text can also be found in Bekker’s Anecdota.19
Its purpose is much the same as that of the Ecloga, but the material more
specialised. Often the aim is to suggest or explain a useful Attic expression,
or give the tick of approval to some feature, without necessarily naming
any equivalent to be avoided.
The Ecloga is primarily a list for the budding Atticist of what to avoid
and what to use instead, mostly with the former first, but sometimes the
reverse. Most of the entries are quite short, consisting of a line or two,
though a few run to five or more. The material is unordered, with only
traces of alphabetical order. As to length, there are 424 lemmata in Fis-
cher’s edition. The simplest type of entry is:
δυσὶ μὴ λέγε, ἀλλὰ δυοῖν.
Do not say δυσί [dat. plural] but δυοῖν [dat. dual]. (Phryn. 180)
Here we see Phrynichus’s typical peremptory style and the extreme

nature of the Atticist program. He tries to revive the use of the dual, which
14
On Phrynichus see Strout and French, ‘Phrynichos’; OCD, s.v.; Dickey, Scholarship,
96–7. Photius gives Phrynichus the epithet ‘Arabius,’ the import of which is unclear.
15
Fischer, Die Ekloge des Phrynichos. Fischer’s introduction (37–50) deals fully with the
questions of date, sources, and whether the Ecloga is abridged (no, says F., 37).
16
Lobeck, Phrynichi Eclogae.
17
Rutherford, The New Phrynichus. For his attitude see, e.g., ix–x (Phrynichus has ‘schol-
arly nerve and wholesome masculine common sense’), 115, 144, 203.
18
De Borries, Phrynichi Sophistae Praeparatio Sophistica.
19
Bekker, Anecdota Graeca, 1:1–74.
even in the word for ‘two’ had begun to be replaced before the Koine
period, and was obsolescent throughout the noun and verb morphology of
Attic by the fourth century BC. The Atticists’ Attic was the most con-
servative variety, more conservative than Attic itself by the end of the
Classical period.20
Like this one, many entries deal with morphology; just as many con-
demn a word entirely, for the simple reason that it cannot be found in a
Classical author. Even the occurrence of a related form is not enough. In
the following example Phrynichus concedes that the Attic writers do use
the verb, but the noun is not found and is therefore not acceptable:
προκόπτειν λέγουσιν, τὸ δὲ ὄνομα προκοπὴ οὐκ ἔστιν παρ᾽ αὐτοῖς.
They [the ancient authors] say προκόπτειν (‘to make progress’), but
the noun προκοπή (‘progress’) is not found in them. (Phryn. 58; cf.
Philet. 84)
Particular uses of words are also freely condemned for the same reason,
though a less rigid observer might have seen that the extension is natural
and could have occurred at any time. So in the next example, in which a
nicety of usage presents an inviting target for the pedant:
τέμαχος κρέως ἢ πλακοῦντος ἢ ἄρτου οὐκ ὀρθῶς ἐρεῖ τις, ἀλλὰ
τόμος κρέως ἢ πλακοῦντος· τὸ δὲ τέμαχος μόνον ἐπὶ ἰχθύος λέγεται.
τέμαχος (‘slice’) of meat or cake or bread is not correct for one to
say, but τόμος (‘piece’) of meat or cake; τέμαχος is said only of fish.
(Phryn. 12)
Another feature is that Phrynichus does not accept just any occurrence
in a Classical author as proof of Attic status; he sets the bar high: the
author must be one of the select group of which he approves. This is why
we sometimes find him declaring a feature unacceptable even if someone
has pointed out an example in an ancient author, as in:
ἀκμὴν ἀντὶ τοῦ ἔτι· Ξενοφῶντα λέγουσιν ἅπαξ αὐτῷ κεχρῆσθαι, σὺ
δὲ φυλάττου χρῆσθαι, λέγε δὲ ἔτι.
ἀκμήν (‘still’) instead of ἔτι: they say Xenophon once uses it, but
you beware of using it and say ἔτι. (Phryn. 93; cf. Antiatt. 77.27;
Moeris α 149)
Here Phrynichus rejects the authority of Xenophon, who frustrated the
Atticists because he refused to fit the mould of an Attic writer.21 When
20
See LSJ for attestation of the two forms. The NT has only δυσί (× 9).
21
Cf. Phryn. 62, where Xenophon is accused of an outrage against his πάτριον διάλε-
κτον in saying ὀδμή instead of ὀσμή. But Antiatt. 92.26 counter-attacks by calling him
ὁ καλὸς Ξενοφῶν.
he says ‘they say’ he is referring to the work of other Atticists, less strict
than himself, who sought to legitimise this use on the basis of an occur-
rence in Xenophon (cf. below on the Antiatticist).
This use of ἀκμήν as an adverb meaning ‘still’ was well established in
Koine Greek, from the time of Polybius on; there is an example in the
NT at Matt 15:16. In the modern language ακόμα/ακόμη is its descendant.
But Phrynichus and others oversimplified the ancient evidence. As Lobeck
showed, in the Xenophon example (Anab. 4.3.26) and elsewhere, ἀκμήν has
the sense of ‘just then,’ a step before the shift to ‘still.’ So in genuine Attic
it is not simply a matter of avoiding ἀκμήν and replacing it with ἔτι.22
Phrynichus may even rebuke an ancient writer for not being more
careful:
ἐμπυρισμός· οὕτως Ὑπερείδης ἠμελημένως, δέον ἐμπρησμὸς λέγειν.
ἐμπυρισμός (‘burning’): so Hyperides carelessly, when one ought to say
ἐμπρησμός. (Phryn. 311; cf. Antiatt. 97)
Menander comes in for a good deal of disapproval of the same kind (he
is mentioned 17 times in the Ecloga). Sometimes the remarks suggest that
Phrynichus actually thought of the ancient writers as guided by a scale of
acceptability just like his own:
αἰχμαλωτισθῆναι· τοῦθ᾽ οὕτως ἀδόκιμον ὡς μηδὲ Μένανδρον αὐτῷ
χρήσασθαι. διαλύων οὖν λέγε αἰχμάλωτον γενέσθαι.
αἰχμαλωτισθῆναι (‘to be captured’): this is so unacceptable that not
even Menander uses it. Separate the words, therefore, and say αἰχμάλω-
τον γενέσθαι (‘to become captive’). (Phryn. 411)
Menander, it seems, might have been tempted to use this verb, but
‘even’ he, who set such low standards, knew it was beyond the pale and
avoided it.
A certain amount of background knowledge is often necessary to appre-
ciate an Atticist prescription fully, as in:
ἔνδον εἰσέρχομαι βάρβαρον· ἔνδον γάρ ἐστι καὶ ἔνδον εἰμὶ δόκιμον.
δεῖ οὖν εἴσω παρέρχομαι λέγειν. ἔσω δὲ διατρίβω οὐκ ἐρεῖς, ἀλλ᾽
ἔνδον διατρίβω.
‘I go ἔνδον’ (‘within’) is barbarous; for ‘he is ἔνδον’ and ‘I am ἔνδον’
are approved. One must say ‘I pass εἴσω’ (‘inside’). But you shall
not say ‘I spend time ἔσω,’ but ‘I spend time ἔνδον.’ (Phryn. 99; cf.
Antiatt. 91.31; Ammon. 169)23
22
Lobeck, Phrynichi Ecloga, 123–4. See further Shipp, MGE, 51–2, with more on the
Modern Greek forms.
23
[It is possible that ἔσω δὲ διατρίβω … is a separate lemma: see Fischer, Ekloge, app. crit.]
The need to prescribe the correct use of ἔνδον and εἴσω arose out of
a general trend of the language, the weakening of the distinction between
rest and motion in adverbs and prepositions. This had already begun in
the Classical period and gathered pace in Koine Greek. Thus ἔνδον origi-
nally applied to rest ‘in’ a place, εἴσω to motion ‘into’; but the differ-
ence gradually ceased to be felt and the words were often interchanged.24
Phrynichus and the other Atticists tackle the trend piecemeal rather than
in a systematic way. Another instance they notice is ποῖ versus ποῦ,25 and
Ammonius was trying to counteract the same trend with his distinction
between ἐκεῖ (‘there’) and ἐκεῖσε (‘thither’).26 But the wide reach of the
phenomenon and many other possible instances are not noticed by the
Atticists, at least in their surviving works, as, e.g., the running together
of εἰς and ἐν, which had the end result that (ει)σ- serves for both ‘in’
and ‘into’ in Modern Greek.
Finally, an example that illustrates the hazards of interpreting the Atti-
cists’ material and how easy it is for faulty argumentation to arise:
αὐθέντης μηδέποτε χρήσῃ ἐπὶ τοῦ δεσπότου, ὡς οἱ περὶ τὰ δικα-
στήρια ῥήτορες, ἀλλ᾽ ἐπὶ τοῦ αὐτόχειρος φονέως.
Never use αὐθέντης for the master, as the orators at the lawcourts do,
but for the murderer with his own hand. (Phryn. 89)
This word occurs in the Greek book of Wisdom at 12:6, with the mean-
ing ‘murderer’ (αὐθέντας γονεῖς ψυχῶν ἀβοηθήτων, ‘parents [who are]
murderers of defenceless souls’). Winston in his commentary sketches
the attestation of the word and its meanings from ‘murderer’ (in Hdt. etc.)
through ‘perpetrator, author’ (in ‘late prose’) to ‘master’ (in Hermas and
Pap.Graec.Mag.), noting that the last usage was condemned by Phrynichus.
He goes on:
A. Dihle has deduced from this that Wisdom was written in the first cen-
tury CE: ‘My point is that in this book the word authentês is used in the
meaning of ‘murderer,’ and this is the Attic meaning of the word which
no Hellenistic writer would have applied in the first century BCE.’27
Is this a valid deduction? I do not think so. It assumes that the mean-
ing ‘murderer’ disappeared altogether and was not known to anyone until
it was revived in the first or second century AD by the Atticists. This is
24
[See Lobeck, Phrynichi Ecloga, 128 for Class. evidence of εἴσω/ἔσω for ἔνδον, none
of which (he says) ‘deterred’ (absterruit) the Atticists.]
25
Phryn. 28; Moeris π 49; Ammon. 392; Philet. 271, 272.
26
Ammon. 159; also ἐνταυθοῖ etc. at 170; cf. Schmid, Atticismus, 1:91.
27
Winston, Wisdom of Solomon, 240–1 (with ref. to source of Dihle’s statement).
an improbable scenario. Old uses, like old words, may reappear at any time
in Greek. They therefore cannot be used to establish a date post quem; only
new uses and words can do that. Furthermore, Dihle overlooks an impor-
tant fact: Wisdom is written in literary Greek full of rhetorical features
and learned vocabulary. It is just the sort of Greek in which an older use of
αὐθέντης might be employed, even if the author was a ‘Hellenistic’ writer
composing in the first century BC; in fact it could be due to Atticistic influ-
ence already at work then.28
5. MOERIS
Little is known of Moeris himself, only that he left a sizeable lexicon,

variously titled, listing Attic features and their non-Attic counterparts.29
His date cannot be determined exactly, but since the work has been shown
to draw on Phrynichus and most of the other Atticists it is probably to
be dated in the third century AD. A welcome new edition, by Hansen,
has appeared quite recently, superseding the older works, which however
offer useful material.30
Moeris’s work, like Phrynichus’s, is clearly a handbook for the would-
be Atticist. Although Moeris mostly does not condemn or recommend
features outright as Phrynichus does, it is clear what the purpose is. The
method of presentation is compact and formulaic. It works with labels,
mostly the same two, Ἀττικοί (= Attic writers) and Ἕλληνες (= later
Greeks). Another label, κοινόν, appears from time to time, but exactly
what it is intended to convey is open to debate.31 The great majority of
the entries deal with forms and words; very few relate to uses of words
(e.g., ε 15). Sometimes entries record other information, such as an occur-
rence in an Attic writer (as δ 33 below) or the meaning of an Attic term
(e.g., ο 36). Occasionally an entry does not simply label features but
28
On the linguistic character of Wisdom see Reese, Hellenistic Influence. Thackeray
(Grammar, 13) labelled Wisdom ‘Literary and Atticistic.’ The meaning ‘murderer’ is
still in use in Ap. Rhod. 2.754 (Epic poetry, III BC); PCairZen 4.532.15 (poetic epitaph
for a dog, III BC) = Lloyd-Jones and Parsons Supplementum Hellenisticum, 977.15.
See Shipp, MGE, 115 [for futher details, including the modern descendant effendi.]
29
On Moeris see Wendel, ‘Moiris’; OCD, s.v.; Dickey, Scholarship, 98; Hansen, Moeris,
36–56. In Sakalis, ‘Ἀνάττικα καὶ Ψευδαττικά’ the inaccuracy of some of Moeris’s
statements about Attic is demonstrated.
30
Hansen, Moeris. The older editions are those of Pierson (1759), Koch (1830), and Bekker
(1833).
31
Cf. Maidhof, Begriffsbestimmung; Wendel, ‘Moiris,’ 2504. Maidhof claims to show
that κοινόν means ‘colloquial speech,’ but I have doubts: cf. ε 59 and η 6 below.
comments in some way and in so doing gives us an unmistakable glimpse

of the Atticist program (see δ 33 below). Some specimens to illustrate:
ἄρρενα Ἀττικοί· ἄρσενα Ἕλληνες. (α 45)
οἶσθα χωρὶς τοῦ σ Ἀττικοί· οἶδας Ἕλληνες. (ο 24)
εἴσεται Ἀττικοί· γνώσεται κοινόν. (ε 59)
ἤρεσέ με Ἀττικοί· ἤρεσέ μοι κοινόν. (η 6)
νώ δυϊκῶς Ἀττικοί· ἡμεῖς Ἕλληνες. (ν 2)
πάντοτε οὐδεὶς τῶν Ἀττικῶν. (π 57)
None of the Attic authors [uses] πάντοτε (‘always’).
διαφορότητος Πλάτων Θεαιτήτῳ· παρ᾽ ἄλλῳ οὐχ εὗρον. (δ 33)
διαφορότης (‘difference’; in gen. sing.) [occurs in] Plato’s Theaetetus;
I have not found it in any other [ancient author].
Some brief notes may be helpful. In ἄρρενα, we see the characteristic

Attic change of original -ρσ-, in contrast to retention in other ancient dia-
lects, from which ἄρσενα etc. entered Koine Greek. The budding Atticist
would need to learn to make the change to -ρρ- in all the words where
-ρσ- was usual in spoken Koine.
The next item relates to the morphology of the verb οἶδα ‘I know,’
whose irregular conjugation in Attic had been regularised in later Greek;
the next relates to its irregular future.
The lemma on ἤρεσε concerns a very minor point of syntax that
would seem hardly worth the trouble. In both these last we have the label
κοινόν, which here at least seems to bear the meaning ‘common,’ i.e., to
both Attic and later Greek, since this is true to the evidence: γνώσεται
and ἤρεσε + dative are well attested in Attic as well as later.
Next comes another attempt to revive the dual, as in our first example
in Phrynichus; here it is the dual first person pronoun (‘we two’).
The word for ‘always’ tended to be replaced in Greek: the newcomer
πάντοτε had already appeared in fourth-century BC Attic (Aristotle,
Menander, Philemon), but these writers were not acceptable models to the
Atticists. Moeris does not tell us what to say instead. Phrynichus recom-
mends ἑκάστοτε and διὰ παντός, without mentioning ἀεί, the other old
word that πάντοτε was replacing.32 It is no surprise that πάντοτε (with
πάντα) has come down as the modern word.
32
Phryn. 74. I think Phryn. is giving acceptable equivalents for πάντοτε as meaning ‘at
all times, on every occasion’; he takes ἀεί for granted. Shipp (MGE, 439) notes the
‘tendency to use strong expressions for the idea.’
The final lemma gives us a picture of Moeris searching the Classics for
the word διαφορότης, of which he was evidently suspicious, and coming
up with examples only in one dialogue of Plato.33 I think the message he
wants to convey is that the word remains under a cloud and the careful
Atticist had better be wary of it.
6. THE ANTIATTICISTA
The author of the work entitled Ἀντιαττικιστής (Latin Antiatticista)

is unknown and the title is not ancient.34 It is also something of a misno-
mer: the compiler was not opposed to Atticism as such; he only adopted
a less strict standard than Atticists like Phrynichus. This meant allowing a
wider canon of approved authors and accepting a single occurrence in a
Classical author as enough to make a feature good Attic, something not
conceded by Phrynichus and others like him. The enterprise belonged fully
within the framework of the Atticist program and reflected controversies
and differences of opinion between Atticists, not a complete rejection of
the program.35
The surviving text of the Antiatticista is a very reduced epitome of
a much ampler work. This is clear from the fact that many of the lem-
mata lack the supporting reference to a Classical author that would have
formed part of the original. One might suppose that it was compiled as
a reaction to Phrynichus’s Ecloga, but it emerges from a study of the
lemmata that Phrynichus made use of the Antiatticista.36 The date of the
work, then, must be the second century AD. The current edition remains
that in Bekker’s Anecdota of 1814.37 A new edition, fully annotated, is a
desideratum.38
33
Plato, Theaet. 209a.5; d.1; e.7; 210a.4, 9; three are gen. sing. There are four more in
Plato, but none in other Classical authors; it is not exactly rare later (see LSJ).
34
It appears the name was given by Ruhnken (1723–1798): see Naber, Photii Patriarchae
Lexicon, 1:95; cf. Bekker, Anecdota Graeca, 3:1074. There is also something wrong
with it: Greek ἀντι- would not normally mean ‘against.’ Ruhnken was thinking in
Latin.
35
On the Antiatticista see Dickey, Scholarship, 97–8; Lee, ‘Future of ζῆν,’ 290–2
[= Essay 4] (but my description ‘a rebel against the Atticizing fashion of his time’ was
overstated).
36
See Fischer, Ekloge, 39–41.
37
Bekker, Anecdota Graeca, 1:75–116; textual and other notes 3:1074–7.
38
De Gruyter reports (email 4 Feb 2009) that they have no edition in prospect. [An edition
by S. Valente (2015) has appeared since the above was written.] See Arnott, ‘Note on
the Antiatticist,’ for an instance of further work to be done on the text.
The work is no less valuable for the study of the Greek language than
the works of Phrynichus and the others, but its purpose, and hence the
layout of the entries, is somewhat different and needs to be understood
if it is to be interpreted correctly.39 The author gives a form, word, or use
censured as un-Attic by other Atticists and then cites an occurrence in a
Classical author to show that it does in fact occur, with the clear impli-
cation that it is acceptable, at least by the standards set by this Atticist.
There are about 800 lemmata in the list, in alphabetical order, with the
condemned feature usually as headword, though not invariably. Many entries
are very abbreviated because of the excision of material, but they are still
readily understood. Where the Classical authority has been omitted, the
intention is still plain; in some extreme cases there is nothing left but a
headword, but even these yield a meaning. But let us begin with an entry
that shows what is going on:
ἤθη· πληθυντικῶς φασι μὴ δεῖν λέγειν, ἀλλ᾽ ἑνικῶς. ἀλλ᾽ Ἀντιφάνης
εἶπε πληθυντικῶς.
ἤθη (‘habits’): they say you should not use this in the plural, but the
singular. But Antiphanes used it in the plural. (Antiatt. 98.15; η 4 V.)40
‘They’ are of course the stricter Atticists, and in this case we actually
have the statement of Phrynichus, who says one must guard against using
ἤθη in the plural, for οἱ δόκιμοι (‘the approved authors’) have the sin-
gular (Phryn. 344). An occurrence in the fourth-century BC comic poet
Antiphanes is cited to prove that it occurs in a Classical author, though
this would not have been enough to satisfy Phrynichus. The author of the
Antiatticista might also have cited Menander’s Thais, quoted by Paul in
1 Cor 15:33 (φθείρουσιν ἤθη χρηστὰ ὁμιλίαι κακαί), but this would have
cut no ice with Phrynichus either (see above on his attitude to Menander).
A more abbreviated type is readily interpreted:
δεῦρο· ἀντὶ τοῦ ἔρχου. Πλάτων Πολιτείας τρίτῳ. (Antiatt. 88.19;
δ 4 V.)
This means: ‘δεῦρο (‘hither’), when used in the meaning ἔρχου (‘come’;
‘come here’), is condemned by some, but there is an example in Plato’s
Republic, third book’ (he refers to 4.445c.1). One would not have expected
39
A misinterpretation of the Antiatticista by G. D. Kilpatrick is taken up in Lee, ‘Future
of ζῆν.’ For Kilpatrick’s response see ‘Atticism and the Future of ζῆν.’ Nigel Turner
completely misunderstands the Antiatticista in ‘Literary Character,’ 113–4.
40
References to page and line of Bekker’s edition, Anecdota Graeca, 1:75–116. [Refs. to
Valente’s edition (‘V.’) are added; there are no substantial differences.]
this use of δεῦρο to be under a cloud, since there are Classical examples,
but we learn that it was. To a rigorous Atticist it was the original function
of δεῦρο, as an adverb, that mattered, and any departure, even if ancient,
was not allowed. The condemned use of δεῦρο is of course widespread
in Koine Greek and occurs in the NT, for example in Mark 10:21 (δεῦρο
ἀκολούθει μοι).
A lemma from Moeris related to noun morphology:
κλεῖν ἀξιοῦσι λέγειν, οὐ κλεῖδα. Δίφιλος Εὐνούχῳ.
(Antiatt. 101.29; κ 17 V.; cf. Phryn. PS 82.17; Moeris κ 45; Philet. 116)
The accusative κλεῖδα, which regularised the anomalous declension,
was rejected by some Atticists, but can be cited from a play of Diphilus, a
comic poet of the fourth to third centuries BC, and is therefore acceptable.
Next an entry on a significant newcomer to the Greek vocabulary:
τρώγειν οὔ φασι δεῖν λέγειν τὸ ἐσθίειν, ἀλλὰ τὸ τραγήματα ἐσθίειν.
They say that one ought not to use τρώγειν to mean ‘to eat,’ but ‘to eat
nuts, fruit, etc.’ (Antiatt. 114.15; τ 2 V.)
The question relates to the use of the word, not to τρώγω itself, which
is as old as Homer. No ancient authority for the meaning ‘to eat’ is cited,
but we can assume that one was originally included, since that is the point
of the entry.41 Nevertheless it is true that τρώγω was not originally the
word for just ‘eat’: that did not become established until Middle Koine.42
τρώγω then became the standard vernacular and finally modern word (only
in the present tense: the aorist remained ἔφαγον). John’s Gospel in the
NT provides significant evidence along the way: τρώγω, Shipp observed,
‘is John’s word for “eat,” one of the features in which he anticipates
modern Greek.’43
My last illustration shows how a short and seemingly cryptic entry may
be unexpectedly useful:
μετάβα· ὥσπερ καὶ ἀνάβα καὶ κατάβα. Ἄλεξις Ἀμφώτιδι. (Anti-
att. 108.10; μ 25 V.)
‘In compounds … -βηθι and -βᾱ alternate [in the NT] without very
clear rationale,’ says Moulton.44 The lemma in the Antiatticista shows that
41
Cf. Philetaerus 231, with an example from Eupolis, which is perhaps the one, though
not very germane.
42
For a proposed division into Early (III–I BC), Middle (I–III AD) and Late Koine (IV–
VI AD) see Lee, ‘Ἐξαποστέλλω,’ 113 n. 31 [= Essay 12].
43
Shipp, MGE, 541. Note that John does not have pres. ἐσθίω/ἔσθω. BDAG, s.v. τρώγω,
still reflects the obsolete idea that τρώγω differs from just ‘eat’ in the NT: ‘to bite or
chew food, eat (audibly).’
44
MH, Grammar, II, 209.
the -βα form was disapproved of by the Atticists, and was therefore the
popular form; the approved form is not named in the lemma, but we can
be certain it was the archaic -βηθι. This gives us the clue to how these
alternatives were perceived by speakers in the first century AD, and may
help to dispel the mystery.45
7. OTHER SOURCES
There were many more Atticist grammarians than those mentioned so

far: producing handbooks was something of an industry once the Atticising
movement took hold. But their works mostly survive only in fragments, if
at all.46 A selection will be covered briefly in this section. The first three,
Aelius Dionysius, the Philetaerus, and Thomas Magister, offer lists simi-
lar in character to the works we have just seen; the other two, Ammonius
and Pollux, compiled works of a quite different shape, though they still
reflect the aims of Atticism. We will also need to notice another important
aspect of our topic, that is, the works of the Atticist writers themselves,
which form a major body of evidence of Atticist practice.
7.1. Aelius Dionysius
Aelius Dionysius, of the second century AD, produced an early lexi-

con of Attic, entitled Ἀττικὰ ὀνόματα. The full work is lost, but substantial
fragments can be recovered from the works of later authors.47 A sample:
εὐώνυμον λέγουσιν, οὐ μόνον ἀριστερόν. καὶ παρὰ Θουκυδίδῃ
πλεῖστον.
They [the ancients] say εὐώνυμον (‘left’), not only ἀριστερόν. And it
is most frequent in Thucydides. (Ael. Dion. ε 77; cf. Antiatt. 97.1)
This note alerts the would-be Atticist to the fact that ancient Attic had
two words for ‘left,’ not just the one familiar in his own time; it implies
45
Details of NT occurrences in MH, Grammar, II, 210; BDF, §95.3; most are affected by
variants (cf. ‘Atticism and the Text of the NT’ below). BDF, §2 (similarly BDR) confus-
ingly describe κατάβα as an ‘Atticism.’ The reason is that BDF use ‘Atticism’ both for
original Attic elements surviving in the Koine (as in this case) and for later Atticising
features revived by the Atticists (as, e.g., in BDF, §3 n. 4). Since -βα forms appear
already in Attic (see MH, Grammar, ibid.; Schwyzer, Grammatik, I, 676 n. 1), BDF can
call it an Atticism in the former sense; but it is the opposite of an Atticism in the latter
sense, since it is the popular form.
46
See Cohn, ‘Griechische Lexikographie,’ 693–5; Kazazis, ‘Atticism,’ 1205; Wendel,
‘Moiris,’ 2505–9; Dickey, Scholarship, 94–6, 98–100.
47
Edition: Erbse, Untersuchungen. Cf. Hansen, Moeris, 42–6; Dickey, Scholarship, 99.
that he should not forget to use εὐώνυμος in his writing or speech-mak-

ing. Its value for us is that it suggests that εὐώνυμος was dying out and
the ordinary word for ‘left’ was ἀριστερός. It follows that where writers
use εὐώνυμος it may be artificial or have an archaic flavour; the original
euphemistic force (‘well-named’) might also still be in play. This seems
to be relevant to the LXX, where there are 19 examples of εὐώνυμος to
65 of ἀριστερός, and to the NT, where εὐώνυμος is found 9 times in
contexts with a special flavour, besides 4 of ἀριστερός.48
7.2. The Philetaerus
The work known as the Φιλέταιρος (Latin Philetaerus) is more substan-

tial, and worth consulting alongside the works of Phrynichus and Moeris.49
It is attributed to Herodian (Aelius Herodianus Grammaticus, II AD) in
the MS tradition, but there is doubt about this.50 The date, too, is uncertain.
Dain dates the surviving redaction to the third century, or even the fourth
or early fifth.51 Two samples of the Philetaerus will suffice:
τῇ ἐπιούσῃ ἐρεῖς μὴ προστιθεὶς ἡμέρᾳ· τῆς δὲ ἐπιούσης ἡμέρας.
You shall say τῇ ἐπιούσῃ (‘on the next day’) without adding ἡμέρᾳ
(‘day’); but τῆς ἐπιούσης ἡμέρας (‘of the next day’) [is acceptable].
(Philet. 151)
A concern for such a nicety as the omission of ‘day’ – but only in the
dative! – shows the same pedantic spirit as the best of them – and another
subtle pitfall for the novice Atticist. It is interesting to compare the NT
evidence. The usual NT expression is τῇ ἐπαύριον (17 times); but Luke
has τῇ ἐπιούσῃ three times (Acts 16:11; 20:15; 21:18) and τῇ ἐπιούσῃ
ἡμέρᾳ once (7:26). Luke scores three out of four in the Atticist game; and
it may not be fanciful to think that in the last instance ἡμέρᾳ is a residue
from Exod 2:13 (τῇ ἡμέρᾳ τῇ δευτέρᾳ).
ἄρκτος σὺν τῷ τ. (Philet. 314)
48
See Chantraine, ‘Les mots designant la gauche,’ in Μνήμης χάριν, 1:61–9. NT occur-
rences of εὐώνυμος: in the Gospels only in the formula ἐξ εὐωνύμων, ‘de résonnance
probablement religieuse’ (69); Acts 21:3 euphemistic in a nautical setting; Rev 10:2
vision of an angel.
49
Edition: Dain, Philétaeros. Cf. Hansen, Moeris, 52–4.
50
Dain, Philétaeros, 9–13. Argyle (‘A New Greek Grammarian’) makes a good case
for attribution to Cornelianus, Imperial Secretary, to whom Phrynichus addresses his
Ecloga.
51
Dain, Philétaeros, 15. Dain’s arguments seem persuasive to me; but if Argyle is right,
II AD is required.
Though the entry is very terse, the interpretation is not difficult if we

bear in mind that the Philetaerus is a guide to Atticism. The entry means
that ἄρκτος (‘bear’) with the τ is the correct Attic form to be maintained;
it is taken for granted there is another form without τ to be avoided. Indeed
ἄρκος exists and the distribution fits: ἄρκτος is the ancient form in Homer,
Attic, and later literary sources; ἄρκος, clearly a more popular form, is
attested first in a third century BC papyrus,52 then in the LXX and a few
other Koine texts. It is no surprise that ἄρκος is what we read in Rev 13:2.53
7.3. Thomas Magister
Next is Thomas Magister, whose date is in the fourteenth century AD.

An Atticist grammarian still at work so late in the day may seem strange
and irrelevant, but not so. Guidance in writing in Attic style was still
needed because there were Byzantine writers doing just that; and Thomas’s
material is based on and continues the work of his much older predeces-
sors.54 This means that an observation of his may be as useful for the his-
tory of the language as one from a second century Atticist. Consider this:
βρέχειν οὐδεὶς τῶν ἀρχαίων εἶπεν ἐπὶ ὑετοῦ, ἀλλὰ ὕειν.
None of the ancients said βρέχειν (‘to wet’) with reference to rain, but
[they said] ὕειν (‘to rain’). (Th. Mag. 57.8)
The entry relates to Koine Greek developments in the word for ‘to rain.’
The old word ὕειν was replaced in everyday language by βρέχειν, which,
though old, had had different senses. Phrynichus has a note of similar
import, warning against this use (Phryn. 255). The new word is standard
in the NT; in the LXX Pentateuch there is an interplay between the two,
based on the difference of formality.55 Thus Thomas Magister’s note is
relevant to a text a millenium and a half older.
7.4. Ammonius
A work of a different character is the Περὶ ὁμοίων καὶ διαφόρων

λέξεων (De adfinium vocabulorum differentia) attributed to Ammonius,
but probably by Herennius Philo. In its original form it is datable to the
52
PMich 1.86.5 (Philadelphia) περὶ τῆς ἄρκου.
53
Technically ἄρκος shows reduction of a cluster of three consonants (-rkt-) by dropping
one, a common phenomenon: cf. artic in older Engl. and Lat. articus. See Shipp, MGE,
97, for modern evidence.
54
Edition: Ritschl, Thomae Magistri ecloga.
55
See Lee, Lexical Study, 122–4.
late first or early second century AD.56 It is quite extensive (525 lemmata)
and many entries contain several lines of discussion. The aim of this work
is to draw distinctions between apparent synonyms, mostly two, but some-
times three or even four.57 The distinctions are mostly well done, but it is
important to realise that they describe Attic norms of an earlier time. The
purpose is didactic, to set out and maintain distinctions (sometimes even
words) that had been lost, or were fading away, in the current language.
The statements are evidence for the loss of these distinctions or words in
the first century AD, not a real picture of how the words were used at that
time. There is also a tendency to find distinctions whether genuine or not,
and some are not true, or are too absolute, even for Attic.
It is impossible to do justice here to the riches of this source; I offer
only one specimen, in which the words for ‘wake up’ and ‘get up’ are the
subject of attention:
ἀναστῆναι καὶ ἐγερθῆναι διαφέρει. ἀναστῆναι μὲν ἐπὶ ἔργον,
ἐγερθῆναι δὲ ἐξ ὕπνου.
ἀναστῆναι (‘to get up’) and ἐγερθῆναι (‘to wake up’) differ. [You say]
ἀναστῆναι for work, but ἐγερθῆναι from sleep. (Ammon. 50; simi-
larly 216)
What makes it necessary for a distinction to be made between ἐγείρο-
μαι (aor. ἠγέρθη), originally ‘I wake up,’ and ἀνίσταμαι (aor. ἀνέστη),
‘I stand/get up’? It is the fact that in Koine Greek the former increas-
ingly invaded the territory of the latter, that is, ἐγείρομαι was used to mean
‘I stand/get up’ (as well as ‘I wake up’).58 Examples are found all over Koine
Greek, as in Mark 2:12 (ἠγέρθη καὶ εὐθὺς ἄρας τὸν κράβαττον ἐξῆλθεν,
‘he got up and …’). Ammonius is aware of this development and is resisting
it by defining the usage of Classical Attic; it is implicit that contemporary
usage is condemned. The diligent Atticist will take note.
7.5 Pollux
The Onomastikon of Pollux (Julius Pollux/ Πολυδεύκης) is well known
and often cited in lexicons and elsewhere. The work is a topical diction-
ary, the first of its kind. It is a list of all, or most, of the words in the Greek
vocabulary, arranged by topics or subjects. Meanings of words, as in a
lexicon, are not the primary goal and are not usually indicated. The original
56
Edition: Nickau, Ammonii Vocabulorum Differentia. Cf. Cohn, ‘Ammonios’; Hansen,
Moeris, 51–2; Dickey, Scholarship, 94–6. Authorship and date: Nickau, lxvi–vii.
57
Rarely homonyms, such as 390 πεῖρα and πήρα, 399 πόμα and πῶμα, despite Dickey,
Scholarship, 95 (her examples are not homonyms).
58
Cf. Shipp, MGE, 75–80.
was compiled in the second century AD and survives in a reduced though

still very extensive form.59 The true nature of this collection is seldom
spelt out and often not understood. It is not a record of the Greek vocabu-
lary as it stood in the second century AD, but a list of acceptable Attic
words and turns of phrase from the golden age of Athens, compiled by a
later scholar looking back. Ordinary Koine Greek vocabulary also appears,
but only to the extent that it is the same as the older vocabulary. The book
is a resource for those wishing to write or speak in the ancient manner;
for this reason it is also in part encyclopedic, including not just words, but
information on the life and culture of ancient Athens, all derived from
the ancient texts. Pollux was an Atticist; he was not an objective observer
of the language of his own time.
That this is so can be readily shown by a few samples, beginning with
some of the words discussed previously. In the list of words for eating,
ἐσθίω appears (6.39) but not τρώγω. Ἀεί is in (1.151), but not πάντοτε.
The word ἀκμή appears several times in various uses (1.60, 61, 177, 180;
2.10; 5.158) but not the adverbial use ἀκμήν, ‘still.’ The form ἄρκτος is
listed (4.157, 159; 5.81, 15), but not ἄρκος. Words condemned by Atti-
cists may be listed, but with a comment. So κράββατος (‘bed, stretcher’),
condemned by Phrynichus (cf. next section, §8), is entered like this (10.35):
καὶ κράββατον εἰρῆσθαι λέγουσιν· ἐγὼ δ᾽ οὐκ ἐντετύχηκα τοῖς δράμα-
σιν (‘they say κράββατος was used [by ancient writers] but I have not met
it in the dramatists’). Most tellingly κοράσιον (‘girl’), also condemned by
Phrynichus, attracts this comment (2.17): τὸ γὰρ κοράσιον εἴρηται μέν,
ἀλλὰ εὐτελές, ὥσπερ καὶ τὸ κορίδιον (‘κοράσιον is said, but it is cheap,
as is κορίδιον’). As for εὐώνυμος, it appears a number of times (along
with ἀριστερός), as one would expect; but once it appears (1.88) in a
context that reveals the purpose of the collection clearly: ὀνομάσαις δ᾽ ἂν
τοῖχον δεξιὸν καὶ εὐώνυμον (‘and you may call the side[s] [of the ship]
δεξιὸν καὶ εὐώνυμον’). This is in a list of the vocabulary of the trireme,
the warship of another era.60
7.6 Atticist writers

Finally in this section we must take note of the fact that the works of
Atticising writers are themselves a major source of data on the practice of
Atticism. As already mentioned, most of the literature from the first cen-
tury BC onwards was affected by Atticism. The degree of effect and the
59
Edition: Bethe, Pollucis Onomasticon. Cf. Bethe, ‘Pollux’; OCD, s.v. Pollux; Dickey,
Scholarship, 96.
60
For an extended example from Pollux, see Lee, History, 200.
exact features deployed vary from one author to another, though with a
solid core of common Atticisms. The complexity of the situation is well
illustrated by Lucian, who was an expert Atticist himself but can vary his
style up and down the scale and even make fun of Atticist pretensions.
The control of this vast body of data is difficult for us for obvious rea-
sons. Schmid studied some important representatives in great detail in his
classic work, but it is not exhaustive.61 The language of Dionysius of Hal-
icarnassus has not yet been examined in depth, though he is key evidence
for the onset of the Atticist phenomenon in the first century BC. A legion
of other Atticising writers over many centuries is available for further
attention.62
I take one brief illustration of how Atticist writers provide evidence
on Atticising features, namely, the substantival neuter adjective with τό
functioning as an abstract noun. Though not explicitly remarked on by any
Atticist grammarian to my knowledge, it is one of the ‘hallmarks’ of Atti-
cist usage.63 There is an instance early on in Dionysius’s De compositione
verborum 1.7 (τὸ περὶ τὰς λέξεις φιλόκαλον, ‘the love of fine language’).
Phrynichus, in the preface to his Ecloga, where he shows off his technique,
has οὐ γάρ τις οὕτως ἄθλιος ὡς τὸ αἰσχρὸν τοῦ καλοῦ προτιθέναι (‘for
no one is so pitiful as to prefer the ugly to the beautiful’). In the NT there
is a concentration in [Luke–]Acts and the Epistles. A suitably contrived exam-
ple is heard in the mouth of Paul before Agrippa in Acts 26:7 (τὸ δωδεκά-
φυλον ἡμῶν = ‘our twelve tribes’). This, Haenchen aptly says, is ‘meant
to sound solemn.’64
8. THE ATTICIST GRAMMARIANS AND THE NEW TESTAMENT
The Atticist grammarians did not, of course, comment directly on the

NT, which would have been beneath their notice if they had even heard
of it. Conversely, the NT writers could not have been directly influenced
61
Schmid, Atticismus. Cf. Anderson, Second Sophistic, 263: ‘the daunting scale of Schmid
… should not blind us to the fact that the linguistic map of the second century AD, for
example, is still far from complete. It is disconcerting to be confronted with the hundreds
of words in Galen unreported in LSJ, for example.’
62
Besides the work of Schmid, short lists of features may be seen in Kazazis, ‘Atticism,’
1207–8; Horrocks, Greek, 83–4. [See Kim, ‘Literary Heritage,’ 472–4 for useful obser-
vations in D.H.]
63
Horrocks, Greek, 83.
64
Haenchen, Acts, 683 n. 7 [I would add ‘learned’]. On this feature cf. MHT, Grammar,
III, 13–4; BDF, §263; Schmid, Atticismus, 4:608; Lee, ‘Some Features,’ 23–4 (τό +
substantival participle). There is no complete list of exx. in the NT, it appears.
by any of our Atticist grammarians, because the NT writers preceded

them. The grammarians’ material is nonetheless of great relevance to the
NT. To begin with, a comparison between Atticist strictures and the lan-
guage of the NT provides proof of its generally Koine Greek character.
Time and again the feature disapproved of by the Atticists is the one that
appears in the NT. Here is just a small selection of condemned words and
forms found in the NT, set alongside the recommended equivalents,
which do not appear in the NT:
γλωσσόκομον γλωττοκομεῖον (Phryn. 70)

γρηγορῶ ἐγρήγορα (Phryn. 88)
ἐμπτύει καταπτύει (Phryn. 8)
κλίβανος κρίβανος (Phryn. 150)
κοράσιον κόριον, κορίδιον, κορίσκη (Phryn. 50)
κράββατος σκίμπους, ἀσκάντης (Phryn. 41; Moeris α 119, σ 33;
Th. Mag. 333.2)65
μύλος μύλη (Moeris μ 5)
οἰκοδεσπότης οἰκίας δεσπότης (Phryn. 349; Th. Mag. 259.3)
σίναπι νᾶπυ (Phryn. 252; cf. Moeris ν 16)
τάχιον θᾶττον (Phryn. 52; Moeris θ 18, τ 7;
Philet. 18)
φάγομαι ἔδομαι (Phryn. 300; Th. Mag. 126.13)
None of this is by accident: the writers of the NT, for the most part,
were using the Koine Greek of their day, and the evidence of the Atticist
grammarians helps to prove it.66
The statements of the Atticist grammarians may also further our under-
standing of NT usage of words. We have seen some straightforward exam-
ples already (such as τρώγω), but it can also happen in quite unexpected
ways. Without the clear statement of Moeris (α 132) that ἀπάτη means
τέρψις (‘enjoyment’) in later popular Greek, such a meaning might not
have been suspected in Matt 13:22, Mark 4:19 and 2 Pet 2:13.
A more remarkable instance arises from an entry of Ammonius in which
he differentiates between παρρησία and παρουσία. As Shipp has shown,
the modern evidence confirms the possibility of semantic confusion of these
seemingly dissimilar words. Ammonius needs to explain the difference
65
Luke and Matt avoid κράββατος where it occurs in Mark, but Luke uses it himself
in Acts 5:15; 9:33; see Shipp, MGE, 106 for a probable explanation of the exx. in
Acts.
66
Similar lists are offered in Triantaphyllides, Γραμματική 1: Ιστορική Εισαγωγή, 411–8
(very extensive), and Browning, Greek, 47–8.
between them because the confusion was occurring in his time. The NT
may reflect this confusion in that it is possible that adverbial παρρησίᾳ
(‘openly, publicly’) has rather the sense of ‘(present) in person,’ in, e.g.,
John 11:54 (οὐκέτι παρρησίᾳ περιεπάτει ἐν τοῖς Ἰουδαίοις).67
9. ATTICISM IN THE NEW TESTAMENT
That the language of the NT is basically Koine Greek is clear. Though

the 27 books of the NT differ in position on the scale from vernacular to
literary, none can be properly described as Atticistic, with the possible
exception of 2 Peter.68 But it is not true that the NT is untouched by Atti-
cism. As we have seen, the phenomenon was well under way in the first
century AD and making its influence felt in the education system; it would
be no surprise if the NT books reflected to some degree what their authors
had learnt in school. Atticistic features do in fact appear in the NT, in a
range of environments. The identification of these features is greatly aided
by the works of the Atticist grammarians, though it is not dependent on
them alone: they may be recognised from the practice of Atticist authors
or from general developments in the language. Important also for under-
standing the NT phenomena is distribution, that is, the kinds of books and
contexts in which significant features are found.
Let us begin with Luke’s Gospel, where the instances are prominent.
Luke, in his rewriting of Mark, frequently makes changes that are in line
with Atticist precepts. Many were noted, along with other changes towards
more stylish or elegant wording, by Cadbury in his notable study of Luke’s
language.69 In the following small selection, the word or expression replaced
by Luke is one censured by the Atticist grammarians:
Mark 5:41 τὸ κοράσιον || Luke 8:54 ἡ παῖς
Mark 5:42 τὸ κοράσιον || Luke 8:55 omitted70
Mark 5:23 ἐσχάτως ἔχει || Luke 8:42 ἀπέθνῃσκεν71
67
Ammonius 401; see Shipp, MGE, 441–2. This example has not yet had any impact on
the lexicons. [An Engl. ex. of the present day (2021) may be mentioned, the confusion
of wonder and wander, probably a result of the misleading spellings.]
68
Cf. Norden, Die Antike Kuntsprosa, 2:479–510; Lee, ‘Some Features,’ 9 (with references);
on 2 Peter MHT, Grammar, III, 30, 126–7.
69
Cadbury, Style of Luke, esp. 182–8, 196–7, 201. Earlier Norden, Die Antike Kunstprosa,
2:485–92.
70
Phryn. 50; Cadbury, Style of Luke, 186.
71
Phryn. 369; Th. Mag. 136.4; Cadbury, Style of Luke, 182.
Mark 10:25 ῥαφίδος || Luke 18:25 βελόνης72

Mark 1:44 + 5 ὕπαγε || Luke 5:14 + 5 all omitted or changed73
Apart from these instances of avoidance, reflections of Atticism can

be found lightly spread through the NT books that are of a higher literary
level, namely, Luke–Acts, the Epistles, and to a lesser extent Matthew.
The features appear generally, but some may also occur in special con-
texts, where the style has been raised to suit the speaker or the situation.
What follows are just a few samples.
The control of particles, whether by employing those that were obso-
lete or by skilled handling of those still available, is a sign of literary taste
inspired by Atticism. On this large topic a small selection must suffice.
τοίνυν appears only in Luke, 1 Corinthians, and Hebrews, once each. The
even more elevated τοιγαροῦν makes two appearances only, in 1 Thes-
salonians and Hebrews. Rare δήπου is used only in Heb 2:16 (‘classical,
literary,’ BDF).74 Beside these very infrequent particles, let us notice τε,
‘an indication of stylistic pretension,’ as Turner puts it, which is very fre-
quent in Acts, found occasionally elsewhere, and not at all with certainty
in Mark, John, and Revelation.75
Next a short list of Atticistic features of various kinds occurring only
in literary books or special contexts:
ἴσασι Acts 26:4 Paul’s speech before Agrippa: Koine οἴδασι × 7.76
ἴστε (indic.) Heb 12:17: Koine οἴδατε × 63.77
χρή Jas 3:3.78
μέγιστος 2 Pet 1:4.79
72
Phryn. 63; Th. Mag. 56.16; Antiatt. 113.14 (ρ 7 V.); Cadbury, Style of Luke, 187. Elliott
(Essays and Studies, 66), rejects the reading βελόνη, even in Luke, because there is a
v.l. ῥαφίς: he does not allow for Luke’s Atticising tendency. He takes the opposite tack
(ibid., 71) with ἦσθα in Matt 26:69, where there is no v.l. (see below).
73
Antiatt. 114.31 (υ 1 V.); Cadbury, Style of Luke, 173; Shipp, MGE, 35. I consider Markan
priority inescapable. Those who reject it face the task of explaining why an author would
revise the language of his source text downwards, against the trend of the time, and that
too not only by replacing ‘correct’ words with vernacular ones (e.g., βελόνη) but by
introducing vernacular words that had no counterpart in the source (e.g., some of ὕπαγε).
For further telling evidence, see Luke’s handling of the participle (Cadbury, Style of
Luke, 133–7).
74
On these three, see MHT, Grammar, III, 339–40; on δήπου also BDF, §441.3.
75
MHT, Grammar, III, 338–9; BDF, §443; Kilpatrick, ‘Atticism and the Text,’ 135; BDAG,
s.v. Cf. μέν: Lee, ‘Some Features,’ 4–6.
76
Moeris ι 22; BDF, §3 n. 4.
77
Phryn., PS 53.
78
BDF, §358.2; Schmid, Atticismus, 4:592.
79
MHT, Grammar, III, 29–30.
μεταξύ × 8 in Matt, Luke, Acts, Rom + John × 1 (ἐν τῷ μεταξύ).80

ἀμῶ Jas 5:4: Koine θερίζω × 21.81
καθότι Luke–Acts × 6; καθό Rom, 2 Cor, 1 Pet × 4; καθάπερ Rom, 1
and 2 Cor, 1 Thess, Heb × 13: Koine καθώς × 182.82
A final point here. The fact that a NT author such as Luke may also use
words and forms censured by the Atticists does not invalidate these obser-
vations. As a general principle, the more popular feature can reappear at
any time, since it is the default form, even if a more ‘correct’ one is usu-
ally deployed. But the reason is often likely to be that the rejection of some
features belonged to higher, stricter levels of Atticism than that to which
the NT authors aspired. It may also be that some of these features had not
attracted Atticist attention in the first century AD. I take three examples
from among many. Both βασίλισσα and ῥύμη (in the meaning ‘lane’), are
deprecated by the Atticist grammarians, but both are used in Matt and
Luke–Acts (besides βασίλισσα once in Rev). The future form φάγομαι,
lambasted by Phrynichus as ‘barbarous,’ occurs in the NT not only in
Revelation where we would expect it, but also twice in Luke and once in
James; probably Attic ἔδομαι was beyond the ken of Luke and James.83
10. ATTICISM AND THE TEXT OF THE NEW TESTAMENT
Atticism has relevance to the NT in another way, one that relates not
to the choice of words at the time of composition but to the transmission
of the text. There is good reason to believe that copyists made changes
to Koine Greek words and forms to match Atticist standards. The process
was not systematic or complete; it shows itself only in the fairly frequent
but sporadic appearance of manuscript variants that have an Atticising
character. It may then be argued that in these cases – whether some or all
is debatable – the Atticising variant is secondary. The matter is of course
80
MHT, Grammar, III, 277 (‘literary and even atticizing’).
81
Moeris α 158; Shipp, MGE, 63.
82
Moeris κ 17; Phryn. 399; Th. Mag. 195.8; MHT, Grammar, III, 319–20. There are many
other features, such as optatives: see scattered observations in MHT, Grammar, III;
BDF; Voelz, ‘Language of the NT.’
83
βασίλισσα Phryn. 197, 231; Moeris β 16; Antiatt. 84.26 (β 16 V.); Philet. 121. NT:
Matt 12:42; Luke 11:31; Acts 8:27; Rev 18:7. ῥύμη Phryn. 383; Antiatt. 113.6 (ρ 2 V.).
NT: Matt 6:2; Luke 14:21; Acts 9:11; 12:10. φάγομαι Phryn. 300, cf. 325. NT:
Luke 14:15, 17:8; Jas 5:3; Rev 17:16; also καταφάγεται John 2:17 (LXX). ἔδομαι is
not used by any NT author. [See Adams, S. A., ‘Atticism, Classicism, and Luke–Acts’ for
a discussion of the placing (and labelling) of Luke’s Greek in the development of Atticism.]
of great importance to the textual critic of the NT and raises questions

that have caused sharp debate. The recognition of this phenomenon was
the work especially of G. D. Kilpatrick, followed by J. K. Elliott, both of
whom have published extensively on the topic.84
To illustrate the point let us take two examples. ἦς was the Koine form
of the second person sing. impf. of the verb ‘to be,’ replacing Attic ἦσθα.
The Atticists naturally ruled the former unacceptable and required the
latter.85 In the NT ἦς occurs six times without variant (Matt 25:21, etc.);
but in Mark 14:67 we find ἦς in a small number of MSS and ἦσθα in
the majority. It may be, in fact it looks rather likely, that ἦς is original and
ἦσθα a scribal change. The former would be expected in Mark’s low-level
Koine; the latter could be due to Atticism as well as assimilation to the
parallel, namely Matt 26:69, where ἦσθα is without variant and likely to
be original in Matthew’s more literary Greek.86
In the case of ἄρρην versus ἄρσην, quoted from Moeris above, we find
that in seven of the nine NT occurrences of ἄρσην there is a variant ἄρρην,
exactly the form recommended by the Atticists and probably a scribal alter-
ation in the course of transmission of the NT text.87
11. CONCLUSION
Atticism was a phenomenon of the utmost significance for the history

of post-Classical Greek. Once it came into the picture, every writer could
not avoid taking, consciously or unconsciously, a position in relation to
it, from complete indifference to complete obedience to its dictates. So we
today cannot fully appreciate any text unless we are aware of its potential
effects and the presence or absence of the features Atticism tried to impose.
The specimens we have seen show that this applies to the books of the
NT just as much as to other texts of the period. Both the Koine features
noted above in §8 ‘The Atticist Grammarians and the NT’ and the Atti-
cising features in §9 ‘Atticism in the NT’ acquire their significance against
the background of Atticism. The Atticist grammarians, suitably interpreted,
provide a direct window into the phenomenon. They not only alert us to
84
See esp. Kilpatrick, ‘Atticism and the Text.’ Elliott’s work is conveniently consulted in
Elliott, Essays and Studies; on Atticism see 30–2, 65–77, 106–11. For the textual debate
see, e.g., Epp and Fee, Studies, esp. chs. 7, 8; Elliott, Essays and Studies, ch. 1.
85
Phryn. 118; Moeris η 4; cf. Ammon. 220.
86
Elliott, Essays and Studies, 70–1. The argument is not accepted by NA27, which reads
ἦσθα in Mark 14:67. [NA28 likewise.]
87
For full NT details see Elliott, Essays and Studies, 75.
Atticising features, but assist us, along with other evidence, in reconstructing
the history of those features and their Koine Greek counterparts. They are
also a guide to assessing the literary standing of any text of the period.
The study of Atticism may seem a dead subject, exhausted by older
researches. But this is not so. There is much room for future work. I sug-
gest the following topics for further attention. (1) A new edition of the text
of the Antiatticista. [Valente’s edition (2015) has met this need.] (2) More
linguistic commentary on the Atticist grammarians, especially in the light
of evidence from other sources; modern editions provide only the text, and
for commentary we have to go back to older scholarship, some from the
eighteenth century. (3) Further study of reflections of Atticism in the ver-
nacular documentary texts; there is a yawning gap here. (4) A systematic
study of the relation of Atticism to the NT; this would lead to a more accu-
rate description of the stylistic character of each book than what we have at
present. (5) The compilation of a list or register of stylistically marked fea-
tures, based on both the Atticists’ statements and the observed practice of
Atticising authors;88 this would be a valuable tool for achieving objectives
(3) and (4), as well as a fuller understanding of the complex phenomenon
known as Atticism.
88
But not by ‘consulting … Attic authors of the fifth and fourth centuries B.C. to form our
own picture of Attic use,’ as suggested by Kilpatrick (‘Atticism and the Text,’ 134): this
would give us real Attic, not the Atticists’ Attic.
20
ETYMOLOGICAL FOLLIES:
THREE RECENT LEXICONS OF
THE NEW TESTAMENT
20131
Abstract
Three smaller lexicons of NT Greek have recently appeared, all introducing

etymological material of certain kinds. The lexicons are those of Trenchard (2003),
Danker (2009), and Newman (rev. ed., 2010). This paper examines the etymo-
logical material in each lexicon and points out serious problems arising from
dependence on obsolete works as well as other weaknesses. In conclusion the
value of including etymologies in NT lexicons is assessed and the future direction
considered.
1. INTRODUCTION
For most of the past century we have not been accustomed to see infor-
mation on etymology or derivation of words in our lexicons of the Greek
NT, nor does it seem to have been greatly missed. In three recent smaller
lexicons a trend towards inclusion of such material has emerged. In prin-
ciple there is no objection to it, and it could prove useful to the beginner
and experienced scholar alike, if suitably handled. But for the lexicogra-
pher it involves excursions outside familiar territory into a specialised field
that may prove hazardous.
Before proceeding to a discussion of these new lexicons, it will be use-
ful to set out a brief analysis of what might be referred to by ‘etymology’
in a Greek lexicon. There are, as it were, three potential levels, correspond-
ing to the depth of penetration into the history of the word or form. The
1
Parts of this paper were presented at the SBL Annual Meeting, New Orleans 2009 (on
Trenchard; also Souter-House), and the Senior New Testament Seminar, Faculty of
Divinity, Cambridge, 2011 (on Trenchard and Danker). I am grateful to Michael Curran,
Anne Thompson, and Steve Walton (see n. 34) for reading this paper, and especially to
Anne for many helpful points.
examples I give simply illustrate the kind of information that might come
under each heading; the manner of presentation varies from one lexicon
to another.
a. Immediate derivation
E.g., λόγιος < λόγος, καταλέγω < κατά + λέγω. Here we are dealing
with simple word-formation and composition within Greek, and the rela-
tionships tend to suggest themselves.2
b. Relationships between stems of the same ‘root’

E.g., λόγος ~ λέγω. The example involves vowel gradation (or ‘Ablaut’);
this one is a common pattern and easy to recognise. But things can get
complicated, as in τέκνον ~ τίκτω, where we have to deal with not only a
zero grade of the root but reduplication and a metathesis (τίκτω < *τί-τκ-ω).
Further, the derivatives of a root may branch off in different directions
semantically, as in ἐκλογή, ‘selection,’ relating to λέγω not in its common
meaning ‘say,’ but in an older sense ‘collect, choose.’
c. Ultimate etymology, i.e., origins and cognates in other languages

E.g., λέγω < IE *leg- : Lat. lego. That is to say, λέγω descends from
Indo-European (IE) into Greek and a cognate is found in Latin lego. This
one is easy, but most of them are not; all sorts of complications can arise
and the etymology often remains obscure. Also, relationships in type b.
often involve going back into much earlier etymology of type c. to find
the link (e.g., εὐδία ~ Ζεύς: see below in 2.b). Loanwords belong here too
under origins (cf. below in 4.b).
The recent history of etymological material in the NT lexicons may be
briefly outlined now. The major lexicon of our time, Bauer (1928–2000)
has not included any until BDAG (2000), which offers very limited data,
mainly with a semantic purpose, in some entries.3 Nor is it found in the
small lexicons of the past century, such as Newman (1971), Gingrich
(1965), Souter (1916), and before that Hickie (1893).4 In Abbott-Smith’s
2
Symbols: ‘<’ = ‘derives from’; ‘~’ = ‘is related to’; ‘*’ = ‘hypothetical form’; ‘:’ = ‘is
cognate with.’
3
Bauer2 (1928); Bauer 3rd to 5th eds. (1937, 1952, 1958) (see Lee, History, 362 for full
details); BRAA (1988); BAG (1957); BAGD (1979); BDAG (2000).
4
Newman, Concise Dictionary; Gingrich, Shorter Lexicon; Souter, Pocket Lexicon; Hickie,
Greek–English Lexicon. In the revised ed. of Gingrich by Danker, ‘English-language
ETYMOLOGICAL FOLLIES 321
medium-sized lexicon (1922), however, there is information of types a.

and b., rarely of c. (loanwords, mainly).5 The information is sound, clearly
presented, and carefully limited: questionable or obscure relationships are
not entered into. Further back, in the nineteenth century, the inclusion of
etymological matter was standard; but this is as far as we need to go.
In ancient Greek generally, the leading lexicon, Liddell and Scott, pro-
vided etymological material in all of the first eight editions (1843–1897),
but in the ninth (completed 1940) it was curtailed (‘reduced to a mini-
mum’).6 The remnants are mostly of types a. and b., rarely c.; the last
underwent some updating, but its quality remains in doubt.7 But the main
resource for Greek etymology is not the lexicons: there is a rich stream
of etymological dictionaries devoted entirely to the subject. It happens that
a new one, Beekes (2010), has just appeared, updating all in the light of the
latest thinking.8 For many years before that we could rely on Frisk (1960–
1972) and Chantraine (1968–1980).9 Both remain valuable, and Chantraine,
with its attention to the later history of the words, seems unlikely to go out
of date. Frisk and Chantraine replaced the earlier standard work of Boisacq
(1916).10 All of these are very large works and admittedly not easy for the
uninitiated, but they are the indispensible source of current knowledge on
Greek etymology. To ignore them is to court disaster.
Now we have the new lexicons of Trenchard (2003), Danker (2009) and
Newman (rev. ed. 2010), all introducing etymological material.11 These
are smaller lexicons, a type with a long history going back centuries. The
smaller lexicon has obviously fulfilled a useful practical role, not only
as a tool for beginners but as a handy reference book for all. This is reason
enough to take them seriously and to assess the new feature now intro-
duced (again). Although subordinate to their main purpose, the etymological
material cannot be regarded as an insignificant extra: it is prominent in the
derivatives’ are added at the end of some entries: e.g., ‘ἀγωνία … [agony].’ In some there
is information of type a. as well: e.g., ‘ἄγνωστος … [agnostic, α priv. + γνῶσις].’
5
Abbott-Smith, Manual Greek Lexicon. Similar information is found in Zorell, Novi
testamenti lexicon and Ebeling, Griechisch–deutsches Wörterbuch.
6
LSJ, p. x.
7
See LSJ, Suppl. (1996), p. vi: ‘The occasional etymological notes in LSJ are frequently
out of date.’ Readers are then referred to Chantraine and Frisk.
8
Beekes, EDG. Beekes’s aim is to update in the light of: (a) the laryngeal theory; (b) Myce-
naean; and (c) the abandonment of the Pelasgian theory (p. vii).
9
Frisk, GEW; Chantraine, DELG.
10
Boisacq, Dictionnaire étymologique. The predecessors of Boisacq were Prellwitz, Ety-
mologisches Wörterbuch (1892), and Curtius, Grundzüge der griechischen Etymologie
(1858–1862; 5th ed., 1879).
11
Trenchard, Concise Dictionary; Danker, Concise Lexicon; Newman, Concise Dictionary.
entries and will be trusted by users, most of whom are not equipped to
judge its accuracy. In what follows I focus on this aspect of the new lexi-
cons, with only brief remarks on the remaining content. I begin with Trench-
ard, who can be credited with starting the new trend, then look at Newman,
whose work will turn out to be closely related; Danker’s work is of a dif-
ferent order. I conclude with general reflections on the purpose and value
of such material in NT lexicons and what the future direction might be.
2. TRENCHARD, A CONCISE DICTIONARY OF NT GREEK (2003)
All lexicons derive material from predecessors, all the more so in the
case of smaller lexicons, which tend to be abridgements of larger works.
Trenchard indicates his main source as BDAG, on which he has drawn
‘extensively.’12 In truth it would have been fairer to call this lexicon a
concise BDAG: the glosses in BDAG (without Danker’s ‘extended defi-
nitions’) have been taken over wholesale with few changes to them. There
are modifications, however, that cause serious problems, especially the
omission of the numbered sense divisions, a move which results in undif-
ferentiated lists of glosses, and the inclusion in the glosses of what were
originally translations of particular occurrences in context (the references
have mostly disappeared as well). These changes impair the value of BDAG’s
material and of the whole exercise. But this aspect of Trenchard’s lexicon
need not be entered into further here.13
Trenchard explains how he presents the etymological material in his
lexicon as follows, under a heading ‘Cognate key word(s)’:
This part of the entry lists up to three key words that share a cognate
relationship with the headword, if there is a cognate word in the NT. The
determination of cognate word groups and their related key words largely
follows those identified by J. Harold Greenlee in A New Testament
Greek Morpheme Lexicon. … An asterisk (*) indicates that the head-
word is itself the key word to which other words are related.14
12
Trenchard, Concise Dictionary, xiii.
13
The reader might like to sample δίδωμι, in which BDAG’s 17 separate senses (plus
subdivisions) are collapsed into one list of 35 glosses, and to notice how BDAG’s con-
textual translations go, venture somewhere (for Acts 19:31) and take pains (Luke 12:58)
have turned into meanings, the latter without its pains, so that δίδωμι appears to mean
both ‘give’ and ‘take’! Rolando Ferri’s long review (BMCR 2004.07.36) raised con-
cerns about the same problems, with full illustrations. None of the seven reviews I
have seen offers any critique of the etymologies. One declares that Trenchard ‘has set
a new standard.’
14
Trenchard, Concise Dictionary, xii. Trenchard’s use of ‘cognate’ is questionable.
The way it works may be illustrated using our earlier examples:

λέγω v. (*)
λόγιος, α, ον adj. (λέγω)
λόγος, ου, ὁ n. (λέγω)
ἐκλογή, ῆς, ἡ n. (λέγω)
καταλέγω v. (λέγω)
τίκτω v. (*)
τέκνον, ου, τό n. (τίκτω)
The asterisk identifies the ‘key word,’ and related words are then cross-
referenced to it. The ‘key word’ is mostly a verb, but sometimes a noun or
other form where the verb derives from it rather than the other way round.
So far all seems well, though we notice that looking under the ‘key word’
gives no indication of the (possibly many) words related to it; that no clue
is given how the form τέκνον is derived from τίκτω; and that the seman-
tics of ἐκλογή (‘selection’) are not explained. However that may be, the
etymologies as such are not incorrect.
But what about this:
μῶμος, ου, ὁ n. (μέμφομαι) blemish, defect.
μέμφομαι v. (*) to find fault with, blame.
The semantics are attractive: if something has a defect you find fault with
it. But it is not enough to find plausible semantic links between words that
look roughly similar in form; the phonetic relationships must be explained.
What about the -φ- in μέμφομαι? Where has it gone in μῶμος? Can these
words really be related? A check reveals that no such connexion is made
in Beekes, Chantraine, Frisk, and Boisacq, and Chantraine even adds for
good measure, s.v. μῶμος: ‘Ce groupe de mots expressifs se distingue
franchement de μέμφομαι etc.’ So where has this ‘etymology’ come from?
It is certainly not from these reference books – which are noticeably absent
from Trenchard’s bibliography. Is it all his own invention? If we look
back to his statement quoted above, we find the beginnings of an explana-
tion: ‘The determination of cognate word groups … largely follows those
identified by J. Harold Greenlee.’ So Trenchard has drawn on an earlier
work by Greenlee.15 Sure enough, Greenlee makes this very connexion:
μῶμος … μέμφομαι.16
But what is Greenlee’s work based on? We have his own statement:
My aim has simply been to record in easily accessible form the ety-
mological analyses of words as they are generally accepted. My basic
15
Greenlee, Morpheme Lexicon.
16
Greenlee, Morpheme Lexicon, 87 (column list); 210–211 (s.v. μέμφομαι).
reference work (except for suffixes) has been the abridged edition of
the well-known Greek-English Lexicon of Liddell and Scott (LS). …
In some instances in which doubt remained, I have consulted other
works including the unabridged editions of LS.17
So Greenlee is based on the ‘abridged edition’ of LS. What is the date

of that work? Greenlee does not say and probably did not realise that it is
1843,18 and that the work was abridged from the first edition of Liddell and
Scott, also 1843.19 We see now what the problem is. Liddell and Scott’s
first edition was compiled before most of the nineteenth century work on
comparative philology, and presents many unscientific etymologies based
on guesswork and folk etymology which have since been jettisoned.20 But
they have gone from Liddell and Scott (1843) into Greenlee (1983) and
from there into Trenchard (2003). Neither Trenchard nor Greenlee thought
to consult any of the more recent etymological authorities mentioned ear-
lier, nor the excellent work of Jacques, who did consult them and covers
precisely the NT.21 The result is that Trenchard’s lexicon is replete with false
etymologies.
We can confirm that LS1 Abridged gives s.v. μῶμος: ‘μῶμος, ου,
ὁ, (μέμφομαι),’ and s.v. ΜΕΜΦΟΜΑΙ the headword printed in capitals
indicates a root word. The parent LS1 makes the connexion more cau-
tiously in a final note s.v. μῶμος: ‘(Prob. from μέμ-φομαι, as if μομφός.),’
and a similar note s.v. μέμφομαι (not in caps.): ‘(Hence prob. μῶμος.)’.
Marshall’s slight tinkering has made the connexion seem certain.
It would be possible to illustrate at length obsolete etymologies of this
kind in Trenchard’s lexicon, derived from Greenlee. I append below a selec-
tion from Trenchard in list form, with brief explanations. Some further
examples are added where the information is not so much wrong as mis-
leading or incomplete, and some where a false or fanciful etymology seems
to originate with Greenlee himself.
17
Greenlee, Morpheme Lexicon, viii. Note ‘generally accepted’: if only it were so! What
can ‘unabridged editions’ (pl.) of LS mean? Not more than LSJ (1940) surely. In his
ensuing discussion (viii–xiv) Greenlee shows no awareness of the changes in Liddell
and Scott’s lexicon over a hundred years and nine editions, nor that there might be other
authorities to consult on etymology (see esp. p. xii on ἄξιος and ἅμαξα). The book con-
tains no bibliography. Kilpatrick’s review in this journal (NovT 1985) said nothing about
Greenlee’s etymologies but much about the NT text.
18
LS1 Abridged.
19
LS1 (1843).
20
Cf. LS1 (1843), pp. viii–ix on sources. LS claim to have ‘departed widely from Passow’
(i.e., 4th ed., 1831) and ‘introduced a little Comparative Etymology’ drawing on Pott’s
Etymologische Forschungen (1833–1836).
21
Jacques, NT Words.
a. Obsolete etymologies
δάκτυλος (δείκνυμι)
= Greenlee 31, 168; LS1 Abridged, s.v. δάκτυλος; LS1, s.v. δάκτυλος;
similarly Et. Mag. (XII AD), with other ideas. Not recognised in Chantr.,
Frisk, Beekes.
δύσκολος (κωλύω), κολάζω (κωλύω), κολοβόω (κωλύω) [+ more]
Greenlee 203 (cf. 38, 75, 77) lists all, including κωλύω, as derivatives
of κόλος ‘curtailed.’ Similarly LS1 Abridged; LS1, s.v. κωλύω. The con-
nexion of κόλος with δύσκολος, κολάζω, κολοβόω is right, but not
with κωλύω. See Chantr., Frisk, Beekes, s.vv.; Chantr. calls the etymol-
ogy of κωλύω ‘inexpliquée.’
θάλασσα (ἅλας)
= Greenlee 60, 154 (ἅλς); LS1 Abridged; LS1, s.v. θάλασσα. Not rec-
ognised in Chantr., Frisk, Beekes.
κοράσιον (κείρω)
= Greenlee 75, 200; similarly LS1 Abridged; LS1, s.v. κόρος, derived
from κείρω. For the age of this etymology see LSJ, s.v. κόρος (B) ‘boy’:
‘Acc. to Eustath. 582.20, al. [XII AD] from κείρω.’ Not cognate in
Chantr., Frisk, Beekes. κόρη/κόρος is usually derived from the same
root as κορέννυμι. The immediate source of κοράσιον is κόρη; this
would have been more helpful information.
κραυγάζω (κράζω)
= Greenlee 76, 204; similarly LS1 Abridged, s.vv. κραυγάζω, κραυγή;
LS1, s.vv. κράζω, κραυγάζω, κραυγή. Not cognate in Chantr., Frisk,
Beekes.
κῶμος (κώμη)
= Greenlee 78, 205; LS1 Abridged; LS1, s.v. κῶμος. Chantr. and Frisk
doubtfully entertain the possibility of a link, very far back via a root *kei-
as in κεῖμαι, but Beekes rejects it; he suggests κῶμος is a ‘Pre-Greek
word.’ The linking of κῶμος with κώμη is very old: cf. Et. Mag., s.v.
κῶμος, and Isidore of Seville, Etymologies VIII.vii.6.
ποταμός (πίνω)
= Greenlee 107, 227; LS1 Abridged, s.v. ποταμός; similarly LS1, s.v.
ποταμός (in a lengthy note). Not countenanced in Chantr., Frisk, Beekes.
χάρτης (χάραγμα)
Similarly Greenlee 139, 256, as derivatives of χαράσσω; LS1 Abridged;
LS1, s.v. χάρτης. See Chantr., Frisk, Beekes: χαράσσω and its deriva-
tive χάραγμα are unconnected with χάρτης. Beekes favours the old
hypothesis of an Egyptian loan.
b. Misleading or incomplete information

αἴνιγμα (αἶνος) riddle … αἶνος (*) praise
= Greenlee 4, 151; cf. LS1 Abridged, LS1, s.vv. αἴνιγμα, αἰνίσσομαι.
There is a link between αἴνιγμα and αἶνος, but only via αἰνίσσομαι,
‘speak in riddles,’ formed from αἶνος in its sense of ‘meaningful

words’ (‘praise’ is a separate track); αἴνιγμα then was a back-forma-
tion from αἰνίσσομαι. See Chantr., Frisk, Beekes, s.v. αἶνος. Greenlee
dropped out αἰνίσσομαι because it is not in the NT. But the user of
Trenchard is likely to wonder how αἴνιγμα can be related to αἶνος.
ἀργός, ή, όν (ἔργον) unemployed, idle. …
In fact ἀργός < ἀ- + ἐργ-. Cf. Greenlee 19, 147, 185 correctly. We
need the ἀ-privative omitted by Trenchard to understand the semantics.
νῆστις (ἐσθίω) not eating. …
The same problem only more so. To make any sense of the connexion
we need to know that νῆστις is made up of νη- ‘not’ + ἔδ-, the root
of ἔδω and ἐσθίω, as correctly stated by Greenlee 89, 175.
εὐδία, ας, ἡ n. (Ζεύς) fair weather
Cf. Greenlee 55 (εὐδία, ἡ … εὖ Ζεύς, Διός, ὁ -α, ἡ). We see here the
hazards of simplification. Yes, εὐδία and Ζεύς are ultimately related,
but we have to go far back to find the link: they derive separately from
an ancient IE root *diu-/dieu-. Thus Beekes, s.v. εὐδία: ‘Compound …
of εὖ and the zero grade of an old word for ‘day, heaven’ (see Ζεύς),
thus εὐ-δίϝ-ᾱ’; and s.v. Ζεύς: ‘The old IE word for ‘heaven’ and the
name of the god of heaven and of daylight … Ζεύς … from IE *diêus.’
To give simply ‘εὐδία (Ζεύς)’ is misleading, but fuller information
would be of doubtful value to most users of the lexicon.
ὄψις (ὁράω), πρόσωπον (ὁράω) [+ many]
= Greenlee 219. But ὄψις, -ωπ- etc. are related to fut. ὄψομαι, aor.
ὤφθην (< *okw-), not the pres. ὁράω, which is from a different root
*uer-. Connecting ὄψις with ὁράω is not etymology but pedagogical
fiction. LS1 Abridged knew better: see ὄψις (ὄψομαι), πρόσωπον
(πρός, ὤψ).
ποῦ (*), ὁποῖος (ποῦ), ὅπως (ποῦ), ποταπός (ποῦ, ἀπό), πότε (ποῦ)
[+ others]
These are indeed related, but it is misleading to tag ποῦ as the key word,
as if the source of the others; they are all derivatives of one stem πο- < IE
*kwo-. Cf. Greenlee 230, all under ‘πός an assumed interrog. pron., who?’;
similarly LS1 Abridged; LS1; Chantr., s.v. πο-; Beekes, s.v. πόθεν. Again,
this is likely to be too much information for the user of a smaller lexicon.
The second element of ποταπός has nothing to do with ἀπό as asserted
by Trenchard: see Chantr., s.v. πο- B.4.; Beekes, s.v. ποδαπός.
c. Fanciful etymologies
δεσπότης (δέω)
= Greenlee 32, 170. Not in LS1 Abridged; LS1, s.v. δεσπότης. See
Chantr., Frisk, Beekes, and LSJ for the real etymology: < *dems- + pot-
(‘master of the house’). Jacques, NT Words, 32 correctly ~ δέμω. It
looks as if Greenlee simply took an unlucky guess (and Trenchard
believed him).
μηδαμῶς adv. (μή, δέ, ἐγω) [ἐγω sic]

= Greenlee 85: μή δέ ἐγώ, ἐμοῦ -ως. The third element is nonsense
made up by Greenlee. See Chantr.; Beekes, s.v. οὐ. Another of a simi-
lar kind is: ἑκάστοτε (ἕκαστος, ὅς, τέ) = Greenlee 41.
πῶς (ὅς)
= Greenlee 113. They are unrelated: πῶς < *kwo-; ὅς < *yo-. See
Chantr., s.vv. πο-, ὅς; Beekes, s.vv. πόθεν, ὅς (IE form *h1i-o- with
laryngeal). Again: πόθεν (ὅς) = Greenlee 105.
ὑετός (ὕδωρ)
Similarly Greenlee 132, 251, as derivatives of ὕω (only ὑετός is), but
not LS1 Abridged, s.vv. ὕδωρ, ὑετός, ὕω. A guess by Greenlee. Attrac-
tive as it may seem to connect the word for ‘water’ with the words for
‘rain,’ the history of the roots does not permit it. ὕδωρ descends from
an old IE noun (cognate with Engl. water) unrelated to the root of ὕω:
see Beekes, etc., and LSJ.
d. A mistake
νίπτω (*), ἄνοια (νίπτω), διάνοια (νίπτω), νοέω (νίπτω), νόημα
(νίπτω), μετάνοια (νίπτω), πρόνοια (νίπτω) … [10 more]
Could the words for ‘mind, thought, think’ etc. really be related to the
word for ‘wash’? Unlikely, surely. A reason for the slip-up may be
that in Greenlee (89) νίπτω … νίζω are followed at once by νοέω …
νοῦς in the alphabetical list. The wonder is that neither the author nor
anyone in the whole editorial process noticed it.
3. NEWMAN, A CONCISE GREEK-ENGLISH DICTIONARY OF THE NT (2010)
Newman’s work was first published in 1971 and has served a generation
as the best small NT lexicon available. As well as being sold separately,
it has usually been published with the UBS NT text, increasing its circu-
lation enormously. It was also the base from which Louw and Nida worked
to create their lexicon arranged by semantic domains. In a short paper
published in 2004 Newman wrote, ‘A revision is long overdue, and I hope
someone will accept the challenge.’22 Newman himself has accepted the
challenge and produced this new edition.
There are some changes, as the preface explains.23 Besides revising the
definitions, Newman has introduced what he calls simply ‘cross-referencing.’
This is how he describes its main features:
22
Newman, ‘Reflections,’ 93.
23
‘Preface to the Revised Edition,’ pp. vii–viii.
Where cross-referencing is used between more than two Greek words,

the simple verb usually serves as the primary entry with the other forms
keyed to it:
ἀγαπάω (ἀγάπη, ἀγαπητός)
ἀγάπη, ης f (ἀγαπάω)
ἀγαπητός, ή, όν (ἀγαπάω)
As a didactic, mnemonic, and phonetic device, compound words are
divided by small upright lines according to their basic components:
ἀντι|καθ|ίστημι and ἀν|εξ|ιχνίαστος.
These are in fact word derivations or etymologies along the same lines
as those in Trenchard. But instead of Trenchard’s ‘key words’ marked
by ‘(*),’ a full list of related words is given under the verb (or alternative
if the verb does not occur), with cross-references back to it. So we find
in Newman ‘λόγιος, α, ον (λέγω); λόγος, ου m (λέγω); ἐκ|λογή, ῆς f
(λέγω),’ as in Trenchard, but under λέγω a list of all the related words, in
this case a very long list of 64 items, with ‘|’ between their components;
meanings follow after the list. This certainly solves the problem mentioned
above, of knowing what words belong to a given verb or other key word,
but at the cost of some unwieldiness. It also means that all the relatives
and derivatives of πρό, for example, are listed under the derived verb πρω-
τεύω, while πρό itself appears as: ‘πρό (πρωτεύω),’ a rather odd result,
having the effect of reversing the order of derivation.24 But this is not the
issue here.
Newman mentions no source for his etymological information, nor
does he include any bibliography. It does not take long to discover that he
has relied heavily on Trenchard: in fact he seems to have used that lexi-
con almost exclusively. The only other possible source is the faithful LS1
Abridged, and instances are rare (see below). In some cases Trenchard’s
etymologies are dropped, but Newman has made up for them by inventing
some of his own. I do not see any sign that Newman went back to Trench-
ard’s fons et origo, Greenlee. The result of all this is that Trenchard’s errors
are not only reproduced but compounded in Newman. Illustrations are
ready to hand from those noted above in Trenchard.
Of the eight obsolete etymologies above (2.a), three (κοράσιον, κραυ-
γάζω, ποταμός) are dropped, but five reappear in Newman thus:
δάκτυλος (δείκνυμι)
δύσκολος (δῦσ- [inseparable prefix] = hard, difficult + κωλύω) [δῦσ- sic]
θάλασσα (ἁλίζω). Cf. ἁλίζω (αἰγι|αλός, ἅλας, ἁλιεύς … θάλασσα …)
24
Similarly, e.g., ἔργον (ἐργάζομαι); θάρσος (θαρσέω); ὁμοῦ (ὁμοιόω).
κῶμος (κώμη)
χάρτης (χάραγμα)
Others of this kind could be added, that is, obsolete etymologies that
descend from LS1 Abridged via Greenlee and Trenchard to Newman. See
γραώδης, ὁμιλέω, ποταπός, ὑγρός, ψῆφος.
Trenchard’s lead is followed also in the examples in the second cate-
gory (2.b), except for an improvement in αἴνιγμα, which omits the refer-
ence to αἶνος and now reads: αἴνιγμα, τος n (αἰνίσσομαι = speak in
riddles). But εὐδία has got decidedly worse:
εὐ|δία, ας f (εὖ + Δία [from gen. of Ζεύς] = godlike, excellent) fair
weather
It appears that Newman thinks the second element of εὐδία (i.e., δία) is
actually the accusative of Ζεύς, turned somehow into a feminine noun in
-ίᾱ. It is probable that consultation of LS1 Abridged played a part in this.
Compare these entries in LS1 Abridged, each of which has contributed
something:
εὐδία, ἡ, (εὔδιος) fair weather
εὔ-διος, ον, (εὖ, Διός gen. of Ζεύς) calm, fine, clear
Διο- διο- in compos. means sprung from Jove … or from the gods,
hence excellent, god-like.25
All the fanciful etymologies in 2.c invented by Greenlee are also carried
over without change. We can be grateful at least that Newman did not fall
for ἄνοια (νίπτω), etc. But his own inventions are bad enough. These three
will suffice as examples:
ἀ|πατάω (πατέω = walk, ἀ|πάτη, ἐξ|α|πατάω, φρεν|α|πάτης) deceive,
lead astray
ἔξ|ωθεν (ἐκ + ὠθέω = push)
καθ|αίρω (κατά + αἴρω, καθ|αρός) clean, prune (of branches)
These are not found in Trenchard or Greenlee, and certainly not in

LS1 Abridged, so I conclude they are Newman’s own. There is no need to
labour the flaws in them.
The new lexicons of Trenchard and Newman represent a step back-
wards, not an advance. If etymological material is to be introduced, it
needs to be done competently and on the basis of reliable authorities. It
is not for amateurs. If the objection is raised that most of Trenchard’s and
25
Another compiled with the help of LS1 Abridged is κατήφεια, plus an egregious mis-
spelling of φάος.
Newman’s material is correct and, say, ten percent is wrong, one might
respond that we would hardly accept a Greek grammar that had ten per-
cent of the forms wrong. As for the question of whether inclusion of such
material is worth the attempt, that will be considered in 5 below.
4. DANKER, THE CONCISE GREEK-ENGLISH LEXICON OF THE NT (2009)
Danker’s motivation in producing a shorter lexicon of his own is not

revealed, but one might suspect that having seen the semantic content of
his monumental BDAG (2000) reused in Trenchard’s lexicon, he deter-
mined to enter the field himself. This new work is said not to be a revision
of any previous small lexicon nor, surprisingly, an abridgement of BDAG,
though Danker does say he has used what he calls ‘common coin.’26 Danker,
it appears, has presented the NT meanings anew in his own way, at the end
of a lifetime of experience as a NT lexicographer. On this aspect of the
lexicon there is no more to say here. But there is something new in the
work. In his own words:
Etymological information following most headwords is brief (ἄγγε-
λος), especially for words with complex histories or origins not readily
discernible, in the hope that users will recognize their status from vari-
ous evaluations about probability of derivation and will be thankful for
being spared theoretical crossfire. The etymological feature, enclosed
in square brackets, includes a gloss or translation equivalent, ordinarily
one word, for terms outside the New Testament vocabulary (τρέπω
in ἀποτρέπω). When no translation is offered, users can conclude that
a headword (γαμίζω) is from a base (γάμος) appearing in the present
volume as a headword and with further etymological detail.27
While Danker is frank about previous NT lexicons from which he has

benefited, he makes no mention of any source for his etymologies. Danker
was too good a scholar not to be aware of the major reference books
noticed before, but he is simply silent on the matter.28 His explanation of
what his etymological information will consist of is also quite vague. What,
then, will we find? To begin with, we find many of types a. and b., as was
to be expected. There are usually no problems here. But Danker has gone
26
Danker, Concise Lexicon, p. v.
27
Danker, Concise Lexicon, v–vi. See also the diagram on p. vii indicating the different
parts of an entry: part 2 is labelled ‘Etymology or derivation in brackets.’
28
Chantraine, Frisk, and Boisacq are listed in the bibliography of BDAG; but there is no
bibliography in the Concise Lexicon. Beekes, EDG (2010) came too late to be used.
much further, into etymology in the fullest sense, and introduced many ety-
mologies of type c. Here some criticism is hard to avoid.
The criticisms are not about fine points of etymology, as if Danker were
compiling an etymological dictionary. They arise from looking at the work
on its own terms, that is, as a smaller NT lexicon that will be used by
students and others inexperienced in Greek etymology. In that light there
are many problems with method and consistency, and the potential impact
of the information on the user. There are also inaccuracies, though not as
serious as in the other two lexicons. My illustrations are divided into three
groups according to the etymology of the words, regardless of how Danker
labels them. The etymological information is given by Danker in square
brackets immediately after the headword.
a. Words with secure Indo-European etymologies

Let us begin with entries of this type:
1 μήτηρ, τρός, ἡ [IE] mother, in central sense Mt 1:18 al.
2 τίς [IE, cp. Lat. quis, quid]
3 πάσχω [IE, cp. πένθος]
4 κάλαμος [IE, cp. Lat. calamus ‘reed, cane’]
The label ‘IE’ is helpful, indicating clearly the ultimate origin, but the
ensuing information varies. In 1 there is nothing more: no cognate such
as Latin mater. In 2–4 the notation ‘cp.’ (= ‘compare’) is used imprecisely,
covering a cognate proper in 2, a related word in Greek in 3, and in 4 a
word that is actually a loanword in Latin from Greek, not a cognate (the
Lat. cognate of κάλαμος is culmus). Further, it is fair to ask what the
user is supposed to make of this. The relationship with Lat. quis is hard
to understand without more knowledge of how the sounds developed
(< IE *kwi-s); similarly puzzling to the beginner is πάσχω ~ πένθος, which
are related in a complicated way not readily discernible; and 4 is mislead-
ing. In all this there has been no explanation of the terms to guide the begin-
ner, nor mention of reference works that might be consulted.
Next a sample of entries in which ‘Skt.’ (= ‘Sanskrit’) is mentioned,
without ‘IE’:
5 πατήρ [Skt. assoc.]
6 βάλλω [Skt. assoc.; of forceful action (esp. of missiles) ‘throw,
cast, hurl’, then in more subdued sense such as ‘put, place, lay’]
7 ἀγρός [Skt. = Lat. ager ‘piece of land’]
8 πούς [cp. Skt. pat, Lat. pes]
9 φέρω [Skt. bhárati ‘bear’, cp. Lat. fero ‘bear’]
10 ἑκατόν [Indo-Germanic origin = Lat. centum]
Entries like this are quite common, but the interpretation is not clear.
What does ‘Skt. assoc.’ etc. imply? In fact it only means that there is a
cognate in Sanskrit. But the beginner may not realise that a Sanskrit cog-
nate only indicates a common origin in IE and may be misled into thinking
that Greek has taken the word from Sanskrit; or worse, that the connexion
with Sanskrit, which is some thousands of years earlier, is still present to
the minds of speakers of NT Greek. In 7, the note seems to imply that ἀγρός
is itself a Sanskrit word (it is a cognate: Skt. form ájra-), and ‘= Lat.’ is not
a good way of saying ‘cognate with Latin.’ Other problems here are: in 5,
why no cognate, when so many are ready to hand (e.g., Lat. pater, Engl.
father)?; in 6, most of the note is about the (mainly later) semantic history;
in 9, the Skt. form given is the third person singular; in 10 the term ‘Indo-
Germanic,’ obsolete in English, is suddenly introduced and never explained.
In all these entries, the missing ‘IE’ would have been helpful.
Next let us see three numerals and another recurrent formula, ‘etym.
complex’:
11 εἷς, μία, ἕν [a numerical term; ‘one’] one …
12 δύο [Skt. assoc.]
13 τρεῖς [etym. complex]
14 ὅς, ἥ, ὅ [etym. complex]
15 πᾶς [etym. complex]
16 χείρ [etym. complex]
All of these have the most secure of IE etymologies, but there is no

indication of that. In 11, the information tells us nothing at all;29 in 12, we
have again the vague ‘Skt. assoc.’ Then comes ‘etym. complex,’ without
anything else, a formula that suggests that the etymology is uncertain or
obscure.30 In fact Danker uses it for etymologies that are difficult to follow
or summarise in a short space and would need the introduction of hypo-
thetical IE forms, which he seems reluctant to touch. In the case of the
numerals at least, it would have been interesting to see some of the cog-
nates that are spread through the IE languages (e.g., δύο : Engl. two). All
this tends to indicate the impossibility of carrying out the task in the way
Danker has set himself to do it.
There are many other unsatisfactory entries of various kinds that might
be added in this section; I note just a few with very brief explanations.
29
Cf. BDAG, s.v. εἷς: ‘a numerical term, “one” (Hom.+).’
30
This formula is the most frequent, but there are variations: ‘of complex origin’ (e.g.,
ὁδός); ‘etym. unclear’ (ξέστης); ‘etymology/ etym. uncertain’ (ὁμείρομαι + many);
‘etym. obscure’ (ὄχλος); ‘origin debated’ (ὀθόνη); ‘derivation unknown’ (οὐ).
17 βοῦς [cp. βόσκω]

Ultimately they may be derivatives of the same IE root, but the IE ety-
mology of βοῦς is clear and interesting (: Lat. bos, Engl. cow).
18 δεσπότης [cp. δεσπόζω ‘be lord/master’]
See 2.c above.
19 δέω [not to be confused with the homograph δέω ‘lack, stand in
need of’, with which the impersonal δεῖ is associated] bind
Why the concern? Is the NT student likely to confuse them? The ‘hom-
ograph δέω’ is in fact always mid. δέομαι ‘ask, beg, pray’ in NT. Noth-
ing on the IE etymol. (cognates in Skt., Hitt.).
20 δηλαυγῶς [δῆλος, αὐγή; ἄγαν φανερῶς = ‘very openly’, states
Hesychius] very clearly Mk 8:25 v.l. for τηλαυγῶς.
Suddenly Hesychius. The Hesychius gloss is quoted in BDAG (2000),
helping to confirm the reality of the rare form. The beginner consult-
ing Danker’s smaller lexicon will be baffled.
21 ἥλιος [cp. Lat. sol]
The relationship is real but very complicated. Without more informa-
tion nothing is gained.
22 τέσσαρες
sic: why not at least ‘IE : Lat. quattuor’?
Finally here the question of sources must be raised. This would not
matter if we could assume (as I did at first) that Danker had used Chan-
traine and Frisk, but it became apparent that another source was in play,
and that was clearly Boisacq (1916), mentioned earlier. The problem with
this is simply that Boisacq’s work, though good in its day, is very obsolete
now and not a reliable authority. This entry is clear evidence of the use
of Boisacq:
23 ἱμάς [ἱμαίνω ‘tie’]
The verb *ἱμαίνω is a hypothetical form (Danker has omitted the aster-
isk), postulated, alongside *ἱμανάω, as a step before the rare noun ἱμανήθρη.
*ἱμαίνω is not mentioned in Chantraine, s.v. ἱμάς, and is buried deep in
Frisk, s.v. ἱμάς. But Boisacq’s entry begins: ‘ἱμάς … présuppose un *ἱμαίνω
“lier, enchaîner”.’ Boisacq is the obvious source. It would have been better
to leave this alone than try to compress the complicated IE links of ἱμάς.
Two more (and others below) confirm that Boisacq was a source:
24 ἱλαρός [cp. Hom. impv. ἵληθι ‘be favourably disposed!’]
Cf. Boisacq, s.v. ἱλαρός: ‘voy. s.v. *ἵλημι’; then s.v. *ἵλημι: ‘*ἵλημι
‘être propice, favorable’, hom. impér. ἵληθι.’ A Homeric imperative is
not helpful to the NT student (cf. Luke 18:13 ἱλάσθητι). ἱλαρός
derives from the same root as ἱλάσκομαι and ἵλεως (Beekes, s.v. ἱλά-
σκομαι).
25 ἴδιος [opp. Hom. δήμιος ‘public’]

Cf. Boisacq, s.v. ἴδιος: ‘… (opp. δήμιος γ 82. δ 314).’ This is no help
with the etymology of ἴδιος, which is complex.
b. Loanwords
Loanwords seem easy but there are traps for the unwary, and informa-
tion needs to be presented in a way that does not mislead the uninformed.
Even the simple statement in 27 below could be misinterpreted: is κῆν-
σος a borrowing of Latin census or the other way round (cf. 28)? Or are
they cognates? The beginner does not necessarily know. A more serious
issue is the age of the borrowing, which in Greek can range from Myce-
naean (32) to around the time of the NT itself (27). Without an indication
of this, the beginner may make false assumptions about what speakers of
the language were conscious of when they used such words.
26 ἀρραβών [Heb.; cp. modern Gk. ἀ. ‘engagement (ring)’]
Danker has followed Boisacq: ‘< hébr. ‘erabon.’ It would be more
accurate to say: Semitic loanword, cf. various forms in Heb. etc. ‘En
grec, il a pu passer par l’intermédiaire du phénicien,’ says Masson.31 In
Greek literature from IV BC on, but may be as old as V BC: see inscr.
in LSJ Suppl. (1996). Is the modern usage relevant? (All NT exx. show
a figurative or transferred sense.)32
27 κῆνσος [Lat. census]
First evidence in I BC. This and the grecised form suggest its origin
would not necessarily have been recalled each time it was used in the NT.
28 κράβαττος [loanword in Lat. grabatus ‘pallet’]
Danker’s note informs us of borrowing into Latin, also implying Greek
as the source. κράβαττος is from a third source (Macedonian suggested,
but doubtful), borrowed independently into Greek and Latin. Evidence
begins in III BC. If Mod. Gk. was to be mentioned this was the place to
do it: κρεβάττι is still the popular word for ‘bed.’
29 νάρδος [Skt. loanword]
Not accurate, rather: Greek < Semit. < Skt.
30 ξέστης [etym. unclear]
Actually from Lat. sextarius with adaptation along the way: see Chantr.,
Beekes. Danker may have relied on MH, Grammar, II, 155, referred to
in BDAG.
31
Masson, Emprunts sémitiques, 31.
32
Trenchard and Newman manage to imply it is an actual Hebrew word in the NT: Trench-
ard: ‘ἀρραβών, ῶνος, ὁ n. Heb. first installment …’; Newman: ‘ἀρραβών, ῶνος m
(Hebrew word) pledge. …’ The statement about modern usage appeared previously in
MM, s.v., and in Abbott-Smith, Manual Greek Lexicon.
31 σουδάριον [Lat. loanword sudarium (sudo ‘to sweat’), a personal

article] a cloth
Is Lat. sudarium the loanword or the source?
32 χρυσός [Semitic loanword]
Borrowed so early, before 1000 BC, that no one would remember its
origin. Phoenician is taken to be the likely source (Masson, Emprunts
sémitiques, 38).
c. Words of unknown etymology, ‘Pre-Greek,’ or formed within Greek
A large proportion of the Greek vocabulary consists of words for which

no etymology can be established. There have always been attempts to
find etymologies, but most of the suggestions have come and gone, and
the etymology of many words remains unknown. Another large group is
formed by words that are not Indo-European but were probably adopted by
the Greeks at a very early time from a non-IE substrate language spoken
in Greece and Western Asia Minor.33 Then there are words of which the
inner-Greek formation can be traced even if the ultimate origin cannot.
Danker’s handling of these types is haphazard.
33 ἄγγελος [Skt. assoc., cp. ἀγγαρεύω]
< Boisacq. The connexion with Skt. is abandoned in Frisk, Chantr.,
and Beekes. A connexion between ἄγγελος and ἄγγαρος, -εύω is also
not made. Both are loans, but the sources are uncertain.
34 ἀριστερός [opp. δέξιος (accent sic) ‘right’]
A formation within Greek, < ἄρισ-(τος) + -τερος = ‘the better (side),’
with euphemistic value. Danker’s information is pointless, though it
had a purpose originally in BDAG (q.v.). BDAG’s ‘(opp. to right)’ >
Trenchard: ‘(opp. to right)’ > Newman: ‘(opposite right).’
35 θησαυρός [orig. uncertain and apparently primitive]
< Boisacq: ‘θησαυρός. … Prob. composé primitif …; origine inexpli-
quée.’ Boisacq’s ‘primitif’ refers to the compound; ‘primitive’ will con-
vey nothing to the beginner. Cf. Beekes: ‘No etymology, but probably
a technical loanword, without a doubt from Pre-Greek.’
36 καθαρός [etym. in dispute (note: no connection with κατά)]
Danker seemed to be warning against a non-existent error, but now we
see it in Newman (3 above). There is no known etymology for καθαρός.
Beekes thinks Pre-Greek.
37 σελήνη [IE]
Formation within Greek, < *σελασ-να. Etymol. of σέλας obscure, not
established as IE.
33
See esp. Beekes, EDG, xiii–xlii.
38 χάρτης [orig. uncertain, cp. χαράσσω ‘engrave’]

Danker has fallen into the same trap as Trenchard (see 2.a above).
These words are unrelated. Where did Danker get this obsolete ety-
mology from? It is not even in Boisacq. Did he happen to glance at
Trenchard?
Danker’s foray into etymology is not a disaster, as are those of Trench-

ard and Newman, but it is ill-advised. The main problem, apart from inac-
curacies, is the randomness of the information given. Danker simply chooses
something, without any consistent system. One can see the difficulty that
he faced: to summarise in a few words a complex and highly variable body
of data on each word. But one could wish he had defined his system and
objective more clearly, for himself as much as for us. Part of doing so would
have been to decide what purpose the etymological information serves and
how it would impact on the user.
5. CONCLUSION
It looks as if there is a trend under way to reintroduce ‘etymologies’ into

NT lexicons. The idea has already caught on and gained momentum in less
than ten years. It is likely to continue and grow in strength. There is also
the likelihood, or rather certainty, that etymological material, from these
three lexicons or others yet to come, will make its way into electronic
resources. All this will be accompanied by the usual copying of material,
of whatever quality, and without scrutiny, from one source to another, as
we see already in Trenchard and Newman. Mistakes, once in the tradition,
will be extremely hard to dislodge.
At this point in the process some reconsideration is desirable. First,
there is the question of what the etymological information is for. What
need does it serve and is it useful? The purpose of information of type a.
is presumably to help the beginner to learn vocabulary by seeing deriva-
tions and linking new words with those already known. This has obvi-
ous pedagogical value and may be useful even for the non-beginner. The
value is clear in Abbott-Smith’s presentation of, for example: ἀπιστέω
(< ἄπιστος), ἀπιστία (< ἄπιστος), ἄ-πιστος (< ἀ- neg., πιστός); Danker’s
is the same. Trenchard and Newman on the other hand reduce the value
by linking the three words to verbs, Trenchard to πείθω and Newman
to πιστεύω: the beginner has to do more work to get anything out of it.
So even the simplest type a. can run into problems if it is not handled
carefully.
The problems with type b. are well illustrated by the same group. When
we look to πιστός we find it linked with πείθω in Abbott-Smith, Danker
and Trenchard (and with πιστεύω in Newman). Is this useful? The begin-
ner, and non-beginner too, will struggle with both the phonetics (πισ-τός
< *πιθ-τός, zero grade of root) and the semantics (how does ‘faithful’
relate to ‘persuade’?). Our earlier instance of τέκνον (τίκτω), given like
that in all four lexicons, is another illustration. In the light of these exam-
ples I personally have doubts that even basic information of types a. and
b. is worth including in a lexicon. If it is to be included, thought must be
given to how to present it in a way that is actually helpful and not mis-
leading or puzzling.
Type c., etymology proper, is even more problematic. There is a practi-
cal difficulty here to begin with, that is, the near impossibility of summa-
rising in a short space the etymology of any word, given the complications
and unpredictable features that constantly turn up, as we have seen so well
illustrated. More serious still is the problem inherent in the whole exercise,
that the beginner, and even the advanced NT scholar, has no framework of
knowledge in which to fit the information and interpret it. A full course of
training is barely enough to prepare one for it. For this reason I consider
the exercise pointless and potentially harmful. But if such material does
continue to be included, it is obvious that it must be prepared competently
and on the basis of the latest etymological authorities, not simply drawn
from predecessors.
There is a further reason for caution about inclusion of etymological
material, and that is the risk of fostering the use of etymology to draw
conclusions based on the ‘original’ or ‘real’ meanings of words. The ‘ety-
mological fallacy’ endures despite repeated warnings from scholars, and
is not confined to students and preachers, or the biblical field. Even the
seemingly innocuous statement ‘ἐκκλησία [ἐκ, καλέω],’ as in Danker,
will continue to feed one of the classics of false exegesis: ἐκκλησία =
‘those called out’; etymologies of type c. offer still more alarming pos-
sibilities. The trend towards inclusion of etymologies in NT lexicons may
have the unintended consequence of adding fuel to this fire.34
This article has had an entirely negative aspect, but that is unavoidable.
There is good in these lexicons, but the mistakes in them vitiate the good.
If some of the etymologies cannot be trusted, then all of them must be
treated with caution. The beginner simply does not have the experience
34
For good discussions see Louw, Semantics, 26–31; Silva, Biblical Words, 44–51;
Carson, Exegetical Fallacies, 28–33. I thank Steve Walton for reminding me of this
important issue.
to sort them out. There is however a positive suggestion with which to

end. The massive etymological reference works mentioned throughout this
discussion are difficult even for the experienced scholar. What we need,
and have long needed, is a concise etymological dictionary of Greek, one
that summarises the information in a targeted, efficient way for the non-
expert. It could cover even basic derivations within Greek, as is already
done in the big works, as well as the difficult questions of ultimate origin.
If this challenge could be met, NT lexicographers could refer to or draw on
this work and be relieved of the burden of dealing with something outside
their competence.
21
THE ONOMA RULE
2014
Abstract
The paper presents an unnoticed ‘rule’ governing the presence or absence of

the article with ὄνομα in naming-constructions of the type ‘the name of x is N.’
It arises from the survival of an archaic pattern with the dative and no article,
alongside a newer pattern with the genitive and the article. The ‘rule’ is shown to
operate not only in the LXX and NT but in ancient Greek generally, from Homer
to Koine Greek. Observance of the ‘rule’ indicates the LXX translators’ and the
NT authors’ familiarity with idiomatic Greek. The ‘rule’ can also contribute to
resolving textual questions, notably Luke 1:63.
1. THE LXX
The grammatical ‘rule’ presented here was first noticed in the LXX
Pentateuch, and then found to operate elsewhere in the LXX, in the NT, in
Koine Greek generally, and in Classical Greek, where it originates. As far
as I can discover, it has not been noticed before.1 Aside from its intrinsic
interest, the ‘rule’ can be used as a proof of natural Greek and as a tool of
textual criticism.
I begin with the example that pointed the way to this discovery:
Exod 15:3 κύριος συντρίβων πολέμους, κύριος ὄνομα αὐτῷ.
The Lord is (the) shatterer of wars, the Lord is his name.
:‫יהוה אישׁ מלחמה יהוה שׁמו‬
In κύριος ὄνομα αὐτῷ we have a ‘possessive’ dative αὐτῷ, meaning

literally: ‘Lord (is) name to him.’ The Hebrew ‫ שׁמו‬is of course somewhat
different, with a possessive suffix, meaning ‘his name.’ While the render-
ing is hardly strange as Greek, a comment by Wevers shows that some
puzzlement has been felt: ‘Exod renders ‫ שׁמו‬by ὄνομα αὐτῷ, though one
1
Cf. KG, Grammatik, II.1, 45, 417; Schwyzer, Grammatik, II, 66, 86; Smyth, Grammar,
§1478; Mayser, Grammatik, II.2, 113.
might have expected a genitive pronoun rather than a dative of posses-

sion.’2 In a recent paper Perkins is similarly puzzled, remarking that ‘the
translator is quite careful normally to render the Hebrew-suffixed pronoun
by a genitive form,’ going on to suggest that the rendering perhaps reflects
an earlier passage, Exod 3:13 τί ὄνομα αὐτῷ;3
But a search of all occurrences of ὄνομα in the Pentateuch shows that
the combination with the dative is not uncommon.4 Besides three with the
pronoun αὐτῷ (Exod 3:13; 15:3 quoted above, and Gen 2:19 τοῦτο ὄνομα
αὐτῷ), there are no less than 28 other instances, by my count, of ὄνομα
with dative of the possessor of the name.5 Examples:
Gen 4:19 ὄνομα τῇ μιᾷ Ἀδά, καὶ ὄνομα τῇ δευτέρᾳ Σελλά.
:‫שׁם האחת עדה ושׁם השׁנית צלה‬
Gen 22:24 καὶ ἡ παλλακὴ αὐτοῦ, ᾗ ὄνομα Ῥεημά.
‫ופילגשׁו ושׁמהּ ראומה‬
As can be seen, the choice of the dative case here is not dependent
on the Hebrew, and this is true of all the instances: the Hebrew has either
‫ שׁם‬with the possessive suffix (‫שׁמו‬/‫ )שׁמהּ‬or ‫ שׁם‬in the construct state +
possessor of the name.
The search confirms that the construction with genitive of the pos-
sessor occurs too, as expected, though less often than with dative – as
Wevers and Perkins might not have expected: I find only 12 instances
(with ὄνομα in the singular).6 Here is one of this type, side by side with
one of the other:
Exod 18:3 ὄνομα τῷ ἑνὶ αὐτῶν Γηρσάμ … 4 καὶ τὸ ὄνομα τοῦ δευ-
τέρου Ἐλιέζερ
‫ ושׁם האחד אליעזר‬4 … ‫שׁם האחד גרשׂם‬
2
Wevers, Exodus, 228.
3
Perkins, ‘The Lord is Warrior,’ 129–30.
4
All LXX and NT searches were made using the program ‘Accordance.’ LXX refs. were
confirmed in the Göttingen ed. where available, otherwise Rahlfs.
5
Full list of ὄνομα + dat. (total 31): Gen 2:11, 13; 2:19; 4:19, 19, 21; 10:25, 25; 11:29,
29; 16:1; 22:24; 24:29; 25:1; 28:19; 29:16, 16; 36:32, 35, 39, 39; 38:1, 2, 6; Exod 1:15;
3:13; 15:3; 18:3; Num 11:26, 26; 25:15. [I do not include Gen 26:33. The Göttingen
ed. has διὰ τοῦτο ἐκάλεσεν ὄνομα τῇ πόλει Φρέαρ ὅρκου, which is not the onoma-rule
pattern; Rahlfs’s text without ἐκάλεσεν is. See Wevers, Genesis, 415: ‘Admittedly, the
verb is somewhat awkward, but the lectio difficilior must here be original.’ In my opinion,
the onoma rule resolves the question in favour of Rahlfs’s reading.]
6
Full list of τὸ ὄνομα + gen. (total 12): Gen 17:5, 15; 32:27(28), 28(29); 35:10, 10;
Exod 1:15; 18:4; Lev 24:11; Num 25:14; 26:30(46); 26:59. There are also 17 of plural
τὰ ὀνόματα + gen. in the expression ταῦτα τὰ ὀνόματα τῶν/αὐτῶν …: Gen 25:13 etc.
THE ONOMA RULE 341
So the case of the possessor can be varied by the translator, regardless

of the Hebrew, which is the same each time. But there is something else
that varies. In the verses just quoted, we see that ὄνομα has no article in
the first instance, but has it in the second. Is this accidental? Is it just the
translator exercising free choice? Investigation shows that there is a pat-
tern here that is adhered to all through the Pentateuch. It is, put simply: in
name-statements of the type ‘the name of x is N,’ if the possessor of the
name is expressed in the dative, ὄνομα has no article; if the possessor of
the name is expressed in the genitive, ὄνομα has the article. This may
seem unexpected, but it turns out to be true; and the likely explanation will
be seen in due course.
The result of an enquiry into other books of the LXX is the same. The
rule is followed in this selection: Joshua, Psalms, Isaiah, XII Prophets, and
Ruth (in 1–4 Maccabees or Tobit there are no relevant cases), with the
dative tending to be more common, as in the Pentateuch. Full data need
not be given here. I simply illustrate with further instances of the two
types, some in close proximity:
Gen 32:27(28) Τί τὸ ὄνομά σού ἐστιν;
Exod 3:13 Τί ὄνομα αὐτῷ;
Exod 1:15 καὶ εἶπεν … ταῖς μαίαις τῶν Ἐβραίων, τῇ μιᾷ αὐτῶν, ᾗ
ὄνομα Σεπφωρά, καὶ τὸ ὄνομα τῆς δευτέρας Φουά
Num 25:14 τὸ δὲ ὄνομα τοῦ ἀνθρώπου τοῦ Ἰσραηλίτου … Ζαμβρί
Num 25:15 καὶ ὄνομα τῇ γυναικὶ τῇ Μαδιανίτιδι … Χασβί
Judg 1:23AB τὸ δὲ ὄνομα τῆς πόλεως (αὐτῶν) ἦν ἔμπροσθεν Λουζα.
Judg 18:29A καὶ ἦν Λαις ὄνομα τῇ πόλει τὸ πρότερον.
Ruth 2:1 καὶ ὄνομα αὐτῷ Βόος.
Ruth 2:19 τὸ ὄνομα τοῦ ἀνδρὸς … Βόος.
Isa 42:8 ἐγὼ κύριος ὁ θεός, τοῦτό μού ἐστι τὸ ὄνομα.
Isa 51:15 κύριος σαβαωθ ὄνομά μοι.
In these cases we are dealing with name-statements of the form ‘the

name of x is N.’ Outside of such name-statements the phrase τὸ ὄνομα +
gen. of the possessor of the name can of course occur, and does so fre-
quently in the Pentateuch, when τὸ ὄνομα is object or subject of a verb.
A small sample out of many:
Gen 38:4 ἐκάλεσεν τὸ ὄνομα αὐτοῦ Αὐνάν.
Exod 23:21 τὸ γὰρ ὄνομά μού ἐστιν ἐπ᾽ αὐτῷ.
Num 11:3 καὶ ἐκλήθη τὸ ὄνομα τοῦ τόπου ἐκείνου Ἐμπυρισμός.
These show the standard pattern of noun + possessive gen. in which
the article is usual, as in ὁ θεός μου, ἡ ψυχή μου etc. In name-statements,
then, the translators are simply going over to this pattern when they depart
from the pattern with the dative. It is the latter that is somewhat unusual
in its syntax. We may already suspect that it was an older, special idiom.
Yet the data show that it was normal for the translators, and that some-
times they use the other pattern, τὸ ὄνομα + gen., only as a varier from it
(Exod 1:15; 18:3; Num 25:14–15 above).7
2. LXX PERIOD
Was this ‘rule’ a peculiarity of the LXX translators’ Greek? It is hardly

likely. But to answer this question it is imperative to investigate Greek
outside the LXX. Contemporary documents are a good place to start. In
documentary papyri of III–I BC, at least 30 name-statements can be found
(including plurals).8 Nearly all show dative of the possessor and con-
form to the rule, that is, ὄνομα has no article. Expressions with the relative
pronoun, like Gen 22:24 ᾗ ὄνομα Ῥεημά above, are commonest, provid-
ing a close parallel to Pentateuch usage (ᾧ/ᾗ ὄνομα Ν occurs nine times
there). The pattern τὸ ὄνομα + gen. of the possessor is also found once
(in plural) in a name-statement; but it is mostly found as subject or object
of a verb. A small selection to illustrate:
PCairZen 1.37.7 (258/7 BC) ὄνομα δʼ ἐστὶ αὐτῶι Ἡδύλος·
PCairZen 2.148.2 (III BC) παῖδα … ὧι ὄνομα Ὀνήσιμος.
PMich 1.103.22 (III BC) γράψον οὖν τῶι ἀρχιφυλακείτηι· ὄνομα δʼ
αὐτῶι Ἁρμυῦσις.
PCairZen 3.379.11 (III BC) ἔστι[ν] τὰ ὀνόματα τῶν [ὑ]φειρημ[έ]νων
[τὰ] ἱερεῖα Κ … Ψε[ν]πεσῶτος …
PTebt 3.793ii.16 (183 BC) … τοῦ Δημητρίου οὗ σοι ὑπόκειται τὸ
ὄνομα
Inscriptions tell the same story.9 The first 500 matches to ὄνομα in
inscriptions of all periods follow the expected pattern: when ὄνομα is
accompanied by a dative it has no article. The type with the relative pro-
noun, ᾧ/ᾗ ὄνομα Ν, is overwhelmingly frequent, being formulaic in manu-
missions. Other examples following the rule do occur, such as Ἀντί[οχος]
ὄνομα αὐτῶι ἐστι.10 The pattern τὸ ὄνομα + gen. as object of a verb can
also be exemplified.
7
I owe this point to Michael Curran. My thanks to him and to Trevor Evans and Terry
Roberts for their helpful comments on this paper.
8
Accessed via ‘Papyri.info.’ Cf. Mayser, Grammatik, II.2, 113.
9
Accessed via ‘PHI Inscriptions.’
10
Rizakis, Achaïe III.2.8 (II BC) = SIG 530.8.
THE ONOMA RULE 343
This evidence already indicates that LXX usage was in accordance with
the Greek of its time. But let us sample the literature of the same period
by looking at Polybius (II BC) and Diodorus Siculus (I BC).11 There
are not many instances in either, but they follow the expected pattern.
(Diodorus much prefers parenthetical ὄνομα + name.) Examples in name-
statements:
Plb. 13.7.6 τοῦτο δ’ ἦν ὄνομα τῇ γυναικὶ τοῦ Νάβιδος.
D.S. 20.33.5 τοῦτο γὰρ ἦν ὄνομα τῇ γυναικί.
Examples with τὸ ὄνομα + gen. as object of a verb:

Plb. 15.24.4 πᾶσι προτείνουσι τὸ τῆς ἐλευθερίας ὄνομα.
D.S. 9.2.3 ἀνεβόα συνεχῶς τὸ τοῦ Σόλωνος ὄνομα.
3. CLASSICAL PERIOD
It is time now to look back into what came before, that is, into Greek
of the Classical period. Where better to start than with Odysseus’s famous
trick answer to the Cyclops:
Homer, Od. 9.366 Οὖτις ἐμοί γ᾽ ὄνομα·
My name is Nobody.
Here is the pattern ὄνομα + dative already established in the oldest

Greek we have after Mycenaean, with absence of the article as in earliest
Greek usage.12 There are some other examples in Homer and the Homeric
Hymns.13 We may reasonably conclude that this was an archaic pattern,
maintained in Greek long after it first appeared. In fact there is a good
chance the pattern originated in Indo-European: there is evidence of it in
Hittite and Latin (though not in Indo-Iranian).14
To fill in the picture of usage in the Classical period after Homer I
have tested (wholly or in part) the works of these authors: Herodotus,
Aristophanes, Xenophon, and Plato. In Herodotus books 1 to 3, I find
56 instances of οὔνομα (Ionic form) + dat. following the rule, of which
11
Searches conducted via TLG.
12
Cf. Chantraine, Grammaire homérique, 2:158–66. The dat. in Od. 9.366 and like instances
Chantraine only says is ‘assez proche du génitif exprimant la possession’ (71).
13
H., Od. 19.183, 409; 24.306. H. Aphrod. 198; Dem. 122. (There are none in Hesiod.)
14
Hahn, Naming-Constructions, 10, 24–5, 116–7, 124–5. It also occurs in Old English
(Hahn, 169–71, 178–80) but I am not certain if any of the exx. are independent of trans-
lation of biblical texts. I am grateful to James Clackson and John Sheldon for assistance
here.
a specimen is: 1.34.2 οὔνομα δέ οἱ ἦν Ἄτυς.15 There are two apparent

exceptions.16 Aristophanes offers a much smaller number of instances (8)
but all follow the rule without exception, as in Thesm. 1200 ὄνομα δέ
σοι τί ἐστιν; Xenophon similarly has a small number (7), all following
the ὄνομα + dat. pattern (he has more of parenthetical ὄνομα + name).17
Plato provides important evidence, with some 273 examples of ὄνομα.18
The Cratylus has more than half the examples and much variation in syn-
tax in Plato’s playful discussion of ‘names’ of things. Even so, I find that
the great majority of the name-statements behave exactly as predicted. By
my count there are 42 in Plato, of which 39 obey the rule. Of these the
usual pattern ὄνομα + dat. prevails (36 times), but there are three instances
of τὸ ὄνομα + gen., confirming the alternative pattern. (Occurrences of the
latter as object of a verb can also be given.) The remaining three are anoma-
lous, involving οὗ ὄνομα, but they may have explanations. I offer brief
illustrations of each type:
ὄνομα + dat.:
Euthyd. 275b.1 ὄνομα δ᾽ αὐτῷ Κλεινίας.
Symp. 209a.7 … ᾗ δὴ ὄνομά ἐστι σωφροσύνη τε καὶ δικαιοσύνη.
Ion 531e.8 τίς οὗτος; τί ὄνομα αὐτῷ;
τὸ ὄνομα + gen.:
Phileb. 12b.8 … τὸ δ᾽ ἀληθέστατον αὐτῆς ὄνομα Ἡδονὴν εἶναι.19
τὸ ὄνομα + gen. as obj. of verb:
Theaet. 144c.8 τὸ δ᾽ ὄνομα οὐκ οἶδα τοῦ μειρακίου.
anomalous ὄνομα + gen.:
Parm. 147e.2 … ἢ ἐκεῖνο οὗπερ ἦν ὄνομα.20
4. THE NT
We now know that the LXX translators were following established

patterns in their observance of ‘the onoma rule.’ Next it will be interesting
15
οἱ = dat. sing., masc./fem. of old pers. prn.
16
Hdt. 1.1.3 … τοῦ βασιλέος θυγατέρα· τὸ δὲ οἱ οὔνομα εἶναι … Ἰοῦν τὴν ᾽Ινάχου.
1.179.4 ἔστι δὲ ἄλλη πόλις …· Ἲς οὔνομα αὐτῇ. ἔνθα ἐστὶ ποταμὸς οὐ μέγας· Ἲς καὶ
τῷ ποταμῷ τὸ (v.l. τῶ) οὔνομα. For τὸ οὔνομα + gen. as obj. of a verb see, e.g., 1.51.4;
2.115.2; 3.64.5.
17
I am not sure if X., Oec. 1.1 (similarly 6.4), with an abstract possessor, is really anoma-
lous: ἆρά γε ἡ οἰκονομία ἐπιστήμης τινὸς ὄνομά ἐστιν, ὥσπερ ἡ ἰατρικὴ …; ‘Is
home management the name of a certain science, like medicine …?’
18
Including ‘Spuria’ but not ‘Dubia.’ TLG search 5.09.13.
19
The two others of this pattern: Crat. 430a.7; Theaet. 144d.1 (sc. αὐτοῦ).
20
Is the text sound?: contrast 147d.4 οὗπέρ ἐστι τοὔνομα, e.6 ἧσπερ ἦν τοὔνομα. The
other two exceptions: Crat. 432e.2; Meno 74e.11.
THE ONOMA RULE 345
to go forward to the NT and see what we find.21 Will things have changed?
Will the NT authors know these patterns? There turn out to be 19 name-
statements of the kind we are looking for, of which no less than 14 adhere
to the old pattern with ὄνομα + dat., while four use the alternative with τὸ
ὄνομα + gen., again confounding expectations. I include one with plural
ὀνόματα, which also follows the rule. The one exception (Luke 1:63) will
be discussed below. It will be useful to see them all, the datives first, then
the genitives:
Mark 5:9 τί ὄνομά σοι;
Mark 5:9 καὶ λέγει αὐτῷ· λεγιὼν ὄνομά μοι
Luke 1:26 εἰς πόλιν τῆς Γαλιλαίας ᾗ ὄνομα Ναζαρέθ
Luke 1:27 πρὸς παρθένον ἐμνηστευμένην ἀνδρὶ ᾧ ὄνομα Ἰωσὴφ
ἐξ οἴκου Δαυὶδ
Luke 2:25 ἄνθρωπος ἦν ἐν Ἰερουσαλὴμ ᾧ ὄνομα Συμεών
Luke 8:30 τί σοι ὄνομά ἐστιν;
Luke 8:41 ἦλθεν ἀνὴρ ᾧ ὄνομα Ἰάϊρος
Luke 24:13 εἰς κώμην … ᾗ ὄνομα Ἐμμαοῦς
John 1:6 ἐγένετο ἄνθρωπος, ἀπεσταλμένος παρὰ θεοῦ, ὄνομα αὐτῷ
Ἰωάννης·
John 3:1 ἦν δὲ ἄνθρωπος ἐκ τῶν Φαρισαίων, Νικόδημος ὄνομα αὐτῷ
John 18:10 ἦν δὲ ὄνομα τῷ δούλῳ Μάλχος.
Acts 13:6 εὗρον ἄνδρα τινὰ … ᾧ ὄνομα Βαριησοῦ
Rev 6:8 καὶ ὁ καθήμενος ἐπάνω αὐτοῦ ὄνομα αὐτῷ [ὁ] θάνατος
Rev 9:11 ἔχουσιν … βασιλέα τὸν ἄγγελον τῆς ἀβύσσου, ὄνομα
αὐτῷ Ἑβραϊστὶ Ἀβαδδών
Matt 10:2 τῶν δὲ δώδεκα ἀποστόλων τὰ ὀνόματά ἐστιν ταῦτα·
Mark 14:32 καὶ ἔρχονται εἰς χωρίον οὗ τὸ ὄνομα Γεθσημανί
Luke 1:5 καὶ τὸ ὄνομα αὐτῆς Ἐλισάβετ
Luke 1:27 καὶ τὸ ὄνομα τῆς παρθένου Μαριάμ.
It is remarkable that these are spread through different books rather

than being the achievement of one author, and that the ‘rule’ is followed
so well by writers supposedly more familiar with Aramaic and, in the case
of Revelation, barely competent in Greek. Attempts to find a Hebraism in
ὄνομα αὐτῷ in Rev 6:8; 9:11 (and in John) can hardly convince, when the
dative is Greek idiom and the supposed Hebrew model ‫( שׁמו‬and Aram.
‫ )שׁמהּ‬has the possessive suffix (= αὐτοῦ).22 In Luke 1:27 Luke achieves
an elegant variation by using both patterns, first dat., then gen.
21
The standard ΝΤ treatments do not remark on the article with ὄνομα: Robertson, Gram-
mar, 460; BDF, §§128.3, 144; MHT, Grammar, III, 230, 295–6, 304–5, 310.
22
See MHT, Grammar, III, 295; Charles, Revelation, 169 n. 1, 246 n. 2; Thompson, S.,
Apocalypse, 89; Burney, Aramaic Origin, 30–2.
5. NT PERIOD
Evidence of Greek contemporary with the NT is to be brought into

the picture here. I have checked three authors of the first to second cen-
turies AD, Josephus (AJ only), Plutarch, and Epictetus, and found confir-
mation that the patterns continued to apply. Both Josephus and Plutarch
use ὄνομα + dat. quite often (I count 27 and 21 each) in accordance with
the rule, though they much prefer the parenthetical ὄνομα + name pattern,
probably because of its more literary tone. I do not find the other pattern
τὸ ὄνομα + gen. in Plutarch, but there is one nice example in Josephus:
AJ 1.36 τὸ δ᾽ ἐκείνης ὄνομα τῆς γυναικὸς Εὔα ἦν.23 Epictetus has only
two relevant examples and both show ὄνομα + dat. as expected.
The evidence from papyri of the first to third centuries AD confirms the
persistence of the ὄνομα + dat. pattern, mainly in the formula ᾧ/ᾗ ὄνομα
Ν (I count about 30 instances); there are none of the pattern τὸ ὄνομα +
gen., except as subj./obj. of a verb (frequent). Plurals match the pattern
too: I find several each of ὀνόματα + dat. and τὰ ὀνόματα + gen. But
there is something new emerging in this documentary evidence. I find
not one – which might not mean much – but five instances of ὄνομα +
gen. without article, even in the formulaic structure with the relative
pronoun:
BGU 2.423.22 + BL 8.27 (II AD) ἔσ[τ]ι[ν] μου ὄνομα Ἀντῶνις Μάξιμος.
POxy 50.3555.5 (I/II AD) θεραπαινίδιόν μου οἰκογενέ[ς], οὗ ἔστιν
ὄνομα Πεῖνα
SB 22.15614.9 (154 AD) \ἀπὸ κοπρίας δο[υ]λ[ι]κὴν/ ἧς ὄνομα Θαή-
σιος24
I interpret this as an indication that the archaic pattern ὄνομα + dat. was
beginning to break down. The more intuitive genitive was replacing the
dative, though the anarthrous ὄνομα remained from the old pattern. Our
documentary examples point to II AD as the time when the shift was under
way in non-literary Greek. The progressive loss of the dative case is sure
to have been a factor.25 There is interesting confirmation of the shift from
another quarter. In Vita Aesopi G, a popular work datable to I AD, there
is one name-statement, which reads (24.2): τί ὄνομά σου; Notable is the
contrast with Mark’s τί ὄνομά σοι;
23
Plutarch has two exceptional exx. of ὄνομα + gen. without article: Is. Os. 354c.8; 357e.8
(both in or. obl.). Cf. the papyrus exx. following.
24
The others are: PLond 2.360.9 (p. 216) (II AD); StudPal 22.40.10 (150 AD).
25
Cf. Browning, Greek, 36–8.
THE ONOMA RULE 347
6. NT TEXT
What of the NT exception to the rule? It is a case where the ‘onoma

rule’ can bring something to textual criticism:
Luke 1:63 ἔγραψεν λέγων· Ἰωάννης ἐστὶν ὄνομα αὐτοῦ.
τὸ ὄνομα αὐτοῦ ‫ א‬A B2 C D W Θ Ψ … ै
ὄνομα αὐτοῦ ो4 B* L Ξ 565. 579. 700. …
Thus NA27, where according to the rule we expect τὸ ὄνομα αὐτοῦ as

in the majority reading, or ὄνομα αὐτῷ. The decision to adopt the minor-
ity reading seems to have been made initially by Westcott and Hort, and
the discovery of ो4 (III AD) would have seemed strong confirmation.
We must resist the temptation simply to restore τὸ ὄνομα αὐτοῦ to the text
in accordance with the rule. If this reading was original, we need to explain
how the other, clearly early, reading ὄνομα αὐτοῦ arose. Simple omission
of τό is possible but not likely. The evidence we have seen of the break-
down around II AD of the archaic pattern, leading to ὄνομα + gen. instead
of dat., might support ὄνομα αὐτοῦ as original, were it not for the consid-
eration that the well-educated Luke, whose Greek accords with the ‘rule’
elsewhere, is unlikely to have made this ‘mistake’: it is more likely to be
the error of a later copyist affected by the new pattern. This argument leads
to the conclusion that the original reading was in fact ὄνομα αὐτῷ. The
attested readings are then explained: ὄνομα αὐτῷ was altered to ὄνομα
αὐτοῦ for the reason suggested, then ὄνομα αὐτοῦ was corrected to τὸ
ὄνομα αὐτοῦ in line with the normal, ‘correct’ pattern of a noun with a pos-
sessive genitive.
7. LXX TEXT
Let us return to where we began, the LXX Pentateuch. The question of

the text arises there too in some instances:
Exod 3:15 τοῦτό μού ἐστιν ὄνομα αἰώνιον καὶ μνημόσυνον γενεῶν
γενεαῖς.
There is a variant μοι in a few MSS (and Patristic citations), which would
remove the anomaly. Wevers thinks it might have arisen ‘under pressure
of the τί ὄνομα αὐτῷ of v. 13.’26 It could also be pressure of the stand-
ard pattern. But the reason for the anomaly may lie in the nature of the
26
Wevers, Exodus, 34.
sentence: τοῦτο refers back to ὁ ὤν of v. 13 and does not so much state

the name – which may not be a name but a description or appellation – as
tell us something further about it, as NETS has it: ‘This [ὁ ὤν] is an ever-
lasting name of mine and a memorial of generations to generations.’
In Gen 2:19 τοῦτο ὄνομα αὐτῷ, there is a variant αὐτοῦ, which Rahlfs
accepted in his text. The rule supports Wevers’s decision not to adopt it
(‘the A + variant αυτου is not to be preferred’).27
8. CONCLUSION
Far more significant than its value as a tool of textual criticism is what
the ‘onoma rule’ indicates about the LXX translators’ command of Greek.
An archaic pattern still current in the language of their time is carefully
adhered to and not confused with a more recent pattern also used, and the
former is used contrary to the structure of the original Hebrew. Their famil-
iarity with this subtlety is strong evidence of their knowledge of idiomatic
Greek. It is on a par with their intermittent adherence to another unobtru-
sive pattern of Greek syntax, the forward placing of an enclitic personal
pronoun (‘Wackernagel’s Law’) against the Hebrew order. This is signifi-
cant enough to have led one recent scholar to conclude that they had native-
speaker competence.28 Their observance of the present rule is as signifi-
cant and even more consistent. The NT phenomena similarly indicate the
NT authors’ familiarity with Greek idiom.
I conclude with a summary of the ‘onoma rule.’ In name-statements of
the type ‘the name of x is N,’ with the components ὄνομα + possessor of
name + name (+ copula), if the possessor is expressed in the dative, ὄνομα
is without the article; if the possessor is expressed in the genitive, ὄνομα
has the article. Thus: ὄνομα αὐτῷ Ν (ἐστιν) / τὸ ὄνομα αὐτοῦ Ν (ἐστιν).
The word-order is variable and the copula is optional. The rule originates
in an archaic pattern with the dative, probably descended from Indo-Euro-
pean, and operates in Greek from the earliest times to at least Middle
Koine (I–III AD), when it begins to break down. Observance of this rule
by writers of Greek is of course to be understood as largely intuitive, not
the product of conscious application of a rule learnt as such.
27
Wevers, Genesis, 32. [See above, n. 5 for another LXX textual question.]
28
Janse, ‘Aspects of Bilingualism,’ 383. [See further Lee, Greek of the Pentateuch, 123–
7.]
THE ONOMA RULE 349
Postscript
Sollamo, in ‘Usage of the Article’ (2020), includes the onoma rule among her
evidence for concluding that the Genesis translator ‘must have been a native Greek
speaker’ (53). In the same paper, Sollamo refers (51 n. 20, 52 n. 21) to a forthcoming
article on a related subject. It will have the title ‘The Onoma Rule, the Vocative,
and Some Other Free Renderings of Hebrew Nouns in the Status Constructus
Defined by Another Noun in the Status Absolutus or by a Pronominal Suffix in the
LXX Genesis.’ (I thank Raija Sollamo for this information.)
22
THE LITERARY GREEK OF
SEPTUAGINT ISAIAH
20141
Abstract
The Greek of LXX-Isaiah was characterised by Thackeray many years ago as

‘Good κοινή Greek.’ This label has been accepted with little question generally
and been very influential among Isaiah scholars. The present paper first sets out
to establish that to classify the books of the LXX we need to give up Thackeray’s
mixed and imprecise categories and to recognise three main criteria of classifi-
cation, namely: ‘translation method,’ ‘style,’ and ‘level of Greek.’ The aim of the
paper is then to present evidence to demonstrate that LXX-Isaiah is appropri-
ately classified as ‘literary’ under the third category, along with other books so
labelled by Thackeray. The implications for a better understanding of the trans-
lator and his methods are explored, especially in the light of the depictions by
A. van der Kooij and R. Troxel.
1. PREAMBLE
The Greek translation of Isaiah has been the subject of intense activity
in recent research, as is well known to any who have taken an interest
in the book. Van der Kooij’s publications have been prominent over more
than a decade and are ongoing.2 Notable monographs have appeared
recently, by Baer, Troxel, De Sousa, and Wagner.3 There is a new collection
of essays by leading scholars in the field,4 and of course numerous other
articles have been written. Work is in progress on the translation and com-
mentary in the series Bible d’Alexandrie. All these build on the efforts of
1
The material of this paper was first presented as the fifth Grinfield Lecture in Oxford in
February, 2012, and again in the lecture series ‘La Bible grecque des Septante’ in Paris
in April, 2013.
2
Van der Kooij, Oracle of Tyre, and, e.g., ‘Septuagint of Isaiah.’
3
Baer, When We All; Troxel, LXX-Isaiah; De Sousa, Eschatology; Wagner, Sealed Book.
Another I have not yet seen is Ngunga, Messianism.
4
Van der Kooij and van der Meer (eds.), Old Greek of Isaiah.
earlier scholars, notably Ottley, Ziegler, and Seeligmann.5 So it is a field

already well occupied. This paper is about one aspect of the book that
has interested me for some years but that I see as neglected by the Isaiah
specialists.
The Isaiah translator’s work has greatly engaged the interest of LXX
scholars – and provided enormous scope for inventive interpretation –
because of his method of translation. It is characterised by inconsistency:
he translates his Hebrew original both accurately and freely, in an unpre-
dictable way. Older responses reflect the disquiet felt at this. Swete said
the Isaiah translation (along with Psalms) shows ‘obvious signs of incom-
petence.’6 Walters called the translator ‘an ingenuous blunderer,’7 and
bequeathed us the very quotable remark: ‘he was completely unequal to
his task.’8 Of course our response is: what was his task? How did he view
it? We ought, surely, to be wary of putting our own expectations onto him.
In recent scholarship this kind of criticism has been dropped and much
more sensitive attempts have been made to understand the translator’s
approach. The focus of much of this research has been on identifying the
translator’s exegetical tendencies. Minute attention has been given to
deviations from a strictly literal rendering and significance extracted from
them. The results have not led to agreement, however, but to ongoing
debate.
2. THE CLASSIFICATION OF LXX-ISAIAH
This is merely background, though not irrelevant, to the main subject

of this paper, which is the Greek of LXX-Isaiah. By this I mean the Greek
viewed as a specimen of the written language in its own time. Its time I
take to be the second century BC, following the general consensus. It is
necessary to begin with Thackeray’s well-known classification of the LXX
books, in which he put Isaiah in his group 1, along with the Pentateuch,
‘Joshua (part),’ and 1 Maccabees. He labelled this group ‘Good κοινή
Greek.’9 The more I have looked at this description the more puzzling,
5
Ottley, Book of Isaiah; Ziegler, Untersuchungen; Seeligmann, Septuagint Version. Cf.
bibliography in van der Kooij and van der Meer (eds.), Old Greek of Isaiah, 211–23.
6
Swete, Introduction, 315–6.
7
Walters, Text, 242.
8
Katz, ‘Septuagintal Studies,’ 200. Troxel (‘βουλή and βουλεύειν,’ 155 n. 6) quotes the
first description as ‘an ingenious blunderer,’ which looks like an ingenuous improvement.
Neither works for me.
9
Thackeray, Grammar, 13. I do not know what Thackeray meant by ‘Joshua (part).’
THE LITERARY GREEK OF SEPTUAGINT ISAIAH 353
even deceptive, it seems. It is not at all clear. ‘Koine Greek’ we under-

stand; good Koine can mean almost anything you want it to mean. It could
be ‘good’ in the sense of idiomatic everyday Greek; ‘good’ in the sense
of high-quality learned Greek; ‘good’ in the sense of solid middle-of-the-
road grammatical Greek; and perhaps more.
The other groups that follow in Thackeray’s classification are different
in character from group 1 and each other. Group 2 ‘Indifferent Greek,’ is,
it appears, about the quality of the Greek as Greek, but the term is very
vague and might well refer to translation method; Thackeray himself says
this group contains ‘books of various styles’ [my emphasis].10 Then fol-
lows group 3, ‘Literal or unintelligent versions,’ which must refer to trans-
lation method. Then the two groups 4 ‘Literary,’ and 5 ‘Literary and Atti-
cistic,’ concern quality of Greek on a scale of ‘literariness,’ and are at least
clear, as is the final group 6, ‘Vernacular,’ indicating a position on the
same scale. Just before he gives his table, Thackeray himself confuses
things by saying that he classifies the books ‘into groups from the point
of view of style’ (p. 12). This is hardly true, except insofar as ‘style’ is
allowed to have multiple meanings. He goes on to say, ‘Isaiah, considered
as a translation, would certainly not be placed in the first class’: now he
seems to be referring to translation method. Similarly, in the preceding
discussion of items of vocabulary used as a test (pp. 9–12), it seems clear
(mostly) that ‘style’ refers to translation method.
All this is getting us nowhere. The problem is that Thackeray is mixing
the categories, as everyone knows and others have pointed out before.11
As I see it, to fix the problem we need to distinguish three main categories
or criteria of classification, which are: (a) ‘translation method’; (b) ‘style’;
(c) ‘level of Greek.’ To these might be added one or two others, notably
‘control/use of idiomatic Greek,’ but this would not change the main group-
ings. The intended meaning of these categories will be more fully explained
in the discussion to follow. Thackeray’s groups do not separate these prop-
erly, and confusion has ensued. The whole matter would not need to be
pursued in such detail if it were not that Thackeray’s classification has
been very influential and is still quoted as the primary authority.
LXX-Isaiah scholars seem to have generally accepted Thackeray’s
‘Good κοινή Greek’ label trustingly, and I think it has been unhelpful
to them. Let us take an example from a recent paper by van der Vorm-
Croughs, a very good one as it happens, on the presence of rhetorical
figures in Isaiah. Her argument that the Isaiah translator was familiar with
10
11
E.g., Sollamo, ‘Some “Improper” Prepositions,’ 775; Barr, Typology, 283.
Classical rhetoric is already convincing, but she adds that it is ‘supported

by the fact [my emphasis] that LXX Isaiah’s translator was writing in good
Koine Greek, as was pointed out by Thackeray more than a century ago.’12
She takes it as a fact; and she evidently understands Thackeray’s descrip-
tion as referring to style: rhetorical devices are a feature of style (as van
der Vorm-Croughs well knows). But if Thackeray’s group 1 is a style group,
then what about the other books in that category, the Pentateuch, Joshua
and 1 Maccabees? Are they similar in style to Isaiah? Clearly not.13
The aim of this paper is not to offer a new classification of the LXX
books – though that is certainly needed – but to demonstrate something
about the Greek of Isaiah that has hitherto not been appreciated by all
parties, including Thackeray. My aim, in brief, is to show that LXX-Isaiah
has features of language that justify classifying it as ‘literary’ Greek. That
of course falls under my category (c) ‘level of Greek’ proposed above.
Before I come to the evidence under that heading, it will be useful to say a
little about the other two categories, (a) ‘translation method’ and (b) ‘style,’
to show what I mean by them. The discussion of these will be illustrative
rather than a full enquiry, and does not claim to be saying anything new.
I will then move on to (c), my main focus.
3. TRANSLATION METHOD
Description of translation method in the LXX has traditionally been

framed in terms of ‘literal’ versus ‘free.’ This will serve for the present
purpose. It at least gives us a means of indicating what we are talking
about under this heading and differentiating the category from the other
two. Nevertheless, the whole matter is complicated. The terms ‘literal’ and
‘free’ are inexact, but it has been hard to replace them so far. Barr in his
much-admired Typology says at the start that they are unsatisfactory,
but at the end of his investigation has nothing better to offer.14 There has
been renewed discussion lately, in an attempt to refine them in some way,
for example by Lemmelijn.15 Barr, while not changing the terminology,
showed that literalism is a complex phenomenon and there are ‘variations
12
Van der Vorm-Croughs, ‘LXX Isaiah,’ 175.
13
Other reiterations of Thackeray’s description: Seeligmann, Septuagint Version, 43; van
der Kooij, ‘Isaiah in the Septuagint,’ 515, 518; De Sousa, Eschatology, 18.
14
Barr, Typology, 6; cf. 49–50. The attempt to identify criteria for literalness made in Tov,
Text-Critical Use (1981), 53–66 illustrates the difficulty.
15
Lemmelijn, Plague of Texts?, 108–14; see esp. Boyd-Taylor, ‘Classification of Literal-
ism.’ Cf. Lee, ‘Accuracy and Idiom’ for a different angle.
within a basically literal approach.’16 In addition to the aspects of literal-

ism that he identified, it seems to me there is another aspect that needs to
be allowed for, that is, whether the translation in any given place is natu-
ral or unnatural Greek. This will be illustrated from Isaiah in a moment.
Another difficulty is that to label or characterise a whole book accurately,
one really needs to analyse every rendering throughout the book. But it has
proved almost impossible to do that (so far), despite the extensive analy-
sis of selected features that has been achieved, especially by scholars of
the Finnish school. So we still rely on a combination of hard data and
impression.
My purpose here is not to solve these problems, but simply to sum
up the Isaiah translator’s translation method. In terms of ‘literal’ versus
‘free,’ it is obvious that the work cannot be characterised overall as ‘lit-
eral’ because ‘free’ renderings occur with great frequency. But this trans-
lator can also be ‘literal.’ Van der Louw has rightly said, ‘literal translation
constitutes his most important tool.’17 The fact is that the Isaiah translator
is both ‘literal’ and ‘free’ – sometimes both within the one sentence. So
it is actually quite difficult to sum him up.18 In addition it seems clear that
the translator has sometimes had problems understanding his Hebrew,
and has simply got an impression and worked from that. This could be the
reason for many of the ‘free’ renderings, rather than a deliberate intention
to depart from a ‘literal’ rendering and introduce an interpretation.19
I offer just two examples to illustrate these remarks and the translator’s
method. Let us begin with an example of ‘literal’ rendering, in a famous
locus:
(1) Isa 6:3
καὶ ἐκέκραγον ἕτερος πρὸς τὸν ἕτερον ‫וקרא זא אל־זה‬
καὶ ἔλεγον Ἅγιος ἅγιος ἅγιος ‫ואמר קדושׁ קדושׁ קדושׁ‬
κύριος σαβαωθ, ‫יהוה צבאות‬
πλήρης πᾶσα ἡ γῆ τῆς δόξης αὐτοῦ. :‫מלא כל־הארץ כבודו‬
The translation is very ‘literal,’ that is, word-for-word or ‘isomorphic’:
the Greek word-order and parts of speech match the Hebrew; and the
Greek has the same meaning as the original. But it is not totally literal.
The translator has changed the verbs (ἐκέκραγον, ἔλεγον) to plural, and
16
Barr, Typology, 7.
17
Van der Louw, Transformations, 243.
18
The same point has of course been made by others: see Troxel, LXX-Isaiah, 87–8;
Wagner, Sealed Book, 113.
19
See Wagner, Sealed Book, 30 n. 151 and 49–50 for a good summary of the difficulties
faced both in reading and in understanding the Hebrew text.
has turned ‫‘( זא אל־זה‬this one to this one’) into natural Greek by using the
idiom ἕτερος πρὸς τὸν ἕτερον, yet without changing the meaning (‘they
cried out to one another’). So we see one of the ways in which ‘literal’ can
vary in itself, in this case between more natural and less natural Greek, and
we see how the Isaiah translator may be literal and idiomatic at the same
time.20
My second example illustrates the Isaiah translator’s ‘freedom,’ com-
bined with literal rendering:
(2) Isa 2:16

:‫… ועל כל־אניות תרשׁישׁ ועל כל־שׂכיות החמדה‬
[The day of the Lord will come] … and on all the ships of Tarshish and on
all the ships(?) of desire.
LXX … καὶ ἐπὶ πᾶν πλοῖον θαλάσσης καὶ ἐπὶ πᾶσαν θέαν πλοίων κάλλους.
[The day of the Lord will come] … and upon every ship of the sea and upon
every spectacle of beauty of ships (or: of ships of beauty).
How exactly he arrived at this rendering we need not analyse in detail.21

What is interesting is that he follows the general shape of the original, but
with the freedom he allows himself, he creates a meaningful and, as I see
it, almost poetic rendering. The turning of the two last Hebrew words into
three (θέαν πλοίων κάλλους) achieves this effect. We notice also the allit-
eration in π, the repetition of πλοῖον, and possible hyperbaton in the final
phrase (θέαν … κάλλους). These elements belong to style, which is our
next category.
4. STYLE
Every writer has his or her own individual style. It can be hard to say
what constitute the elements of style, and even harder to say what is good
or bad style. It is an elusive category, and very subjective; but it is real.
Style relates to the features which depend on the personal choice of the
writer, and can include almost anything. Even literal translation can be seen
as a style choice; but style is a separate entity from translation method.
20
For the idiom see LSJ, s.v. ἕτερος I.3. In Muraoka, Lexicon, s.v. ἕτερος 5, it is mis-
takenly marked as unattested before the LXX, perhaps from a misreading of Thackeray,
Grammar, 45. It occurs three more times in Isa at 13:8; 34:14, 16, with the varier
ἀλλήλων in 34:15.
21
Cf. Ottley, Book of Isaiah, 2:114. Aramaic is likely to be involved in the interpretation
of ‫שׂכה‬.
My own impression of the Isaiah translator is that he is a skilled stylist,

very conscious of what he is doing, and that the style of his version is a
most important, perhaps overriding, concern to him. His skill as a stylist
has of course been noted by others.
What are the indicators of an interest in style? One sure indicator
is the use of rhetorical devices. I mean ancient Greek rhetorical devices
(though we still use many of them today). In the context of translation we
mean those not in the original but introduced or contrived by the trans-
lator. These are clearly present in LXX-Isaiah. Van der Vorm-Croughs has
noted examples of inclusio, epiphora, anaphora and parallelism, as well
as avoidance of excessive use of some of these, especially repetition.22 I
made some slight observations myself some years ago of variatio, allitera-
tion and asyndeton in Isaiah.23 I have not seen the dissertation by Philippe
Le Moigne, which seems to be an important contribution; I know that van
der Kooij refers to it for a treatment of chiasmus in Isaiah, which of course
is a notable stylistic tool.24 Van der Louw in his study of Isaiah chapter 1
has identified a remarkable instance of double chiasmus purposely created
by the translator (Isa 1:25).25 I select here just one passage to illustrate the
Isaiah translator’s interest in style:
Isa 30:13–14
διὰ τοῦτο ἔσται ὑμῖν ἡ ἁμαρτία αὕτη
ὡς τεῖχος πῖπτον παραχρῆμα πόλεως ὀχυρᾶς ἑαλωκυίας,
ἧς παραχρῆμα πάρεστι τὸ πτῶμα,
(14) καὶ τὸ πτῶμα αὐτῆς ἔσται ὡς σύντριμμα ἀγγείου ὀστρακίνου, . . .
Therefore will be to you this sin
as a wall falling immediately, of a strong city captured,
of which immediately comes the fall,
and its fall will be like the breaking of an earthen jar.26
‫ לכן יהיה לכם העון הזה‬13
‫כפרץ נפל נבעה בחומה נשׂגבה‬
:‫אשׁר־פּתאם לפתע יבוא שׁברהּ‬
‫ ושׁברהּ כשׁבר נבל יוצרים‬14
22
Van der Vorm-Croughs, ‘LXX Isaiah,’ 174–88. More exx. of anaphora (repetition of a
word or phrase at the beginning of successive clauses): Isa 14:5; 32:5; 59:19–20. For
terminology one can consult Smyth, Grammar, 671–83; Rowe, ‘Style,’ as well as van
der Vorm-Croughs, ibid.
23
Lee, ‘Translations,’ 777–8, 780.
24
Van der Kooij, ‘Septuagint of Isaiah,’ 34. Le Moigne, ‘Le livre d’Esaïe.’
25
Van der Louw, Transformations, 219–22; cf. 244–5 for summary of stylistic features in
Isaiah 1.
26
My translations are deliberately literal: stylish English is not an issue here.
Therefore will be to you this iniquity

like a breach falling, bulging in a wall raised high,
(of) which suddenly in an instant will come its breaking;
and one will break it like the breaking of a jar of potters.
The Greek is a loose rendering of the original, conveying a meaning

not quite the same but similar. The steps by which the translator reached
this result are not clear; but it seems to me that the style is as important
to him as the meaning. He certainly does not aim at an exact rendering.
The stylistic features we can notice are: (a) alliteration in π, achieved by
choosing certain words instead of other equivalents; (b) anastrophe, that
is, when a word at the end of a clause is repeated at the beginning of
the next, in ‘τὸ πτῶμα, καὶ τὸ πτῶμα’: this of course twice matches the
Hebrew (‫)שׁבר‬, but the third time the translator declines to repeat it, using
instead σύντριμμα; (c) the chiasmus ‘πάρεστι τὸ πτῶμα … τὸ πτῶμα
αὐτῆς ἔσται’; (d) repetition of παραχρῆμα; (e) hyperbaton in ‘τεῖχος …
πόλεως.’ None of this is accidental: the translator has contrived this out-
come, at the expense of a completely literal rendering. And the result is a
pleasing one.
This matter of style is very relevant to our third and main topic, the
‘literary level’ of the Greek. The Isaiah translator clearly had some train-
ing in the rhetorical arts. Such a person would also have had a training
in writing good-quality Greek: the two go together. So we would not be
surprised to find evidence of skill in the latter.
5. LEVEL OF GREEK
Level of Greek is about the kind of Greek a writer uses. Greek, like
every language, offers a range of variations from vernacular to learned.
Post-Classical Greek was particularly sensitive to this kind of differen-
tiation. Every text reveals degrees of ‘literariness,’ that is, the presence
(or absence) of features that reflect the level of education of the writer
and his success in deploying them. A Greek education beyond the basics
encouraged the use of features of good-quality or ‘correct’ Greek, and taught
the elements of writing in a good style. The level aimed at became higher
as the student progressed. Every aspect of the language was involved: pho-
nology, morphology, syntax, vocabulary, and at the higher levels, rhe-
torical figures. The approved features contrasted with those of every-
day spoken Greek, which were disparaged. The educational model was of
course not static but developed over time. The second century BC was
not yet the era of Atticism, which began to exert influence in the first
century BC, calling for the use of strict features of Attic (like -ττ-). But
Atticism had its roots in earlier approaches to education and ideas of
what constituted ‘good’ Greek, and these are reflected in the texts of the
third and second centuries BC. The labels we use are inevitably impre-
cise, and the details of the features involved are still the subject of study,
but the phenomenon is real.
Thackeray himself, as alluded to above, classified certain books as
‘literary’ or ‘literary and Atticistic.’ He never explained or justified his
decisions, but his judgement was generally first-rate, based as it was on
a deep knowledge of Classical and later Greek. Experience with the LXX
confirms his decisions about the books that he did place in these cate-
gories, but I believe that in the case of Isaiah he was not quite on the mark
(as also with Sirach).27 It is interesting to see that in an earlier article,
before he published his Grammar, Thackeray said that the Isaiah trans-
lator ‘employs a Greek which much more nearly approaches the classi-
cal style than the Greek of the more painstaking translators of the other
prophetical books.’28 He went on to illustrate this by the translator’s use
of ‘connecting particles.’ But for some reason Thackeray did not follow
these remarks through to their natural conclusion. Since that time, no one
seems to have looked into the question.29
A study of the language of Isaiah many years ago (in 1934) by Ziegler
showed, from observation of parallels in documentary papyri, that the
Isaiah translator was very familiar with contemporary Egyptian Greek
vocabulary and drew on it extensively.30 More recently van der Meer has
pursued similar enquiries with good results.31 But in some ways this work
has been misleading. It has allowed the impression to be gained that the
Greek of Isaiah is everyday, even vernacular Koine Greek, rather than
anything else. While the parallels are good, this is not the full story by
any means.
27
On Sirach see Aitken, ‘Literary Attainment.’ The list of ‘literary’ books then is: 1 Esd,
Dan (LXX), Esth, Job, Prov, Wis, Ep Jer, Bar, 2–4 Macc, Isa, Sir (+ Eccl?).
28
Thackeray, ‘Greek Translators,’ 583.
29
Olley (Righteousness, 9) perceptively remarked that ‘little attention has been given to
Thackeray’s conclusion that “he employs a Greek which … the other prophetical books”’
[full quotation as above].
30
Ziegler, Untersuchungen, 175–212 (conclusions, 211).
31
See van der Meer, ‘Papyrological Perspectives,’ 107–33. Note the observation (109–10)
that Ziegler’s study covered only five percent of the whole vocabulary of LXX-Isaiah
and was based only on the lexicon of Preisigke.
6. EVIDENCE FOR A HIGHER LEVEL OF GREEK IN ISAIAH
The evidence for my case follows now, under a number of headings.

Only the strongest examples are included: others have been considered
and dropped. Even so, I am sure that the list could be extended. A full
discussion is not given for every item but could readily be added.
6.1. A Tragic phrase: ὦ τάλας ἐγώ

Isa 6:5 καὶ εἶπα Ὦ τάλας ἐγώ, ὅτι κατανένυγμαι, ὅτι ἄνθρωπος ὢν
καὶ ἀκάθαρτα χείλη ἔχων ἐν μέσῳ λαοῦ ἀκάθαρτα χείλη ἔχοντος
ἐγὼ οἰκῶ.
I said, ‘O wretched am I! for I am cut to the heart, that being a man
and having unclean lips in the midst of a people having unclean lips I
dwell’
… ‫ואמר אוי־לי כי־נדמי כי אישׁ טמא־שׂפתים אנכי‬
I said, ‘Woe is me! for I am destroyed/struck dumb, for a man unclean
of lips am I …’
The phrase is highly recognisable as an expression from Attic Tragedy,

found in Sophocles and Euripides. It is sure to have been encountered at
some point in the course of a Greek education. Let us see two examples
from popular plays:
Sophocles, Phil. 744 δύστηνος, ὢ τάλας ἐγώ. | ἀπόλωλα, τέκνον·
Euripides, Hippol. 1090 ὦ τάλας ἐγώ, | ὡς οἶδα μὲν ταῦτ,’ οἶδα δ’
οὐχ ὅπως φράσω.
This expression cannot be found in contemporary documents. Even the
word τάλας is rare and used only in poetic epitaphs, as one would expect.
In the LXX τάλας occurs only in Wis 15:14; 4 Macc 12:4, both literary
books. This does not amount to proof that the translator knew ὦ τάλας
ἐγώ from reading Tragedy, though I myself think it probable. But even
if the phrase was sometimes used in speech by the educated, its literary
provenance would have been known. It is notable that the only other
example in the LXX is in the literary 4 Maccabees (8:17 ὦ τάλανες ἡμεῖς).
No other LXX translator thought of using it; and of course it is inde-
pendent of the Hebrew (the common rendering of ‫ אוי‬is οὐαί).32
32
The connexion with Tragic language goes unnoticed in: Ziegler, Untersuchungen, 108;
Ottley, Book of Isaiah, in loc.; Seeligmann, Septuagint Version, 53; LXX.D Komm., in loc.
But Walters (Text, 230) spotted it: ‘an isolated classical reminiscence, well fitted to the
sublime note of this passage.’ Schleusner (Lexicon in LXX, s.v. τάλας) was best of all:
after Isa 6:5 he immediately quoted a parallel in Euripides, Phoenissae 1345.
6.2. A proverbial expression

Isa 44:4 … ἐπὶ τὰ τέκνα σου, (4) καὶ ἀνατελοῦσιν ὡσεὶ χόρτος ἀνὰ
μέσον ὕδατος καὶ ὡς ἰτέα ἐπὶ παραρρέον ὕδωρ.
… upon your children, and they will spring up like grass in the midst
of water and like a willow beside water flowing by.
:‫וצמחו בבין חציר כערבים על־יבלי־מים‬
and they will spring up in the midst of grass, like willows by streams
of water.
The Greek rendering ἐπὶ παραρρέον ὕδωρ does not follow the origi-
nal exactly but catches the sense of it while converting it into a pleasing
poetic phrase. The verb παραρρέω is rare in the LXX, occurring only once
elsewhere (Prov 3:21), in a figurative, developed sense (intrans. in pass.)
‘drift away,’ hence ‘be neglectful’ (of something), also found in the NT
(Heb 2:1). It is a word with Classical and literary antecedents, in which the
figurative uses already predominate. So the choice of word is interesting to
begin with. But why this phrase? Is it his own creation? A search in Greek
literature establishes that there was a proverb, first attested in a fragment
of the Old Comedian Cratinus (V BC), and explained as referring to those
who promise to undertake something whatever the odds. The proverb reap-
pears centuries later in the renowned rhetorician Libanius (IV AD).33
In between, the phrase recurs at long intervals, in a scholiast on Homer,
then in Plutarch, then much later in Patristic literature.34 All this suggests
that it was a set expression with a literary provenance, known to those
with a good education, not one made up by the Isaiah translator. He takes
it and puts it to work for his own purpose, without allusion to the sense
of the proverb.
6.3. Literary or poetic vocabulary
Here are grouped a number of words, and a phrase, with literary or

poetic antecedents. It is not intended to be a complete list. A full study
of the vocabulary of LXX-Isaiah, which is a desideratum, would produce
more examples like these and interesting results overall.35
33
Cratinus, Fr. 64 ὕδωρ παραρρέει. Libanius, Ep. 109.2 κἂν ὕδωρ, φασί, παραρρέῃ.
34
Schol. in Hom., Il. 2.307 . . . Πλάτων πλάτανον καὶ τὸ παραρρέον ὕδωρ φησί (cf. Pl.,
Phaedr. 230b). Plut., Cato Ma. 19.1 (I–II AD) τὸ παραρρέον δημόσιον ὕδωρ. In Patristic
lit., e.g., Asterius, Homil. 2.3.3 (IV–V AD) τὸ παραρρέον ὕδωρ.
35
Cf. Baer, When We All, 281.
(1) κῦδος, ‘renown’: only here in LXX.

Isa 14:25 καὶ ἀφαιρεθήσεται ἀπ’ αὐτῶν ὁ ζυγὸς αὐτῶν, καὶ τὸ κῦδος
αὐτῶν ἀπὸ τῶν ὤμων ἀφαιρεθήσεται.
And will be removed from them their yoke, and their renown from
their shoulders will be removed.
:‫וסר מעליהם עלו וסבלו מעל שׁכמו יסור‬
And will fall off/depart from them his yoke and his burden from on
his shoulder will fall off/depart.
The Greek rendering differs somewhat from the original, and it is not
clear why the translator used κῦδος to translate ‫‘( סבל‬burden’), but that is
not the point.36 This is quite an unusual choice of word. κῦδος is used only
here in the LXX; it is an Epic and poetic word, obviously not an everyday
one; and in contemporary documents it is found only in inscriptions, often
epitaphs and other poetic texts.37 So it seems to me a clear indicator of the
translator’s command of the poetic ‘register’ of vocabulary. The context
(verses 24–25) might be relevant to his choice of κῦδος here: God is speak-
ing, in a noticeably lofty style. (The elegant chiasmus is in the original and
the translator naturally does not disturb it.)
(2) ἐν ᾅδου, ‘in Hades’: Isa 38:18, 18.
Other LXX × 4: Job 1; Sir 3.
The similar expression εἰς ᾅδου (‘to Hades’) appears first in the LXX in
the Pentateuch and then occasionally in other books following their lead.
It originates in Homer and then appears frequently in Classical authors,
especially Tragedy. It contains an archaism (εἰς + gen.) that could not
have been created by the translators. It seems to me a sign of the Penta-
teuch translators’ Greek education. The Isaiah translator also uses εἰς ᾅδου
(14:11, 15, 19), and ἕως ᾅδου (57:9), but he goes a step further, as it were,
and shows knowledge of another form of it that similarly appears in Attic
Tragedy (and Plato) and goes back to Homer’s εἰν Ἀΐδαο. It is signifi-
cant that the only other translators to use it are Job and Sirach.
(3) ἀγαυρίαμα, ‘pride’: Isa 62:7.
Other LXX × 3: Job 1; Jer 1; Bar 1. Not attested outside LXX. Cf.
ἀγαυρός Hesiod, Hdt., Nicander, rare; ἀγαυριῶμαι Job 1; Aquila.
36
See an attempted explanation in Ottley, Book of Isaiah, 2:181–2; Seeligmann, Septuagint
Version, 50.
37
E.g., IEgVers 38.8 (Egypt, late Ptol.) καὶ γὰρ ἐνὶ φθιμ<έ>νοισι θ<ε>ουδείης γέ<ρ>ας
ἔσχον, | ἡμιθέων ἀλόχων κῦδος ἐνεγκαμένη·
(4) ἀπαμαυρῶ, ‘make dim, make fade away’: Isa 44:18.

Not elsewhere in LXX. Rare: Hymns to Isis (I BC), Orphica (date?),
+ later.
(5) ἀποσκορακίζω, ‘reject with contempt’: Isa 17:13.
Other LXX × 2: 1 Macc 1; Ps 1. First in LXX, then Plutarch +. Cf.
σκορακίζω, Demosth., Phld., Philo, + later;
(6) ἀποσκορακισμός, ‘rejection with contempt’: Isa 66:15.
Not attested again. Cf. σκορακισμός Sir 41:19, then Plut.38
(7) κύω, metaph. ‘conceive in the mind’: Isa 59:4, 13.
Not elsewhere in LXX. The word Homer (κυέω) +; this use Plato,
Xen., rare.
(8) μακρόβιος, ‘long-lived’: Isa 53:10.
Other LXX: Wis 3:17. Hdt., Hippocr. +. Cf. μακροβίωσις Bar 3:14,
only here.
(9) ὁμίχλη, ‘mist, fog’: Isa 29:18.
Other LXX × 9: Ps 1; Job 2; Wis 1; Sir 2; XII Proph 3. Homer +.
(10) πυρίκαυστος, ‘burnt by fire’: Isa 1:7; 9:4; 64:10.
Not elsewhere in LXX. Homer +.
6.4. New creations
LXX-Isaiah presents a number of compounds that are hapax legomena

not only in the LXX but in all of Greek (as so far attested). It is always
risky to assert that a word is entirely new, but these certainly have the look
of learned neologisms coined by the Isaiah translator to meet his purpose
at certain points. They indicate a confident command of the processes of
forming literary compounds.
(1) ἄνισχυς, ‘strengthless’: Isa 40:30.
(2) βαθύφωνος, ‘deep-voiced’: Isa 33:19.39
(3) κατάκλιτος, ‘downward-flowing’: Isa 3:23.40
(4) πριστηροειδής, ‘saw-like’: Isa 41:15.
38
The origin of this group is in the Attic expression (Ar. +) ἐς κόρακας ‘(go) to the ravens’
= ‘go to hell’: see Beekes, EDG, s.v. κόραξ.
39
The full list of -φωνος compounds (about 125) contains only a few common words; most
are poetic or learned creations, with limited attestation: see Buck-Petersen, Reverse Index,
287–8.
40
As a descriptor of θέριστρα: see van der Meer, ‘Trendy Translations,’ 592, 596.
6.5. Particles
The reduction of the Classical range of particles in post-Classical Greek

and the significance of particles as a sign of education are well known.41
The following is a list of those occurring in LXX-Isaiah that contribute
to my case, though they are not all of the same weight.
(1) τοιγαροῦν
Isa 5:26 οὐκ ἀπεστράφη ὁ θυμός, ἀλλ’ ἔτι ἡ χεὶρ ὑψηλή. τοιγαροῦν
ἀρεῖ σύσσημον ἐν τοῖς ἔθνεσι τοῖς μακρὰν καὶ συριεῖ αὐτοῖς ἀπ’
ἄκρου τῆς γῆς
His anger has not turned back, but his hand is still raised high. For that
reason he will raise a standard among the nations that are far off and
will whistle to them from the end of the earth.
(2) τοίνυν: Isa 3:10 + 3. Other LXX × 16: literary books 14; others 2.
(3) μέν: Isa 6:2; 41:7. Other LXX × 216: Pent. 19; literary books 184;
others 13.
(4) οὖν: Isa 30:8; 40:25; 57:6. Other LXX × 243: Pent. 79; literary books
124; others 40.
(5) τε: Isa 37:12. Other LXX × 265: Pent. 54; literary books 189; others
22.
(6) καὶ γάρ: Isa 23:10. Other LXX × 81: Pent. 5; literary books 42;
others 34 (1, 2 Chr 8; Ps 23; Zeph 1; Ezek 2).
(7) δή: Isa 3:1 + 5. Other LXX × 280 +: Pent. 4.
The word that stands out in this list is τοιγαροῦν. It is rare elsewhere
in the LXX and then only in literary books (Prov × 2; Job × 2; Sir × 1;
2–4 Macc × 5). No examples can be found in Ptolemaic papyri. Thackeray
noted it (and τοίνυν) in his list illustrating the ‘classical style’ of Isaiah.42
Here in Isaiah it corresponds to a simple ‫ו‬, so it is the translator’s choice.
Also quite significant is τοίνυν, which again is a rarity elsewhere. The
other particles listed are more common, and the Pentateuch, in middle-level
Greek, uses μέν, οὖν, and τε quite often. But their use there is an indicator
of education too, and they still occur most often in the literary books. καὶ
γάρ is also favoured in literary books, except in the case of Psalms, where
some stereotyping as an equivalent of ‫ גם‬is involved (13 times out of 23).
I add δή to the list with caution, because it had become a stereotyped
equivalent of ‫ נא‬for many translators. It is suggestive however that the
41
See, e.g., Blomqvist, Greek Particles, 132, 144–5.
42
Thackeray, ‘Greek Translators,’ 583. Cf. Blomqvist, Greek Particles, 130.
Isaiah translator introduces it independently of ‫( נא‬or any Hebrew word)

four out of six times.
I see this collection of particles as evidence of the Isaiah translator’s
intention to compose his work in Greek of a higher literary level. To see
τοίνυν and τοιγαροῦν only as showing ‘the translator’s interest in mark-
ing sequences of thought,’ as Troxel does, is to miss half the story. Of
course they do that, but why does he choose these particular words?43
6.6. Literary word-choices

There are many instances in which the Isaiah translator chooses the more
literary equivalent of some everyday word. This is not to say that he will
never use the everyday one; in fact he may do so, and no particular sig-
nificance should be seen in it when he does. But it is the use of the higher-
level, ‘better’ word that is significant. Contrast is the key, that is, between
the ‘better’ word and the more everyday one. In the list I have given the
latter in each case (with ‘≃’ indicating practical equivalence rather than
complete synonymy). For νοῦς and οἶμαι, there is no single equivalent: it
is rather a choice between using the word and not using it by saying it in
a different way (or not at all). For example, instead of νοῦς he could have
used καρδία, the more Semitic expression of the idea, which he does use
often. I have listed below the vocabulary that seems to me significant;
there is probably more. Some are selected for brief remarks after the list.
ἀεί, ‘always’: Isa 42:14; 51:13. Other LXX × 12: literary books 10;
Judg 1; Ps 1.
≃ διὰ παντός Isa × 10; LXX × c.130.
ἀμῶ, ‘I reap’: Isa 17:5; 37:30. Other LXX × 3: Pent. 2; Mi 1.
≃ θερίζω Isa × 0; LXX × c.27.
ἀροτήρ, ‘ploughman, farmer’: Isa 61:5.
≃ γεωργός Isa × 0; LXX × 9.
αὔξω, ‘I increase’: Isa 61:11.
≃ αὐξάνω.
βούλομαι, ‘I wish, want’: Isa 1:11 + 12. Other LXX × 108: Pent. 14;
literary books 39; others 55.
≃ θέλω Isa × 9; other LXX × 144.
43
Troxel, LXX-Isaiah, 93, in a discussion of ‘conjunctions.’ Similarly on γάρ (92) Troxel
notes an ‘unusually high’ frequency in Isaiah, but only concludes that it ‘attests his inter-
est in creating smoother connections between clauses.’ That is what γάρ does if you use
it; but using it a lot indicates literary intent. Wagner (Sealed Book, 30) distils Troxel into
a bland: ‘He fosters cohesion in the translated text with his adept use of Greek particles.’
δέδοικα, ‘I am afraid’: Isa 60:14. Other LXX: Job 7.

≃ φοβοῦμαι.
ἐπίσταμαι, ‘I know’: Isa 29:11 + 7. Other LXX × 50: Pent. 14; literary
books 18; others 18.
≃ γινώσκω, ἐπιγινώσκω, οἶδα.
ἐρυθρός, ‘red’: Isa 63:2 ἐρυθρὰ τὰ ἱμάτια. Other LXX × 29, all ἐρυθρὰ
θάλασσα.
≃ κόκκινος Isa × 2; other LXX x 41.
νοῦς, ‘mind’: Isa 10:7 + 3 (41:22 ἐφίστημι τὸν νοῦν). Other LXX × 25:
literary books 22; others 3.
οἶμαι, ‘I suppose’: Isa 57:8 (ᾤου). Other LXX × 21: Gen 4; 1 Macc 2;
literary books 15.
πόρρω, ‘far off’: Isa 17:13 + 4. Other LXX × 12: literary books 8; others 2.
≃ μακράν Isa × 7; other LXX x c.65.
πόρρωθεν, ‘from afar’: Isa 10:3 + 7. Other LXX × 8: literary books 4;
others 4.
≃ μακρόθεν Isa × 2; other LXX × 35.
πράσσω, ‘I do, act’: Isa 57:10. Other LXX × 36: Pent. 1; literary books 33;
Josh 1; 1 Macc 1.
≃ ποιῶ Isa × c.153; other LXX × 3000+.
ἀεί. It may seem surprising that ἀεί is significant, but changes were
happening in words for ‘always’ in early post-Classical Greek. ἀεί was in
decline except in set expressions or more stylish Greek, while διὰ παντός
was coming in strongly at this time. The frequency of διὰ παντός in the
LXX itself reflects this, and the Isaiah translator himself uses it more
often than ἀεί. The word that prevailed later, πάντοτε, had only recently
appeared.44
ἀμῶ. This word is the subject of an observation by the Atticist Moeris:
ἀμᾶν Ἀττικοί· θερίζειν Ἕλληνες.45 Though much later, this indicates the
direction in which things were going. Shipp says that ἀμῶ was ‘already
rare in Attic.’46 It is true that ἀμῶ also occurs in the Pentateuch (Lev 25:11;
Deut 24:19) and in Micah (6:15), and that does not put those books in
the ‘literary’ category; it only shows that their translators had some edu-
cation. It is a more significant choice in Isaiah when combined with the
other evidence.
44
Cf. Lee, ‘Atticist Grammarians,’ 293–4 [= Essay 19]; Meecham, Letter of Aristeas, 204.
45
Moeris α 158 (ed. Hansen).
46
Shipp, MGE, 63.
ἐρυθρός. The point is that a new word κόκκινος had come in to replace
ἐρυθρός, and ἐρυθρός was confined to the set phrase ἐρυθρὰ θάλασσα,
the ‘Red Sea.’ The Isaiah translator uses the old-fashioned word on its own
to mean ‘red’ in this one instance.47
πράσσω. This is another word that was going out of use, except in
restricted senses.48 In papyri of the third century BC it is nearly always
‘exact payment.’ When it does occur as a word for ‘do’ it is a more formal
word, and usually refers to bad actions, as in the one example in the Pen-
tateuch (Gen 31:28). In Josh 1:7 it occurs in God’s speech. The figures
for the LXX show how infrequent it is compared with the standard word
ποιῶ, and how much it is favoured by the literary books.
6.7. Other features

Noted here with short discussion are miscellaneous items: an Attic
form, two features of syntax, a poetic expression, and a special item of
vocabulary.
(1) ἡττῶμαι Isa 8:9 + 11. Other LXX: 2 Macc 1; Sir 1; Jer 1; Dan 2.
The word is an official, sometimes legal term that is rare in the LXX
outside literary books, but used often in Isaiah.49 I think it is significant
that it appears in the Attic form ἡττῶμαι, instead of ἡσσῶμαι. It used to
be thought that only the Attic form was available in early Koine Greek,
but we now have documentary evidence of both forms. So it was possible
to make a choice.50
(2) Comparative optative, mostly with ὡς, ὡς ἄν: Isa 11:9; 21:1; 66:20.
Other LXX: Pent. 9; JudgAB 2; Ps 7 (5 in one verse); Prov 3; Ezek 1.
The comparative optative is a special and rare use in Greek. It is best
explained as derived from Homeric language.51
47
Evidence in Lee, Lexical Study, 111–2. Another documentary ex. of κόκκινος can be
added: PCairZen 4.696.4 (III BC). Cf. Shipp, MGE, 325.
48
See Shipp, MGE, 461–8; cf. Gerleman, Job, 15: πράσσω ‘is much more literary than
the ordinary ποιεῖν.’ [See also Lee, Greek of the Pentateuch, 114–7.]
49
A good study of its usage in Isaiah, esp. in the light of documentary papyri, is found in
van der Meer, ‘Papyrological Perspectives,’ 113–20.
50
Verb forms: -σσ- in PHal 1.54, 115, 118, 188 (259 BC); PHib 2.197.53 (III BC); -ττ- in
PZenPestm 51.27 (257 BC). In the adj., ἧσσ- (60+) is commoner than ἧττ- (5) in III–I BC
papyri (as at 23.1.14). Obsolete discussion in: Thackeray, Grammar, 121 n. 2; Walters,
Text, 260.
51
See Evans, Verbal Syntax, 190–7; cf. Joosten, ‘Elaborate Similes,’ 230–3. No support
here for Baer’s (unconvincing) argument that ch. 66 is by a different hand: Baer, ‘What
Happens in the End?’
(3) Conditional optative with εἰ: Isa 49:15. Other LXX: Pent. 0; 1 Kgdms 1;
4 Macc 8; Job 2.
The conditional optative also looks significant. It is rare and favoured
in Job and especially ‘the literary essay’ 4 Maccabees, as Thackeray calls
it in remarks on rare uses of the optative.52 This and the type of optative
in (2) are remnants left over from the older fuller range of use, as the mood
became desystematised in early Koine Greek.53 The volitive optative per-
sisted longest, but these two types were certainly obsolescent and are
therefore signs of a higher education.
(4) βουλεύομαι βουλήν: Isa 3:9; 7:5; 8:10; 31:6 (cf. ἡ βουλὴ /τὴν
βουλὴν ἣν βεβούλευται 14:26; 19:17). Elsewhere Ezek 11:2.
The Isaiah translator has a fondness for this particular figura etymo-
logica, with its solemn, poetic effect. In one place (8:10) it can be said to
be motivated by the Hebrew (‫ ;)עצו עצה‬but in the other three he contrives
to introduce it himself. A search in Greek literature finds the combination
seven times in Homer and once in Plato, not elsewhere.54 It does not seem
unlikely that acquaintance with Homer contributed to the Isaiah transla-
tor’s inclination to use it.55
(5) Isa 43:14 Ἕνεκεν ὑμῶν ἀποστελῶ εἰς Βαβυλῶνα καὶ ἐπεγερῶ
πάντας φεύγοντας, καὶ Χαλδαῖοι ἐν πλοίοις δεθήσονται.
In Greek of an earlier period there was a distinct use of φεύγω to mean
‘be in exile,’ especially in the participial form. It is almost the default mean-
ing when (οἱ) φεύγοντες is encountered in Classical texts.56 It seems to
me that it is intended here: ‘I will rouse up all exiles’ (not ‘all who are
fleeing’).57 It is triggered naturally by the mention of Babylon, the place
of exile; the obscurity of what comes next (‘and Chaldaeans will be bound
in boats’) does not matter. In this use of φεύγω the translator reveals
familiarity with Classical usage.
52
53
See Evans, Verbal Syntax, 175–8.
54
Homer, Il. 9.75; 10.147, 327, 415; 23.78; 24.652; Od. 6.61 (βουλεύω act. in all, βουλάς
pl. except βουλήν once); Plato, Politicus 298b.8 εἰ δὴ ταῦτα διανοηθέντες βουλευ-
σαίμεθα περὶ αὐτῶν βουλήν τινα.
55
Troxel, in ‘βουλή and βουλεύειν,’ seeks reasons for the ‘insertion’ of βουλή but works
entirely within the framework of LXX-Isaiah and its Hebrew original.
56
See LSJ, s.v. φεύγω III.
57
So NETS (‘all who are fleeing’); similarly Ottley, Book of Isaiah, 1:235 (‘all that flee’).
But cf. LXX.D: ‘alle Verbannten aufwecken.’ Muraoka, Lexicon does not separate this
sense. The Modern Greek translations of Kolitsaras (1928) and Giannakopoulos (1986)
interpret as ‘I will stir up all (the Babylonians) to flee,’ which [does not seem] possible.
7. CONCLUSION
A strong case can be made for classifying LXX-Isaiah as ‘literary’ in

its level of Greek. It should be bracketed with other ‘literary’ books. This
is not to say that all those labelled ‘literary’ are exactly the same: there are
certainly differences between books, and further study might place Isaiah
somewhat lower than, say, Job and Proverbs. Nor does it mean that high-
level Greek is consistently maintained in Isaiah: there is of course vari-
ation downwards on the scale and there is interference in the quality of the
Greek from the Hebrew original.58 But I think it is mistaken to put Isaiah
into a vague ‘good Koine Greek’ category, as Thackeray did, and helpful
to know that its place is in the ‘literary’ category. More work needs to be
done, certainly, to refine what we mean by ‘level of language,’ to securely
identify ‘literary’ features, and to establish criteria for classifying books
on that basis. But we have some definite evidence in front of us now for
Isaiah, beyond Thackeray’s mainly intuitive placing of books in the ‘liter-
ary’ category.
This result may seem unimportant, but it has value for linguistic study
of the LXX. To appreciate properly the Greek of any LXX translated book
(or free composition), and to connect it with other Greek of the time, we
need to know where it stands on the literary scale manifested in the whole
of post-Classical Greek. We then understand better the choices the trans-
lator makes from the range of features available, and we can gain a bet-
ter understanding of his aims and of the translator himself. In the case of
a book where the translator and his aims are the subject of so much debate,
an appreciation of the kind of Greek employed may have a significant
contribution to make.
The finding that the Greek of LXX-Isaiah is high on the literary Greek
scale tells us something important about the translator. It indicates that
he had had a Greek education to an advanced level. We do not have exact
knowledge of that education, but it cannot be doubted that a Greek educa-
tion was accessible to Alexandrian Jews of the second century BC and that
they availed themselves of it. If we had nothing else, the Exagoge of the
Jewish tragedian Ezekiel and the polished Greek of the Letter of Aristeas
and the Prologue to Sirach would be sufficient evidence.59 How well does
58
I am grateful to Cameron Boyd-Taylor for input here and other comments on the paper.
My thanks also to Michael Curran for valuable feedback and improvements.
59
I believe the author of the Prologue refers to his education in the knotty εὑρὼν οὐ μικρᾶς
παιδείας ἀφόμοιον (l. 29), which I read as ‘having acquired the likeness/shape of no
small education.’ This is modesty clothed in artistry.
a translator with an advanced level of Greek education accord with the

portraits of the Isaiah translator that have been propounded in recent
study? It is not possible for me to deal with this question in a thorough
or even adequate way; I only put forward some observations that I think
are worth considering. I will take as representative the portraits drawn by
two leading Isaiah scholars.
Arie van der Kooij has led the way in characterising the Isaiah translator
as a ‘scribe.’60 I have to admit to some difficulty understanding the reason
for choosing this term and the precise meaning van der Kooij intends it
to have.61 Variations such as ‘scribe-translator’ do not help. ‘Scribe’ seems
to be no more than a convenient label to cover whatever it is the Isaiah
translator does. Wilk expands its scope by reference to ‘scribal artistry,’
and even extends this description to Paul’s handling of quotations from
Isaiah.62 Nevertheless it is fairly clear what van der Kooij’s conception of
this translator is: he is a Jewish scholar trained in the reading and interpre-
tation of the Hebrew text.
The observations made in this paper may raise some questions about
van der Kooij’s picture. In regard to translation method, one might have
expected a consistent (even ‘slavish’) literal method from one accustomed
to a close reading of the Hebrew, rather than a method that is basically
literal but takes great liberties within that framework. Similarly, in the
style category, one does not expect to see Greek stylistic and rhetorical
devices playing such a significant role, often at a tangent to the meaning of
the original. And thirdly, the choice of the literary Greek register requires
some explanation if the translator’s focus was on conveying the meaning
of the Hebrew text. The person who made this translation seems rather to
have been focused on turning the text not just into meaningful Greek, but
into stylish ‘biblical’ Greek, that is, Greek with a ‘biblical’ flavour on the
pattern of the Pentateuch, but in a better style and on a higher level. Such
a translation would appeal to the Greek-educated Jewish reader and dis-
play the translator’s own Greek learning. To me it seems that this was an
aim that at times overrode even the basic purpose of rendering the meaning
of the Hebrew text. Whether the translator also reveals himself to be an
60
Van der Kooij, Oracle of Tyre, esp. 112–23. Cf., e.g., Troxel, LXX-Isaiah, 20; Wilk,
‘Between Scripture and History,’ 202–4; De Sousa, Eschatology, 157.
61
Is it equivalent to ‫ ?ס ֵֹפ ר‬In what sense? As in Ezra ch. 7? Or is it the γραμματεύς
(‘scribe’) of the NT? Perhaps the rendering γραμματικοί in 33:18 plays a role too in
its meaning: cf. van der Kooij, Oracle of Tyre, 19: ‘it is reasonable to assume that our
translator is to be seen as a scholar (‘scribe’).’
62
Wilk, ‘Between Scripture and History,’ 204–8.
exegete with certain tendencies such as ‘actualisation,’ is a question for

others to decide; personally I have difficulty being persuaded.63
Another leading LXX-Isaiah scholar, Ronald Troxel, offers a portrait
that seems at first sight to accord rather well with the findings of this paper.
Troxel rejects van der Kooij’s ‘scribe,’ but on the basis of the rendering
of ‫ ספר‬by γραμματικοί in Isa 33:18, which van der Kooij sees as a self-
reference, Troxel builds an argument of his own. While he explicitly rejects
calling the translator a γραμματικός, he uses the term to develop a con-
nexion with the γραμματικοί of the Alexandrian Museum as a direct influ-
ence on the translator’s work, and to locate him within ‘the literati of the
ancient world.’64 We might expect then that the translator’s high-quality
Greek would be used to support this proposal. Troxel comes close to it, but
then qualifies it away: ‘The Isaiah translator appears to have been familiar
enough with literary Greek to formulate sentences in its style, suggesting
that he was among the well educated, although we must suspect that to
be true, in any case, for a person undertaking the task of translation.’65 In
that case all the LXX translators were ‘well educated.’ Earlier, the discus-
sion of particles, as already noted, is used by Troxel only to show that the
translator was interested in creating ‘smoother syntax for his Greek-reading
audience.’66 Furthermore, the link to the methods of the Homer scholars
of the Museum seems strained and is not really necessary to explain the
translator’s procedure.67 The idea at the heart of all this, that the choice
of rendering in 33:18 is significant and tells us something about the trans-
lator’s attitude to γραμματικοί, seems to me very speculative.68 In the end
Troxel’s portrait of the Isaiah translator is not well focused: a better appre-
ciation of the language could do much to sharpen it.
My main aim has been to show that the Isaiah translator’s Greek should
be classified as ‘literary’ and his language connected with other works
composed in that kind of Greek, both in the LXX and outside it. If I have
succeeded, the observation may also contribute to unfolding the secret of
the Isaiah translator and his methods.
63
Cf. Wagner, Sealed Book, 34 for a summary of proponents and critics.
64
Troxel, LXX-Isaiah, 290–1.
65
Troxel, LXX-Isaiah, 132.
66
Troxel, LXX-Isaiah, p. x.
67
Cf. Hugo, Review of Troxel, LXX-Isaiah (2008).
68
I agree with Pietersma’s discussion of the rendering in ‘A Panel Presentation,’ 354–5.
23
THE SAYINGS OF THE DESERT FATHERS:
THEIR EVIDENCE FOR LATE KOINE
GREEK
20151
Abstract
The Sayings of the Desert Fathers, in their Greek form, date from the fourth
to fifth centuries AD and are written in the Greek language of their time. They
contain both sayings and simple narratives, composed for the most part in every-
day, not literary Greek. These texts therefore have the potential to provide useful
evidence for the ongoing development of Greek in the late Koine period (IV–
VI AD). There is a shortage of evidence of this kind and the Sayings have
received little notice. The aim of this paper is to draw attention to the linguistic
value of the Sayings, concentrating on the Alphabetic collection. One Saying
(Paphnoutios 2) is sampled first, with observations on features of interest there,
then other features found elsewhere in the corpus are listed with brief comments.
The list is extensive but not exhaustive or systematic: a full study still needs to
be done. The objective of the paper is linguistic, not theological, nor does it aim
to resolve textual questions pending until a critical text of the Alphabetic collec-
tion is achieved.
1. INTRODUCTION
The Desert Fathers were those first Christian ascetics who withdrew
from the world and went out into the Egyptian desert to focus on the
spiritual life in solitude. Abba Antonios is still commemorated today as
the one who led the way. The Sayings are a collection of both memorable
sayings and simple narratives about the Fathers that were thought worth
remembering. They always contain some point of significance related to
1
Most of the material of this article was first presented at the ‘Images of Egypt’ conference
at Macquarie University, Sydney in April, 2004. I am grateful to Trevor Evans for the
invitation to present at the conference, and also to the students of St Andrew’s Theo-
logical College who over the years 1988–2000 joined me in reading the Sayings in the
‘Advanced Koine’ course.
the spiritual life and practice. They range in length from a line or two to
stories extending over two pages. In the form of the collection studied here,
the Greek Alphabetic, there are some 130 named monks, and the number of
Sayings attached to each ranges from one to over 200 in the case of Abba
Poimen.
In most cases, the material has an Egyptian setting, that is, almost all
of the Fathers are located in Egypt and the various sayings and narra-
tives take place in Egypt. The texts survive in a number of languages, but
it appears that the Greek form was the earliest, from which translations
were made into Syriac, Armenian, Sahidic, and so on. We know that some
of these monks were Egyptians and may not have spoken Greek, so some
translation into Greek is likely to have occurred at an early stage. The Greek
collection is in three forms, the Alphabetic, the Anonymous, and the System-
atic, with a lot of material in common. The Alphabetic arranges the material
by the names of the Fathers (or Mothers) with whom the saying or story
is associated; the Anonymous collects sayings that have no name attached;
and the Systematic arranges much of the same material by topics or themes,
21 in all. It appears that the Alphabetic collection came first. That is the
one on which this paper is based.2
Who first put this material together, and who did any necessary trans-
lating, say from Coptic into Greek, we do not know. But the nature of the
Greek itself indicates something. It is not learned or literary Greek, nor
is it entirely vernacular, but it is middle-level Koine Greek of a century
before the beginning of early Medieval Greek (600–1100).3 There is no
evidence of interference from Coptic, but biblical Greek, that is, the Greek
of the NT and the Greek OT (‘the Scriptures’), has a significant influence
on the language and there is much use of scriptural quotations and allu-
sions.4 Much of this influence could come, and in some instances clearly
does come, from liturgical Greek, which itself is deeply influenced by the
Greek of both Testaments.5 Apart from these influences, the narrative frame-
work and most of the Sayings material are in ordinary Greek of the period
and as a result are of great interest linguistically.
2
For the foregoing points see: Guy, Recherches; idem, Les Apophtegmes: Collection sys-
tématique; Ward, The Sayings, xvii–xxxi; Wortley, ‘How the Desert Fathers’; Rönnegård,
Threads and Images, 5–11, 197–8; Butler, Lausiac History, 1:208–15.
3
See Browning, Greek, 53, 55 for the term (‘early middle ages’) and dates.
4
See Rönnegård, Threads and Images.
5
See Lee, ‘LXX in Liturgy.’
THE SAYINGS OF THE DESERT FATHERS 375
2. TEXT AND TRANSLATIONS
The ruling edition of the Greek text of the Alphabetic collection is that
found in Migne’s great collection of patristic texts.6 It is basically a reprint
of an edition of 1677 by Jean-Baptiste Cotelier, derived from one twelfth-
century manuscript.7 A modern critical text of the Alphabetic is obviously
needed, but in the meantime we have no choice but to use Migne’s text.
There is also a more recent edition by Paschos, which is a convenient work
to use but is in essence the same text as in Migne.8 In the following dis-
cussion, Migne’s is the text quoted, with a reference to Paschos as well (the
numbering sometimes differs). Despite the lack of a proper critical edition,
linguistic studies can and should proceed on the basis of the text available
(as the translations have done), on the understanding that when a critical text
is produced, earlier comments on language might need some correction
or revision. The alternative is to delay such study indefinitely. Further-
more, observations on the language may themselves contribute to estab-
lishing the text when the task is undertaken.
There are modern translations of the Alphabetic collection, into English
by Sr Benedicta Ward; into French by Jean-Claude Guy and Lucien Reg-
nault; and into Modern Greek by Basileios Pentzas.9 In Migne’s edition
there is a Latin parallel version. The author is not stated, but presumably
it was Migne or Cotelier. This is not the same version as the sixth-century
translation of the Systematic collection to be noticed in a moment.
3. DATE
The date of the material is important for linguistic study, though precise
dating is not necessary: a date within a century or even longer is adequate.
6
Migne, PG, vol. 65. Text also available electronically in TLG (no. 2742).
7
Guy, Recherches, 8 n. 2, 13, 18–19. The MS is Paris grec 1599; Guy gives the date ‘XIIe siè-
cle’ without comment. Migne gives the title of his source at PG 65:71–72 (below heading).
8
Paschos, Τὸ Γεροντικόν (1981). As far as I am aware, there is no work in progress on a
critical edition of the Alphabetic. Guy’s critical edition of the Systematic reached com-
pletion in 2005 (Les Apophtegmes: Collection systématique). This includes an index of
words (3:251–463). [The research for this paper was done before Guy’s edition of the
Systematic appeared; otherwise it would have been possible (and perhaps better) to base
my study on it. Even so, there is not a great deal of difference in the sayings themselves
in the two texts.]
9
Ward, The Sayings; Guy, Les Apophtegmes: Série alphabétique; Regnault, Les sen-
tences; Pentzas, Εἶπε Γέρων. Ward’s translation, on which everyone relies, is not as accu-
rate as it is assumed to be (for an example, see n. 32). A revision is needed.
We have one indication of date in the fact that the MS of the Alphabetic
collection on which the first printed edition was based is from the twelfth
century. Moreover, there are still older MSS of this collection dating from
the tenth to eleventh centuries AD.10 This is useful to begin with, as estab-
lishing that there can have been no intrusion of late Medieval or Modern
features that arose in the language after the eleventh century. But we can
narrow the date much further. A terminus ante quem is provided by the
Latin translation of the Systematic collection made by two Popes, Pelagius I
and John III.11 This was made in the early sixth century, so the date of com-
pilation of the Systematic must precede the translation, and the compilation
of the Alphabetic preceded the Systematic. Guy finds a date around 530 AD
for both collections assured, and argues further that both can be dated to
480–490 AD.12
The history of the material before then is hazy. We do not know how
it came to be written down or how much adjustment (and translation)
there was between the original saying and its appearance in written form,
or at the final stage of compilation. Oral transmission is obviously likely
to have had a role. But we know that the first generation of monks belong
to the fourth century. Antonios is said to have gone out into the desert
around 285, and he died in 356. Makarios the Egyptian, the founder of
Scetis, overlapped with Antonios and died in 390. The last generation,
which included Arsenios (d. 449) and Poimen (d. after Arsenios), are in
the fifth century.13 So we can say with some confidence that the material
originates in the fourth to fifth centuries and that the compilation date
is late fifth or early sixth century, let us say around 500 AD. The Sayings
therefore could provide good evidence of Greek of the fifth century, right
in the middle of ‘Late Koine,’ that is, Greek of IV–VI AD.14
There is however the question of changes in the course of transmission,
as the texts were copied between 500 and the surviving MSS. First as
regards modernisation, that is, intrusion of more recent features of the
popular language, it is evident that this has had no major effect. To take
specific cases, the phonological change of the loss of unaccented initial
vowels (except ἀ-), which took place along with related changes in early
10
Guy, Recherches, 16.
11
On their dates see Guy, Les Apophtegmes: Collection systématique, 1:80, and for details
of editions see, e.g., Wortley, ‘How the Desert Fathers,’ 328.
12
Guy, Les Apophtegmes: Collection systématique, 1:80–3.
13
For these dates, see ibid., 47–8, 76, 78.
14
For the term see Lee, ‘Ἐξαποστέλλω,’ 113 [= Essay 12].
Medieval Greek, is not in evidence; nor is the loss of final -ν.15 The later
morphological changes to the noun and verb that might have intruded are
also not to be seen.16 All this does not rule out the possibility, even like-
lihood, that minor spelling changes or slips occurred. ‘Improvement’ of
the text, that is, from vernacular to more educated, is a different possibility
to consider, and eliminate. The presence of so many lower-register items
of vocabulary suggests that no, or at least very little, ‘improvement’ of
this kind occurred in transmission: words like νερόν (‘water’) and περνῶ
(‘cross over’) and others noticed below would be strong candidates for
replacement in that case. It seems rather that the character of the text was
fixed at the time of compilation around 500 AD.
Further evidence that the Greek text goes back to the early sixth cen-
tury is found in the fact that the sixth-century Latin translation mentioned
above appears to have been made from the same Greek text as the one
we have. That translation is of the Systematic collection, but much of
the same material appears in the Alphabetic, though in a different order,
and the Greek texts can be compared. In the Paphnoutios text discussed
below, for instance, the text in Migne (Alphabetic) differs only in very
minor details from that in Guy’s edition of the Systematic;17 the Latin ver-
sion can be seen to translate this Greek.18 The translation therefore indicates
the existence in the early sixth century of a Greek text of the Alphabetic
like the one we now have, even though the current MSS are later.
4. AIM
The aim of this paper is philological, to draw attention to the value

of the Sayings as a witness to Late Koine Greek. Their potential contri-
bution for this purpose has been generally overlooked in study of the his-
tory of the Greek language. While some of the vocabulary has been noted
in Lampe’s lexicon,19 and is often well observed in Sophocles’s,20 there
15
Browning, Greek, 57–8; cf. Horrocks, Greek, 274–7.
16
Browning, Greek, 58–60; 63–6; cf. Horrocks, Greek, 284–323.
17
Guy, Les Apophtegmes: Collection systématique, 3:19–20; XVII 15.
18
Migne, PL, 73:975A–B. The first few lines: Dicebant de abbate Paphnutio, quia non
cito bibebat vinum. Ambulans autem aliquando iter, supervenit in conventu latronum, et
invenit eos bibentes: cognovit autem eum qui primus erat latronum, et sciebat quia non
biberet vinum. Videns ergo eum de multo labore fatigatum, implevit calicem vini, et in
alia manu tenuit gladium evaginatum, et dicit seni: Si non bibis, occidam te.
19
Lampe, PGL.
20
Sophocles, Lexicon. See, e.g., s.vv. νερόν, φαγίον.
is no mention of the Sayings in a standard work such as Browning and

there appears to be just one in Horrocks.21 But they offer riches that ought
to be better known. In a short paper such as this, the thorough study that
needs to be done is not possible. My paper is to be seen as a first foray
into a subject that awaits a much larger enquiry.
Other sources of Greek of the same period may be mentioned in pass-
ing. There are of course numerous patristic works contemporary with the
Sayings. But their Greek is usually of a much higher literary level, often
following Attic models, and provides little information about spoken Greek.
One work of a similar linguistic level and similar in content, a little later in
date, is the Meadow of John Moschos.22 One known parallel between its
language and that of the Sayings is the occasional use of θέλω as a future
auxiliary.23 It too could do with more thorough study. But the other main
source of everyday Greek in this period, as in all earlier periods from the
time the evidence begins in the third century BC, is the body of documen-
tary texts preserved on papyrus and stone. They of course vary in linguistic
level, but parallels with the evidence of the Sayings are evident (e.g.,
νερόν, μισθοφορία, below). Making the links and exploiting the full value
of the combined evidence is something still to be done.24
In the following discussion, I start with a sample text and note the fea-
tures of interest occurring in it. This shows at a glance what a typical saying
is like and its linguistic potential. Then follow lists of features from else-
where in the Sayings, with brief remarks on their significance. A full dis-
cussion of each item and its place in the longer-term history of the language
cannot be attempted here. Many of the items will be recognised by students
of various aspects of Koine Greek usage, including that of the NT and even
21
Browning, Greek; Horrocks, Greek, 292 on κἂν ἕν at Migne, PG, 65:261B [not 45]
(the book has no Greek index). I do not find the Sayings (Ἀποφθέγματα …) in the list of
works covered in the Mega Lexikon.
22
Text in Migne, PG, vol. 87 (TLG no. 2856). See Horrocks, Greek, 253–6 for an extract
with comments. The Historia Monachorum (TLG no. 2744), is another similar text, earlier
in date (V AD), but somewhat more literary in language: text in Festugière, Historia
Monachorum; Engl. translation Russell, Lives of the Desert Fathers.
23
Lee, ‘Auxiliary θέλω,’ 23.
24
Millar (Greek Roman Empire, 16) presses the claims of the records of the fifth-century
Church Council debates as representing ‘for spoken Greek by far the best evidence for the
Greek language that survives from Antiquity.’ A record of what was spoken in the Coun-
cils they may be, but Millar overlooks the educated quality of the Greek used by the
learned bishops. For better representatives of ‘spoken Greek,’ see the letter of condolence,
POxy 16.1874 (VI AD) or the catalogue of complaints against a husband, POxy 6.903
(IV AD) – or the Sayings. (I am grateful to Trevor Evans and Genevieve Young-Evans
for bringing Millar’s remark to my attention.)
the LXX. Some are early instances of developments that were to become
established later in Greek and to continue into the modern language; others
look back to older usage on which they throw new or additional light.
5. A SAMPLE TEXT WITH FEATURES OF INTEREST
5.1 Paphnoutios 1/2 (P105, M377C–380A)25
Ἔλεγον περὶ τοῦ ἀββᾶ Παφνουτίου, ὅτι οὐ ταχέως ἔπινεν οἶνον.

Ὁδεύων δέ ποτε εὑρέθη ἐπάνω κοληγίου λῃστῶν, καὶ εὗρεν αὐτοὺς
πίνοντας οἶνον. Ἐγνώριζε δὲ αὐτὸν ὁ ἀρχιλῃστής, καὶ ᾔδει ὅτι οὐ
πίνει οἶνον. Καὶ θεωρῶν αὐτὸν ἀπὸ μεγάλου κόπου, ἐγέμισε ποτήριον
οἴνου, καὶ τὸ ξίφος ἐν τῇ χειρὶ αὐτοῦ, καὶ λέγει τῷ γέροντι· Ἐὰν μὴ
πίῃς, φονεύω σε. Γνοὺς δὲ ὁ γέρων ὅτι ἐντολὴν Θεοῦ θέλει ποιῆ-
σαι, βουλόμενος αὐτὸν κερδῆσαι, ἔλαβε καὶ ἔπιεν. Ὁ δὲ ἀρχιλῃστὴς
μετενόησεν αὐτῷ, λέγων· Συγχώρησόν μοι, ἀββᾶ, ὅτι ἔθλιψά σε. Καὶ
λέγει ὁ γέρων· Πιστεύω τῷ Θεῷ, ὅτι διὰ τὸ ποτήριον τοῦτο ποιεῖ
μετὰ σοῦ ἔλεος καὶ ἐν τῷ νῦν καὶ ἐν τῷ μέλλοντι αἰῶνι. Λέγει ὁ
ἀρχιλῃστής· Πιστεύω τῷ Θεῷ, ὅτι ἀπὸ τοῦ νῦν οὐ μὴ κακοποιήσω
τινά. Καὶ ἐκέρδησεν ὁ γέρων ὅλον τὸν κολήγιον, ἀφεὶς τὸ θέλημα
αὐτοῦ διὰ τὸν Κύριον.
They said concerning Abba Paphnoutios that he drank wine reluctantly.
Once when he was on a journey he found himself face to face with a
gang of robbers, and he found them drinking wine. The robber chief
recognised him and knew that he did not drink wine. Seeing that he
was tired from great exertion, he filled a cup of wine and with his
sword in his hand said to the elder, ‘If you don’t drink, I’ll kill you.’
The elder, knowing that he was about to perform a command from God,
and wishing to win over the robber, took it and drank. The robber chief
made a metania to him, saying, ‘Forgive me, Abba, for distressing you.’
The elder said, ‘I trust in God that because of this cup he will show
you mercy both now and in the age to come.’ The robber chief said,
‘I trust in God that from now on I will do no harm to anyone.’ And so
the elder won over the whole gang, by giving up his own will for the
Lord.
This is a simple narrative in ordinary Greek, with occasional biblical

flourishes. It is literate but not literary. Even a short extract like this is full
of interesting linguistic material. Some features are routine Koine usage,
25
Key: name of Abba + saying no. (x/x if different in P and M); P = Paschos page; M =
Migne, PG 65, col. no. This saying is also found in the Systematic collection at Guy, Les
Apophtegmes: Collection systématique, XVII 15.
though useful as confirming the continuance of the word or use to this

date; others open up a whole line of enquiry. The notes indicate the point
of interest, without going into full detail.
5.2 Features of Paphnoutios 1

οὐ ταχέως. This is evidence of ταχέως in the sense of ‘readily, without
hesitation,’ a weakening of the original sense ‘quickly.’ It could support
a similar interpretation of NT Gal 1:6; 2 Thess 2:2; 1 Tim 5:22. The attes-
tation noted in BDAG is quite limited.
εὑρέθη, ‘found himself.’ That is, ‘he came to be, arrived at,’ as in the
similar English expression, and French se trouver (cf. Guy’s translation
il se trouva). There is no actual finding involved. This has relevance to the
debated occurrence in Acts 8:40, where even Semitic influence has been
suggested: see BDAG, s.v. 1.b. There are others in the Sayings, e.g., Mile-
sios 2: P77, M297B.
ἐπάνω is normally ‘above’; here it is like Engl. ‘on top of,’ in effect ‘face
to face with.’ Similarly ἦλθόν ποτε ἐπάνω αὐτοῦ τρεῖς λῃσταί (Theo-
doros of Pherme 29: P41, M196A). Cf. from further afield: LXX Gen 18:2
ἰδοὺ τρεῖς ἄνδρες εἱστήκεισαν ἐπάνω αὐτοῦ. POxy 6.903.20 (IV AD)
καὶ ἐποίησεν | τὰς ἔξω θύρας αὐτοῦ ἐνκλισθῆναι ἐπάνω μου, ‘he
had the outer doors of it (the church) closed in my face.’26 There is more
to find out about the history of this interesting word, now standard Mod.
(ε)πάνω.
κολ(λ)ήγιον, a Latin word (collegium, ‘college’; ‘guild,’ etc.) borrowed
into Greek. LSJ cite two older examples; Lampe, PGL has others con-
temporary. The meaning ‘gang,’ not in Latin covered by the OLD (to end
II AD), is clearly attested here.
ἐγνώριζε. The modern meaning ‘recognise’ (= ἀνα-) fits best; similarly
Kronios 5: P60, M249C. No instances are recorded in LSJ; Lampe, PGL;
BDAG.
θεωρῶν. The standard Koine use continues.
ἀπό. A special use of the preposition: ‘just arrived from’; similarly ἦν
γὰρ ἀπὸ καμάτου πολλοῦ (Makarios 4: P65, M264C). This is the same
as in Mark 7:4 ἀπ᾽ ἀγορᾶς ‘when they come in from the agora,’ on which
there has been (unnecessary) debate: see BDAG, s.v. ἀγορά.
ἐγέμισε. Standard Koine use of the verb applied to filling anything,
not just ships, as originally.
26
I was led to this last example by LSJ (s.v. ἐπάνω I.3), though their meaning in the
presence of is not right.
φονεύω. Present with future meaning: ‘I kill you’ = ‘I will kill you.’
(Cf. below.)
θέλει. ‘Wanted, wished’ does not fit. Rather, the sense is: ‘was about
to do/would be doing.’ It is an instance of θέλω beginning to function as
a future auxiliary. (Cf. below on the future tense.)
βουλόμενος. At an earlier time (e.g., in the NT) this was the more
formal or stylish equivalent of θέλω in the meaning ‘wish, want.’27 While
βούλομαι is not uncommon in the Sayings (c.18 times), θέλω remains far
commoner (200+). A fuller study of the two in this period would be inter-
esting. Perhaps βούλομαι had made a comeback under learned influence.
κερδαίνω (twice). What meaning is intended? Originally the verb means
‘gain (money), profit, get for oneself’; hence ‘gain (for God)’ > ‘win
over, convert’ (as 1 Cor 9:19, 20, 22); or ‘win over’ could be taken simply
to mean ‘win the favour of, avert the hostility of.’ The meaning ‘convert’
is perhaps more likely. A study of the semantic history of the word, with
its modern descendant κερδίζω, would shed more light.28
μετενόησεν. Not ‘expressed repentance to him,’ but a physical action,
‘made a metania to him.’ Cf. the frequent expression ἔβαλε/ ἔβαλον μετά-
νοιαν, ‘made a metania’ (Kronios 5: P60, M249B +). Similarly Lampe,
PGL, s.v. μετανοέω C.
συγχωρῶ, ‘forgive’ is post-NT usage, but becomes significant in liturgi-
cal and patristic language (see Lampe, PGL, s.v. 5; Sophocles, Lexicon,
s.v.).
θλίβω. Here it shows a weakened meaning, like ‘trouble, bother,’ not
‘afflict’ etc., as earlier. Ward’s ‘made you unhappy’ is on the right track.
Modern θλίβω may have a similarly mild sense.
ποιεῖ. Another example of present in future sense.
ποιεῖ μετὰ σοῦ ἔλεος. An OT Greek expression derived from Hebrew.
‘Do mercy with’ = ‘show mercy to.’ This and the following καὶ ἐν τῷ νῦν
καὶ ἐν τῷ μέλλοντι αἰῶνι are biblical and liturgical in origin and flavour.
ἀπὸ τοῦ νῦν is also a biblical phrase: in the LXX c.20 times, notably in
the Psalms, and in the NT 8 times including the extremely familiar ἀπὸ
τοῦ νῦν μακαριοῦσί με πᾶσαι αἱ γενεαί (Luke 1:48).
οὐ μή + fut., aor. subj. This construction was originally Classical, but is
found only sporadically in Koine Greek. Its adoption for use in the LXX,
27
See Shipp, MGE, 166–7.
28
Ward (The Sayings, 202) translates the first occurrence of κερδαίνω in our text as ‘to
win the confidence of’ and the second as ‘converted’: but the meaning ought to be the
same in both. I do not think the meaning ‘avoid, spare oneself,’ that is found in BDAG
(s.v. 2) for Acts 27:21, and in LSJ (s.v. III), and might fit here, is soundly established.
however, where it is very frequent as a solemn-sounding negative, ensured

its entry into the NT and liturgical Greek.29 It probably continues to have
a ‘biblical’ flavour in the Sayings, though it is not a ‘Hebraism.’ There are
many examples.
διὰ τὸν Κύριον. Here διά = ‘for the sake of,’ a good example of exten-
sion in the meaning of διά, leading to Mod. για, with a range as wide as
Engl. ‘for.’ Cf. already in Mark 2:27 τὸ σάββατον διὰ τὸν ἄνθρωπον
ἐγένετο.
6. SIGNIFICANT FEATURES ELSEWHERE IN THE SAYINGS
6.1 Morphology
ἀναβάζω, ‘bring up’ = ἀναβιβάζω, cf. Mod. ανεβάζω ‘bring up’

(Makarios 30: P69, M273C).30
ἀφίω (or ἀφιῶ), δίδω (-ῶ), τίθω (-ῶ) are frequent, providing further
evidence of the new thematic forms that appeared early in Koine Greek.31
ἔσο ‘be’ (imper.): ἅλατι δὲ ἠρτυμένος ἔσο ἐν Κυρίῳ, ‘be seasoned
with salt in the Lord’ (Or 13: P127, M440C; also Antonios 6: P2, M77A).
This echoes Col 4:6 ὁ λόγος ὑμῶν... ἅλατι ἠρτυμένος.32 Only rare
glimpses of this replacement for ἴσθι are available. Some few examples
occur in literature, starting with Plutarch.33 PMert 1.45r.6 (V–VI AD) ἔσο
δὲ καὶ αὐτὸς μέρος τῆς καθ[- - -] is a possible instance.
ληκύνθιον, ‘flask,’ for ληκύθιον: not recorded in LSJ or Lampe,
PGL, but in Sophocles, Lexicon (Makarios 3: P64, M261A). There is
no reason to doubt it is a real form, influenced by other words ending in
-ίνθιον.34
μελανός, ‘black,’ for older μέλας, continued in Modern Greek (Moyses 4,
8: P72, M284B, 285A). See LSJ; Lampe, PGL; Sophocles, Lexicon.
ῥιγᾶτε, ‘shiver,’ originally ῥιγέω, -όω (Makarios 2: P64, M260D).35
29
See Lee, ‘Some Features,’ 18–23.
30
On loss of reduplication see Gignac, Grammar, 2:242–3.
31
Gignac, Grammar, 2:380–3.
32
Ward has gone quite wrong: ‘Let your salvation be founded in the Lord.’ This seems
to be a flawed version of the Latin transl. in Migne: Sale autem conditus sis in Domino.
33
LSJ, s.v. εἰμί, introd. In Thackeray, Grammar, 257 n. 1, it is noted in Symmachus
(without ref.).
34
See Buck-Petersen, Reverse Index, 73.
35
See Gignac, Grammar, 2:363–5 on the merging of inflectional types in contract
verbs.
6.2 Vocabulary: common Koine
αἴρω, ‘take’: δεῦρο σὺ ἆρόν με πρὸς τὸν βασιλέα (Poimen 109:

P95, M349A +).
ἀκμήν, ‘still’ (Sisoes 7: P110, M393B; Arsenios 24: P7, M93C +).
ἀφανίζω, ‘ruin, spoil’ (Karion 2: P61, M252C).
δέρω, ‘hit’ (Zacharias 4: P36, M180B).
διαφέρω, ‘belong to’ (Karion 2: P60, M252A).
εὑρίσκω, ‘be able’ (Poimen 111: P95, M349C +).36
ἴδιος as simple possessive (e.g., Karion 2: P60, 61, M249D, 252A).
καιρός, ‘period of time’ (Karion 2: P60, M249D).
παραβάλλω, ‘come to, visit’ (Antonios 12: P2, M77C + many).
πλήν, ‘but’ (Matoes 7: P75, M292B +).
πληροφορῶ, ‘inform’ (Eucharistos: P32, M169A).
τρώγω, ‘eat’ (Sisoes 31: P113, M401C +).37
φεύγω, ‘go away’ (Sisoes 27: P113, M401A; Kopris 3: P61, M252D).
φθάνω, ‘reach, arrive at’ (Agathon 13: P13, M113A +).
6.3 Vocabulary: new examples of uncommon or unknown words and

uses
ἀρχάριος, ‘novice, beginner’ (Esaias 1: P36, M181A).

βράζω / βράσσω, ‘boil’ (food) (Esaias 5/6: P37, M181C).
γεοῦχος, ‘landowner’ (Esaias 4/5, P36, M181B).
γραΐς, ‘old woman’ (Sarra 4: P119, M420C +).
δράξ, ‘handful’ (Esaias 5/6: P37, M181C).
ἑαυτόν as 1st pers. sing. reflexive, ‘myself’ (Simon 2: P117, M412D).
ἐπαίρω, ‘take away’ (Antonios 5: P1, M77A). A further contribution
to the evidence for the derivation of Mod. παίρνω from ἐπαίρω (not
ἀπαίρω).38
θεωρία, ‘vision’ (Zacharias 4: P36, M180B).
καθάπαξ / κατὰ ἅπαξ, ‘each time’ (Kronios 5: P60, M249B; Poimen 93:
P93, M344D).
κρούω, ‘run,’ ‘spring up’: οἴδατε καὶ τὸ χωρίον τοῦ δεῖνος ὅπου ὁ
ποταμὸς κρούει; (Makarios 30: P69, M273C). There is no record of this
use elsewhere. Cf. κρουνός, ‘spring.’
36
Cf. Lee, ‘A Non-Aramaism.’
37
Cf. Shipp, MGE, 540–1.
38
See Shipp, MGE, 243. Exx. from other sources in Lampe, PGL, s.v. ἐπαίρω B, and
Sophocles, Lexicon, s.v. ἐπαίρω.
λαγύνιον, ‘flagon’ (Moyses 13: P73, M288A).

μισθοφορία, ‘rent; payment,’ here a special payment in return for
freedom (Mios 2: P78, M301C). Lampe, PGL (service for wages) and
Ward (‘his wages’) are on the wrong track. The evidence in late documen-
tary texts (BGU 12.2139.12, 13 [432 AD], 17.2696.27 [616 AD] +) helps
a better understanding of this old but rare word. Cf. LSJ, Suppl. (1996),
s.v. (rent, for BGU 12.2139), Sophocles, Lexicon (‘payment by a freedman
to his former master,’ for Mios 2).
νερόν, ‘water’ (Ioannes o Kolobos 7: P45, M205B). The standard
Mod. word νερό already in place. If Ioannes’ own words are preserved,
we could date this example to early V AD (he died in 409, according to
Guy). Other evidence around this time is secure but meagre.39
περνῶ (-άω), ‘cross over’; ‘pass through’: φύσει οὐ περνᾷ τὴν φάραγγα
ταύτην, καὶ βλέπει με (Or 6: P126, M440A). Here we have a rare early
attestation of what would become a common Modern verb.
ποῖον; ‘which? what?’ (= older τί;): ἕν ἐστι μόνον ἐν ᾧ νικᾷς με.
Λέγει … ποῖον τοῦτο; ‘There is only one thing in which you beat me.’
He said, ‘What’s that?’ (Makarios 11: P66, M268C +). Evidence of pro-
gress towards the standard modern use. The beginnings are much earlier:
see LSJ, s.v. ποῖος IV; BDAG, s.v. ποῖος 2.
σαλός, ‘half-witted’ (Moyses 8: P72, M285A).40
σπλαγχνίζομαι, ‘have pity’ (Moyses 18: P74, M289A).
συλῶ (-άω), ‘steal; rob (someone)’ (Saio: P119, M420B).
τελῶ pass., ‘become ripe’ (Arsenios 19: P6, M92C).
τίποτε, ‘something, anything’ (Sisoes 7: P110, M393B). Mod. τίποτα.
φαγίον, τό, ‘food’ (Sisoes 52: P115, M408B).
ψιχίον, τό, ‘crumb’ (Phokas 2: P124, M433C). Found in the NT
(Matt 15:27; Mark 7:28), but not well attested (cf. BDAG; Lampe, PGL).
6.4 Syntax
6.4.1 The future tense. New ways of expressing futurity are appearing:
(a) The present indicative, which is frequent (e.g., Paphnoutios 1/2 quoted
above).
(b) θέλω as future auxiliary: πῶς οὖν θέλεις σῖτον λαβεῖν, μὴ θερί-
σας; ‘how then will you receive grain if you have not harvested?’
39
See note on POxy 56.3865.35 (late V AD) by Sirivianou, with refs. there. For the date
of Ioannes: Guy, Les Apophtegmes: Collection systématique, 67.
40
Cf. POxy 56.3865.57 with Sirivianou’s note.
(Esaias 4/5: P36, M181B); ἤθελε σχίσαι ‘he was about to tear it
in two’ (Arsenios 29: P8 M97C; simil. Makarios 1: P64, M260B;
Silouanos 1: P115, M408C +).41
(c) ἔχω + aor. infin. (Sisoes 15: P111, M397A; Arsenios 24: P7,
M96A +).42
6.4.2 Other auxiliary uses:43

(a) εἶχον + infin. = ‘would have’ (Makarios 21: P67, M272A).
(b) ἔχω + pres. part.: ἔχεις πολλὴν ὥραν κρούων; ‘have you been
knocking for a long time?’ (Arsenios 27: P7, M96C).
(c) εἰμι + part.: ἤμην ἀναβαίνων, ‘I used to go up’ (Poimen 110: P95,
M349C).
6.4.3 Aoristic perfect: ἔλεγον περὶ τοῦ Ἀββᾶ Μακαρίου τοῦ μεγάλου,
ὅτι γέγονε καθώς ἐστι γεγραμμένον θεὸς ἐπίγειος, ‘They said about
Abba Makarios the Great that he was (or became), as it is written, God on
earth’ (Makarios 32: P69, M273D; Arsenios 32: P9, M100C; Karion 2:
P61, M249D +).
6.4.4 Other features

τό as relative: ἐκεῖνο τὸ ἐφάγομεν τῆς ἀγάπης ἦν, ‘that (food) which
we ate was for love’ (Silouanos 1: P115, M408C).
εἷς = ‘a’: καὶ ἰδοὺ μία χήρα καλαμωμένη ἦν ὀπίσω ἡμῶν, ‘and there
was a widow gathering reeds behind us’ (Makarios 6/7: P65, M265A;
simil. Sisoes 47: P115, M405C +).
ἐπάνω + gen. as prep.: frequent.
ὀπίσω + gen. as prep.: frequent.
αὐτός as demonstr.: ἀπῆλθεν εἰς ἀγρὸν καὶ ἐκρύβη … οἱ πατέρες
… φθάσαντες εἰς αὐτὸν τὸν ἀγρόν …, ‘he went into a field and hid. …
The fathers, arriving at that field. …’ (Isaak 1: P51, M224B +).44
ἵνα temporal: ἔρχεται καιρός, ἵνα οἱ ἄνθρωποι μανῶσι, ‘a time is
coming when men go mad’ (Antonios 24/25: P4, M84C). Cf. NT exam-
ples noted in BDAG, s.v. ἵνα 2.d.
ἵνα imperatival or jussive: εἰ δὲ ἔτι ζῶντα, ἵνα μεταλάβω τῆς ἁγίας κοι-
νωνίας, ‘but if (you find me) still alive, let me receive Holy Communion’
41
Cf. Lee, ‘Auxiliary θέλω.’
42
Cf. Browning, Greek, 33; Horrocks, Greek, 130.
43
Cf. Browning, ibid.
44
On the development of αὐτός see Horrocks, Greek, 128–9.
(Phokas 2: P124, M433A +). This phenomenon has a long history,

reaching back to NT times and even earlier, and resulting in imperatival
uses of να + subj. in the modern language.45
6.5 Idioms and colloquial expressions
ἄρχομαι + infin., e.g., ὡς δὲ ἤρξατο ἰδιάζειν αὐτόν, ‘when he came

to be/was alone with him’ (Makarios 3: P65, M264A +). This is relevant
to the debate about ‘pleonastic’ ἄρχομαι in the NT, allegedly due to
Semitic influence. A proper study of the evidence outside the NT has not
been done.46
βάλλω + infin. ‘begin (to)’: ἔβαλε κράζειν ‘he began to shout’ (Phila-
grios: P124, M436A +). Cf. Lampe, PGL, s.v. C.5.
εἰς τὰ ἴδια ‘to one’s home’: ἀπῆλθε μετὰ χαρᾶς εἰς τὰ ἴδια, ‘she went
home with joy’ (Arsenios 28: P8, M97B). Cf., e.g., John 16:32 ἔρχεται
ὥρα … ἵνα σκορπισθῆτε ἕκαστος εἰς τὰ ἴδια.
ἐφέτος / ἐπέτος adv., ‘this year’: εὐχαριστοῦμεν τῷ Θεῷ, ὅτι ἔβρεξεν
ἐπ᾽ ἔτος [sic M] πολλά (Ioannis o Kolobos 10: P45, M208A; Or 4: P126,
M437C). This is the Mod. (ε)φέτος, ‘this year’ making its appearance.
Other evidence around this time can be seen in Sophocles, Lexicon, s.v.
ἐπέτος, and earlier in LSJ, Supplement (1996), s.v. ἐφέτος, improving the
LSJ entry.
ἦν ἰδεῖν αὐτὸν ὡς ἄγγελον ‘he was like an angel to look at’ (Kronios 5:
P60, M249B).
ἰδού is frequent, and not only in the sense of ‘behold, look’ but also
‘here is/was, there is/was’ (Arsenios 36: P10, M101C; Ammonas 6: P15,
M120D +).
κέλευσον, ‘give the word,’ in effect ‘please allow’: κέλευσον ἵνα
κἀγὼ βρέξω ἐμαυτῷ, ‘please let me also soak (some palms) for myself’
(Makarios 4: P65, M264C).
λοιπόν, ‘then, therefore’ (Sisoes 3: P110, M392D).
μὴ οὐ …; strengthened form of οὐ introducing a question expecting the
answer ‘yes’: μὴ οὐκ εἰμὶ μήτηρ ὑμῶν; ‘am I not your mother?’ (Poi-
men 76: P92, M341A). NT instances like 1 Cor 9:4 are clearly similar:
μὴ οὐκ ἔχομεν ἐξουσίαν φαγεῖν καὶ πεῖν; cf. BDAG, s.v. μή 3.a.
45
See, e.g., BDAG, s.v. ἵνα 2.g; Mandilaras, Verb, §§585–9.
46
There are numerous examples in the popular I AD text Vita Aesopi G. [Hunkin’s 1924
study ‘Pleonastic ἄρχομαι’ goes a good way towards correcting old misunderstandings.]
ὅλως in several idiomatic senses: ‘at all’ (Agathon 14: P13, M113A;
Isidoros the Priest 1: P55, M236A); ‘just’ (Esaias 5/6: P37, M181C;
Poimen 154: P99, M360C); ‘really, actually’ (Poimen 76: P92, M341B).
The interpretation of the NT examples at 1 Cor 5:1 and 15:29 might be
aided by this evidence (cf. BDAG, s.v. ὅλως).
οὗτος οὐκ ἔχει πρᾶγμα, ‘he has nothing to do with it’ (Makarios 1:
P64, M260A).
ποῦ ἔνι μοι …; ‘how can I …?’ (Makarios 3: P64, M261B).
πῶς ὁ κόσμος; ‘how’s the world?’ (Makarios 2: P64, M260D).47
πῶς τὰ κατὰ σέ; ‘how are things with you?’ (Makarios 3: P65, M264A).
σωθείης, ‘good health!’ (Makarios 3: P64, M261B; +).
τί ἔχεις, πάτερ; ‘what’s up, father?’ (Sisoes 47: P115, M405C).
τόπον ἐκ τόπου, ‘to one place after another’ (Synklitiki 6: P120,
M424A). The remarkable longevity of this expression is demonstrated.
It is first attested in a III BC papyrus (PLond 7.2049.8), and sporadi-
cally thereafter. The same syntactic pattern is seen in ἡμέραν ἐξ ἡμέρας
(LXX etc.) and other phrases.48
φύσει as an interjection or particle (cf. Engl. ‘naturally’), with a range
of uses hard to define precisely: ‘in fact,’ e.g., καὶ πόσον χρόνον ἔχεις
ὧδε; ὁ δὲ ἔφη· Φύσει, Ἀββᾶ, ἔχω ἕνδεκα μῆνας ἐν τῷ ὄρει τούτῳ,
‘… Actually, Abba, I’ve been on this mountain eleven months’ (Sisoes 7:
P110, M393B); ‘really,’ ‘truly,’ e.g., ἠρώτησαν αὐτὸν οἱ ἀδελφοί· Πῶς
ἡ πόλις; ὁ δὲ εἶπε· Φύσει, ἀδελφοί, ἐγὼ πρόσωπον ἀνθρώπου οὐκ εἶδον,
εἰ μὴ μόνον τοῦ ἀρχιεπισκόπου, ‘… Truly, brothers, I did not see the
face of any man except the Archbishop’ (Isidoros 8: P50, M221B; simil.
Arsenios 25: P7, M96A +). Cf. Lampe, PGL, s.v. φύσις II.D.2; Sophocles,
Lexicon, s.v. φύσις 5. This idiomatic φύσει seems to have arisen in Late
Koine, not before. There does not appear to be any direct modern descend-
ant of φύσει; but there is φυσικά.49 A full study would be useful.
The ancient ‘onoma rule’ is still being followed, i.e., ὄνομα without
article and the possessor of the name in dat., in naming-constructions of
the type ‘the name of x is N’: ὄνομα δὲ αὐτῇ Παησία (Ioannis o Kolo-
bos 40: P49, M217B).50
47
This expression (in the form πῶς εἶναι ὁ κόσμος;) has continued in use right up to
today. I am grateful to Fr Miltiades Chryssavgis for confirming this and demonstrating
it in conversation with others (on 11th April, 2004, Pascha).
48
[See discussion of LXX exx. of this pattern in Lee, Greek of the Pentateuch, 141–5.]
49
I thank Emmanuel Roumanis for drawing my attention to this.
50
See Lee, ‘Onoma Rule.’
CONCLUSION
The linguistic value of the Sayings material is established with little

difficulty. We have been able to observe a wide range of features, from
basic Koine Greek continuing in use to striking instances of new develop-
ments. All of these help to fill out the record of the language in this period,
a record that is limited for this level of Greek. Besides completing the
record for possible inclusion in lexicons and other reference books, value
will be found in this evidence for studies of particular features, such as
the development of the future tense. The study of the history of the Greek
language is not yet complete by any means.
Two further observations may be made. This study shows that there are
untapped sources of evidence still available. Periods for which we thought
we had little information turn out to be witnessed to by texts into which
little or no enquiry has been made. Work needs to continue on these less
obvious sources. Essential to further progress is the combining of evidence
from all available sources and eras, even those distant in time. Electronic
databases and searching tools have changed the nature of such work and
made it immensely faster and easier than ever before.
The second point concerns the recording of the results. How will it
happen? Printed lexicons (and grammars) take many years to be updated
and even then cannot include all the material available. Recording of data
is haphazard and slow. New observations are scattered through multi-
ple publications and easily missed. Moreover, there is as yet no lexicon that
is devoted to the Koine Greek period: Lampe’s lexicon is focused on theo-
logical vocabulary and leaves the rest to be taken care of by LSJ; LSJ
itself is focused on the Classical period, with haphazard coverage of later
Greek;51 the Mega Lexikon covers everything from Classical to Modern,
but has many gaps (and is unlikely to be updated); the excellent lexicon
of Kriaras begins its coverage later; and the specialist lexicons of the NT
and LXX are focused on what concerns them. But the era of these works
is coming to an end. Electronic databases will serve the same purpose in
a far more efficient way. This is not to say that printed lexicons will cease
to exist, only that the major storehouse of lexicographical data will be an
electronic database. This is a development not yet achieved, but one to
be sought and welcomed.52
51
DGE provides better coverage of the post-Classical period than LSJ but still does not
cover Koine Greek in depth, and has a long way to go to reach completion.
52
Cf. Lee, ‘Lexicographical Database.’
24
THE VOCABULARY OF THE SEPTUAGINT
AND DOCUMENTARY EVIDENCE
2016
Abstract
This chapter surveys concisely (a) the history of study of the relation between
documentary texts and the LXX; (b) the nature and value of the documentary
evidence; and (c) practical aspects of accessing the evidence. Three examples are
then presented, namely, ποτίζω, εὐίλατος, and ἀγαπῶ, showing how the docu-
mentary evidence impacts on our understanding of these LXX words. A consid-
eration of future directions completes the chapter.
1. INTRODUCTION
The Septuagint translators translated their Hebrew Bible into Greek.

That Greek was the Greek of their time, the third to first centuries BC.
Our knowledge of it depends on evidence in surviving Greek texts. Much
of that evidence comes from sources that have long been known and have
formed the basis of the material in our lexicons, namely, literary texts of
both the translators’ time and the preceding period of ancient Greek (Clas-
sical Greek). These texts comprise a range of genres, styles, dialects, and
levels of Greek, from the Epic of Homer to the historical prose of Polybius.
Everyday language is not well represented, though much can be gleaned
from writers like Plato and Aristophanes. The literary sources may provide
a fairly accurate picture of the usage of a word in the Greek known to the
LXX translators, and then again may not.
The discovery in modern times of documentary texts brought forth a
different and illuminating source of evidence. Documentary texts are those
inscribed or written on stone or papyrus, surviving from the day they were
set down. The majority are post-Classical, though there is a large body of
inscriptions that are earlier. They are mainly not literary but official or
everyday, using the kind of Greek appropriate to their purpose. They thus
give access to another dimension of the language, and fill out much of
the incomplete picture that we worked with before. The absorption of

this new evidence into our reference books is still in progress today and
will be ongoing as more documentary texts are found and published.
They have made, and will continue to make, a major contribution to our
knowledge of LXX Greek.
2. HISTORY OF STUDY
Investigation of the documentary texts in relation to the LXX began in

earnest with Adolf Deissmann, shortly after the first major finds of papyri.
But there were forerunners. A full list of those who earlier made the
connection would be tedious, but the names of Friedrich Sturz,1 Heinrich
Thiersch,2 and Heinrich Anz3 deserve mention. All had little documen-
tary material to work with, but realised its importance for the LXX. John
P. Mahaffy, the first editor of the Petrie papyri, contemporary with the
LXX Pentateuch,4 was struck by the similarities.5 Henry B. Swete also,
following up Mahaffy’s insight, pointed out a number of LXX words found
in PPetrie and the just-published BGU, and quoted in full a PPetrie text.6
These were early straws in the wind.
Deissmann’s researches were not systematic but a collection of the most
telling data he could find to demolish the notion of a ‘Biblical Greek,’
which had long held sway.7 He pointed out documentary parallels to the
language of both the NT and the LXX, but the NT was the ultimate target
and grows more prominent in the later studies.8 Deissmann drew on what
was currently available, both inscriptions and papyri, and though they were
few in comparison with subsequent discoveries, they were usually enough
to establish his point. His work was a breakthrough at the time: it showed
that the documentary texts were the evidence needed to bring the lan-
guage of the NT and LXX out of the isolation that had been imposed on
them. But there was a long way to go to exploit these sources fully, and
1
Sturz, De dialecto (1808).
2
Thiersch, De pentateuchi versione (1841), 87–90.
3
Anz, Subsidia (1894).
4
Mahaffy, Flinders Petrie Papyri.
5
See Swete, Introduction, 21 for quotations from Mahaffy and refs.
6
Swete, Introduction, 292–3.
7
Deissmann, Bible Studies (transl. of Bibelstudien, 1895 and Neue Bibelstudien, 1897);
idem, Light from the Ancient East (transl. of Licht vom Osten (4th ed., 1923).
8
On the LXX (with NT) see, in particular, Bible Studies, 86–169 (about 70 lexemes); with
more focus on the NT, 194–267 (= Neue Bibelstudien), and, generally, Light (LXX 81–
116 passim).
THE VOCABULARY OF THE SEPTUAGINT AND DOCUMENTARY EVIDENCE 391
some of Deissmann’s conclusions have not worn well, especially his view
that the Greek found in the NT was the language of ‘simple, unlearned
folk.’9 Even his statement that the language of the Seventy was ‘the Egyp-
tian Greek of the period of the Ptolemies’ needs considerable nuancing.10
James Moulton took up the challenge of gathering documentary parallels
to the vocabulary of the NT, and with his (eventual) collaborator George
Milligan produced the now-famous ‘MM.’11 As Frederick W. Danker puts
it, ‘Their work helped open up even more the curtains that Deissmann had
drawn aside to expose an exciting new world for New Testament explor-
ers.’12 But MM had limited application to the LXX: its focus was the NT,
even if LXX vocabulary was often mentioned along the way. And there
were other limitations: it is not a lexicon, and the analyses of meaning are
not well done; it deals with a selection of NT words, not the whole vocabu-
lary; it draws mostly on papyri, not inscriptions; and its material is inevi-
tably out of date and incomplete now.13 The ‘new world’ has expanded to
undreamt-of horizons. Those who use MM today need to be aware that it
must be supplemented by further investigation, and that its conclusions are
not necessarily reliable.
An ‘MM’ for the LXX was on no one’s agenda when MM reached
completion in 1929. Observations on selected parallels may be found in
Joseph Ziegler’s 1934 study of the book of Isaiah,14 and in a survey article
by Orsolina Montevecchi in 1964.15 That was the scene when the present
author’s 1970 dissertation on the LXX Pentateuch was undertaken.16 Its
aim was to combat once more the idea of a special ‘Jewish Greek,’ which
had shown a resurgence, by gathering evidence for the links between the
vocabulary of the Pentateuch and the contemporary language. Use was
made especially of third-century BC papyri from Egypt, which had grown
in quantity but been barely exploited. Only a selection of words and uses
was studied, namely, new developments in Koine Greek, as being especially
9
Deissmann, Light, 62; cf. 143–5 (‘Book of the People,’ etc.).
10
Deissmann, Bible Studies, 70. On the one hand, there is variety within the Greek of
Ptolemaic Egypt, and it is not a distinct entity from Greek elsewhere; on the other, the
LXX translators made adjustments to the language to suit their purpose and chose dif-
fering styles and literary levels.
11
MM; originally published in 5 vols. in 1914–1929, with a one-volume edition in 1930.
12
Danker, Multipurpose Tools, 117.
13
See Hemer, ‘Towards a New MM’; Horsley, New Documents, 5:83–92; Lee, History,
123 with n. 9.
14
Ziegler, Untersuchungen, 175–212. See assessment of Ziegler in van der Meer, ‘Papy-
rological Perspectives,’ esp. 109–10.
15
Montevecchi, ‘Continuità ed evoluzione.’
16
Lee, Lexical Study.
good indicators of the translators’ familiarity with the Greek of their time.
Widening the field to other vocabulary and other books of the LXX remained
to be done.
The years following saw minor studies on selected words or parts of the
LXX vocabulary by Raija Sollamo,17 Hélène Cadell,18 and the Italian papy-
rologists Montevecchi, Francesco Vattioni, and Anna Passoni dell’Acqua.19
More recently there has been renewed activity in short studies by this
author20 and Trevor Evans,21 and significant results are being obtained in
Michael van der Meer’s work on Isaiah, notable for its close engagement
with the papyrus documents (and even pictorial remains).22 All this has con-
tinued to show the value of the documentary evidence for elucidating the
vocabulary of the LXX, but there has been no systematic exploration, and
a means of gathering the results of these piecemeal studies into one place
has not yet emerged. The two recent lexicons of the LXX, by Lust, Eynikel,
and Hauspie and Muraoka, have to some extent performed this role, noting
or using any material up to the date of compilation.23 But the rate of pro-
duction has outstripped the ability of any printed work to keep an up-to-
date record. These lexicons, moreover, designedly made no attempt at
new researches into the documentary evidence, but took in what was on
hand, rightly considering further work on that front a task for the future.
3. NATURE AND VALUE OF THE EVIDENCE
Documentary texts were engraved on stone, or written on papyrus,

ostraka, wooden tablets, and other material. They survive from the time
when they were written or inscribed, and are in varying states of preser-
vation, from complete to very fragmentary, from perfectly legible to illegi-
ble. Decipherment, i.e., simply reading them, is a specialist task, because
of differences in the writing of Greek as compared with today, as well
as other factors. The range of dates is from the late eighth century BC for
inscriptions and from the late fourth century BC for papyri, down to the
17
Sollamo, Semiprepositions, 308–51 (parallels in documents and literary texts).
18
Cadell, ‘Vocabulaire de l’irrigation’; ‘Vocabulaire de la législation.’
19
Detail in van der Meer, ‘Papyrological Perspectives,’ 110; see also HDM, La Bible
grecque, 243.
20
Lee, ‘Thirty Years On’; ‘Ἐξαποστέλλω’; ‘Lexicographical Database’. Note also Lee,
‘Aposkeuê in the Septuagint,’ and O’Connor and Lee, ‘A Problem.’
21
Evans, ‘Court Function of the Interpreter.’
22
Van der Meer, ‘Trendy Translations’; ‘Bridge over Troubled Waters?’; ‘Papyrological
Perspectives.’
23
LEH, Lexicon; Muraoka, Lexicon.
end of the Koine period (600 AD) and beyond. The find locations also
range greatly and the extremes are remarkable, but it is sufficient to say
that Greek inscriptions have been found in quantity all over ancient Greece,
Asia Minor, Egypt, and the rest of the eastern Mediterranean as well as
Rome, while papyri have survived mainly in upper Egypt, with a small
number in Syria, Palestine, and elsewhere; ostraka also come mainly from
Egypt, though not all. The number of inscriptions is in the thousands;
papyri number more than a million, of which less than a tenth have been
published.24
The nature of the documents is endlessly variable and any generalisa-
tion is liable to be contradicted by a particular text, but a summary will be
useful. The inscriptions tend to be of a formal and public character – not
surprisingly, since they were set up for public viewing. They include: decrees
of civic bodies, treaties, honorific decrees, dedications, epitaphs, manu-
missions, sacred laws, records of benefactions, accounts, catalogues, and
even milestones.25 Texts on papyrus may be said to vary even more. At the
formal end there is, say, a letter of the emperor, at the other extreme a shop-
ping list or a brief receipt on an ostrakon. In between there is an almost
inexhaustible range of public and private documents: legal proceedings,
contracts, accounts, petitions, tax receipts, reports, lists, and private let-
ters, the last often the most interesting of all for their human appeal and
language.26
The Greek of the documents is almost as varied as their content.27
We must first separate the texts of the Classical period (to 300 BC), which
are written in varieties of Classical Greek, including of course local dia-
lects. The inscriptions of Athens are in Attic, those of Delphi in Delphian,
and so on. Once Koine Greek, which grew primarily out of Attic, emerged,
the dialects were gradually eroded, but not entirely: inscriptions in dialect
continue to appear as late as the second century AD, usually with an admix-
ture of Koine features. The poetic genres are also an important part of the
picture, beginning in the earlier period and continuing right through the
later. Epitaphs and the like will often display metrical forms and poetic
vocabulary and morphology, though they may also contain Koine features.
The researcher needs to be able to discern these differing genres and the
features of language that accompany them.
24
Bagnall, Handbook of Papyrology, 645 (P. van Minnen). [Thousands of ostraka have
also been found and are still being found and edited.]
25
For a full survey see McLean, Greek Epigraphy, 181–210.
26
See survey in Bagnall, Handbook of Papyrology, 358–94 (B. Palme).
27
Besides Greek, there are of course documents in other languages, notably Latin, Egyptian
Demotic, and Coptic, and also bilingual texts.
The great majority of the texts of the later period, both inscriptions
and papyri, are in Koine Greek, but not of one single variety. It is a mis-
take to regard them all uniformly as witnesses to vernacular or ‘popular’
Greek. While they may and do witness to that, their language offers a range
from everyday to literary, and what features appear will depend on the
choice of register and the degree of education of the writer. Even an ‘uned-
ucated’ writer may introduce learned features and use stereotyped expres-
sions containing obsolete forms. A sensitivity to the possible variations of
‘literariness’ is a prerequisite in the researcher.
Works of literature, whether texts of Classical or of later authors, also
survive on papyrus (less often stone) and may be included in modern edi-
tions of documents.28 They are by definition not ‘documentary,’ but in terms
of language the line is not so easily drawn. The language of literary works
overlaps with that of everyday language. For enquiry into the LXX vocabu-
lary this source of evidence is not to be neglected, both for everyday words
and for literary vocabulary if ‘literary’ books or sections of the LXX are
in view. In any case, any occurrence of a word in any source may be useful
for filling in the history of that word.
The value of the documentary evidence for study of the vocabulary of
the LXX is essentially that it enables us to understand the LXX better. All
evidence is valuable, but this body of additional and growing evidence,
as was said above, gives access to a dimension of the language that was
inadequately known before. The material that is closest in time to the LXX
is the most valuable, and there is by good fortune a large body of Ptole-
maic papyri. Elucidation involves a quest for parallels, and this has the
additional outcome of enabling us to conclude (up to a point) whether a
word or use belongs to normal Greek or is peculiar to the LXX. A further
aspect of such researches is important. The traffic is not all one way: the
LXX itself, used with due caution, is a witness to Koine Greek. By bring-
ing the LXX and documentary evidence together we may elucidate or
support either by the other.
4. PRACTICALITIES
If access to the documentary texts were just a matter of consulting a

printed lexicon or other tool, this section would be unnecessary. As it is,
the constant growth of the material and rapid developments in technology
28
The categories ‘subliterary’ and ‘paraliterary’ are also used: see Bagnall, Handbook of
Papyrology, 282–3 (T. Renner).
have meant that printed works could not possibly keep pace. Scholarship
itself cannot keep pace. Consulting any of the existing reference works,
such as LSJ, Preisigke, Wörterbuch, or MM, gives only a partial and often
misleading view. On the other hand, full access to the documentary texts
themselves has become easy, transformed in recent years by electronic
databases and search tools. Any word can be searched for and found elec-
tronically with great speed. But therein lies a difficulty. The results are raw
material, a vast body of it, requiring expertise and sheer hard work to make
use of. A list of occurrences is just the beginning. The data need thorough
analysis if they are to be useful for any purpose. Quick and easy answers
will not be found.
Until the arrival of digitisation, documentary texts were accessible only
in the printed volumes in which they were published. Searching could be
done only via the indexes supplied (or not) by the editors, or by continuous
reading. Whatever the drawbacks of this form, the book had, and still has,
the advantage that discussion, notes, and translation (if provided) can all
be readily consulted in one place on the printed page.29 With the transfer
of the texts to electronic form much of this assistance has been omitted
(though no doubt it will in time be added). Print publication of texts is
however still the norm and is the source from which electronic texts are
derived (though this too could change). The upshot at present is that for
an accurate report of the text and aids to comprehension, it is necessary,
or at least desirable, to consult the printed volume.30 Published volumes of
documentary texts number many hundreds, and to manage the abundance,
lists and abbreviations have been created.31 The secondary bibliography,
itself enormous, will be found listed to some extent in the printed volumes,
but complete only up to the date of publication; the electronic media are
in the process of bringing this under better control.
29
Translations are an important aid. Besides those in the editions, there are selections of
translated documents, which can usually be located via one of the websites (see notes 32
and 34).
30
This of course is easier said than done: it is a rare library that possesses even a majority
of the volumes, let alone all. Epigraphic volumes are especially intractable. [Reliance on
the electronic form is increasingly the reality whether it should be or not, as access to the
printed volumes (often now in ‘storage’) becomes harder.]
31
For papyri: Oates et al., Checklist. Epigraphic abbreviations are in disarray. A start was
made on a unified system in Horsley and Lee, ‘Preliminary Checklist.’ This is adopted
in McLean, Greek Epigraphy, 387–472, with additions to 2002. The system in PHI Greek
Inscriptions is different; CLAROS (<moderna1.ih.csic.es/claros/cnc/2cnc.htm>) offers
another; LSJ and LSJ, Suppl. (1996) inevitably have another. A useful guide to epi-
graphic works is Bérard et al., Guide de l’épigraphiste.
Electronic searching, which until recently was possible only via CDs,
can now be done online. The current main search locations may be noted
here,32 but further change and development are inevitable. There are also
multiple sites gathering different types of information.33 A significant devel-
opment has been the creation of the ‘Trismegistos’ database, which aims
to gather ‘metadata’ on all published papyrological and epigraphic texts
from Egypt, in the period 800 BC to 800 AD.34 It also sets out to pro-
vide links to all the main projects, and has begun the process of assigning
a unique number to each documentary text, an essential device for man-
agement of the proliferating information.
There are good general introductions to the two fields, by B. Hudson
McLean in epigraphy35 and Roger S. Bagnall in papyrology.36 For an
introduction to Koine Greek, works by Robert Browning37 and Geoffrey
Horrocks38 are the best on offer. Important reference grammars are those
of Edwin Mayser39 and Francis T. Gignac,40 but vocabulary is treated only
incidentally in these works. A lexicon or extended treatment of the Koine
Greek vocabulary is non-existent, though MM has a useful collection of
data, provided it is not regarded as the final word. As for the LXX, the
lexicon of Muraoka does not offer documentary evidence as such, or
undertake new research in the field, but simply uses the results of existing
studies; but it is the essential guide to the meanings of words in the LXX,
based on a fresh and complete lexical analysis. LSJ’s material on the LXX
is patchy and often unreliable; but its coverage of documentary evidence,
though limited, is more trustworthy, at least as a starting-point.
The need for expertise has been mentioned. Of course this can be said
of any subject, but work on documentary texts can be especially challeng-
ing for those coming to it from other disciplines. There is first the need
for familiarity with a wide range of types of Greek, from Classical to late
Koine, from learned to colloquial. This includes awareness of the pho-
netic and other changes in the later period; to approach Koine Greek as
32
<papyri.info> and <epigraphy.packhum.org/inscriptions/>.
33
See further McLean, Greek Epigraphy, 383–5; Bagnall, Handbook of Papyrology, xv–
xvi.
34
<www.trismegistos.org>
35
McLean, Greek Epigraphy. See also Woodhead, Greek Inscriptions.
36
Bagnall (ed.), Handbook of Papyrology. See also E. G. Turner, Greek Papyri.
37
Browning, Greek, chapter 2.
38
Horrocks, Greek.
39
Mayser, Grammatik.
40
Gignac, Grammar. Epigraphy lacks any equivalent to Gignac and Mayser.
a sort of debased Classical Greek is a serious mistake. Even with this

knowledge, to understand the technical and other terminology, or even
just to grasp the purpose and background of a text, can be demanding.
The word being studied will not necessarily yield its meaning quickly,
and extensive enquiry may be needed.
Some particular difficulties deserve notice. Incomplete texts invite res-
toration, but not every restoration found in the published editions is to be
trusted. The researcher needs to be wary of accepting any word fully or even
partially restored, unless it is part of a formula or other predictable expres-
sion.41 Another problem is ‘correction’ of originals by editors, that is to
say, when the original presents a non-standard spelling or other feature, the
editor offers the ‘correct’ form. Some of the corrections go beyond indi-
cating the standard spelling and alter forms or syntax that actually belong
to the Greek of the time.42 Finally, dating. This is of vital concern to the
lexical enquirer, but there are many obstacles put in the way of ascertaining
the date of a document quickly. These include total omission of ‘BC’ and
‘AD’ or equivalents, the use of a variety of dating formulae (some in Latin),
and most frustrating of all, failure to offer any date whatever, even within
wide parameters.43
5. ILLUSTRATIONS
5.1 ποτίζω
Let us begin with a straightforward example of a new use in the

LXX paralleled in contemporary Egyptian papyri.44 In the Classical period
ποτίζω is well attested, but not in the meaning ‘irrigate’ (land); the word
for this was ἄρδω. In the Pentateuch and occasionally in later books ποτίζω
occurs in this sense, ἄρδω not at all.45 An example:
41
Cf. E. G. Turner, Greek Papyri, 67–70; McLean, Greek Epigraphy, 35–9. Acquaint-
ance with the editorial sigla (square brackets, subscript dots, etc.) is also a requisite: see
Bagnall, Handbook of Papyrology, 203 (P. Schubert); McLean, Greek Epigraphy, 27–
35.
42
An example: in PCairZen 2.155 χερός is corrected to χειρός, but the former is a genuine
form, not a misspelling.
43
Some variations border on the ridiculous, e.g.: ‘sac’ = saeculum ante Christum, i.e.,
BC/BCE, but ‘ac’ = after Christ, i.e., AD/CE; but then ‘a.C.’ = ante Christum, i.e., BC/
BCE.
44
Cf. Lee, Lexical Study, 118–22.
45
ποτίζω ‘irrigate’: Gen 2:6, 10; 13:10; Deut 11:10; Sir 24:31; Joel 4:18.
Gen 2:6 πηγὴ δὲ ἀνέβαινεν ἐκ τῆς γῆς καὶ ἐπότιζεν πᾶν τὸ πρόσω-
πον τῆς γῆς.
The evidence of the Ptolemaic papyri shows that this use was standard
in the time of the LXX translators. There are now thirty or so occur-
rences, of which one sample will suffice:
PZenPestm Suppl. B p. 266.2, 3 (256 or 255 BC) Ζήνων Κλέωνι χαί-
ρειν. τὸ ὕδωρ τὸ ἐ[ν τῆι διώρυγι οὐκ ἀνα]βέβη[κ]εν πλείω ἢ [πῆ]χυν
| ὥστε μὴ δύνασθαι ἀπ᾽ αὐτῆς ποτίζε[σθαι τὴν γῆν. καλῶς ἂν ο]ὖν
π[ο]ήσαις ἀνοίξας | τὰς θύρας, ἵνα ποτίζηται ἡ γῆ.
Zenon to Kleon greetings. The water in [the canal has not] come up
more than a cubit, so it is not possible for [the land] to be irrigated from
that source. Please, then, open the sluice-gates, so that the land may be
irrigated.
In Deut 11:10 the word appears in a special expression, which is

worth looking into briefly: ὅταν σπείρωσιν τὸν σπόρον καὶ ποτίζωσιν
τοῖς ποσὶν ὡσεὶ κῆπον λαχανείας (‘when they sow the seed and irrigate
it with their feet like a vegetable garden’). The documentary evidence does
not provide an exact parallel, but it comes tantalizingly close: there are
examples of ποτίζω / ποτισμὸς ἀπὸ ποδός (‘irrigate/-ation by foot’) in
papyri of the first and second centuries AD.46 This looks like the same prac-
tice, but what method of irrigation was it? The debate has been drawn-out,
but there is a probable answer.47
The evidence of the papyri makes an important contribution to tracing
the history of ποτίζω. Without it the literary record is not very informa-
tive: if we trusted LSJ this use appears to occur only in Lucian (II AD).48
The inscriptions also happen to offer only late occurrences of ποτίζω,
none of them in the sense of ‘irrigate.’
46
The current tally of exx. is five: PSoter 1.32 (69 AD), 2.26 (71 AD); PFlor 3.369.7
(139/149 AD); PRyl 2.157.21 (135 AD); PMilVogl 7.308.97 (II AD); similarly Cadell,
‘Vocabulaire de l’irrigation,’ 114–5; this has grown from two in Lee, Lexical Study,
121–2. (SIFC 13.366 in LSJ Suppl. (1996) = PFlor 3.369.)
47
See Lee, Lexical Study, 121–2; Dogniez and Harl, Le Deutéronome, ad loc.; Cadell,
‘Vocabulaire de l’irrigation,’ 110–7; Wevers, Deuteronomy, ad loc.; LSJ Suppl. (1996),
s.v. πούς; Muraoka, Lexicon (2002, 2009), s.v. ποτίζω. Cadell examines all the pos-
sibilities and returns to the suggestion (in Lee, from Driver) that it involves use of the
feet to control the flow of water on to the garden through channels in the earth, also
known in modern parallels.
48
Luc. Abd. 27 ‘etc.’ (LSJ, s.v. 3). The truth of course is that there are more, as, e.g., in
Strabo the Geographer (I BC/I AD) at 17.1.3, 1.4. These were found by a search in the
TLG.
5.2 εὐίλατος
This adjective (‘kindly disposed,’ ‘merciful’) appears twice in the LXX:

1 Esd 8:53 καὶ πάλιν ἐδεήθημεν τοῦ κυρίου ἡμῶν κατὰ ταῦτα καὶ
εὐιλάτου ἐτύχομεν.
And again we entreated our Lord in these matters and found him merciful.
Ps 98:8 ὁ θεός, σὺ εὐίλατος ἐγίνου αὐτοῖς καὶ ἐκδικῶν ἐπὶ πάντα
τὰ ἐπιτηδεύματα αὐτῶν.
O God, you were merciful to them, even while taking vengeance on
all their practices.
What attestation does εὐίλατος have outside the LXX? The results of
a thorough search are remarkable. Aside from the LXX itself, it does not
appear in literature until much later, namely, Patristic texts from the sec-
ond century AD onwards (mostly quoting Ps 98:8) and the Byzantine lexi-
cographers;49 its attestation before then is entirely documentary. And this
evidence reveals what we would otherwise not have known, that εὐίλα-
τος was in extensive use in the time of the LXX, and not just in Egypt. LSJ
(1925–40) were able to cite six occurrences in documents; the tally today
is almost five times as many. The bulk of the texts fall in the period from
the third century BC to the first century AD; geographically they range
from Egypt, through Asia Minor (notably Knidos in Caria), Attica, and
Thrace, to Pompeii.50 A small selection must suffice:
IKnidos 1.150.A.9 (Knidos, II–I BC) μὴ τύχοι Δάματρος καὶ |
[Κ]όρας μηδὲ θεῶν τῶν παρὰ Δάματρος εὐιλάτων
May he not find Demeter and Kore nor the gods with Demeter well-
disposed.51
IEgVers 175.I.36 (Narmouthis, I BC?) κλῦθι ἐμῶν εὐχῶν, μεγαλο-
σθενὲς οὔνομ᾽ ἔχουσα | εὐείλατος ἐμοί τε γείνου, λύπης μ᾽ ἀνάπαυ-
σον ἁπάσης.
Hear my prayers, O mighty-named one [= Isis-Thermouthis] | and be
merciful to me, give me rest from all pain.
49
Results from TLG search. An example attributed to Antimachus (V/IV BC) in Et. Mag.
s.v. Ἀδόροις is to be read as εὐήλατος (Frag. 109, ed. Wyss).
50
There is also a later example in a magical papyrus, PapGraecMag 2.166 (IV AD?):
δέσποτα Μουσῶ[ν], ἵλαθί μοι, τῷ σῷ ἱκέτῃ, καὶ ἔσο εὐμενὴς καὶ εὐίλατος, φάνηθί
μοι καθαρῷ τῷ προσώπῳ. [There is even an occurrence in Seneca!: Aitken, No Stone
Unturned, 101 n. 46.]
51
Doric α for η in the goddesses’ names illustrates the continuation of older dialect:
Knidos was an old Doric foundation. There are about ten other similar texts from Knidos.
[For παρά + gen., ‘attached to, with’ see Essay 26.]
IG II2 1365.25 (Attica: Sunium, I AD) = CMRDM 1.12.25 (n. d.)

εὐίλατος γένοιτο ὁ θεὸς τοῖς | θεραπεύουσιν ἁπλῇ τῇ ψυχῇ.
May the god [Men Tyrannos] be merciful to those who tend him with
simplicity of soul.
PSI 4.392.6 (242/1 BC) … τοῦ βασιλέως εὐειλάτου γενομένου καὶ

ἐπιγράψαντος τὴν ἄφεσιν.
… if the king is merciful and signs the release.
Something further is evident. This is a word with an elevated, religious

tone, applied in the documentary texts only to gods or King Ptolemy (also
a god). The Psalms and Esdras translators knew what they were doing when
they chose it; even the expressions match.52
The LXX also has the verb εὐιλατεύω, derived from εὐίλατος, in
Deut 29:19, Jdt 16:15, and Ps 102:3, all with the same religious tone. So
far there is no other attestation, except in later Patristic texts and lexicog-
raphers. But it would be unwise to conclude that it was a neologism of
the LXX translators; and if past experience is any guide, there is a good
chance that it will turn up.
5.3 ἀγαπῶ
A full study of this word would be a large subject. The point to be made
here is that documentary evidence supports the view that ἀγαπῶ began
to supplant φιλῶ as the word for ‘love’ already in the fourth century BC
and continued on the same path later. This was demonstrated by Robert
Joly from the literary evidence.53 It follows that the LXX translators used
ἀγαπῶ (c.250 times) because it was current usage, not because they
intended to reject ‘pagan’ usage and create a new vocabulary for ‘love’
words, passed on in due course to the NT.
Documentary examples from the time of the LXX are very few, but
they provide evidence that ἀγαπῶ ‘love’ was used not just in Egypt but
in the furthest corners of the Greek-speaking world:
OGI 1.90.9 (Rosetta, 196 BC) ἐπειδὴ βασιλεὺς Πτολεμαῖος αἰωνό-
βιος, ἠγαπημένος ὑπὸ τοῦ Φθᾶ, θεὸς Ἐπιφανὴς Εὐχάριστος …
52
Talshir (1 Esdras, on 8:53) missed an opportunity: she quotes just one example, taken
from LSJ (without the reference, PPetr 2 p. 45). [See Aitken, No Stone Unturned, 98–102
for a full treatment of this word, with interesting results, including that it appears to be a
typically Egyptian term and may have an Egyptian resonance even outside Egypt.]
53
Joly, Le vocabulaire de l’amour. A valuable exposition is found in Swinn, ‘Ἀγαπᾶν in
the Septuagint.’
Whereas King Ptolemy, living eternally, loved by Phtha, god Mani-

fest, Gracious …54
IGExtrEast 292.9 (Kandahar, mid III BC)55 ἔδει … διδάσκαλον καὶ
πατέρα καὶ μητέρα | ἐπαισχύνεσθαι καὶ θαυμάζειν, φίλους καὶ ἑταί-
ρους ἀγαπᾶν καὶ μὴ διαψεύδεσθαι, | δούλοις καὶ μισθωτοῖς ὡς κου-
φότατα χρᾶσθαι.
They were to … show modesty and respect to their teacher and father
and mother, to love their friends and companions and not deceive them,
to treat slaves and hired servants as gently as possible.
6. FUTURE DIRECTIONS
The time of establishing the relevance of the documentary evidence to

the language of the LXX is past: it has been well and truly proven. Numer-
ous studies over the past century have made forays into this resource and
produced good findings. But a thorough, systematic exploration is still
needed, one that is ongoing, as the material continues to grow and become
even more accessible via electronic tools. For vocabulary we now have
a basic tool of investigation, namely, a full lexical analysis of the LXX,
provided by Muraoka’s lexicon. The next step is systematic searching of the
documents for parallels (or not) to the whole gamut of usage of the LXX.
We are also in great need of a record of the published discussions on LXX
words, kept up to date as they appear. At present there is no reliable way
of keeping track, and each researcher begins the quest all over again.
A way of gathering the data into one accessible resource will be essen-
tial. In the past this might have been attempted by a printed lexicon, as
it was for example by Bauer/BDAG for the NT, but the drawbacks of that
type of resource are evident. The obvious tool now is an electronic data-
base, able to gather all the material and keep it up to date. This is the new
desideratum of LXX studies. In the long run, benefits to all scholarship on
the Greek vocabulary will ensue from such a project. Not only will the
vocabulary of the LXX be seen more fully in its context, but its own
witness to the language of its time will be more evident.56
54
There are several more instances of this formula in the Rosetta stone and copies. In
PCairZen 4.580.3 (mid III BC) ἀγαπῶ occurs without context.
55
The abbreviation problem illustrated. The volume is F. Canali De Rossi (ed.), Iscrizioni
dello estremo oriente greco (IK 65; Bonn, 2004). Abbreviation in PHI Greek Inscriptions:
‘IK Estremo oriente’; in CLAROS: ‘IEOG.’ I have suggested ‘IGExtrEast’ on the pat-
tern of the Horsley-Lee list.
56
My thanks to Trevor Evans for useful comments on this chapter.
25
THE PUZZLE OF JOHN 21:15–17:
A FORMALITY SOLUTION
2017
Abstract
The alternation of the words for ‘love’ in John 21:15–17 is a long-standing

puzzle that has defied solution. It is suggested in this note that ἀγαπῶ and φιλῶ
carry a slight difference of ‘formality’ as a result of their history. This accounts
for the interplay and makes it meaningful compared with simple variation.
Many attempts have been made to explain the mysterious alternation of

the two verbs for ‘love,’ ἀγαπῶ and φιλῶ, in John 21:15–17. This short
note will not rehearse them but move as speedily as possible to a sugges-
tion that as far as I know has not been made before.
In the famous series of questions addressed by Jesus to ‘Simon son of
John,’ Jesus twice asks, ‘Do you love me?’, ἀγαπᾷς με; to which Simon
replies both times, ‘You know that I love you,’ φιλῶ σε, using a different
verb. Jesus asks the question a third time, using now the same verb as
Simon, φιλεῖς με; to which Simon replies as before, ‘You know that I
love you,’ φιλῶ σε. So Jesus uses ἀγαπῶ twice, then switches to Simon’s
φιλῶ, while Simon persists with φιλῶ throughout the exchange.
Attempts to find a difference of meaning between the two verbs have
proved futile, and rightly so. It is not possible to show that each verb indi-
cates a different type of love, or that such a difference can make the exchange
more meaningful.1 Barrett declared it ‘impossible to doubt that they are
synonyms.’2 The best alternative explanation, in fact the only one currently
available, is that this is a case of simple variation between words of simi-
lar meaning, a phenomenon seen elsewhere in John, indeed in this very
1
For survey and discussion see Morris, John, 871–3; TDNT, 9:134–5 (Stählin); Brown,
John, 2:1102–3. I am grateful to Trevor Evans for reading this paper and offering helpful
comments.
2
Barrett, John, 486.
passage.3 In that case the words have the same lexical meaning, that is,
approximately love. Lexical meaning as such, however, is not all that may
be involved; there is also ‘connotation’ or ‘tone,’ that is, the additional
semantic elements that lie outside lexical meaning and attach to many if not
most words. If these two words do not differ in lexical meaning, might they
differ in ‘tone’? To find the answer we need to look into their history.
The two verbs ἀγαπῶ and φιλῶ were in competition in the Classical
period, and in the fourth century BC ἀγαπῶ began to supplant φιλῶ as the
word for ‘love.’ This has been demonstrated by Robert Joly from the evi-
dence of Classical and later literature.4 The verb φιλῶ, while retaining its
meaning ‘love,’ had already developed two additional meanings, namely,
‘be accustomed (to)’ (VI–V BC +) and ‘kiss’ (V BC +). These competed
with the meaning ‘love’ and were maintained through its later history.
‘Kiss’ prevails as the usual modern meaning. In addition, presumably as a
development from its meaning ‘kiss,’ φιλῶ shows signs of occasional use
as a euphemism for ‘have sexual intercourse with.’5 Joly did not notice
this but he came close: he thought that the meaning ‘kiss’ killed (‘a tué’)
the meaning ‘love.’6 The sexual meaning would have been even more
deadly – the kiss of death, one might say.7 These changes were enough to
ensure that a new word for ‘love’ would make its way into the vocabulary.
φιλῶ did continue to be used with its old meaning ‘love,’ but not entirely
on the same footing as ἀγαπῶ. A highly significant fact is that φιλῶ, while
frequent in documentary papyri, occurs mostly in formulae, which are of
course a standard location of old-fashioned, formal language.8 Another
3
Argued with many exx. in Morris, Studies, 293–319. In this passage: βόσκε τὰ ἀρνία
μου (15), ποίμαινε τὰ πρόβατά μου (16), βόσκε τὰ πρόβατά μου (17); οἶδας, γινώ-
σκεις (17).
4
Joly, Vocabulaire de l’amour, 10–29. See also Swinn, ‘ἀγαπᾶν in the Septuagint’; Silva,
Biblical Words, 96–7.
5
The evidence is limited but persuasive: Hesych. βαίνειν· φιλεῖν (noted in LSJ, s.v.
φιλῶ I.3.b ‘of sexual intercourse’); Schol. Theocr. 4.58 μύλλει· ἀντὶ τοῦ φιλεῖ; Schol.
Ar., Plutus 1093 καὶ νῦν φιλῶ σημαίνει διακορεύω ἢ βινῶ. See Shipp, MGE, 126–7 with
n. 72 for these examples and discussion. Continuation into vernac. mod. Greek is indi-
cated by the saying quoted in Lex. Proia, s.v. φιλῶ: [ὅπου βρίσκει καὶ φιλεῖ, μοῦντζές
του κι᾽ ἂν παντρευτῇ, ‘wherever a man finds and makes love (to a woman), mountzes
(a gesture of contempt) to him, even if he marries her.’ I am grateful to Tasos Kalogerakis
and John Fountotos for help with the translation.]
6
Joly, Le vocabulaire de l’amour, 31.
7
French baiser shows parallel development. The earliest Greek word for ‘kiss,’ κυνῶ, had
suffered a similar fate: see Joly, ibid., 33. καταφιλῶ ‘kiss’ was an early addition to the
mix (Xenophon +), no doubt to disambiguate φιλῶ.
8
Joly, ibid., 27–9. Most are in the epistolary closing formula ἀσπάζομαι / ἀσπάζου τοὺς
φιλοῦντας ἡμᾶς / σε, attested from I AD onwards. Examples: POxy 17.2148.12 (27 AD);
OClaud 1.146.7 (100–120 AD); PMich 3.221.19 (296 AD). Also formulaic is the cliché
THE PUZZLE OF JOHN 21:15–17: A FORMALITY SOLUTION 405
observation is that φιλῶ often appears with μισῶ, in a more or less fixed
pair.9 Further data pointing in the same direction could be added.
This evidence leads to the conclusion that in early Koine Greek, and
certainly by the time of the New Testament, φιλῶ was obsolescent in the
meaning ‘love.’ The standard word for the idea was ἀγαπῶ. Obsolescent
words tend to carry greater formality. Thus we can say that φιλῶ was the
slightly more formal word, while ἀγαπῶ had no extra connotation of that
kind, but was ‘unmarked’ for formality. It should be emphasised that the
difference, though noticeable to a native speaker, would be slight. English
offers no parallel in words for ‘love,’ but we might illustrate by ‘Do you
wish to go?’ and ‘Do you want to go?’: wish is slightly more formal than
want.10
In the conversation in John 21, Simon Peter is being formal and polite.
He draws back from using Jesus’s plain word, which might imply a certain
familiarity. This does not mean that he loves any less, only that he feels
unable to express it so directly. When Jesus the third time uses the same
word as Simon, he accedes, as it were, to Simon’s level of formality, but
without changing his meaning. Simon persists with his chosen politeness
in his final answer.
How did this feature come to be part of the conversation now recorded
in John’s Gospel? We can only guess, but some things are more likely
than others. It is only possible in the Greek form of the exchange. If the
conversation was translated from Aramaic, someone introduced it, either
at the point of translation or later. Introduction at the point of translation
is unlikely, unless Aramaic offers a similar pair of variables.11 If the con-
versation took place originally in Greek, with the alternation of words as
it is now, it is unlikely that it was noted and preserved exactly. More likely
the compiler of this Gospel, John, introduced the subtle detail of Peter’s
choice of word to portray Peter’s respectful attitude to Jesus.
οἶδα πῶς / ὅτι με φιλεῖς, in PCol 10.279.13 (III AD) etc. In addition, Joly (29) notes
some ‘formules religieuses’ with φιλῶ.
9
Joly, ibid., 11, 16, 18, 30. The pairing is found in John 12:25; 15:19.
10
On formality cf. Lee, ‘Some Features,’ esp. 7, 10 [= Essay 5].
11
Brown (John, 2:1103) notes that it does not (nor does Hebrew).
26
THE BRILL DICTIONARY OF ANCIENT GREEK:
NEW CLOTHES FOR AN OLD EMPEROR
2017
Abstract
This review article offers a critical assessment of the new Brill dictionary.
Several important aspects of the work are considered and illustrated by examples:
the provenance of the dictionary, going back through Italian predecessors to LSJ;
the structure of verb entries; the lexical analysis; the coverage of documentary
evidence; the coverage of the LXX and NT; and the treatment of a sample particle.
In the conclusions an overall assessment of GE is given, especially in comparison
with LSJ, and remarks are offered on the state of lexicography of Ancient Greek
and its future direction.
1. PREAMBLE
The publication of a new ‘major’ lexicon of Ancient Greek is a sig-

nificant event.1 The new Brill dictionary (‘GE’)2 is of the same size and
scope as the lexicon of Liddell and Scott, the current standard work. More-
over, the new lexicon has been published complete in one volume, unlike
the other major lexicon project of our time, DGE, still progressing through
publication in parts.3 Though GE is based, it is true, on a previous lexicon,
the completion of the project in just four years is a prodigious achievement.
The co-ordination of the team that produced this result – there are at least
15 names mentioned in the preface – was a major undertaking in itself.
Similarly, the Italian predecessor (‘GI’) involved at least 60 collaborators
under the leadership of Franco Montanari.4
1
This review is a much-expanded version of my presentation in a panel on the new lexicon
at the SBL Annual Meeting, San Antonio, 2016. My thanks to Michael Curran for read-
ing it and offering helpful comments.
2
Montanari, et al. (eds.), The Brill Dictionary of Ancient Greek (2015).
3
Adrados, et al. (eds.), Diccionario Griego-Español (7 vols. to 2009).
4
Montanari, (ed.), Vocabolario della Lingua Greca (2013). For the figures mentioned,
see GE, p. vii; GI, pp. i–ii.
The character and merits of the new lexicon are indicated in the preface
to GE by the three leaders of the team that produced it.5 Though many
statements are made and most are clear, some reading between the lines
is required to understand the presuppositions of the editors and what this
lexicon attempts to achieve. The editors naturally draw attention to the
improvements (new evidence; new lemmata; corrections to GI; inclusion
of proper names), but they also appear to recognise that GE is not, and
could not possibly be, the equivalent of a completely revised LSJ (cf.
below, §8). They do claim that ‘this Greek-English project is presented as
an enhancement of lexicography,’ and that the dictionary ‘presents a criti-
cal approach to lexicography in and of itself,’ statements that are difficult
to interpret, if not meaningless. Two further assertions will be commented
on in the next section.
It is of course impossible to assess the entire contents of a massive
reference tool such as a lexicon: any review necessarily involves sam-
pling. My samples and discussion will have in view three main themes
or aspects: (a) the provenance of GE; (b) GE as compared with LSJ; and
(c) the current state of the lexicography of Ancient Greek. The samples are
presented under various headings, but all will be found to illustrate these
three topics.
2. PROVENANCE OF GE
Dependence on predecessors is a general phenomenon of lexicogra-

phy. Rarely has a lexicon been created afresh, that is, on the basis of a new
collection of data subjected to a new lexical analysis. Only if the compilers
are prepared to spend up to a century can it be done. The Oxford English
Dictionary was created in this way (70 years); so was the Oxford Latin
Dictionary (50) and the recently completed Dictionary of Medieval Latin
from British Sources (100). LSJ was not; neither was GE. Both descend
from a line of predecessors, which in GE’s case include LSJ itself. The
provenance of GE may be summarised as follows:
GE 2015
< GI (1st ed. 1995; 2nd ed. 2004); 3rd ed. 2013
< Rocci (1st ed. 1939; 2nd ed. 1941); 3rd ed. 1943
< LSJ 9th ed. in 10 parts, 1925–1940; + Bailly 1894
< LS 8th ed. 1897 …
5
G. Nagy, L. Muellner, M. Goh; see GE, p. vii.
THE BRILL DICTIONARY OF ANCIENT GREEK 409
< LS 1st ed. 1843

< Passow 4th ed. 1831
< Schneider-Passow (1st ed. 1819–23); 3rd ed. 1827
< Schneider (1st ed. 1797–98); 3rd ed. 1819
GE derives from GI, which was based on an Italian predecessor by
Rocci,6 compiled on the basis of LSJ, with occasional help from Bailly.7
So GE actually derives from LSJ, via translation into Italian and back
again into English. LSJ’s own history goes back through a line of revi-
sions to a German ancestor, the work of Passow; and Passow itself rests on
earlier predecessors (of which enough).8 It cannot be said that any of these
lexicons was a completely new work.
All this is acknowledged in Montanari’s preface to GE (and its prede-
cessor, the Prefazione to GI), and in a further preface by Nagy, Muellner,
and Goh.9 Despite this, the latter editors make a surprising admission:
When Brill asked the three of us in 2010 to undertake the editing of an
English version of the Greek-English dictionary of Franco Montanari …
we had not yet fully reckoned with the fact that the original ‘Liddell and
Scott,’ later transformed into ‘Liddell-Scott-Jones’ or ‘LSJ’ … cannot
really be called an ‘original’ Greek-English lexicon.
Furthermore, after summarising the history of LSJ back to Schneider,
they characterise the process as ‘cross-fertilization,’ adding ‘we are con-
vinced that such cross-fertilization is at the very core of lexicography.’
Some ‘cross-fertilization’ there may be, but this is hardly the term to
describe what is in fact a chain of dependence. Still more surprising is this
statement of the editors regarding their own lexicon:
[I]t is to be emphasized that the lexicon is not a translation of the Italian
definitions in and of themselves.
Evidence to disprove this claim can be found on every page of the
lexicon when set beside its predecessor GI.
Does any of this matter? It does, because not everything in the Liddell
and Scott tradition is reliable. There is a strong possibility that at every
stage older material has been simply taken on trust and never re-examined.
Decisions about the lexical analysis, the definitions, and the selection of
data were often made long ago and have not been questioned when they
6
Rocci, Vocabolario. As I understand it, Rocci’s work was primarily intended as a manual
for Italian students rather than a major lexicon to rival LSJ.
7
Bailly, Dictionnaire. See Rocci’s preface (1943) for acknowledgement of LS and men-
tion of Bailly.
8
For details of editions see Lee, History, 347–8; 355–6; DGE vol. I, p. x.
9
GE, pp. v–vi; p. vii.
ought to have been. The revision process to which Liddell and Scott was
subject guaranteed this result: for practical reasons most of the material was
carried on unchanged in each edition and a thorough revision could not
be undertaken. Supplementary material added piecemeal each time often
did no more than create confusion. Most users of the great lexicon – and
even compilers of new lexicons – are unaware of these problems, though
there has been no lack of remark on them.10 A first example to illustrate
the point:
2.1. μεταδετέον
GE: μεταδετέον [μετά, δέω] verb. adj. it is necessary to unbind
XEN. Hip. 4.4.
GI: μεταδετέον [μ., δέω] vb. bisogna slegare XEN. Hip. 4.4.
Rocci3 (1943): μετα-δετέον, vrb. da δέω, si deve sciogliere, ἀπὸ τῆς
φάτνης, dalla mangiatoia, SEN. Eq. 4, 4.
LSJ Suppl. (1996): μεταδετέον, for ‘one must untie’ read ‘one must
change the tethering (from one place to another)’.
LSJ: μεταδετέον, one must untie, X.Eq. 4.4.
LS8 (1897): μεταδετέον, verb. Adj. one must untie, Xen. Eq. 4, 4.
LS1 (1843): μεταδετέον, verb. Adj., one must untie, Xen. Eq. 4. 4:
from μεταδέω, δήσω, to tie differently; to untie.
Passow4 (1831): μεταδέω, δήσω, (δέω, Irr.) umbinden, d. i. anders od.
anderswohin binden.
In this case a mistake has been carried on undetected all the way from
the first edition of LS to GE. The meaning ‘untie’ is simply wrong for
μεταδέω: it appears to have arisen from a mistranslation of umbinden as
unbinden. The correction in LSJ Supplement was missed and the mistake
sailed on. This is a trivial example – or maybe not, to those interested
in ancient horse management – but it illustrates the phenomenon.11 The
next case is more serious.
2.2. κάμινος
GE: κάμινος … oven or furnace, forge … || flue for heating a
room… || alcove VT Num. 25.8.
GI: κάμινος … forno o fornace, forgia … || condotto per riscaldare
una stanza … || alcova VT. Num. 25.8.
10
See, e.g., Chadwick, ‘Replacing Liddell and Scott’; idem, Lexicographica Graeca, esp. 6–
9; Lee, ‘Releasing LSJ’ [= Essay 17]. The revision process is graphically illustrated
by a sheet of the galley proofs of LS 6th ed. (1869): see Imholtz, ‘Liddell and Scott,’
130.
11
I owe this example to Glare, ‘From the Wrong End,’ 38–9.
Rocci: κάμινος … forno; fornace; fucina … alcova, SET. Num. 25, 8.

LSJ: κάμινος … oven, furnace, or kiln … alcove, Lxx Nu. 25.8.
The meaning ‘alcove’ for Num 25:8 has descended to GE from LSJ. It
has no connexion with the known meanings of κάμινος or justification in
the context of Num 25:8. What is the evidence for it? Investigation reveals
that it is an alternative rendering offered in the margin of the Revised
Version, that is, an English rendering of the Hebrew word translated by
κάμινος at Num 25:8 in the LXX.12 The underlying Hebrew is not a certain
guide to the meaning of the Greek word that translates it; a guess at the
meaning of the Hebrew in an English translation is no guide at all. By
what process this proposal came into LSJ we do not know, but it has
no right to be there.13 The evidence for the meaning of κάμινος is to be
sought in Greek, in usage contemporary with the LXX, if available. It hap-
pens that we have good material in documents of the third century BC.
These not only show κάμινος in its usual meaning ‘furnace, kiln, oven,’
but also reveal that κάμινοι in Ptolemaic Egypt could be sizable structures
used for other purposes, such as storage of wine and the keeping of ani-
mals.14 From this it becomes clear that a κάμινος could have provided just
the right secluded space for the illicit activity described in Num 25:8.15
3. STRUCTURE OF VERB ENTRIES
Let us take a straightforward common verb for a view of the general

structure of an entry in GE and a comparison of GE with LSJ.
12
RV (1885); Num 25:8: text pavilion, margin alcove (KJV tent; RSV inner room;
NRSV tent). The Heb. word is ‫ק ׇבּה‬,ֻ a hapax of uncertain meaning; many suggestions
have been made. See Gray, Numbers, 385: ‘alcove (RV. marg.) is derived from the
corresponding word in Arabic.’ Gray I suppose means that the meaning ‘alcove’ was
derived by connecting ‫ קבה‬with the Arabic cognate; but one suspects that the deriva-
tion of English alcove itself from Arabic (al-qubbah, via Fr., Sp.) might have played
a part.
13
Presumably it came via A. H. McNeile or A. Llewellyn Davies, named in Stuart Jones’s
1925 preface to LSJ (p. ix) as having ‘advised’ him on the LXX. The dubious quality of
LXX material in LSJ is well known: see Caird, ‘Towards a Lexicon. I, II’; Lee, ‘Note
on Septuagint Material’ [= Essay 1].
14
See Dorival, Les Nombres, 463–4; Wevers, Numbers, 424–5; Evans, ‘Numbers,’ 60. But
NETS has fallen victim to the authority of LSJ (‘alcove’). κάμινος is found c.40 times
in the LXX meaning ‘furnace, kiln.’
15
Another example of the trap of relying on translations: LSJ Suppl. (1996), s.v. ἀνα-
καλέω inserts proscribe for App., BC 4.25. This comes from an egregious misreading
of the Loeb translation. GE has swallowed it without checking. (I owe this example to
Terry Roberts.)
3.1. κλαίω
GE: κλαίω … impf. ἔκλαιον … || fut. κλαιήσω …; mid. κλαύσομαι …
1 act. A to weep, lament IL. 1.362 … B to suffer ill-treatment,
undergo a punsihment [sic]: κλάοις ἄν, εἰ ψαύσειας you will be
punished, if you touch them AESCHL. Suppl. 952 … C to weep for,
bewail, grieve over > with acc. IL. 20.210 OD. 1.363 … || to
call by crying: κλάειν μάμμας καὶ τιτθάς to cry for mummies
and nurses ARR. … 2 mid. A to weep, lament IL. 18.340 … B
to suffer, undergo punishment or harm, usually fut. ARISTOPH.
Ve. 1327 etc.; … C to weep for, bewail > with acc. … 3 pass. to
be bewailed AESCHL. Ch. 687 … | impers. μάτην ἐμοὶ κεκλαύ-
σεται … ARISTOPH. Nub. 1436. …
LSJ: κλαίω … I. intr., cry, wail, lament, of any loud expression of
pain or sorrow, … Od. 10.201 … 2. αὐτὸν κλαίοντα ἀφήσω I
shall send him home crying, howling, i.e. well beaten, Il.2.263:
freq. in Att., κλαύσεται he shall howl, i.e. he shall suffer for it,
Ar. V.1327 … II. trans., weep for, lament … Od.1.363 … : – Pass.,
to be mourned or lamented …: impers., μάτην ἐμοὶ κεκλαύσε-
ται … Ar.Nu.1436. 2. cry for, of infants, μάμμας καὶ τιτθάς Arr. …
III. Med., bewail oneself, weep aloud, A.Th. 920 … 2. trans.
bewail to oneself. … S.Tr. …
The general layout in GE is very similar to LSJ’s, with the usual tumble
of glosses, divided by a numbering system that requires concentration.
GE gains in clarity by using bold for the meanings; but its numbering
system is harder to pick up and somewhat counter-intuitive. The letters
A, B, etc. (enclosed in boxes) are subordinate to the numbers 1, 2, etc.
(enclosed in black blobs). While the blobs catch the eye, the numbers in
them are very small; the boxed letters likewise are quite hard to spot. In
addition, there are subdivisions introduced by double and single bars (||, |).16
The explanation of these given in the endpapers is clear with regard to mark-
ing tenses and moods in the ‘principal parts section’ of the entry (as in
‘|| fut. κλαιήσω’ above), but is much less clear about why some senses
are introduced in this way: the unsatisfying explanation given is that they
are ‘further subdivisions of the entries.’ In the case of κλαίω above, it
looks rather as if ‘|| to call by crying’ ought to have been another num-
bered sense.
These are somewhat minor matters. More importantly, we see in κλαίω
a feature carried through systematically in all verb entries, a division into
active, middle, and passive voices (marked by numbers 1, 2, 3). This is
16
There are further subdivisions using a solid arrow, dot, and diamond (not exemplified
above): see the endpapers of GE for their explanation.
a significant innovation, introduced by Montanari in GI. While it may be

helpful to the beginner, and even to some advanced scholars, it may not
be a wise one from the point of view of lexical analysis. It has the effect of
changing the upper level of the analysis from semantic to morphological;
that is, lexical meaning is subordinated to the active–middle–passive distinc-
tion. The result is that in verbs (though only verbs) the primary divisions
into 1, 2, and 3 do not indicate different senses but different voices, and
consequently the same meanings may appear, indeed must often appear,
twice or even three times in the entry. So in κλαίω, the meanings in 1 act. A,
B, C are the same as those in 2 mid. A, B, C (‘weep,’ ‘suffer,’ ‘weep for’);
and sense 3 pass. gives a meaning which is just the passive of 1C and 2C
(‘weep for’).
This innovation can also lead to trouble of a different kind in the many
Greek verbs in which form and meaning do not necessarily match.17 Exam-
ples: (a) λαμβάνω is normally middle in form in the future (λήψομαι),
but has an ‘active’ meaning; so GE has to present this as ‘1 act. (with
fut. mid.).’ (b) The usual aorist of βούλομαι is passive in form but not
in meaning; GE gives up the division by form for this verb and simply
says, in the principal parts section, ‘aor. only pass. ἐβουλήθην or ἠβ-.’
(c) The verb βιάζομαι has the annoying characteristic of being usable in
the present middle form in both an ‘active’ and a passive sense; GE’s
solution is ‘1 act. … 2 usu. mid. … 3 pass.’; but this means that 2 mixes
form and meaning.
As for verbs that have suppletives with different voices, major diffi-
culties are unavoidable. GE in some cases simply gives up the division
by voice, as in ἔρχομαι and its aorist ἦλθον, which are wrapped together;
in others various compromises are required.18 In ἐσθίω, all these prob-
lems come home to roost. The entry is simply divided into ‘1 act. … 2
pass.’ Under the first, the active ἔδω is mentioned, but we are not told
that the usual future of ἐσθίω is actually the middle ἔδομαι (in Class.,
but later φάγομαι); the standard aorist ἔφαγον, formed from a different
root, is also not mentioned: one has to know that there is a separate entry
for ἔφαγον, where its relation to ἐσθίω is noted; and although coverage
17
The extent of these variations (just for Classical Greek) can be appreciated from the
lists in Smyth, Grammar, §§805–818.
18
In the case of πίπτω, πεσοῦμαι is noted as the fut. form, but the mid. category does not
appear. In ὁράω we find ‘1 act. and mid.’ combined, then 2 pass. In ἐρωτάω we have
only 1 act., and 2 pass. but no middle; the usual (Attic) aor. to ἐρωτάω, the middle
ἠρόμην, is not mentioned (it is found under εἴρομαι). Similarly in πωλέω: no mention
of the standard aor. ἀπεδόμην.
of later Greek is said to be a feature of this lexicon, nothing at all is said

about τρώγω, which came in as the present to replace ἐσθίω (cf. below).
One would not want to suggest that LSJ has solved all these problems,
but its presentation of the voices and suppletives is generally much
clearer than GE’s, and its entry for ἐσθίω covers everything with great
efficiency.
All in all, as regards layout and convenience for the user, there is not
much to choose between GE and LSJ. Both present us with a packed
entry offering a range of possible meanings, arranged in sense-divisions
that may convey little to most readers. The selection of examples is over-
loaded and mostly from Classical authors. We have become used to find-
ing our way through this in LSJ; we will have to do the same with GE.
How are such entries actually used? I suspect that most readers simply
pluck out from the offerings a gloss that seems to fit the passage they are
reading; whether it is the right one is often a matter of chance (unless
they can locate a citation of the passage).
Let us follow up τρώγω, just mentioned in relation to ἐσθίω, and see
how it has fared:
3.2. τρώγω
GE: τρώγω … 1 act. to gnaw on, nibble at, of animals: τ. ἄγρωστιν
to nibble on weeds OD. 6.90; … | of sick people … || to eat
(raw), chew on, of pers., esp. vegetables and sweets … || later
simpl. to eat (= ἐσθίω) … 2 pass. to be eaten: τρώγεται ἁπαλὰ
ταῦτα καὶ αὖα the fresh and dry are eaten HDT. 2.92.4; …
LSJ: τρώγω … [I.] gnaw, nibble, munch, esp. of herbivorous animals,
as mules, τ. ἄγρωστιν Od.6.90; of swine … of cattle … of human
beings in disease … II. of men, eat vegetables or fruit … of des-
sert, eat fruits, as figs, almonds, etc. … of small fish as hors-
d’oeuvres … III. later, simply eat, serving as pres. to ἔφαγον
instead of ἐσθίω …
The basic idea of this verb is ‘gnaw, nibble,’ as stated, an action typi-
cal of animals but readily extended to humans. But it is rather the kind
of eating, not who is eating or what is being eaten, that characterises
τρώγω and distinguishes it from ἐσθίω, as Chadwick pointed out.19
LSJ’s division into I., applied to animals and II., applied to humans (with
leakage in ‘of human beings …’ in I.) is not necessary and the two could
have been collapsed into one lexical meaning. Later on, from about the
19
Chadwick, Lexicographica Graeca, 288; discussion of τρώγω, 287–90. Chadwick says
the human/animal distinction was made in antiquity.
first century AD, by the same process by which Lat. manducare (‘chew’)
developed into Fr. manger (‘eat’), this verb came to mean simply ‘eat,’
making it equivalent to ἐσθίω, as LSJ correctly report.
In GE the distinction between animals and humans is kept, with the
meaning stated more or less satisfactorily (though ‘to eat (raw)’ is not a
good idea); but the insertion of the later meaning ‘eat’ into this section,
with only || to introduce it, underplays its significance as a distinct devel-
opment in meaning, which leads in the long run to the Modern Greek word
for ‘eat.’ Then when the active–passive distinction is imposed, it creates
problems. It looks to the user as if the passive voice is only used with the
(sub-) meaning ‘eat,’ a meaning said to be ‘later’ (in 1.), yet cited from
Herodotus (in 2.). In fact this example is the passive of ‘gnaw, nibble,
chew’ not ‘eat’: Hdt. 2.92.4 τρώγεται δὲ καὶ ἁπαλὰ ταῦτα καὶ αὖα =
‘and these [edible kernels] are nibbled both fresh and dried.’20 In short,
GE’s entry, though ultimately derived from LSJ’s, has lost more than it
has gained in the process of revision and translation.
4. LEXICAL ANALYSIS
The core of the lexicographer’s task is to state the meanings of words,

that is, to analyse the available data for each word into its ‘lexical mean-
ings,’ with suitable definitions. This is a very challenging exercise and
good results are not guaranteed; they are usually worse when an entry has
passed through many hands in its history. The samples in the previous
section have already illustrated problems in GE (especially τρώγω), and
others will be found in later sections (notably ὀψάριον, §5). Here I explore
a typical noun frequently attested and with a range of meanings.
βίος
Just as in the verb κλαίω, both GE and LSJ present crowded entries
for βίος, with divisions into an array of senses without paragraphing.
GE has five major divisions (A–E), and within these, nine further sub-
groups marked by ||. LSJ’s major sense-divisions run from I to VIII, with
subdivisions 1, 2, 3 in the first. Let us summarise and compare the main
features of these two entries, with a look back at their predecessors.
20
GE’s translation, ‘the fresh and dry are eaten,’ is not accurate; nor does it appear to be
an accurate translation of GI: ‘questi vengono mangiati freschi e secchi.’ For ταῦτα =
τρωκτά ‘edible kernels’ see Waddell, Herodotus Book II, 205.
LSJ begin with a ham-fisted attempt at distinguishing βίος from ζωή:

‘life, i.e. not animal life (ζωή), but mode of life …, manner of living
(mostly therefore of men, v. Ammon. …; but also of animals …’. This
is difficult enough to grasp; a look at ζωή only makes it harder: there is
nothing remotely like ‘animal life’ to be seen there. The distinction is based
on a statement in Ammonius, and was added in LS7 (1890).21 Offering a
distinction is not a bad thing in itself, but relying on an ancient Atticist
grammarian is ill-advised. Ammonius is doing what he does, that is, sepa-
rating two apparent synonyms (in Classical Greek), by some simple and
rigid distinction, in this case that βίος is used only of humans, ζωή of
humans and beasts. A more reliable guide to the distinction may be found
in Chantraine’s DELG, which in a nutshell is: βίος = life as lived, ζωή =
life itself.22 Differences in application or reference are incidental. This open-
ing statement in LSJ makes the whole section I.1 more difficult to follow.
The rest of LSJ’s entry, however, is more or less satisfactory and clear.
Before turning to GE, it will be useful to note the content of Rocci’s
entry. Rocci does not mention or try to apply the human-animal distinction.
His division into five meanings a) – e) is simple but good. (One can see
clearly the influence of Bailly.) GI’s entry, however, shows enough differ-
ences from Rocci’s to indicate that extensive revision was undertaken for
GI. Comparison with LSJ shows that LSJ was the main source, not a new
analysis. Most of the senses in LSJ have been taken over, mostly in the
same places, sometimes rearranged. In GI and hence GE, the attempt to
distinguish βίος and ζωή is abandoned (and Ammonius is not quoted);
yet LSJ has left its mark in the survival of the human-animal distinction,
now set up as two separate senses, with LSJ’s examples relocated: ‘A life,
existence, of a human … B rar. life, of other beings.’23 Some other odd
things occur and the use of numbered sense-divisions versus subdivi-
sion by || is erratic. In particular there is a slip-up at ‘D later a life, biog-
raphy … || the world, mankind, humanity, people (the surroundings
in which one lives),’ which presents the latter as if it were a sub-sense of
the former. LSJ rightly separated these on an equal level, as ‘III. the world
we live in, ‘the world’’ and ‘V. a life, biography.’ Overall, then, the results
21
Ammonius 101 (ed. Nickau): βίος μὲν ἐπὶ τῶν λογικῶν τάσσεται ζῴων, τουτέστιν
ἀνθρώπων μόνων, ζωὴ δὲ καὶ ἐπὶ ἀνθρώπων καὶ ἐπὶ τῶν ἀλόγων ζῴων. LSJ’s quota-
tion from Plotinus 3.7.11 (III AD) is not a lot of help.
22
Chantraine, DELG, s.v. βίος: ‘Sens, non le fait de vivre, mais la manière de vivre … d’où
« moyens de vivre, ressources »’; s.v. ζωή: ‘« propriété d’être vivant, vie » par oppo-
sition à « mort ».’ There is a fair bit of overlap in actual usage.
23
A hare that was in LSJ’s I.1 has managed to slip through the net and stay in GE’s ‘A …
of a human’ (‘DEMOSTH. 18.263’).
of the revision in GI and GE are mixed, and it cannot be said that the entry
is an advance on LSJ.
A close look at the entries for βίος in GE and GI reveals something
further: they are identical in content. The only changes are the translations
from Italian into English and minor alterations in reference style (e.g.,
LUC. 51.25 > LUC. Pseudol. 25). It is a fair guess that the editors of GE
were able to work from the electronic files of GI, making only minimum
changes. There was no need to retype the Greek or the references; only
the translations needed to be done, addenda inserted, if any, and some
information relocated.24 The consequence of this process was of course
that little or no revision to the structure of the entry in GI could be, or was
made in GE.25
5. DOCUMENTARY EVIDENCE
The main purpose of lexicons is to supply meanings, but we also con-

sult them for other information, especially attestation. Innumerable discus-
sions in the scholarly literature offer statements like ‘first found in …,’
‘common in …’. On what are they based? On a lexicon usually, and usu-
ally LSJ. Lexicons thus have an important role to play in this regard. Yet
it cannot be said that the present state of play is satisfactory. In the LSJ
tradition the attestation or evidence for each word has accumulated unsys-
tematically over centuries. It is primarily employed, as is right, to establish
and illustrate the meanings, with the additional aim of supporting the read-
ing of standard texts. But there is no guarantee that the picture given is a
reliable guide to the spread and currency of the word. Until there is a
systematic gathering of evidence, with some sort of statistical information,
this situation will continue. We cannot expect GE to have solved this prob-
lem, but we would expect at least to see no loss of existing data, and the
addition of new data where significant. Let us take as a test documentary
evidence, that is, inscriptions and papyri, of which it is said GI/GE ‘makes
substantial use.’26
24
In verbs, some morphological material is moved from the end of GI’s entry to the
‘principal parts’ section after the headword in GE, leaving only ‘dialectal, poetic …
forms’ at the end.
25
For further samples of GE’s handling of lexical analysis, the reader might like to
compare ἰσχυρός with Chadwick, Lexicographica Graeca, 165–70, and ἕξις with Lee,
‘Hebrews 5:14’ (summary 165–6).
26
GE, p. vi; see also p. vii: ‘In addition to the updated language of our definitions, the
strengths of this volume include the incorporation of new evidence, especially from
epigraphical sources and papyri.’
5.1. εὐίλατος
GE: εὐίλατος -ον [εὖ, ἱλάομαι] kind, favorable, merciful with dat.
VT Ps. 98.8 etc.
GI: εὐίλατος -ον [εὖ, ἱλάομαι] benigno, propizio, misericordioso,
con dat. VT. Ps. 98.8 ecc.
Rocci: εὐ-ίλατος, ον [ῑ, ἱλάομαι] propizio; pietoso, I.; SET.
LSJ: εὐίλατ-ος [ῑ], ον, (ἵλημι) very merciful, of deities, PCair.Zen. 34.19
(iii B.C), IG … GDI … etc.; … Lxx Ps. …; also … PPetr. …;
later written … GDI… UPZ 109.6 (i B.C.).
The documentary evidence for this word, mainly in inscriptions, is

extensive and important, showing as it does that the LXX was using a cur-
rent term and how it was used.27 LSJ had quite a good collection. What
has happened to it in GI/GE? It has disappeared. Rocci at least had ‘I.’
(= ‘Iscrizioni’) to alert the reader; GI let it go. Hence in GE it looks as if
εὐίλατος is a word used only in the LXX.
5.2. ἀποσκευή
GE: ἀποσκευή … A baggage … || household goods VT Gen. 14.12,
al. … | all members of the house VT Exod. 10.24 | the whole
people, aside from the adult males VT Exod. 12.37 | in the army
attendants VT Iud. 7.2 || excrement, filth STRAB. … B suppres-
sion, elimination IOS. A.I. 18.2.4(41).
LSJ Suppl. (1996): … add ‘2 soldier’s encumbrances, i.e. family,
PBaden 48.9 (ii BC), UPZ 110.199 (ii BC); dependants, Lxx
Ge. 46.5, al.’
GE’s analysis is very odd to begin with; but the main point is that
no documentary evidence is mentioned, even though it is important for
the history of the word and indispensible to determining the meaning
in the LXX. LSJ Suppl. (1996) supplied some, enough to show the new
meaning family, dependants that appears in Ptolemaic Greek.28 Not only
ought some of this evidence to have been reported in GE but even a
brief look at it would have greatly improved GE’s account of the LXX
uses.29
27
See Aitken, No Stone Unturned, 98–102.
28
See further Lee, ‘Ἀποσκευή in the Septuagint’ [= Essay 2].
29
In Exod 10:24 ‘all members of the house’ is inexact: ‘family, dependants’ is more
like it. In Exod 12:37 what GE says is nonsense: the text is οἱ ἄνδρες πλὴν τῆς ἀπο-
σκευῆς, i.e., ‘the men apart from their dependants.’ In Jdt 7:2 it looks like ‘baggage
train.’
5.3. ὀψάριον
GE: ὀψάριον -ου, τό [ὄψον] course of a meal, a little dish, esp. of
fish ARISTOPH. fr. 45 PLAT.1 102.2 PHERECR. 32 etc.; ἔχει πέντε
ἄρτους κριθίνους καὶ δύο ὀψάρια he has five barley loaves and
two fishes NT John 6.9.
LSJ: ὀψ-άριον, τό, Dim. of ὄψον, Ar.Fr.45, Pl.Com.95, Pherecr.27,
Philem.32, Test.Epict.6.11, PPetr. 3 p.327 (iii B.C.), PCair.
Zen.440.3 (iii B.C.), etc.; … a jar of pickled fish, BGU1095.17
(i A.D.), cf. PRyl.229.21 (i A.D.), Ev.Jo.6.9, al., OGI484.12
(Pergam., ii A.D.).
LSJ Suppl. (1996): ὀψάριον, after ‘dim. of ὄψον’ insert ‘foodstuff,
esp. fish’ and add … SEG 26.382 (Athens)’.
Rocci: ὀψάριον, ου, τό, dim. di ὄψον, propr. piccolo companatico,
ma, in Atene, spec. pesciolino; pesce, vari COM. in AT. 385; P.;
δύο ὀψάρια,VOLG. duos pisces NT. Io. 6, 9.
As well as a complete lack of documentary evidence in GE, we find a

very unsatisfactory statement of meaning. LSJ had good data to work with
(added in LSJ), but committed the sin of not telling us the meaning, instead
relying on ‘Dim. of ὄψον’ to do the job. This is worse than useless: the
user is given the impression that ὀψάριον is a small ὄψον – whatever that
is – and turns to ὄψον only to see a range of meanings, any of which could
be relevant. LSJ Suppl. (1996) tried to remedy this but didn’t do it well:
‘foodstuff’ covers all food; ‘esp. fish’ vacillates. GE has made it worse.30
Rocci was good within his limits. What is needed, instead of guesses based
on a preceding lexicon, is a fresh assessment of all the data and a clear
statement of the meaning so far as it can be known.
5.4. Further examples
The three examples above barely touch on the amount of under-reporting

of documentary evidence in GE. In regard to the papyri alone, the extent
of it is dumbfounding given the claims made. Some are standard and
frequent technical terms, such as ἔντευξις (‘petition’). Here are 50 well-
attested words in papyri that lack any reference to that evidence in GE:
ἀμάω, ἄμφοδον, ἀξίνη, ἀξιόχρεως, βῆμα, διαθήκη, διάπρασις, δια-
πράσσω, διαστέλλω, διάστημα, διαστολή, διαστολικόν (no entry),
διασῴζω, διαταγή, διάταγμα, διάταξις, διατάσσω, ἐγγύη, ἐγγυητής,
ἔγγυος, ἐλαιών, ἔντευξις, ἐντυγχάνω, ἐξακολουθέω, ἐπικεφάλαιον,
ζωή, ζῷον, ἡμίονος, κάμινος, κοινωνός, κόκκος, λογίζομαι, μέρος,
30
GE’s ‘course of a meal, a little dish’ translates piatto di pietanza, un po’ di pietanza
in GI.
μεταβολή, μοναστήριον, ναύκληρος, νομός, ὀφείλημα, παράδεισος,

παραθήκη, πράκτωρ, πρόγραμμα, προσευχή, προστάσσω, πυρός,
ῥόα, σιτομέτρης, ὑποζύγιον, φερνή, χωρίον.
Of course this simply reflects the reality, that to try to graft onto a
lexicon of Classical Greek adequate coverage of even a part of the vast
body of later Greek is impossible without a giant effort.
6. NT AND LXX
Coverage of the NT and LXX in GE is difficult to assess fully (and

fairly). It is evident that references to these texts are incorporated fre-
quently throughout GI/GE (as they were in Rocci’s Vocabolario).31 More-
over, much of the NT/LXX vocabulary is no different from that of other
ancient texts, so most words and meanings are as it were automatically
covered. Even so, GE comes nowhere near providing the in-depth treat-
ment available in the specialist lexicons,32 and in some cases there are
significant gaps. This is the natural outcome of a lexicon tradition in which
the NT and LXX were treated as an adjunct to a lexicon of Classical
Greek. The coverage was never really satisfactory in LSJ or any of its pre-
decessors, even though these texts received some notice from the first edi-
tion onwards. GE’s material has come into existence in a similar way and
the results are similar.
6.1. θυμός
The old word θυμός has a range of meanings rather difficult to analyse,
but one, ‘anger,’ begins to predominate in Classical Greek and contin-
ues into Modern Greek.33 In the LXX there are 300+ examples of this
sense. A careful search in GE’s long and packed entry finds one citation:
‘2Kgs. 5.12,’ under ‘D. … || extens.’.34 Why this one was chosen (in GI)
is somewhat mystifying when so many others, e.g., from the Pentateuch
and Psalms, were available. (The reference is not taken from Rocci or
LSJ.) The LXX also has a few examples of θυμός in the sense of ‘mental
31
Cf. GE, vi: ‘Significant consideration is also given to later forms of Greek, in particu-
lar Greek of the imperial age … (Old and New Testament, Patristics, etc.), up to the
VIth century.’
32
BDAG for the NT; Muraoka, Lexicon for the LXX.
33
Chadwick, Lexicographica Graeca, 143, 149; full analysis of θυμός, 143–50.
34
= 4 Kgdms 5:12. This happens to be in the ‘Kaige recension’ of Kingdoms, not exactly
a good representative of LXX Greek.
disposition, ‘temper’,’ as Muraoka defines it, but there is no mention in GE.

As for the NT, where ‘anger’ is the predominant meaning too (16 or so
examples), GE cites none at all.
6.2. νουθετέω
Again, GE manages one example in the LXX (Job 36:12), under ‘2. pass.
to be admonished, reproved, corrected,’ rightly enough. But it gives none
from the NT, where the word occurs eight times. The NT evidence would
have been a useful – one might think indispensible – addition to sense 1. act.
‘advise, admonish’ etc., where currently all the citations are Classical except
two from Philo and Lucian.
6.3. λογικός
This is a difficult word, found some 9,000 times in Greek literature to
VI AD. It is commonest in philosophers and theologians (which is warn-
ing enough). The two NT examples present special challenges, especially
Rom 12:1, which is the subject of ongoing debate:
παραστῆσαι τὰ σώματα ὑμῶν θυσίαν ζῶσαν ἁγίαν εὐάρεστον τῷ θεῷ,
τὴν λογικὴν λατρείαν ὑμῶν.
to present your bodies as a sacrifice living, holy, pleasing to God, your
λογική service /(act of) worship.
LSJ steered clear of this one. GE has plunged in and come up with ‘||
later spiritual ORIG. … etc.; ἡ λ. λατρεία the spiritual cult NT Rom. 12.1’
(GI ‘il culto spirituale’).35 The other example, 1 Pet 2:2 τὸ λογικὸν ἄδο-
λον γάλα, escapes notice, perhaps fortunately.
6.4. ὑπομονή
‘Endurance, patience’ in the face of adversity, ὑπομονή, is an impor-
tant Christian concept, with parallels in Jewish and ‘secular’ thought. GE
has improved on LSJ by at least including one NT reference (Rom 2:7),
out of 30 or so. The LXX examples, however, do not rate a mention.36
35
Rocci did not include any NT example; but under λατρεία ‘2) culto; adorazione’ he
notes exx. in Plato and adds ‘Set.; NT.’ In Lampe, PGL, λογικός fills two columns.
Some renderings of Rom 12:1: KJV reasonable (< Tyndale < Vg. rationabile); RSV,
NRSV spiritual; NEB fn. for such is the worship which you, as rational creatures, should
offer.
36
LSJ had none either, but LSJ Suppl. (1996) makes some rather complicated LXX addi-
tions, including a new meaning hope (which may not be for the better).
More noteworthy, in the light of the statement already quoted (n. 31), is
the absence of any subsequent evidence like that seen in Lampe, or that
can be found by TLG searching (2000+ exx. to VI AD).
6.5. Other observations
In contrast with the above examples, GE’s treatment of the LXX

does sometimes show advances. An instance is seen in καθιζάνω, which
has a transitive use (‘cause to sit’) that LSJ missed. GI/GE includes it,
with Prov 18:16 as the example (also found in Job 12:18).37 The attes-
tation of the word given in GE, however, could have been fleshed out
significantly.38
To conclude this section, I have taken a sample of 50 NT words from
αὐτόπτης to ἄψυχος. In GE, 19 have a NT reference; 31 do not. Why the
variation I cannot discern: it appears that the NT material in GI/GE was
collected unsystematically. In the same sample the LXX is slightly more
often referred to, that is, in those words occurring in the LXX (31 of the
50): 14 have a LXX reference; 17 do not. So the hit rate for both NT
and LXX is less than half.
7. A PARTICLE
Should we expect lexicons to help with particles? Whether we should

or not, we certainly do. We hope to get a quick indication to help us out.
Few of us are ready to face the ordeal of reading Denniston’s famous book
or some other specialist study (inaccessible to us anyway). The lexicons
do try to cover particles, usually in some depth. In GE μέν, for instance,
is given a very extensive treatment filling three columns; the material
in LSJ is of similar length; Rocci was shorter but still thorough. We have
no reason not to trust most of this. Rather than try to examine μέν, let us
take a less frequent particle, καίτοι, to see (a) whether GE is an advance
on LSJ; and (b) how well both lexicons cover the data and usage of
καίτοι.
37
I owe this example to Claude Cox. For his discussion of GE, focused especially on
LXX Job, see his page at <academia.edu>. Cox has noted an important failing of GE,
the use of the LXX text of Rahlfs (1935), not the standard Göttingen edition where
available.
38
TLG searching produces many more exx. of καθιζάνω (c.60) than the 7 in GE, in a much
greater range of authors. There is also one in UPZ 1.78.20 (159 BC, record of a dream).
καίτοι
GE: καίτοι or καί τοι adv. certainly, truly, indeed: κ. ἐμοί yes, for
me also IL. 13.267 cf. EUR. Med. 344; καίτοι σοφοῦ παρὰ φωτὸς
εἰρημένον indeed said by a wise man SIM. 37.12 || advers. how-
ever, yet, nonetheless: κ. τί φημι and yet what am I saying
AESCHL. Pr. 101; καίτοι γε nevertheless XEN. Mem. 3.12.7 | con-
cess. although, even if: καίτοι περ although HDT. 8.53.
Rocci: καίτοι, certo; a dire il vero; veramente, κ. ἐμοί, certo, anche a
me, IL.13, 267; v. 1, 426; EU. Med. 344. – b) tuttavia; pure; per
altro; nondimeno, ATT.; κ. τί φημι; ma che dico? ESCHL. Pr. 101;
καίτοι γε, SEN. Mem. 3, 12, 7. – c) quantunque; sebbene, ATT.:
κ. περ, quantunque, ER. 8, 53.
Bailly: καίτοι, adv. 1 et certes, et en vérite [sic], Il. 13,267; ISOCR.
Pan. 67, etc. || 2 quoi qu’il en soit, cependant, toutefois, ATT.;
… κ. τί φημι; ESCHL. Pr. 101; … || 3 quoique, devant un part.
… renforcé par περ . καίτοι περ, HDT. 8,53.
LSJ: καί τοι, and indeed, and further, freq. in Hom. with one or more
words between, Il. 1.426, al.; καὶ σύ τοι E.Med.344; καὶ τἆλλά
τοι X.Cyr.7.3.10: once in Hom. as one word, Il.13.267. II. after
Hom. usu., and yet, to mark an objection introduced by the speaker
himself, freq. in Rhetorical questions, καίτοι τί φημι; A.Pr.101 …:
without a question …: strengthd., καίτοι γ᾽ … etc.: mostly sepa-
rated, καίτοι...γε … X.Mem.3.12.17, etc. …; so καίτοι περ v.l. in
Hdt. 8.53. III. with a participle, much like καίπερ, Simon. 5.9. …
It will give us some perspective to start from outside the lexicons. Den-
niston’s analysis finds four main uses of καίτοι (adversative; continua-
tive; logical; combined with other particles), with subgroups in some, a
total of 14 sections. Each is illustrated, in Denniston’s usual fashion, by
a wealth of examples, all from the Classical period except one. At a rough
count he gives 250 examples.39 Blomqvist covers the same ground for his
selection of Hellenistic authors (330–30 BC). He includes a table of fre-
quency in his Hellenistic texts and some Classical authors. It shows that
while there are fewer in later texts (70+), the numbers in Classical authors
add up to hundreds; these include high numbers in prose writers such as
Demosthenes (190+), Plato (120+), and Aristotle (200+).40
It is obvious that our lexicons offer an exceedingly limited selection
of evidence for καίτοι. GE cites only six examples. Only two of these are
from prose (Xen., Hdt.) and none are post-Classical. LSJ has more, a total
39
Denniston, Greek Particles, 555–64. At 561 he includes some totals for Isocrates
(× 125) and Lysias (× 106).
40
Blomqvist, Greek Particles, 35–45; table at 36. There are c.50 in papyri; c.15 in inscrr.
The TLG tally just to II AD is 5,000+.
of 27 (with ‘etc.’ sometimes), and they are slightly more representative,

including more from prose (D., Pl.) and even a couple of late authors (Plb.,
Philo). But the coverage is still very inadequate in LSJ, and in GE entirely
unsatisfactory.
LSJ appears to recognise three meanings or uses of καίτοι. The second
and third are clearly described, but the first is rather strange and unclear:
it seems as if and indeed, and further is frequent in Homer, but only if καί
τοι is split by another word (except once), and that this use occurs only a
couple of times later. The fit to the noted examples does not seem right
either: for instance, the first, Iliad 1.426 καὶ τότ᾽ ἔπειτά τοι εἶμι, (‘and
then after that I will go’), can hardly be an example of καί + τοι. Close
scrutiny of the tradition suggests the first meaning was never intended to
be a separate sense back in Passow and LS1 where it originated, but was
an ambient meaning to cover the word as a whole; treatment of Homer
was simply the next item in Passow. Proof of this is that sense I. was
originally and yet, yet in LS1 (< Passow doch, und doch, doch auch, den-
noch). It was changed in the 5th edition (1861) to and indeed, and further,
the meaning we have now, at the same time as a new sense II. and yet was
introduced. (LSJ’s sense III., the concessive use, was separated from the
beginning.) These changes caused confusion in sense I. The match between
these apparent three meanings in LSJ and Denniston’s analysis is a worry
too. Sense II. matches D.’s (1) ‘adversative’ use, and III. matches (1.vii)
the ‘concessive’ use; but where are D.’s (2) ‘continuative’ and (3) ‘logi-
cal’ uses? Various other problems could be pointed out but need not be
gone into here.
GE’s entry is no more than an attenuated remnant of its predecessors.
The first sense or use (there are no numbers, only || and |) has been fresh-
ened up with input from Bailly via Rocci, but the illustrations doubtfully
support it (one, SIM. 37.12, has migrated from sense [3]).41 The second
sense, again from Bailly via Rocci, is an inexact catch-all, and the trans-
lation of Xen. Mem. 3.12.7 (included in LSJ only to illustrate καίτοι
separated from γε) is simply wrong. The third is real but the example is
a dubious v.l. (many others were available).42
41
GE begins with Iliad 13.267, but Denniston (Greek Particles, 555) says ‘This com-
pound is not found in Homer …’; in Il. 13.267 (he says) ‘καί goes closely with ἐμοί
(‘Know that for me too’)’. In E., Med. 344, GE’s next ex., Denniston (541) takes τοι as
attached to σύ (i.e., ‘you are their father, you know’). [Simonides 37.12 (= earlier 5.9)
was in LSJ’s III., moved up in GI, hence in GE.]
42
The OCT (Hude) at Hdt. 8.53.1 prints καίπερ. On concessive καίτοι see Denniston,
Greek Particles, 559; Blomqvist, Greek Particles, 39–43.
This look into καίτοι has enabled us to see not just what is wrong with
the current entries but how their history has controlled the outcomes. The
major underlying problem is reliance on an existing lexicon entry as the
basis for the next, however much ‘revised.’ Unless the old entry is com-
pletely laid bare and all its data are reassessed, its problems will be passed
on or made worse. Hence our conclusion that (a) GE’s entry for καίτοι is
not an advance on LSJ’s; and (b) the coverage of data and usage is inade-
quate in both lexicons, but much worse in GE than in LSJ.
8. CONCLUSIONS
The foregoing samples may be a small selection but they are repre-
sentative of widespread phenomena in the new lexicon. They justify an
assessment as follows. (a) Is GE a ‘better’ lexicon than LSJ? No. (b) Is
it a competitor to LSJ? Yes; but only because it is a major lexicon aimed
at covering the same territory. (c) Is it the equal of LSJ? No. (d) Can it
replace LSJ? No. (e) Will it serve the purposes of the ordinary student
beyond elementary level? Yes, up to a point. (f) Can it serve the purposes
of scholars? Again, up to a point. (g) Does it apply an improved method
of defining meaning? No. (h) Is it better at lexical analysis than LSJ?
No. (i) Does it offer generally better coverage of meanings and attestation
than LSJ? No. (j) Does it actually provide better coverage of post-Classical
Greek, as it claims to do? No, very seldom. (k) Does it omit information
that LSJ offers? Yes, often. (l) Is its presentation easier to follow than
LSJ’s? To a slight extent only. (m) Is its numbering system easy to fol-
low? No. (n) What does it offer that LSJ does not? Limited additions to
the attestation of words and some meanings. (o) Is it basically a translation
of its predecessor GI? Yes. (p) Is GI basically a revision of a predecessor?
Yes.
GE is a useful lexicon, but it is not the great advance that we long for.
The shape of the lexicography of Ancient Greek has not been changed
by GE, but remains where it was. GE simply supplies another lexicon of
similar size and coverage to LSJ, doing the job no better and often rather
worse. The principal reason is that GE is a descendant of LSJ itself, with
some revisions, but with most of the faults as well as merits of LSJ pre-
served. The major, far-reaching overhaul that LSJ’s material needs has not
been undertaken, either by the editors of GE or by anyone.
Montanari rightly describes the situation in his preface to GE:
The Liddell-Scott-Jones dictionary is used by scholars of antiquity as

the reference dictionary for Greek, although the need for a profound
revision of LSJ itself or for a completely new dictionary of Ancient
Greek has been put forward many times and is indeed increasingly felt,
in order to take into account more organically the advances of knowl-
edge achieved over the years.43
I read these words as a true statement of where we are, not a claim that
GI/GE is itself the ‘profound revision’ that is needed. Montanari wisely
refrains from making such a claim. If after all he intended to imply it, or
others take his words in that way, it needs to be stated that there is no
justification for it.44
What, then, is to be done? Lexicons like GE are a legacy of the past.
They deal primarily with Classical Greek, and contain both too much and
too little. They retain old material and have difficulty adding new. They
also preserve the mistakes of their predecessors and have difficulty remov-
ing them. The coverage of post-Classical Greek is out of control because
there is too much material. Trying to graft small portions of it onto exist-
ing lexicons is unworkable. Lexicons like GE are now a holding opera-
tion. Rather than continue to produce new lexicons based on old ones,
the next generation of Greek lexicographers must rethink how to gather
and present their material. Gathering of data by electronic means is now
the only option; and a reappraisal of the meanings is essential. The whole
corpus of Ancient Greek is too vast to deal with at once, but a systematic
approach could move us forward. If the corpus were divided into manage-
able segments, the data could be gradually combined into one major elec-
tronic resource. Printed lexicons would still be needed, but could be slim-
mer and more selective, and based on better data.
43
GE, p. v.
44
The second preface by Nagy, Muellner, and Goh (p. vii) apparently sees GE as the
lexicon that incorporates ‘recent advances made in scholarship on the ancient Greek
world.’
27
GREEK IDIOM IN THE LXX-PENTATEUCH:
THE PREPOSITION PARA
2019
Abstract
This paper examines the preposition παρά in the LXX-Pentateuch, where it

occurs 223 times. A full analysis of the Hebrew matches shows that παρά renders
a great variety of equivalents. A lexical analysis, with parallels in non-LXX Greek,
is presented, showing that all the senses of παρά are in accordance with Greek
usage. The conclusion is drawn that the choice of παρά is not due to consistent
matching with any Hebrew equivalent but is determined by context and Greek
usage alone. The hypothesis is put foward that the same may be true of all prepo-
sitions in the LXX-Pentateuch. The value of the LXX as evidence for contem-
porary Koine Greek is also noted.
1. INTRODUCTION
Greek and Hebrew each have their own prepositions, and in each lan-
guage they are part of a web or system of relationships proper to that lan-
guage. While there are many ready matches between prepositions in the
two languages, the two systems do not exactly align. Even when a prepo-
sition in one language is the natural equivalent of one in the other, they
will almost certainly differ in their range of usage. It is also likely that there
will be some spare capacity on the Greek side. The number of Greek prepo-
sitions used in the Pentateuch is 17 (that is, ‘proper’ prepositions). The
Hebrew ‘simple’ prepositions are about the same in number, but most
of the Greek ones can be used with more than one case: about half are
used with two or three cases. The second most common preposition in
the Pentateuch, ἐπί, appears with three cases, making it in effect three
prepositions, not one. It also has a great variety of uses with each case.1 One
1
There are 39 senses (10 + 17 + 12) in Muraoka, Lexicon, s.v. ἐπί. The most frequent
preposition is ἐν (2,200 +), then ἐπί (1,400 +), followed by εἰς (1,200 +). Hebrew of
course expands its repertoire by ‘complex’ prepositions such as ‫לפני‬.
might wonder what Hebrew match or matches call for ἐπί to be used over
1,400 times. Only systematic enquiry can find out. The subject of how the
LXX translators use the Greek prepositions and match them to the Hebrew
has received limited attention.2 Investigation of this question is likely to
produce new insights into the translators’ methods and their familiarity
with Greek.
An examination of the preposition παρά in the Pentateuch is the sub-
ject of this paper.3 παρά occurs 223 times, with all three cases. This is
rather a lot for a preposition whose immediate Hebrew match is not obvi-
ous. One first wonders if some of the frequency could be due to a match
to Hebrew ‫אצל‬, which has the meaning ‘beside’ like παρά + acc. In fact
‫ אצל‬occurs only 10 times in the Pentateuch, only 7 of these are translated
by παρά, and only 4 are παρά + acc., ‘beside’ (the other 3 are παρά + dat.,
‘with, near’). That explanation does not take us far. What then are the
matches to παρά? The aim of this paper is to answer that question, and
to find out how παρά is actually used and why so frequently. Also to be
addressed is the question whether the usage of παρά in the Pentateuch is
in accordance with natural Greek usage. Some general conclusions from
the evidence will complete the paper.
2. THE HEBREW MATCHES OF ΠΑΡΑ IN THE LXX-PENTATEUCH
Let us first see a summary of all the Greek–Hebrew matches of παρά,

that is, what παρά corresponds to in the Hebrew text (of MT) at each
occurrence. The data are based on a full list I have compiled, which need
not be presented in its entirety here.4 The matches for the three cases are
as follows:
2
Some questions are explored in five papers by Soisalon-Soininen in Studien zur
Septuaginta-Syntax: ‘Wiedergabe hebräischer Zeitangaben mit ‫‘ ;’ב‬Comparative Expres-
sions with ‫‘ ;’מן‬ἐν für εἰς in der Septuaginta’; ‘Wiedergabe des ‫ ב‬instrumenti’; and
‘Wiedergabe des partitiven ‫מן‬.’ Cf. also Sollamo’s classic Semiprepositions, and ‘Repeti-
tions of Prepositions.’ Johannessohn, Präpositionen has good observations but is only a
beginning.
3
Revised version of a paper presented in the Greek Bible Section at the SBL Annual Meet-
ing, Denver 2018. My thanks to Jim Aitken for comments at the session and to Trevor
Evans for reading the final version; also to Siegfried Kreuzer for retyping the Hebrew
and an anonymous reviewer for helpful suggestions.
4
Prepared with the aid of ‘Accordance’ and checked against the printed Göttingen text.
The possibility of a different Heb. original of course arises, but affects very few
cases.
GREEK IDIOM IN THE LXX-PENTATEUCH: THE PREPOSITION PARA 429
2.1. παρά + acc.

Key: + = added without match in MT. >MT = text lacking in MT. ~ = corre-
sponds to.
‫אל‬ ‫ב אצל‬ ‫לפני‬ ‫מן מן‬# ‫ †על־ על‬other* + free >MT

Gen 1 1 3 2 3 2 1
Exod 1 2 1 2 1 3 2 3 6 2 1
Lev 6 2 3 4
Num 2 2 1 1 5 1 4 5 2
Deut 1 2 8 1 2 1
Totals 10 4 9 4 13 3 13 4 9 16 2 3
# In ‫מבל)ת(י‬ [Total 90]
† ‫על־יד‬, ‫על־שׂפת‬, (‫על־דבר )אשׁר‬
* Gen: ‫ל‬, ‫עם‬. Exod: ‫אחר‬, ‫בקצה‬, ‫מדי‬. Num: ‫עלי‬, ‫מעבר ל‬, ‫ ~( ימה‬παρὰ θάλασσαν)
2.2. παρά + gen.

‫ על מעם מן מלפני מיד מאת מאת ל את‬other* + free >MT
‫פניו‬
Gen 1 6 1 4 4 2 1
Exod 9 1 2 2 1 3 1 1
Lev 8 1 2 1 1
Num 1 16 1 1 1 6 1
Deut 5 1 1 4 1
Totals 1 1 44 1 5 4 9 8 2 1 12 2 1
* ‫ ~( מלבי‬παρ᾽ ἐμαυτοῦ) [Total 91]
2.3. παρά + dat.

‫אצל‬ ‫את‬ ‫ביד‬ ‫בין‬ ‫בית‬ ‫בעיני‬ ‫עם‬/‫עמדי‬ + free >MT
Gen 3 5 2 3 1+2 1 3
Exod 4 1 2 1 1
Lev 1 5
Num 1 2
Deut 2+1 1
Totals 3 10 2 1 3 3 14 2 1 3
[Total 42]
2.4. Comments
The striking thing is the great variety of the Hebrew matches. For all
three cases of παρά there are multiple Hebrew equivalents. The total, if
one unpacks the ‘other’ category, is over 30. We note also that παρά is
quite often added without a match in the Hebrew (30 times), and there are
a few ‘free’ renderings as well. Are there any constant matches? The only
one that stands out is παρά + gen. ~ ‫ ×( מאת‬44). But what is its signifi-
cance? It hardly constitutes a case of stereotyping, when the total is no
more than 69% of the renderings of ‫( מאת‬total 64), and there are signifi-
cant and deliberate variations from παρά.5 It is not in any case an exact
match or calque representing the two elements of ‫( מאת‬lit. ‘from with’):
παρά equates only to the ‘from’ element.6 Can we detect interference, that
is, use of παρά in an unGreek way caused by the match to ‫ ?מאת‬Hardly:
παρά + gen. (receive) ‘from’ (a person) is an entirely natural Greek use and
is exactly the right expression for the contexts in which it appears.
None of the other fairly frequent matches comes to a different result.
For instance, παρά renders 22 ‫ מן‬times (with two cases). But παρά is not
the obvious equivalent: that would be ἀπό or ἐκ. So interference can-
not be a factor. On examination it turns out that (a) παρά + acc., which
renders 31 ‫ מן‬times, mostly renders ‫ מן‬meaning ‘more than, beyond’, for
which it is a natural Greek expression;7 and (b) παρά + gen., which ren-
ders ‫ מן‬9 times, is mostly used for the meaning ‘from’ (a person), again
natural Greek,8 the same as the rendering of ‫מאת‬. Similarly, the match of
παρά + dat. to ‫עם‬/‫עמדי‬, which is found 14 times, comes about because
certain uses of παρά are more suitable in the contexts than the predicta-
ble equivalent μετά.9 In short, the choice of παρά is not governed by the
5
Summary of other renderings of ‫מאת‬: ἀπό × 12; ἐκ × 4; ὀπίσω × 1 (not = MT); omitted
× 3; total 20. Specimens: Exod 5:20 ἐκπορευομένων αὐτῶν ἀπὸ Φαραώ (‫;)מאת פרעה‬
Num 31:2 ἐκδίκει τὴν ἐκδίκησιν … ἐκ τῶν Μαδιανιτῶν (‫)מאת המדינים‬. These are
contextual choices.
6
Cf. Lee, Greek of the Pentateuch, 244 for further comments on ‫ מעם‬and ‫מאת‬.
7
Ten times: Gen 37:3; 43:34b; Exod 18:11; 33:16b; Num 12:3; Deut 7:6, 7a, b, 14;
10:15. The other three are: Deut 2:8 (I.2 ‘beside’); 7:8 (I.4 ‘on account of’); 9:28b
(I.4 ‘on account of’). Cf. Soisalon-Soininen, ‘Comparative Expressions with ‫מן‬,’ dis-
cussed in Lee, Greek of the Pentateuch, 30–1.
8
Seven times ‘from’: Gen 18:14; 23:13; 24:50; Exod 3:22; 12:35; Lev 25:33; Num 3:12.
The other two are: Gen 49:24(25) (unclear: II.2 ‘from’?) and Deut 22:3 (II.4 ‘by’).
9
παρά meaning ‘with, near, in the hands of’ (III.1): Gen 24:25; 31:32c; Exod 22:25(24);
Lev 47a, b; Deut 29:17(16); 32:34; meaning ‘in the household of’ (III.2): Gen 29:27;
Lev 25:35, 39, 40; Num 22:8, 9; Deut 15:16. This is not to say μετά is not used for ‫עם‬:
see Lee, Greek of the Pentateuch, 317–8 for full lists of ‫ ~ עם‬μετά + gen. and ‫ עם‬with
other renderings.
Hebrew word in the original but by other factors, namely context and
Greek usage (see further below, §5).
3. SAMPLES OF THE USAGE OF ΠΑΡΑ
For a closer look at how παρά is used, let us see the occurrences with
each case in three different books. The second-last column shows the
Hebrew match; the number in the last column refers to the lexical analy-
sis of παρά in §4.1 below.
3.1. παρά + acc. in Genesis
13:18 κατῴκησεν παρὰ τὴν δρῦν τὴν ‫וישׁב באלני ממרא‬ ‫ ב‬I.2
Μαμβρή
19:1 ἐκάθητο παρὰ τὴν πύλην Σοδόμων ‫ישׁב בשׁער־סדם‬ ‫ ב‬I.2
22:17 ὡς τὴν ἄμμον τὴν παρὰ τὸ χεῖλος ‫וכחול אשׁר על־שׂפת הים‬ ‫ על‬I.2
τῆς θαλάσσης
24:11 ἐκοίμισεν τοὺς καμήλους... παρὰ ‫אל־באר המים‬ ‫ אל‬I.2
τὸ φρέαρ τοῦ ὕδατος
25:11 κατῴκησεν... παρὰ τὸ φρέαρ τῆς ‫עם־באר לחי ראי‬ ‫ עם‬I.2
ὁράσεως
29:20 παρὰ τὸ ἀγαπᾶν αὐτὸν αὐτήν ‫באהבתו אתה‬ ℶ I.4
36:37 ἐκ Ῥοωβὼθ τῆς παρὰ ποταμόν ‫מרחבות הנהר‬ + I.2
37:3 ἠγάπα τὸν Ἰωσὴφ παρὰ πάντας ‫אהב את־יוסף מכל־בניו‬ ‫ מן‬I.3
τοὺς υἱοὺς αὐτοῦ
41:3a καὶ ἐνέμοντο παρὰ τὰς βόας ‫ אצל ותעמדנה אצל הפרות‬I.2
41:3b παρὰ τὸ χεῖλος τοῦ ποταμοῦ ‫על־שׂפת היאר‬ ‫ על‬I.2
41:17 ᾤμην ἑστάναι παρὰ τὸ χεῖλος τοῦ ‫הנני עמד על־שׂפת היאר‬ ‫ על‬I.2
ποταμοῦ
43:34b ἐμεγαλύνθη δὲ ἡ μερὶς Βενιαμὶν ‫ותרב משׂאת בנימן ממשׂאת‬ ‫ מן‬I.3
παρὰ τὰς μερίδας πάντων ‫כלם‬
49:13 Ζαβουλὼν παράλιος κατοικήσει, ‫והוא לחוף אניות‬ ‫ ל‬I.2
καὶ αὐτὸς παρ᾽ ὅρμον πλοίων
In most of these παρά has the one meaning, ‘beside’ (I.2), yet it corre-
sponds to six different Hebrew prepositions: ‫ב‬, ‫על‬, ‫עם‬, ‫אל‬, ‫עצל‬, ‫ל‬. In 49:13
there is some interpretation by the translator: the Hebrew has ‘and he
(will be) as/for a shore/haven(?) of ships,’ but the translator has made it
‘and he (will be) beside an anchorage of ships.’ There is interpretation

similarly in 36:37, which is a good example of added παρά: the Hebrew
says ‘Rehoboth of the river,’ but the translator clarifies with ‘R. beside
the river.’10 In all of the examples the translator says what he wants to
say using παρά to do it, instead of representing the Hebrew by a stock
equivalent.
In 29:20 there is an example of παρά + articular infinitive meaning
‘on account of,’ a use that may be unfamiliar to some but is well attested
in documents (see sense I.4 below). As can be seen, the translator has
declined to use ἐν as the match to ‫ ב‬and also converted the noun (‫)אהבתו‬
into an infinitive. The match παρά + acc. ~ ‫‘ מן‬more than,’ discussed
above, is exemplified in 37:3 and 43:34b.
3.2. παρά + gen. in Deuteronomy
2:6a βρώματα ἀγοράσατε παρ᾽ ‫אכל תשׁברו מאתם בכסף‬ ‫ מאת‬II.2

αὐτῶν
2:6b καὶ ὕδωρ μέτρῳ λήμψεσθε ‫וגם־מים תכרו מאתם בכסף‬ ‫ מאת‬II.2
παρ᾽ αὐτῶν ἀργυρίου
3:4 οὐκ ἦν πόλις, ἣν οὐκ ‫לא היתה קריה אשׁר לא־לקחנו‬ ‫ מאת‬II.2
ἐλάβομεν παρ᾽ αὐτῶν ‫מאתם‬
10:12 τί κύριος ὁ θεός σου ‫מה יהוה אלהיך שׁאל מעמך‬ ‫ מעם‬II.2
αἰτεῖται παρὰ σοῦ ἀλλ᾽ ἢ ‫כי אם־ליראה‬
φοβεῖσθαι...;
17:18 καὶ γράψει ἑαυτῷ τὸ ‫ מלפני וכתב לו את־משׁנה התורה הזאת‬II.2
δευτερονόμιον τοῦτο εἰς ‫על־ספר מלפני הכהנים‬
βιβλίον παρὰ τῶν ἱερέων
18:3a, b καὶ αὕτη ἡ κρίσις τῶν ἱερέων, ‫וזה יהיה משׁפט הכהנים מאת‬ ‫ מאת‬II.2,
τὰ παρὰ τοῦ λαοῦ, παρὰ τῶν ‫העם מאת זבחי הזבח‬ ‫ מאת‬II.2
θυόντων τὰς θυσίας...·
18:16 κατὰ πάντα, ὅσα ᾐτήσω ‫ככל אשׁר־שׁאלת מעם יהוה‬ ‫ מעם‬II.2
παρὰ κυρίου τοῦ θεοῦ σου
22:3 ὅσα ἂν ἀπόληται παρ᾽ αὐτοῦ ‫אשׁר־תאבד ממנו ומצאתה‬ ‫ מן‬II.4
καὶ εὕρῃς
23:15 οὐ παραδώσεις παῖδα τῷ ‫לא־תסגיר עבד אל־אדניו‬ ‫ מעם‬II.2
(16) κυρίῳ, ὃς προστέθειταί σοι ‫אשׁר־ינצל אליך מעם אדניו‬
παρὰ τοῦ κυρίου αὐτοῦ
10
παρὰ ποταμόν is a set phrase that does not require the article, as we might say ‘Rooboth-
on-river’: cf. Henley-on-Thames, Southend on Sea (and many others). Cf. Wevers,
Genesis, 607, dubiously: ‘One might have expected του ποταμου [sic], but the word
is here rationalized as meaning ‘beside a river’.’
23:21 ὅτι ἐκζητῶν ἐκζητήσει ‫כי־דרשׁ ידרשׁנו יהוה אלהיך‬ ‫ מעם‬II.2

(22) κύριος ὁ θεός σου παρὰ σοῦ ‫מעמך‬
33:23 καὶ ἐμπλησθήτω εὐλογίαν ‫ומלא ברכת יהוה‬ + II.2
παρὰ κυρίου
Here παρά ‘from’ (a person) is frequent (× 11) because it is, as was

said above (§2.4), the best fit in the contexts. In this sample it renders
‫ מאת‬often (× 5), but note that it renders ‫ מעם‬nearly as often (× 4). If we
try out ἀπό or ἐκ as an alternative to παρά in these places, we can see
that they are not well-suited to the contexts. In 10:12, for example, ἀπὸ
σοῦ might just be possible, but ἐκ σοῦ does not seem right at all; a check
confirms that αἰτῶ/-οῦμαι ‘ask for’ (from a person) is never followed in
the Pentateuch by ἀπό or ἐκ, only παρά, regardless of what the original
has.11 The ‘from’ words used with λαμβάνω, like παρά in 2:6b and 3:4,
are a longer story, but when ἀπό and ἐκ are used with this verb, which
they are, they have meanings different from παρά.12 Similar data could be
gathered for the other verbs in this sample.
Two other examples here are worth comment. In 18:3 the translator
has twice translated ‫ מאת‬by παρά, but in the first instance has inserted
τά before it. Why? To make the meaning clear as he sees it: ‘This is the
decision for the priests as regards the things from the people (τὰ παρὰ τοῦ
λαοῦ), from those sacrificing (παρὰ τῶν θυόντων).’ The first phrase gives
the subject-matter, the second specifies. The last occurrence in the table,
33:23, is another case of added παρά reflecting an interpretation by the
translator: ‘blessing from the Lord,’ instead of ‘blessing of the Lord.’13
3.3. παρά + dat. in Exodus
2:21 κατῳκίσθη δὲ Μωυσῆς παρὰ ‫ויואל משׁה לשׁבת את־האישׁ‬ ‫ את‬III.2

τῷ ἀνθρώπῳ
16:18 ἕκαστος εἰς τοὺς καθήκοντας ‫ אישׁ לפי־אכלו לקטו‬free III.2
παρ᾽ ἑαυτῷ συνέλεξαν
22:25 ἐὰν δὲ ἀργύριον ἐκδανείσῃς τῷ ‫אם־כסף תלוה את־עמי‬ ‫ עם‬III.1
(24) ἀδελφῷ τῷ πενιχρῷ παρὰ σοί ‫את־העני עמך‬
11
αἰτῶ/-οῦμαι + παρά: Exod 3:22 (‫ ;)מן‬11:2 (‫ ;)מאת‬12:35 (‫ ;)מן‬22:14(13) (‫;)מעם‬
Deut 10:12 (‫ ;)מעם‬18:16 (‫)מעם‬.
12
λαμβάνω + ἀπό or ἐκ in Deut: 1:15 (ἐξ ὑμῶν, ‘from among’ > MT), 23 (ἐξ ὑμῶν,
‘from among’ ~ ‫)מכם‬, 52 (ἀπὸ τοῦ καρποῦ, ‘some of’ ~ ‫ ;)מפרי‬2:62 (ἀπὸ τῆς ἀπαρ-
χῆς, ‘some of’ ~ ‫)מראשׁית‬, 4 (ἐκ τῶν χειρῶν σου, ‘out of’ ~ ‫)מידך‬.
13
It is not evident why παρά is inserted here in particular, when the same phrase ‫ברכת‬
‫ יהוה‬is translated by a simple genitive in Gen 39:5; Deut 12:15; 16:17; 33:13 (‫)מב ֶֹר ֶכת‬.
ְ
31:13 ἔστιν γὰρ σημεῖον παρ᾽ ἐμοὶ ‫כי אות הוא ביני וביניכם‬ ‫ בין‬III.1
καὶ ἐν ὑμῖν
33:12b οἶδά σε παρὰ πάντας, καὶ χάριν ‫ בעיני ידעתיך בשׁם וגם־מצאת חן‬III.3
ἔχεις παρ᾽ ἐμοί ‫בעיני‬
33:16a ὅτι εὕρηκα χάριν παρὰ σοί ‫ בעיני כי־מצאתי חן בעיניך‬III.3
33:21 ἰδοὺ τόπος παρ᾽ ἐμοί, στήσῃ ‫הנה מקום אתי ונצבת‬ ‫ את‬III.1
ἐπὶ τῆς πέτρας ‫על־הצור‬
35:23 καὶ παρ᾽ ᾧ εὑρέθη βύσσος ‫וכל־אישׁ אשׁר־נמצא אתו‬ ‫ את‬III.1
‫תכלת‬
35:24 καὶ παρ᾽ οἷς εὑρέθη ξύλα ‫וכל אשׁר נמצא אתו עצי‬ ‫ את‬III.1
ἄσηπτα ‫שׁטים‬
To render ‫ עם‬and ‫את‬, both roughly equivalent to English ‘with,’ the

translators of course frequently use μετά + gen.; but there are many other
renderings, and μετά is used for only 64% of the occurrences of ‫עם‬, and
47% of ‫את‬.14 For the meaning ‘with, near, in the hands of’ (III.1 below),
παρά + dat. is more appropriate. How often this is the case in the Pen-
tateuch overall cannot be pursued here, but in this sample we see it several
times. Thus in 33:21, ‘Here is a place near me, you shall stand on the
rock,’ παρ᾽ ἐμοί is more suitable than μετ᾽ ἐμοῦ (= ‘in company with
me’), which would hardly give the right sense.
The sample above also shows two instances of an interesting use of
παρά + dat. that emerged from the evidence of the documentary papyri,
namely, ‘in the household of’ (sense III.2 below). This is identifiable in
2:21 and 16:18 above, and several times elsewhere, including two notable
instances in Genesis (see below, §4.2). A case of interpretative rendering
appears in 31:13, where the Exodus translator changes ‘[my Sabbaths] are
a sign between me and between you’ (‫ )אות ביני וביניכם‬into ‘a sign with
me and among you’ (σημεῖον παρ᾽ ἐμοὶ καὶ ἐν ὑμῖν).15
4. LEXICAL ANALYSIS OF ΠΑΡΑ, WITH PARALLELS IN NON-LXX GREEK
I next offer a lexical analysis of the occurrences of παρά in the Penta-

teuch, with parallels in non-LXX Greek. Parallels were readily found in
III BC papyri (one II BC) for all the meanings except the first, one of the
oldest and most basic; but it was easily found in older sources from Homer
14
Lee, Greek of the Pentateuch, 317–20.
15
The standard rendering of ‫ בין‬is ἀνὰ μέσον. The Exodus translator again avoids it in
31:17 with reference to the Sabbath: διαθήκη αἰώνιος (17) ἐν ἐμοὶ (‫ )ביני‬καὶ τοῖς υἱοῖς
(‫ )ובין בני‬Ἰσραήλ· σημεῖόν ἐστιν αἰώνιον. There has to be an exegetical reason for this.
onwards. There can be no doubt about its currency in the language of the
third century BC.16 In this analysis the method of indicating meaning is
by gloss (in italics) and collocation (in brackets, plain text). An improved
method could be achievable.17 For each of the senses one example in the
Pentateuch is offered (sometimes two), with the Hebrew match to παρά.
4.1. Lexical analysis
I. παρά + acc.
1. (movement) alongside (a place)
Exod 2:5 αἱ ἅβραι αὐτῆς παρεπορεύοντο παρὰ τὸν ποταμόν. (‫)על־יד‬
Homer, Iliad 1.34 βῆ δ᾽ ἀκέων παρὰ θῖνα πολυφλοίσβοιο θαλάσσης.
And he went in silence along the shore of the loud-sounding sea.
Plato, Gorgias 511e.5 ταῦτα διαπραξάμενος ἐκβὰς παρὰ τὴν θάλατταν καὶ
τὴν ναῦν περιπατεῖ.
After carrying out (this service), he goes ashore and walks back and forth
beside the sea and the ship.
2. (to or at a position) beside, at (a place)
Exod 15:27 παρενέβαλον δὲ ἐκεῖ παρὰ τὰ ὕδατα. (‫)על‬
Lev 19:21 προσάξει... τῷ κυρίῳ παρὰ τὴν θύραν τῆς σκηνῆς... κριὸν πλημ-
μελείας. (‫)אל־פתח‬
PCairZen 1.119.6 εἰς βο|τανισμὸν τῆς παρ᾽ αἰ|γιαλὸν ἀρ(ουρῶν) ιε.
For weeding of the 15 arouras of land beside the shore.
PHeid 8.421.10 (II BC) ἐν τῷ τρ[ί]τῳ καταβαίνουσι | ἀπὸ τῆς . . υφης π[α]ρὰ
τὴν θύραν τοῦ | ἀνδρῶνος.
In the third (room) they come down from the ceiling(?) beside the door of the
men’s apartment.
3. more than, beyond
Num 12:3 Μωυσῆς πραῢς σφόδρα παρὰ πάντας τοὺς ἀνθρώπους. (‫)מן‬
PSI 4.422.34 ὁ πατήρ | μου προετέραι (= -ᾳ) παρὰ πάντας | τοὺς ἐκεῖ.
My father was superior to all the (farmers) there.
4. on account of (+ artic. inf.)
Exod 14:11 παρὰ τὸ μὴ ὑπάρχειν μνήματα ἐν γῇ Αἰγύπτῳ. (‫)מבלי‬
PCairZen 3.377.5 παρὰ τὸ μή σε χορηγεῖν ἡμῖν | τὰ κατ∖ὰ∕ τὴν συνγραφήν.
Because you have not supplied us with what was stated in the contract.
II. παρά + gen.

1. (movement) from (a place, person)
Lev 9:24 καὶ ἐξῆλθεν πῦρ παρὰ κυρίου. (‫)מלפני‬
16
A later example spans the gap: Mark 1:16 παράγων παρὰ τὴν θάλασσαν τῆς Γαλιλαίας.
17
As is being explored in the project ‘A Greek-English Lexicon of the Zenon Archive’
conducted by Trevor Evans and this author. The Pent. data might seem to require many
senses of παρά, but that corpus produces even more.
PMich 1.34.16 ἀπόλωλέμ μοι ὄνος θήλεια | λευκὴ ἐκπηδήσασα παρὰ |

Νικίου νυκτός.
I have lost one white female donkey which escaped from Nikias at night.
2. (obtain, receive) from (a person)
Lev 25:36 οὐ λήμψῃ παρ᾽ αὐτοῦ τόκον οὐδὲ ἐπὶ πλήθει. (‫)מאתו‬
PCairZen 1.120.2 διὰ τὸ προλαβεῖν παρ᾽ αὐτοῦ κερμάτιον εἰς ἐφόδια.
Because we received in advance from him money for travel expenses.
Idiom: παρ᾽ ἐμαυτοῦ (‘from my own resources’/ ‘on my own account’)
Gen 31:39 ἐγὼ ἀπετίννυον παρ᾽ ἐμαυτοῦ κλέμματα. (‫)מידי‬
Num 24:13 οὐ δυνήσομαι παραβῆναι τὸ ῥῆμα κυρίου ποιῆσαι αὐτὸ πονη-
ρὸν ἢ καλὸν παρ᾽ ἐμαυτοῦ. (‫)מלבי‬
PMich 1.58.27 προστέ|θεικα δὲ καὶ παρ᾽ ἐ|μουτοῦ (= ἐμαυ-) ἐρίων μν(ᾶν) α.
I have also added from my own resources 1 mina of wool.
3. (a person) attached to (a person)
Num 31:49 οἱ παῖδές σου εἰλήφασιν τὸ κεφάλαιον τῶν ἀνδρῶν τῶν πολε-
μιστῶν τῶν παρ᾽ ἡμῶν. (‫)אשׁר בידנו‬
PCairZen 1.33.7 ἠξίουν κατα|σταλῆναί τινα τῶν παρὰ σοῦ πρὸς Λυσίμαχον.
I asked that one of your people be sent down to Lysimachos.
PCairZen 4.532.24 τῶι παρ᾽ Ἀπολλωνίου Ζήνωνι.
To Zenon who is Apollonios’s agent.
4. (action) by (a person)
Deut 22:3 ὅσα ἂν ἀπόληται παρ᾽ αὐτοῦ καὶ εὕρῃς. (‫)ממנו‬
PCol 3.46.3 ...τὴν πα|ρὰ σοῦ δοθῖσαν γρα|φὴν τῶν ἱερέων.
[Contrary to] the list of pigs given by you.
III. παρά + dat.

1. with, near, in the hands of (a person)
Gen 39:15 καταλιπὼν τὰ ἱμάτια αὐτοῦ παρ᾽ ἐμοὶ ἔφυγεν. (‫)אצלי‬
Lev 19:13 οὐ κοιμηθήσεται ὁ μισθὸς τοῦ μισθωτοῦ παρὰ σοὶ ἕως πρωί. (‫)אתך‬
PCairZen 1.120.1 παρὰ Διονυσοδώρωι τῶι Διονυσίου υἱῶι κατελίπομεν |
βατιάκια β ἐν θήκηι.
With Dionysodoros son of Dionysios we left 2 dishes as security.
2. in the household of (a person)
Gen 40:3 καὶ ἔθετο αὐτοὺς ἐν φυλακῇ παρὰ τῷ ἀρχιμαγείρῳ εἰς τὸ δεσμω-
τήριον. (‫)בית שׂר בטבחים‬
Gen 40:7 ...τοὺς εὐνούχους Φαραώ, οἳ ἦσαν μετ᾽ αὐτοῦ ἐν τῇ φυλακῇ παρὰ
τῷ κυρίῳ αὐτοῦ. (‫)בית אדניו‬
PCairZen 2.192.5 ...γράψαι | [σο]ι περ[ὶ αὐ]τ[οῦ], ἵνα κατατάξῃς που αὐτὸν
παρὰ σοί.
[They asked me] to write to you about him, to give him some position with you.
3. in the estimation of (a person)
Exod 33:16a ὅτι εὕρηκα χάριν παρὰ σοί. (‫)בעיניך‬
PCairZen 3.481.18 γίνωσκε δὲ καὶ | παρὰ τοῖς κεραμεῦσιν | διαβολὴν
ἔχοντά με·
Know also that I am subject to malicious accusation among the potters.
4.2. Remarks
Brief comments may be made on some meanings and examples here.

The approach to idioms (i.e., fixed expressions that have a meaning
beyond that of the parts) is to place them under the meaning of the prepo-
sition itself, as one can best understand it. The idiom παρ᾽ ἐμαυτοῦ turns
up in the Pentateuch and is able to be paralleled in a contemporary text.
It is placed here under sense II.2, as the closest to what seems to be the
meaning of παρά itself.18
A further point related to the lexical analysis. The lexicographically-
inclined reader may notice that sense I.2 wraps together two meanings, one
implying movement to an end-point, i.e., ‘to’ a position beside something,
the other without movement, ‘at’ a position beside something. This is about
‘lumping’ and ‘splitting.’ The lexicographer has to decide if the difference
is significant enough to justify splitting, that is, setting up separate senses,
or not. In this case many of the occurrences are not clearly one sense or the
other, or may imply a bit of both, so separate senses are difficult to maintain.
Sense II.3 ‘attached to’ is a somewhat counter-intuitive meaning: one
expects παρά + gen. to mean ‘from.’ But this sense is well-attested in
documents, and the transition is illustrated by PCairZen 1.33.7 (in II.3
above), which could possibly be taken as ‘from.’ Num 31:49 is clear and
represents the meaning of the original: ‘Your servants have taken the total
of the fighting men who belong/-ed to us’ (τῶν παρ᾽ ἡμῶν ~ ‫אשׁר בידנו‬,
‘who are/were under our control’).19
Sense III.2 ‘in the household of’ has been noticed above in Exod (§3.3).
Most of the examples correspond to ‫עם‬, and some can be hard to dis-
tinguish from III.1 ‘with.’20 But in the two quoted (Gen 40:3, 7), the
original has ‫בית‬, ‘in the house’ (of), leaving little doubt how the Greek
was intended.21 Knowing that παρά has this meaning makes the narrative
18
The parallel is exact for the first, literal use (Gen 31:39); the other (Num 24:13) is
metaphorical. The next exx. in documents are BGU 8.1882.7 (I BC), SEG 25.501.56
(c.85 BC), both literal; there are others later. I have not searched literary Greek. LSJ,
s.v. παρά A.II.2, has exx. with 3rd pers. reflexive ἑαυτοῦ.
19
Other III BC examples of sense II.3: PCairZen 1.4.46; 2.247.2; 2.217.5.
20
Full list of definite and possible exx. of ‘in the household of ’: Gen 24:23 (‫)בית‬, 25 (‫;)עם‬
29:27 (‫ ;)עמדי‬39:2 (+); 40:3 (‫)בית‬, 7 (‫ ;)בית‬Exod 2:21 (‫ ;)את‬16:18 (free); Lev 25:35
(‫)עם‬, 39 (‫)עם‬, 40 (‫)עם‬, 47a (‫)עם‬, 47b (‫ ;)עם‬Num 22:8 (‫)עם‬, 9 (‫ ;)עם‬Deut 15:16 (‫)עם‬
[total 16]. Muraoka (Lexicon, s.v. παρά) includes four of these under II.1.a (I give M.’s
translations): Gen 40:3 (‘at the chief cook’s’), Num 22:9 (‘at your place’), Gen 29:27
(‘you will work with me’), 39:2 (‘he entered … the Egyptian master’s service’); but they
are combined with others under one meaning stated as: ‘indicates proximity, ‘beside,
at’.’ Other documentary exx. are: PCairZen 1.15.vrp.4; 3.491.19.
21
Also in Gen 24:23, a rather different context but evidently intended in the same sense:
εἰ ἔστιν παρὰ τῷ πατρί σου τόπος ἡμῖν καταλῦσαι; (...‫)הישׁ בית־אביך מקום‬.
clear, and proves that the translator was fully cognizant of what his origi-
nal meant to say.22
5. CONCLUSIONS
Let us begin with a summary of characteristics of the usage of παρά

in the LXX-Pentateuch, based on the foregoing data:
(a) Inconsistency in Greek–Hebrew matches
(b) Limited consistency in ‫ ~ מאת‬παρά + gen.
(c) No evident cases of Hebrew interference
(d) Full range of uses of παρά
(e) All natural Greek uses
(f) Used in free or interpretative renderings
(g) Added without any Hebrew match
The evidence for (a) is seen in the summary of the Hebrew matches in
§§2.1–3 and in the samples in §§3.1–3. It is hard to detect any consistent,
one-to-one matches, except in παρά + gen. ~ ‫ ;מאת‬but this match only
reflects the fact that παρά + gen. is the natural Greek way of expressing
the idea in the contexts (§2.4). The same is true of other somewhat fre-
quent matches (§2.4). Hebrew interference (c) cannot be identified if the
use of παρά is normal Greek – which it is in every case. The evidence for
(d) and (e) was found in the analysis of the uses of παρά in the Pentateuch,
with parallels in contemporary documentary texts (or earlier) (§4.1). The
significance of (f) and (g) is that they show that the translators were able
to deploy παρά idiomatically without being constrained by the original.
A final point may be added. Though a fine-grained study of each translator
would be needed for finality, the impression is strong that there are no
differences between the (five) translators in their usage of παρά. That there
were five translators I think likely on other grounds, but παρά does not
seem to offer help to prove it.
The conclusions I draw are the following.
(1) The Pentateuch translators do not follow any fixed equivalence in
the use of παρά but use it, more or less instinctively, when it best suits the
context in accordance with natural Greek.
(2) The evidence I think justifies putting this conclusion in a stronger
(perhaps surprising) form: it does not matter what the Hebrew preposition
is; the translator’s choice of παρά is determined by context and natural
22
Wevers (Genesis, 663) claims that the Hebrew original in Gen 40:3 ‘troubled’ the trans-
lator; but his own understanding is troubled by unawareness of this meaning of παρά.
Greek usage alone. This is not to say that the Hebrew preposition plays
no role: obviously the translator reads (or hears) the Hebrew text with the
preposition in it and takes in its meaning before choosing a rendering; but
which Greek preposition will be chosen (or whether there will even be one)
is not determined by the Hebrew preposition.
(3) On the basis of this result for παρά and the similar findings of
studies of κατά and the renderings of ‫ עם‬and ‫את‬,23 I here put forward a
somewhat bold proposal, namely, that the same conclusion as reached for
παρά applies to all prepositions in the LXX-Pentateuch: that is, the choice
of Greek preposition is generally if not always determined by the needs of
context and Greek usage, not the preposition used in the Hebrew original.
This is a hypothesis to be tested by further enquiry. It is intended to chal-
lenge the easy assumption that the translators’ usual practice was simply
to replicate the Hebrew preposition by a standard Greek equivalent. This
is not true of παρά, as we have seen, nor of κατά and the renderings of
‫ עם‬and ‫את‬.
(4) The Pentateuch translators’ use of παρά is further proof of their inti-
mate acquaintance with Greek to the point of native-speaker competence.
(5) The value of studying LXX usage alongside that of contemporary
Koine Greek is again demonstrated. There is a further outcome of that
approach: the LXX Pentateuch itself may provide evidence for contem-
porary Greek.24 Its value in this respect varies. It may do no more than
confirm an already well-attested meaning or use, as is the case with most
of the uses of παρά, or it may provide useful support, as in its attestation
of the idiom παρ᾽ ἐμαυτοῦ (II.2) and its confirmation of the meaning ‘in
the household of’ (III.2).
23
Lee, Greek of the Pentateuch, 154–9 (κατά written May 2015); 242–7 (‫ עם‬and ‫ את‬July
2014).
24
Cf. Lee, Greek of the Pentateuch, 5.
28
BACK TO THE QUESTION OF
GREEK IDIOM
20201
Abstract
The paper first discusses Soisalon-Soininen’s recognition of the presence of

idiomatic Greek renderings in the LXX, although he takes these to be departures
from the translators’ usual strict word-for-word method. The aim of the paper is
to identify clear cases of idiomatic Greek and to establish a method for recognising
them. An analysis of all renderings into three categories is proposed. Two cases
of idiomatic Greek are then presented in detail, the enclitic pronoun/ adjective τις
and the particle ἄν. The conclusions drawn are: (a) these features are evidence for
the translators’ native-speaker competence in Greek; (b) the LXX-Pent. should be
viewed as in essence Greek with Hebrew interference; and (c) its Greek has value
as evidence for Koine Greek. The last is illustrated by the ‘iterative’ use of ἄν +
indicative.
1. SOISALON-SOININEN AND GREEK IDIOM
The title of my paper intentionally echoes that of a 1990 paper by

Soisalon-Soininen, ‘Back to the Question of Hebraisms,’ providing both a
link to his work and a starting-point for the present discussion.2 That paper
was, naturally, about Hebraisms, which he saw as neglected in recent work
on literal versus free rendering in the LXX. He did not only discuss Hebra-
isms, however, but made significant mention of idiomatic renderings:
‘… nearly all the translators occasionally produce very skilful, idiomatic
renderings.’ He also said that ‘our work is based on (the question) which
1
This paper was presented at the Soisalon-Soininen Symposium on the Septuagint, held
in Helsinki in 2017. I very much appreciate the invitation to attend and the opportunity
to visit Helsinki for the first time, and most of all the kindness and hospitality of all those
involved in organising and conducting the Symposium.
2
Soisalon-Soininen, ‘Zurück zur Hebraismenfrage.’ English translation by Theo van
der Louw (‘Back to the Question of Hebraisms’) in Kauhanen and Vanonen, The Legacy,
283–300. Page refs. are to the original; translations are van der Louw’s with occasional
revisions.
renderings are Hebraistic and which are idiomatic.’ By idiomatic ren-

derings he evidently meant those that use natural, normal Greek; and the
work to which he referred was ‘to study the differences between individual
translators.’3
Soisalon-Soininen thus recognised the presence of idiomatic Greek in
the LXX, and his paper gave some illustrations, as well as examples of what
he termed Hebraisms. But it seems to me he did not really set out a means
of identifying the two, though he did give a definition of a Hebraism.4 As
to idiomatic renderings, he only said that they are ‘renderings which involve,
to a greater or lesser degree, a break with the customary word for word
approach’5 – which is rather limited and I think not true: a ‘word for word’
rendering may still be idiomatic Greek. (I admit to some trouble following
his thinking here, as at times elsewhere).
In other publications also, Soisalon-Soininen made observations on
the presence of idiomatic Greek in the LXX. I take two further papers that
show this and reveal more about the significance he attached to these fea-
tures. A 1973 paper dealt with the genitive absolute.6 This, as Soisalon-
Soininen of course knew well, is a purely Greek construction without any
equivalent construction in Hebrew, and therefore a case of idiomatic Greek.
It appears only sporadically in the LXX as a whole, though more frequently
in books like Job and Proverbs (as one would expect). When it does appear,
it gives us (he says) a glimpse of ‘the true language competence of the trans-
lators.’7 Though he does not quite spell it out, by ‘competence’ Soisalon-
Soininen evidently means competence or skill in Greek, not competence
in translation. He adds, moreover, in his summing up, that ‘we should give
3
‘Auf der anderen Seite bringen fast alle Übersetzer gelegentlich sehr geschickte idioma-
tische Wiedergaben’ (Soisalon-Soininen, ‘Zurück,’ 43); ‘[Ob wir es geradeheraus sagen
oder nicht,] so gründet sich unsere Arbeit dann darauf, welche Wiedergaben hebraistisch,
welche idiomatisch sind’ … ‘[Haben wir denn überhaupt die Möglichkeit,] die Unter-
schiede zwischen den verschiedenen Übersetzern zu untersuchen?’ (‘Zurück,’ 51).
4
‘Ein Hebraismus in der Septuaginta ist ein aufgrund des Hebräischen entstandener Aus-
druck, der nicht mit dem Sprachgebrauch des gleichzeitigen Griechisch im Einklang steht
oder der vom Hebräischen aus erklärliche Gebrauch möglicher Vokabeln oder Ausdrücke
im Griechischen in Zusammenhängen, in denen sie nach dem griechischen Sprachgebrauch
nicht passend sind’ (Soisalon-Soininen, ‘Zurück,’ 39).
5
‘… Wiedergaben, die einen größeren oder kleineren Eingriff in die übliche Wiedergabe
der einzelnen Teile des Ausdrucks voraussetzen’ (‘Zurück,’ 43).
6
Soisalon-Soininen, ‘Gebrauch des genetivus absolutus.’ English translation by Theo
van der Louw (‘The Use of the Genitive Absolute in the Septuagint’) in Kauhanen and
Vanonen, The Legacy, 223–8. Page refs. are to Studien; translations are van der Louw’s.
7
‘[D]ie wirklichen Sprachkenntnisse der Übersetzer’ (Soisalon-Soininen, ‘Gebrauch,’ 180).
He also speaks of ‘ihr wirkliches Sprachgefühl and ihr Können’ (176) und ‘die Geschick-
lichkeit … der Übersetzer’ (180).
BACK TO THE QUESTION OF GREEK IDIOM 443
more weight to these rare cases,’8 a sentiment that is in harmony with the
aim of this paper – though I would not agree that they are rare.
Clearly, then, Soisalon-Soininen recognised the presence of Greek idiom
in the occasional appearances of the genitive absolute. Yet he adds the
(to me) curious interpretation that ‘in those cases, the constraining factors
stepped back for some reason.’9 What are those ‘constraining factors’?
What does he mean? Evidently they are the constraints imposed by attempt-
ing to produce a close rendering of the original Hebrew. This is deducible
from Soisalon-Soininen’s preceding discussion, where he speaks of the
syntax of the LXX as ‘translationese’ and asserts that ‘one could say that
it is not Greek but rather Hebrew with Greek words.’10 Free renderings,
then, are not expected, they are not supposed to occur; the translators
were supposed to render the Hebrew exactly and not depart from it. And
why ‘for some reason’? Could the answer be in doubt? Soisalon-Soininen
does not explain further or seem to want to explore why a translator might
depart from the strict letter of the original.
My third sample is Soisalon-Soininen’s 1977 paper on the relative pro-
noun.11 The main interest here is in what Soisalon-Soininen calls ‘nominal
relative clauses,’ that is, those in which a Hebrew relative clause introduced
by ‫ אשׁר‬is translated as a noun phrase with the article, as for example:
Gen 43:16 ‫ = יאמר לאשׁר על־ביתו‬he said to the (one) who (was) over his
house.
καὶ εἶπεν τῷ ἐπὶ τῆς οἰκίας αὐτοῦ = he said to the (one) over his house.
Gen 19:11 ‫ = את־האנשׁים אשׁר־פתח הבית‬the men who (were) at the door
of the house.
τοὺς ἄνδρας τοὺς ὄντας ἐπὶ τῆς θύρας τοῦ οἴκου = the men the (ones)
being at the door of the house.
The renderings are natural Greek idiom; the literal equivalent with a rela-
tive pronoun would have been possible but less natural. Soisalon-Soininen
describes such forms of expression as ‘idiomatic, one might even say stylish
from the view-point of the Greek language’ and ‘good Greek style’ (57).
His appreciation of these as Greek idiom is thus plainly and well stated.
8
‘[M]üssen wir das Gewicht eben auf diese seltenen Fälle legen’ (Soisalon-Soininen,
‘Gebrauch,’ 180).
9
‘Es sind aus irgendeinem Grunde die hemmenden Faktoren zurückgetreten’ (Soisalon-
Soininen, ‘Gebrauch,’ 176).
10
‘Übersetzungssprache’; ‘Man könnte sagen, es ist kein Griechisch, sondern Hebräisch
mit griechischen Worten’ (Soisalon-Soininen, ‘Gebrauch,’ 175).
11
Soisalon-Soininen, ‘Hebrew Relative Clause.’ Page refs. are to Studien; the original is
in English.
But he nevertheless seems to want to push – he does push – the Greek into
a closer match to the Hebrew. He does this by an argument that equates
the article with the relative pronoun: ‘the [Greek] relative pronoun is
replaced by the almost identical article’ (56), and ‘in some cases the form
is actually identical’ (57). One may respond that to equate the Greek arti-
cle and relative pronoun is to ignore their difference of function; and it
is not really true that the forms are ‘almost identical.’12 But on this basis
Soisalon-Soininen goes on to assert that ‘this type of translation is not to
be taken as a significant indication of the liberties taken by the Septuagint
translators’ (57). We are thus back to the position that the translator is fol-
lowing his Semitic original exactly and not really departing from it, even
when he turns a Hebrew relative clause into a natural Greek equivalent
with different syntax – and does depart from it.
2. A DIFFERENT VIEW
This brief exploration of three of Soisalon-Soininen’s papers suffices

to give an indication of his thinking on the subject of Greek idiom in the
LXX. Now my aim here is not to criticise everything he said, or to dimin-
ish the value of all the fine work he did. Rather, I wish to bring the focus
on to the element that he himself identified, the idiomatic Greek ren-
derings. He and I agree that they are present; how we ought to view them
is where we differ. He evidently sees them as somehow unexpected or
anomalous, even difficult to explain, in translations presumed to follow the
original as exactly as possible. To me this is to approach the matter around
the wrong way. The LXX translators, at least in the Pentateuch, were not
tied to producing a close, exact translation, which they departed from only
when forced to – for some strange reason. Rather, natural Greek, of which
they had an intimate knowledge, was their starting-point, and they devi-
ated from it because of their generally literal method. They follow this
method with great inconsistency: they are just as likely to produce natu-
ral Greek as to produce Hebraic Greek. That is why natural Greek idiom
keeps appearing.
It may be said, of this or that Hebrew feature, that an idiomatic render-
ing is statistically infrequent compared with a Hebraistic one, and there-
fore the former is of little or no significance: the translators’ aim was still
to make the translation as literal as possible, and deviations are in the nature
12
The only forms that are the same are ὁ, ἡ, οἱ, αἱ, and these are distinguished by the accent
(rel. pron. ὅ, ἥ, οἵ, αἵ). Further, ὅ is neuter nom./acc., so cannot function like masc. ὁ.
of anomalies. Again I take a contrary view. That a certain feature of Greek

idiom appears at all is significant, all the more so if it involves an unre-
quired or unforced change, as in the case of a shift of word-order. One
instance of that kind is just as telling as many.
Whether one agrees with all this or not, the question of the presence of
idiomatic Greek in the LXX is a very important one, and I suggest is worth
pursuing for its own sake. The aim of this paper is to demonstrate the pres-
ence of incontestable cases of Greek idiom in the LXX-Pentateuch. The
enclitic pronoun/adjective τις and the particle ἄν will be our examples
for this purpose. But we need to begin with sound criteria for identifying
Greek idiom.
3. IDENTIFYING GREEK IDIOM
The question of Hebrew interference in the Greek of the LXX has

been the subject of enduring discussion, in various guises, since the days
of Thackeray. But the sort of methodological framework suited to our
purpose has not yet been proposed (as far as I know). I suggest that it is
helpful for the present purpose to divide all LXX renderings into three
categories, as set out below. This is not, of course, a complete system for
analysing all features of a rendering: it is aimed at finding renderings free
of interference. I recognise also that the categories are broad and the ter-
minology imprecise. But further refinements can follow, and some will
emerge in the discussion below. The illustrations all involve Hebrew ‫אישׁ‬
in some combination, and all have been chosen from the Leviticus trans-
lator, deliberately so, to show that he has no less control of Greek idiom
than his fellow-translators. These are the three categories of rendering that
I propose:
1. The rendering matches the Hebrew and is not natural Greek: inter-
ference is present. Example:
(1) Lev 17:3 ἄνθρωπος ἄνθρωπος τῶν υἱῶν Ἰσραήλ, ὃς ἂν σφάξῃ μόσχον …
(4) καὶ λογισθήσεται τῷ ἀνθρώπῳ ἐκείνῳ αἷμα·
A man a man of the sons of Israel who slaughters a calf … (4) then blood-
shed shall be reckoned to that man.
‫( דם יחשׁב לאישׁ ההוא‬4) ...‫אישׁ אישׁ מבית ישׂראל אשׁר ישׁחט שׁור‬
2. The rendering matches the Hebrew and is consistent with natural

Greek: interference may or may not be present. Example:
(2) Lev 25:17 μὴ θλιβέτω ἄνθρωπος τὸν πλησίον· καὶ φοβηθήσῃ κύριον τὸν
θεόν σου.
Let a man not oppress his neighbour; and you shall fear the Lord your God.
‫ולא תונו אישׁ את־עמיתו ויראת מאלהיך‬
3. The rendering does not match the Hebrew and is natural Greek:
interference is not present. Example:
(3) Lev 6:40(7:10) καὶ πᾶσα θυσία ἀναπεποιημένη ἐν ἐλαίῳ καὶ μὴ ἀναπε-
ποιημένη πᾶσιν τοῖς υἱοῖς Ἀαρὼν ἔσται, ἑκάστῳ τὸ ἴσον.
And every sacrifice made up with oil or not made up shall be for all the sons
of Aaron, to each an equal portion.
:‫וכל־מנחה בלולה־בשׁמן וחרבה לכל־בני אהרן תהיה אישׁ כאחיו‬
Brief explanations will be useful. Example (1) shows a rendering that
is not acceptable as natural Greek, even though the meaning would not
be totally opaque to a Greek speaker. Comprehension would be aided by
‘distributive doubling,’ a known though rare phenomenon in Greek, as
well as by context. But a natural idiomatic Greek equivalent would be dif-
ferent, and one is produced by the Leviticus translator himself in Lev 20:2
ἐάν τις ἀπὸ τῶν υἱῶν Ἰσραὴλ … (‫)אישׁ אישׁ מבני ישׁראל‬. Renderings in this
first category can be labelled ‘Hebraisms,’ if one wants to use the term.
In renderings that fall into category 2 – and they would probably be
the great majority of all renderings – the question of interference remains
open. Although they match the Hebrew, they are not inconsistent with
natural Greek, that is, they are capable of being read as natural Greek, and
are therefore not certainly Hebraisms. Example (2) is in fact not an entirely
natural Greek form of expression: we would be unlikely to find it in a free
Greek text. But one cannot say it would be impossible in terms of Greek
grammar, syntax, or word-usage.
Renderings of type 3 are by definition one thing and not the other.
They are undoubted cases of Greek idiom because they do not match
the Hebrew. In example (3) the Hebrew literally means ‘a man as/like his
brother.’ The translation departs from a literal rendering and gives an idio-
matic Greek equivalent that conveys the sense of the original in different
words, literally ‘to each the equal (portion).’ The insertion of τό is Greek
idiom (where English uses ‘an’). No better proof of the Leviticus transla-
tor’s grip on Greek could be found than this rendering.
Some further observations arising from the discussion in the Helsinki
Symposium and afterwards may be added here.13 As regards the term
13
Discussion of LXX syntax may advance even in an ice-cream shop or at a tram-stop
(thank you Theo van der Louw and Jim Aitken).
‘Hebraism,’ I would reserve it for individual renderings of type 1 alone,

where natural Greek is by definition not present; I would not use it for
those of type 2, where there is nothing wrong with the rendering as Greek.
This means that a great many renderings are left open; but methodologi-
cally that is sounder than classing everything that matches the Hebrew as
a ‘Hebraism.’ Other considerations may come into play to help decide
the question. Many of the renderings of type 2 seem to me to be a matter
of style. That is, they reflect the style of the original, which like the content
is carried over into the translation. It would be surprising if this were not
so – unless we think that a Hebrew book can be easily converted into a
Greek one. Example (2), as we have noted, may not be the sort of thing a
Greek writer would say, but it is acceptable Greek. The reason for the odd-
ity is that the form of expression, like the content, comes from the original.
Style is also the answer, in my opinion, to the perennial issue of fre-
quency. If a Hebrew expression rendered by a natural Greek equivalent
appears more frequently than in a normal Greek text, is it a ‘Hebraism’?
My answer is definitely no. Since each individual instance is not a Hebraism,
the frequency only reflects that of the original, so it is a feature of style.
Soisalon-Soininen himself expressed it well in his discussion of ἰδού: ‘So,
even ἰδοὺ φωνὴ μεγάλη is not a Hebraism per se. But the frequency of
ἰδού gives the text a Hebraizing flavour.’14 The term ‘Hebraizing flavour’
is of course inexact, but I am not sure of a better alternative. ‘Stylistic inter-
ference’ from the Hebrew original might work and is worth a trial.
It is only renderings of type 3 that provide us with undoubted idiomatic
Greek. We turn now to examples.
4. THE ENCLITIC PRONOUN/ADJECTIVE ΤΙΣ
The enclitic pronoun/adjective τις is found 79 times in the LXX-

Pentateuch.15 A full analysis appears in my book and the data presented
here are derived from it.16 First let us consider the matches between τις and
the Hebrew original. Hebrew has no exact equivalent, since it does not
14
‘So ist sogar ἰδοὺ φωνὴ μεγάλη an sich kein Hebraismus. Das sich oft wiederholende
ἰδού gibt aber dem Text ein hebraisierendes Kolorit’ (Soisalon-Soininen, ‘Zurück,’ 41).
Similarly: ‘… eine wiederholte Verwendung … dem Ganzen ein hebräisches Kolorit
geben kann’ (39). Cf. Sollamo, Semiprepositions, 299: ‘… exceptionally frequent usage
[of the studied expressions] in the LXX is Hebraistic (a stylistic Hebraism).’
15
Initial search via Accordance©; readings checked in the Göttingen edition. I do not
include Exod 22:9(8) ὅ τι οὖν [sic Gött. ed.], which is an instance of the relative pronoun
ὅστις + -οὖν.
16
employ enclitics like this at all. The Hebrew word for the idea is ‫אישׁ‬, which
has the primary meaning ‘man,’ with a subsidiary function to mean ‘some-
one, a person.’ The Pentateuch translators do sometimes use the ‘literal’
equivalent ἀνήρ or ἄνθρωπος to render ‫ אישׁ‬meaning ‘someone, a person,’
as in the renderings by ἄνθρωπος in examples (1) and (2) above. Mostly,
however, they employ τις or ἕκαστος (‘each’), both of which are a move
away from a ‘literal’ rendering in favour of more natural Greek.17
Summary data for τις and the corresponding Hebrew in the Pentateuch
are as follows:
Total of τις in the Pentateuch: 79
In clauses: 54
In noun phrases: 22
In μή τι: 3
With equivalent in Hebrew: 43
~ ‫אישׁ‬: 40
~ ‫אחד‬, ‫מאומה‬: 3
Without equivalent in Hebrew, i.e., added: 36
Position of τις:
In same position as Hebrew equivalent: 19
Repositioned: 24
The tabulation includes data on the position of τις in its clause – and
this is where the story becomes really interesting. Like other enclitics, τις
conforms to certain position rules. The main one is that it tends to gravitate
to the second position in its clause, or the ‘second slot’ (to use a term of
my own). Position in the second slot is a manifestation of Wackernagel’s
Law, best known for its effects on the position of the enclitic personal pro-
noun.18 For τις I have formulated 11 position rules, based on a close analy-
sis of the 79 occurrences in the Pentateuch, supported by parallels in non-
LXX Greek. This list of rules is drawn on here and presented in full in my
book.19
The question of interest is how much the position of τις matches that
of its equivalent ‫אישׁ‬. Some might expect it to match consistently, being
changed only where ‫ אישׁ‬is first word and τις has to be moved, or perhaps
not even then. In fact, while the position of τις often matches that of ‫אישׁ‬,
it is more often changed (24 times), the new position being of course in
17
ἕκαστος is found 80 times in the LXX-Pent. and corresponds to ‫ )אשׁה( אישׁ‬or some
combination thereof in 77 of them. This word ἕκαστος, too, could be used to demon-
strate the presence of idiomatic Greek.
18
See, e.g., Horrocks, Greek, 108–9; Sollamo, ‘Enclitic Personal Pronouns.’
19
comformity with the Greek position rules. Moreover, in a large proportion

of the examples (36 out of 79), τις is added without any match to a word
in the original. Why would this be? The answer can only be that the addi-
tion of τις is natural Greek idiom.
Let us now consider τις in relation to the categories of rendering defined
above, with selected examples and parallels. There are no instances of
τις in category 1, that is, matches to the Hebrew that are also not natural
Greek. In all the cases where τις keeps the same position as ‫אישׁ‬, none
are unnatural Greek positions. Where keeping the Hebrew order would
have led to an unnatural position, as in example (7) below, the translators
avoided it by moving τις to a new position.
There are instances of τις in category 2, that is, those in which τις keeps
the same position as ‫( אישׁ‬or another Hebrew equivalent), and is in a natu-
ral Greek position. According to my analysis there are 19 of these. Two
examples of τις in category 2:
(4) Exod 16:20 ἀλλὰ κατέλιπόν τινες ἀπ᾽ αὐτοῦ
‫ויותרו אנשׁים ממנו‬
BGU 6.1280.17 (210 BC) ἐὰ]ν δὲ μὴ βεβαιοῖ ἀλλʼ ἐγβάλλῃ τις Διοκλῆν
Pl. Theaet. 175c.1 ὅταν δέ γέ τινα αὐτός, ὦ φίλε, ἑλκύσῃ ἄνω, καὶ
ἐθελήσῃ τις αὐτῷ ἐκβῆναι
(5) Exod 22:5(4) ἐὰν δὲ καταβοσκήσῃ τις ἀγρόν
‫כי יבער־אישׁ שׂדה‬
BGU 14.2367.17 (III BC) ἐὰν δʼ ἐπάναγκες δανείζῃ τις μετα[---]
Pl. Symp. 221e.1 εἰ γὰρ ἐθέλει τις τῶν Σωκράτους ἀκούειν λόγων
Theoretically these could be due to interference, but other considerations

apply that make this unlikely to be the main reason for the placing of τις.
In example (4), the translator has good reason not to reposition τινες before
the verb, though this is a standard position in the case of ἐὰν δέ at the head
of the clause, as in examples (6) and (7) below: τις cannot be placed imme-
diately after a co-ordinating conjunction such as ἀλλά. In example (5), in
which τις is ‘postponed’ till after the verb when it could have gone before,
there are two possible factors to consider: (a) Exod 22:5(4) occurs in a
rather lengthy string of regulations, many of them expressed in the same
form as in example (6) below, that is, ἐὰν δέ τις + verb.20 It is possible
the translator placed τις in a different position simply for variety; (b) the
second factor is a possible link to historical developments in the placement
20
Exod 21:7, 14, 20, 26, 33; 22:1(21:37), 7(6), 10(9). With the order ἐὰν δέ + verb + τις:
Exod 22:5(4), 14(13), 16(15).
of enclitics. Over time Greek shows a tendency towards relocation of

enclitics to a position after the verb.21 As can be seen in the parallel above,
this was already happening in the time of Plato. It is possible the translator
was unconsciously influenced by this undercurrent. Of course it could have
been a little of both factors. At any rate, both these examples are definitely
natural Greek and this explains them better than a resort to interference
from the Hebrew. This is true of τις throughout the Pentateuch.
But it is the instances that fall into category 3 that provide incontestable
cases of Greek idiom. In these, τις is repositioned or added, contrary to
the Hebrew, while being at the same time natural Greek. Examples of τις
in category 3:
a. τις corresponds to ‫אישׁ‬, but is repositioned

(6) Exod 21:7 ἐὰν δέ τις ἀποδῶται τὴν ἑαυτοῦ θυγατέρα
‫וכי־ימכר אישׁ את־בתו‬
(7) Lev 25:29 ἐὰν δέ τις ἀποδῶται οἰκίαν οἰκητήν
‫ואישׁ כי־ימכר בית מושׁב‬
SB 18.13256.9 (III BC) ἐὰν δέ τις παραθῆται, ἀποτίνειν αὐτόν …
Pl. Phaedo 60b.6 ἐὰν δέ τις διώκῃ τὸ ἕτερον …
b. τις is added, without correspondence in the Hebrew

(8) Gen 6:5 καὶ πᾶς τις (‘everyone’) διανοεῖται ἐν τῇ καρδίᾳ αὐτοῦ
‫וכל־יצר מחשׁבת לבו‬
UPZ 1.60.17 (179–168 BC) ἀλλὰ πᾶς τις πειρᾶται
Pl. Theaet. 178a.5 ἔτι τοίνυν ἐνθένδε ἂν μᾶλλον πᾶς τις ὁμολογήσειεν
In (6) the translator moves τις to the standard position in the second
slot, where the Hebrew equivalent ‫ אישׁ‬came later, after the verb. There is
another word there as well, the particle δέ, and τις is placed after it. This
order is not accidental: there are word-order rules within the second slot.
(More will be said on these under ἄν.) In example (7) the position of ‫אישׁ‬
could not have been kept without breaking one of the basic rules for τις,
that it is not placed first in its clause. Even καί τις (~‫ )ואישׁ‬would not have
been idiomatic. In (8) τις is just a natural addition to πᾶς in the manner of
idiomatic Greek since Classical times.
So τις is a good example of Greek idiom in the LXX-Pentateuch, indu-
bitably in the cases in category 3 and potentially in all cases in category 2.
21
See Horrocks, Greek, 108–9.
We may note again the point made earlier that the mere choice of τις as
opposed to a ‘literal’ equivalent of ‫ אישׁ‬is a step towards more idiomatic
Greek.
5. THE PARTICLE ἌΝ
The particle ἄν occurs just over 400 times in the Greek Pentateuch.22
It is an entirely Greek feature without any counterpart in Hebrew, and there-
fore an instance of pure Greek idiom. It occurs in the LXX as a component
of various clause types, as required by Greek syntax. These are familiar
to us, even if we do not always notice ἄν or realise that it is not there by
accident. Most of these clause types are illustrated in the examples below;
parallels in contemporary (III BC) papyri are added. The ‘indefinite con-
struction’ in temporal and relative clauses is common, as in examples (10),
(13), (14), (15), (18). In examples (9) and (17) we have the stylish addition
of ἄν to ὅπως, a feature often seen in Ptolemaic officialese. In (11) and
(16) there is the highly idiomatic Greek use of ἄν with a verb in the indica-
tive to convey an ‘unfulfilled’ condition (‘I would have sent you …’). In
example (12) we see the rare use of ὡς ἄν + optative in a simile.
(9) Gen 12:13 ὅπως ἂν εὖ μοι γένηται
BGU 8.1738.32 (72 BC) ὁπηνίκʼ ἂν εὖ συνθῶμεν
(10) Gen 21:6 ὃς γὰρ ἂν ἀκούσῃ
PCairZen 1.33.7 ἐπόρισεγ γὰρ ἂν αὐτοῖς
(11) Gen 31:27 ἐξαπέστειλα ἄν σε μετ᾽ εὐφροσύνης
PMich 1.29.3 ἀπέστ[ειλα] | ἄν ∖σοι∕ αὐτήν
(12) Gen 33:10 εἶδον τὸ πρόσωπόν σου, ὡς ἄν τις ἴδοι πρόσωπον θεοῦ
PCairZen 1.93.18 κέχρηται ἡμῖν ὡς ἂν εἴ τις ἐχθρῶι χρήσαιτο.
(13) Exod 32:34 ᾗ δ᾽ ἂν ἡμέρᾳ ἐπισκέπτωμαι
PHib 1.29.10 ἧι δʼ ἂν ἡμ[έρ]αι
(14) Lev 13:51 κατὰ πάντα ὅσα ἂν ποιηθῇ
PCairZen 5.815.4 καὶ ὅσα ἂν εὑρεθῆι
(15) Lev 16:17 ἕως ἂν ἐξέλθῃ
PPetr 2.3a.1 διὸ ἐπέχω ἕως ἂν ἐξέλθηι
(16) Num 22:29 εἰ εἶχον μάχαιραν … ἤδη ἂν ἐξεκέντησά σε
PCairZen 4.599.8 καὶ εἰ μὴ … ἐπεβοήθει … πάλαι ἂν ἠσχημόνουν
(17) Num 16:40(17:5) ὅπως ἂν μὴ προσέλθῃ μηθείς
PLond 7.2033.6 ὅπως ἂν μὴ κατέχηται ὁ Καλλικῶν
22
In the current Göttingen text 394 of ἄν (as counted by Accordance©) + 12 of ἐάν =
particle ἄν. [Cf. discussion of ἄν now in Lee, Greek of the Pentateuch, 137–9.]
(18) Deut 12:26 ἃ ἄν σοι γένηται

PCairZen 3.358.3 ὅσα ἄν σοι δόξῃ
These uses of ἄν in the Pentateuch are interesting enough as instances of

Greek idiom; but there is more. This particle exhibits word-order behaviour
very similar to that of τις, even though it is not an enclitic. Like τις it
gravitates to the second slot in its clause, but it has an even stronger ten-
dency to do so than τις; it also precedes some words that τις does not, and
τις itself. The rules can be seen played out in our selection: ἄν follows the
particles γάρ (10) and δέ (13), but otherwise comes immediately after the
headword of its clause, preceding the enclitic pronouns σε (11), σοι (18),
the negative μή (17), and τις (12). These word-order rules can be traced
back as far as Homer.23 This of course does not mean that the LXX trans-
lators learnt them from Homer; it means that the patterns were estab-
lished in early Greek and continued to operate all the way to the time of
the translators.
6. CONCLUSIONS
The aim of identifying Greek idiom in the LXX-Pentateuch can be read-

ily achieved. As we have seen, there are undoubted instances according
to the criteria we have established. In the case of ἄν, every use falls into
this category. In the case of τις, we have found clear instances in which
Hebrew interference cannot be involved (60 in all); in the rest the pos-
sibility of interference remains open according to our criteria, but other
factors apply that can tip the balance against interference. The approach
through three defined categories of rendering (§3) provides a method appli-
cable to all cases of potential interference, in effect to all renderings in
the LXX.
There are three further conclusions that I think are warranted. First, the
evidence of idiomatic Greek woven into the fabric of the Greek Pentateuch
demonstrates the translators’ complete competence in Greek. If one also
takes into account other evidence beyond τις and ἄν – though these are
significant enough – it is not too much to assert that the Pentateuch trans-
lators had native-speaker competence.24 It is hard to see what other expla-
nation there is for their intimate knowledge of the usage of τις and ἄν,
23
See Ruijgh, ‘La place des enclitiques,’ 213–33.
24
Further evidence supporting this conclusion is offered in Lee, Greek of the Pentateuch,
esp. ch. 4 (123–72).
including the word-order patterns. We can say further that this observation
applies to all the translators of the Pentateuch, however many there were.
I believe that there were five, but the point is not affected even if this is
not true: the spread of τις and ἄν through the Pentateuch shows that they
were all equally competent in Greek. This is not the same as competence
as translators. Nor do these two features enable us to characterise the
techniques of individual translators: as far as I can see, they all use τις and
ἄν in the same way.
A second conclusion is that the presence of Greek idiom as in τις and
ἄν supports the contention that the Greek translation of the Pentateuch
should be viewed as in essence Greek with Hebrew interference, rather
than as Hebraic Greek into which idiomatic Greek occasionally intrudes,
for no evident reason. This is to some extent just a matter of viewpoint,
but there is a genuine issue underlying it. What we expect to see will have
an impact on what we find, when we try to analyse the Greek of the Penta-
teuch and the translators’ methods. Soisalon-Soininen’s approach to ‘nomi-
nal relative clauses’ might be seen as an illustration (see §1).
A third conclusion relates to Koine Greek as a whole. Our usual approach
to the Greek of the LXX, founded on a long-standing assumption that there
is something anomalous about it and every feature needs to be proved to
be acceptable Greek, has been to pursue a quest for parallels outside the
LXX to provide the proof. What we have found in τις and ἄν in the Pen-
tateuch, that is, normal usage of characteristically Greek idiom, suggests
that the Pentateuch itself has value as evidence for Koine Greek (with of
course the usual cautions and provisos). In regard to the uses of τις and ἄν
that are already well-attested, as in those presented above, the Pentateuch
has little new to offer. But things are different in the case of another use
of ἄν, the ‘iterative’ use with indicative in subordinate clauses, as seen in
this example:
(19) Exod 33:8 ἡνίκα δ᾽ ἂν εἰσεπορεύετο Μωυσῆς εἰς τὴν σκηνήν. …
And whenever Moyses went into the tent. …
… ‫והיה כצאת משׁה אל־האהל‬
(Lit.) And it happened at the going out of Moses to the tent. …
Though this new Koine Greek construction is well attested later, there
are no examples in third-century BC documents and there is a general
lack of evidence for it in early Koine Greek.25 The Pentateuch, with some
15 instances, actually provides the (so far) missing proof that it had begun
25
See Mayser, Grammatik, II.1, 295; II.3, 79–80, 94; Muraoka, Syntax, 773.
in the third century BC. As can be seen from the example, it has nothing to
do with the Hebrew original, nor can there be any reason to think the trans-
lators were using something that was not current Greek.
To conclude. The study of Hebraisms has its place, but rather than go
back to the question of Hebraisms I would like to see us go forward to the
question of Greek idiom in the LXX. Our evidence has shown that Greek
idiom is undoubtedly present in the LXX-Pentateuch, but the pursuit of
such evidence has been slow up till now. We could now turn our attention
to the many ways in which the translators followed Greek idiom, rather
than the ways in which they departed from it.
BIBLIOGRAPHY
ABBOTT-SMITH, G., A Manual Greek Lexicon of the New Testament (Edinburgh,

1922; 2nd ed. 1923; 3rd ed. 1937).
ADAMS, J. N., Bilingualism and the Latin Language (Cambridge, 2003).
ADAMS, S. A., ‘Atticism, Classicism, and Luke–Acts: Discussions with Albert
Wifstrand and Loveday Alexander,’ in Porter and Pitts (eds.), The Language
of the New Testament (2013), 91–111.
ADRADOS, F. R., ‘Sobre los orígenes del vocabulario ático. II,’ Emerita 25 (1957),
81–121.
AEJMELAEUS, A., ‘Translation Technique and the Intention of the Translator,’ in
C. E. Cox (ed.), VII Congress of the International Organization for Septua-
gint and Cognate Studies, Leuven 1989 (Atlanta, 1989), 23–36.
AITKEN, J. K., ‘Phonological Phenomena in Greek Papyri and Inscriptions and
their Significance for the Septuagint,’ in J. Corley and V. Skemp (eds.), Studies
in the Greek Bible: Essays in Honor of Francis T. Gignac, SJ (CBQMS 44;
Washington DC, 2008), 256–77.
— ‘The Literary Attainment of the Translator of Greek Sirach,’ in J.-S. Rey and
J. Joosten (eds.), Texts and Versions of the Book of Ben Sira: Transmission
and Interpretation (Leiden, 2011), 95–126.
— ‘The Characterization of Speech in the Septuagint Pentateuch,’ in D. J. A. Clines
and J. C. Exum (eds.), The Reception of the Hebrew Bible in the Septuagint
and the New Testament: Essays in Memory of Aileen Guilding (Sheffield,
2013), 9–31.
— No Stone Unturned: Greek Inscriptions and Septuagint Vocabulary (Winona
Lake, 2014).
— ‘The Origins of καί γε,’ in Aitken and Evans (eds.), Biblical Greek in Context
(2015), 21–40.
— (ed.), The T&T Clark Companion to the Septuagint (London, 2015).
AITKEN, J. K., and T. V. EVANS (eds.), Biblical Greek in Context: Essays in Honour
of John A. L. Lee (Leuven, 2015).
ALAND, K., Vollständige Konkordanz zum griechischen Neuen Testament, II
(Berlin, 1978).
ALAND, K., and B. ALAND, The Text of the New Testament (trans. Erroll F. Rhodes;
Grand Rapids, 1987).
ALLEN, W. C., S. Matthew (ICC; Edinburgh, 1912).
ANDERSEN, F. I., ‘Advances in the Lexicography of Biblical Languages: A Review
Essay,’ ANES 45 (2008), 235–59.
ANDERSON, G., The Second Sophistic: A Cultural Phenomenon in the Roman Empire
(London, 1993).
ANZ, H., Subsidia ad cognoscendum graecorum sermonem vulgarem e Pentateuchi
versione alexandrina repetita (Diss. Phil. Halenses, 12.2; Halle, 1894).
ARGYLE, S., ‘A New Greek Grammarian,’ CQ 39 (1989), 524–35.
ARGYRIADES, G. P., Νεοελληνική Γλώσσα· Ιστορικές και Γλωσσολογικές Διαστά-

σεις (Θεσσαλονίκη, 1990).
ARNOLD, B. T., ‘Luke’s Characterizing Use of the Old Testament in the Book of
Acts,’ in B. Witherington III (ed.), History, Literature, and Society in the
Book of Acts (Cambridge, 1996), 300–23.
ARNOLD, W. R., ‘The Meaning of ‫בתרון‬,’ AJSL 28 (1912), 274–83.
ARNOTT, W. G., ‘The Confrontation of Sostratos and Gorgias,’ Phoenix 18 (1964),
110–23.
— ‘A Note on the Antiatticist (98.17 Bekker),’ Hermes 117 (1989), 374–6.
ARRIBAS ARRANZ, F., Paleografía documental hispánica, I. Láminas; II. Trans-
cripciones (Vallodolid, 1965).
AUWERS, J.-M., Concordance du Siracide (Grec II et Sacra Parallela) (CahRB, 58;
Paris, 2005).
BABINIOTES, G. D., Λεξικό της Νέας Ελληνικής Γλώσσας (Athens, 2002).
BAER, D. A., When We All Go Home: Translation and Theology in LXX Isaiah 56–
66 (Sheffield, 2001).
— ‘What Happens in the End? Evidence for an Early Greek Recension in LXX
Isaiah 66,’ in van der Kooij and van der Meer (eds.), Old Greek of Isaiah
(2010), 1–31.
BAGNALL, R. S. (ed.), The Oxford Handbook of Papyrology (Oxford, 2009).
BAILLY, A., Dictionnaire Grec-Français (Paris, 1894; rev. eds. 1950, 1963).
BARDY, G., ‘Basilius von Caesarea,’ RAC 1 (1950), 1263.
BARKER, N., Aldus Manutius and the Development of Greek Script and Type in
the Fifteenth Century (Sandy Hook, Conn., 1985).
BARR, J., Comparative Philology and the Text of the Old Testament (Oxford,
1968).
— The Typology of Literalism in Ancient Biblical Translations (Göttingen,
1979).
BARRETT, C. K., The Gospel According to St John (London, 1955).
— The Acts of the Apostles (ICC; Edinburgh 1994, 1998).
BARTHÉLEMY, D., Les devanciers d’Aquila (VTSuppl. 10; Leiden, 1963).
BATAILLON, M., Érasme et l’Espagne (1st ed., Paris, 1937. Nouvelle édition en
trois volumes, ed. D. Devoto and C. Amiel. Genève, 1991).
BAUER, H., and P. LEANDER, Historische Grammatik der hebräischen Sprache
des Alten Testamentes (Halle, 1922).
BEENTJES, P. C., The Book of Ben Sira in Hebrew: A Text Edition of All Extant
Hebrew Manuscripts and a Synopsis of All Parallel Hebrew Ben Sira Texts
(Leiden, 1997).
BEKKER, I., Anecdota Graeca (3 vols.; Berlin, 1814–1821).
BENTEIN, K., ‘Perfect Periphrases in Post-Classical and Early Byzantine Greek:
An Ecological-Evolutionary Account,’ Journal of Greek Linguistics 12
(2012), 205–75.
— ‘Particle-usage in Documentary Papyri (I–IV AD): An Integrated, Sociolin-
guistically Informed Approach,’ GRBS 55 (2015), 721–53.
— ‘Dimensions of Social Meaning in Post-Classical Greek,’ JGL 19 (2019),
119–67.
BENTLEY, J. H., Humanists and Holy Writ: New Testament Scholarship in the
Renaissance (Princeton, 1983).
BIBLIOGRAPHY 457
BÉRARD, F., et al., Guide de l’épigraphiste: Bibliographie choisie des épigraphies

antiques et médiévales (3rd ed.; Paris, 2000).
BETHE, E. (ed.), Pollucis Onomasticon (2 vols.; Lexicographi Graeci 9:1–2;
Leipzig, 1900, 1931).
— ‘Pollux,’ in Pauly-Wissowa, RE 10:773–9.
BETZ, H. D., Galatians (Hermeneia; Philadelphia, 1979).
BLACK, M., An Aramaic Approach to the Gospels and Acts (3rd ed.; Oxford, 1967).
BLOMQVIST, J., Greek Particles in Hellenistic Prose (Lund, 1969).
BOISACQ, É., Dictionnaire étymologique de la langue grecque (Paris, 1916).
BOLING, R. G., Judges (New York, 1975).
BOLINGER, D., Aspects of Language (2nd ed.; New York, 1975).
BOWIE, E. L., ‘Greeks and Their Past in the Second Sophistic,’ Past and Pre-
sent 46 (1970), 3–41.
BOYD-TAYLOR, C., ‘A Place in the Sun: The Interpretative Significance of LXX-
Psalm 18:5c,’ BIOSCS 31 (1998), 71–105.
— ‘The Evidentiary Value of Septuagintal Usage for Greek Lexicography: Alice’s
Reply to Humpty Dumpty,’ BIOSCS 34 (2001), 47–80.
— ‘Lexicography and Interlanguage – Gaining our Bearings,’ BIOSCS 37 (2004),
55–72.
— Reading Between the Lines: The Interlinear Paradigm for Septuagint Studies
(BTS 8; Leuven, 2011).
— ‘The Classification of Literalism in Ancient Hebrew–Greek Translation,’ in
LXX.H, 3:139–60.
BOYD-TAYLOR, C. (ed.), A Question of Methodology: Albert Pietersma, Collected
Essays on the Septuagint (BTS 14; Leuven, 2013).
BRANDWOOD, L., A Word Index to Plato (Leeds, 1976).
BRATCHER, R. G., and E. A. NIDA, A Translator’s Handbook on the Gospel of Mark
(Leiden, 1961).
BRAUN, A., ‘Nota sui verbi greci del “volere”,’ AIV 98 (1938/9), 337–55.
BREITENSTEIN, U., Beobachtungen zu Sprache, Stil und Gedankengut des Vierten
Makkabäerbuchs (Basel, 1978).
BROCK, S. P., ‘The Phenomenon of the Septuagint,’ OtSt 17 (1972), 11–36.
BROWN, R. E., The Gospel According to John: Introduction, Translation and Notes
(AB; 2 vols.; Garden City, 1966–1970).
BROWNE, G. M., Sortes Astrampsychi, Volumen I: Ecdosis Prior (Leipzig, 1983).
— ‘The Origin and Date of the Sortes Astrampsychi,’ Illinois Classical Studies 1
(1976), 53–8.
BROWNING, R., Medieval and Modern Greek (2nd ed.; Cambridge, 1983).
BRUCE, F. F., The Epistle of Paul to the Galatians (NIGTC; Exeter, 1982).
BÜCHNER, D. L., Review of LEH, Lexicon, BIOSCS 37 (2004), 139–47.
BUCK, C. D., ‘Greek ἄμϕοδον, Oscan amvíanud, and the Oscan eítuns-inscriptions,’
CP 17 (1922), 111–8.
— A Dictionary of Selected Synonyms in the Principal Indo-European Languages
(Chicago, 1949).
BUCK, C. D., and W. Petersen, A Reverse Index of Greek Nouns and Adjectives
(Hildesheim, 1970).
BURGUIÈRE, P., Histoire de l’infinitif en grec (Paris, 1960).
BURNEY, C. F., The Aramaic Origin of the Fourth Gospel (Oxford, 1922).
BURRIDGE, R. A., ‘The Gospels and Acts,’ in Porter (ed.), Handbook (1997), 507–
32.
BUTLER, Dom C., The Lausiac History of Palladius (2 vols.; Cambridge, 1898,
1904).
CADBURY, H. J., The Style and Literary Method of Luke (Cambridge, Mass., 1920).
CADELL, H., ‘Vocabulaire de l’irrigation: la Septante et les papyrus’ in B. Menu
(ed.), Les problèmes institutionnels de l’eau en Égypte ancienne et dans
l’antiquité méditerranéenne (Cairo, 1994), 103–17.
— ‘Vocabulaire de la législation ptolemaïque: problème du sens de dikaiôma
dans le Pentateuque,’ in Dorival and Munnich (eds.), Κατὰ τοὺς Οʹ (1995),
207–21.
CAIRD, G. B., ‘Towards a Lexicon of the Septuagint. I,’ JTS n.s. 19 (1968), 453–
75.
— ‘Towards a Lexicon of the Septuagint. II,’ JTS n.s. 20 (1969), 21–40.
CARSON, D. A., Exegetical Fallacies (2nd ed.; Grand Rapids, MI, 1996).
CASE, S. J., ‘The Use of ἄμϕοδον in the Septuagint and the New Testament,’ in
Buck, ‘Greek ἄμϕοδον,’ 117 n.
CAVALLO, G., and H. MAEHLER, Greek Bookhands of the Early Byzantine Period,
A.D. 300–800 (ICS Bulletin Suppl. 47; London, 1987).
CHADWICK, J., ‘Semantic History and Greek Lexicography,’ in F. Létoublon (ed.),
La langue et les textes en grec ancien: Actes du Colloque Pierre Chantraine
(Grenoble – 5-8 Septembre 1989) (Amsterdam, 1992), 281–8.
— ‘The Case for Replacing Liddell and Scott,’ BICS 39 (1994), 1–11.
— Lexicographica Graeca: Contributions to the Lexicography of Ancient Greek
(Oxford, 1996).
CHANTRAINE, P., La formation des noms en grec ancien (Paris, 1933).
— ‘Les mots designant la gauche en grec ancien,’ in H. Kronasser (ed.), Μνήμης
χάριν: Gedenkschrift Paul Kretschmer (2 vols.; Vienna, 1956–1957), 1:61–
9.
— Grammaire homérique (2 vols.; Paris, 1958, 1963).
CHARLES, R. H., The Revelation of St. John (ICC; Edinburgh, 1920).
COBET, C. G., Variae lectiones (Leiden, 1873).
COHN, L., ‘Griechische Lexikographie,’ in K. Brugmann, Griechische Grammatik,
4th ed., rev. Albert Thumb (München, 1913), 679–730.
— ‘Ammonios,’ in Pauly-Wissowa, RE 1:1866–7.
COLLARD, C., M. J. CROPP, and K. H. LEE / J. GIBERT, Euripides: Selected Frag-
mentary Plays, with Introductions, Translations and Commentaries (2 vols.;
Warminster, 1995, 2004).
CONNOLLY, A. L., ‘Atticism in Non-literary Papyri of the First Seven Centuries AD:
A Study in Several Features of Orthography and Syntax’ (BA diss., Univer-
sity of Sydney, 1983).
CONTOPOULOS, N., Νέον Λεξικόν Ελληνοαγγλικόν (3rd ed.; London, 1887).
COSTACHE, D., and P. KARIATLIS (eds.), Cappadocian Legacy: A Critical Appraisal
(Sydney, 2013).
COSTAS, P. S., An Outline of the History of the Greek Language: With Particular
Emphasis on the Koine and the Subsequent Stages (Chicago, 1936).
[CRASTONUS, Joannes,] Dictionarium graecum copiosissimum secu[n]dum ordi-
nem alphabeti cum interpretatione latina (Venetiis, 1497).
BIBLIOGRAPHY 459
CRIBIORE, R., Gymnastics of the Mind: Greek Education in Hellenistic and Roman
Egypt (Princeton, 2001).
CROSS, F. M., and S. TALMON, Qumran and the History of the Biblical Text (Cam-
bridge, MA, 1975).
CUNNINGHAM, I. C. (ed.), Συναγωγὴ Λέξεων Χρησίμων (Berlin, 2003).
CURTIUS, G., Grundzüge der griechischen Etymologie (1st ed., Leipzig, 1858–1862;
5th ed., Leipzig, 1879).
DAIN, A. (ed.), Le ‘Philétaeros’ attribué à Hérodien, (Paris, 1954).
DANIEL, S., Recherches sur le vocabulaire du culte dans le Septante (Paris, 1966).
DANKER, F. W., Multipurpose Tools for Bible Study (Minneapolis, 1993).
— The Concise Greek-English Lexicon of the New Testament (Chicago, 2009).
— Review of Muraoka, Lexicon (2009), RBL 04/2010.
DARDANO, M., and P. TRIFONE, La lingua italiana (Bologna, 1985).
DARIS, S., ‘I Quartieri di Ossirinco: Materiali e Note,’ ZPE 132 (2000), 211–
21.
DARLOW, T. H., and H. F. MOULE, Historical Catalogue of the Printed Editions
of Holy Scripture in the Library of the British and Foreign Bible Society
(2 vols.; London, 1911).
DAVIES, W. D., and D. DAUBE (eds.), The Background of the New Testament and
its Eschatology: Studies in Honour of C. H. Dodd (Cambridge, 1956).
DE BORRIES, J. (ed.), Phrynichi Sophistae Praeparatio Sophistica (Leipzig, 1911).
DEBRUNNER, A., and A. SCHERER, Geschichte der griechischen Sprache (2 vols.;
2nd ed. Berlin, 1969).
DECKER, R. J., ‘Markan Idiolect in the Study of the Greek of the New Testament,’
in Porter and Pitts (eds.), Language of the New Testament (2013), 43–66.
DEFERRARI, R. J. (trans.), Basil: Letters 1–58 (LCL 190; Cambridge, MA, 1926).
— Basil: Letters 249–368 (LCL 270; Cambridge, MA, 1934).
DEFERRARI, R. J., and M. R. P. MCGUIRE (trans.), Basil: Address to Young Men
on Reading Greek Literature (LCL 270; Cambridge, MA, 1934).
DEISSMANN, A., Bible Studies: Contributions Chiefly from Papyri and Inscriptions
to the History of the Language, the Literature, and the Religion of Hellenistic
Judaism and Primitive Christianity (trans. A. Grieve; 2nd ed.; Edinburgh,
1909).
— Light from the Ancient East (trans. L. R. M. Strachan; New York, 1927).
DE JONGE, H. J., ‘De nieuwe Nestle: N26,’ NedTT 34 (1980), 307–22.
— Opera Omnia Desiderii Erasmi Roterodami, IX.2, Apologia respondens ad
ea quae Iacobus Lopis Stunica taxaverat in prima duntaxat novi testamenti
aeditione (Amsterdam, 1983).
DENNISTON, J. D., The Greek Particles (2nd ed.; Oxford, 1954).
DE SOUSA, R. F., Eschatology and Messianism in LXX Isaiah 1–12 (New York,
2010).
DE TROYER, K., T. M. LAW, and M. LILJESTRÖM (eds.), In the Footsteps of Sher-
lock Holmes: Studies in the Biblical Text in Honour of Anneli Aejmelaeus
(CBET 72; Leuven, 2014).
DHONT, M., ‘Multicausality in Septuagint Studies’ [forthcoming in JSCS].
DICKEY, E., Ancient Greek Scholarship: A Guide to Finding, Reading, and Under-
standing Scholia, Commentaries, Lexica, and Grammatical Treatises, from
Their Beginnings to the Byzantine Period (Oxford, 2007).
DIETERICH, K., Untersuchungen zur Geschichte der griechischen Sprache: von der
hellenistischen Zeit bis zum 10. Jahrhundert n. Chr (Leipzig, 1898. Repr.
Hildesheim, 1970).
DINES, J. M., The Septuagint (London, 2004).
DOGNIEZ, C., and M. HARL, La Bible d’Alexandrie: Le Deutéronome (Paris, 1992).
DORIVAL, G., La Bible d’Alexandrie: Les Nombres (Paris, 1994).
DRIVER, G. R., ‘Linguistic and Textual Problems: Jeremiah,’ JQR 28 (1937),
97–129.
DU BOUCHET, J., ‘Quartiers, Îlots et Rues: Remarques sur ἄμϕοδον,’ RPhil 78
(2004), 43–55.
EBELING, H., Griechisch-deutsches Wörterbuch zum Neuen Testamente (Hannover,
1913; 3rd ed. 1929).
EDMONDS, J. M., The Fragments of Attic Comedy: After Meineke, Bergk, and Kock
(3 vols. in 4; Leiden, 1957–1961).
EISSFELDT, O., ‘Ein gescheiterter Versuch der Wiedervereinigung Israels,’ La Nou-
velle Clio 3 (1951), 110–27.
ELLIOTT, J. K., ‘The Text and Language of the Endings of Mark’s Gospel,’ TZ 27
(1971), 255–62.
— ‘The United Bible Societies Greek New Testament: An Evaluation,’ NovT 15
(1973), 278–300.
— Essays and Studies in New Testament Textual Criticism (Córdoba, 1992).
ELLIS, J., and J. N. URE, ‘Language Varieties: Register,’ in A. R. Meetham (ed.),
Encyclopedia of Linguistics, Information and Control (Oxford, 1969).
EPP, E. J., and G. D. FEE, Studies in the Theory and Method of New Testament
Textual Criticism (Grand Rapids, 1993).
ERASMUS, Desiderius, De recta latini graecisque sermonis pronuntiatione (Basle,
1528. Facsimile ed., Menston, 1971).
ERBSE, H., Untersuchungen zu den Attizistischen Lexika: Aelii Dionysii et Pau-
saniae atticistarum fragmenta (AAWB 1949.2:95–221; Berlin, 1950).
EVANS, T. V., Verbal Syntax in the Greek Pentateuch: Natural Greek Usage and
Hebrew Interference (Oxford, 2001).
— ‘The Court Function of the Interpreter in Genesis 42:23 and Early Greek
Papyri’ in J. K. Aitken, J. Dines, S. Pearce, and T. Rajak (eds.), Jewish Per-
spectives on Hellenistic Kingship (Berkeley, 2007), 238–52.
— ‘Numbers,’ in Aitken (ed.), T&T Clark Companion (2015), 58–67.
EVANS, T. V., and D. D. OBBINK (eds.), The Language of the Papyri (Oxford,
2010).
EXLER, F. X. J., The Form of the Ancient Greek Letter (Washington, 1923).
FABRICIUS, C., ‘Der sprachliche Klassizismus der griechischen Kirchenväter. Ein
philologisches und geistesgeschichtliches Problem,’ JAC 10 (1967), 187–
99.
FARMER, W. R., The Last Twelve Verses of Mark (Cambridge, 1974).
FEE, G. D., Papyrus Bodmer II (P66): Its Textual Relationships and Scribal
Characteristics (Studies and Documents 34; Salt Lake City, 1968).
— ‘Rigorous or Reasoned Eclecticism—Which?,’ in J. K. Elliott (ed.), Studies
in New Testament Language and Text: Essays in Honour of G. D. Kilpatrick
(NT Suppl. 44; Leiden, 1976), 174–97.
— The First Epistle to the Corinthians (NICNT; Grand Rapids, 1987).
BIBLIOGRAPHY 461
FERNÁNDEZ-ARMESTO, F., ‘Cardinal Cisneros as a Patron of Printing,’ in D. W. Lomax

and D. Mackenzie (eds.), God and Man in Medieval Spain: Essays in Honour
of J. R. L. Highfield (Warminster, 1989), 149–68.
FESTUGIÈRE, A.-J., Historia Monachorum in Aegypto (Subsidia Hagiographica 34;
Brussels, 1961).
FEYERABEND, K., Langenscheidt’s Pocket Greek Dictionary: Classical Greek–
English, (Berlin, 1910).
FISCHER, E. (ed.), Die Ekloge des Phrynichos (SGLG 1; Berlin, 1974).
FITZMYER, J. A., ‘The Study of the Aramaic Background of the New Testament’
(1975), in A Wandering Aramean: Collected Aramaic Essays (SBLMS 25;
Missoula, MO, 1979), 1–27.
— The Gospel According to Luke I-IX (AB; New York, 1981).
FLEISCHMAN, S., The Future in Thought and Language: Diachronic Evidence
from Romance (Cambridge, 1982).
FOX, W., ‘βούλεσθαι und (ἐ)θέλειν,’ Berliner Philologische Wochenschrift 37
(1917), 597–606, 633–9.
FRAME, J. E., Thessalonians (ICC; Edinburgh, 1912).
FRÄNKEL, H., Einleitung zur kritischen Ausgabe der Argonautika des Apollonios
(Göttingen, 1964).
FRÖSÉN, J., Prolegomena to a Study of the Greek Language in the First Centu-
ries A.D.: The Problem of Koine and Atticism (Helsinki, 1974).
GALLAY, P. (ed.), Saint-Grégoire de Nazianze: Lettres (2 vols.; Paris, 1964).
GARCÍA SANTOS, A. A., Diccionario del Griego Bíblico: Setenta y Nuevo Testa-
mento (Estella, Navarra, 2011).
GARLAN, Y., ‘Cités, armées et stratégie à l’époque hellénistique d’après l’oeuvre
de Philon de Byzance,’ Historia 22 (1973), 16–33.
— Recherches de poliorcétique grecque (Athens, 1974).
GEANAKOPLOS, D. J., Byzantium and the Renaissance: Greek Scholars in Venice
(Harvard, 1962; repr. Hamden, 1973).
GEHMAN, H. S., ‘Adventures in Septuagint Lexicography,’ Textus 5 (1966), 125–32.
GERLEMAN, G., Studies in the Septuagint I: The Book of Job (Lund, 1946).
GIANNAKOPOULOS, I., Ἡ Παλαιὰ Διαθήκη κατὰ τοὺς Ο´ (26 vols.; 4th ed.; Thes-
saloniki, 1955–68).
GIGNAC, F. T., A Grammar of the Greek Papyri of the Roman and Byzantine
Periods (2 vols.; Milan, 1976, 1981).
GINGRICH, F. W., Shorter Lexicon of the Greek New Testament (Chicago, 1965;
2nd ed. rev. F. W. Danker, 1983).
GLARE, P., ‘Starting from the Wrong End,’ in G. Cigman and D. Howlett (eds.),
Birthday Celebration for Naky Doniach (Yarnton, Oxford, 1991), 35–41
GOETZ, G., and GUNDERMANN, G., Glossae Latinograecae et Graecolatinae, in
G. Loewe and G. Goetz (eds.), Corpus Glossariorum Latinorum (vol. 2;
Leipzig, 1888).
GOMME, A. W., and F. H. SANDBACH, Menander: A Commentary (Oxford, 1973).
GOODSPEED, E. J., Index Patristicus (Leipzig, 1907).
GRAY, G. B., Numbers (ICC; Edinburgh, 1912).
GREENLEE, J. H., A New Testament Greek Morpheme Lexicon (Grand Rapids, MI,
1983).
GREGORY, C. R., Textkritik des Neuen Testamentes (3 vols.; Leipzig, 1900–1909).
GREGORY, M., ‘Aspects of Varieties Differentiation,’ JLing 3 (1967), 177–98.

GREGORY M., and S. CARROLL, Language and Situation: Language Varieties and
their Social Contexts (London, 1978).
GUY, J.-C., Recherches sur la tradition grecque des Apophthegmata Patrum (Sub-
sidia Hagiographica 36; Brussels, 1962; repr. with suppl. 1984).
— Les Apophtegmes des Pères du Désert: Série alphabétique: Traduction fran-
çaise (Paris, 1967–1968).
— Les Apophtegmes des Pères: Collection systématique (3 vols.; Paris, 1993–
2005).
HAENCHEN, E., The Acts of the Apostles (ET; Oxford, 1971).
HAHN, E. A., Naming-Constructions in Some Indo-European Languages (APA
Monograph 27; Cleveland, 1969).
HALL, B., ‘The Trilingual College of San Ildefonso and the Making of the Com-
plutensian Polyglot Bible,’ in G. J. Cuming (ed.), Studies in Church History
(Leiden, 1969), 5:114–46.
HANSEN, D. U. (ed.), Das attizistische Lexikon des Moeris: Quellenkritische
Untersuchung und Edition (SGLG 9; Berlin, 1998).
HARL, M., La Bible d’Alexandrie: La Genèse (Paris, 1986).
HARL, M., G. DORIVAL, and O. MUNNICH, La Bible grecque des Septante: du
judaïsme hellénistique au christianisme ancien (Paris, 1988).
HARRISON, N. V., Saint Basil the Great: On the Human Condition (Crestwood, 2005).
HART, J. H. A., Ecclesiasticus: The Greek Text of Codex 248, edited with a Tex-
tual Commentary and Prolegomena (Cambridge, 1909).
HATZIDAKIS, G. N., Einleitung in die neugriechische Grammatik (Leipzig, 1892).
— Μεσαιωνικὰ καὶ νέα Ἑλληνικά (2 vols.; Athens, 1905, 1907. Repr. Amsterdam,
1989, 1990).
HAUSPIE, K., ‘The LXX Quotations in the LSJ Supplements of 1968 and 1996,’
in Taylor et al. (eds.), Language and Lexicography (2004), 108–25.
HELBING, R., Die Kassusyntax der Verba bei den Septuaginta (Göttingen, 1928).
HEMER, C. J., ‘Towards a New Moulton and Milligan,’ NovT 24 (1982), 97–123.
HENDERSON, J., ‘A1–Zythum: Domimina Nustio Illumea, or Out with the OLD
(1931–82),’ in C. Stray (ed.), Classical Dictionaries: Past, Present and Future
(London, 2010), 139–76.
HENNIG, D., ‘Staatliche Ansprüche an privaten Immobilienbesitz in der klassi-
schen und hellenistischen Polis,’ Chiron 25 (1995), 235–82.
HERKENNE, H., De Veteris Latinae Ecclesiastici capitibus I–XLIII, una cum notis
ex eiusdem libri translationibus Aethiopica, Armeniaca, Copticis, Latina
altera, Syro-hexaplari depromptis (Leipzig, 1899).
HICKIE, W. J., Greek-English Lexicon to the New Testament (New York, 1893).
HOCK, H. H., and B. D. JOSEPH, Language History, Language Change, and Lan-
guage Relationship: An Introduction to Historical and Comparative Lin-
guistics (Berlin, 1996).
HOHLWEIN, N., L’Égypte Romaine: Recueil des termes techniques relatifs aux
institutions politiques et administratives de l’Égypte romaine, suivi d’un
choix de textes papyrologiques (Bruxelles, 1912).
HOLLANDER, H. W., and M. DE JONGE, The Testaments of the Twelve Patriarchs:
A Commentary (Leiden, 1985).
HOLLEAUX, M., Études d’épigraphie et d’histoire grecques, III (Paris, 1942).
BIBLIOGRAPHY 463
HOLTON, D. W., ‘The Formation of the Future in Modern Greek Literary Texts
up to the 17th Century,’ in N. M. Panayotakis (ed.), Αρχές της νεοελληνικής
λογοτεχνίας / Origini della litteratura neogreca (Venice, 1993), 1:118–
28.
HOLTON, D., P. MACKRIDGE, and I. PHILIPPAKI-WARBURTON, Greek: A Compre-
hensive Grammar (2nd ed., rev. V. Spyropoulos; London, 2012).
HOMZA, L. A., Religious Authority in the Spanish Renaissance (Baltimore, 2000).
HOPPER, P. J., and E. C. TRAUGOTT, Grammaticalization (2nd ed.; Cambridge,
2003).
HORROCKS, G. C., Greek: A History of the Language and its Speakers (1st ed.
1997; 2nd ed.; Chichester, 2014).
HORSLEY, G. H. R., and J. A. L. LEE, ‘A Preliminary Checklist of Abbreviations
of Greek Epigraphic Volumes,’ Epigraphica 56 (1994), 129–69.
— ‘A Lexicon of the New Testament with Documentary Parallels: Some Interim
Entries, 1,’ FilNT 10 (1997), 55–84.
HUDSON, R. A., Sociolinguistics (Cambridge, 1980).
HUGO, P., Review of Troxel, LXX-Isaiah (2008), JHebS 10 (2010) (electronic).
HUMBERT, J., Syntaxe grecque (3rd ed.; Paris, 1960).
HUNKIN, J. W., ‘Pleonastic ἄρχομαι in the New Testament,’ JTS 25 (1924), 390–
402.
IMHOLTZ, A. A., ‘“Liddell and Scott”: Precursors, Nineteenth-Century Editions,
and the American Contributions,’ in Stray (ed.), Oxford Classics (2007),
117–34.
JACQUES, X., List of New Testament Words Sharing Common Elements (Rome,
1969).
JAEGER, W., Early Christianity and Greek Paideia (Cambridge, MA, 1961).
JANNARIS, A. N., An Historical Greek Grammar: Chiefly of the Attic Dialect as
Written and Spoken from Classical Antiquity down to the Present Time
(London, 1897).
JANSE, M., ‘Aspects of Bilingualism in the History of the Greek Language,’ in
J. N. Adams, M. Janse, and S. Swain (eds.), Bilingualism in Ancient Society:
Language Contact and the Written Text (Oxford, 2002), 332–90.
JELLICOE, S., The Septuagint and Modern Study (Oxford, 1968).
JEREMIAS, J., Unknown Sayings of Jesus (London, 1964).
JOBES, K. H., and M. SILVA, Invitation to the Septuagint (Grand Rapids, 2000).
JOHANNESSOHN, M., Der Gebrauch der Präpositionen in der Septuaginta (MSU III.3;
Berlin, 1926).
JOHNSON, A. F., The First Century of Printing at Basle (London, 1926).
— The Italian Sixteenth Century (London, 1926).
JOLY, R., Le vocabulaire chrétien de l’amour est-il original? Φιλεῖν et Ἀγαπᾶν
dans le grec antique (Bruxelles, 1968).
JOSEPH, B. D., Morphology and Universals in Syntactic Change: Evidence from
Medieval and Modern Greek (New York, 1990).
JOSEPH, B. D., and P. A. PAPPAS, ‘On Some Recent Views Concerning the Devel-
opment of the Greek Future System,’ BMGS 26 (2002), 247–73.
JOUGUET, P., La vie municipale dans l’Égypte romaine (Paris, 1911; repr. 1968).
JOÜON, P., ‘Θέλειν au sens d’“être sur le point de” dans Jean 1, 43,’ RSR 29 (1939),
620–1.
— ‘Les verbes βούλομαι et θέλω dans le Nouveau Testament,’ RSR 30 (1940),

227–38.
— A Grammar of Biblical Hebrew (trans. and rev. T. Muraoka; 2 vols.; Rome,
2006).
JOYAL, M., I. MCDOUGALL, and J. C. YARDLEY, Greek and Roman Education: A
Sourcebook (London, 2009).
JUDGE, E. A., ‘Paul’s Boasting in Relation to Contemporary Professional Practice,’
in D. M. Scholer (ed.), Social Distinctives of the Christians in the First Cen-
tury: Pivotal Essays by E. A. Judge (Peabody, MA, 2008), 57–71.
KARRER M., and W. KRAUS, with M. MEISER (eds.), Die Septuaginta: Texte, Kon-
texte, Lebenswelten. [1.] Internationale Fachtagung veranstaltet von Septua-
ginta Deutsch (LXX.D), Wuppertal 20.–23. Juli 2006 (WUNT 219; Tübingen,
2008).
KATZ, P., Philo’s Bible (Cambridge, 1950).
— ‘Septuagintal Studies in the Mid-Century: Their Links with the Past and their
Present Tendencies,’ in Davies and Daube (eds.), Background of the New
Testament (1956), 176–208.
KAUHANEN, T., and H. VANONEN (eds.), The Legacy of Soisalon-Soininen: Towards
a Syntax of Septuagint Greek (DSI 13; Göttingen, 2020).
KAZAZIS, J. N., ‘Atticism,’ in A. F. Christidis (ed.), A History of Ancient Greek:
From the Beginnings to Late Antiquity (Cambridge, 2007), 1200–12.
KIEFFER, R., Au delà des recensions? L’évolution de la tradition textuelle dans
Jean VI, 52-71 (Lund, 1968).
KIESSLING, E., ‘Die Aposkeuai und die prozeßrechtliche Stellung der Ehefrauen
im ptol. Ägypten,’ APF 8 (1927), 240–9.
KILPATRICK, G. D., ‘Atticism and the Text of the New Testament,’ in J. Blinzler
et al. (eds.) Neutestamentliche Aufsätze: Festschrift für J. Schmid (Regens-
burg, 1963), 125–37.
— Review of Rydbeck, Fachprosa (1967), JBL 88 (1969), 354–5.
— ‘What John Tells Us about John,’ in Studies in John Presented to Dr. J. N. Seven-
ster (NT Suppl. 24; Leiden, 1970), 75–87.
— ‘Eclecticism and Atticism,’ ETL 53 (1977), 107–12.
—‘Some Thoughts on Modern Textual Criticism and the Synoptic Gospels,’
NovT 19 (1977), 275–92.
— ‘Atticism and the Future of ζῆν,’ NovT 25 (1983), 146–51.
— Review of Greenlee, Morpheme Lexicon (1983), NovT 27 (1985), 382–4.
KIM, L., ‘The Literary Heritage as Language: Atticism and the Second Sophistic,’
in E. J. Bakker (ed.), A Companion to the Ancient Greek Language (Chiches-
ter, 2010), 468–82.
KINZIG, W., ‘The Greek Christian Writers,’ in Porter (ed.), Handbook (1997), 633–
70.
KLAMP, D., ‘Das Testament des Taharpaesis,’ ZPE 2 (1968), 81–150.
KOLITSARAS, I. Th., Ἡ Παλαιὰ Διαθήκη κατὰ τοὺς Ἑβδομήκοντα (5 vols.; Athens,
1989).
KOOIJ, A. van der, ‘Isaiah in the Septuagint,’ in C. G. Broyles and C. A. Evans
(eds.), Writing and Reading the Scroll of Isaiah (Leiden 1997), 513–29.
— The Oracle of Tyre: The Septuagint of Isaiah XXIII as Version and Vision
(VTSuppl. 71; Leiden, 1998).
BIBLIOGRAPHY 465
— ‘The Septuagint of Isaiah and the Issue of Coherence: A Twofold Analysis

of LXX Isaiah 31:9b–32:8,’ in van der Kooij and van der Meer (eds.), Old
Greek of Isaiah (2010), 33–48.
KOOIJ, A. van der, and M. N. VAN DER MEER (eds.), The Old Greek of Isaiah:
Issues and Perspectives, Papers Read at the Conference on the Septuagint
of Isaiah, held in Leiden, 10–11 April 2008 (CBET 55; Leuven, 2010).
KOTSEVALOV, A., ‘Koine Syntax of Greek Colonies on the Black Sea,’ in G. E.
Mylonas and D. Raymond (eds.), Studies Presented to D. M. Robinson (2
vols.; Saint Louis, MO, 1951, 1953), 2:434–42.
LADAS, G. G., and A. CHATZIDIMOS, Προσθῆκες, διορθώσεις καὶ συμπληρώσεις
στὴν Ἑ λληνικὴ Βιβλιογραϕία τοῦ Émile Legrand γιὰ τοὺς αἰῶνες XV, XVI καὶ
XVII (Athens, 1976).
LANE, W. L., The Gospel According to Mark: The English Text with Introduction,
Exposition and Notes (London, 1974).
LAGRANGE, M.-J., Evangile selon Saint Marc (5th ed.; Paris, 1929).
LA ROI, E., ‘A Rephilologized Diachronic Analysis of “Post-Classical Greek”:
Pitfalls and Principles for Progress,’ JGL 20 (2020), 213–38.
— ‘The Insubordination of If- and That-Clauses from Archaic to Post-Classical
Greek: A Diachronic Constructional Typology,’ Symbolae Osloenses 95
(2021), 1–63.
LE BOULLUEC, A., and P. SANDEVOIR, La Bible d’Alexandrie: L’Exode (Paris, 1989).
LEE, J. A. L., ‘A Note on Septuagint Material in the Supplement to Liddell and
Scott,’ Glotta 47 (1969), 234–42.
— ‘Ἀποσκευή in the Septuagint,’ JTS n.s. 23 (1972), 430–7.
— A Lexical Study of the Septuagint Version of the Pentateuch (SCS 14; Chico,
Calif., 1983).
— ‘The Future of ζῆν in Late Greek,’ NovT 22 (1980), 289–98.
— ‘Equivocal and Stereotyped Renderings in the LXX,’ RB 87 (1980), 104–17.
— ‘Some Features of the Speech of Jesus in Mark’s Gospel,’ NovT 27 (1985),
1–26.
— ‘Συνίστημι: A Sample Lexical Entry,’ in Melbourne Symposium (1990), 1–
16.
— ‘A Non-Aramaism in Luke 6.7,’ NovT 33 (1991), 28–34.
— ‘The United Bible Societies’ Lexicon and its Analysis of Meanings,’ FilNT 5
(1992), 167–89.
— ‘Translations of the Old Testament I. Greek,’ in Porter (ed.), Handbook of
Classical Rhetoric (1997), 775–83.
— ‘Hebrews 5:14 and Ἕξις: A History of Misunderstanding,’ NovT 39 (1997),
151–76.
— A History of New Testament Lexicography (New York, 2003).
— ‘The Present State of Lexicography of Ancient Greek,’ in Taylor et al. (eds.),
Biblical Greek Language and Lexicography (2004), 66–74.
— ‘Ἐξαποστέλλω,’ in J. Joosten and P. J. Tomson (eds.), Voces Biblicae: Sep-
tuagint Greek and its Significance for the New Testament (Leuven, 2007),
99–113.
— ‘A Lexicographical Database for Greek: Can it be Far Off? The case of
amphodon,’ in Karrer et al. (ed.), Die Septuaginta: Texte, Kontexte, Lebens-
welten (2008), 214–20.
— ‘Auxiliary θέλω,’ in Evans and Obbink (eds.), Language of the Papyri (2010),
15–34.
— ‘Why Didn’t St Basil Write in New Testament Greek?,’ Phronema 25 (2010),
3–20.
— Review of Muraoka, Lexicon (2009), BIOSCS 43 (2010), 115–25.
— ‘Releasing LSJ from its Past,’ in C. Stray (ed.), Classical Dictionaries (London,
2010), 119–38.
— ‘The Atticist Grammarians,’ in Porter and Pitts (eds.), Language of the New
Testament (2013), 283–308.
— ‘Accuracy and Idiom: The Renderings of mittahat in the Septuagint Pentateuch,’
in De Troyer, Law, and Liljeström (eds.), In the Footsteps (2014), 79–99.
— The Greek of the Pentateuch: Grinfield Lectures on the Septuagint 2011–2012
(Oxford, 2018).
— ‘The Use of the Septuagint in the Liturgy and Lectionary of the Greek Orthodox
Church,’ in A. G. Salvesen and T. M. Law (eds.), The Oxford Handbook of
the Septuagint (Oxford, 2021), 573–87.
LEE, J. A. L., and G. H. R. HORSLEY, ‘A Lexicon of the New Testament with
Documentary Parallels: Some Interim Entries, 2,’ FilNT 11 (1998), 57–84.
LEGG, S. C. E., Nouum Testamentum Graece: Euangelium secundum Marcum
(Oxford, 1935).
LEGRAND, É., Bibliographie hellénique ou description raisonnée des ouvragres
publiés en grec par des Grecs au XVe et XVIe siècles (4 vols.; Paris, 1885–
1906. Repr. Paris, 1962).
LEMMELIJN, B., A Plague of Texts? A Text-Critical Study of the So-Called ‘Plagues
Narrative’ in Exodus 7:14–11:10 (Leiden, 2009).
LE MOIGNE, P., ‘Le livre d’Esaïe dans la Septante: ecdotique, stylistique, linguis-
tique, ou esquisse d’une poétique de la Septante’ (Diss. École Pratique des
Hautes Études, Paris, 2001).
LIETZMANN, H., Griechische Papyri (Bonn, 1910).
LLOYD-JONES, H., and P. PARSONS (eds.), Supplementum Hellenisticum (Berlin,
1983).
LOADER, J. A., ‘An Explanation of the Term prosêlutos,’ NovT 15 (1973), 270–7.
LOADER, W., The Septuagint, Sexuality, and the New Testament (Grand Rapids,
2004).
LOBECK, Chr. A. (ed.), Phrynichi Eclogae nominum et verborum atticorum: cum
notis P. J. Nunnesii, D. Hoeschelii, J. Scaligeri et Cornelii de Pauw, partim
integris partim contractis (Leipzig, 1820).
LOLOS, Y., ‘Greek Roads: A Commentary on the Ancient Terms,’ Glotta 79 (2003),
137–74.
LONGENECKER, R. N., Galatians (WBC; Dallas, 1990).
LOUW, J. P., Semantics of New Testament Greek (Philadelphia, 1982).
— ‘The Analysis of Meaning in Lexicography,’ FilNT 6 (1993), 139–48.
— (ed.), Lexicography and Translation (Cape Town, 1985).
LOUW, J. P., and E. A. NIDA et al., Greek-English Lexicon of the New Testament:
Based on Semantic Domains (2 vols.; 2nd ed., New York, 1989).
LOUW, T. A. W. van der, Transformations in the Septuagint: Towards an Inter-
action of Septuagint Studies and Translation Studies (CBET 47; Leuven,
2007).
BIBLIOGRAPHY 467
LOWRY, M. The World of Aldus Manutius: Business and Scholarship in Renais-

sance Venice (Oxford, 1979).
LYELL, J. P. R., Cardinal Ximenes: Statesman, Ecclesiastic, Soldier and Man of
Letters, with an Account of the Complutensian Polyglot Bible (London,
1917).
MAAS, P. M., Greek Metre (trans. Hugh Lloyd-Jones; Oxford, 1962).
MAHAFFY, J. P., The Flinders Petrie Papyri (3 vols.; Dublin, 1891–1905).
MAIDHOF, A., Zur Begriffsbestimmung der Koine: besonders auf Grund des Atti-
zisten Moiris (Würzburg, 1912).
MANDILARAS, B. G., The Verb in the Greek Non-Literary Papyri (Athens, 1973).
MARICHAL, R. (ed.), Quart Livre de Pantagruel (Geneva, 1947).
MARKOPOULOS, T., The Future in Greek: From Ancient to Medieval (Oxford, 2008).
MARSHALL, I. H., The Gospel of Luke (NIGTC; Exeter, 1978).
MARTINI, C. M., ‘Eclecticism and Atticism in the Textual Criticism of the Greek
New Testament, in M. Black and W. A. Smalley (eds.), On Language, Cul-
ture and Religion in Honor of E. A. Nida (The Hague/Paris, 1974), 149–56.
MASSON, É., Recherches sur les plus anciens emprunts sémitiques en grec (Paris,
1967).
MAUERSBERGER, A., Polybios-Lexikon (Berlin, 1956–2006).
MAYSER, E., Grammatik der griechischen Papyri aus der Ptolemäerzeit (2 vols.
in 6 parts; Berlin and Leipzig, 1906–1938; repr. 1970).
MCLEAN, B. H., An Introduction to Greek Epigraphy of the Hellenistic and
Roman Periods from Alexander the Great down to the Reign of Constantine
(323 B.C.– A.D. 337) (Ann Arbor, MI, 2002).
MEECHAM, H. G., The Letter of Aristeas: A Linguistic Study with Special Reference
to the Greek Bible (Manchester, 1935).
MEER, M. N. van der, Review of NETS (2007), BIOSCS 41 (2008), 114–21.
— ‘Bridge over Troubled Waters? The gephura in the Old Greek of Isaiah 37:25
and Contemporary Greek Sources’ in M. K. H. Peters (ed.), XIII Congress
of the International Organization for Septuagint and Cognate Studies, Lju-
bljana 2007 (Atlanta, 2008), 305–24.
— ‘Trendy Translations in the Septuagint of Isaiah: A Study of the Vocabulary of
the Greek Isaiah 3:18–23 in the Light of Contemporary Sources,’ in Karrer,
Kraus, and Meiser (eds.), Die Septuaginta (2008), 581–96.
— ‘Papyrological Perspectives on the Septuagint of Isaiah,’ in van der Kooij and
van der Meer (eds.), Old Greek of Isaiah (2010), 107–33.
MELANCTHON, Philip, Grammatica graeca (Hagenau, 1527. Facsimile ed., Men-
ston, 1969).
METZGER, B. M., The Text of the New Testament: Its Transmission, Corruption,
and Restoration (Oxford, 1st ed. 1964; 2nd ed. 1968).
— A Textual Commentary on the Greek New Testament (London/New York
1975).
— Manuscripts of the Greek Bible: An Introduction to Greek Palaeography
(New York/Oxford, 1981).
MIGLIORINI, B., Storia della lingua italiana (3rd ed.; Firenze, 1961).
— The Italian Language (rev. and trans. T. G. Griffith; London, 1966).
MIHEVC-GABROVEC, E., Études sur la syntaxe de Ioannes Moschos (Ljubljana,
1960).
MILLAR, F., A Greek Roman Empire: Power and Belief under Theodosius II (408–
450) (Berkeley, 2006).
MILLIGAN, G., Selections from the Greek Papyri (Cambridge, 1912).
MIRAMBEL, A., ‘Essai sur l’évolution du verbe en grec byzantin,’ BSL 61 (1966),
167–90.
MONTEVECCHI, O., ‘Continuità ed evoluzione della lingua greca nella Settanta
e nei papiri,’ Actes du Xe Congrès international de Papyrologues (Warsaw,
1964), 39–49.
MORGAN, T., Literate Education in the Hellenistic and Roman Worlds (Cambridge,
1998).
MORISON, S., Politics and Script: Aspects of Authority and Freedom in the Devel-
opment of Graeco-Latin Script from the Sixth Century B.C. to the Twentieth
Century A.D (Oxford, 1972).
MORRIS, L., Studies in the Fourth Gospel (Grand Rapids, 1969).
— The Gospel According to John (Grand Rapids, 1971).
MORWOOD, J., and J. TAYLOR, The Pocket Oxford Classical Greek Dictionary
(Oxford, 2002).
MOULE, C. F. D., An Idiom Book of New Testament Greek (Cambridge, 1963).
MURAOKA, T. (ed.), Melbourne Symposium on Septuagint Lexicography (Atlanta,
1990).
— A Greek-English Lexicon of the Septuagint (Twelve Prophets), (Louvain,
1993).
— A Greek-English Lexicon of the Septuagint: Chiefly of the Pentateuch and the
Twelve Prophets (Louvain, 2002).
— ‘Recent Discussions on the Septuagint Lexicography,’ in Karrer et al. (eds.),
Die Septuaginta (2008), 221–35.
— A Greek-English Lexicon of the Septuagint (Louvain, 2009).
MUSSIES, G., The Morphology of Koine Greek as used in the Apocalypse of
St. John (Leiden, 1971).
NABER, S. A. (ed.), Photii Patriarchae Lexicon (2 vols.; Leiden, 1864–1865;
repr. Amsterdam, 1965).
NEBRIJA, Antonio de, Gramatica castellana. Reproduction phototypique de l’édi-
tion princeps (1492). Publiée avec une préface par E. Walberg (Halle a. S.,
1909).
— Elio Antonio de Nebrija: Diccionario Latino-Español (Salamanca 1492) (ed.
G. Colón and A.-J. Soberanas; Barcelona, 1979).
— Vocabulario español-latino (Salamanca ¿1495?): sale nuevamento a luz
reproducida en facsímile por acuerdo de la Real Academia Española
(Madrid, 1951).
NESTLE, E., Introduction to the Textual Criticism of the Greek New Testament
(trans. W. Edie from 2nd ed.; London, 1901).
NEWMAN, B. M., A Concise Greek-English Dictionary of the New Testament (Lon-
don, 1971; rev. ed.; Stuttgart, 2010).
— ‘A Concise Greek-English Dictionary of the New Testament: Reflections
and Ruminations,’ in Taylor et al. (eds.), Biblical Greek Language and Lexi-
cography (2004), 91–3.
NGUNGA, A. T., Messianism in the Old Greek of Isaiah: An Intertextual Analysis
(Göttingen, 2013).
BIBLIOGRAPHY 469
NICKAU, K. (ed.), Ammonii qui dicitur liber De Adfinium Vocabulorum Differ-

entia (Leipzig, 1966).
NIDA, E. A., ‘Varieties of Language,’ BT 23 (1972), 316–22.
NIDA, E. A., and J. P. LOUW, Lexical Semantics of the Greek New Testament
(Atlanta, 1992).
NORDEN, E. E., Die antike Kunstprosa: Vom VI. Jahrhundert v. Chr. bis in die
Zeit der Renaissance (2 vols.; 5th ed. [= 2nd ed. 1909]; Stuttgart, 1958).
NORTON, F. J., Printing in Spain 1501–1520 (Cambridge, 1966).
— A Descriptive Catalogue of Printing in Spain and Portugal 1501–1520 (Cam-
bridge, 1978).
OATES, J. F., et al., Checklist of Editions of Greek, Latin, Demotic and Coptic
Papyri, Ostraca and Tablets (5th ed.; 2001: updates at <papyri.info>).
O’CONNELL, S., From Most Ancient Sources: The Nature and Text-Critical Use of
the Greek Old Testament Text of the Complutensian Polyglot Bible (OBO 215;
Fribourg, 2006).
O’CONNOR, M., and J. A. L. LEE, ‘A Problem in Biblical Lexicography: The
Case of Hebrew tap and Greek aposkeuê,’ ZAW 119 (2007), 403–9.
Ó FEARGHAIL, F., ‘The Imitation of the Septuagint in Luke’s Infancy Narrative,’
PIBA 12 (1989), 58–78.
OLIN, J. C., Catholic Reform from Cardinal Ximenes to the Council of Trent, 1495-
1563 (New York, 1990).
OLLEY, J. W., ‘Righteousness’ in the Septuagint of Isaiah: A Contextual Study
(SCS 8; Missoula, 1979).
ORLANDOS, A. K., and I. N. TRAVLOS, Λεξικὸν Ἀρχαίων Ἀρχιτεκτονικῶν Ὅρων
(Athens, 1986).
O’SULLIVAN, J. N., A Lexicon to Achilles Tatius (Berlin/New York, 1980).
OTTLEY, R. R., The Book of Isaiah According to the Septuagint (Codex Alexan-
drinus) (2 vols.; Cambridge, 1904, 1906).
— A Handbook to the Septuagint (London, 1920).
PAGE, D. L., Euripides Medea: The Text Edited with Introduction and Commen-
tary (Oxford, 1938).
PALMER, L. R., The Latin Language (London, 1954).
PAPPAS, P. A., ‘The Microcosm of a Morphological Change: Variation in thelô
+ infinitive futures and êthela + infinitive counterfactuals in Early Modern
Greek,’ Diachronica 18 (2001), 59–92.
PARKE, H. W., Greek Mercenary Soldiers (Oxford, 1933).
PASCHOS, P. B., Τὸ Γεροντικόν, ἤτοι Ἀποφθέγματα Ἁγίων Γερόντων. Πρόλογος,
Κείμενον, Γλωσσάριον, Σχόλια, Εὑρετήριον θεμάτων (3rd ed.; Athens, 1981).
PASSOW, F., Handwörterbuch der Griechischen Sprache (4th ed.; Leipzig,
1831).
PAUL, S. M., et al. (eds.), Emanuel: Studies in Hebrew Bible, Septuagint, and
Dead Sea Scrolls in Honor of Emanuel Tov (Leiden, 2003).
PELÁEZ, J., ‘New Testament Lexicography: A Critical Introduction,’ in Peláez
and Mateos, New Testament Lexicography (2010), 3–54.
PELÁEZ, J., and MATEOS, J., New Testament Lexicography: Introduction, Theory,
Method (trans. A. Bowden, ed. D. S. du Toit; FSBP 6; Berlin, 2010).
PELLETIER, A., ‘Le vocabulaire du commandement dans le Pentateuque des LXX
et dans le Nouveau Testament,’ RSR 41 (1953), 519–24.
PENTZAS, B., Εἶπε Γέρων· Τὸ Γεροντικὸν σὲ νεοελληνικὴ ἀπόδοση (Athens, 1983).

PERKINS, L. J., ‘The Lord is Warrior,’ BIOSCS 40 (2007), 121–38.
PERRY, B. E., Aesopica (Urbana, 1952).
PIETERSMA, A. (trans.), A New English Translation of the Septuagint: The Psalms
(New York/Oxford, 2000).
— ‘A Panel Presentation on Ronald Troxel’s LXX-Isaiah,’ in Boyd-Taylor (ed.),
A Question of Methodology (2013), 339–58.
— ‘A Response to Muraoka’s Critique of Interlinearity,’ in Boyd-Taylor (ed.),
A Question of Methodology, 315–37.
— ‘Context is King in Septuagint Lexicography—Or Is It?,’ in Aitken and
Evans (eds.), Biblical Greek in Context (2015), 165–75.
PINKSTER, H., ‘The Development of Future Tense Auxiliaries in Latin,’ Glotta 63
(1985), 186–208.
PORTER, S. E. (ed.), Handbook of Classical Rhetoric in the Hellenistic Period
330 B.C.–A.D. 400 (Leiden, 1997).
— ‘Paul of Tarsus and His Letters,’ in Porter (ed.), Handbook (1997), 533–85.
— ‘Lexical Semantics and New Testament Greek: A Review Article of Some
Major Works,’ FilNT (2019), 113–55.
PORTER, S. E., and A. W. PITTS (eds.), The Language of the New Testament:
Context, History, and Development (Leiden, 2013).
PREISIGKE, F., Fachwörter des öffentlichen Verwaltungsdienstes Ägyptens in den
griechischen Papyrusurkunden der ptolemäisch-römischen Zeit (Göttingen,
1915).
PRELLWITZ, W., Etymologisches Wörterbuch der griechischen Sprache (Göttingen,
1892).
PROCTOR, R., The Printing of Greek in the Fifteenth Century (London, 1900).
PSICHARI, J., Quelques travaux de linguistique, de philologie et de littérature hel-
léniques (1884–1928) (vol. 1; Paris, 1930).
QUASTEN, J., Patrology, Vol. 3 (Utrecht, 1960; repr. Allen, Tex., n.d.).
RAJAK, T., Translation and Survival: The Greek Bible and the Ancient Jewish
Diaspora (Oxford, 2009).
REESE, J. M., Hellenistic Influence on the Book of Wisdom and its Consequences
(AnBib 41; Rome, 1970).
REEVE, A., and M. A. SCREECH (eds.), Erasmus’ Annotations on the New Testa-
ment: Acts–Romans–I and II Corinthians (Leiden, 1990).
REGNAULT, L., Les sentences des pères du désert: Collection alphabétique
(Solesmes, 1981).
REITERER, F. V., ‘Review of Recent Research on the Book of Ben Sira,’ in
P. C. Beentjes (ed.), Book of Ben Sira in Modern Research (Berlin, 1997),
23–60.
RENGSTORF, K. H., A Complete Concordance to Flavius Josephus (Leiden, 1973–
1983).
RENOUARD, A. A., Annales de l’imprimerie des Alde: ou histoire des trois Manuce
et de leurs éditions (3rd ed.; Paris, 1834).
RIESENFELD, H., Zum Gebrauch von θέλω im Neuen Testament (Uppsala, 1936).
RINK, H., Strassen- und Viertelnamen von Oxyrhynchus (Giessen, 1924).
RITSCHL, F. (ed.), Thomae Magistri sive Theoduli monachi ecloga vocum attica-
rum (Halis Saxonum, 1832).
BIBLIOGRAPHY 471
ROBERT, L., À travers l’Asie Mineure: Poètes et prosateurs, monnaies grecques,

voyageurs et géographie (Paris, 1980).
ROBERTSON, A. T., A Grammar of the Greek New Testament in the Light of His-
torical Research (4th ed.; Nashville, 1934).
ROCCI, L., Vocabolario Greco Italiano (3rd ed.; Rome, 1943).
RÖDIGER, R., ‘βούλομαι und ἐθέλω,’ Glotta 8 (1917), 1–24.
RÖNNEGÅRD, P., Threads and Images: The Use of Scripture in Apophthegmata
Patrum (CB, NT Series 44; Winona Lake, IN, 2010).
ROSS, W. A., ‘Lexical Possibilities in LXX Research: Revision and Expansion,’
in W. Kraus et al. (eds.), XV Congress of the International Organization
for Septuagint and Cognate Studies Munich, 2013 (Atlanta, 2016), 341–
59.
ROUSSEAU, P., Basil of Caesarea (Berkeley, 1994).
ROWE, G. O., ‘Style,’ in Porter (ed.), Handbook of Classical Rhetoric (1997),
121–57.
RUIJGH, C. J., ‘La place des enclitiques dans l’ordre des mots chez Homère d’après
la loi de Wackernagel,’ in H. Eichner and H. Rix (eds.), Sprachwissenschaft
und Philologie: Jacob Wackernagel und die Indogermanistik heute (Wies-
baden, 1990), 213–33.
RUSSELL, N., The Lives of the Desert Fathers (London, 1981).
RUTHERFORD, W. G. (ed.), The New Phrynichus: being a revised text of The Ecloga
of the Grammarian Phrynichus (London, 1881).
RYDBECK, L., Fachprosa, Vermeintliche Volkssprache und Neues Testament
(Uppsala, 1967).
SÁENZ-BADILLOS, Á., La filología bíblica en los primeros helenistas de Alcalá
(Estella, 1990).
SAKALIS, D. T., ‘Ἀνάττικα καὶ Ψευδαττικὰ ἀπὸ τὸν Ἀττικιστὴ Μοίρη,’ Δωδώνη 6
(1977), 441–70.
SATTERTHWAITE, P. E., ‘Judges: To the Reader,’ NETS, 195–200.
SCHLEUSNER, J. F., Novus Thesaurus Philologico-criticus: sive Lexicon in LXX. et
Reliquos Interpretes Graecos ac Scriptores Apocryphos Veteris Testamenti
(2nd ed.; Glasgow, 1822).
SCHMID, W., Der Atticismus in seinen Hauptvertretern von Dionysius von Halikar-
nass bis auf den zweiten Philostratus (5 vols.; Stuttgart, 1887–1897).
— Review of LSJ Part 4. II, PhilWoch 51 (1931), 705–18.
SCHMIDT, J. H. H., Synonymik der griechischen Sprache (4 vols.; Leipzig, 1876–
1886.)
SCHOLDERER, V., Greek Printing Types, 1465–1927: Facsimiles from an Exhibi-
tion of Books Illustrating the Development of Greek Printing Shown in the
British Museum 1927 (London, 1927).
SCHRENK, G., ‘θέλω, θέλημα, θέλησις,’ in TDNT, 3:44–62.
SCOTT, J. A., ‘The Vocative in Homer and Hesiod,’ AJP 24 (1903), 192–6.
— ‘The Vocative in Aeschylus and Sophocles,’ AJP 25 (1904), 81–4.
— ‘Additional Notes on the Vocative,’ AJP 26 (1905), 32–43.
SCREECH, M. A., Rabelais (London, 1979).
SCRIVENER, F. H. A., A Plain Introduction to the Criticism of the New Testament:
For the Use of Biblical Students (1st ed., 1861; 4th ed., 2 vols.; ed. E. Miller,
London, 1894).
SEELIGMANN, I. L., The Septuagint Version of Isaiah: A Discussion of its Problems

(Leiden, 1948).
SHAW, F., Review of Muraoka, Lexicon, BMCR 2010.04.20.
SHIPP, G. P., ‘Some Observations on the Distribution of Words in the New Testa-
ment,’ in E. C. B. MacLaurin (ed.), Essays in Honour of Griffithes Wheeler
Thatcher 1863–1950 (Sydney, 1967), 127–38.
SILVA, M., Biblical Words and their Meaning: An Introduction to Lexical Seman-
tics (Grand Rapids, 1983).
— ‘Esaias: To the Reader,’ NETS, 823–5.
— Review of Muraoka, Lexicon, WTJ 72 (2010), 435–8.
SIMPSON, E. K., Words Worth Weighing in the Greek New Testament (London,
1946).
SKEHAN, P. W., and A. A. DI LELLA, The Wisdom of Ben Sira: A New Transla-
tion with Notes (AB 39; New York, 1987).
SKINNER, J., A Critical and Exegetical Commentary on Genesis (ICC; 2nd ed.;
Edinburgh, 1930).
SMEND, R. (ed.), Die Weisheit des Jesus Sirach: Hebräisch und Deutsch, mit
einem hebräischen Glossar (Berlin, 1906).
— Griechisch-Syrisch-Hebräischer Index zur Weisheit des Jesus Sirach (Berlin,
1907).
SMITH, H. P., The Books of Samuel (ICC; Edinburgh, 1899).
SMYTH, H. W., Greek Grammar, rev. G. M. Messing (Cambridge, MA, 1956).
SOISALON-SOININEN, ‘Der Gebrauch des genetivus absolutus in der Septuaginta’
(1973), in Studien zur Septuaginta-Syntax (1987), 175–80.
— ‘The Rendering of the Hebrew Relative Clause in the Greek Pentateuch,’ in
Proceedings of the Sixth World Congress of Jewish Studies Held at the Hebrew
University of Jerusalem, 13–19 August, 1973 (vol. 1; Jerusalem, 1977), 401–
6. Repr. in Studien zur Septuaginta-Syntax (1987), 55–61.
— ‘Die Wiedergabe einiger hebräischer Zeitangaben mit der Präposition ‫ ב‬in
der Septuaginta’ (1978), in Studien zur Septuaginta-Syntax (1987), 107–
15.
— ‘Renderings of Hebrew Comparative Expressions with ‫ מן‬in the Greek Pen-
tateuch’ (1979), in Studien zur Septuaginta-Syntax (1987), 141–53.
— ‘ἐν für εἰς in der Septuaginta’ (1982), in Studien zur Septuaginta-Syntax (1987),
131–40.
— ‘Die Wiedergabe des ‫ ב‬instrumenti im griechischen Pentateuch’ (1983), in
Studien zur Septuaginta-Syntax (1987), 116–30.
— ‘Die Wiedergabe des partitiven ‫ מן‬im griechischen Pentateuch’ (1985), Studien
zur Septuaginta-Syntax (1987), 154–71.
— Studien zur Septuaginta-Syntax: Zu seinem 70. Geburtstag am 4. Juli 1987
herausgegeben von Anneli Aejmelaeus und Raija Sollamo (AASF 237;
Helsinki, 1987).
— ‘Zurück zur Hebraismenfrage,’ in D. Fraenkel, U. Quast and J. W. Wevers
(eds.), Studien zur Septuaginta. Robert Hanhart zu Ehren (Göttingen, 1990),
35–51.
SOLLAMO, R., ‘Some “Improper” Prepositions, such as ἐνώπιον, ἐναντίον, ἔναντι,
etc., in the Septuagint and Early Koine Greek,’ VT 25 (1975), 773–82.
— Renderings of Hebrew Semiprepositions in the Septuagint (Helsinki, 1979).
BIBLIOGRAPHY 473
— ‘Repetitions of Prepositions in the Septuagint of Genesis,’ in F. García Martínez

and M. Vervenne (eds.), Interpreting Translation: Studies on the LXX and
Ezekiel in Honour of Johan Lust (BETL 192; Leuven, 2005), 371–84.
— ‘The Place of the Enclitic Personal Pronouns in the Old Greek Psalter,’ in
M. K. H. Peters (ed.), XII Congress of the International Organization for
Septuagint and Cognate Studies, Leiden 2004 (Atlanta, 2006), 153–60.
— ‘The Usage of the Article with Nouns Defined by a Nominal or Pronominal
Genitive in LXX Genesis,’ in Kauhanen and Vanonen (eds.), The Legacy of
Soisalon-Soininen (2020), 43–53.
SOPHOCLES, E. A., Greek Lexicon of the Roman and Byzantine Periods (New York,
1887).
SOTIROPOULOS, D., ‘Diglossia and the National Language Question in Modern
Greece,’ Linguistics 197 (1977), 5–31.
SOUTER, A., A Pocket Lexicon to the Greek New Testament (Oxford, 1916).
— ‘Ἀγαπητός,’ JTS 28 (1926–1927), 59–60.
SPARKS, H. D. F., ‘The Semitisms of St. Luke’s Gospel,’ JTS 44 (1943), 129–38.
— ‘The Semitisms of the Acts,’ JTS n.s. 1 (1950), 16–28.
SPEISER, E. A., Genesis (New York, 1964).
SPOTTORNO, V., ‘The Relative Pronoun in the New Testament: Some Critical
Remarks,’ NTS 28 (1982), 132–41.
STAIKOS, K. S., and T. E. SKLAVENITES, Πεντακόσια Χρόνια· Εντύπης Παράδοσης
του Νέου Ελληνισμού (1499–1999) (Athens, 2000).
STASINOPOULOS, M. D., Μορφὲς ἀπὸ τὸν Τέταρτον Αἰῶνα μ. Χ.: Ἱστορικὴ Εἰσα-
γωγὴ στὸν Λόγο πρὸς τοὺς Νέους τοῦ Μεγάλου Βασιλείου (Athens, 1972).
STEEN, H. A., ‘Les clichés épistolaires dans les lettres sur papyrus grecques,’ Class
et Med 1 (1938), 119–76.
STEPHANUS, H., Thesaurus Graecae Linguae, ed. C. B. Hase et al. (Paris, 1831–
1865).
STOLK, J. V., ‘Dative and Genitive Case Interchange in Greek Papyri from Roman–
Byzantine Egypt,’ Glotta 93 (2017), 182–212.
STRAY, C., (ed.), Oxford Classics: Teaching and Learning 1800–2000 (Oxford,
2007).
— Classical Dictionaries: Past, Present and Future (London, 2010).
STRAY, C., et al. (eds.), Liddell and Scott: The History, Methodology, and Lan-
guages of the World’s Leading Lexicon of Ancient Greek (Oxford, 2019).
STROUT, D., and R. FRENCH, ‘Phrynichos,’ in Pauly-Wissowa, RE 20:920–5.
STURZ, F. G., De dialecto macedonica et alexandrina (Lipsiae, 1808).
SWAIN, S., Hellenism and Empire: Language, Classicism, and Power in the
Greek World, AD 50–250 (Oxford, 1996).
SWETE, H. B., An Introduction to the Old Testament in Greek (rev. R. R. Ottley;
Cambridge, 1914).
SWINN, S. P., ‘Ἀγαπᾶν in the Septuagint,’ in Muraoka (ed.), Melbourne Symposium
(1990), 49–81.
TABACHOVITZ, D., Die Septuaginta und das Neue Testament: Stilstudien (Lund,
1956).
TAGLIAFERRO, E., in ‘Nota linguistica filoniana (Cher. 1–10),’ Helikon 18–19 (1978–
1979), 415–424.
TALSHIR, Z., 1 Esdras: A Text Critical Commentary (Atlanta, 2001).
TARAZI, P. N., Galatians (OBS; Crestwood, 1994).

TASKER, R. V. G., ‘The Complutensian Polyglot,’ CQR (1953), 197–210.
TAYLOR, B. A. (ed.), IX Congress of the International Organization for Septuagint
and Cognate Studies, Cambridge 1995 (Atlanta, 1997).
— X Congress of the International Organization for Septuagint and Cognate
Studies, Oslo 1998 (Atlanta, 2001).
TAYLOR, B. A., et al. (eds.), Biblical Greek Language and Lexicography: Essays
in Honor of Frederick W. Danker (Grand Rapids, 2004).
TAYLOR, V., The Gospel According to St. Mark (2nd ed.; London, 1966).
THACKERAY, H. St. J., ‘The Greek Translators of the Prophetical Books,’ JTS 4
(1903), 578–85.
— A Grammar of the Old Testament in Greek According to the Septuagint (vol. 1;
Cambridge, 1909).
— The Septuagint and Jewish Worship: A Study in Origins (Schweich Lectures
1920; London 1921).
THESLEFF, H., Review of Rydbeck, Fachprosa (1967), Gnomon 42 (1970), 551–
5.
THIERSCH, H. G. J., De Pentateuchi versione Alexandrina libri tres (Erlangae,
1841).
THOMAS, H., Spanish Sixteenth-Century Printing (London, 1926).
THOMPSON, A., Review of Lee, History, BIOSCS 36 (2003), 113–27.
— ‘Βάπτω: An Illustration of the State of our Ancient Greek Dictionaries,’ in
Stray et al. (eds.), Liddell and Scott (2019), 353–94.
— ‘Response: Thompson on Goldhill on Stray, Clarke and Katz, Liddell and
Scott,’ BMCR 2021.06.35.
THOMPSON, S., The Apocalypse and Semitic Syntax (SNTSMS 52; Cambridge,
1985).
THRALL, M. E., Greek Particles in the New Testament: Linguistic and Exegetical
Studies (Leiden, 1962).
THUMB, D., Die griechische Sprache im Zeitalter des Hellenismus (Strassburg,
1901).
— Handbook of the Modern Greek Vernacular: Grammar, Texts, Glossary
(trans. S. Angus; Edinburgh, 1912).
THURÉN, L., ‘The General New Testament Writings,’ in Porter (ed.), Handbook
(1997), 587–607.
TORREY, C. C., Our Translated Gospels (London, n.d.).
TOV, E., The Septuagint Translation of Jeremiah and Baruch (HSM 8; Missoula
1976).
— ‘Three Dimensions of LXX Words,’ RB 83 (1976), 529–44.
— ‘Compound Words in the LXX Representing Two or More Hebrew Words,’
Biblica 58 (1977), 189–212.
— The Text-Critical Use of the Septuagint in Biblical Research (1st ed. 1981;
3rd ed.; Winona Lake, 2015).
— ‘Did the Septuagint Translators Always Understand Their Hebrew Text?,’ in
A. Pietersma and C. Cox (eds.), De Septuaginta: Studies in Honour of John
William Wevers on his Sixty-fifth Birthday (Mississauga, 1984), 53–70.
— The Greek and Hebrew Bible: Collected Essays on the Septuagint (Leiden,
1999).
BIBLIOGRAPHY 475
— ‘The Septuagint,’ in M. J. Mulder and H. Sysling (eds.), Mikra: Text, Trans-

lation, Reading, and Interpretation of the Hebrew Bible in Ancient Judaism
and Early Christianity (Peabody, 2004), 161–88.
TOY, C. H., Proverbs (ICC; Edinburgh, 1899).
TREGELLES, S. P., An Account of the Printed Text of the Greek New Testament
(London, 1854).
TRENCHARD, W. C., A Concise Dictionary of New Testament Greek (Cambridge,
2003).
TRIANTAPHYLLIDES, M. A., Νεοελληνική Γραμματική, Πρώτος Τόμος· Ιστορική
Εισαγωγή (Αθήνα, 1938).
— Νεοελληνική Γραμματική· Αναπροσαρμογή της Μικρής Νεοελληνικής Γραμ-
ματικής (Αθήνα, 1986).
TRONCI, L., ‘Future Forms in Postclassical Greek: Some Remarks on the Sep-
tuagint and the New Testament,’ in D. Rifayenko and I. A. Seržant (eds.),
Postclassical Greek: Contemporary Approaches to Philology and Linguistics
(TiLSM 335; Berlin, 2020), 111–43.
TROXEL, R. L., LXX-Isaiah as Translation and Interpretation: The Strategies of
the Translator of the Septuagint of Isaiah (Leiden, 2008).
— ‘βουλή and βουλεύειν in LXX Isaiah,’ in van der Kooij and van der Meer
(eds.), Old Greek of Isaiah (2010), 153–71.
TRUDGILL, P., Sociolinguistics (Harmondsworth, 1974).
[TURNER, C. H.], ‘Ἀγαπητός,’ JTS 20 (1919), 339–44.
TURNER, C. H., ‘Ὁ υἱός μου ὁ ἀγαπητός,’ JTS 27 (1925–1926), 113–29.
— ‘Marcan Usage: Notes, Critical and Exegetical, on the Second Gospel,’ JTS 28
(1926–1927), 349–62.
— The Early Printed Editions of the Greek Testament (Oxford, 1924).
TURNER, E. G., Greek Papyri: An Introduction (Oxford, 1968).
— Greek Manuscripts of the Ancient World (2nd ed., rev. P. J. Parsons; ICS Bul-
letin Suppl. 46; London, 1987).
TURNER, G. W., Stylistics (Harmondsworth, 1973).
TURNER, N., ‘The Literary Character of New Testament Greek,’ NTS 20 (1973–
1974), 107–14.
VALENTE, S., The Antiatticist: Introduction and Critical Edition (Berlin, 2015).
VEITCH, W., Greek Verbs Irregular and Defective (Oxford, 1866).
VOELZ, J. W., ‘The Language of the New Testament,’ ANRW II 25.2 (1984), 893–
977.
VORM-CROUGHS, M. van der, ‘LXX Isaiah and the Use of Rhetorical Figures,’ in
van der Kooij and van der Meer (eds.), Old Greek of Isaiah (2010), 173–88.
WADDELL, W. G. (ed.), Herodotus Book II (London, 1939).
WAGNER, C., Die Septuaginta-Hapaxlegomena in Buch Jesus Sirach: Untersu-
chungen zu Wortwahl und Wortbildung unter besonderer Berücksichtigung
des textkritischen und übersetzungstechnischen Aspekts (Berlin, 1999).
WAGNER, J. R., Reading the Sealed Book: Old Greek Isaiah and the Problem of
Septuagint Hermeneutics (Tübingen, 2013).
WALBANK, F. W., A Historical Commentary on Polybius (3 vols.; Oxford, 1957–
1979).
WALTERS, P., The Text of the Septuagint: Its Corruptions and their Emendation
(ed. D. W. Gooding; Cambridge, 1973).
WARD, B., The Sayings of the Desert Fathers: The Alphabetical Collection, Trans-
lated, with a Foreword (rev. ed.; London, 1984).
WEINFELD, M., Book of Genesis (Tel Aviv, 1975) [in Hebrew].
WEINREICH, U., Languages in Contact (New York, 1953).
WELLHAUSEN, J., Einleitung in die drei ersten Evangelien (2nd ed.; Berlin, 1911).
WENDEL, C., ‘Moiris,’ in Pauly-Wissowa, RE 15:2501–12.
WESSELY, C., Papyrorum Scripturae Graecae Specimina lsagogica (Leipzig, 1900).
WEVERS, J. W., Notes on the Greek Text of Genesis (SCS 35; Atlanta, 1993).
— Notes on the Greek Text of Exodus (SCS 30; Atlanta, 1990).
— Notes on the Greek Text of Numbers (SCS 46; Atlanta, 1998).
— Notes on the Greek Text of Deuteronomy (SCS 39; Atlanta, 1995).
WHITMARSH, T., The Second Sophistic (G&R: New Surveys in the Classics 35;
Oxford, 2005).
WIFSTRAND, A., ‘Die griechischen Verba für wollen,’ Eranos 40 (1942), 16–36.
— ‘Stylistic Problems in the Epistles of James and Peter,’ Studia Theologica 1
(1948), 170–82.
WILCKEN, U., ‘III. Referate: Papyrus-Urkunden,’ AFP 8 (1927), 63–167.
WILCOX, M., ‘Semitisms in the New Testament,’ ANRW II.25.2 (1984), 978–1029.
WILK, F., ‘Between Scripture and History: Technique and Hermeneutics of Inter-
preting Biblical Prophets in the Septuagint of Isaiah and the Letters of Paul,’
in van der Kooij and van der Meer (eds.), Old Greek of Isaiah (2010), 189–
209.
WILLI-PLEIN, I., Review of Lee, Lexical Study, ThLZ 110 (1985), 94–5.
WILSON, N., Mediaeval Greek Bookhands: Examples Selected from Greek Manu-
scripts in Oxford Libraries (Plates, Text; Cambridge, MA, 1972–1973).
WILSON, N. G. (ed.), Saint Basil on the Value of Greek Literature (London, 1975).
— ‘Vettor Fausto, Professor of Greek and Naval Architect,’ in A. C. Dionisotti
et al. (eds.), The Uses of Greek and Latin: Historical Essays (London, 1988),
89–95.
WINSTON, D., The Wisdom of Solomon: A New Translation with Introduction
and Commentary (AB; New York, 1979).
WITHERINGTON III, B., The Acts of the Apostles: A Socio-Rhetorical Commentary
(Grand Rapids, 1998).
WOODHEAD, A. G., The Study of Greek Inscriptions (2nd ed.; Cambridge, 1981).
WOODY, K. M., ‘A Note on the Greek Fonts of the Complutensian Polyglot,’
Papers of the Bibliographical Society of America 65 (1971), 143–9.
WORTLEY, J., ‘How the Desert Fathers “Meditated”,’ GRBS 46 (2006), 315–28.
WRIGHT, B. G., No Small Difference: Sirach’s Relationship to its Hebrew Parent
Text (SBLSCS 26; Atlanta, 1989).
— ‘The Septuagint and Its Modern Translators,’ in Karrer et al. (eds.), Die Sep-
tuaginta (2008), 102–14.
YOUNG, F., L. AYRES, and A. LOUTH, The Cambridge History of Early Christian
Literature (Cambridge, 2004).
ZERWICK, M., Biblical Greek: Illustrated by Examples, trans., rev. J. Smith (Rome,
1963).
ZGUSTA, L., Manual of Lexicography (The Hague/Paris, 1971).
ZIEGLER, J., Untersuchungen zur Septuaginta des Buches Isaias (Münster, 1934).
BIBLIOGRAPHY 477
— ‘Zum Wortschatz des griechischen Sirach,’ in J. Hempel and L. Rost (eds.), Von
Ugarit nach Qumran: Beiträge zur Alttestamentlichen und Altorientischen
Forschung. Otto Eissfeldt zum 1. September 1957 dargebracht (BZAW 77;
2nd ed.; Berlin, 1961), 274–87.
— ‘Die Vokabel-Varianten der O-Rezension im griechischen Sirach,’ in D. Winton
Thomas and W. D. McHardy (eds.), Hebrew and Semitic Studies Presented
to Godfrey Rolles Driver in celebration of his seventieth birthday, 20 August
1962 (Oxford, 1963), 172–90.
— ‘Ursprüngliche Lesarten im griechischen Sirach,’ in Mélanges Eugène Tisserant
(Studi e Testi 231; 7 vols.; Città del Vaticano, 1964), 1:461–87.
— (ed.), Sapientia Iesu Filii Sirach (Septuaginta Vetus Testamentum Graecum
Auctoritate Academiae Scientiarum Gottingensis editum XII.2; 2nd ed.;
Göttingen, 1980).
ZORELL, F., Novi testamenti lexicon graecum (Paris, 1911; 5th ed. Rome, 1999).
— Lexicon Hebraicum Veteris Testamenti (Rome, 1984).
NOTICE OF ORIGINAL PUBLICATION
[WITH PERMISSIONS]
1 ‘A Note on Septuagint Material in the Supplement to Liddell and

Scott,’ Glotta 47 (1969), 234–42.
By permission of Brill Deutschland GmbH, Paderborn.
2 ‘Aposkeuê in the Septuagint,’ JTS 23 (1972), 430–7.
By permission of Oxford University Press.
3 ‘Equivocal and Stereotyped Renderings in the LXX,’ RB 87 (1980), 104–17.
By permission of Revue Biblique.
4 ‘The Future of zên in Late Greek,’ NovT 22 (1980), 289–98.
By permission of Koninklijke Brill NV, Leiden.
5 ‘Some Features of the Speech of Jesus in Mark’s Gospel,’ NovT 27 (1985),
1–26.
6 ‘A Non-Aramaism in Luke 6.7,’ NovT 33 (1991), 28–34.
7 ‘The United Bible Societies’ Lexicon and its Analysis of Meanings,’ FilNT 5
(1992), 167–89.
By permission of the editor of Filologia Neotestamentaria and Ediciones El Almendro.
8 ‘A Lexical Study Thirty Years On, with Observations on “Order” Words in
the LXX Pentateuch,’ in S. M. Paul et al. (eds.), Emanuel: Studies in Hebrew
Bible, Septuagint and Dead Sea Scrolls in Honor of Emanuel Tov (Leiden:
Brill, 2003), 513–24.
9 ‘The Present State of Lexicography of Ancient Greek,’ in B. A. Taylor et al.
(eds.), Biblical Greek Language and Lexicography: Essays in Honor of
Frederick W. Danker (Grand Rapids, MI: Eerdmans, 2004), 66–74.
Reprinted by permission of the publisher, Wm. B. Eerdmans Publishing Co.
10 Review of: T. Muraoka, A Greek-English Lexicon of the Septuagint: Chiefly
of the Pentateuch and the Twelve Prophets (2002), BIOSCS 37 (2004), 127–
39.
By permission of the editor of BIOSCS.
11 ‘Dimitrios Doukas and the Accentuation of the New Testament Text of the
Complutensian Polyglot,’ NovT 47 (2005), 250–90.
Image of page of the Complutensian Polyglot vol. 5 (1514) from the holdings of the
Biblioteca Nacional de España.
12 ‘Exapostellô,’ in J. Joosten and P. J. Tomson (eds.), Voces Biblicae: Septua-
gint Greek and its Significance for the New Testament (CBET 49; Leuven:
Peeters, 2007), 99–113.
By permission of Peeters, Leuven.
13 ‘A Lexicographical Database for Greek: Can it Be Far Off? The Case of

amphodon,’ in M. Karrer et al. (eds.), Die Septuaginta – Texte, Kontexte,
Lebenswelten: Internationale Fachtagung veranstaltet von Septuaginta
Deutsch (LXX.D), Wuppertal 20.–23. Juli 2006 (Tübingen: Mohr Siebeck,
2008), 214–20.
By permission of Mohr Siebeck GmbH & Co., Tübingen.
14 ‘The Complutensian Polyglot, the Text of Sirach, and a Lost Greek Word,’
BIOSCS 42 (2009), 95–108.
15 Review of: T. Muraoka, A Greek-English Lexicon of the Septuagint (2009),
BIOSCS 43 (2010), 115–25.
16 ‘Auxiliary thelô,’ in T. V. Evans and D. D. Obbink (eds.), The Language of
the Papyri (Oxford: OUP, 2010), 15–34.
By permission of Oxford University Press.
17 ‘Releasing LSJ From Its Past,’ in C. Stray (ed.), Classical Dictionaries: Past,
Present and Future (London: Duckworth, 2010), 119–38.
Bristol Classical Press, by permission of Bloomsbury Publishing Plc.
18 ‘Why Didn’t St Basil Write in New Testament Greek?,’ Phronema 25 (2010),
3–20.
By permission of Phronema, Journal of St Andrew’s Greek Orthodox Theological
College, Sydney.
19 ‘The Atticist Grammarians,’ in S. E. Porter and A. W. Pitts (eds.), The Lan-
guage of the New Testament: Context, History and Development (Leiden:
Brill, 2013), 283–308.
20 ‘Etymological Follies: Three Recent Lexicons of the New Testament,’ NovT 55
(2013), 383–403.
21 ‘The onoma Rule,’ NovT 56 (2014), 411–21.
22 ‘The Literary Greek of Septuagint Isaiah,’ Sem et Class 7 (2014), 135–46.
DOI: 10.1484/J.SEC.5.103523
By permission of Brepols Publishers, Belgium.
23 ‘The Sayings of the Desert Fathers: Their Evidence for Late Koine Greek,’
Phronema 30 (2015), 23–42.
By permission of Phronema, Journal of St Andrew’s Greek Orthodox Theological
College, Sydney.
24 ‘The Vocabulary of the LXX and Documentary Evidence,’ in E. Bons and
J. Joosten (eds.), Handbuch zur Septuaginta: Handbook of the Septuagint,
vol. 3 Die Sprache der Septuaginta: The Language of the Septuagint (Güter-
sloh: Gütersloher Verlagshaus, 2016), 98–108.
Author’s copyright.
25 ‘The Puzzle of John 21:15–17: A Formality Solution,’ NovT 59 (2017), 27–
30.
26 ‘The Brill Dictionary of Ancient Greek: New Clothes for an Old Emperor,’
NovT 59 (2017), 415–35.
NOTICE OF ORIGINAL PUBLICATION 481
27 ‘Greek Idiom in the LXX-Pentateuch: The Preposition para,’ JSCS 52 (2019),

39–51.
By permission of the editor of JSCS.
28 ‘Back to the Question of Greek Idiom,’ in T. Kauhanen and H. Vanonen (eds.),
The Legacy of Soisalon-Soininen: Towards a Syntax of Septuagint Greek
(DSI 13; Göttingen: V&R, 2020), 13–25.
Vandenhoeck & Ruprecht, by permission of Brill Deutschland GmbH.
INDEX OF GREEK WORDS
The following are not indexed: illustrations from Muraoka, Lexicon in Essays 10 (pp. 138–
42) and 15 (p. 225); illustrations of accents in Essay 11 (pp. 147–8); lists of examples
from the Atticist grammarians in Essay 19 (p. 313); the list of common Koine vocabulary
in Essay 23 (p. 383); the list of words in papyri in Essay 26, §5.4 (pp. 419–20).
ἀγαπάζω 259 ἀποσκορακίζω 363

ἀγαπητός 258–65, 272–6 ἀποσκορακισμός 363
ἀγαπῶ 259, 400–1, 403–5 ἀποστέλλω 83, 183–96
ἀγαυρίαμα 362 ἀπὸ τοῦ νῦν 381
ἄγγελος 335 ἀργός 326
ἀγρός 331 ἀριστερός 307–8, 335
ἀδόκιμος 6 ἄριστον 82–3
ἀεί 303, 311, 365 ἄρκος 308–9, 311
ἀθετῶ 6 ἄρκος ἀπορουμένη 235–6
αἴνιγμα 325–6, 329 ἄρκτος 308–9, 311
αἰτῶ 83, 433 ἀροτήρ 365
αἰχμαλωτίζω 300 ἄρουρα 230 n. 23
ἀκμήν 299–300, 311 ἀρραβών 334
ἄλλος 85–6 ἄρρην 303, 317
ἄμφοδον 201–6 ἄρσην 303, 317
ἀμῶ 316, 365 ἀρτύματα 39
ἄν 49, 67, 100, 451–2, 453–4 ἀρχάριος 383
ἀνάβα 306–7 ἄρχω/-ομαι 83, 386
ἀναβάζω 382 ἄρωμα 213–6
ἀναβαίνω 90–1 ἀρωματίζω 215–6
ἀνακαινοτομῶ 157 ἀσπάλαθος 209–16
ἀνδρὸς ἀνδρός 231–2 ἄτοπος 69
ἄνθρωπος ἄνθρωπος 445 αὐθέντης 301–2
ἀνίσταμαι 310 αὐξάνω/-ομαι 84
ἄνισχυς 363 αὔξω 365
ἀνομία 98 αὐτός 157, 385
ἀντιπίπτω 6 ἀφίω/-ιῶ 382
ἀπαμαυρῶ 363
ἀπατῶ 329 βαθύφωνος 363
ἀπάτη 313 βάλλω 331, 386
ἀπιστέω/-ία/-ος 336 βάρβαρος 98
ἀπό 380 βᾶρις 204
ἀπόγραφος 158 βάσανος 5–6
ἀποκρίνομαι 39 βασίλειον 204
ἀπὸ λιβός 230 n. 23 βασίλισσα 316
ἀποσκευή 11–20, 418 βελόνη 315
484 INDEX OF GREEK WORDS
-βηθι 306–7 δικαίωμα 30

βιάζομαι 413 δίκην 157
βίος 415–7 διχόγνωμος 158
βιώσομαι 40, 42–3 δράξ 383
βλέπω 94 δύναμις 22 n. 4
βόλβιτον 39 δύο 332
βουλεύομαι βουλήν 368 δυοῖν 298–9
βούλομαι 365, 381, 413 δυσί 298–9
βοῦς 333 δύσκολος 325, 328
βοῶ 69
βράζω 383 ἑαυτόν 383
βρέχω 309 ἐβίων 43 n. 29
ἐγείρομαι 310
γαμῶ 86 n. 13 ἐγκολλῶ 6
γε 49 ἔδομαι 316, 413
γεμίζω 380 ἔζησα 43 n. 29
γέμω 38–9 εἰ 93, 368
γεοῦχος 383 εἰμι 385
για 382 εἰρήνη 29
γίνομαι 97 εἷς 332, 385
γνωμόνιον 157 εἰς 101, 301
γνωρίζω 380 εἴσεται 303
γνώσεται 303 εἰς τὰ ἴδια 386
γραικοτύπωσις 158 εἴσω 300–1
γραΐς 383 εἶχον 385
γραμματεύς 370 n. 61 ἕκαστος 448
γραμματικός/-οί 370 n. 61, 371 ἑκάστοτε 303
ἑκατόν 331
δακρύω 69 ἐκβάλλω 97
δάκτυλος 325, 328 ἐκδέχομαι 28–9
δέδοικα 366 ἐκεῖ 301
δεῖπνον 82–3 ἐκεῖσε 301
δεσπότης 326, 333 ἐκκαίω 2–3
δεῦρο 305–6 ἐκκλησία 337
δέω 333 ἐκκρούω 6
δή 364 ἐκλογή 320, 323
δηλαυγῶς 333 ἐκπνέω 69
δήπου 315 ἔκπρακτος 218
διά 382 ἐλπίζω 83
διαθήκη 31 n. 29 ἐλπίς 83
διὰ παντός 303 ἐμπρησμός 300
διαστέλλω 136 ἐμπυρισμός 6, 300
διαστολή 136 ἐν 231 n. 25, 301
διὰ τί; 95 ἐν ᾅδου 362
διάφορος 39 ἔνδον 300–1
διαφορότης 303 ἐνδόσιμον 158
δίδω/-ῶ 382 ἐν ἐμοί 233–4
δίδωμι 5 ἔνθεν ἑλών 281
INDEX OF GREEK WORDS 485
ἐντέλλομαι 109–16 ζῶ 94
ἐντυπῶ 158 ζωή 416
ἐντύπωσις 158
ἐξαποστέλλω 183–96 ἡγοῦμαι 94, 98
ἐξηγοῦμαι 109 n. 12 ἤθη 305
ἕξις 49 ἥκω 86
ἔξωθεν 67, 329 ἥλιος 333
ἐπαίρω 383 ἡμέραν ἐξ ἡμέρας 387
ἐπάνω 380, 385 ἤν 291
ἐπί 427–8 ἤρεσε 303
ἐπιθυμῶ 140 ἦς 317
ἐπιούσης/-ῃ 308 ἦσθα 317
ἐπίπεμτος 3 ἡττῶμαι 367
ἐπίσταμαι 366
ἐπιτάσσω 109 n. 12 θάλασσα 228, 230, 325, 328
ἐργασία 157 θέλετε 255
ἐργολαβία 158 θέλησον/-ήσατε 242–3, 253–4
ἐρεθίζω 34 θέλω 91–3, 237–55, 378, 381, 384–5
ἐρυθρός 366 θεωρία 383
ἔρχομαι 413 θεωρῶ 380
ἐρωτῆσαι εἰς εἰρήνην 34–5 θησαυρός 335
ἐσθίω 84, 311, 413–4 θλίβω 381
ἐσθίω τὸν ἑαυτοῦ ἄρτον 84, 85 θυμός 420–1
ἔσο 382 θύραι 67
ἐσχάτως ἔχω 314
ἔσωθεν 67 ἴδιος 334
ἕτερος 86 ἰδού 386, 447
εὖ 56–8 ἱκανός 34 n. 34
εὐδία 326, 329 ἱλαρός 333
εὐίλατος 399–400, 418 ἱμάς 333
εὐλογία 83 ἵνα 100, 385–6
εὐπορῶ 77 n. 18 ἴσασι 315
εὑρίσκομαι 76, 380 ἱσπανία 158
εὑρίσκω 71–8, 94, 380 ἴστε 315
εὐρυθμία 157
εὐώνυμος 307–8, 311 καθαίρω 329
ἔφαγον 413 καθάπαξ 383
ἐφέτος 386 καθάπερ 316
ἔχω 77, 243, 385 καθαρός 329, 335
ἔχω πρᾶγμα 386 καθιζάνω 422
καθό 316
Ζεύς 326 καθότι 316
ζῆλος 169–70 καὶ γάρ 364
ζήλωσις 169–70 καινοτομῶ 157
ζηλωτής 169–70 καίτοι 422–4
ζῆν 37–46 κάλαμος 214, 331
ζήσω/-ομαι 37–46 καλός 226
ζητῶ 91 καλῶς 56–8
κάμινος 410–1 μεθεκτέον 291

κάνθαρος 6–7 μελανός 382
κατάβα 306–7 μέλλω 242
καταβαίνω 95 μέμφομαι 323–4
κατάκλιτος 363 μέν 47–52, 97 n. 27, 364
καταλέγω 320 μετά 434
καταστρώννυμι 25–6 μετάβα 306–7
κάτεργον 134–5 μεταδετέον 410
κέλευσον 386 μετάνοια 381
κελεύω 109–16 μετανοῶ 381
κεραία 156 μεταξύ 316
κερδαίνω 381 μή 95
κῆνσος 334 μηδαμῶς 327
κλαίω 69, 412–3 μὴ θέλε 255
κλεῖν 306 μὴ θελήσῃς/-ήσητε 242–3, 254
κοινόν 302–3 μὴ οὐ; 386
κολλήγιον 380 μήτηρ 331
κόλος 325 μήτι 95
κόλπος 3–4 μιμητής 169–70
κοράσιον 311, 314, 325 μισθοφορία 384
κορίδιον 311 μῶμος 323–4
κράβαττος 286, 311, 334
κράζω 69 νάρδος 334
κραυγάζω 325 νερόν 384
κρούω 383 νῆστις 326
κῦδος 362 νίπτω 327
κυνῶ 404 n. 7 νουθετέω 421
κύω 363 νοῦς 366
κῶλον 5, 230 n. 22 νύξ 32
κωλύω 325 νώ 303
κῶμος 325, 329
ξέστης 334
λαγύνιον 384 ξύν 280, 284
λαλητός 219
λαμβάνω 413, 433 οἶδας 303
λαμπάδιον 135–6 οἴδασι 315
λέγω 87–9, 109 n. 12, 320 οἶμαι 366
ληκύνθιον 382 οἶσθα 303
λογικός 421 ὅλως 386
λόγιος 320 ὁμιλία 23 n. 5
λόγοι 289 n. 36 ὁμιλῶ 23 n. 5
λόγος 133–4 ὁμίχλη 363
λοιπόν 386 ὄνομα 339–48, 387
λούω/-ομαι 84 ὀπίσω 385
λυπῶ/-οῦμαι 84 ὀπωροφυλάκιον 9
ὁρῶ 67, 326
μακρόβιος 363 ὅς, ἥ, ὅ 332, 444
μέγιστος 315 ὅστις 95
INDEX OF GREEK WORDS 487
ὅτι; 95 πράσσω 366

οὐαί 67 πριστηροειδής 363
οὐ μή 63–7, 381–2 προκοπή 299
οὖν 67–8, 364 προκόπτειν 299
ὄφελον 97 προσήλυτος 30, 235
ὀψάριον 419 προσπορεύομαι 24
ὄψις 326 προστάσσω 109–16
πρόσωπον 326
παιδίον 135 προφέρω 99
παλάθη 208 πρωτοποίητος 157
πάλαθος 208–221 πυργόβαρις 204
παλαιῶ 156 πύργος 204
πανταχόθεν 157 πυρίκαυστος 363
παντοκράτωρ 31, 32, 36 πύρωσις 6
πάντοτε 303, 311 πῶς 327
παρά 95, 428–39 πῶς ὁ κόσμος; 387
παρακαλῶ 86, 89–90
παραπικραίνω 33–4, 36 ῥαβδίσκος 156
παρατείνω 8 ῥαφίς 315
παραρρέω 361 ῥιγάω 382
παρρησία 313–4 ῥύμη 316
παροξύνω 34
παρουσία 313–4 σαλός 384
πᾶς 332 σελήνη 335
πάσχω 331 σκελίζω 7–8
πατήρ 30–1, 331 σκίμπους 286 n. 25
πείθω 98 σουδάριον 335
πέμπω 183–96 σπάλαθος 217
περίκειμαι 99 σπλαγχνίζομαι 384
περιψήχω 219–20 σπουδαῖε 156
περνῶ 384 στέργω 259
πιστός 337 συγγηρῶ 220
πλατύνω 9 συγγίνομαι 22–3, 26–7
πνοή 6 σύγκρισις 157
ποῖ; 301 συγχωρῶ 381
ποῖον; 384 συλῶ 384
ποιῶ 97 n. 27 σύν 280
ποιῶ ἔλεος 381 σύνειμι 23 n. 5
πολλά 95–6 συνομιλῶ 158
πολύλαλος 219 συνουσία 23 n. 5
πολύς 96, 100 συντάσσω 109–16
πόρρω 366 συντείνω 157
πόρρωθεν 366 συντελῶ 157
ποταμός 325 σωθείης 387
ποτίζω 397–8 σωτήριος 6
ποῦ 301, 326
ποῦ ἔνι; 387 τάλας 360
πούς 331 ταχέως 380
τε 315, 364 ὑπομονή 421–2

τέκνον 320, 323 ὑποσκελίζω 8
τελῶ 384 ὕω 309, 327
τέμαχος 299
τέρψις 313 φαγίον 384
τέσσαρες 333 φάγοι 59–60
τί; 7 φάγομαι 316, 413
τίθω/-ῶ 382 φέρω 331
τίκτω 320, 323 φεύγω 368
τίποτε 384 φιλῶ 403–5
τίς 331 φράσις 158
τις 447–51 φύσει 387
τό 385
τοιγαροῦν 315, 364 χαίρειν 233
τοίνυν 315, 364 χάρτης 325, 329, 336
τόκος 82 χείρ 332
τόπον ἐκ τόπου 387 χρή 97, 315
τρεῖς 332 χρυσός 335
τρώγω 306, 311, 414–5
τυγχάνω 232–3 ψιχίον 384
τυπῶ 158 ψύλλα 296–7
ψύλλος 296–7
ὕβρις 98 ψυχή 1–2
ὕδωρ 327
ὑετός 327 ὦ 60–3
υἱός 5 ὤ 61, 62–3
ὕπαγε 315 ὡς, ὡς ἄν 367
ὑπάγω 98
‫‪INDEX OF HEBREW WORDS‬‬
‫א‬ ‫מ‬
‫‪ 30–1‬אב‬ ‫‪ 430, 438‬מאת‬
‫‪ 360‬אוי‬ ‫‪ 430‬מן‬
‫‪ 232, 445–6, 448‬אישׁ‬ ‫‪ 232‬מצא‬
‫‪ 445‬אישׁ אישׁ‬ ‫‪ 33–4‬מרה‬
‫‪ 446‬אישׁ כאחיו‬ ‫‪ 33‬מרר‬
‫‪ 428‬אצל‬
‫‪ 201, 204–5‬ארמון‬ ‫נ‬
‫‪ 230 n. 23‬אשׁל‬ ‫‪ 364–5‬נא‬
‫‪ 443‬אשׁר‬ ‫‪ 230 n. 23‬נגבה‬
‫‪ 439‬את‬ ‫‪ 2‬נפשׁ‬
‫‪ 5‬נתן‬
‫ב‬
‫‪ 233‬בי אדוני‬ ‫ס‬
‫‪ 5‬בן‬ ‫‪ 362‬סבל‬
‫‪ 2‬בער‬ ‫‪ 370 n. 61‬ספר‬
‫‪ 31 n. 29, 234 n. 35‬ברית‬
‫‪ 216‬בשׂם‬ ‫ע‬
‫‪ 8‬בתרון‬ ‫‪ 430, 439‬עם‬
‫‪ 368‬עצו עצה‬
‫ג‬
‫‪ 364‬גם‬ ‫פ‬
‫‪ 5‬פגר‬
‫ה‬
‫היה‬ ‫‪23‬‬ ‫צ‬
‫‪ 31, 32‬צבאות‬
‫ז‬ ‫‪ 109‬צוה‬
‫זה אל־זה‬ ‫‪356‬‬ ‫‪ 115‬צוה מצוה‬
‫‪ 3–4‬צלחת‬
‫ט‬
‫‪ 15–7, 19, 20‬טף‬ ‫ק‬
‫‪ 7–8‬קלע‬
‫י‬ ‫‪ 24‬קרב‬
‫‪ 27‬ידע‬
‫‪ 228‬ים‬ ‫ר‬
‫‪ 232‬יסף‬ ‫רכושׁ‬ ‫‪16, 19‬‬
‫כ‬ ‫שׁ‬
‫‪ 5–6‬כלמה‬ ‫‪ 358‬שׁבר‬
‫‪ 33‬כעס‬ ‫‪ 25‬שׁחט‬
‫‪ 6–7‬כפיס‬ ‫‪ (Aram.) 71‬שׁכח‬
‫‪ 29, 35, 233‬שׁלום‬
‫ל‬ ‫‪ 188 n. 15‬שׁלח‬
‫‪ 22–3‬להיות‬ ‫‪ 339–40, 345‬שׁם‬
INDEX OF BIBLICAL REFERENCES
References to main discussions are given. Mentions in footnotes are not noted unless
significant. The list of LXX-Pent. examples on p. 451 is not indexed.
Genesis 1:17 110

2:6 398 2:13 308
2:19 340, 348 2:14 241, 251, 255
3:1 64 3:13 340, 341
3:11 110 3:15 347
4:19 340 7:6 110
6:5 450 9:18 69
8:10–12 232 10:8–11 14
11:10 5 10:24 15
12:5 2 12:35 110
12:20 110 12:37 14, 20
13:14 228, 230 15:3 339, 340
14:12 16 16:4 69
15:11 5 n. 16:20 449
15:14 14 18:3 340
18:2 380 18:19 134
19:5 26 20:5 64
19:11 443 20:17 140
22:7 135 21:7 450
22:8 135 22:5(4) 449
24:23 437 n. 22:9(8) 447 n.
28:1 110 22:20 30
29:20 432 23:21 341
30:28 136 24:14 27 n.
31:28 367 30:16 134
32:27(28) 341 30:20 24
36:37 432 30:23 214, 216
38:4 341 30:34 214
39:10 22 31:17 434
40:3, 7 437 33:8 453
42:2 192 33:21 434
43:8 14, 20 35:21 134
43:16 443 36:2 24
43:29 135 39:22 18
44:32 28
46:5 17 Leviticus
49:13 431 5:4 136
5:16 3
Exodus 6:40(7:10) 446
1:15 341 13:54 111
492 INDEX OF BIBLICAL REFERENCES
15:2 232 n. 3 Kingdoms

17:3 445 3:17, 26 234
20:2 232 n. 20(21):21 2
25:17 446
25:29 450 4 Kingdoms
18:32 41
Numbers
2:34 111 2 Chronicles
5:12 231 20:25 14
5:20, 29 27 n.
11:3 341 Psalms
14:16 25 98:8 399
14:29 5 137:7 41
16:27 17 142:11 41
19:2 136
25:8 411 Proverbs
25:14, 15 341 19:21(24) 3
30:7 136 24:43(28) 9
32:26 18 26:15 3
Deuteronomy Song of Songs

5:21 140 4:13 5
10:12 433
18:3 433 Isaiah
19:5 232 1:7+ 363
20:13–1 16 1:11 365
22:20 134 2:16 356
27:1 110 3:1 364
31:27 34 3:9+ 368
32:16 33 3:10 364
3:23 363
Joshua
5:26 364
1:7 367
6:2 364
6:3 355
Judges
6:5 360
1:23 341
8:9 367
6:13, 15 234
10:3 366
13:8 234
10:7 366
18:29 341
11:9 367
Ruth 14:25 362
2:1, 19 341 17:5 365
17:13 363, 366
1 Kingdoms 21:1 367
1:26 233, 234 23:10 364
2:1–10 191 29:11 366
29:18 363
2 Kingdoms 30:8 364
2:29 8 30:13–14 357
INDEX OF BIBLICAL REFERENCES 493
33:19 363 Habakkuk

37:12 364 2:11 6
37:30 365
38:18 362 ———
40:25 364
1 Esdras
40:30 363
8:53 399
41:7 364
41:15 363 Tobit
42:8 341 3:10 250, 263
42:14 365 6:15 250
43:14 368 10:12 263
44:4 361
44:18 363 Judith
48:22 233 7:2 19
49:15 368
51:13 365 Wisdom
51:15 341 12:6 301
53:10 363
57:6 364 Sirach
57:8 366 6:35 255
57:10 366 7:13 255
57:21 233 10:8 218
59:4, 13 363 11:16 220
60:14 366 18:33 219
61:5 365 21:25 219
61:11 365 24:15 209, 212–6
62:7 362 30:7 219
63:2 366 41:19 363
66:15 363
66:20 367 Baruch
3:14 363
Jeremiah
10:18 7 1 Maccabees
17:27 201, 204 10:33 2
23:12 8
2 Maccabees
30:16 201, 204
12:21 19
39:29 33
———
Ezekiel
16:52, 54 5 Matthew
32:24, 30 5 4:3 88
5:34 88
Hosea 6:13 169
12:12 231 n. 9:14 96
13:8 235 10:2 345
13:22 313
Micah 15:16 300
3:12 9 15:27 384
494 INDEX OF BIBLICAL REFERENCES
15:28 62 1:27 345

17:17 62 1:53 191
21:19 60 1:63 347
25:21+ 317 6:7 71
26:14 92 9:31 88
26:15 248 9:41 62
26:53 90 11:7 67
26:69 317 11:54 71
12:15 67
Mark 13:24 71
1:16 95 16:23 67
1:44+ 67 20:10, 11 191
2:27 382 23:22 69
4:4 50 23:41 69
4:19 313 23:46 69
5:9+ 345 24:25 62
5:10 96 24:38 90
5:14 68 24:49 194
6:48 92, 248
7:4 380 John
7:11 67 1:6+ 345
7:15 67, 68 1:43 249
7:18 67, 68 4:33+ 95
7:20 68 6:21 249
7:21, 23 67 8:6 77
7:28 384 11:35 69
8:12 93 11:54 314
9:12 51 13:22 94
9:19 62 16:19 91
10:32 68 16:32 386
11:14 59 18:17, 25 95 n.
12:2 59 n. 21:15–17 403–5
12:5 51
12:44 68 Acts
13:20 67 1:1 62
13:29 67 7:12 192
14:7 57 7:26 308
14:21 51 8:26 95
14:38 51 8:40 380
14:67 317 9:30 193
15:34 69 9:36 88
15:37, 39 69 11:22 193
16:8 194 12:11 193
16:19 51 13:6 345
13:26 192
Luke 14:13 249
1:1–4 283 15:29 57
1:26+ 345 16:11+ 308
INDEX OF BIBLICAL REFERENCES 495
17:14 193 2 Thessalonians

17:28 284 2:2 380
19:2 93
19:33 249 1 Timothy
21:3 308 n. 5:22 380
22:21 192 6:2 89
24:19 59
26:4 315 Titus
26:7 312 2:15 89
26:29 59
Romans Hebrews
8:3 193 2:16 315
12:1 421 12:17 315
1 Corinthians James
5:1 387 3:3 315
9:4 386 5:4 316
10:5 26
15:29 387 2 Peter
15:33 284, 305 1:4 315
2:13 313
Galatians
1:6 380
Revelation
4:4, 6 193
6:8 345
Ephesians 9:11 345
6:3 57 10:2 308 n.
13:2 309
Colossians 18:13 2
4:6 382
INDEX OF MODERN AUTHORS
Abbott-Smith, G. 320–1, 334 n., 336–7 Bauer, H. 177 n.

Adams, J. N. 169 n., 230 n., 234 n. Bauer, W. 320 n.; see also BAG, BAGD,
Adams, S. A. 316 n. BDAG, BRAA
Adler, A. 208 n., 211 Bauer2 (1928) IX, 121
Adrados, F. R. 239 n.; see also DGE Baumgartner, W. see HALOT; KB, Lexicon
Aejmelaeus, A. V, XV, 207 n., 230 n. BDAG IX, 6 n., 30, 72 n., 75 n., 121,
Aitken, J. K. V, 35 n., 70, 127, 197 n., 126 n., 127, 139 n., 156–8, 184, 185 n.,
255, 359 n., 399 n., 400 n., 418 n., 428 n., 187 n., 202, 240 n., 267 n., 268 n.,
446 n. 306 n., 315 n., 322, 332 n., 381 n.,
Aland, B. 145 n.; see also BRAA 386 n., 401, 420 n.
Aland, K. 50 n., 145 n.; see also BRAA BDB IX, 7, 8, 15, 23 n., 204
Alimeni, G. 179 n. BDF IX, 40 n., 43 n., 50 n., 51 n., 54 n.,
Allen, W. C. 69 n. 56 n., 57 n., 58 nn., 59 nn., 61 n., 65 n.,
Andersen, F. I. 234 n. 65 n., 95 n., 284 nn., 307 n., 312 n., 315,
Anderson, G. 294 n., 296 n., 312 n. 316 n.
Anz, H. 111 n., 184 n., 390 BDR IX, 59 n., 61 n., 307 n.
Argyle, S. 308 nn. Beek, L. van IX
Argyriades, G. P. 149 n. Beekes, R. IX, 125 n., 321, 335–6, 363 n.
Arndt, W. F. see BAG, BAGD, BDAG Beentjes, P. C. 216 n.
Arnold, B. T. 190 n., 192 n. Bekker, I. 38 nn., 39 nn., 298 n., 304,
Arnott, W. G. 69 n., 304 n. 305 nn.
Arnold, W. R. 8 n. Bentein, K. 70, 197 n.
Arribas Arranz, F. 179 n. Bentley, J. H. 145 n., 165, 166, 167,
Auwers, J.-M. 209 nn. 170 n. 37, 181
Bérard, F. 395 n.
Babiniotes, G. D. 195 n. Bethe, E. 311 n.
Baer, D. A. 351, 361 n., 367 n. Betz, H. D. 193 n.
BAG IX, 25 n., 30, 72 n. BHS 26 n.
BAGD 40 n., 43 n., 50 n., 51 n., 54 n., Bihlmeyer, K. 194 n.
57 n., 61 n., 64 n., 72 n., 75, 84 n., 82–6, Black, M. 71
88–95, 99, 102–3 Blass, F. 62 n.; see also BDF, BDR, KB,
Bagnall, R. S. 393 nn., 394 n., 396, 397 n. Grammatik
Bailly, A. 409, 423 Blomqvist, J. 47 n., 364 nn., 423, 424 n.
Balz, H. see EDNT BM IX, 26 n.
Barber, E. A. see LSJ Suppl. Boisacq, É. 321, 333–4
Bardy, G. 282 n. Boling, R. G. 27 n.
Barker, N. 149 n. Bolinger, D. 52 n.
Barr, J. 7 n., 9, 233 n., 353 n., 354–5 Boned Colera, P. see RBLG
Barrett, C. K. 95 n., 190 n., 403 Bons, E. 235; see also LXX.H, 3
Barthélemy, D. 41 nn. Bowie, E. L. 294 n.
Bastianini, G. 185 n. Boyd-Taylor, C. 136, 137 n., 228 n.,
Bataillon, M. 165, 166 n., 168 n., 181 229 nn., 236, 354 n., 369 n.
498 INDEX OF MODERN AUTHORS
BRAA 72 n. 4, 75 n. 11, 127 Christchev, Th. 169 n.

Brandwood, L. 42 n. Chryssavgis, Fr M. 387 n.
Bratcher, R. G. 92 n. Clackson, J. 343 n.
Braun, A. 239 n. Clarysse, W. 239 n.
Breitenstein, U. 61 n. Clines, D. J. A., see DCH
Brenton, L. C. L. IX, 23 n. Cobet, C. G. 43 n.
Brock, S. P. 26 n., 106 Cohn, L. 270 n., 307 n., 310 n.
Briggs, C. A., see BDB Coles, R. A. 247 n.
Brooke, A. E. see BM Collard C. 247 n.
Brown, F., see BDB Connolly, A. L. 294 n.
Brown, R. E. 403 n., 405 n. Contopoulos, N. 156–8
Browne, G. M. 75 Costache, D. 279 n. 1, 290
Browning, R. 40 n., 53 n., 54 n., 58 n., Cotelier, J.-B. 375
238, 239, 285 n., 286 n., 290 n., 294 n., Cowley, A. E., see GKC
296 n., 313 n., 346 n., 374 n., 377 nn., Cox, C. E. 422 n.
378 n., 385 nn., 396 Crastonus, J. 210 n., 211
Bruce, F. F. 193 n. Cribiore, R. 286 n., 187 nn., 294 n.
Büchner, D. L. 200 n. Cross, F. M. 41 n.
Buck, C. D. 113 n., 202; see also Buck- Cullen, M. 178 n.
Petersen Cunningham, I. C. 208 n.
Buck-Petersen 157, 217 n., 363 n., 382 n. Curran, M. 144 n., 189 n., 207 n., 293 n.,
Burguière, P. 239 n. 319 n., 342 n., 369 n., 407 n.
Burney, C. F. 345 n. Curtius, G. 321 n.
Burridge, R. A. 283 n.
Butler, Dom C. 374 n. Dain, A. 308 nn.
Daniel, S. 6 n.
Cadbury, H. J. 50 n., 69 n., 184 n., Danker, F. W. 226 n., 234 n., 320 n., 321,
192 n., 314 nn., 315 nn. 330–7, 391; see also BAGD, BDAG
Cadell, H. 108 n., 392, 398 n. Dardano, M. 179 n.
Caird, G. B. 135 n., 136, 204 n., 205 n., Daris, S. 203 n.
411 n. Darlow, M. 145 n., 170 n.
Cañas Reíllo, J. M., see BGS DCH 20, 32 n., 204
Cannan, C 265 n. De Borries, J. 298 n.
Carroll, S. 52 n. Debrunner, A. 39 n., 58 n.; see also BDF,
Carson, D. A. 337 n. BDR; Schwyzer, Grammatik
Case, S. J. 205 n. Decker, R. J. 66 n.
Cavallo, G. 150 n. Deferrari, R. J. 281 n., 282 n., 288 n.,
CGL IX, 126, 278 289 n., 291 n.
Chadwick, J. 9, 106, 119–20, 139 n., 187 n., Deissmann, A. 4 n., 106, 390–1
258–260, 263, 265 n., 267 n., 268, 269 n., De Jonge, H. J. 250 n.
276 n., 277, 278, 410 n., 414, 417 n., Denniston, J. D. 50 n., 51 n., 422–4
420 n. Depauw, M. 272 n.
Chantraine, P. X, 30, 56 n., 125 n., 213 n., De Rossi, F. Canali 401 n.
217 n., 275 n., 297 n., 308 n., 321, 323 De Sousa, R. F. 351, 354 n., 370 n.
343 n., 416 n. De Troyer, K. V
Charles, R. H. 345 n. DGE X, 120, 124, 136, 156–8, 184, 186 n.,
Chatzidimos, A. 166 n. 202, 240, 274 n., 278, 388 n., 407, 409 n.
CHB 145 n. DGENT X, 127
INDEX OF MODERN AUTHORS 499
Dhont, M. 35 n. Fraser, P. M. X, 297 n.

Dickey, E. 298 n., 302 n., 304 n., 307 nn., French, R. 298 n.
310 nn., 311 n. Frisk, H. X, 125 n., 321
Diego Sigeo 173 n. Frösén, J. 54 n., 55 n., 295 n.
Diels, H. 270 Funk, R. W., see BDF
Dieterich, K. 239 Furguson, R. A. 293 n.
Diggle, J. IX
Dihle, A. 301 Gaisford, T. 208 n., 211, 219 n.
Di Lella, A. A. 212 n., 214, 215 n., 220 n. Gallay, P. 282 n.
Dines, J. M. 226 n. García Santos, A. A. 127
DMLBS X, 139, 169 n., 408 Garlan, Y. 203 n.
Dogniez, C. 233 n., 398 n. GE X, 126, 202 n., 278, 407–26
Dorival, G. 109 n., 136, 187 n., 411 n.; Geanakoplos, D. J. 145, 149 n., 150 nn.,
see also HDM, La Bible grecque 160 n., 165, 166 n., 167, 168, 173 n., 181 n.
Doukas, D. 150 n., 153, 158, 164–82 Gehman, H. S. 230 n.
Driver, G. R. 7, 9 Gerleman, G. 367 n.
Driver, S. R., see BDB Gerth, B. see KG, Grammatik
Du Bouchet, J. 202, 204 n., 205 n. Ghisalberti, A. M. X
GI X, 126, 202, 278, 407–26
Ebeling, H. 321 n. Giannakis, G. K., see EAGLL
Edmonds, J. M. 40 n., 274 n. Giannakopoulos, I. 214, 368 n.
EDNT X, 193 n. Gignac, F. T. 60 n., 239, 382 nn.
Eissfeldt, O. 7 n. Gingrich, F. W. 320, 396; see also BAG,
Elliott, J. K. 37 n., 51 n., 315 n., 317 BAGD, BDAG
Ellis, J. 52 n. 22 Glare, P. G. W. 410 n.; see also LSJ Suppl.;
Epp, E. J. 317 n. OLD
Erasmus, Desiderius 178 n. Glaser, O. 184 n.
Erbse, H. 307 n. Goetz, G. 219 n.
Evans, T. V. V, 58 n., 105 n., 144 n., 237 n., Goh, M. 409, 426 n.
241, 342 n., 367 n., 368 n., 373 n., 378 n., Gomme, A. W. 274 n.
392, 401 n., 403 n., 411 n., 428 n., 435 n. Goodspeed, E. J. 42 n.
Exler, F. X. J. 49 n. Gray, G. B. 26 n., 411 n.
Eynikel, E., see LEH, Lexicon Greenlee, J. H. 323–7
Greenslade, S. L. X
Fabricius, C. 281 n. Gregory, C. R. 145 n.
Farmer, W. R. 50 n. Gregory, M. 52 nn.
Fausto, V. 166 Gundermann, G. 219 n.
Fee, G. D. 37 n., 284 n., 317 n. Guy, J.-C. 374 n., 375–6, 377 n., 384 n.
Fernández-Armesto, F. 170 n.
Festugière, A.-J. 378 n. Haenchen, E. 57 n., 65 n., 190 n., 312
Feyerabend, K. 269 n. Hahn, E. A. 343 n.
Fischer, E. 298 n., 300 n., 304 n. Hall, B. 145, 159 n.
Fitzmyer, J. A. 71, 72, 190 nn., 191 n. HALOT X, 32 n., 204
Fleischman, S. 239 n. Hamer, E. 150 n.
Fountotos, J. 404 n. Hanhart, R. XII
Fox, W. 239 n. Hansen, D. U. 302, 307 nn., 308 n., 310 n.
Frame, J. E. 65 n. Harl, M. 112 n., 228 n., 233 n., 236, 398 n.;
Fränkel, H. 30 see also HDM, La Bible grecque
Harris, G. 50 n. Johannessohn, M. 61 n., 428 n.

Harrison, N. V. 290 Johnson, A. F. 178 nn.
Hart, J. H. A. 210 n. Joly, R. 265 n., 400 n., 404, 405 nn.
Hatch, E. see HR, HR Suppl. Jonge, M. De 46, 165 n., 170 n., 181 n.
Hatzidakis, G. N. 238 n., 295 n. Joosten, J. 236, 367 n.; see also LXX.H, 3
Hauspie, K. 10; see also LEH, Lexicon Joseph, B. D. 238, 238 n., 239
HDM, La Bible grecque X, 106 n., 107 n., Jouguet, P. 203 n.
108 n., 392 n. Joüon, P. 177 n., 239
Helbing, R. 24 n. Joyal, M. 288 n.
Hemer, C. J. 391 Judge, E. A. 169 n., 279 n., 283 n.
Henderson, J. 278
Hennig, D. 203 n. Kalogerakis, A. 404 n.
Herkenne, H. 216 Karalis, V. 144 n.
Hernitschek, E. V Kariatlis, P. 279 n.
Hickie, W. J. 320 Karrer, M. M. see also LXX.D, LXX.D
Hill, C. 178 n. 47 Komm.
Hock, H. H. 238 n. Katz, P. 41 n., 278, 352
Hohlwein, N. 203 n. Kauhanen, T. 441 n.
Hollander, H. W. 169, 250 n. Kautzsch, E., see GKC
Holleaux, M. 11 n., 12 Kazazis, J. N. 285 n., 294 n., 295 n., 296 n.,
Holton, D. 145 n., 147 n., 148 n., 238 n. 307 n., 312 n.
Homza, L. A. 170 n., 176 n., 181 n. KB, Grammatik X, 40 n.
Hopper, P. J. 237 n., 240 n., 243 n. KB, Lexicon X, 4 n., 7, 8, 23 n., 204
Horrocks, G. C. 157, 187 n., 238 n., 239, Kefallinos, E. 149 n.
294 n., 295 n., 312 nn., 377 nn., 378 nn., KG, Grammatik X, 50 n., 63 n., 339 n.
385 nn., 396, 448 n., 450 n. Kieffer, R. 37 n.
Horsley, G. H. R. IX, 122, 126, 391, Kiessling, E. 11 n. 13 n.; see also Preisigke,
395 n.; see also New Documents Wörterbuch
Hort, F. J. A. 347 Kilpatrick, G. D. 37–40, 46, 54 n., 54 n.,
Howard, W. F. see MH, Grammar, II 55 n., 56 n., 59 n., 305 n., 315 n., 317,
Howlett, D. R. 169 n.; see also DMLBS 318, 324
HR X, 64 n., 209 n. Kim, L. 312 n.
HR Suppl. X, 209 n. Kinzig, W. 282 n., 284 nn.
Hudson, R. A. 293 n. KJV 95 n., 219 nn., 220 n.
Hugo, P. 371 n. Klamp, D. 203 n.
Humbert, J. 60 n., 76 n. Koehler, L. see HALOT, KB, Lexicon
Hunkin, J. W. 386 n. Kolitsaras, I. Th. 195 n., 214, 368 n.
Kooij, A. van der 351, 352 n., 354 n., 357,
Imholtz, A. A. 268 n., 410 n. 370–1
Kotsevalov, A. 294 n.
Jacques, X. 324 Kraus, W. see LXX.D, LXX.D Komm.
Jaeger, W. 290 n. Kreuzer, S. 428 n.
James, A. W. 30 Kriaras, E. XI, 123, 195 n., 208 n., 213 n.,
James, P. 265 n. 215 n., 217 n., 388
Jannaris, A. N. 48 n., 62 n., 238 n. Kühner, R. see KB, Grammatik; KG,
Janse, M. 347 Grammatik
Jellicoe, S. 4 n., 107, 212 n.
Jeremias, J. 65 n. Ladas, G. G. 166 n.
Jobes, K. H. 106 nn., 107 n., 236 Lagrange, M.-J. 65 n., 66 n.
Lampe, PGL XI, 30, 75 n., 123, 156–8, LS5 (1861) 201
184, 202 n., 218 n., 219, 239, 263 n., LS8 (1897) 135 n., 261–2
289 n., 377, 383 n., 388, 421 n., 422 LSJ 6 n., 23 n., 24 n., 25 n., 27 n., 40 n.,
Lane, W. L. 60 n. 43 n., 63 n., 72, 74 n., 76 n., 111 n., 119–
La Roi, E. 197 n., 255 20, 124, 135 n., 136, 156–8, 184, 200,
Lascaris, C. 149 n. 201, 213 n., 219, 220 n., 239, 257–78,
Latham, R. E. see DMLBS 299 n., 321, 380 n., 388, 396, 407–25
LBG XI, 123 LSJ Suppl. (1968) 1–10, 17 n., 119
Leander, P. 177 n. LSJ Suppl. (1996) 10, 119, 201, 263 n.,
Le Boulluec, A. 109 n., 135 321 n., 398 n., 411 n., 410, 418, 419,
Lee, J. A. L. passim 421 n.
Lee, K. H. 247 n. Lust, J. see LEH, Lexicon
Legg, S. C. E. 50 n., 51 n., 59 n. LXX.D XI, 214 n., 368 n.
Legrand, É. 150 n., 153, 154, 157, LXX.D Komm. XI, 360 n.
165 n., 166 n., 168 n., 169 n., 171 n., LXX.H, 3 XI
173 nn. Lyell, J. P. R. 159
LEH, Lexicon XI, 106 n., 108, 109 n.,
121–2, 130, 133, 188 n., 201, 202, Maas, P. M. 281 n.
218 n., 235, 392 MacAlister, S. 64 n., 73 n.
Lemmelijn, B. 354 Maehler, H. 150 n.
Le Moigne, P. 357 Mahaffy, J. P. 390
Lewis and Short XI, 215 Maidhof, A. 302 n.
Lex. Acad. XI, 213 n., 215 n., 217 n. Mandilaras, B. G. 40 nn., 58 n., 158, 239,
Lex. Proia XI, 156–8, 195 n., 297 n., 294 n., 386 n.
404 n. Markopoulos, T. 244 n., 254
Liddell, H. G. see LS, LSJ, LSJ Suppl. Marshall, G. 258 n.
Lietzmann, H. 74 n. Marshall, I. H. 191 n., 192 n.
Lindars, B. 106 Martini, C. M. 37 n., 40 n.
Llewellyn Davies, A. 411 n. Masson, É. 334 n.
Llewelyn, S. R. see New Documents Mateos, J. see DGENT
Lloyd-Jones, H. 302 n. Matthews, E. see Fraser, LGPN
LN XI, 79–103, 112 n., 121, 127, 139 n., Mauersberger, A. 112 n.
187 n., 190 n., 240 n. Mayser, E. 48 n., 50 n., 56 n., 60 n., 63 n.,
Loader, J. A. 30 241 n., 339 n., 342 n., 396, 453 n.
Loader, W. 140 n. McGuire, M. R. P. 291 n.
Lobeck, Chr. A. 298 n., 300, 301 n. McLean, B. H. 393 n., 395 n., 396, 397 n.
Lolos, Y. 202 McLean, N. 106 n. 4; see also BM
Longenecker, R. N. 193 n. McNeile, A. H. 411 n.
Louth, A. 290 n. Meecham, H. G. 366 n.
Louw and Nida see LN Meer, M. N. van der 227 n., 351 n., 352,
Louw, J. P. 80 n., 81 n., 84 n., 85 n., 103, 359, 363 n., 367 n., 391 n., 392
240 n., 268 n., 337 n.; see also LN Mega Lexikon X, 120, 156–8, 184, 195 n.,
Louw, T. A. W. van der 355n., 357, 202, 218 n., 220 n., 378 n., 388
441 n., 442 n., 446 n. Melancthon, P. 178 n.
Lowry, M. 165 n., 173 n. Melandinos, D. 149
LS1 (1843) 259–60, 324, 424 Metzger, B. M. 144, 150 n., 174 n., 194 n.
LS1 Abridged 324 MH, Grammar, II XI, 54 n., 306 n., 307 n.
LS2 (1845) 273 n. MHT, Grammar, III XI, 22 n., 50 nn.,
LS4 (1855) 260–1 57 n., 58 n., 59 nn., 60 n., 61 nn., 63 n.,
95 n., 312 n., 314 n., 315 nn., 316 nn., NETS XII, 23 n., 214, 218 n., 219 n., 220 nn.,
345 nn. 226–34, 348, 368 n., 411 n.
MHT, Style XI, 54 n., 65 n. New Documents XII, 102 n., 203 n.
Migliorini, B. 179 Newman, B. M. 99, 121, 320, 321, 327–
Migne, J.-P., PG XII, 375, 377, 378 nn. 30, 336–7
Migne, J.-P., PL XII Ngunga, A. T. 351 n.
Mihevc-Gabrovec, E. 238 n. Nickau, K. 310 n.
Millar, F. 378 n. Nicolaus de Lyra 161
Milligan, G. 202, 391; see also MM Nida, E. A. 52 n., 80 n., 85 n., 92 n., 268 n.;
Minnen, P. van 393 n. see also LN
Mirambel, A. 239 n. NIV 95 n.
MM XII, 30, 56 n., 64 n., 66 n., 72 n., Norden, E. E. 284, 314 nn.
102–3, 107, 122, 184, 202, 334 n., 391, Norton, F. J. 161 n.
396 NRSV 95 n. 24, 219 nn., 220 n., 227
Moffitt, D. M. 235 Núñez de Guzmán, H. 181
Montanari, F. 407–9, 425–6; see also GE,
GI Oates, J. F. IX, 395 n.
Montevecchi, O. 107, 391, 392 Obbink, D. D. 237 n.
Morgan, T. 286 n. OCD 302 n., 311 n.
Morison, S. 149 n., 150 n. O’Connell, S. 210 nn., 216 n.
Morris, L. 403 n., 404 n. O’Connor, M. P. 20
Morwood, J. 269 n. OED XII, 139, 268, 271, 408
Moule, C. F. D. 54 n., 65 n. Ó Fearghail, F. 190 n.
Moule, H. F. 145 n., 170 n. OGELD XII, 156–8
Moulton, J. H. 391; see also Moulton, OLD XI, 139, 268, 408
Grammar, I; MH, Grammar, II; MHT, Olin, J. C. 170 n.
Grammar, III; MHT, Style; MM Olley, J. W. 359 n.
Moulton, Grammar, I XI, 40 n., 58 n., Orlandos, A. K. 202 n.
59 n., 60 n., 65 nn., 66 n., 196 n. O’Sullivan, J. N. 73 n.
Mousouros, M. 157, 158, 171 Ottley, R. R. 22 n., 352, 356 n., 360 n.,
Muellner, L. 409, 426 n. 362 n., 368 n.
Munnich, O. see HDM, La Bible grec-
que Page, D. L. 76 n.
Muraoka, T. 106 n., 108, 121, 127, 129– Palme, B. 393 n.
42, 177 n., 187 n., 202 n., 223–6, 356 n., Palmer, L. R. 69 n.
368 n., 392, 396, 398 n., 401, 420 n., 421, Pappas, P. A. 238 n.
427 n., 437 n., 453 n. Parke, H. W. 12 n.
Musaeus 158, 171 n. Parker, B. 23 n.
Muses, C. A. XII, 23 n. Parsons, P. 302 n.
Mussies, G. 60 nn. Paschos, P. B. 375
Passoni dell’Acqua, A. 392
NA27 347 Passow, F. 257, 260, 409, 424
NA28 XII, 77 n. Pauly-Wissowa, RE XII, 42 n.
Naber, S. A. 38 n., 304 n. Peláez, J. 103
Nagy, G. 409, 426 n. Pelletier, A. 112 n.
NEB 15 n., 95 n., 214 Pentzas, B. 375
Nebrija, A. de 167, 170, 178, 180–1 Perkins, L. J. 340
Nestle, E. 145 n. Perry, B. E. 73 n.
Petersen, W., see Buck-Petersen Rutherford, W. G. 298 n.

Philippaki-Warburton, I. 145 n. RV 411
Pietersma, A. 5 n., 10, 226–30, 371 n.; Rydbeck, L. 54 nn.
see also NETS
Pinkster, H. 239 n. Sáenz-Badillos, Á. 165 n., 173 n., 180,
Poliziano, A. 162, 180–1 181 nn., 211 n.
Porter, S. E. 103, 283 n. Sakalis, D. T. 302 n.
Preisigke, F. XII, 6 n., 11 n., 74, 122, 135 n., Sandbach, F. H. 274 n.
202, 203 n., 239 n. Sandevoir, P. 109 n., 135
Prellwitz, W. 321 n. Satterthwaite, P. E. 230 n.
Proctor, R. 149 n., 150 n., 159–61, 162 nn. SBLHBS 126 n.
Psichari, J. 239 n. Scherer, A. 39 n., 58 n.
Schleusner, J. F. 7, 8 n., 108, 121, 129,
Quasten, J. 282 n., 290 n. 208 n., 360 n.
Schmid, W. 40 n., 56 n., 294 n., 301 n.,
Rabelais 169 312, 315 n.
Rahlfs, A. XII, 41 n., 111 n., 209, 348 Schmidt, J. H. H. 112 n.
Rajak, T. 226 n. Schneider, G. see EDNT
RBLG XII, 206 n. 23 Scholderer, V. 149 n., 150 n., 159 nn.,
Redpath, H. A. see HR, HR Suppl. 162 nn., 163 n.
Reese, J. M. 302 n. Schrenk, G. 239
Reeve, A. 145, n. Schubert, P. 397 n.
Regnault, L. 375 Schwyzer, Grammatik XII, 43 nn., 44 n.,
Reichmann, V. IX 60 n., 63 n., 150 n., 217 n., 307 n., 339 n.
Reiterer, F. V. 215 n. 26 Scott, J. A. 60 n.
Rengstorf, K. H. 42 n. 27 Scott, R., see LS, LSJ, LSJ Suppl.
Renner, T. 394 n. Screech, M. A. 145, 169 n.
Renouard, A. A. 150 n. Scrivener, F. H. A. 144
Richardson, M. E. J., see HALOT Seeligmann, I. L. 352, 354 n., 360 n., 362 n.
Riesenfeld, H. 239 Shaw, F. 234 n.
Rink, H. 203 n. Sheldon, J. 178 n., 343 n.
Ritschl, F. 309 n. Shipp, G. P. XII, 30, 69 n., 74 n., 106,
Robert, L. 203 n. 300 n., 302 n., 303 n., 306 n., 309 n., 310 n.,
Roberts, T. 342 n., 411 n. 313 n., 314 n., 315 n., 316 n., 366, 367 n.,
Robertson, A. T. 58 n., 65 n., 345 n. 381 n., 383 nn., 404 n.
Rocci, L. 408–24 Silva, M. 84 n., 106 nn., 107 n., 227 n.,
Rödiger, R. 239 230 n., 234 n., 236, 337 n., 404 n.
Rönnegård, P. 374 nn. Simpson, E. K. 69 n.
Rosenbach, Z. 129 Sirivianou, M. G. 384 nn.
Ross, W. A. 109 n. Skehan, P. W. 212 n., 214, 215 n., 220 n.
Roumanis, E. 387 n. Skinner, J. 15 n.
Rousseau, P. 280 n., 282 n., 289 n., 290 n., Sklavenites, T. E. 149 n.
291 n. Smend, R. 209 n., 214, 214 n., 216 n.
Rowe, G. O. 357 n. Smith, H. P. 15 n.
RSV 90, 95 n. Smyth, H. W. 43 n., 76 n., 154, 339 n.,
Ruijgh, C. J. 452 n. 357 n., 413 n.
Ruhnken, D. 304 n. Soisalon-Soininen, I. 428 n., 430 n., 441–
Russell, N. 378 n. 4, 447, 453
Sollamo, R. 349, 353 n., 392, 428 n., 447 n., Thompson, S. 345 n.
448 n. Thomson, C. XII, 23 n.
Sophocles, E. A. 75 n., 123 n., 156–8, 202, Thrall, M. E. 47 n.
208 n., 218 n., 239, 377, 383 n. Thumb, D. 43 n., 44 n., 196 n., 238 n.
Sotiropoulos, D. 54 n. Thurén, L. 284 n.
Souter, A. 265 n., 278 n., 320 TLG XII, 119, 187 n., 270 n., 375 n., 422 n.
Sparks, H. D. F. 65 n., 190 n. TLG Canon XII, 126, 187 n., 279 n.
Speiser, E. A. 27 n. TLL XII, 7 nn., 215
Spottorno, V. 95 n. Torrey, C. C. 51 n.
Staikos, K. S. 149 n. Tov, E. 21–36, 105 n., 202, 204–5, 229 n.,
Stamm, J. J., see HALOT 354 n.
Stasinopoulos, M. D. 280 n., 288 n., 291 n. Toy, C. H. 4 n.
Stavropoulos, D. N. see OGELD Trapp, E. see LBG
Steen, H. A. 49 n. Traugott, E. C. 237 n., 240 n., 243 n.
Stephanus, H. 40 n., 252 n. Travlos, I. N. 202 n.
Stolk, J. V. 197 n. Tregelles, S. P. 144 n., 145 n., 148 n., 160 n.
Stray, C. 258 n., 265 n., 278 Trenchard, W. C. 321–9, 336–7
Strout, D. 298 n. Triantaphyllides, M. A. 145 n., 149 n.,
Strugnell, J. 30 n. 284 n., 313 n.
Stuart Jones, H. 270, 411 n.; see also LSJ, Trifone, P. 179 n.
LSJ Suppl. Tronci, L. 254
Sturz, F. G. 390 Troxel, R. L. 351, 352 n., 355 n., 365,
Stylianos, Arbp. 279 n. 368 n., 370 n., 371
Swain, S. 294 n. Trudgill, P. 52 nn.
Swete, H. B. 4 n., 107, 209, 352, 390 Turner, C. H. 159 n., 259–60, 264–5
Swinn, S. P. 400 n., 404 n. Turner, E. G. 107 n., 150 n., 396 n., 397 n.
Turner, G. W. 52 nn.
Tabachovitz, D. 62 n., 65 n. Turner, N. 55 n., 305 n., 315; see also
Tagliaferro, E. 184 n., 185 n., 190 n. MHT, Grammar, III; MHT, Style
Talmon, S. 41 n.
Talshir, Z. 400 n. Ure, J. N. 52 n.
Tarazi, P. N. 193 n.
Tasker, R. V. G. 144, 159 n. Valente, S. 38 n., 304 n., 318
Taylor, B. A. 121 n. Vanonen, H. 441 n.
Taylor, J. 269 n. Vattioni, F. 392
Taylor, V. 47, 60 n., 65 n., 66 n. Veitch, W. 40 n.
TDNT 184, 403 n. Vergara, Juan de 211, 212 n.
Thackeray, H. St. J. 4 n., 34 n., 40 nn., Vilaras, Y. 149
48 n., 203 n., 230 n., 302 n., 352–4, 356 n., Voelz, J. W. 316 n.
359, 364, 367 n., 368 n., 369, 382 n. Vorm-Croughs, M. van der 353–4, 357
Thayer, J. H. 267 n.
Theophilos, M. 218 n. Wackernagel, J. 448
Thesleff, H. 54 n. Waddell, W. G. 415
Thiele, W. 216 Wagner, C. 209 n.
Thiersch, H. G. J. 390 Wagner, J. R. 351, 355 n., 365 n., 371
Thomas, H. 150 n., 178 nn., 178 n. Walbank, F. W. 13 n.
Thompson, A. A. 138 n., 257 n., 278, Walters, P. 33 n., 61 n., 62 n., 106, 135,
319 n.; see also LSJ Suppl. 352, 360 n., 367 n.
Walton, S. 319 n., 337 n. Wilson, N. 150 n.

Ward, B. 374 n., 375, 381 n., 382 n. Wilson, N. G. 166, 280 nn., 281 nn., 288 n.,
Watt, M. 178 n. 290 n., 291 n.
Weber, R. 215 n. Winston, D. 301
Weinfeld, M. 23 n. Witherington III, B. 190 n.
Weinreich, U. 5 n. Woodhead, A. G. 396 n.
Wellauer, A. 30 Woody, K. M. 160, 162, 176 n., 180
Wellhausen, J. 71 Wortley, J. 374 n.
Wendel, C. 302 n., 307 n. Wright, B. G. 214 n., 215 n.; see also NETS
Westcott, B. F. 347
Wevers, J. W. 20, 109 n., 111 n., 135, 136, Ximénes de Cisneros, F. 170–2
339, 340, 347–8, 398 n., 411 n., 432 n.,
438 n. Young-Evans, G. 378 n.
Whitmarsh, T. 286 n., 287 nn., 294 n.
Whittaker, M. 194 n. Zervos, I. S. see Mega Lexikon
Wifstrand, A. 55 n., 239 Zerwick, M. 61 n., 66 n.
Wilcken, U. 11 n. Zgusta, L. 135 n.
Wilcox, M. 71 Ziegler, J. 209, 209 nn., 210 nn., 212–3,
Wilk, F. 370 nn. 215 n., 216, 352, 359, 360 n., 391
Willi, A. 239 n. Zorell, F. 204, 321 n.
Willi-Plein, I. 108 n. Zúñiga, López de, D. 170, 181
Contributions to Biblical Exegesis and Theology
1. J.A. Loader, A Tale of Two Cities, Sodom and Gomorrah in the Old Testament,
early Jewish and early Christian Traditions, Kampen, 1990
2. P.W. Van der Horst, Ancient Jewish Epitaphs. An Introductory Survey of a Millen-
nium of Jewish Funerary Epigraphy (300 BCB-700 CE), Kampen, 1991
3. E. Talstra, Solomon’s Prayer. Synchrony and Diachrony in the Composition of
1 Kings 8, 14-61, Kampen, 1993
4. R. Stahl, Von Weltengagement zu Weltüberwindung: Theologische Positionen im
Danielbuch, Kampen, 1994
5. J.N. Bremmer, Sacred History and Sacred Texts in early Judaism. A Symposium in
Honour of A.S. van der Woude, Kampen, 1992
6. K. Larkin, The Eschatology of Second Zechariah: A Study of the Formation of a
Mantological Wisdom Anthology, Kampen, 1994
7. B. Aland, New Testament Textual Criticism, Exegesis and Church History: A Discus-
sion of Methods, Kampen, 1994
8. P.W. Van der Horst, Hellenism-Judaism-Christianity: Essays on their Interaction,
Kampen, Second Enlarged Edition, 1998
9. C. Houtman, Der Pentateuch: die Geschichte seiner Erforschung neben einer
Auswertung, Kampen, 1994
10. J. Van Seters, The Life of Moses. The Yahwist as Historian in Exodus-Numbers,
Kampen, 1994
11. Tj. Baarda, Essays on the Diatessaron, Kampen, 1994
12. Gert J. Steyn, Septuagint Quotations in the Context of the Petrine and Pauline
Speeches of the Acta Apostolorum, Kampen, 1995
13. D.V. Edelman, The Triumph of Elohim, From Yahwisms to Judaisms, Kampen, 1995
14. J.E. Revell, The Designation of the Individual. Expressive Usage in Biblical Narrative,
Kampen, 1996
15. M. Menken, Old Testament Quotations in the Fourth Gospel, Kampen, 1996
16. V. Koperski, The Knowledge of Christ Jesus my Lord. The High Christology of
Philippians 3:7-11, Kampen, 1996
17. M.C. De Boer, Johannine Perspectives on the Death of Jesus, Kampen, 1996
18. R.D. Anderson, Ancient Rhetorical Theory and Paul, Revised edition, Leuven, 1998
19. L.C. Jonker, Exclusivity and Variety, Perspectives on Multi-dimensional Exegesis,
Kampen, 1996
20. L.V. Rutgers, The Hidden Heritage of Diaspora Judaism, Leuven, 1998
21. K. van der Toorn (ed.), The Image and the Book, Leuven, 1998
22. L.V. Rutgers, P.W. van der Horst (eds.), The Use of Sacred Books in the Ancient
World, Leuven, 1998
23. E.R. Ekblad Jr., Isaiah’s Servant Poems According to the Septuagint. An Exegetical
and Theological Study, Leuven, 1999
24. R.D. Anderson Jr., Glossary of Greek Rhetorical Terms, Leuven, 2000
25. T. Stordalen, Echoes of Eden, Leuven, 2000
26. H. Lalleman-de Winkel, Jeremiah in Prophetic Tradition, Leuven, 2000
27. J.F.M. Smit, About the Idol Offerings. Rhetoric, Social Context and Theology of
Paul’s Discourse in First Corinthians 8:1-11:1, Leuven, 2000
28. T.J. Horner, Listening to Trypho. Justin Martyr’s Dialogue Reconsidered, Leuven,
2001
29. D.G. Powers, Salvation through Participation. An Examination of the Notion of the
Believers’ Corporate Unity with Christ in Early Christian Soteriology, Leuven, 2001
30. J.S. Kloppenborg, P. Hoffmann, J.M. Robinson, M.C. Moreland (eds.), The Sayings
Gospel Q in Greek and English with Parallels from the Gospels of Mark and Thomas,
Leuven, 2001
31. M.K. Birge, The Language of Belonging. A Rhetorical Analysis of Kinship Language
in First Corinthians, Leuven, 2004
32. P.W. van der Horst, Japheth in the Tents of Shem. Studies on Jewish Hellenism in Anti-
quity, Leuven, 2002
33. P.W. van der Horst, M.J.J. Menken, J.F.M. Smit, G. van Oyen (eds.), Persuasion and
Dissuasion in Early Christianity, Ancient Judaism, and Hellenism, Leuven, 2003
34. L.J. Lietaert Peerbolte, Paul the Missionary, Leuven, 2003
35. L.M. Teugels, Bible and midrash. The Story of ‘The Wooing of Rebekah’ (Gen. 24),
Leuven, 2004
36. H.W. Shin, Textual Criticism and the Synoptic Problem in Historical Jesus Research.
The Search for Valid Criteria, Leuven, 2004
37. A. Volgers, C. Zamagni (eds.), Erotapokriseis. Early Christian Question-and-
Answer Literature in Context, Leuven, 2004
38. L.E. Galloway, Freedom in the Gospel. Paul’s Exemplum in 1 Cor 9 in Conversation
with the Discourses of Epictetus and Philo, Leuven, 2004
39. C. Houtman, K. Spronk, Ein Held des Glaubens? Rezeptionsgeschichtliche Studien
zu den Simson-Erzählungen, Leuven, 2004
40. H. Kahana, Esther. Juxtaposition of the Septuagint Translation with the Hebrew
Text, Leuven, 2005
41. V.A. Pizzuto, A Cosmic Leap of Faith. An Authorial, Structural, and Theological
Investigation of the Cosmic Christology in Col 1:15-20, Leuven, 2005
42. B.J. Koet, Dreams and Scripture in Luke-Acts. Collected Essays, Leuven, 2006
43. P.C Beentjes. “Happy the One Who Meditates on Wisdom” (SIR. 14,20). Collected
Essays on the Book of Ben Sira, Leuven, 2006
44. R. Roukema, L.J. Lietaert Peerbolte, K. Spronk, J.W. Wesselius (eds.), The Interpre-
tation of Exodus. Studies in Honour of Cornelis Houtman, Leuven, 2006
45. G. van Oyen, T. Shepherd (eds.), The Trial and Death of Jesus. Essays on the Passion
Narrative in Mark, Leuven, 2006
46. B. Thettayil, In Spirit and Truth. An Exegetical Study of John 4:19-26 and a Theo-
logical Investigation of the Replacement Theme in the Fourth Gospel, Leuven,
2007
47. T.A.W. van der Louw, Transformations in the Septuagint. Towards an Interaction of
Septuagint Studies and Translation Studies, Leuven, 2007
48. W. Hilbrands, Heilige oder Hure? Die Rezeptionsgeschichte von Juda und Tamar
(Genesis 38) von der Antike bis zur Reformationszeit, Leuven, 2007
49. J. Joosten, P.J. Tomson (eds.), Voces Biblicae. Septuagint Greek and its Significance
for the New Testament, Leuven, 2007
50. A. Aejmelaeus, On the Trail of the Septuagint Translators. Collected Essays, Leuven,
2007
51. S. Janse, “You are My Son”. The Reception History of Psalm 2 in Early Judaism
and the Early Church, Leuven, 2009
52. K. De Troyer, A. Lange, L.L. Schulte (eds.), Prophecy after the Prophets? The Con-
tribution of the Dead Sea Scrolls to the Understanding of Biblical and Extra-Biblical
Prophecy, Leuven, 2009
53. C.M. Tuckett (ed.), Feasts and Festivals, Leuven, 2009
54. M. Labahn, O. Lehtipuu (eds.), Anthropology in the New Testament and its Ancient
Context, Leuven, 2010
55. A. van der Kooij, M. van der Meer (eds.), The Old Greek of Isaiah: Issues and
Perspectives, Leuven, 2010
56. J. Smith, Translated Hallelujehs. A Linguistic and Exegetical Commentary on Select
Septuagint Psalms, Leuven, 2011
57. N. Dávid, A. Lange (eds.), Qumran and the Bible. Studying the Jewish and Christian
Scriptures in Light of the Dead Sea Scrolls, Leuven, 2010
58. J. Chanikuzhy, Jesus, the Eschatological Temple. An Exegetical Study of Jn 2,13-22 in
the Light of the Pre 70 C.E. Eschatological Temple Hopes and the Synoptic Temple
Action, Leuven, 2011
59. H. Wenzel, Reading Zechariah with Zechariah 1:1–6 as the Introduction to the Entire
Book, Leuven, 2011
60. M. Labahn, O. Lehtipuu (eds.), Imagery in the Booky of Revelation, Leuven, 2011
61. K. De Troyer, A. Lange, J.S. Adcock (eds.), The Qumran Legal Texts between the
Hebrew Bible and Its Interpretation, Leuven, 2011
62. B. Lang, Buch der Kriege – Buch des Himmels. Kleine Schriften zur Exegese und
Theologie, Leuven, 2011
63. H.-J. Inkelaar, Conflict over Wisdom. The Theme of 1 Corinthians 1-4 Rooted in
Scripture, Leuven, 2011
64. K.-J. Lee, The Authority and Authorization of Torah in the Persion Period, Leuven, 2011
65. K.M. Rochester, Prophetic Ministry in Jeremiah and Ezekiel, Leuven, 2012
66. T. Law, A. Salvesen (eds.), Greek Scripture and the Rabbis, Leuven, 2012
67. K. Finsterbusch, A. Lange (eds.), What is Bible?, Leuven, 2012
68. J. Cook, A. van der Kooij, Law, Prophets, and Wisdom. On the Provenance of Trans-
lators and their Books in the Septuagint Version, Leuven, 2012
69. P.N. De Andrado, The Akedah Servant Complex. The Soteriological Linkage of
Genesis 22 and Isaiah 53 in Ancient Jewish and Early Christian Writings, Leuven,
2013
70. F. Shaw, The Earliest Non-Mystical Jewish Use of Ιαω, Leuven, 2014
71. E. Blachman, The Transformation of Tamar (Genesis 38) in the History of Jewish
Interpretation, Leuven, 2013
72. K. De Troyer, T. Law, M. Liljeström (eds.), In the Footsteps of Sherlock Holmes. Studies
in the Biblical Text in Honour of Anneli Aejmelaeus, Leuven, 2014
73. T. Do, Re-thinking the Death of Jesus. An Exegetical and Theological Study of Hilasmos
and Agape in 1 John 2:1-2 and 4:7-10, Leuven, 2014
74. T. Miller, Three Versions of Esther. Their Relationship to Anti-Semitic and Feminist
Critique of the Story, Leuven, 2014
75. E.B. Tracy, See Me! Hear Me! Divine/Human Relational Dialogue in Genesis, Leu-
ven, 2014
76. J.D. Findlay, From Prophet to Priest. The Characterization of Aaron in the Penta-
teuch, Leuven, forthcoming
77. M.J.J. Menken, Studies in John’s Gospel and Epistles. Collected Essays, Leuven, 2015
78. L.L. Schulte, My Shepherd, though You Do not Know Me. The Persian Royal Propa-
ganda Model in the Nehemiah Memoir, Leuven, 2016
79. S.E. Humble, A Divine Round Trip. The Literary and Christological Function of the
Descent/Ascent Leitmotif in the Gospel of John, Leuven, 2016
80. R.D. Miller, Between Israelite Religion and Old Testament Theology. Essays on Archae-
ology, History, and Hermeneutics, Leuven, 2016
81. L. Dequeker, Studia Hierosolymitana, Leuven, 2016
82. K. Finsterbusch, A. Lange (eds.), Texts and Contexts of Jeremiah. The Exegesis of
Jeremiah 1 and 10 in Light of Text and Reception History, Leuven, 2016
83. J.S. Adcock, “Oh God of Battles! Steal My Soldiers’ Hearts!” A Study of the Hebrew
and Greek Text Forms of Jeremiah 10:1-18, Leuven, 2017
84. R. Müller, J. Pakkala (eds.), Insights into Editing in the Hebrew Bible and the Ancient
Near East. What Does Documented Evidence Tell Us about the Transmission of
Authoritative Texts?, Leuven, 2017
85. R. Burnet, D. Luciani, G. van Oyen (eds.), The Epistle to the Hebrews. Writing at the
Borders, Leuven, 2016
86. M.K. Korada, The Rationale for Aniconism in the Old Testament. A Study of Select
Texts, Leuven, 2017
87. P.C. Beentjes, “With All Your Soul Fear the Lord” (Sir. 7:27). Collected Essays on
the Book of Ben Sira II, Leuven, 2017
88. B.J. Koet, A.L.H.M. van Wieringen (eds.), Multiple Teachers in Biblical Texts,
Leuven, 2017
89. T. Elgvin, The Literary Growth of the Song of Songs during the Hasmonean and
Early-Herodian Periods, Leuven, 2018
90. D.C. Smith, The Role of Mothers in he Genealogical Lists of Jacob’s Sons, Leuven,
2018
91. V.P. Chiraparamban, The Manifestation of God’s Merciful Justice. A Theocentric Read-
ing of Romans 3-21-26, Leuven, 2018
92. P. Paul, Beyond the Breach. An Exegetical Study of John 4:1-42 as a Text of Jewish-
Samaritan Reconciliation, Leuven, 2021
93. I. Fröhlich, David in Cultural Memory, Leuven, 2019
94. M. Langlois, The Samaritan Pentateuch and the Dead Sea Scrolls, Leuven, 2019
95. A. Livneh, Studies on Jewish and Christian Historical Summaries from the Hellen-
istic and Early Roman Periods, Leuven, 2019
96. M. Rotman, The Call of the Wilderness. The Narrative Significance of John the
Baptist’s Whereabouts, Leuven, 2020
97. O. Lukács, Sabbath in the Making. A Study of the Inner-Biblical Interpretation of
the Sabbath Commandment, Leuven, 2020
98. J.J. Spoelstra, Life Preservation in Genesis and Exodus. An Exegetical Study of
the Tebâ of Noah and Moses, Leuven, 2020
99. T. Havukainen, The Quest for the Memory of Jesus: a Viable Path or a Dead End?,
Leuven, 2020
100. K. De Troyer, The Ultimate and the Penultimate Text of the Book of Joshua, Leuven,
2018
103. Angela Kim Harkins and Barbara Schmitz (eds.), Selected Studies on Deuterocanonical
Prayers, Leuven, 2021
104. Torben Plitt, Wachstumsgesetz oder Kürzungstendenz? Gedächtnispsychologische
Erkenntnisse zu den inhaltlichen Überhängen in den synoptischen Paralleltraditionen,
Leuven, 2021
105. Ian Wilson, Praying to the Temple: Divine Presence in Solomon’s Prayer, Leuven,
2022
PRINTED ON PERMANENT PAPER • IMPRIME SUR PAPIER PERMANENT • GEDRUKT OP DUURZAAM PAPIER - ISO 9706
N.V. PEETERS S.A., WAROTSTRAAT 50, B-3020 HERENT

John Lee, Collected Essays On The Greek Bible and Greek Lexicography

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

John Lee, Collected Essays On The Greek Bible and Greek Lexicography

Uploaded by

Copyright:

Available Formats

BIBLICAL EXEGESIS & THEOLOGY 112

Reimund Bieringer (Leuven)

COLLECTED ESSAYS ON THE GREEK

© 2022 — Peeters, Bondgenotenlaan 153, B-3000 Leuven

All rights reserved. No part of this publication may be reproduced, stored in a

Whether one ought to publish a collection of one’s essays is a question

1 A Note on Septuagint Material in the Supplement to Liddell and

22 The Literary Greek of Septuagint Isaiah . . . . . . . . . . . . 351

BAG = W. Bauer, trans. W. F. Arndt and F. W. Gingrich, Greek-English Lexi-

KJV = King James Version (1611).

Migne, PG = J.-P. Migne, (ed.), Patrologiae cursus completus: Series Graeca

Note abbreviations of the following biblical books:

There is no need to give a summary of the contents of the essays at

in principle the same, just with a difference in time period. Corresponding

At the same time, a wide range of language phenomena are investigated.

In 1968 a Supplement to LSJ was published, containing 153 pages of ‘correc-

In the recently published Supplement to Liddell and Scott’s Greek-

The two LXX passages are as follows:

‘Slave’ is a possible translation of ψυχήν in these contexts, but it does

The passage is:

enmity is suggested by the general tone of the passage. Admittedly the

Lev 5:16 reads as follows:

That the fifth is additional is indicated by προσθήσει. The use of ἐπί-

A somewhat different error of method, that of supposing without ade-

The passages are:

The proposed meaning is based solely on the meaning of the Hebrew

in an instance where the Hebrew word has a sense unparalleled in the

Here and in Ezek 16:54; 32:24, 30 βάσανος corresponds to MT ‫כלמה‬,

equivalent of ‫כלמה‬. The evidence we have for the meaning of βάσανος in

below. The evidence at present available to us indicates clearly that σκε-

In another type of entry the Supplement classifies certain occurrences

The separation of the example under a new heading, suggesting that

In conclusion we may notice the Supplement’s entry s.v. ὑφοράω:

In early Koine Greek ἀποσκευή, a new formation attested first in the

The primary sense is, as the etymology leads us to expect, ‘movable

Similarly Plb. 1.68.3; 2.70.5; 2.3.8, and probably in PCairZen 1.93.9

Other examples of ἀποσκευή in the sense of ‘baggage-train’ are found

Consequently the Carthaginians compel the mercenaries to take τὰς

Similarly in UPZ 1.110.90; 2.199, 206 (164 BC); SB 5.8009.3 (I BC).

It is to be noticed that in senses 3 and 4 ἀποσκευή refers specifi-

In other passages, however, it is clear that we must recognise a sense

Two further examples may be seen in Exod 10:8–11. Pharaoh agrees

Exod 10:10 καθότι ἀποστέλλω ὑμᾶς, μὴ καὶ τὴν ἀποσκευὴν ὑμῶν; …

The LXX translators’ renderings of ‫ טף‬also indicate that they recog-

A number of LXX passages raise a special problem. They are those in

What is the meaning of our word here? ‘Baggage,’ ‘possessions’ will

sense. In my view, however, there is no compelling reason for seeing a

The second is:

Ἀποσκευή is hardly possible in this context. If read it will have to bear an

In a recent article Emanuel Tov has tackled some basic questions of

with reference to the original Hebrew, by means of which the translator’s

If we read the Greek on its own, the meaning of συγγενέσθαι is unmis-

My next example is:

Of course προσπορεύεσθαι is a more or less literal equivalent of ‫לקרבה‬.

A third example is:

In MT the word corresponding to κατέστρωσεν αὐτοὺς is ‫וישׁחטם‬

Moreover, a NT passage provides evidence that κατέστρωσεν was under-

The word of interest is once again συγγίνομαι. But here we find a

available. καταστρώννυμι is unusual (only here in Pent.) and seems an unmistakable

to lead one to understand συγγενώμεθα in anything but a vague sense

The construction in the Greek is misleading: παρὰ τοῦ πατρός seems to

I turn now to a different aspect of the problem of determining the

that of the Hebrew word, because it follows its Hebrew equivalent in