Liddell and Scott: The History, Methodology, and Languages of the World's Leading Lexicon of Ancient Greek 9780198810803, 0198810806

Table of contents :
Title_Pages
Frontispiece
List of Figures
List of Tables
List of Abbreviations
List of Contributors
A Note on the History of the Lexicon
Liddell and Scott in Historical Context. Victorian Beginnings, Twentieth-Century Developments
Dictionaries as Translations. English in the Lexicon
Latin in the Lexicon
Obscenity. A Problem for the Lexicographer
Etymology and Etymologies
Incorporating New Evidence. Mycenaean Greek in the Revised Supplement
A Canonical Author. The Case of Hesiod
Philosophy and Linguistic Authority. The Problem of Plato’s Greek
Medical Vocabulary, with Especial Reference to the Hippocratic Corpus
The Greek of the New Testament
The Ancient, the Medieval, and the Modern in a Greek-English Lexicon, or How To Get Your Daily ‘Bread’ in Greek Any Day Through the Ages
Greek Dialects in the Lexicon
Between Cunning and Chaos. μῆτις
Looking for Unity in a Dictionary Entry. A Perspective from Prototype Theory
Discourse Particles in LSJ. A Fresh Look at γε
LSJ and the Diachronic Taxonomy of the Greek Vocabulary
Literary Lexicography. Aims and Principles
Lessons Learned During my Time at the Lexikon des frühgriechischen Epos
Diminishing Returns and New Challenges
Βάπτω - An Illustration of the State of our Ancient Greek Dictionaries
Liddell and Scott and the Oxford English Dictionary
Bibliography
Greek Index

Citation preview

LIDDELL AND SCOTT

Frontispiece. Liddell and Scott’s Greek English Lexicon, first edition (1843), page 1.

Liddell and Scott The History, Methodology, and Languages of the World’s Leading Lexicon of Ancient Greek

E 

CHRISTOPHER STRAY, M I C H A E L C L A R K E , A N D JO S H U A T . K A T Z

1

3 Great Clarendon Street, Oxford, OX2 6DP, United Kingdom Oxford University Press is a department of the University of Oxford. It furthers the University’s objective of excellence in research, scholarship, and education by publishing worldwide. Oxford is a registered trade mark of Oxford University Press in the UK and in certain other countries © Oxford University Press 2019 The moral rights of the authors have been asserted First Edition published in 2019 Impression: 1 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, without the prior permission in writing of Oxford University Press, or as expressly permitted by law, by licence or under terms agreed with the appropriate reprographics rights organization. Enquiries concerning reproduction outside the scope of the above should be sent to the Rights Department, Oxford University Press, at the address above You must not circulate this work in any other form and you must impose this same condition on any acquirer Published in the United States of America by Oxford University Press 198 Madison Avenue, New York, NY 10016, United States of America British Library Cataloguing in Publication Data Data available Library of Congress Control Number: 2019950829 ISBN 978 0 19 881080 3 DOI: 10.1093/oso/9780198810803.003.0001 Printed and bound by CPI Group (UK) Ltd, Croydon, CR0 4YY Links to third party websites are provided by Oxford in good faith and for information only. Oxford disclaims any responsibility for the materials contained in any third party website referenced in this work.

Frontispiece. Liddell and Scott’s Greek English Lexicon, first edition (1843), page 1.

List of Figures Frontispiece. Liddell and Scott’s Greek English Lexicon, first edition (1843), page 1. 1.1. The monstrous letter Π, depicted by Henry Liddell in his letter to Robert Scott of July 1842 (Thompson 1899, 75).

ii 12

12.1. Beginning of the entry for ἄρσην ‘male’, LSJ.

204

12.2. Beginning of the entry for νᾱός ‘temple’, LSJ.

205

12.3. Beginning of the entry for ἀνδρεῖος ‘of or for a man’, LSJ.

206

12.4. Entry for Περσεφόνη ‘Persephone’, LSJ.

206

12.5. Beginning of the entry for σωφροσύνη ‘soundness of mind’, LSJ.

207

12.6. Beginning of the entry for Ῥέα ‘Rhea’, LSJ.

207

12.7. Entry for ψάμμη ‘sand’, LSJ.

207

12.8. Beginning of the entry for ψάμμος ‘sand’, LSJ.

207

12.9. Beginning of the entry for ἕννυμι ‘put clothes on (another)’, LSJ.

208

12.10. Entry for ποτιμάστιος ‘at the breast’, LSJ.

208

12.11. Entry for συστᾰθεύω ‘help to roast’, LSJ.

208

12.12. Entry for ἄεθλος, ‘contest’, Passow 1830.

215

12.13. Entry for εὐφροσύνη ‘mirth’, Passow 1830.

215

12.14. Entry for ἐχθάνομαι, ‘incur hatred’, Passow 1830.

216

12.15. Entry for ἐπιλοιβή, ‘drink offering’, Passow 1830.

216

12.16. End of the entry for ἐϋκτίμενος, Passow 1830.

221

12.17. End of the entry for προτιόσσομαι, Passow 1830.

221

12.18. Eton Greek Grammar (1819, 218).

223

14.1. Chadwick’s account of μένος in Lexicographica Graeca.

253

14.2. μένος plotted according to its (hypothetical) diachronic development.

255

14.3. A possible prototype based account of ἄλσος.

259

16.1. Watkins’ schema for moveable property in Early Greek.

291

List of Tables 1.1. Editions of the Lexicon.

13

12.1. Forms labelled ‘common’ in LS.

213

12.2. Instances of the label ‘gewöhnlich’ used in Passow 1830, Vol. 1 to distinguish a form from one or more dialectal variants, where LS does not label the same form ‘common’.

217

12.3. Instances of the label ‘gewöhnlich’ used in Passow 1830, Vol. 1 to distinguish a form from one or more variants, where the variation has nothing to do with dialects and LS does not label the same form ‘common’.

218

12.4. Instances of the label ‘gemein’ used in Passow 1830 to distinguish a form from one or more variants.

220

List of Abbreviations BDAG

Danker, F.W. (2000) A Greek English Lexicon of the New Testament and Other Early Christian Literature (BDAG). Based on W. Bauer’s Griechisch deutsches Wörterbuch zu den Schriften des Neuen Testaments und der frühchristlichen Literatur, Sixth edition, ed. K. Aland and B. Aland with V. Reichmann, and on Previous English Editions by W.F. Arndt, F.W. Gingrich, and F.W. Danker (Chicago).

CGL

J. Diggle et al. (2019) Cambridge Greek Lexicon (Cambridge).

DELG

P. Chantraine (1968 80) Dictionnaire étymologique de la langue grecque (Paris).

DELG Suppl.

P. Chantraine et al. (2009) Dictionnaire étymologique de la langue grecque (Paris).

DGE

F.R. Adrados (ed.) (1980 ), Diccionario griego español (Madrid).

GE

F. Montanari (1995) Vocabolario della lingua greca (Turin).

GI

F. Montanari (2015) The Brill Dictionary of Ancient Greek (Leiden).

IGL

H.G. Liddell (1889) An Intermediate Greek English Lexicon (Oxford).

LS1–8

H.G. Liddell and R. Scott (1843; 8th edn 1897), A Greek English Lexicon (Oxford).

LfgrE

Lexikon des frühgriechischen Epos (1955 2010) (Göttingen).

LSJ

H.S. Jones and R. McKenzie (1940) A Greek English Lexicon Compiled by H.G. Liddell and R. Scott. A New Edition, Revised and Augmented Throughout (Oxford).

LSJ Suppl.

H.G. Liddell, R. Scott, H. S. Jones, and R. McKenzie (1940), Greek English Lexicon: A Supplement, ed. E.A. Barber with the assistance of P. Maas, M. Scheller, and M.L. West.

LSJ Revd Suppl.

H.G. Liddell, R. Scott, H. S. Jones, and R. McKenzie (1996), Greek English Lexicon: Revised Supplement, ed. P.G.W. Glare, with A.A. Thompson (Oxford).

OED

J. Simpson and E. Weiner, (eds) (1989; CD ROM 1992, online 2000), Oxford English Dictionary, 2nd edn (Oxford).

OLD

P.G.W. Glare, (ed.) (1982), Oxford Latin Dictionary (Oxford).

Passow¹

F. Passow (1819 24), Handwörterbuch der griechischen Sprache, 1st edn. (Leipzig).

Passow⁴

F. Passow (1831), Handwörterbuch der griechischen Sprache 4th edn (Lepizig).

SEG

Supplementum Epigraphicum Graecum (1923 ) (Amsterdam).

TLG

Thesaurus Linguae Graecae.

TLL

Thesaurus Linguae Latinae.

List of Contributors Evelien Bracke is Language Lector in Greek Literature at Ghent University, Belgium. David Butterfield is Senior Lecturer in Classics at the University of Cambridge and Fellow and Director of Studies in Classics at Queens’ College, Cambridge, UK. James Clackson is Professor of Comparative Philology at the University of Cambridge and Fellow and Director of Studies in Classics at Jesus College, Cambridge, UK. Michael Clarke is Professor of Classics at the National University of Ireland, Galway, Ireland. Amy Coker is an Honorary Research Fellow, Dept of Classics & Ancient History, University of Bristol and Teacher of Classics at Cheltenham Ladies’ College. John Considine is Professor of English at the University of Alberta, Canada. Elizabeth Craik is Honorary Professor of Classics at the University of St Andrews, UK. David Goldstein is Associate Professor of Linguistics and Indo-European Studies at the University of California, Los Angeles, USA. Patrick James was a Research Associate in the Faculty of Classics at the University of Cambridge, UK, and served as an Assistant Editor for the Cambridge Greek Lexicon Project. He now teaches Greek, Latin, and Classical Civilisation at Haileybury in Hertfordshire. Mark Janse is BOF-ZAP Research Professor in Ancient and Asia Minor Greek at Ghent University, Belgium. Joshua T. Katz is Cotsen Professor in the Humanities and Professor of Classics at Princeton University, USA. Tom Mackenzie is a Leventis Research Fellow in the Department of Greek and Latin at University College London, UK. Michael Meier-Brügger is Professor of Comparative and Indo-European Linguistics at the Free University of Berlin, Germany. Philomen Probert is Professor of Classical Philology and Linguistics at the University of Oxford, UK.

xvi

List of Contributors

Christopher Rowe is Emeritus Professor of Greek at Durham University, and Leverhulme Emeritus Fellow. Michael Silk is Emeritus Professor of Classical and Comparative Literature at King’s College London, UK. Christopher Stray is an Honorary Research Fellow in the Department of Classics, Ancient History, and Egyptology at Swansea University, UK. Anne Thompson was Assistant Editor for the Revised Supplement to Liddell and Scott, and the first Editor of the forthcoming Cambridge Greek Lexicon. Brent Vine is Distinguished Professor of Classics and A. Richard Diebold, Jr. Professor of Indo-European Studies at the University of California, Los Angeles, USA. †

Martin L. West was an Honorary Fellow of All Souls College, Oxford, UK, and a Member of the Order of Merit. Margaret Williamson is Associate Professor Emerita of Classics and Comparative Literature at Dartmouth College, USA.

A Note on the History of the Lexicon Christopher Stray The Lexicon now universally known as Liddell and Scott first appeared in 1843, the work of two young Oxford graduates. Henry George Liddell (1811–98) and Robert Scott (1811–87) had graduated with first-class honours in Classics in 1833; Liddell became a student (fellow) of Christ Church, where they had both been undergraduates, Scott a fellow of Balliol. Both later became college heads: Scott was Master of Balliol 1854–70, Liddell Dean of Christ Church 1855–91. They were not only collaborators but firm friends, though their friendship must have been challenged by the chasm between their political views: Liddell was a liberal, Scott a conservative of the deepest dye. The Lexicon was commissioned from the two young graduates in 1836 by a local bookseller and publisher, David Talboys, who had published several translations of German works on classical subjects, some of them translated by himself. The Lexicon too was to begin with a translation, based on a GreekGerman lexicon by Franz Passow, who had died in 1833. The first edition of Passow’s book had been published in 1819–24, the fourth and final edition in 1831. The first three editions of Liddell and Scott’s Lexicon bore the title, A Greek-English Lexicon based on the German Work of Francis Passow. In the fourth edition of 1855, this became A Greek-English Lexicon compiled by Henry Liddell and Robert Scott, the editors explaining in their preface that they had added so much material of their own that the Lexicon was now a new work. The seventh edition of 1882, intended at the time to be the final, definitive edition, was the last to be revised by both editors, Liddell alone being responsible for the eighth edition (1897) after Scott’s death in 1887 and just before his own in 1898. An abridged edition, ‘chiefly for the use of schools’, was published a few months after the first edition of the larger work; it was revised several times and is still in print a century after its last revision. An Intermediate Lexicon assembled by Liddell appeared in 1889, to cater for the needs of sixth forms and undergraduates. This too is still in print, but remarkably, it is unchanged from its original publication. After Liddell’s death, a scheme was hatched to move on from Liddell’s rather lightly revised final edition to a new work which would take account of recent discoveries by epigraphers and papyrologists. After some false starts, the project was taken over by Henry Stuart Jones (1867–1939), a capable and versatile scholar whose career included work on ancient history and archaeology as well as on Greek texts. The philologist Roderick McKenzie

xviii

A Note on the History of the Lexicon

(1861–1934) was recruited to update Liddell and Scott’s etymologies, and a stream of material was sent in by British and Continental scholars. LSJ, as it is now generally called, was published in ten fascicles from 1925 onwards, and then in two volumes in 1940. Its title page balanced inheritance, collaboration, and outside assistance: A Greek-English Lexicon compiled by Henry George Liddell and Robert Scott. A New Edition, Revised and Augmented Throughout by Sir Henry Stuart Jones D.Litt (1867–1939), with the assistance of Roderick McKenzie, and with the co-operation of many Scholars. New material, including evidence provided by the decipherment of Linear B, was later published in two Supplements in 1968 and 1996.

1 Liddell and Scott in Historical Context Victorian Beginnings, Twentieth-Century Developments Christopher Stray

Liddell and Scott is so massively familiar a part of the lives of students of Greek at all levels that we tend to take it for granted: it is just there on the shelf, desk, or table, waiting to be consulted by the uncertain and appealed to in cases of dispute. Behind this monumental and impartial familiarity, however, is a complex history of scholarly controversy and commercial book production. This chapter sets the first eight editions of the Lexicon (1843–97) in a number of contemporary contexts: the institutional and intellectual world of Oxford in the 1830s and 1840s; the emergence of classical dictionaries using vernaculars rather than Latin for glosses; the relationship of the Lexicon with other dictionaries; the development of the book through successive editions and abridgements; the reputation of the Lexicon; and its printing and publishing history. The chapter will aim to explore these separate contexts, and to suggest how they interacted.

1 . 1 . E A R L Y V I C T O RI A N O X F O R D : I N S T I T U T I O N A L AND I NTELLECTUAL CONTEXTS The Lexicon began life as a commission not from its eventual publisher, Oxford University Press, but from a local bookseller and publisher, David Talboys, who since his arrival in the city in 1814 had built up an impressive list including several translations of German academic works, some of them translated by himself.¹ Oxford University Press did not have a book in the ¹ Various conjectural dates have been given for the commission. Thompson 1899, 66 7 quotes from a letter in which Liddell tells his friend H.H. Vaughan that he and Scott are about Christopher Stray, Liddell and Scott in Historical Context: Victorian Beginnings, Twentieth-Century Developments In: Liddell and Scott: The History, Methodology, and Languages of the World’s Leading Lexicon of Ancient Greek. Edited by: Christopher Stray, Michael Clarke, and Joshua T. Katz, Oxford University Press (2019). © Oxford University Press. DOI: 10.1093/oso/9780198810803.003.0001

4

Christopher Stray

field, though this was not for want of trying. In 1827 the Delegates of the Press had enthusiastically accepted an offer (which they may have encouraged) from a young Christ Church man, J.A. Cramer, to translate a seventeenth-century Greek-Latin/Latin-Greek lexicon into English, but in 1830, after printing had begun, the project was cancelled, for reasons that remain unclear.² In 1829 the press commissioned another recent Oxford graduate, John Riddle, to translate the Latin-German dictionary of Immanuel Scheller; this emerged in 1835 as a single heavy and unwieldy volume printed in three columns. Riddle went on to produce three progressively abridged versions (1836–43), published not by OUP but by Longmans. Both this commission and Cramer’s offer were probably encouraged by Thomas Gaisford, a Press Delegate since 1807 and regius professor of Greek since 1812; he himself worked extensively on ancient lexica (Stray 2018, 53–81). Since 1831 Gaisford had been Dean of Christ Church, the college to which Liddell and Scott belonged as undergraduates; his support is acknowledged in the preface to the first edition of the Lexicon. Liddell and Scott, like Cramer and Riddle, were recent graduates with firstclass degrees when commissioned to produce their books—a common pattern in OUP publishing in this period. The Abridged Lexicon was published shortly after the large book in 1843, so it had clearly been planned for some time (cf. note 21 below). This unusual, indeed unique arrangement may have stemmed from previous abridgements such as Riddle’s Latin dictionaries and the Tyro’s Lexicon of John Jones (1825), an abridgement of his larger Greek-English lexicon of 1823, the first work of its kind. On Liddell’s death in 1898, Thomas Hardy published a poem entitled ‘Liddell and Scott, on the completion of their lexicon’, cast in the form of a dialogue between the two men. At one point Scott remarks: . . . how I often, often wondered What could have led me to have blundered So far away from sound theology To dialects and etymology . . .

Hardy’s contrast between sound theology and speculative philology is dramatically effective, but does not do justice to Liddell and Scott’s situation in the midst of Anglican controversy. It is not just that theology and philology were heavily intertwined in the 1830 and 1840s, but that theology was becoming a contested area just at the point when Liddell and Scott began to close an engagement with Talboys to edit the Lexicon. The letter, which Thompson does not date, was written on 21 November 1836 (Bodleian Library, Ms Eng Let d 435). ² The source lexicon was probably William Robertson’s Thesaurae Graecae Linguae (1676), a Greek Latin, Latin Greek dictionary This was based on the 1654 lexicon of Cornelis Schrevel, itself an enlarged revision of the 1562 Greek Latin lexicon of Robert Constantin, which was more useful than Estienne’s Thesaurus both in its manageable size and in its alphabetical ordering of entries. For other lexica derived from Schrevel, see n. 13 below.

Liddell and Scott in Historical Context

5

work. From 1833 Oxford was haunted by the Tractarian controversy, which culminated in J.H. Newman’s conversion to Catholicism in 1845 (Nockles 1997). In 1838, in a letter to his pupil Henry Acland, Liddell reported that ‘I am still able to plod my weary way through the never ending reams of Passow’, before telling him that ‘the University’s principle [sic] topic is what the Cambridge scoffers call “New-mania” . . . There are men who have (as it were) a call to higher things . . . most of us . . . knew (thank God for it) what we ought to believe, & what . . . we engaged steadily in following & doing, without constant speculation . . .’³ Liddell’s account reflects a concern to avoid controversy in which the steady pursuit of philological projects surely played a useful role. One of his and Scott’s final degree examiners was William Sewell, an eccentric high Anglican who at first supported the Tractarians, and who has been credited with suggesting the Lexicon project. In his biography of Liddell, Henry Thompson hinted that Sewell proposed the project as a way of avoiding the obsession with theology in the University (Thompson 1899, 66–7). Thompson quotes from a letter from Liddell to H.H. Vaughan (referred to above, n.1) in which he wrote that ‘Sewell thinks the Oxford mind is running too much to pure Theology: if you think so too, you will be glad to hear that some of us are—in all likelihood—about to close an engagement with Talboys for a lexicon founded chiefly on Passow.’ The nexus of theology and philology has been explored by Linda Dowling in her Language and Decadence in the Victorian Fin-de-Siècle (Dowling 1986), where she discusses the work of Julius Hare, Connop Thirlwall, and their Philological Museum (1831–3). She also mentions Passow’s lexicon (p. 58), but as a precursor to the Oxford English Dictionary: Liddell and Scott do not appear. Yet in the period from Passow’s death in 1833 to R.C. Trench’s proposal of a new English dictionary in 1857, it was the Greek lexicon which carried the Passow standard, with entries which laid out the biographies of words.⁴ The links between theology and philology can be seen in the careers of William Wordsworth’s nephews Charles and Christopher Wordsworth, schoolmasters who became bishops. They had been brought up as Tories and high Anglicans, and in the 1830s, when liberalism and political reform was in the air and the Tractarian controversy was at its height, were concerned to bolster established religion and conservative morality. As scholars, they felt that classical literature should be used to defend revealed religion; as schoolmasters, that (in Christopher’s words) ‘uniformity in grammar is no inconsiderable step towards uniformity in religion’ (italics in original: see Stray 2016). The brothers vigorously refought the religious battles of the late Roman Empire and the Reformation, but their isolation was heightened when their erstwhile ally Gladstone went over to liberal policies in the late 1840s. In 1847, ³ Liddell to Acland, 23 April 1838. Liddell papers, Christ Church library, MS 348. ⁴ For useful context, see Aarsleff 1983, 249 58; Zgusta 2006, 27 38.

6

Christopher Stray

Charles Wordsworth wrote to J.R. Hope, ‘WEG has let us down, abandoned the high ground and the sure ground, the mission to save Church and State’, and went on to urge uniformity in religion in Britain. D.C. Latham commented on this statement, ‘Wordsworth was left in the fortress raised by his own imagination, secure, had he but known it, in the fact that it would never again be thought worthy of a serious attack’ (Latham 1910, I.372–3). Philology also loomed large in the work of such liberal Anglicans as Julius Hare, Connop Thirlwall, and Thomas Arnold (see Brent 1983; Morris 2004). The impact of the Tractarian debates can be seen in the measured pronouncement of the Royal Commissioners on Oxford. In their report, issued in 1852, they supported the establishments of professorships, remarking that: The presence of men eminent in various departments of knowledge would impart a dignity and stability to the whole institution . . . whilst from within it would tend above all other means to guard the university from being absorbed, as it has been of late years, by the agitations of theological controversy. (Royal Commission 1852, 52; cf. O.Chadwick 1970, 440)

Looking back on the Tractarian era and its aftermath, Mark Pattison commented that: If any Oxford man had gone to sleep in 1846 and had woke up again in 1850 he would have found himself in a totally new world. In 1846 we were in Old Tory Oxford; not somnolent because it was fiercely debating, as in the days of Henry IV., its eternal Church question. . . . In 1850 this was all suddenly changed as if by the wand of a magician. The dead majorities of head and seniors, which had sat like lead upon the energies of young tutors, had melted away. Theology was totally banished from Common Room, and even from private conversation.⁵

Even after the Tractarian controversy had died away, divisions between religious conservatives and liberals played a part in Oxford appointments. In 1855 Benjamin Jowett was appointed to the regius chair of Greek. Soon afterwards, Arthur Stanley wrote to him that he had not pressed Jowett’s claims to the chair on Liddell, since he did not want Jowett to get lost in that sort of work: it would have been better to wait for something more important. Stanley went on, ‘I should have been inclined to let Scott have it. He would be less dangerous there than in a theological field.’⁶ (Stanley was alluding to Scott’s theological conservatism.) He and Jowett must have been unhappy at Scott’s election in 1861 to the chair of Biblical exegesis, which came in the wake of renewed religious controversy after the publication of a volume of liberal theological studies, Essays and Reviews, to which Jowett contributed (Ellis ⁵ Mark Pattison, Memoirs (London, 1885), 244; see also 212 15, 244 5. For the veiled role played by OUP reprints of sermons by Anglican divines, see Ledger Lomas 2013. ⁶ Oxford, Balliol College Archives, MS 410. Stanley to Jowett, 29 June 1855. Quoted by permission of the Master and fellows, Balliol College, Oxford.

Liddell and Scott in Historical Context

7

1980; Hinchliff 1982). My conclusion is that the motivation of Sewell, and perhaps of Liddell and Scott too, was one of avoidance, and that once the currently controversial realm of theology had been escaped, lexicography could proceed in an autonomous fashion. Thomas Hardy would thus have been more historically accurate had he contrasted ‘sound philology’ with ‘speculative theology’. Greek lexicography was seen as relatively uncontentious—except when it dealt with New Testament Greek. There had been vigorous debates in the 1820s about the role played by theological agendas in the making of dictionaries in that area, for example in that of Parkhurst.⁷ Liddell and Scott have not, however, escaped posthumous denunciation from those who believe that the King James version of the Bible is alone inspired. Gail Riplinger, who belongs to the wilder shores of this movement, has alleged that Cecil Rhodes, who travelled with the Lexicon in Natal in the 1870s, had his faith corrupted by it. For her, in fact, the Lexicon, as the source of later dictionaries used in Bible study, is ‘the whorish mother of all harlot lexicons’.⁸

1.2. THE VERNACULAR TURN One of the most striking aspects of the Preface to the first edition (1843) is Liddell and Scott’s defence of their decision to use English rather than Latin for their glosses and explanations. ‘It may be asked’, they write, ‘whether such a Lexicon should not be in Latin, as in the old times; whether the other is not an unworthy condescension to the indolence of the age.’ Their response distinguishes between a lexicon and notes to classical authors. The latter, they claim, are best couched in Latin, which has an established technical vocabulary and is universally understood; English, however, is far better equipped to render the ‘richness, boldness, freedom, and variety of Greek words’. They conclude that ‘A Frenchman may have reason for using a GreekLatin lexicon; an Englishman can have none’ (LS¹, iii). Their distinction between lexicons and commentaries constitutes an intervention in a contemporary debate about the use of English in classical books. This was to become common over the next two decades but in 1843 was controversial, denounced by conservatives as a surrender to modernity and populism. The controversy ⁷ Parkhurst’s lexicon of New Testament Greek (1769) was Greek English, and his preface makes clear his reformation agenda; the preface to the large scale revision of 1829 by the conservative theologian Hugh James Rose is heavily critical of Parkhurst’s theological/philological views. The first Greek English lexicon published in Britain (a lexicon of NT Greek) was that of Thomas Cokayne of Corpus Christi College, Oxford; it appeared in 1658, twenty years after his death. On Cokayne’s and Parkhurst’s lexicons, see Lee 2003, 88 95. ⁸ Riplinger 2008, 208, 83. The quality of her scholarship is evident from her belief that the Lexicon was the first ever Greek English lexicon, and that ‘football’ derives from ‘Baal’.

8

Christopher Stray

was largely focused on the language used in schoolbooks, and Liddell and Scott’s preface aims to deflect potential conservative objections by distinguishing between this genre and that of lexicons. They bolster their case by listing English precursors, including the Cambridge classical scholar Charles Blomfield, by this time Bishop of London, and Alexander Nicoll, a former professor of Hebrew at Oxford (LS¹, iv n.a).⁹ Liddell and Scott go on to refer to the Greek-German lexicon on which they based their work, that of Franz Passow (first edition 1819–24). It was Passow who had urged that a dictionary entry for a word should tell its history, a principle adopted not only by Liddell and Scott but also later on by James Murray for the OED (Considine 2015). They also refer to the earlier book by Johann Gottlob Schneider on which Passow had drawn (last edition 1819). Though they do not make the point explicitly, this represents another justification for their decision to use the vernacular: that their German predecessors, working within the dominant European scholarly formation of the era, had followed the same path. The shift to the vernacular in Germany formed part of a wider movement in Europe involving changes of fashion in publishing and the emergence of large-scale dictionaries fuelled by ideologies of romantic nationalism.¹⁰ Liddell and Scott’s choice of the vernacular belonged to a wider revalorizing of English and Englishness which led to the exploration of regional dialects, the study of Anglo-Saxon by John Kemble and others (Frantzen 1990, 50–61), and the celebration of Shakespeare as a national treasure (Taylor 1990, 162–230).

1.3. THE LEXICON AND OTHER DICTIONARIES Liddell and Scott made no bones about basing their book on that of Franz Passow, or about discussing this in their preface. It was common for lexiconmakers to select as source a dictionary whose author was dead, and it may well have been Passow’s death in 1833 which prompted David Talboys, who had published (and himself translated) several German works, to approach Liddell (or Scott) to make the Lexicon. Passow’s name remained on the title page until the fourth edition, when Liddell and Scott justified its removal by referring to

⁹ It is worth noting that politically or religiously radical precursors like Gilbert Wakefield, who had planned a Greek English lexicon in the 1790s, and John Jones, whose pioneering lexicon of 1823 has already been referred to, are not mentioned in the 1843 preface (see Stray 2010b, 102). ¹⁰ For publishing, see Febvre and Martin 1997; for dictionaries, Hass 2012; cf. Leerssen 2006, 200 1. The nationalist current also interacted with the tradition of academy dictionaries which had begun in the early seventeenth century (Considine 2014).

Liddell and Scott in Historical Context

9

the substantial expansion and revision they had carried out; they pointed out in their preface that Schneider’s name had disappeared from the fourth edition of Passow’s lexicon for the same reason. The other lexicon which must have loomed large in Liddell and Scott’s thinking, however, is not named in their preface. In declining to list other Greek-English lexicons, on the grounds of lack of space and a concern to avoid invidious comment, they nevertheless remark that ‘the most popular of these Lexicons now abroad’ closely resembles their own, but that this is because both books draw on Passow—though their rival has made ‘slow and scanty acknowledgment of the amount of his debt’ (LS¹, iv). The allusion is to the lexicon of the Irish physician James Donnegan, whose first edition had appeared in 1826 (Donnegan 1826); a fourth edition was published in 1842.¹¹ Donnegan’s book, which had been revised and enlarged in each edition (first edition 1148pp, fourth edition 1743pp), was the market leader in Britain when Liddell and Scott’s first edition came out.¹² Donnegan’s book was based on the Greek-German lexicon of I.J.G. Schneider (1797–8), unlike two other Greek-English dictionaries which appeared in 1826, which were both based on the 1654 Greek-Latin lexicon of the Dutch scholar Cornelis Schrevel.¹³ As can be seen, Liddell and Scott take the moral high ground, but an earlier draft of their preface had included a highly critical examination of Donnegan’s book which took up almost half of their preface.¹⁴ They begin this by declaring that it was ‘insufficient’ for both beginners and scholars because the arrangement of material was ‘random, disorderly and perplexed’ and the book was inaccurate throughout. They go on to consider several samples from Donnegan’s lexicon, of which I quote one as an example. Having noted twenty-five false references in Schweighaeuser’s lexicon to Herodotus, they checked them in Donnegan’s text, and found that only one

¹¹ Little is known about Donnegan, an Irishman who received an MD in Edinburgh in 1809 and probably died in the early 1840s. In the 2nd edition of his book (1831), he acknowledges help from Karl Hase, who was beginning to assemble a new edition of Estienne’s Greek thesaurus in Paris (see Petitmengin 1983). ¹² His 3rd edition (1837) is referred to in a contemporary satire as the ‘costing a pound more Done Again Lexicon’ (Anon. 1837). ¹³ Schrevelius’ Greek Lexicon, translated into English (London: Baldwin & Cradock, 1826) included a Latin Greek dictionary; it was probably edited by John Richardson Major, whose name appears in the 2nd edition of 1831, where the Latin Greek dictionary of the original replaced by an English Greek section. The Greek Lexicon of Schrevelius, Translated into English, with Many Additions (Boston: Cummings and Hilliard, 1826) was edited by John Pickering and Daniel Oliver. George Dunbar and Edmund Barker’s A Greek English, English Greek Lexicon (Edinburgh: Maclachlan and Stewart, 1831) was also based on Schrevel. ¹⁴ This reached proof stage, and survives in the OUP Archive (OUP/PB/12954). Quotations from the Archive are given by permission of the Secretary to the Delegates of Oxford University Press. When Aarsleff (1983, 252) referred to Liddell and Scott’s ‘admirably brief 6 page Preface’, he was presumably unaware of the suppressed preface with its lengthy denunciation of Donnegan.

10

Christopher Stray

had been corrected by him. They go on remorselessly to pile up evidence against their hapless rival, adding that most of the errors were present in his second edition of 1831, but had still not been corrected in his fourth (1842).¹⁵ In the draft preface, Liddell and Scott follow their declaration that an Englishman has no reason to use a Greek-Latin lexicon (quoted above) with the following words: Nor is this a mere opinion of our own. A Greek English Lexicon has been demanded often and by high authorities; has been undertaken more than once by able scholars; nay, there at present several before the world which command a large sale. And this brings us to a second objection. 2. For, it may be asked, are there not Greek English lexicons enough? Is not Dr. Donnegan’s in its Fourth Edition? Why write another? To answer this objection will be an invidious task. But we will not therefore decline it.

In the published preface, the last two sentences are replaced by these: It might be expected that we should here take such notice of these Lexicons as to justify our adding another to the list. We could easily do so. But at this time and in this place we decline the task; partly because it is an invidious one, and might be attributed to other motives than serving the cause of truth and good Scholarship . . .

After adding that they also reluctant to overload their preface with a detailed critique, Liddell and Scott take a parting shot at Donnegan, though without naming him: if in the most popular of these Lexicons now abroad, there are found resem blances to ours . . . , the reason hereof is that we have both been indebted to Passow, though the Author of the Lexicon we allude to has made slow and scanty acknowledgment of his debt.

What are we to make of the somewhat cryptic phrase ‘at this time and in this place’? I suggest that this is a veiled allusion to the Tractarian controversy raging in Oxford in 1843: neither the suppressed nor the published preface is dated, but it is unlikely to have been composed much before publication in the summer of 1843. Donnegan’s name was Irish, and so might have suggested that he was a Catholic; sufficient reason perhaps to avoid a controversy which could be glimpsed only in Liddell and Scott’s use of ‘partly’.

¹⁵ J. Enoch Powell also had a poor opinion of Schweighaeuser’s lexicon, which he called ‘a pretence of a lexicon . . . I have counted more than twelve hundred words used by Herodotus which it omits . . . ’: J.E. Powell, A Lexicon to Herodotus (Cambridge, 1938), viii.

Liddell and Scott in Historical Context

11

1.4. THE DEVELOPMENT OF THE LEXICON THROUGH I TS EIGHT VICTORIAN EDITIONS ‘Lexicography is an endless task’. (Robert Scott to Wilhelm Dindorf, 9 April 1862¹⁶)

After contracting with Talboys, the two editors divided up the essential tasks. Passow had based his own book on a reading of Homer, and by the time of his death in 1833 had extended its coverage only to post-Homeric epic and Herodotus, the latter rather thinly. Liddell and Scott began by working on Herodotus and Thucydides, taking one author each. They worked on the book together almost every day; looking back in the 1890s, Liddell remembered that ‘Robert Scott (then Fellow of Balliol) used to come to my rooms in Christ Church every evening at 7 o’clock to work at our Greek lexicon.’¹⁷ Liddell stayed in Oxford till his marriage in 1846 obliged him to give up his fellowship; he returned as Dean of Christ Church in 1855. Scott had married and left Oxford in 1840 for a rural living, returning on his election as Master of Balliol in 1854. The final stages of work on the first edition were thus probably carried out largely by Liddell, while Scott held the Balliol living of Duloe in Cornwall (1840–50). In his 1898 poem ‘Liddell and Scott, on the completion of their lexicon’, discussed above, Thomas Hardy has Scott say: ‘I almost wished we’d not begun. Even now, if people only knew My sinkings, as we slowly drew Along through Kappa, Lambda, Mu, They’d be concerned at my misgiving, And how I mused on a College living Right down to Sigma, But feared a stigma If I succumbed, and left old Donnegan For weary freshmen’s eyes to con again.’¹⁸

In fact, it was the letter π which most alarmed Liddell: in July 1842 he wrote to Scott (Fig. 1.1.): You will be glad to hear that I have all but finished Π, that two legged monster, who must in ancient times have worn his legs a straddle, Λ else he could never ¹⁶ Robert Scott letter books, OUP Archive. ¹⁷ Oxford, Balliol College Archives. Jowett Papers IIA.21.95. Undated memories of Jowett, written after Jowett’s death in 1893. Quoted by permission of the Master and Fellows, Balliol College, Oxford. ¹⁸ Hardy had bought a second hand copy of the second (1832) edition of Donnegan’s lexicon in 1858, at a time when he was a poorly paid apprentice architect; as an old copy of a superseded book, it must have been much cheaper than the current (1855) edition of Liddell and Scott, which was selling at 31s. 6d.

12

Christopher Stray

Fig. 1.1. The monstrous letter Π, depicted by Henry Liddell in his letter to Robert Scott of July 1842 (Thompson 1899, 75). have strode over such enormous a space as he has occupied and will occupy in Lexicons. Behold the monster, as he has been mocking my waking and sleeping visions for the last many months.¹⁹

David Talboys died in May 1840, and in October that year OUP took the book over; Liddell and Scott were given £500 each on account, and in the following month they were given an additional £150 for translating Passow’s German, and promised £1.10s per sheet for correcting proofs.²⁰ In the process of revision, Liddell and Scott were helped by several other scholars, most of them outside Oxford. In Germany, Karl Wilhelm Dindorf of Göttingen, a prolific scholar who had been publishing with OUP for some time, became involved at an early stage, and supplied large amounts of material for revision (Stray 2013b, 447–8). He and his brother Ludwig August had been recruited in the 1830s to work on the new edition of Estienne’s Thesaurus Linguae Graecae, published by the Paris firm of Didot from 1831 to 1865, and thus had substantial material to offer (see Petitmengin 1983). In Scotland, William Veitch, whose Greek Irregular Verbs was published by OUP in 1866, also sent comments. For later editions, material was received from leading American scholars, notably Henry Drisler (Columbia), William Goodwin (Harvard), and Basil Gildersleeve (Johns Hopkins). In Oxford, Liddell’s ex-pupil George Marshall of Christ Church checked all the references for the first edition, and also produced the Abridged version (1843).²¹ Another important figure was the Press’s reader Philip Molyneux, whom we shall meet later on. ¹⁹ Thompson 1899, 74 5. The letter π occupied 221 pages in the first edition, second only to α (236pp). ²⁰ £500 is the equivalent of £53,000 at 2018 values. ²¹ It is not clear when the decision was made to assemble this. The order to print 3000 copies of the larger lexicon was given by the OUP Delegates on 25 March 1841; on 7 May, they were

Liddell and Scott in Historical Context

13

Through the second half of the nineteenth century a cycle of revision, publication, and distribution was developed, Liddell keeping a watchful eye till his death in 1897: ‘from its first appearance in 1843 interleaved copies always lay on a high-standing desk in his study for continuous amendment and correction’.²² In Liddell’s lifetime eight editions were published, as Table 1.1 shows.²³ Liddell and Scott spent a considerable amount of time working on revised editions, though Scott became less involved after his appointment to the chair of Biblical exegesis in 1861.²⁴ The history of the revision process can be traced from their prefaces. In the brief Advertisement to the second edition, they regretted that the previous edition had sold so fast that they had been able to make very few changes. They had, however, drawn on Pape’s Greek-German lexicon, though they complained of its ‘countless false references’.²⁵ The third edition carried no additional preface. The fourth edition was the first to incorporate substantial revisions and enlargement, including material on ‘comparative etymology taken from Pott’s Etymologische Forschungen (1833–6)’.²⁶ As can be seen from Table 1.1, the pagination was slightly lower, this made possible by an increase in page size. Liddell and Scott remarked that this expansion explained the disappearance of the ‘apologetic tone’ of their first preface. The fifth edition, they declared, had been ‘very much improved’, largely through consulting the Greek-German lexicon of Rost and Palm (Passow 1841–57). The increased print run for this edition also gave Liddell and Scott breathing space for revision, by Table 1.1. Editions of the Lexicon. Edition

date

pp.

copies

price

1 2 3 4 5 6 7 8

1843 1845 1849 1855 1861 1869 1882 1897

1586 ?* 1623 1617 1644 1865 1776 1776

3000 6000 6000 8000 10,000 15,000 15,500 15,000

42s 42s 42s 30s 31s 6d 36s [standing type] 36s [electrotype: 1882, 1885, 1890] 36s [electrotype: 1897, 1901, 1908, 1922, 1928]

* This edition was not paginated.

shown a specimen of the Abridged Lexicon and ordered 2000 to be printed. Marshall also compiled an even smaller book, the Copious Greek English Vocabulary (Marshall 1850). ²² Anon.1899a, 241. ²³ To give an idea of modern equivalents: the 2018 price of the first edition of 1843 would be about £235; of the eighth edition of 1897, about £125. ²⁴ Scott to W. Dindorf, 9 April 1862 (Scott letter books, OUP archive). ²⁵ In a footnote, Liddell and Scott suggested that Pape’s handwriting was to blame, his 3, 5, and 8 being indistinguishable, and also his 1 and 4. ²⁶ Liddell apparently asked Max Müller to work on the lexicon’s etymologies, but Scott disap proved and so the suggestion was blocked: see Anon. 1898. Scott’s objection may have stemmed from his conservative political and theological views, which were opposed to those of Max Müller.

14

Christopher Stray

prolonging the interval before they needed to go to press. Each edition took on average two to three years to print, during which period, though they might be collecting new material, they had also to correct proofs. The increased pagination of the sixth edition of 1869 indicates how much new material was incorporated; the seventh shows a distinct decrease, but was printed on larger pages. Liddell and Scott, who had both entered their seventies, wrote in their preface that this edition was ‘the last which we can hope to see published’; they saw it as the final and definitive edition. The revision process was a delicate one, involving not just semantic definition, exposition, and organization, but also composition and page layout, type and paper selection, proof-reading, and decisions on print runs, pricing, advertising, and marketing. Some of the issues involved can be seen from Liddell’s correspondence with Bartholomew Price, Secretary to the OUP Delegates, about the preparation of the seventh edition, published in 1882, thirteen years after its predecessor. By this time attempts to use other readers had proved unsatisfactory, and Liddell prepared all the proofs himself and then passed them to the OUP compositors. In 1877 a deal was struck with Harper and Brothers of New York, who had been publishing American editions, that these would be suppressed in favour of Oxford editions exported to Harpers in sheets and bound and sold by them.²⁷ This arrangement brought to an end a long history of unauthorized reprinting by the Harpers. In 1846 they had published a version of the Oxford first edition, edited by Henry Drisler of Columbia College, who had provided additional material, including proper names.²⁸ Drisler’s version sold well, not just because of its additions, but because its price was less than half that of the imported copies of the Oxford edition. Over the next thirty years, while Liddell and Scott brought out revised editions, Harper simply reprinted their 1846 book.²⁹ By the late 1870s, when they opened negotiations with OUP, their advantage had been significantly reduced, since their own book had been outpaced by the updated Oxford versions. Because of the uncertainty that the deal with Harpers created about the extent of future sales, the Delegates decide to electrotype the next edition. This meant that the printing plates would be durable and unit costs lower, but the text would not be revisable except in small details.³⁰ The revision process

²⁷ These were technically not ‘pirated’ editions, as international copyright did not exist. Some English authors, however, managed to make gentlemen’s agreements with American publishers. ²⁸ As his preface reveals, the second English edition reached Drisler just in time for him to incorporate Liddell and Scott’s changes in his own book. ²⁹ After Liddell and Scott changed their entry for the theologically sensitive word βαπτίζω (on which, see James’s chapter), Drisler altered his own entry on the original plates to follow them. ³⁰ As a result, the eighth edition of 1897 incorporated only the minor changes which could be made on the plates, and relegated other changes to a short appendix.

Liddell and Scott in Historical Context

15

preceding electrotyping thus became crucially important. In July 1878 Price wrote to Liddell: I send herewith copies of the first 4 pages of the Lexicon, made up to the proposed size, and printed on paper of the proposed size viz. the ordinary demy 4to; also with the new type and letters altered as suggested. It has been composed three times with the Copy, the last time by Molyneux, and he has also read it ‘by the eye’, i.e. as an Editor would . . . ³¹

Three years later, the revision was still in progress, and Price wrote in October 1881: I am sorry that so few sheets of the lexicon should have reached you of late the main reason for the short supply has been the illness of Molyneux who caught a cold which settled on his stomach and entirely incapacitated him for work. He has however returned this morning, but he looks very ill and weak altho’ his Doctor pronounces him well. We find him nearly indispensable to the progress of the lexicon. The [ . . . ] stock of the 4to lexicon now consists of only 650 copies, so that we must print the new edition as quickly as possible.³²

As this suggests, in the ongoing and not always predictable cycle in which one edition was being sold as the next was bring prepared, a check in the warehouse sometimes led to alarm and acceleration. If copies ran out before the revised edition was ready, sales and reputation suffered. If a new edition was issued before stocks of its predecessor were exhausted, a lot of books suddenly became almost unsaleable. For the early editions, the type had been redistributed after publication and set up anew for the next edition. The first change in this system came with the sixth edition (1869), for which the type was left standing. This cut costs by facilitating speedy reprinting, and allowed for revision, as long as it was not radical. This was in fact planned to be the final revision, and the increased print run of 15,000 was designed to last for eleven years (Thompson 1899, 78). The process of revision was complicated by the parallel production of two smaller versions of the Lexicon. The Abridged Lexicon (‘chiefly for the use of schools’), published just after the first edition of the Lexicon in 1843, was later revised several times on the basis of revisions to the larger Lexicon, most recently in 1901. An Intermediate Lexicon, prepared by Liddell, appeared in 1889. This was based on the seventh edition of the Lexicon (1882), which was intended at

³¹ Secretary’s Letter books, Price to Liddell, 5 July 1878. The reference is probably to reading for sense, rather than letter by letter, as a compositor would. ³² Ibid., Price to Liddell, 3 October 1881.

16

Christopher Stray

the time to be the final edition. The Intermediate Lexicon, which like its smaller sibling is still in print, has never been revised.³³

1.5. THE L EXICON AS A BOOK The account of printing and marketing given above reminds us that the Lexicon was a book as well as a text. Historians of lexicography have usually concentrated, understandably, on the texts of dictionaries, rather than on the material and commercial realities of making the book which carries the text (but see Considine 2017). The core activity in the making of the Lexicon was of course a matter of exploring classical usage, analysing it, classifying usages and putting that all on paper. But the paper it was put on had to be made or bought, printed on, bound, distributed, and sold; and the type used for printing selected, made or bought, and set up for use, and then redistributed, kept standing, stereotyped, or electrotyped.³⁴ The bolded headwords which must have been in 1843 such a startling part of the look of the Lexicon’s pages were printed with type which had to be chosen, bought, or made. Boldlooking (fattened face) lemmata were used in the fifth edition of the French Academy’s dictionary (1835), but that was in roman type. A supplement to the previous edition, published in 1827, had also used such headings (Twyman 1993). In Britain they were almost unheard of in the 1840s, when bolded type (‘Clarendon’ or ‘Ionic’) was only just appearing in the printing of roman fonts (Twyman 1997). Liddell and Scott’s bolded headwords were probably the work of David Talboys. Some of his other publications have inventive type and layout, notably his Oxford Chronological Tables of Universal History (1835), which makes extensive use of typographic differentiation (roman, italic, bolding). Another source of influence was Liddell and Scott’s teacher Thomas Gaisford. Hugh Lloyd-Jones’s pronouncement that Gaisford ‘was not interested in typography’ (Lloyd-Jones 1982, 99) is undercut by substantial evidence that he was a hands-on Delegate who had a long-term concern with ‘the look of the book’ (Stray 2018, 53–81). The bolded Greek in the Lexicon, the first to be used in ³³ In the twentieth century the Press took the view that the book, though not ideal, was good enough for its readers in the upper forms of schools. Proposals for a new edition of the Intermediate Lexicon, urged by John Chadwick, were dropped in the 1990s. The project then passed to Cambridge; the Cambridge Greek Lexicon is planned to appear in 2019. ³⁴ Standing type was held in formes (cases) and stored ready for reuse. Type was kept in this way if a demand for reprinting was expected, and if the type was not needed for other books. This strategy allowed for revision, as the type could be reset to a certain degree. Stereotyping and electrotyping involved the making of plates from which reprints could be made; revision was still possible, but on a much more limited scale. Stereotyping was used at OUP from the 1800s; electrotyping, a later invention permitting a more finely detailed impression on the page, from about 1868.

Liddell and Scott in Historical Context

17

Britain, was one of two such founts used by the Press, both of which were known within the Printing House as ‘Gaisford type’.³⁵ In January 1824, Gaisford had already been already thinking of how to make his books user-friendly. Writing to his friend Henry Fynes Clinton, he explained about his edition of Suidas³⁶ that: my book will make . . . two volumes in folio . . . the Latin version will not appear, being as I conceive, useless, or nearly so. The size, or rather the height of the page will be shorter considerably than that of Kuster’s edition;³⁷ I propose to use what is called foolscap paper a very convenient form, much used by the early printers. I shall introduce, what I cannot help but think will present many facilities to those who may consult the lexicon, a method of printing the glosses which has not hitherto been adopted in any edition of a Greek lexicon. The heads of the glosses, as they are technically termed, i.e. the word or phrase to be explained, will be printed in small capitals the explanation in the ordinary character, & the example in another line and in a less character. In short the book will wear an appearance somewhat similar to a modern, say Johnson’s, dictionary.³⁸

Successive revisions led to the gradual enlargement of the text. The number of pages increased from 1586 in the first edition to 1776 in the eighth, and page size was enlarged in the fourth and sixth editions. These changes in turn led to the adoption of thinner paper to make a book which was strongly bound and manageable by the reader. In the past, such matters have rarely been considered in studies of academic writing; yet a full understanding of a book like the Lexicon is impossible without bringing them into relation with the study of its content.³⁹ The list of prices given above for the Lexicon’s successive editions reminds us that its assembly as both text and object involved payments to its editors and those involved in its physical production, and that some or all of these expenses were recouped by the press through sales. The Lexicon was in fact one of the two books which were the mainstay of OUP’s learned publishing in

³⁵ ‘List of ancient and modern Greek and oriental types in use at OUP’ (1959: copies in OUP archive and at St Bride Printing Library); cf. Bowman 1992, 58 60; 1998, 156. ³⁶ The name of the compiler of this tenth century encyclopedic lexicon is unknown; ‘Suidas’, or more commonly nowadays ‘the Suda’ (‘fortress’) is its title. ³⁷ This is the Cambridge edition of 1705; Gaisford’s copy contained annotations by Jonathan Toup. ³⁸ Gaisford to Clinton, 16 February 1824. Christ Church library, Ms 498. The American editions of Liddell and Scott did not have bolded headwords. W. Freund’s Latin dictionary (1834 45) has no bolding; headwords are mostly lower case, with some entries in capitals. Andrews’ (1850), based on Freund, has bolded headwords, as does William Smith’s (1855); in 1856 Andrews’ British publishers Sampson Low claimed that Smith had copied their typography (Stray 2007c, 48). ³⁹ For examples within the history of classical scholarship, see Kenney 1974; Stray 2007b; for vernacular dictionaries, Considine 2017; for the publishing history of historiography, Howsam 2009.

18

Christopher Stray

the mid-nineteenth century, the other being Wordsworth’s Greek grammar, (Sutcliffe 1978, 12; Stray 2013b, 450–7; Stray 2016). The Lexicon fulfilled the major criterion for a profitable book: it was expensive and was bought in large numbers. The Order Books of the OUP Delegates list a series of payments to Liddell and Scott, some of them very large: £400 in 1841, £1000 in 1845, £600 in 1846, £1500 in 1849. In 1861 they were given £1300 as half-profits on editions of the Abridged Lexicon, and £2750 in instalments over the next five years for their work on the new (sixth) edition of the large book. Most of the work on the seventh edition was done by Liddell, Scott having left Oxford in 1870; in 1882, the year of its publication, he was paid £1000 for preparing the book and £650 for proofing the text. These large sums were, in his case, superadded to an income as Dean of Christ Church of about £3000 a year.⁴⁰ The Deanery at Christ Church contains a splendid wooden staircase Liddell installed there, paid for by his profits: it is known as the Lexicon Staircase.

1 . 6 . ‘O U R TW O F R I E N DS’ At the end of 1869, the Delegates awarded the two men £600 as a partpayment for profits on the sixth edition. Scott had contributed relatively little, and asked to be paid less than £300, but Liddell wanted the sum to be divided equally. As this indicates, Liddell and Scott’s relationship stands as an exemplary case of lexicographical amity. Their friendship and collaboration led to the indissoluble linking of their names, and as Peter Sutcliffe said in his history of OUP, ‘they went hand in hand into eternity’ (Sutcliffe 1978, 11). The friendship is the more striking because the two men held very different political views, Liddell being a Liberal and Scott a Tory. Both men also belonged to other partnerships, since both were married: hence Scott’s departure from Oxford in 1840 and Liddell’s in 1846, marriage having ended their college fellowships. The effect of preferment on scholarly work was considerable in this period. Thomas Gaisford, who encouraged Liddell and Scott to assemble the Lexicon, had managed to avoid preferment almost entirely, and as Dean of Christ Church was uniquely free for his classical work, as ruler of an institution which was both college and cathedral (Stray 2018, 53–81). Both the authority of the Lexicon and the affection with which it was regarded were disseminated as it circulated in the abridged version (1843) and the intermediate version (1889) as well as the eight full editions of Liddell’s lifetime (1843–97). A pupil could progress from the abridged, ⁴⁰ Both income and payments can be followed in Liddell’s (and Scott’s) bank records, now held at the Barclays Group Archives, Wythenshawe, Manchester.

Liddell and Scott in Historical Context

19

through the intermediate, to the full Lexicon. It was this last to which, in the last quarter of the nineteenth century, Edmund Morshead of Winchester College used to refer in his classroom as ‘our two friends’.⁴¹ The collaborative nature of the Lexicon provoked discussion over the division of labour between the two editors. On the neutral ground of Winchester College they could be seen as ‘our two friends’, but when one of them used the book in his own classroom, as Liddell did as headmaster of Westminster 1846–55, the question of who wrote what came to the fore. Hence the well-known doggerel offered by one of his pupils. Two men wrote a lexicon, Liddell and Scott; Some parts were clever, but some parts were not. Hear, all ye learned, and read me this riddle, How the wrong parts wrote Scott and the right parts wrote Liddell.⁴²

Liddell apparently provoked this by his habit, when controverted by pupils who cited the lexicon, of saying ‘Scott wrote that part.’ Second, the clear trend in revisions of the lexicon was that it got not just better, but bigger. New words were found, old words were found to have been omitted. To improve and expand was a scholarly duty, but it posed problems for profitability. In the nineteenth century, the solution was to increase page size so as to control the rising page numbers (the Seventh was about ¾” taller and wider than the Sixth, but down from 1865pp to 1776pp). In the twentieth century, the pages of the ninth edition (LSJ) were reorganized to reduce the number of separate headwords—thus making life harder for the reader (cf. Chadwick 1994, 2). In some cases, existing material was cut to make room for new words or new evidence.⁴³ Finally, the correspondence throws light on the usually obscure process of composition. OUP employed dozens of compositors, but the best of them tended to be confined to jobs suited to their special aptitudes.⁴⁴ Among them was Philip Molyneux, who composed and checked the proofs of the Lexicon from the late 1870s into the 1900s, and whose illness, as we have seen, brought ⁴¹ See Stray 1996, whose cover reproduces a photograph of 1898 showing Morshead teaching, with lexica piled up on his desk. ⁴² Cited in Kitchell 1988. In a follow up to Kitchell’s article (Calder 1989), W.M. Calder III notes that Kitchell’s ultimate source is H.L. Thompson’s biography, and mentions an alternative version preserved by Liddell’s distant cousin, Augustus Hare: ‘Two men wrote a lexicon,/Liddell and Scott;/One half was clever,/And one half was not./ Give me the answer, boys,/Quick to this riddle,/Which was by Scott/ And which was by Liddell?’ (Hare 1896, 10). Calder gives reasons for preferring Hare’s version. Kitchell responded, giving reasons for preferring Thompson’s version (Kitchell 1989). The appearance of the first part of LSJ in 1920 gave rise to a flurry of reported versions, several of which were collected in The Periodical (OUP) 8 (April 1921): 40 1. ⁴³ For example, some examples were cut from the entry for θέλω: see Lee 2010a, 18 n. 12. ⁴⁴ The compositors were organized in small groups known as ‘ships’ (companionships), one of which concentrated on OED, whose composition was begun by three men on 3 June 1882. One of them spent the next thirty years on its successive parts. See Belson 2003, 315.

20

Christopher Stray

printing to a halt. If Liddell, the well-bred, often haughty, and certainly distinguished Dean of Christ Church, was a hands-on reviser and editor for several decades, he met his match and his complement in Molyneux. In Britain, copies of the Lexicon were distributed at first by the Clarendon Press’s partner J.H. Parker of Oxford, and after the partnership was ended in 1863, by Longmans and by Alexander Macmillan, who was publisher to OUP from 1863 to 1880 and who exported copies to the USA for sale by Harpers. Until the USA joined the Berne Convention by passing the Chace Act in 1891, issues of copyright often arose.⁴⁵ In that year, in fact, the Economy Book House of New York began to sell photographic copies of the Abridged Lexicon for $1.25, much less that the price charged by Harper and Bros.⁴⁶ Harper decided not to take them to court but to undercut them by printing from ‘block copies’ (unbound copies, now called ‘text blocks’) and selling at $1.⁴⁷ Material for revision came in from several sources. Readers sent in corrections to the current text. Scholars offered revisions—the Americans Goodwin, Gildersleeve, and Drisler among them. In September 1879, Harper sent Price the final proofs of Gildersleeve’s articles on οὐκ and μὴ οὐ for the seventh edition, promising that ὅπως would follow shortly. The American scholars were not always happy with the treatment of the articles they sent. Reviewing the eighth edition in his editorial ‘Brief Mention’ column in the American Journal of Philology, Gildersleeve commented that: Liddell and Scott were even greater sinners than the average lexicographer, and complaints enough were heard in their lifetime. In the seventh edition they not only kept in mistakes of their own, but spoiled other people’s work, as I pointed out AJP III 515, and my article on μή is no exception to their arbitrary processes.⁴⁸

1.7. AFTER L IDDELL: TWEN TIETH-CENTURY REVISION Liddell’s death in 1898 prompted reflection at the Press: what was to be done next? They had been used to a running process of revision—six editions since 1843, by Liddell alone since about 1870.⁴⁹ Thompson’s memoir of Liddell ⁴⁵ The Act specified that books had to be printed on American soil in order to claim US copyright. ⁴⁶ The Library of Congress has copies (OC PA445.E5 L7 1897/1901) tentatively dated to 1897 and 1901, probably from accession dates. ⁴⁷ A larger scale piracy took place in Greece, where a four volume translation of the 7th edition appeared between 1901 and 1907: Stray 2007b, 107 8. ⁴⁸ Gildersleeve 1898, 233. In the earlier volume to which he referred, Gildersleeve had written of ‘the unfair and inconsiderate manner in which [my] work has been treated’ (Gildersleeve 1882, 515). ⁴⁹ The 7th edition of 1882 was the first to be revised by him alone,

Liddell and Scott in Historical Context

21

might loyally proclaim that the Lexicon ‘has never become out of date’ (Thompson 1899, 65), but it was widely known that this was not so. As the Athenaeum pointed out, ‘It is notoriously so . . . in more ways than one, and a thorough revision, without the limitations imposed on the last recension by “electrotype plates”, is sadly needed’ (Anon. 1899b). Since the seventh edition of 1882, thought of as ‘final’, the prospect of revision had been closed down, just as new words and new evidence had been merging. The matter was made more pressing by the publication of new papyrus material in the 1890s: Aristotle’s Constitution of Athens, Herondas, Bacchylides, and in 1898, the year of Liddell’s death, the first volume of the Oxyrhynchus papyri. In 1903 OUP briefly considered commissioning a Greek equivalent of the Thesaurus Linguae Latinae project, which had begun in 1894 (Flow 2015), but the German scholar Hermann Diels pointed out that the five German academies had sunk so much money into the Thesaurus that they could not contemplate another large project.⁵⁰ The Press therefore decided to commission a new lexicon on roughly the same scale as the eighth edition, and appointed the sixty-three-year old Arthur Sidgwick, Reader in Greek at Oxford, to edit the new edition of the Lexicon. Sidgwick however did not get very far with the task, and when asked in 1911 to report on progress, confessed his failure and resigned.⁵¹ In his place OUP appointed Henry Stuart-Jones, one of the most talented and versatile scholars of his generation. Under his editorship the ninth edition appeared in a series of parts, the first in 1925, the last in 1940, just after his death. From 1920 Stuart-Jones was helped by the comparative philologist Roderick McKenzie, who died in 1937. The editors of Oxford Classical Texts were asked to send in material on their own authors, and foreign scholars, including Ada Adler and Kurt Latte, also sent in material.⁵² Patristic and Byzantine Greek were not included, a decision justified by Stuart-Jones in the preface to the first part (1925) on the ground that a separate lexicon of patristic Greek was in preparation.⁵³ In 1920, five years before the first part of the ninth edition (‘LSJ’) appeared, a letter was published in the Classical Review from F.M. Cornford of Trinity College, Cambridge, asking if there was a permanent body to which corrections and additions could be sent. This drew a response from Stuart-Jones, ⁵⁰ Diels’s views are cited in LSJ, preface p. v. The British Academy itself received no state support until 1924. ⁵¹ Sidgwick’s copy of the 8th edition, now in the library of Corpus Christi College, Oxford, shows no sign of annotation. ⁵² British scholars of varying competence also sent in suggestions. Among them was John Isaac Beare, Regius Professor of Greek at Trinity College Dublin 1902 15, who first wrote to Stuart Jones in 1912; Stuart Jones’s replies are pasted in to Beare’s copy of LS⁸, now at Trinity College, Cambridge (Adv. b.25.16). ⁵³ This had been first mooted in 1906, but was proceeding very slowly. It was eventually completed under the editorship of G.W.H. Lampe and published by OUP in parts from 1961, and as a volume in 1969.

22

Christopher Stray

announcing the imminent appearance of the first part of the Lexicon and welcoming comments.⁵⁴ The editor of the Review, Cornford’s colleague Ernest Harrison, kept a watching brief. Having offered suggestions both in reviews of parts and in private letters to OUP, he allowed a note of discontent to creep into his welcoming review of the final part in March 1941 (CR 55: 28–30): he had been thanked by the editors for his comments, but ‘I must mention a few of the words on which I appear to have written in vain’. Similar complaints were heard from other scholars, and the window of opportunity for corrections was effectively closed when reprinting took place in 1948. In a sense, the state of the Lexicon recapitulated that of the eighth edition of 1897: the logistics and economics of book publishing operated as an obstacle to effectual revision. In the 1950s, however, a small team was appointed to prepare a supplementary volume; this was published in 1968, and a revised version in 1996.

1.8. CONCLU SION Liddell and Scott is not just a lexicon but a literary and cultural phenomenon. It was not long after the appearance of the first edition that references to the Lexicon began to appear in contemporary literature. The first reference, not surprisingly, was in a work written by an Oxford scholar: Arthur Hugh Clough’s Bothie of Toper-na-fuosich (1848), a mock-epic account of an Oxford undergraduate reading party in Scotland, written in experimental hexameters. After a few weeks, one of the undergraduates proposes that they abandon their Greek texts and travel about: Slumber in Liddell and Scott, O musical chaff of old Athens, Dishes, and fishes, bird, beast, and sesquipedalian black guard! Sleep, weary ghosts, be at peace and abide in your lexicon limbo! (lines 401 3)

The ‘lexicon-limbo’ imagined by Clough reminds us of the curious nature of the text of a lexicon, a corpus shaped for reference rather than experience; an artefact reflecting and refracting the original patterns of usage on which it is based. In it the ghosts of a dead language live a half-life which enables them to act as witnesses to their antique selves. The best-known literary reference to the Lexicon comes from 1898, just after Liddell’s death, when Thomas Hardy wrote the light-hearted poem, from ⁵⁴ CR 34 (1920): Cornford, May/June p.79; Stuart Jones, Aug/Sept p. 127. The individual sections of LSJ, like those of OED, were officially referred to as Parts; those of the Oxford Latin Dictionary (1968 82) as Fascicles.

Liddell and Scott in Historical Context

23

which I quoted above, about the completion of the original edition of 1843. As the dialogic structure of Hardy’s poem suggests, part of the book’s charm has always stemmed from its being edited by two friends. Such lexicographic collaboration, amicable or not, was in fact unusual, most dictionaries being single-editor books, except in the sense that they were commonly revisions of earlier works.⁵⁵ The Lexicon’s readers not only used the book as a dictionary, they looked within it for evidence of the humanity of its editors. The search for foible and jest some have conducted is surely driven by the sense of a fallible humanity, and this is only reinforced by the assumption provoked by editorial plurality of a division of labour and of disagreement. The best-known (supposed) jest is the claim s. v. συκοϕάντης (sycophant, informer) that the derivation from legislation forbidding the import of figs into Attica was ‘a figment’. In fact the first four editions call the story ‘an invention’; the fifth and sixth editions substitute ‘a mere figment’; and later editions revert to the original text—perhaps after the sound of unseemly mirth reached the editors’ ears. A writer commenting on Liddell’s retirement as Dean of Christ Church in 1891 referred to ‘the austere gravity of the two Deans’ and added that ‘No one . . . ever ventured in their presence to comment on their explanation of ἄλοχος. The first meaning given is “bedmate”, the initial alpha being explained as “copul[ative]”’ (Anon. 1891, 464). Liddell and Scott was made by Liddell and Scott. In this chapter, I have tried to tease out some of the implications of this editorial plurality, which made their Lexicon both authoritative and yet also in a sense reassuringly fallible. ‘Which parts wrote Liddell?’ looks forward strikingly to John Henderson’s postmodern inversion ‘Juvenal’s Mayor’.⁵⁶ The author (or here, editor(s)) does not disappear, but his/their relationship to the published text is no longer straightforward. To consider the book as the material product of a process of printing and publishing further complicates its identity: the ‘text’ is realized as an impression made by specific types of specific paper, chosen and costed to make a profit for a press which had one foot in its university and another in an international commercial market.⁵⁷ Editorial plurality co-existed, in the history of the Lexicon, with the collaboration of a wider community: not just other scholars, British, German, and American, but also the unsung heroes (Marshall and Molyneux) without

⁵⁵ An exception in this period is the German revision of Passow by Rost and Palm (1841 57). On collaborative editing, see Stray 2020. ⁵⁶ See Henderson 1998, an engaging account of the editor of ‘Mayor’s Juvenal’. ⁵⁷ The printing specifications for LSJ show that the paper stock used was ‘Liddell and Scott’. For a discussion which stresses the instability of printed texts, see McKitterick 2003, esp. pp. 217 30. The tension between OUP’s local and wider identity bases are discussed in Eliot 2013. For a stimulating account of theories of the nature of text, see Greetham 1999.

24

Christopher Stray

whom the published text would not have seen the light.⁵⁸ Yet this was not simply a republic of (Greek) letters: the editors decided what was to be printed, subject to technological constraints, and the evidence cited above shows that neither contributions nor corrections could be guaranteed to find a place in the revised text. The Lexicon is thus a complex phenomenon; it is implicated in questions of lexical meaning and lexicographic technique, but it also belongs to a history of social relations, links with other lexica, economic calculation, and the mythicizing of scholarship.

⁵⁸ Another hero, not sung above, was the young Falconer Madan, who sent in a number of corrections to the 6th edition in the early 1870s. He was invited by Liddell to succeed him as editor, but declined, and went on to become Bodley’s Librarian 1912 19.

2 Dictionaries as Translations English in the Lexicon Margaret Williamson

As one of the editors of this volume has memorably written, any reading of a Greek text is an act of translation. ‘To read an ancient text is to translate it, mapping its words one by one onto the semantic units of our own mother tongue vocabulary’ (Clarke 2010, 120). The scenario implied here, involving reader and text in dialogue, will be the one most readily conjured up for a habitually solitary student or scholar of Greek. But it is heavily mediated by those previous dialogues represented by the tools of scholarship, which we might regard as a kind of secondary speech community for a language no longer spoken. Foremost among these tools for an Anglophone scholar, of course, has been Liddell and Scott’s Greek-English Lexicon. The conversation in which we engage in consulting its glosses is at the very least three-way, involving not only ancient Greek and our own ‘mother tongue’, but also the English of Liddell and Scott. This essay seeks to bring into sharper focus the vocabulary of the first edition’s glosses—a task rendered urgent not only by the passage of time but also by the fact that the editors had an axe to grind in selecting it. The agenda they announced in 1843 remained influential through subsequent editions, even though one of its key elements was almost immediately withdrawn. Some years ago I was approached by a student in an advanced Greek class with a question about a gloss in the ‘Middle Liddell’ (the Intermediate Lexicon). Could I, he asked, explain to him what the English terms ‘huckster’ and ‘higgler’, provided as renderings of the Greek κάπηλος, meant? His bafflement at Liddell and Scott’s distinctive lexical choices has been shared by many of their readers up to and including the present day. It is true that the Middle Liddell has remained unchanged since 1889, when it was abridged from the then-current seventh edition of 1882. But the peculiarity of these glosses cannot be explained solely by their date. The seeds of the archaizing Margaret Williamson, Dictionaries as Translations: English in the Lexicon In: Liddell and Scott: The History, Methodology, and Languages of the World’s Leading Lexicon of Ancient Greek. Edited by: Christopher Stray, Michael Clarke, and Joshua T. Katz, Oxford University Press (2019). © Oxford University Press. DOI: 10.1093/oso/9780198810803.003.0002

26

Margaret Williamson

idiolect of which they are examples were already sown in the first edition, and as we shall see it is very much a product of its time, evoking a nationalist politics of language that seeks to claim superiority for the English language and its speakers as interpreters of ancient Greece. Although Liddell and Scott’s was not the first Greek-English lexicon, the idea of using English rather than Latin (or more recently German) to interpret Greek was still sufficiently novel that they felt obliged to address it in the preface to the first edition. They begin by characterizing Greek in full-blown Romantic vein: at least as deployed in poetry and oratory, it is ‘full of living force and fire, abounding in grace and sweetness, rich to overflowing’ (i). A brief discussion then follows of the merits of two other languages as media for interpreting (by which they clearly mean translating) Greek. Latin, the traditional choice, is judged superior as a medium for critical writing about classical authors on grounds of its universal currency, its brevity, and its clarity. But they consider it ‘feeble and defective for purposes of Lexicography’, because it lacks the ‘richness, boldness, freedom and variety’ of Greek vocabulary. The hint of national pride implicit in this declaration is amplified when they turn to a competing modern vernacular. French is granted the distinction of being the ‘language of Mathematics’, but it too lacks the ‘richness’ and ‘freedom’ that make English ‘not unworthy to compare with the Greek’.¹ Consequently, they conclude, ‘the best lexicon an Englishman can use to read Greek with, will be in English’ (ii). Significantly, this chauvinism about the English language is linked to pride in its literature: French, they declare, is also inadequate to the task of ‘giv[ing] any adequate conception of Milton or Shakespeare’ (iii–iv). Liddell and Scott’s focus on France as a competitor echoes political tensions between the two nations at this period; and the literary interests through which their rivalry is expressed are, as we shall see, a key influence on their linguistic choices. In both respects they differ from other lexicographers such as Donnegan, their main competitor, whose New Greek and English Lexicon first appeared in 1826. Like Donnegan, however, they based their work on that of a German predecessor: in his case Schneider, in theirs Passow, who are respectfully acknowledged as forerunners rather than rivals. A second discernible trend in their use of English is announced a few pages later in the preface, in the following statement: In the Translations of the Greek terms we have been anxious to use genuine Saxon English words, rather than their Latin equivalents. (ix)

¹ Claims of this kind were not new: as Trapp points out, ‘practically every nation in Europe has entertained the notion that its vernacular had a special relationship with Greek’ (Trapp 1971, 239). His examples include Italian, French, Spanish, Flemish, and German as well as English.

Dictionaries as Translations: English in the Lexicon

27

This tantalizing sentence stands alone, with neither preamble nor explanation, and the subject of translation is never referred to again. In order to understand what it means, therefore, we must turn to other sources of the time, as well as to the editors’ own linguistic practice. Liddell and Scott’s celebration of the ‘richness’ and ‘variety’ of English on one hand, and on the other their emphasis on its Saxon, or Germanic, elements, reflect views that were widespread at the time. In a near-contemporary discussion, an anonymous commentator on John Wesley’s sermons offers a standard explanation of the richness of English vocabulary: . . . its copiousness of terms . . . is in great measure the result of the peculiar composition of the English tongue, which, like a river, whose abundant volume of water is made up of two distinct streams, flows partly from a Saxon and partly from a Norman source.²

Many others echoed both the pride and the explanation for it. In 1839, just before the Lexicon’s publication, Henry Hallam countered opposition to the ‘forced introduction of French words’ by Chaucer and the ‘harsh latinisms’ this entailed by declaring that they have ‘given the English language a copiousness and variety that perhaps no other language possesses’ (Hallam 1839, in Matthews 2000, 207). A few years earlier, Macaulay had argued for the use of English, rather than any of the native vernaculars, in the Indian Empire on the grounds that English language and literature had been rendered superior to any native Indian language by their encounter with classical languages.³ Yet at the same time, an influential strand of thinking tended to privilege the Saxon, or Germanic, element of the language as the more fundamental, and more authentically English, of the two—a view that both predated and outlasted Liddell and Scott. For writers in earlier centuries such as Hare (1647) or Pinkerton (1786), the arrival of other languages, especially Norman French, marked the adulteration of a previously pure language, and even as late as 1871 John Earle lamented the ‘dissolution of the old cultivated language of the Saxons’ under the impact of the ‘sinister and ill-favoured words’ imported from French (Matthews 2000, 43, 45). Even R.G. Latham, one of the more philologically sophisticated analysts of the time and influenced by continental philologists such as Jacob Grimm and Rasmus Rask, was not immune to it. Although he gives full weight to the various components of the language, he is also capable of evoking binary oppositions remarkably similar to those of Liddell and Scott, in which an original ‘SaxonEnglish’, or some variant thereof, is distinguished from later additions. In his fourth chapter, on ‘Relations of the English to the Anglo-Saxon Language’, he characterizes Anglo-Saxon as foundational, describing it as ‘less . . . an element of ² ‘Plain Words in the Pulpit’, The Wesleyan Methodist Magazine, no. 16, February 1837: 108 11. ³ Momma 2013, 96 9.

28

Margaret Williamson

the present English than . . . the Mother-Tongue; upon which a few words or phrases from other languages have been engrafted’ (Latham 1841, 41). The bias displayed here had more than one determinant. For many, questions about the primacy of Saxon flowed from debates about English national identity and its origins. As Christopher Hill showed in a classic essay, the idea of the Norman Conquest as an unwelcome intrusion on native political freedom and self-determination had been a rallying-cry for political activists of all stripes since the seventeenth century (Hill 1958, 50–122). This characterization was not always uncontested, nor was the Conquest the only key moment: the Magna Carta was another contender.⁴ But the Middle Ages, and in particular the Anglo-Saxon inheritance, remained key elements in the nation’s self-definition. And although the image of the Norman Yoke as a weight on the necks of previously free Anglo-Saxons had lost some of its charge in political theory by Liddell and Scott’s time, it came to renewed life as a literary and historiographical idea in the nineteenth century, in the context of the Gothic Revival. The influence of this movement can be traced at all levels of cultural life, from architecture to historiography, literature, and theatre. Two historical works that reflect its importance, bookending the production of the Lexicon, are Sharon Turner’s The History of the Anglo-Saxons (1799–1805) and J.M. Kemble’s The Saxons in England (1849). Writing in (of all years!) 1848, Kemble in his preface introduces his history with an explicit declaration of the importance of the Saxon past for the English present. Saxon history is, he says, ‘the history of the childhood of our own age—the explanation of its manhood’ (v), and he goes on to attribute the stability which Britain, unlike other nations, is enjoying to the institutions inherited from the Saxons. In literature, we might single out the historical novels of Walter Scott, whose 1819 novel Ivanhoe drew on the work of Turner, and the popularity of plays on historical subjects, including antiquarian stagings of Shakespeare.⁵ In the context of language and literature, the political dimension of ‘Saxon English’, though less marked, is still discernible. Almost a century earlier, for example, Samuel Johnson in the preface to his Dictionary had framed the balance between the two components of English thus: I have studiously endeavoured to collect examples and authorities from the writers before the restoration, whose works I regard as the wells of English undefiled, as the pure sources of genuine diction. Our language, for almost a century, has . . . been gradually departing from its original Teutonick character, and deviating towards a Gallick structure and phraseology, from which it ought to ⁴ In addition to Hill 1958, see Burrow 1981, 97 125 and Simmons 1990. ⁵ See Schoch 1998, ch. 4. Other historical novels celebrating resistance to the Conquest include Bulwer Lytton’s 1848 Harold, Last of the Saxon Kings and Charles Kingsley’s 1866 Hereward the Wake. Scott’s Waverley (1814) had also deployed Homeric parallels.

Dictionaries as Translations: English in the Lexicon

29

be our endeavour to recall it, by making our ancient volumes the groundworks of stile . . . (Preface, Johnson 1755)

Johnson takes the composite origins of the language as a given, but he regards the balance as having tilted too far away from its Teutonic, or Germanic elements and towards the Gallic, which includes the Latinate.⁶ The nationalism implicit here anticipates not only the Anglo-French rivalry expressed in the first edition’s preface but also the polemical thrust of their emphasis on ‘genuine Saxon-English.’ In many sources close to Liddell and Scott’s time, however, the discussion is also framed in more purely aesthetic terms. While some linked Saxon English style with political freedoms,⁷ others valued it for its intrinsic merits, for the ‘simple majesty’ of a way of writing in which Romance and Latinate loanwords, while not absent, did not predominate.⁸ For Wesley’s anonymous admirer in the Wesleyan-Methodist Magazine, this style is to be preferred not only because it is (appropriately for a sermon) intelligible to all, but also for its force, beauty, and musicality. And although terms of Latin origin are not excluded, the approved style will avoid Latinisms in favour of ‘good old Saxon English’, even ‘words purely Saxon in their origin’. He illustrates the difference between this and an excessively Latinate style with reference to passages from the King James Bible, and hammers the point home by translating one such passage ‘literally’ into ‘modernized’ language, transforming: If God so clothe the grass of the field, which to day is, and tomorrow is cast into the oven, shall he not much more clothe you, O ye of little faith?

into: If the Deity so adorn the vegetable productions, which continue for a very limited period, and are subsequently applied to the most ordinary domestic uses, will he not provide you with the necessary adornments? &c.

This is, of course, to take the contrast to a parodic extreme; yet the features it highlights—the simplicity of the original as against the florid and polysyllabic ⁶ Ironically, Johnson himself came to be criticized for excessive Latinity: for example ‘good Saxon English’ is contrasted with ‘the sesquipedalia of Johnson’ in Fraser’s Magazine, March 1838. Reprinted in the Preston Chronicle (Preston, England), Saturday, 10 March 1838; Issue 1332. ⁷ In 1844 the Hull Packet and East Riding Times, covering Lord John Manners’ dinner address to the Birmingham Athenic Institution, reports that he commended ‘good, sound, plain Saxon English’ to his audience, and added ‘nothing was better . . . than plain, simple words; they alone were suited to the manly spirit of a free people’, 6 September 1844, issue 3116. The Institution had been founded by Christian Chartists in 1841 to promote intellectual culture and ‘the enjoyment of old English manly sports by respectable young men’: Addresses delivered at the Birmingham Athenic Institute [sic] . . . in 1844 (London, 1845), 3. ⁸ ‘Saxon English’, reprinted from the Quarterly Review in The Mirror of literature, amusement and instruction 30 (1838), 93.

30

Margaret Williamson

Latinity of the paraphrase—find echoes in Liddell and Scott’s lexical choices, to which we now turn. The term with which we began—κάπηλος—may serve as an example. It is used first, according to them, in Herodotus, and thereafter mainly in prose, and the full 1843 entry reads: κάπηλος, ὁ, (κάπτω, κάπη) strictly one who sells provisions: then any retail dealer, petty tradesman, huckster, higgler, Lat. caupo, propola, Hdt. I.94, etc.; opp. to ἔμπορος, the whole sale dealer, importer, Lat. mercator, Xen. Cyr. 4.5.42: esp. a tavern keeper, publican: frequ. in compds., βιβλιο , ἱματιο , σιτοκάπηλος. II. from the popular character of κάπηλοι, a cheat, rogue, knave: hence 2. as Adj. ος, ον, cheating, knavish, κ. τεχνήματα Aesch. Fr. 328.

Comparison of this with the entry in LS’s main competitor, Donnegan,⁹ reveals some clear differences. Donnegan’s entry reads: properly, one who sells by retail in a market, Herod. ii.141 m. Darius so termed contemptuously, iii.89. f. contrasted with ἔμπορος, a wholesale dealer, a merchant, Xen, Cyr. iv. 5.42. a retailer of small wares, in general, a huckster, but most freq. a low tavern keeper, or vintner, hence, from their proverbial dishonesty in adulter ating wines, the Adj. adulterated, falsified; fraudulent, perfidious, Aesch. Fragm. 339.W.

Setting aside the variations in interpretation of the Greek, the most obvious difference between the lexicographers is Donnegan’s greater willingness to use words of clearly Latin origin, especially for the adjectival sense. All of his glosses for the adjective—‘adulterated’, ‘falsified’, ‘fraudulent’, and ‘perfidious’—wear their Latinity on their sleeve: their connection with Latin morphology, and not just roots, is easily recognized.¹⁰ Liddell and Scott’s glosses, after the first two descriptive ones, not only tend to be of earlier origin, but most importantly they avoid the Latinate forms found in Donnegan, and even, for the most part, any Latin origins. In their general preference for words with Germanic origins, illustrated here by ‘tradesman’ and ‘knave’ as well as ‘huckster’, which they share with Donnegan, they, not Donnegan, seem to be the outliers, as we can see not only in other lexica of the time¹¹ but also in contemporary translations of the authors cited by the Lexicon: the κάπηλοι ⁹ Donnegan’s was not the only other Greek English lexicon available, but as Stray (Chapter 1, this volume) points out, it was the market leader until Liddell and Scott appeared, and their original preface critiqued it, though in a passage eventually omitted from the published version (OUP archive, PB/ED/012952.) For reasons of space Donnegan (1842⁴) has been treated here as the main comparandum; unless otherwise noted examples are taken from his fourth (1842) British edition. ¹⁰ ‘Adulterated’ is from the English verb ‘adulterate’, derived directly in the sixteenth century from Latin; ‘falsified’ and ‘fraudulent’ come from Latin via French. The OED’s first citation for the verb ‘falsify’ in this sense is from 1502; in 1449 it was used to mean ‘to prove false’. ¹¹ Groves (1839) has ‘(fr. κάπη victuals; or, fr. κακύνω to adulterate, and πηλὸς, wine) a keeper of a tavern or inn, host, landlord; a retailer, dealer in small wares, huckster, pedler’; in Giles (1840)

Dictionaries as Translations: English in the Lexicon

31

found in Herodotus I.94—the sole citation given in 1843, though it was soon supplemented in subsequent editions—are rendered as ‘retail[ers] of merchandise’ ‘retailers’, and ‘venders by retail’ by translators in, respectively, 1818, 1842, and 1846.¹² This contrast between Liddell and Scott and Donnegan is repeated many times, to varying degrees. Let us sample some entries adjacent to that for κάπηλος. The next word to be treated in any detail is καπνός, together with its compounds and cognates. Where Donnegan has ‘to make a fumigation; to produce a smoke; to kindle a fire’ for καπνίζω, Liddell and Scott resolutely avoid any word related to the Latinate ‘fumigation’ throughout, even preferring ‘a smoking’ for the deverbal (that is, formed directly from the verb) noun κάπνισις. Their preference can be explained by the contrast between the Germanic origins of ‘smoke’, described in a forthcoming revised OED entry as of Old English origin,¹³ and ‘fumigation’ which is a borrowing from Latin traceable to Middle English. Here their avoidance of Latinity, both actual and apparent, is absolute. But this degree of etymological purism is unusual. For κανονίζω, on the same page as κάπηλος, their main entry is: to make, form by rule, to measure or judge by rule, to rule or establish, Arist. Eth. N.

as against Donnegan’s: to make according to rule and plummet; to make according to a rule, or a model to trace, or delineate after a rule, or model to judge, or determine after a rule established to establish as a rule, model, or pattern.

While there is still a perceptible gap between Liddell and Scott and Donnegan, the difference is in large part a matter of style. Donnegan’s ‘delineate’ is, unlike any of Liddell and Scott’s terms, a direct borrowing from Latin, and first cited a century later than any of the glosses particular to them.¹⁴ More striking, however, is his willingness to pile up multisyllabic and clearly Latinate terms. Here it is the stylistic appearance of Latinity that weighs heaviest in determining the choices in the the Lexicon. In yet another entry on the same page, the distinction between Liddell and Scott and Donnegan is even more lightly marked, yet it is still perceptible. For Κάνωβος, they have ‘a town . . . notorious for its luxury’, for Donnegan’s the glosses are ‘a merchant; huckster; vintner; adulterated; fraudulent’. Jones (1825) has only ‘dispenser of food, victualler, innkeeper’, omitting any hint of negative connotations. ¹² Littlebury 1818, 38 ‘They . . . first practised the way of retailing merchandise’; Wheeler 1842, 337 κάπηλοι glossed as ‘retailers’; Laurent 1846, 50 ‘They . . . were the first venders by retail’. ¹³ Grateful acknowledgment is due to Eleanor Maier of the OED for this information. ¹⁴ ‘delineated’ is similar in origin and date to ‘adulterated’ (n. 8); the OED’s citations for the various uses of ‘form’, from Latin via French, begin in the thirteenth century; the OED’s 2011³ entry for ‘measure’, a borrowing from French, suggests that it begins to be used in the sense (2a) intended here in the late thirteenth century.

32

Margaret Williamson

‘a city . . . infamous for the profligacy of its inhabitants’; the respective glosses for the related verb κανώβιζω are ‘to live like a Canobian, live luxuriously’ (LS¹) and ‘to live after the manner of the inhabitants of Canopus’. There is little to choose between the two in terms of etymology—both use almost exclusively Latinate terms—yet Donnegan still distinguishes himself stylistically by piling on the syllables. While it cannot, therefore, be said that the Liddell and Scott’s allegiance to ‘Saxon English’ always results in a strongly marked difference between them and Donnegan, it is nonetheless often present, and sometimes marked; and the contrast rarely goes in the opposite direction.¹⁵ We have already seen Saxon English being commended for its intelligibility. For writers such as the Wesleyan-Methodist Magazine commentator, discussing the efficacy of sermons, this concern was directed particularly at the ‘labouring classes’ and at one point he describes Saxon as ‘homely’. In other contexts it sometimes just meant plain speaking, as when, in an 1844 fiction, one character says to another ‘I say you may get out of the house . . . Isn’t that plain Saxon English? Get you gone’ (Mathews 1843, 334.) But it is clear from other discussions that ‘Saxon English’ style could also be a feature of poetic language, perhaps in a version influenced by the Romantic turn to ‘the language really used by men’. The anonymous writer who alludes to its ‘simple majesty’ also argues that ‘some of the finest bursts of our literature are in almost pure Saxon’, and includes among his models not only poems by Crabbe and Cowper but also several by Milton.¹⁶ Both intelligibility and literary style are appealed to by a writer in the Cincinnati Journal who recommends ‘plain sense in plain Saxon’, and finds it exemplified in Milton’s ‘noble sonnet on the massacre in Piedmont’, as well as in the Liturgy and in ‘our English Bible’.¹⁷ Liddell and Scott express a similar pride in the literary heritage of English which, combined with advocacy of Saxon English style, also influences some of their lexical choices. The vocabulary examined so far consists mostly of fairly workaday words, where the lexicographer’s main concern is to establish what the word denotes. But this is not the only consideration for a translator. The 1843 preface, with its lyrical appreciation of Greek, has already suggested that

¹⁵ An unusual and rather startling example is ‘unbeloved’, the first gloss for ἄϕιλος in one of the entries singled out by the publisher of Donnegan’s fourth (1846) American edition to demonstrate his superiority to both Dunbar (1844²) and Liddell and Scott (who both gloss it as ‘friendless’). ‘Unbeloved’ was a later addition: Donnegan’s first (1826) edition had ‘wanting friends; unfit for friendship; unbecoming a friend; unfriendly, Sophoc. disagreeable’. But despite its rather archaic appearance it does not seem to have been used before the latter part of the sixteenth century. Both Liddell and Scott and Donnegan, however, use the Middle English ‘beloved’ quite frequently. ¹⁶ ‘Milton is never greater than when he is speaking [Saxon English]. His noble sonnet on the massacre in Piedmont contains scarcely a word which is not Saxon. His ode on the Nativity is of the same stamp: so are his Allegro and Penseroso.’ See n. 9 above. ¹⁷ Reprinted from the Cincinnati Journal in the Connecticut Courant (Hartford, Connecticut) on 9 September 1838: Supplement p. 127.

Dictionaries as Translations: English in the Lexicon

33

the range of their interests in the expressive potentialities of language is wider than that of Donnegan, whose attention to translation begins and ends with a promise on the title page that his illustrative quotations shall be ‘literally translated’. Liddell and Scott’s concern to match one language to another can also include register and tone as well as denotation. Not surprisingly, given their source material, this includes some attention to poetic language, and sometimes we can detect in their glosses a kind of internal translation between registers. A straightforward example can be found in their treatment of archaic forms of pronouns and of second-person verbs. No longer current in the speech or prose of the time,¹⁸ archaizing forms such as ‘thou’/ ‘thee’, and the second person ending -st were still a feature of poetic language. Both Liddell and Scott and Donnegan, accordingly, make some use of them. Donnegan, for example, uses ‘thou’ and ‘you’ as glosses in the σύ and ὑμεῖς entries in order to distinguish singular and plural, but in translating quotations he can use either for the second-person singular, with ‘thou’ being used primarily in translations of verse, but not with complete consistency.¹⁹ There are also some, though relatively few, instances of the second- and third-person endings -st and -th in Donnegan: for example he uses ‘hast’ in translating quotations almost twenty times—most often, again, for verse and for early sources. In the Lexicon, however, these archaizing pronoun and verb forms appear more frequently, and most importantly with apparent self-awareness (and perhaps a little more consistency). In the entry for ὀνίνημι they translate a line from the Iliad, per their usual practice, using ‘thee’, but then paraphrase it in, as it were, their own voice using ‘you’: τί σευ ἄλλος ὀνήσεται; what enjoyment will others have of thee, i.e. what good will you have done them? Il. 16.31.

in what amounts to an internal translation between registers. The same phenomenon can be found in their treatment of various heroic accoutrements, found repeatedly in Homer. Take for example the knightly attire denoted by the Homeric words ἀορτήρ and θώρηξ. Donnegan renders these in fairly neutral terms as follows: ἀορτήρ belt, thong from which something hangs, a sword belt

and: ¹⁸ OED s.v. you (a revised entry, published in 2012): ‘In late Middle English and in the 16th cent. a common pattern was that forms in th were used towards social inferiors or children, or to others to mark either intimacy or contempt, but forms in y were used in most other functions. These gradually became the neutral, usual forms. The forms in th became much less frequent in the standard language in the 17th cent.’ ¹⁹ To take a more or less random instance, in the phrases quoted under ἄν, Donnegan renders the second person singular in Sophocles’ Ajax 550 with ‘thou’, but in OT 1438 with ‘you’.

34

Margaret Williamson

θώρηξ a breast plate, a cuirass, a coat of mail.

Liddell and Scott’s glosses are not dissimilar, but with a crucial difference: in addition to ‘a strap over the shoulder to hang anything to . . . a sword-belt’ and ‘a breast-plate, cuirass’ for the second, they add the more distinctive terms ‘baldrick’ for ἀορτήρ and ‘corslet’ for θώρηξ. The same is true of πήληξ and κορύς, where in addition to Donnegan’s ‘helmet’ they also offer ‘casque’ and ‘helm’. It is worth looking more closely at what motivates the editors’ decision to offer these additional options. A useful analogue for their interests is the translator made famous by Matthew Arnold’s critique: F.W. Newman, whose translation of Homer’s Iliad was published in the decade following their first edition. Like Liddell and Scott, he is concerned to match one language to another on more than one level, including style as well as denotation. Arguing in his preface that a translator of Homer is ‘necessitated . . . to adopt a more or less antiquated style’ in order to match the qualities of the original, he spells out the implications of this in a formulation remarkably close to that of the Lexicon: the entire dialect of Homer being essentially archaic, that of a translation ought to be as much Saxo Norman as possible, and owe as little as possible to the elements thrown into our language by classical learning. [my emphasis] (Newman 1856, vi)²⁰

Newman agrees with Liddell and Scott, therefore, in advocating a version of Saxon English—here ‘Saxo-Norman’ – which as for them entails avoiding egregious (‘thrown in’) Latinisms. Newman also, however, articulates one further consideration, and assumes a third. Implicit in his discussion is the assumption that a poetic register, or ‘dialect’, in one language should be matched with something similar in another: although he includes the Bible, along with Shakespeare, as an example of a text whose cultural authority for the English compares with Homer’s in ancient Greece, he never considers translating Homer into anything other than verse. The additional glosses ‘baldrick’, ‘corslet’, ‘casque’, and ‘helm’ offered by Liddell and Scott belong, like the ‘thou’/‘you’ distinctions, to a register still current in the verse of the time. Though archaic, or at least obsolescent, in prose,²¹ they are found repeatedly in contemporary poetry, especially in works

²⁰ Newman is aware that Saxon style is not universally commended: later in the essay he refers to ‘good Saxon words, which only a false taste counts ignoble’ (1856, x). ²¹ ‘Baldrick’ and ‘helme’ [sic] are listed in Halliwell Phillipps 1852 with no regional designa tion, implying that he regards them as archaic; the OED’s 1898 entry labels ‘helm’ in this sense ‘poet. and arch.’ H P lists only ‘cask’ thus, but the OED’s 1888 entry for ‘casque’ designates it as ‘now only historical, poetical, or foreign’.

Dictionaries as Translations: English in the Lexicon

35

on classical or historical subjects.²² This raises the question of whether, in addition to aiming for a match between poetic registers, they share Newman’s aspiration to distinctively ‘antiquated language’. Is archaism simply an element of the generally elevated language deemed appropriate to verse, or do they share Newman’s declared commitment to archaism per se, especially in translating Homer? It is true that Newman showcases the ancient, and predominantly Germanic, origins of some of his lexical choices by providing a list, with glossary, of ‘the more antiquated or rarer words’ he has used—but none of the above figures in it (xxi–xxii).²³ Yet elsewhere these vocabulary choices are staples in prose with an explicitly archaizing agenda: Walter Scott’s prose descriptions of armour, used for example to give an archaic flavour to his medievalizing novel Ivanhoe. While Newman aims for both poetic affect and archaism, Scott is interested only in composing prose that gives a general air of antiquity to his narrative, which he achieves through the judicious use of period vocabulary and syntax. It seems likely, therefore, that several considerations converge when it comes to their appearance in the Lexicon: in addition to Liddell and Scott’s concern to represent elevated poetic language, the association of this vocabulary with England’s own heroic past (itself often treated at the time in archaizing verse), and especially with things Saxon, or as Newman would have it, Saxo-Norman. We may note in passing that Walter Scott’s novel Ivanhoe, set in the late twelfth century, manages to evoke the period evoked by this composite term in a distinctly slanted way. The central conflict between Norman rulers and subjugated Saxons is expressed linguistically: the Norman characters speak Saxon, also referred to on occasion as English, only with disdain and under compulsion, while their Saxon subject Cedric has sworn to express his defiance by never speaking French. The association of the Saxon language with the subjugation of its speakers is comically spelled out near the beginning of the novel in a dialogue between the swineherd Gurth and the jester Wamba, who explains how a ‘good Saxon’ swine ‘becomes a Norman, and is called pork [porc], when she is carried to the Castle-hall to feast among the nobles’ (Scott 2000 [1820], 21). But the novel’s proud, free Saxons such as Cedric, and later Robin Hood, can be celebrated because the Normans they are pitted against ²² Examples include: for ‘baldric’, L.E.L.’s ‘Song of the hunter’s bride’ (1831) 20, and Tennyson’s ‘The Lady of Shalott (1832) III.15; for ‘corslet’, Felicia Hemans’ ‘The Storm of Delphi’ (1839), 58 and Walter Scott’s ‘Lay of the Last Minstrel’ (1841); for ‘helm’, Charlotte Brontë’s ‘Blondel’s Song’ (1833) 18, 41 and Felicia Hemans’ ‘Troubadour Song’ (1839), 6; for ‘casque’, Wordsworth’s ‘Struggle of the Britons’ (1821 2) 5 and Elizabeth Barrett Browning’s ‘The Dead Pan’ (1844), 85. Uses of all four drop off sharply after the nineteenth century. ²³ Among the most striking items are ‘eld’, ‘wight’, and ‘foeman’, characterized by Halliwell Phillipps respectively as ‘A[nglo]. S[axon]’ (eld, wight) and ‘now obsolete’ (foeman). The second two are not direct translations of a Homeric word, but appear in supplementary phrases added by Newman, giving the impression of a kind of decorative sprinkling of archaic Germanic words, not fully integrated with the rest.

36

Margaret Williamson

are rebels and villains, and these virtuous Saxons end up allied with the legitimate and victorious ruler Richard Lionheart. Scott’s account of the linguistic situation parallels this wish-fulfilment ending. Saxon language is dignified by the association of the characters who speak it with resistance to the corrupt Normans, but according to Scott it has been incorporated into a lingua franca rather than being suppressed.²⁴ This linguistic sleight of hand parallels the equally unrealistic reconciliation of outlaws and Norman king in the narrative, and echoes the emphasis on the Saxon that is implicit even in Newman’s term ‘Saxo-Norman’. The archaism of some of Liddell and Scott’s Homeric glosses has, therefore, more than one determinant, but it is in keeping with the backward-looking orientation of advocates of ‘Saxon English’, whose models were often drawn, as Johnson put it, from ‘ancient volumes’. But, as we have seen, their antiquity need not extend all that far back: in practice many follow Johnson in privileging the sixteenth and seventeenth centuries. The most often cited exemplar of Saxon English from this period was the King James Bible; but others included the editors’ favourites Milton and Shakespeare, both mentioned in the preface as the untranslatable acme of English literature. Newman echoes their focus on Elizabethan literature when he commends the Homeric translations of Shakespeare’s contemporary George Chapman over those of later, more ‘polished’ translators like Cowper and Pope (1856, iv). We have already seen selected passages from Milton used to illustrate the style, and Shakespeare, who was hailed by Carlyle in his 1840 essay ‘The Hero as Poet’ as ‘an English King’ who would unify the ‘Saxondom covering great spaces of the Globe’ (2013, 102), is an even more obvious choice. The 1844 obituary of Sir Francis Burdett declares that his diction ‘was, without being florid or diffuse, eloquent . . . and as full of genuine Saxon English as that of Shakespeare himself ’;²⁵ a decade and a half later Mary Cowden Clarke declared that ‘Shakespeare may be taken as a standard for language . . . his is genuine Saxon English’ (1860, v). In view of Liddell and Scott’s literary interests it is not surprising, therefore, to find occasional echoes of Elizabethan English and sometimes specifically Shakespearean language in the Lexicon’s glosses. In the entry for κέντρων, for example, they follow up the descriptive account of its meaning, ‘one who bears the marks of the κέντρον’ with the far more colourful phrase ‘(and so) a spur-galled jade’. This phrase seems to have come into use around the turn of the sixteenth and seventeenth centuries: it was used, for example, in ²⁴ While the Norman/Saxon division is, according to a modern commentator on Scott, reasonably accurate for the period, Scott’s reference to ‘the lingua Franca, or mixed language, in which the Norman and Saxon races conversed with each other’ is less well founded. Scott 2000 [1820], 404. ²⁵ Trewman’s Exeter Flying Post or Plymouth and Cornish Advertiser (Exeter, England), Thursday, 25 January 1844; Issue 4079.

Dictionaries as Translations: English in the Lexicon

37

George Gascoigne’s The Noble Art of Venerie (1575), John Day’s Lavv-trickes (1608), John Stephens’ Cinthia’s Revenge (1613), and William Bishop’s Disproofe of D. Abbots counterproofe against D. Bishop . . . (1614), and later in the seventeenth century.²⁶ While the complete phrase does not appear in Shakespeare, two of its components—‘spur-galled’ and ‘galled jade’—are found in, respectively, Richard II²⁷ and Hamlet, and their fortunes are closely linked with the dissemination of Shakespeare: the second in particular is very often part of an allusion to Hamlet’s ‘let the galled jade wince’ (III.ii.236). By the nineteenth century ‘galled jade’ was almost always used in this context, while ‘spur-galled’ was rarer and mainly used in archaizing sources such as Walter Scott’s The Antiquary (1815). Even more unusual was the complete phrase; a rare example is the allusion in ‘lest the jade be spur-galled’ in a New York newspaper in 1828.²⁸ Given the fact that the vocabulary used was thought to require glossing in a dictionary for schoolboys (Webster 1833¹¹), and given the steep rise in the publication and staging of Shakespeare, it seems reasonable to conclude that the phrase had a Shakespearean as well as an archaic ring by the time Liddell and Scott used it in 1843. Certainly by the time Madden published his study of Shakespeare and sport at the end of the century it was seen unequivocally as Shakespearean (Madden 1897, 301). Another term distinctive to Liddell and Scott and demonstrably Shakespearean to nineteenth-century ears is ‘murrain’. In the entry for ϕθείρου they explain that it: was a common imprecation, go and be hanged! away, with a murrain on thee! Lat. abi in malam rem! Ar. Ach. 460, Plut. 598, 610.

The word ‘murrain’, of Anglo-Norman vintage and probably ultimately Latin origin, was still current in the nineteenth century to denote disease among animals, but it had both archaic and Shakespearean overtones when used as an imprecation. Shortly after the already-mentioned conversation between Saxon peasants in Ivanhoe, the swineherd Gurth uses the term to express his impatience with Wamba—‘“A murrain take thee!” rejoined the swineherd’ (2000 [1820], 22)—and though not glossed by Scott himself the term was thought by his 1903 editor to require explanation (1903, 52). Shakespeare’s characters wish murrains, or plagues, upon each other seven times in his surviving plays; but they are outdone in a fictional rendering of Shakespeare and his times published only a few years before the first edition of the Lexicon. In Robert Folkestone Williams’ 1838 historical novel Shakespeare and His Friends: or, The Golden Age of Merrie England, the phrase appears no fewer than twenty times. ²⁶ Later uses of the phrase include Richard Carter’s The schismatick stigmatized, 1641 and the anonymous Newes out of the west (1647.) ²⁷ Richard II V.5.94. In modern editions the phrase reads ‘Spurr’d, gall’d: ‘Spur galled’ is the normalized reading in the First Folio (Gurr 2003, ad loc.). ²⁸ Free Enquirer 1 (1828): 496.

38

Margaret Williamson

Another example of the editors’ familiarity with Elizabethan English speaks, again, to the complexity of their linguistic interests as compared with those of Donnegan. Their entry for ἐπίκλοπος reads: ἐπίκλοπος, ον (κλέπτω, κλοπή) : thievish, given to stealing, tricksy, wily, Od. 11.364, 13.291, Hes. Op. 67. Sometimes also c. gen., ἐπίκλοπος μύθων cunning in speech, Lat. callidus, peritus, Il.22.281, and so prob. ἐπίκλοπος τόξων, cunning in archery, Od. 21.397, where however others make it = ἐπιθυμητής.

Here Donnegan has: thievish; secret; furtive cunning; artful; deceitful; with prejudice to another, Od γ, 364. ν, 291. Aesch. Eum. 144. with a genit. clever, dexterous, skilful in, not in a bad s. ϕ, 281. Il. ϕ, 397.

This comparison is a particularly interesting one because the structure of the entries, and even the particular references, corresponds quite closely. Generally, the contrast is close to that found in the κάπηλος entries. Liddell and Scott share only two glosses with Donnegan: ‘thievish’ and ‘cunning’. Otherwise they favour earlier vocabulary with no evident connection to Latin, where Donnegan’s glosses are generally later and obviously Latinate.²⁹ But what is particularly striking here is LS’s use of a gloss they share: ‘cunning’. Like Donnegan they use this for the pejorative sense of ἐπίκλοπος; they differ from him in also using it for the second, neutral sense of the term, when it is used with a genitive and means simply ‘skilful.’ Donnegan takes pains to distinguish the two senses of the term, both by offering the glosses ‘clever, dexterous, skilful in’ for the second meaning, and by adding ‘not in a bad sense’. Why did Liddell and Scott repeat ‘cunning’ for the second meaning, with all the potential for ambiguity that entails? As Donnegan’s entry makes clear, there were alternatives that would have made the meaning plain, and at least one of those, ‘skilful’, has a non-Latin etymology just as ancient as that of ‘cunning’.³⁰ (Compare Bracke’s Chapter 13 in this volume, on the continuing difficulty of pinning down words in this semantic field.) The editors were clearly attracted to the term both by its antiquity and by its Germanic origins, which they highlight in the entry for δαιδάλλω by adding to

²⁹ Where their glosses differ, Liddell and Scott have ‘steal[ing]’, Old English; ‘tricksy’, first listing in the OED’s sense 3 in 1766, but derived from Old French ‘trick’; ‘wily’, origin obscure, used in the fourteenth century but cf. the twelfth century ‘wile’. First uses of the glosses distinctive to Donnegan range from the fifteenth century (‘deceitful’, first citation 1483 in the OED’s 1991 entry) to the seventeenth (‘dexterous’, ‘artful’, ‘secret’) and possibly the nineteenth (‘furtive’, though only, as here, applied to a person: actions could be performed ‘furtively’ in the late fifteenth century). Most of these are ultimately derived from Latin by way of French: for ‘dexterous’ and ‘furtive’ the derivation is direct and obvious. See below on Donnegan’s ‘skilful’. ³⁰ ‘cunning’ and ‘skilful’ are both Middle English. The OED relates ‘cunning’ to Old English cunnen and ‘skilful’ to Old Norse skil.

Dictionaries as Translations: English in the Lexicon

39

their discussion of the root δα- the note ‘as our cunning, from to ken’.³¹ But although it and its cognates are fairly common in Middle English, its appeal probably arises from a later period. The attraction of using ‘cunning’ for the neutral sense of ἐπίκλοπος, even though it was well on the way to being archaic, side by side with the pejorative one may arise from the fact that both meanings were current in Shakespeare, who uses the adjective repeatedly— over one hundred times.³² Although pejorative uses predominate, he also occasionally uses it in its neutral sense, and most importantly, he plays on the two senses. When, for example, in Romeo and Juliet Capulet calls for ‘twenty cunning cooks’ (IV.ii.2), the meaning ‘skilful, clever’ is clear: but Othello’s address to Desdemona as ‘thou cunning’st Patterne of excelling Nature’ (V.ii.11) plays on both. Liddell and Scott’s entry replicates the ambiguity of the term, and it seems likely that another of the attractions for them was the opportunity to match the multivalence of the Greek word with an equally multivalent English one. The final aspect of the Lexicon’s lexical choices that will be examined here is their attention to tone. Here too there is a perceptible difference between Liddell and Scott and Donnegan. While Donnegan’s glosses tend to the neutral and descriptive, Liddell and Scott are sometimes, as in the spheres discussed earlier, obviously trying for a register and tone through which to render the force of the Greek terms as well as what is denoted. This is especially clear in their treatment of the language of opprobrium and insult. An example of this can be found in Donnegan’s and Liddell and Scott’s versions of the vocabulary used by Strepsiades in Aristophanes’ Clouds when he expresses his determination to outwit his creditors by using the skills he hopes to learn in Socrates’ school: εἴπερ τὰ χρέα διαϕευξοῦμαι, τοῖς τ᾽ ἀνθρώποις εἶναι δόξω θρασὺς εὔγλωττος τολμηρὸς ἴτης βδελυρὸς ψευδῶν συγκολλητὴς εὑρησιεπὴς περίτριμμα δικῶν κύρβις κρόταλον κίναδος τρύμη μάσθλης εἴρων γλοιὸς ἀλαζὼν κέντρων μιαρὸς στρόϕις ἀργαλέος ματιολοιχός (443 51)

³¹ Johnson has: ‘[from ‘connan’, Sax.]’; ken appears in Newman’s list of ‘antiquated or rarer’ words, glossed as ‘to espy, to take cognizance’ (1856, xxi). ³² Halliwell Phillipps’s listing of the meaning ‘skilful, knowing’ without regional designation suggests it is included for its archaism (cf. n. 21 above). The meanings in Webster’s Dictionary for Primary Schools (1833) ‘artful, crafty’ tend to the negative, but ‘cunning’ is used to mean ‘intricate’ in the historical novel Shakespeare and his Friends (Williams 1838, 90.) OED’s 1893 entry characterizes it as ‘now only a literary archaism’.

40

Margaret Williamson

In a translation first published in 1798 but still influential in Liddell and Scott’s time, Richard Cumberland translates this passage as follows: So that I may but fob my creditors, Let the world talk; I care not though it call me A bold fac’d, loud tongu’d, over bearing bully; A shameless, vile, prevaricating cheat; A tricking, quibbling, double dealing knave; A prating, pettyfogging limb o’th’law; A sly old fox, a perjurer, a hang dog, A raggamuffin made of shreds and patches, The leavings of a dunghill Let ’em rail, Yea, marry, let ’em turn my guts to fiddle strings, May my bread be my poison! if I care.

The dramatic force of this passage depends not only on the bravura of what Cumberland calls ‘this torrent of terms’,³³ but most importantly on their strongly negative connotations. Strepsiades makes his defiance of social norms clear by declaring his indifference to any and all of the insults he expects to incur through his antisocial behaviour, and this extravagant chronicle of them sets him up for comeuppance later in the play. Cumberland renders the passage accordingly, inserting a litany of pejorative terms some of which, like ‘hang-dog’, have an exact analogue in the text, while others—‘bully’, ‘vile’, ‘cheat’, ‘knave’—are used more generally to intensify the negative tone of the passage.³⁴ Cumberland’s use of this latter group of terms draws attention to the particular challenges of a lexicographer in glossing one word at a time. Tone is often, as here, built up in a sequence of words, and may not be easy to attach to each and every word. But even bearing that constraint in mind, the extent to which tone is indicated in the glosses provided for the Greek of this passage by Liddell and Scott on one hand and Donnegan on the other differs in some interesting ways. One—between Liddell and Scott’s plainer, sometimes more Germanic vocabulary and Donnegan’s greater tendency to Latinity—has already been discussed. Where the Lexicon has ‘hard, painful, troublous, grievous’ and ‘troublesome’ for ἀργαλέος, for example, Donnegan offers ‘vexatious; troublesome; distressing; painful; difficult’. For the metaphorical sense of γλοιός two of Liddell and Scott’s three glosses—‘slippery, trickish,

³³ ‘This torrent of terms, nearly if not quite synonymous, forms one of the most curious passages in this very singular author, and is such a specimen of the versatility and variety of the language, as almost defies translation’ (Cumberland 1812, 43). ³⁴ It seems likely that in Greek this tone builds up gradually, beginning with terms whose connotations are ambiguous and leading to a typical Aristophanic παρὰ προσδοκίαν effect when they turn unambiguously negative: see Dover 1968 ad loc. But Cumberland (rightly, in view of its dramatic importance) goes for the most important affect from the start.

Dictionaries as Translations: English in the Lexicon

41

knavish’—overlap with Donnegan’s, but Donnegan has no fewer than eight additional ones, almost half of which are obviously Latinate. The more striking and interesting contrast between them, however, has to do with tone. Donnegan does not include κέντρων, discussed earlier, in this negative metaphorical sense at all, though Cumberland’s rendering ‘hang-dog’ shows that that meaning was clear to some readers. For μιαρός, Donnegan, explicating its ritual connection with defilement, offers mainly descriptive terms that are neutral in tone: thus ‘impious’ is the closest he gets to a translation that could be used as an insult. The Lexicon’s last three glosses, however—‘brutal, coarse, blackguard’—are closer to the tone of the word in Aristophanes, and in ‘blackguard’ it is the insulting tone of the word that carries most weight, independently of denotation. The translations offered by each lexicon for στρόϕις, finally, exemplify clearly both the characteristic differences between them: in place of Liddell and Scott’s ‘a twisting, slippery fellow’, Donnegan has ‘an artful man; one who is dexterous in finding expedients’. A Wesleyan-Methodist magazine-style translation into Donnegan-ese of the line in which these last two words appear would convey little to nothing of Strepsiades’ devil-may-care attitude. It seems clear from the samples discussed in this essay that Liddell and Scott’s target audience included another group besides the would-be practitioners of Greek prose composition identified by Silk (Chapter 17, this volume). The awareness of the requirements of translation that they express in the preface, although limited, is borne out by their attention to aspects of meaning that Donnegan neglects: that is, to style, register, and tone as well as denotation. Together with the influence of ‘Saxon English’ style on their practice, it has important consequences for the lexicon that we have inherited, some of which they could not have anticipated. We can gauge some of them by returning briefly to the entry for κάπηλος as it appears in 1882, the edition that formed the basis for the Middle Liddell. By this time the emphasis announced in 1843 had changed radically. The original preface had disappeared after the fourth edition of 1855, to be replaced in 1861 with one much influenced by new developments in the discipline of philology. Gone is the awareness that what the Lexicon is undertaking is an approximate, imperfect match between two different languages and cultures. In the new preface what is stressed is coverage, exhaustiveness, grammatical precision, and ‘advances in the science of Comparative Philology’. The new orientation is reflected in the 1882 κάπηλος entry, which is much longer than in 1843 and includes fuller citations, with much additional detail on comparative etymology. Yet given the effort that has gone into other aspects of the revision, it is striking how little the English glosses have changed since 1843. For the noun, the main changes are the addition of ‘hawker’ and ‘peddlar’, replacing ‘petty tradesman’; the adjectival glosses are now ‘cheating, cozening, knavish’ as opposed to ‘a cheat, rogue, knave’. Not only is the

42

Margaret Williamson

distinctive flavour of the surviving glosses from 1843 still in place; it is intensified by some of the new ones, with the same avoidance of obviously Latinate vocabulary, preference for Germanic roots, and liking for archaicand Shakespearean-sounding language.³⁵ Liddell and Scott’s stylistic preferences not only outlived them but were sometimes heightened in later editions, not only by the passage of time but also as a matter of editorial choice. To some extent the momentum their distinctive style acquired is inevitable, going hand in hand with the translations they influenced. For example, the 1853 Clouds translation of William James Hickie, which draws several phrases directly from Liddell and Scott, is still in circulation as the online translation offered by Perseus. The fashion for archaizing translations of Greek literature, especially of Homer, continued long after Newman: a prominent example is the prose translations of Lang and his various associates, who take biblical English as the closest analogue to Homeric Greek (Butcher and Lang 1878; Lang, Leaf and Myers 1883). Another factor is suggested by the evolution of the κάπηλος entry, and would perhaps have been harder to foresee in 1843. The first edition offered, in addition to the tone of the glosses, information that contextualized that tone. In section II of the entry the glosses ‘cheat, rogue, knave’ are introduced by the phrase ‘from the popular character of κάπηλοι’: the disparaging tone of many uses of the word and its cognates is named and explained rather than being conveyed solely through glosses. In 1882 this explanatory phrase has been displaced by the new stress on comparative philology, casting the burden of conveying tone onto the glosses and their connotations. But connotation changes more rapidly than denotation, and the passage of time can only intensify the question of which overtones and associations are relevant, and which are later accretions. When we read in the notes to a 2004 edition of Sophron that κάπηλος can mean ‘knavish’, the adjective’s pejorative associations are now likely to be swamped by its now rather arch and quaint archaism (Hordern 2004, 124). All of the features of Liddell and Scott’s approach discussed here—their awareness of contemporary stylistic ideals, the Shakespearean echoes in their glosses, their attention to register and tone—can be seen to flow from the aspirations announced in the preface. The most specific, the preference for Saxon English, was perhaps the most influential, short-lived though its expression was: the disappearance of the sentence announcing it is one of the minor but unacknowledged amendments made when the preface was reprinted in the second edition (1845). But all, however patchily implemented, ³⁵ ‘Hawker’ is in use from the sixteenth century, and in 1898 the OED traced it tentatively to Middle Low German; ‘peddlar’ (‘pedlar/ peddler’ in the OED) is Middle English. The derivation of ‘cozen’ is uncertain, but it is certainly widely used in Elizabethan English, with twenty two uses in Shakespeare. None appears to be Latinate.

Dictionaries as Translations: English in the Lexicon

43

are part and parcel of their aspiration to provide translations that would represent the literary qualities of Greek by drawing on those of English, and thus help to position the English as true and worthy heirs of the glory that was Greece. As the preface also shows, other European nations, and especially the French, were also contenders.³⁶ The editors’ emphasis on ‘Saxon English’, together with their dependence on Passow and his predecessors, had the implicit effect of aligning the English with the Germans and their decadeslong dominance of Greek scholarship. It is ironic that Donnegan’s glosses sometimes stand the test of time better than Liddell and Scott’s: his κάπηλος entry, for example, focusing on the neutral, descriptive, and where necessary Latinate, is now more accessible and less quaint-sounding than theirs. But that is because he had fewer aspirations, as revealed by the inadequacy of his glosses for translating Aristophanes. Albeit in a limited way, Liddell and Scott were attempting something more difficult. As another nineteenth-century translator puts it: Translation is one of the most difficult kinds of literary work; and requires, for a decided success, such a combination of learning, judgment, perseverance, enthu siasm and taste, as is seldom found in the same person. (Blackie 1861, 268)

How much more difficult to achieve this combination when its practitioners are dealing with the multitude of other considerations discussed in this volume—and in the atomized format of a lexicon to boot. Interpretation, especially of anything beyond simple denotation, cannot in fact proceed, or be represented, on a word-by-word basis, as the nineteenth-century theatrical translators sampled in this essay knew.³⁷ If Liddell and Scott succeeded only partially in an area where Donnegan hardly ventured at all, their shortcomings are not theirs alone: they also point to the limits of translation as a medium of interpretation in a dictionary. Two conclusions follow. One is that descriptive, explanatory glosses of the kind that were abandoned in the 1882 κάπηλος entry continue to be required as a supplement to the translations. The other is that the translations

³⁶ A patriotically inspired re edition of Henri Estienne’s sixteenth century Thesaurus Grae cae Linguae was published in nine volumes between 1831 and 1865 by the Parisian firm of Didot. Ironically, its initial slow progress led Didot to bring in German scholars to take charge of the project: see Petitmengin 1983. ³⁷ For one example among many, see J.H. Frere’s 1820 discussion of the importance of variations in style to Aristophanes’ comedy (1911 [1820], xiv.) Later in the essay he focuses on tone in his argument for using ‘huckster’, one of Liddell and Scott’s κάπηλος glosses, to translate ἀρτόπωλις: it can, he says, legitimately be used interchangeably with ‘market wom[a]n’ because the most important aspect of the Greek word is its pejorative association with ‘the race of Market Scolds’ (xviii). In his free translation of a section of Clouds, published in 1836 as ‘The Possums of Aristophanes’, the Strepsiades character is called ‘Hucksterides’ [Son of Huckster], a petty tradesman of Athens’ (1836, 286).

44

Margaret Williamson

themselves need constant revision. The reasons why a twenty-first century reader should no longer be content to view his or her Greek through a nineteenth-century lens are obvious. The emphasis in 1882 was in many ways a retreat from the idea of meaning as dynamic, part of dialogue, towards a more empiricist idea of classical texts as already there, pre-constituted, waiting to be discovered and then represented by the epiphenomenal addition of translation. This is the model that has dominated classical scholarship until recently, and in some quarters still does. But there are signs that we are entering a new era, in which it is possible to recognize translation as integral and interior to interpretation, not subsequent to it. To study the history of Liddell and Scott’s English is at once to see that process in action and to understand the constant need to revisit it. No language or text is dead: it lives through—and only through—repeated readings, which is to say through translation. It is always time to re-translate.³⁸

³⁸ Heartfelt thanks to the following, whose comments on drafts of this chapter saved me from many errors and pointed new ways forward: John Considine, Stuart Gillespie, Eleanor Maier, Monika Otter, and above all Chris Stray. I am also grateful to the librarians at Dartmouth College, especially Wendel Cox and Daniel Abosso, for their cheerful guidance through some dark and winding text mining tunnels.

3 Latin in the Lexicon David Butterfield

‘An OCT with a preface in English! This is the end of civilization as we have known it.’ (M.L. West¹)

In that venerable year of 1843 more than one toe was left throbbing in the drawing rooms, college libraries, and club lounges of Britain as copies of the newly published Greek-English Lexicon of Liddell and Scott tumbled at pace from tweed, stuff, and cassock-cloth onto unsuspecting feet. A Greek dictionary issuing from an ancient university—but translated into English? quidnam proximum? The decision to abandon the scholarly garb of Latin and opt for direct expression in the vernacular was, in its time, a remarkable one. Of this the editors were acutely aware.² The very first page of their preface tackles head on the ‘objection that may be taken, at starting, to the notion of an English Lexicon of the Greek Tongue. It may be asked, whether such a Lexicon should not be in Latin, as in the old times; whether the other is not an unworthy condescension to the indolence of the age’ (LS¹, iii). Having brought the issue to the fore, the editors are careful to acknowledge the indisputable superiority of Latin in certain contexts: We hold that Critical Notes on [Classical] authors will always be best in the Latin Tongue. No other will be found so brief, clear, and easy of remembrance; no other has the advantage of technical terms and phrases which all Scholars have agreed to use; no other will be so readily understood by Readers of all countries and all ages.

¹ Review of H. Lloyd Jones and N.G. Wilson (eds), Sophoclis Fabulae (Oxford, 1990), CR 41.2 (1991) 299 301, at 301. West’s tone was ironic, but other scholars must have held seriously the view he presented with tongue in cheek, after almost a century of Latin prefaces to Oxford Classical Texts. Some of the issues this raises are discussed in this chapter. ² The complex reasons for preferring English have been carefully explored by Margaret Williamson in Chapter 2. David Butterfield, Latin in the Lexicon In: Liddell and Scott: The History, Methodology, and Languages of the World’s Leading Lexicon of Ancient Greek. Edited by: Christopher Stray, Michael Clarke, and Joshua T. Katz, Oxford University Press (2019). © Oxford University Press. DOI: 10.1093/oso/9780198810803.003.0003

46

David Butterfield

We shall see that the concision, and indeed precision, of such Latin terminology retains a role in constructing Liddell and Scott’s Lexicon. However, the editors go on to make clear the language’s unsuitability for their own lexicographical undertaking: But though this is our opinion in regard to Critical Notes, it by no means follows that we should hold the same in regard to Lexicography. The chief business of Lexicography is one, to interpret words; of Criticism another, to unravel the idioms and intricacies of language. The Latin Tongue may be the best organ for the latter work, yet very unequal to the due execution of the former. And quite unequal it is. For just as impossible is it to render the richness, boldness, freedom and variety of Greek by Latin words, as it is to give any adequate conceptions of Milton or Shakspere by French translations. Yet French is, confessedly, the language of Mathematics. So Latin is the language of Classical Criticism. But we hold it feeble and defective for purposes of Lexicography. And when we add to this the fact that, in richness at least and freedom (though certainly not in beauty or exactness) our own language is not unworthy to compare with the Greek, we conclude confidently that the best Lexicon an Englishman can use to read Greek with, will be in English. A Frenchman may have reason for using a Greek Latin Lexicon; an Englishman can have none. (LS¹, iii iv)³

This is a cogent case to give both the conservative and the sceptic pause for thought: for all its brevity and clarity of expression, Latin (the implication goes) is unhelpfully austere and constrictive for the lexicographer. Latin—the unquestionable language of scholarship—is deemed to lack sufficient flexibility and flair to do the job properly for Greek; no weightier support is given for the claim, which implicitly reflects Victorian notions of the liberal ingenuity of the Greek genius and the derivate rigidity of Roman culture. And yet, despite almost every Greek lemma being rendered by Liddell and Scott into English, Latin plays a significant role in their Lexicon. In fact, every single page of the work, at least in its first few editions, contains some Latin. The purpose of this chapter is to explore how and why Latin continued to play a significant role for Liddell and Scott, notwithstanding the impropriety of its allegedly ‘feeble and defective’ character. The uses of Latin in Liddell and Scott are manifold. Some reflect the genuine demands of precise and professionalized lexicography, some a dependence on the previous scholarly tradition, and some the exigencies of Victorian prudery. Although the collective tapestry is rich and complex, it is possible to draw out ³ The first substantial reviewer of the Lexicon (J.R. Fishlake in Quarterly Review 75 [1845] 303 24) shared this opinion that progress beyond Latin was necessary: ‘The time for Greek and Latin Lexicons is gone by in Germany as well as in England’ (p. 303). Nevertheless, he also expressed serious doubts that a Greek German dictionary could be transmuted into a Greek English format: ‘We assert unhesitatingly that no scholarship however high, no experience however tried, no knowledge of Greek and German however accurate, can translate successfully a Greek and German into a Greek and English Lexicon. It is a literary impossibility’ (p. 309).

Latin in the Lexicon

47

seven separate threads of how Latin is deployed throughout the Lexicon’s pages and history. Latin is used: (i) to illustrate the etymology of a Greek word; (ii) to explain a Greek word transliterated directly from it; (iii) to illustrate a direct Latin derivation from a Greek word; in particular, when a sense of the Greek word, or indeed the Greek word itself, is only attested in Latin literature; (iv) to clarify a Greek term designed to translate or represent a specific Roman entity; (v) to elucidate the idiom of a Greek word, phrase or syntactic construction; (vi) to serve as the language of scholarship; (vii) to avoid the undesirably direct expression of the English vernacular. This can only be an approximate taxonomy, since in some cases Latin is deployed for more than one of these purposes simultaneously. It will, however, provide a useful map for plotting a way through the complex and copious foliage of the Liddell-and-Scott jungle.⁴ Nor are all of these seven instances equally common or obvious: the first is a ubiquitous feature of modern lexicography, but took time to earn its place in classical dictionaries; the second, third, and fourth are essential parts of the process; the fifth can convey some benefit, if deployed methodically and intelligently; the sixth and seventh, by contrast, are vestiges of a Victorian age which have been largely obliterated in modern classical scholarship. For what follows in this paper, I will trace a course through these different roles played by Latin in the Lexicon, before turning to the more protean—and somewhat purposeless—deployment of the language before its time was called by the editors of the ninth edition (1925–40).

3.1. THE LATIN OF ETYMOLOGICAL EXCURSUS Given the close relationship between Greek and Latin, through their shared descent and intimate interaction, it is no surprise that Latin words appear frequently among the etymological data, typically given in brackets. In the first edition, the etymological material has an amateurish feel to it, both in its sketchy content and inconsistent treatment. Its secondary value is reflected by its location, a final parenthesis provided almost as an afterthought.⁵ ⁴ There is much to be said for the claim, made in an anonymous survey of the work and its legacy, that ‘Liddell and Scott’ should be treated as a univerbation (The Spectator, 20 October 1906, p. 10). ⁵ For discussion of the place that etymology has held in the history of Liddell and Scott, see Chapter 5 by Joshua Katz in this volume.

48

David Butterfield

Nevertheless, once the etymological can of worms has been opened, Latin must play a starring role. It appears frequently enough in the first edition, at least in the most obvious instances: for example, ‘Lat. pater’ appears s.v. πατήρ, ‘Lat. no, nato’ s.v. νέω (B) and ‘Lat. ne-re, nec-tere’ s.v. νέω (C).⁶ By the time of the fifth edition (1861), the etymological material has expanded considerably. Advances in the field, particularly by Georg Curtius and other German philologists,⁷ had successfully unlocked a far broader range of relevant material. As further pieces of the puzzle fell into place, the ninth edition was able to cherry-pick reflexes from a remarkably wide range of Indo-European material. Such terms are generally listed in a schematic order: the Proto-Indo-European root, if reconstructable, is given; there then follow several languages that take priority over Latin, such as Sanskrit, Avestan, Old Irish, Old Slavonic, Old Norse, and (Old) English. Still, Latin words appear among these cognate sets more commonly than not. Among the many hundreds of examples, the following cases from the ninth edition are typical: s.v. ἄρκτος: ‘(Cf. Skt. :rkshas, Lat. ursus, etc.)’ s.v. ἵστημι: ‘(From I.-E. sthā-, cf. Skt. sthā- (aor. á-sthā-t), Lat. stare, etc.; Gr. redupl. pres. and pf. fr. si-sthā-, se-sthā-.)’ s.v. μύρμηξ: ‘(Cf. βόρμαξ, ὅρμικας, Skt. vamrás, Avest. maoiriš, OIr. moirb, OSlav. mravĭji, ONorse maurr, Engl. (pis-)mire, Lat. formica: I.-E. forms perh. momro-, momrī-, memro- ‘ant’.)’ s.v. ταῦρος: ‘(Cf. Lat. taurus, Lith. tauras, Slav. turŭ, etc.)’ Nevertheless, like the role of etymology itself, this remains a comparatively minor part for Latin to play in the Liddell–Scott Lexicon.

3.2. THE LATIN OF DIRECT TRANSLITERATION Although Greek influence upon the Roman world was more profound, in cultural and linguistic terms, than in the reverse direction under the Empire, many Latin words were transposed wholesale into later phases of ancient Greek. Such words underwent the mandatory—but minor—changes in script and morphology to function in their new host language. Particularly fecund authors in this process of Hellenizing Latinity were Dionysius of Halicarnassus, Plutarch, Appian, and John the Lydian, although there is naturally a ⁶ In some cases ‘Lat.’ is removed to introduce the Latin reflex, perhaps for reasons of space, e.g. ‘u mbil icus’ s.v. ὀμϕαλός. ⁷ The publication of Curtius’ Grundzüge der Griechischen Etymologie (2 vols, 1858 62) ushered in a new wave of technically proficient etymological enquiry, which took time and often translation to lap against British shores.

Latin in the Lexicon

49

wealth of evidence from the sub-elite milieu attested by epigraphic and papyrological data.⁸ It is manifest that, if the Greek word is a direct transliteration of an existing Latin word, the original Latin term needs citation in the article. Some examples (again drawn from LSJ) are as follows: Ἀβοριγῖνες, οἱ, = Lat. aborigines, D.H.1.9, al. δούξ, δουκός, ὁ, = Lat. dux, PLond. 2.141.18 (iv AD), Just.Edict.13.18Intr., etc.⁹ οὐά or οὐᾶ, Lat. vah! exclam. of admiration, or of astonishment, Arr. Epict.3.22.34, D.C.63.20; of irony, Ev.Marc.15.29. πρίγκῐπες, οἱ, = Lat. principes, Plb.6.21.7. For such clear-cut cases, it is deemed sufficient to give reference to the Latin term, in the expectation that the reader will already be familiar with it. In some instance, however, it was thought necessary to provide broader context about the original term: σπεκουλάτωρ, ορος, ὁ ,= Lat. speculator, prop. scout: but in the Roman Imperial army, 1. one of the principales or head quarters’ staff of a legionary commander or provincial governor (whose duties include the carrying out of executions), Ev. Marc.6.27, POxy.1193.1 (iv A.D.), etc. 2. one of the Imperial body guard (spec ulatores Augusti), = δορυϕόρος, Suid.

At any rate, this category really reflects Latin lexicography crystallized in Greek script. All told, it is a relatively small field within the array of Latin terms in the Lexicon.

3.3. THE LATIN OF GREEK DESCENT A much commoner occurrence is the reverse scenario, i.e. when a Greek term is reflected by a direct transliteration into Latin. Again, the changes are usually limited to script and (in most cases) inflection. Any competent Roman writer was intimate with the Greek classics; nevertheless, only a portion of the Latin canon proves to be a regular and reliable source for the Nachleben of Greek vocabulary. Among the (predominantly Greek!) list of ‘Authors and Works’ prefacing the ninth edition (pp. xvi–xxxviii) there appear twenty-five Latin writers. These include poets and litterateurs (Ausonius, Claudian, Fronto, Gellius, Juvenal, Martial, Persius, Petronius), philosophers (Cicero, Seneca), grammarians (Festus, Isidore, Macrobius, Martianus Capella, Nonius, ⁸ Since this last class of material did not find a major place in the Lexicon until the twentieth century, many cases in this category appear first in the ninth edition. ⁹ These two terms were omitted from earlier editions, perhaps in the belief that neither was a truly Greek meriting a place in a Greek English Lexicon.

50

David Butterfield

Quintilian, Terentianus Maurus, Terentius Scaurus, Varro, Velius Longus) and other technical writers (Justinian, Rutilius Lupus, Servius, Suetonius, Vitruvius). It would of course be possible to cite Latinized Greek terms from among these authors ad nauseam. For the most part, however, such cases tend to be restricted to those where the Latin evidence can helpfully inform the picture painted by the extant Greek. In some instances, the Latin evidence may clarify technical matters of orthography and prosody, e.g.: s.v. ἀπηλιώτης: ‘[The form] ἀϕηλιώτης [appears] on a later table of the winds, IG14.1308, and in Latin authors, Catull.26.3, Seneca QN5.16.4, Gell.2.22.8.’¹⁰ s.v. δρῶπαξ: ‘ᾰ in Lat. gen., Mart.3.74, 10.65.’ In some cases, by contrast, the dictionary will accommodate semantic developments that may—so far as the evidence goes—have occurred exclusively in the Roman world. A particularly well-known case is presented by ϕάσηλος (‘bean’). Although the word’s primary meaning is amply attested in Greek authors, the transferred sense of ‘boat’ or ‘skiff ’ is only supported by Latin texts. The Lexicon must therefore range entirely beyond Hellenic waters: ‘II hence Lat. phaselus, a light boat, canoe, skiff, from its likeness to a bean-pod, Catull.4, Hor. Od.3.2.29.’ Analogously, under πυτίζω (‘spit frequently’) is given the more precise continuation ‘hence Lat. pȳtissare, spit out wine after tasting, Ter.Heaut.457; pytisma, spittle, Vitr.7.4.5, Juv.11.175.’ The latter of these Latin words is without parallel in Greek, despite its manifestly legitimate Hellenic form. It will be worthwhile to turn up a few more examples where Latin texts significantly develop the lexicographical picture. Under ἔμβλημα (‘insertion’) are given seven senses, of which the fourth is attested only in Latin: ‘4. Lat. emblema, mosaic, Lucil.85 Marx, Varro RR3.2.4.’ Likewise, under ἴον (‘violet’) the fifth of five senses is supported only by Pliny the Elder, viz ‘a precious stone of dark colour’. Towards the end of the entry for λῆμμα (III) we read of a particular semantic development: ‘hence, title or argument of an epigram, Lat. lemma, Mart.14.2; theme or thesis, Plin.Ep.4.27.3, Mart.10.59.’ Many of these insights from Latin literature were present from the first edition of the Lexicon onwards. For instance, under ἀρτέμων (‘foresail’), the first edition gives as a second sense ‘the principal pulley in a system, Vitruv.’. Likewise, under κατάληψις, the technical Stoic sense of ‘perception’ is supported by reference to Cicero alone: ‘2. in Stoic. philosophy, comprehension, perception, Lat. comprehensio, Cic. Fin. 3. 5.’¹¹

¹⁰ These references to Latin texts were added in the ninth edition. ¹¹ A case with unusual detail drawn from Latin is provided by κήρωμα. After the most basic sense (‘wax salve’), the entry reads: ‘layer of mud or clay forming the floor of the wrestling ring in

Latin in the Lexicon

51

More importantly, several words of manifestly Greek origin are attested only in Latin. For instance, θερμοπώλιον (‘cook-shop’) merits a place, although it is attested only in ‘Lat form thermopolium, Plaut. Curc.292, Trin.1013’. Given its non-Greek patronage, the term took some time to enter the Lexicon by the back door: it was not included until the fourth edition (1855). Likewise, ὀρνιθοβοσκεῖον (‘bird-feeder’) has the sole entry ‘in Lat. form ornithoboscion, aviary, poultry-house, Varro RR3.9.2, etc.’¹² In some cases there was a wider evidential gap to be bridged. The entry for the apparently innocuous adjective παθικός is forced by lack of Greek data to shift language instantly, as well as gear:¹³ ‘remaining passive: hence Lat. pathicus, i.e. qui muliebria patitur, Juv.2.99.’¹⁴ I will return to the deployment of Latin as a veil for the obscene, evidenced in those fuzzy last three words, in section 3.7. Some Greek lemmata in the Lexicon are still more hopeful in the process of reconstruction followed. In the case of χαιρέϕυλλον the form has to be awarded a second lambda, since in its sole occurrence, the Latin chaerephylon at Colum. 10.110, it has lost one to secure choriambic scansion.¹⁵ Perhaps bolder is the case of νηνία, which is entered as a regular first-declension noun. It is defined, with an almost Balderdash-worthy mix of vaguery and specificity, as a ‘public eulogy on great men, sometimes accompanied by the flute: hence, lament, dirge, only in Lat. nenia, a Gr. word acc. to Cic.Legg.2.24.62.’ And yet, if Cicero (and Pollux 4.79) is mistaken—or bluffing—this is a word that has managed to crash this most august of lexicographical parties.

3.4. THE LATIN OF TECHNICAL TERMINOLOGY In very many cases an established Greek word is chosen to translate a particular, often technical, Latin term. In scores of instances a neologism was required to perform this task satisfactorily. In theory, Liddell and Scott should equate all such Greek terms with their Latin original by the equals sign (=).¹⁶ In practice, the sign is not used with anything approaching consistency until the ninth edition, where we find examples such as: the times of the Empire, Lat. ceroma, locus exercitii utilis . . . aequali et molli ceromate stratus, Cael.Aur.Salut.Praec.35 (ed. V. Rose Anecd.Gr.2.199).’ ¹² Similar cases can be found under, e.g. ῥύτιον (Martial) and τορευτικός (Pliny the Elder). ¹³ The case was made by David Bain, ‘Two submerged items of Greek sexual vocabulary from Aphrodisias’, ZPE 117 (1997) 81 4, at 81 2, that the word may survive in a Greek graffito. ¹⁴ This entry is present from the first edition, which is still coyer in expression: ‘usu. subject to unnatural lust, Lat. qui muliebria patitur’. ¹⁵ Pliny the Elder developed the word further along Latin lines, viz. the hybrid caerifolium. ¹⁶ This avowedly mathematical symbol of identity is glossed in the prefatory list of ‘Signs, etc.’ as ‘equal or equivalent to, the same as’.

52

David Butterfield s.v. ἀρχιέρεια: ‘misspelt ἀρχειέρια IG5(2).313 (Mantinea, ii A.D.): = Lat. virgo Vestalis maxima, D.C.79.9.’ s.v. ἵππαρχος: ‘= Lat. magister equitum, D.S.12.64, Plu.Cam.5, etc.; = praefectus equitum, App.Hisp.47.’ s.v. μύχιος: ‘II οἱ μ. θεοί, = Lat. Penates, D.H.1.67.’ s.v. πλατύσημος: ‘with broad border, π. χιτών, = Lat. tunica laticlavia, D.S.36.7, Str.3.5.1; ἡ π. ἐσθής Hdn.3.11.2; συνθέσεις PHamb.10.15 (ii A.D.): abs., ἡ π. Arr.Epict.1.24.12.’

In a few cases, some Greek authors render the same term in different ways, either through deliberate variatio or unwitting inconsistency. In such cases the Lexicon is content with supplying the simple Latin, rather than seeking to explain the new-fangled diversity in Greek. For instance, four similar but distinct terms used by Dionysius of Halicarnassus—οἰωνιστής, οἰωνόμαντις, οἰωνοπόλος, and οἰωνοσκόπος—are all glossed simply with ‘= Lat. augur’. Although this is a relatively common category among later Greek authors, the editors are typically happy—as with category (ii)—to regard the citation of a Latin term as providing sufficient detail, without the need for further clarification in English.

3.5. THE LATIN OF ANCIENT I DIOM Much the most frequent role of Latin in the Lexicon is to provide parallel information to support the Greek evidence. In many cases these Latin additions closely reflect or echo the Greek idiom requiring illustration. For instance, we find in the ninth edition: s.v. δέμω: ‘δ. ὁδόν, Lat. munire viam, Hdt. 2.124’ s.v. διαπτύω: ‘δ. τὸν χαλινόν, Lat. frenum respuere, Philostr.Im.2.5’ s.v. θέλεος: ‘θ. ἀθέλεος, Lat. nolens volens, A.Supp.862 (lyr.)’ s.v. ὀϕρύκνηστον: ‘ἐρυθριῶντα, οἱ γὰρ ἐρυθριῶντες κνῶνται τὰς ὀϕρῦς, Hsch. (Cf. Lat. homo fronte perfricta, one who has rubbed his brow so often that he can blush no more)’ In these instances it is true to say that Latin phrasing provides a parallel that allows a reader to see that the tenor of Greek thought is not isolated to that language. A less clear case is provided within the entry for ἐκτρίβω, under section II: ‘βίον ἐ. bring life to a wretched end, = Lat. conterere vitam, S.OT248, cf. 428’. This case, though cited from the ninth edition, shows a curious deployment of ‘=’. In earlier editions the sign does not introduce this Latin idiom, which allows the Latin to serve as a supporting gloss. When prefaced by

Latin in the Lexicon

53

the equals sign, however, the phrase raises the unwelcome (and absurd) anachronism that Sophocles’ Oedipus is echoing a specifically Latin idiom. It may be assumed that such illustrative Latin material was present from the appearance of the first edition. In fact, although much of it was indeed deployed from the outset, a considerable proportion was introduced gradatim during subsequent expansions, particularly in the fourth and fifth editions. For reasons that remain unclear, Liddell and Scott believed that including more material from Latin sources furthered the clarity of their lexicography and the utility of their book. In fact, it was only with the arrival of the ninth edition that much of this accretive Latinate material was removed from the Greek-English Lexicon. While this exercise cleared plenty of superfluous matter, as we will see at the close of this survey, it also removed details of independent interest. For instance, τρικυμία, described as ‘the third wave, for the third was supposed to be the largest’, was supported in the first eight editions by the useful comparison ‘as in Lat. the fluctus decumanus’. Likewise, under ὄνος (I.2) the first eight editions compared the phrase ‘περὶ ὄνου σκιᾶς, for an ass’s shadow, i.e. for nothing at all’ with ‘Lat. de lana caprina’; in the ninth, the Latin phrase finds no place. As one final example, under ἄγροικος, Plato’s phrase ἄγροικος σοϕία (Phdr. 229e) is glossed with ‘Lat. crassa Minerva’; despite its interest as a parallel, this Roman idiom is dropped in the ninth edition. This same process of paring back Latin curtailed other more discursive comments. For instance, under ϕλέγμα (II.2), the first edition makes the following observation about later usage: ‘The Latin medical writers retained flegma in the signf. of phlegm, but for inflammation and swelling they said flemen and plemen.’ By the time of the seventh edition (1882) this had been reduced to the curt, but still helpful, observation, ‘The Latin medical writers retained flegma in the same sense.’ In the ninth edition, however, there is no account of the term’s fate in Latin—which may either reflect a decreased interest in its Roman deployment, or a real need to make the most of the limited space available. In grammatical and syntactical surveys, in particular, Latin words are regularly deployed to help orientate the reader. A clear example may be seen under ἄν (IIIc), where the Latin subjunctive is deployed to clarify the sense of both the Greek and the English: ‘βουλοίμην ἄν I should like, Lat. velim (but ἐβουλόμην ἄν I should wish, if it were of any avail, vellem)’. Such clarificatory material, introduced in the seventh edition, still survives in the pages of the ninth. Similarly, under ὅταν (II.4) we still find ‘ὅ. τάχιστα, Lat. cum primum, Ar.Th.1205’; and under ὅπου (II.2) the phrase ὅκου γε is still glossed with ‘Lat. quandoquidem or quippe, Hdt.7.118’.¹⁷ Most strikingly of all, the entry for ὡς (and ὥς), contained in the seventh edition no fewer than sixty Latin words

¹⁷ In the first edition only quandoquidem was given.

54

David Butterfield

throughout its various subdivisions, typically providing the closest parallel to the adverb and conjunction. In the ninth edition, these Latin auxiliaries were removed entirely, leaving English—though far simpler in its variety of constructions—to fend for itself. This is evidently a broad category that covers all cases where Latin words or phrases can provide helpful illustration of the Greek. Nevertheless, it is not broad enough to cover the copious amounts of Latin in earlier editions of the Lexicon which play little or no explanatory role. To that welter of material I will return at the end of this survey.

3 . 6. TH E L A T IN O F SCH O L A R SH IP As boldly announced in the preface to the first edition of the Lexicon, Latin is a language preternaturally well suited for concise and precise annotation. This original edition, just like all its Victorian successors, uses Latin in this fashion throughout, deploying curt Latin abbreviations as a matter of course. Even with the arrival of the ninth edition, such efficient and condensed use of Latin remains common, e.g.: c(um) gen(itivo) pers(onae), v(ide) h(anc) v(ocem), sub (audi), s(i) v(era) l(ectio). All such Latin abbreviations were only at last removed from the Lexicon’s armoury in the Revised Supplement (1996). It is not just this language of the classical apparatus criticus that fills earlier editions of the Lexicon. Alongside these Latin abbreviations, we also find the technical parlance of Latin grammar, itself fossilized in Latin. On the very first page, for instance, the prefix ἀ- is subdivided into four Latin terms, named not only in Greek but also in Latin as alpha privativum, alpha copulativum, alpha intensivum, and alpha euphonicum. The major progresses in grammar, philology, and pedagogy of the late nineteenth century left such practices significantly outdated; it is no surprise that the ninth edition of the Lexicon eradicated all such archaic deployments of grammatical Latin. Perhaps a more surprising presence of Latin is to be found in the citation of contemporary scholarship, which was of course written in this international language of philology. Such instances, buried deep within more complex or controversial definitions, are relatively rare in the first edition but become commoner as the nineteenth century proceeds. For example, under ἀμϕίπολις, the Aeschylean phrase ἀνάγκη ἀμϕίπτολις (Cho. 72) is glossed in the seventh edition as ‘necessitas urbi circumdata’ (Blomf.)’, i.e. ‘compulsion surrounds the city’. Charles Blomfield’s commentary, published in 1824,¹⁸ is treated both as a ¹⁸ Aeschyli Choephoroe ad fidem manuscriptorum emendavit, notas et glossarium adiecit C.J.B. (Cambridge, 1824). The phrase appears in the glossarium at p. 118: the presence of such text specific mini lexica no doubt made Blomfield particularly quotable.

Latin in the Lexicon

55

mark of authority and as one sufficiently accessible as not to require further illustration in English. In a similar fashion, Blomfield is also adduced under ἀμείβω (B IV), where the verb’s appearance at A. Theb. 856 is glossed first as ‘convoys, accompanies it’ before the Latin parenthesis ‘(deducit Blomf.)’.¹⁹ Such citations of contemporary scholarship were not restricted to Englishmen. For instance, under δίψιος, the obscure sense of this adjective at A. Cho. 185 includes the observation that ‘Herm. explains it plenae desiderii, ποθειναί’. Both the Latin and Greek are drawn from the note in Gottfried Hermann’s posthumous edition of 1852. Likewise, under ἄγκαθεν, after the definition ‘on the top’, the editors observe that ‘Herm. interprets it cubito presso, with bent arm, resting on the arm’, before dismissing it as improbable.²⁰ Despite its ambiguous phrasing, this Latin phrase is, of course, not Gottfried Hermann’s, but Horace’s (Odes 1.27.8). The roll-call of scholars whose Latin is worthy of citation on the pages of Liddell and Scott is revealed by the handful of names that merit a place in the list of abbreviations: Bekk(er), B(er)gk, Blomf(ield), Br(unck), Elmsl(ey), Erf(urdt), Gaisf(ord), Herm(ann), Lob(eck), Pors(on), Schneid(er), Schw(eighauser), and Valck(enaer).²¹ The Latin of scholarship is not restricted to classical exegesis. It also plays a role in technical terminology, especially in botany, biology, and astrology. For instance, under ἄγνος (2) is recorded the specific classification Vitex agnuscastus, a precise designation present from the first edition.²² Turning to the sea, σαρδίνη sees the English definition ‘pilchard or sardine’ immediately qualified with the Latin ‘Clupea pilchardus’. As for the stars, under νότιος the phrase νότιος Ἰχθῦς is translated as ‘the constellation Piscis Australis’.

3.7. THE LATIN OF OBSCENITY Perhaps the most notorious role played by Latin in Liddell and Scott is to be found in the context of obscenity. From the first edition through to the ninth, Latin provides a suitably decent veil to hide English expressions that modesty forbids the mouth. This forelock-tugging practice of Liddell, Scott, and several ¹⁹ Aeschyli Septem contra Thebas ad fidem manuscriptorum emendavit, notas et glossarium adiecit C.J.B. (Cambridge, 1812); p. 170 gives the gloss ‘Deduco. Quo sensu nescio an alibi occurrat.’ ²⁰ In the first edition, published a decade before Hermann’s Aeschylus, we find ‘on the arm, i.e. resting on it, Lat. cubito presso’. By the time of the ninth edition, the definition had gone full circle, and Hermann’s interpretation is silently accepted. ²¹ In the ninth edition there still survive a venerable octet of scholars B(er)gk, Blomf(ield), Elmsl(ey), Herm(ann), Lob(eck), Pors(on), Ruhnk(en), and Schw(eighauser) but their writings are no longer cited in Latin. ²² Other botanical terms have arrived later: for instance, under κάνναβις the ninth edition introduces both Cannabis sativa and its wild variant Althaea cannabina (hemp mallow).

56

David Butterfield

of their successors is explored in rich detail elsewhere in this book.²³ For our present purposes, it will suffice to outline the different ways in which Latin is corralled into this often unsavoury task. To set the scene, we may begin with some examples of the grubby uses to which Latin is put where English simply will not do: αἰσχροποιΐα: ‘euphem. for fellatio, Sch.Ar.Nub.296’²⁴ ἀνδροβατέω: ‘= Lat. paedico’ ἀνταποπέρδω: ‘Lat. oppedere’ γλωττοδεψέω: ‘= Lat. fellare, Com.Adesp. 32’²⁵ εὐρύπρωκτος: ‘wide, loose-breeched’, strictly pathicus, catamitus’²⁶ κακοστόματος: ‘foulmouthed, sens. obsc., = Lat. fellator, AP 11.155’²⁷ Λεσβιάζω: ‘do like the Lesbian women’, Lat. fellare’²⁸ παθικός: ‘remaining passive: hence Lat. pathicus, i.e. qui muliebria patitur, Juv.2.99’ πρωκτίζω: ‘= Lat. paedico, Ar. Th. 1124’²⁹ σαλαγέω: ‘trans., sens. obsc., = subagito, Orac.ap.Luc.Alex.50’³⁰ ϕιλοπυγιστής: ‘= Lat. paedico, PTeb. 1.17’ These sexual and scatological cases were evidently thought to be better handled indirectly by direct Latin. And yet, when it comes to the crux of these matters, the Latin deployed must also be indirect. For instance, pudenda (lit. ‘things to be ashamed of ’) occurs many times to describe female genitalia, usually in the phrase pudenda muliebria. On one occasion (ἵππος IV b, ninth ed.) this phrase is expanded to pudenda muliebria et virilia, a surprising collocation in lieu of the Victorian verenda. Likewise, instead of the Latincum-English ‘penis’ we typically find ‘membrum virile’ (e.g., s.vv. ἀνδρεία, καυλός, μύρσινος,³¹ ῥόπαλον, ψωλή).³² Perhaps the clearest instance of this barge-pole poking, as Coker (Chapter 4, this volume) discusses, is to be seen with ψωλή, used of a fully erect penis: from the first edition to the ninth the

²³ In particular, see the following Chapter 4 by Amy Coker. ²⁴ In the earliest editions, only the simple Latin ‘fellatio’ is given. The same development is found s.v. αἰσχροποιός, where ‘euphem. for fellator’ replaced ‘esp. fellator’. ²⁵ In the first edition, the Latin is preceded by the English ‘obscene word’. ²⁶ The Revised Supplement approaches the matter directly: ‘read having a wide anus’. ²⁷ This word is absent from the first eight editions. ²⁸ Λεσβιάς is given no definition in the ninth edition, nor in the seventh; in the first, however, it read ‘a Lesbian woman, Lat. fellatrix.’ ²⁹ This word fails to appear until the ninth edition. ³⁰ The same verb subigitare is used to gloss senses of διασποδέω and κατελαύνω. ³¹ It is an amusing detail that τό μύρρινον is glossed in the first eight editions as ‘the lower part of the membrum virile’, before sober contemplation of an erection allowed a different perspective for the ninth edition, in which it has become the ‘upper part of membrum virile’. ³² πέος rejoices in a brace of members: not only is it explained as ‘membrum virile’ but the Sanskrit cognate pásas is translated into the same phrase.

Latin in the Lexicon

57

word is described solemnly as a ‘membrum virile praeputio retracto’.³³ In a similar vein, ἐντονία, which for the first eight editions was glossed as ‘tension, force’, enjoys more precise focus in the ninth edition—but it is glossed strictly in Latin guise: ‘= Lat. distentio penis, Horap. 1.46’.³⁴ The ninth edition is, at least in this respect, at one with the first. When a new, obscene word is discovered—or an established word is found to be obscene—Jones and Mackenzie reached for Latin to obviate vulgarity.³⁵ For instance, the passive form ἀντιοχεύομαι was described in the first eight editions in the desperately vague terms of being ‘driven against’; in the ninth edition the verb (attested at AP 11.284) is rendered as ‘contrario more futuo’. We see something similar with αἰσχρουργέω, rendered in the first edition simply as ‘to act obscenely’. Later in the century this has been expanded: ‘to act obscenely, masturbare’. The minor contribution of the ninth edition is to correct the active form of the Latin verb to the more typical middle: ‘to act obscenely, esp. = masturbari’.³⁶ It is also noteworthy that ‘Lat.’ has been omitted, which implies that the Latin word is not being cited as analogous, but rather that the lexicographers are choosing to express themselves in Latin. The same direct deployment of Latin occurs elsewhere in the ninth edition: under ϕλάω (II), ‘sens. obsc., = masturbari’;³⁷ under πυγίζω, ‘paedico’; under βινέω: ‘inire, coire, of illicit intercourse’. Although ἐγχέζω and ἐκχέζω are glossed respectively with ‘= Lat. incacare’ and ‘= Lat. ecacare’, κακκάω is translated directly as ‘cacare’. Likewise, when in the ninth edition we find the introduction of αἰδοιολείκτης (‘genital-licker’), it is rendered straight into Latin: ‘= cunnilingus, Hsch. s.v. σκερός’. The presence of the equals sign leaves the reader in no doubt that this is a Latin word—and not (perish the thought!) English. A further Latin marker of prudery is given by the regular deployment of sens obsc, a phrase that appeared frequently throughout the dictionary from its first edition onwards but was studiously omitted from the list of abbreviations. Only in the ninth edition did the phrase at last enter the list of abbreviations, now with the improved orthography of sensu obsceno.³⁸ The ³³ Some changes are most dainty: βάλανος is tweaked from the ‘glans membri virilis’ of earlier editions to ‘glans penis’ in the ninth. ³⁴ A similar change is found s.v. τέτανος (2), where the previous ‘strain, tension’ is clarified, in Latin, as ‘erectio penis’. ³⁵ As a snapshot into the serious head scratching/mopping that lay behind these revisions of more vivid terms, see the letter of A.E. Housman to Stuart Jones about the meaning of λαικάζω (Burnett 2007, II.228). Housman had already earned his stripes in the field by researching his celebrated but necessarily Latin dressed article ‘Praefanda’ (Hermes 66 [1931] 402 12). For more on this verb, see esp. H.D. Jocelyn, ‘A Greek indecency and its students: ΛΑΙΚΑΖΕΙΝ’, PCPS 30 (1980) 12 66. ³⁶ The same correction occurs s.v. ἀναϕλάω. The Revised Supplement has no truck with such diversions: and replaced this entry with ‘to perform sexual perversions’. ³⁷ However, δέϕω is glossed with ‘sens. obsc., = Lat. masturbari’. ³⁸ This is perhaps as good a place as any to apologize for a game that kept my friends and me amused at school among more trying spells of teenagerhood. Flicking through our battered

58

David Butterfield

phrase goes on to be deployed ninety-two times throughout that edition of the Lexicon. By the time of the Revised Supplement—emerging amidst the caterwaul of the Spice Girls—it had lost its purpose and was cast aside.

3 . 8. Q U O D S U P E R ES T : L ATIN WITHOUT A CAUSE If we leave aside these seven principal fields in which Latin is deployed, there still remain very many cases—at least in the first eight editions—where the language crops up. However, for most of these, it is hard to see what role it actually plays—beyond its comforting presence to the British Greek scholar necessarily imbued with Latin. Under the long entry for the preposition πρό, for example, the phrase ὀλέσθαι π πόληος is cited from the Iliad (22.110). Given its simplicity, the phrase is not translated into English; yet it is instead followed by the Latin phrase pro patria mori. Although this Horatian tag is achingly old-hat, it remains a strange and unnecessary gloss. Other cases are similarly odd. Under προσβάλλω (AI2) is found the phrase ‘π. δεῖμα πατρί’ from Euripides’ Ion (585), which is glossed by ‘Lat. incutere amorem alicui’. Under εὐχή the phrase εὐχὴν ἐπιτελέσαι, which is not attributed to any text, is glossed by ‘Lat. vota persolvere’. Under ἀγχώμαλος, the neuter plural adverb is glossed as ‘ἀγχώμαλα ναυμαχεῖν, Lat. aequo Marte pugnare, Th.7.71’. As a final example, under καταβαίνω the absolute use is glossed, irregularly, with ‘= Lat. in certamen descendere, Pi.N.3.42, S.Tr.504(lyr.)’. In all of these cases, the Latin is otiose, and in some cases distracting. It was therefore a mark of progress that one of the major changes made throughout the Lexicon in its ninth edition was the removal of superfluous Latin. This cull of material that had been gathered increasingly over earlier editions was done methodically, replacing the omitted Latin either with additional English or (much more commonly) with nothing at all. The effects of this process are neatly exemplified by the case of the adjective ἐναίσιμος. I reproduce below the first half of the entries from the first and ninth editions (omitting Greek citations): LS¹: ‘fated, sent by destiny, fateful, Lat. fatalis, Il. 2. 353, Od. 2. 159, 182; esp. in good signf., seasonable, Lat. opportunus, Il. 6. 519; in gen. lucky, favourable, boding good, Lat. faustus, Ap. Rh.’ Biddell Liddells, we would seek out ‘sens. obsc.’ as a sure fire marker of words that looked lively. We would then take turns to feign ignorance and ask our Classics master why, in such a case, the ‘sense was obscure’; the more detail and colour you could extract from poor old Mr Lonsdale a man as pure as the driven snow the better.

Latin in the Lexicon

59

LSJ: ‘ominous, fateful, Il. 6.519; Od. 2.159; ib. 182; esp. in good sense, seasonable, of omens, Il. 2.353; generally, favourable, boding good, A. R. 1.438.’ The clarificatory role of these Latin adjectives has been dismissed as entirely superfluous. A similar attitude permeates the edition. A small selection from hundreds of other examples is as follows. The eighth edition reads: s.v. ἀγακλεής: ‘Very glorious, famous, Lat. inclytus’ s.v. ἐνταϕιοπώλης: ‘an undertaker, Lat. libitinarius, Artem. 4. 56’ s.v. ἐπιλογίζομαι: ‘οὐδὲν τοῦτο ἐπελογίσαντο, nullam hujus rei rationem habuerunt, Xen. Hell. 7. 5, 16.’³⁹ s.v. ἔρρω: ‘ἔρρε, a curse, like Lat. abi in malam rem, go with a plague on thee’ ibid.: ‘ἔρρει τὰ ἐμὰ πράγματα, Lat. actum est de me! Xen. Symp. 1. 15.’ s.v. κωτίλλω: ‘to prattle, chatter, chat, Lat. garrire’ For all of these cases the Latin, whatever its explanatory value, is removed without any trace in the ninth edition. The same development is evident in grammatical and syntactical entries. However, despite these concerted efforts to remove these otiose Latinisms, several managed to survive the cull. These vestigial cases do not seem to reflect particular merit or arcane lexicographical precision. Instead, they seem haphazard instances of chance survival into the ninth edition, e.g.: s.v. ἀναγορεύω: ‘= Lat. renuntiari, ὕπατος ἀνηγορευμένος Plu. Mar. 45’ s.v. ἄνεργος: ‘ἔργα ἄ., Lat. facta infecta, E.Hel.363’ s.v. γηθέω: ‘part. γεγηθώς, like χαίρων, Lat. impune’ s.v. ἐξονειδίζω: ‘simply, bring forward, Lat. objicere, τὸ τόλμημ᾽ οἷον ἐξωνεί δισεν E. Ph. 1676’ s.v. μηδαμόθεν: ‘μηδεὶς μ., Lat. nullius filius, D. 21.148’ s.v. ξενία: ‘ξεινίην τινὶ συντίθεσθαι, Lat. hospitium facere cum aliquo, Hdt.1.27, 3.39’ s.v. ξίϕος: ‘2. power of life and death, Lat. jus gladii, Philostr. VA 4.42.’ In some cases the Victorian practice of citing small poetic tags has survived even into the ninth edition. Under θυμόω, the phrase θυμοῦσθαι εἰς κέρας is glossed ‘vent fury with the horns, Virgil’s irasci in cornua’. Similarly, under ὄνυξ the phrase ἐξ ἁπαλῶν ὀ. is glossed as ‘from childhood, Horace’s de tenero ungui’. The great majority, however, have been removed: for instance, under ἀμείβω B1, the phrase ἀμείβεσθαι ὁπλαῖς is glossed in the first edition with ‘of a horse, like Virgil’s sinuat alterna volumina crurum’; by the time of the ninth edition, any mention of Virgil has disappeared. Similarly, under ὡς DI3 the ³⁹ In the first edition the Latin gloss was given in the generic infinitival form: ‘rationem habere alicuius’.

60

David Butterfield

sense of ‘when . . . how’ is glossed ‘so Virgil, Ecl. 8. 41, ut vidi, ut perii, ut me malus abstulit error’ in the eighth; in the ninth Virgil has vanished. The adjective ἀγάλακτος in its occurrence at A. Ag. 718 is supported, for the first eight editions, by ‘Horace’s jam lacte depulsus’, without further reference (= Odes 4.4.15); the ninth edition dispenses with any reference to Latin at all.

3.9. CONCLU SION The first preface of Liddell and Scott’s Lexicon began with a strident defence of their setting Latin aside for English. We have seen that it proved more difficult to remove Latin wholesale than perhaps either eponymous editor anticipated; even the heroic labours of Jones and Mackenzie could not dispense with the language entirely. The first preface went on to close with some Latin. Reflecting on the demanding efforts that their dictionary required, they touch upon the lexicographical work of Samuel Johnson: His labours have been compared to ‘those of the anvil and the mine;’ or even worse— condendaque Lexica mandat Damnatis, poenam pro poenis omnibus unam. These lines are drawn from one of Johnson’s own poems, under the Greek title ‘ΓΝΩΘΙ ΣΕΑΥΤΟΝ’, written on the completion of the fourth, revised edition of his English dictionary in 1772: Lexicon ad finem longo luctamine tandem Scaliger ut duxit, tenuis pertaesus opellae, Vile indignatus studium, nugasque molestas, Ingemit exosus, scribendaque lexica mandat Damnatis, poenam pro poenis omnibus unam. Johnson sees the exhausted Scaliger leave dictionary-writing to the damned; Liddell and Scott concur gravely that the founding (condenda, rather grandiosely) of a new dictionary is indeed a grim punishment. Johnson found that the Latin hexameter best channelled his mixed emotions. Channelling Johnson, Liddell and Scott also found that Latin has advantages that are hard to shake: not only does it reveal the life of the Greek language in the Roman world but it allows an indirect route for the lexicographer to say what one either cannot, or will not, say in the direct dress of English.

4 Obscenity A Problem for the Lexicographer Amy Coker¹

I think it might be curtly dismissed, of course without a quotation, a thing presque introuvable. I think print the word, and say ‘An obscure term applied to the private parts of a woman’. The thing itself is not obscene. (J. Dixon to J.A.H. Murray, editor of the OED MP/17/2/1891 (Murray papers); reprinted at Mugglestone 2000b, 11)

4.1. INTRODUCTION In 1843 Liddell and Scott published the first edition of the Lexicon which would in time become, with eight revisions—and two Supplements—the work familiar to all Classical scholars as LSJ.² In direct contrast with most earlier lexica of Greek, this Lexicon was from its conception a Greek-English work, and some 170 years ago it was the use of English which was stressed in particular in the original Preface.³ This clarion call for the vernacular therefore makes it all the more surprising when we find Latin rather than English used in the translation of certain words.⁴ ¹ This work was funded in part by the Leverhulme Trust in the form of an Early Career Fellowship hosted by the University of Manchester; the support of the Trust is here warmly acknowledged, as is the unfailing editorial enthusiasm and generosity of Chris Stray. ² The abbreviation ‘LS⁹/LSJ’ is used in this chapter where changes between editions are being discussed to stress the importance of the history of revisions. ³ See Stray, this volume, pp. 7 8, and Williamson, pp. 395 400 for the choice of English; Considine, this volume, pp. 26 7 discusses the contemporary intellectual climate. ⁴ On the famous phrase ‘the decent obscurity of a learned language’, see Bain and Coker 2014, 404 n. 48. Amy Coker, Obscenity: A Problem for the Lexicographer In: Liddell and Scott: The History, Methodology, and Languages of the World’s Leading Lexicon of Ancient Greek. Edited by: Christopher Stray, Michael Clarke, and Joshua T. Katz, Oxford University Press (2019). © Oxford University Press. DOI: 10.1093/oso/9780198810803.003.0004

62

Amy Coker

To give a single example, consider the following lines from Aristophanes’ Lysistrata, spoken by the Spartan woman Lampito in response to Lysistrata’s revelation of her plan for peace (ll. 142–3, text from Wilson OCT 2007): χαλεπὰ μὲν ναὶ τὼ σιὼ γυναῖκάς ἐσθ᾽ ὑπνῶν ἄνευ ψωλᾶς μόνας. By the Two Gods, it’s tough for women to get to sleep when they’re on their own without psolas If a reader was unfamiliar with the emboldened word—which is a rare one— and thus consulted the most recent edition of LSJ to complete an emerging translation, this would be the result of the enquiry: LSJ (2029 i⁵): ψωλή, Dor. ψωλά, ἡ, prop. fem. of ψωλός, membrum virile praeputio retracto, Ar.Lys.143, Av.560 (anap.), Suppl.Epigr.3.596 (Panticapaeum, v B.C.).

To find Latin here rather than English is a striking example of lexicographical prudery. This aspect of the Lexicon is well known to its users, and most will have encountered this phenomenon or similar examples of lexicographical chicanery (no doubt with their own choice examples), found when the editors deal with material they deem in some way inappropriate.⁶ Several central texts of the classical canon, such as the plays of Aristophanes, and indeed Homer, contain many references to topics such as sex, defecation, and bodily functions which in the England of the eighteenth and nineteenth centuries had come to be seen as inappropriate for public discussion. For a lexicon with the claims implicit in ours, this means that the societal constraints of propriety clash with the principles of comprehensiveness and vernacular clarity.⁷ To write a lexicon in English is a striking break with tradition, but one which thus presented the editors with a problem.⁸ If every Greek word should be allowed ‘to tell its own story’, how are the stories of words for censured facets of life to be told, and censured words rendered, in a lexicon where the target language is English? This is the problem Dixon wrestles with at the head of this chapter: he is at pains to stress that ‘the thing itself ’, i.e. the female genital region, is ‘not obscene’, but that words indicating this body part must be carefully chosen. The word in question is English ‘cunt’, still a highly offensive word and a problem for the OED, for ⁵ i or ii with a page number refers to column number throughout. ⁶ In addition to those discussed in the rest of this chapter, see Butterfield, this volume, pp. 55 8, and Craik, this volume, p. 149 on ὀργάω. ⁷ As explored for example in the collection of essays on the expurgation of Greek and Latin texts edited by Harrison and Stray 2012a; see also Dover’s classic discussion, with a focus on Greek (Dover 1981). ⁸ As posed by C.S. Lewis, in an oft quoted remark: ‘As soon as you deal with it explicitly, you are forced to choose between the language of the nursery, the gutter and the anatomy class’ (quoted in Hughes 1998, 1).

Obscenity: A Problem for the Lexicographer

63

reasons which will be given shortly. It is just these same choices of inclusion and labelling which the editors of Liddell and Scott faced—and with which lexicographers of all languages continue to struggle. This chapter is my own exploration in the Lexicon, into the less-savoury material which one might assume the editors would have wanted to avoid. This chapter is not concerned with whether, for example, the entry for ψωλή accurately represents what we can say about the distribution and meaning of the word in Greek, or the utility of the adherence by the work to the ‘historical principle’;⁹ rather here the interest is in the decision by the editors to identify this word for a specific part of the male genital apparatus in Latin, and not English. The use of Latin to bowdlerize the more ‘difficult’ aspects of Classical texts, or indeed of sex more broadly when discussed in English, is not unusual in the nineteenth and even twentieth centuries. Krafft-Ebing’s influential Psychopathia Sexualis (1886), for example, was written partly in Latin to conceal its contents,¹⁰ and as late as the 1970s the word ‘cunt’ was identified as the prudish ‘pudendum’ when its inclusion in the OED was discussed in K. M. E. Murray, Caught in the Web of Words: James A.H. Murray and the Oxford English Dictionary (1977).¹¹ The repeated revision of Liddell and Scott, together with its debts to earlier lexica,¹² means that the Lexicon has to be considered as an historical artefact. Just what is viewed as ‘obscene’—in the sense of matter which should be hidden from public view or from certain groups—of course shifts according to the (pliable) rules of good conduct and verbal propriety dictated by society, or at least those most vocal in their claims to moral authority.¹³ As will be seen ⁹ For an assessment of the treatment of some words from medical texts, see Craik, Chapter 9, this volume. ψωλή itself is still rarely attested; Henderson 1991, 110 collects some examples with a focus on Old Comedy, to which can be added among others a label on a drawing in the ‘Indecent Proposal’ letter (P.Oxy 3070), a mock epitaph on an ostracon dated to the 3rd century BC which appears to describe Cleitorius’ sticky end (SH 975 = FGE 147), and several examples from graffiti. This distribution suggests that the tone of ψωλή was offensive in the Classical and Hellenistic periods and beyond, quite apart from the fact that it refers to the erect penis; see Coker forthcoming 2019, ‘Looking for Dysphemistic Language in Ancient Greek (or: how filthy was Cleopatra?)’, Journal of Historical Pragmatics 20.2 (Special issue, ed. K. Ridealgh), on the methodological problems present in assessing connotative value in historical languages. ¹⁰ See Orrells 2015, 92 7, who stresses in addition the utility of precision granted by the use of Latin, as well as the cultural influence of Krafft Ebing’s adherence to the Roman conceptualiza tion of sex acts as active and passive. ¹¹ Noted by Mugglestone 2000b, 11 n. 14. Despite the revolutions in the attitudes to sex and sexuality during the twentieth century, some aspects of life are still subject to censure in certain contexts, particularly the classroom; see the final notes in this chapter for some observations on two recent Greek textbooks for the school market. ¹² See the sketch in Stray 2010b (especially pp. 103 11) and Stray, Chapter 1, this volume, on debts to nineteenth century and earlier lexica; ancient lexicographical material from Hesychius and the Atticist lexicographers was also taken on verbatim (see Chadwick 1994 and 1996). ¹³ This chapter is thus interested in prudery rather than probity; it is not the aim to investigate the private lives of those who worked on the dictionary, but rather their composite public face as manifested in this monumental work of scholarship. ‘Obscenity’ as a concept is difficult to pin

64

Amy Coker

throughout this chapter, each new edition is subject to contemporaneous concerns about the publication of obscene material, but the lack of comprehensive revision of the work means that such changes may not be systematically applied.¹⁴ In short, to find Latin in the ninth edition does not necessarily imply that in 1940 terms such as ‘penis with foreskin drawn back’ or ‘erection’ were viewed as unsuitable for print per se in English: rather, we must view these terms as possible historical relics from earlier incarnations of the text. What follows in this chapter considers the topics and words which were classed as requiring censorship in the Lexicon, and traces the various strategies employed to shield the reader from material deemed to be morally dubious or dangerous. While at times an increasing prudishness about sexual matters can be observed in the Lexicon, we shall see that in many respects its contents are nonetheless an astonishingly long way from the Victorian ‘buttoned-up’ stereotype, and that the work contains some surprises for the assiduous reader.

4.2. THE N INETEENTH-CENTURY CONTEXT The Victorians, especially those of the end of the nineteenth century, have long had a reputation for prudery, reinforced through oft-repeated anecdotes about the unmentionability of ankles and trousers, even if this is somewhat of a mirage.¹⁵ However, despite this image of strict moral rectitude, individuals indulged in their own private affairs of various biological kinds.¹⁶ Consider the following short excerpt from My Secret Life by ‘Walter’, a (probably nonfiction) diary of the author’s sexual exploits (quoted in Mason 1995, 46–7): down; Heath (2007, 512 14) provides a useful discussion, stressing that it is the reaction of the ‘viewing subject’ which is crucial (p. 514). ¹⁴ Constant revision has had both positive and negative effects, and Lee’s comment that ‘LSJ’s entry for ἀγαπητός is a mess’ (Lee 2010b, 125) could stand for many others. Many of the reasons he posits (pp. 126 8) for this messiness control of the evidence, dependence on predecessors, piecemeal revision, control of secondary literature, physical constraints, and method of indicat ing meaning are apposite to what follows. ¹⁵ Mason 1995 challenges this still pervasive view. Nonetheless the word ‘trousers’ was sometimes avoided, for example by the thirteen year old R.C. Jebb in 1854, writing home from school to his father: ‘I have been very busy indeed lately, and have not had time to write. As to the “nefanda” I told the “propeller” of the College Car to call for them but he said he could not get them without a note. So I suppose I must execute an Epistle, which shall be short and pithy, containing a peremptory injunction to deliver ‘The “Unspeakables” ’ and not keep me in longer suspense’ (Letter to father of 27 March 1854, in Stray 2013a, 16). A few days later (31 March), he ventures ‘breeches’ (17). There will be more references to trousers shortly. ¹⁶ Hughes 1998, 151 6 stresses this ‘schizophrenic quality’ of Victorian culture, and observes how the original OED ‘included all the religious and racial swear words, but [has] omitted several of the most genital’ (238). Sex may be a particular moral concern, but excretory habits are hidden too (and continue to be).

Obscenity: A Problem for the Lexicographer

65

all this did not give her an excess of sexual pleasure, with all her fucking, frigging and gamahuching, she looked the very picture of health and strength.

Note the language used here: this text not only talks directly about sex, but does so using vocabulary which is offensive (e.g. ‘fucking’), including some words which are now all but obsolete in English (e.g. ‘gamahuching’).¹⁷ This reminds us that, quite apart from any difference in social mores, the English of the nineteenth century is different to that of the contemporary anglophone world, and that when reading the Lexicon we must bear this in mind. When we use this work we are essentially reading an ancient, alien society through another more recent, but still to some extent foreign, cultural lens (as stressed by Williamson, Chapter 2, this volume). There was also a large amount of material circulating in the nineteenth century in public which counters this Victorian stereotype, from slang dictionaries to ‘fast’ (in the contemporary parlance¹⁸) publications such as Paul Pry or The Town.¹⁹ Crucially, many of these publications were inexpensive and thus accessible to a wide public:²⁰ the vast increase in literacy through the mid-eighteenth and nineteenth centuries, with primary education becoming compulsory in the 1880s, also led for the first time to a mass reading public. For this public, of multiple social layers, sex sold. In turn, this led to an increasing concern that vulnerable people—especially women and children—were encountering harmful material. In short, material of a sexual nature—‘obscene’ material— became democratized, and became increasingly viewed as requiring control. That there was a strong class element to attitudes to this material, and in the rhetoric surrounding appropriate linguistic behaviour, is highlighted by a short anecdote from a collection of often illicit folklore called the Κρυπτάδια (1883–1911):²¹ A young lady was out riding, accompanied by her groom. She fell off her horse and in so doing displayed some of her charms; but jumped up very quickly and said to her groom: ‘Did you see my agility, John?’ ‘Yes, miss,’ said he, ‘but I never heard it called by that name before!’ Another version has it: ‘Yes, miss, but we calls it cunt in the kitchen!’

¹⁷ Glossed by the OED (s.v. gamahuche, v.) as ‘To practise fellatio or cunnilingus on or with (a person). Also intr.’. Here the word is stated to be in current use, although rare: it is not part of the author’s own UK English idiolect, nor any of her peers’. ¹⁸ OED s.v. fast, adj. 10. ¹⁹ For examples see Mason 1995, 163 4; 220 1, including the erotically charged ‘Anonyma’ novels of the mid 1860s (pp. 86, 95). Hughes 1998, 158 62 discusses nineteenth century dictionaries of slang, with choice examples; see Coleman 2004 10 for a full account of such works. ²⁰ This is in part driven by improvements in printing technology, coupled with the abolition of stamp tax in 1855 and the tax on paper in 1861 (Lewis 2003, 144), and the advent of popular newspapers in the 1880s. ²¹ Volume IV, pp. 394 5; reprinted at Fryer 1963, 47.

66

Amy Coker

Here laid out are a striking set of polarities: the unintentionally euphemistic language of the high-class female (‘agility’) versus the crude or dysphemistic language of the servant male groom (‘cunt’); note too the narrator’s ‘charms’, a playfully vague metaphor. The story was told for its humour, but the anxiety at this period was that the two sets of behaviours should become confused, and the young lady be exposed to, and thus begin to use, the groom’s kitchen vernacular rather than the diction of the drawing room. The Obscene Publications Act of 1857, for example, was aimed almost entirely at the suppression of the activities of printers on Holywell Street in central London,²² not only because this street was the main site of the publication of illicit material in the city, but also because it was a public thoroughfare running parallel to the Strand, on which such printed wares were openly on display and visible to any pedestrian, male or female (Nead 2000, 182).²³ It was one thing for ‘the middle class man to indulge himself with expensive works of erotica in the privacy of his study. It was quite another for cheaper, milder forms to be visible openly in the public streets, or to fall into the hands of women and children within the sanctity of the home itself ’ (Lewis 2003, 154). Note in this light that the story of the groom, the lady, and her ‘agility’ is one among many tales from range of European countries collected in a set of four small volumes given the Greek title Κρυπτάδια. As its editors state in their preface, the title was chosen in part because ‘it will not tempt the hand of the curious on the bookshelf of the library’.²⁴ The intended audience was presumably those educated people with enough Greek to be able to decipher the letters, at the exclusion of the Greek-less, for whom the contents of the work could be a dangerous thing. The Greek letters here—like the Latin ones in Liddell and Scott—were used for their segregational, cryptic effect. This veiling effect is often achieved through the use of a foreign language. For English writers French may be used—or often (potentially long-winded) circumlocution, and these strategies were also employed by two contributors to the dictionary now known as the Oxford English Dictionary (see Considine, Chapter 21, this volume, pp. 395–411).²⁵ The OED was published later in the

²² On the Act see St. John Stevas 1956, 18, and Lewis 2003, 153. Prior to the Act, there were prosecutions for ‘obscene libel’ in England as early as 1757 (the case of Edmund Curll for Venus in the Cloister, or the Nun in her Smock); for the earlier ‘Act to Restraine Abuses of Players’ of 1606 and similar controls on language, especially that of the theatre, see Hughes 1998, 103 5. ²³ The street was demolished at the end of the nineteenth century to make way for Aldwych and Kingsway; Nead 2000, 161 89 gives a short history of the street and its place within the nineteenth century cityscape. ²⁴ Volume 1, p. x: ‘Et d’abord notre titre de Kruptadia, c’est à dire les sujets secrets, nous l’avons choisi précisément pour son aspect hirsute, barbare et rébarbatif, parce qu’il est intelli gible au plus grand nombre et qu’inscrit sur le dos d’un livre, il ne tentera pas la main d’un curieux sur le rayon de bibliothèque.’ ²⁵ The next few paragraphs draw heavily on Lynda Mugglestone’s work, in particular Mugglestone 2000b, 2005 (esp. 84 7), 2007.

Obscenity: A Problem for the Lexicographer

67

nineteenth century than the earlier editions of Liddell and Scott, but nonetheless its form and contents are indicative of some of the same decision-making processes faced by our Greek lexicographers at this period.²⁶ Despite the OED’s aims of objectivity, impartiality and descriptive—not prescriptive— lexicography (Mugglestone 2000b, 10), some words considered obscene were excluded, unlike some earlier dictionaries of English.²⁷ Such omissions did not however go unnoticed, and a John Hamilton ‘ventured to send [ . . . ] a word that is not found’, as follows (reprinted in Mugglestone 2000b, 11):²⁸ It is an old English word of Teutonic origin, & is just as good English (though by the nature of things not so much used in polite society) as the words, leg, arm, heart, stomach, & other parts of the body . . . It has cognates in French, German, & Dutch . . . The mere fact of its being used in a vulgar way, does not ban it from the English language & its absence from your Dictionary is brought forcibly before one’s eyes as it was before mine in rapidly turning over its pages by the presence of the same syllable in big letters as a contraction of Contra. In reality it is no more vulgar than bowels or womb, & hardly so vulgar as a certain word inserted in your Dictionary to indicate the posterior under letter B. (Murray Papers 1/9/1899)

It is an additional ‘c-word’ which is commented on by another contributor, the surgeon James Dixon. He urged that not to include obscenities would be ‘cowardly’ (Mugglestone 2000b, 11), and offers the compromise for ‘cunt’ reproduced at the head of this chapter.²⁹ These same broad concerns over readership—particularly who is reading— in the late Victorian period can be seen to play out too in classical scholarship. The late nineteenth and early twentieth centuries saw the publication of huge numbers of cheap school editions of Latin and Greek texts, and editors and publishers ‘felt obliged to balance scholarship with propriety’ (Harrison and Stray 2012b, 3). In a small-scale study of Aristophanic school texts, Ian Ruffell notes that in particular sex, homosexual acts, sexual organs, and female sexuality are suppressed: as Ruffell says, ‘whatever else you can say about Aristophanic gender stereotypes, they are a mile away from the ideologies of ²⁶ Chris Stray has written (2010b: 112 13 n. 11; also Stray, this volume, p. 27) of a nice hint of the ‘naughty’ humour of some of Liddell and Scott’s contemporaries at the lexicographers’ explanation of the initial alpha of ἄλοχος (a word usually translated as ‘wife’) as ‘copul[ative]’, even if the famous pair themselves did not see the funny side. ²⁷ For example ‘piss’ and ‘fart’ are frequently found in such earlier works, ‘fucke’ and ‘cunt’ more rarely (Mugglestone 2005, 85 6). ²⁸ Implying that John Hamilton, like so many other historical and contemporary users of lexica, had looked up the taboo words in question; Hughes 1998, 157 8 retells a wonderful anecdote about Dr Johnson’s response to two ladies who had criticized his omission of ‘naughty words’ in his own dictionary ‘What! my dears! then you have been looking for them?’ ²⁹ As Mugglestone notes, ‘cunt’ was finally included in the 1972 Supplement (2000b, 11 n. 14); similar concerns can be seen over ‘twat’, which was originally defined in the OED with Latin ‘pudendum muliebre’ (Mugglestone 2005, 85).

68

Amy Coker

the period’ (Ruffell 2012, 39).³⁰ The comments of Deana Heath, in reference to censorship and translation throughout the British Empire in the midnineteenth century, are equally apposite if we think of ancient Greece as one of the territories of Empire: ‘it was common for inexpensive or translated editions of works to be censored but not expensive ones, or those in their original languages’ (Heath 2007, 511). This can be distilled into attempts by the foreign-language reading elite of high, ‘disinterested’ culture to maintain distance between themselves and the ‘obscenity’ of the low culture, associated with the working classes and predicated upon pleasures of the appetite (2007, 514). Classics too could be democratized, but not too far. Towards the end of the nineteenth century some voices were particularly critical of women reading Greek, especially with the first admissions of women to Classics degrees. Such students would be expected to be using Liddell and Scott, and as we shall see this may have played a part in how the contents of its constituent entries were reshaped by revisers of the Lexicon at this period.

4.3. PLAIN S PEAKING IN THE VERNACULAR: ENGLISH W ORDS IN THE LEXICON It comes as no surprise that LSJ avoids the use of offensive English words in its glosses, even for Greek words likely to have had such tone originally.³¹ A search of the English text of LSJ³² reveals no uses of a number of ‘bad words’ with which many Greek words could have been translated, including a number of obsolete terms.³³ Similarly, some other English words not included are ‘dildo’ and ‘cundum’/‘condom’: it can be suggested that these topics were ³⁰ Some simply refused to see anything which contradicted their conception of the Greeks as representing the peak of civilization: Gladstone, for example, writing in 1858 on elite ladies bathing men in Homer states (quoted in Fryer 1963, 196): ‘It is almost of itself incredible, that habitually, among persons of the highest rank and character, and without any necessity at all, such things should take place. And, as it is not credible, so neither, I think, is it true.’ So too, (p. 312, n. 19): ‘The statement that water was poured over [Odysseus’] head and shoulders, as he sat in the bath, evidently implies that what might be called essential decency was preserved [ . . . ]. The statements of Homer give no ground whatever for sinister or disparaging imputation.’ As Blanshard 2010, 14 15 puts it, Gladstone was a keen Homeric scholar, but his Greeks were always ‘decorously clad’. Dover 1981 comes to similar conclusions. ³¹ Compare Williamson, Chapter 2, this volume, for the attempts in the Lexicon to capture ‘style’ in English. ³² A systematic search is possible only because the English text of LSJ is now searchable electronically for English/Latin words via several digital platforms, e.g. that provided by the Thesaurus Linguae Graecae at http://stephanus.tlg.uci.edu/lsj (last accessed 21/12/2015). ³³ There is no evidence at least for the following words: ‘frig’, ‘gamahuch’, ‘fuck’, ‘cunt’, ‘twat’, ‘dick’, ‘shag’, ‘tit’, ‘boobs’, ‘shit’ (compare the sketch of words about which the OED was anxious given at Mugglestone 2005, 84 5). I imagine this list could be greatly expanded.

Obscenity: A Problem for the Lexicographer

69

considered inappropriate for mention in English, especially since they were also subject to correspondingly moralizing entries in the OED.³⁴ In contrast, however, there are many English terms which we might perhaps be surprised to find in a lexicon with its origins in the nineteenth century, when there are at least some individuals who are rumoured to have claimed in the latter decades of Victoria’s reign that the uttering of ‘trousers’ or ‘breeches’ is inappropriate in polite company.³⁵ We find for example the word ‘sex’ used a number of times, but almost always only in the sense of biological gender,³⁶ and there is frequent use of the phrase ‘sexual intercourse’.³⁷ ‘Bastard’ is similarly used a number of times, usually in the core sense of the word as indicating illegitimacy.³⁸ ‘Whore’, ‘prostitute’, ‘wench’,³⁹ ‘strumpet’, ‘bawd’, ‘brothel’, and ‘harlot’ are used throughout the work, in both the ninth and first editions, with varying degrees of frequency.⁴⁰ ‘Menstruation’ and ‘menses’ also both appear in the first and ninth editions. Present also, however, are other English terms for genitalia and other body parts about which we may predict anxiety: there are many examples of ‘testicles’, ‘scrotum’, and ‘breast’, as well as ‘nipple’ and ‘clitoris’.⁴¹

³⁴ See Mugglestone 2005, 84 7 on the OED. On the discovery at the end of the nineteenth century of a new Greek word for ‘dildo’ (βαυβών) in the Herodas papyrus (Mimiamb 6), and reactions to this poem in the context of a growing number of female classicists, see Orrells 2015, 135 41; βαυβών appears no earlier than LS⁹/LSJ, glossed as ‘= ὄλισβος’. ³⁵ See the examples collected at OED s.v. ‘unmentionable’, B 1, ‘inexpressible’, B 2 and ‘ineffable’, B 1 for euphemistic avoidance of the words of this period, and the comic vignettes by Dickens noted in Hughes 1998, 153; and cf. note 15 above. It is pleasing to note that the word ‘trousers’ does appear in LS⁹/LSJ in the glosses for ἀναξυρίδες, θύλακος, and σαράβαρα, all three of which go back to LS¹ (for the final example in LS¹ s.v. σαραπάραι); in LS¹ the spelling ‘trowsers’ is found for the first two words. The word ‘trews’, s.v. LSJ/LS⁹ βράκαι, is a relic from this earlier era. ‘Breeches’ is also found in several places, often in reference to the anus as well as male apparel, e.g. λακκόπρωκτος ‘loose breeched’. ³⁶ This is true for fourteen words in LSJ (e.g. ἄρσην, γένος, γύνανδρος, διϕυής, στυγάνωρ), and also for the same set in LS¹; the exception is a reference to ‘sexual intercourse’ in διϕυής in LS¹, which is subsequently revised out of the Lexicon. ³⁷ E.g. μίξις (II), συνουσιάζω (Ι), ϕιλέω (Ι 3 b). ³⁸ E.g. νόθος, σκότιος, ψευδάγχουσα; there are also a number of zoological and botanical names in LSJ which include the word, e.g. κύτισος ‘bastard ebony’. LS¹ in contrast uses ‘bastard’ in the glosses for these words indicating illegitimacy only once, for μητρόξενος. Other Greek words for which LSJ uses ‘bastard’ are either not present in this earlier edition, or are circumlocuted, e.g. νόθος is given as ‘illegitimate, born out of wedlock’. ³⁹ ‘Wench’ as a verb is infamously associated with the entries for λαικάζω and λαικαστής it only appears elsewhere in LSJ in λήκημα; the Revised Supplement, in its rewritten entry for λαικάζω, replaces the ‘wench’ with ‘practice fellatio on’. See in particular Jocelyn 1980 for a detailed account of the ‘indecency’ λαικάζω, and Bain 1991, 74 7; Henderson 1991, 153 collects examples from Old Comedy. ⁴⁰ ‘Whore’ and ‘bawd’ only appear a handful of times in the editions examined; ‘strumpet’ is more common in LS¹ among the searches performed, e.g. also appearing in the entries for χαμαιτυπία, πόρνη and ἑταῖρος. ⁴¹ Of the four lemmata in which ‘clitoris’ appears in LSJ, only two are found in LS¹, νυμϕή and κλειτόρις; the first of these is glossed with ‘part of the pudenda muliebria’, with ‘clitoris’ thus appearing only in the entry for κλειτόρις. The English word ‘clitoris’ is attested from 1615

70

Amy Coker

‘Prepuce’,⁴² ‘foreskin’, and ‘phallus’ are also all present a handful of times.⁴³ The verb ‘masturbate’ does not appear, but there is a single example of ‘masturbation’ (ἀναϕλασμός), and an ‘ejaculate’ (ἐκμιαίνω).⁴⁴ English ‘penis’ does appear in LSJ, although the Greek words defined thus are treated differently in earlier editions (see below p. 73); this case is complicated by the fact that this string of five letters can represent an English or Latin word, with the distinction not always clear. In a similar vein the inclusions of ‘fellatio’ (αἰσχροποιΐα, αἰσχρότης), and ‘cunnilingus’ (αἰδοιολείκτης, λείκτης), which might be thought at first glance to be English terms, are particularly surprising.⁴⁵ Of the first pair, αἰσχροποιΐα is so glossed in LS¹, and αἰσχρότης (II) acquires the gloss ‘euphem. for fellatio’ in LS⁵ (1861). Given that the OED appears to give the first English example of ‘fellatio’ (s.v.) as dating to 1892, the Lexicon strictly speaking uses Latin here, and uses a word which, although familiar to speakers of contemporary English, had little or no currency in the language of the nineteenth century.⁴⁶ Something different and more curious operates for αἰδοιολείκτης and λείκτης: these do not appear in the Lexicon until LSJ, by which time the word ‘cunnilingus’ had become domiciled in English, but for both words the sense intended by the gloss is the Latin meaning of cunnilingus as ‘one who performs oral sex on a woman’ (equivalent to OED s.v. ‘cunnilingus’, sense 1), rather than the now more common sense in English of ‘cunnilingus’ as the performance of the sex act in abstract terms (sense 2). This reminds us again that it is difficult to make any concrete statements about the connotative effects, or tone, of English words to a reader at any given period other than the present (and indeed sometimes Latin words may have ‘become’ English, with a concomitant shift in the likely reaction of a reader): their inclusion is however positive evidence that the topics they indicate were mentionable in English in the work, even if the dysphemistic counterparts to these words were not.

according to the OED. ‘vagina’ appears three times in LSJ, although none of these lemmata are so glossed in LS¹. ⁴² English ‘prepuce’ can refer to the fold of skin covering the clitoris as well as the foreskin, although the examples in LSJ all appear to refer to male, not female, anatomy. ⁴³ All these English words appear at least once too in the first edition, although many Greek words so glossed in LSJ have shorter or different definitions in LS¹. Interestingly, there is a general shift in spelling from ‘phallos’ in the first edition to ‘phallus’ in LSJ (although ‘phallus’ appears in LS1, σκύτινος). On the addition of examples of ‘foreskin’, see below. ⁴⁴ In all editions earlier than the ninth ἀναϕλασμός is glossed instead with ‘Lat. masturbatio’, and ἐκμιαίνω with ‘effluxu seminis pollui’. ⁴⁵ Although λείκτης in LSJ is glossed ‘= Lat. cunnilingus’. ⁴⁶ It is indicative of the initial restriction of the word to scholarly circles that earlier examples in the lemma in the OED (s.v. ‘fellatio, n.’), both of ‘fellation’, have specifically classical reference. The next two, from the English translation of von Krafft Ebing’s Psychopathia Sexualis (1897) and Havelock Ellis (1897), are also deeply imbued with classical learning.

Obscenity: A Problem for the Lexicographer

71

4.4. OVERT MARKERS OF O BSCENE CO NTENT: ‘S E N S . OB S C. ’ The device ‘sens. obsc.’, standing for ‘sensu obsceno’⁴⁷ ‘with obscene meaning’ is the most explicit way the Lexicon indicates a matter which is viewed as obscene or lewd. The abbreviation is found ninety-one times in the text of LSJ.⁴⁸ One example is found in the entry for δέϕω, where sens. obsc. is added to clarify the metaphor of manipulation with the hand as an indicator of masturbation: (LSJ 382 ii) δέϕω, soften by working with the hand: δ. ἑαυτόν, sens. obsc., = Lat. masturbari, Eub.120.5: Med., Ar. Eq.24, Pax 290

Very many of these examples are often accompanied by a ‘metaph.’ or ‘Ar.’ [= Aristophanes], implying a transferred use of the basic meaning of the word—usually innocuous—to matters sexual or scatological. The reader is left to untangle the meaning at some level, but nonetheless these obscene metaphors or uses are marked as existing in the extant canon. The fact that LSJ is willing to highlight such examples, which could often easily pass without mention to the casual reader, is striking: the Lexicon does not shy away from alerting its readers to the presence of innuendo and sex in ancient texts, even if those matters are to some extent obfuscated. The fact that many such examples are found in Aristophanes, a canonical author despite his ‘vulgarity’, together with an ethic of comprehensivity, can be seen as drivers of this process. This is similar to the frequent addition of the terms membrum virile or pudenda muliebria to otherwise seemingly innocuous words, as discussed below.

4.5. COVERT MARKERS OF OBSCENITY: L ATIN I N THE LE XI CON The use of Latin in translation in place of English⁴⁹ is one of the clearest indicators that some bowdlerization is at work, as in the example of ψωλή at the beginning of this chapter.⁵⁰ Of sixteen Greek words belonging to the ⁴⁷ As given in LSJ’s General List of Abbreviations; the Latin is untranslated, but this is in common with other Latin abbreviations. ‘sens. obsc.’ is not listed as an abbreviation in editions earlier than the ninth, nor in the Abridged or Intermediate versions. The phrase sensu obscoeno is found written in full in a number of entries in earlier editions, e.g. LS¹ τέτανος, LS² στύω; in the first edition the use of the phrase ‘freq. in Ar.’ (e.g. s.v. χέζω and πεός), is perhaps also specifically used imply likely offensive tone. ⁴⁸ For example in γαμέω, διακροτέω, ἐπιβλής, κακοστόματος, περαίνω, ϕλάω, ψηνίζω, etc. ⁴⁹ For the material in this section compare also Butterfield, this volume, pp. 55 8. ⁵⁰ Note that Latin may also be used for technical terms, e.g. animal or plant names, astro nomical terms, etc. It is my impression that such lemmata will also contain English, rather than the exclusive use of Latin (see for example κτείς ‘comb’, but sense 8 ‘caruncula lachrymalis’, part

72

Amy Coker

sexual or scatological vocabulary taken as a sample,⁵¹ nine use some Latin in their entries in LSJ/LS⁹ (βινέω, δέϕω, κύσθος, πέος, πόσθη, πυγίζω, σάθη, στύω, ψωλή) and five of these Latin only (κύσθος, πέος, πυγίζω, σάθη, ψωλή). There is little substantive variation in the treatment of most of these words in earlier editions of the Lexicon, which speaks to the age of the material.⁵² However, we can observe a change in the use of Latin over the lifetime of Liddell and Scott for some words as follows: κύσθος LS¹ – LS⁶: ‘any hollow: the pudenda muliebria’ (1843–69) LS⁷ onwards: ‘pudenda muliebria’ (1882–) πέος LS¹ – LS6 ‘membrum virile, penis’ (1843–69) LS⁷ onwards: ‘membrum virile’ (1882–) πυγίζω LS¹ – LS⁴ ‘strike on the buttocks, II. paedicari’ (1843–55) LS⁵ onwards: ‘paedico’ (1861–) As seen here, the only changes are the removal of material, specifically English words. The result is a complete retreat to Latin by the early 1880s for κύσθος, πέος and πυγίζω, perhaps linked to increasing anxieties about sexual propriety at this period.⁵³ This is in contrast with what is seen at a slightly later period with the inclusion of specialized medical vocabulary, discussed towards the end of this chapter: note that despite the traces of a more liberated approach to the vocabulary of human genitalia in LSJ, the word ‘penis’ is not reincorporated into the entry for πέος here, nor in the Revised Supplement.⁵⁴ Returning to our sample list of Greek words, those where only Latin is used indicate the male or female genitalia, and anal intercourse, suggesting perhaps that these were the matters about which the editors were most anxious; the additional four where some Latin is found are also all either nouns indicating

of the eye). The reverse may of course also be true: a generic Latin tag could be a way of avoiding having to elicit precise meaning from a difficult example: see the discussion in the final section of this chapter. ⁵¹ That is: βινέω, δέϕω, λαικάζω, πέρδομαι, πυγίζω, στύω, χέζω (verbs); κύσθος, πέος, πόσθη, πρωκτός, σάθη, τιτθίον, ψωλή (body parts); κόπρος, σκῶρ (excrement). The sample is made up of those given by Henderson 1991, 35 as likely inherently offensive (his ‘primary obscenities’), plus additional vocabulary for other body parts and bodily functions. ⁵² δέϕω shows a shift in interpretation between the first and second editions, from ‘moisten, soften by moisture’ (LS¹) to ‘soften by working with the hand’ (LS²) but the middle or reflexive sense is always marked as obscene. ⁵³ For πυγίζω, these entries represent a shift in the interpretation of the word from a passive to active action. The use of Latin to gloss πυγίζω mirrors its use by Passow; the entry in that lexicon reads ‘auf den Steiss schlagen. 2) paedicari, τινά’; cf. the discussion below. ⁵⁴ Of the sample set of sixteen words, the Revised Supplement rewrites the entries for βινέω and λαικάζω; this reflects a strand of scholarly interest in sex and sexuality in the ancient world over the last few decades, as opposed to other bodily functions which are perhaps viewed as non culturally constructed and thus of less import. Bain and Coker 2014, 403 10 provide some additional anecdotal discussion of the tribulations of Greek lexicographers of the last one hundred years.

Obscenity: A Problem for the Lexicographer

73

genitalia, or verbs for intercourse or masturbation. Sexual activities appear to be the locus of greater censure than other bodily functions, for which an English translation is offered, albeit often a metaphorical one (e.g. χέζω ‘ease oneself ’).⁵⁵ These observations however should not give the impression that LSJ is consistent in its application of a veil of Latin to all matters sexual. The following pair of entries stand for a wider phenomenon, in that Latin is used inconsistently across the Lexicon; note that these Greek words are not only given identical meanings, with the Latin in this case a close translation of the English, but are also separated by only a single intervening word:⁵⁶ ψωλή (2029 i) LSJ ‘membrum virile praeputio retracto’ ψωλός (2029 i) LSJ ‘with the prepuce drawn back’ Similarly, in the entry for πόσθη, both Latin and English are found:⁵⁷ πόσθη (1452 ii) LSJ ‘I. membrum virile, II. foreskin’ This means that the use of Latin per se does not indicate that a denotatum or lexical token was considered to be obscene. This reminds us too that constant revision of individual constituent entries in the Lexicon, but not of the work in its entirety, gives the text an uneven texture.⁵⁸ The difficulties of maintaining consistency across such a vast work as LSJ— and the use of Latin as a masking device—are highlighted by the case of the word ‘penis’. This word can be Latin or English, with the difference often indicated only by context: of the seven times ‘penis’ appears in the text of LSJ, two—possibly three—are English, with four Latin.⁵⁹ The two certain English examples refer to precise medical terms (‘duct of the penis’ and a medical

⁵⁵ Compare the Latin phrase offered in definition of γονοπώτης (357 i) ‘qui semen bibit’. ⁵⁶ LSJ removes the phrase ‘properly fem. of ψωλός’ which immediately precedes the Latin gloss in the entry for ψωλή. The entry for ψωλός in LS¹ contains the phrase ‘one circumcised’, with the reference to the ‘prepuce’ appearing from LS² onwards. ⁵⁷ The entry for πόσθη is identical in earlier editions, except for an additional ‘the’, hitherto immediately anteposing ‘foreskin’, but removed in the revision of the eighth edition: this suggests a desire to conserve space through the removal of superfluous words, as elsewhere. ⁵⁸ Though note that the two terms ‘membrum virile’ and ‘foreskin’ coexist too in the entry for πόσθη in the very first edition. It is possible that ‘foreskin’ was sanctioned because of its frequency in the King James Bible (my thanks are due to an anonymous reviewer for this observation). Compare here Lee’s comments on the state of LSJ, quoted above in n. 14.To the above examples can also be added ἐναποπατέω (LSJ 555 ii), glossed only with ‘ventrem exonerare in . . . ’, contrasted with the shorter compound of the same word ἀποπατέω (212 i) translated as English ‘retire to ease oneself, . . . II pass with [sic] the excrement, void’; a similar range of different translations can be seen in the compounds of χέζω (ἐκχέζω ‘ecacare’, ἐπιχέζω ‘ease oneself upon’, καταχέζω ‘befoul’), and elsewhere. ⁵⁹ Compare the use of the word as both Latin and English in the Revised Supplement where, in reference to the same poem (AP 12.3, Strato), we find λαλοῦ and κωκώ both given as names for the penis, but once English ‘penis’ (κωκώ) and once Latin ‘penis’ (λαλοῦ).

74

Amy Coker

disorder),⁶⁰ whereas the Latin uses seem to be applied to a word or use of a word viewed as inappropriate by the editor(s), for example an erection in Lysistrata (τέτανος) or a dildo (ὄλισβος, ‘penis coriaceus’ = ‘leather penis’).⁶¹ Note that the dysphemistic Latin terms for the penis are not employed here as pragmatically equivalent terms: verpus appears only once, in the entry for δρῖλος (449 ii), and mentula, the other word one might think of in Latin, and particularly familiar from the poet Catullus, famed for his obscenity, not at all.⁶² Despite the lexicographical prudery here, marked by a retreat from English, it is possible that Latin was sometimes intended to contain an element of assistance to the target reader: we could imagine a situation where a reader was unfamiliar with what a word like ‘dildo’⁶³ meant, or the finer details of the genitalia, but for whom a succinct Latin description like penis coriaceus, or to use a different example labia majora pudendorum for μυρτόχειλα (1155 ii) would speak volumes. For the editors, a few choice Latin words may also fulfil the important criterion of economy of space, especially important in the everexpanding text of successive editions of the Lexicon. Final proof, however, of the use of Latin as a deliberate way of masking obscenity is the nature of the Latin employed. While earlier lexica of Greek used Latin as standard, Liddell and Scott chose not to re-use these pre-existing Latin glosses, but rather to bowdlerize actively the German of Passow’s Handwörterbuch.⁶⁴ In contrast to these English lexicographers, Passow was truer to his aims of producing a vernacular lexicon: for the sample set of sixteen Greek words, Latin is found in Passow only in the entries for δέϕω and πυγίζω.⁶⁵ Surprisingly explicit German translates the others, of which the following are a sample; the Latin of earlier Greek-Latin lexica is given for comparison: κύσθος Hedericus (1810), Schrevelius (1730) podex; pudendum muliebre LS ‘any hollow: the pudenda muliebria’ < Passow ‘jede Höhlung, bes. a) der After. b) die weibl. Schaam (‘any hollow, esp. a) the anus, b) the female private part(s?)’)

⁶⁰ καυλός I 3 (932 i) ‘duct of the penis’ (with membrum virile also under sense III) and περιϕίμωσις (1392 ii) ‘disorder of the penis in which the prepuce cannot be drawn forward’; ‘penis’ at νεῦρον (1171 i, sense V) is probably English, but it is difficult to say. In LS¹ there is no reference to the penis under νεῦρον, and περιϕίμωσις is interpreted as not relating to the penis, and defined instead as ‘an unnatural obstruction of the bowels’; for καυλός, sense III simply states ‘= πόσθη’. ⁶¹ βάλανος (304 i) (‘glans penis’), ἐντονία (576 ii) (‘distentio penis’), ὄλισβος (1216 i) (‘penis coriaceus’), and τέτανος (1779 i) (‘erectio penis’). ⁶² See Adams (1982, 9 14) for mentula and verpa/verpus. Latin cunnus, which might have had utility as an equivalent for κύσθος, does not appear in LSJ either. ⁶³ The earliest example of this word is found in 1598, according to the OED. ⁶⁴ The following statements are based upon the fourth edition of Passow’s lexicon (1831), the earliest to which I could gain ready access, via a digital copy. ⁶⁵ For δέϕω, after a number of definitions, Passow adds ‘Med. masturbari’; Latin paedicari is given as a second meaning for πυγίζω (see n. 53).

Obscenity: A Problem for the Lexicographer

75

πέος, πόσθη, σάθη LS membrum virile < Passow ‘das männliche Glied’ (‘the male member’) χέζω LS ‘to ease oneself, do one’s need’ < Passow ‘scheissen, kacken, seine Nothdurft verachten’ (‘to shit, crap, to do one’s necessary’) ψωλή Hedericus (1810), Schrevelius (1730) glans penis denudata; mentula LS membrum virile praeputio retracto < Passow ‘die aufgerichtete und entblösste männliche Ruthe’ (‘the erect and denuded male penis’) For χέζω, German words judged inappropriate for rendering into English are replaced with a euphemistic phrase which is given as their semantic, if not pragmatic, equivalent. Similarly in many—if not all—places where pudenda muliebria is found in LSJ, Passow has ‘die weibliche Schaam’;⁶⁶ the same can be said for membrum virile and ‘das männliche Glied’. If LSJ can be characterized (at least in its early editions) as an English translation of Passow, these are points at which English is actively avoided in preference for Latin glosses, but glosses which are produced by Liddell and Scott, as opposed to being a continuation of the existing Greek-Latin tradition. The use of the then-conventional phrases pudenda muliebria and membrum virile as designations of female and male genitalia respectively is however widespread in the Lexicon, and beyond.⁶⁷ However, as with the use of sens. obsc., the distribution of these phrases throughout the Lexicon is not restricted to words which solely indicate those body parts, but is added to the entries for a large number of others as follows (the brackets indicate each word’s nonmetaphorical meaning): membrum virile⁶⁸ ἀνδρεία (manliness), ἐρέβινθος (chick-pea), καυλός (stem), κέρκος (tail), κωλῆ (ham), μύκης (mushroom), μύρρινον (s.v. μύρσινος) (myrtle),⁶⁹ ὅπλον ⁶⁶ This is certainly the case for κτείς, κῆπος, and κέλης, for example. ⁶⁷ For example, the singular pudendum muliebre appears in the lexicographical tradition of English, e.g. in Nathan Bailey’s Universal Etymological Dictionary of the early eighteenth century; it is also used by the Kryptadia, e.g. Vol. 3 (1886) p. 381, and see also the example in n. 29.Lewis and Short’s Latin English dictionary uses both the singular and plural forms. See Adams 1982, 93 for uses of muliebria and related terms in classical Latin. ⁶⁸ The phrase is used in addition in defining πέος, πόσθη, σάθη, and ψωλή from the sample list of Greek words; it is also used as the sole definition for ληκώ, πριαπισκωτός, σάννιον, and ϕαλλός, and the primary one for στῆμα (‘the exterior part of the membrum virile’), where three other meanings are also isolated (stamen, a nautical term, and a shaft or bearing). Here the first, medical use is privileged, being in Rufus and Pollux, the second and third in Hesychius, and the fourth in Heron of Alexandria. ⁶⁹ See Butterfield, this volume, p. 56, n. 31 for a note on the changing interpretation of the meaning of this word.

76

Amy Coker (weapon), πάσσαλος (peg), ποδεών (ragged ends of a skin where the feet were), ῥόμβος (bullroarer), ῥόπαλον (club), σαύνιον (javelin), σαύρα (lizard), σχοινίον (small rope), σωλήν (channel), τερμέρειον (misfortune one brings on oneself ), τύλος (callus), ϕλέψ (blood vessel) pudenda muliebria⁷⁰ γλωσσόκομος (sarcophagus), δελϕάκιον (sucking-pig), Εὐρώτας (the Eurotas), ἵππος (horse),⁷¹ κέλης (riding-horse), κῆπος (garden), κνέωρον (spurge-flax), κόκκος (seed, berry), κτείς (comb), λειμών (meadow), Μηριόνης (Meriones), μύρτον (myrtle-berry), πεδίον (plain), πλατίστακος (the mullos fish), πτερύγωμα (wings), ῥόδον (rose), σαβύττης (a type of haircut), σέλινον (celery), σῦκον (fig), ταῦρος (bull), χελιδών (swallow), χοῖρος (young pig) pudendum muliebre⁷² τιτίς (small bird), τόπος (place)

Although here again the Lexicon does not use English to express references to genitalia, it does specifically mark metaphorical references which otherwise could easily have been skipped over.⁷³ This is the same principle which is at work when the Lexicon provides a Greek gloss such as ‘αἰδοῖα’ or ‘αἰδοῖον’,⁷⁴ or provides an internal cross-reference. Sense three of LSJ’s entry for ταῦρος ‘bull’, given below, shows the range of possibilities; here found are: (i) a cross-reference (= κοχώνη—in turn given s.v. in the singular as meaning ‘perineum’), (ii) a Latin gloss (pudenda muliebria) and (iii) a Greek gloss (αἰδοῖον): ταῦρος (1761) ‘III. = κοχώνη, Poll.2.173, Gal.14.706: also the pudenda muliebria, Phot.: the male αἰδοῖον, Suid.’

⁷⁰ Used in defining κύσθος from the sample list, and in addition is the sole definition of ἱπποκλείδης, σαβαρίχις, σάκανδρος, σάραβος, and ὕσσακος. It is also found in the entry for παραϕάσσω, glossing παραϕάσιες ‘interior of the pudenda muliebria’, and in μυλλός B. ‘cake in the shape of pudenda muliebria’. ⁷¹ Glossed with pudenda muliebria et virilia I can find no other examples of pudenda virilia in LSJ. ⁷² This phrase is also the sole gloss for both δορίαλλος and ἴκταρ. ⁷³ Many of these metaphorical expressions come from a semantically similar range to those found in Fryer’s catalogue (1963, 45 54), or still ‘make sense’ in contemporary UK English. There is a warning here that we have found in the ancient texts allusions to genitalia because they are similar to our own metaphors, and not necessarily because they represent ancient sensibil ities; see Sommerstein’s comments, for example, on the likely metaphor for the penis contained in the phrase ‘foot of time’ at Frogs 100 (2012, 117 18), despite the unfamilarity within modern English of a metaphor linking feet and the genitals. ⁷⁴ I counted some seven example of the plural used as a gloss, and forty of the singular. Many of these examples in LSJ essentially reproduce entries from Hesychius, in whose lexicon αἰδοῖον in particular is used as a label for words, many of which are hapax legomena, which refer to genitalia, e.g. LSJ (966) κόθημα, τό, = αἰδοῖον, Id. [= Hsch.], reproducing Hesychius κ 3213 (Latte) κόθημα· ἐπὶ τοῦ αἰδοίου. καὶ κότιλον ὁμοίως.

Obscenity: A Problem for the Lexicographer

77

In many cases such as this the offending subject could have been avoided or expurgated, but it is not. This is evidence of the desire on the part of the editors of the Lexicon to be true to their aims of producing a work which was complete in all aspects (all words, and all senses of those words)—but which also attempts to maintain some degree of decency.

4.6. GENITALIA AND MEDICAL TERMINOLOGY Throughout this chapter, again and again it is anxieties surrounding the female and male reproductive organs which have been to the fore, in line with nineteenth-century worries about promiscuity and the control of female sexuality. This is perhaps why we find that even in some nineteenth-century slang dictionaries, the ‘c-word’ is too offensive to be printed in full, although of course (as in the complex circumlocutions given by the contributors to the OED), all fluent speakers know to which word reference is being made.⁷⁵ Consider the following entry from A Dictionary of Buckish Slang, University Wit and Pickpocket Eloquence (1811): C**T. The κοννος [sic] of the Greek, and the cunnus of the Latin dictionaries; a nasty name for a nasty thing: un con Mierge.

Here too, curiously, there is a retreat to Latin (cunnus), and also to Greek. However, κοννος (with whichever accent) is not attested anywhere as a Greek word equivalent in meaning, and suggests some kind of schoolish transliteration into Greek from Latin, rather than any thorough-going knowledge.⁷⁶ Greek is perhaps here being used as a cipher for illicit sexual activity. At the other end of the scale from the ‘language of the gutter’, genitalia and other aspects of the human body and its processes are an important component of Greek literature because of the survival—and continued influence—of works by medical authorities such as Hippocrates and Galen. It has already been noted that English terms for genitals in LSJ seem to be associated with technical usages, i.e. καυλός ‘duct of the penis’ and περιϕίμωσις ‘disorder of the penis in which the prepuce cannot be drawn forward’. We know from the preface to the ninth edition that Dr Edmund Theodore Withington, who ⁷⁵ Towards the end of the century in 1891 John Farmer, the author of Slang and its Analogues (1890 1904), took his printers to court and lost for refusing to set the entries for ‘cunt’ and similar words (Mugglestone 2007, 30). Similarly note that while πέος is included in the Pocket Oxford Classical Greek Dictionary (2002), translated as ‘penis’ (p. 252), as is βινέω, given as ‘fuck’ (p. 64), κύσθος is excluded. ‘Cunt’ is generally agreed to be one of the most offensive words in UK English, joined now increasingly by ‘nigger’ and other racial slurs; a short account of the history of ‘cunt’ is found at Hughes 1998, 27. ⁷⁶ A word κόννος is given by LSJ meaning a ‘trinket’ (Suda) or a ‘beard’ (Lucian), but this looks like pure happenchance.

78

Amy Coker

published widely on ancient medicine and produced the Loeb editions of several Hippocratic texts, did substantial work on the medical vocabulary as part of the revision between LS⁸ and LSJ. Many of these non-prudish and precise English terms entered the lexica at this time as a result of the work of Withington and similar experts (see Craik, Chapter 9, this volume). Perhaps this reflects not so much a change in attitude to sexuality and to the body, as a greater familiarity with and understanding of its reproductive parts and processes as compared with the original editors. For example, this revision sees the wholesale addition of entries for ‘clitoris’ words of the νύμϕη group in the ninth edition (νυμϕοτομέω, νυμϕοτομία 1185 i). νύμϕη (1184 ii) was also rewritten to include the overt meaning ‘clitoris’ (sense IX), whereas the only indication in LS⁸ of νύμϕη (usually meaning ‘bride’) as a body part was under sense VII with ‘= μύρτον ΙΙ Galen, etc.’⁷⁷ There is also a drive towards precision between LS⁸ and LSJ with ‘foreskin’ in two places replacing ‘skin’ (ἀντανατρέχω, λιπόδερμος), and the use of ‘penis’ rather than more oblique formulations, for example LSJ ‘glans penis’ for LS⁸ ‘glans membri virilis’ (βάλανος) and LSJ ‘erectio penis Ar.’ for LS⁸ ‘sens. obsc.’ (τέτανος). Hence this movement away from coyness therefore does not just apply to the vocabulary of the anatomy class, though it is perhaps influenced obliquely by the presence of that category of vocabulary.⁷⁸ We still find, however, that not all words for genitals are precisely defined at the level of the individual part: yet this is not just down to prudery and a general desire to hide behind a Latin cover-all term. Many of these Greek words are not, and never were, precise in an ‘anatomical’ way to their original users. Many of the metaphorical words listed above under pudenda muliebria and membrum virile allude to the external general appearance the male or female genitals (‘stem’ or ‘javelin’, or ‘fig’), rather than to a precise part of them or their functions as part of the human reproductive apparatus.⁷⁹ English ‘minge’ for example is in no way an anatomical term, but does refer to a specific area of the female body, but not to the female body as reproductive ⁷⁷ Note, however, that in the first four editions (LS¹ ⁴, 1843 55) we are told that the νύμϕη is ‘part of the pudenda muliebria’, before the phrase is expunged in the fifth edition (1861). ⁷⁸ Curiously, the statement in κλειτορίς (957 i) telling the reader the part is ‘in pudendis muliebribus’, present throughout all earlier editions, is removed in LSJ: this may reflect a change in expected readership (with an increased familiarity with female anatomy and its vocabulary?), or perhaps more likely and prosaically may be an exercise in space saving. ⁷⁹ Some however do, and there is an important distinction to be drawn between technical anatomical terms and generic perhaps one off, creatively humorous metaphors. Of the female terms listed above, Skoda for example observes that κτείς indicates the pubis of both men and women in a number of texts of the Hippocratic corpus (1988, 156 7, §5.118), and πτερύγωμα the labia majora (‘grandes lèvres’) in Galen, Soranus and Rufus (1988, 175, §5.143). LSJ indicates this distinction only by the citations attached to each meaning: the catch all term pudenda muliebria thus indicates words ranging in connotative meaning equivalent to English ‘pussy’ or ‘cunt’ (offensive) or ‘clunge’ (humorous, idiosyncratic) to ‘labia minora’ or ‘vulva’ (technical).

Obscenity: A Problem for the Lexicographer

79

anatomy per se. The following rhyme ‘It’—as aposiopetic as other texts quoted in this chapter—allegedly circulated orally in medical circles around a century ago, and makes exactly this point about the need for precision in medical circles for the ‘delicate and complicated matter’, as opposed to the short word used when ‘common people chatter’ (cited from Fryer 1963, 49):⁸⁰ The portions of a woman that appeal to man’s depravity Are constructed with considerable care, And what at first appears to be a simple little cavity Is in fact a most elaborate affair. Physicians of distinction have examined these phenomena In numerous experimental dames; They have tabulated carefully the feminine abdomina, And given them some fascinating names. There’s the vulva, vagina, and the jolly perineum, And the hymen, in the case of many brides, And lots of other little things you’d like, if you could see ’em, The clitoris, and other things beside. So isn’t it a pity, when we common people chatter, Of these mysteries to which I have referred, That we use for such a delicate and complicated matter Such a very short and ordinary word!

In light of this, LSJ’s use of a broad, non-specific, albeit Latin term has some utility: it marks a Greek word as indicating either the male or female genitalia, but avoids the need to gloss that word with a term which may not be pragmatically appropriate, especially since it may not always be clear what that stylistically-equivalent word should be in English.

4.7. CONCLUDING REMARKS It is all too easy in the present to be critical of a work which, some 170 years since its creation in a cultural and intellectual environment far removed from ⁸⁰ Fryer (l.c.) credits a Mr T.B. Thompson as the source of his version (p. 266, n. 103). The original composer was however the humorist A.P. Herbert, who confessed in his autobiography to writing the first version of these lines on a ship in the Red Sea in 1927 or 1928, finding inspiration from a book he found in the office of the ship’s doctor (Herbert 1970, 63 7). Herbert calls the poem ‘Lines on a book borrowed from the ship’s doctor’, but it is known more popularly as ‘The Portions of a Woman’, or ‘The Doctor’s Lament’; some longer versions also circulated with replies from ‘The Layman’ and ‘The Woman’ (one version was sung in Hong Kong Rugby Club at least as late as 1950, http://members3.boardhost.com/tigersongs/msg/1306600121.html, last accessed 19/2/2016), and Herbert himself (67) recalls being recited a reply ‘from the point of view of the “experimental dames” ’, among the many other versions which dogged him, by the Royal Navy China Station before the Second World War.

80

Amy Coker

our own, remains in use, and yet is at odds with our ‘modern’ sensibilities. Despite the age of some of the material still residing in the Lexicon, and the fact that it avoids English dysphemistic terms, it is remarkably free in its use of English words for matters which we may think are unacceptable to the Victorian stereotype. There are hints that in the 1860s–80s there was a move towards removing some ‘difficult’ material, but by LSJ the substance of the Lexicon is comprised of an increasingly frank and technical exposition of genitalia, building upon what was often already present in editions in the earlier nineteenth century. The revisions have gone some way to rehabilitate this area of the Greek lexicon (to date the most comprehensive single work to handle this material is Henderson’s ground-breaking The Maculate Muse, 1991 [1st edn 1975]), but there is more to be done in this neglected subset of the Greek anatomical and dysphemistic vocabulary. Liddell and Scott’s enterprise can be viewed as democratizing Greek to some extent through the use of English, but some of these sexual matters were still too risqué to stand completely unveiled. We must also remember though that the work in its full form is an elite product. Alongside there also existed the Abridged (1843¹, 1871²) and Intermediate (1889 [= IGL]) versions, written for use in schools, and which include most of the original list of sixteen Greek words, with the exception of πυγίζω, σάθη, στύω, and ψωλή (and κύσθος for the Intermediate), all rare words which the pupil—the intended audience— was unlikely to encounter. Of the rest, all give very much the same information as LSJ: exceptions are κύσθος in the Abridged version, for which the entry simply reads ‘any hollow’ (ironically the words which would be deleted by the revision to LS⁷ some forty years later in 1882), and δέϕω, which has been sanitized by the removal of the ‘sens. obsc.’, so liberally applied throughout the unabridged text.⁸¹ This perhaps indicates that female privy parts and anal sex were to be hidden from the schoolboy and masturbation discouraged, but again, excrement and breaking wind are somehow less taboo, as is some limited discussion of sex.⁸² A glance at those words listed above which are made into metaphorical expressions for the genitals by the addition of pudenda muliebre and similar terms (pp. 71–7) reveals that none are so designated in the Abridged and Intermediate, even though these words are included in

⁸¹ The first edition of the Abridged gives the meaning of δέϕω as ‘to moisten’, subsequently revised in the second to ‘soften’ in line with changes in the full text of the Lexicon: see p. 72, n. 52. ⁸² In this light, note how JACT’s Reading Greek textbook (2007, second edition), in offering a version of the lines of Aristophanes quoted at the beginning of this chapter replaces Lampito’s more literal ἄνευ ψωλᾶς with the euphemistic ἄνευ τῶν ἀϕροδισίων (§10A, line 31, glossed as ‘without sex’). These words are provided specifically for translation, and the bowdlerization highlights a current anxiety about talking about sex in the vernacular in the (potentially under 18) classroom which continues to echo the concerns of our predecessors.

Obscenity: A Problem for the Lexicographer

81

those versions.⁸³ This is true even for those words which are key to understanding scenes in canonical texts, such as the episode with the Megarian and his daughters/pigs (χοῖρος) in Aristophanes’ Acharnians. There is some indication that in IGL ‘Ar., etc.’ may be being used in the same way as membrum virile, e.g. in καυλός (‘II. the stalk of a plant, Ar., etc.’), although this is not certain. It seems, then, that sexual metaphor is suitable for the educated scholar, but not for the young; economy of space may perhaps again however be a relevant criterion, particularly in a work which is short by design.⁸⁴ We find thus throughout the history of LSJ the widespread application of devices marking obscenity to otherwise innocent words, even though such metaphorical uses would be the easiest aspect to expurgate. Even though these strategies aim to shield the casual reader, for its Latinate user they function on a practical level as excellent covert signposts that something unsavoury is present both in the text of the Lexicon and in the ancient work which prompted the reader to reach for it—even if not all these matters would now be subject to such prudery. While sens. obsc., pudenda muliebria, and other devices may mask meaning for some readers without Latin, this very act of evasion—like the asterisks or dashes of a censored word—indicates something which requires covering up, and that act of evasion speaks volumes. This elegant, albeit accidental, system has maintained the longevity of the Lexicon, without the need for contemporary dysphemistic terms, even where such terms might be considered appropriate on the basis of the context or distribution of the word. The reader is thus left free and unfettered to supply an appropriately obscene term from his—and now also, her—own vernacular.

⁸³ Of those checked (male, ἀνδρεία, ἐρέβινθος, καυλός, κέρκος, ὅπλον, ῥόμβος, σαύρα; female δελϕάκιον, ἵππος, κέλης, κῆπος, κτείς, λειμών, μύρτον, σῦκον, χοῖρος) only δέλϕακιον was ex cluded, and then only from the Intermediate. ⁸⁴ This can be seen in the revision and enlargement of the Abridged version, where it looks as if βινέω ‘inire, coire, of illicit intercourse’ (Abridged¹) is pared down to ‘coire, of illicit intercourse’ (Abridged²) so that the entire entry can be made to fit on a single line.

5 Etymology and Etymologies in the Lexicon Joshua T. Katz

More often than not, when colleagues have a question about a word in Ancient Greek, their first stop is a large reference work that sits on the desk, or is within easy digital reach, of every one of them: LSJ, more fully known as A GreekEnglish Lexicon, whose first edition Henry George Liddell (1811–98) and Robert Scott (1811–87), having plundered the work of the German lexicographers Johann Gottlob Schneider and Franz Passow, published with the Clarendon Press in 1843.¹ This seems to be true even when what these colleagues wish to discover is an etymology—though I rely here mostly on anecdote, including from members of the audience at the colloquium at Corpus Christi College, Oxford, where a version of this chapter was delivered in November 2013. To those of us who have a professional interest in the prehistory of Greek lexica, such a move is surprising, and not merely because quite a number of other large reference works are readily available, ones devoted specifically to what the French call histoire des mots. The fact is that prior to writing these words, I would have spoken not about attention but about apparent inattention to etymology in LSJ, for I had a hard time visualizing any etymological comment in the dictionary, despite dipping into it myself almost every day. There is, to be sure, a good reason for my difficulty. Sir Henry Stuart Jones (1867–1939), the Camden Professor of Ancient History at Oxford whose last name² provides the final letter of the familiar abbreviation, has this to say in

¹ For comments and encouragement I am grateful to Michael Clarke and Chris Stray, as well as to Jonathan Katz, John Ma, Philomen Probert, and the late Martin West (1937 2015). ² According to the Oxford Dictionary of National Biography, where he is listed as ‘Jones, Sir Henry Stuart ’, ‘Stuart was his second forename, but after his marriage he and his wife generally prefixed it to their surname: when he was knighted in 1933 he legally assumed the name Stuart Jones’ (Blakiston 2004, 521). Joshua T. Katz, Etymology and Etymologies in the Lexicon In: Liddell and Scott: The History, Methodology, and Languages of the World’s Leading Lexicon of Ancient Greek. Edited by: Christopher Stray, Michael Clarke, and Joshua T. Katz, Oxford University Press (2019). © Oxford University Press. DOI: 10.1093/oso/9780198810803.003.0005

Etymology and Etymologies in the Lexicon

83

his ‘Preface [of ] 1925’ about the role of etymology in the revision of the Lexicon that he was supervising, a revision that would result, a decade and a half later, in the year after his death, in the completed ninth edition that we use still today: When I became Camden Professor at the beginning of 1920 it became necessary to provide me with assistance in my editorial work, and Mr. R. McKenzie of Trinity College (now Fereday Fellow of St. John’s College) was appointed Assistant Editor by the Delegates of the Press. Apart from his arduous labour in putting my drafts into final shape and in arranging and working in a large mass of accumulated material, Mr. McKenzie has been able to render inestimable service to the Lexicon on the philological side. After careful consideration it was decided that etymological information should be reduced to a minimum. A glance at Boisacq’s Dictionnaire étymologique de la langue grecque will show that the speculations of etymologists are rarely free from conjecture; and the progress of comparative philology since the days of G. Curtius (whose Griechische Etymologie was the main source drawn upon by Liddell and Scott) has brought about the clearance of much rubbish but little solid construction. Some assured results, however, have been attained, and the etymologies presented in the text have in almost every case been approved by Mr. McKenzie.³

The decision to clear etymological trash without at the same time engaging in much sensible ‘construction’ (or, as it might be better put, reconstruction) is certainly worthy of discussion, but it is presumably not controversial that the reference of record for the Ancient Greek word-hoard⁴ should strive to report ‘assured results’ and that any questionable material should be clearly marked as such. Because etymological proposals are indeed ‘rarely free from conjecture’ and the arguments for non-obvious etymologies are often too complicated to distill into a few words,⁵ it is easy to see why compilers of a standard ³ Jones 1940, x. The references are to Georg Curtius’s Grundzüge der griechischen Etymologie and Émile Boisacq’s Greek etymological dictionary. The former was first published in two volumes between 1858 and 1862, with the fifth edition, revised by Ernst Windisch, appearing in 1879 (Curtius 1879); an English translation, Principles of Greek Etymology, appeared in 1875 and a revision based on the fifth German edition in 1886 (Curtius 1886). The latter was first published in 1916 and received its fourth and final edition in 1950 (Boisacq 1950). ⁴ The Madrid based Diccionario griego español (Adrados 1980 ) has not yet made it through epsilon; it will be interesting to see to what extent colleagues in the years ahead make active use of The Brill Dictionary of Ancient Greek (Montanari 2015; based heavily on an Italian work first published in 1995 and most recently revised in 2013), in addition to or instead of LSJ, on which, it must be stressed, it is consciously dependent (see Lee 2017, 417 18, 434, and passim). On the forthcoming Cambridge Greek Lexicon (Diggle et al. 2019), which is less ambitious (it is ‘aimed primarily at students, but taking account of the most recent textual and philological scholarship’: http://www.classics.cam.ac.uk/Research/projects/glp), see Hire 2005. ⁵ In his recent review of Montanari 2015, Filos (2018) writes, ‘A reference dictionary is not supposed to deal with subtle etymological explanations, which are often complicated and subject to regular revision (note e.g. the problems with the many outdated etymologies in LSJ, as also acknowledged in the Preface to the 1996 Supplement).’ Adopting much the same point of view that I do in this paper about how etymologies should and should not be presented in

84

Joshua T. Katz

dictionary might avoid adding a great deal of information of this kind.⁶ The mass removal of ill-advised etymological material in the preparation of LSJ makes sense, too, but is undeniably a striking move: for obvious reasons, dictionaries tend to expand rather than contract, not always for the better,⁷ and the title page of the work under discussion explicitly claims that this is ‘A New Edition Revised and Augmented throughout by . . . Jones’ (italics added).⁸ In any case, it would appear that the ‘seeming inattention’ to etymology in the Lexicon on which I remarked above is a mirage: however exactly history judges McKenzie’s ‘inestimable service’, there was in fact quite a lot of behind-thescenes attention.⁹ For this reason alone it seems right to pay attention to the etymological material that remains and, to the extent that this is possible without engaging in a full survey, to try to figure out why it is there. Enter Roderick McKenzie (1887–1937), the missing M of LSJ,¹⁰ a man who worked with and complemented Jones and of whom it is written in the anonymous prefatory ‘Postscript [of] 1940’ that he was responsible for the ‘ultimate form’ of most of the lemmata and that what ‘gave his contribution inestimable value’ (note the use of the word ‘inestimable’ again) were ‘his great

standard dictionaries, he goes on to describe ‘the non uniform approach adopted by the editors’ as ‘rather prudent’. ⁶ Jones’s jaundiced view of ‘the progress of comparative philology’ remains a familiar opinion among scholars today. It does not come as a surprise that the dictionary by Boisacq, which Jones nearly a century ago cited as the latest, but obviously problematic, source (see above, with fn. 3), is these days rarely referred to; what some will find startling, however, is that no single work has successfully superseded it. Four new etymological dictionaries of Ancient Greek have appeared since LSJ was completed, and anyone today who is making a serious argument is certain to consult two of them: Frisk 1960 72 (especially good for older bibliography) and Chantraine 1968 80 (the overall best bet, though not to be used uncritically), the latter of which needs to be looked at together with the ‘Chroniques d’étymologie grecque’, fourteen of which have appeared so far in issues of Revue de Philologie (latest: no. 14, in RPh 87 (2013, publ. 2016) 157 202), with the first ten incorporated into the newly typeset 2009 edition of Chantraine. (The other two are Van Windekens 1986, which has understandably sunk largely without trace, and Beekes 2010, which needs to be used with far greater caution than non experts generally realize.) For my views on the practice of etymology today and the role of etymological dictionaries, see Katz 2010c, as well as Katz 2016. ⁷ For example, Chadwick (1996, 8) has this to say about the sorts of things for which most people consult LSJ or any other standard dictionary: ‘[T]he increase in both the number of entries and the diversity of senses over the immense period regarded as Ancient Greek render the new lexicon [i.e. LSJ] less rather than more serviceable for the average user’. ⁸ A proper accounting of the ways in which the etymological material in LSJ differs from what appears in earlier editions (i.e. LS1 8) would be instructive but that would be another paper. Some brief remarks appear in Imholtz 2007, 126 and 128 9. ⁹ John Chadwick has disputed the level of attention as well as its success; see the quotation below on p. 86. ¹⁰ ‘The abbreviation LSJ is now hallowed by usage . . . , though I have often wondered whether LSM might not be fairer’ (Chadwick 1996, 7). McKenzie, who is given modest credit on the title page of LSJ in the form of the words ‘with the assistance of ’ in small print, finally becomes the fourth proper author (as though ‘LSJM’) at the head of the title page of the 1996 Revised Supplement (= LSJ Revd Suppl.).

Etymology and Etymologies in the Lexicon

85

knowledge of comparative philology, his laborious accuracy, and his tireless patience’.¹¹ McKenzie was not a prolific scholar in his own right, and information is not easy to come by, though his obituary in The Scotsman offers the sad report of ‘[t]he death of a brilliant Edinburgh man who was engaged at Oxford on the revision of the Liddell and Scott Greek Lexicon . . . . On Thursday [i.e. 24 June 1937] he was found in a bath at his lodgings in Museum Road with throat and wrist injuries. His death took place some time later at the Infirmary at Oxford’.¹² Aside from his 1910 Gaisford Prize-winning translation of part of the second book of the Aeneid into Homeric hexameters¹³ and a few dozen reviews and short notices in Classical Review,¹⁴ he appears to have written eleven articles, all published between 1918 and 1936 in Classical Quarterly or Classical Review and all short, sometimes very, very short (four are under a page in length, one of them a single sentence that compares a phrase in the Iliad to something in Sir Walter Scott¹⁵).¹⁶ These papers, all of them linguisticky in nature, nearly all of them in the first place about Greek rather than Latin, and some of them considering and then reconsidering the same forms (ἔρχομαι is a favorite), show that McKenzie was particularly interested in etymology; in fact, three of them are titled simply ‘Etymologies’. They also show that McKenzie had what was for someone in his position an unusual fascination with the Baltic languages Lithuanian and Latvian (or, as he calls it, Lettish) and, to a slightly lesser extent, Slavic.¹⁷ Perhaps it is not surprising that, as far as I have been able to ascertain, no etymological ideas of McKenzie’s own made it into LSJ (compare James, Chapter 10, this volume, p. 168, n. 118). In view of the extent of the criticism that scholars have levelled against both the editors and the publisher of LSJ, as well as of the two supplements, particularly the first one of 1968 (= LSJ Suppl.), another thing that is not surprising is that some of the objections have had to do with the treatment ¹¹ Anon. 1940, xiii. The fuller quotation explains the relationship between Jones and McKenzie: ‘The work done by these two men could not be overrated. Sir Henry was the ideal Editor . . . [, who] gave the work at once its consistency and its elasticity. McKenzie . . . provided a fine complement.’ ¹² Anon. 1937. ¹³ Virgil, Aeneid II 268 385 Translated into Homeric Hexameters (Oxford: Blackwell, 1910). ¹⁴ Beginning with CR 35 (1921) 180 1 and going up to 51 (1937) 38 and 147. From 1928 until his death he also contributed the section ‘Lexicography and grammar’ to the annual papyro logical ‘Bibliography: Graeco Roman Egypt’ in the Journal of Egyptian Archaeology. ¹⁵ ‘The phrase [of the title, οὖλον κεκλήγοντεϛ] may be rendered “uttering thick screams,” these English words bearing the same sense as in Scott, Heart of Midlothian, Chapter XXV.: “She proceeded to raise the family by her screams of horror, uttered as thick as if the Brownie had been flaying her” ’ (McKenzie 1928). ¹⁶ ‘An unnoticed suppletive verb’, CQ 12 (1918) 57 8 (on ἔρχομαι, etc.); ‘The Greek adjectives ending in ηϛ’, CQ 13 (1919) 141 8; ‘Graeca’, CQ 15 (1921) 44 8 (in part on ἔρχομαι); ‘Graeca (continued)’, CQ 15 (1921) 186 8; ‘Etymologies’, CQ 18 (1924) 23 (in part on ἔρχομαι); ‘Etymologies’, CQ 19 (1925) 108 10; ‘Etymologies’, CQ 19 (1925) 208 10; ‘Οὖλον κεκλήγοντεϛ’, CQ 22 (1928) 206 (see above, with fn. 15); ‘Latin lugeo’, CQ 24 (1930) 54; ‘A note on Julius Africanus’, CR 47 (1933) 9; and ‘A rare imperative form’, CR 50 (1936) 60. ¹⁷ See especially the papers from 1924, 1925 (the first), and 1930 cited in fn. 16 above.

86

Joshua T. Katz

of etymology. Towards the end of his life, John Chadwick—best known as a Mycenologist, he toiled for six years after the War on what would become the Oxford Latin Dictionary and was later a member of the committee for the Revised Supplement, which incorporates Mycenaean forms—railed against the failures of LSJ, etymological and otherwise,¹⁸ thoroughly damning the linguistic and lexicographical competence and even attention (see above) of not just the polymathic historian Jones but also McKenzie: As far as I know, neither of the editors [sc. Jones and McKenzie] had any claim to competence in lexicography, and the quality of their work must sometimes call in question their ability to analyse correctly passages of the ancient authors. One obvious deficiency is shown by the frequently incorrect and misleading etymological notes. Any trained Greek linguist could have eliminated the worst of their errors. But here, as elsewhere, they were content to take over the text of the 8th edition, which they treated as a draft to be improved by insertion, deletion and minor alterations. As far as I can discover, they hardly ever scrapped an existing entry and rewrote it ab initio. Most of the criticisms that must be levelled at them arise from that simple decision, which may have been that of the publisher, to emend and patch, rather than attempt a recasting of the material.¹⁹

It is perhaps interesting to speculate how the history of etymologizing in the dictionary might have been different—I believe it is fair to say that it would not necessarily have been better—had the idiosyncratic Max Müller been involved in nineteenth-century editions, a scenario about which P.G.W. Glare has this to say: Liddell & Scott also intermittently included etymological notes. Max Müller, who in other respects was full of praise for the Lexicon, criticized these as being unscientific. Liddell admitted his lack of competence in this field and asked Müller to revise them, but afterwards, with some embarrassment, cancelled the arrangement. Müller learned much later that this was at the instance of Scott, who for some reason had taken a dislike to him.²⁰

The non-participation of Müller notwithstanding, I am sympathetic to Glare’s view that the ‘occasional etymological notes [that] remain in LSJ . . . are usually so brief and haphazard in their choice of cognates that they are of little use to the reader’ and that ‘[i]t would probably be better as a general rule to abandon them, and refer the user to one of the standard etymological dictionaries’²¹—though the problem with this last idea is that the would-be standard etymological ¹⁸ Though Chadwick could be cranky, his ‘Introduction’ to Lexicographica Graeca (Chadwick 1996, 1 30) is essential reading, as is his article ‘The case for replacing Liddell and Scott’ (Chadwick 1994); see also Chadwick 1992b. ¹⁹ Chadwick 1994, 1 2. ²⁰ Glare 1987, 11; for Scott’s motivation, see Stray, Chapter 1, this volume, p. 13, n. 26. ²¹ Glare 1987, 11. The authors of the ‘Preface’ to LSJ Revd Suppl. write that ‘[t]he occasional etymological notes in LSJ are frequently out of date. No attempt has been made to bring them up to date and readers are referred to the standard etymological dictionaries of Chantraine . . . and Frisk’; see fnn. 5 and 6 above.

Etymology and Etymologies in the Lexicon

87

dictionaries do not really provide a benchmark. Or maybe the problem lies rather with members of my tribe, who do not like to admit that there is some truth to the infamous waggish remark about etymology, popularized by none other than Müller, that it is a science in which consonants count for little and vowels for nothing at all?²² Let us return to Chadwick. In Lexicographica Graeca, he chastises the editors of LSJ for ‘fail[ing] to keep the etymological notes up to date’, writing that ‘[b]y 1925 a great deal was known on that subject which could not have appeared at least in the earlier editions of Liddell and Scott’.²³ True, though by the standards of 1996 or 2019, certain kinds of etymological information in the dictionary could not possibly have been successful—the work could hardly have contained anything useful about Hittite and Tocharian or about laryngeals, all of which were obscure in 1940, never mind in 1925—and one might anyway suppose that Chadwick would approve of the removal of obviously incorrect information even if he would also wish for greater highlighting of positive results. However, he demonstrates a surprising misconception of McKenzie—surprising, for one thing, because McKenzie died less than a decade before Chadwick himself arrived in Oxford—when he writes that ‘[o]ne can only suppose that Stuart Jones and McKenzie were not interested in this aspect of their work; but since it often has implications for the development of senses, it was obviously important not to keep quoting discarded theories’.²⁴ Etymology, we have seen, is precisely what McKenzie was most interested in—and something for which his immediate boss and the Clarendon Press as a whole went out of their way to praise him. This brings me to the question of the nature of the etymological material that is actually found in LSJ. Etymologies are placed into round brackets (parentheses),²⁵ which usually come at the end of a given lemma, though sometimes, for longer lemmata, they are tucked away at the end of a significant section. The contrast with the Oxford English Dictionary is instructive: in that dictionary of record, the etymological section is up front, as indeed it often is in earlier editions of the Greek dictionary as well. Of course, the OED is famously based on ‘historical principles’, whereas LSJ is not—a fact about which Chadwick liked to complain.²⁶

²² For what is known about the history of the phrase, generally but incorrectly attributed to Voltaire, see Considine 2009. ²³ Chadwick 1996, 8. ²⁴ Chadwick 1996, 8. The meaning of the end of this quotation (from ‘but since’) and its connection to what comes before the semicolon is opaque. Could it be a pragmatically odd indication that Chadwick does actually recognize that McKenzie (and Jones) removed etymolo gies from the 8th edition? ²⁵ ‘( ) Between these brackets stand the Etymological remarks’ (LSJ, p. xlviii). ²⁶ In addition to the quotation cited towards the end of the previous paragraph, see especially Chadwick 1992b.

88

Joshua T. Katz

Two questions occupy me in what follows: (1) What lemmata have etymological brackets? and (2) When these brackets exist, what sort of information lies between them? One way to arrive at answers would be to read through the dictionary. I have not done this. While the main reason comes from a combination of lack of time and lack of patience, it is, I think, important not to lose sight of the fact that LSJ is meant not to be read but to be used—used and used and used some more. It is a reference work, not a monograph, and as such it is supposed to advance understanding through intelligent arrangement and codification; it is not in the first place supposed to ‘produce knowledge’, if I may use that voguish turn of phrase. In Chadwick’s words, ‘It is all too often forgotten that a dictionary is a tool, and needs therefore to be designed with the user in mind; what is suitable for a student will not serve an advanced scholar and vice versa’.²⁷ The material in the pages ahead comes from using the dictionary in a more or less normal fashion, that is to say, as an ordinary, albeit obsessive, Hellenist might—someone without a background in comparative philology, whether undergraduate or senior figure in the field. Suppose one picks up the Iliad and starts reading it with the goal of knowing all there is to know about each and every Homeric word. Here as my opening test case are the opening two verses of the poem: Μῆνιν ἄειδε, θεά, Πηληϊάδεω Ἀχιλῆοϛ οὐλομένην, ἣ μυρί ’ Ἀχαιοῖϛ ἄλγε’ ἔθηκε.

All eleven words, including the names, are to be found in LSJ. While different people understandably have different views of what constitutes ‘interest’, I am confident that the word that nearly everyone would say is the least interesting is ἥ, with ἔθηκε taking second place. And yet these are the only two words for which LSJ supplies etymological information that moves outside Greek. Ignoring for the moment Πηληϊάδηϛ, we are left with eight words on whose pre-Greek background LSJ is silent. It is sobering that for six or seven of these, what is today widely believed to be the ‘right’ story, or at any rate an important part of the right story, could not have been known to McKenzie. There is, to be sure, always the chance that our stories today are not right either; furthermore, what does and does not count as right is sometimes a matter of perspective.²⁸ Still, it will be immediately obvious that quite a number of these words have putative cognates of more than passing significance in Hittite and the other Anatolian languages (none of which McKenzie could have known about beyond the barest trace) and/or are reconstructed for Proto-Indo-European with one or more laryngeals (ditto). Here are the eight words (I provide the headwords as they appear in the dictionary), together with my brief (bibliographically minimal) assessments: ²⁷ Chadwick 1996, 8.

²⁸ See Katz 2010a, 2010c, and 2016, all with references.

Etymology and Etymologies in the Lexicon

89

μῆνιϛ The probable etymology could NOT have been sketched even in rough terms in LSJ. The word is usually connected to Proto Indo European *men ‘think’ (cf. Latin mēns, English mind, etc.) via its derivative *mneh₂ ‘keep in mind’. The most important scholarly advances from the past four decades were made by Calvert Watkins and Leonard Muellner. For references and further discussion, see Katz 2013a, 17 21 and 2019, 165 8. ἀείδω The probable etymology could NOT have been sketched in LSJ, though there is a brief and perhaps not wholly accurate low level parenthetical note about a digamma: ‘(ἀϝείδω, cf. αὐδή, ὑδέω.)’. My view is that the verb goes back to the root *h₂weid , which is not found outside Greek; see Katz 2013a, 16 23 and Katz 2013b, both with references. If this is so, then the connection with αὐδή (cf. Sanskrit vad ‘speak’) is folk etymological, and also with ὑδέω (whose derivational link to αὐδή is anyway disputed). There is, however, a chance (see Katz 2019, 165, n. 51, with references) that the root is instead *h₂wed(H) , in which case ἀείδω and αὐδή are indeed sisters. The point at hand stands either way. θεά and θεόϛ The etymology could NOT have been sketched authoritatively in LSJ. The details of the reconstruction *dhh₁ s ó , which provides the Greek god with exact or nearly exact cognates in the Anatolian languages Luwian, Lycian, Lydian, and Palaic (and near cognates in Armenian and Italic: e.g. Latin fānum, fēriae, and fēstus), have only recently been clarified, above all by Melchert (1997) and Meier Brügger (2006). For the specifically feminine form θεά, see Katz 2018. Ἀχιλλεύϛ The probable etymology could NOT have been sketched in LSJ. The most coherent etymological statement of this difficult name, most likely an old compound, was published only a dozen years ago: Nikolaev 2007. οὐλόμενοϛ and ὄλλυμι The connection with Latin ab oleō has long been known and COULD thus have been noted in LSJ. However, the phonological details of the root could not have been: *h₃elh₁ . (I note in passing that Homer’s οὐλομένην has in my opinion been consistently misanalysed, but the demonstration of this must await another occasion.) μυρίοϛ The probable etymology could NOT have been sketched in LSJ. See Weiss 1996 for a convincing etymological argument: μυρίοϛ is intimately linked to Hittite mūri ‘cluster (of grapes or other fruit)’, with both going back to *meuh₁/3 . Ἀχαιόϛ The etymology could NOT have been sketched in LSJ.²⁹ The word’s precise preform remains a matter of dispute; of enormous interest these days is the nature of the (almost certainly historically real) connection to the Hittite toponym Aḫ ḫ iyawa, a form that was not even a glimmer in anyone’s eye

²⁹ There is a parenthetical folk etymological comment s.v. Ἀχαία, an epithet of Demeter: ‘(Acc. to Hsch. from ἄχοϛ grief for the loss of her daughter . . . )’.

90

Joshua T. Katz

just a few generations back. Fischer (2010) provides an overview of what has been said. ἄλγοϛ Because the prehistory of the word remains uncertain to this day, its etymology could NOT reasonably have been sketched in LSJ. The connection with ἀλέγω (and, then, Latin neg legō, intel legō, etc.) is old, and some scholars today accept it. I avoid taking a stand.

The contrast between these words and the two for which LSJ does provide etymological commentary is stark: ὅϛ, ἥ, ὅ—‘(With Gr. Relat. ὅϛ, ἥ, ὅ cf. Skt. Relat. yas, yā, yad, Lith. jis, ji (he, she), OSlav. i, ja, je (he, she, it).)’ τίθημι—‘(Cf. Lith. dēti “lay (eggs, etc.)”, Skt. dádhāti “lay down, place”, Lat. -do in con-do, etc., Engl. do, doom.)’ Such information seems sharply different from the ‘useful’ knowledge contained elsewhere in the same and similar lemmata: about authors, contexts, and idioms; about periods, dialects, and meters; and of course about sense. The use of ‘cf.’ does not help: a weaselly device—as scholars who think about commentaries have sometimes admitted as well³⁰—for it is a quick way to indicate that one has, or believes one has, a point to make without actually being forced to spell out what that point is. Even aside from this, though, it is hard to imagine that many classicists could, or for that matter should, do anything with the unadorned knowledge of what some third-person pronouns look like in Balto-Slavic; it is not clear why the order of the cognates in the one entry is Sanskrit-Lithuanian-Old Church Slavonic while in the other Lithuanian comes before Sanskrit (and why Old Church Slavonic, which could have been cited, is omitted); and although anyone who gets so far will presumably be intrigued by the link between do and doom in English, the epistemological status of this unexplained juxtaposition in connection with τίθημι—not to mention the foregrounding of a Lithuanian verb that means ‘lay eggs’—can hardly be anything but confusing. All of this is just to say at greater length what Glare wrote more than three decades ago, namely that etymological notes in LSJ tend to be ‘brief and haphazard in their choice of cognates’. If there is one thing here that is not haphazard, it is that both the note for the relative pronoun and the one for τίθημι cite Sanskrit. This is not unexpected: on account of antiquity, prestige, the rise of comparative philology in the nineteenth century, and the presence of the Raj, if there is one ancient IndoEuropean language other than Greek and Latin to which a British classicist of a hundred years ago might choose to refer, it is Sanskrit. What might appear surprising, by contrast, is that both notes also cite Lithuanian—but I suggest that once one knows something about McKenzie’s predilections, as far as they ³⁰ Cf. e.g. Gibson 2002.

Etymology and Etymologies in the Lexicon

91

can be gleaned from his slim publications, the attention to Baltic can be read as a kind of signature.³¹ What of Πηληϊάδηϛ? The parenthetical information that LSJ gives in this instance, s.v. Πηλεύϛ, is of a different sort, being folk-etymological rather than (strictly) etymological (see fn. 29 above): Πηλεύϛ—‘(Popularly derived from πηλόϛ [“clay, earth; mud”] (which however has ᾱ in Dor.): hence prov. μὴ δεῖν τὸν Οἰνέα Πηλέα ποιεῖν don’t make wine into lees, Ath. 9.383c, cf. Demetr. Eloc. 171.)’³² Unlike pronominal parallels in Lithuanian, this is the sort of material classicists can understand and might be able to use, and while I believe there are many occasions when it is important to distinguish ‘scientific’ etymology from the folk variety, both are important, though sometimes for different reasons, and the line between them can be fuzzier than many are willing to admit.³³ I cannot resist adding that McKenzie missed a trick and an opportunity in the lemmata for θεόϛ and θεά, for what an excellent place a major reference work would be for students of Greek and Latin to discover in simple brackets that the popular connection with deus and dea is just that: popular, that is to say, ‘only’ a folk-etymology.³⁴ The impression given by the Homeric exercise in which I have just been engaged is, I imagine, as Jones would have wished: these are ‘assured results . . . [that] have been attained . . . [and] approved by Mr. McKenzie’. What the results actually mean is almost irrelevant. Play with clay is one thing—that’s literature, not linguistics; that’s art, not science—but the relative pronoun and the common μι-verb belong to the core vocabulary of Greek, and if their derivations are frankly not that exciting, that is beside the point. In a characteristically eccentric review of one of the fascicles of what would turn into LSJ, Joshua Whatmough suggested that etymology has something to it of ‘the spirit of scholarly adventure’ and that Jones and McKenzie are ‘altogether out of sympathy’ with this spirit;³⁵ true or not, the fact remains that all responsible historical and comparative linguists believe that the excitement of this simultaneously theory- and data-driven field is made possible by its strong foundations—foundations that it is the job of a ³¹ While the Lithuanian cognates of the Greek relative pronoun are noted in LS⁸, the addition of dēti (actually dėti) ́ is new in LSJ as are the Baltic forms in the lemmata quoted below in connection with Plato. ³² Right before this comes a comment on the Iliadic personal name Πηλεΐων: ‘(also as name of a frog, with play on πηλόϛ, Batr. 206)’. ³³ I consider the distinction less fundamental than most of my linguistically minded col leagues do: see the references in fn. 28 above. ³⁴ I have to correct even tenured colleagues on this point a couple of times a year. It is a matter of chagrin to classical linguists that there are few Greco Latin equations better known than ‘θεόϛ/ θεά = deus/dea’ and yet the equation is a mirage! ³⁵ Whatmough 1937, 168.

92

Joshua T. Katz

reference work to observe and not disturb. Put another way, etymological information in a tome like LSJ is not to be absorbed, except possibly cumulatively over a lifetime, but is there to provide reassurance to any reader who notices it. What the parenthetical ‘cf.’ at the end of a lemma announces—or should announce, at any rate—is in effect: ‘I’m a “solid” word of Greek and you can rely on me.’ Let me test this with a second gobbet, equally famous but very different: the opening of Plato’s Republic (327a). Socrates speaks, and what he says is in the clearest and most elementary—one might say prosaic—Greek to be found anywhere: κατέβην χθὲϛ εἰϛ Πειραιᾶ μετὰ Γλαύκωνοϛ τοῦ Ἀρίστωνοϛ προσευξόμενόϛ τε τῇ θεῷ καὶ ἅμα τὴν ἑορτὴν βουλόμενοϛ θεάσασθαι τίνα τρόπον ποιήσουσιν ἅτε νῦν πρῶτον ἄγοντεϛ. The passage contains twenty-three words (twenty-five tokens, but the three forms of the article—τοῦ, τῇ, and τήν—belong to the same lemma and thus count as one and the same thing). Every one of them is just the sort of basic—i.e. uninteresting—vocabulary item that one would expect to have clear cognates across the Indo-European languages and thus potentially receive parenthetical mention in LSJ. And, indeed, if one keeps in mind that etymology plays only a small role in the dictionary, then it seems improbable that many other passages have a greater percentage of ‘hits’ than this one. In brief, only eight of the twenty-three words receive no etymological information, and two of the eight hardly count: the personal names Γλαύκων and Ἀρίστων, neither of which is in the dictionary at all, though if one wished to connect them to γλαῦξ or γλαυκόϛ and to ἄριστοϛ, respectively, one would find no etymological satisfaction under those lemmata. As for the other six, they are the place name Πειραιεύϛ;³⁶ the nouns θεόϛ (see above), ἑορτή, and τρόποϛ (there is also no etymological information s.v. τρέπω); the conjunction καί; and the verb ἄγω.³⁷ I will not speculate on why certain ‘solid’ words receive no etymological notice, but the absence in the case of ἄγω is most peculiar since its Latin cognate, the like-meaning agō, is essentially homophonous. As it happens, however, there is also nothing s.v. ἔδω/ἔδομαι, for which the situation is basically the same (Latin edō), and while the Sanskrit cognate of the pronoun ἐγώ is mentioned (ahám), Latin ego is not. It may be that while solid words are among the most likely ones to receive etymological comment, some words are too solid (!): if an etymological connection between Greek and

³⁶ It is unclear whether there is a connection between Πειραιεύϛ and either or both of the nouns πεῖρα and πεῖραρ, for each of which LSJ supplies some etymological information. ³⁷ The only word Plato uses that most scholars today would agree is truly etymologically obscure is ἑορτή.

Etymology and Etymologies in the Lexicon

93

Latin would be entirely obvious even to a first-year undergraduate, then it may not be noted (but compare Butterfield, Chapter 3, this volume, p. 48). This leaves fifteen words in the first sentence of the Republic that LSJ considers worthy of etymological reportage. What is attractive for the present purposes is that the kinds of information contained in the brackets range widely, and so the Platonic passage goes a long way towards answering not just the first question I posed before—Which lemmata have etymological brackets?—but also the second, namely, When these brackets exist, what lies between them? Aside from speculating about folk-etymology and giving details of foreign loans,³⁸ of which there are none in the passage, lemmata in LSJ can do one or more of the following three things: they can provide (1) lowlevel information about the history of the word in question inside Greek; (2) lists of cognates (what I call ‘cf.-entries’, along the lines of what we have seen for the relative pronoun and for τίθημι); and (3) information about a word’s Proto-Indo-European, or at least pre-Greek, reconstruction, in this last instance taking the ‘cf.’-ed cognates and working backwards from them to build up a picture of how Greek sounded before—sometimes long, long before—it was anything that would conventionally be called Greek. In the first category are only two words: θεάομαι—‘(Orig. prob. θᾱϝέομαι and θᾱϝάομαι, cf. θαῦ-μα.)’ εἰϛ—‘(Orig. ἐνϛ, as in IG 4.554.7 (Argos), GDI 4986.11 (Crete); cf. ἐν, ἰν. The diphthong is genuine in Aeol. εἰϛ, but spurious in Att.-Ion.)’ In the former, as also in ἀείδω (see above), the reader gets to see a digamma, a linguistic morsel that rarely fails to intrigue students; in the latter, the reader learns how the preposition is morpho-phonologically related to the semantically similar ἐν. Curiously, the lemma for ἐν is devoid of etymological information, perhaps because the fact that Latin in and English in are cognates is too obvious to need pointing out—another example of the ‘ἄγω phenomenon’. The second category (‘cf.-lemmata’), which includes as well occasional nuggets of low-level information for which an asterisk is employed (e.g. a lost final delta and a hidden digamma),³⁹ is larger and no less mundane. Into it, with references to such languages as Latin, Sanskrit, Avestan, Old High German, and in one instance Latvian, fall two temporal adverbs (χθέϛ and νῦν) and one other kind of adverb (ἅτε); the definite article; a preposition (μετά); the preposition(al preverb) πρόϛ(-) together with the verb εὔχομαι; one more common verb (ποιέω); and two further basic forms that have etymological

³⁸ Consider e.g. the non Indo European word χρυσόϛ: ‘(Borrowed from Semitic, cf. Hebr. chārūts, Assyr. ḫ urāšu “gold”, Aram. hara “yellow”.)’. ³⁹ The dictionary uses asterisks for low level reconstructions (e.g. τό < ‘*τόδ’) but (as we will see in the next and final section) not for Proto Indo European.

94

Joshua T. Katz

information highlighted in a self-standing section, the preposition πρό and the copula τε: χθέϛ—‘(Cf. Skt. hyás “yesterday”, Lat. heri, hesternus, OHG. gestaron “yesterday”, etc.)’ νῦν—‘(Cf. Skt. nú, nū,́ nūnám, OE. nū “now”, etc.)’ ἅτε—The reader is referred to ὅστε, and from there to ὅϛ (see above) and τε (see below). ὁ, ἡ, τό—‘(With ὁ, ἁ, cf. Skt. demonstr. pron. sa, sā, Goth. sa, sō, ONorse sá, sú, Old Lat. acc. sum, sam (Enn.) :—with τό [from *τόδ] cf. Skt. tat (tad), Lat. is-tud, Goth. þata :—with τοί cf. Skt. te, Lith. tĩe, OE. þá, etc. :—with τάων cf. Skt. tāsām, Lat. is-tarum :—the origin of the relative ὅϛ, ἥ, ὅ (q.v. [see above—JTK]) is different.)’ μετά—‘(Cf. Goth. miþ, OHG. miti, mit “with”.)’ προσεύχομαι —πρόϛ—‘(With προτί, πρόϛ cf. Skt. práti “towards, near to, against, back, etc.”, Slav. protivŭ, Lett. pret “against”, Lat. pretium : ποτί (q.v.) and πόϛ are not cogn.)’ —εὔχομαι—‘(Cf. Skt. óhate “to (be able to) boast that one is”, “to brag”, Avest. aog- “declare solemnly”.)’ ποιέω—‘(ποιϝέω perh. from *ποι-ϝό-ϛ (in κλινοποιόϛ, νεωποιόϛ, τραπεζοποιόϛ, etc.) “builder”, “maker”, cf. Skt. cinóti, cáyati “arrange in order”, “build”.)’⁴⁰ plus πρῶτοϛ—The reader is sent to πρότεροϛ, and from there to πρό, which has a special non-parenthetical etymological section at the end of the entry—‘Etymology: cf. Lat. prŏ-, Slav. pro-, Skt. pra-, etc., in compounds.’ τε—Again, there is a special non-parenthetical etymological section at the end of the entry—‘Etymology: signf. A is found also in Skt. ca, Lat. -que; for signfs. B and C cf. Skt. ca in yáḥ káś ca “whosoever (with following verb)”, Lat. -que in quisque, ubique, plerique, usque, neque, nec (= non in necopinans, etc.), Goth. ni-h “not” (also “and not”), Lat. namque (= nam).’ Finally, we come to the deeper reconstructions, which understandably sometimes have ‘cf.’-s as well. In this category are the prehistories of the prepositional adverb ἅμα, which contains a syllabic nasal, and of three words that have in

⁴⁰ Whatever the precise morphological background is of ποιέω, it is cognate at the level of the root with Sanskrit cinóti: *kwei . Without an explicit note about the labiovelar, it is unclear how non linguists could be expected to understand what makes the Greek and Indic words related. (There is a labiovelar in the reconstruction of εὔχομαι as well, *gwh, but the connection with the Avestan verb may look plausible to the casual reader.)

Etymology and Etymologies in the Lexicon

95

common that they go back to a labiovelar (see above, with fn. 40), the common verbs βαίνω and βούλομαι and the interrogative pronoun (τίϛ, etc.): ἅμα—‘(Root sm̥ -, cf. A α II.)’⁴¹ καταβαίνω —κατά—There is no etymological comment, so this really belongs with words like ἑορτή and καί. —βαίνω—‘(For βάμ-yω, cf. Lat. venio, Skt. gamyáte; βάσκω corresponds to Skt. gácchati (gṷm̥ -sk-); root gṷem- in OHG. quëman “come”; ἔβην, βήσομαι fr. gṷā-, Skt. jigāti, aor. ágāt.)’ βούλομαι—‘(gṷel- gṷol-, cf. the dialectic forms.)’⁴² τίϛ (contained s.v. τιϛ)—‘(I.-E. qṷi-, cf. Lat. quis, quid, etc.; for σά, τά, v. ἄσσα, σά μάν; with τέο (v. infr. B) cf. OSlav. gen. česo.)’ If there is one thing that is as exciting as a digamma, it is a labiovelar. What have we learned from these exercises? Because pronouns, prepositions, temporal adverbs like ‘now’ and ‘yesterday’, and basic verbs like ‘want’ and ‘come’ are part of everyday vocabulary, they are likely to be inherited; because they are likely to be inherited, they are likely to be inherited into other languages as well, thereby giving evidence of cognates and allowing for the reconstruction of preforms from the distant past; and because of all this, it is easy enough for the compilers of a dictionary to provide at least minimal—and typically semantically unexciting—etymological information about them. Such words are the core of Greek, they are ‘solid’, they are there to be learned from the start of one’s studies. Are the etymological brackets in LSJ useful to anyone? Probably not—but better that they should be there than material that is manifestly incorrect. Though not an especially distinguished linguist himself, Roderick McKenzie, the generally unsung fourth member of the LSJ team, deserves credit for largely keeping etymological nonsense out of the work. While what remains is undeniably haphazard and could for the most part have been compiled even by someone without philological training, I conclude that the dictionary has many more problematic aspects than the specific etymologies to be found within its more than 2100 pages. We await the next edition of the Lexicon.

⁴¹ The cross reference notes, ‘α ἀθροιστικόν . . . , properly ἁ since it represents sm̥ (cf. ἅμα, εἷϛ = sems)’; there is a reference to ‘I. Eur. sem ’ s.v. εἷϛ. ⁴² Among the interesting dialectal forms is Coan δηλ .

6 Incorporating New Evidence Mycenaean Greek in the Revised Supplement Brent Vine

The editors of LSJ, whose final fascicle appeared in 1940, could scarcely have imagined the 1952 decipherment of Linear B and the wealth of information it provided about the Greek language and Mycenaean culture. The Preface to the 1968 Supplement justified the absence of Mycenaean Greek material with the assertion that ‘[t]he scholarly world is at present divided on the validity of the Ventris decipherment’. While it is true that pockets of resistance to the decipherment persisted until surprisingly late,¹ the alleged scholarly division seems overstated for the mid-1960s, when major handbooks like those of Ventris and Chadwick, Vilborg, and Palmer² were very much part of mainstream scholarship. However that may be, a decision was finally reached, a full two decades after the foundational handbook of Mycenaean Greek had reached its second edition (Ventris and Chadwick 1973), to incorporate Mycenaean Greek forms into the 1996 Revised Supplement. More than two decades on, we can take stock of that decision, first so as to assess some aspects of its implementation in LSJ Revd Suppl., and then to consider what this may mean for the treatment of Mycenaean Greek in future editions.

6.1. ON THE M YCENAEAN MATERIAL IN LSJ REVD SU PPL. : OMISSIONS AND I NCONSISTENCIES For the selection of material, the Preface to LSJ Revd Suppl. says only (p. ix): ‘The more speculative additions to the vocabulary have not been entered.’ Yet ¹ In the 1992 postscript to his classic account of the decipherment, Chadwick noted that there were at that time ‘still a few unbelievers’ (Chadwick 1992a, 142). ² Ventris and Chadwick 1956; Vilborg 1960; Palmer 1963. Brent Vine, Incorporating New Evidence: Mycenaean Greek in the Revised Supplement In: Liddell and Scott: The History, Methodology, and Languages of the World’s Leading Lexicon of Ancient Greek. Edited by: Christopher Stray, Michael Clarke, and Joshua T. Katz, Oxford University Press (2019). © Oxford University Press. DOI: 10.1093/oso/9780198810803.003.0006

100

Brent Vine

there are numerous omissions and inconsistencies, all the more striking because of the large quantity of accurately recorded material. Personal and place names. Especially problematic is the treatment of personal and place names, given their profusion in the Mycenaean lexicon. Evidently some onomastic material was deemed worthy of inclusion—perhaps mainly, as it seems, forms with exact correspondents in alphabetic Greek (e.g., Myc. o-re-ta s.v. Ὀρέστης). But this elides much of importance, for which the following examples may be cited: (i) a-re-ka-sa-da-ra appears s.v. ἀλέξανδρος, but the ‘Kurzname’ a-re-ke-se-u /Alekseus/ goes unmentioned, despite the prominence of this formation in Mycenaean onomastics, one of a series of morphological patterns that differentiate the Mycenaean system from first-millennium onomastic usage (Morpurgo Davies 1999). (ii) LSJ Revd Suppl. users have no way of knowing that first-millennium ‐ϕοίτης or -ϕόντης names and epithets (Hom. Ἀργεϊϕόντης, etc.) correspond to Myc. -qo-i-ta, -qo-ta names.³ (iii) Myc. ne-ti-ja-no /Nestiānōr/ prefigures Thessalian Νέσσανδρος and is closely related to Hom. Νέστωρ (Peters 1986, 547; García Ramón 1992, 243). (iv) While Myc. ra-wo-do-ko is cited s.v. λαός as equivalent to Λαόδοκος, there is no mention of Myc. names with second member -ra-wo, including e-ti-ra-wo /Ertilāwos/, a remarkable counterpart to Hom. Λᾱέρτης (Mühlestein 1968, 113 = 1987, 24; García Ramón 1992, 240). (v) Perhaps most striking, among personal names, is the absence (s.v. ἄϕθιτος) of the woman’s name a-qi-ti-ta, which implies or encapsulates the earliest attestation of the Homeric formula (or formulaic component) κλέος ἄϕθιτον (Risch 1987; Lejeune 1987, cf. Watkins 1995, 174; Katz 2010b, 361; García Ramón 2011, 225). (vi) Most toponyms with clear first-millennium correspondent are duly recorded; but while e-ko-me-no appears s.v. Ὀρχομενός (cf. Ἐρχομενός [Hes.+]), there is no mention of the man’s name o-ko-me-ne-u, despite the interesting appearance of the vowel alternation /erkho-/ ~ /orkho-/ already in Mycenaean, and the parallel with Hom. Ὀρχομενός as a man’s name (Il. 2.511, 605). Other compounds and derivatives. The treatment of non-onomastic compounds and derivatives is likewise problematic, especially where an alphabetic Greek counterpart does not exist. To be sure, a lemma like *τοιχοδόμος—an unattested word constructed in order to register Myc. to-ko-do-mo ‘builder, mason’—indicates a measure of attention to Mycenaean compounds and a willingness (cf. LSJ Revd Suppl. Preface pp. viii–ix) to record forms not documented in first-millennium Greek. Yet gaps and inconsistencies abound: ³ E.g., among many striking examples, Myc. a no qo ta /Anorkwhontās/ or /Anr̥kwhontās/, cf. Hom. ἀνδρεϊϕόντης (García Ramón 2011, 221).

Incorporating Mycenaean Greek in the Revised Supplement

101

(vii) While te-u-ke-pi and te-tu-ko-wo-a₂ are recorded (s.vv. τεῦχος, τεύχω), there is no trace of o-pi-te-u-ke-e-we (nom.pl., cf. nom.sg. o-pi-teke-e-u) ‘equipment-overseer’, displaying ‘a remarkable use of the suffix -ευς’ (Meißner 2006, 168), based on an s-stem prepositional governing compound. Nor can one find the following (to list only a few): (viii) a-mo-ra-ma /āmōrāmar/ ‘day to day’ (cf. Hom. ἦμαρ); (ix) do-po-ta ‘master’ (cf. δεσπότης); (x) me-ta-ki-ti-ta /metaktitai/ ‘fellow settlers’ (vs. simplex ki-ti-ta s.v. κτίτης); (xi) o-po-qo /opōkwois/ ‘blinders’ to block a horse’s sight (cf. πρόσωπον, μέτωπον); (xii) ko-to-no-o-ko /ktoinohokhoi/ ‘land-holders’ (cf. simplex ko-to-i-na/ko-tona s.v. κτοίνα; the lemma ἔχω lists Mycenaean finite forms, participles, and the first compound member e-keo, but not the second compound member oo-ko). It may be added that this last form carries a good deal of interest in connection with Grassmann’s Law, one of the best-known features of Greek historical and synchronic grammar. This regular pattern of dissimilation of aspiration is typically illustrated by citing the contrast between Gk. (present) ἔχω and (future) ἕξω: in the somewhat quaint formulation by Smyth (1956, 32, §125e.), ‘ἔχω have stands for ἕχω . . . , the rough changing to the smooth breathing before a rough stop. The rough breathing reappears in the future ἕξω.’ As it happens, a similar process is found in Sanskrit, also known under the name of ‘Grassmann’s Law’; and for a long time it was thought (and continues to be taught in some quarters) that the two processes, despite some differences of detail, are ultimately one and the same, and go back to a remote period not far removed from Proto-Indo-European itself.⁴ Yet the evidence of Mycenaean forms such as ko-to-no-o-ko, with its internal ‘hiatus’ (as far as the Linear B syllabic spelling goes),⁵ now makes clear that this view is mistaken. The orthographic ‘hiatus’ indicates that the second member of the compound began with /h-/ (thus -hokho-); and this, in turn, points to the likelihood that Grassmann’s Law in Greek is a post-Mycenaean innovation, and thus an independent development as compared with Grassmann’s Law in Indic. The omitted form ko-to-no-o-ko, then, turns out to be an important data point in the analysis of the relative chronology of Grassmann’s Law in Greek.⁶ As a corollary, one may note that the practice of transcribing Mycenaean forms like ⁴ For a useful historical conspectus of the question and related issues, see Collinge 1985, 47 61. ⁵ Similarly Mycenaean pu ko so , e ke e /puksohekhehe/ ‘with boxwood struts (vel sim.)’ (also nowhere in evidence in LSJ Revd Suppl.), with second member built to the same root as in o ko. (On the orthography of Mycenaean compounds with word divider, indicated here with a comma, see Melena 2014, 127 8.) ⁶ For further discussion of Grassmann’s Law in Greek and its chronology (as well as the typology of such dissimilatory processes involving aspiration), see Vine 2014, with additional references.

102

Brent Vine

e-ke ‘he has’ or e-ke-e ‘to have’ as /ekhei/ and /ekhehen/ (still common, as in Duhoux 2008 passim, a prominent handbook treatment) is to be abandoned, since such forms most likely reflect pre-Grassmann /hekhei/, /hekhehen/. Finally, as is now abundantly clear, first-millennium dialectal forms with double aspirate (such as θυϕλος ‘blind’, in a famous early West Ionic inscription) do not reflect hoary pre-Grassmann archaisms or words that somehow miraculously escaped the operation of Grassmann’s Law, but instead arise from trivial assimilatory processes.⁷ Equally numerous omissions involve derivational patterns: (xiii) While the man’s name ko-ma-we appears s.v. κομήεις (along with gen. ko–ma–we-to, though not dat. ko-ma-we-te), the theonym ko-ma-we-teja, with complex morphological derivation (García Ramón 2011, 218, n. 8), is missing. Other lexical categories. Similar problems affect verbs, nouns, and adjectives, as in the following miscellaneous examples: (xiv) Under ἄπειμι, 3pl. pres. a-pe-e-si and the participle a-pe-o (nom.sg.m.) are listed (though not pl. a-pe-o-te); yet the important participial form a-pe-a-sa (nom.sg.f., cf. Arcadian/Messenian ἔασσα vs. Att. οὖσα) is given only s.v. εἰμί. (xv) The verb form a-pe-e-ke /apehēke/ ‘he has released’ or ‘he has excused’ or ‘he has sent away’—important as one of the only Mycenaean perfect or (as in this case) aorist indicative forms that may have an augment—does not appear s.vv. ἵημι or ἀϕίημι (or anywhere else, so far as I can determine). (xvi) There is no mention of the basic verb form pe-re ‘ϕέρει’ s.v. ϕέρω (or anywhere else), despite its clear usage at least in the important religious text PY Tn 316, interestingly coordinated with a-ke ‘ἄγει’ (itself duly recorded s.v. ἄγω). (xvii) Also unmentioned (s.v. ϕθίω) is the perfect active participle (nom.pl. or nom.du.) e-qi-ti-wo-e ‘dead’ (TH Wu 75, published in 1990; see García Ramón 1990 on the extraordinary linguistic interest of the form). (xviii) Recorded s.v. ἄναξ is nom.sg. wa-na-ka, but not dat.sg. wa-na-ka-te ~ wa-na-ke-te. (xix) Given Cretan ἐρευτάς, one expects to find e-re-u-te-re ‘inspector’ (dat.sg.). (xx) The lemma υἱός (which in LSJ Revd Suppl. records dialectal and other variants) ignores both dat.sg. i-je-we (in a transparent context in the religious text PY Tn 316, already cited—among the most famous of Mycenaean texts and one of central importance for early Greek religion: di-wo i-je-we ‘to the son of Zeus’) as well as its variant i-jo (now augmented by i-ju; see Hajnal 2006, 64–6). (xxi) One looks in vain s.v. πᾶς for Mycenaean forms of this word (including pa-te ‘πάντες’, dat.pl. pa-si, /pan/ in to-so-pa ‘so much altogether’), ⁷ For the West Ionic text and the interpretation of θυϕλος via anticipation of aspiration, see most conveniently Colvin 2007, 124.

Incorporating Mycenaean Greek in the Revised Supplement

103

although pa-si-te-o-i (dat.pl., probably ‘to the All-Gods’ rather than ‘to all the gods’—another item of importance for Greek religion) is listed under θεός, and ku-su-pa is recorded s.v. σύμπας.

6.2. WHITHER M YCENAEAN IN LSJ? The above exercise could easily be extended; but that would serve little purpose, and might seem to impugn unfairly the overall high level of accuracy of the substantial Mycenaean Sprachgut that is recorded in LSJ Revd Suppl., which amounts (by my rough count) to well over 300 lexical items. But the striking omissions and inconsistencies raise broader questions, some of which may profitably be listed here, with minimal comment, as a stimulus for future discussion. There is, first of all, a basic cui bono question: what is the audience to which Mycenaean material in an edition of LSJ is directed, and for what purpose(s) is such material provided? Indeed, the Preface to the 1968 Supplement had judged that Mycenaean material might be ‘better left to special lexica’. There is perhaps room for debate on this fundamental point. But my own view (compare Katz, Chapter 5, this volume, on etymology) is that the middle course chosen by the editors of LSJ Revd Suppl. was the correct one for that time and remains correct for the future—in other words, substantial lexical coverage in the dictionary, with readers referred to more specialized bibliography in the ‘Preface’ (as in LSJ Revd Suppl. pp. viii–ix).⁸ What, further, is meant by ‘substantial’, in the above phrase ‘substantial lexical coverage’? Once again, although one may quibble with various details, I think that the editors of LSJ Revd Suppl. achieved approximately the right level of coverage overall, at least with respect to non-onomastic data. The policy to exclude ‘[t]he more speculative additions to the vocabulary’ (LSJ Revd Suppl. p. ix, cited above at the outset) was of course appropriate, and was generally well implemented; judgments of this kind as to what is ‘speculative’, moreover, will change (indeed, have already changed) with the accumulation of research on Mycenaean vocabulary. The less speculative material, however, was generally well covered, with accurate grammatical identifications and appropriate lexical comparisons with first-millennium material, when available. The major exception, in my view, concerns onomastic (and to some extent also toponymic and ethnonymic) data: although,

⁸ Such resources, of course, can now be updated in ways that would be useful for non specialist users, ill equipped as they are to navigate what has become a forbiddingly voluminous secondary literature.

104

Brent Vine

as already mentioned, some such items were incorporated,⁹ there are significant omissions (cf. (ii)–(vi) and (xiii) above). In this category, one could even argue that some of the more ‘speculative’ material may be worthy of inclusion: who, for example, would deny a place to the obscure and (at least among specialists) much-discussed figure di-ri-mi-jo, dat.sg. of a ‘*Δρίμιος’ (PY Tn 316.10), explicitly identified in the celebrated Pylian religious text already cited twice above as ‘the son of Zeus’ (cf. (xx))?¹⁰ There are, to be sure, subsidiary questions that concern the integration of Mycenaean material with other elements of the Greek lexical coverage, such as dialectal data (see Probert, Chapter 12, this volume), etymological notices (see Katz, Chapter 5, this volume), and the ‘diachronic taxonomy’ of Greek vocabulary (see Clackson, Chapter 16, this volume, with discussion of Myc. qe-to-ro-po-pi ‘quadrupeds’), to say nothing of useful comparisons with a range of other lexica, as considered by Meier-Brügger, Chapter 18, this volume (including remarks on Myc. do-e-ro ‘δοῦλος’ and related forms) and West, Chapter 19, this volume. Left aside here entirely, as it must be, is the question of how Mycenaean Greek might appear in a digital or online version of LSJ, should such an entity ever materialize: see West, Chapter 19, this volume, for some preliminary comments on that eventuality. For the foreseeable future, specialists trained in Linear B and Mycenaean Greek language—especially classical linguists interested in the history of the Greek language and classical archaeologists specializing in Bronze Age culture and history—will no doubt access Mycenaean vocabulary primarily via specialized lexica, handbooks, or other dedicated resources. Given the nature of the subject and the state of current research, this is quite as it should be. But Mycenaean Greek, as the oldest surviving form of the language that is otherwise treated so magisterially in LSJ and therefore the oldest direct window onto Greek culture, is (as specialists well know) a treasure trove of lexical riches, notwithstanding the uninspiring list format of the language’s documentation. This is material that deserves to be made available to any classicist—or indeed any lover of the Greek language—who is inclined to consult, or even just to peruse, the other riches to be found in LSJ.

⁹ Cf. above on o re ta, a re ka sa da ra, ra wo do ko, e ko me no, ko ma we; also recorded, e.g., are familiar names and places like a ki re u ‘Ἀχιλλεύς’, e ko to ‘Ἕκτωρ’, ze pu₂ ro ‘Ζέϕυρος’, a pa i ti jo (cf. Ἥϕαιστος), ka to ‘Κάστωρ’, ko ri to ‘Κόρινθος’, and a number of others, not to mention theonyms: a ta na ‘Ἀθήνη’, etc. ¹⁰ On di ri mi jo, see, e.g., García Ramón 2011, 230, with n. 58 and (in more detail) 2012, 450 6.

7 A Canonical Author The Case of Hesiod Tom Mackenzie

7.1. INTRODUCTION As any user of LSJ will know, its method is to provide a diachronic history of the meanings of each lexeme by listing the attested usages from oldest to latest.¹ This method, at core, has remained substantially unchanged throughout the nine editions of Liddell and Scott, and is adopted from Passow’s Handwörterbuch der griechischen Sprache, on which the first edition of Liddell and Scott was based. Passow’s Handwörterbuch itself was based on the earlier dictionary of Schneider, who also adopted such an approach.² Unsurprisingly, this historical principle has not gone unchallenged over the past two centuries: words are not unanimously felt to have a discrete and finite set of meanings which are clearly distinguishable from one another.³ Theoretical and historical aspects of this problem are discussed in other chapters of this book; the present chapter presents a case study of the Lexicon’s treatment of a single author, Hesiod. Early Greek poetry in general, and that of Hesiod in particular, presents certain difficulties for the historical approach which do not arise for texts of later periods. The main body of this chapter will be divided into three sections, treating different respects in which LSJ, and the historical principle it adopts, may seem problematic for the modern reader of Hesiod. The first shall consider some ways in which LSJ conflicts with current beliefs concerning the text and dating of Hesiod. The second section shall outline some respects in which the

¹ Hesiodic fragment numbers are taken from Solmsen et al. 1990. ² See Imholtz 2007, Stray, Chapter 1, this volume, Clarke, Chapter 14, this volume. ³ See, with regard to LSJ, Clarke 2010, and on lexicography in general, Kilgarriff 1997. For specific criticism of Liddell and Scott see especially Chadwick 1996, 3 28 and Glare 1997. Tom Mackenzie, A Canonical Author: The Case of Hesiod In: Liddell and Scott: The History, Methodology, and Languages of the World’s Leading Lexicon of Ancient Greek. Edited by: Christopher Stray, Michael Clarke, and Joshua T. Katz, Oxford University Press (2019). © Oxford University Press. DOI: 10.1093/oso/9780198810803.003.0007

106

Tom Mackenzie

historical principle may be inadequate for dealing with early Greek hexameter in general, given more recent scholarship on the nature and semantics of formulaic verse. The third shall treat more idiosyncratic features of Hesiod’s poetry that are particularly noteworthy for the lexicographer. Throughout, the main focus will be on the Theogony and the Works and Days, but reference will also be made to the fragments, especially those of the Catalogue of Women, which, of the other works attributed to Hesiod, has the strongest claim to authenticity.⁴ I shall investigate the treatment of particularly problematic lexemes in LSJ and earlier editions of the Lexicon. The in-progress Diccionario griego-español (DGE) and the recently-completed Lexikon des frühgriechischen Epos (LfgrE) will provide further comparanda for LSJ’s approach. It is hoped that this analysis will be of value both for the historical question of why LSJ treats Hesiod in the way that it does, and for the broader question of how any future edition might best serve the needs of students and researchers of the sage of Ascra.

7 . 2 . L S J AN D T H E MO D E R N HI S T O RY OF HESIODIC VOCABULARY In the preface to the first edition, Liddell and Scott planned ‘to make each Article a History of the usage of the word referred to’.⁵ There are certain obvious respects in which LSJ’s ‘history’ of each word seems incomplete, or conflicts with modern views about the chronology and text of Hesiod.

7.2.1. Hesiodic Citations LSJ does not cite every word Hesiod uses, and is especially unlikely to cite a Hesiodic word if it has already been used by Homer.⁶ This is in accordance with the method made explicit in the preface to the first edition, according to which, Liddell and Scott were satisfied to ‘give first the earliest authority for [each word’s use]. Then, if no change was introduced by later writers, to leave it with that general authority alone,—adding, however, whether it continued in general use or not.’⁷ However, as Hesiod is one of the earliest and most influential Greek authors, we might want a lexicon to cite his usage, especially for particularly rare words, even if it does not differ substantially from that of ⁴ See most recently on this Janko 2012. ⁵ Preface p. vii. ⁶ E.g. κρήνη (Th. 3, 6; Op. 595, 736a, 758; frr. 188A.2, 380.1); ἐσθής (Th. 574) and κερτομέω (Th. 545). ⁷ Preface p. viii.

A Canonical Author: The Case of Hesiod

107

Homer. For instance, κερτομέω occurs in both Homer and Hesiod (Th. 545) but is relatively rare in later Greek; a reader of LSJ’s entry for the verb, which lacks the Hesiodic citation, might mistakenly infer that the usage of the verb at (e.g.) A. Pr. 986 is specifically in imitation of Homer, rather than an appropriation of epic diction in general. LSJ’s occasional neglect of Hesiod may derive from a wider neglect of the poet in nineteenth-century British scholarship. It is symptomatic of that trend that Gaisford’s text of the poet, used by the first eight editions of the Lexicon, was from a collection entitled Poetae Minores Graecae (Oxford 1814, Leipzig 1823). The only other nineteenthcentury British edition of Hesiod,⁸ that of Paley (London 1861, revised 1883) describes the Hesiodic poems as ‘by a kind of common consent, or long established fashion, in this country at least, so little studied in modern times’.⁹ This nineteenth-century British distaste for Hesiod may, then, still be felt in LSJ, in cases where Hesiodic citations are lacking. Of course, a dictionary of LSJ’s size and purpose could not list every attested citation for every lexeme. However, the neglect of Hesiod in particular means that LSJ cannot be relied upon to determine whether a word is unique to Homer in early Greek poetry, even if no other early references are supplied.

7.2.2. Relative Chronology LSJ continues the practice of Liddell and Scott (and Passow) in citing the Hymns before Hesiod, as if they are earlier compositions (see, e.g. the entries for αἱμύλιος, βαρύκτυπος, and ῥοδόπηχυς). Yet this conflicts with the modern understanding of the relative chronology of early Greek hexameter poetry, whereby Hesiod is usually considered to be later than the Iliad and Odyssey, but earlier than the Homeric Hymns.¹⁰ The linguistic evidence shows that certain archaic features decrease, whilst neologisms increase, from Homer to Hesiod and then to the Homeric Hymns (with the exception of the Hymn to Aphrodite, which may be older than Hesiod).¹¹ The archaisms can now be identified as such with greater confidence since the decipherment of Linear B, as certain features—such as observation of the digamma, genitive singulars ending in -αο (first declension) and -οιο (second declension), and genitive plurals in -αων (first declension)—have been shown to go back to the

⁸ Contrast the seven nineteenth century German editions listed in the bibliography of West 1966. ⁹ Paley 1883 p.v, who also remarks of the Theogony that ‘it is certainly a dull poem’ (p. vi). ¹⁰ On the relative chronology of early Greek epic, see the collection Andersen and Haug 2012. ¹¹ As demonstrated by Janko 1982; updated in Janko 2012. The Homeric Hymn to Aphrodite is generally accepted as being the oldest of the Homeric Hymns. Janko 2012, 21 suggests it may be close in date to Homer. For further discussion see Faulkner 2008, 47 50 and Olson 2012, 10 15.

108

Tom Mackenzie

Mycenean period. LfgrE, accordingly, treats Hesiod as earlier than the Hymns.¹² Indeed, even in the nineteenth century, the Hymns in general seem to have been thought to be later than Hesiod, as Wolf ’s still influential theory that they were later rhapsodic compositions, designed as proems to longer epic performances, gained influence.¹³ Liddell and Scott may have maintained this order for the sake of convenience, or it may be a product of the influence of Passow. In any case, we might want future editions of the Lexicon to list Hesiodic citations before those of the Hymns, if the historical principle is maintained. However, the relative chronology of early Greek hexameter is highly controversial. The linguistic method for dating the poems has been challenged: the data may be explained by differences in space or genre rather than time; a poet in his seventies may compose a poem at the same time as a poet in his twenties, but still use more archaic features.¹⁴ Martin West has famously argued that Hesiod is to be dated earlier than the Iliad and Odyssey, on the basis of the ancient tradition and apparent influences of the former on the latter (although few have agreed with him).¹⁵ No chronological ordering of early hexameter is likely to be universally accepted. In light of this uncertainty, it might seem preferable, as far as early hexameter is concerned, to abandon such a strict chronological ordering. We might, for instance, prefer to order the citations according to how close they seem to an all-encompassing prototypical sense of the word.¹⁶ Alternatively, the citations could be ordered according to context in discourse or syntactical usage before date, as we find in LfgrE.¹⁷

7.2.3. The Text of Hesiod A further, obvious respect in which the historical picture presented by LSJ differs from more recent attitudes is in the very texts it cites. LSJ cites ¹² DGE is inconsistent in its relative chronology of Hesiod and the Hymns: for αἱμύλιος it cites hMerc.317 before Th. 890, despite the fact that the Hymn to Hermes has always been regarded as one of the latest of the Hymns (see Vergados 2012, 130 53); whereas the entry for βαρύκτυπος cites the Hesiodic examples (Op. 79, Th. 388, Sc. 318) before those of the Hymn to Demeter (3, 334, 441, 460). ¹³ Outlined in Wolf 1795, English translation Wolf 1986, of which see pp. 112 3 for discussion of the Hymns. This view accepted in the major editions of the Hymns by Ilgen 1796 and Hermann 1806. The first eight editions of Liddell and Scott use Wolf ’s edition for the Homeric Hymns. ¹⁴ See, e.g. West 2012, 227 8. ¹⁵ A view recently defended in West 2012. ¹⁶ See Clarke 2010. ¹⁷ For example, LfgrE’s citations under κελεύω 1cαaa refer to sense ‘command, order, direct’ (1) addressed to human or divine subordinates (c), used with an infinitive (α), specifically directed towards ἑταῖροι (aa). Of course, practical limitations may preclude a single volume dictionary that aims to cover the entire corpus of ancient Greek, such as LSJ, from including such detailed information.

A Canonical Author: The Case of Hesiod

109

Rzach’s edition of 1913. Since then, the most significant developments in our understanding of the text of Hesiod have been various new papyrus discoveries, especially of the Catalogue of Women.¹⁸ These have modified our understanding of Hesiodic vocabulary in a number of respects. The papyri have produced two new lexemes: ἰσαίων (fr. 1.8 from P.Oxy. 2354, first published in 1956, of which note is taken in the Supplements) and εὐσχεθής (fr. 33a.32 from P.Oxy. 2481, first published in 1962, of which note is taken in the Revised Supplement). Occasionally, we now find that Hesiod is the earliest attested author for a particular word. In some cases (ἠλοσύνη, fr. 37.15; σπαρνός, fr. 66.6; μίτρα for a woman’s girdle at fr. 1.4), note is taken in the Supplements. For others, however, the Supplements provide no new assistance (συνοπηδός, fr.26.10; μαψίδιος, fr.10a.87; possibly ὁμοπά τριος, fr. 280.18 a fragment of doubtful authorship). An active form of στεϕανόω is found at fr. 185.5, in spite of the fact that LSJ described the word as ‘used by Hom. and Hes. only in Pass.’, and no correction is made in the Supplements. Of course, LSJ could not be faulted for failing to take account of papyri that were discovered later than 1940. However, in at least one instance, LSJ fails to cite Hesiod where, on the basis of the manuscript tradition, he is the earliest attested source for a lexeme. ἄπλαστος occurs at Th. 151 and Op. 148 in all modern editions of Hesiod (including Rzach) where it refers to the ‘Hundredhanders’ and the men of the age of bronze respectively; but Aristotle (Mete. 385a15) is the earliest citation for the lexeme supplied by LSJ, which translates the first sense of the adjective as not capable of being moulded. However, LSJ also describes the adjective as a ‘v.l. for ἄπλατος’. The Revised Supplement adds the Hesiodic citations: ‘after “II” read “app. = ἄπλατος, (w. which it frequently coexists as a variant), Hes. Th. 151, Op. 148”’. The Supplement, then, accepts the validity of the variant adopted by modern editors, and treats it as an alternative form of ἄπλατος, a word meaning, ‘unapproachable, always with a notion of terrible, monstrous’ (LSJ s.v., related to πελάζω). However, as West points out, the translation, ‘unmoulded’—in the sense rudis, informis—may be more appropriate for the Hesiodic examples, given that πλαστός means ‘shaped’ by a craftsman at Th. 513.¹⁹ If West is correct, LSJ and the Revised Supplement are simply wrong in treating Hesiod’s ἄπλαστος as interchangeable with ἄπλατος. The possible mistranslation of ἄπλαστος may appear to be simply a failure to keep up to date with modern scholarship’s understanding of the adjective.

¹⁸ For an overview of the contribution of Hesiodic papyri to our understanding of early hexameter language, see West 2008. ¹⁹ West 1978 ad Op. 148 9, revising the view expressed in West 1966 ad Th. 151 2.

110

Tom Mackenzie

However, the first edition of Liddell and Scott already entertains the possibility of West’s interpretation: ‘usu. taken as = ἄπλατος, monstrous, huge, v. l. Hes. Op. 147, Th.151, Soph. Fr. 350 : if in this signf. also it be not better referred to πλάσσω, shapeless, monstrous, like Lat. informis: cf. ἄπλατος, ἄπληστος᾽

In this case, the first edition actually gives a better idea of the problem posed by Hesiod’s text than does LSJ. The extra clause, which poses the possibility that the adjective is related to πλάσσω, is included in the first six editions of the Lexicon, but drops out in the seventh. As the Lexicon developed across its various editions, lengthy explanations of the kind supplied for this entry in the first edition dropped out in favour of concise definitions, to create space for more, and more detailed, citations.²⁰ This seems to be especially the case for Homer and Hesiod, who were given particular prominence in the first edition, but who had to make way for later authors in subsequent editions.²¹ The reduction in the space given to Hesiodic entries, in this instance, has elided an alternative interpretation of a word—which is now thought to be the correct one.

7.3. LEXICOGRAPHY AND E ARLY GREEK EPIC Developments in scholarship, changes in the main editions used, and discovery of new evidence are all aspects which would eventually render any Greek lexicon out of date for any author. There are, however, certain problems more specific to this period and type of literature.

7.3.1. Obscurity of Early Vocabulary Because this is the earliest period of surviving Greek literature, there are many words whose meanings are uncertain to us, and which were already disputed by ancient commentators. Occasionally, LSJ gives an indication that this is the case (see, e.g. s.v. ἀνοπαῖα). However, in other instances, LSJ misleadingly offers a confident definition without indicating that the meaning of the term was already uncertain in antiquity. An example of particular relevance to Hesiod is the epithet of men, ἀλϕηστής. This term is given various ²⁰ See in greater detail Imholtz 2007. ²¹ The preface to the first edition claims that ‘Passow indeed had done all that was necessary for Homer and Hesiod’ (p. vi) whilst Passow’s dictionary only covered Greek literature from Homer to Herodotus.

A Canonical Author: The Case of Hesiod

111

contradictory definitions by the scholiasts.²² LSJ simply gives ‘lit. earners (ἀλϕάνω), i.e. enterprising men’, without indicating that the sense is uncertain. More recent scholars, including the author of the entry in the Revised Supplement, see the epithet as a compound consisting of ἄλϕι-, ‘barley’ + *ed- ‘eat’ (cf. ὠμ-ηστής), so that it means ‘barley-’ or ‘bread-eating’.²³ This sense is significant thematically within the Works and Days: men are by definition eaters of bread, therefore they must work the land to produce barley. It will be a hindrance to the student of the poem if she does not find that definition in the dictionary, and does not take the initiative to look in the Supplement. Liddell and Scott can hardly be blamed for not identifying a sense that would only find acceptance in the following century. But more importantly, the Lexicon would be more useful if it mentioned that the word is of disputed meaning, as, for instance, DGE does. Indeed, the entry for the term in the first edition (which itself is virtually a translation of Passow’s entry) supplies more information, as it makes explicit that the translation ‘inventive, reasoning’ derives from Eustathius. Again, a definition has been slimmed down to make way for more entries and citations; consequently, LSJ does not indicate that its definition is partly derived from a late authority, a detail which might suggest to the reader that the sense is questionable.

7.3.2. Formulaic Language In this last example, LSJ’s definition justifies itself by making use of the attested contexts of the epithet and the verb from which it was believed to be derived. However, this method is made inherently problematic by the fact that ἀλϕηστής is a formulaic epithet. Milman Parry famously placed emphasis on the ‘essential idea’ of formulas, given their deployment on the principle of ‘thrift’ or ‘economy’, so that the composer freely selects from a set of formulas which ‘having the same metrical value and expressing the same idea, could replace one another’.²⁴ We might, therefore, prefer to see the formulaic phrase ἀνδράσιν ἀλϕηστῇσιν (Op. 82) as expressing merely the ‘essential idea’ of ‘men’, whilst the epithet fills out the expression so that it occupies space up to the third-foot caesura.²⁵ In that case, the context would be irrelevant for our understanding of the sense of ἀλϕηστής in this formula. Recent commentators ²² As reported in LfgrE s.v. A scholiast at Od. 1.349 defines it as ἐϕευρέται, εὑρετικοί, ἐπινοητικοί, whilst scholiast D at Il. 18.593 defines it as ἔντιμοι, βασιλεῖς. The epithet is also used at Aesch. Th. 770 and S. Ph. 709. ²³ E.g. LfgrE and Beekes 2010 s.v.; West 1966 ad Th. 512; cf. Il. 6.142 βροτῶν οἳ ἀρούρης καρπὸν ἔδουσιν. ²⁴ Parry 1971, 276 (originally published in 1930). ²⁵ On the ‘essential idea’ see Parry 1971, liv, 131 4. On the related principle of economy in Hesiod, see Edwards 1971, 55 73.

112

Tom Mackenzie

tend to take a more nuanced view of stock epithets, allowing that they can be deployed with relevance to particular contexts,²⁶ but few would deny that a formulaic collocation such as ἀνέρες ἀλϕησταί constitutes a distinct lexical unit. J.M. Foley, in one of many influential contributions to the oral-formulaic theory, even goes so far as to describe such expressions as ‘words’ which belong to a Homeric epic lexicon.²⁷ Taken literally, this implies that formulas might be given their own entries in a dictionary of ancient Greek. Such a practice would add significantly to the weight of the dictionary. However, it might not be too much to want a dictionary to adopt explicitly a definition of a formula, and indicate the formulas in which lexemes are used accordingly. In most cases, LSJ does provide information which enables the user to infer that a word is used in a formula: we learn that ἀλϕηστής is used in the Odyssey ‘in phrase ἀνέρες ἀλϕησταί’.²⁸ In other entries, we are sometimes told when a word is an epithet (Hesiodic examples are πολύχρυσος and μητιόεις), but a more consistent and systematic approach to terms used in epic formulas might be desirable. We are not, for instance, told that πολυγηθής is an epithet of Dionysus (Th. 941, Op. 614) or that πελώρη is one of Γαῖα (Th. 159, 173, 479, etc.). A consistent approach to marking formulaic expressions might help the user to decipher obscurities which arise when formulaic expressions are adapted to new contexts. An example is the description of the snake that guards the golden apples at Th. 333–5: Κητὼ δ’ ὁπλότατον Φόρκυι ϕιλότητι μιγεῖσα γείνατο δεινὸν ὄϕιν, ὃς ἐρεμνῆς κεύθεσι γαίης πείρασιν ἐν μεγάλοις παγχρύσεα μῆλα ϕυλάσσει. Ceto mingled in love with Phorcys and gave birth to her youngest, a terrible snake, which, in the depths of the dark earth guards the all golden apples in great limits.

The relevance of the phrase ‘in great limits’ is unclear, and seems to jar with the context. West explains that we are to understand that the ‘limits’ are of the earth, as the expression seems to be an adaptation of μεγάλης ἐν πείρασι γαίης, ‘in the limits of the great earth’, which occurs later in the poem (Th. 622). For West, the omission of γαίης in 335 is made easier by its occurrence in the previous line. The expression πείρατα γαίης is conventional in Homer (Il. 14.200, 14.301; Od. 4.563, 9.284; cf. Il. 8.478–9). The poet, then, seems to have adapted a conventional expression, but in the process of adaptation he has omitted a significant detail which would have made it more readily comprehensible. An audience familiar with epic diction may have had little trouble in supplying the detail. One problem that this poses for lexicography is ²⁶ See, e.g. Nagy 1990, 18 35; Foley 1997; Graziosi and Haubold 2005. ²⁸ DGE, for instance, refers to ἀλϕηστής as an epithet.

²⁷ Foley 1997, 167.

A Canonical Author: The Case of Hesiod

113

whether to include, under πεῖραρ, an expression such as ‘esp. of the Earth’ to aid the modern reader with the decipherment of this passage,²⁹ or whether to supply the meaning ‘limit’ and leave it for the reader of Hesiod to work out the precise significance of the word in that passage. LSJ has, as the word’s first sense, end, limit, and quotes part of a line (Il. 8.478) in which the expression πείρατα γαίης is used; however, the first edition of the Lexicon has the extra detail, after the start of the word, ‘poet. esp. Ep. for πέρας, an end, usu. in plur. πείρατα γαίης’. This is, essentially, a translation of the opening of Passow’s entry for the word.³⁰ The entry for the earlier editions of Liddell and Scott, and for Passow, would actually be more helpful for deciphering this Hesiodic passage, as they make it more explicit that πείρατα is often used with γαίης in epic. The detail that in epic the word is mostly used in the plural is included in the first eight editions of the Lexicon, but drops out in LSJ. As we saw with ἄπλαστος and ἀλϕηστής, extra information that would have been useful for deciphering particularly obscure usages of Hesiod has been lost.

7.3.3. Glosses Oral composition (or written imitation of oral composition) seems to encourage the preservation of archaisms, as the artificial Kunstsprache retains forms and vocabulary that have dropped out of the vernacular. As a result, in some cases, Homer and Hesiod use words of which the meaning may already have been uncertain to them. Michael Silk has coined the term ‘iconym’ to deal with certain instances of this kind, where an obsolete word, which no longer carries a specific denotative meaning, is used for its sound or literary associations.³¹ An example he cites, used by Hesiod (Th. 611), is ἀλίαστος, which LSJ translate as not to be turned aside, unabating, and as derived from λιάζομαι (go aside, recoil, shrink from). Although Silk accepts the etymology, he argues that, semantically, ‘the word is clearly at some remove from its parent’, as in epic it is applied to battle (μάχη), the din of battle (ὅμαδος), wailing (γόος), and then, in Hesiod, to distress (ἀνίη). As a result, Silk converges with Erbse’s judgement that the word becomes ‘an empty epithet’.³² If Silk is correct, this may present a peculiar problem for lexicography: instead of supplying specific definitions, it may be more appropriate, with these sorts of lexemes, for the ²⁹ We sometimes receive information about the typical context of a word in Homer. E.g., for μαλερός we are told ‘in Hom. always epith. of fire’, a detail which will help our interpretation of the word in later authors. ³⁰ ‘poet. bes. Ep. st. πέρας, das Ende, das Aeusserste, die äusserste grenze, gew. in plur. πείρατα γαίης, πόντου, Ωκεανοῦ’ ³¹ Silk 1983; cf. Silk, Chapter 17, this volume. ³² Silk 1983, 316, citing Erbse 1953.

114

Tom Mackenzie

lexicographer to provide the etymology with citations of uses, and leave the reader to make up her own mind.³³ A further possible example may be the epithet ὁμοίιος (Op. 182): οὐδὲ πατὴρ παίδεσσιν ὁμοίιος οὐδέ τι παῖδες οὐδὲ ξεῖνος ξεινοδόκῳ καὶ ἑταῖρος ἑταίρῳ, οὐδὲ κασίγνητος ϕίλος ἔσσεται, ὡς τὸ πάρος περ. (Op. 182 4) Father will not be like minded with sons, nor sons at all, nor guest with host, nor comrade with comrade, nor will the brother be dear, as he once was.

Here, as LSJ identify, the word is used for ὁμοῖος.³⁴ However, in Homer, ὁμοίϊος seems to be a separate adjective of uncertain meaning, applied to old-age, death, strife, and war.³⁵ This sense has been supported by the Sanskrit noun ámīvā [f.], ‘misery, vexation, suffering’, so that ὁμοίϊος may come from the putative forms *ὁμοιϝιος and *ὁμοι-ϝα.³⁶ Possibly, Hesiod has used the term in a manner contrary to established epic usage. He may simply have misunderstood the word. Alternatively, the adjective may simply be a distended form of ὁμοῖος: it is unclear whether we should treat the instance as one of ὁμοῖος used homophonously with ὁμοίιος, or as a distinct usage of the separate lexeme ὁμοίιος. LSJ provides an ‘A’ definition of ὁμοίιος and a ‘B’ definition (which cites Hesiod) whereby it is synonymous with ὁμοῖος. This approach enables the student to make sense of the passage, but obscures the uncertainty over the relationship between the two sub-lexemes.

7.3.4. Misunderstandings? In other cases, words seem to acquire a secondary sense as a result of a poet misunderstanding or reinterpreting an earlier usage. This hypothesis was first applied to early Greek epic in Manu Leumann’s influential Homerische Wörter (Leumann 1950).³⁷ A Hesiodic example may be the adjective μεταχρόνιαι, ³³ Silk concludes (p. 330), ‘I would wish to see LSJ labelling iconyms as such, with the instances arranged into groups defined by contexts or associations, not by meanings, and presented for the most part without reference to meanings, and with sufficient discussion of when the word seems to have attained its iconymic status.’ ³⁴ Exactly what sense of ὁμοῖος is used here whether ‘at one with’ or ‘physically similar’ is uncertain and was disputed in the scholia vetera. See West 1978 ad loc. LSJ opts for the former translation. ³⁵ γῆρας (Il. 4.315); θάνατος (Od. 3.236); νεῖκος (Il. 4.444); πόλεμος (9.440, 13.358, 15.670). LSJ supplies the examples with the possible translations of ‘distressing’ (supported by the fact that it = κακός acc. to Anon. ap. Apollon.Lex.) and ‘common to all’. ³⁶ This explanation seems to go back to G. Christ’s 1884 edition of the Iliad: see Catenacci 1996, 134, who objects that the etymology does not account for the aspiration of the Greek word. ³⁷ See more recently Reece 2009.

A Canonical Author: The Case of Hesiod

115

applied to the Harpies at Th. 269. Morphologically, the word seems to be formed from μετὰ χρόνον, and on that basis should mean ‘happening afterwards’ or ‘delayed’, the first senses given by LSJ (citing Tryph. 1, Luc. Alex. 28, Gal. 19.522). However, in the Theogony and other poetry the word seems to mean ‘high in the air’, and is glossed by the ancient commentators as such.³⁸ West explains this unexpected sense using the ‘méthode Leumannienne’:³⁹ in the Catalogue of Women, the word is used in the expression, μεταχρονίοισι πόδεσσιν (fr. 150.34, possibly of a Harpy at fr. 76.18), apparently of someone pursuing someone else, where the sense of ‘with following feet’ would be appropriate. In the instance from fr. 150, the pursuit seems to take place δι’ αἰθέρος (fr. 150.35); if such a passage occurred in a pre-Hesiodic epic, a poet may have misunderstood the expression to mean ‘with high-flying foot’, hence the epithet came to mean ‘high-flying’. A further example along these lines is the name for the flower, ἀσϕόδελος, which first occurs as a noun at Op. 41 (although its adjectival form occurs at Od. 11.539, 24.13), in a couplet describing the ‘gift-eating kings’ (Op. 40–1): νήπιοι, οὐδὲ ἴσασιν ὅσῳ πλέον ἥμισυ παντὸς οὐδ’ ὅσον ἐν μαλάχῃ τε καὶ ἀσϕοδέλῳ μέγ’ ὄνειαρ. Fools, they do not know how much the half is more than the whole nor how much benefit there is in the mallow and the asphodel.

Much ink has been spilled over the significance of the ‘mallow and the asphodel’:⁴⁰ traditionally they have been thought to refer to the humble fare of the farmers, which the kings fail to appreciate.⁴¹ An alternative possibility (not incompatible with the traditional explanation), suggested by Bruno Currie, is that the mallow and asphodel are mentioned because of their mystical and eschatological associations.⁴² For Hesiod, then, eating the ‘mallow and asphodel’ may be part of becoming a θεῖος ἀνήρ (Op. 731) who will enjoy eschatological benefits, just like the θεῖον γένος (Op. 159) of heroes, who will end up on the isles of the blessed (Op. 170–3). Steve Reece has recently suggested a diachronic development, along Leumannian lines, of ἀσϕόδελος which explains how the flower came to have these eschatological connotations.⁴³ In the Odyssey, the adjectival form of the word occurs, in both of its occurrences, in the expression κατ’ ἀσϕοδελὸν λειμῶνα. Reece suggests that this formula was originally κατὰ σποδελὸν λειμῶνα, where σποδελός was an adjectival form from the root, σποδ-, also found in the Homeric σποδός, meaning ‘ashes’, and the ³⁸ The Et. Mag. explains it as meaning μετέωρος; a scholiast on the line explains that χρόνος is really a word for οὐρανός. ³⁹ West 1966 ad loc., recapitulated in West 2008. ⁴⁰ See the detailed explanation of the lines by Currie 2007, 172 5, who cites extensive bibliography. The topic was a quaestio in antiquity (Plut. Sept. sap. conv. 157d 158c; Gellius 18.2.13; a Proclean scholion on Op. 41). ⁴¹ See, e.g. West 1978 ad loc. ⁴² Currie 2007. ⁴³ Reece 2009, 261 71.

116

Tom Mackenzie

common suffix –ελος. The ‘ashy meadow’ would be appropriate for the underworld in a society which cremates its dead. Here, the etymology could support Currie’s interpretation of the controversial line, and affect our interpretation of the meaning of the poem as a whole. It is difficult to assess the extent to which a lexicon ought to take account of such complex and speculative etymologies. LSJ, of course, does not, as it was published before Leumann’s Homerische Wörter, and long before these particular etymologies were suggested. LfgrE provides the different possible etymologies and cites West here, without offering his explanation for the shift in full. Reece’s explanation of ἀσϕόδελος is too recent for any of the major dictionaries to take account of, although LfgrE gives an indication of the plant’s mystical and eschatological associations by quoting a substantial part of Proclus’ comment on the line, and mentioning that the line was ‘déjà énigmatique dans l’Antiquité’.⁴⁴ LSJ gives an indication that this may be the case by mentioning that it refers to ‘the asphodel mead which the shades of heroes haunted’. The first edition adds the extra information that the noun refers to, ‘a plant of the lily kind, the roots of which were eaten’. This might help the reader of Hesiod make some sense of the line. The possibility that poets reinterpret, ‘mistranslate’, or ‘mis-use’ certain terms raises a problem for the lexicographer: most users will want a lexicon of ancient Greek to describe the sense of words as they are found in ancient Greek texts, rather than to prescribe usages that are accepted as being correct. However, users would also want to know when ancient poets use words in a manner that conflicts with normal usage, in order to reconstruct ancient ways of reading the text in question. If an author seems to use a word in a manner that (as far as can be reconstructed) appears to be contrary to established usage, we might interpret such a usage as ‘erroneous’ on the understanding that the author misunderstood the word, and that his usage would have been obscure to his audience. Alternatively, we might choose to be more charitable, and so read such deviant usages as in some way functional. This dilemma seems to be particularly acute in the case of Hesiod, given his reputation as a rough or primitive poet.

7.4. LEXICOGRAPHY AND HESIOD

7.4.1. Hesiod gets ‘muddled’? Hesiod, perhaps more than most early Greek poets, has been thought to make mistakes, both in the form of linguistic errors—thought to arise as a ⁴⁴ DGE offers an ‘Etim. dub.’ (doubtful etymology) of the noun from σϕυδόω, ‘be in full health’.

A Canonical Author: The Case of Hesiod

117

consequence of the fact that he is a Boeotian poet using the predominantly Ionic dialect of epic⁴⁵—and in the form of thematic or narrative inconsistencies.⁴⁶ Both these alleged deficiencies are of relevance to the lexicographer. I have already mentioned Hesiod’s possibly erroneous usage of ὁμοίιος (see above). A further possible example of Hesiod ‘muddling’ the established sense of a word is at Op. 240: πολλάκι καὶ ξύμπασα πόλις κακοῦ ἀνδρὸς ἀπηύρα, (240) ὅστις ἀλιτραίνῃ καὶ ἀτάσθαλα μηχανάαται. Often even a whole city suffers because of an evil man who sins and devises wicked deeds.

The Homeric participle ἀπούρας seems to mean ‘taking away’ or ‘wresting from, robbing of ’ (LSJ s.v. I). However, it seems to make little sense to say that ‘the whole city takes away from a bad man who sins and contrives wickedness’. A form of ἐπαυρίσκω/ἐπαυρέω, ‘partake of, have the enjoyment of ’ (LSJ s.v. I; cf. Op. 419) may seem more appropriate, and indeed, Triclinius suggested the emendation ἐπαυρεῖ, but otherwise the tradition is unanimous as far back as Aeschines (Ctes. 135) that ἀπηύρα is the correct reading. Possibly in imitation of this passage, Euripides and the author of Prometheus Bound also seem to have used ἀπηύρα to mean, ‘partake of ’ (Andr. 1030; PV 28), a fact which may further confirm the reading in Hesiod. As a result, LSJ provide a second sense of ἀπούρας as ‘receive good or ill, enjoy or suffer’ citing the examples from Hesiod and Euripides. West suggests that Hesiod, and possibly the two tragedians, ‘got the words muddled’. Such ‘muddling’ may be paralleled elsewhere in early epic.⁴⁷ A central reason for supposing this to be the case is that Hesiod’s usage here is unparalleled in archaic poetry. This reason is an application of the principle (usually applied to textual criticism) attributed to Wilamowitz, that einmal ist keinmal und zweimal heisst immer. That is to say, if one other such instance were to be found in early hexameter, we might assume that this was established usage in that tradition, and so not an instance of ‘muddling’ or textual corruption; but since only one instance is found, it is to be considered an error ⁴⁵ On Hesiod’s dialect which in fact seems to observe certain Ionicisms more consistently than the Homeric poems see Cassio 2009. ⁴⁶ Kirk 1962, 66 remarks that ‘old formulas derived from the Ionian tradition . . . are combined with each other in a rather clumsy, redundant or colourless manner’. West 1966, 23 is similarly damning of Hesiod’s style; his commentary on the Works and Days (West 1978) frequently points out narrative or logical inconsistencies. He remarks on p. 41, ‘To anyone who expects an orderly and systematic progression of ideas, it is liable to appear a bewildering text.’ More recent critics, such as Lardinois 1998, Clay 2003, and Scully 2015 tend to be more sympathetic towards Hesiod’s style, and read the inconsistencies as serving literary functions. ⁴⁷ Note Il. 7.434 and 24.789, where ἔγρετο comes from ἀγείρω, not ἐγείρω. Conventionally, the relevant from of ἀγείρω would be ἤγρετο. Janko 1992, 35 n. 65 suggests that bards could have confused the two forms. See also West 2001, 23.

118

Tom Mackenzie

of the original author or the result of textual corruption (as Triclinius thought). The principle, however, is hardly infallible: it is quite possible that Hesiod’s example represents an established usage of which no other instance happens to survive. The parallels from Andromache and Prometheus Bound suggest that this usage at some point became an established one, at least among tragedians. It may even be the case that the two verbs ἐπαυρέω and ἀπαυράω had become ‘muddled’ at an earlier point in the epic tradition, but that this was already normalized by the time of Hesiod. There is little objective evidence on which to determine whether Hesiod’s use of ἀπαυράω here is deviant or not, and our response to this question will depend largely upon fallible a priori methodological assumptions, such as that einmal ist keinmal. In addition to possible linguistic errors such as this, the two canonical Hesiodic poems have often been thought to contain inconsistencies of narrative and logic which may affect our understanding of particular lexemes. Perhaps the most famous instance is in the myth of Pandora, where Hope is found in the jar of evils (Op. 90–105).⁴⁸ There are at least two apparent inconsistencies in the story: firstly, ἐλπίς is thought to be a good thing, yet it is in the jar of ‘evils’; and secondly, ἐλπίς is among men because it remains in the jar, whilst the evils are among men because they escape it. These have troubled commentators since antiquity.⁴⁹ There are many ways in which Hesiod could be defended against the charge of illogicality: it has been argued that he is being ironic, or we could interpret the symbolism of the jar in such a way as to make it coherent.⁵⁰ Relevant to our immediate purposes, however, is the question of how we choose to translate ἐλπίς: we are accustomed to translating the noun as ‘hope’, a word which has positive connotations in English; and we would not expect to find ‘hope’ in a jar of ills. However, as is well known, ἐλπίς can often mean ‘anticipation’ or ‘anxious thought about the future’ rather than the positive sense of ‘Hope’.⁵¹ ‘Empty hope’ (κενεὴν . . . ἐλπίδα) seems to be a bad thing at Op. 498. It is possible, then, that ἐλπίς is used in a negative or ambiguous sense, so that it is presented as one of the ‘evils’ after all. LSJ simply cite the instance under the sense of hope, expectation, and so does not address the interpretive problem. Possibly, Hesiod is using an already established, negative or ambiguous sense of ἐλπίς. Alternatively, he may be making the unexpected point, contrary to the normal usage ⁴⁸ Cf. West 1978, 170, ‘It is of course illogical to make the same jar serve both purposes at once [sc. imprisoning the evils and maintaining Hope]. But that is what Hesiod has done, and we must not distort his meaning for the sake of better logic.’ ⁴⁹ Note the scholiast on Op. 97, who reports that Aristarchus claimed that there were two Hopes, a good and a bad one, the former of which flew out of the jar whilst the latter remained. ⁵⁰ Nisbet 2004 argues for irony in Hesiod; for a detailed summary of different views on the symbolism of the jar, see Verdenius 1985, 66 70. ⁵¹ See, e.g. LSJ s.v. II ‘anxious thought on the future, boding’.

A Canonical Author: The Case of Hesiod

119

of ἐλπίς, that it is a bad thing.⁵² Either possibility is more charitable to Hesiod than to regard him as illogical. Our understanding of the semantics of Hesiod’s use of ἐλπίς depends in part on whether we apply a principle of charity when interpreting his text, or whether we see him as an imperfect maker, struggling to express ideas and sometimes slipping into confusion.

7.4.2. Aetiologies The myth of Pandora and the jar is part of a wider tendency of Hesiod to provide aetiological explanations that hint at nuances of the meaning of certain terms. Given Hesiod’s immense influence in antiquity, we might especially want a dictionary to supply detailed references to his usage of terms for which he provides aetiological or explanatory myths.⁵³ However, for certain key terms which fall into this category, LSJ’s impression of Hesiod’s usage is somewhat misleading. One example is the description of the two ‘Strifes’ (ἔριδες) that feature at Op. 11–25. Here, Hesiod revises the claim made at Th. 225, that there was a single ‘Strife’. The two ‘Strifes’ have different functions: one is blameworthy (Op. 13), and responsible for war and slaughter (Op. 14) while the other is ‘much better’ (Op. 19) and stirs a lazy man to work when he sees a wealthier man (Op. 21–2). This latter ἔρις is perhaps more accurately translated as ‘emulation’ or ‘rivalry’. LSJ has, as sense III, ‘personified Eris, a goddess who excites to war’ (and cites Th. 225 here), whilst, as sense IV, it has ‘contention, rivalry’, at the end of which is written, ‘cf. Hes. Op.24.’ This may provide adequate help for deciphering the Greek; however, it is incomplete, and arguably misleading, for readers interested in the history of the concept of Eris. We might want to find reference to the two ‘Strifes’ under sense III, and a citation of this passage after the Homeric references under sense IV. A further example is in the case of δίκη. Prominently, in the Works and Days, Δίκη is personified at 212ff. and her typical behaviour is described in some detail: she wins out over hubris in the end (217); and when she is ‘dragged’ by gift-eating men who make crooked judgements (220–1), she follows the city and tribes of men weeping, and bringing trouble to them (223–4). Later, we are told that whenever a mortal harms her by scorning her, she sits by her father Zeus and tells him of the unjust mind of men (259–60). The description of Δίκη in the Works and Days would be influential in later ⁵² Cf. Clay 2003, 102 3, who argues the Hesiod’s hope is the ‘ultimate kalon kakon’, and like the jar, and Pandora herself, is ostensibly attractive but harbours ills for man. Cf. also Theognis 1197, where personified ἐλπίς is described as the only good god among men. ⁵³ Hesiod’s influence on Greek religious thought is perhaps most explicit at Hdt. 2.53. The reception of Hesiod in antiquity has been the subject of much recent research: see Koning 2010; Boys Stones and Haubold 2010; Hunter 2014; and Van Noorden 2014.

120

Tom Mackenzie

literature.⁵⁴ However, LSJ does not include a separate sense for the personified Δίκη as it does for ‘Strife’; instead, it simply includes, under sense II ‘order, right, personified, Hes.Th. 902, A.Th. 662, etc.’ We might wonder whether some elements of Hesiod’s description here, or at least reference to it, ought to be included in a dictionary definition of the term. As it is, the only citations for Hesiod under the entry for Δίκη are that to Th. 902, and reference to Op. 219, 250 for the phrase δίκαι σκολιαί (‘crooked judgements’). Surprisingly, there is no reference to Hesiod under sense IV.3 ‘the object or consequence of the action, atonement, satisfaction, penalty’, in spite of the fact that the word seems to have this sense at Op. 272, and the first source cited for this sense is Herodotus. Because of the immense influence of Hesiod for the ancient understanding of δίκη and ἔρις, we might expect a lexicon to privilege references to him; instead, the opposite seems to be the case: he is neglected at particular points where he seems to have been most influential. This may be a residual consequence of the nineteenth-century neglect of Hesiod, relative to Homer.

7.4.3. Personifications and Figurative Language Part of the difficulty in both of these cases is that the terms are sometimes used as abstract nouns, and sometimes used as divine personifications. This raises a wider and difficult question of whether there is a conceptual distinction in early Greek thought between divine personifications and the abstract nouns that they personify.⁵⁵ LSJ is not consistent in distinguishing divine personifications from the non-personified usages: for ἔρις, the personified use receives a separate sense distinction, whilst for Δίκη, the personification is absorbed under the sense of ‘order, right’. A similar case is τύχη. Here again the personified Tyche is not given a separate category of her own, but among the different senses (e.g. ‘II.1 fortune’; ‘III.2 ill fortune’) examples appear where the Lexicon provides glosses indicating full personification. This latter example is particularly pernicious, as Hesiod is in fact the earliest attested source for this being, giving her as a daughter of Ocean at Th. 360. Poets often use personifying language with abstract nouns, but that does not necessarily mean that the abstract noun is conceived of as a god; it is up to the editor to determine whether the language used is sufficiently personifying to justify capitalizing the term and treating it as a deity.⁵⁶ A problem case is Op. 230, which describes the just city: ⁵⁴ Cf., e.g. A. Fr. 530 M (282 Ll. J.) and X. Cyn. 12.21 with Hunter 2014, 59 63. See also Koning 2010, 172 7. ⁵⁵ For the worship of such personified ‘abstractions’ in cult, see especially Stafford 2000. ⁵⁶ For criteria for distinguishing such personifications, see Stafford 2000, 9.

A Canonical Author: The Case of Hesiod

121

οὐδέ ποτ’ ἰθυδίκῃσι μετ’ ἀνδράσι λιμὸς ὀπηδεῖ nor does famine attend straight judging men

Λιμός is one of the children of Eris at Th. 227, and the verb ὀπηδεῖ, (follow, accompany, attend) may be thought to have a personifying force. As a result, Solmsen and West capitalize the noun here, treating it as a personification. However, ὀπηδέω is frequently used of things, rather than people or animals (LSJ s.v. II; cf. Il. 5.216; Od. 8.237), and Hesiod sometimes uses nouns in a clearly un-personified manner which he personifies elsewhere.⁵⁷ Perhaps as a result of this, other editors, such as Rzach and Most, leave λιμός uncapitalized at 230. The choice of whether or not to treat a noun as a true personification is a difficult one, not least because authors personify nouns for literary effect, without conceiving of them as deities. In the case of λιμός, no edition of Liddell and Scott includes a personified sense for it, in spite of its clearly personified status at Th. 227. The editorial dilemma concerning whether or not to capitalize certain abstract nouns in Hesiod is part of a larger problem of when to distinguish metaphorical from literal uses of language. Few areas of semantics have been more discussed in recent decades than metaphor.⁵⁸ A clear division between metaphorical and literal uses of language has sometimes been seen to be problematic: metaphors are much more ubiquitous in everyday speech than was once realized, and, according to some definitions of the term, all language is seen to be metaphorical. Some have rejected altogether a dichotomy between metaphorical and non-metaphorical uses and preferred to see a sliding scale between core and more ‘stretched’ usages.⁵⁹ For many entries, LSJ includes a ‘metaph.’ sense. Methodologically, some might object to making such a clear distinction between a metaphorical and a literal usage in a dictionary at all, although most critics of classical literature still seem to employ a traditional, intuitive distinction between metaphorical and literal uses of language. In light of these developments, LSJ’s approach to the drawing of separate sense-divisions for personifications and for ‘metaph.’ usages seems somewhat arbitrary. A Hesiodic example of the latter is the use of μαυρόω. The LSJ entry (substantially unchanged in all editions) cites Hesiod’s use of the verb at Op. 325 under a secondary sense of ‘metaph., make dim or obscure’; yet the verb is also marked as synonymous with ἀμαυρόω for which the primary sense is given as ‘make dim, faint, or obscure’. The choice to label ‘make faint, obscure’ as a metaphorical, secondary sense for μαυρόω, but a non-metaphorical, primary ⁵⁷ See especially δίκη at Op. 270 2. ⁵⁸ For a recent introduction, see Kövecses 2010. On metaphors in classical literature, see Boys Stones (ed.) 2003 and Silk 1974. ⁵⁹ Note, for instance, Geoffrey Lloyd’s preference for the concept of ‘semantic stretch’, as opposed to a clear distinction between metaphorical and literal uses, outlined in Lloyd 1987, 172 214 and defended in Lloyd 2003.

122

Tom Mackenzie

sense for ἀμαυρόω seems inconsistent. The problem of how a lexicon ought to deal with figurative and non-figurative sense-divisions is a complex one, which will apply to any author, but the particular issue of the division between personified and non-personified usages is especially pertinent to Hesiod. We might at least hope for a consistent and explicit method for dealing with such distinctions in future dictionaries.

7.5. CONCLUDING THO UGHTS Aside from the basic criticisms that Hesiod ought to be cited before the Homeric Hymns, and that it cites editions and fragment numbers that are now obsolete, LSJ is still extremely useful for the student in translating Hesiod. There are remarkably few instances (perhaps most egregiously, in the cases of ἀλϕηστής and ἄπλαστος) where it conflicts with the modern understanding of the Greek. Certain misleading entries seem to arise in part as a consequence of the fact that later editions of Liddell and Scott abridge discussions from earlier editions of problematic passages, and so neglect to mention instances of uncertainty. A further area in which it could be more useful for the student of early Greek poetry is if more information were provided on the use of words in formulaic expressions. Of course, a dictionary the size of LSJ could not hope to be exhaustive in this respect, in the way that LfgrE is, but occasionally, a lack of this information limits the value of the dictionary in deciphering certain passages of Hesiod. Yet, LSJ is used by students and scholars not only to help with translation from Greek authors, but also to trace the history of particular words and concepts. After all, in the preface to the first edition, Liddell and Scott planned ‘to make each Article a History of the usage of the word referred to’.⁶⁰ For this use, LSJ seems more deficient: Hesiod’s usage of particular words, even when he uses them in an idiosyncratic way, or in a manner which offers an explanation for their meaning, is often neglected. This is particularly problematic given Hesiod’s immense influence in the development of the ancient understanding of particular terms and concepts, such as ἔρις and δίκη. Some of these complaints—the lack of citations of Hesiod for relevant lexemes, the lack of descriptions of the formulaic phrases in which Hesiodic terms are found—are simply sins of omission, and it could be objected that the inclusion of the extra relevant information would render any dictionary unwieldy. We should remember Chadwick’s principle that the efficiency of a dictionary is equal to its usefulness divided by its weight:⁶¹ the more ⁶⁰ LS¹, vii.

⁶¹ Articulated at Chadwick 1996, 19.

A Canonical Author: The Case of Hesiod

123

information a dictionary contains, the heavier it will be and therefore the less efficient. However, the weight-variable is becoming less significant now that users are increasingly using online and digital dictionaries, as opposed to hard copies: the days of the hard-copy dictionary, especially for a ‘dead’ language such as ancient Greek, may well be numbered. Moreover, any researcher is now able to discover almost all attested usages of a lexeme in a matter of seconds, thanks to the online ‘TLG’.⁶² In light of this technology, we may be justified in demanding more information from future lexicographical projects. In any case, thanks to these developments, the definitions written and sensedivisions drawn by future lexicographers will be far more easily scrutinized than those of previous generations.⁶³

⁶² http://stephanus.tlg.uci.edu. ⁶³ I am most grateful to Henry Mason for passing this project onto me (after it had initially been offered to him), for sharing some of his notes on Hesiodic vocabulary, and for commenting on an earlier draft of this chapter. Comments from the editors were also invaluable. Remaining deficiencies should be attributed to me alone.

8 Philosophy and Linguistic Authority The Problem of Plato’s Greek Christopher Rowe

To compare the ninth edition of Liddell and Scott (LSJ) with its predecessors in relation to the treatment of Plato is instructive in a number of different ways. The present chapter focuses on two particular aspects of the back history of LSJ. First, Plato is, for LSJ, still one of the most important writers of ‘the best’ Attic prose, as he was, for its predecessors, of ‘correct’ Attic; he tends to be one of the benchmarks for Attic usage. LSJ in this respect continues to reflect the priorities of those who valued the practice of prose composition—perhaps not surprisingly, given that Greek prose composition retained its high status not just in pedagogical practice but, in some quarters, as an art form up until the 1960s and beyond. But secondly, at the same time LSJ typically treats Plato as belonging to a breed apart: that of the Philosopher, with his own technical or semi-technical vocabulary, his own special ‘philosophical’ ways of thinking, as if these were quite separate from those of non-’philosophical’ authors. The chief purpose of the present chapter is to illustrate some of the difficulties attaching to these two approaches, both of which are rooted firmly in earlier editions, stretching back even to the first in 1843.

8.1. PLATO AS L EXICOGRAPHICAL BENCHMARK For Liddell and Scott, throughout much of its history, Plato was either the most important writer of Attic prose, or at least one of the most important. The range of the vocabulary Plato uses is more extensive, and the registers in which he writes more varied, than those of any other Attic writer; it is no surprise, then, that he should be the lexicographer’s darling. Platonic examples would in any case have tended to come either first or at least early in any Christopher Rowe, Philosophy and Linguistic Authority: The Problem of Plato’s Greek In: Liddell and Scott: The History, Methodology, and Languages of the World’s Leading Lexicon of Ancient Greek. Edited by: Christopher Stray, Michael Clarke, and Joshua T. Katz, Oxford University Press (2019). © Oxford University Press. DOI: 10.1093/oso/9780198810803.003.0008

Philosophy and Linguistic Authority: The Problem of Plato’s Greek 125 list illustrating a particular lexical item, given that examples are cited in chronological order. The fact that citations of Plato tend to come in early lends him even more prominence. If one wanted to know how a given word is used, or what it ‘means’,¹ Plato was both by design and by accident an authority one could rely on. This is of course simplistic. No lexicographer, and certainly not Liddell, Scott or their successors, could possibly be blind to the fact that any particular use of a word² will at least partly be determined by context: a reference to a given passage should in principle be no more than an invitation to go and look up that passage and see for oneself how the word is in fact used. But at the same time a lexicon would be failing in its purpose if it did not also make a stab at saying what the nearest equivalent or equivalents of the word in the destination language would be; and that then tends to tie both the lexicographer and the user of the lexicon to a particular reading of the passage or passages referred to. Lexicographers try to mitigate the damaging effects of this inevitable bind by offering more and more sophisticated subdivisions of the ‘meanings’ of words, and more and more examples; and LS7-8 and LSJ³ offer endless examples of this process. With Plato, however, the problems become so acute as to be critical. Plato is one of the most inventive and innovative of writers, capable on occasion of stretching the Greek language in ways comparable to the inventions of the lyric or tragic poets;⁴ and no one would presume to use them as a guide to prose usage⁵ in the way that Liddell and Scott and LSJ inevitably try to use Plato. One should clearly not overstate the difficulties: for much or most of the time Plato’s Greek will actually be as ‘standard’ as that of any (‘correct’?) writer can be. One of the beauties of the dialogues is that they give us an extraordinarily extended exposure to—a written, carefully composed, and cultivated version of—conversational Greek. To that extent, using Plato as benchmark is entirely unexceptionable. But at the same time it is difficult to avoid the fact⁶ that, through his Socrates, Plato is trying to change, and not merely to reflect,

¹ ‘Meaning’ (a slippery term at the best of times) here refers to the ‘dictionary definition’, and nothing more. ² Or, better, ‘lexical item’, since entries are not always for single words; but ‘word’ will do here as shorthand. ³ On the basis of my sampling of the various editions, LS7 8 tend to be closer in the relevant respects to LSJ than any of their predecessors. ⁴ Cf. M. Silk, ‘Pindar meets Plato: theory, language, value, and the classics’, in S.J. Harrison (ed.) Texts, Ideas, and the Classics (Oxford: OUP, 2001), 26 45. ⁵ For the purposes of prose composition (as its few surviving expert practitioners will testify) the distinction between prose and poetic idiom was crucial. ⁶ However hard some, especially literary readers, attempt to avoid it: for a recent, and extreme, example see A.K. Cotton, Platonic Dialogue and the Education of the Reader (Oxford: OUP, 2014: see my review in Clio: A Journal of Literature, History, and the Philosophy of History, published online in August 2017).

126

Christopher Rowe

people’s attitudes, namely the way they think; and changing the way they think also means changing the language they think and talk with—or more strictly, in Plato’s case (as I shall explain in the next paragraph), changing the way they use their language. And to the extent that this is so, he will become a minefield for the lexicographer instead of a treasury. Take one central example:⁷ ἀγαθός, which typically translates into English as ‘good’. Now fourth-century Athenians who read Plato would be likely to think themselves quite at home with the way he uses the term: an ἀγαθός person will in the relevant context be someone who possesses all or some of the virtues, whichever they may be, while an ἀγαθόν thing will be something useful, or good of its kind, or something one is fortunate to have, and so on. Plato’s Socrates, however (to cut a long story short), proceeds first of all to reduce all the virtues to wisdom, and then to make wisdom itself not only the one uncomplicatedly good thing, but the one thing we all truly want, the one thing that makes us truly fortunate. The good person, in brief, will be good at getting what is good for him- or herself. Not everyone will accept this as a statement of what Plato’s Socrates is saying, even in a restricted number of dialogues, but at least most contemporary interpreters of Plato would be prepared to accept that he sometimes says things that are consistent with some version or other of what is contained in the preceding two sentences.⁸ The reader of Plato, faced with this, receives the following guidance from LSJ under ἀγαθός: ‘I. of persons’, ‘1. well born, gentle’ [with a number of examples, one from Plato], ‘2. brave, valiant, since courage was attributed to Chiefs and Nobles, Il.1.131, al.’, ‘3. good, capable, in reference to ability’ [examples include a number from Plato], ‘4. good, in moral sense, first in Thgn., cf. Heraclit.104, S. El.1082, Xen.Mem.1.7.1, Pl. Ap. 41d, etc. [5. . . . , 6. . . . ] II. of things, 1. good, serviceable . . . 2. of outward circumstances . . . , 3. morally good . . . ’ [limited examples given in each case, none Platonic], then ‘4. ἀγαθόν, τό, good, blessing, benefit, of persons or things . . . ’ with a long list of examples, none Platonic, which are then followed by ‘ . . . : τὸ ἀ. or τἀ., the good, Epich.171.5, cf. Pl.R.506b, 508e, Arist. Metaph.1091a31, etc.: n pl., ἀγαθά, τά, goods of fortune, treasures, wealth . . . ; ἀγαθὰ πράττειν, fare well . . . ; also, good things, dainties . . . : good qualities . . . ; good points, of a horse . . . ’.

What we find here under I. may be fine enough for the other authors cited, but to suggest under I.4 that ‘good’ in ‘a good man’ at Plato, Apology 41d (‘nothing bad can happen to a good man whether in life or after he has died, nor are his

⁷ In the interests of brevity I shall restrict myself in this section to three examples from early in the alphabet; there could have been any number of others. ⁸ For the status quaestionis on the ethics of Plato’s Socrates, see, e.g., Thomas C. Brickhouse and Nicholas D. Smith, Socratic Moral Psychology (Cambridge: CUP, 2010), together with my response: ‘Socrates on Reason, Appetite and Passion: A Response to Thomas C. Brickhouse and Nicholas D. Smith, Socratic Moral Psychology’, in The Journal of Ethics 16 (2012): 305 24.

Philosophy and Linguistic Authority: The Problem of Plato’s Greek 127 affairs neglected by the gods’) is the same ‘good’, i.e. ‘moral’ good, as in Theognis, in Heraclitus, in Sophocles, in Xenophon, or in Greek writers generally,⁹ is unhelpful insofar as it closes down the very question Socrates is trying, here as elsewhere, to open up: namely, what goodness actually is.¹⁰ One can understand how the gods can look after a ‘morally’ good person after death, e.g. by transporting him to some or other Isles of the Blest; but given that Socrates is presently counting himself among the good (why else would he mention the good man?), how exactly are they supposed to be looking after him, in his lifetime (‘whether in life . . . ’)?¹¹ The real point of his claim about the good man depends on the identification of goodness with wisdom. The wise man will discount all conventional views about what is bad; for him only ignorance is a truly a bad thing, and since he is wise, nothing bad can possibly affect him either after life or during it. In any case, Ap.41d is not a good choice to illustrate the allegedly standard usage referred to at LSJ s.v. ἀγαθός I.4; nor, I add, would it be easy to find a better passage in Plato for the purpose, unless it is when he is referring to the views of others (perhaps Simonides’ in the Protagoras?).¹² Section II.4 raises further problems. In the string ‘τὸ ἀ. or τἀ., the good, Epich.171.5, cf. Pl.R.506b, 508e, Arist. Metaph.1091a31, etc.’ the reference is to Epicharmus DK 23 B3. This fragment comes from Diogenes Laertius, III.14, where Diogenes is reporting the claims of a certain Alcimus that Platonic thinking about forms owes a debt to Epicharmus. This passage cites twelve extraordinarily Platonic-sounding lines allegedly from Epicharmus, including this: So do you think / it would be like that about the good [τὸ ἀγαθόν] too? That / the good is a certain thing in itself, and that whoever / comes to know that [sc. the good] is already become good . . . ? (lines 4 7)

The fact that LSJ cites Epicharmus before Plato immediately suggests that we are being asked to take this Alcimus seriously, and accept that Plato and his successors may actually have been following this apparently early fifth-century

⁹ As the ‘etc.’ after ‘Thgn., cf. Heraclit.104, S. El.1082, Xen.Mem.1.7.1, Pl. Ap. 41d’ proposes. ¹⁰ I pass over the general question whether the modern category of the ‘moral’ should be translated back into the ancient Greek context. Certainly some of the concerns of ancient Greek writers seem to mimic those that we label as ‘moral’, but it is not so clear that ‘morality’ is what interests Plato, or indeed Aristotle; what they offer, or sketch, is something more neutrally described as a set of ethical views, these being entirely innocent of the kind of ‘morality system’ famously criticized by Bernard Williams (Ethics and the Limits of Philosophy, London: Fontana, 1985). ¹¹ Even if he explicitly denies that his present situation is bad for him, and trusts that his death will actually be good for him, he can hardly want to claim that being on trial for one’s life is a gift from god to be treasured. ¹² It follows that I should want to raise similar queries about the use of Platonic examples to illustrate the use of ἀρετή (s.v.) as ‘moral virtue’.

128

Christopher Rowe

poet (if that is how we are to interpret the ‘cf.’ in ‘cf. Pl.R. . . . ’); which, as most would agree, is quite unlikely. But there is a more serious objection to the entry. It claims to separate off a specific use of the neuter with definite article, as ‘the good’, from the use of the phrase to indicate an individual ‘good, blessing, benefit’ (and then from various uses in the plural); but to go by the examples offered, it actually combines what, by LSJ’s own standards, ought to constitute at least two separate uses. In the second Republic context referred to, 508e, τὸ ἀγαθόν is in the genitive, bracketed in the phrase τὴν τοῦ ἀγαθοῦ ἰδέαν: ‘the good’ being talked about here is the ‘ideal form, archetype’, to cite LSJ’s own interpretation of the phrase s.v. ἰδέα (see section 8.2.7 below).¹³ The Aristotle passage cited also seems to be about the Platonic, or at least Platonist, forms. But the other Republic passage cited, 506b, is quite different: Adimantus is asking Socrates ‘Do you say the good is knowledge, or pleasure? Or something else again?’ Here, clearly, the phrase τὸ ἀγαθόν cannot refer, at least in the first instance, to the form of the good. Socrates cannot be asking ‘is the form of the good knowledge, or pleasure?’ ‘The good’ here is ‘the goal we set ourselves in our lives (vel sim.), sc. whatever it may be’. The context of the passage is one in which Socrates and his partners in the conversation raise the question, which is in principle entirely open, about just what it might be for a thing to be good or beneficial. This is the way Aristotle himself tends to use the expression τὸ ἀγαθόν when developing his own argument, as do other philosophers. How LSJ got itself into this situation may perhaps be explained by its relationship to preceding editions.¹⁴ At the equivalent location in their entry for ἀγαθός, LS1-6 identify a Platonic use of τὸ ἀγαθόν as ‘the highest good, summum bonum, Plat., etc.’; then in LS7-8 ‘the highest good, summum bonum, Plat., etc.’ is replaced by ‘the good, Cicero’s summum bonum, Plat. Rep. 506B, 508E, 534C, al.’ ‘Highest’ is still implicitly retained through the reference to Cicero, but this particular usage is now apparently identified as specifically Platonic, with introduction of ‘al.’ after the three Platonic references in place of LS1-6’s ‘Plat., etc.’. This is odd: it makes it look as if ἡ τοῦ ἀγαθοῦ ἰδέα at Republic 508e is Socrates’ answer to Adimantus’ question, so that the summum bonum is actually the form of the good. In fact, as closer analysis shows, Socrates is only saying that we will only find out what our good is if we study the form (ἰδέα) of the good.¹⁵ The editors of LSJ see the problem, rightly removing the reference to Cicero and the highest good; and that now enables ¹³ To be clear: the ἰδέα here is identical with ‘the good’, not something separate from it; the ἰδέα τοῦ ἀγαθοῦ is the ‘form’ that itself is, and can be referred to as, αὐτὸ τὸ ἀγαθόν. ¹⁴ The idea that there is a ‘moral’ good in Plato specifically goes back to LS⁶, which has ‘[I].4: good, in moral sense, first perhaps in Theogn. 438, but not freq. till the philos. writers, as Plato’ (LS¹: ‘Att. usu. in moral signf., good, virtuous’). ¹⁵ Which of course will still be ‘the highest good’, but in rather different way from that in which Cicero’s summum bonum is summum.

Philosophy and Linguistic Authority: The Problem of Plato’s Greek 129 them to restore the ‘etc.’ from LS1-6 at the end of their list of examples, because what they now appear to be talking about (‘the good’ in the abstract?) is quite a common usage among philosophical authors. But at the same time their actual examples are centred around Platonic/Platonist usage, with pole position allocated to Epicharmus; and even if we granted the authenticity and/or early date of his supposedly Platonizing lines, his treatment of the good as ‘a certain thing in itself ’ already takes us beyond the more general usage of the phrase τὸ ἀγαθόν that would justify that ‘etc.’—one that is actually illustrated by the first of the two references to Plato retained from LS7-8.¹⁶ LS1-6’s ‘the highest good, summum bonum, Plat., etc.’ is at least less misleading. Another smaller example of the mismatch between Plato’s linguistic flexibility and the demands of ‘dictionary definition’: the case of ἀδολεσχία and ἀδολέσχης. LSJ gives three senses for ἀδολεσχία, of which the first two are as follows: [I.] prating, garrulity, Ar Nu.1480, Isoc.13.8, Pl. Tht.195e, Arist.Rh. 1390a9, Thphr. Char. 3 . . . II. keenness, subtlety, Pl.Phdr.269e.

Similarly, while an ἀδολέσχης is a ‘prater, idle talker, esp. of reputed sophists: Σωκράτην, τὸν πτωχὸν ἀ., Eup.352, cf, Ar.Nu.1485’, he is also ‘II. in good sense, subtle reasoner, Pl. Cra. 401b’. LS1-5 quote Heindorf as the authority for the ‘good sense’ of ἀδολέσχης/-ία, and add Parmenides 135d as another illustration of it. Now it is true that the Phaedrus and Cratylus passages use ἀδολέσχης/-ία, paired with μετεωρολόγος/-ία, ‘in a good sense’. But it by no means follows that the ‘meaning’ of the words as used there is exhausted by the English ‘keenness, subtlety’. The origin of the pairing of ἀδολεσχία and μετεωρολογία, and of Plato’s use of ἀδολεσχία itself, surely lies in the kinds of abuse levelled at Socrates himself by the comic poets. That will make their ‘bad sense’ permanently dominant, so that when they are apparently used in a positive way, as in Phaedrus 269e, they combine their acquired ‘good’ meaning with the other, negative one—or, as we might put it, their appearance there is inevitably qualified by scare quotes.¹⁷ Having turned its back on the expositions common in early editions, LSJ probably lacks the resources to express these subtleties. It would have been better, perhaps, if it had just left out the ¹⁶ LSJ s.v. ἀγαθός perhaps also reveals its roots in that to me jarring reference to ‘Chiefs and Nobles’ (‘brave, valiant, since courage was attributed to Chiefs and Nobles’), which first appears in LS⁶ as a replacement for the extraordinary ‘brave . . . [nearly =] noble . . . Compare the Lat. Optimates, French Prudhommes, Saxon Good men, etc., applied to Nobles or Freemen generally, as opposed to the lewd people, base kinds (κακοί ), Welcker Theogn. praef. p. xx1 sq.’ in LS⁴ (in earlier editions, the comparison was restricted to ‘Lat. optimates’). ‘Chiefs and Nobles’ is better than that, but ought surely to have felt outdated even in 1925. ¹⁷ This is what I think Heindorf is actually saying in the passage cited by Liddell and Scott from his Platonis dialogi selecti [Lysis, Charmides, Hippias Major, Phaedrus] of 1802 (Berlin), p. 236, ad Phaedrus 270a; in other words, they are also mistaken in using him as an authority as are those editions that use Parmenides 135d as an example of sense II.

130

Christopher Rowe

Phaedrus and Cratylus passages, and made readers of Plato supply the larger context—stretching from Eupolis and Aristophanes to the μετεωρολόγον, ἀδολέσχην τινὰ σοϕιστήν of Politicus 299b—which will explain how a set of terms with so peculiarly negative connotations could ever come to be used in such a way as (apparently, but only apparently) to lose them altogether. My third example of the difficulty of using Plato as a lexicographical benchmark: the case of βλοσυρός. At Theaetetus 149a, Socrates refers to himself as the son of Phainarete, μαίας μάλα γενναίας τε καὶ βλοσυρᾶς. For βλοσυρός, we find the following in successive editions of the Lexicon: LS¹: ‘awful, awe-inspiring. Hom. only in Il., of the look and mien of heroes: also in Plato manly, noble. II. later, terrible, stern.’ LS²: ‘aweful, awe-inspiring, of Ajax, μειδιόων βλοοσυροῖσι προσώπασι, Il.7.212; βλ. ὀϕρύς, of Hector, Il.15.608:—in Plat., manly, noble, Rep. 535B, Theaet.149A. II. later . . . [with examples].’ LS3-4: ‘grim, awful, of Ajax . . . ; of Hector, . . . ;—later of anything terrible . . . : —in Plat., manly, noble’ . . . :—also coarse, rough, Theophr.’. LS⁵ adds a couple more examples, including ‘so, Nicostr. Incert.4’ after Theaetetus 149a.¹⁸ In LS6-8, however, βλοσυρός at Republic 535b (γενναίους τε καὶ βλοσυροὺς τὰ ἤθη) is rendered as ‘bluff, burly, valiant’,¹⁹ and in the Theaetetus passage as ‘stout, stark’. LSJ, by contrast, goes for ‘hairy, shaggy’ for the two Homeric instances, with a number of supporting examples, then continues as follows: . . . later,²⁰ grim, fearful, ἄγος A.Eu. 167 (lyr.) . . . 2. virile, burly, . . . Pl. R. 535b; β. τὴν ψυχὴν ἔχεις Nicostr.35; of a woman, μαῖα γενναία καὶ β., Pl.Tht. 149a; βλοσυρωτάτη τὸ εἶδος, of Boudicca, D. C. 62.2; also, coarse, πίττα Thphr. . . . 3. solemn, dignified, . . . Aelian . . . ’ We have plenty of choice here: according to which edition we read, Socrates is calling his mother either ‘manly, noble’, or ‘stout, stark’, or ‘masculine’. But all of these alternatives look unappealing, and LSJ’s perhaps particularly so: why exactly should Socrates want to call his mother ‘masculine’? ‘Muscular’, perhaps,²¹ or ‘imposing’, as in LS1-2’s ‘awe-inspiring’? LSJ seems to have ruled out the Homeric connection, having made Ajax and Hector ‘hairy, shaggy’; but the fact that Hesiod uses βλοσυρός of lions (Sc. 175), and of the Κῆρες (ib.250: both passages cited in LSJ) surely allows it back in again. In the

¹⁸ LS⁴ also appends to the Theaetetus reference ‘(ubi v. Heindf.)’ = v. Platonis dialogi duo: Gorgias, Theaetetus (1829), p. 283 ad Theaetetus 149a; Heindorf ’s note adds little except references to Athenaeus and Aelian. ¹⁹ In my own Penguin translation of the Republic (2012), I translate it as ‘enterprising’. ²⁰ I.e., later than the original Homeric examples? ²¹ My own choice in Plato, Theaetetus and Sophist (in the series Cambridge Texts in the History of Philosophy, 2015).

Philosophy and Linguistic Authority: The Problem of Plato’s Greek 131 context,²² the main point seems to be that Phainarete was someone who looked as if she meant business,²³ someone who dealt firmly with her charges. In any case, as its career in Liddell and Scott and LSJ shows, βλοσυρός here in the Theaetetus is a peculiarly suggestive and elusive term, which defies any attempt to tie it down to a single ‘sense’ or ‘meaning’. Predominantly a poetic word, it is an appropriately striking choice for a surprising turn in the dialogue—‘Did you know my mother was a midwife? And that I am a midwife too?’²⁴

8.2. PLATO AS PHILOSOPHER (A BREED APART) In this section I propose to discuss Liddell and Scott/LSJ’s handling of a small number of terms, each of which is central to Plato’s thinking and to the business of philosophy as he sees it: αὐτός, γένος, διαίρεσις, διαλέγομαι, εἶδος, εἰρωνεία, ἰδέα, μέθεξις, ὁρίζω, ψυχή. I shall make individual comments on each item, and then draw those comments together in my concluding paragraphs in order to characterize the lexicon’s treatment overall of Plato as Philosopher.²⁵

8.2.1. αὐτός LSJ s.v. αὐτός, I.3: ‘by oneself or itself, alone . . . ’; I.4: ‘in Philosophy, by or in itself, of an abstract concept or idea, δίκαιον αὐτό Pl.Phd.65d; αὐτὸ τὸ ἕν Id. Prm.143a, al., cf. Arist. Metaph.997b8: neut. αὐτό is freq. in this sense, attached to Nouns of all genders . . . ’. LS7-8 are similar, but with the important difference that I.4 there begins ‘Plato used αὐτός to signify a thing by itself . . . ’. Section I.4 in LS5-6 is essentially concerned with the grammatical point: ‘Plato uses αὐτό with substantives of all genders, to signify a thing by or in itself ’, referring to Republic 363a (‘ubi v. plura apud Stallb.’) together with two contexts in the Symposium, and continuing ‘hence, later [specified in V.6 as from Aristotle onwards], not only αὐτοαγαθόν, etc., but αὐτοάνθρωπος . . . ’. ²² And context is all: thus I have no inclination to repent of my rendering of βλοσυρός, similarly paired with γενναῖος, in the Republic passage (see n. 19 above); though I would also guess that the use of the same pairing was deliberate, and deliberately provocative. ²³ Like Socrates e.g., as portrayed by Alcibiades, in Plato’s Symposium, in the retreat from Delium? But anyway like Ajax, Hector (and, later, Boadicea). ²⁴ I should stress that while this first section of the present chapter has been aimed specifically at Liddell and Scott and LSJ, its general argument applies equally, and in principle, to any attempt to confine Plato within the pages of a dictionary. LSJ, in particular, remains a massive achieve ment, and an absolutely indispensable tool. The following section, by contrast, will be directed exclusively at Liddell and Scott and LSJ themselves. ²⁵ I retain the capitalization of ‘philosopher’ because, as we shall see, the various editions tend to treat this aspect of Plato as a thing apart; I shall argue that this is a mistake.

132

Christopher Rowe

LS1-4 do not refer even to the grammatical point, picking up on the ‘later use’ of αὐτοαγαθόν etc in a separate section on αὐτός ‘in Compos.’ LS1-8 all conclude with the acknowledgement that ‘the chief authority’ for their entries is ‘Hermann, Dissertatio de Pron./pron autos’, so Hermann was evidently not interested in what the Lexicon identifies as the Platonic usage (‘to signify a thing by itself ’); LSJ is the first edition to make that usage ‘philosophical’ rather than Platonic. The reference to Aristotle, first introduced in LS7-8 I.4, actually concerns Platonists, and forms/ideas.

8.2.2. γένος The entry develops as follows: LS¹: V ‘kind, genus, opp. to εἶδος, species . . . ’ LS2-4: V ‘a class, sort, genus, opp. to εἶδος (species), Plat. Parm. 129C, etc., Arist. Top. . . . ’ LS5-6: V ‘a class, sort, kind . . . :—in Logic, opp. to εἶδος (species) . . . ’ LS7-8: V.2 ‘in Logic, . . . ’, adding a reference to Aristotle, Metaphysics, then ‘but a γένος may become an εἶδος to a more comprehensive γένος, and vice versa . . . ’ [i.e. in Aristotle] Intermediate Greek Lexicon: V.2 ‘in Logic, genus, opp. to εἶδος (species), Plat.’; LSJ V.2 ‘in Logic, opp. εἶδος (species), Pl. Prm.129c, al., Arist. Top.102a31, 102b12, al., [as in earlier editions, then] τὰ γ. εἰς εἴδη πλείω καὶ διαϕέροντα διαιρεῖται Id. Metaph. 1059b36’, like LS7-8 Earlier editions also assimilate Plato’s use of γένος (‘in Logic’) to Aristotle’s; LSJ omits the references in 7–8 to the variations in the relationship between γένος and εἶδος that occur even in Aristotle.

8.2.3. διαίρεσις LS¹ does not refer to Plato in its entry for διαίρεσις: ‘ . . . esp. in Logic, division of a class into its constituent parts, Arist.Org.’). LS² has ‘III. In Logic, division of genus into species, Plat. Soph.267D, Arist. Anal. Pr.I.31; opp. to συναγωγή, Plat. Phaedr. 266B’. This is more or less what we find under διαίρεσις in all later editions, including LSJ.

8.2.4. διαλέγομαι LSJ (s.v. διαλέγω B.2) has ‘in Philosophy, practise dialectic, elicit conclusions by discussion, οὐκ ἐρίζειν ἀλλὰ δ. Pl. R. 454a, cf.511c, Tht. 167e, etc.’. LS1-8 all

Philosophy and Linguistic Authority: The Problem of Plato’s Greek 133 refer—under ‘discourse, argue’, with an ‘esp.’—to ‘the dialectic method of the Socratics’, explaining ‘where the conclusions were not drawn absolutely (or “directly”) from known premisses (or “by the speaker”), but elicited by questions (or “discussion”)’, the original source of this explanation being cited as ‘Heind. Plat. Phaedo 84C’.

8.2.5. εἶδος The following are the principal data: LSJ, s.v. εἶδος III: ‘[1.] class, kind, Isoc, cf. D[em.]: freq. in Pl. [three examples from Tht.], etc.; logical species, Sph. 235d; . . . Plt. 262e . . . , 285b, al., cf. Arist. Metaph. 1057b7, al., Cat. 2b7 . . . 2. = II.2 [see in section 8.2.7 below], Pl.Phd.103e, R.596a, Prm.132a, al., Arist. Metaph.990b9, al., etc.’. LS¹, II: ‘in gen. a form, figure, fashion, sort, particular kind [examples from Hdt., Thuc., etc.]: esp. species, opp. to γένος, genus, hence also = ἰδέα, Plat., and Arist., cf. Ritter Hist. of Philos. 2.265, sqq.’. LS2(-4): ‘II. generally, a form, species, sort, particular kind . . . III. in Logic, species, opp. to γένος, genus, hence also = ἰδέα, Plat. Parm.129C, etc., and freq. in Arist.; cf. Ritter . . . ’. LS⁵: ‘III. in Plat., a general form, imperfectly represented in particular individuals, more commonly called ἰδέα, Phaed. 103C, Parm. 129C, etc.; τὸ ἐπ’ εἴδει καλόν ideal beauty, Symp. 210B. IV. In the Logic of Arist., species, opp. to γένος, genus’. LS⁶: III = LS⁵, with the following added: ‘2. εἶδος ἔχειν τινός Arist. Pol. 3.15, 2:—more generally, a class, species, [three examples from Pl., Tht., as in LSJ], etc.:—hence a logical form or species, Plat. Soph.246C, Polit.262E, 285B, etc. v. Grote Plat. 2. pp. 467 sqq.; adopted and more precisely defined in the Logic of Arist., v. Categ. 45.’ LS7-8: ‘III. a class, kind, sort, whether genus or species, [the same three examples from the Theaetetus], etc.:—a logical form or species . . . ’ This is as in LS⁶, but then there is a new item: ‘2. in Plat. εἴδη were often used = ἰδέαι (v. ἰδέα II.2), Phaedo 103E, Rep.597A, Parm.132D, etc., cf. Arist. Metaph.I.6, 3 sq., al; τὸ ἐπ’ εἴδει καλόν . . . ’. Intermediate Greek Lexicon: ‘3. a class, kind, sort, whether genus or species, Plat., etc.’. Particularly interesting here is the way LSJ drops the ‘whether genus or species’ that we find in LS7-8 III.1, and the ‘cf.’ before the reference to Aristotle, Metaphysics in LS7-8 III.2: see my general comments below.

134

Christopher Rowe

8.2.6. εἰρωνεία LSJ begins as follows: ‘[I.] dissimulation, i.e. ignorance purposely affected to provoke or confound an antagonist, a mode of argument used by Socrates against the Sophists, Pl. R.337a, cf. Arist. EN 1124b30, Cic. Acad.Pr. 2.5.15 . . . ’

This is similar to what we find in LS¹: ‘dissimulation, esp. ignorance purposely affected . . . an antagonist, irony, used esp. by Socrates’. The ‘i.e.’ in LSJ is strange, insofar as there are other kinds of dissimulation than the one attributed to Socrates (in Plato too: see Ast, Lexicon Platonicum s.v. εἰρωνεύομαι). ‘Esp.’, as in LS¹, thus seems preferable, though ‘e.g.’ would be even better. LS¹’s second ‘esp.’, however, is just as strange, since it is hard to find anyone other than Socrates who is supposed to deploy this ‘mode of argument’. Indeed, it is not clear that even Socrates does: Republic 337a simply has Thrasymachus accusing Socrates of ‘purposely affecting’, etc., and that Socrates does anything of the sort, even in Republic I, is hotly disputed. In short, none of the editions of Liddell and Scott/LSJ does very well by εἰρωνεία in Plato.

8.2.7. ἰδέα In LS1-3 we find the following in the entry for this word: ‘3. a nature, species, kind, sort . . . a way, manner . . . 4. = εἶδος, a class, species, under which individuals or smaller species are ranged: an idea or general principle for such classification; but in the Platonic Philosophy the ἰδέαι were not only εἴδη, but something more, viz. the perfect archetypes, models or patterns (Lat. formae), of which, respectively, all created things were the imperfect anti types or repre sentations, v. esp. Rep. 596 sq.; εἶδος therefore might be used for ἰδέα, but not ἰδέα for εἶδος, Stallb. . . . , but v. Ritter . . . ’.

LS⁴ adds ‘Arist. Metaph. 6.14, 13.1, sqq., Eth.N.I.6’ after ‘Rep. 596 sq.’, and omits the reference to Stallbaum and Ritter. LS5-6, for ‘4. = εἶδος,’ etc., have ‘II. in Logic, = εἶδος . . . ’; LS6-8, after ‘ . . . representations’ add ‘the eternal forms of Being, opp. to their material forms, subjects of thought, but not of sight’, then continuing ‘τὰς . . . ἰδέας νοεῖσθαι μέν, ὁρᾶσθαι δ’ οὔ Rep. 507B, cf. 508E, and esp. 596 sq.; cf. also Arist. Metaph. 1.6, 3., 6,14. 12.10., al., Eth.N.1.6:– εἶδος therefore might be used . . . ’ LSJ then shortens the entry from LS7-8: ‘3. kind, sort . . . II. in Logic, = εἶδος, class, kind hence, principle of classification [three Platonic examples]. 2. pl. in Platonic Philosophy, ideal forms, archetypes, . . . Id. R. 507b, cf. 596b, al., Arist. Metaph. 990a34, al., EN 1096a17: also in sg., . . . Pl R. 508e, al., cf. εἶδος.’

Philosophy and Linguistic Authority: The Problem of Plato’s Greek 135 But it adds its own twist, when it suggests that ἰδέα as ‘archetype’ (etc.) appears in Plato mainly in the plural.

8.2.8. μέθεξις The entry evolved as follows: LS¹: ‘participation, Plat. Soph. 256A: τινὶ τινός with another in a thing, Id. Parm.132D’. LS2-5: ‘participation, Plat.Soph. 256A : esp. of the communication between the εἴδη (ἰδέαι) and earthly objects, Plat. Parm.132D, cf. Arist. Metaph.1.6, 3’ LS6-8: ‘[I.] participation, ταὐτοῦ of or in the same, Plat. Soph.256A; μ. οὐσίας Id. Parm.151D; χρόνου Ib. 141D . . . II. in the Platonic philosophy, participation in the ideas, ἡ μ. τοῖς ἄλλοις τῶν εἰδῶν the participation in the ideas by the earthly objects, Plat. Parm.132D, cf. Arist. Metaph.1.6, 3 . . . ’ LSJ: ‘[I.] participation, οὐσίας μετὰ χρόνου participation of being in time, Pl. Prm. 151e; χρόνου in time, ib.141d . . . II. in Platonic philosophy, participation in the ideas, ἡ μ. τοῖς ἄλλοις . . . τῶν εἰδῶν Pl. Prm. 132d, cf. Arist. Metaph. 987b10; ταὐτοῦ in the same, Pl.Sph. 256b.’

The notion of ‘[the] Platonic philosophy’ in LS6-8 and LSJ as apparently excluding ‘sense’ (?) I clearly requires comment.

8.2.9. ὁρίζω The interest here lies in the fact that LS1-8 all specify that when ὁρίζω (‘mainly mid.’) in Plato = ‘define’, it is a matter of defining a word; LSJ avoids this singular error, and firmly says ‘define a thing’ (III.2, IV.3).

8.2.10. ψυχή LS7-8 divide ψυχή as follows: ‘[I.] breath . . . life . . . II. in Hom., the life or spirit of man which survives after death and dwells in Hades, the departed soul, spirit, ghost . . . 2. the abstract notion of the soul or spirit of man, Lat. anima, first in the Physical Philosophy . . . , [as in] Arist. De An. I.2 . . . 3. The ψυχή was²⁶ the seat of θυμός, i.e. of the will, desires, and passions, the soul, heart . . . 4. sensual desire, propension, appetite . . . 5. used sometimes also of animals . . . III. as the organ of νοῦς, i.e. of thought and judgment, the soul, mind, reason, understanding . . . ; freq. in Plat. . . . IV. the ²⁶ A misprint for ‘as’?

136

Christopher Rowe

vital principle, defined by Arist . . . 2. In the most ancient Philosophers, the anima mundi or animating spirit of the Universe . . . ’

Compare and contrast LSJ: ‘[I.] life . . . II. in Hom., departed spirit . . . III. the immaterial and immortal soul . . . IV. the conscious self or personality as centre of emotions, desires and affec tions . . . 2. of various aspects of the self . . . 3. of the emotional self. 4. of the moral and intellectual self . . . 5. of animals . . . 6. of inanimate things . . . V. Philosophical uses: 1. In the early physicists, of the primary substance, the source of life and consciousness . . . 2. The spirit of the universe . . . 3. In Pl. the immaterial principle of movement and life [where Platonic references eventually give way to refer ences to Aristotle, Stoics, Epicureans, and Neoplatonists].’

This is a fairly complete reorganization of the structure of LS7-8, which itself goes back, in essence, all the way to LS¹. The case of ψυχή offers a particularly extreme example of one of the oddest features of Liddell and Scott/LSJ throughout: the tendency to separate off certain uses as specifically ‘philosophical’. It happens, in the case of ψυχή, to be particularly noticeable in LSJ (‘V. Philosophical uses’), with traces in earlier editions (‘the Physical Philosophy’, ‘the most ancient Philosophers’), but in the other cases, and indeed in general, it is common across all editions, as even the small selection of entries above, from αὐτός to ψυχή, tends to confirm: sections beginning ‘in Philosophy’, ‘in Logic’, or ‘in Platonic philosophy’ are ubiquitous. The case of ψυχή also shows with particular clarity why this can be a problem, at any rate for Plato. Plato, after all, inevitably figures prominently under both ‘III. the immaterial and immortal soul’ and ‘IV. the conscious self or personality as centre of emotions, desires and affections . . . 2. of various aspects of the self . . . 3. of the emotional self . . . 4. of the moral and intellectual self . . . ’, as well as under ‘V. Philosophical uses’; and ‘the philosophical uses’ of ψυχή in Plato are prima facie actually inseparable from the allegedly non-philosophical ones. The intention was, presumably, to separate off the exclusively philosophical uses of terms, i.e. those uses that appear to lie outside and beyond ‘ordinary’ literary usage. The entry for ϕιλόσοϕος that survives more or less unchanged throughout LS1-8 is illuminating: [T]he pecul. signf. philosopher, i.e. one who speculates on the nature of things, man, freedom, truth (ὁ τῆς ἀληθείας ϕιλοθεάμων, Plat. Rep. 475E; ἡ ϕ. ἐπιστήμη τῆς ἀληθείας Arist. Metaph. I min. 1, 5., 3. 3, 1) first came into general use with the various philosophical schools, from which time ϕιλόσοϕος is a philosopher of the schools, one who teaches science, etc., according to his own system.²⁷

LSJ offers no more than ‘philosopher, i.e. one who speculates, οἱ ἀληθινοὶ ϕιλόσοϕοι, defined as οἱ τῆς ἀληθείας ϕιλοθεάμονες, Pl.R.475e’, but this is ²⁷ Cited from LS¹.

Philosophy and Linguistic Authority: The Problem of Plato’s Greek 137 evidently just a matter of abbreviation, not of a change of approach. The assumption is that all the ‘philosophers’, including Plato, have their own ‘systems’, and it is the policy to corral them, together with their specialist vocabularies, in their own separate compartment. This approach works well with Aristotle, or the Stoics, or the Epicureans (the latter two schools being given due attention for the first time in LSJ), but it works much less well with Plato. Aristotle and the Hellenistic schools selfconsciously develop their own technical vocabularies. Plato is different. Variatio is the norm with him, even when he is apparently involved in highly technical discussions. Thus, pace Liddell and Scott, εἶδος and ἰδέα are used quite interchangeably for Platonic ‘forms’ or ‘ideas’; and Plato has a range of terms for ‘sort’, ‘kind’, class’, etc.—not just εἶδος, ἰδέα, and γένος but ϕῦλον, ϕύσις. He can use the term μέθεξις of the relationship between particulars and forms (he has other ways of expressing that relationship too), but he can also use it of the relationships between kinds (or forms). He sometimes uses the term διαίρεσις in connection with a process resembling that of the division of genera into species, but he also treats it—along with its counterpart, συναγωγή—as what enables us to sort out the relationships between things in general, whether kinds or particulars. Even when διαίρεσις does appear to be of genus into species, γένος is sometimes used of what is being divided, and εἴδη are sometimes what it is being divided into, but not always; there is no basis for the general claim that γένος in Plato is genus, εἶδος species.²⁸ It is also my view that Plato makes no clear distinction between εἶδος and ἰδέα, as used of what we call Platonic forms, and the εἴδη and ἰδέαι (and γένη) that appear in the Theaetetus, Sophist, and Politicus.²⁹ This is, to say the least, a controversial view: the standard position is, or has been, that Platonic forms as such figure in none of these three dialogues.³⁰ Most contemporary students of Plato will not, then, object to the fact that LSJ reflects a version of this position, insofar as neither it nor Liddell and Scott, in any edition ever—so far as I know—cites a single passage from any of these three dialogues to illustrate the use of εἶδος and ἰδέα for a Platonic form, despite the fact that they teem with examples of both terms. But what anyone now would be likely to find disconcerting is Liddell and Scott/LSJ’s more or less unswerving allegiance to a particularly one-sided version of ‘the Platonic philosophy’, in which the ἰδέαι, or εἴδη (LSJ abandons the claim found in Liddell and Scott that εἶδος cannot be substituted for ἰδέα in this role) are to be regarded exclusively as archetypes ‘of which, respectively, all created things were the imperfect anti-types or ²⁸ For the interchangeability of γένος and εἶδος (and ἰδέα) in such contexts, see, e.g. Sophist 253d e. ²⁹ See, e.g., the Introduction to Plato, Theaetetus and Sophist (n. 21 above). ³⁰ Except, of course, in the argument with the ‘friends of the forms (εἴδη)’ in Sophist 246a 248a. (I note in passing that this use of εἶδος is distinguished from others in the Sophist only by interpreters, not by the speakers in the dialogue.)

138

Christopher Rowe

representations, and were conceived of as the eternal forms of Being, opp. to their material forms, subjects of thought, but not of sight’. This is LS⁸; LSJ drops all of it, and restricts itself to ‘ideal forms, archetypes’ for ἰδέα (‘pl. in Platonic Philosophy’, though ‘also in sg.’), but the introduction of ‘ideal forms’ alongside ‘archetypes’ in itself suggests the contrast with something like LS⁸’s ‘material forms’, and in general there is little sign that LSJ is proposing to diverge radically from its predecessors. But Plato does not only treat his forms as archetypes. This will be true regardless of how we decide to treat ἰδέαι, εἴδη, and γένη in Theaetetus, Sophist, and Politicus; indeed it is implicit in the various different metaphors Plato uses for the relationship between forms and particulars: likeness, participation/sharing in, presence.³¹ A Platonic form will sometimes appear as an archetype, but it will often appear as something else: something that will usually be ill-defined, in the context, but will certainly not be (just?) an archetype. All this is no more than the consequence of treating Plato as just another ‘philosopher of the schools’. To do so is a misapprehension, as I think would now be widely recognized. In truth, Plato becomes a ‘philosopher of the schools’ only after he is dead. Until then, his thought is as elusive as it is exploratory: like his Socrates, he has certain general ideas he is sure of; sure enough to want to persuade us of them, even as he tries to pin them down in detail and find ever better arguments for them. Because he writes to persuade as well as to explore,³² he uses a literary style, accessible for the most part insofar as it employs ordinary language, appropriating it, stretching it, but rarely leaving its more usual connotations entirely behind. Plato lacks a technical vocabulary, then, both in practice and in principle: to adopt one would restrict his audience in a way that he rarely shows the desire to do, even in the later dialogues. It is Aristotle, above all, who turns Plato into a school philosopher—like himself. And it is to Aristotle that Liddell and Scott/LSJ repeatedly turn for supporting evidence about the supposed technical vocabulary of ‘the Platonic philosophy’. Take the case of παράδειγμα. LSJ has: ‘pattern, model: of an architect’s model of a building . . . ; a sculptor’s or painter’s model, Pl.Ti. 28e, R.500e . . . ; of the divine exemplars after which earthly things are made, ἐν οὐρανῷ ἴσως π. ἀνάκειται Pl.R.592b; of the Platonic ideas, opp. εἰκών, Arist.Metaph.991a21, 1013a27 . . . ’.

LSJ does not claim that Plato himself uses παράδειγμα for forms (as indeed he does not); but the plural ‘exemplars’, followed by an example with παράδειγμα

³¹ I.e., the presence of forms in particulars. ³² Not always both in the same measure; but the general point still holds.

Philosophy and Linguistic Authority: The Problem of Plato’s Greek 139 in the singular, reflects the fact that from the perspective of Liddell and Scott/ LSJ in general παράδειγμα would be a natural term for Plato to have used. Hence LS7-8’s ‘ . . . Pl.R.500b; so [my italics] in Arist. of the Platonic ideas’; LSJ may drop the ‘so’, but the sub-text remains the same. The same phenomenon is visible in the treatment of most of the more important items discussed above: ἀγαθός (‘cf. Pl.R.506b, 508e, Arist. Metaph.1091a31’, on the special use of τὸ ἀγαθόν), αὐτός (‘αὐτὸ τὸ ἕν Pl. Prm.143a al., cf. Arist.Metaph.997b8’), διαίρεσις (‘in Logic, division of genus into species, . . . Pl.Sph.267b, . . . Arist. APr.46a31’), εἶδος (‘= ἰδέα II.2, Pl.Phd.103e, R.596a, Prm.132a, al., Arist. Metaph.990b9, al.’), ἰδέα (‘pl. in Platonic Philosophy, ideal forms, archetypes, . . . [Pl.] R. 507b, cf. 596b, al., Arist. Metaph. 990a34, al., EN 1096a17’), μέθεξις (‘in Platonic philosophy, participation in the ideas, ἡ μ. τοῖς ἄλλοις . . . τῶν εἰδῶν Pl. Prm.132d, cf. Arist. Metaph.987b10’). In some of these cases—notably those of διαίρεσις, γένος and εἶδος—the process goes a stage further: Plato is actually assimilated to Aristotle. It is an open question whether this has anything to do with the affiliations of those on whose contributions the whole enterprise of Liddell and Scott/LSJ depended.³³ But the plain fact is that Plato is a quite different kind of writer from Aristotle. He is a writer who can apologize for introducing a novel term, even in a highly technical context, as LS6-8/LSJ note he does at Theaetetus 182a: having introduced the word ποιότης, Socrates says to the mathematician Theodorus ‘well, perhaps “quality” strikes you as a strange word, and it isn’t helping your understanding, either, to have the point put collectively, so here are some individual examples . . . ’. Theodorus has intimated³⁴ that he is himself not entirely at home with the sorts of arguments that interest Socrates; he may well represent the kind of audience—intelligent, but non-specialist—for whom Plato was chiefly writing. This tendency to read Plato through Aristotle looks distinctly odd, at any rate from a contemporary point of view. Even the most superficial reading of Aristotle’s frequent engagements with Plato’s ideas and arguments—in the Metaphysics, say, or in the Politics—will show at once that Aristotle is much more concerned with debating with his teacher than he is with getting him right. That may partly be a matter of choice, but surely also has to do with the fact that Plato makes himself so hard to pin down.³⁵ Nevertheless Aristotle has a pretty clear idea of what he thinks Plato was about, whether in metaphysics, ethics, politics, or

³³ Thus it may or may not be important that the main advisor on philosophical vocabulary for LSJ appears to have been the Aristotelian J.L. Stocks. ³⁴ Theaetetus 165a. ³⁵ Especially, of course, since he writes in the form of conversations in none of which is he an actual participant.

140

Christopher Rowe

whatever area it may be; and that is what forms the basis for his own arguments with Plato, and begins Plato’s transformation into one of Liddell and Scott/LSJ’s ‘philosophers of the schools’. If the one-sidedness of the lexicon’s Plato resembles Aristotle’s, that is because in many important respects the Lexicon’s Plato is derived directly, and deliberately, from the Plato of Aristotle.

9 Medical Vocabulary, with Especial Reference to the Hippocratic Corpus Elizabeth Craik

The first three volumes of Littré’s magisterial ten-volume edition of the Hippocratic works (Littré 1839–61) appeared in successive years: in 1839, 1840, and 1841.¹ Soon afterwards, in 1843, before Littré’s work had made its mark, came the first edition of our Lexicon, Liddell and Scott. Even in the standard ninth edition (LSJ), with its main preface dated 1925, postscript 1940, little reference is made to Littré; by contrast the works of Galen are cited with reference to Kühn’s twenty-volume edition (Kühn 1821–30) and to later editions of certain texts, where available. In the list of Authors and Works, Hippocrates is designated Hippocrates Medicus [Hp.] and the works, all uniformly and rather uncritically dated to the fifth century (‘V B.C.’), are listed in a clear but somewhat misleading sequence. There, used tout court, the titles Morb., Mul., and Vict. are used to denote respectively the totally distinct four books of On Diseases, two of the three books of the rambling compilation On Diseases of Women (the third being separately listed as Steril.), and the four well-integrated books of the organic composition On Regimen; at the same time such composite works as On Fractures–On Joints and On Generation–On the Nature of the Child are separately designated as Art. and Fract., Genit. and Nat. Pue.² We must be grateful that for the ninth edition there was a new cognizance of the need to incorporate, in revision and amplification, ‘technical’ vocabulary from such ‘scientific’ fields as Medicine (with the advice of E.T. Withington), Mathematics (with the advice of Sir Thomas Heath), and Natural History

¹ I gratefully acknowledge the generous editorial input of Christopher Stray. ² It must be conceded, however, that there is still no consensus on the best way to list and enumerate Hippocratic texts; for different approaches, see Anastassiou and Irmer 1997 2012; Craik 2015. Elizabeth Craik, Medical Vocabulary, with Especial Reference to the Hippocratic Corpus In: Liddell and Scott: The History, Methodology, and Languages of the World’s Leading Lexicon of Ancient Greek. Edited by: Christopher Stray, Michael Clarke, and Joshua T. Katz, Oxford University Press (2019). © Oxford University Press. DOI: 10.1093/oso/9780198810803.003.0009

142

Elizabeth Craik

(with the advice of D’Arcy Thompson). Meantime, Withington collaborated with W.H.S. Jones on the series of Hippocratic texts published in the Loeb Classical Library from 1923 onwards. The Supplement of 1968 and Revised Supplement of 1996 made some important additions and revisions, but these were not primarily on medical material. Rather, much attention was paid to matter gleaned from papyri. The study of medical texts, and to a limited extent the study of medical vocabulary, has attracted much interest recently, and can be viewed as an academic growth area. The proliferation of new subject areas—women’s studies, sociology, environmental issues, ethnography—which impinge on issues of human health and well-being has caused scholars to look anew at the rich source material to be found in medical writings. In the special field of papyrology, recent finds have contributed much new medical material for study. The situation has changed greatly since the first edition of the Lexicon in 1843 and even since the ninth in 1925–40: ancient medicine, once viewed as the esoteric pursuit of eccentrics or the province of retired physicians, has become a mainstream area of academic specialization. There is now a large and growing constituency of young, and not so young, scholars who need a lexicon to consult for aid in reading these unfamiliar and frequently difficult texts. A word of warning to the Hippocratic novice: do not waste time searching in the abridged Intermediate Greek Lexicon (1889 and reprints), as in this field it is skimpy at best, and deficient even in essentials. This specialized area of the lexicographical tradition has a long history. The early emergence of Hippocratic lexicography is an aspect of the weight already attached to writings regarded as Hippocratic in an era soon after the lifetime of their putative author, the historical Hippocrates. Hippocrates came to be viewed as a prose author with authority parallel to that of Homer in verse. Erotian (1st century CE), whose Hippocratic lexicon is the earliest extant, gives much information about the activity of his predecessors among the grammarians of Alexandria. Galen too, writing about a century after Erotian, compiled a collection of Hippocratic glosses. In the same era of general retrospection, Rufus collected and explained medical terms and Pollux followed, with a medical section in his larger lexicon. In subsequent scholarship, medical language attracted much attention from scholiasts and commentators. The great lexicographer Hesychius contributes much to our understanding of medical terms, in which he seems to have had a particular interest. Also, it is important to remember that, in addition to a complete Hippocratic edition, the sixteenth-century physician Foesius produced a valuable work of explanation and annotation of Hippocratic language, based on a combination of medical expertise and prodigious philological acumen, with immense mastery of the texts (Foesius 1588). In the nineteenth century, Greenhill pioneered philological study of Hippocratic texts, as an adjunct to medical comment

Medical Vocabulary and the Hippocratic Corpus

143

upon them (Greenhill 1872).³ In 1881 H.G. Liddell sent a postcard to ‘Dr Greenhill, Hastings’ seeking enlightenment on the ‘English equivalents’ of six terms, five from Soranus and one from Galen. These include, from Soranus, πρόρρηγμα ‘foetal membrane’ and ῥωχμός ‘wheezing’. In the seventh edition, whose Preface is dated 1882, all six words appear for the first time, and all save one—a philosophical rather than a medical term, an adjective readily related to the substantive ‘hyposyllogism’—survive without major change in the ninth edition. Greenhill is not the only medical éminence grise who can be traced in the Lexicon. With him may be counted Donnegan (J., M.D.) who planned to include among words in his dictionary ‘a considerable number from the works of the Greek Physicians, expressly admitted to render a service to Medical Students’; that he did so with some success may be seen in his gloss on ἄγγος ‘a vase, an urn or any vessel—a cavity of the human body, as the womb, etc. a blood vessel’.⁴ Perhaps specialists’ interest in the vocabulary of medicine has waned in recent decades. However, some important studies have appeared, among them those of van Brock, Skoda, and Langslow (van Brock 1961; Skoda 1988; Langslow 2000). Editors and commentators too have contributed much incidentally in the course of wider works. Finally, modern means of compilation, distant from the ‘slips’ of past generations and facilitated by all the appurtenances of computerized search, have revolutionized the field, with provision of a full Hippocratic concordance as well as a lexicon. In turning to Liddell–Scott–Jones (hereafter LSJ) we attempt to assess the accuracy and usefulness of the Lexicon with regard to the translation and interpretation of vocabulary arranged in categories according to these main subject areas: Anatomy; Physiology; Pathology; Nosology; Diet and Regimen; Surgery; Gynaecology and Embryology; Scientific Theory; Ideas and Ideals. In the course of this discussion, which is perforce highly selective, particular attention will be paid to entries on some key terms and concepts. The approach followed is pragmatic, rather than lexicographically principled: the aim is not to present a detailed critique of LSJ’s editorial practice or to suggest alternative arrangements, merely to isolate and discuss some questions relating to medical vocabulary.⁵ As many of these comments may seem to have a critical tenor, it is proper to stress at the outset the obvious truth that such quibbles in no way minimize the huge debt owed by all Greek scholars, of all areas of specialization, to our monumental Lexicon.

³ Greenhill’s 100 page book collected thirteen ‘Adversaria medico philogica’ contributed between October 1864 and October 1872 to the British and Foreign Medico chirurgical Review, the forerunner of the British Medical Journal. ⁴ I owe valuable information on these sources to Christopher Stray. ⁵ This contrasts with the more philological approach of Chadwick (Chadwick 1996).

144

Elizabeth Craik

Liddell and Scott initially (LSJ Preface, 1925) followed the dictate of Passow that ‘citations should be chronologically arranged in order to exhibit the history of each word and its uses’. It followed from this historical principle that epic usage received particularly full treatment and this salient policy set out at the inception of the work has remained a constant feature. Homer’s medical, especially anatomical, knowledge was considerable and celebrated already in antiquity, and very few important anatomical terms postdate Homer for first attestation: as a result, the Lexicon’s treatment of anatomy and physiology is on the whole markedly successful. However, while Homer’s detailed descriptions of war wounds display a very good grasp of bone structure and bodily organs, there is a less good knowledge of bodily makeup and tissues. And naturally the other topics addressed here—pathology, surgery, and so on—are less prominent in epic. The advances of medical thought, with new approaches and new insights, brought linguistic changes in their turn. These changes are not confined to medical texts, but pervade writing in other genres also, notably tragedy. The historical principle, which entails the general placing of ‘Hippocrates’ after certain authors and before others, conspires with the ‘technical’ principle, whereby language viewed as medical, even though it is not exclusively so, is prefaced ‘Medic.’; the combined effect is to conceal the interpenetration between different areas of language use, and this often detracts from the Lexicon’s utility and accuracy.⁶ It may be added that medical terms from different authors of different dates are somewhat uncritically listed together, without allowance for change over the centuries between Hippocrates and Paul of Aegina, or even between Hippocrates and Galen. In LSJ, the word ϕλέψ is properly translated and explained ‘blood-vessel, whether vein or artery’ and reference is properly made to Il. 13. 546: a hero dies when the single ϕλέψ said to travel right up the back and through the neck is severed. However, this, the sole occurrence in Homer, is problematical, as no blood vessel corresponds in its course to that described. There is, however, a corresponding channel: the spinal cord, carrying cerebro-spinal fluid, and allusion is surely to this. Here as in later Greek the sense of ϕλέψ is simply ‘vessel’ (carrying fluid, but not necessarily blood), ‘duct’ or ‘channel’. Thus, as noted in the Lexicon, the term can denote the ureters. One more point may be made about the entry on ϕλέψ. The term ϕλέψ κοιλή, translated vena cava, did certainly come eventually to carry this particular anatomical sense; but in two of the three instances cited it is not actually present. In the first (Hp. Vict. 1. 9) the usage is plural, and in the second (E. Ion 1011) it is simply a non-specific ‘hollow vessel’ (a vessel of the neck, whether vein or artery, swiftly collapsing and appearing empty when the throat of a sacrificial victim is cut). ⁶ Compare Rowe, Chapter 8, this volume, on the equivalent problems following from the treatment of Plato’s Greek as a separate variety, keyed to philosophy as a profession apart.

Medical Vocabulary and the Hippocratic Corpus

145

In LSJ, the term ἄρθρον is properly translated ‘joint’ and the usage (none here in Homer or Aeschylus) of Empedocles, Sophocles, Euripides and others is well set out. Knowledge of the joints was slow to follow knowledge of the bones, and was possibly not well formulated until early Hippocratic writing on the subject. Erotian finds it necessary to give a particularly extended explanation of the term (A 59). The great Hippocratic work on the subject of joints, On Articulations (Art. or Artic.), pays special attention to dislocations of the shoulder, a type of injury doubtless commonly sustained in wrestling. In LSJ the term ἐπωμίς, extensively and precisely used by the ancient orthopaedic surgeon, is translated ‘the point of the shoulder, where it joins the collar-bone’. So far so good, but in two passages of Euripides (E. Hec. 558 and IT 1404) where the word is used with the same anatomical precision, the sense ‘tunic’ or ‘[part of ] tunic’ is wrongly imputed to the tragedian. It is noted that τὰ ἄρθρα used alone has the sense ‘genitals’ in passages of Herodotus and Aristotle, but not that the singular is so used in Hippocratic gynaecology. The Revised Supplement (1996) adds further examples of this plural usage from Aelian and Lucian. The lexicographers completely miss the use of ἁρμός in this same sense (‘genital joint’ or ‘cleft’): the dative is mistranslated adverbially ‘just’, although the true sense is clear in several passages in On Diseases of Women (Mul. 1. 4, 36 and 213). Before leaving joints, we may look at the Lexicon on two related words used of their content, μύξα and μυελός. The former term, μύξα, is most commonly applied to peccant matter, as correctly noted by the translations ‘discharge from the nose . . . generally, mucus, mucous discharge’; but the translation ‘synovial fluid ’, suggested for the word when used of a fluid said to occur naturally in the joints (Hp. Loc. Hom. 7) is less correct, anachronistically imputing modern knowledge to the writer. Rather, μύξα, like σίαλον (similarly treated in LSJ), can be applied to a range of different body fluids. The latter term, μυελός, is more complex, referring both to bone marrow and—more commonly—to cerebral or spinal fluid (that implicit in the Homeric passage noted above): this inconsistency in nomenclature was castigated in the Hippocratic work On Flesh (Hp. Carn. 4). Some of the complexities inherent in the term are recognized in LSJ, s.v.: ‘Brain’ is offered as an alternative to ‘marrow’, and also for μ ῥαχίτης we are given ‘spinal cord’. In the context of a goose, μυελός is translated ‘fat’; instances are cited of ‘marrow’ as good food, and ‘marrow’ in metaphorical usage is also covered. The alternatives offered are, however, not quite correct. More properly, it might be said that the term is applied (1) to bone marrow and (2) to the contents of a channel believed to course from brain to spine to genitals (in men) or to womb and genitals (in women) or to the lower limbs, especially the joints (in both sexes); this, however, is extended description rather than restricted definition. And it is fair to add that the entry in LSJ displays a proper awareness that μυελός is not always and not simply ‘marrow’, as it is often loosely rendered in translation.

146

Elizabeth Craik

Anatomy merges with physiology in questions relating to the nature and function of bodily fluids. Humoral theory became dominant in later thinking, but this dominance is not Hippocratic. In LSJ the entries on χυλός and χυμός show an impressive awareness of the general sense of these key terms. The former entry, on χυλός, begins (I) ‘juice in general’ and refines this to 1 ‘juice of plants’, 2 ‘animal juices’, 3 ‘juice produced by the digestion of food ’, 4 ‘barleywater, gruel’; then (II) there is a palmary analysis of the equivalence in, despite grammarians’ supposed distinction between, the usage of χυλός and χυμός. The latter entry, on χυμός, is prefaced (I) ‘used much like χυλός, though sts. distd. fr. it’ and continues 1 ‘juice of plants’, 2 ‘animal juices’, ‘humours’; juice in a wider sense covering I. 1 and 2’ then for sense 3 gives the gloss of Hesychius conveying the sense ‘saliva’: χυμός∙ σίαλον Hsch. There is a clear perception of the inherent significance of the root verb χέω in both substantives and the interrelation in sense is made clear. There is a good treatment also of ἰκμάς, another weasel word used of bodily moisture, not confined to medical texts, and not uniformly used by all medical writers. The entry on ϕλέγμα contains some content of doubtful validity. The first sense is given (I) as ‘flame, fire, heat’ with citation of Il. 21. 337. Then (II) we have a heading ‘Medic.’ explained 1 ‘inflammation, heat’, 2 ‘phlegm, one of the four humours in the body’; 3 and 4 are more particular instances. The Homeric passage cited is extraordinary: Hera sets out to stir up a storm from the sea, carrying with it ϕλέγμα κακόν, to attack the Trojans. This is the only instance of ‘phlegm’ in the epics. Nothing necessitates an association with fire; rather it may relate to a deadly swelling in the sea, analogous to a rush of noxious phlegm in the body. The medical instances are misleading. Although ϕλέγμα, like the verb ϕλεγμαίνειν, may be associated with heat, it is more commonly associated with cold. And the status of phlegm as one of four bodily humours is by no means canonical. The entry on χολή makes more allowance for a variety of senses and for metaphorical usage. In the pathology of the Hippocratic writers, illness most frequently is thought to arise from the motion or blockage of peccant fluids in the body, especially fluids descending from the head. The key word for such a descent is κατάρροος, contr. -ρους. In LSJ the adjective is correctly rendered ‘downflowing’; the substantive (II) less happily as ‘running from the head, catarrh’. This mistranslation is unhappily recurrent in commentators on medical texts, perhaps through unthinking trust in the authority of LSJ. Much might be said about names for particular illnesses and the dangers of retrospective diagnosis. Even where there seems to be an equation between an ancient illness and a modern disease, such as πλευρῖτις (tr. ‘pleurisy’ LSJ) this is inexact; here, a noncommittal ‘disease of the lungs’ would be safer. Even more difficult is εἰλεός or ἰλεός, boldly rendered ‘intestinal obstruction’ and related to other diseases such as ‘nephritis’ and ‘jaundice’; here the transliteration eileos with the gloss ‘an intestinal disease’ would be safer. It would be easy, but pointless, to multiply

Medical Vocabulary and the Hippocratic Corpus

147

such instances. There is much use in medical texts of the adverbs ἄνω and κάτω applied to the upper and lower regions of the body, the thoracic and abdominal cavities: above and below the diaphragm rather than, as in our parlance, the belt. This very common idiom is not fully represented in LSJ. The most common usage in prescribing the diet and regimen appropriate for particular medical conditions in particular individuals and at particular seasons is based on words allied in root sense with δίαιτα. The substantive itself is ubiquitous in medical texts (193 occurrences listed in the Hippocratic concordance). When we turn to the entry δίαιτα in LSJ, we find the translation ‘way of living, mode of life’ illustrated by various tragic citations and instances from Herodotus and Thucydides. Only after II (1) ‘dwelling, abode’; ‘room’; ‘sailors’ quarters in a ship’ do we come to ‘2. Medic.’: ‘prescribed manner of life, regimen . . . esp. of diet’. Only two Hippocratic examples are given, the first (Hp. Vict. 1. 1) from the beginning of On Regimen, a work where the concept is central and much reiterated, the other (Hp. Fract. 36) from On Fractures, a work where it is of peripheral concern. The chronological principles followed are partly responsible for the skewing of emphasis; but the separation of ‘medical’ usage from that perceived as non-medical plays a part also. There are many words for foodstuffs, both animal and vegetable, in the Hippocratic corpus. LSJ is here helpful in pointing out parallel usage in Theophrastus and other relevant writers. The regular inclusion of glosses from Hesychius is less helpful than might have been hoped, as the information given often lacks precision. Here the convention of consistency in citation militates against utility. In many medical texts, δίαιτα is not a matter simply of diet but embraces other aspects of daily life, such as sleeping, bathing, and especially walking or taking exercise. There is some pairing in medical texts of the twin concepts κόπος and πόνος, both of which convey the sense ‘exertion’, but with the distinction that the exertion of the former is commonly fatigue and of the latter rather the chosen exertion of physical exercise. This distinction does not emerge in the Lexicon entries. In surgery, the verbs τέμνω and καίω describe the activities of cutting and burning, both used—though different physicians at different times seem to have favoured one or the other—to treat wounds. Surgical cutting, or lancing, is prominent in the LSJ entry on τέμνω: at (I) 3. ‘of a surgeon, cut’ . . . ‘abs., use the knife as opp. to cautery . . . ’ However the entry on καίω postpones mention of surgery until IV ‘of surgeons, cauterise’. The principles and practices followed in cautery remain somewhat obscure, as do the instruments and materials involved. Cautery by vegetable matter, μύκης, is advocated in On Internal Affections. LSJ unhelpfully offer ‘mushroom or other fungus’; the plant used may be herbal and similar to the mugwort burned on the skin in oriental moxibustion of the present day. A different type of omission can be seen in the entry κονδύλωμα, κονδύλωσις. LSJ give ‘knob, callous lump’, following on κόνδυλος ‘knuckle’. In context it is evident that the connotation is precise

148

Elizabeth Craik

and the reference is to ‘genital warts’, for which the modern medical term is condylomata acuminata. Hippocratic gynaecology can be seen as separate from other areas of medical activity. Expression as well as content is in many ways idiosyncratic. We begin with some general terms. The general adjective γυναικεῖα (neuter plural) can refer to all manner of things viewed as pertaining to women, some being listed in LSJ where ‘female disorders, titles of works by Hp. and Sor. . . . ’ is oddly postponed to II 2 d. The neuter plural adjective παρθένια is similarly used substantivally of all manner of things pertaining to girls. The term παρθένος is a social rather than a biological term, properly referring to a girl who is not (yet) married rather than to one who is a virgin.⁷ When we turn to more specific terms, much variation in usage presents itself. Here the separation of medical from literary—and especially from metaphorical— terminology has particularly unfortunate results. Such words as κρημνοί and δρόσος carry associations that are lost in translation. Euripides in particular adapted the language of the gynaecological treatises to suit his own creative purposes. A key word in embryology is γονή. Here, the entry in LSJ seems to err in proliferation of meanings yet does not quite catch the overriding sense: (I) 1 ‘offspring . . . ’, 2 ‘race, stock, family . . . parentage’; 3 ‘generation’; II (1) ‘that which engenders, seed’, 2 ‘organs of generation’, generally; III (1) act of generation, 2 of the mother, childbirth, 3 of the child, birth. There is a single reference, at II 1, to the key Hippocratic work on the topic. Erotian glosses it as sometimes σπέρμα (seed), sometimes μήτρα (womb). The latter, more unusual, meaning is seen in a Euripidean fragment (fr. 839). Lexical discussion of scientific theory, ideas, and ideals is beset by peculiarly intractable recurrent difficulties. Here, above all, we come up against questions relating to the development of Greek linguistic usage and to the interplay of different authors and genres. The very definition of terms to be viewed as ‘technical’ is fraught with difficulty: even in modern parlance, the notion of technicality is relative to the user. In Greek usage, when the development of a medical vocabulary from ordinary language was inchoate and new technical terms were being forged from pre-existing usage, the divide is peculiarly blurred. In addition to this difficulty of discrimination between general and technical terms, a distinction between imagery and literal speech is often hard to maintain. For example, the adjectives βαρύς, ῥᾴδιος, ἐλαϕρός, and ἰσχυρός can all be used in ways both technical and metaphorical when applied to medical conditions or symptoms. The situation was fluid: physicians adapted the existing language to their needs; literary authors might then adopt or adapt medical terms; physicians in turn borrowed back. The fifth century was a time of great semantic latitude and ambiguity.

⁷ Cf. Chadwick 1996, 226 9.

Medical Vocabulary and the Hippocratic Corpus

149

Some instances may be given. The verb τείνω, used both literally and metaphorically in the senses classified with some otiose prolixity in LSJ as (I) ‘stretch by force, pull tight’; 2 ‘stretch or strain’; 3 ‘stretch out, spread’; II ‘stretch out in length, lay’ B intr. ‘stretch out or extend’ . . . are commonly used in medical texts. In medical usage the verb has two different, but related, senses: the perfect participle passive, as applied typically to vessels (ϕλέβες) may signify ‘extending’, that is having a particular course through the body; or alternatively ‘tensed’, that is stretched taut. A more complex case is the substantive καιρός. Medical καιρός is a matter of balance and discrimination: the right time, the right amount, the right proportion, commonly all in conjunction.⁸ The separation of senses—with reference to place, time, season etc.—made in LSJ is arbitrary. There is only brief passing reference, at the end of section I, to an important long discussion of medical καιρός in the treatise Places in Man (Hp. Loc. Hom. 44) and other passages too might have been cited. The section II of place, ‘vital part of the body’ is illustrated not as it might naturally have been by medical usage, but by a single tragic citation (E. Andr. 1120). The strong connection with the adjective καίριος, which is common in Homer, is noted. The case of ὀργάω is difficult. Chadwick discusses the repeated Hippocratic conditional clause ἢν ὀργᾷ, regarding this as an example of impersonal use, surely rightly, but interpreting in the sense ‘if the need is pressing’, surely wrongly.⁹ The semantic range of the verb is implicitly indicated in the entry of LSJ II, though masked by some Victorian prudishness: ‘of men, like σϕριγάω, swell with lust, wax wanton, be rampant . . . ’ The analogy with σϕριγάω is apt and to it may be added the similarly analogous verbs σπαργάω and, less commonly found but preferred in certain medical texts, ϕλεγμαίνω. All are used of the vital sap in nature and vital fluid in animals, especially in a state of turgescence and change. Such words are used not only of vigorous healthy swelling, such as ripening fruit, breasts full of milk, sexual tumescence, but also of sick unhealthy oedema, such as fluid gathering in wounds. In all the passages cited by Chadwick, the sense is quite literal: ‘if there is swelling’. The word ὀξύς presents a challenge to translators, commentators, and lexicographers alike. LSJ begin: ‘(I) sharp, keen . . . brought to a point . . . ’ and continue ‘(II) in reference to the senses, 1 of feeling, sharp, keen . . . 2 of the sight . . . 3 of sound . . . 4 of taste . . . ; III metaph., of the inner sense, sharp, keen, hasty . . . 2 sharp, quick . . . ; IV of motion, quick, swift . . . ’ Here, as often, a proliferation of similar meanings is unhelpful. The pervasive attempt to address and distinguish ‘intersensual’ aspects of the adjective is apt for literary texts (especially epic and tragedy) but does little to illumine prose usage. The ordering of examples under the head II 1 is lamentable: we begin and continue ⁸ On the development of the word, see the wide ranging study of Trédé (Trédé 1992). ⁹ Chadwick 1996, 217 18.

150

Elizabeth Craik

with epic and lyric verse before there are a few Hippocratic examples. As an apparent afterthought to Sophoclean and Pindaric citations we find a reference to the title of the significant Hippocratic work On Acute Diseases: ‘cf. Hp. Acut. tit.’; in this almost parenthetical reference the important medical sense of acute, as opposed to chronic, with reference to disease is not brought out. And although, under IV, we find a late Greek instance of the adjective qualifying καιρός ‘an urgent crisis’, we miss the Hippocratic instance of the same phrase (‘occasion is fleeting’) in the famous first aphorism. Much mention has been made above of the dangers of regarding literary and medical usage as separable or separate, and of making an artificial distinction between use of the same terms, according to author and genre. One final example of general application may be given: the distinction made in LSJ between χλωρός in tragedy (‘of blood or tears, fresh’) and in medicine (‘of body fluids, pale, yellow, biliouslooking’) lacks justification. The ordering of entries is often problematic, due to the frequently unrecognized relation between texts in medicine and in other genres. There was undoubtedly a complex two-way process at work and priority is impossible to determine. Thus, it is surely correct to note, s.v. II, that πρόϕασις ‘external exciting cause’ is used ‘esp. as a medical t. t.’; but is this different from Thucydides’ famous use of the word, with regard to the Peloponnesian war, and is Thucydides simply adopting a medical term? The conventional use of the Ionic dialect in scientific writing doubtless makes classification difficult at times; but this is perhaps beyond the reach of lexicography. Attention has been paid in this chapter to the ways in which medical and especially Hippocratic language is presented, represented, and at times misrepresented. Little attention has been paid to the many medical writers, not all minor, known from fragmentary sources. There is no mention in LSJ of, for instance, Praxagoras, undoubtedly a major figure in the post-Hippocratic tradition. It may seem ungrateful to point out such omissions in a very large lexicon. However as space is found for extensive citation of papyri and of minor post-classical authors, one may lament the underlying choices made and emphasis laid by the editors.

10 The Greek of the New Testament* Patrick James

‘We have also been careful to notice such words as occur first, or in any unusual sense, in the Alexandrian version of the Old Testament, and in the New Testament.’¹ ‘ . . . and especial care has been taken to explain all words contained in the New Testament. . . . ’ ‘All tenses and forms of words in the Gospels that presented any difficulty have been inserted in their place.’² ‘Besides these [lexemes from Homer to “the close of Classical Attic Greek”], will be found words used . . . by the writers of the New Testament.’³ * The content of this chapter and my assessment of LSJ’s treatment of the New Testament and its Greek were formulated during my work both for the Cambridge Greek Lexicon Project (2007 16) and with Peter Williams and Dirk Jongkind (2016) on the text of the New Testament for The Greek New Testament Produced at Tyndale House, Cambridge: see Jongkind, Williams, Head, and James 2017. I am very grateful to Anne Thompson for all that she has taught me about lexicography and to James Diggle and Bruce Fraser for asking for my assistance with New Testament forms and citations as well as to Simon Westripp, with whom I have discussed many of the points considered here. I am grateful too for several exchanges with Jim Aitken and John Lee. Jane Mclarty very kindly read and commented on a draft. Other indirect contributions have been made by several Readers at Tyndale House and others with whom I have read the New Testament in Greek and the Septuagint or with whom I have discussed their vocabulary. Their questions influenced my selection of entries. In particular, I wish to thank †Sam Barker, Jonathan and Laura Beatham, Bruce Clark, Robert Crellin, Daniel Eng, Simon Gathercole, Bobby Jamieson, Toby Hudson, Matthew Ralph, and Howard Spencer. The abstract for this chapter was submitted shortly before the birth of my son, Joshua, who has turned five. I wish to thank the editors, and the other contributors, for their μακροθυμία throughout. ¹ LS¹, vii. ² [Marshall] 1846 [1843], ‘Advertisement’. Although this lexicon abridged from Liddell and Scott’s Greek English Lexicon was published anonymously, Thompson (1899, 75 6) identified its editor as George Marshall, on whom see Thompson 1899, 39, 74, and 168 n. 1, and Stray, Chapter 1, this volume. This statement seems to be the precedent for the coverage of the New Testament by the Cambridge Greek Lexicon. ³ IGL, Preface. The claim was not met in that several New Testament words were omitted from IGL (e.g. ἀποϕορτίζομαι, ἀρτιγέννητος, θηριομαχέω, and ὑπογραμμός), while others were Patrick James, The Greek of the New Testament In: Liddell and Scott: The History, Methodology, and Languages of the World’s Leading Lexicon of Ancient Greek. Edited by: Christopher Stray, Michael Clarke, and Joshua T. Katz, Oxford University Press (2019). © Oxford University Press. DOI: 10.1093/oso/9780198810803.003.0010

152

Patrick James ‘Liddell and Scott’s Lexicon . . . also covers the early Greek of Homer and Hesiod as well as New and Old Testament Greek.’⁴

These statements, which frame the Liddell and Scott tradition, remind us that the treatment of the Greek of the New Testament has been central both to the Lexicon from its first edition and to its 1843 abridgement⁵ and to IGL. The stated aim was to provide students and scholars with a full inventory of New Testament words and their meanings ‘with especial care’. The Abridgement sought to provide help with problematic forms of words, at least for the Gospels. Apart from these statements, very little has been said about how non-classical words and non-classical meanings that are attested in the New Testament were handled by Henry Liddell and Robert Scott and by their successors, Henry Stuart Jones and Roderick McKenzie. Even John Lee’s masterful History of New Testament Lexicography mentions the various editions only in passing.⁶ The story of the treatment of the Greek of the New Testament⁷ in the Lexicon will be filled out in this chapter alongside a critical assessment of that treatment. The internal evidence of a selection of entries for words attested in the New Testament⁸ will be examined as well as the external evidence from discussions of the Lexicon (including, for the present purpose, the various Prefaces). The focus of this chapter will be the development from the eighth edition of Liddell and Scott (henceforth, LS⁸) to LSJ. LSJ marked something of a new beginning,⁹ not only in its coverage, as we shall see next, but also, as we shall see later, in its approach both to the New Testament’s vocabulary and to its Greek in general. By contrast, LS⁷ was in effect reprinted as LS⁸, the last edition cited from authors later than the New Testament (e.g. ἀποθησαυρίζω cited from Lucian). My colleague, Simon Westripp, has compiled a comprehensive list of omitted New Testament headwords. ⁴ LSJ Revd Suppl., back flap of the dust jacket. ⁵ [Marshall] 1843. ⁶ Lee 2003, 3, 12, 16, 20, 79, 98, 134, 208, 221, 227, 270 1, 281, and 295. ⁷ Although this chapter will concentrate on the lexicon of the New Testament, in line with the editors’ invitation to contribute, in places I will touch upon the lexicon of the Septuagint, a convenient term for the collection both of compositions in Greek and of translations into Greek of the Hebrew Bible that comprise the Old Testament in Greek as preserved chiefly in Codices Vaticanus and Alexandrinus and edited by Rahlfs 1935a and 1935b. For the history of the term ‘Septuagint’, see Dines 2004, 1 3 and Williams 2012, 173 8. ‘LXX’ stood for ‘LXX interpretes’ until LSJ, when ‘LXX’ became ‘Vetus Testamentum’. See Jones 1940, xxviii and xxxvii xxxviii. For LSJ’s treatment of the Greek of the Septuagint, readers should consult Lee 1969 and Caird 1968 and 1969, which remain immensely valuable, as well as Glare (1997, 212) and Hauspie 2004. The unpublished lecture that was delivered at Oxford on 10 April 1942 by Peter Walters (né Katz) and is preserved as PW/4/1 in the Library of the Faculty of Asian and Middle Eastern Studies, Cambridge, also deserves close attention. ⁸ Since any selection is partial by definition and, therefore, is open to accusations of not being representative, readers may like to consult the discussions by Chadwick (1996, especially 32 4, 45, 138 40, 226 9, and 307 11) and the passing comments made by Lee (2003, 208, 221, 227, 270 1, and 281), but it should not be assumed that I endorse the conclusions of those scholars. ⁹ See Stray 2010b, 108 11 for the background and Lee 2010b, 122 4 for a specific example.

The Greek of the New Testament

153

from Liddell himself (Scott died in 1887 and, even if he had not always been the ‘junior partner’,¹⁰ the seventh and eighth editions were the work of Liddell alone):¹¹ ‘All corrections and additions that could be made without altering the pagination have been inserted in the text’, Liddell wrote in retirement at Ascot in 1897.¹² More substantial alterations were confined to a few pages of Addenda et Corrigenda¹³ and LS⁸ was not regarded as a significant advance on LS⁷.¹⁴ For these reasons, I shall refer to LS7-8 in my discussion of particular entries.

1 0. 1 . CO V E R A G E LSJ marked a significant departure from previous editions in that the vocabulary of Patristic texts and Byzantine literature (‘Eccl.’ and ‘Byz.’ in LS⁸) was relocated to the Patristic Greek Lexicon. At the time, this was in preparation under Darwell Stone and was eventually published from 1961 to 1968 under G.W.H. Lampe.¹⁵ Jones announced that ‘Christian poetry and inscriptions’ would be included with the Church Fathers in that lexicon (henceforth, ‘Lampe’).¹⁶ It is often remarked that the result of this decision was that the works of Nonnus, in particular, were parted:¹⁷ the Dionysiaca remained in LSJ,

¹⁰ For the extent of Scott’s involvement, especially in LS⁵ and LS⁶, see Stray, Chapter 1, this volume. We should not draw firm conclusions about Scott’s contribution from the ‘Westminster Epigram’ quoted and discussed by Kitchell 1988 and 1989 and by Calder 1988, but we should take into account Scott’s activities away from Oxford. ¹¹ Stray 2010b, 108 reports that Liddell had worked alone ‘since about 1870’. See Thompson 1899, 78 9 n. 1 and Jones 1925, iv. ¹² LS⁸, vi. ¹³ See Stray 2010b, 106 and LSJ s.v. Ἀραβάρχης, which reads ‘v. Addenda’, a relic of the revision process. ¹⁴ Stray 2010b, 108 quotes the verdict of Richard Garnett of the British Museum. Chadwick 1994, 1 gave his own assessment and noted (1996, 3) that, in part, his impressions had been shaped by his ‘admirable Greek teacher, Mr (later Professor) G.E. Bean’. (George Bean, who taught Chadwick and Kenneth Dover at St Paul’s School, was later professor of Greek at the University of Istanbul.) LS⁷ had included words from CIG, such as ἀνθυποϕαίνω from CIG III 4958 (Egypt), which proved to be a ‘ghost word’, when the inscription was re read and printed with γένος ὑποϕήνας, not γένος ἀν[θ]υποϕήνας: see Wagner 1987, 69, no. 12 and SEG VIII 792. However, LS⁸ was distinguished by greater attention to inscriptions. See Jones 1925, iv, for Liddell’s ‘lexical study of inscriptions’ between LS⁷ and his ‘later years’. ¹⁵ For Lampe himself, see Moule 1981 and 1982. For the Patristic Greek Lexicon, see Moule 1981, 405 6; Chadwick 1982; and the reviews by Chadwick 1962; Fabricius 1967 and 1968; and Wifstrand 1963 and 1964. ¹⁶ Jones 1925, xi. For a retrospect, especially on Lampe’s lack of attention to papyri, to inscriptions, and, within Lampe’s Patristic texts, to ‘a lot which appears to have no immediate theological application’, see Glare 1997, 215 16. For theologians’ perspectives, see Richards 1925 and Chadwick 1982, 66 7 and 71. ¹⁷ See, for example, Fabricius 1968, 314 15.

154

Patrick James

while the Paraphrase of the Gospel According to John found a new home in Lampe.¹⁸ Since debate continues about whether and how both hexameter poems could be the work of a single poet, such a separation is justifiable.¹⁹ The division affected other authors with less justification.²⁰ The epigrams of Paul the Silentiary continued to be treated by LSJ in isolation from his metrical Ecphrasis of Hagia Sophia.²¹ The division between classical and classicizing Greek, on the one hand, and ‘Christian Greek’ literature and documents, on the other, was not consistently practiced. Some ‘Christian’ words remained in LSJ, such as ἀρχιεπίσκοπος ‘archbishop’, because they could be cited from the Novellae of the sixth-century emperor Justinian. Origen is cited s.vv. ἀνταγείρω and παραμηχανάομαι, words which are attributed to his pagan opponent Celsus and, hence, are justifiably included. A less justifiable addition to LSJ is στεϕανοσταύριον ‘cross in a wreath’, which is cited from an inscription published in 1931,²² years after the separation of material between LSJ and Lampe had begun, after the period of ‘copy-writing’, and long before the publication of the relevant part of Lampe. There is, at least, no duplication in Lampe. By contrast ἁγνευτήριον is cited, without a date, by LSJ from P.Oxy. V 840.8, an apocryphal gospel fragment, as a ‘sacristy’, and in Lampe, dated to the fourth or fifth century, joined by the occurrence in line 13, and explained more accurately as ‘of inner court of Temple’ [sic]. The duplication may be justified if Lampe intended to correct LSJ. Inclusion in LSJ may be justified on the grounds that the citation came from a papyrus, not a ‘Patristic’ text. However, since the papyrus had been published in 1908,²³ it should simply have been assigned to Lampe and omitted by LSJ. A more significant concern, to my mind, is that, although LSJ retained the Greek vocabulary of the Septuagint as a precursor to that of the New Testament, the New Testament was separated from the writings of those who sought to explain it and to live by it, as we shall see in the discussions of ἐπερώτημα and ϕωτίζω below.²⁴ One exception illustrates the utility of such ¹⁸ The vocabularies of the Dionysiaca and Paraphrase are being reunited in DGE. Peek 1968 75 treated the Dionysiaca in isolation. ¹⁹ See Sherry 1996, 411 13 for a case for two poets and, in favour of a single author, see Franchi 2016. ²⁰ Fabricius 1968, 314 15 also drew attention to Aeneas and Procopius, both of Gaza, as victims of this policy, Chadwick (1982, 66 7) to Synesius, an ‘undistinctively Christian’ writer. ²¹ See Jones 1940, xxxi and Lampe 1961 8, xxxix. ²² Fitzgerald 1931, 66 7 and Plate VII. The inscription was published in full by Starr 1937, 87 8; cf. SEG VIII, 39 40. This entry also indicates that words from some texts published after ‘the period of copy writing’ of 1911 24 were included. For ‘the period of publication’ and ‘the period of copy writing’, see Anon. 1940, xiii. ²³ Grenfell and Hunt 1908, 1 10, who translated ἁγνευτήριον as ‘place of purification’ and discussed its application to the Temple in Jerusalem at length (1908, 7 8). ²⁴ Chadwick (1982, 66) commented that ‘LSJ excluded all Christian texts other than, para doxically, the New Testament’.

The Greek of the New Testament

155

data. The reference to Origen, De Oratione 27.7, s.v. ἐπιούσιος, which concerns the extreme rarity of that adjective in Origen’s time, is significant for assessing its interpretation in Matthew 6.11 and Luke 11.3.²⁵ It would be valuable to have the vocabulary of the Apostolic Fathers²⁶ and of the non-canonical gospels and acts presented alongside the New Testament evidence, since they are closer to one another in language and genre than either is to the Church Fathers of later centuries, as Lampe himself had suggested.²⁷ To take an example to which we shall return: although Lampe has ἐπερώτημα both as ‘inquiry’ from the Shepherd of Hermas 11.2 and as glossed by ὁμολογία ‘pledge’ by Cyril of Alexandria (Paschal Homilies 30.3/5².346D) in reference to 1 Peter 3.21, he did not include the instance from the Apophthegmata Patrum.²⁸ That instance was included neither by E.A. Sophocles²⁹ nor by Frederick Danker³⁰ (henceforth, ‘BDAG’), a reflection of the absence of that citation from New Testament studies. Neither LSJ nor Lampe mentions the two instances of this noun in the Greek versions of the Hebrew Bible: Daniel 4.14(17) Theodotion, which has no parallel in the ‘Old Greek’ of Daniel,³¹ and Siracides 33.3 (Codex Sinaiticus inter alios).³² Although the second-century Egerton Gospel had been published in Bell and Skeat 1935, its use of ξένον ἐπερώτημα was not included by Lampe. The removal of ‘Christian’ words also resulted in several anomalies such as empty cross-references and false citations. The entry for πλάστης suggests that the meaning in Plutarch’s Life of Dion 9 should be equated with that of τριχοπλάστης. Be that as it may,³³ since LSJ has no entry for τριχοπλάστης, the reader will have to turn to Lampe s.v. τριχοπλάστης to understand πλάστης in that Plutarch passage.³⁴ This situation arose because the very rare noun ²⁵ Another exception is suggestive. S.v. ἀκτημοσύνη Crates of Thebes is cited ‘ap. Epiph. Haer.3.2, Poll.3.111, 6.197’. The noun may well have been used by Crates in his verse, but its prominence as a technical term in Patristic texts from Clement of Alexandria and Origen onwards and the nature of its occurrence in Epiphanius, who uses it for himself elsewhere, suggests that it may not have been used by Crates himself: ἔλεγεν ἐλευθερίαν εἶναι τὴν ἀκτημοσύνην. The instances in Pollux are not attributed to Crates, as LSJ may be taken to imply. ²⁶ Their absence was a ‘pity’ according to Katz 1942. ²⁷ Lampe 1961 8, viii. ²⁸ Abba Sisoes 17: Alphabetical Collection 397 ιζʹ/Systematic Collection 8.21. ²⁹ Sophocles 1870, s.v. ³⁰ Danker 2000, s.v. ³¹ For the Greek versions of Daniel, see Dines 2004, 2, 23 4, and 84 7. ³² BDAG mentions both these instances, but with no more precision than ‘v.l.’ for Siracides 33.3. Muraoka 2009, 262 and 293 does not cite the variant reading at Siracides 33.3. ³³ Anne Thompson would strongly argue against ever equating one word with another. ³⁴ Although it may be reasonable to have to consult another lexicon either for a particular word or for a technical meaning, especially in a particular author, it seems unreasonable to be directed to another lexicon for one word (e.g. τριχοπλάστης) for the meaning of another (e.g. πλάστης). Lampe contains an entry for πλάστης, but only as ‘modeller, shaper’ in reference to God or Logos ‘in Creation’. LSJ could have simplified this situation by giving an explanation of πλάστης in that Plutarch passage, rather than ‘perh. = τριχοπλάστης’, an equivocal equation with an absent headword. This fault was rectified by LSJ Revd Suppl., by deleting the equation and by glossing with ‘hairdresser (cf. κεροπλάστης)’.

156

Patrick James

τριχοπλάστης is attested first in Synesius’ Encomium of Baldness 21,³⁵ which is hardly ‘Patristic’, but was relegated to Lampe because Synesius was a Neoplatonist who converted to Christianity and became a bishop.³⁶ The entry for ἀρχάγγελος has ‘LXX Da.10.13, al.’ as its first citation, followed by ‘Ep.Jud.9’ (but not 1 Thessalonians 4.16). The word ἀρχάγγελος does not occur in that verse of Daniel in any Greek version, elsewhere in that book, or anywhere in the Septuagint. Instead, the reference to Daniel 10.13 seems to have originated as a reference to Theodoret’s commentary on Daniel 10.13 (PG LXXXI 1497 A) among other occurrences (1472 B and 1477 B; hence, LSJ’s ‘al.’), perhaps through confusion of an abbreviation for Theodoret (‘Thdt.’ in Lampe) with the abbreviation for Theodotion (‘Thd.’ in LSJ).³⁷ Since the translation of Daniel in the Septuagint is that associated with the name Theodotion,³⁸ the step from Th(eo)d(otion) to ‘LXX’ would be easy. However, why such a confusion would happen remains mysterious. Since LS7-8 simply cites ‘N.T., Eccl.’ and, rightly, does not mention the Septuagint at all, there was no instance of ‘LXX’ to expand and no abbreviation of Theodoret to be the occasion for confusion. It may be the case that here we see traces of the influence either of Professor A.H. McNeile and the Reverend A. Llewellyn Davies, who advised Jones ‘in matters relating to the LXX, Hexapla, etc.’³⁹ or of one of the ‘well-meaning persons’⁴⁰ and ‘social inadequates and reprobates, including defrocked clergymen, who were so useful to learned presses well into the twentieth century’.⁴¹

1 0. 2 . Τ Ο Ο Ν Ο Μ Α Τ Ο Υ Π Ε Ρ Π ΑΝ Ο Ν Ο Μ Α : NE W TESTAMENT W ORDS AND THEIR MEANINGS The entry for the personal name Jesus (Ἰησοῦς) provides a good starting-point for illustrating the nature of LSJ’s treatment of the lexicon of the New Testament. ³⁵ Translated ‘haircrimpers’ by Kendal 1985, 35. Three occurrences of this noun in the Suda involve quotations of the Synesius passage. Since Hesychius used τριχοπλάστης to gloss κικιννᾶς, both should have been included in LSJ, but neither was. ³⁶ Lampe 1961 8, vi. See also Chadwick 1982, 66 7 and 71. The Encomium of Baldness, On Dreams, and letters of Synesius subsequently returned to the Liddell and Scott tradition (Liddell, Scott, et al. 1968: ix). ³⁷ The clue to this explanation came from the commentary to P.Oxy. LXXXII 5306.5, in which ‘Thdt. Dan. 9.24, 25 (PG LXXXI 1472B, 1477B)’ was mentioned in relation to the phrase ὁ ἅγιος ἀρχάγγελος. See Gonis, Maltomini, Henry, and Slattery 2016, 81. Then, I consulted Theodoret on Daniel 10.13 to confirm that ἀρχάγγελος was there too. ³⁸ For Theodotion, see Dines 2004, 84 7. Rahlfs 1935b, 870 936 printed both the text of Theodotion from Codices Vaticanus and Alexandrinus and the text of the ‘Old Greek’ from his 88 (his P967 is another witness to the ‘Old Greek’ of Daniel). See Rahlfs 1914, 278 80 and 341. ³⁹ Jones 1925, ix. For the Hexapla, see Dines 2004, 97 103. ⁴⁰ For this phrase in relation to the sources of Septuagint material and the weaknesses of its treatment, see Chadwick 1994, 7 8. ⁴¹ Stray 2010b, 110.

The Greek of the New Testament

157

‘Ἰησοῦς, οῦ, dat. οῖ, Joshua, LXX Jo.1.1, al., Act.Ap.7.45; in NT ’, with dat. οῦ, JESUS, Ev.Matt.9.27, al.’

Its precursor from LS7-8 follows to show the development to the entry’s present form. LS⁶, the first to include this name, differed only in that it parenthesized the comment about the form of the dative. ‘Ἰησοῦς, οῦ, dat. οῖ, LXX, but οῦ, N.T., JESUS; Greek form of Hebrew Joshua or Jehoshua, saviour.’

Liddell and Scott were very selective in their inclusion of personal names: ‘We have left such only as had in themselves some force and significance, or presented anything remarkable in their grammatical forms . . . And here it may be observed that the proper names of the mythological and heroic times contain elements of the language which sometimes can and sometimes cannot be traced elsewhere; cf. Ζεύς, Σείριος, etc.’⁴² As examples of names with remarkable ‘grammatical forms’, Ἀγαμέμνων and Ἡρακλῆς were cited and, from LS⁴, they were joined by Ὀδυσσεύς.⁴³ Liddell and Scott’s claim to ‘have left’ names from Homer and Hesiod betrays the origin of their lexical inventory in Passow, whose foundations were a lexicon to Homer and Hesiod.⁴⁴ The inclusion of Ἰησοῦς may be justified similarly by the complexities of its inflection. In particular, the loss of a distinct dative form between the earlier rendering and that of the New Testament is noteworthy.⁴⁵ However, the treatment of the morphology is overly simplistic. Exodus 17.14 has Ἰησοῖ as a dative after the imperative δός,⁴⁶ as LSJ would have us expect, but Exodus 17.9 has Ἰησοῦ in the function of a dative after τῷ,⁴⁷ a precursor of the name’s treatment in the New Testament. There is no mention of plural forms of this name,⁴⁸ an example of which will be quoted below. The accusative and vocative are left to be inferred on the basis of the genitive -ου, as if that of a second-declension noun: -ος, -ου, -ον, -ῳ, and -ε. That is misleading in relation ⁴² LS¹, ix. ⁴³ The pair appeared in a footnote in LS¹, ix, LS², ix, LS³, ix. LS⁴, iv promoted the three into the main text. ⁴⁴ LS¹, vii. ⁴⁵ For the dative in post Classical Greek, see Browning 1983, 36 8; Horrocks 2010, 97, 107 8, 116 17, 138, 152, 179 80, 184 5, 186, 246 7, 262, 264, and 284 5; and Stolk 2015, 2017a, and 2017b. ⁴⁶ No instance of Ἰησοῖ as a dative is attested in documentary papyri or inscriptions. Wevers 1990, 268 and 271 did not discuss this variation. Wevers printed Ἰησοῦ at 17.14 (1991, 222) and noted (1990, 271) that ‘presumably what Exod(us) means here is “recite in the hearing of (Jesous)” ’, implying, but not guaranteeing, analysis as a genitive. ⁴⁷ So too Joshua 10.17, Joshua 17.14, and 1 Chronicles 24.11. Wevers (1991, 220) printed τῷ Ἰησοῦ at Exodus 17.9. His apparatus criticus mentions some manuscripts that have πρὸς Ἰησοῦν (and so avoid using the dative) and that Codex Vaticanus was later altered to τῷ Ἰησοῖ. ⁴⁸ LSJ cites several such plurals of personal names, especially in their function as ‘people like so and so’: e.g. s.v. Θερσίτης (s.v. θερσιεπής) and s.v. Πήγασος: ‘pl. Πήγασοι, as a sample of prodigies, Pl.Phdr.229d, . . . ’.

158

Patrick James

to the vocative and accusative (not to mention for the dative), since the vocative is Ἰησοῦ,⁴⁹ the accusative Ἰησοῦν.⁵⁰ A more accurate analysis of the morphology would be as follows. We have Ἰησοῦ as an indeclinable rendering of a Semitic name, which is used as a vocative (usually nothing other than the noun’s stem, except in second declension singular), as a genitive, and, by some at first and by all later, as a dative. To that extent, this rendering is comparable with Ἀβραάμ and Ἰωσήϕ, both of whom are indeclinable in the New Testament.⁵¹ Indeed, in 2 Esdras 22.8 and 22.24, the stem Ιησου, uninflected and unaccented,⁵² must function as a nominative.⁵³ The indeclinable Ἰησοῦ was treated as a stem (perhaps on the basis of the vocative) which could be marked either as nominative (by -ς) or as an accusative (by -ν).⁵⁴ Some, and only for a time, used a dative (Ἰησοῖ), which coincides with that of the πειθώ and Σαπϕώ type. From another point of view, we have here a name that functions in accord with the Modern Greek rule that various masculine paradigms involve a genitive formed by removing the final sigma of the nominative.⁵⁵ Comparison with Modern Greek, which has no dative case,⁵⁶ also makes the loss, or (for many) the absence, of a distinct dative in the history of this name particularly noteworthy. LSJ’s entry for Ἰησοῦς contains several curiosities indicative of the status of the New Testament in the development of the Liddell and Scott tradition. Its use of small capitals for an English equivalent is unique, to my knowledge.⁵⁷

⁴⁹ See, e.g. in Mark: Ἰησοῦ Ναζαρηνέ (1.24) and Ἰησοῦ, υἱέ τοῦ θεοῦ τοῦ ὑϕίστου; (5.7). ⁵⁰ Although the accusative form had not been mentioned explicitly in LS⁶, it was included by Marshall in the Abridged Lexicon of 1871, 328, the revision derived from LS⁶. The entry read: ‘Ἰησοῦς, οῦ, dat. also οῦ, acc. οῦν, Greek form of the Hebrew Joshua, Saviour.’ The name was included as one of those ‘words in the Gospels that presented any difficulty’. ⁵¹ However, an inflected form Ἀβράαμος is found in the papyri (Moulton and Milligan 1914: 1). The Jewish historian appears as Ἰώσηϕος and Ἰώσηπος in his own writings. See BDAG s.v. Ἰωσήϕ. ⁵² Rahlfs 1935a and 1935b did not accent indeclinable renderings of Semitic names. ⁵³ I am grateful to Peter Myers of Tyndale House for sharing with me some of his data on Greek renderings of Semitic personal names in 2 Esdras. Rahlfs 1935a, 945 6 reported that some manuscripts have Ἰησοῦς as a nominative at 22.8 and he printed Ἰησοῦς as a nominative in 22.10: καὶ Ἰησοῦς ἐγέννησεν τὸν Ιωακιμ (cf. Ἰησοῦ as a genitive in 22.26: ἐν ἡμέραις Ιωακιμ υἱοῦ Ἰησοῦ υἱοῦ Ιωσεδεκ καὶ ἐν ἡμέραις Νεεμια). ⁵⁴ Another genitive, Ἰησοῦτος is found in the papyri. See Moulton and Milligan 1920, 301 2. ⁵⁵ Cf. Μωυσῆς, genitive (and dative) Μωυσῆ. See Holton, Spyropoulos, Mackridge, and Philippaki Warburton 2012, 55 7, 60 3, 72, and 83; and Horrocks 2010, 228. ⁵⁶ See Holton, Spyropoulos, Mackridge, and Philippaki Warburton 2012, 85 for the ‘learned’ relics of the dative. ⁵⁷ LS7 8, s.v. χριστός II 2, repeated the headword as ΧΡΙΣΤΟΣ and translated as ‘the Anointed One, the CHRIST ’, with the explanation, ‘transl. of the Hebr. Messiah, N.T. passim;’. LSJ recast this entry and removed the uppercase Greek and Roman type: ‘3. in NT, ὁ χ. the Messiah, Ev. Matt.2.4, etc.; ὁ χ. Κυρίου Ev.Luc.2.26; then used as pr. n. of Jesus, Ἰησοῦς χ. Ev.Matt.1.1, etc.; Ἰησοῦς ὁ λεγόμενος χ. ib.16.’ Type faces are otherwise limited in LSJ to (1) bold for headwords, (2) italics for translations, and (3) Roman for explanatory comments. Lee (2010b, 123 4) discusses a problematic instance of this practice.

The Greek of the New Testament

159

The choice of ‘Ev.Matt.9.27, al.’ appears somewhat random in comparison with ‘Jo(shua) 1.1, al.’, especially since this name appears in Matthew 1.1, cited s.v. χριστός II 3. ‘Ev.Matt.9.27’ ( . . . τῷ Ἰησοῦ . . . ) probably was cited as an example of Ἰησοῦ in the function of a dative. The citation of ‘Jo.1.1’, in fact, exemplifies Ἰησοῖ as a dative: εἶπεν κύριος τῷ Ἰησοῖ υἱῷ Ναυη τῷ ὑπουργῷ Μωυσῆ. LSJ’s entry shows no interest in this name outside the Septuagint and the New Testament. However, Joshua and Jesus were not the only bearers of this name. Indeed, the works of Josephus name as many as twenty-one Jesuses and John of Damascus commented in the seventh or eighth century: πολλοὶ μὲν γὰρ Χριστοὶ καὶ Ἰησοῖ, ἀλλ᾽ εἷς ὁ ἐσταυρωμένος.⁵⁸ As of the time of writing, the Lexicon of Greek Personal Names knows of seven bearers of the name Jesus, the majority of whom have been identified as Jews. BDAG (s.v. Ἰησοῦς) refers to a few inscriptions, ostraca, and papyri in which the name is attested and mentions the Letter to Aristeas. Of the papyri, the description of P.Oxy. IV 816, published in 1904,⁵⁹ certainly was available to Jones and McKenzie and would also have been known to them through Moulton and Milligan.⁶⁰ Further, the coverage of the name is incomplete even within the New Testament and Septuagint. BDAG (s.v. Ἰησοῦς) lists five New Testament bearers of this name, adding the son of Eliezer (Luke 3.29: Papyrus 4 and Codices Sinaiticus and Vaticanus), [Jesus] Barabbas (Matthew 27.16–17),⁶¹ and Jesus also known as Justus (Colossians 4.11). As for the Septuagint, beside Joshua the son of Nun, there was also the son of Jehozadak in Ezra-Nehemiah, Haggai, and Zechariah. LSJ made no comment on the appearance and identity of Ἰησοῦς at Jude 5 in Codices Alexandrinus and Vaticanus and was wise to do so, in my view, given the textual issues there.⁶² This case study illustrates four features of the treatment of New Testament words throughout the Liddell and Scott tradition. First, New Testament words, Apart from Greek, non Roman scripts are presented in transliteration. The only exception of which I am aware is the entry for Χημία: ‘Black land, Chemmi, Egyptian name for Egypt, Plu.2.364c. (Egypt(ian) Kmt, Copt(ic) κημε, χημι “Egypt”.)’. Although Coptic used Greek script with six additional letters, none of which feature in this word, in this entry LSJ employed an Alexandrian majuscule font for the Coptic spellings, not its usual Greek typeface. For that ‘Greek Uncial of Coptic Type’, see Cavallo and Maehler 1987, 5. ⁵⁸ Expositio Fidei 84 on ἡμεῖς δὲ κηρύσσομεν Χριστὸν ἐσταυρωμένον (1 Corinthians 1.23). ⁵⁹ Only part of P.Oxy. IV 816 verso (6/5 BC), one of the ‘descriptions of miscellaneous documents’, has been published:]ς Ἰσιδώρου καὶ Ἰησοῦς. If the second sigma of Ἰησοῦς really is part of that word and not the first letter of the next, in the absence of further context, we may see Ἰησοῦς used as a genitive after the certain genitive Ἰσιδώρου, as implied by the translation of Tcherikover and Fuks 1960, 179: ‘of Isidoros and Iesous’. ⁶⁰ Moulton and Milligan 1920, 301 2. Cf. Jones 1925, ix. ⁶¹ The name Ἰησοῦς is absent as the name of Barabbas in Matthew 27.16 from Codices Sinaiticus, Vaticanus, Alexandrinus, Bezae, and Washingtonianus, but was printed by Nestle Aland 2012. ⁶² Metzger 2000, 657 8 quotes ‘the Committee’ on Ἰησοῦς here as ‘difficult to the point of impossibility’. Note the reassessment between Nestle Aland 1993, 629 and Nestle Aland 2012, 731 such that κύριος became Ἰησοῦς. As James Diggle taught me, ‘Some lectiones difficiliores are too difficiles.’

160

Patrick James

including some proper names, were included, but not necessarily from the first edition (Ἰησοῦς from LS⁶).⁶³ Second, entries for New Testament words are not necessarily comprehensive in their treatment of the New Testament attestations. Third, New Testament instances of Classical Greek words are not always noted. The entry for ἐπερώτημα did not mention its occurrence in 1 Peter 3.21 until LSJ. Herodotus and Thucydides were cited without any indication that the word is attested elsewhere.⁶⁴ Omission prior to LSJ may have been deliberate, and wise, given the questions of interpretation that concern that word, the verse as its immediate context, and the broader context of 1 Peter 3.18–22. The publication of papyri and inscriptions that contained ἐπερώτημα, and related words, may have prompted the cautious citation of 1 Peter 3.21 in LSJ: ‘hence prob., pledge . . . ’. Fourth, New Testament words are not always situated in the broader context of relevant literature, inscriptions, and papyri. The entry for the adjective ὑπέρακμος exemplifies this point. Until LSJ, apart from Epiphanius and Eustathius, only 1 Corinthians 7.36 was cited and only one interpretation was mentioned: ‘sexually well-developed’. LSJ added ‘Soran(us) 1.22’, but did not comment on whether a man or, as in Soranus, a woman was being described. The discussion of 1 Corinthians 7.36 by A.C. Thiselton, to take one commentator as an example, illustrates how little is said about the difficulty of this adjective in this context. He has one excursus entitled ‘“parthenos” and “hyperakmos”’ and another entitled ‘The meaning of ὑπέρακμος’, which summarize the range of interpretations preserved in the Greek and Latin Fathers and their reception by the Reformers.⁶⁵ He noted that other commentators include a ‘special appendix’⁶⁶ on this adjective. His note begins with a quotation that the verse is ‘one of the most difficult and controversial in the N(ew)T(estament)’.⁶⁷ Further, lines 11 and 18 of the one hundred lines of Praecepta Salubria, which may date to the first century BCE,⁶⁸ remain uncited, even in LSJ Revd Suppl. Both lines read ἂν δ’ ὑπέρακμος, τῆσδε τὸ πλῆρες σκόπει. The first is followed by καὶ γαστέρος κένωσιν ἐκ τῶν σκυβάλων,/τό τ’ ⁶³ Richards 1925, 76 noted the ‘arbitrary attitude’ to personal names. LSJ’s σύζυγος is a relevant curiosity: ‘masc., yoke fellow, comrade, . . . , Ep.Phil.4.3 (unless pr(oper) n(ame));’ LS7 8 did not mention this Philippians passage. The idea that the second word in γνήσιε σύζυγε should be capitalized as a personal name was presented in the margin by Westcott and Hort 1881. Such an interpretation is attested as early as John Chrysostom (PG LXII 260). ⁶⁴ This reflects the attention that one of Liddell and Scott paid to Herodotus and that the other paid to Thucydides in the preparation of the first edition. See LS¹, vi. ⁶⁵ Thiselton 2000, 594 and 594 8. ⁶⁶ Thiselton 2000, 594. ⁶⁷ Thiselton 2000, 593. ⁶⁸ Winter 1998, 74 drew attention to this text in his analysis of ‘important evidence . . . avail able on the Thesaurus Linguae Graecae CD Rom’. Since the Praecepta Salubria was published by U.C. Bussemaker in 1862, the edition cited by the TLG, it could have been cited from LS⁶ onwards, even if no edition of the text had been published earlier.

The Greek of the New Testament

161

ἀκριβῶς ἄδιψον ἄπικρον στόμα. The second is followed by τήρει τε γαστρὸς ἐκϕόρησιν καὶ στόμα / δίψους ἄγευστον, εἰ καὶ μή γε πικρίας. Although the Liddell and Scott tradition should not be expected to discuss New Testament hapaxes and interpretative cruces as thoroughly as a commentary or a specialized lexicon, some note of caution would be appropriate, as with ἐπερώτημα, and more comprehensive citation would be welcome. It would be wiser to separate Soranus and 1 Corinthians than to leave the implication that the gloss suits both passages equally well. From these observations on Ἰησοῦς, we are left with the impression that the treatment of the New Testament in the Liddell and Scott tradition is somewhat superficial, whether deliberately or unintentionally. On the one hand, it is as if no uncertainty surrounded ὑπέρακμος and, on the other, it is as if attestations in difficult passages could be ignored for convenience, as with ἐπερώτημα.⁶⁹

10.3. PERSONALITIES Both Henry George Liddell and Robert Scott, like all fellows of Oxford and Cambridge colleges of their day, were ordained as priests in the Church of England. The development of the Liddell and Scott tradition is thus situated in a certain confessional setting and distinguished from the Greek-English lexica and lexical projects of non-conformists, such as the dissenter/Unitarian Gilbert Wakefield⁷⁰ or the Congregationalist Alexander Souter, and from the Lutheran tradition of Walter Bauer⁷¹ and Frederick Danker. The ecclesiastical careers of Liddell and Scott differed significantly, as did their involvement in the Oxonian theological controversies of their time. While Liddell, the son of an Anglican minister,⁷² remained at Oxford as Dean of Christ Church (1855–91) after a spell at Westminster School (1846–55),⁷³

⁶⁹ Cf. s.vv. διαλογισμός and διηγέομαι and see the discusssion in section 10.4 below. ⁷⁰ On Wakefield, see Clarke 1945, 78 81 and 139 40; Hiscock 2010; and Stray 2010 11, 530. ⁷¹ See Lee 2003, 143 51. ⁷² Indications that Liddell was devout appear throughout Thompson 1899 and in its extracts from Liddell’s autobiography and correspondence. For example, Liddell reflected on the lasting impact of his confirmation throughout his time at Charterhouse (Thompson 1899, 10). In general, see Thompson 1899, 14 15, 20, 21, 40 1, 49 50, 83, 129, 251, and 272 3. Long before ‘Desert Island Discs’, Liddell wrote of Boswell’s Life of Johnson, ‘I think if I was allowed only one book (beside the Bible) to take with me to a desert island, that would be the book’ (Thompson 1899, 254). For those who influenced Liddell, see Thompson 1899, 26 7, 42 4, and 186 7. ⁷³ Thompson 1899, 52 4: ordained (June 1838); Select Preacher (1842 and 1847); Whitehall Preacher (1845); Domestic Chaplain to Prince Albert (1846 ). Liddell’s marriage necessitated that he vacate his studentship at Christ Church, but in Gaisford’s words, ‘love and lexicography were not incompatible’ (Thompson 1899, 59). Likewise, Scott left Balliol to marry in 1840.

162

Patrick James

Scott’s appointments were mainly ecclesiastical and outside Oxford.⁷⁴ Liddell’s academic activities certainly involved biblical scholarship: he taught Hebrew at Westminster, he taught Lord Wemyss to read the New Testament in Greek ‘like a Christian’ and, while working on the Lexicon, he lectured weekly on Acts.⁷⁵ Scott, the son of the Rector of Bondleigh in Devon, returned to Oxford as Master of Balliol (1854–70), because Benjamin Jowett’s theological orthodoxy was in question,⁷⁶ and he was Dean Ireland’s Professor of Exegesis (1861–70). He was a prominent member of the New Testament revision company (1870–81).⁷⁷ While Liddell published his History of Ancient Rome,⁷⁸ Scott published two sets of sermons, one of which I have seen,⁷⁹ and a commentary on the Epistle of James, which I have not seen. Scott’s appointments and publications represent a great amount of engagement with current affairs in the Church of England. In contrast, as Christopher Stray, Chapter 1, this volume, has discussed, Liddell saw lexicography (with his college duties and artistic pursuits) as a way to escape from the doctrinal controversies⁸⁰ around John Henry Newman, the Tractarians, and the Oxford Movement, rather than to address them on a philological basis. The treatment of New Testament words does not consistently reflect the Anglicanism of Liddell and Scott. There is no claim that an ἐπίσκοπος was a ‘bishop’ in the New Testament. (LSJ, like LS7-8, gives ‘ecclesiastical superintendent’). A πρεσβύτερος was always an ‘elder of the Christian church, a presbyter’, not ‘a priest’. However, an Anglicanism is reflected in the terminology of baptism. Christopher Stray has discussed how Liddell and Scott’s entries for βαπτίζω featured in debates about baptism and how the entry was altered as a result of those debates.⁸¹ The point at issue was the mode of baptism, not the theology behind the practice.

⁷⁴ Ordained in 1835; Balliol College living of Duloe in Cornwall (1840 50); Rector of South Luffenham, Rutland (1850 4); Prebendary of Exeter (1845 66); Select Preacher (1853 4 and 1874 5); Dean of Rochester (1870 87). On the other hand, Scott was Master of Balliol 1854 70. See Craik and Smail 2004. ⁷⁵ Thompson 1899, 47, 50, and 107. ⁷⁶ See Stray, Chapter 1, this volume, Atherstone 2003 and Thompson 1899, 185. Scott’s involve ment in the controversy can be followed through the correspondence edited by Prest 1966. ⁷⁷ He attended 337/405 of the company’s meetings, fewer than the textual critics F.H.A. Scrivener (399/405) and F.A.J. Hort (362/405), but more than B.F. Westcott (304/405) and J.B. Lightfoot, then Hulsean Professor of Divinity at Cambridge (290/405). See Chapman 2016. ⁷⁸ Thompson 1899, 121. ⁷⁹ Scott 1851. I have also read two individually published sermons: Scott 1839 and 1841. ⁸⁰ See Thompson 1899, 45, 66 7, 187, 188 9, 246 7, 248, and 275 6. Thompson 1899, 172 5 reports that Liddell and E.B. Pusey took action against ‘a Roman ecclesiastic’, who ‘was in the habit of frequenting undergraduates’ rooms . . . and trying to unsettle their minds’. Liddell was ‘dissatisfied’ with the Revised Version, according to the recollections of the Secretary to the Delegates of Oxford University Press. See Thompson 1899, 203 4. ⁸¹ In an unpublished paper, ‘Philology and religion in 19th century Oxford and Cambridge’, Jan. 2017. I am grateful to Christopher Stray for drawing my attention (pers.comm.) to the difference between the content of the earliest entry for βαπτίζω and the entry as it remained from LS² onwards. The glosses in LS¹ s.v. βαπτίζω I ‘to pour upon, drench’, for which no Greek author was cited, were removed in LS², never to return.

The Greek of the New Testament

163

It should be noted at the outset of this discussion that LS7-8 shows greater theological rigour than LSJ in this regard. The fuller wording in LS7-8 s.v. ϕωτίζω, which was based on its wider coverage, was not retained. LSJ reads (s.v. II 4): ‘illuminate with spiritual light, ὁ ὑπὸ τοῦ θεοῦ πεϕωτισμένος Corp.Herm. 9.3, cf. 13.18 (Pass.). b. in a special sense, baptize, in Pass., Ep.Heb.6.4, 10.32.’ LS7-8 read (s.v. II 4): ‘to enlighten spiritually, and (in a special sense) to baptize, Eccl., cf. Ep.Hebr.6.4, 10.32 and v. ϕώτισμα.’ Comparison with the fuller wording in LS7-8 and lack of sub-division in II 4 shows that LSJ’s II 4 b is to be understood from II 4 a and clarifies ‘in a special sense’. Its ‘cf.’ does not translate the Hebrews instances. The force of this arrangement is confirmed by the invitation in LS7-8 (necessarily omitted by LSJ) to consult the entry for ϕώτισμα, which read: ‘an enlightening:—but only found in Eccl. sense, baptism or (properly) the enlightenment and inward grace of baptism, for οἱ αἱρετικοὶ βάπτισμα ἔχουσιν, οὐ ϕώτισμα, Jo. Chrys.; v. Suicer., and v. ϕωτίζω II. 4.’ Since Patristic material had been relegated to Lampe, LSJ was left with only ‘ϕώτ-ισμα, ατος, τό, phase, of the moon, Eustr. in EN31.33’ and interpreters of ϕωτίζω in Hebrews lost the benefits of LS7-8’s qualification (‘properly’) and John Chrysostom’s comment. LSJ’s treatment of παλιγγενεσία in Titus 3.5 retains the wording of LS7-8 and earlier editions unaltered: ‘regeneration by baptism, διὰ λουτροῦ παλιγγενεσίας’.⁸² Many commentators associate or equate λουτρόν with βάπτισμα, but such an interpretation is not beyond question.⁸³ Note that LSJ’s entry takes λουτρόν as dependent on παλιγγενεσία (i.e. λουτροῦ παλιγγενεσία ‘regeneration associated with washing’) rather than παλιγγενεσία as a genitive that qualifies λουτρόν (i.e. λουτρὸν παλιγγενεσίας ‘washing related to regeneration’). Since the text continues οὗ ἐξέχεεν ἐϕ᾽ ἡμᾶς and since the distant λουτρόν is less likely to be the antecedent of the relative pronoun οὗ than the adjacent πνεῦμα ἅγιον, the outpouring of the Holy Spirit is in view (as in Acts 2.17 and

⁸² The gloss ‘resurrection’ for Matthew 19.28 is also questionable. It is likely that ‘renewal’ or ‘regeneration’ of all of creation is in view, given Jewish and Christian expectation of ‘a new heaven and a new earth’ (as in Isaiah 65.17 and 66.22, and continued in 2 Peter 3.13 and Revelation 21.5). The meaning ‘renewal, regeneration’ would fit more closely with Philo’s reference to Noah and his sons as παλιγγενεσίας ἡγεμόνες, cited by LSJ s.v. 1 and to the Stoic use of the term, as in LSJ s.v. 2: ‘rebirth of the κόσμος’. ⁸³ See, e.g., Knight 1992, 350.

164

Patrick James

2.33), not subsequent baptism with water itself.⁸⁴ A less interpretative translation of the context and a fuller quotation would have resulted in a more appropriate treatment: ‘regeneration , διὰ λουτροῦ παλιγγενεσίας καὶ ἀνακαινώσεως πνεύματος ἁγίου’. Fuller wording leaves open the possibility, itself plausible, that the phrase διὰ λουτροῦ παλιγγενεσίας concerns not baptism with water specifically, but the washing depicted by baptism with water,⁸⁵ if baptism with water is in view at all. The only other instance of λουτρόν in the New Testament is qualified by τοῦ ὕδατος ἐν ῥήματι (Ephesians 5.26). The instance in Titus 3.5 may likewise not involve physical water. We may now consider Sir Henry Stuart-Jones himself. Said to be the ‘ideal Editor’ and noted for ‘his wide range of knowledge and his exact scholarship, his persistent devotion to his task even in periods of ill health, his tactful assiduity in consulting experts and his skill in co-ordinating their results’,⁸⁶ Jones made contributions on points of New Testament semantics and exegesis. One of Jones’ contributions is found s.v. ἀπαρχή 7: ‘birth-certificate of a free person, PTeb.316.10 (i AD), PGnom.131 (ii AD):⁸⁷ perh. metaph. in Ep. Rom.8.23.’. Here, Jones incorporated, with appropriate reservation, a parallel between section 47 of the Gnomon of the Idios Logos and Romans 8.23. Wilhelm Schubart had already cited Romans 8.23 in his 1919 publication of the Gnomon.⁸⁸ Jones advocated its exegetical value in his inaugural lecture in March 1920 (‘Fresh Light on Roman Bureaucracy’) and, in print, in 1922. It is noteworthy that in this entry a New Testament passage was interpreted in the light of papyrological discoveries, a subject with which this chapter will close. Sections 5 and 6 are similar to the content of the entry in Moulton and Milligan,⁸⁹ but the references to the papyri in section 7 are independent. All is not well in this entry. The end of the final section of an entry should always attract suspicion and, in this instance, that suspicion of a late addition and imperfect revision is justified. One of the citations, P.Tebt. II 316.10,⁹⁰ also appears as the sole explicit citation in section 5: ‘entrance fee, PTeb.316.10 (i A.D.), al.’.⁹¹ LSJ here followed Moulton and Milligan, but omitted the

⁸⁴ Whether the genitive (καὶ) ἀνακαινώσεως (πνεύματος ἁγίου) is to be construed with διά or, like παλιγγενεσίας, with λουτροῦ is debated: see Wright 1992, 342 4. ⁸⁵ For comparison, note the fuller wording of Souter’s entry (1916, 187) for παλινγενεσία (spelled in accordance with the manuscript evidence): ‘(b) rebirth of the individual life following on or typified in baptism, Tit. iii 5’. ⁸⁶ Anon. 1940, xiii. ⁸⁷ The Gnomon was published in Schubart 1919 as BGU V 1210. Line 131 reads ἐὰν δὲ καὶ ὑπὸ ἀμϕοτ̣έ̣ρ̣[ων ἀπ]αρχὴ τέκνων τεθῇ, τηρεῖται τοῖς τέκνοις ἡ πολιτε̣ία: ‘But if the children’s [ἀπ]αρχή is presented by both (parents), citizen status is retained for their offspring.’ ⁸⁸ Schubart 1919, 23. ⁸⁹ Moulton and Milligan 1914, 54. ⁹⁰ Grenfell, Hunt, and Goodspeed 1907, 116 20 and Mitteis and Wilcken 1912, 173 5. ⁹¹ At the very least, ‘al.’ here refers to the other instances of this word on this papyrus whose text is a series of declarations by ephebes that they had not appropriated another’s ἀπαρχή (lines 49 and 80) and that they met the other necessary conditions for enrolment.

The Greek of the New Testament

165

suggestion (and no more than that) of the editors of the papyrus that the fee was paid by ephebes on enrolment in the Alexandrian demes.⁹² After the publication of the Gnomon, this papyrus and others were reassessed. Although section 7 was added for ‘birth-certificate’, section 5 was not removed, despite the lack of evidence for it among the papyri cited.⁹³ As for Romans 8.23, ‘birth-certificate’ makes some sense as a translation, but is not unequivocally suitable. The text reads: οὐ μόνον δέ, ἀλλὰ καὶ αὐτοὶ τὴν ἀπαρχὴν τοῦ πνεύματος ἔχοντες, ἡμεῖς καὶ αὐτοὶ ἐν ἑαυτοῖς στενάζομεν υἱοθεσίαν ἀπεκδεχόμενοι, τὴν ἀπολύτρωσιν τοῦ σώματος ἡμῶν.

Many translations use ‘first-fruits (of the Spirit)’, the usual translation word for ἀπαρχή throughout English versions, but ‘the birth-certificate that consists of (or “is from”) the Spirit’ has some merit in the context of adoption to sonship (υἱοθεσία) and in the broader context of being brought to life, of inheritance, of liberation from condemnation and slavery (Romans 8:1, 11, 14–17, and 21). The notion of the ‘documentary’ evidence given by the Spirit fits with αὐτὸ τὸ πνεῦμα συμμαρτυρεῖ τῷ πνεύματι ἡμῶν ὅτι ἐσμὲν τέκνα θεοῦ (Romans 8.16), as Jones remarked.⁹⁴ By contrast, there is nothing in the context that chimes with the notion of first-fruits of dough (cf. Romans 11.16) or as an offering (LSJ s.v. 2) or, as a metaphor, in reference to the first instance that guarantees that others will follow (cf. Romans 16.15, 1 Corinthians 15.20, 15.23, and 16.15, 2 Thessalonians 2.13, and James 1.18). Paul’s idea seems to be that the Spirit is the ἀπαρχή of future redemption of the body just as the Spirit is the ἀρραβών (2 Corinthians 1.22 and 5.2), specifically of the inheritance that awaits Christians (Ephesians 1.14).⁹⁵ The Holy Spirit is the pledge of the redemption and inheritance to come. If Paul was making the same point that ἀρραβών makes elsewhere, it remains to be asked why he did not use ἀρραβών in Romans 8.23 as well. What was the force that ἀπαρχή conveyed in that context that ἀρραβών would not convey? Although BDAG reported the proposal associated with Jones as a possibility,⁹⁶ the idea has not been warmly welcomed in lexicography⁹⁷ or in New Testament ⁹² See Grenfell, Hunt, and Goodspeed 1907, 120 for the suggestion for this context, the meaning of ἀπαρχή in some papyri, and its obscurity in others. ⁹³ For more on ἀπαρχή in the New Testament and in the papyri, see Aune 2003; Erlemann 2001; and Taubenschlag 1959, 211 21. Aune 2003, 108 registers a deficiency in LSJ’s treatment. ⁹⁴ Jones 1922, 283. ⁹⁵ This analysis is the consensus of the commentators cited in n. 99 below. See also Oke 1957, 456 7. ⁹⁶ BDAG s.v. 2 ‘ “birth certificate” also suits the context of Romans 8.23’. BDAG s.v. 1bβ also includes Romans 8.23, as its only New Testament citation (hence, s.v. 2 ‘also suits’). BDAG s.v. 2 incorporates M.Chr. II/2 372.4 and 7 as well as P.Flor. I 57.81 (and 86 and 89) from Moulton and Milligan 1914: 54. DGE s.v. ἀπαρχή did not accept Jones’s interpretation. Romans 8.23 is cited s.v. ἀπαρχή B I 3, but the birth certificates of P.Gnomon 131 and P.Flor. I 57.81 are s.v. B II. ⁹⁷ Erlemann 2001, 70 3 does not consider ‘birth certificates’ at all. Taubenschlag 1959, 220 1) cited Jones 1922, but as by Preuschen and as dated to 1921. Lampe 196 8, 177 s.v. ἀπαρχή B 5 b (on Romans 8.23) contains no hint of ‘birth certificates’ and quotes John

166

Patrick James

studies. Of English commentators in the second half of the twentieth century,⁹⁸ C.K. Barrett, C.E.B. Cranfield, D. Moo, and J. Murray did not mention it, J.D. G. Dunn rejected it,⁹⁹ and F.F. Bruce, the most positive, only went so far as to comment: ‘Although [an identification card] is not precisely what Paul means, [Paul’s early readers] would not have been far astray if they did make this inference, as something of this sort is implied with the “sealing” of the Spirit’ (he then cites 2 Corinthians 1.22 and Ephesians 1.14 and 4.30). Of German scholars, Ernst Käsemann rejected the meaning ‘birth-certificate’ as a ‘seltene Bedeutung’ and ‘ganz unwahrscheinlich’.¹⁰⁰ In his rejection and by citing Horst Balz, Käsemann indicated that the proposal was still debated by some, not only in Germany, but also in England. Balz¹⁰¹ does not refer to Jones directly, but via C.C. Oke.¹⁰² Jones’s other clear contribution to New Testament lexicography concerns σπιλάς (B) in Jude 12 and arose in dialogue with A.D. Knox.¹⁰³ The treatment in LS7-8 was as follows:¹⁰⁴ ‘σπῐλάς (B), άδος, ἡ, = σπίλος (ὁ), spot, κατάστικτον σπιλάδεσσι πυρσῇσιν Orph. L. 614:—in Ep.Jud.12, either σπιλάς spot, or σπιλάς rock¹⁰⁵ will suit the sense.’ Chrysostom’s explanation of τὴν ἀπαρχὴν . . . ἔχοντες as τῶν μελλόντων ἤδη γευσάμενοι (wording that recalls Hebrews 6.4 6). ⁹⁸ Barrett 1957, 167; Bruce 1985, 164; Cranfield 1975, 417 18; Dunn 1988, 473 4; Moo 1996, 519 20; and Murray 1967, 156 7. It might be unreasonable to expect the proposal to have been discussed in a commentary as early as Dodd 1932, but Jones had published it in 1922 and the relevant part of LSJ had been published in 1925. ⁹⁹ Dunn 1988, 474 argued that despite its ‘apparent attractiveness’, a birth certificate here would ‘throw the thought in some confusion a birth certificate already issued while the birth travail is still in progress!’ Mixed metaphors are not unknown in Paul, as Wright 2005, 74 noted on 1 Thessalonians 5.3: since the thief will come in the night and the woman will go into labour, do not get drunk, but stay awake and put on your armour. Cf. the nexus in Ephesians 4.14, itself already in the context of the image of the body of Christ. Paul’s point is that the groaning of all creation is ἀχρὶ τοῦ νῦν (Romans 8.23), the time when Christians have the ἀπαρχή already, but wait, with the rest of creation (Romans 8.19 21 and 25), for adoption and redemption (8.23: two more metaphors) and continue to groan (which may, or may not, be a birthing metaphor). Romans 8.24 5 reflects the temporal tension of ‘the now and not yet’. ¹⁰⁰ Käsemann 1973, 226. ¹⁰¹ Balz 1971, 56. ¹⁰² Oke 1957. Balz actually referred to an abridgement of this article by Greenlee 1960. Oke attributed the suggestion to Jones and referred to Milligan 1922. ¹⁰³ Alfred Dillwyn ‘Dilly’ Knox (1884 1943) of King’s College, Cambridge, was the son of an evangelical and the brother of an Anglo Catholic priest and a Roman Catholic priest. In cryptography, he is known for his services in both world wars, especially at Bletchley Park. In Classics, he is known for the completion of Walter Headlam’s edition and commentary on the recently rediscovered Herodas. See Batey 2006. The scope of his Greek interests is captured in the dedication to Knox in Wright (1932, v), which proclaims, as if in Knox’s own words, ‘I would as soon read Dio as Xenophon, Aristides as Demosthenes, Ælian as Aristotle.’ ¹⁰⁴ The single entry in LS⁶ had sections for a rock (σπιλάς (A) in LSJ), for clay (σπιλάς (C) in LSJ), and for a spot (σπιλάς (B) in LSJ). ¹⁰⁵ That is, σπιλάς (A) in LS7 8.

The Greek of the New Testament

167

The formatting of σπιλάς (B) is remarkable. Instead of listing Jude 12 as a citation for the meaning ‘rock’ or ‘spot’, an addition restates the headword for each translation word. Jones’s slight addition for LSJ becomes obvious through comparison. ‘σπῐλάς (B), άδος, ἡ, = σπίλος (ὁ), spot, κατάστικτον σπιλάδεσσι πυρσῇσιν Orph.L.620:–in Ep.Jud.12, σπιλάς spot is prob. in view of 2 Ep.Pet.2.13; cf. also σπιλάς (C).’ The ‘rock’ has gone. 2 Peter 2.13 tipped the balance against it.¹⁰⁶ A crossreference to ‘σπιλάς (C)’ has appeared, but the headword and translation are still repeated. LSJ’s new entry for σπιλάς (C), a feminine noun, reads as follows: ‘storm, squall, Plu.2.476; ἐκραγείσης ὥσπερ ἐν εὐδίᾳ σπιλάδος ib.101b; ἄνδρας αἰϕνιδίῳ σπιλάδι κατασεισθέντας Hld.5.31, cf. AP7.382.4 (Phil.): cf. κατασπιλάζω II.’ Knox had argued that σπιλάς either referred to a wind or described a wind as ‘foul’ or ‘dirty’ in two passages in Plutarch and other later authors.¹⁰⁷ (LS7-8 had reported no such meaning.) On that basis, he had argued that ‘in such company [with clouds driven by the winds, withered trees, waves, and planets] winds are more naturally mentioned than spots or rocks’ in Jude 12. Jones clearly accepted Knox’s comments, but only to a certain extent. LSJ’s citations for σπιλάς (C) and its cross-reference to κατασπιλάζω II reflect Knox 1913. However, Jude 12 is not mentioned s.v. σπιλάς (C), even as a possibility for consideration. It is clear that Jones had revised, or was in the process of revising, the treatment of the σπιλάς homonyms and of κατασπιλάζω by 1922 (long before those entries were published), since the entry for κατασπιλάζω II contains a cross-reference to σπιλάς (C) and since Jones himself responded at that time to articles of the previous decade¹⁰⁸ to supply a further citation and to correct the references to Patristic and Byzantine sources in Dindorf ’s Thesaurus Linguae Graecae, the source of Knox’s citations. While Knox had noted that the root in κατασπιλάζω gave indirect evidence for σπιλάς as a storm or a squall, Jones adduced evidence for that verb as early as Philo of Alexandria from a text recently published by J.R. Harris¹⁰⁹ and, thus, before the time of Jude. Jones strengthened Knox’s case, but LSJ only reflects it partially.

¹⁰⁶ Both epistles describe οὗτοι at length. Jude 12 refers to οἱ ἐν ταῖς ἀγάπαις ὑμῶν σπιλάδες συνευωχούμενοι ἀϕόβως which corresponds to οὗτοι . . . σπίλοι καὶ μῶμοι ἐντρυϕῶντες ἐν ταῖς ἀπάταις αὐτῶν συνευωχούμενοι ὑμῖν in 2 Peter 2.12 13. ¹⁰⁷ Knox 1913, 547 9. ¹⁰⁸ Jones 1922, 282 cited Knox 1915, 78, needlessly. ¹⁰⁹ Harris 1886, 28.

168

Patrick James

We must note that the reformulation of the σπιλάς homonyms involved very little change to what had been printed in LS7-8. The editorial intervention was minimal, perhaps a reflection of LSJ’s origins as ‘a paste-up of the eighth edition column by column’.¹¹⁰ From the headword to ‘in Ep.Jud.12’, the changes are factual (‘620’ for ‘614’) or typographic (italics and the division of πυρ-σῇσιν). Then, the alterations become more extensive, but the repetitions of the headword and gloss remain. Although the parallel and the crossreference were added at the end of σπιλάς (B), the use of σπιλόω (Jude 23)¹¹¹ was not mentioned as further corroborative evidence. Other parts of Knox’s case are absent. Although σπιλάς (C) was added, it was added as a feminine noun, while Knox had taken it as an adjective that qualified the masculine noun ἄνεμος that was to be understood in the context.¹¹² LSJ directs the reader of Jude from σπιλάς (B) to (C), but offers no confirmation there that ‘storm, squall’ is a possibility in the epistle. Although ‘storm, squall’ had the implicit support of LSJ and ‘miry wind’ alone is presented by Alexander Souter,¹¹³ it has not found favour more widely. BDAG gives only two meanings (‘a rocky hazard hidden by waves, . . . ’ and ‘that which soils or discolours’), but does cite Knox 1913 and 1915 and Jones 1922 at the end of its entry: ‘(dirty, foul wind)’. As an example of a recent commentator, Richard Bauckham reports Knox’s suggestion and Jones’ response as the third of his four hypotheses, only to reject the idea in favour of ‘rocks’ or ‘dangerous reefs’, because that is ‘the natural meaning of the word and makes excellent sense’ and because ‘it occurs too early to be connected with the later wind metaphor’.¹¹⁴ Storms appear in no English translation that I have been able to consult.¹¹⁵ There is less opportunity to consider the impact of Roderick McKenzie on the lexicography of the New Testament. Although he touched upon points of New Testament philology in his ‘Lexicography and Grammar’ contributions to ‘Bibliography: Graeco-Roman Egypt—Part I: Papyrology’,¹¹⁶ he is credited chiefly, and perhaps to his shame,¹¹⁷ for approving ‘the etymologies presented . . . in almost every case’.¹¹⁸ ¹¹⁰ Jones 1941, 10. ¹¹¹ See Bauckham 1983, 85. ¹¹² Knox 1913, 548 was reluctant to ‘build much’ on the masculine plural definite article, ‘though σπιλάς is adjectival and probably masculine of a wind’. ¹¹³ Souter 1916, 239 followed Knox (1913, 548 9) closely: ‘adjectivally used with ἄνεμος understood, . . . perhaps of its effect on the water’. ¹¹⁴ Bauckham 1983, 77 and 85. ¹¹⁵ That is, by means of http://biblehub.com/jude/1 12.htm (accessed 30 May 2017). ¹¹⁶ See McKenzie 1932a, in which he refers to the ‘ἄρτος ἐπιούσιος controversy’, 1936, and 1937, in which he notes a study of the perfect in John. McKenzie 1932b also reviewed Thack eray’s Lexicon to Josephus. ¹¹⁷ See Chadwick 1994, 2 3 and 1996, 8. ¹¹⁸ Jones 1925, x. There and in Anon. 1940, xiii, McKenzie received greater acknowledgement in print than Chadwick 1996, 7 recognized. For etymology in LSJ, see Katz, Chapter 5, this volume. Even McKenzie’s plausible proposal (1925, 210) for ὑπο(σ)σείω at Odyssey 9.385 as ‘set in rapid motion’ on the basis of a Sanskrit cognate, does not appear in LSJ (‘rotate, spin round’ ).

The Greek of the New Testament

169

10.4. SOURCES AND INFLUENCES The Preface names among New Testament lexica that were ‘generally sufficient’ for the ‘ordinary purposes of revision’ only those of Heinrich Ebeling (1913) and Francisco Zorell (1911).¹¹⁹ The former, which reached a third edition (1929) and has been described by John Lee as ‘the forgotten lexicon of the twentieth century’, made use not only of literary texts, but also of the growing body of documentary evidence.¹²⁰ The latter, in Latin for a Roman Catholic readership,¹²¹ reached a fifth edition (1999) and likewise made use of attestations in papyri. Before considering the other resource that Jones named in the Preface, there is at least one conspicuous absence: the 1910 Handwörterbuch of Erwin Preuschen (1910), the predecessor of the Bauer tradition, first as PreuschenBauer in 1928, and then of BDAG and its most recent offspring, the 2009 Concise Greek-English Lexicon of Danker and Krug. Jones and McKenzie are unlikely to have been able to conclude, from its humble beginnings with Preuschen, that the then nascent Bauer tradition, in time, would lead to the most widely used New Testament lexicon, BDAG.¹²² This work may not have been ‘generally sufficient’ because its lack of citations from literature and documentary texts made it less useful than lexica that served up such material for Jones and McKenzie. That possibility would be of interest in relation to the place of New Testament Greek in LSJ, a subject to which we will soon turn. Another candidate as a conspicuous absence is Souter’s Pocket Lexicon, like LSJ an early twentieth-century product of Oxford University Press. Souter himself, later the Editor of the Oxford Latin Dictionary (1933–9),¹²³ was named by Jones as one to whom non-specific ‘special thanks’ were due.¹²⁴ Although the scale of Souter’s Pocket Lexicon eliminated its utility as a source of citations, it remains exemplary as an English-language lexicon based on modern critical editions via the 1897 Concordance of W.F. Moulton and A.S. Geden, assisted by the 1908 Prolegomena of Moulton, and drawing on papyrological parallels certainly by means of Moulton and Milligan, but also apparently from Ebeling and from Adolf Deissmann.¹²⁵ Milligan called it ‘a marvellous multum in parvo’.¹²⁶ Souter’s use of ‘information culled from the

It was reported, with approval, by W.B. Stanford (1958, 360) and is reflected in LSJ Revd Suppl. s.v. (‘set in violent motion, twist back and forth’), but is treated only indirectly by Heubeck and Hoekstra 1989, 34. ¹¹⁹ Jones 1925, ix. ¹²⁰ Lee 2003, 139 and 141 2. ¹²¹ Lee 2003, 139, 140 1, 152 n. 3, and 364. ¹²² Since Bauer’s revision of Preuschen is mentioned as an aid to Moulton and Milligan (1929, 5), it may have had an indirect influence on Jones and McKenzie. ¹²³ Glare 1968 8, v and vii. ¹²⁴ Jones 1925, x. ¹²⁵ Souter 1916, v vii. ¹²⁶ Milligan 1929, v; cf. 1915, 5 and 1919, 5. Lee 2003, 285 thought that Souter, in his independence from other lexica, was correct about ἕξις, but had ‘no effect on the main stream’.

170

Patrick James

Latin Fathers’ was characteristic and his goals noble: ‘unless I am mistaken, the newer knowledge sheds a flood of light on passages hitherto misunderstood or regarded as unprofitable (e.g. I Cor. x. II, James i. 3, I Pet. ii. 2), and sweeps into the dustbin a deal of the well-meant but hair-splitting theology of the past . . . ’.¹²⁷ Souter’s Pocket Lexicon seems not to have influenced LSJ, to judge by the passages quoted just above.¹²⁸ (The same examples can also be used as evidence that the contributions of Moulton and Milligan and of Deissmann were not systematically incorporated into LSJ.) However, Souter’s personal influence on the treatment of New Testament vocabulary may be tentatively explored. In LSJ’s entry for ἀγαπητός, John Lee has rightly detected the detrimental influence of C.H. Turner.¹²⁹ Since Souter wrote in support of Turner’s argument,¹³⁰ he may have been an additional advocate of Turner’s proposal to Jones and McKenzie.¹³¹ There is also the case of ὕσσωπος. The 1940 Addenda and Corrigenda and the 1968 Supplement have s.v. ὑσσωπός (sic) ‘after “Ev.Jo. 19.29” add “(nisi leg. ὕσσῳ)”’. Souter certainly supported this conjecture without reservation (s.v. ὕσσωπος): ‘In John xix 29 ὑσσώπῳ is a graphic error for ὑσσῷ (pilum), pike’.¹³² However, since, for example, the New

I once had the opportunity (July 2009) to ask Frederick Danker for his opinion of Souter’s Lexicon. When I showed him my copy, he seemed unfamiliar with the work. Indeed, Souter is absent from the abbreviations listed in Danker 2000, lvii. However, the 1920 reprint of Souter 1916 is cited in full s.v. ψάλλω. ¹²⁷ Souter 1916, v. ¹²⁸ I think that Souter had in mind (respectively): καταντάω ‘of property, I come down (descend) by inheritance to an heir’ and τέλος ‘of the spiritual revenues of the ages, I Cor. x. II’ (note that Souter gave citations only in exceptional circumstances); δοκίμιον ‘(neut. of δοκίμιος genuine, as opposed to alloyed, counterfeit), what is genuine, the approved part, the pure part’; λογικός ‘(b) metaphorical, as contrasted with literal, I Pet. ii. 2’. LSJ’s entries for τέλος and for λογικός make no reference to such interpretations or, indeed, to any New Testament passage. However, LSJ s.v. καταντάω 5 has ‘of an inheritance, κ. εἴς τινα fall to one’s share, 1 Ep. Cor.10.11, POxy.75 (ii A.D.), etc.’ Although LSJ’s entry for δοκιμεῖον does mention James 1.3 (and 1 Peter 1.7) in the context of δοκίμιος as a variant reading, it does so only as a ‘test’ or ‘means of testing’. Anne Thompson would object, rightly, that LSJ’s treatment is in error in that it conflates as one word a formation in εῖος and another in ιος, as if they were (iotacistic) spelling variants of each other. LSJ has no entry either for δοκίμιος or for δοκίμιον, listed as a variant reading s.v. δοκιμεῖον. For the suffixes, see Buck and Petersen 1945, 43 4 and 45. Following Deissmann, Moulton and Milligan 1930, xvi, 167, and 378) used δοκίμιον and λογικός as incontrovertible examples of the merits of using the evidence of the papyri. The entry for τέλος by Moulton and Milligan (1928, 630 1) was published after Souter’s Pocket Lexicon and does not refer to 1 Corinthians 10.11. Moulton and Milligan’s entry for καταντάω (1920, 330) would not have been available, but Moulton 1901, 272 3 had deemed the ‘legal sense’ to be ‘exceedingly appropriate’ in relation to 1 Corinthians 10.11. ¹²⁹ Lee 2010b, 124 5. ¹³⁰ Souter 1926 7, 59 60. ¹³¹ A possibility that Lee (2010b, 136 n. 16) recognized, at my suggestion. ¹³² Souter 1916, 272.

The Greek of the New Testament

171

English Bible admitted this widely-accepted conjecture into its text,¹³³ Souter was by no means its only proponent,¹³⁴ but he is one known to have been associated with LSJ. The addition to the entry for ὕσσωπος was deleted for LSJ Revd Suppl. Jones described Moulton and Milligan’s Vocabulary of the Greek Testament Illustrated from the Papyri and Other Non-Literary Sources¹³⁵ as ‘a most valuable aid’.¹³⁶ The overlap in papyrological (and epigraphic) citations between LSJ and Moulton and Milligan is self-evident. Jones acknowledged his dependence on them, but given the numbers of papyri that were available to Jones and McKenzie, overlap (perhaps)¹³⁷ would be inevitable. LSJ’s entry for διαλογισμός discusses none of the fourteen New Testament instances (LS7-8 cited none), but added five of Moulton and Milligan’s eleven papyri.¹³⁸ Jones expressed his gratitude to Milligan for advance proofs as far as the fifth part (published in 1924).¹³⁹ However, some of the overlap may reflect instead the influence of Jones on Moulton and Milligan, since he is thanked among ‘various friends’.¹⁴⁰ The Preface does not mention E.A. Sophocles’ 1870 Greek Lexicon of the Roman and Byzantine Periods, but at least two points suggest its direct or indirect influence. LSJ contains an entry for ἀρχιεταῖρος: ‘chief friend or companion, LXX 2 Ki.16.16 (due to mistranslation of pr. n.¹⁴¹ ‘Arkī)’.¹⁴² LS7-8 had no more than ‘a chief friend or companion, LXX (2 Regg. 16.16, cf. 1 Par. 27.33). LS⁶, like all preceding editions, lacked the reference and the comparandum. The additional comment in LSJ may be derived from a disquisition in Sophocles: ¹³³ ‘ . . . so they soaked a sponge with the wine, fixed it on a javelin’ with a note ‘So one witness; the others read: on marjoram’. The one witness must be the eleventh century minuscule 476* mentioned by Metzger (2000, 217). ¹³⁴ Moulton and Milligan 1929, 661 included a headword for ὑσσός, a word not elsewhere in the New Testament, and regarded ὑσσώπῳ as ‘probably a graphic error for ὑσσῷ’. I will save further discussion of this conjecture’s history for my forthcoming defence of the transmitted reading. Details of this particular conjecture, one of several for ὑσσώπῳ in this verse, can be found at the Amsterdam Database of New Testament Conjectural Emendation, http:// ntvmr.uni muenster.de/nt conjectures?conjID=cj10089 (accessed 6 March 2018). ¹³⁵ Moulton and Milligan 1914 29. ¹³⁶ Jones 1925, ix. Souter 1916, vi was unrestrained in his enthusiasm for this work. ¹³⁷ To take the Oxyrhynchus papyri for example: P.Oxy. XVII 2156, published in 1927, is the most recently published edition cited by LSJ. Publication over the following eighty nine years had reached P.Oxy. LXXXII 5343. That is, 40 per cent of the Oxyrhynchus papyri available to us were available to Jones and McKenzie. ¹³⁸ For διαλογισμός in the papyri and New Testament, see Lee 2003, 125 and 138, who notes that LSJ ‘offered the correct meaning’. ¹³⁹ Jones 1925, ix. ¹⁴⁰ Moulton and Milligan 1926, 5; cf. 1920, 5. W.G. Waddell, the editor of the Loeb Manetho and, in 1932, author of The Lighter Side of the Papyri (New Castle), appears most consistently: Moulton and Milligan (1920, 5, 1924, 5, 1926, 5, 1928, 5, and 1929, 5). ¹⁴¹ A toponym, to be precise. ¹⁴² Muraoka 2009, 753 included ἀρχιεταῖρος only in his ‘List of Lexemes not Included in this Lexicon’.

172

Patrick James

ἀρχιεταῖρος, ου, ὁ, (ἑταῖρος) chief companion. Sept. Reg. 2, 15, 32. 37. 2, 16, 16.¹⁴³ [This word seems to be a figment. The true reading in the passages referred to most probably is Χουσὶ ὁ Ἀρχί (‫ )ארכי‬ἑταῖρος Δαυίδ. Par. 1, 27, 33 Χουσὶ ὁ πρῶτος ϕίλος τοῦ βασιλέως, where ‫ ארכי‬was apparently mistaken by the transcribers¹⁴⁴ for the Greek prefix ἀρχι . 2, 17, 5. Χουσὶ τὸν Ἀραχί. the Archite.]’

The first edition of Sophocles (1870) appeared in time for LS⁷ (1882), but slightly too late for LS⁶ (1868). The other possible instance of the direct or indirect influence of Sophocles is the treatment of ἐπερώτημα. LSJ added a section (s.v. 3), which ends ‘hence prob., pledge . . . 1 Ep.Pet.3.21’. That suggestion follows an equation with Latin stipulatio on the basis of P.Cair.Preis.1.16 and Cod. Just.8.10.12.3. Only Sophocles refers to the Justinian Code in relation to this noun and his lexicon seems to have been the first to suggest a comparison with 1 Peter 3.21. The papyrus citation may well have come from Moulton and Milligan (1919: 231–2), which repeats Wilcken’s equation of ἐπερώτημα = stipulatio.¹⁴⁵ Ebeling refers only to Herodotus, Thucydides, and ‘UntersuchgByz. . . . Si(racides) 36.3’. Zorell’s entry glosses with stipulatio and sponsio, but gives a non-specific reference to ‘pap. scr. Byz.’ (presumably, this includes the second-century AD text P.Cair.Preis. 1.16). We may add ‘Herr Pfarrer P. Katz of Coblenz’¹⁴⁶ as a contributor to the Addenda and Corrigenda. This is Peter Walters (died 1962), an Anglicized refugee from Nazi Germany, whose library and papers (1929–61), including an annotated proof of Bauer, Arndt, and Gingrich,¹⁴⁷ now reside in the Faculty of Asian and Middle Eastern Studies, Cambridge.¹⁴⁸ Since Katz’s research concerned Philo, the Septuagint, and its versions, we may assume that his contributions related more to those fields than to the New Testament.¹⁴⁹

10.5. NINETEENTH-CENTURY NEW TESTAMENT PHILOLOGY AND LSJ ’S BAS E TE XT Jones and McKenzie were able to draw on significant discoveries and advances in New Testament textual criticism, codicology, palaeography, and papyrology ¹⁴³ Of these passages, only 2 Kingdoms 15.32 and 16.16 contain the sequence ΟΑΡΧΙΕΤΑΙΡΟΣ that could give rise to ἀρχιεταῖρος and, in the former, the Hebrew has no justification for the presence of ἑταῖρος. ¹⁴⁴ The date of the mistake is clarified by occurrences of ἀρχιεταῖρος in discussions of Χουσι in the commentaries of Eusebius, Athanasius, Basil of Caesarea, Asterius, and Didymus the Blind (all fourth century AD), and by Michael Syncellus (eighth/ninth century AD) and Nicetas Seides (twelfth century) in other contexts. ¹⁴⁵ See Preisigke 1911, 1. ¹⁴⁶ Anon. 1940, xiv. ¹⁴⁷ Bauer, Arndt, and Gingrich 1957. ¹⁴⁸ See http://www.ames.cam.ac.uk/library/archive/walters (accessed 15 June 2017). ¹⁴⁹ That much is reflected in Katz 1942.

The Greek of the New Testament

173

in the decades that saw the preparation and publication of the previous eight editions of the Lexicon. In 1843, the year of the publication of LS¹, Constantin von Tischendorf published his transcription of the New Testament text preserved as the underwriting in the Codex Ephraemi Rescriptus. The following year, he encountered part of what would become known as the Codex Sinaiticus. Tischendorf published his first critical edition in 1849. An eighth edition was published between 1869 and 1872 and was accompanied by three volumes of Prolegomena (1884, 1890, and 1894). Samuel Prideaux Tregelles produced a critical edition (1857–72) and the edition of B.F. Westcott and F.J.A Hort was published (1881). The first edition of the New Testament by Eberhard Nestle was published in 1898 and, to date, has reached its twentyeighth edition, as Nestle-Aland 2012. Before LSJ, there was no explicit statement about which critical edition of the New Testament was used for the Lexicon. When that statement came, the base text was not one of the critical editions of Tischendorf, Tregelles, Westcott and Hort, or Nestle, but the 1910 Oxford edition by Souter,¹⁵⁰ whose influence on LSJ has already been considered. The goal of that edition was not to present a critical text, but the Greek text behind the Revised English Version (1881) with a supporting apparatus criticus of readings from the Church Fathers and the early versions as well as of Greek manuscripts (including papyri). Not even LSJ Revd Suppl. acknowledged the status of the Nestle-Aland editions as the standard reference: ‘add K. Aland et al., Stuttgart 1968²’¹⁵¹ refers to the second edition of the United Bible Societies’ Greek New Testament (i.e. Aland et al. 1968), whose third edition (1975) was the first to converge completely with that of the twenty-sixth Nestle-Aland edition (1979).¹⁵² Either of these editions could have been cited by LSJ Revd Suppl. from its conception in 1979. Even then, that ‘K. Aland et al.’ was an addition to Souter, not a replacement. Various LSJ entries show that the Textus Receptus continued to be the base text for LSJ, despite the discoveries and labours of nineteenth-century scholars and that, despite the statement that Souter’s edition was the base text, LSJ’s entries were not brought into conformity with that edition. LSJ cites the compound ἀντιμετρέω both from Matthew 7.2 and from Luke 6.38, but Souter’s text, like Nestle-Aland 2012, has μετρηθήσεται in the former place (the apparatus is silent).¹⁵³ LSJ’s entry for δευτερόπρωτον, the neuter of an adjective that describes a σάββατον, is particularly instructive: ‘prob. corrupt in Ev.Luc.6.1 (no expl. is satisfactory)’. In Souter’s edition, this baffling adjective was removed from the text on the authority, chiefly, of a third-century papyrus

¹⁵⁰ Jones 1940, xxx. ¹⁵¹ LSJ Revd Suppl., xvi. ¹⁵² Aland, Aland, Karavidopoulos, Martini, and Metzger 2000, xi. ¹⁵³ Nestle Aland 2012 marks a variant at Luke 6.38.

174

Patrick James

and of Codices Sinaiticus and Vaticanus.¹⁵⁴ If LSJ had to report this adjective at all, it should have been as a ‘dub. l.’ or perhaps a ‘v. l.’. Section II of LSJ’s entry for κεϕαλαιόω cites Mark 12.4 alone for the meaning ‘smite on the head’. That section is suspicious because no parallels have been found¹⁵⁵ and because it is so different from ‘sum up’ (section I). LSJ does not signal any doubt about the reading in that verse, even though Souter, like the critical editions, had accepted ἐκεϕαλίωσαν (third-person plural aorist of κεϕαλιόω) on the basis of Codices Sinaiticus and Vaticanus. The verb κεϕαλιόω remains unparalleled, but is tentatively reported by LSJ Revd Suppl.: ‘hit on the head (v.l., v. κεϕαλαιόω II).’¹⁵⁶ That is, its treatment depends on that of the reading of outdated editions and not those on which LSJ and LSJ Revd Suppl. claim to rest. However, a hint of the influence of the critical editions is found s.v. ἀνίλεως. LSJ cited ‘Ep.Jac.2.13 (s.v.l.)’. On the evidence of Codices Sinaiticus, Vaticanus, Alexandrinus (ἀναίλεος, with for ), and Ephraemi Rescriptus, nineteenth-century editors, and their successors, accepted ἀνέλεος instead of the Attic form ἀνίλεως (the reading of the Textus Receptus, as retained by Souter’s edition). Although LSJ (s.v. ἀνέλεος) reads simply ‘unmerciful, Ep.Jac.2.13.’, LS⁶ included ‘Lachm(ann) (Vulg. ἀνίλεως)’ before the full stop, evidence that the critical edition of Karl Lachmann had been consulted,¹⁵⁷ at least on this point perhaps because it involved an otherwise unattested word and, perhaps, only by an anonymous contributor.¹⁵⁸

¹⁵⁴ Souter himself (1916, 60) commented ‘a ghost word which has crept into the text of many authorities at Lk. vi 1, by mistake’. ¹⁵⁵ There is an attempt (BDAG s.v. κεϕαλίζω 2) at explanation by means of comparison with ‘our “total” in the sense of demolish’. ¹⁵⁶ I am grateful to Simon Gathercole for asking me about κεϕαλίοω. ¹⁵⁷ See Reynolds and Wilson 2013, 188 9. ¹⁵⁸ Other reservations about the Textus Receptus are implied or found s.vv. ἀνθυπατεύω, ἐπισύστασις, and ἐπίστασις. LSJ s.v. ἀνθυπατεύω inserted a citation of Act.Ap.18.12 between the two citations in LS7 8 to bring the entry in line with the Textus Receptus (and, in this instance, Souter’s edition). (Souter 1916, 23 4) does not have this verb.) Its absence from LS7 8, which cites only Plu.Comp.Dem. Cic.3 and Hdn.7.5, is remarkable given that the verb is not particularly common (24 instances in the whole TLG, of which three are within LSJ’s coverage) and given that the Acts citation might well predate Plutarch. Perhaps an edition that read not ἀνθυπατεύοντος, but ἀνθυπάτου ὄντος, with the nineteenth century editions, was used. Although ἀνθυπατεύοντος was printed by Robinson and Pierpont 2005, that reading does not appear even in the variae lectiones minores in Nestle Aland 2012. LSJ’s entry for ἐπισύστασις cites Acts 24.12 in its section 1, but with caution: ‘(nisi leg. ἐπίστασις)’. The headword is the reading of the Textus Receptus, the alternative that of Souter’s edition and of critical editions based on Codices Sinaiticus, Vaticanus, and Alexandrinus. (Section 3 reads, ‘v. l. for ἐπίστασις, 2 Ep.Cor.11.28’, the headword again being the reading of the Textus Receptus, but the word ἐπίστασις, for which it is a variant, is the reading of the critical editions.) The entry for ἐπίστασις likewise has ‘III. onset, LXX 2 Ma.6.3; ὄχλου Act.Ap.24.12 (nisi leg. ἐπισύστασις)’. Souter 1916, 94 5 has an entry for ἐπίστασις only, the reading he printed in 1910 (the variant was retained in the apparatus).

The Greek of the New Testament

175

The text preserved in Codex Bezae is a particular problem and one that was neither well nor consistently handled. Bezae, which preserves longer versions of Luke and Acts, is named in four entries: s.vv. ἐπιχείρησις (in Acts 12.3, not, as LSJ has it, 12.1), ὀχετός (Mark 7.19), τέσσερες (Acts 11.5), and ὅσος (Luke 5.3). Of those entries, only ἐπιχείρησις is cited from one of Bezae’s additional passages. Some entries, such as those for διαλιμπάνω (Acts 8.24, cf. 17.13), προβιβάζω (Acts 19.33), and συντεχνίτης (Acts 19.25), cite lexical items from additional passages, but they do so only as ‘v. l.’, without specific reference to Bezae. The treatment of the various spellings of κράββατος shows that, although the most recent manuscript discoveries received attention, the results were particularly faulty. The spelling κράβακτος is erroneously cited as a ‘v. l.’ from ‘cod(ex) Alex(andrinus). In fact, it is Codex Sinaiticus¹⁵⁹ that has κραβακτ- in eight of nine instances (the exception is Acts 5.15). Codex Washingtonianus (LSJ’s ‘cod. W’), which was purchased in November 1906 by Charles Lang Freer, a Detroit industrialist, and first published comprehensively by H.A. Sanders in 1912, is cited for the spelling κράβαττος, as if it is found only in that New Testament manuscript. That spelling could have been cited from Codices Vaticanus, Alexandrinus, Ephraemi Rescriptus, or Bezae and, since it was printed throughout by Westcott and Hort and Nestle(-Aland),¹⁶⁰ it is a strong candidate for the headword spelling.¹⁶¹ Other spellings, such as γραβάττοις and κρεβάττοις at Mark 6.55 in Bezae and Washingtonianus respectively, were not reported, even though the former is the spelling used by Latin authors, of whom LSJ cites ‘Virg.Mor.5’.

1 0 . 6. F R O M ‘N .T . ’ TO NT’: THE NEW TESTAMENT A S A CO RP U S A N D C A N O N I N L S J Two innovations in the treatment of New Testament words are evident in LSJ. One was global and reflects one of the distinguishing features of the revision as a whole.¹⁶² The other was the extension of ‘NT’ to a linguistic category as well as a theological one. The first innovation has already been seen in our discussion of the entry for Ἰησοῦς: the inclusion of at least one ¹⁵⁹ The only explicit reference to Codex Sinaiticus, as far as I can tell, in LSJ itself occurs s.v. ψαλμῳδός: ‘Si(racides) 47.9 cod.Sin.’. The Addenda and Corrigenda to LSJ s.v. ξαίνω report the erased text of Sinaiticus at Matthew 6.28 as read by T.C. Skeat (1938) with the aid of ultra violet light. This comment was expanded for LSJ Revd Suppl. ¹⁶⁰ Westcott and Hort 1881 and Nestle Aland 2012. ¹⁶¹ Souter 1910 retained κραββατ throughout, but gave κράβαττος as his headword (1916, 139), the Latin equivalent grabbatus, and a note: ‘Spelling κράβακτος in Egyptian documents’. ¹⁶² Jones 1941, 10.

176

Patrick James

precise citation from the New Testament instead of the vague reference in LS7-8 by means of ‘N.T.’.¹⁶³ Precise citations had been given in LS⁷ for words that occurred only once¹⁶⁴ (and as early as LS⁶ for ὑπέρακμος). LS7-8 used either ‘Byz.’ or ‘Eccl.’¹⁶⁵ (or both as ‘Eccl., Byz.’)¹⁶⁶ to refer to any number of instances in the vast swathes of Greek literature from Late Antiquity and beyond. This practice reflects a similar lack of precision prior to LSJ.¹⁶⁷ LSJ’s replacement of the ‘awkward’¹⁶⁸ references in LS7-8 to the orators by page numbers with references by speech number and section number was the background to this innovation in the treatment of New Testament citations. The second innovation reveals the revisers’ attitude to the Greek of the New Testament. In LS7-8, various words of theological and ethical significance had sections that were introduced by ‘N.T.’.¹⁶⁹ For example, s.v. πνεῦμα section III (‘the afflatus or inspiration of the prophets’, without a citation) and section V (the Holy Spirit, angels, and evil spirits), but not section IV (‘esp. the highest, noblest part’, opp. to ψυχή, 1 Ep.Thess. 5.23, cf. Rom. 2.29, 8.2 sqq., . . . ’).¹⁷⁰ The treatment of σάρξ remained essentially the same between LS7-8 and LSJ as did that of the ‘Christian’ meaning of ταπεινός, to give an ethical example. This practice was extended in LSJ (but with ‘in NT ’’) to entries for words without

¹⁶³ This revision had begun in LS⁷ with words cited from ‘LXX’. The first appearance in the Septuagint was introduced by ‘LXX’, cited in parentheses, and followed by ‘al.’ s.v. ὑπερασπίζω and s.v. ὑπερασπισμός. Similar formatting still can be found in LSJ for a New Testament word s.v. σκοτία: ‘NT (Ev.Matt.10.27, al.)’. Still, LSJ s.v. ὑπερδικάζω has no more precise reference than ‘Aquila V.T.’. ¹⁶⁴ Some entries show that this process began before LSJ. Jones 1941, 8 related how a false citation (‘Mark 5.2’) had been added in LS⁷ s.v. δαιμονίζομαι through a misreading of the Thesaurus. LSJ gave the first attestation in canonical order and an indication that the verb appeared elsewhere in the New Testament: ‘Ev.Matt.4.24, al., . . . ’. ¹⁶⁵ The respective scope of these terms is not entirely clear. Of the compound adjectives in LS7 8 that have Χριστο as their first element, some are cited from ‘Byz.’ (e.g. Χριστό γραϕος), others from ‘Eccl.’ (e.g. Χριστο δίδακτος). ¹⁶⁶ For both labels in conjunction, see, e.g. s.v. ὑπολιμπάνω. There, LSJ substituted citations from papyri for ‘Eccl., Byz.’. ¹⁶⁷ Some entries in LS7 8 did specify a particular author, but not a specific citation: ‘Basil.’ appears s.v. ὑπεραπολαύω as do ‘Greg.Nyss.’ s.v. ὑλακώδης, ‘Greg.Naz.’ s.v. ὑλάκτης, and ‘Procop(ius)’ s.v. ὑπεράριθμος. ¹⁶⁸ Jones 1941, 10 described this task as ‘the chief aim’ of the ‘first revision’. ¹⁶⁹ Entries for words that do not obviously belong to these categories, such as the treatment s.v. ἔξω as ‘the heathen’ were also preceded by ‘in N.T.’. ¹⁷⁰ The treatment of this word is substantially, and rightly in my view, different at this point in LSJ. Section III now makes no reference to the New Testament. Section V continues to have a ‘spiritual or immaterial being, angel’ as its concern, but does not refer to ‘NT ’ in general. The section now lacks an explicit mention in English of an evil spirit as the referent of this noun itself. Section IV was divided between ‘the spirit of God’ and ‘2. spirit of man: in NT, opp. ψυχή’. The citations LS7 8 were further analysed into uses ‘opp. τῷ σώματι’ and ‘opp. γράμμα, Ep.Rom. 2.29’. Substantial reduction is also evident in LSJ’s revision of LS7 8 s.vv. παραβολή 3 (‘a fictitious narrative by which some religious or moral lesson is conveyed’) and προϕήτης III and 2 (‘one who possesses the spiritual gift of προϕητεία . . . ’). Cf. on ϕωτίζω above.

The Greek of the New Testament

177

such theological or ethical¹⁷¹ significance: s.vv. ἄνω, ἀποκρίνω, ἄρσην, αὐξάνω, δηλονότι, ἐθέλω, ἐνεργέω, ἐχθές, ἔχω, ἵνα, καταγγέλλω, καρπός, μέν, μηθείς, ναῦς, νόος, προστίθημι, πρόθεσις, σκοτία, σκότος, σύν, σύνδουλος, σῶς, τε, τις, τράχηλος, ὑπό, χθές, and ψεῦδος. Several of these additions involve comments on syntax and morphology in the Septuagint as well as in the New Testament. The use of προστίθημι either in the active or in the middle with an infinitive (s.v. III 5 and B III 5) is one such application of ‘in NT’. LSJ does not go so far as to describe this phenomenon as a ‘Hebraism’, ‘Jewish Greek’ or ‘Biblical Greek’, but others do so.¹⁷² The use of the dative case with ὑπό receives a qualification (s.v. B 1): ‘esp. in Poets, never in LXX (Jb.12.5 is dub. l.) or NT, not common in Arist., Ptolemaic papyri, or Plb.’. Discussion of the use of ἀπεκρίθην for ἀπεκρινάμην (s.v. ἀποκρίνω IV 3) involves its condemnation by the Atticist lexicographer Phrynichus, its absence from ‘earlier Attic’, its appearance in Hellenistic papyri and inscriptions and in Koiné writers (Polybius and Josephus), that it is ‘regular in LXX’ (exceptions are found in ‘solemn language’), and that it ‘prevails in NT’. Elsewhere (s.vv.), we are told that both αὐξάνω and αὔξω are found ‘in NT’, while αὔξω is found in ‘Att. Inscrr. and Ptolemaic Pap.’, that ἐθέλω, in contrast to θέλω, is not found ‘in LXX or NT’, that ἵνα after εἶπον is ‘freq. in NT’, that ἐχθές is more common than χθές in LXX, is the only form ‘in NT’, and is ‘freq. in papyri of all periods’,¹⁷³ that the use of ἵνα for ὅπως after ‘Verbs of command and entreaty’ (‘common only in later Gr.’) is ‘freq. in NT’,¹⁷⁴ that ‘heterocl(ite) forms’ of νόος are found ‘in NT’ and ‘later writers’, and that σκότος is ‘always neut. in LXX and NT’, while the masculine is said to be ‘the Att. form’. Some of these additional ‘in NT’ comments concern orthography. The spelling ἀρσ- is said (s.v. ἄρσην) to ‘prevail in LXX and NT’, while ἀρρ- ‘is more common in Pap. (exc. Pap. Mag.)’. The entry for μηθείς reports that the spelling with theta is ‘freq. in Inscrr. and Pap. from iv B.C., IG2².43.37, al., Men. Epit.145, Pk.129, PPetr.1p.80, PCair.Zen.18.7, al. (iii B.C.), etc.: but rarely after the Christian era, once in NT, Act.Ap.27.32, cf. POxy.495.17 (ii A.D.)’. A fuller discussion is found s.v. οὐθείς. Although the New Testament’s evidence does not feature there, we are told, after a discussion both of inscriptions and

¹⁷¹ ‘In NT ’ and a context specific translation were added to some ethical entries, e.g. s.vv. ϕιλαδελϕία and ϕιλάδελϕος, and to some ‘theological’ entries, e.g. s.v. τέρας. After ‘in NT ’ s.v. ἀπόλλυμι B 2, ‘in theol. sense’ was added (uniquely, as far as I can ascertain). ¹⁷² Blass, Debrunner, and Funk 1961, 224, BDAG s.v. 1c, and Muraoka 2009, 599. ¹⁷³ A more complicated account is given s.v. χθές. The three New Testament occurrences are cited with a note that ‘codd. vary betw. ἐχθές and χθές’ and that ‘χθές is not found in Ptolemaic papyri, but in PLond.2.161.8 (iii A.D.)’. ¹⁷⁴ The discussion continues with an explanation of ἐν τούτῳ , ἵνα καρπὸν πολὺν ϕέρητε (John 15.8) by ἐν τῷ ϕέρειν.

178

Patrick James

Ptolemaic papyri and of the manuscripts of Classical and Hellenistic authors that: ‘the frequency of θ forms in the uncials of LXX varies roughly according to the date (known or probable) of the translation of the book in question (though the δ forms are in a large majority in the LXX as a whole)’.¹⁷⁵

Seven instances of οὐθείς have sufficient manuscript support to appear in Nestle-Aland 2012.¹⁷⁶ Five of these occur in quotations of direct speech in Luke and Acts, in which οὐδείς is also found. The other two are in Paul’s Corinthian correspondence (1 Corinthians 13.2 and 2 Corinthians 11.9). That data, and particularly such a restricted pattern of attestation, belongs with the story LSJ tells of a spelling which appears in the fourth century BC, predominates in the early Hellenistic period, but is ousted by the earlier spelling οὐδείς in the first century AD, the time of the New Testament’s events and documents. Another group of ‘in NT’ additional comments concern word choice. The particle μέν is said (s.v. A II) to be ‘less freq. in later Gr. (rare in NT)’ and the use of μενοῦν and μενοῦνγε to begin a sentence is noted. Only the cryptic ‘cf. Phryn.322, Hsch.’ signals that this is abnormal from the point of view of Classical Attic. The entry for τε is discursive (s.v. A II 5): the copulative τε becomes rare in later Gr.; it is found about 340 times in LXX, mostly in the Pentateuch and 1 4 Ma., only 3 times in Ps.; in the NT it is found about 150 times in Act. Ap., 20 times in Ep.Hebr., and very rarely in the other books.

Here, the New Testament, like the Septuagint, is used to substantiate a comment about ‘later Gr(eek)’. Other bodies of evidence are not cited and a distinction is made between Acts and Hebrews on the one hand (texts that show greater stylistic sophistication)¹⁷⁷ and the rest of the New Testament on the other. The univerbation δηλονότι is said to occur ‘once in NT, 1 Ep. Cor.15.27.’¹⁷⁸ The history of the preposition σύν is told with reference to the New Testament: ‘in Pap., NT, and later Prose its use is much less restricted’. The entry for ναῦς relates that it is ‘rare in non-literary Hellenistic Greek, once in NT, Act.Ap.27.41, πλοῖον being generally used’. This comment exemplifies a tendency in the Koiné for ‘anomalous substantives’ to be replaced by more

¹⁷⁵ These remarks seem to be derived from Thackeray 1909, 58 62. ¹⁷⁶ Souter 1910, the edition that LSJ claims to follow, printed οὐθέν at 1 Corinthians 13.2, but not in any of the other places. At Acts 27.33, he printed μηδέν with the result that there is another discrepancy between the New Testament evidence that LSJ cites and that presented by the source text it claimed to use. ¹⁷⁷ See Turner 1976, 45 63 and 106 13. ¹⁷⁸ Souter 1910, LSJ’s base text, printed this as two words in this instance (as does Nestle Aland 2012). Again, a discrepancy between LSJ’s citations and their base text emerges.

The Greek of the New Testament

179

easily inflected synonyms.¹⁷⁹ From this comment, it may be inferred that Acts serves to corroborate this claim as one (indeed, the only) cited example of ‘non-literary Hellenistic Greek’. Instead, what should be underlined is not the uniqueness itself of this instance of ναῦς, but that when ναῦς does occur in the New Testament, it occurs in Acts, some of the most stylistically sophisticated Greek in the New Testament,¹⁸⁰ and in a ‘purple passage’ at that, the storm and shipwreck.¹⁸¹ The entry for τράχηλος contains a similar note about the obsolescence of a word in the Koiné: ‘in LXX, NT, and Pap. τ(ράχηλος) is more freq. than αὐχήν)’. There is no such comment s.v. αὐχήν. We are told (s.v. σῶς) that, although inscriptions and papyri provide a few attestations, ‘the word is rare in LXX . . . ; not found in NT’. The indefinite genitive του (for τινος) is said (s.v. τις A I) to be ‘rare after 300 B.C., never in LXX or NT’. Citations from Hellenistic inscriptions and papyri, from Polybius, and from its revival ‘by the Atticists’ are given. It will have become clear that many additional ‘in NT’ comments relate the Greek of the New Testament not only to that of the Septuagint, but also to that of inscriptions and that of the documentary papyri. That is, LSJ follows the path of Deissmann, Moulton and Milligan, Souter, and others in seeing the Greek of the New Testament not as a Semiticized Jewish variety (Nietzsche’s ‘the language of the Holy Ghost’) to be studied in isolation, but as a specimen of the Koiné of its time and comparable with the language of inscriptions and of the documentary papyri.¹⁸² Jones’ assessment of the Vocabulary of Moulton and Milligan, ‘which (within its natural limits) may almost be regarded as a Lexicon of the κοινή as a whole’, is borne out by LSJ’s entries.

10.7. CONCLUSION As we have considered the place of the New Testament in the Liddell and Scott tradition, the coverage of Christian texts, typical features of LSJ’s entries, the personalities involved and their contributions, sources and influences, and the impact of nineteenth-century New Testament philology, we have seen plenty ¹⁷⁹ Browning 1983, 28. ¹⁸⁰ See Turner 1976, 45 63. ¹⁸¹ On this basis, the list of examples in Browning (1983, 28) needs correction: ναῦς is found in the New Testament, as it is in the Septuagint (1 Kingdoms 5.6, 3 Kingdoms 9.26, 10.11, 10.22, 16.22, 2 Paraleipomena 9.21, 4 Maccabees 7.1, Proverbs 30.19, 31.14, Job 9.26, Wisdom 5.10, Daniel 11.40 Theodotion (πλοῖον ‘Old Greek’), but its use in Acts still proves to be very instructive for Browning’s point. ¹⁸² For the debate between those who took New Testament Greek to be ‘Jewish Greek’ or similar and those who saw it as the continuation of Classical Greek in everyday use, see Aitken 1999 and 2014 a, b; George 2010; Horrocks 2010, 106 8 and 147 52; Horsley 1989, 5 40; and Joosten 2005, 2014, and 2016.

180

Patrick James

to substantiate the opinion that I expressed to John Lee and that met with his approval:¹⁸³ LSJ’s treatment of New Testament Greek is superficial, shoddy, and shambolic. The New Testament seems never to have been treated fully and that treatment seems to have received little by way of systematic revision to keep in step with New Testament philology contemporary with the successive editions of Liddell and Scott. Entries show the idiosyncrasies of the editors and suffer both from having to be overly concise and isolated from other Christian texts in Greek for practical reasons and as a reflection of less coverage of later Greek in general. These criticisms echo those made about LSJ in general. Any lexicon on such a scale, revised piecemeal over so many decades, and even with the aid of digital humanities would show such flaws. If the weakness of the Liddell and Scott tradition was that the Greek of the New Testament had been isolated from New Testament scholarship and was later separated from other Christian texts, LSJ’s strength, in general and over its previous editions, was the development, albeit imperfect, towards relating the language of the papyri to that of the New Testament.

¹⁸³ Personal communication, 14 April 2017.

11 The Ancient, the Medieval, and the Modern in a Greek-English Lexicon, or How To Get Your Daily ‘Bread’ in Greek Any Day Through the Ages Mark Janse

In Memoriam Martin Litchfield West (1937–2015)¹, ² LSJ has always been entitled A Greek-English Lexicon, from the first (1843) through the ninth edition (1940), the falsely modest indefinite article lacking from both cover and spine, and omitted altogether from the Supplement (1968) and the Revised Supplement (1996).³ Clearly no need was felt, not even in 1996, to add Ancient to the title, even though LSJ is not and never was intended to be a comprehensive lexicon of the Greek language in its entire history. One year later, however, the term ‘Greek’ appeared, again without further modification, in the main titles of two important monographs: Holton, Mackridge, and Philippaki-Warburton 1997 and Horrocks 1997. Only in the subtitles does it become clear that the former is a comprehensive grammar of the modern language and the latter a comprehensive history of the language from the second millennium BC to the present day. As Holton et al. note in their preface: ‘It has long been customary to use the term “Greek” to refer to the ancient language, while the contemporary language is known as “Modern ¹ Special thanks are due to Michael Clarke and Christopher Stray for their comments, patience and cheerfulness. I would further like to thank my colleagues Klaas Bentein, Filip De Decker, and Joanne Stolk for assistance of various sorts. ² I had the special privilege and, indeed, great pleasure of getting to know Martin very well when I was a visiting fellow of All Souls College in Michaelmas 2007 and again in Michaelmas 2014. I am very grateful to the Warden(s) and Fellows of the College for the honour of electing me twice to a visiting fellowship. ³ The definite article in The Online Liddell Scott Jones Greek English Lexicon appears to define the online version rather than the LSJ itself, although TLG director Maria Pantelia justly calls it ‘the premier lexicon of classical Greek’ in the ‘About’ section (stephanus.tlg.uci.edu/lsj). Mark Janse, The Ancient, the Medieval, and the Modern in a Greek-English Lexicon, or How To Get Your Daily ‘Bread’ in Greek Any Day Through the Ages In: Liddell and Scott: The History, Methodology, and Languages of the World’s Leading Lexicon of Ancient Greek. Edited by: Christopher Stray, Michael Clarke, and Joshua T. Katz, Oxford University Press (2019). © Oxford University Press. DOI: 10.1093/oso/9780198810803.003.0011

182

Mark Janse

Greek”’, adding: ‘We believe that, as a living language, contemporary Greek does not need to be qualified by an adjective which implies that it is somehow secondary to the ancient language. For this reason, we use “Greek” throughout this book to refer to the modern language, adding the adjective “Ancient” or “Modern” only when these two chronological stages need to be distinguished’ (1997, xiii). In the same vein, Horrocks notes in the preface to the second edition of his work that it ‘was never intended to be primarily a history of Ancient Greek’ (2010, xvi—my italics). The list of basic reference grammars, dictionaries, and handbooks of Ancient Greek that got away without Ancient in their title is too long and surely too familiar to mention here. It is nevertheless significant that a number of important recent publications depart from this honourable tradition.⁴ For the purpose of this paper, the most important of these are Chadwick’s Lexicographica Graeca (1996), which has Ancient in its subtitle,⁵ and the English revision of Montanari’s Vocabolario della lingua greca (2013), retitled The Brill Dictionary of Ancient Greek (2015).⁶ Apart from the addition of Ancient, one may note the presence of the definite article (and of the publisher’s instead of the author’s name).⁷ The question I want to address in this contribution is not whether Ancient should be added to the title of LSJ but whether its scope should be extended to include later stages, particularly the Medieval and the Modern,⁸ given the remarkable continuity of the Greek language stressed by Chantraine (DELG, viii) and others.⁹ With the availability of the online LSJ this is an option which should be seriously considered, although the editorial problems of a continuously updated online version may seem forbidding (as pointed out by the late and much lamented Martin West, Chapter 19, this volume). A comprehensive lexicon of the entire Greek language, however, is not what I have in mind, as such an enterprise would not entail a revision of LSJ, but the creation of an entirely new lexicon.¹⁰ What could be envisaged, rather, is the addition, wherever relevant, of

⁴ E.g., Christidis’ History (2007), Bakker’s Companion (2010), Giannakis’ Encyclopedia (2014) and Colvin’s Brief History (2014). Colvin’s Historical Greek Reader (2007) lacks the Ancient, probably because the book was originally commissioned from James Hooker, who had prepared a first draft of the manuscript at the time of his death in 2001 (Colvin 2007, v). ⁵ Chadwick was a member of the Committee established by the British Academy to oversee their project for the Revised Supplement (1996) and made a large number of suggestions for improvements (1996, 1). ⁶ The forthcoming Cambridge Greek Lexicon (Diggle et al. 2019) is explicitly announced as an intermediate Ancient Greek English Lexicon (www.classics.cam.ac.uk/research/projects/glp). ⁷ Compare the titles of the online versions: The Online Liddell Scott Jones Greek English Lexicon vs. The Brill Dictionary of Ancient Greek Online. ⁸ As is well known, Stuart Jones decided, ‘[a]fter due consideration’, to exclude both Patristic and Byzantine literature from the ninth edition (LSJ, x). ⁹ See, e.g., Joseph (2009), 349 and Horrocks (2010), xiii. ¹⁰ One may compare the magnitude of multi volume dictionaries such as Dimitrakos’ Great Lexicon (1936 50; abridged form 1964), which covers Ancient, Medieval and Modern Greek (both Katharevousa and Demotic) or Kriaras’ Dictionary of Medieval Vulgar Greek Literature

Ancient, Medieval, and Modern in the Lexicon

183

brief references to the semantic and/or grammatical development of particular words in Medieval and/or Modern Greek. This is, in fact, what Chantraine had in mind when he decided to add Histoire des mots to the main title of his DELG. His justification for doing so is worth quoting in full: [L]’étymologie devrait être l’histoire complète du vocabulaire dans sa structure et son évolution et c’est pour l’histoire du vocabulaire, reflet de l’histoire tout court, que je me suis donné le plus de peine. [ . . . ] Le grec présente une histoire continue et [ . . . ] le grec d’aujourd’hui sous sa forme démotique ou puriste continue directe ment le grec d’Homère et de Démosthène, la langue byzantine fournissant l’anneau qui unit les deux morceaux de la chaîne. Il va de soi qu’il ne pouvait être question de donner ici une idée de l’étymologie du grec moderne, enrichi d’emprunts de toute sorte: slaves, turcs, italiens et autres. En revanche, il pouvait être utile d’indiquer à l’occasion comment un mot ancien a subsisté en grec d’aujourd’hui (DELG, v viii)

While I fully endorse Chantraine’s point of view and firmly believe that it could be fruitfully applied to the online LSJ, it should nevertheless be noted that Medieval Greek, which Chantraine so aptly calls ‘the ring that unites the two pieces of the chain’, does not figure at all in his dictionary, so that the historical chain connecting Ancient and Modern Greek is actually broken, or at least appears to be so. I will show how this methodology is bound to distort the historical picture and how this could be remedied. To illustrate the survival of ancient words in Modern Greek, Chantraine, without further discussion, mentions ψωμί ‘bread’ and κρασί ‘wine’ as ‘cas classiques’ (DELG, viii). They are rightly so called, as any classicist without knowledge of Medieval and/or Modern Greek will have encountered these words to her or his frustration instead of the ancient ἄρτος and οἶνος when ordering bread or wine on her or his first visit to Greece—only to find that the former does appear on the labels of the various types of bread in the bakery, still called today αρτοποιείο instead of *ψωμοποιείο,¹¹ and the latter on the labels of wine bottles (if indeed the wine was served in a bottle at all). To add to the confusion, ‘red wine’ is now called κόκκινο κρασί, ‘white wine’ άσπρο κρασί, but the labels still have the archaic ερυθρός οίνος and λευκός οίνος. Maybe Chantraine did have wine in mind when he singled out ἄσπρος as an interesting case of semantic change in addition to his classic cases as cited above. In fact, νερό ‘water’ should be added to the mix: never order ὕδωρ /ˈiðor/ in a Greek tavern or coffee house, or you risk ending up drinking D₂O instead of H₂O.¹² (1968 , abridged as Kazazis and Karanastasis 2001 ; the English translation follows the title given in the English preface to the online version). ¹¹ Instead of *ψωμοποιείο one encounters the words ψωμάδικο and, more generally, ϕουρνάρικο or simply ϕούρνος, lit. ‘oven’ < Lat. furnus (cf. Ath. 3.113c etc.). ¹² The ancient word ὕδωρ is used in Modern Greek only in learned collocations such as βαρύ ύδωρ ‘heavy water’ (D₂O), which is opposed to κοινό νερό ‘normal water’ (H₂O).

184

Mark Janse

To be sure, ψωμί, κρασί and νερό are not listed in LSJ, at least not in their modern appearances, but the information provided by Chantraine is still historically rudimentary, not to say incomplete. Compare, for instance, the short line on the survival of words derived from the Proto-Indo-European root *bhs-eh₁- (Gr. °ψη-) or, alternatively, *bhs-eh₂- (Gr. °ψᾱ-): ‘De ψωμός, ψωμίον, a survécu en grec modern le nom usuel du pain, ψωμί ’ (DELG s.v. *ψήω).¹³ Fortunately, Chantraine in the entry refers the interested reader to Kretschmer 1926 for more information on the ‘spécialisation sémantique’.¹⁴ He should have repeated the reference elsewhere, as Kretschmer provides ample information on the history of the Modern Greek words for not only ‘bread’, but ‘wine’ and, although not mentioned in the title, ‘water’ as well. With all the ingredients for an admittedly basic Greek meal at hand, we should now be ready to embark on a historical journey centred around eating and drinking through the ages and see how this information could be incorporated in the online LSJ. For reasons of space, however, I will concentrate on the word for ‘bread’ and leave the wine and the water for other occasions. It will be useful briefly to present the resources that are indispensable for complementing and updating the online LSJ. First of all, there are LSJ’s recent competitors, both of which are broader in scope and based on recent advantages in Ancient Greek lexicography:¹⁵ Montanari’s Dictionary of Ancient Greek (2015) and Adrados’ Diccionario griego-español (1980–), the latter having reached the letter Epsilon (2009). Lampe’s Patristic Greek Lexicon (1961–8) needs no further introduction to the classicist, its specific aim being ‘to form a supplement, or companion, to the ninth edition of Liddell and Scott [LSJ]’, particularly ‘the listing of all words occurring in the Fathers which were either not contained in [LSJ] or but poorly attested there’ (Lampe 1961–8, v–vi). It covers the post-biblical patristic literature from the first to the ninth century (Lampe 1961–8, vii). The focus of Trapp’s Lexikon zur byzantinischen Gräzität (1994–2017) is on the subsequent period from the ninth to the twelfth century inclusive, but it explicitly intends to complement LSJ and Lampe as well (Einführung, n.p.). It covers high-register literary and nonliterary texts, the low-register Byzantine texts being covered in Kriaras’ monumental Λεξικό της Mεσαιωνικής Eλληνικής Δημώδους Γραμματείας (1968–), which focuses on the ‘vulgar literature’ from the twelfth to the seventeenth ¹³ Beekes, following Frisk (1960 72 s.v. ψῆν), tentatively derives this family of words from PIE. *bhs h₂ , only to conclude that ‘the IE explanations must be given up; the group of words is probably Pre Greek’ (2016 s.v. ψῆν). A[lain] B[lanc] calls the question whether the root should be analysed as *bhs eh₁ (Gr. °ψη ) or as *bhs eh₂ (Gr. °ψᾱ ) ‘irritante’ (DELG Suppl. s.v. *ψήω). ¹⁴ The reference in Chantraine (DELG s.v. *ψήω) should be corrected: Kretschmer’s article was published in 1926, not 1927. ¹⁵ Chadwick’s Lexicographica Graeca (1996) constitutes a fundamental contribution to Ancient Greek lexicography.

Ancient, Medieval, and Modern in the Lexicon

185

century inclusive.¹⁶ Following the classification of Holton and Manolessou (2010, 541), essentially Trapp covers Early Medieval Greek (500–1100) and Kriaras Late Medieval Greek (1100–1500). As Holton and Manolessou acknowledge, the start and end dates of the various periods vary according to the criteria adopted: historical, literary, linguistic, or a combination of any of these. Their periodization ‘does not ignore external (historical, literary, etc.) criteria, but gives more weight to internal (linguistic) ones on the basis of clusters of significant linguistic changes’ (2010: 541).¹⁷ This also applies to the subsequent period called Early Modern Greek (1500–1700) by Holton and Manolessou, but the similarities between ‘vulgar’ Late Medieval Greek and ‘vulgar’ Early Modern Greek are such that any boundaries are bound to obfuscate the fundamental continuity.¹⁸ E.A. Sophocles goes as far as to claim that ‘[t]he popular dialect of the twelfth century was essentially the same as the Romaic or modern Greek of the present day, and may with propriety be called the early modern Greek’ (Sophocles 1900, 10).¹⁹ What is called here ‘Romaic or [M]odern Greek’ is in fact the ‘vulgar’ or colloquial variety commonly known as Demotic (δημοτική) and not to be confused with Standard Modern Greek, which is a compromise between Demotic and Katharevousa (καθαρεύουσα).²⁰ The Demotic Greek of Sophocles’ time has been described in a magisterial way by the Panhellenist Albert Thumb (1910), unfortunately with little attention to its historical development, especially the continuity between Late Medieval and Demotic Greek, but fortunately with due attention to the Modern Greek dialects. The importance of these for the continuity of the Greek language was highlighted by Shipp, whose ground-breaking Modern Greek Evidence for the Ancient Greek Vocabulary (1979) shows how the modern dialects can be used to elucidate obscure words or meanings in Ancient Greek.²¹ A fundamental reference work from the Modern Greek perspective is the Ἱστορικὸν λεξικὸν τῆς νέας ἑλληνικῆς published by the Academy of Athens, which has now, after eighty-three years, reached the letter Delta: the sixth volume (2016) appropriately ends with the entry διάλεκτος. It contains a wealth of dialectal and historical data tracing the semantic development of the Modern Greek vocabulary. In what follows I will show how the above mentioned resources can be profitably used to complement and update the

¹⁶ The period is defined more narrowly in the subtitle: from 1100 to the completion of the Ottoman conquest of Crete in 1669). On ‘high’ and ‘low’ in Medieval Greek see Horrocks (2017). ¹⁷ Compare, e.g., the different periodization adopted by Joseph (2009, 349). ¹⁸ For discussion see Janse and Joseph (2014). ¹⁹ Sophocles subtly remarks that ‘[i]mbecility, pedantry, childishness and self conceit are the characteristics of the last epoch of the Byzantine period’ (Sophocles 1900, 10). ²⁰ On the Greek γλωσσικό ζήτημα ‘language question’ see Mackridge (2009). ²¹ Andriotis (1974) is a monumental collection of archaisms in the Modern Greek dialects.

186

Mark Janse

online version of LSJ, using Chantraine’s ‘classic example’ ψωμί and related words. It goes without saying that the TLG has been constantly consulted. The word ψωμίον is listed in LSJ and the lexicographical information provided there is quite intriguing: ‘Dim. of ψωμός, of a bun for a crocodile.’ This explanation may seem rather perplexing, given our current knowledge of the dietary preferences of modern crocodiles, especially those of the niloticus species, being the apex predators they are. The context is worth quoting in full (P.Tebt. 1.33.13–4, 112 BC):²² τὸ γεινόμενον τῶι Πετεσούχωι καὶ τοῖς κροκο(δείλοις) | ψωμίον ‘the customary ψωμίον for Petesouchos and the crocodiles’

The phrase refers to the Egyptian crocodile cult well known from Herodotus and Strabo.²³ Herodotus (2.69) informs us that crocodiles were considered sacred around Lake Moeris in the Fayum (Arsinoite nome)²⁴ and fed with σιτία ἀπότακτα, usually translated as ‘food set aside’ (Montanari s.v. ἀπότακτος ad loc.), ‘special food’ or ‘fodder’, not ‘bread’ (Powell 1938 s.v. σιτία ad loc.).²⁵ Strabo (17.1.38) adds that the crocodile was called Σοῦχος:²⁶ it was fed with σιτίοις καὶ κρέασι καὶ οἴνῳ ‘σιτία and meat and wine’. The juxtaposition of σιτίοις καὶ κρέασι suggests that the former is here to be taken in a narrower sense than in σιτία ἀπότακτα. This is confirmed by the sentence immediately following, in which Strabo’s host brings from the dining table πλακουντάριόν τι καὶ κρέας ὀπτὸν καὶ προχοίδιόν τι μελικράτου ‘a kind of cake and roast meat and a small pitcher of melicrate’.²⁷ Then some of the priests go up to the sacred crocodile, open its mouth (and keep it open, one presumes), while another priest puts in τὸ πέμμα καὶ πάλιν τὸ κρέας, εἶτα τὸ μελίκρατον ‘the baking and then the meat, followed by the melicrate’.

²² Papyrological metadata, including abbreviations, are taken from the Papyrological Navi gator (www.papyri.info). ²³ On the crocodile cult in the Nile Valley see the excellent chapter ‘Crocodiles and Papyrus’ in Verhoogt (1998, 7 21). ²⁴ The former name of the city Arsinoë was Κροκοδείλων Πόλις ‘Crocodilopolis’ (Hdt. 2.148; Str. 16.2.27, 17.1.38 & 47). ²⁵ LSJ (s.v. ad loc.) translate ἀπότακτος here as ‘set apart for a special use’. ²⁶ Σοῦχος is the Greek name of the Egyptian god Sobek (Eg. sbk), originally a water and fertility god (Verhoogt 1998, 8). Πετεσουχος (e.g. pȝ di sbk ‘given by Sobek’) is the name of a crocodile god worshipped in several villages in the Arsonoite nome (Verhoogt 1998, 9). In its Latinized form, suchus is used to refer to several clades and genera in the crocodylian genealogy, e.g. deinosuchus ‘terrible crocodile’, sarcosuchus ‘flesh crocodile’ and the clade eusuchia ‘true [sic] crocodiles’, to which the crocodylians belong (Grigg and Kirshner 2015, ch. 2 ‘The Crocodylian Family Tree’, 43 79). ²⁷ The term μελίκρατον is worthy of a lexicographical study in its own right, which I hope to offer in the near future. (Melicrate is a honey based drink.)

Ancient, Medieval, and Modern in the Lexicon

187

Πλακουντάριον is of course a diminutive of πλακοῦς ‘flat’ which, either in combination with ἄρτος (Ath. 14.645d, cf. 14.646c) or, more commonly, without, obviously means ‘flatbread’ (Montanari) or ‘flatcake’ (LSJ).²⁸ As Wilkins and Hill note, ‘[t]here was great variety in the “cakes” offered in religious rituals’ and ‘also possibly much variation between communities over what was meant by terms such as plakous’ (Wilkins and Hill 2006, 127), which is perhaps why Strabo unspecified, so to speak, what kind of cake was served to the croc. Not surprisingly, cake-making was considered an art in antiquity.²⁹ Book 14 of Athenaeus’ Deipnosophistae is devoted entirely to cakes (more of which are named in passing in book 3). He mentions two obscure authors of lost treatises (or should we say, treats?) called Περὶ πλακούντων ‘On cakes’: Iatrocles (14.647a–b) and Harpocration of Mendes (14.648b). Athenaeus further informs us that Callimachus in his Pinakes lists four other, equally obscure, writers of πλακουντοποϊικὰ συγγράμματα ‘writings related to cakemaking’: Aegimius, Hegesippus, Metrobius and Phaetus (14.643e). The popularity of cakes is further demonstrated by the variety of Hellenistic terms for ‘caker’:³⁰ LSJ list πλακουντάριος, πλακουντᾶς, and πλακουντοποιός, translated as ‘pastry-cook’ (‘pastry-chef ’ in the English Montanari). The first two are still attested in Early Medieval Greek: πλακουντᾶς ‘Kuchenbäcker’ and the adjectives πλακουντάρικος and πλακουντάριος ‘kuchenförmig’ (Trapp). Πλακοῦς has survived into Modern Greek as πλακούντας, meaning ‘cake’ as well as ‘placenta’ (the latter from French placente and ultimately from the learned Neo-Latin placenta, cf. Babiniotis 2011 s.v.). Πλακοῦς and its derivatives must have become obsolete in the spoken language, as there are no traces of it left in Late Medieval Greek,³¹ nor in the Modern Greek dialects.³² Cakes, of course, are baked, as when Trygaeus’ slave announces that everything is ready for his master’s wedding (with one notable but predictable exception):³³ ὁ πλακοῦς πέπεπται ‘the cake is baked’ (Ar. Pax 869). From this ²⁸ Πλακουντάριον, like its equivalent πλακούντιον, is left untranslated in LSJ, but Montanari has ‘little flatbread, pastry’. ²⁹ Useful modern surveys of ancient cakes can be found in Dalby (2003), 69 71 and Wilkins and Hill (2006), 126 30. ³⁰ The word caker is not included in the OED, but even a superficial Google search reveals that the term is actually used in the same sense as cake chef, i.e. (professional) ‘cake maker’. ³¹ Kriaras lists πλακόπιτα ‘a kind of (sweet) cake’, which is derived from πλάκα, °πλακο rather than πλάξ, °πλακο . It is used in an interesting context by Planudes: ὁ δὲ ξένος, ἐπειδὴ ποτέ του δὲν ἔϕαγε πίτες, ἐμάζωνεν αὐτὰς καἰ ἔτρωγέν τες ὡσὰν ψωμίν ‘but the stranger, since he never in his life ate cakes, used to steal them and eat them as if they were bread’ (Planudes, Life of Aesop 4.106.40). ³² There is one possible exception, attested in the Asia Minor Greek dialect of Pharasa: ϕκακκούδες (Dawkins 1916, 635 s.v. πλακούς ‘cake’; 158f.) or ϕκακούδες (Andriotis 1948, 62 s.v.). The sound change /pl/>/fk/ is well documented in this dialect (Dawkins 1916, 158 9; Andriotis 1948, 31), but the meaning given by Andriotis does not seem to make sense: ‘στραγάλια’, i.e. ‘roasted chickpeas’ (cf. Andriotis 1974, 454 s.v. πλακοῦς). ³³ Sc. τοῦ πέους δὲ δεῖ ‘it’s just the cock that’s missing’ (Ar. Pax 870).

188

Mark Janse

remark we may safely conclude that Strabo’s πέμμα is another word for a kind of cake, as in fact it is (Dalby 2003, 69). LSJ translate it as ‘any kind of dressed food’, but add that it is mostly used in the plural, meaning ‘pastry, cakes, sweetmeats’. The diminutive πεμμάτιον is used by Athenaeus (14.645e), who also coins the term πεμματολόγος as an epithet of Chrysippus of Tyana (14.648a), credited as the author of the lost Ἀρτοποιϊκόν or ‘Baking Bread’ (3.113a, 14.647c; spoonerism intended and recommended—MJ).³⁴ Πεμματολόγος is translated as ‘discoursing of cakes’ in LSJ but, perhaps more appropriately, as ‘learned in cakes’ by Wilkins and Hill (2006, 127).³⁵ Πέμμα and πεμμάτιον are attested as late as the twelfth century in learned Byzantine authors such as John Tzetzes and Eusthatius of Thessalonica (cf. Trapp). In the Grottaferrata version of the Byzantine epic Digenis Akritis, the plural πέμματα occurs with reference to all kinds (παντοῖα) of burning spices (6.38 Jeffreys),³⁶ perhaps best translated as ‘incenses’.³⁷ The combined evidence of Strabo’s use of πλακουντάριόν τι and πέμμα to specify the previously mentioned σιτία suggests that LSJ were right in translating ψωμίον in the Tebtunis papyrus as ‘bun’ (for a crocodile). But before discussing the semantic development of ψωμίον, we will consider that of σιτίον or rather that of σῖτος, from which it is derived and with which it shares the following basic meanings according to LSJ: 1: ‘grain, comprehending both wheat (πυρός) and barley (κριθή)’, 2: ‘food made from grain, bread, opp. flesh-meat’, 3: ‘in a wider sense, food, as opp. to drink’ (cf. Montanari s.v.). Meaning 1: ‘grain’ or ‘corn’ is certainly the original meaning, as Mycenaean si-to is well attested in ration lists in the Linear B tablets, with ideograms representing both barley and wheat (Chadwick and Baumbach 1963, 244; cf. DMic s.v. si-to).³⁸ The distinction between meanings 2: ‘bread’ and 3: ‘food’ is already blurred in Homer (LfgrE s.v.). When Odysseus and Eumaeus have satisfied their desire for food after feasting upon the boar roasted by Eumaeus and the bread served by Mesaulius, we read the following (Od. 14.455–6):³⁹

³⁴ The afore mentioned Iatrocles is also credited for having written a treatise with the same name (Ath. 14.326e). ³⁵ Liddell himself may have been ambivalent about cakes, since an undergraduate of his college, Christ Church, in 1874 circulated a satirical pamphlet entitled Cakeless about Mrs Liddell’s attemots to marry off her daughters. The author, John Howe Jenkins, was identified and expelled. ³⁶ Jeffreys (1998, 155) translates πέμματα as ‘spices’, Mavrogordato (1956, 165) as ‘confections’. ³⁷ Compare Kriaras’ gloss of πέμμα: ‘κάθε είδος αρωματικής ουσίας που καίγεται για να ευωδιάσει η ατμόσϕαιρα’, i.e. ‘any kind of aromatic substance which is being burnt to make the air smell sweetly’. ³⁸ It should be noted that κριθή is attested in Mycenaean (Chadwick and Baumbach 1963, 213; DMic s.v. ki ri ta), but πυρός is not (Vittiglio 2017, 5; Fischer 2017, 15): pu ro being Πύλος (DMic s.v. pu ro). ³⁹ |4c denotes the position of the bucolic diaeresis in the notation of Janse (2003).

Ancient, Medieval, and Modern in the Lexicon

189

σῖτον μέν σϕιν ἀϕεῖλε Μεσαύλιος, |4c οἱ δ’ ἐπὶ κοῖτον σίτου καὶ κρειῶν κεκορημένοι |4c ἐσσεύοντο ‘Mesaulius took away the food from them, | and they to bed sated with bread and meat, | hastened’

Any English translation reads ‘food’ for σῖτον in 455 and ‘bread’ for σίτου in 456, although 455 is a variation on 449, where σῖτον refers to the bread served to accompany the meat of the boar just roasted and divided into seven portions (ἕπταχα, 434) by Eumaeus (Od. 14.448–50):⁴⁰ ἐν χείρεσσιν ἔθηκεν, |3b ὁ δ’ ἕζετο ᾗ παρὰ μοίρῃ σῖτον δέ σϕιν ἔνειμε Μεσαύλιος, |4c ὃν ῥα συβώτης αὐτὸς κτήσατο οἶος |3b ἀποιχομένοιο ἄνακτος ‘he placed [the cup] in [Odysseus’] hands | and he sat by his own portion and Mesaulius served them bread, | which the swineherd had acquired himself alone | while his master was gone’

Moritz (1955), naturally unaware of the Mycenaean evidence to be published one year after his article (Ventris and Chadwick 1956),⁴¹ erroneously assumed that ‘it is difficult to decide whether an original meaning of “food generally” was becoming restructured to “cereal food” or whether the word properly denoted cereal food and was being loosely used to cover other food also’ (1956, 136). Given the Mycenaean evidence, the relation between the three meanings distinguished in LSJ is clearly metonymic. The linguistic change involves the metonymic extension or ‘semantic broadening’ of the meaning of the basic ingredient of bread, viz. ‘grain’ (whether wheat or barley) to ‘bread’ as the staple food prepared on the basis of this basic ingredient. The next step is not ‘food generally’, as Moritz (and LSJ and Montanari) would have it, but rather ‘food including bread and other food items’ and then ‘food generally’. The gist of Moritz’s article, however, is not the distinction between ‘cereal food’ and ‘food generally’, but the semantic narrowing of the original meaning of σῖτος from ‘grain’ to ‘wheat’. Moritz refers to the entry in the Suda, which deserves to be quoted in full: πᾶς ὁ σιτικὸς καρπός, οὐχ ὁ πυρὸς μόνον. Moritz concludes that ‘Suidas implicitly admits that in common speech the meaning of σῖτος had become restricted to the narrower of the two senses of “corn”, and that in this restricted meaning it denoted πυρός’ (Moritz 1955, 136). As a matter of fact, the same gloss is found in Photius’ lexicon and probably goes back to Aelius Dionysius, which would seem to imply that σῖτος was used to refer to ‘wheat’ exclusively already in the second century AD (Shipp 1979, 499).

⁴⁰ |3b denotes the position of the trochaic or ‘feminine’ caesura in the notation of Janse (2003). ⁴¹ The evidence for si to is presented in Ventris and Chadwick (1956), 408, cf. Chadwick and Baumbach 1963, 244.

190

Mark Janse

The semantic narrowing is further confirmed by the many scholia which gloss πυρός as σῖτος.⁴² Moritz had looked in vain for papyrological data pre-dating his tenthcentury evidence, but Cadell (1973) confirms ‘la disparition presque complète du mot πυρός après 340 et, parallellement, la diffusion rapide de σῖτος au sens restreint de blé dès le début du IVe s.’ (cf. Shipp 1979, 499). In fact, the parallel mention of artabae of πυρός and κριθή in receipts and orders from Egypt decreases sharply after AD 300, by which time σῖτος was already regularly employed instead of πυρός, as in the following example (P.Tebt. 2.204.3–6):⁴³ λόγος ἀνν[ώ]νης ⟦ ̣⟧ σίτ̣[ου] κ̣αὶ κριθῆς στατῆρες ι ̣ϛ, εἰς λόγον σ̣[ίτο]υ καὶ κριθ[ῆς] στατῆρες λβ καὶ (δραχμαὶ) β ‘[on] account of annona for wheat and barley 16 staters, on account of wheat and barley 32 staters and 2 (drachmas)’

Other evidence comes from the compound κριθόπυρον ‘mixture of barley and wheat’ (LSJ), attested thirty-nine times in the Papyrological Navigator, all dated BC (hence not listed in Trapp).⁴⁴ This was replaced by σιτόκριθον ‘mixture of wheat and barley’ (LSJ) or ‘Steuer auf Weizen und Gerste’ (Trapp), attested forty-four times after AD 300 and still current in Late Medieval Greek (Kriaras). Σῖτος continued to be used as the learned word for ‘wheat’ and occasionally for ‘food generally’ in Late Medieval Greek, but the colloquial word for ‘wheat’ is σιτάρι(ν) or, syncopated, στάρι(ν) (Kriaras).⁴⁵ In Modern Greek both words are still in use: σῖτος is the learned (Katharevousa), σ(ι)τάρι the ‘vulgar’ (Demotic) variant, both meaning ‘wheat’ (exclusively). Σ(ι)τάρι is of course the later outcome of the diminutive σιτάριον, glossed in LSJ in a rather disorderly way as ‘a little corn or bread, a bit of corn or breadstuff ’ and for the plural also ‘bits of food’. Montanari is better organized: ‘a little food || a little grain or grain (= σῖτος)||piece of bread).’ However, the semantic narrowing from ‘grain’ to ‘wheat’ can be traced back to Hellenistic times in the case of σιτάρι(ο)ν, as in the case of σῖτος. Preisigke 1925–31 (s.v. σιτάριον) quotes σιτάρια καὶ κριθάρια (BGU 1.33.10–11, ii–iii AD) and rightly observes: ‘hier scheint σιτάρια = πυρός zu sein’. Trapp glosses σιτάριον and its variant σιτάρι(ν) as ‘Weizen, Getreide’, but does not differentiate between

⁴² See, e.g., the scholia on Homer, Aristophanes and Oppian recorded in the TLG. ⁴³ The APIS translation in the Papyrological Navigator reads ‘grain and barley’ [sic]. A similar example is quoted in Preisigke (s.v. σῖτος 2: ‘Weizen’). Sophocles glosses σῖτος as ‘wheat’, quoting just one example: τὸν ἐπὶ τοῦ σίτου ὄντα (Epict. 1.10.2), translated as annonae praefectum. The point was already made by Shipp (1979), 499. ⁴⁴ See also Shipp (1979), 499. ⁴⁵ Shipp quotes τὸ σειτάριον καὶ τὴν κρειθήν (BGU 249, ii AD) ‘which seems to show that κριθάριον, mod. κριθάρι, is the later of the two diminutives’ (Shipp 1979, 499).

Ancient, Medieval, and Modern in the Lexicon

191

the two meanings. Sophocles simply equates σιτάριν (sic) with σῖτος and glosses ‘wheat’. This is the only meaning left in Late Medieval Greek: Kriaras simply glosses σιτάριον as ‘σιτάρι’. Both Trapp and Kriaras, however, record a more technical sense of σιτάριον: ‘Weizenkorn (als Gewichtsmaß, = 1/4 κεράτιον = 0,046 g)’ (Trapp; cf. Kriaras 1968 s.v. σιτάριον 2: ‘Μονάδα μέτρησης βάρους’). The weight measure is explained in the undatable collection De ponderibus et mensuris falsely attributed to Galen: τὸ κεράτιον σιτάρια δˊ (fr. 58.26, cf. 53.8). The diminutive σιτίον, usually used in the plural, has the same basic meanings as σῖτος according to LSJ: 1: ‘grain, corn’, 2: ‘food made from grain, bread’, 3: ‘generally, victuals, provisions . . . food’. The relation between the three meanings is again metonymic, but the semantic narrowing observed in σῖτος and its diminutive σιτάριον is poorly if at all attested in later Greek: Preisigke glosses ‘Speise, Kost, Nahrung’, whereas Sophocles and Trapp do not list the word at all. Kriaras 1968 has a separate entry for σιτίον with very few references (and only in the quoted form, not σιτί(ν) and glossed as 1: σιτάρι, 2: τρόϕιμα, ϕαγώσιμα, προμήθειες (= LSJ’s meaning 3), but it is doubtful that σιτάρι here refers to ‘wheat’ rather than ‘grain’. Babiniotis lists σιτίο, which is glossed as η τροϕή που χρειάζεται ο άνθρωπος για να ζήσει ‘the food that a man needs to live’, but the word is surely learned and in any case obsolete, as it is not listed in Triantafyllidis. The feminine σιτία is not mentioned in LSJ, but listed in Lampe with a single reference (Didache Apostolorum 13.5; ii AD): ἐὰν σιτίαν ποιῇς, τὴν ἀπαρχὴν λαβὼν δὸς κατὰ τὴν ἐντολήν ‘when you make σιτία, take the first fruit and give it according to the commandment’

Lampe translates σιτία here as ‘bread’, Bauer-Aland 1988 as ‘Teig’. Trapp, who does not mention this particular instance, distinguishes three different meanings which can be linked by consecutive metonymic extensions: 1: ‘Teig’, 2: ‘Brot’, 3: ‘Speise’. The first meaning seems to be clearly represented in Trapp’s first reference to the Apophthegmata Patrum (PG 65.192A): καὶ λαβὼν σιτίαν εἰς τὸ ἀρτοκοπεῖον, ἐποίησεν ἄρτους ‘and having taken the dough to the bakery, he made breads’

Remarkably, the second meaning is equally clearly represented in the same collection of Apophthegmata (PG 65.196B): ἀπῆλθον οὖν εἰς τὸ ἀρτοκοπεῖον ποιῆσαι δύο σιτίας ‘So I took off to the bakery to make two breads’

Sophocles refers to both passages (and only these) and translates σιτία in either case as ‘batch’, which perhaps best catches the ambiguity. Trapp’s only

192

Mark Janse

reference for his third meaning is to a probably corrupt passage in John of Damascus’ (vii–viii AD) Sacra Parallella (PG 65.196B): διὰ τριῶν σείεται ἡ γῆ, τὸ δὲ τέταρτον οὐ δύναται ϕέρειν ἐὰν οἰκέτης βασιλεύσῃ, καὶ ἄϕρων πλησθῇ σιτίαν, καὶ οἰκέτις ἐὰν ἐκβάλῃ τὴν ἑαυτῆς κυρίαν, καὶ μισητὴ γυνὴ ἐὰν τύχῃ ἀνδρὸς ἀγαθοῦ ‘by three things the earth is shaken, the fourth it cannot bear up: when a servant becomes king, and a fool is filled with food, and a servant displaces her mistress, and when a contemptible woman gets a good man’

This is obviously a literal quotation from Proverbs (30.21–3), which has the expected genitive σιτίων instead of the ungrammatical accusative σιτίαν. In the absence of other evidence it seems better to exclude the meaning ‘food’ altogether. The putative meaning ‘dough’, however, is equally problematical. Niederwimmer calls the Didache passage a crux interpretum in his commentary and hesitatingly translates the expression σιτίαν ποιεῖν as ‘einen Teig(?) anmachen’ (Niederwimmer 1993, 233). Reviewing the remaining evidence, he admits: ‘Nicht ausschließen möchte ich, daß das auffällige σιτίαν der Did.-Stelle auf das Konto des Kopiisten geht, während der ursprüngliche Text gelautet haben köntte: ἐὰν σιτία (! von σιτίον) ποιῇς’ (1993, 233 n. 17). He adds that the Ethiopian version reads ‘bread’ and that the unknown compilator of the Didache labelled Const[itutor] (1993, 45–7) paraphrases σιτία(ν) as ἄρτοι θερμοί. This leaves us with just a single instance in which σιτία is glossed as ‘dough’ in Trapp: the first example from the Apophthegmata quoted above. I would like to argue that the meaning here is ‘grain’ instead of ‘dough’ (unless σιτίαν is again a scribal error for σιτία) and that σιτία had the same meanings as σιτίον, except that the meaning ‘food’ is not securely attested. I would further argue that the feminine singular σιτία was reanalysed from the neuter plural σιτία, which often has singular or collective interpretations, perhaps to accommodate singular and plural uses for the meaning ‘bread’. As a matter of fact, the second example from the Apophthegmata deserves to be quoted in its full context (PG 196B-C): ὅτε ἤμην νεώτερος, εἰς τὴν ἔρημον ἔμενον. ἀπῆλθον οὖν εἰς τὸ ἀρτοκοπεῖον ποιῆσαι δύο σιτίας, καὶ εὗρον ἐκεῖ ἀδελϕὸν θέλοντα ποιῆσαι ἄρτους, καὶ οὐκ εἶχέ τινα δοῦναι αὐτῷ χεῖρα. ἐγὼ δὲ ἀϕῆκα τὰ ἐμὰ, καὶ ἔδωκα αὐτῷ χεῖρα. ὡς δὲ ἐσχόλασα, ἦλθεν ἄλλος ἀδελϕὸς, καὶ πάλιν ἔδωκα αὐτῷ χεῖρα, καὶ ἐποίησα τὰ ψωμία. καὶ πάλιν τρίτος ἦλθε, καὶ ἐποίησα ὁμοίως· καὶ οὕτως ἕκαστον τῶν ἐρχομένων ἐποίουν· καὶ ἐποίησα ἓξ σιτίας. ὕστερον δὲ ἐποίησα τὰς δύο σιτίας τὰς ἐμὰς, ἀποσχόντων τῶν ἐρχομένων. ‘when I was younger, I lived in the desert. So I took off to the bakery to make two breads, and there I found a brother who wanted to make breads, and he didn’t have anyone to give him a hand. So I left my stuff, and gave him a hand. As I was at it, another brother came, and again I gave him a hand, and I made the breads.

Ancient, Medieval, and Modern in the Lexicon

193

And then a third one came, and I did the same, and so I treated each of those who came, and I made six breads. Afterwards I made my own two breads, while those who came kept off.’

In this passage we actually have three words that denote ‘bread’: σιτία, ἄρτος, and ψωμίον. They are all used in combination with the verb ποιεῖν, so it seems as if the speaker treats them as if they were synonymous. This brings us back to ψωμίον and the word from which it derives, ψωμός. The latter literally means ‘bite, bit’ (cf. supra on the etymology of *ψήω). This ‘bite’ could be a ‘bit’ of anything. When the blind-drunk Cyclops (pun intended) vomits, wine and ψωμοὶ ἀνδρόμεοι ‘human bits’, i.e. ‘morsels of human flesh’ spurted out from his throat (Od. 9.374), echoing the previously mentioned ἀνδρόμεα κρέ(α) (Od. 9.297). More often than not, however, ψωμός is used to refer to a piece of bread, whence Montanari’s gloss ‘small piece, of bread or sthg. else’. This is particularly evident when ψωμός is contrasted with ὄψον ‘cooked or otherwise prepared food, a made dish, eaten with bread and wine’ (LSJ) or ‘food (cooked on a fire), dish, esp. of meat; post-Hom. usu. dish, also not of meat, consumed with bread’ (Montanari). In an oft-quoted passage from Xenophon’s Memorabilia (3.14.1–6), Socrates observes that some people eat more ὄψον than bread, or no bread at all, whence they are called ὀψοϕάγοι (3.14.2). The words for bread used by Socrates are σίτος (3.14.2–4, six times), ἄρτος (3.14.4, once) and ψωμός (3.14.5–6, twice). It is clear that a distinction is made between σῖτος ‘bread’ as a generic term, ἄρτος as a ‘loaf of bread’ (LSJ) and ψωμός as a ‘piece of bread’. The collocation ψωμὸς ἄρτου occurs four times in the Septuagint (Jd. 19.5; 1Ki. 28.22; 3Ki. 17.11; Pr. 28.21) and many more times in Christian authors who quote or comment on these passages. Preisigke does not mention ψωμός, but the word is attested in two papyri (BGU 1.52.15–6 = 1.53.17–8, 161 BC): ἀπὸ Χοίαχ ι ἕως τῆς σήμερον ἡμέρας οὐκ ἰλήϕασι ψομόν ‘from Choiach 9th up to the present day they haven’t received ψωμός’

The abstract mentions ‘Forderung nach Zahlung der Rückstände (Gehalt) der Zwillinge in Brot’ (my italics), but it is more likely that οὐκ ἰλήϕασι ψομόν simply means ‘they haven’t received a bit’ in light of the immediately preceding clause (BGU 1.52.14-5, cf. 1.53.16–7): ἀπὸ Θῶυθ α ἕως Χοίαχ θ οὐκ ἰλήϕασι ἄλλ’ ἢ τὸ ἥμισυ ‘from Thouth 1st until Choiach 8th they have received but half of it’

Kretschmer (1926, 60), with reference to Sophocles, notes that the semantic narrowing from ‘piece’ (of bread) to ‘bread’ is attested for ψωμός only after 400 AD, as in the following example from Evagrius Ponticus (Sent. Mon. 15.4): ϕάγε τὸ ψωμόν σου μετ’ αὐτοῦ ‘eat your bread with him’

194

Mark Janse

The Papyrological Navigator has no attestations of ψωμός at all, except for the two instances just discussed. By the twelfth century ἄρτος had to be glossed as ψωμός in the Etymologicum Magnum as well as in a scholion to Aristophanes (Pl. 320), although the Suda glosses ἄρτος as ψωμός as well as ψωμός as ἄρτος. The diminutive ψωμίον has a similar fate. Its original meaning is a ‘little bite’ or a ‘little bit’, but its use in the gospel according to John is noteworthy in light of the immediately preceding quotation (Ev.Jo. 13.18 = Ps. 41.10): ὁ τρώγων μου τὸν ἄρτον ἐπῆρεν ἐπ’ ἐμὲ τὴν πτέρναν αὐτοῦ ‘he who eats my bread has lifted up his heel against me’

Moments later, when Jesus is asked who will betray him, he replies (Ev.Jo. 13.26, cf. 13.27, 13.30): ἐκεῖνός ἐστιν ᾧ ἐγὼ βάψω τὸ ψωμίον καὶ δώσω αὐτῳ βάψας οὖν τὸ ψωμίον δίδωσιν Ἰούδᾳ Σίμωνος Ἰσκαριώτου ‘it is he for whom I will dip the piece (of bread) and give to him, so after dipping the piece (of bread) he gives it to Judas, son of Simon Iscariot’

The use of βάψω indicates that the ψωμίον is indeed a piece of the ἄρτος just mentioned, whereas ἄρτος is used in the synoptic accounts of the last supper because Jesus has to break it before giving the pieces to his disciples (Ev.Luc. 22.19, cf. Ev.Matt. 26.26; Ev.Marc. 14.22): καὶ λαβὼν ἄρτον εὐχαριστήσας ἔκλασεν καὶ ἔδωκεν αὐτοῖς ‘and he took the bread, gave thanks, broke it and gave (it) to them’

Although much attention has traditionally been given to John’s use of ψωμίον in the passage just quoted, as if it were announcing, so to speak, the semantic narrowing of the word, it should be noted that the usual word for ‘piece’ (of bread) is κλάσμα (ἄρτου), as in the account of the feeding of the five thousand off five breads (Ev.Matt. 14.13–21; Ev.Marc. 6.22–44; Ev.Luc. 9.10b–17: Ev.Jo. 6.1–15).⁴⁶ LSJ and Trapp list ἀρτόκλασμα ‘morsel of bread’, which seems to have been used exclusively by John Tzetzes with reference to the miraculous multiplication of the loaves (Hist. 8.49; cf. Ep. 46.6). Because of the religious symbolism of the breaking of the bread at the last supper (Ev.Matt. 26.26; Ev.Marc. 14.22; Ev.Luc. 22.19), the term ἀρτοκλασία, first attested in the fifteenth century according to Trapp and Kriaras, is still used in the Greek orthodox liturgy today as a religious re-enactment of the feeding of the five thousand. As with ψωμός, the meaning ‘bread’ is well attested for ψωμίον from the fourth century onwards, as Kretschmer observes (1926, 60). Lampe distinguishes 1: ‘morsel of bread’ (with reference to the passage from John just quoted) and 2: ‘loaf ’. For the latter meaning he refers to the (longer) passage ⁴⁶ Similarly in of the feeding of the four thousand (Ev.Matt. 15.32 9; Ev.Marc. 8.1 10).

Ancient, Medieval, and Modern in the Lexicon

195

from the Apophthegmata cited above and to the following papyrus fragment (P.Lond. 6.1914, after 335 AD): μὴ ἀμελήσηται οὖν περὶ ἡμῶν, ἀδελϕοί, ἐπιδὴ τὰ ψωμία ἀϕῆκαν ὀπίσω . . . ἐγὼ γὰρ ἀγοράζων ἄρτους εἰς διατροϕὴν ἠγώρασα ἀρτάβην σίτου . . . ἐπὰν οὖν εὕρηται εἰδήμωνα ἀποστίλατέ μοι ὀλίγα ψωμία ‘Don’t be neglectful of us, brothers, as we have left the breads behind . . . For I, who buys breads for nourishment, bought an artaba of wheat . . . When you have found an expert, send me a few breads’

The Papyrological Navigator, however, contains a number of pre-fourth century documents in which the meaning ‘bread’ is secured, including a λόγος ψομίων ‘account of breads’ (O.Did. 96; ii–iii AD), another account which lists ψομίων (διώβολον) ‘two obols of bread’ and πλακουντίου (τετρώ βολον) ‘two obols of cake’ (P.Dubl. 17.15–6; ii–iii AD) and an another one with an (illegible) amount of artabae τοῖς ἀρτοπ(οιοῖς) εἰς ψωμία ‘for the bakers to make breads’ (SB 20.14197.130–1; 253 AD). We can therefore safely assume that the meaning ‘bread’ was already current in the third, perhaps even in the second century AD The earlier attestations do not really refute Kretschmer’s reliance on the in his time available fourth-century evidence, as he cautiously writes: ‘die Verwendung des Wortes für Brot began spätestens im 4. nachchristlichen Jahrhundert’ (1926, 60). As a matter of fact, Kretschmer cleverly invokes a word rarely attested in the literature but very similar in formation to the μελίκρατον mentioned in Strabo’s account of the feeding of the crocodiles in the Arsinoite nome: βουκάκρατον, translated as ‘bread steeped in wine’ in Lampe and as ‘in Wein getunktes Brot’ in Trapp. Kretschmer quotes the passage from Pseudo-Athanasius’ Historia de Melchisedech, a retelling of Abraham’s encounter with Melchizedek, who offers him ἄρτους καὶ οἶνον ‘breads and wine’ (Gen. 14.18). The wording of the passage is very interesting and amusing (PG 28.529C): ἐπέδωκεν αὐτῷ ποτήριον οἴνου ἀοράτως ἐπιβαλὼν αὐτῷ καὶ κλάσμα ἄρτου, ὃ λέγεται βουκάκρατον ἕως τῆς σήμερον ἡμέρας ‘he gave him a cup of wine, in which he had surreptitiously steeped a piece of bread, which is called βουκάκρατον up to the present day’

The use of the phrase κλάσμα ἄρτου instead of ψωμίον is noteworthy, but even more so the suggestion that the word βουκάκρατον had been in use from the days of Abraham, whenever he lived if he ever lived, until the days of the author. Kretschmer’s second (and last) mention is taken from the Acts of Cosmas and Damian (Act. Cosm. 119.41). In addition, βουκάκρατον is used

196

Mark Janse

four times in the Vita Sancti Dosothei (Dor. Gaz. Dos. 11.21–9), where it is described as a ϕιάλην ἔχουσαν οἶνον καὶ ἄρτον (11.23), which Dorotheus (of Gaza, vi AD) took as an act of hospitality (ὡς ξενιζόμενος, 11.25) on the part of his disciple Dositheus, who actually expected a blessing (ὡς ἵνα λάβῃ εὐλογίαν, 11.24). The scene is referred to by Theodorus Studita (Cat. Mag. 79.12; viii–ix AD), who is the last one to use the word, as far as we know. Judging from the scarce evidence it would seem that the word was used exclusively in the Byzantine Diocese of the East.⁴⁷ The word is explained by Kretschmer as a so-called dvandva-compound, i.e. a copulative compound combining the words βούκα and ἄκρατον (explained in the same way in Sophocles), not therefore not from βούκα and -κρατον (as in μελί-κρατον): ‘weil für jenen Imbiß auch anderwärts ungemischter Wein bezeugt wird’ (1926, 61).⁴⁸ Indirect confirmation for this interpretation can be found in the answer of Barsanuphius (of Gaza, vi AD) to the question ‘What should I give to the poor who go from house to house?’ (Resp. 635):⁴⁹ καθὼς εὑρίσκει ἡ χείρ σου, εἴτε μικρὸν βουκίον, εἴτε ἄκρατον ‘whatever comes to your hand, whether a little piece of bread, or some unmixed wine’

Now βούκα is obviously a Latin loanword. The TLL provides the following information s.v. bucca: ‘forma vulgaris buca’, ‘vulgariter pro ore . . . inde transl. i.q. offa’, the latter meaning ‘bite, bit’, literally a ‘mouthful’. Augustus writes in a letter: duas buccas manducavi ‘I’ve eaten two mouthfuls’ (Suet. Aug. 76),⁵⁰ translated as ‘two mouthfuls of bread’ in J.C. Rolfe’s 1914 Loeb translation.⁵¹ The phrase bucca panis is used by Petronius (44.2) and Martial (7.20.8), so the similarity with ψωμίον should be obvious, both semantically and phraseologically. Bucca was borrowed in Greek as βούκα, but the rare attestations are all very late. A scholion (ix–xiii AD) on the origin of the term βουκελλάριος in the ninth-century Byzantine code of law called Basilica (after its initiator Basil I) reads (Bas. B3536.11): ἐκ δὲ τοῦ βούκα ὀνόματος, ὅ ἐστιν ὁ ἄρτος, ἐκλήθησαν βουκελλάριοι ‘after the word βούκα, which means bread, they were called bucellarii’

⁴⁷ Cosmas and Damian were born in Syria and died as martyrs in Cilicia. ⁴⁸ This explains the life threatening activity of the children in the Acts of Cosmas and Damian: τρία παιδία βουκάκρατον ἐσθίοντα δαψιλῶς, i.e. ‘in too great quantity’ (Act. Cosm. 119.41). ⁴⁹ See below for discussion of the diminutive βουκίον. ⁵⁰ Bread was Augustus’ preferred food according to the fragments of his letters quoted by Suetonius (Aug. 76). ⁵¹ The translation is left unchanged in the revised and updated 1998 edition.

Ancient, Medieval, and Modern in the Lexicon

197

The word occurs several times in the works of Theodore Prodromos (1.352, 2.203, 4.249) and is glossed as ‘μπουκιά’ in Kriaras and as ‘mouthful, morsel’ in Sophocles, who lists it under the unattested βούκκα.⁵² Related to βούκα is βουκία, glossed ‘Feinbrot’ in Preisigke 1925–31. It is not listed in LSJ, but the Supplement glosses it as ‘a kind of cake or biscuit’. Both refer to an early first-century, extremely lacunary, papyrus fragment which reads: βουκίαι, κολλύραι (P.Oxy. 2.397.32),⁵³ where LSJ ‘roll or loaf of coarse bread’ and Preisigke ‘großes Brot’ again differ about the exact interpretation of κολλύρα. The etymological clarification in LSJ Revd Suppl. that βουκία ‘perh. = Mod. Gr. βούκα fr. Lat. bucca’ misses the point in that βουκία is actually borrowed from Vulgar Latin buccea, attested as a varia lectio in the letter from Augustus quoted by Suetonius (and as such listed in Lewis and Short). The word has survived into Modern Greek as βουκιά and is listed in Triantafyllidis as ‘λαϊκότροπος’, which is a technical term to characterize a word as substandard but with a wide, supradialectal, geographical distribution. Finally, there is the diminutive βουκίον (βουκίν), which is not listed in LSJ Revd Suppl., but again glossed as ‘Feinbrot’ in Preisigke, ‘biscuit, morsel of bread’ in Lampe and, nicely, ‘little βούκκα’ in Sophocles. The earliest attestations are sixth-century, e.g. Barsanuphius (quoted above), who also testifies that a βουκίον fits the mouth (Resp. 512): ἔβαλον τὸ βουκίον εἰς τὸ στόμα μου ‘I put the (mouthful of ) bread into my mouth’

There is one later attestation which deserves to be quoted in full (Anastasius Sinaita, Hod. 2.3 = PG 89.57A; vii AD): ϕύσις καὶ οὐσία καὶ γένος καὶ μορϕὴ ἓν καὶ τὸ αὐτό εἰσιν ἐν τοῖς ἐκκλησιαστικοῖς δόγμασιν, ὥσπερ ὁ ἄρτος καὶ τὸ ψωμίον καὶ βουκίν ‘nature and substance and kind and form are one and the same in the doctrine of the Church, like the (loaf of) bread, the (piece of) bread and the (mouthful of) bread)’

A literal translation of the three words for ‘bread’ in the passage just quoted perhaps does not do full justice to the religious connotations of the first two: ἄρτος as the bread from the ἀρτοκλασία and the last supper, ψωμίον as the piece of bread Jesus gave to Judas. The juxtaposition nevertheless confirms that all three could be used to refer to ‘bread’ and that ψωμίον and βουκίον by this time had followed similar paths of semantic narrowing but at a different pace, which explains their near but not total synonymity.

⁵² Sophocles ignores the ‘vulgar’ Latin spelling buca and deems βούκα ‘incorrect for βούκκα’. ⁵³ The first thirty one lines are missing.

198

Mark Janse

The final question to be answered is why ψωμί(ον) eventually prevailed and became the Modern Greek word for ‘bread’, both as a countable and as an uncountable noun. The answer was formulated almost a century ago by Kretschmer: ‘das Brot kam in jener Zeit dem Publikum, den Speisenden vielfach nicht als ganzes, sondern zerstückelt zu’ (1926, 62). He refers to the semantic narrowing of Latin bucca and especially its diminutive buc(c)ella, which is generally glossed as ψωμός or ψωμίον (cf. TLL for references). Kretschmer notes how the ψωμίον in the Judas passage (Ev.Jo. 13.26–30) is translated as bucellam in the Itala (Vetus Latina), but as panem in the Vulgate, and as hlaif in the Gothic translation of Wulfila who, in his translation of the feeding of the five thousand, renders κλάσματα as drauhsnos (6.12) and κλασμάτων as gabruko (6.13). He concludes that the semantic narrowing from ‘piece (of bread)’ to ‘bread’ must have taken place between the second and the fourth century, which agrees even with the pre-fourth century evidence quoted above. Kretschmer is aware of the fact that Jerome does not translate ψωμίον consistently as panem (1926, 61 fn. 1). As a matter of fact, it is rendered as panem with reference to Jesus’ dipping the ψωμίον (13.26 bis) and as bucellam with reference to Judas’ accepting it (13.27, 30). Interestingly, the Old English Lindisfarne Gloss on bucellam in the Judas passage is bréad, that of panem is laf. The Anglo-Saxon translation reads bitan and hláf respectively. This ‘seems to show that bréad was not yet identified with panis’ (OED s.v. bread). Indeed, the later Blickling Gloss has bréadru for Latin frusta ‘pieces, bits’. The OED concludes that English bread and its West Germanic cognates ‘appear . . . to have originally meant ‘piece, bit, fragment, Latin frustum’: but already in Old Saxon and Old High German it had acquired the sense of ‘bread’’, against the traditional and still commonly entertained derivation from Proto-Germanic *breuwan- ‘brew’ (e.g. Kroonen 2013, 74 s.v. *brauda-). The West Germanic words for ‘bread’ again illustrate the metonymic extension from ‘piece’ to ‘piece of bread’ to ‘bread’, whereas the Germanic words for ‘loaf ’ illustrate the semantic narrowing from ‘bread’ to ‘loaf of bread’.⁵⁴ Kretschmer, ignoring the Germanic evidence, nevertheless rightly concludes: ‘Diese Verwendung der Worter für Bissen, Stück Brot stammt offenbar aus der Wirtshaus- und Familiensprache. Der Gast verlangte und erhielt ein ψωμίν, nicht einen ganzen ἄρτος. Für das Bukakraton brauchte man ebenfalls einzelne Brotstücke, nicht einen ganzen Laib’ (1926, 63). Kretschmer mentions other words used in the same context which have undergone similar

⁵⁴ Several South Slavic languages illustrate the same metonymic extension from ‘piece’ to ‘bread’, e.g. Serbian, Croatian and Slovenian kruh ‘bread’, cf. Old Church Slavonic кроухъ ‘piece’ < PIE. *krous (Derksen 2008, 252 s.v. *kruxъ; DELG s.v. κρούω). The OED also mentions the use of piece as referring to ‘[a] portion of bread, esp. eaten on its own rather than as part of a meal’ in Irish, Scottish and other northern English dialects.

Ancient, Medieval, and Modern in the Lexicon

199

metonymic extensions such ὄψον ‘a made dish’ but ‘at Athens, esp. of fish, the chief delicacy of the Athenians’ (LSJ, with reference to Plu. 2.667f. and Ath. 7.276e), whence ὀψάριον, uninformatively glossed as ‘Dim. of ὄψον’ in LSJ, at Athens again especially of fish (Kretschmer 1926: 63, with references in footnote 1) which eventually developed into Modern Greek ψάρι. Or, with Kretschmer’s touch of humour: ‘Jetzt schwimmen aber für den Griechen die “Zuspeisen” auch im Meer umher, wie schon Euripides in der Κρῆσσα bei Athen. XIV 640 B von ὄψων ποντίων und Hippokrates von ὄψα θαλάσσια redet’ (1926, 63f.).⁵⁵ All this explains why ψωμίον came to be used for ‘bread’ instead of ἄρτος, but why is the latter still used in Modern Greek? The answer is to be found in Triantafyllidis, where ἄρτος is characterized on the one hand as λόγ[ιος] ‘learned’, on the other as εκκλ[ησιαστικός] ‘ecclesiastical’. As mentioned in the introduction, ἄρτος has survived in a number of compounds thanks to the Katharevousa such as the above mentioned αρτοποιείο ‘bakery’, glossed as ϕούρνος, the most commonly used word, in Babiniotis and Triantafyllidis, who adds the much less frequently used ψωμάδικο. It is also used in a number of phrases to indicate special varieties of bread such as κρίθινος ‘barley’, πολύσπορος ‘multigrain’, λευκός ‘white’ or μέλας ‘dark’. The last two are called άσπρο and μαύρο ψωμί in Demotic. Shipp notes that ‘it has been suggested that the loss of ἄρτος in the common speech was caused by the church use, a kind of tabu, as also the loss of ὕδωρ. It would fit in well with this view that the words survived in the part of the Greek world which remained pagan into the middle ages’ (Shipp 1979, 101–2.).⁵⁶ This may well be the case, as Greek children in primary school are taught how to ask for their daily bread in accordance with the Lord’s prayer: τὸν ἄρτον ἡμῶν τὸν ἐπιούσιον δὸς ἡμῖν σήμερον (Katharevousa) το καθημερινό μας ψωμί δώσ’ μας για σήμερα (Demotic)

⁵⁵ As the title of Kretschmer’s article indicates, the development of Modern Greek κρασί is also briefly discussed, but I leave it for another occasion to delve into that question (see footnote 27 on μελίκρατον above). ⁵⁶ Shipp refers specifically to Tsakonian and some northern varieties (Shipp 1979, 102).

12 Greek Dialects in the Lexicon Philomen Probert

12.1. ABSTRACT This chapter considers the treatment of dialect forms in LSJ and the origins of LSJ’s practice.¹ It begins by illustrating a well-known fact: although Aeolic and Doric forms of words sometimes get their own LSJ entries, the main or most informative entry (to which others may cross-refer) is for a non-Aeolic or nonDoric form wherever this is reasonably sensible. At first sight, the obvious conclusion is that some non-Aeolic and non-Doric dialect, such as Attic, functions as the basic dialect for LSJ. On closer inspection, however, it turns out that the Lexicon is not built on a principle of treating any one variety of Greek as ‘basic’. Instead LSJ operates with the notion of a normal or default form: a form judged to be available for use in the widest range of Greek texts. In some entries, the label ‘common form’ appears as a label for forms of this type. Importantly, by designating a form as the ‘common form’ or choosing it as the basic dictionary entry, LSJ make a judgement about the wide availability of the form in principle, not about its actual attestation. Although ‘common forms’ are often the most widely attested, they need not actually be attested at all. The chapter goes on to show how this concept of a ‘common form’ is reminiscent of a concept found in some ancient grammatical texts. In the modern lexicographical tradition, the concept is found already in the first edition of Liddell and Scott, and there are precursors in Passow’s lexicon. However, Liddell and Scott appear to have moved closer to the ancient model than Passow. An important source of influence is likely to have been the ¹ I am grateful to the organizers of the conference from which this volume originates, as well as to the organizers of the VII Colloque sur les dialectes grecs anciens (Basel and Lausanne, 2017) and to the Classics Department of the University of California, Irvine, for opportunities to present work relating to this chapter and for valuable discussion. I am further grateful to Eleanor Dickey and to the editors of this volume for feedback on a draft written version, and to Elizabeth Tucker, David Raeburn, and Peter Parsons for personal insights into approaches to the Greek dialects in the English classical tradition. Philomen Probert, Greek Dialects in the Lexicon In: Liddell and Scott: The History, Methodology, and Languages of the World’s Leading Lexicon of Ancient Greek. Edited by: Christopher Stray, Michael Clarke, and Joshua T. Katz, Oxford University Press (2019). © Oxford University Press. DOI: 10.1093/oso/9780198810803.003.0012

Greek Dialects in the Lexicon

201

didactic tradition in Victorian England, and a historical link to the ancient and medieval tradition can be traced via the Renaissance.

12.2. INTRODUCTION: LOOKING UP DIA LECT W ORDS IN LSJ Suppose we were trying to read the following fragments of Sappho and Alcaeus, and we wanted help with ὂν δ’ ἔψυξας in the Sappho and πεδαλευόμενος in the Alcaeus: ἦλθες, †καὶ† ἐπόησας, ἔγω δέ σ’ἐμαιόμαν, ὂν δ’ ἔψυξας ἔμαν ϕρένα καιομέναν πόθῳ. (Sappho, fr. 48 V.) κάτω γὰρ κεϕάλαν κατί σχε[ι] τὸν ϝὸν θάμα θῦμον αἰτιάμενος, πεδαλευόμενος τά κ’ εἴ πη (Alcaeus, fr. 358.4 7 V.)

We might reasonably turn to the current edition of Liddell and Scott: at present the ninth edition with Revised Supplement (LSJ). But what exactly do we look up? If we are really stuck with ὂν δ’ ἔψυξας, we might look for an entry for ὄν. Helpfully, we find an entry for ὀν-, telling us that ὀν- is ‘Aeol., Cypr., etc. for ἀν-’. Armed with this knowledge, we might work out that ὂν…ἔψυξας is a prefixed verb in tmesis. There is no entry ὀμψύχω, but once we know that ὀν- is a form of ἀνά we can look under ἀναψύχω, where we learn that this verb means (for example) ‘cool’, ‘refresh’, or ‘revive’. If we take the trouble to look up ἀναψύχω in the Revised Supplement, we even find our fragment of Sappho cited (albeit with a typographical error, ὃν for ὂν). Turning to πεδαλευόμενος, we might now try to be clever. Remembering that πεδα- is a dialectal equivalent of μετα-, we look under μεταλεύω or μεταλεύομαι. But there is no such entry. In fact we have to look under πεδα-: the participle πεδαλευόμενος turns out to have its own entry, telling us that Hesychius glosses the word with μεταμελόμενος and μεταδιωκόμενος.² Why then does πεδαλευόμενος get its own entry, when ὀμψύχω has to be looked up under ἀναψύχω? Part of the answer is that ἀναψύχω is attested (indeed very well attested), but no verb μεταλεύω is attested, and quite possibly none ever existed. However, this difference in attestation is not the whole story. Firstly, some LSJ entries do have unattested forms as their headwords; this is a point to which we shall return. Secondly, it is a more general principle that prefixed verbs with ὀν- need to be looked up under ἀνα-. To find out about ὀνέτροπε in Alcaeus, for example, we have to look under ἀνατρέπω. To find out about ὀνγράψαντας in ² Hesychius π 1180 Hansen. Incidentally, one might wonder if μεταδιωκόμενος in the text of Hesychius is corrupt for μεταδοκούμενος.

202

Philomen Probert

a Thessalian inscription, we have to look under ἀναγράϕω. An exception is the Thessalian word ὀνάλα, meaning ‘expenditure’. This has its own entry, because if we replace ὀν- with ἀν- we do not get the non-Thessalian form: the nonThessalian form is ἀνάλωμα (as indeed we learn from the entry for ὀνάλα), not ἀνάλα. But the general rule is that ὀν- needs to be turned into ἀν- or ἀνα- in order to give an LSJ headword. On the other hand, no such rule applies to words beginning with πεδα-. Any word beginning with πεδα- has its own entry in LSJ, even if in some instances this entry just tells us to look under μετα-. The entry for πεδαίρω, for example, reads ‘Aeol. or Dor. for μεταίρω (q.v.)’. The general principle is that if we are trying to look up an Aeolic or Doric word, the main, most informative dictionary entry will have a non-Aeolic or non-Doric form as its headword wherever one can reasonably be produced. But there are some operations that users are expected to be able to perform in order to find this headword (such as turning ὀν- into ἀν- or ἀνα-), and others (such as turning πεδ- or πεδα- into μετ- or μετα-) for which help is provided in the form of a subsidiary entry with the Aeolic or Doric form as its headword. To explore the general principle a bit further, suppose we are interested in a word that contains a Doric ᾱ. In general, we are expected to turn this ᾱ into an η in order to locate a headword. But where a Doric ᾱ may be difficult to identify as such, LSJ provides some help. For example, if we are reading Pindar and come across the word ἀβοατί (Nemean 8.9), we might be tempted to look under ἀβοατί—and indeed there is such an entry, telling us (in effect) to look under ἀβοητί. If we then look under ἀβοητί, we find an entry telling us that the word has a Doric form ἀβοατί, that it means ‘without summons’, and that it occurs in Pindar, Nemean 8.9. This is the very passage of Pindar we were reading, and it is essentially the only place this word occurs.³ As this example illustrates, it is a very strong principle that a word with a Doric ᾱ should have a main entry with the non-Doric form as its headword. A slightly different case is that of the Doric word σταμαγορίς. The word is attested only in this very form, in Hesychius’ lexicon. It is uncertain whether the second alpha is long or short, and hence also uncertain what the non-Doric form should be. Because the non-Doric form cannot be confidently produced, the Doric form this time provides the headword for the main dictionary entry: ‘στᾱμᾱγορίς or -ᾰγορίς, ίδος, ἡ, Dor. for *στημηγ- or *στημᾰγ- (στήμων 1, ἀγείρω) twisting of several threads of the warp into one, Hsch.’. But just in case anybody, trying to be clever, looks the word up under something starting with στημ-, LSJ also provides an entry for στημαγορίς, telling users to look under σταμ-. Even here, where LSJ refrains from putting the main entry under a nonDoric headword because the right form cannot confidently be produced, the user’s awareness of a general need to turn Doric ᾱ into η is helpfully recognized. ³ In the form ἀβοητί, the word is transmitted as a lemma in a textually problematic entry at Hesychius α 154 Latte/Cunningham (or, more relevantly for LSJ, Hesychius α 153 Schmidt), but LSJ makes no allusion to this attestation.

Greek Dialects in the Lexicon

203

In making the main entry a non-Aeolic or non-Doric form whenever such a form can confidently be produced, LSJ unsurprisingly treats non-Aeolic and non-Doric forms as more basic than their Aeolic and Doric counterparts. But what do we mean by ‘non-Aeolic’ or ‘non-Doric’? Up to now the terms ‘nonAeolic’ and ‘non-Doric’ have served as a convenient way of postponing our main question: what counts as the basic or default dialect for LSJ?

12.3. A BASIC OR DEFAULT DIALECT? At first sight, the answer to our question may seem to be obvious: surely Attic is the basic dialect. Most of the time, the forms LSJ treats as basic are indeed the ones we might think of as Attic. For example, we have seen that ὀμψύχω needs to be looked up under ἀναψύχω, and ἀβοατί under ἀβοητί. Both ἀναψύχω and ἀβοητί are Attic forms, even if ἀναψύχω is a well-attested Attic form and ἀβοητί a hypothetical one. On closer inspection, however, the principle is not as simple as this. For example, the Attic for ‘male’ is ἄρρην, and yet the main entry in LSJ has ἄρσην as its headword. The Attic words for ‘temple’ and ‘people’ are νεώς and λεώς (following the ‘Attic declension’), yet the main entries in LSJ are for νᾱός and λᾱός. The Attic for ‘sea’ is θάλαττα, yet the main entry in LSJ is for θάλασσα, and the same applies to other words with ττ in Attic and σσ in most other dialects. Attic has ξύν as the preposition and the verbal prefix, but in LSJ we always have to look under συν-. These entries suggest that the basic dialect for LSJ is not straightforwardly Attic. But if it is not straightforwardly Attic, what is it? A second possibility, one might think, is that the basic dialect is Attic not as we know it from Attic inscriptions of the fifth century BC, but as we know it from Attic tragedy and other serious Attic literature of the fifth and (to a lesser extent) fourth centuries BC. Like any other Greek literary dialect, literary Attic is not just Attic as it was spoken on the street, but an Attic that has been (as it were) toned down a little. That is to say, literary Attic tends to avoid the features that Attic shares with few other dialects—the features that would have struck other Greeks as especially peculiar. For this reason the tragedians avoid the forms ἄρρην and νεώς; they do use λεώς, but also λαός; they avoid Attic ττ; and they use σύν more often than ξύν. Thinking of LSJ’s basic or default dialect as literary Attic takes us closer to LSJ’s actual practice, and yet this cannot be an accurate account of the theory behind the practice. The entry for θάλασσα, for example, begins ‘θάλασσα [θᾰ], Att. -ττα IG 1².57 (but θάλασσα 2².236 (338/7 B.C.))’. If the headword θάλασσα were thought of as ‘Attic’, with ‘Attic’ meaning the language of Attic tragedy, it would hardly make sense to suggest that the Attic form is usually θάλαττα, or to imply that Attic forms are the ones we find on inscriptions.

204

Philomen Probert

Fig. 12.1. Beginning of the entry for ἄρσην ‘male’, LSJ. (The asterisk in a circle means the word also appears in LSJ Revd Suppl.)

An apparent third possibility is that LSJ treats the Koiné as the basic or default dialect. Like literary Attic, the Koiné is heavily based on spoken Attic, but the most peculiarly Attic features tend to have been replaced by features from Ionic or (less often) further afield. Once again we typically find ἄρσην rather than ἄρρην; ναός and λαός rather than νεώς and λεώς; -σσ- rather than Attic -ττ-; and σύν rather than ξύν. Thinking of the Koiné as LSJ’s basic or default dialect once again brings us fairly close to LSJ’s actual practice, but it would be surprising if post-classical Greek were really the point of reference. Not only does LSJ devote considerably more space to examples from classical authors than to examples from postclassical authors, but the structure of entries such as the one shown in Fig. 12.1 suggests that postclassical Greek does not provide the reference point. The headword here is ἄρσην, and this form is first of all said to occur in epic, Ionic, and tragedy. We are then told that the Attic form is ἄρρην and that a form ἔρσην is found in various other varieties of Greek. Only then do we come onto postclassical Greek: ἄρσην prevails in the Septuagint and the New Testament, while ἄρρην is more common in papyri. Although the form ἄρσην is identical to the headword, postclassical texts were not mentioned at the beginning of the entry along with epic, Ionic, and tragedy. Thus although the form ἄρσην happens to occur in some postclassical texts, this does not appear to be the reason for choosing ἄρσην as the headword.⁴

12.4. THE ‘ COMMON F ORM’ So far we have seen three varieties of Greek that do not consitute LSJ’s basic or default dialect: the search for a basic or default dialect begins to appear fruitless. Yet it would be a mistake to conclude that there is no system governing the forms treated as basic. If we want to know why the main entry for ἄρσην/ἄρρην/ἔρσην, for example, goes under the headword ἄρσην ⁴ Compare the comments on postclassical authors in the preface to the first edition of Liddell and Scott (LS¹), pp. vi vii.

Greek Dialects in the Lexicon

205

Fig. 12.2. Beginning of the entry for νᾱός ‘temple’, LSJ.

rather than ἄρρην, a first clue comes from the beginning of the entry (Fig. 12.1). As a form that occurs in epic, Ionic, and tragedy (not to mention the Septuagint and the New Testament), ἄρσην occurs in a considerable variety of different kinds of Greek. If we read a variety of Greek texts—especially texts that are traditionally widely read—we might well feel that ἄρσην is the ‘normal’ form, even though it is not the Attic form. The entry for νᾱός (Fig. 12.2) suggests the same point. The form νᾱός is here said to occur in ‘Doric, Thessalian, etc.’; to be used also in ‘Tragedy (even dialogue) to the exclusion of νεώς’; to be rare in Attic prose and comedy but more frequent in Xenophon; to be found in Attic inscriptions from the third century BC, and in Hellenistic and later Greek. Importantly, the list of varieties of Greek in which we find νᾱός is not the same as the list for ἄρσην, but again the impression is that νᾱός occurs all over the place—that νᾱός is the ‘normal’ form, even if it is not the Attic one. A further clue to the thought behind LSJ’s approach to dialects comes from a whole series of entries featuring the phrase ‘the common form’.⁵ In almost all

⁵ Relevant entries were tracked down via a search for the word ‘common’ in the electronic version of LSJ available via the Perseus Project (http://www.perseus.tufts.edu). In addition to the

206

Philomen Probert

instances, the form intended by this phrase is the one chosen as the headword for the main dictionary entry. For example the entry for ἀνδρεῖος tells us that the Ionic form is ἀνδρήιος, but the manuscripts of Herodotus have the ‘common form’ in the comparative and superlative. What this means is that the comparative and superlative forms found in these manuscripts are built on the very form ἀνδρεῖος that provides the headword for this main dictionary entry (Fig. 12.3). The entry for Περσεϕόνη tells us there is an epic form Περσεϕόνεια, and then tells us where the ‘common form’ first occurs. Once again the ‘common form’ is the very form Περσεϕόνη that provides the headword for this main dictionary entry (Fig. 12.4). The entries for σωϕροσύνη and Ῥέα illustrate the same point again (Figs 12.5 and 12.6). The first of these tells us about an epic form, and then tells us where the ‘common form’, that is to say σωϕροσύνη itself, first occurs. The entry for Ῥέα again tells us about an epic form, and then tells us that the ‘common form’, i.e. Ῥέα itself, also occurs in the Iliad.

Fig. 12.3. Beginning of the entry for ἀνδρεῖος ‘of or for a man’, LSJ.

Fig. 12.4. Entry for Περσεφόνη ‘Persephone’, LSJ.

LSJ entries discussed below, see also those for τοσόσδε (τοσόσδε is termed the ‘common’ form, by contrast with the epic form τοσσόσδε); τοσοῦτος (τοσοῦτος is termed the ‘common’ form, by contrast with the epic form τοσσοῦτος); τρέχω (ἔδραμον is termed ‘the common aor.’, by contrast with Old Attic ἔθρεξα); ὑδροχόος (the dative ὑδροχόῳ is termed ‘the common (sc. dative form)’, by contrast with the epic form ὑδροχοῆϊ); ϕθίω (ϕθίνω is called ‘the common pres.’, by contrast with the present ϕθίω and imperfect ἔϕθιον, which are said to occur once each in Homer and are probably considered epic forms). The word ‘common’ (without ‘form’) is also used in the LSJ entries for ἀναβιόω, ἀντιλέγω, ἄπειμι (B), and ἠπάομαι to distinguish one form from one or more variants, but in these instances the variation in question has nothing to do with dialects. For uses of the word ‘common’ that are not relevant in the current context see n. 13 below, on LS¹; many of the uses listed there can also be found in LSJ.

Greek Dialects in the Lexicon

207

Fig. 12.5. Beginning of the entry for σωφροσύνη ‘soundness of mind’, LSJ.

Fig. 12.6. Beginning of the entry for Ῥέα ‘Rhea’, LSJ.

In the entry for ψάμμη (Fig. 12.7) we see a slightly different pattern. ψάμμη is said to be a rarer form of ψάμμος, occuring in Herodotus—but Herodotus also uses the ‘common form’, i.e. ψάμμος. This time the ‘common form’ is not the headword of the entry in which we find the phrase ‘common form’ (i.e. ψάμμη), but a form treated as a different word with its own entry (Fig. 12.8).

Fig. 12.7. Entry for ψάμμη ‘sand’, LSJ.

Fig. 12.8. Beginning of the entry for ψάμμος ‘sand’, LSJ.

The phrase ‘rarer form’ in the entry for ψάμμη, contrasting with the phrase ‘common form’, suggest that in this instance the ‘common form’ is simply the form that occurs most often. However, two entries suggest that when the phrase ‘common form’ is used in the way that is more typical in LSJ (the pattern of use seen under ἀνδρεῖος, Περσεϕόνη, σωϕροσύνη, and Ῥέα), the ‘common form’ is not necessarily the form that occurs most often. Under the entry for ἕννυμι (Fig. 12.9), LSJ tells us there is an epic aorist ἕσσα. In addition to this epic form one might have expected an ordinary aorist with one sigma, but we learn that this ‘common form’ occurs only in compounds.

208

Philomen Probert

Fig. 12.9. Beginning of the entry for ἕννυμι ‘put clothes on (another)’, LSJ.

This time, then, the ‘common form’ is not attested at all for the simplex verb ἕννυμι whose dictionary entry we are looking at. At this point one might ask whether the aorist with one sigma is so common in compounds that it is still considered the most commonly occurring form. But the entry for ποτιμάστιος makes it clear that the notion of a ‘common form’ cannot be based even indirectly on frequency of occurrence (Fig. 12.10).

Fig. 12.10. Entry for ποτιμάστιος ‘at the breast’, LSJ.

The form ποτιμάστιος begins with a dialectal variant of προσ-, and so one might expect there to be a form προσμάστιος as well. But LSJ tells us the word is ‘not found in the common form προσμ(άστιος)’. The idea that the common form can be unattested is also implicit in entries such as the one shown in Fig. 12.11.

Fig. 12.11. Entry for συστᾰθεύω ‘help to roast’, LSJ.

We have seen that a verb with the prefix ξυν- or συν- has to be looked up under συν- (or where appropriate συ-), and this entry for συσταθεύω is no exception. LSJ tells us the word occurs in Aristophanes’ Lysistrata, and we can learn from the TLG that the word also occurs in the Suda⁶ but in no other texts ⁶ Suda ξ 165 Adler.

Greek Dialects in the Lexicon

209

in the TLG. The form found in Aristophanes and the Suda is ξυσταθεύω, and in all likelihood the form συσταθεύω never occurs at all, yet συσταθεύω is still the headword. There is of course a perfectly practical reason for this: once we know that we always have to look under συν- and not ξυν-, it would be confusing if LSJ listed some words under ξυν- instead, just because the form with συν- is unattested: how is the user to know that this form is unattested? For the present purposes the important point is that a form labelled the ‘common form’ or chosen as the dictionary entry is not necessarily the most frequent form, and may not even be attested at all. Instead the concept is that this form could potentially occur in the widest range of Greek texts (with especial weight given to the kinds of texts that are widely read), irrespective of whether it actually occurs in many texts or indeed any.

12.5. A S IMILAR CONCEPT IN ANCIENT GRAMMATICAL TEXTS LSJ’s concept of ‘common forms’ is reminiscent of a concept that appears in some ancient Greek grammatical texts,⁷ where a distinction is made between forms described as κοινά and others described as ἴδια. The following passage, for example, discusses dialects that have double consonants where other dialects do not:⁸ παρ’ Αἰολεῦσι δὲ τῶν ἀμεταβόλων γίνεται ἀγείρω ἀγέρρω, ἐγείνατο ἐγέννατο, στειλάμεναι στελλάμεναι, εἷμα ἔμμα, καὶ τοῦ σ, τοσοῦτον τέσσουτον. ἄλλου δὲ συμϕώνου παρ’ αὐτοῖς οὐ γίνεται … ἐὰν οὖν ἕτερον διπλασιασμὸν παρ’ αὐτοῖς εὕρῃς, τῷ κοινῷ ἐχρήσαντο καὶ οὐκ ἰδίως ἐποίησαν (Ep. Hom. alph. ο 94 Dyck) Among the Aeolians (doubling) of resonant consonants occurs: ἀγείρω (becomes) ἀγέρρω, ἐγείνατο (becomes) ἐγέννατο, στειλάμεναι (becomes) στελλάμεναι, εἷμα (becomes) ἔμμα. And doubling of σ: τοσοῦτον (becomes) τέσσουτον. But there is no doubling of any other consonant among them…. So if you find any other doubling among them, they have used the common (form) and not created it by themselves.

The Aeolians are said to double resonant consonants and sigma. After this point has been illustrated with a series of examples, we are told that if any other doubling of consonants is found among the Aeolians, they have ‘used the common form and not created it by themselves’. The adjective κοινός here does not denote the Koiné, but the opposite of ἴδιος: some forms are ⁷ For more detail see Consani 1991, 27 53, to which the following discussion is indebted. ⁸ On this passage see further Consani 1991, 27 8.

210

Philomen Probert

considered specific to particular dialects while other forms are considered available for use in various dialects or varieties of Greek. The latter are κοινά, or ‘common forms’, in the same sense as in LSJ. In the following passage, Theodosius discusses the paradigm of the word ὄϕις ‘snake’: Ἑνικά. Ὁ ὄϕις τοῦ ὄϕιος καὶ ὄϕεως: τὰ εἰς ΙΣ προσηγορικὰ μὴ ὄντα παρώνυμα μακροπαράληκτα διὰ καθαροῦ τοῦ ΟΣ κλίνεται, ‘μάντιος’, ‘ὄϕιος’· Ἀττικοὶ δὲ ‘μάντεως’ καὶ ‘ὄϕεως’ ϕασίν· εὑρέθη δὲ καὶ ‘ὄϕεος’ ‘δράκοντος ἔσπειρ’ ὄϕεος ἐν γαίᾳ θέρος’ καὶ ‘ὃς ἂν δύνηται πόλεος ἔν τ’ ἀρχαῖσιν ᾖ’. (Theodosius, Canones 10.7 11) Singular forms: nominative ὄϕις, genitive ὄϕιος and ὄϕεως: common nouns in ις are inflected with ος preceded by a vowel (as long as they are not derivatives of nouns having a long vowel in the penultimate syllable): μάντιος, ὄϕιος. But Attic speakers say μάντεως and ὄϕεως. And ὄϕεος is also found, as in δράκοντος ἔσπειρ’ ὄϕεος ἐν γαίᾳ θέρος and ὃς ἂν δύνηται πόλεος ἔν τ’ ἀρχαῖσιν ᾖ.

After telling us that there are two genitives, ὄϕιος and ὄϕεως, Theodosius gives a general rule suggesting that he takes ὄϕιος as the basic one. But in Attic, he says, the form is ὄϕεως, and ὄϕεος is also found. Theodosius here does not use the term κοινός, but Choeroboscus explains in his commentary on this passage that Theodosius takes the form ὄϕιος as ‘common’ and ὄϕεως as Attic:⁹ Ἰστέον δὲ ὅτι ὁ τεχνικὸς δοξάζει τὴν μὲν διὰ τοῦ Ι καὶ Ο γενικὴν κοινὴν εἶναι, τὴν δὲ διὰ τοῦ Ε καὶ Ω Ἀττικήν κρεῖττον δέ ἐστι τὴν διὰ τοῦ Ι καὶ Ο Ἰωνικὴν εἶναι λέγειν γενικήν, ἐπειδὴ οἱ Ἴωνες αὐτῇ κέχρηνται, τὴν δὲ διὰ τοῦ Ε καὶ Ο κοινήν, εἰ καὶ ὁμοίως τοῖς Ἀττικοῖς διὰ τοῦ Ε καὶ Ω δεῖ γράϕειν τὰς τοιαύτας γενικάς. (Choer. Th. 1.201.24 8) One must know that the grammarian thinks the genitive in ιο is common, and the one in εω is Attic. But it’s better to say that the genitive in ιο is Ionic, since the Ionians use it, and that the one in εο is common, even if one ought to write such genitives in εω , like the Attic speakers do.

In Choeroboscus’ opinion, Theodosius is wrong about the status of some of the genitive forms: Theodosius took the genitive in -ιο- to be the common one, but it is preferable to say that this genitive is Ionic and that the one in -εο- is common. The final comment of the quoted passage makes clear, in addition, that for Choeroboscus the notion of a ‘common form’ is different from the notion of a correct or preferable form: the forms in -εο- are common, but one ought really to use those in -εω-.

⁹ I take ὁ τεχνικός to mean Theodosius rather than Herodian here; differently Consani 1991, 38, with n. 67.

Greek Dialects in the Lexicon

211

It is also clear that when Choeroboscus labels a form κοινός, he does not mean that this is the Koiné form. The passage just quoted is preceded by a comment on the usage of the Koiné, which is labelled ἡ κοινὴ συνήθεια and ἡ κοινὴ διάλεκτος: ἡ δὲ κοινὴ συνήθεια μᾶλλον ἀττικίζουσα διὰ τοῦ Ε καὶ Ω γράϕει ταύτας τὰς γενικάς εἰς τὰ πολλὰ γὰρ ἡ κοινὴ διάλεκτος τοῖς Ἀττικοῖς ἕπεται. (Choer. Th. 1.201.20 2) But the κοινὴ συνήθεια, which is more atticizing, writes these genitives in εω . Because in general the κοινὴ διάλεκτος follows the Attic speakers.

Choeroboscus here shows himself perfectly well aware that the Koiné uses the genitive in ‐εως. There is thus no necessity for the ‘common form’ and the Koiné form to be identical to one another.¹⁰ Another medieval commentator on Theodosius takes the same view as Choeroboscus on the status of the different forms (ὄϕιος is Ionic, ὄϕεως is Attic, and ὄϕεος is common) but offers an additional reflection on the infrequent attestation of ὄϕεος. We learn that there is no necessity for ‘common forms’ to occur frequently, or even to be attested at all:¹¹ Ἰστεόν δὲ ὅτι ἡ μὲν ‘ὄϕιος’ γενικὴ Ἰώνων ἐστίν, ἡ δὲ ‘ὄϕεως’ Ἀττικῶν ἡ δὲ σπανίως εὑρεθεῖσα ‘ὄϕεος’ κοινή, ὡς ἐν ταῖς Εὐριπίδου χρήσεσιν, ‘ἔσπειρ’ ὄϕεος ἐν γαίᾳ θέρος’, καὶ ἄλλῃ ‘ὃς ἂν δύνηται πόλεος ἔν τ’ ἀρχαῖσιν ᾖ’. Εἰ δέ τις εἴποι, πῶς τὸ κοινὸν οὐ πολιτεύεται, ϕαμὲν ὡς οὐ ξένον πολλὰ γὰρ ἔστιν, ὧν τὰ μὲν κοινὰ σεσίγηται, τὰ δὲ ἐθνικώτερα λέγεται… (Sophronius’ excerpts from Charax’ commentary on Theodosius’ Canones = Grammatici Graeci IV.ii, 389.4 10) And one must know that the genitive ὄϕιος belongs to the Ionians, and ὄϕεως to the Attic speakers. And the rarely found ὄϕεος is common, as in the examples from Euripides (ἔσπειρ’ ὄϕεος ἐν γαίᾳ θέρος, and elsewhere ὃς ἂν δύνηται πόλεος ἔν τ’ ἀρχαῖσιν ᾖ). And if somebody were to ask why that which is common isn’t customary, we’ll say that’s nothing strange. For there are many words whose common forms aren’t in use, while the forms more proper to specific dialects are in use….

The parallel with LSJ is very striking here, but continuity with ancient times is not entirely straightforward. In order to see why, we turn now to the first edition of Liddell and Scott (LS¹) and to its immediate predecessor, the fourth edition of Passow’s Handwörterbuch der griechischen Sprache (Passow 1830).

¹⁰ Cf. Consani 1991, 28 9, 38.

¹¹ Cf. Consani 1991, 36.

212

Philomen Probert

12.6. THE FIRST E DITION OF LIDDELL AND S COTT The concept of a ‘common form’, in the sense we have been considering, is clearly in evidence already in LS¹. Forms are labelled ‘common’ in at least forty-three entries (more often,¹² incidentally, than in LSJ), as summarized in Table 12.1.¹³ Readers will notice that some forms in this table are underlined; the reason for this will concern us in section 12.7. ¹² See section 12.4, with n. 5. ¹³ These entries were tracked down by searching for ‘common’ in the electronically searchable copy of the US edition of 1848 available at https://archive.org/details/greekenglishlexi00lidd, and the electronically searchable copy of the fourth UK edition (published in 1855) available at https://archive.org/details/b22651500, and then cross checking relevant instances of the word ‘common’ with LS¹. Relevant instances of the word ‘common’ were thus uncovered in LS¹ only where these survived into the US edition of 1848 and/or the fourth UK edition. LS¹ entries were taken to be relevant if they use the word ‘common’ (with or without ‘form’) as a way of singling out one form among two or more variants. Occasionally a form is called ‘common’ in the absence of any contrasting (i.e. not ‘common’) form: e.g. ‘the common syncop. forms τέθνᾰμεν, τέθνᾰτε, τεθνᾶσι’, under θνήσκω (there is no clear contrast with non syncopated τεθνήκαμεν, τεθνήκατε, τεθνήκασι); ‘both aorists are common’, under τίθημι. The word ‘common’ probably conveys frequency of attestation here, and such instances were treated as falling outside the category of interest. LS¹ also uses the word ‘common’ in a wide variety of further ways that are not relevant for present purposes: to characterize a frequently occurring collocation or construction (e.g. ‘WITH OPTAT., with which ἄν is most common’, under ἄν); a frequently occurring termination (e.g. in the entry for the letter Ι); a meaning that occurs frequently in a particular variety of Greek, or occurs frequently in absolute terms (e.g. ‘the common signif. in Att.’ under ἀναγιγνώσκω); a vowel or syllable capable of being treated as long or short (e.g. under ἀεί); a noun that can take masculine or feminine gender (e.g. under αἰθήρ); objects, activities, or language belonging to everyday life (e.g. ‘the language of common life’, under ἅλας; ‘our common wine bottles’, under ἄμβων); an everyday pronunciation (under ποιέω); a concept taken to be widespread in antiquity (e.g. ‘from the common notion of goodness in early times’, under ἐσθλός); an expression characteristic of a frequently occurring social interaction (e.g. ‘in Att. usu. absol., esp. as the common form on meeting, ἀσπάζομαί σε or ἀσπάζ. alone’, under ἀσπάζομαι); a common noun as opposed to a proper noun (e.g. under διο ); a name by which something was often known (e.g. under νεκυία); a name by which several different things are known (under Ὄλυμπος); a verb common to two clauses (under ὁ, ἡ, τό, B III); a shared element of meaning taken to link apparently disparate senses of a word (e.g. under ἀπειλέω); an element of meaning common to words thought to be etymologically related (e.g. under ἀλέγω); a root taken to link etymologically related words (e.g. under γείνομαι); a sound change taken to occur frequently (e.g. under ἀσελγής); a frequently occurring variant reading (e.g. under κλίσιον); a reading, punctuation, orthography, or scholarly opinion adopted by many modern scholars (e.g. under ἀλλήλων); and widely used editions of Greek works (under ὀκρυόεις). In some entries a form is said to be ‘common’ in a specific variety of Greek (e.g. ‘common in later Att.’, under ἀετός); here the word ‘common’ has the sense ‘frequently encountered’. In some entries a comparative ‘more common’ (e.g. under τᾰνύπους), ‘commoner’ (e.g. under γνώστης), or ‘less common’ (e.g. under λυχναψία) conveys that one form is more frequently encountered than another, either in absolute terms or under particular circumstances; the phrases ‘very common’ (e.g. under εἶεν), ‘most common’ (e.g. under κωχεύω), ‘one of the commonest Greek words’ (under μάλᾰ), and ‘in common use’ (e.g. under ἐχθρός) likewise convey frequency of occurrence. The phrase ‘common Greek’ occurs in the entries for ἀλήθω, ἀναγιγνώσκω, Ἀτλᾱγενής, θέμις, πόσε, ῥῑπίς, and ϕάτνη. Under ἀλήθω, Ἀτλᾱγενής, θέμις, and πόσε, ‘common Greek’ is probably to be understood as the core of forms available for use in a wide variety of different dialects. Under ἀναγιγνώσκω, ῥῑπίς, and ϕάτνη, on the other hand, the same phrase probably denotes the Koiné. Under ῥῑπίς, the phrase notably

Greek Dialects in the Lexicon

213

Table 12.1. Forms labelled ‘common’ in LS¹. Headword (whether the main or a subsidiary headword)

Form actually labelled ‘common’

Form(s) contrasted with the ‘common’ one

ἄεθλος ἁνδάνω ἀποχέω ἐπιχέω ἐϋκτίμενος εὐφραίνω εὐφροσύνη κεῖσε Κῶς (ἡ) μονόω ξένιος ὄνομα

ἆθλος ἥδομαι ἀποχέω ἐπιχέω εὐκτίμενος εὐφραίνω εὐφροσύνη ἐκεῖσε the acc. Κῶν μονόω ξένιος ὄνομα

ὀπίσω ὁποῖος οὐθείς Περσεφόνη

ὀπίσω ὁποῖος οὐδείς Περσεφόνη

πηός

πίεξις πλέω

πᾱός (‘Dor. πᾱός, which became the common form’) πίεσις πλέω

ἄεθλος ἁνδάνω ἀποχεύω ἐπιχεύω ἐϋκτίμενος ἐϋφραίνω ἐϋφροσύνη κεῖσε contracted forms like nom. Κόως μουνόω ξείνιος The most immediate contrast is with οὔνομα, but also ὄνῠμα. ὀπίσσω ὁπποῖος οὐθείς The most immediate contrast is with Περσεφόνεια, but also Περσέφασσα. πηός

πολλαπλάσιος προτιόσσομαι ῥέγκω ῥέγχω σπόνδῠλος¹ σύρβη σφόνδῠλος σωφροσύνη ταύσιμος τέσσᾰρες

πολλαπλάσιος προσόσσομαι ῥέγχω ῥέγχω σπόνδῠλος σύρβη σπόνδῠλος σωφροσύνη ταύσιμος and ταύσιος τέσσᾰρες

πίεξις The most immediate contrast is with πλώω, but also πλείω. πολλαπλήσιος προτιόσσομαι ῥέγκω ῥέγκω σφόνδυλος τύρβη σφόνδῠλος σᾰοφροσύνη τηΰσιος The most immediate contrast is with πίσυρες, but also τέττᾰρες, τέσσερες, τέττορες, and τέτορες. (continued )

paraphrases Pseudo Draco’s (i.e. Jacob Diassorinus’) ἡ…κοινὴ συνήθεια ‘the Koiné’ (see [Draco] (1812) at De metris poeticis 23.15 16); compare the entry for σῑρός, where the phrase ‘common language’ paraphrases Pseudo Draco’s ἡ…συνήθεια (De metris poeticis 82.1). The phrase ‘the later common dialect’ occurs in the entry for προσπέτομαι, and denotes the (later) Koiné. In the entry for the letter Υ, the phrase ‘the common written language’ perhaps denotes the core of forms available for use in a wide variety of dialects, but the intended meaning could also be the Koiné.

214

Philomen Probert

Table 12.1. Continued Headword (whether the main or a subsidiary headword)

Form actually labelled ‘common’

Form(s) contrasted with the ‘common’ one

τεῦτλον τηλία τηλοτάτω τῆτες τόσος τοσόσδε τοσοῦτος ὑπώβολος φημί φθίω φρέᾰρ φύλοπις ψάμμη

σεῦτλον σηλία πορρωτάτω σῆτες τόσος τοσόσδε τοσοῦτος ὑπόβολος εἶπον present φθίνω φρέᾰρ acc. φύλοπιν ψάμμος

ὠκύς

ὠκεῖα

τεῦτλον τηλία τηλοτάτω τῆτες τόσσος τοσσόσδε τοσσοῦτος ὑπώβολος or ὑπήβολος εἶπα pres. φθίω and imperf. ἔφθιον φρεῖαρ φυλόπιδα The essential contrast is with ψάμμη, but Doric ψάμμᾱ is also mentioned. ὠκέᾰ

¹ Where quantity marks appear in LS¹, these are reproduced here.

In all these entries except the ones for ψάμμη and perhaps ὑπώβολος,¹⁴ the variation between ‘common’ forms and others is understood as a manifestation of Greek dialectal variation, if dialectal variation is conceived broadly enough to include the categories ‘epic’ and ‘poetic’. Thus, the forms standing in contrast to the ‘common’ ones are variously labelled ‘Ep. and Ion.’ (ἄεθλος, πλώω, πλείω, ὠκέᾰ), ‘Ion. and Ep.’ (κεῖσε), ‘Ep.’ (ἐπιχεύω, ἐϋϕραίνω, ἐϋϕροσύνη, Κόως, ὀπίσσω, ὁπποῖος, προτιόσσομαι, τόσσος, τοσσόσδε, τοσσοῦτος), ‘Ion.’ (μουνόω, ξείνιος, πολλαπλήσιος, τηΰσιος), ‘more Ion.’ (εἶπα), ‘Ion. and poet.’ (οὔνομα, ϕρεῖαρ), ‘poet., and esp. Ep.’ (Περσεϕόνεια), ‘mostly Ion. and poet.’ (ἁνδάνω), ‘poet.’ (ἀποχεύω, σᾰοϕροσύνη), ‘in Ion. prose’ (τέσσερες), ‘in the new Ion. of Hipp(ocrates)’ (πίεξις), ‘Att.’ (ῥέγκω, σϕόνδυλος, τύρβη, τέττᾰρες, τεῦτλον, τηλία, τῆτες), ‘Dor.’ (τέττορες, τέτορες), ‘Aeol.’ (ὄνῠμα, πίσυρες), and ‘later’ (οὐθείς, Περσέϕασσα). In a few instances no specific label is given for a form standing in contrast to the ‘common’ one, but all the attestations cited for this form are from Homer and/or Hesiod (ἐϋκτίμενος, πηός, τηλοτάτω, the present ϕθίω and imperfect ἔϕθιον, and the acc. ϕυλόπιδα); these forms are likely to have been considered epic and/or Ionic.¹⁵ ¹⁴ Under ψάμμη, the main contrast is between the ‘common form’ ψάμμος and the ‘rarer form’ ψάμμη (as in the LSJ entry, already discussed). The form ὑπώβολος (or the alternative conjecture ὑπήβολος, for the transmitted ὑπόβολος) is cited as occurring in Pherecrates Comicus (fr. 64 K. A.); the form is perhaps considered Attic, by contrast with the ‘common form’ ὑπόβολος. If this is the thought, however, the LS¹ entry does not make it very clear, as there is no explicit mention of Attic. ¹⁵ The situation is less than clear with regard to the accusative ϕυλόπιδα, however, because all forms of ϕύλοπις are found primarily in epic (as the entry makes clear), including the ‘common’

Greek Dialects in the Lexicon

215

12 . 7 . GEWÖHNLICHE F ORMEN AND GEMEINE FO RM EN IN PASSOW 1830 If we now turn to LS¹’s immediate predecessor, the fourth edition of Passow’s Handwörterbuch (Passow 1830),¹⁶ and examine Passow’s entries under the forty-three headwords just listed, we find that twenty-four of the forms that LS¹ calls ‘common’ are called ‘gewöhnlich’ (‘usual’); these forms are underlined in Table 12.1.¹⁷ At first sight, Passow’s notion of a ‘gewöhnliche Form’ might seem to be an exact equivalent of LS¹’s ‘usual form’. In the entry for ἄεθλος, for example, the form ἆθλος is labelled ‘die gew(öhnliche) Form’, by contrast with the epic and Ionic form ἄεθλος; in the entry for εὐϕροσύνη, the form with four syllables (εὐϕροσύνη) is labelled ‘die gew(öhnliche) Form’ by contrast with the epic form with five (ἐϋϕροσύνη) (Figs 12.12 and 12.13).

Fig. 12.12. Entry for ἄεθλος ‘contest’, Passow 1830.

Fig. 12.13. Entry for εὐφροσύνη ‘mirth’, Passow 1830.

accusative ϕύλοπιν. Cf. Passow 1830’s presentation, s.v. ϕύλοπις: ‘oft bey Hom. der neben dem gew. acc. ϕύλοπιν Einmal den seltnern, ϕυλόπιδα, braucht’. ¹⁶ See LS¹, pp. iv vi. ¹⁷ The characterization of εἶπον as ‘gew(öhnlich)’ is included here, although it appears under the headword εἶπα rather than under ϕημί, as in LS¹. I have not included Passow 1830’s treatment of the form παός (under πηός), because Passow does not directly call the form itself ‘gewöhnlich’; instead he says ‘das Dor. παός ging auch in die gew. Sprache über’.

216

Philomen Probert

On closer inspection, however, Passow’s use of the label ‘gewöhnlich’ is broader than the use we find in LS¹. As well as using the word ‘gewöhnlich’ to distinguish a form from one or more dialectal variants, Passow also—and apparently somewhat more often¹⁸—uses the same label to distinguish a form from one or more variants in situations that have nothing to do with dialects. In the following entries, for example, the prefixed form ἀπεχθάνομαι and the non-prefixed form λοιβή are labelled ‘gewöhnlich’, by contrast with the rarer forms ἐχθάνομαι and ἐπιλοιβή (Figs 12.14 and 12.15).

Fig. 12.14. Entry for ἐχθάνομαι ‘incur hatred’, Passow 1830.

Fig. 12.15. Entry for ἐπιλοιβή ‘drink offering’, Passow 1830.

Neither ἀπεχθάνομαι and λοιβή nor ἐχθάνομαι and ἐπιλοιβή are associated with particular dialects: we simply have pairs of variant forms, with each pair comprising one form that we encounter more frequently and one that we encounter less frequently. Tables 12.2 and 12.3 give a fuller list,¹⁹ for the first volume of Passow 1830, of entries in which the word ‘gewöhnlich’ is used to distinguish a form from one or more variants.²⁰ For the entries summarized in Table 12.2, it is fairly clear that the variation in question is regarded as reflecting broader patterns of ¹⁸ See Tables 12.2 and 12.3 below, with n. 19. ¹⁹ This list was compiled by searching for the letters ‘gew’ in the electronic version of Passow 1830, Vol. 1, available at https://books.google.co.uk/books?id=Vmrn 5DYBiQC. Experimenta tion showed that searches of this copy do not find all instances of the sequence being searched for; our list of relevant instances of the word ‘gewöhnlich’ (or abbreviations such as ‘gew.’) is therefore unlikely to be exhaustive. ²⁰ Other uses of the word ‘gewöhnlich’ occcur, but are not relevant to our purposes. Many of these are similar to the uses of ‘common’ found in LS¹ and listed in n. 13. Entries in which a particular paradigmatic form is said to be ‘gewöhnlich’ (for example, the future middle ἀγνοήσομαι as opposed to ἀγνοήσω, under ἀγνοέω) are not taken into account. Some entries include a comment to the effect that one particular meaning is ordinarily (‘gewöhnlich’) conveyed by a different word (thus the entry for ἄγχῐ includes a subsection ‘2) von der Zeit, nächstens, bald, statt des gew. τάχα, Od. 19, 301’); these too are not considered relevant here. The entry for κυλλόπους (‘κυλλόπους, ὁ, ἡ, πους, τό, gen. ποδός, (κυλλός, πούς) krummfüssig, gew. = Κυλλοποδίων, w. m. s.’) is omitted from Table 12.2 because it is not clear whether κυλλόπους or Κυλλοποδίων is being called ‘gewöhnlich’.

Greek Dialects in the Lexicon

217

Table 12.2. Instances of the label ‘gewöhnlich’ used in Passow 1830, Vol. 1 to distin guish a form from one or more dialectal variants. Headword

Form actually labelled ‘gewöhnlich’

Form(s) contrasted with the one labelled ‘gewöhnlich’

ἀδεής ἀέθλιον ἄεθλον ἄεθλος ἀντίξοος ἀξιόχρεος ἁρπάζω ἀτλαντογενής δένδρεον εὔδμητος ἕδνον ἐπιανδάνω ἐπιπνέω ἐπιχέω (ἐπόπτω) εὔληρα εὐμελίας ἐϋμμελίης εὐρύς εὐφραίνω εὐφροσύνη εὔφρων Ζάν ἧμαι

ἀδεής ἄθλιον ἆθλον ἆθλος ἀντίξοος or ἀντίξους ἀξιόχρεως fut. ἁρπάξω ἀτλαντογενής δένδρον εὔδμητος ἕδνα ἐφανδάνω ἐπιπνέω ἐπιχέω fut. ἐπόψομαι ἡνία εὐμελίας εὐμελίας acc. εὐρύν εὐφραίνω εὐφροσύνη εὔφρων Ζεύς 3. pl. ἧνται, ἧντο

ἡμεῖς

gen. ἡμῶν, acc. ἡμᾶς

Ἰαωλκός καθαρπάζω Κάρπαθος κεῖσε κρᾰταιός κτάομαι Κῶς (ἡ)

Ἰωλκός fut. καθαρπάξω Κάρπαθος ἐκεῖσε κρατερός perf. κεκτῆσθαι the acc. Κῶν

ἀδδεής and ἀδειής ἀέθλιον ἄεθλον ἄεθλος ἄντιξος ἀξιόχρεος ἁρπάσω ἀτλαγενής δένδρεον and δένδρειον ἐΰδμητος ἔεδνα ἐπιανδάνω ἐπιπνείω ἐπιχεύω fut. ἐπιόψομαι εὔληρα ἐϋμμελίης ἐϋμμελίης εὐρέᾰ ἐϋφραίνω ἐϋφροσύνη ἐΰφρων Ζάν and Ζήν εἵᾰται, εἵᾰτο and ἕᾰται, ἕᾰτο gen. ἡμέων and ἡμείων, acc. ἡμέᾰς, ἧμᾰς, and ἄμμε Ἰαωλκός⁴ καθαρπάσω Κράπαθος κεῖσε κρᾰταιός perf. ἐκτῆσθαι contracted forms like nom. Κόως

LS¹’s label (if any) for the form that Passow labels ‘gewöhnlich’

‘the common form’ ‘the usu.’ ‘in common Greek’¹ ‘the usu.’

‘the common form’ ² (see below) ‘the prose form’³ ‘the common form’ ‘common…form’ ‘usu.’ ‘the usu. form’

⁵ ‘usu.’ ‘the usu. form’ ‘the common form’ ‘the usu.’ ‘the common acc.’

¹ This label is found in LS¹ under the lemma ἀτλᾱγενής, not ἀτλαντογενής. ² Unlike Passow 1830, LS¹ does not use the hypothetical form ἐπόπτω as a lemma, but ἐπόψομαι and ἐπιόψομαι appear as lemmata. Neither entry gives a specific label to the form ἐπόψομαι. ³ Unlike Passow 1830, LS¹ does not use the form εὐμελίας as a lemma, but the entry for ἐϋμμελίης includes the comment ‘the prose form εὐμελίας only occurs in Gramm.’. ⁴ Under the headword Ἰωλκός, however, the form Ἰωλκός is called not simply the ‘gew. Form’ but the ‘gew. Att. Form’. ⁵ LS¹ has no entry for Ἰαωλκός or Ἰωλκός.

218

Philomen Probert

Table 12.3. Instances of the label ‘gewöhnlich’ used in Passow 1830, Vol. 1 to distin guish a form from one or more variants, where the variation has nothing to do with dialects. Headword

Form actually labelled ‘gewöhnlich’

Form(s) contrasted with the one labelled ‘gewöhnlich’

LS1’s label (if any) for the form that Passow labels ‘gewöhnlich’

ἄβροτος αἱμύλιος ἀλείφω ἁλικώδης

ἄμβροτος αἱμύλος aor. pass. ἠλείφθην ἁλυκώδης

ἄβροτος αἱμύλιος aor. pass. ἐξηλίφην ἁλικώδης

‘more freq.’ ‘more usu.’ ‘usu.’ (ἁλικώδης = ‘less good form’)

ἀντίτυπος ἀστερίζω ἀτενής αὐχέω ἀφίδιτος βράγχος βράσσω

adv. ἀντιτύπως καταστερίζω adv. ἀτενές καυχάομαι φειδίτιον βράγχιον βράσσω and Attic βράττω ὑποβρύχιος δῃόω nom. pl. fem. δίδυμαι fut. (ἐκ)δρᾰμοῦμαι fut. ἐμμαχοῦμαι fut. (ἐμ)παιήσω fut. ἐνδρᾰμοῦμαι μετεξέτεροι λοιβή ἀνεσταλμένως στέρφος and τέρφος ἀπεχθάνομαι 2. sg. ᾔδεισθα τηλικοῦτος (answering πηλίκος) gen. ἡμίσεως ἱερόσυλος δοίδυξ θύννη, ἡ adv. ἰδίως ἰώγα and ἰώνγα καθάπερ βαρβαρίζω Καππάδοξ Καππάδοξ κεάζω pl. κλοιά πολύκμητος πάρνωψ

adv. ἀντίτυπα ἀστερίζω adv. ἀτενῶς αὐχέω φιδίτιον βράγχος βράζω

βρύχιος δῃόω δίδυμος ἐκτρέχω ἐμμάχομαι ἐμπαίω ἐντρέχω ἐξέτεροι ἐπιλοιβή ἐπιστολάδην ἔρφος ἐχθάνομαι ᾔδειν ἡλίκος ἥμῐσυς θεοσύλης θῠέστης θύννος, ὁ ἴδιος ἰώ, ἰών καθά καμψικίζω Καππᾰδόκης Καππάδοξ κείω κλοιός κμητός κόρνωψ

βρύχιος δηϊόω nom. pl. fem. δίδυμοι fut. (ἐκ)θρέξομαι² fut. ἐμμαχέσομαι fut. (ἐμ)παίσω fut. ἐνθρέξομαι ἐξέτεροι ἐπιλοιβή ἐπιστολάδην ἔρφος ἐχθάνομαι 2. sg. ᾔδεις τηλίκος (answering πηλίκος) gen. ἡμίσεος and ἡμίσους θεοσύλης θῠέστης θύννος, ἡ adv. ἰδίᾳ ἰώ and ἰών καθά καμψικίζω Καππᾰδόκης Καππαδόκης κείω³ pl. κλοιοί κμητός κόρνωψ or κόρνοψ

‘usu.’ ‘more freq.’ ¹ ‘usu. form’ ‘usu.’ ‘usu. contr. form’ ‘usu.’ ‘but usu.’ ‘usu.’ ‘usu.’ ‘usu.’ ‘usu.’ ‘usu.’ ‘usu.’ ‘usu.’ ‘usu.’ ‘usu.’ ‘usu.’ ‘more usu.’ ‘the more usu. fem.’ ‘the usu. Adv.’ ‘usu.’ ‘more usu.’ ‘usu.’ ‘usu.’ ‘usu. form’ ‘the usu.’ ‘the usu.’ ⁴ ‘usu.’

Greek Dialects in the Lexicon κορυμβίας κόρυμβος κοχλιάριον κύβιτον

κισσός pl. κόρυμβα λιστρίον ὠλέκρανον

κορυμβίας pl. κόρυμβοι κοχλιάριον κύβιτον

219 ‘the regul. form’ ‘usu.’ ‘more usu.’

¹ The entry for φειδίτιον gives a discussion of the form φιδίτιον, but neither form is straightforwardly labelled. ² The entry gives the future forms as δρᾰμοῦμαι and θρέξομαι, leaving the prefix ἐκ- to be understood from the lemma. Similarly for the future forms of ἐμπαίω, listed below. ³ The entry reads ‘κείω, spalten, Od. 14, 425. sonst nicht vorkommende Grundform des gew. κεάζω’. The point that κείω occurs only in a Homeric passage may suggest that dialectal variation was thought relevant to the alternation between κείω and κεάζω, but κεάζω too is primarily found in epic. Moreover, the characterization of κείω as the ‘Grundform’ for κεάζω suggests that the two forms were primarily conceived as related to one another qua derivational base and derivative, rather than qua dialectal variants. ⁴ LS¹ gives no straightforward label for the form πολύκμητος, but after noting that κμητός occurs in Hesychius the authors comment ‘elsewh. only in compds. πολύκμητος, etc.’.

Greek dialectal variation. For the entries summarized in Table 12.3, on the other hand, the variation is probably not thought to have anything to do with dialects. As the right-hand column of Table 12.2 shows, six of the ‘gewöhnlich’ forms listed in that table are called ‘common’ in LS¹ (and a seventh is said to occur ‘in common Greek’). By contrast, none of the ‘gewöhnlich’ forms listed in Table 12.3 is called ‘common’ in LS¹; the label we most often find for these forms in LS¹ is ‘usu(al)’. To a considerable extent, then, LS¹ reserves the concept of a ‘common’ form (as a way of distinguishing a form from one or more variants) for situations involving dialectal variation, whereas Passow uses the notion of a ‘gewöhnlich’ form for dialectal and non-dialectal variation alike. On occasion, Passow (1830) uses what looks like a more literal equivalent of the ancient term κοινός, namely ‘gemein’, to distinguish a form from one or more variants. Table 12.4 lists all the relevant instances of the word ‘gemein’ that I have been able to gather,²¹ this time taking both volumes of Passow 1830 into account. In three of these twenty-two entries (under the headwords ἄνοργος, βρύχω, and πάθνη) the information that a specific form is the ‘gemeine Form’ is attributed to the ancient atticist Moeris. In Moeris’ own terms, Moeris attributed the forms ἀνόργητος, βρύχω, and πάθνη to the Ἕλληνες: to Koiné

²¹ These instances were found by searching for the letters ‘gem’ in the electronic version of Passow 1830, Vol. 1 available at https://books.google.co.uk/books?id=Vmrn 5DYBiQC, and in the electronic version of Vol. 2 available at: https://books.google.co.uk/books?id= mgQBAAAAYAAJ. Experimentation showed that searches of these copies do not find all instances of the sequence being searched for; our collection of relevant instances of the word ‘gemein’ (or the abbreviation ‘gem.’) is therefore unlikely to be exhaustive.

220

Philomen Probert

Table 12.4. Instances of the label ‘gemein’ used in Passow 1830 to distinguish a form from one or more variants. Headword

Form actually labelled ‘gemein’

Form(s) contrasted with the one labelled ‘gemein’

ἀλήθω ἀνόργητος ἄνοργος βρύχω ἐϋκτίμενος ξενεῖον πάθνη πεῖν προτιόσσομαι ῥέγκω ῥέγχω σεῦτλον σῆτες σπόνδυλος σύρβη σφόνδυλος τεῦτλον τηλία τῆτες τύρβη ὑγεία φάτνη

ἀλήθω ἀνόργητος ἀνόργητος βρύχω εὐκτίμενος ξενεῖον πάθνη πεῖν προσόσσομαι ῥέγχω ῥέγχω σεῦτλον σῆτες σπόνδυλος σύρβη σπόνδυλος σεῦτλον σηλία σῆτες σύρβη ὑγεία πάθνη

ἀλέω ἄνοργος ἄνοργος βρύκω ἐϋκτίμενος ξεινήϊον φάτνη πιεῖν πίνειν προτιόσσομαι ῥέγκω ῥέγκω τεῦτλον σᾶτες/τῆτες σφόνδυλος τύρβη σφόνδυλος τεῦτλον τηλία τῆτες/σᾶτες/τῆδες τύρβη ὑγίεια φάτνη

speakers as opposed to Atticists.²² In a further nine entries (under the headwords ἀνόργητος, ῥέγκω, ῥέγχω, σεῦτλον, σῆτες, τεῦτλον, τῆτες, ὑγεία, and ϕάτνη), we may suspect that Passow’s use of the word ‘gemein’ also derives ultimately from Moeris’ attribution of the form in question to the Ἕλληνες.²³ In these entries the choice of the word ‘gemein’ appears inspired by the ancient term κοινῶς (whether this is taken to mean ‘in general’ or ‘in the Koiné’), even though Moeris himself uses the term Ἕλληνες in all the relevant entries. For another entry (for πεῖν), the form labelled ‘gemein’ is clearly postclassical.²⁴ Here the word ‘gemein’ appears to be inspired by the term κοινή (‘Koiné’) rather than by κοινός ‘common’. ²² Moeris, α 21, β 24, and ϕ 19 Hansen. ²³ For Moeris’ attribution of the forms ἀνόργητος, ῥέγχω, σεῦτλον, σῆτες, ὑγεῖα, and πάθνη to the Ἕλληνες, see Moeris, α 21, ρ 7, σ 1, τ 2, υ 11, and ϕ 19 Hansen. In the entries in question there is no explicit mention of Moeris’ view, although Moeris is mentioned obliquely under the headwords τῆτες and ὑγεία, and the forms labelled ‘gemein’ under ἀνόργητος and ϕάτνη are (as we have just seen) also mentioned in other entries (for ἄνοργος and πάθνη) in which Moeris is explicitly invoked. ²⁴ See LSJ, s.v. πίνω.

Greek Dialects in the Lexicon

221

Fig. 12.16. End of the entry for ἐϋκτίμενος ‘well founded’, Passow 1830.

Fig. 12.17. End of the entry for προτιόσσομαι ‘look at’, Passow 1830.

For a further three entries (for ἐϋκτίμενος, ξενεῖον, and προτιόσσομαι) the form labelled ‘gemein’ is very rare or (under προτιόσσομαι) actually unattested, and Passow’s entry mentions or alludes to this fact.²⁵ Thus under ἐϋκτίμενος Passow mentions one occurrence of the quadrisyllabic form εὐκτίμενος (the only one known in his day)²⁶ while the entry for προτιόσσομαι makes it clear that the form προσόσσομαι is never attested at all (Figs 12.16 and 12.17). Under the headwords σύρβη, τηλία, and τύρβη, the forms labelled ‘gemein’ (σύρβη and σηλία) are attested only in grammatical and lexicographical works, and were plausibly considered rare too. In entries such as those we have just seen, Passow’s use of the word gemein comes close to LS¹’s use of the word ‘common’, but Passow makes much more limited use of gemein in this sense than LS¹ of ‘common’. In at least some instances Passow’s gemein appears to be motivated by the rarity or lack of attestation of the form in question (a situation in which one might reasonably hesitate to call the form ‘usual’), but in general Passow prefers the broader term gewöhnlich. LS¹ thus extended the notion of a ‘common form’ beyond the limits it had seen in Passow, and strengthened the sense of a consistent distinction between

²⁵ Cf. LSJ, s.vv. ἐϋκτίμενος, ξεινήϊον, and προτιόσσομαι. (According to LSJ, ξεινήϊον is ‘only found in Ion. and Ep. form’. A TLG search for word forms beginning with ξενει turns up a Byzantine instance of the noun ξενεῖον in Eustathius, Od. 1416.20 = 1.57.9 Stallbaum.) ²⁶ Today two further attestations can be cited from Bacchylides (Epinicia 5.149, 11.122).

222

Philomen Probert

dialectal variation on the one hand and variation tout court on the other. In so doing the authors are likely to have been influenced by the didactic tradition of Victorian England, where the notion of ‘common forms’ shared between various varieties of Greek is reflected in the so-called Eton Greek Grammar, in the various versions in which this influential work circulated in the nineteenth century.²⁷ The following introduction to the principal dialects of Greek appears, for example, in the edition of 1819, and in H.J. Tayler’s English edition of 1843: QUATUOR numerantur Græcæ linguæ dialecti, sive loquendi formæ præcipuæ, pro numero præcipuarum Græciæ gentium, quæ à linguâ communi in nonnullis deflectuntur, nimirum Attica, Ionica, Dorica, Æolica ; præter quas, suam dialec tum, et idioma, in quibusdam habuerunt poëtae: quæ omnes, quod ad termina tiones partium orationis spectat, in sequentibus tabellis subjiciuntur: In reliquis verò, cùm tanta sit varietas, ut regulis comprehendi vix possit, singularium proprietates insigniores breviter proponere sufficiat. I. Attici mutant σ in ξ ut, ξὺν pro σὺν σσ in ττ ut, θαλάττα pro θαλάσσα [sic] σ in ρ ut, ἄῤῥην pro ἄρσην, vir: (Eton Greek Grammar (1819, 210)) The Greek language has four dialects, or principal ways of speaking, correspond ing to the number of the principal nations of Greece, which vary in some things from the common language, namely, the Attic, Ionic, Doric, and Æolic: besides which, the Poets had a dialect and idiom peculiar to themselves: all which, as far as relates to the terminations of the parts of speech, will be found subjoined in the tables which follow. The variations, however, in the other portions of words being so numerous, that they can scarcely be all comprised in any set of rules, it may be sufficient briefly to explain the more remarkable peculiarities of each. I. The Attics change σ into ξ as, ξὺν for σύν σσ into ττ as, θαλάττα for θαλάσσα [sic] σ into ρ as, ἄῤῥην for ἄρσην, a male: (Eton Greek Grammar, tr. H.J. Tayler (1843, 198))

There follow tables of dialect forms of the kind shown in Fig. 12.18, in which the forms that are considered shared or non-dialectal are labelled com(munes) in the Latin edition of 1819 (shown here), and com(mon) in the English edition of 1843.

²⁷ Compare Clarke 1945, 16.

Greek Dialects in the Lexicon

223

Fig. 12.18. Eton Greek Grammar (1819, 218).

The Eton Greek Grammar originated as a grammar written for Westminster School by William Camden (1595),²⁸ itself an adaptation of a grammar by Edward Grant (1575) in question-and-answer format.²⁹ The introduction to the principal dialects that appears in nineteenth-century editions of the Eton Greek Grammar is already present in the first edition of Camden’s work, in a form very similar to those we have already seen: Quatuor numerantur Græcæ linguæ Dialecti siue loquendi formæ præcipuæ pro numero præcipuarum Græciæ gentium, quæ à linguâ communi in nonnullis variant, nimirum Attica, Ionica, Dorica, Æolica. Præter quas suam Dialectum & idioma in quibusdam habuerunt Poetæ, quæ omnes quod ad terminationes partium orationis spectat, sequentibus tabellis subijciuntur. In reliquis vero cùm tanta sit varietas, vt regulis comprehendi vix possit: singularum proprietates insigniores breuiter proponere sufficiat. (Camden 1595: 66)

There follows a list of principal deviations of specific dialects, and tables similar to those we have already seen, with the label ‘com.’ for the forms considered shared or non-dialectal. The link with the ancient and medieval concept can be seen more clearly, however, in Grant’s work. Here Greek is said to have five main dialects, Attic, Ionic, Aeolic, Doric, and Common. The ‘Common’ dialect is defined as comprising words shared between dialects and used equally by everyone: Δ. Sed antequam ad nomina inflectenda maturemus, ne fortasse dialectorum insolentia pueri perplexi hæreant, operæ pretium erit, obiter annotare Grecos (quorum monumenta per manus quodammodo ad nos sunt transmissa) non omnes vna voce, vna linguæ proprietate, & scribendi genere usos esse: nam

²⁸ See e.g. Watson 1908, 495 6, 497; and cf. Stray 2016. ²⁹ See e.g. Littlefield 1904, 268; Watson 1908, 497.

224

Philomen Probert

diuersarum regionum, vt diuersi fuerunt mores & ritus, ita apud eas, dissimiles fuerunt loquendi & scribendi formae. Quot sunt igitur praecipuæ linguæ, dialecti, seu loquendi idiômata, quæ in Græcorum scriptis paßim notantur?

M. Quinq’ præcipuę:

ἀττίς

Attica

omnium comptissima, & optima, dicta ab Atthide Cranai filia, qva usi sunt attici: vt Demost.Isocrat. zenoph. Plato, Aristophanes.

ἰωνίς

Ionica

ab Ione, filio Apollinis appellata, qva usi sunt Herodot. Hippocrates, & maxime Homerus.

ἀιολίς

Æocila [sic]

Ab Aeolo Hellenis filio nominata, qua vsi sunt Alcæus, Sappho Lesbia.

δωρίς

Dorica

A Doro Deucalionis & Pyrrhæ filio dicta, qva usi sunt Siculi, & Rhodii: vt Theocritus & Pindarus.

κοινὴ

Cõmunis

qua vsi sunt omnes promiscue, & communiter, vti nomen significat, nulli Græcorum genti peculiaris, vt quae cõmunia comprehendat vocabula, & ab omnibus pariter vsurpata. (Grant 1575, f. 28r 28v)

TEACHER: But before we hasten to the inflections of nouns, lest boys perhaps get stuck in confusion through lack of experience of dialects, it will be worthwhile to note in passing that the Greeks (whose records have in some way been handed down to us) did not all use one speech, one quality of language, and one way of writing: for just as the customs and habits of different regions were different, so the forms of speech and writing were dissimilar among them. What then are the main languages, dialects, or peculiarities of speech that are widely noted in the writings of the Greeks?

STUDENT: 5 main ones:

ἀττίς

Attic

ἰωνίς

Ionic

ἀιολίς

Aeolic

δωρίς

Doric

κοινὴ

Common

The most comely of all, and the best, named after Atthis, daughter of Cranaus; used by the Athenians like Demosthenes, Isocrates, Xenophon, Plato, Aristophanes. Called after Ion, the son of Apollo; used by Herodotus, Hippocrates, and mostly Homer. Named after Aeolus the son of Hellen; used by Alcaeus and Sappho of Lesbos. Named after Dorus, son of Deucalion and Pyrrha; used by the Sicilians and Rhodians like Theocritus and Pindar. Used by all indiscriminately and in common, as the name implies; particular to no people among the Greeks, since it comprises the words that are common and used equally by everyone.

Greek Dialects in the Lexicon

225

The idea that Greek has five dialects, with the Koiné as the fifth, is attested already in Clement of Alexandria around 200 AD,³⁰ and enjoyed a very long history. The definition of the Koiné was contested and transformed in various ways over the centuries,³¹ but the notion that the Koiné can be equated with the stock of ‘common forms’ has its roots in the Byzantine period.³² The history of thought on Greek dialects cannot be pursued further here, and has been admirably taken up to the modern period by Consani (1991), but LSJ has led us to glimpse some ramifications of this history in England.

³⁰ Clemens Alexandrinus, Stromata I.142.3 4; see Consani 1991, 21 2. ³¹ See Consani 1991, 50 3, 60 7, 75 94. ³² See Consani 1991, 60 7.

13 Between Cunning and Chaos μῆτις Evelien Bracke

Composing a lexicon is an undertaking fraught with complications, not only linguistic but also contextual. Political agendas, socio-economic background, education, health, and personal relationships all play a role in any translator’s interpretation of another language.¹ In the composition of this Lexicon, the friendship between Liddell and Scott has often been romanticized, and its merits and imperfections have been humorously encapsulated in the well-known rhyme already mentioned by Stray, Chapter 1, this volume (p. 19). Although hyperbolic, the rhyme highlights the core issue: while in some parts solid research underpins translations, other parts appear chaotic in interpretation and arrangement. One the one hand we see the editors’ explicit lexicographical strategy and incorporation of scholarly advances in the nine editions; on the other, the progressive slipping-in of seemingly arbitrary alterations. While invariably paying respect to the herculean labour that went into the Lexicon’s creation, reviewers of its different editions were quick to point out their many shortcomings.² Although Liddell and Scott were acutely aware of these,³ they never succeeded in reversing this trend, and upon publication of the first fascicle of the ninth edition, Jones (1941: 1) acknowledged the many ‘sins of omission and commission’ still apparent.

¹ For a summary of complications in the composition of the Liddell and Scott Lexicon, see Stray 2010a, 94 118. ² See, e.g., Keene 1893, 329; Leeper 1893; Wayte 1894; Gildersleeve 1898; Knapp 1921, 136; Harrison 1935, 226; Jay 1940, 86; Harrison 1941; Whatmough 1942; Dunbabin 1946, 8; Fletcher 1947, 3. Note that none of the reviewers mention the entry for μῆτις. ³ E.g. LS¹, x. See also LS⁴, iii: ‘we have not grown blind to [the lexicon’s] imperfections’. Evelien Bracke, Between Cunning and Chaos: μῆτις In: Liddell and Scott: The History, Methodology, and Languages of the World’s Leading Lexicon of Ancient Greek. Edited by: Christopher Stray, Michael Clarke, and Joshua T. Katz, Oxford University Press (2019). © Oxford University Press. DOI: 10.1093/oso/9780198810803.003.0013

Between Cunning and Chaos: μῆτις

227

The aim of this chapter is to explore this tension running through the nine editions by focusing on the development of one particular entry, that for μῆτις.⁴ This term is best known from the still-authoritative 1974 book by Marcel Detienne and Jean-Pierre Vernant, Les ruses de l’intelligence: la métis des Grecs. The ambiguity inherent in ruses—also present in the English title ‘cunning intelligence’—stands in contrast to LSJ’s broader definition as 1) ‘wisdom, skill, craft’, and 2) ‘counsel, plan, undertaking’. This chapter therefore sets out to explore both the intratextual tensions within the various editions of Liddell and Scott as well as the extratextual tensions between the Lexicon and contemporary scholarship. By comparing the entry for μῆτις with those of its derivatives, such as αἰολόμητις and ἀγκυλομήτης, as well as of related concepts such as σοϕία and δόλος, we will aim to shed light on the skilful yet chaotically evolving entity that is the Lexicon.

13.1. OF GREEK, GERMA N, AND E NGLISH: L IDDELL AND SCOTT AND THEIR PREDECESSORS As Stray (Chapter 1, this volume) has already mentioned, Liddell and Scott was not the first Greek-English lexicon to be published: the same venture had been undertaken successively by Jones (1825), Pickering (1826), Donnegan (1826), Dunbar and Barker (1831), and Giles (1840).⁵ Donnegan provides a key point of reference because his lexicon was Liddell and Scott’s strongest competitor when the first edition was published.⁶ Donnegan compiled A New Greek and English Lexicon based on J.G. Schneider’s Kritisches griechisch-deutsches Handwörterbuch (1797–8). Donnegan’s entry for μῆτις reads as follows: Μῆτις, ιος, Att. ιδος, ἡ, s./s. as μῆδος, wisdom; penetration; skilfulness a counsel; an expedient; a contrivance in general, discernment; prudence; reflection. Th. μάω, Lennep., μήδω, obs. Schn. L.

Compare this with the same entry in Schneider: Μῆτις, ιδος, ἡ, (μήδω, f. μήσω, perf. μέμηται) f. v. a. μῆδος, Klugheit, Einsicht, Geschicklichkeit: Rath, Hülfsmittel

⁴ Zgusta 2006, 35 7; Imholtz 2007; and Lee 2010a also explore specific entries throughout the editions and come to similar conclusions. There are many studies of specific entries, e.g. Clark 1944 and Prato 1988. ⁵ See Imholtz 2007, 118 23. ⁶ See Stray, Chapter 1, this volume, and Stray 2010a, 103.

228

Evelien Bracke

Donnegan did not adhere slavishly to Schneider, adding the non-Attic genitive and more information about cognate verbs.⁷ However, he did complicate Schneider’s translation. First, while Schneider maintained a binary distinction between aptitude (the first three terms) and application (the final two terms), Donnegan created a tripartite definition of aptitude, application, and a third ‘general’ characteristic uniting them. It is difficult to discern any significant distinction between the first and third definitions, since all three ‘general’ translations are rather vague descriptions. Secondly, for his first translations, Donnegan evidently relied wholly on Schneider: wisdom, penetration, and skilfulness are translations of the German Klugheit, Einsicht, and Geschicklichkeit, although penetration—while it entails an aspect of ‘insight’—is an impractical overtranslation of Einsicht. This entry highlights the key issue surrounding the birth of Greek-English lexica: the use of German translations as an intermediary level of interpretation. The same issue cast a shadow over LS¹, since it of course relied heavily on German scholarship in the form of Franz Passow’s 1819–24 Handwörterbuch der griechischen Sprache. Passow had first edited but then departed from Schneider’s lexicon in his own and laid the groundwork for nineteenthcentury lexicography in his 1812 essay.⁸ His departure from Schneider is visible in the LS¹ entry for μῆτις: ΜΗΤΙΣ, ιος, ἡ, Att. gen. ιδος, Ep. dat. μήτι for μήτιι, Hom., acc. μῆτιν, the faculty of advising, wisdom, skill, cunning, craft, freq. in Hom.; opp. to βία, Il. 23.315. II. advice, counsel, a plan, undertaking, Hom.; esp. μῆτιν ὑϕαίνειν, cf. μῆδος. III. as fem. prop. n., the first wife of Zeus, mother of Athena, Hes. Th. 886. Ep. word, used also by Pind. and Aesch. (Cf. Sanscr. Mati, thought, counsel, from man to think, cf. Germ. Muth, with Lat. mens.)

In comparison with Donnegan’s translations, the Lexicon’s translations are a noticeable improvement. If one juxtaposes the two entries for the same term, it becomes clear why LS¹ rapidly supplanted Donnegan’s lexicon among Englishspeaking scholars. The formatting is clearer: rather than a confusing tripartite definition, a clearer definition of ability and application is present, this being retained from Schneider; and reference is made to specific authors. While Liddell and Scott relied heavily on Passow,⁹ they also departed from him: they added not only a reference to the mythological figure of Athena’s mother,¹⁰ but also an

⁷ Schneider’s derivation of μῆτις from μήδω stems back to Stephanus ad μῆτις. Donnegan’s addition of μάω presumably relates to the contemporary connection of μῆτις and the Latin mens, see also LS¹ and Curtius. ⁸ See Stray, Chapter 1, this volume, pp. 8 9, Zgusta 2006, 27 31 and Considine 2010, ix x for Passow’s importance in the creation of the Lexicon. ⁹ See also Liddell’s correspondence: Thompson 1899, 66 7. ¹⁰ In their reference to texts other than the Homeric epics and Hesiod, they also departed from Passow. See LS¹, vi vii.

Between Cunning and Chaos: μῆτις

229

etymological discussion which reflected contemporary (albeit later refuted) scholarship.¹¹ Many of the improvements, however, reflect developments in Germanic more than in English scholarship, and herein lies Liddell and Scott’s greatest weakness: its original English translations do not always engage directly with the Greek, but are often translations via German. In Passow, the translations of μῆτις are as follows: I die Fähigkeit zu rathen, Verstand, Klugheit, Einsicht, Geschicklichkeit, Anstelligkeit; II. Rath, Rathschluss, Rathschlag, Anschlag. It thus appears that every German translation has been rendered into an English equivalent; the faculty of advising is a particularly recognizable rendition of die Fähigkeit zu rathen. However, while Liddell and Scott’s translations from German are an improvement on Donnegan’s, wisdom is no accurate translation of Verstand, nor skill of Klugheit, and the same issue applies to the other translations. Each of these German terms is polyvalent, and thus their precise translation into English is as problematic as that from Greek into English. Moreover, the English translations are equally ambiguous, particularly craft and cunning. In Victorian times, craft was not only associated with handicraft through the Arts and Crafts movement which began around 1880, but also bore a connotation of generic intellectual ‘skill’—related to the Greek τέχνη— and indeed also of ‘wile’.¹² These notions of handicraft and wiliness were not, however, diametrically opposed: Walker Gore (2015) argues that, in nineteenth-century novels, female heroines who knitted, embroidered, or practised other crafts considered to be conventionally feminine, were represented as wilier and more subversive than women who did not, since their craft encapsulated ‘their ability to weave lies, entangle men in their meshes, or plot against them’. The translation cunning is equally ambiguous. Used as a synonym for ‘craft’ and ‘wile’, it was generally considered a more ‘sinister’ wisdom with an ‘implication of deceit and underhandedness’,¹³ in Victorian times particularly associated with criminals and policemen.¹⁴ Additionally, the term ‘cunning’ was linked to ‘cunning folk’, practitioners of magic in Britain up to the middle of the twentieth century.¹⁵ It is thus clear that neither of these

¹¹ As clarified at LS¹, ix, the lexicon’s notes on comparative philology were based on Pott’s 1833 Etymologische Forschungen. Wharton’s 1890 Etyma Graeca, which relied heavily on Curtius’ 1858 Grundzüge der griechischen Etymologie, although published after LS¹, argued that μῆτις derived from μένος (‘mind’) via compensatory lengthening. Liddell and Scott’s explanation was thus in line with contemporary scholarship. See further Katz, Chapter 5, this volume. ¹² Kroesch 1929, 437 for the meaning of ‘skill’; ibid., 436 for ‘wile’. ¹³ Chamberlain 1903, 299 as synonym of ‘craft’, ibid., 296 for its sinister implications. ¹⁴ For Jack the Ripper, see, e.g., the New York Times article ‘Whitechapel startled by a fourth murder’ on 9 September 1888: ‘The assassin . . . is as cunning as he is daring . . . Both the character of the deed and the cool cunning alike exhibit the qualities of a monomaniac.’ For policemen, see Cox 2012, 228. ¹⁵ De Blécourt 1994; Moran 2004, 123.

230

Evelien Bracke

English words was unequivocal at the time of the publication of LS¹.¹⁶ The use of both terms might be explained by Liddell and Scott’s political agenda favouring Anglo-Saxon translations over those with Romance roots (though note advice and counsel in the translation of μῆτις).¹⁷ For modern scholars, when considering application of these translations to Greek texts, the question arises to what extent one is to foreground the ‘wily’ rather than (or as well as) the ‘skilful’ theme. Moreover, for the modern student the ‘wily’ connotation of ‘craft’, although still mentioned (last) in its Oxford English Dictionary entry, is no longer in common use—the translation ‘craftiness’ rather than ‘craft’ might be more practical today. In short, editors’ default culture and world views could overshadow their scholarly practice, and they often sought to bypass the arduous road of lexical assembly by relying heavily on German authorities. This in itself was nothing new, as no lexicographer since Antiquity has started ex nihilo.¹⁸ Nevertheless, although in the fourth edition Passow’s name was deleted from the front page of the Lexicon, since by that stage the influence of other sources had significantly increased,¹⁹ it was never quite able to shed the shackles of its Germanic origins. As we will see, problematic and outdated translations continued to be used throughout the nine editions.

1 3 . 2. Μ ΗΤ Ι Σ FROM LS¹ TO LSJ: BETWEEN SCHOLARSHIP AND IDIO SYN CRA SY Between the first and ninth editions, scholarship advanced dramatically. When compared with LS¹, the entry for μῆτις in LSJ bears testimony to the alterations and additions made in response to these advances: μῆτις, ἡ, gen. ιος Pi.N.3.9; acc. pl. μήτιας h.Ven.249; also gen. ιδος A.Supp.61 (lyr.); acc. pl. ιδας Id.Ch.626 (lyr.); dat. μήτιδι Orac. ap. Hdt.7.141; Ep. μήτι for μήτιι, Hom. (v. infr.); pl. μητίεσσι Pi.O.1.9; acc. μῆτιν Il.2.407, S.Ant.158 (lyr.): wisdom, skill, craft, Διὶ μῆτιν ἀτάλαντος (cf. μητίετα) Il. l.c., al.; βροτείη μ. Emp.2.9; τὰν Διὸς γὰρ οὐχ ὁρῶ μῆτιν, ὅπᾳ ϕύγοιμ’ ἄν A.Pr.906 (lyr.); μήτι.. καὶ κέρδεσιν Od.13.299; μήτι.. μέγ’ ἀμείνων ἠὲ βίηϕι Il.23.315; μῆτιν ἀλώπηξ a fox for

¹⁶ Their definitions are still similar: examination of their OED definitions reveals only minor differences. Cunning is defined as ‘now usually in a bad sense: skill employed in a secret or underhand manner, or for purposes of deceit; skilful deceit, craft, artifice’ as well as ‘personal quality: disposition to use one’s skill in an underhand way; skilfulness in deceiving, craftiness, artfulness’. Craft, on the other hand, is described as ‘skill or art applied to deceive or overreach; deceit, guile, fraud, cunning’. ¹⁷ See LS¹, ix, and Williamson, Chapter 2, this volume. ¹⁸ See Dickey 2010, 5 24. ¹⁹ See LS⁴, iii.

Between Cunning and Chaos: μῆτις

231

craft, Pi.I.4(3).47; of a poet’s craft, Id.N.l.c. II. counsel, plan, undertaking, ὑϕαίνειν μῆτιν Il.7.324, cf. Od.4.678, etc.: pl., σοϕῶν μητίεσσι Pi.O.l.c.; γυναικοβούλους μήτιδας A.Ch.l.c. Poet. word. (Cf. Skt. Mimāti, pf. part. Pass. Mitá ‘measure’, Lat. metior, OE. mæþ ‘measure’.)

Grammatical details have been added in a contextualized manner, quotations have replaced mere references, and focus has moved away at least partly from the Lexicon’s original centring on Homer and Athenian authors. A dual definition has been maintained, and the etymological discussion at the end is updated to reflect contemporary scholarship. However, it will turn out that although some changes were innovations in LSJ itself, in which the editors consciously departed from LS⁸,²⁰ many of them had in fact been introduced gradually in the intervening editions over the years.

13.3. CHANGES BETWEEN TH E F IRST AND THE EIGHTH EDITION The second and third editions of the Lexicon (1845 and 1849) appeared shortly after the first, and only minor changes were made. This is reflected in the entry for μῆτις, in which only one meaningful change was made (in LS³). Running against the political agenda stated in the preface to the first edition, where the authors had made the conscious decision to write a Greek-English rather than the traditional Greek-Latin lexicon,²¹ the etymological explanation was suddenly rendered from English back into Latin: this was only deleted again in the fifth edition. While Latin is often referred to in other entries as comparison with the Greek (Zgusta 2006, 35–6) or as ‘masking device’ due to the editors’ ‘lexicographical prudery’ (Coker, Chapter 4, this volume), I have only found one other example of the change from English into Latin in entries with etymological discussions, namely in the entry for *μάω: this suggests that one editor took an idiosyncratic decision to re-impose Latin on a group of entries.²² The input of an idiosyncratic editor might also explain the rather random addition of a reference to the specific craft of a fox and a poet. While both fox and poet were typical possessors in Greek literature of a specific cleverness, it is puzzling that other

²⁰ See LS⁹ vi. ²¹ LS¹, iii. The etymology in LS³ reads: Cf. Sanscr. mati consilium; from man cogitare. ²² Other entries, such as those for ἀπαϕίσκω and ἐμϕανής, include Latin among their translations throughout the various editions (palpare and res notae/in propatulo respectively), which emphasizes the arbitrariness of the editorial process.

232

Evelien Bracke

professions such as seafarers and politicians, as well as other animals such as the octopus,²³ were omitted. This example reveals that the editorial quest for progress and clarity is inevitably shaped by personal preference. In LS⁵, which announced major changes to entries because of new information gained from the Paris papyrus, the fifth edition of Passow’s GreekGerman lexicon edited by Rost and Palm, and the integration of Curtius’ Grundzüge der griechischen Etymologie,²⁴ only one alteration was made to the μῆτις entry: one reference was changed from the nominative to the dative case to reflect the actual text: LS⁵ read μήτι opp. to βίηϕι, which reflects its use in Il. 23.315. Again contrary to major revisions throughout,²⁵ LS⁶ changed little in the μῆτις entry. First, a reference to Od. 4.678 was added to the existing reference to Il. 7.324, perhaps to clarify the context of the phrase μῆτιν ὑϕαίνειν, ‘to weave metis’. This is understandable: the Iliadic reference had been included since it was the first instance in which this phrase was used,²⁶ and referred to Nestor advising the Greeks to ask the Trojans for a day to bury their dead. Nestor is described as οὗ καὶ πρόσθεν ἀρίστη ϕαίνετο βουλή, ‘whose counsel before too appeared best’ (Il. 7.325), which emphasizes that his weaving of plans must be interpreted as a constructive act. The Odyssean reference to the suitors weaving a plot to dispose of Telemachus upon his return from Sparta, by contrast, is represented in a negative light by Medon, the messenger, who describes their plan as πολὺ μεῖζόν τε καὶ ἀργαλεώτερον, ‘far greater and more distressing’ than Penelope’s fear that the suitors are wasting Telemachus’ wealth. The addition of the Odyssean reference thus adds to the interpretation of μῆτις as a complex term, whose connotation shifts with the context. It is difficult to establish, however, which LS translation of μῆτις would be appropriate: a connotation of deception is implied in the metaphorical application of the back-and-forth motion of weaving, but in both cases, the μῆτις in question is a practical plan rather than a generic skill. Since the options offered by LS are either ‘cunning’/’craft’ or ‘plan’ (without the ‘wily’ connotation), the binary definition of μῆτις cannot be maintained. There are many examples of emphatically ‘wily plans’: Nestor, for example, defines as μῆτις his advice to his son Antilochus that he should cheat in the chariot races during Patroclus’ funeral games (Il. 23.313–18). A second alteration in the LS⁶ entry was the deletion of the notes on comparative philology at the end. The editors might plausibly have acknowledged that their proposed etymology was incorrect yet

²³ For politics, see Detienne and Vernant 1978, 307 16; for navigation, ibid., 215 48; for the octopus, ibid., 27 54. ²⁴ LS⁵, iii. ²⁵ LS⁷, iii. ²⁶ See LS¹, vii: according to the Passowian model, each entry was designed to provide a history of the word usage and mention the earliest authority of each use; Zgusta 2006, 37.

Between Cunning and Chaos: μῆτις

233

lacked the time to revise it or disagreed on the details, and hence decided to delete the notes altogether. The preface to the seventh edition announced a thorough revision with fuller references to classical authors.²⁷ In the μῆτις entry, however, only two changes were made: the full quotation of Διὸς γὰρ οὐχ ὁρῶ μῆτιν, ὅπα ϕύγοιμ’ ἄν, ‘for I do not see the μῆτις of Zeus, such as I might flee’ was given rather than merely τὰν Διὸς μῆτιν. Furthermore, μήτι, opp. to βίηϕι was extended to a fuller quote μήτι.. μέγ’ ἀμείνων ἠὲ βίηϕιν. That this detail regarding the contrast between cleverness and violence was adapted three times, while other references were left untouched throughout the editions, again points towards idiosyncratic editing. In short, it appears that, for each new edition, a general agenda was set out by the editors regarding which changes that were to be made; nonetheless, this agenda was not applied uniformly to all entries. It is difficult to ascertain the reasons behind these inconsistencies, but it is likely a combination of factors, including lack of time, oversight, and conscious decisions by individuals to edit as they saw fit.

13.4. EXPLORING THE CHANGES I N T H E NI N T H E D I T I O N ( L SJ ) With Liddell’s passing in 1898, new editors were brought on board to prepare the ninth edition. Henry Stuart Jones and Roderick McKenzie started with the revision of the first fascicle in 1920, and the ninth edition (LSJ) was finished in 1940. Drastic revisions were made to reduce the number of separate headwords, and existing material was deleted to make room for new words and evidence on areas such as medicine and botany.²⁸ This is reflected in the μῆτις entry: new information about cases and their usage was added, and quotations were written in full.²⁹ The reference to Μῆτις as personal name was deleted, perhaps when it was detected that this had been omitted in the editorial process of LS⁸.³⁰ Moreover, new etymological evidence was included: whereas μῆτις had initially been connected to the notion of ‘mind’ (Latin mens, LS¹), new comparative philological findings, published as early as Walter Prellwitz’ 1905 Etymologisches Wörterbuch der griechischen Sprache, made a connection with the Sanskrit māti-ṣ, ‘measure’, which became incorporated in LS⁹.³¹ As ²⁷ LS⁷, iii. ²⁸ See LS⁹, vii xi for a full discussion. ²⁹ See p. 230 above for the full entry. ³⁰ See Mackenzie Chapter 7, this volume, pp. 118 22 on the general inconsistency in distin guishing divine personifications from non personified usages in the LS. ³¹ Henceforth this connection became the consensus omnium, e.g. Boisacq 1916, Frisk 1960, Chantraine 1968, and Beekes 2010. Curtius 1858, 277 did in fact already question his own

234

Evelien Bracke

the preface to LS⁹ announced a reduction of etymology to a minimum, again the entry for μῆτις can be seen to defy the general approach. With regard to the actual translation, the only change made was the removal of the translations the faculty of advising and cunning. The former was consistent with the Lexicon’s implicit agenda, as it is descriptive rather than a translation and a move away from description can be detected particularly between the sixth and ninth edition.³² A rationale for the deletion of cunning is more difficult to trace. As discussed above, μῆτις can be applied to constructive and destructive cleverness and planning. The translations remaining in LS⁹, however—wisdom, skill, and counsel, plan, undertaking—are predominantly positive or neutral. The only translations which provided insight into a destructive application of μῆτις were craft and cunning. It is conceivable that the translation cunning was deleted since—as explored above—there was a degree of connotational overlap between cunning and craft, and the editors might have wished to avoid duplication. In P.M. Roget’s Thesaurus of English Words and Phrases (1805 edition), however, both the fox (in the phrase ‘as cunning as a fox’) and Odysseus are mentioned in the entry for cunning, emphasizing the nineteenth-century connection of this particular English term with the Greek notion of μῆτις which makes its deletion—rather than that of craft—more peculiar. Moreover, an exploration of the entries for similar concepts and cognates of μῆτις reveals that between LS¹ and LSJ they did indeed maintain the ‘cunning’ element of their definition. This ran through in the entries for τέχνη and σοϕία, which both share with μῆτις the translations skill and cunning, as well as in the entry for δόλος, which shares the craft and cunning translations. The noun μῆδος, referred to in the μῆτις entry until LS⁸, also retained the cunning translation in LS⁹.³³ In compound adjectives derived from μῆτις, however, such as ἀγκυλομήτης (crooked of counsel), αἰολόμητις (full of various wiles), δολόμητις (crafty-minded), and ποικιλόμητις (full of various wiles), the translation cunning was never used from the first edition onward. These epithets were all translated instead as wily and crafty, and their translations were not altered between LS¹ and LSJ.³⁴ Wily is comparable to cunning, and wiliness would have been an appropriate translation for μῆτις.

etymology of μῆτις and hinted at what was later proved to be the correct etymology however, this was ignored by Liddell and Scott. ³² See MackenzieChapter 7, this volume, pp. 118 22. Zgusta 2006, 37 regrets this ‘disintegra tion of the Passowian model’, since Passow’s approach to linking the various translations was abandoned and no links between translations were made explicit. ³³ Until it was deleted in LS⁷, indeed, the entry for μῆδος also referred to the μῆτις entry for the definition of ‘care, anxiety’, a definition completely absent from μῆτις and possibly based on the incorrect comparative etymology connecting μῆτις with ‘mind’. ³⁴ There is one exception: the adjective πολύμητις is translated as of many counsels, which does not incorporate any connotation of wiliness like the other derivative adjectives: there is no logic to its neutral translation which focuses on the second (practical) rather than the first

Between Cunning and Chaos: μῆτις

235

Examination of terms with meanings similar to μῆτις as well as of its derivatives thus reveals firstly that its adjectives were translated in a more negative sense than μῆτις itself;³⁵ secondly, that simple μῆτις was the only entry from which the translation cunning was deleted. When we consider this in the context of the nine editions, it exemplifies the often mystifying adjustments that were made. The recurring changes made to the opposition of μῆτις and βίη in the entry similarly point towards idiosyncratic editorial work. General trends announced in the prefaces of the respective editions, by contrast, were often ignored in the entry for μῆτις or acted on only in a later edition. There is a discrepancy between the inclusion of new scholarly findings on the one hand, and idiosyncratic changes (and failures to change) on the other.

13.5. DESCENDANTS AND RIVALS Following the succesful publication of the Lexicon, editions were rapidly published in other countries, such as the US, Italy, and Greece. Interestingly, in the American edition, published in 1846 and largely based on LS¹, mythological detail was added rather than deleted: Μῆτις was described as the daughter of Oceanus and Tethys, first wife of Jupiter, and mother of Minerva.³⁶ In the modern Greek translation of LS⁷, Μέγα Λεξικόν τῆς Ἑλληνικῆς Γλώσσης, μῆτις was translated as ἡ ἱκανότης περὶ τὸ συμβουλεύειν, σοϕία, σύνεσις, εὐϕυία, πανουργία, a word by word translation from the English. By contrast, in the Italian Dizionario illustrato greco-italiano, based on LSJ, the translations wisdom, skill, and craft were rendered as ingegno, scaltrezza, and saggezza. The latter corresponds well to wisdom, and scaltrezza approximates craft, but ingegno cannot be equated comfortably to skill; it rather means ‘intelligence’ or ‘perspicacity’. The three other translations (counsel, plan, undertaking) were compacted into two (progetto, disegno), both of which are closer to plan than counsel. The process of lexical compilation through the medium of another language is necessarily complex. Concepts differ from one language to another, and hence accurate translation—even between modern European languages—is problematic. The American, Italian, and Greek descendants of Liddell and (conceptual) translation of μῆτις. Initially the entry also featured the baffling translation ever ready, but this was deleted in the ninth edition. ³⁵ Further evidence of this trend can be found in the related verb μήδομαι, to be minded, to intend, resolve, which was specifically stated to occur mostly in bad sense. ³⁶ Drisler, the editor of the first American edition, indeed made many changes to LS. See Imholtz 2007, 126 7.

236

Evelien Bracke

Scott were based on different editions, and thus different translations of the same words were used in different languages. The Italian and modern Greek translations were now also one further step removed from the original language, since they approached ancient Greek via the English interpretation of a German lexicon. Moreover, every translator of LS now faced the same dilemma concerning translation versus interpretation: while the Greek translation adhered closely to LS⁷, the Italian translation was rather an interpretation. It is noteworthy, however, that both the Greek and Italian translations maintained a stronger sense of the negative potential of μῆτις than had LSJ: scaltrezza and πανουργία in particular are similar to cunning while avoiding the ambiguities of the term craft. These discrepancies in translation and equivalence highlight the differing connotational systems of similar terms across the languages of modern academia. Because of all the issues surrounding the LS in its various editions and translations, editors of twenty-first century dictionaries generally aim to detach themselves from Liddell and Scott.³⁷ The Brill Dictionary of Ancient Greek, published in September 2015, was based instead on the third edition of Franco Montanari’s Vocabolario della Lingua Greca (2013). Since it is almost a word-by-word translation, it is useful to start by looking at Montanari’s lexicon, which claims to be based on a variety of scholarship as acknowledged in the preface.³⁸ Montanari has the following entry for μῆτις: μῆτις -ιος, ἡ [cf. lat. mētior, ai. māti-, got. mēl, aatd. māl, ags. moēđ, asl. mĕra] (a) sagezza, prudenza, abilità IL. 7.447, al. AESCHL. Pr. 906 ecc.; ἐν πᾶσι θεοῖσιν μήτι τε κλέομαι καὶ κέρδεσιν fra tutti gli dèi eccello per prudenza e per artifizi Op. 13.299 (Atena)//personif. Μῆτις, ἡ Metis HES. Th. 358, 886 ecc. (b) consiglio, disegno, piano IL. 9.93 AESCHL. Suppl. 971, al. SOPH. Ant. 159 ecc.; ἄλλην μῆτιν ὕϕαινε ordiva un altro disegno HES. Sc. 28 • sg. gen. –ιδος AESCH. Suppl. 61/dat. ep. μήτι contr. da μήτιι, rar. μήτιδι HDT. 7.141.3; acc. μῆτιν/pl. dat. μητίεσσι; acc. μήτιας e μήτιδας. In the case of μῆτις, at least, Montanari has not taken any radical steps away from LSJ: the translations and references hardly deviate from it. Schneider’s dual definition, moreover, is still maintained and again the translations are kept indeterminate: scaltrezza, the only ambivalent term in the Dizionario, is not included. Moreover, Donnegan’s vague term prudence has been reintroduced (prudenza). The Brill Dictionary is almost a word-for-word translation of Montanari, offering (1) ‘wisdom’, ‘prudence’, ‘capability’ and (2) ‘counsel’, ‘design’, ‘plan’. It is in one of the examples that a difference between Montanari and the Brill Dictionary emerges: while the former translates the example μήτι τε κλέομαι ³⁷ See, e.g., Chadwick 1994. The discussion regarding updating LSJ does continue; see e.g Glare 1987. ³⁸ Montanari 1995, iii.

Between Cunning and Chaos: μῆτις

237

καὶ κέρδεσιν as ‘eccello per prudenza e per artifizi’, the Dictionary translates as ‘I am renowned for wit and wiles’, replacing prudenza/prudence with ‘wit’, a translation not in their list. This deviation from Montanari cannot be accidental, and must have been made deliberately by an editor in an effort to specify the meaning. This change in translation highlights the difficulty of translating μῆτις, and demonstrates that generic and vague terms such as ‘prudence’ and ‘capability’ fall short in rendering such a complex term. Ultimately, it is interesting to note how little lexical entries have changed since Stephanus’ 1572 Thesaurus Graecae Linguae: he already suggested a binary definition of intelligentia, prudentia, and sapientia on the one hand, and consilium on the other, with an added note that, in some cases, a more pejorative meaning—rendered by him as sollertia—seems implied. One more lexicon merits mention: the ongoing Cambridge Greek Lexicon Project—with a projected publication date of 2019—aims to create definitions through close examination of the Greek text, without reference to Liddell and Scott or other existing lexica. A preliminary look at the draft of the μῆτις entry (with thanks to its editor) reveals the immense improvement on LSJ: the binary definition of quality and action has—rightly—been abandoned, the Passowian principle of providing explanations linking translations has been reintroduced, and thoughtful translations bypass the ambiguity of the LS translations of craft and cunning. The lack of such terms does mean that the harmful application of μῆτις is largely missing: only expedient and scheme are pejorative translations. In general, however, the improvements are considerable.³⁹

13.6. TWO OTHER FRIENDS: DETIENNE AND VERNANT Marcel Detienne and Jean-Pierre Vernant, in their seminal book Les ruses de l’intelligence: la métis des grecs (1974)—translated as Cunning Intelligence in Greek Culture and Society (1978)—offer a more complex interpretation of the term than the generic translations in LSJ and other lexica, and identify a lexical field of other terms in relation to it.⁴⁰ They consider μῆτις to be the opposite of βίη: it entails an indirect approach to attack, deploying trickery (δόλος),⁴¹ craft (τέχνη), deceit, lying, and treachery. Μῆτις is able to adapt itself to any shifting situation (πολύτροπος),⁴² using rich, dense (πυκινός)⁴³ knowledge from past experiences and premeditation, waiting (δοκεύειν) for the right moment

³⁹ See Hire 2005, 179 80 for a discussion of the CGL. ⁴⁰ See Detienne and Vernant 1978, 11 for a discussion of earlier scholarship on μῆτις. ⁴¹ E.g. Aesch. Pers. 107. ⁴² E.g. h.Herm. 13. ⁴³ E.g. h.Dem. 414.

238

Evelien Bracke

(καιρός) to move fast (αἰόλος).⁴⁴ It can transform itself, and is many-faceted (ποικίλος)⁴⁵ and pliable, while remaining impenetrable (ἄπορος) itself. Its form masks rather than reveals: as such, it can create the illusion that it is not μῆτις, as in the example of Odysseus pretending to be No Man, and the Trojan Horse which is disguised as a gift but is in fact a trap.⁴⁶ Its special weapon is the bond, by means of which it can encircle (ἐγκυκλεῖνοῦν) and thus trap others. Expressions of this bond are the net, the web, the trap, and above all the circle. Indeed, the circle ‘is perfect, because it completely turns back on itself, is closed in on itself, with neither beginning nor end, front nor rear, and [ . . . ] in rotation becomes both mobile and immobile, moving in both directions at once’ (Detienne and Vernant 1978, 46). Detienne and Vernant’s work is not unlike Liddell and Scott’s: the two colleagues also commenced on a much-praised herculean task, delivered ‘a programme of research that is far from complete’ (Fisher 1980, 228), and were criticized on many accounts, for example for their perceived lack of focus and methodology (Walcot 1976, 212) and lack of differentiation between μῆτις and magic (Turcan 1976, 224).⁴⁷ Fisher (1980, 228) argued against their ‘unhistorical’ approach, since it presents ‘an increasingly complex and interlocking, yet static portrait of the structures of the ‘Greek mind’, and ignor[es] or underplay[s] origins, influences, changes, or divergences’—which is similar to criticism Liddell and Scott faced regarding their often random ordering of translations.⁴⁸ Detienne and Vernant too faced issues of translation, from French to English, criticized by Fisher (1980, 229); indeed, the English title which renders les ruses de l’intelligence as cunning intelligence is a translation of ‘kluge Massnahme’, a German translation used by Frisk in his etymological dictionary, rather than of the French title. The French ruse is more practical than the English translation as intelligence, and refers to the act of trickery. Throughout the book, however, μῆτις is translated inconsistently—for example as ‘prudence avisée’ (1974: 17; ‘informed prudence’ (1978, 11) and ‘intelligence rusée’ (1974, 32); ‘wily intelligence’ (1978, 27))—which exposes the fluidity of the Greek word yet, without clarification by the authors regarding choice of terminology, suggests that they, and their translator, struggled to pin it down. There have been some developments in scholarship on μῆτις since Detienne and Vernant: Kofman 1988, for example, argued against their understanding of μῆτις in Platonic thought;⁴⁹ and Hawhee 2004, furthermore, argues for μῆτις as a corporeal as well as a cognitive category. In spite of the criticism, however, modern studies still rely heavily on their scholarship, and ⁴⁴ ⁴⁶ ⁴⁷ ⁴⁸

E.g. Hes. Theog. 511. ⁴⁵ E.g. Od. 7.168. Odysseus: Od. 9.366; the Trojan horse: Od. 8.509. Other reviews were more positive, e.g. Lloyd Jones 1979 and Klein 1986, 2. See, e.g., Whatmough 1938. ⁴⁹ See Chanter and DeArmitt 2008, 55.

Between Cunning and Chaos: μῆτις

239

methodology has scarcely been updated from their structuralist approach. Publications on μῆτις in the decades following Detienne and Vernant instead applied their findings to the then current analytical approaches to Classics. In the nineties particularly, μῆτις was a popular topic in gender studies.⁵⁰ In the 2000s, it was connected to a varying degree to the concept of magic in the flourishing studies on magic in antiquity.⁵¹ However, the structuralist approach was never challenged. It is nevertheless clear that a number of issues require attention. First, Detienne and Vernant’s structuralist approach amplifies the contrast between μῆτις as the weapon of the weak on the one hand, and βίη as tool of the strong on the other. However, narratives such as the confrontation of Odysseus and the Cyclops, or that of the Trojan horse, demonstrate that there is no such dichotomy. Mῆτις rather causes, and goes hand in hand with, βίη: Odysseus tricks the Cyclops in order to blind him, and the Greeks cajole the Trojans into opening their city gates so they may slaughter the population. Zeus, furthermore, is not only called μητίετα,⁵² but also has κράτος and βίη as his companions.⁵³ Indeed, Detienne and Vernant argue (1978, 78) that the thunderbolt created for Zeus by the Cyclopes was made ‘with their craft (mēchanai) as much as with their strength’. In the use of the net, web, or other implements, again, μῆτις goes hand in hand with violence: fish are caught, wild animals are trapped, and opponents are captured. Plenty of textual examples can be found in which both concepts go hand in hand.⁵⁴ These examples reveal an inherent connection between the two supposed opposites rather than a strict dichotomy. While the two may be described as mutually exclusive in specific contexts,⁵⁵ the dichotomy with βίη thus does not delineate the universal meaning of μῆτις. The contrast between μῆτις as weapon of the weak and βίη as that of the strong is equally tenuous. Zeus may have been the weaker party in his battle with Cronus, but he continued to use μῆτις after he attained sovereignty. Athena, Hermes, and Hephaestus are other examples of deities who use μῆτις in situations in which they are the stronger party. Figures of μῆτις indeed possess it as inherent mode of life rather than a quality they only use when confronted by a stronger opponent. ⁵⁰ See, e.g., Clayton 2004; Holmberg 1997 and 1998; Lee 2004; Marquardt 1993; Pellizer 1979; Pucci 1986; and Winkler 1990. ⁵¹ See, e.g., Carastro 2006 and Bracke 2009. ⁵² E.g. Il. 1.175. I disagree on the translation of μητίετα by Faraone and Teeter 2004, 205 as ‘wise in counsel’. They contrast the epithet with ἀγκυλομήτης, but since no adjectival prefix is used to delineate the connotation of μῆτις, the adjective must simply mean ‘possessing μῆτις’. While I agree that there is an element of moral advice in Hesiod’s description of Metis’ ingestion (Th. 900: ὡς δή οἱ ϕράσσαιτο θεὰ ἀγαθόν τε κακόν τε), this does not exclude a reading of agathos and kakos as strategic (Faraone and Teeter 2004, 204). ⁵³ Hes. Theog. 383 403; Aesch. PB 12 13. ⁵⁴ E.g. Il. 14.109. ⁵⁵ E.g. Dunkle 1987 on the dichotomy in the Iliad.

240

Evelien Bracke

Finally, Detienne and Vernant never explicitly address the differentiation between μῆτις and related terms such as σοϕία, δόλος, and τέχνη. Yet in LSJ, these terms share translations of ‘wisdom’, ‘craft’ and ‘art’ with μῆτις. They also ignore the issue regarding the decline of references to μῆτις in Greek literature from the sophists onwards (1978, 5). How does one bridge the gap between their definition and Liddell and Scott’s? It may be worthwhile to ask again to what extent a negative connotation is present in μῆτις, how it relates to similar Greek terms, and how it might be rendered in English. It is not the purpose of this chapter to offer a revolutionary new interpretation of μῆτις, but I will end with suggestions regarding the kind of revised reconstruction that might inform a new lexical entry.

͂ ΤΙΣ 13.7. CON STRUCTING A DEFINITION OF ΜΗ The first reference to μῆτις in Greek literature—if one adheres to the standard chronology of Archaic poetry—is Iliad 2.169-70: εὗρεν ἔπειτ᾽ Ὀδυσῆα Διὶ μῆτιν ἀτάλαντον /ἑσταότ᾽.⁵⁶ With the Argives about to depart from Troy, disheartened by Agamemnon’s speech, Athena—goaded by Hera—seeks out Odysseus so he may persuade the Greeks to remain in Troy. The narrator’s description of Odysseus as ‘equal to Zeus in μῆτις’ is repeated thrice more in the Iliad.⁵⁷ Both characters indeed have their own identifiable epithet derived from μῆτις: Zeus is called μητίετα sixteen times in the Iliad; Odysseus is described as πολύμητις a total of eighteen times.⁵⁸ Neither epithet, however, provides information about the meaning of μῆτις. It is only in book 23 of the Iliad that μῆτις is explained (Il. 23.315–18): μήτι τοι δρυτόμος μέγ᾽ ἀμείνων ἠὲ βίηϕι, μήτι δ᾽ αὖτε κυβερνήτης ἐνὶ οἴνοπι πόντῳ νῆα θοὴν ἰθύνει ἐρεχθομένην ἀνέμοισι, μήτι δ᾽ ἡνίοχος περιγίγνεται ἡνιόχοιο. By μῆτις the wood cutter is made much better, and by strength; by μῆτις indeed the helmsman on the wine faced sea directs his swift ship, when it is battered by the winds; by μῆτις one charioteer wins victory over another.

Its parallel and hyperbatic position in the instrumental dative depicts μῆτις as a means to an end for specific professions: wood-cutters, helmsmen, and charioteers. The application differs for each: while the wood-cutter applies ⁵⁶ The phrase Διὶ μῆτιν ἀτάλαντον is also applied to Hector, at Il. 7.47 and 11.20. ⁵⁷ Also at Il. 2.409, 2.636, and 10.137. ⁵⁸ The epithet πολύμητις is also applied to Hephaestus once, at Il. 21.355.

Between Cunning and Chaos: μῆτις

241

his experience to know where to cut the wood so as to fell a tree in the least tiring manner, the helmsman applies his knowledge of the winds to steer a ship with the help of oars from the harbour into favourable winds, and the charioteer applies his understanding of the route and opponents to win in a contest. The constructive ability to plan ahead through strategy and experience is foregrounded. The context of these lines, however, is Nestor’s speech to his son Antilochus who is about to take part in the funeral games for Patroclus. Nestor urges his son to use μῆτις, ἵνα μή σε παρεκπροϕύγῃσιν ἄεθλα (Il. 23.314). An element of deviousness can thus be perceived in this speech since Nestor acknowledges the inferior quality of Antilochus’ horses (Il. 23.311) and wishes to alter the outcome of the race by stealth. Menelaus’ anger at Antilochus’ behaviour (Il. 23.439–41) betrays the questionable connotation of μῆτις, and through his direct speech, the intended audience is invited to share his resentment. A positive translation of μῆτις as ‘skill’, ‘intelligence’, ‘prudence’, or ‘capability’ must therefore be avoided, since the effective use of μῆτις and its duplicitous effects are presented to the audience in juxtaposition through two direct speeches. Even in the Odyssean Cyclops episode, known for celebrating Odysseus’ μῆτις in juxtaposition with the Cyclops’ anthropophagous violence, μῆτις cannot be interpreted as a wholly positive concept. Odysseus has full narratorial control over this episode, and goes to great pains to manipulate the perceptions both of his internal audience at the Phaeacian court, and of the audience listening to the Odyssey: while his companions are described as puppies with entrails falling out when devoured by the Cyclops, the blinding of Polyphemus is represented in terms drawn from the manufacture of a ship and the shaping of a sword by skilful craftsmen, echoing Nestor’s simile.⁵⁹ It is only through other narrators that Odysseus’ deviousness is exposed. When Polyphemus has unsuccessfully called for the other Cyclopes, the primary narrator mentions that Odysseus laughs inwardly (ἐγέλασσε ϕίλον κῆρ, Od. 9.413) at the success of his deception. His companions, too, uncover the ambiguity of Odysseus’ μῆτις, when they beg the hero not to provoke (ἐρεθιζέμεν, Od. 9.494) the Cyclops. Narrators other than Odysseus thus reveal the disreputable aspect of his μῆτις. Both examples apply similes with imagery from craftsmanship—in which craftsmen manipulate nature to create a useful tool or path—to the field of social conflict. While the craft-related similes are constructive images focusing on one man’s solitary toil, in conflict, by contrast, the narratorial voices diverge, and what is one man’s helpful plan is another’s downfall; the connotation depends on the narrator.⁶⁰ Positive terms such as ‘wisdom’, ‘skill’, ‘advice’, and ‘prudence’ must thus not be considered as primary translations of μῆτις. ⁵⁹ For the simile of the companions as puppies, see Od. 9.289; for the Polyphemus similes, see Od. 9.384 94. ⁶⁰ Faraone and Teeter 2004, 203 downplay the ambiguity of Nestor’s speech.

242

Evelien Bracke

The differentiation between μῆτις and similar terms such as σοϕία, βουλή, τέχνη, and κέρδος is another contentious issue, since in the Lexicon they share translations such as ‘craft’, ‘cunning’, and ‘plan’. Athena’s speech to Odysseus upon his arrival on Ithaca exemplies the issue (Od. 13.297–9): ἐπεὶ σὺ μέν ἐσσι βροτῶν ὄχ᾽ ἄριστος ἁπάντων βουλῇ καὶ μύθοισιν, ἐγὼ δ᾽ ἐν πᾶσι θεοῖσι μήτι τε κλέομαι καὶ κέρδεσιν Since you are much the best of all mortals in planning and words, but I among all the gods am famed for μῆτις and profitmaking . . .

Athena draws an explicit parallel between her own capacities and those of her favourite hero, yet juxtaposes them by means of the μέν and δέ construction. The use of two singulars (βουλῇ and μήτι) and two plurals (μύθοισιν and κέρδεσιν) emphasizes the parallel. Because of the parallelism, it is tempting to translate the two combinations as hendiadys.⁶¹ While Odysseus is the best as ‘plotting stories’ (rather than plotting and stories) to persuade others, Athena excels at ‘clever tricks’ which confound the suitors and the Phaeacians. The four terms are very similar and complement each other. Yet while βουλή is a neutral term for a practical plan and κέρδος is primarily destructive in aim, μῆτις drifts uneasily in between these as well as σοϕία and τέχνη. Since there is significant overlap between μῆτις and other terms, and none of them encapsulates the others, μῆτις can be interpreted as one among several types of intelligence, modes of living, and means of attaining a desired end. Translations such as ‘knowledge’, ‘resourcefulness’, ‘scheme’, ‘trickery’, and ‘deception’ ought therefore to be integrated into a lexical entry, in order to emphasize the ambiguous and fluid nature of the term.

13.8. CONCLU SION I hope to have demonstrated that μῆτις is a complex term, and indeed not all entries are so complex in their history and significance. Even if they translate it pejoratively in specific contexts, most scholars consider ‘cunning intelligence’ as a largely positive construct, and translations such as wisdom and skill prevail. However, much of this may be connected with the pro-Odyssean agenda of the Odyssey’s primary narrator, and even brief examples demonstrate that μῆτις was at the least morally ambiguous, particularly in combat against others; it is pivotal for modern lexica to foreground this complex connotation. Ultimately, the exposure of narrative manipulation by secondary ⁶¹ One might argue the same for ὀάρους καὶ μήτιας at h.Aphr. 249.

Between Cunning and Chaos: μῆτις

243

narrators draws attention to the hidden primary narrator’s ability to control and manipulate the narrative, and thus to the inherent presence of μῆτις in the storytelling process.⁶² Any text which engages with the notion of μῆτις thus acknowledges its awareness of its manipulative potential at a narrative level; the reader must therefore carefully consider the context of each occurrence, without assuming a unilaterally positive (or indeed pejorative) connotation. To return to Liddell and Scott, it is clear that, from the outset, the editors’ political and personal agenda impeded progress. The insistence on translations drawn from the Anglo-Saxon word-stock meant that the approximately 60 per cent of the English language which is based on Latin was marginalized, and opaque terms such as ‘craft’ and ‘skill’ were used in unhelpful contexts. Moreover, the dogged reliance on Germanic scholarship—albeit accepted practice in lexical assembly since the Renaissance—created not only a shortcut but also a barrier between English scholarship and the Greek language which was never wholly overcome, as the μῆτις entry reveals: though grammatical details were added, LSJ translations hardly developed from LS¹ and Passowian clarity was lost. The explicit agenda imposed on the Lexicon by its editorial team, combined with the implicit agenda of individual editors when editing specific entries, shaped its formation. The μῆτις entry demonstrates that the Lexicon is not a fixed representation of the Greek into English, but a fluid entity transforming according to its creators’ wishes while responding to scholarly innovations and reviews of previous editions. It is difficult to gauge to what extent the process was a deliberate one or a product of chaotic and reactive editing. Future lexica may be able to shed Liddell and Scott’s shackles—as future research may be able to move beyond Detienne and Vernant—but the term μῆτις, owing to its fluidity and dependence on narratorial voices, will continue to resist definition. The entry for μῆτις indeed exemplifies Liddell and Scott just as the assembly process of the Lexicon exemplifies μῆτις: both balance uncomfortably between cunning and chaos, impenetrable yet simultaneously adapting to changing contexts and weaving their way through the text to persuade the reader of the validity of their projected image.

⁶² Pindar explicitly extolls the μῆτις of poets at O. 1.9 and N. 3.9. See Nisetich 1975 on O. 1.9.

14 Looking for Unity in a Dictionary Entry A Perspective from Prototype Theory Michael Clarke

The resources available today for engaging with ancient Greek thought are vastly more sophisticated than those of the nineteenth century;¹ yet our principal authority for understanding the words of the literature is still recognizable as an outgrowth of the work published in 1843,² and successive layers of improvement and adjustment have not decisively altered the guiding scheme that structures its word-entries. This anomaly has been brought about by more than each generation’s fear of seeming presumptuous or arrogant in challenging the authority of a book that one was taught to trust. It is also bound up with the ambiguity in the status of any printed dictionary: is it meant to be a compendium of scholarly insight into the way the language communicates meaning, or just a convenient tool when one is facing a hard word in a text? The uncertainty of that question shows that dictionary-writing does not fit well into our expectations about academic production. In the mental lexicon of a native speaker, the meaning of a word is the cumulative trace of all the times that speaker has heard and used it: as such it will be fuzzy and diffuse, potentially even chaotic. The lexicographer’s task is to reduce this to manageable order. As such, even the best possible dictionary entry is no more than an εἰκὼς μῦθος, a ‘likely story’ in Plato’s original sense:³ it is the shadowy approximation of a semantic shape that cannot be adequately expressed in words, precisely because that shape is prior to the verbal form in which any given thought is expressed.⁴ As a scholar one is supposed to criticize, to problematize, ¹ I am grateful to Christopher Stray, Michael Silk, and Eric Cullhed for comments on earlier drafts of this chapter. ² See Stray 2010b and Chapter 1, this volume. ³ Plato, Timaeus 29d2. On the meaning of εἰκὼς μῦθος and εἰκὼς λόγος in this passage see Cornford 1937, 28 3; Rowe 2007, 259 60; and Rowe, Chapter 8, this volume. ⁴ See further Clarke 2010, 122. Michael Clarke, Looking for Unity in a Dictionary Entry: A Perspective from Prototype Theory In: Liddell and Scott: The History, Methodology, and Languages of the World’s Leading Lexicon of Ancient Greek. Edited by: Christopher Stray, Michael Clarke, and Joshua T. Katz, Oxford University Press (2019). © Oxford University Press. DOI: 10.1093/oso/9780198810803.003.0014

248

Michael Clarke

to rely on what is knowable and to distinguish it from speculation: but such discipline is a poor friend to one trying to write a dictionary entry, where the challenge is less like academic research and more like structural engineering—the need is to build a structure that works, given the constraints of the site and the needs of its users, as much as to approach truth in the way that professionalism seems to demand. Clearly, however, a good dictionary entry should represent a real step towards entering the patterns of thought among the ancient speechcommunity. This need not entail adopting any dogmatic version of the belief that the semantic shapes of a particular language determine and constrain the shaping of thought among its speakers, as in the standard handbook formulation of the ‘Sapir–Whorf hypothesis’.⁵ But as a matter of practical method, it must be a good policy to proceed on the basis that those shapes may be useful to us, as aliens and outsiders, for guiding our path as we try to find a route into that world of thought and expression.⁶ To say this is to suggest that the lexicographer’s engagement with a word is fundamentally a search for unity: a search for the essential idea that holds together a group of things (referents, concepts, senses, call them what you will) that may not be straightforwardly united in the modern mother tongue that provides our metalanguage and our default assumptions. In the present chapter I will approach the problem with the help of perspectives from prototype theory, one of the richest areas in the relatively young discipline of cognitive linguistics.⁷ Typically, work in this field is presented as a contribution to the understanding of meaning as such, of the workings of the mental lexicon as an aspect of the human language faculty.⁸ My use of prototype theory will be more limited, avoiding any claims to truthvalue and treating it as a sounding-board for new possibilities in modelling and describing the behaviour of words.⁹ It should be admitted from the outset that the ‘Sapir–Whorf problem’—that is, the problem of incommensurate semantic structures in the language studied and the lexicographer’s mother tongue—is likely to prove more pressing with some kinds of words than others, even if we restrict ourselves for

⁵ For a sustained and up to date dismissal of the ‘Sapir Whorf hypothesis’ see McWhorter 2014. In contrast, for revised and nuanced versions of ‘ “soft” Whorfianism’ see for example Lucy 1992; Wierzbicka 1997; and for an up to date account of the corresponding issues in the discourse of cognitive linguistics see Kövecses 2006. ⁶ Cf. Slobin 2003. ⁷ For convenient practical introductions see the essays in Cuyckens and Geeraerts 2008, noted individually below, with Taylor 2002, 440 84. ⁸ The classic statement remains Lakoff 1987, conveniently distilled into conventional aca demic language in Lakoff 1999. For overviews see Taylor 1995; Geeraerts 2010. ⁹ The present study is an attempt to propose a more practical application of approaches to the lexical semantics of ancient Greek that I have attempted in a number of previous case studies and essays, synthesized in Clarke 2010.

Looking for Unity in a Dictionary Entry

249

argument’s sake to simple nouns and adjectives at the basic level of core word-stock.¹⁰ In general, we will find it easier to dodge the difficulties of semantic reconstruction when we face a word whose class of possible referents is defined by their visible characteristics in the world of experience: so, for example, it is easy enough to deal with a word like κύων ‘dog’ or ἐρυθρός ‘red’, a word whose observable behaviour and even its etymology suggests that it is underpinned by an objectively verifiable phenomenon—here, respectively an animal species and an area of chromatic space.¹¹ Less obvious, perhaps, is the converse principle: a word whose range of potential reference is entirely defined by intangible cultural constructs, as τιμή or αἰδώς or ἥρως, can conveniently be separated off into a small and limited group of ‘culturewords’ whose study becomes part of an anthropological study of early Greek thought and belief. As a result, one’s embarrassment at failure to decide in these three cases between pairs of apparently unconnected translationwords—‘honour’ and ‘payment’, ‘shame’ and ‘concern for self-respect’, ‘warrior’ and ‘minor deity’—becomes simply an invitation to turn instead to an anthropology of the relevant area of the ancient world-picture: in these cases, respectively reciprocity, ethics, or popular religion. But the need to give an orderly lexical account becomes unavoidable with words that seem (at least from the modern observer’s point of view) to be poised indeterminately between the visible or tangible realm and the realm of social construction. In what follows I will examine a series of words that fall into this especially challenging category, in the hope that they will provide an opportunity for a fresh approach to lexical interpretation in general.

14.1. PUZZLING POLYSEMY I N EARLY GREEK I begin with a familiar example, the word αἰών. As is well known, in the early epic this word often refers less to a temporal concept (as in later Greek) than to something like an essence of vitality, which underlies or motivates life and which at death is lost (or the subject ‘taken away’ from it). But a significant number of instances survive in which the word αἰών identifies a tangible bodily substance. Two of these, curiously, are in the Homeric Hymn to Hermes.

¹⁰ For the definition of basic level vocabulary, and its special openness to prototype based interpretation, see Geeraerts 2010, 199 203. ¹¹ For the etymology of κύων see Wotko et al. 2008, 436 40; for the contrast between the simplicity of ἐρυθρός and the complexity of certain other Greek words in the same semantic field, such as ἀργός and χλωρός, see Clarke 2004.

250

Michael Clarke

In the first, the god has taken up a living tortoise and goes to separate its shell from its body: ἔνθ’ ἀναμηλώσας γλυφάνῳ πολιοῖο σιδήρου αἰῶν’ ἐξέτορησεν ὀρεσκῴοιο χελώνης. (hHerm 42) There he probed with a chisel of grey iron and gouged out the αἰών of the mountain dwelling tortoise.

At a pinch, one can translate with a term like ‘life-force’, implying that the moist vitality of the flesh makes it an embodiment or representation of the abstract essence of life. But a little later on in the Hymn, we find the noun again in collocation with the same verb: the god, slaughtering two cattle, rolls them over onto their backs δι’ αἰῶνας τετορήσας ‘after he had pierced their αἰῶνες [plural]’ (hHerm 119). If you can gouge out or pierce αἰών, and in particular if the word is a ‘count noun’ that becomes plural when several people are involved, the only sensible conclusion seems to be that the word names some localized part of the body. Since Antiquity it has been held that the word sometimes refers to spinal marrow. Hesychius cites it in this sense, ὁ νωτιαῖος μυελός¹²—but as a scientific term rather than an archaism,¹³ citing a Hippocratic passage specifying the location of a fatal illness (Epid. 7.122). Erotian in his Hippocratic glossary corroborates the existence of this medical term by citing Pindar’s account of how Heracles’ crushed Antaeus’ ribs and αἰὼν . . . δι’ ὀστέων ἐρραίσθη, ‘the αἰών broke out through his bones’ (fr. 111.5 Snell-Maehler)¹⁴—which in turn looks very close to the examples that we observed from the Hymn to Hermes. On the face of it, then, the word belongs in two areas of meaning which in our own language and world-view have virtually nothing to do with each other. The instances assigned to the ‘bone-marrow’ group could perhaps be explained as a series of mutually referenced oddities, all based on the survival in the epic tradition of a usage whose motivation was no longer understood:¹⁵ in this case the lexicographer would be facing no more than a trivial choice between homonymy and polysemy. LSJ responds accordingly and divides the word into two lettered sections, assigning to A everything that can be reasonably linked to the temporal idea ‘period of existence’ and corralling off the bone-marrow instances under B. If we extract the guiding words of the lexicon entry according to the structure implied by the layout, they look like this:¹⁶

¹² Hsch. s.v. αἰών. μυελός is supplied editorially from the alternative formulation later in the entry, ὁ ἐν παντὶ τῷ σώματι μυελός. ¹³ τινὲς δὲ τῶν νεωτέρων . . . ἀπέδωκαν (Hsch. s.v. αἰών). ¹⁴ For the original context in Erotian see Race 1997, 344 5. On Erotian see Dickey 2007, 45. ¹⁵ For such approaches to archaizing poetic language, cf. Silk, Chapter 17, this volume. ¹⁶ I have re inserted the omitted first letter or number in each series.

Looking for Unity in a Dictionary Entry

251

A. period of existence: I. 1. lifetime, life 2. age, generation 3. one’s life, destiny, lot II. 1. long space of time, age; eternity, opp[osed to] χρόνος 2. space of time clearly defined and marked out, epoch, age; hence in pl[ural the ages, i.e. eternity 3. Αἰών, personified 4. Pythag[orean], = 10 B. spinal marrow (perh[aps] regarded as seat of life) This arrangement goes back to the Lexicon’s first edition, and it is easy to see how it seemed unavoidable; but a transverse crack runs through the structure. Section B ends with the uneasy words ‘perh[aps] also Il. 19.27’: this refers back to an instance cited in section A, where Achilles, speaking of his fear that maggots will grow in the corpse of Patroclus, uses the words ἐκ δ’ αἰὼν πέφαται, which must translate as something like ‘the αἰών has been murdered out of him’.¹⁷ A scholion matches this instance to the medical meaning νωτιαῖος μυελός,¹⁸ and the phrasing of the Iliadic words seems in turn to have served as model for the Pindaric passage cited above. If this one instance bridges the gap and refuses to be assigned confidently either in part A or part B, this should encourage at least the suspicion that the presentation is flawed overall. True enough, in poetic or otherwise creative language there is always the possibility that a word is being used artfully in two senses at once, and that the unification is itself anti-traditional;¹⁹ but such an explanation seems not to be available here, given the overall pragmatic urgency of Achilles’ speech. The lexicographer’s unease about assigning this instance to this or that subsection suggests that the structure as a whole has failed to distribute the word’s attestations across a series of constituent areas that match those of the word’s ancient behaviour. Why, then, was the LSJ entry framed in this way? The underlying tactic is clear: the entry begins from the array of English words that one might deploy for translating αἰών in various kinds of contexts, then it establishes a set of senses or sub-meanings on the basis of those glosses, and finally it fits them into a hierarchical structure with three levels—capital letters, roman numerals, arabic numerals. Two English-language concepts keep appearing, time and

¹⁷ Edwards 1991 ad loc. ¹⁸ Schol. bT on Il. 24.725, attributing this interpretation to the γλωσσογράφοι. ¹⁹ For a possible example, see Bacchylides 17.108 where the Nereids dancing in their undersea palace are described as ὑγροῖσι ποσίν ‘with ὑγρός feet’. Famously, in the poetic tradition this adjective was ambiguous between something like ‘moist’ and something like ‘vigorous’ (cf. the parallel example of διερός): here the feet are simultaneously ὑγρός in both senses. In such a case it seems irresistible to guess that deliberate ingenuity has been applied to the inherited poetic language (cf. Maehler 2004, 185).

252

Michael Clarke

life, but the relationship between them is configured differently in different places along the series: and the initial words period of existence seem to be an attempt to hold time and life together in a single idea. The sequence is designed so that at each level of the hierarchy the user can move from one unit to the next by jumping across the narrowest possible gulf, moving each time over what Ladislav Zgusta called a ‘semantic bridge’.²⁰ The problem is not the translation-words themselves—it is hard to improve on them as a store of possible resources to draw on if faced with translating αἰών in a passage of a Greek text—but the internal dispositions of the structure in which they have been arranged. The range of meaning that the word occupied in the mental lexicon of an ancient Greek speaker has been sectioned up and redistributed according to the thought-associations of the age and time of the dictionarywriter and, ideally, of the dictionary-user who follows after him. English, rather than Greek, has had the upper hand throughout the procedure, and this is the heart of the difficulty.

14.2. UNITY AND DISUNITY: THE CASE OF ΜΕΝΟΣ This problem has not gone away with time. Chadwick’s Lexicographica Graeca,²¹ a remarkable example in recent times of an impeccable knowledge of ancient Greek applied to the relatively lowly task of framing dictionary entries, offers a particularly telling example of the issue in its account of μένος.²² Schematically, the logic of Chadwick’s exposition can be fairly represented as in Fig. 14.1. Chadwick’s starting-point, chosen apparently for intuitive reasons, is the balloon of meaning placed here at the centre and marked by a double ellipse. From this a series of other areas of meaning are plotted as neighbours, linked to the first by the movements of thought-association represented here as arrows, with the bolder arrows representing the more decisive movements that take Chadwick from one paragraph of his definition to the next. The numbers that I have imposed are an attempt to trace his intended sequence of movement between them. I have depicted it as an archipelago, with the islands linked by bridges as in Zgusta’s image mentioned above; but of course there are many other ways to represent such a scheme—Chadwick himself imagines a word’s senses as the branches of a tree, though in terms that are hard to visualize.²³ The bridges represent the distinctions or subdivisions that dictionaries typically represent as numbered or lettered sections marked by bold type or indentation, or through the use of terms like fig., hence, metaph.; but what ²⁰ Zgusta 2006, 267 8. ²² Chadwick 1996, 189 95.

²¹ Chadwick 1996. ²³ Chadwick 1996, 12.

Looking for Unity in a Dictionary Entry

of things

of natural phenomena personified

of abstract concepts

of active, dangerous forces

253

6

of people and animals

(6) effectiveness, powerful force

(5) [physical] strength, might

5

[mental not physical] determination to impose one’s will, spirit

[of animals] wildness

1

also of animals [distinguished by heavy breathing] prople “breathe menos”

2 4 [In Homeric periphrasis with genitive of personal name] forcefulness that commands respect

[temporary] aggressiveness, reckless courage, temper

3

...and other bodily substances (as drimu menos, Od. 24.319) its loss is injury or feebleness it becomes milder if calmed down

aggressiveness realised as hostile activity, fighting

it is lost at death, as life-force

Fig. 14.1. Chadwick’s account of μένος in Lexicographica Graeca.

do they signify in the analysis? One might think of them purely as a matter of convenience; but if they are more serious, they must be read either synchronically or diachronically. If synchronically, they must purport to represent the associations that made it possible for a speaker to apprehend why the range of meaning of the word included this particular set of referents. Alternatively, they may be a proposal as to the historical sequence of semantic change that the word’s uses followed over time—in the case of Liddell and Scott, a timespan of up to a thousand years. If the bridges are meant to function in all three ways at once, the unresolved ambiguity means that none of them can be trusted at all.²⁴ It is a curiosity of the history of lexicography that homage was paid for so long to the principle ascribed to Passow, ‘that every word should be made to tell its own story’, in other words that the internal arrangement of each entry

²⁴ For recent discussions of lexicography invoking this problem under different guises compare Hawke 2016, 180 2, with Durkin 2016b, 242 6.

254

Michael Clarke

should plot a path of development over time.²⁵ The ‘historical principles’ of the OED²⁶ corresponded in the Greek lexicon to a slow process of accretion, beginning for each word with a structure dictated by the Homeric evidence and fitting later evidence around it.²⁷ Chadwick invokes this same principle in his introductory essay;²⁸ but in the case of μένος the logic of his arrangement seems to be strictly synchronic—the validity of his account depends on the perceived shortness of the crossings from island to island. It still leaves one wondering why determination to impose one’s will, spirit was chosen as the starting-point. If one does think in diachronic terms, a very different pattern suggest itself. As is well known (and as Chadwick himself acknowledges), the Indo-European prehistory of μένος is not in doubt: Sanskrit mánas- and Avestan manah- are only the most obviously close members of a large set of cognates in other languages that all refer to phenomena of mental agency— mind, intention, will: and even within Greek the early derivations seen in the adjectives δυσμενής and εὐμενής belong in the same territory.²⁹ If one begins from the hypothesis that among the users of the early or prehistoric Greek language thought was understood as the exercise of mental energy, and this in turn as the essence of self-propelled motion, then it is easy to see how the name of that activity could be generalized to encompass other forms of kinetic and forceful action—including, ultimately, the embodiment or representation of such action in concrete bodily essences such as blood, tears, or semen.³⁰ In this way it should be possible to plot the relationship between the senses along a diagram which would look something like Fig. 14.2.

²⁵ The original phrase in the Preface to Passow’s Handwörterbuch was ‘die Lebensgeschichte jedes einzelnes Wortes in bequem geordetneter Überschaulichkeit entwerfen’, and this in turn was underpinned by the argument in Passow’s essay Über Zweck, Anlage und Ergänzung griechischer Wörterbücher of 1812 (see Aarsleff 1983, 252 4, and cf. Considine 2014b). ²⁶ The link was explicit in Herbert Coleridge’s paper in the Transactions of the Philological Society in 1857: ‘The theory of lexicography we profess is that which Passow was the first to enunciate clearly and put in practice successfully, viz., “that every word should be made to tell its own story”, the story of its birth and life, and in many cases of its death, and even occasionally of its resuscitation’ (cited by Willinsky 1994, 30; see also Mugglestone 2000 and 2005, 4 5). ²⁷ Stuart Jones wrote as follows in the 1925 Preface to LSJ, p. iii: ‘Passow had laid down . . . the canons by which the lexicographer should be guided, amongst which the most important was the requirement that citations should be chronologically arranged in order to exhibit the history of each word and its uses. In obedience to this principle, Passow based his work on a special study of the Early Epic vocabulary, and the relatively full treatment of Homeric usage is a legacy bequeathed by him to Liddell and Scott which has persisted throughout the successive editions of their work.’ ²⁸ ‘The arrangement of senses in the dictionary should, so far as possible, mirror the word’s development [in time]’ (Chadwick 1996, 12). ²⁹ The survey of forms by Euler 1979, 222 4 is still the best starting point for surveying the data. For the related problem of the archaic Homeric periphrasis ἱερὸν μένος, see Katz 2010b, 360 1. ³⁰ See for example Il. 1.103 4 (blood), Od. 24.318 19 (nasal mucus?), Archilochus fr. 196a.52 West and perhaps Od. 2.270 2 (semen).

Looking for Unity in a Dictionary Entry

virtually personified (menos of x = x himself)

255

kinetic thought, will, urge to action

violent or aggressive behaviour

force of selfpropelled movement

its source or essence realised as breath

...or another vital substance, moist or flowing

Fig. 14.2. μένος plotted according to its (hypothetical) diachronic development.

The differences here from Chadwick’s approach are twofold. The first is that the area of meaning chosen as the starting-point is shaped as closely as possible by the observable etymology: in the case of a word whose ancestry is so well known, it is reasonable to proceed on the hope that, barring accidents, the polysemy of the word approximately replicates a diachronic development. Within the diagram itself, however, no claim is made as to a sequence over time—note that the lines are not numbered sequentially—but their positioning in relation to each other suggests the possible association of ideas that prompts the affinity between each of these ideas and one or more of its fellows. Additionally, their placing relative to the central area indicates a quantitative judgment as to whether each is relatively close to or distant from the original idea at the centre. If one asks what are the benefits of such a graphic arrangement over the traditional numbered list of sub-senses, the answer is simple: it allows us to include nuanced ambiguities as an inherent aspect of the structure, and it avoids the necessity of pronouncing a judgment as to what is primary and what is secondary in terms of the range of ways we see the words used in the ancient sources. In the same way, such an arrangement could be configured slightly differently for two different members of the speech-community with different varieties of linguistic experience, but without losing sight of the common ground between them and their ability to communicate effectively with each other.

14.3. PROTOTYPE S EMANTICS So far, I have tried to approach μένος without going beyond the terminology sanctioned by traditional lexicography and Classical commentaries. However,

256

Michael Clarke

the reconstruction proposed above can easily be converted into the terminology of cognitive linguistics, in a way that may allow for some useful new insights.³¹ In the scheme favoured by workers in this school, the balloon of meaning from which the diagram begins represents the semantic prototype; each of the areas of meaning plotted around it is an instantiation of the prototype—in other words, a group of points of reference to which the word was applied on the basis of a perceived affinity between each such referent and the concept represented by the prototype. The lines between them represent these linkages as radial: to locate a given instantiation more or less close to the prototype is to indicate the degree to which it is focal—that is, the degree to which it approaches a typical or ‘classic’ exemplification of the prototypical concept. Just how to define or characterize the central space for a given word is the abiding problem of prototype semantics. Some writings on the subject imply that the ‘best example’ is itself identifiable as the prototype, others that no such one-to-one identification is possible and that instead any chosen example will be more or less focal than another, but as a difference of degree alone. At its most extreme and most ambitious level, the theory may even propose that the prototype has no real content at all, and is merely an abstraction from the way native speakers would assign particular instances of the word to a range of near or distant points, clustered around a focal centre and more or less closely associated with each other in groups.³² In practice, however, the lexicographical application of this approach demands that we sketch the prototypical concept in a way that can be printed on a page; and this can have surprising results, even when we rely for examples on our own native-speaker intuitions. The classic demonstration of the potential of this method is advanced in a paper by Fillmore and Atkins on the polysemy of the English verb crawl.³³ Their approach is to propose a prototypical concept, visualized rather than lexically defined—‘moving while holding one’s body close to the ground’—and to arrange attested uses according to their proximity to the prototype, which itself gains clarity and focus as a function of the strategies that are used to locate them in relation to it. Clearly in the baby is crawling along the floor we have an instantiation very close to the prototype; further out we would locate the use represented by this train is moving at a crawl, where the characteristic

³¹ The terminology introduced here is based on an adaptation of the approach of Geeraerts 1997, customized in the light of Lakoff 1987 and Fillmore and Atkins 2000, and adapted for the study of Greek in a number of earlier case studies of mine (Clarke 2004, 2005, 2010). The state of the debate on each issue is well summed up in the corresponding chapter of the exceptionally useful handbook edited by Cuyckens and Geeraerts 2007, among which see esp. Lewandowska Tomaszczyk 2007. The overview by Geeraerts 2010, 182 272 passim, represents the most up to date consensus view of the field known to me. ³² Cf. Lakoff 1999. On the issue of defining a prototype see Geeraerts 2010, 183 9. ³³ Fillmore and Atkins 2000.

Looking for Unity in a Dictionary Entry

257

motion of human limbs is absent and only the characteristic slowness of movement is invoked from the prototype concept. Further out towards the periphery would be placed a use like my flesh crawls at the sound of the Vicechancellor’s voice or the ambitious lecturer crawled in front of the Dean, where the logic of association with the prototype depends on more ambitious kinds of linkage, which could be given labels like metonymy and metaphor—to which we will return below—or may even depend on allusion to an established or lexicalized formula, approaching the status of a suppressed quotation.

14.4. DIACHRONIC OR S YNCHRONIC—OR BOTH One advantage of the prototype-based approach is that it enables a rapprochement between synchronic and diachronic axes. The diagram proposed in Fig. 14.2 answers well to the semantic behaviour of μένος as observed in Homer, Hesiod, and early lyric poetry; but in Greek of the fifth and fourth centuries the word is overwhelmingly applied to the force of violent motion as an abstract or virtually abstract phenomenon, with little or no reference to the old prototypical concept—except in so far as the word continues to be known and used in a high literary register (usually by poets) in ways that recall the Homeric patterns. In this way, we could propose a shape for the fifth-century diagram that deletes or marginalizes those areas of meaning that have become obsolete, and is now dominated by the concept labelled ‘self-propelled movement’. To continue with the terminology of prototype semantics, this area of meaning has now become entrenched and potentially becomes salient as a conceptual unity in its own right.³⁴ Once that is done, one assumes, the other established points of instantiation will progressively lose their connection to the centre, and their relationships to the prototype will be entirely renegotiated: in which case revised lines of connection will need to be drawn to explain how the instantiations are supported. Ultimately, then, a new concept may acquire prototype status as the centre of the word’s communicative life. However, there is a pitfall here. If the account one gives of the prototype is treated as if it were a traditional lexical definition—that is, as a statement of what the word means in the public lexicon of an ancient speech-community— it is liable to seem hopelessly abstract. In earlier case studies on this topic, I found myself setting up for τρέφω the prototypical concept ‘make the unrealized coagulate into fullness’, and for χλωρός ‘fecund oozing green vitality’.³⁵ Similarly, ³⁴ Cf. Schmid 2007. ³⁵ On χλωρός see Clarke 2004, 135, with discussion by Bradley 2013, 130 1; on τρέφω Clarke 2010, 126 7, largely following Benveniste 1966, 292 4; and cf. the nuanced and updated presentation of this approach to τρέφω in Griffith 2010.

258

Michael Clarke

in the case of αἰών above one might be tempted to propose a prototype along the lines of ‘essence of vitality’ and propose that this concept is instantiated potentially in the time-period of life, potentially in substances that embody vitality, and potentially in eternity conceived as vitality unbounded by limitations of time—a proposal that was indeed made in a famous study by Benveniste, to whom we will return below.³⁶ Each such form of words is an attempt to put a finger on an elusive entity: but if they were treated as if they had the status of dictionary definitions, it could reasonably be objected that real-life vocabulary does not work like this—that a robust and historically meaningful account of this word should be tighter, more practical and more effectively constrained. Here a provisional answer, at least, will be that the verbal form given to the prototype does not and cannot claim one-to-one equivalence with the word-meaning articulated on the surface of an utterance: it is not a definition purporting to be able ‘to replace the definiendum both grammatically and semantically’.³⁷ Rather, it is an attempt to pinpoint a pre-verbal or sub-verbal concept around which any or all applications of the word are clustered, and against which any such application would be intuitively ranked by a native speaker. By the same token, the chase after the semantic unities of an ancient language is much more difficult than for those of a modern language, where the task is eased because native speakers are alive among us and because, except with an increasingly slight minority of non-Western languages, the world of its speakers’ experience is vastly closer to our own than is that of speakers separated from us by two millennia and more. If our attempt to verbalize an ancient Greek semantic prototype turns out vague and inadequate, this is best taken as a sign either that we have oversimplified it by eliding the complexities of the mappings between prototype and instantiations, or that we need to work harder to enter into the patterns of ancient thought that provided the centre and the coherence of the whole.

14.5. DEFINING A P ROTOTYPE CONCEPT: THE CA SE OF Α ΛΣ ΟΣ To clarify the challenge posed by applying this method, let me try an example which must correspond to a linguistic exchange that really happened in the 460s BCE. In Aeschylus’ Suppliants the Chorus of young women are railing against the Egyptian men who have come to ravish them, and the singer wishes for the drowning of her persecutor:

³⁶ Benveniste 1937.

³⁷ Hawke 2016, 184.

Looking for Unity in a Dictionary Entry D

259

ἄλσος applied to

E

(A) a grove dedicated to a deity

B C

(B) [any] clump of trees (C) [any] enclosed or demarcated space

A a space imbued with divine vitality

(D) [any] level expanse (E) the sea as a level expanse

Fig. 14.3. A possible prototype based account of ἄλσος. αἲ γὰρ δυσπαλάμως ὄλοιο δι’ ἁλίρρυτον ἄλσος (Suppl. 867 8, following Page’s OCT text) If only you were sent helplessly to death through the sea flowing grove . . .

I have rendered ἄλσος temporarily as ‘grove’, the traditional gloss that heads the entry in LSJ. It is not in dispute that (as the LSJ examples show) many of the early attestations of this word label a clump of trees, and among those several are characterized in context as sacred; others seem simply to refer to a broad expansive space, perhaps characteried by rich vegetation and perhaps not; and within that group there are several that refer specifically to the sea. Let us try to model how a member of the audience of the Suppliants might have made sense of the word when used in this way by the Chorus. A prototypebased reconstruction might take a form like that in Fig. 14.3. Assuming that this is a reasonably accurate representation, it invites us to guess that the listener may have apprehended the word in any of several ways. He or she may simply have recognized the word as one conventionally applied to the sea, which seems to have been already moving towards entrenchment in the variety of language used in this poetry (E in the diagram: see, e.g. Bacch. 17.85, Aesch. Pers. 112). Less passively, the listener may have taken this as an extension or specification of the word’s application to any broad level area, potentially characterized by rich vegetation and vitality (C, D), or may have understood its application to the sea as the metaphorical transference of a meaning primarily associated with literal clumps of trees, either understood in purely vegetative terms (B) or specifically a sacred grove in a holy precinct, as in the instantiation marked here as (A) and placed very close to the prototype. But by taking that bolder step, the listener implicitly moves closer to an apprehension of the prototype concept itself: in other words, a deeper or more poetically rich awareness of language takes one further and further from the initial impression of a chaotic network of disparate meanings, and more towards a relatively abstract conceptual centre. Vitally, several spectators

260

Michael Clarke

sitting alongside each other in the Theatre of Dionysos might each have heard the word in a different way within the range of possibilities encompassed by the diagram. The shape given to the diagram above depends on an implicit decision on how to arrange and rank the instantiations: although this may seem arbitrary, the proposed structure can be easily judged for credibility, elegance, and cohesion—just like any other attempt to reconstruct ancient realities. Further, it can be seen as a commentary on the very practical problem of how to understand and translate the word in its attested contexts. In the Suppliants passage, my initial sea-flowing grove was only a placeholder and is close to nonsense; but I am not sure that Sommerstein was right to reduce the words to ‘the flowing expanse of the salty sea’,³⁸ any more than Gilbert Murray was justified in going to the other extreme of rendering πόντιον ἄλσος at Persians 112 as ‘the holy places of the deep’.³⁹ To stay in much the same speech-community, the verse epitaph for Aeschylus famously declares his war record: ἀλκὴν δ’ εὐδόκιμον Μαραθώνιον ἄλσος ἂν εἴποι (Vita Aeschyli 11; Athenaeus 14.627c, Pausanias 1.14.5) The ἄλσος of Marathon could speak of his glorious valour.

Does this imply the image of a (sacred?) grove of trees at this locale, or does our word simply refer to the level plain running down to the seashore where the battle was said to have been staged, as Sommerstein himself has forcefully argued?⁴⁰ Similarly, when the Chorus of the Persians sing of how they have dared to gaze⁴¹ on the πόντιον ἄλσος, this might imply a specific reference to the Hellespont, if the idea of enclosure is essential to the meaning;⁴² it might suggest the wild and potentially hostile energy of the sea, if we follow the formulation that I propose for the prototype concept in the diagram above; and if Sommerstein is right, the ellipse of meaning marked as D has become salient and this instance communicates nothing more than would (say) πεδίον ‘plain’ used of the sea.⁴³ This is not the place to propose a definitive answer to the question of ἄλσος, or to the equivalent questions posed by the other words discussed earlier. The key point is that in each case the challenge faced by the commentator or the translator can be succinctly expressed in terms of two complementary ³⁸ Sommerstein 2008, 399. ³⁹ Murray 1940, 112. ⁴⁰ ‘Thus invariably [in such passages as the epitaph] alsos denotes, not a (relatively small) area covered with trees, nor a (relatively small) area consecrated to a divinity, but a level area regardless of its size . . . ’ (Sommerstein 2010, 199). ⁴¹ For a defence of the transmitted reading ἐσορᾶν (110) see Garvie ad loc.: Garvie 2009, 82. ⁴² This possibility is explored in a shrewd note ad loc. by Broadhead 1960, 55 6, reviewed without firm conclusions by Garvie, ibid. ⁴³ See, e.g. Aesch. fr. 150.

Looking for Unity in a Dictionary Entry

261

questions: how to impose a wording on the prototype concept, and how to arrange the instantiations around it. The former question is the more vexed, and there is a real danger here that the verbal formulation will be reached by simply splitting the difference between the meanings seen in the attested uses of the word. This problem is inherent in the overall project of linguistic reconstruction: for example, our issue here chimes with problems thrown up by Benveniste’s celebrated essay on the semantic reconstruction of ProtoIndo-European words.⁴⁴ Benveniste’s strategy of ‘looking for a usage in which the disparate senses recover their unity’⁴⁵ led in practice to the positing of extraordinarily abstract areas of ‘original’ meaning: the derivation of words for ‘fear’ and the number ‘two’ from the same root *dwey- was explained on the basis of an original meaning of ‘doubt, ambiguity, doubleness’,⁴⁶ and the perceived relationship between words for ‘oak’ and similar trees on the one hand, and for truth and honour and honesty on the other, was traced to the concept of stability, toughness, reliability plotted for a stem reconstructed as *derwo- and realized in various forms with a range of vocalisms.⁴⁷ The sense of drifting into fantasy here comes from the peculiar difficulty of reconstructing meaning (as opposed to form) for a language whose enactment in real-life communication is unknowable;⁴⁸ but the same can be said, potentially, of any attempt to write up a prototypical concept as if it once had an objective existence in someone’s mental lexicon.

1 4 . 6 . M ET A P H O R A N D ME T O N Y M Y IN COGNITIVE S EMANTICS There is little hope of success here without a disciplined approach to plotting the radial lines that link a prototype to its instantiations. This applies above all to any account of innovative ‘shift’ that works with the terms metaphor and metonymy. There are endlessly variant working definitions for these terms, but it may be useful to give as an example one formulation that has won wide acceptance in the cognitive linguistics school.⁴⁹ Metaphor depends on a shift from one part of the world-picture to a decisively different part, moving a ⁴⁴ Benveniste 1966. ⁴⁵ ‘ . . . En présence de morphèmes identiques pourvus de sens différents, on doit se demander s’il existe un emploi où ces deux sense recouvrent leur unité’ (Benveniste 1966, 290). ⁴⁶ Benveniste 1966, 294 5. ⁴⁷ Benveniste 1966, 298 301. ⁴⁸ On the methodological issue see also Clackson 2007, 189 91, 187 95, 209 13. ⁴⁹ For an overview of metaphor and metonymy in cognitive linguistics see Geeraerts 2010, 203 22. The classic essay towards a working definition of metonymy remains Lakoff 1987, 77 90, building on Lakoff and Johnson 1980. See now Panther and Thornburg 2007 for a review of the theory and its variants.

262

Michael Clarke

lexical item (or group of such items) from a referent in one conceptual domain to a referent in another;⁵⁰ metonymy, in contrast, is a shift in the linguistic form under which lexical items apply to their referents, with no sense of movement between discrete areas of reference.⁵¹ This working definition of the distinction between the two terms is different in spirit and motivation from many that are current in literary theory, where the concern is less with semantic change in the collective language than with once-off poetic innovation,⁵² but the results converge once it is agreed that the key criterion is relative conceptual distance: metaphor as change by ‘analogy or similarity’ involves an implicit or explicit leap, metonymy as change by ‘association or contiguity’ does not.⁵³ Two simple examples will be enough to clarify the contrast. κύων, to judge both by its behaviour within Greek and by its close cognates in other languages, is and always was the word for a dog, and the prototypical image defining its use is not problematic.⁵⁴ When one insults another’s humanity or dignity by calling him a dog, as the Homeric Achilles does when he addresses the cowering Hector as κύον (Il. 22.345) or as the reformed Helen does when she refers to her old disgraceful self as κυνῶπις (Od. 4.145), these are unambiguously instances of metaphor, because clearly the early Greek view of the world (like any other) makes a sharp conceptual distinction between humans and animals. It is not in doubt that there is a leap from one domain of experience to another. Contrast this with one of the more fraught entries in the dictionary, that for the noun δόξα. Since the earliest editions of Liddell and Scott, this word has been subdivided with most of its uses separated into two distinct parts—roughly, δόξα as ‘opinion [without sure knowledge]’ versus δόξα as ‘the experience of the [typically good] opinion of others, reputation, glory’. Clearly the first group refers to the action of forming an opinion and the second refers to the experience of enjoying the good opinion of others. Guided

⁵⁰ The question of the relationship between metaphor and thought, prominent in all writings on cognitive linguistics, is less relevant in the present discussion. ⁵¹ It must be emphasized here that the Aristotelian sense of metaphora subsumes under a single label the two kinds of phenomena that are here distinguished as metaphor and metonymy. See Silk 2003a for a full contextualization. ⁵² For example, Silk’s summary formulation (Silk 2003b) locates its working definition of metaphor to once off creative innovations, in direct relation to Aristotle’s treatment and the later ancient system of tropes. In the present discussion of lexicography, the dissonance between this kind of approach and that encouraged by prototype semantics becomes serious only in the borderline case of a metaphorical usage which has become familiar and conventional. For Silk, such a usage is ‘non deviant’ and therefore does not deserve the name of (living) metaphor at all (Silk 2003a, 122 7, and Chapter 17, this volume, pp. 301 3); for prototype semantics, such a usage can be described as an entrenched metaphorical instantiation and the question of its vitality or otherwise can be left open. ⁵³ Silk 2003a, 120; see further Silk, Chapter 17, this volume, pp. 307 8. ⁵⁴ See Wodtko et al. 2008, 436 40.

Looking for Unity in a Dictionary Entry

263

by the diachronic story and the transparently close relationship of this noun to the verb δοκέω,⁵⁵ one can easily see how these are instantiations of the semantic unity underlying the verb—roughly the concept or bundle of concepts represented by the English words seeming, likelihood, estimation. In both major instantiations of δόξα, the radial relationships are metonymic and not metaphorical, because the mappings are within a single domain—the difference is simply that one is subjective, recognizing the prototypical concept in human thought-processes, and the other is objective, recognizing it in human experience. Over time, as we have seen, one or other of the clusters of application of such a word may become so deeply entrenched that it emerges as an independent focus of meaning with its own prototypical concept underpinning it. Certainly in the Greek of the New Testament and the Fathers, δόξα referring to the manifestation of divine glory seems to take on a life of its own that bears only the most distant relation to the Classical word in its subjective instantiation. In this case, we can say that this application of the word has become salient and constitutes a distinct lexical structure in its own right.⁵⁶ Here, as often, the increasing complexity and even diffuseness of the structure is a function of contact between languages (Greek and Hebrew?), as much as of autonomous internal development.⁵⁷ It is significant here that κύων, for which we posit a nominal prototype concept, is extended by metaphor, while the dependence of δόξα on a verbal concept means that the prototype is not restricted to a single domain and can be reapplied much more easily by metonymy. These examples are simple, but others are less so. Consider the case of σπέρμα.⁵⁸ Recognizing the word as a transparent nominal formation in *-mn: from the root seen in the verb σπείρω,⁵⁹ the prototype could be characterized in two ways. On the one hand it could be seen as an entirely indefinite and abstract image of scattering and spreading across any of a potentially endless range of domains of experience; on the other, it could be pinned precisely on an activity-sequence from the realm of plant life and arable farming—the sowing, scattering, propagating of seed. This question is forced upon us by the famous simile from the Odyssey in which the exhausted hero has lain down to sleep under a pile of leaves and he is compared to the scene of a farmer preserving σπέρμα πυρός, the ‘seed of fire’, by burying a red-hot ember in the earth (Od. 5.490). Is this usage metaphorical? Such a judgment is implied once one translates it into English ⁵⁵ Cf. Ruijgh 2004, 58 9. The relationship with δέκομαι and related words, whose synchronic relationship with δοκέω is made less immediate by the difference of vowel grade, is beyond the scope of the present discussion. ⁵⁶ For the data and an excellent discussion see Bauer et al. 2000, s.v. ⁵⁷ Cf James, Chapter 10, this volume. ⁵⁸ For a contrasting but not irreconcilable approach, cf. Silk, Chapter 17, this volume, 306 7, 309 10. ⁵⁹ For a survey of forms based on Homeric data see Risch 1974, 49 51.

264

Michael Clarke

as ‘seed’, and this is echoed in the scholia;⁶⁰ but it remains equally possible that it is not a metaphor at all, but simply a metonymic reapplication of the prototypical action of ‘scattering, distributing’ to a particular point of instantiation—and one guesses that the application to the ‘scattering’ of fire by distributing hot embers is relatively peripheral rather than focal. The same question applies with equal force whenever the language of ploughing, sowing, sprouting is used of human sexuality and reproduction: are we dealing with a systematic mapping from one domain of experience to another, or does the world-picture embedded in ancient Greek treat both sets of phenomena as manifestations of a single process?⁶¹

14.7. WORDS AN D WORLD-PICTURE When the problem of metaphor and metonymy is formulated in this way, it becomes clear that any useful judgment depends not only on linguistic engagement but also on engagement with the world-picture of the original speakers of the ancient language. For each and every lexical item, the approach demands sensitivity first to the characteristic mechanisms of metonymic extension, and then to the characteristic forms of the conceptual domains in which the world-picture is organized. In other words: once a dictionary seeks to plot structures based on semantic prototypes, it ceases to be merely an account of words and becomes an account of the bundles of ideas and associations that motivated and underlay those words.⁶² It is because traditional dictionaries have failed to distinguish between these two types of lexicographical information— translation-glosses on the one hand, reconstructed conceptual unities on the other—that their archipelago-shaped definitions so often fall short of a convincing account of the way a word conveyed meaning in Antiquity. Here, I suggest, classical scholarship has something to learn from the discourses that have developed around the equivalent problems in the study of Biblical Greek. Biblical scholarship has been an outstanding arena for the development of a scientific approach to translation studies,⁶³ perhaps because so many wars have been fought and souls won and lost over the word-by-word interpretation of scripture; and for the same reason lexicography has been problematized in this field with greater energy and urgency than in the study of pagan varieties of the classical languages. In an important essay, Frederick ⁶⁰ ἐστὶ γὰρ ὁ τρόπος μεταφορά (sch. on Od. 5.490). The context implies that the sense of μεταφορά here is close to that of the English derivative. ⁶¹ Cf. DuBois 1988, 39 85; Vernant 1990 [1974], 189 91, for varying ways to make anthropo logical sense of this and similar mappings. ⁶² For a recent case study exemplifying the ambiguity and acknowledging the resultant danger of circularity in argument, see Cairns 2014. ⁶³ See for example Nida 1964, 2003.

Looking for Unity in a Dictionary Entry

265

W. Danker has shown how New Testament scholars since the early twentieth century gradually learned to separate out the two kinds of information and to embody the distinction in lexicographical presentation.⁶⁴ The most spectacular example of this is Louw and Nida’s Lexicon of the New Testament Based on Semantic Domains,⁶⁵ which takes the radical step of refusing to arrange words alphabetically and instead clusters them into ninety-three groups based on their perceived interrelationships in terms of the world-picture of their users. In so far as this lexicon functions as a cultural encyclopedia, it becomes nothing less than an anthropological guide to the patterning of ancient thought; but its componential approach makes it less radical in practice than in theory, and it invites the criticism that the domains thus established are inadequately motivated, making the work problematic unless one agrees in advance to a wholesale acceptance of the model of cultural reconstruction adopted by the authors.⁶⁶ A more low-key response was applied by Danker himself and his team in BDAG, his monumental reworking of Bauer’s GreekEnglish Lexicon of the New Testament. The decisive step here was underpinned by a simple practice, the rigorous application of a system of typographical categorization to the English-language material in the entries: One of the most important innovations in BDAG is the use of boldface roman to distinguish meanings as well as functions of headwords from translation equiva lents or glosses, which appear in boldface italics.⁶⁷

To these Danker also added a third level, the surrounding discursive material that remains in default roman type. At first sight this tripartite division seems like a mere matter of housekeeping: but its consistent application allows the Lexicon to take on a whole new role as a repository of data not only about words in action but also, and separately, about the cultural underpinnings of their semantic behaviour. The entries for words that have been battlegrounds of religious dissension—as παρθένος, σάρξ, or υἱός in the collocation υἱὸς θεοῦ⁶⁸—allow the contributor and the reader to separate out the provision of lexical glosses from the discussion of syntactic and contextual information and then of the wider and often fraught matters of cultural and theological context. I will restrict myself here to one seemingly innocent example, the entry for the verb μισέω, whose essential parts read as follows:⁶⁹ Depending on the context, this verb ranges in mng. from ‘disfavour’ to ‘detest’. The Eng. term ‘hate’ generally suggests affective connotations that do not always do justice to esp. to some Semitic shame honour oriented use[s] of μισέω [sc. calqued onto Hebrew substrate vocabulary] in the sense ‘hold in disfavour, be disinclined to, have relatively little regard for’. ⁶⁵ Louw and Nida 1988. ⁶⁴ Danker 2004. ⁶⁶ For a more positive evaluation cf. Vorster 1998. ⁶⁷ Danker 2004, 16. ⁶⁸ Bauer et al. 2000, 1025 6. ⁶⁹ Bauer et al. 2000, 652 3 (excerpted).

266

Michael Clarke

(1) to have a strong aversion to, hate, detest (b) w. acc. of thing (c) abs. (2) to be disinclined to, disfavour, disregard, in contrast to preferential treatment For clarity’s sake I omit the many quotations, digressions, and bibliographical references: but the skeleton shown here is enough to demonstrate this lexicon’s power to direct the reader towards a reconfigured sense of how the word made sense in the ancient conceptualization of emotive behaviour. There is an instructive contrast with the corresponding entry in LSJ: [H]ate, once in Hom., c. acc. and inf., μίσησεν δ’ ἄρα μιν δηΐων κυσὶ κύρμα γενέσθαι Zeus hated (would not suffer) that he should become a prey . . . , Il. 17.272

The entry seeks to make do with the one-word gloss-translation, but immediately falls foul of it in the first cited example, where the uneasy and ambiguous parenthesis amounts to an admission that hate cannot cope with the ancient realities. If the Lexicon is to function successfully it cannot restrict itself to mappings from words in one language to words in the other: it must attempt mappings between world-pictures, acknowledging the need to explain the ancient realities as a system of thought that must be studied sensitively before its lexical forms can be understood. If taken to extremes, the approach proposed here from Danker might seem like a counsel of perfection. Against this it should be emphasized that the proposed cultural-anthropological discussions are actually called for only in a relatively small number of words, typically simple and radical words or even families of words whose internal semantic and formal kinship is transparent. Thus, the discussion accompanying a new entry for μισέω would not only include and subtend the entries for compounds and derivatives, but would also be correlated with the entries for words in a complementary or contrastive relationship with it. A case in point here is the verb στέργω, whose semantic range is something like a reverse image of that of μισέω, involving a range of variation from emotive love to submissive acceptance.⁷⁰ Consider a classic example of the verb in action, from the opening scene of the Prometheus Bound. Zeus’s henchmen Force and Power declare that they are punishing Prometheus in order to re-educate him, ὡς ἂν διδαχθῇ τὴν Διὸς τυραννίδα στέργειν (Aesch. PV 10 11) . . . so that he will be taught to love the princedom of Zeus ⁷⁰ Interesting, the LSJ entry for στέργω is much richer and more subtle than that for μισέω, which does nothing to nuance or extend the one word gloss hate.

Looking for Unity in a Dictionary Entry

267

How to hear the tone and connotations of στέργειν here would make for a classic problem in literary exegesis. It can be argued that the deliberate perversion of the vocabulary of emotional relationships evokes the grotesque brutality of the tyrant, and forms part of the play’s programme of mapping the kingship of Zeus onto narratives of abnormal autocracy among humans; or that the peculiar deployment of the vocabulary of friendship is meant to assimilate Zeus’ power to Greek images of oriental tyranny; or conversely that this is a ‘normal’ and unmarked use of the verb in a weak or emotionally neutral sense—which, as it happens, is the line followed in LSJ, where our text appears under ‘be content or satisfied, acquiesce’.⁷¹ But in terms of the lexical issues that we have broached, much of the debatable space is in the lexicology: how do we plot the area the semantic life of στέργω, how in effect do we give an account of the prototypical concept at the centre of its radial diagram? To begin to attempt such a project is to plot the shapes of ancient assumptions about affection, submission, and power: and in this way the challenge demands that the student of word-meanings should also be a student of the ancient construction of social and psychological concepts and categories.⁷² If the presentation encouraged by prototype theory makes it a little easier to clarify such challenges and develop a discourse for talking about semantics and world-picture in the same breath, then it may encourage a new suppleness in the investigation of ancient words—above all, perhaps, for the simple reason that the heuristic possibilities of a diagram made of arrows and ellipses are so much more rich than those of a column of printed text broken up by letters and numbers.

⁷¹ LSJ s.v. στέργω III. ⁷² ‘The importance of prototypicality effects for lexical structure . . . blurs the distinction between semantic information and encyclopedic information’ (Geeraerts 2007, 1160).

15 Discourse Particles in LSJ A Fresh Look at γε David Goldstein

15.1. INTRODUCTION The outdated quality of LSJ is more evident in its treatment of discourse particles than perhaps anywhere else.¹ In recent decades, our understanding of natural language meaning has witnessed dramatic advances and reached a level of sophistication and detail that was unimaginable under Queen Victoria. Recent work in formal semantics and pragmatics in particular has shed new light on the heterogeneous class of words known as discourse particles (e.g. Kaplan 1999; Gutzmann 2015; Szabolcsi 2015).² The goal of this chapter is to contrast the LSJ account of the particle γε with an approach that takes advantage of some of the conceptual tools of twenty-first-century semantics and pragmatics. The rest of this article is structured as follows. Section 15.2 discusses the question of why describing the meaning of discourse particles is so challenging. From here I home in on the particle γε, ‘one of the subtlest and most

¹ I am indebted to a number of people for comments on earlier versions of this chapter, including Michael Aubrey, Gabriel Bertilson, Nicolas Bertrand, Stephen Carlson, Ben Cartlidge, Carlo DaVia, Marc Greenberg, Dieter Gunkel, Dag Haug, Jesse Lundquist, Pura Nieto Hernán dez, Tom Recht, Jessica Romney, Brent Vine, Anthony Yates, the audience at the Melpomene Chair Greek Studies Conference, and last but certainly not least the editors. ² In the interests of making the ideas in this chapter as accessible as possible, I have kept formalism to an absolute minimum. I fully agree with the following view of McReady 2012, 785, however: ‘There is an extensive literature on these expressions [viz. discourse particles] in traditional Japanese grammar, much of which is insightful; but in many cases the literature suffers from its lack of formalism. It is difficult to make precise what one takes the meaning of an expression that completely lacks truth conditional content to be without the requirements and clarity imposed by a formal model.’ On the advantages of formal semantics for the classical languages, see Devine and Stephens 2013 with Goldstein 2013a and Goldstein 2016b. David Goldstein, Discourse Particles in LSJ: A Fresh Look at γε In: Liddell and Scott: The History, Methodology, and Languages of the World’s Leading Lexicon of Ancient Greek. Edited by: Christopher Stray, Michael Clarke, and Joshua T. Katz, Oxford University Press (2019). © Oxford University Press. DOI: 10.1093/oso/9780198810803.003.0015

Discourse Particles in LSJ: A Fresh Look at γε

269

elusive particles’, according to Denniston (1954, 115). Section 15.3 critically reviews its article in LSJ. Beginning with section 15.4, I present the results of a fresh examination of the particle in two Platonic dialogues, Meno and Cratylus (examples from other Platonic dialogues and tragedy will play a supporting role at various places). As constraints of space do not permit a full exposition of the particle even in these two dialogues, I concentrate on the most salient aspects of its meaning, in particular phenomena that LSJ does not mention. I begin by clarifying the relationship between the distribution of γε (i.e. where it occurs in the clause) and its meaning, which is actually more straightforward than the literature would lead one to believe. Section 15.5 provides background information on the semantics of questions and focus, which are crucial for understanding the meaning of γε. In section 15.6, I argue that γε is characterized by two semantic properties, scalar interpretation and non-at issue semantics. Three readings illustrate the first property: that of a superlative modifier ‘at least’ (section 15.6.2); of a scalar exclusive ‘just’ (section 15.6.3); and of particularizer ‘in fact, in particular’ (section 15.6.4). The nonat issue character of γε is presented in section 15.7. Section 15.8 concludes with prospects and directions for future research.

15.2. THE CHALLENGES OF PARTICLE MEANING Elucidating the meanings of discourse particles, both in Greek and crosslinguistically, is notoriously challenging (Krifka 1993; McReady 2012). Trying to specify the meanings of a discourse particle in a corpus-bound language such as Ancient Greek is all the more difficult, as witnessed by the decision of the editors of the Lexikon des frühgriechischen Epos (LfgrE) not to offer definitions of particles after the first volume (in fact, volumes two through four decline to offer glosses for any function words). But what is it exactly that has led to this aporia on the subject of particles? The study of Greek particles is beset by (at least) the following two fundamental problems. First, we are often at a loss for a suitable English paraphrase. It is not always clear what a discourse particle contributes to an utterance. I have often heard classicists (both linguists and non-linguists alike) say that Greek particles are ‘untranslatable’.³ This is true to an extent: it is quite difficult to describe the meaning of a discourse particle with everyday language. Other scholars have wrongly equated this untranslatable character of discourse particles with meaninglessness, however: Neil (1901, 201) and Reece (2009) ³ It is in fact an insight that scholars investigating discourse particles in other languages have also come to, e.g. McReady (2012, 779): ‘[P]articles have a meaning which is, in some sense, ineffable in that it does not admit of any satisfying paraphrase’.

270

David Goldstein

consider certain particles in metrical contexts to be mere prosodic fillers.⁴ Worse yet, the view that particles do not mean much of anything in Greek has led to the view that they are unworthy of serious study.⁵ Discourse particles appear not to mean anything because they typically do not affect the truth conditions of a proposition. Truth conditions are the conditions that must hold in the world for a proposition to be true (Lewis 1970). Switching lexical items within a proposition often yields different truth conditions. For instance, one perceives an immediate difference between It’s raining and It’s snowing. The former requires that it be raining at the moment of the utterance (in some contextually defined space), while the other requires that it be snowing (again in some contextually defined space). Likewise, when it comes to function words, we can readily characterize the difference in meaning between It is raining and It was raining. It is precisely this effect that appears to be absent when it comes to discourse particles. That is, we are (often, but certainly not always) hard pressed to say how the appearance of a discourse particle affects the meaning (broadly construed to include both semantics and pragmatics) of an utterance. While discourse particles typically do not contribute truth-conditional meaning, they certainly do contribute meaning (see recently Soltic 2014 on this very point). Roughly speaking, discourse particles comment on the utterance, or a portion of the utterance, in which they occur (Zimmermann 2011). Their meanings often revolve around the relationship between other propositions in the discourse or the relationship between a proposition and an interlocutor (which can of course also be modelled as relations between propositions, or sets of propositions, in as much as speaker attitude can be formalized as sets of propositions).⁶ Consider for instance the AustroBavarian discourse particle eh (the examples are from Zobel 2015): ⁴ This view has been around for millennia. Dionysius Thrax (Ars grammatica 96 100 Uhlig) also held this view with the class of particles that he called παραπληρωματικοί ‘expletives’, of which γε is a member. ⁵ The standard reference work, Denniston 1954, is now woefully out of date. Further investigations of Greek particles include Hoogeveen 1788 and 1829; Hartung 1832 3; von Bäumlein 1861; Bakker 1988; Sicking and van Ophuijsen 1993; Rijksbaron 1997; Bonifazi 2009a, 2009b; Revuelta Puigdollers 2009; Bonifazi 2012. Páez (2012) collects more recent bibliography on particles. The landscape has changed dramatically with the publication of Bonifazi, Drummen, and de Kreij 2016. Outside of Classics, discourse particles have been investigated from a number of different perspectives: see e.g. Fraser 1996; Cinque 1999; Blakemore 2002, 2004; McReady 2005; Fraser 2009; Aijmer and Simon Vandenbergen 2011; McReady 2012; Martín Zorraquino and Portolés Lázaro 1999. This sample does not do justice to the wealth of literature available. Within Indo European more broadly, see Berenguer Sánchez 2000; Widmer 2009; Dunkel 2014. ⁶ Wakker (1997, 211) argues that particles contribute nothing to the description of an event or action, but rather serve as ‘road signs’ to help an addressee understand the structure of an utterance.

Discourse Particles in LSJ: A Fresh Look at γε

271

(1) Austro-Bavarian eh i. Ist dein Katzerl eh drinnen? is your kitten eh inside Your kitten is inside, I hope? ii. Ist dein Katzerl eh nicht drinnen? is your kitten eh not inside Your kitten is not inside, I hope? Roughly speaking, the particle eh in polar interrogatives contributes a preference for the positive answer (which I have here tried to highlight with ‘I hope’ at the end of the question). So in (1i), the speaker prefers that the proposition your kitten is inside will hold. In (1ii), the expectation is that the proposition your kitten is not inside holds. With both examples we see how the meaning concerns the speaker’s attitude toward the proposition. With γε, we will see below that it makes crucial reference to the relationship between the host proposition and other propositions in the discourse. The upshot of all this is that one typically has to look beyond the utterance that contains a discourse particle to understand what the particle contributes. In section 15.7, I further characterize discourse particles as contributing non-at issue meaning (roughly speaking, non-at issue meaning refers to the content of an utterance that is not asserted). We can thus say that discourse particles contribute use-conditional meaning (cf. Gutzmann 2015): discourse particles seem not to affect the truth conditions of their host utterance, but they do affect the conditions under which the host utterance can be observed in use. The second problem that we face in the investigation of discourse particles is that, even when we have intuitions about the contribution of a discourse particle, which we capture with an English paraphrase, it is difficult to explain what the paraphrase itself means. Consider the classic paraphrase of μέν . . . δέ as ‘on the one hand’ . . . ‘on the other hand’. The paraphrase gives us some sense as to what the meaning of these particles is and when the construction can be used. But this approach also comes with drawbacks. By relying on paraphrases, we shift the question of the meaning(s) of μέν and δέ onto the English paraphrases (which in itself is risky, because they are only partial synonyms). In other words, if μέν and δέ mean ‘on the one hand’ . . . ‘on the other hand’, how do we define the meaning of the paraphrases? Without this additional step, we do not have an adequate account of the meanings of μέν and δέ. We have only identified English phrases that we can (sometimes) substitute for Greek particles. The sole way to avoid this problem is with a formal metalanguage. Indeed, discourse particles cross-linguistically often exhibit highly idiosyncratic meanings such that it would be difficult to find any suitable paraphrase. In the account of γε below, paraphrases play only a secondary role.

272

David Goldstein

1 5 . 3. ΓΕ IN LSJ Before turning to the lemma for γε in LSJ, I want to provide a few concrete examples of the particle to ground the discussion: (2) Declarative Socrates οἶσθα οὖν τίνας ϕησὶν Ἡσίοδος εἶναι τοὺς δαίμονας; Meno οὐκ ἐννοῶ. Socrates οὐδὲ ὅτι χρυσοῦν γένος τὸ πρῶτόν ϕησιν γενέσθαι τῶν ἀνθρώπων; Meno οἶδα τοῦτό⸗γε. Do you know who Hesiod says the daimones are? No, I don’t. Do you also not know that he says the first race of men was a golden race? I know this⸗γε. Plat. Crat. 397e (3) Interrogative Cassandra ἒ ἒ παπαῖ παπαῖ, τί τόδε ϕαίνεται; ἦ δίκτυόν⸗τί⸗γ’ Ἅιδου; ἀλλ’ ἄρκυς ἡ ξύνευνος, ἡ ξυναιτία. Ah! Ah! What apparition is this? Is it a net⸗γε of death? No, it is a snare that shares his bed, that shares the guilt of murder. Aesch. Ag. 1114–16 (Denniston 1954, 124–5) (4) Directive Electra ἔπειθ ̓ ἑλοῦ⸗γε θάτερ ̓, ἢ ϕρονεῖν κακῶς ἢ τῶν ϕίλων ϕρονοῦσα μὴ μνήμην ἔχειν. Then choose⸗γε one or the other: either be imprudent or prudent without regard for your loved ones. Soph. El. 345–6 (Denniston 1954, 125–6) These three examples establish the ability of γε to occur in declarative, interrogative, and directive contexts. The particle is far more frequent in declaratives than in the other two contexts, however. The following is the entry for γε in LSJ with the examples removed (for other lexicographic accounts, see Cunliffe 1924, s.v.; Slater 1969, s.v.; Schwyzer 1988, 561; Dunkel 2014, 2.279–83; DGE, s.v.; Denniston 1954, 114–62 remains the most detailed account of the particle to date): I. with single words, at least, at any rate, but often only to be rendered by italics in writing, or emphasis in pronunciation 2. with Pronouns 3. after Conjunctions, to emphasize the modification or condition introduced by the subjoined clause

Discourse Particles in LSJ: A Fresh Look at γε

273

4. after other Particles 5. when preceding other Particles, γε commonly refers to the preceding word, while the Particle retains its own force: but sts. modifies the sense of the following Particle, γε μήν nevertheless II. exercising an influence over the whole clause 1. epexegetic, namely, that is 2. in dialogue, in answers where something is added to the statement of the previous speaker 3. to heighten a contrast or opposition a. after conditional clauses b. in disjunctive sentences to emphasize an alternative 4. in exclamations, etc. 5. implying concession III. γε freq. repeated in protasis and apodosis IV. POSITION: γε normally follows the word which it limits; but is freq. placed immediately after the Article Starting from the top, Roman numerals mark the highest-order groupings. Strangely, these first categories do not contain the same type of information. Sections I and II divide the attestations of γε according to their scopal behaviour, that is, whether they modify a single word or a clause. It is not clear where phrases are supposed to fit in: either γε does not scope over phrases, or these examples are contained under I. Headings III and IV, however, have nothing to do with scope. III contains an out-of-the-blue aside about the ability of γε to appear in both a protasis and an apodosis. IV notes a frequent mismatch between surface syntax and semantics: γε can be hosted by a word that it does not modify (more will be said about this in a moment). It is not clear from the above outline whether the meaning of γε is sensitive to scope domain, or whether it has the same meaning when it modifies individual words as it does when it operates on larger units such as clauses. The gloss at least, at any rate from I does not appear under II. Working downwards, the Arabic numerals under I group attestations of γε into various co-occurrence classes (such as pronominals and complementizers⁷) and interactions with other particles. Under II, the classes are less uniform: three describe the functions of γε (1 epexegetic, 3 contrast heightening, and 5 concessive), one is based on discourse type (2 dialogue), and the last concerns utterance type (4 exclamations). The design of this entry is problematic. First, although the distinction between categories I and II is real, it is not acknowledged that it can be difficult

⁷ Complementizers are words that introduce clauses, such as if, whether, and that. The meaning of this term overlaps with the traditional term subordinating conjunction.

274

David Goldstein

to discern when a particular example belongs to one category and not the other. Consider, for instance, clauses that begin with a sequence of complementizer-γε, such as εἴ γε. Is the scope of the particle the complementizer or the whole clause? This surface ambiguity is not limited to clauses that begin with complementizers, but rather extends to all clauses in which γε occupies second position. Second, to characterize the functions of γε, Liddell and Scott use descriptors such as emphasize, lay stress on, and strengthen an assertion. It is unclear what any of these phrases actually means.⁸ Furthermore, the categories above are not designed to be mutually exclusive. That is, assignment to one category (i.e. meaning) does not mean that it is not also a member of another category. One is then left wondering why a particular example is cited in one category and not another. If users of the dictionary want to know what the meaning space of γε is, and how to fit a particular token into this framework, it is not easy to do this. Finally, there is a theoretical question lurking in the background: what is a distinction in ‘meaning’ when it comes to discourse particles? We observed in section 15.2 above that discourse particles typically do not have truthconditional effects. Truth-conditional effects are, however, one of the surest ways to know that we are dealing with a difference in meaning. To take II above as a concrete example, do meanings 2 (‘in dialogue, in answers where something is added to the statement of the previous speaker’) and 3 (‘to heighten a contrast or opposition’) really belong on the same level? Could one not legitimately classify 3 as a subtype of 2? I raise this issue not because I think such an adjustment should be made to the structure of the article, but rather to say that we as readers of dictionaries need to be clear what we are looking at when it comes to the way lexicographers carve up the meaning space of words.⁹

15.4. DISTRIBUTION Turning to the distribution of the particle in the clause, Liddell and Scott write that ‘γε normally follows the word which it limits; but is freq. placed immediately

⁸ It is not clear to me whether these labels refer to prosodic properties in addition to semantic ones. I am unaware of any evidence that would link e.g. sentence stress with the host of γε. ⁹ It is worth noting that while the lemma for γε in the DGE is much richer and more fine grained than the above account of LSJ, it too suffers from many of the faults mentioned above. It is unfortunate that the editors of the DGE did not take advantage of contemporary methods in corpus linguistics, as have become standard in the lexicography of modern languages (see e.g. Biber 2005; Atkins and Rundell 2008; Hanks 2013).

Discourse Particles in LSJ: A Fresh Look at γε

275

after the Article’. The use of but to open the second part of this description suggests that the appearance of γε after the definite article is somehow unexpected or at odds with the first part of the description. Whichever one was intended, both are incorrect. The basic distributional generalization is simpler than LSJ (or Denniston 1954, 146–50) would have us believe: γε is hosted by the first word of its argument.¹⁰ If the surface argument of γε is a single word, then that word will host γε (see Ar. Nub. 401, with Denniston 1954, 119). If the argument of γε is a complex unit such as a phrase or a clause, then γε will be hosted by its first word (contra Denniston 1954, 146): (5) ἡ δὲ [θηλὴ]⸗ἆρά⸗γε, ὦ Ἑρμόγενες, ὅτι τεθηλέναι ποιεῖ ὥσπερ τὰ ἀρδόμενα; Hermogenes, is [θηλή]⸗γε so called because it makes things flourish (τεθηλέναι), like plants wet with showers? Plat. Crat. 414a (6) ἤδη γάρ σϕι [τό⸗γε Δέλτα], ὡς αὐτοὶ λέγουσι Αἰγύπτιοι καὶ ἐμοὶ δοκέει, ἐστὶ κατάρρυτόν τε καὶ νεωστὶ ὡς λόγῳ εἰπεῖν ἀναπεϕηνός. For we have seen that, as the Egyptians themselves claim and I judge, [the Delta]⸗γε is alluvial land and has only lately (as it were) come into being. Hdt. 2.15.2 In (5), γε is hosted by the noun θηλή, because this is its argument. By contrast, in (6) the particle is hosted by the determiner τό, because the particle takes as its argument the entire phrase [τό Δέλτα]. Typically the position of γε makes it clear what its argument is. There is one persistent ambiguity, however: when the host of γε is the first word of a clause, then one has to decide whether its scope is restricted to the host or is in fact the entire clause: (7) Socrates ἐγώ σοι ἐρῶ. τῇ γάρ που ὑστεραίᾳ δεῖ με ἀποθνῄσκειν ἢ ᾗ ἂν ἔλθῃ τὸ πλοῖον. Crito ϕασί⸗γέ⸗τοι δὴ οἱ τούτων κύριοι. I will tell you. I must die on the day after the ship comes in, must I not? So those say who have charge of these matters. Plat. Crit. 44a

¹⁰ I use the term word here to mean specifically ‘morphosyntactic word’, that is, a word as a syntactic element. Other clitics in Greek select prosodic words as hosts, so it is important to distinguish this property of γε. We can think of words such as γε as functions (in the mathem atical sense) that map inputs to outputs. I use the term argument here and below to refer to the input of a word. So γε will take an argument as an input on the basis of which it will return a particular output. If the argument of γε is a phrase composed of multiple words, then it will be hosted by the first word of the argument.

276

David Goldstein

Here one has to decide whether γε takes as its argument the entire clause, or only the verb ϕασί. On either construal, the same surface position of the particle is predicted. The particle does not appear to the left of its host because it is a secondposition clitic, the distribution of which is governed by ‘Wackernagel’s Law’ (Wackernagel’s Law is neither a law nor was it discovered by Wackernagel, as Wackernagel himself acknowledged; see further Goldstein 2010, 2014, 2016a). The precise mechanisms of this phenomenon would take us too far afield. For our purposes, all that is relevant is that γε needs to incorporate phonologically with an element on its left. So we can think of its distribution as the result of two competing constraints: to appear as far to the left within its scope domain as possible while still occurring after its host. There is one deviation from this generalization. There are cases in which the focus of the utterance is ellipsed but γε is nevertheless present and takes as its argument the ellipsed focus (subscript ‘F’ abbreviates ‘focus’): (8) Meno εἴπερ [ἕν]F⸗γέ⸗τι ζητεῖς κατὰ πάντων. Socrates ἀλλὰ μὴν ζητῶ⸗γε. If you’re looking for just [one]F definition for all the examples. In fact I am after just/precisely (that). Plat. Men. 72d In Socrates’ response, ζητῶ is the host of γε, although the particle takes as its argument an ellipsed anaphoric expression referring to the single definition of Meno’s remark. Since this phenomenon concerns the nature of ellipsis more than the use of γε, I have nothing more to say about it. I wanted to at least call attention to its existence, since the standard handbooks betray no awareness of the phenomenon.

15.5. ALTERN ATIVES: QUESTION S AND FOCUS As alternatives will play a crucial role in the three readings of γε below, we begin with the semantics of alternatives. Within formal semantics, questions are standardly held to denote sets of propositions (Hamblin 1958, 1973; Groenendijk and Stokhof 1984). The meaning of a question such as Who did Fatima invite to the party? is the set of its possible answers¹¹ (the Oxford brackets ⟦ ⟧ refer to the semantic denotation of the expressions they contain, and the curly braces {} denote a set): ¹¹ Other analyses restrict the denotation to the set of true answers. This distinction carries no significance for our purposes.

Discourse Particles in LSJ: A Fresh Look at γε

277

(9) Propositional Approach to Questions ⟦Who did Fatima invite to the party?⟧ ⇝ {⟦Fatima invited Henry⟧, ⟦Fatima invited Jack⟧, ⟦Fatima invited Noa⟧, ⟦Fatima invited Wilson⟧, . . . } The question thus denotes a set. The answers here differ in the value that they supply for the interrogative pronoun who. Essentially the question is an open proposition (that is, a proposition with a variable): (10)

Fatima invited x to the party.

Here we have swapped out the interrogative pronoun for the variable x, which represents the set of values (people, in this case) that makes the sentence true. The semantics of questions is so important because discourse is thought (under some models at least) to be organized around questions (and sets of questions) that are under discussion (Roberts 2012). According to this view of discourse, focus is then the information that fills in a variable of a question: (11)

A: Who did Fatima invite to the party? B: Fatima invited [Henry]F.

The question Who did Fatima invite to the party? was answered by picking one of the alternatives that it denotes. Henry provides a value for the variable in the open proposition and is thus the focus of the utterance. If we do not pick one of the alternatives, then we end up with discourse incoherence: (12)

Question-answer incongruence A: Who did Fatima invite to the party? B: I really like [wine]F.

The response I really like wine does not qualify as an answer to the question because it does not lie within its set of alternatives. The set of possible values that could fill in the variable (i.e. the x in the open proposition above) is the set of focus alternatives (Rooth 1985, 1992, 1996).¹² Two dimensions of meaning are typically recognized, the so-called ordinary meaning and the focus meaning (superscript ‘o’ abbreviates ‘ordinary meaning’; superscript ‘f ’ abbreviates ‘focus meaning’): (13) Ordinary meaning Fatima invited [Henry]Fo = Fatima invited Henry (14) Focus meaning (unordered) Fatima invited [Henry]Ff = {Fatima invited Henry, Fatima invited Jack, Fatima invited Noa, Fatima invited Wilson, . . . }

¹² The variable filling approach to focus goes back at least to Kvíčala 1870 and Paul 1920; within Generative Grammar, the locus classicus is Jackendoff 1972.

278

David Goldstein

The focus meaning of an utterance is thus very similar to that of questions (Beck and Gergel 2015, 246).

1 5 . 6 . AN UP D A T E D L O O K In this section, I offer an updated look at the meaning and use of γε. In lieu of a complete account of the use of the particle, the discussion will focus on the following three aspects. The first is to delimit the scope of what we should be trying to explain in presenting an account of the meaning of γε (section 6.1). The second is to clarify the relationship between the distribution of γε and its meaning; neither Denniston (1954, 114–62) nor LSJ provides an account of this relationship. Finally, concerning the meaning, I argue that γε is a scalar operator: it imposes a scalar interpretation on the material in its scope domain. That is, the particle serves to rank elements according to a parameter, which will be supplied by context.¹³ I offer three examples of this scalar behaviour: superlative modifier ‘at least’ (6.2); scalar exclusive ‘just’ (6.3); particularizer ‘in fact’ (6.4).

15.6.1. Focusing on the Compositional One reason why γε is ‘one of the subtlest and most elusive particles’ (Denniston 1954, 115) is its diversity and frequency. Concerning the former, the particle occurs across various speech act types including declaratives, questions, and directives. The number of elements that host γε seems unconstrained, as it includes nouns, adjectives, adverbs, verbs, determiners, and complementizers (we will see below in section 15.6.6, however, that interrogative pronouns do not host γε). Given this diversity, it is important to have a clear target of what should be explained. To my mind, previous accounts of γε have tried to include too much. There are at least three phenomena that should be treated separately: response phrases with γε (e.g. πάνυ γε, σϕόδρα γε); particle combinations with γε (e.g. μέντοι⸗γε, γοῦν, γάρ < *γε⸗ἄρ; see further Denniston 1954: 119–20, 150–62); and the pronominal forms ἔγωγε and ὅγε (on the latter, see the detailed survey of Bertrand 2015). It is not the case that these issues are unworthy of treatment. But if our goal is to explain the contribution that γε makes to an utterance, we need to set (at least) these three phenomena aside, because they could very well represent non-compositional ¹³ Consider the following example: ‘John cannot ride a bicycle, let alone a motorcycle.’ Here bicycle and motorcycle are ordered on a scale of difficulty. See further Toosarvandani 2010; Goldstein 2013b.

Discourse Particles in LSJ: A Fresh Look at γε

279

or lexicalized phenomena. That is, it is no trivial matter to figure out whether πάνυ γε, for instance, is built up from its component parts or better viewed as a word-like entity. Likewise, it is not clear with certain particle combinations whether they can be decomposed into the meaning of their parts or whether the two together have an indivisible meaning. One wonders whether it is the non-compositional nature of certain combinations of γε that lead to its showing up twice: (15) οἶμαι ἔγωγε πάντων⸗γε μάλιστα ϕίλον. I think that (τὸ ὅσιον) is most dear of all. Plat. Euthphr. 15b (16) μὴ μέντοι⸗γε μὴ σπανίσας⸗γε ἄλλως ἀνοίξῃ. With the exception of being in need, do not however open (this tomb) under other circumstances. Hdt. 1.187.2 Each example contains two instances of γε in one clause. While prima facie it looks as if the particle is being iterated, it may be the case that one of the tokens of γε is the ‘real’ one, while the other is simply part of a lexicalized word (ἔγωγε, μέντοι⸗γε).¹⁴ Examples such as (15) and (16) are not common, however, and more research would be required to know whether they can all be explained away in this manner (see further Denniston 1954, 144; Smyth 1956, §2822).

15.6.2. Reading 1: Superlative Modifier ‘at least’ Alternatives are crucial to the meaning of γε, because the particle comments on the status of its argument among alternatives. In its role as a superlative modifier,¹⁵ which can be glossed with ‘at least’, the particle locates its argument at the lower bound of a scale of alternatives:¹⁶ (17 [= 2]) Socrates οἶσθα οὖν τίνας ϕησὶν Ἡσίοδος εἶναι τοὺς δαίμονας; Meno οὐκ ἐννοῶ. ¹⁴ Joshua Katz calls my attention to a parallel phenomenon in γάρ⸗ῥα strings (see further Katz 2007). ¹⁵ ‘At least’ is characterized as ‘superlative’ because its argument ranks as low as or lower than all its scalar alternatives. If someone is said to be at least forty years old, that statement is true if that person is forty or older, i.e. forty is the minimum age that the person might be. Morpho logically, least is of course the superlative form of comparative less. ¹⁶ Liddell and Scott actually come very close to saying something similar in I.5, where they write of γε τοι that it implies ‘that the assertion is the least that one can say’ (italics theirs). (It is not clear to me what role they envisioned for τοι.) Space constraints unfortunately preclude a full presentation of scalar models and their principles. I refer the reader to Fillmore, Kay, and O’Connor 1988 and Israel 2011.

280

David Goldstein

Socrates οὐδὲ ὅτι χρυσοῦν γένος τὸ πρῶτόν ϕησιν γενέσθαι τῶν ἀνθρώπων; Meno οἶδα [τοῦτό]F⸗γε. Do you know who Hesiod says the daimones are? No, I don’t. Do you also not know that he says the first race of men was a golden race? [This]F at least I know. Plat. Crat. 397e Before trying to understand the contribution of γε in Meno’s answer, we first have to understand the denotation of Socrates’ question: (18) The denotation of Socrates’ question {You do not know that Hesiod says that the first race of men was a golden race, You do know that Hesiod says that the first race of men was a golden race} Meno picks the second alternative, but his answer does more than this. It also comments on the position of the answer on a scale of alternatives. Meno locates his answer against the backdrop of the following conceptual scale, which ranks the obscurity of Hesiodic knowledge: (19) Scale of Hesiodic knowledge ⟨Hesiod says the initial race of men was a golden race, who Hesiod says the daimones are, . . . ⟩ What Meno is saying is that knowing that Hesiod says that the initial race of men was a golden race is the minimum of what he knows about Hesiod. Here we also have a negated higher bound, namely the identity of the Hesiodic daimones.¹⁷ It need not be the case that a higher scalar value is rejected, however: (20) Socrates ἢ οὐ μανθάνεις ὅτι λέγω; Meno [δοκῶ]F⸗γέ⸗μοι μανθάνειν. Or are you not taking in what I am saying? I at least [think]F I understand. Plat. Men. 72d Socrates’ question presents two alternatives, {you are not taking in what I am saying, you are taking in what I am saying}. The question divides the answer space into two equivalence classes: either you understand or you do not understand. Meno’s answer is not, however, within this answer space. He needs the gradience of a scale, which is what γε affords. Here γε takes the verb δοκῶ ‘(I) think’ as its argument (that is, δοκῶ is the word to which the particle applies), which I take to be the focus of the ¹⁷ I am grateful to Dieter Gunkel for this insight.

Discourse Particles in LSJ: A Fresh Look at γε

281

utterance. The effect of the particle is again to locate its value at the lower bound of a scale of alternatives: (21) A scalar model ⟨I do not understand, I think I understand, I understand, . . . ⟩ Here the scale is the degree to which Meno understands what Socrates is saying. The least that can be said is that Meno thinks he understands. The higher value ‘I understand’ is neither ruled out nor committed to. What Meno excludes with his answer is the lower proposition, I do not understand, which is what Socrates seemed to expect was the case. This reading has two pragmatic effects. The first is that the question that is being discussed remains open. To take example (20) as illustrative, Meno does not offer a specific answer to Socrates’ question, but rather a range of answers whose minimum (i.e. thinking that he understands) is the focus of the utterance. The answer does not resolve the questions with the alternatives presented by Socrates’ question, however. The second effect is a suspension of implicatures (Kay 1992). In a typical cooperative conversation, speakers offer as much information as the context requires (this is known as Grice’s Maxim of Quantity: see Grice 1975). On the basis of this behaviour, interlocutors routinely draw inferences to the effect that stronger utterances do not obtain. For instance, if someone answers the question How many children do you have? with three, this is interpreted to mean ‘three and no more’. One line of analysis holds that the component ‘and no more’ is derived by pragmatic inference and not part of the meaning of three. Logically speaking, if one actually had five children, the answer three would still be true, but insufficiently informative to the point that one would be accused of lying by anyone who knew the true number. To return to γε, with this reading, upper-bound implicatures of the sort just described are not licensed. That is, one is not to infer a meaning ‘and no more’.

15.6.3. Reading 2: Scalar Exclusive ‘just’ The second reading of γε also positions the host of the particle at the lower bound of a scale. The difference between this reading and the previous one is that higher alternatives are excluded rather than included: (22) εἴπερ [ἕν]F⸗γέ⸗τι ζητεῖς κατὰ πάντων. If you’re looking for just some [one]F definition for all the examples. Plat. Men. 73d This clause is an extract from Meno and Socrates’ discussion of virtue, in which Meno tells Socrates that if he is looking for just one definition of virtue, it is the ability to govern people. As with the ‘at least’ reading, this use of γε

282

David Goldstein

positions its host at the lower bound of a scale. The natural numbers form a ready scale, so it is easy to see that with the value one we are at the lower bound of a scale in this context. In contrast to the ‘at least’ reading, however, here γε excludes higher scalar alternatives: Meno’s conditional refers to the prospect of finding one and only one definition of virtue, not at least one.

15.6.4. Reading 3: Particularizer ‘in fact’ The third reading differs from the preceding two in that it offers a stronger alternative to a proposition that is under discussion. More specifically, the proposition with γε unilaterally entails some proposition active in the discourse: (23) Socrates Ἕλλην μέν ἐστι καὶ ἑλληνίζει; Meno πάνυ γε σϕόδρα, [οἰκογενής]F⸗γε. Is (he) Greek and able to speak Greek? Absolutely, (he was born) [in this house]F in fact/specifically. Plat. Men. 82b Socrates’ question sets up the alternatives {he was born in Greece, he was not born in Greece}. What Meno does in his answer is to offer an informationally stronger answer than either of these alternatives. Meno’s answer is informationally stronger because it entails one of the alternatives in the question: ‘He was born in this house’ entails ‘He was born in Greece.’ The reverse does not, however, hold. The following examples further illustrate this reading: (24) Creon καὶ ταῦτ ̓ ἐπαινεῖς καὶ δοκεῖς παρεικαθεῖν; Chorus [ὅσον γ᾽, ἄναξ, τάχιστα]F συντέμνουσι γὰρ θεῶν ποδώκεις τοὺς κακόϕρονας βλάβαι. And you recommend this? You think that I should yield? [As quickly as possible]F in fact, my lord. For harm of the gods makes short work of the misguided. Soph. Ant. 1102–4 (25) Crito ἄτοπον τὸ ἐνύπνιον, ὦ Σώκρατες. Socrates ἐναργὲς μὲν οὖν, ὥς γέ μοι δοκεῖ, ὦ Κρίτων. Crito [λίαν]F⸗γε, ὡς ἔοικεν. A strange dream, Socrates. A clear one, Crito, at least as it seems to me. In fact, [too]F (clear), as it seems. Plat. Crit. 44b In both examples, we find not just affirmation, but strengthened affirmation. In (24), the chorus urge Creon not simply to yield, but to yield immediately.

Discourse Particles in LSJ: A Fresh Look at γε

283

Likewise in (25), Crito’s response is affirmation with a narrowing of the reference, as a dream that is too clear entails a dream that is clear.

15.6.5. Non-focal Sets of Alternatives Thus far the examples we have looked at all involve the set of focal alternatives—that is, the set of alternatives of a question under discussion. Given how common this type is, one can easily take away the impression that γε is a focus quantifier. This description of the particle is too narrow, however, as γε can also operate on non-focal alternatives: (26) Socrates ἢ ταὐτὸν πανταχοῦ εἶδός ἐστιν, ἐάνπερ ὑγίεια ᾖ, ἐάντε ἐν ἀνδρὶ ἐάντε ἐν ἄλλῳ ὁτῳοῦν ᾖ; Meno [ἡ αὐτή]F μοι δοκεῖ ὑγίειά⸗γε εἶναι καὶ ἀνδρὸς καὶ γυναικός. Or, wherever we find health, is it of the same character universally, in a man or in anyone else? Health at least seems to me to be [the same]F, both that of a man and that of a woman. Plat. Men. 72d–e (27) Oedipus τίνων τὸ σεμνὸν ὄνομ ̓ ἂν εὐξαίμην κλύων; Stranger [τὰς πάνθ ̓ ὁρώσας Εὐμενίδας]F ὅ⸗γ ̓ ἐνθάδ ̓ ἂν εἴποι λεώς νιν, ἄλλα δ ̓ ἀλλαχοῦ καλά. Whose awful name might I hear and invoke in prayer? [The all-seeing Eumenides]F the people here at least would call them: but other names please elsewhere. Soph. OC 41–3 In example (26), Socrates asks whether health is the same everywhere or not, to which Meno replies that it is the same. We thus know that ἡ αὐτή is the focus of his answer. Crucially, γε is hosted not by this phrase but by ὑγίεια ‘health’. Example (27) further illustrates this possibility with a constituent question: Oedipus is asking the name of the goddesses to whom the area on which he is treading belongs. The focus is τὰς πάνθ ̓ ὁρώσας Εὐμενίδας, as it supplies a value for the interrogative pronoun of the question. This phrase does not host γε, however. It is hosted later in the clause by ὁ ἐνθάδ ̓ . . . λεώς. We established in section 15.5 that questions denote sets of alternative propositions. This set of alternatives appeared to be the source of the alternatives at the lower bound of which γε locates its argument. Examples such as (26) and (27) raise the question of the source of the set of the non-focal alternatives.¹⁸ The ¹⁸ Dieter Gunkel calls my attention to Krifka 2008, 267 8, which defines contrastive topics as containing a focus. On this analysis, if one were to treat examples such as (26) and (27) as contrastive

284

David Goldstein

function of γε in this context seems to be to indicate that the focus of the utterance does not fully answer the question. So in example (27), Oedipus inquires about a single name for the area, while the stranger’s answer suggests that it has more than one designation.¹⁹

15.6.6. No Interrogative Hosts Amidst the robust cross-categorial behaviour among the examples in the preceding sections, there is one noticeable gap: interrogative pronouns do not host γε (Denniston 1954: 124–5 lists some possible exceptions to this generalization but deems them textually suspect; a TLG search yielded no hits). On my analysis this is predicted because γε needs as its input two arguments: the ordinary semantic value of its host and a scale (i.e. an ordered set of alternatives). Interrogative pronouns cannot provide γε with the first of these arguments. As we saw above in section 15.5, questions denote sets of alternatives: (28 [= 9]) Propositional Approach to Questions ⟦Who did Fatima invite to the party? ⟧ ⇝ {⟦Fatima invited Henry⟧, ⟦Fatima invited Jack⟧, ⟦Fatima invited Noa⟧, ⟦Fatima invited Wilson⟧, . . . } An interrogative pronoun such as who thus denotes a set, and not a member of a set. But the surface argument γε has to be a member of a set. Thus this syntactic restriction results directly from the meaning of the particle.

15.7. NO N-AT ISSUE MEANING It has become increasingly common to recognize that utterances often correspond to more than one proposition and that these propositions belong to different ‘dimensions’ of meaning (Bach 1999; Potts 2005). Recent research has focused on the distinction between at-issue meaning and projective content. Roughly speaking, at-issue meaning is the primary, asserted meaning of an utterance (cf. the proffered content of Roberts 2012). This content is topics, they would at heart contain a focus constituent and γε would associate with focus after all. On this analysis of information structure, it seems that any set of alternatives would have to be analysed as ‘focal’. The distinction between (26) and (27) and its use with focal alternatives of constituent questions would then amount to a difference in whether or not the focus is embedded (as it would be with contrastive topics: i.e. [[argument]Focus]Topic). ¹⁹ This class is reminiscent of the implicational topics of Büring 1997. But the use of γε does not seem designed to signal an interest in a different topic, but rather to indicate that more may be relevant than is contained in the speaker’s question.

Discourse Particles in LSJ: A Fresh Look at γε

285

described as at-issue because it addresses the question under discussion (Simons et al. 2011; see also Matos Amaral, Roberts, and Smith 2007 and Tonhauser 2012). Projective content is non-asserted additional content that is communicated by either lexical items or particular syntactic constructions. Presuppositions and conventional implicatures appear to belong to this class. After reviewing a couple of semantic properties of projective meaning, I argue below that γε (with other Greek discourse particles) belongs to this class (cf. Potts 2005, 16; Gutzmann 2015, 41). While various tests have been developed to distinguish at-issue and projective content, perhaps the central property of the latter is its inability to be directly denied. Consider the appositional relative clause who stole from the FBI in the following mini-dialogue (Potts 2005, 13): (29)

A: Ames, who stole from the FBI, is now behind bars. i. B: No, that’s wrong. ii. B: No, Ames never stole from the FBI.

With the response in example (29i), speaker B can only negate the proposition Ames is now behind bars. To negate the appositional relative clause, the speaker has to explicitly mention its content, as in example (29ii). Certain types of projective content also project beyond logical operators that are part of the at-issue content. In the following example, the appositional relative clause escapes the negation of the main clause (Potts 2005, 114): (30)

It’s false that Alonzo, a big-shot executive, is now behind bars.

The predicate it’s false negates the proposition Alonzo is now behind bars. It does not and cannot negate the appositional phrase a big-shot executive. The following example illustrates the inability to directly deny γε: (31) Socrates εἰ δὲ ἀεὶ ὡσαύτως ἔχει καὶ τὸ αὐτό ἐστι, πῶς ἂν τοῦτό⸗γε μεταβάλλοι ἢ κινοῖτο, μηδὲν ἐξιστάμενον τῆς αὑτοῦ ἰδέας; Cratylus οὐδαμῶς. And if it (that which is never in the same state) is always so (i.e. in the same state) and the same, how could this at least change or move and yet not give up its form? It wouldn’t at all. Plat. Crat. 439e What Cratylus denies is the possibility that something that is never in the same state changes or moves without giving up its form. He is not denying the contribution of γε to the meaning of τοῦτο. This inability to negate γε results from the simple fact that it outscopes the question operator. That is, since γε cannot be questioned, it cannot be denied. On this analysis, we expect that γε also cannot be assented to. This prediction is borne out:

286

David Goldstein

(32) Socrates εἰργάζετο⸗δέ⸗γε ἡ περιττή; Cebes ναί. This was just the result of being an odd number? Yes. Plat. Phaed. 104d Here Phaedo is agreeing with the proposition ‘The result was produced by the concept of the odd’ and not to the position on the verb εἰργάζετο on the relevant scale. The question operator is not the only logical operator that γε outscopes. It exhibits the same behaviour with negation: (33) Socrates ἦ καὶ ὁμολογοῦσιν, ὦ Εὐθύϕρων, ἀδικεῖν, καὶ ὁμολογοῦντες ὅμως οὐ δεῖν ϕασὶ σϕᾶς διδόναι δίκην; Euthyphro οὐδαμῶς [τοῦτό]⸗γε. Yes, but do they acknowledge, Euthyphro, that they have done wrong and, although they acknowledge it, nevertheless say that they ought not to pay the penalty? [That] at least they certainly do not do. Plat. Euthphr. 8c Although negation outscopes (i.e. precedes) γε, semantically the particle outscopes negation. That is, the reading ‘That is not the least that they do’ is not available. My account predicts that γε will never fall within the scope of logical operators such as negation. In other words, one cannot negate or otherwise modify γε, owing to its background quality. This prediction appears to hold.

15 .8 . C O N CLU S I O N AN D P R OS PE C TS This chapter has offered a new analysis of the particle γε as a scalar operator, which has at least the following three readings: (34) Semantic dossier i. Superlative modifier ‘at least’ ii. Scalar exclusive ‘just’ iii. Particularizer ‘in fact’ In each of these readings γε typically associates with the focus of the utterance, but it need not. Assuming that we are on the right track in identifying this set of meanings, we have a considerable way to go before we can claim to have anything that approaches an adequate account of the meaning of this word. Perhaps the first and most pressing is the question of a Gesamtbedeutung. I have concentrated here on explicating the basic readings of γε with little

Discourse Particles in LSJ: A Fresh Look at γε

287

attention to the question of whether these three readings are synchronically related, i.e. whether we can set up one single basic meaning from which the individual readings are derived. Turning from γε to its textual environment, Denniston (1954, 116) observed that γε is most at home in ‘lively dialogue’. It is not clear what constitutes ‘lively dialogue’, but it is true that γε is most frequently attested (among classical authors) in Aristophanes and Plato. I highly doubt that there is a direct relation between the use of the particle and the text type: that is, γε does not make dialogue more ‘lively’, and Greek speakers did not feel a need to add the particle to contexts that qualified as ‘lively’. At this point, I would advance the idea that the frequency of γε correlates specifically with the frequency of (information-seeking) questions. Looking beyond Greek, there is the question of the comparative and diachronic evidence. And here there are tantalizing possibilities to connect γε with function words elsewhere in Indo-European (Frisk 1960–72, s.v.; Beekes 2010, 263; Dunkel 2014, 2.279–83). One would like to know what kinds of source constructions give rise to scalar operators and, furthermore, why the phonetic erosion that accompanies grammaticalization can lead to irregular sound changes (see e.g. Longobardi 2001). If nothing else, I hope that I have been able to elevate Greek particles from the troughs of mockery and thereby to demonstrate that they are worthy of (and will repay) serious investigation. Our understanding of natural language meaning has advanced so much since the composition of LSJ, and it is time that our field met twenty-first-century standards (cf. Devine and Stephens 2013, 3).

16 LSJ and the Diachronic Taxonomy of the Greek Vocabulary James Clackson

Aristotle’s dictum that ὁ ἄνθρωπος φύσει πολιτικὸν ζῷον (Politics 1253a), usually translated as ‘man is by nature a political animal’ or sometimes ‘man is by nature a social animal’, is both well known and not well understood, and consequently might even persuade a non-classicist to look up the meaning of the phrase πολιτικὸν ζῷον in LSJ. The LSJ definition of πολιτικός is clear: 4. living in a community. But the entry for ζῷον is perhaps more intriguing, since here the dictionary states that ‘The word [ζῷον] is post-Hom., no generic word used for animal being found till after the middle of the fifth cent. B.C.’ (this is corrected in the Revised Supplement to read ‘no generic word used for animal being found before Semonides’). The interested reader may then become curious about the ways in which the ancient Greeks before Semonides conceived of the world if they had no general term for animal, and what it was in the fifth century which precipitated the development of a new taxonomic term. If our mythical interested reader wanted to follow up the development of the conceptual world of terms relating to ‘animal’, there is currently no way in which to do so. It is, of course, possible to search through the digital version of LSJ in order to find a number of different definitions which contain words such as ‘animal’, ‘beast’, or ‘creature’, but there is nothing in any of these other entries that follows up the lead of the parenthetical statement made at the end of the entry for ζῷον. Indeed, if one looks further afield through the entries of LSJ there is little that is helpful for classifications or the history of taxonomic relations. There are only occasional statements relating to generic terms; an example is found s.v. νομεύς: ‘generic term including the special αἰπόλος, βουκόλος, ποιμήν, συβώτης’. Furthermore, there is no definition known to me in the whole LSJ that records how taxonomic or generic terms expand or narrow their meaning throughout the history of the Greek language. Indeed, James Clackson, LSJ and the Diachronic Taxonomy of the Greek Vocabulary In: Liddell and Scott: The History, Methodology, and Languages of the World’s Leading Lexicon of Ancient Greek. Edited by: Christopher Stray, Michael Clarke, and Joshua T. Katz, Oxford University Press (2019). © Oxford University Press. DOI: 10.1093/oso/9780198810803.003.0016

LSJ and the Diachronic Taxonomy of the Greek Vocabulary

289

even when LSJ relates the meaning of one word to another, it does so in largely ahistorical terms. To illustrate this aspect of LSJ, one can compare its entry for βούλομαι with that of Chantraine 2009, both of them contrasting the sense of βούλομαι with that of ἐθέλω. At the LSJ entry for βούλομαι (A) will, wish, be willing, we are told that it opposed to ἐθέλω, consent, but then later that ‘ἐθέλω is also used = “wish”’ (with examples). LSJ further adds a section on Homeric usage, noting that ‘Hom. uses βούλομαι for ἐθέλω in the case of the gods, for with them wish is will; ἐθέλω is more general, and in sts. used where βούλομαι might have stood’. Chantraine begins his account of the two verbs with Homer, noting that ἐθέλω is more common in epic, and is the default verb meaning vouloir, whereas βούλομαι is better understood as désirer, préférer. In Attic prose, βούλομαι encroaches on the ‘want’ and ‘wish’ meanings of ἐθέλω, which is specialized in the sense ‘être disposé à, accepter’. In Ionic, however, ἐθέλω was more common, and becomes the standard verb for ‘want’ in koiné and in later Greek, up until the modern language, although βούλομαι is also retained in use. The account given by Chantraine is easier for the reader to make sense of, and shows the advantage of considering the semantics of a word through diachrony. It is possible to imagine a reference work on Greek vocabulary, organized on semantic grounds, which would show how terms developed over time, expanding or contracting their meanings. Indeed, this is exactly what is done for English in the Historical Thesaurus of the Oxford English Dictionary (Kay et al. 2009, searchable and in a more user-friendly format on the website www.oed.com/thesaurus). For example, under the term ‘animal’ the reader of the Historical Thesaurus is able to trace the rise and fall of terms in English including neten, deer, wight, beast, creature, all of which have taken on the generic meaning at some time. Compilation of the English Historical Thesaurus took over forty-five years and a large team of researchers to complete, and is itself the object of research and popular accounts: see Crystal 2014 for one work describing explorations of English semantics through the thesaurus. Even in this age of large research grants providing the means to employ teams of scholars, it is very unlikely that such a task will ever be undertaken for the ancient Greek language. In the absence of any imminent publication of a Historical Thesaurus of Greek, the LSJ entry for ζῷον suggests that something might be gained from looking at the history of a concept through the lexicon, and particularly to show the rise and fall of different categorizations. The Ancient Greek case is not just of interest to lexicographers and classicists: linguistic ethnobotanists have proposed that a number of different human societies have arrived at similar classifications of the natural world (see Brown 2000), and the history of Greek terms may be seen as a contribution to this wider debate. The Greek vocabulary is recorded for over a three-thousand-year time-span, and through

290

James Clackson

reconstruction of the immediate ancestor of Greek, Proto-Indo-European, it is possible to go back further still. Examination of the different classificatory terms in Greek, with their ancestry, thus allows us to test some of the hypotheses proposed by linguistic ethnobotanists. One such hypothesis concerns the effect on the terms for animals made by the development from a hunter-gatherer to a sedentary, agricultural lifestyle: Brown (2000, 68) claims that taxonomies ‘of hunter gatherers tend to only have only one level, consisting entirely of generic classes’, and that over time ‘folk taxonomies have tended to expand up and down, adding more inclusive life-form and less inclusive specific classes to pre-existing generic categories’. In the conclusion to this chapter I shall sketch out the sort of contribution Greek can make to such debates, but here it should be noted that this is not the place for a full investigation of the Greek classifications of animals, including the development of Aristotelian categories of animal classification; the interested reader is referred to the excellent discussion of these matters by Zucker 2005. In my survey of the Greek terms, I shall start with the earliest Greek records, the Mycenaean archives preserved on tablets written in the Linear B script from the second millennium. In the Mycenaean archives there is no occasion to use a generic term for ‘animal’, but one word used by the Linear B scribes at Pylos in a formulaic expression might give an insight into an early categorical scheme, particularly as it has a clear etymology with cognates in other IndoEuropean languages. Furthermore, since the Albanian cognate, shtazë, actually means ‘animal’, we should therefore consider the possibility that this meaning is inherited also into Greek. The Mycenaean word in question is qe-to-ro-po-pi, a noun generally understood to be in the instrumental case, reconstructed as *κwετρόποπφι. It occurs several times in the following formula: A (nominative) o-pi B (genitive) qe-to-ro-po-pi o-ro-me-no Here A and B stand for men’s names, with the meaning ‘A watching over the quadrupeds of B’ (Aura Jorro 1985–93 II. 203). qe-to-ro-po-pi is obviously a compound of the numeral ‘four’ and the word for leg or foot, and it can consequently be translated ‘quadruped’. Context and etymology suggest that this is a generic term covering live-stock, in this case most likely sheep and goats. Later Greek also knows the compound τετράπους with the meaning ‘four-footed’, and a neuter τετράπουν (most often in the plural τετράποδα) is attested in Herodotus, Plato, and later authors also to mean ‘quadruped’. The use of the term for livestock or for animals in general is also continued in dialect inscriptions: note in particular τετράπος from Gortyn IC 4.41 col. 3. lines 7–8, where it is paired with ὄνν[ι]θα ‘bird’, and Arcadian τετραπόδων πάντων opposed to ὑποπτέρων ‘winged creatures’ at IG V 2 159–60. The word τετράπος must also lie behind the division in the longest Gortyn lawcode (IC 4.71 col. IV 35–6) between τὰ πρόβατα καὶ τὰ καρταίποδα ‘the small and large cattle’, with καρταίποδα meaning literally something like ‘sturdy-footed’.

LSJ and the Diachronic Taxonomy of the Greek Vocabulary

291

Both elements of Greek τετράπους are inherited terms, with cognates in most branches of Indo-European. Reflexes of the same compound, reconstructed *kwet(w)r-pod-, are found also in other Indo-European languages, including Albanian shtazë ‘animal’ mentioned above. More interesting for the purposes of this chapter are the terms in the Italic and Indo-Iranian branches of the Indo-European language family, particularly Umbrian peturpursus and Sanskrit cátuṣpad. The Umbrian language is known principally from a set of seven bronze tables surviving in the Italian town of Gubbio (ancient Iguvium), which record ritual activities and utterances of a priestly college, many of which are expressed in formulaic language. In one prayer on these tables, the sacrifice is offered dupursus peturpursus ‘for the bipeds and the quadrupeds’ of the town. Since the nineteenth century, scholars have noticed that there is an exactly cognate phrase surviving in the Vedic hymns: dvipáde cátuṣpade ‘for the biped and quadruped’ (see Watkins 1979, 275 and 1995, 15 for references to texts and earlier scholarship). In Sanskrit texts these two terms—biped, referring to humans, and quadruped, referring principally to stock animals—can together form a division of a superordinate term paśu- which denotes mobile property and wealth (on these terms and the Indo-European background to them see Watkins 1979; Clackson 2007, 207–9). Watkins has argued that the Mycenaean Greek qe-to-ro-po-pi and dialectal and later Greek τετράπους can be fitted into the same schema of categorization of humans and livestock that is found in Umbrian and Sanskrit, as well as attested in other Indo-European traditions, principally Hittite. In an illuminating passage (1979, 286), Watkins builds up a taxonomy of wealth in early Greek on the basis of different passages of the Iliad, incorporating the Mycenaean word qe-to-ro-po-pi (since τετράπους does not occur in this sense in Greek). I repeat a section of his schema with some adaptations below in Fig. 16.1. The highest-order term in Fig. 16.1, Homeric πρόβασις, is also a much discussed term; here it will suffice to say that it is taken to denote mobile wealth as a whole, divided between animals and slaves. The Greek word for ‘slave’ ἀνδρόποδον, itself probably a back-formation from a plural ἀνδρόποδα (see Beekes 2010, 101), can add some weight to Watkins’ reconstructed parallel between ἀνδραπόδεσσι and Mycenaean qe-to-ro-po-pi, even though άνδραπόδεσσι

(ίπποι)

(πρόβασις) βόεσσι (Mycenaean qe-to-ro-po-pi)

(βόες) πρόβατα

Fig. 16.1. Watkins’ schema for moveable property in Early Greek.

292

James Clackson

these two words (or their later forms) do not appear together in any surviving Greek text known to me. Etymologically, ἀνδρόποδον means ‘human-footed’, but Greek slaves were not distinguished from the free by the fact that they had human feet. The best way to explain the existence of this curious compound is as a replacement of an earlier form such as δίπους with the more transparent stem ἀνδρο-, much as the dialectal form καρταίποδα ‘larger cattle’ mentioned above reflects τετράποδα. The reconstruction of a pair ‘two-footed’ (human slaves) and ‘four-footed’ (livestock animals) for Greek would then fit in with the same contrast as found in the other Indo-European languages. It should be noted that in later Greek the terms τετράπους and δίπους were used to refer to four- and two-legged items of furniture or pots, and also sometimes to denote measures of length, which may be connected to the loss of meaning ‘slave’ and ‘livestock animal’. Note also that, according to Watkins, ‘livestock’ is itself divided into larger (cows and horses) and smaller animals (sheep and goats), even though in Homer the word βοῦς does double duty to mean both ‘cow’ and its immediate hyperonym ‘larger livestock animal’. In most dialects of later Greek the term for ‘livestock’ is replaced by the cover term πρόβατα, the standard term in Homer, and only attested there in the plural. Benveniste 1969 tried to show that the word in Homer is not a general term for cattle, but instead a word for property in general, which then became restricted in use to cattle alone. In support of this Benveniste (1969, but already 1949) argued against the traditional etymology (accepted in LSJ s.v., and deriving ultimately from the ancient scholia, see citations at LfrgE s.v.) that the term originally referred to smaller livestock which went ahead of the other animals in mixed herds. Instead, Benveniste argued that the term is derived from προβαίνω which means little more than ‘go forward’, giving parallels for the use of a term ‘movable (property)’ for sheep or livestock from other languages. Benveniste’s theory works better for Homeric πρόβασις than πρόβατα, but is nevertheless a most plausible account for the latter word (Beekes 2010, 1235). When Homer or others writing in the epic dialect want to refer to the entire living world, they had the option to employ a structure which refers to birds of the air, land-animals and sea-creatures, such as Od. 24.291–2: ἠέ που ἐν πόντῳ φάγον ἰχθύες, ἢ ἐπὶ χέρσου θηρσὶ καὶ οἰωνοῖσιν ἕλωρ γένετ᾽ . . . Maybe the fish in the sea have eaten him or he has become food for the animals on the land or for the birds . . .

or Hesiod Op. 276–7 τόνδε γὰρ ἀνθρώποισι νόμον διέταξε Κρονίων ἰχθύσι μὲν καὶ θηρσὶ καὶ οἰωνοῖς πετεηνοῖς For the son of Cronos has laid down this law for men, that for fish and beasts and winged birds . . .

LSJ and the Diachronic Taxonomy of the Greek Vocabulary

293

or the Homeric Hymn to Aphrodite 3–6 καί τ᾽ ἐδαμάσσατο φῦλα καταθνητῶν ἀνθρώπων οἰωνούς τε διιπετέας καὶ θηρία πάντα, ἠμὲν ὅσ᾽ ἤπειρος πολλὰ τρέφει ἠδ᾽ ὅσα πόντος And she subdues the tribes of mortal men, the fast flying birds and all the wild animals, as many as the land nurtures and as many as the sea . . .

(The first two of these passages are cited in LSJ s.v. θήρ, the third s.v. θηρίον; all three passages are discussed at Zucker 2005, 25.) In these examples, humans are seen as a separate category to beasts of the air, the sea, and the land, rather than as sharing any characteristic with animals. In the Odyssey passage, birds, fish, and wild beasts are seen as potential threats to man, but the Hesiod and Homeric Hymn passages list the three categories as encompassing the natural world as subordinate to man and also god. Even so, it is questionable whether these collections of terms are really underpinned by a coherent taxonomy of living creatures. Zucker 2005, 28–30 argues against the thesis of Körner (1917) that saw Homer as prefiguring a classificatory system such as is found later in Aristotle. Zucker is surely right to say that the terms θήρ and θηρίον should be seen as ‘titles’ or ‘surnames’ when used in apposition to λέων ‘lion’ or ἔλαφος ‘deer’, and do not function as taxonomic terms. However, both θήρ and θηρίον are applied in Homer only to non-domesticated animals, showing that there is some categorization of ‘tame’ versus ‘wild’ at work, as we have already seen in the discussion above. It is probably best to see the epic division of animals of the air, of land, and sea, as an expansion of the merism of the different parts of world, air, land, and sea, which is depicted, for example, on the shield of Achilles: ἐν μὲν γαῖαν ἔτευξ᾽, ἐν δ᾽ οὐρανόν, ἐν δὲ θάλασσαν (Iliad 18.483) Thereupon he fashioned the earth, thereupon the heaven, thereupon the sea.

Later Greek authors continue this broad division of animals into three types according to their habitat. Note for example Empedocles (Fr. B 21, 9–12), who shows the same framework for describing the created world, here with the addition of plants and gods as well. ἐκ τούτων γὰρ πάνθ' ὅσα τ' ἦν ὅσα τ' ἔστι καὶ ἔσται, δένδρεά τ' ἐβλάστησε καὶ ἀνέρες ἠδὲ γυναῖκες, θῆρές τ’ οἰωνοί τε καὶ ὑδατοθρέμμονες ἰχθῦς, καί τε θεοὶ δολιχαίωνες τιμῆισι ϕέριστοι. For from these comes to life everything that has been, that is, and that will be, trees, and men and women, beats and birds and fish that live in water, and long lived gods, who are the most worthy of worship.

In Empedocles the plural θῆρες is used for land animals only, in opposition to trees, birds, fish, gods, and men and women.

294

James Clackson

As we have seen in the Homeric Hymn to Aphrodite, θηρίον can have a wider application than θήρ, and be used both for land animals and fish. The word θηρίον approaches close to a generic term for animal not including humans in classical authors as well. Thus in Herodotus, θηρίον is applied to sharks (6.44), in the comic poet Antiphanes to eels (147.7) and in Aristotle more generally to fish (HA 598b1). In a passage of the Politics just after the ‘man is a political animal’ section, Aristotle uses the phrase ἢ θηρίον ἢ θεός (Pol.1253a29), which LSJ take to be a proverbial expression for referring to forms of life conceptualized either as ‘above’ or ‘below’ the nature of humankind. In only one passage known to me is the semantic range of Greek θηρίον taken even further, and reckoned to include ‘humans’ as well—I discount here other passages where men or women are downgraded to the status of beasts as terms of abuse or opprobrium (in the LSJ entry these are classed together as sense III ‘as a term of reproach’). The passage in question is a fragment of the Sophist Antiphon (48, Diels-Kranz), as reported by Photius: ἄνθρωπος πάντων θηρίων θεειδέστατον

‘a human is the most godlike of creatures’. If this is not a deliberate play on the notion that humans stand in the chain of creation between animals and gods, it would imply that Antiphon is aware of a different way of conceptualizing the animal world, with humans as members of the animal kingdom, rather than apart. It is certainly the case that Greek thinkers in the fifth century BC developed a new conception of the animal kingdom which is not previously found. This is the idea that all living things, including humans and animals (and sometimes also plants) share ψυχή ‘life’ which can be seen in evidence in a fragment of Anaxagoras (B278, cited in Zucker 2005, 47): ἀνθρώπους . . . καὶ τὰ ἄλλα ζῷα ὅσα ψυχὴν ἔχει. humans . . . and all other creatures that have life

Antiphon, in the passage cited above, is exceptional in using θηρίον to refer to the collective of humans and other creatures which share life. In other authors, the term ζῷον is used in this sense. The history of this word is relevant. Mallory and Adams (1997, 23) see ζῷον as an inherited term from IndoEuropean, comparing the Tocharian B word for ‘sheep’, śaiyye, to arrive at an Indo-European lexeme *gwyéh₃wyom meaning ‘animal, living being’. This view of the antiquity of the word might be thought to receive some support from the attestation of forms zo-wo and zo-wi-jo in Mycenaean. However, the Mycenaean terms are men’s names, and have nothing to do with animals, and their derivation must remain uncertain (Aura Jorro 1985–93 II.460). The first attestation of ζῷον is in fact later in Greek; it is absent from epic (although some texts of Hesiod print ζῴοισιν at Th. 584, it is better to keep the reading of

LSJ and the Diachronic Taxonomy of the Greek Vocabulary

295

the manuscripts, ζωοῖσιν ‘living (things)’, see West 1966, 328). The word first occurs in fragment 11 (Diehl) of Semonides: τὸ δ᾿ ἥμιν ἑρπετὸν παρέπτατο τὸ ζωΐων κάκιστον ἔκτηται βίον. The beetle (ἑρπετόν) has flown to us, which has the worst life of all creatures . . .

where a ἑρπετόν (on which see below) is judged as the worst of all ζῷα. It seems best to follow Leumann (1945, 7) in taking ζῷον (which is a contracted form of Semonides’ ζωΐον), as a formation within the history of Greek. Leumann sees the word as a derivative of an adjective ζώς ‘alive’, much as θηρίον relates to θήρ; the absence ζῷον from Homer can be thus explained as reflecting the later formation of the word. The meaning of ζῷον itself undergoes some fluctuation in its early attestations (see Zucker 2005, 47–8). Democritus always uses the term to mean anything living (including plants) and that sense is also found in Empedocles; but Empedocles is also able to oppose plants and men to ζῷα, implying a more limited reference. In the Timaeus, Plato employs ζῷον to refer to a wide spectrum of creation, including divinities (39e) and plants (77a), although this is in the context of the account of the origin of the universe, in which ‘life’ is parcelled out and repackaged by the Craftsman. Plato seems to defend his wider definition of ζῷον in the Timaeus (77b): πᾶν ὅ τι περ ἂν μετάσχῃ τοῦ ζῆν ζῷον ἂν λέγοιτο. In other works of Plato the word ζῷον is used to cover humans and other animals in opposition to plants, for example at Phaedo 70d: μὴ τοίνυν κατ᾽ ἀνθρώπων, ἦ δ᾽ ὅς, σκόπει μόνον τοῦτο, εἰ βούλει ῥᾷον μαθεῖν, ἀλλὰ καὶ κατὰ ζῴων πάντων καὶ φυτῶν ‘So don’t just think about this issue’ he said, ‘as it regards humans, if you want to understand it easily, but also consider it with regard to all animals and plants.’

An anecdote preserved by Diogenes Laertius (6.6.41) gives a definition attributed to Plato, Ἄνθρωπός ἐστι ζῷον δίπουν ἄπτερον ‘a human is a twolegged animal without feathers’, in response to which the cynic Diogenes plucks a chicken. This story may belong with other school-room jokes concerning Diogenes the cynic rather than have any validity as a historical fact; one can compare the collection of chreiae concerning Diogenes surviving on papyri, collected at Adorno et al 2015, 2.363–425, one of which (CHR 9) may also involve the definition of ἄνθρωπος. It does, however, serve as an illustration of the continued broad sense of ζῷον, subsuming humans, in the philosophical and educated tradition. After Plato there also seems in general to be no dispute that plants are not considered to belong to the class of ζῷα. Aristotle’s History of Animals includes all the creatures that we would consider as animals and no plants, and the word is still in use in the same sense in Modern Greek.

296

James Clackson

It is worth here considering a few other terms which have wide taxonomic reference in Greek. First, ἑρπετόν, which is clearly derived from the verb ἔρπω, and thus naturally has a tendency to be applied to creeping things such as snakes, reptiles and insects. The Semonides fragment, cited above, seems to be referring to a specific creature such as a beetle. In general in early Greek the word has a wider application, although always with a connotation of something to be looked down upon or perhaps feared. For example, in the description of Proteus which Eidothea gives to Menelaos in the Odyssey, ἑρπετόν refers to the different shapes Proteus will assume: ὅσσ᾽ ἐπὶ γαῖαν/ἑρπετὰ γίγνονται (Od. 4.417–18); and the word occurs in Herodotus’s description of the way the Magi kill all sorts of living creatures, unlike Egyptian priests, with their bare hands: οἱ δὲ δὴ Μάγοι αὐτοχειρίῃ πάντα πλὴν κυνὸς καὶ ἀνθρώπου κτείνουσι, καὶ ἀγώνισμα μέγα τοῦτο ποιεῦνται, κτείνοντες ὁμοίως μύρμηκάς τε καὶ ὄφις καὶ τἆλλα ἑρπετὰ καὶ πετεινά. (1.140) The Magi, to be sure, kill every creature except a dog or a man with their bare hands, and they make a great show of it, killing alike ants and snakes and all the other creepy crawlies and flying things.

Note here how the full context is more revealing than LSJ’s simple ‘opp. πετεινά’. A term with wider reference is κνώδαλον, which is not found in prose. There are no certain cognates to κνώδαλον outside Greek, but the connection to κνώδαξ ‘pin or pivot’ and κνώδων ‘projecting teeth (as of a spear)’ seems unproblematic, and suggests that the word initially referred to animals that had the ability to bite or wound (Beekes 2010, 726–7). In Homer and Hesiod κνώδαλον can be used for any wild creature of land or sea: κνώδαλ᾽ ὅσ᾽ ἤπειρος πολλὰ τρέφει ἠδὲ θάλασσα (Hesiod Th. 582), of the animals depicted on Pandora’s headband. In one hexameter passage from the end of the sixth century, the Homeric Hymn to Hermes (188), κνώδαλον appears as a beast that Hermes finds being shepherded by an old man, and thus presumably refers to a domesticated animal such as an ox or donkey: ἔνθα γέροντα / κνώδαλον εὗρε νέμοντα παρ᾽ ἔξοδον ἕρκεος αὐλῆς There he found an old man tending a beast beside the exit from the fenced courtyard.

LfrgE argues that the text must be corrupt, since the confusion between terms referring to wild and tame animals is elsewhere avoided in epic (and LfrgE also dismisses any interpretation of the passage in which γέροντα and κνώδαλον agree). But it is possible that the author of the Hymn to Hermes was already aware of notions that both tame and domesticated animals can be categorized together as ζῷα, and chose the more high-sounding word κνώδαλον as suited

LSJ and the Diachronic Taxonomy of the Greek Vocabulary

297

to the genre. Pindar (P. 10.36) also refers to asses as κνώδαλα. The suggestion of Gildersleeve (1885, 354) that the poet has in mind the word κνάω ‘gnaw’, because ‘[a]sses gnaw each other in their play’, meets with the objection that κνάω means ‘scratch’ rather than ‘gnaw’. In Aeschylus’ Choephoroe (601) the chorus use the merism κνωδάλων καὶ βροτῶν to describe the creatures over whom ἔρως holds sway. The distinction between tame and domesticated can still be called upon in the fifth century, however; in Prometheus (462) κνώδαλον is the word chosen to describe wild animals when they are first domesticated by Prometheus: ἔζευξα πρῶτος ἐν ζυγοῖσι κνώδαλα. Less common, and also mainly restricted to poetry, is the word βοτόν, a derivative of βόσκω ‘feed’ or ‘graze’, glossed as ‘beast’ by LSJ, and applied both to domesticated grazing animals and also occasionally birds. Another word for domesticated herd-animals is κτῆνος, derived from the verb meaning ‘possess’, κτάομαι, with the same semantic development from possessions to cattle as discussed above. LSJ claim that at Heraclitus fragment 29 (Diels-Krantz) the word is used of beasts in general, but the context argues against this. Heraclitus contrasts the ἄριστοι, who choose one thing, fame, above everything else to οἱ δὲ πολλοὶ κεκόρηνται ὅκωσπερ κτήνεα. It seems likely that he has in mind here the flock animals which are content when finished with grazing, rather than a wider context (note the translation given by Kirk, Raven, and Schofield 1983, 211, ‘The majority are glutted like cattle’). θρέμμα is a word properly used of the young of any animals, but can have a wider extension to animals or humans more generally, particularly when a creature is described as the offspring or native of a place, and sometimes φύσις can be used also to describe creatures, such as Sophocles’ πόντου εἰναλία φύσις (Ant. 345, lyric). Finally, the language of the Septuagint and the Greek New Testament has two terms κτίσις and κτίσμα which can both be translated as ‘creature’, and may be used to refer to animals. For example, at Rev. 5.13: καὶ πᾶν κτίσμα ὃ ἐν τῷ οὐρανῷ καὶ ἐπὶ τῆς γῆς καὶ ὑποκάτω τῆς γῆς καὶ ἐπὶ τῆς θαλάσσης καὶ τὰ ἐν αὐτοὶς πάντα ἤκουσα λέγοντας . . . And every creature which is in heaven, and on the earth, and under the earth, and such as are in the sea, and all that are in them, heard I saying . . . (King James Version)

These meanings of κτίσις and κτίσμα are specific to biblical texts, and rely on the fact that κτίζειν is the preferred translation term in most of the Septuagint for Hebrew bārā’ ‘create’. κτίσις and κτίσμα, like English creature, emphasize living things as creations of God, but the sense may also be wider than living things and refer to anything that is created. This short survey is enough to show that the ways in which animals as a broad class were conceptualized changed greatly over the history of Greek. In the earliest texts, the division between domesticated and wild animals looms larger than any sense that living creatures as a whole share anything in

298

James Clackson

common. If we are right in the suggestion that there was an early, possibly prehistoric Greek association between two-footed property (slaves) and fourfooted (domestic animals), then the division between owned and free, domestic and wild, may have originally been as salient, if not more so, than the division between human and beast. In Homer and other surviving early works of Greek poetry, the conception of the world of living things seems to have placed a priority on the separation between humankind and everything else. Even so, the overarching categories of ‘domestic’ and ‘wild’ remain in place as a possible framework for viewing the animal world in later centuries. By the fifth century, a different conceptualization of the natural world can be traced, in which all animals belong together, and humans are included amongst them. In later works, the Judaeo-Christian concept of animals as creatures of God is also reflected in the lexicon. The survey of the Greek evidence also casts doubt on some of the suggestions from ethnolinguistic studies about universals of folk taxonomy, and a possible diachronic shift between a hunter-gatherer and agricultural classification of animals. There is an observable change in the ways in which the Greek lexicon categorizes fauna, but it does not appear to tally closely with social change. For the period under consideration in this chapter, most speakers of Greek lived in agrarian societies with a greater or lesser degree of human slavery. Changes in the intellectual climate and religious belief seem to have been important elements in the way in which the world was represented within the Greek lexicon. A putative Historical Thesaurus of Greek would no doubt add further examples through which it would be possible to track the way in which the Greeks’ view of the world developed and mutated over time, and indeed, such a thesaurus may present this information more straightforwardly than can be done here.

17 Literary Lexicography Aims and Principles Michael Silk

How should a historical dictionary of a dead language¹ deal with literary, especially poetic, language? My chapter will attempt to clarify the issues and set out some principles for ‘literary lexicography’, with special reference to LSJ and ancient Greek poetry,² and Greek usage in the early and classical periods.³ The issues that concern me apply equally to LSJ and the Revised Supplement,⁴ and for the most part my discussion will subsume both. In what follows, unless the reference is explicitly to one and not the other, ‘LSJ’ should be taken to refer to the Revised Supplement as well.

17.1. ELEVATION AND HEIGHTENING Poetic language in the Western traditions subsumes two distinct categories of usage: (a) elevation (whereby usage conforms to a conventional ‘high style’) and (b) heightening (whereby meaning is enriched, often by mechanisms of defamiliarization).⁵ For lexicographical purposes, whether in the context of ¹ ‘A language . . . no longer used as a natural daily means of spoken communication within a community’: Crystal 1999, 80. ² The literary lexicography of ancient Greek has been among my active interests for many years, and the present chapter draws freely on (esp.) Silk 1974 and 1983. The phrase ‘literary lexicography’ itself I first used in print, in connection with LSJ and ancient Greek, in Silk 1974, xiii. ³ In this chapter, accordingly, dates and centuries are , unless specified as . ⁴ In these notes, two specialist lexica are referred to in abbreviated form: Ebeling = H. Ebeling, Lexicon Homericum, 1885; Kühn Fleischer = J. H. Kühn and U. Fleischer, Index Hippocrati cus, 1989. ⁵ See, e.g. Silk 2014, 434 9. Michael Silk, Literary Lexicography: Aims and Principles In: Liddell and Scott: The History, Methodology, and Languages of the World’s Leading Lexicon of Ancient Greek. Edited by: Christopher Stray, Michael Clarke, and Joshua T. Katz, Oxford University Press (2019). © Oxford University Press. DOI: 10.1093/oso/9780198810803.003.0017

300

Michael Silk

ancient Greek or (for instance) modern English, various manifestations of both categories are irrelevant,⁶ and in practice the categories resolve into: (a) conventional high-style vocabulary or grammatical forms and (b) innovative tropes (specifically, innovative metaphor and metonymy)⁷ along with certain other mechanisms, from activation of connotations (most obviously in puns) to (in Greek) ‘iconymy’.⁸ For lexicographical purposes, both (a) and (b) are significant, but (a) is unproblematic in principle, whereas (b) is problematic per se. Conventional high-style forms or vocabulary are usually easy to recognize, whether ubiquitous, as in most Greek verse, or rare and out of fashion, as in modern English verse. Such forms and such vocabulary occur only or largely or primarily in verse,⁹ which usually means the more ‘serious’ types of verse. In current English usage, the noun ‘mortal(s)’ (unlike, e.g. ‘person’/‘people’/ ‘human being(s)’) exists largely as high-style verse usage from the past.¹⁰ One associates it with past poetry (‘Lord, what fools these mortals be!’),¹¹ even if one meets it—exceptionally—in a relatively recent and largely unelevated context (‘You’ve opened heaven’s portal | Here on earth for this poor mortal.’).¹² In Greek usage, likewise, βροτοί, ‘mortals’ (unlike, e.g. (οἱ) ἄνθρωποι, ‘human beings’/‘people’) carries elevating connotations, because in early and classical Greek the word is common in—but only common in—‘serious’ verse, even in the mouth of a ‘low’ speaker, like Phaedra’s nurse in Euripides’ Hippolytus (σὺν πολλοῖς βροτῶν: ‘along with many mortals’).¹³ The same speaker uses the simplex grammatical form ὄλλυμι (ἔρωτος οὕνεκα ψυχὴν ὀλεῖς;—‘will you stroy

⁶ E.g.: for (a) in Greek, omission of the definite article with nouns; for (a) in English, inversion of noun and adjective (‘Ye myrtles brown’: Milton, ‘Lycidas’, 2); for (b) in either language, aural effects with enactmental outcomes. ⁷ Where ‘metonymy’ subsumes all the contiguity tropes (synecdoche, enallage et al.): Silk 2003a, 124, 132 4 (following Jakobson), and, comprehensively, Matzner 2016. Formal defin itions of metaphor and metonymy: Silk 2003a, 124, 132, cf. Silk 2012. ‘Innovative’: n. 23 below. ⁸ See pp. 318 26 below. ⁹ Unless also, exceptionally, in high style prose (cf. nn. 19, 35, 56, and p. 309 below, on Plato) or e.g. in parodic comedy. In English, such vocabulary may occur in religious usage, in the tradition of the King James Bible and the Book of Common Prayer (p. 301 below, with n. 17). ¹⁰ Though also in fossil phrases (‘a mere mortal’) and, in recent years, in quasi grandiloquent science fantasy idiom. ¹¹ Shakespeare, A Midsummer’s Night’s Dream, 3.2. ¹² From ‘You Are My Lucky Star’, a 1935 popular song (lyrics by Arthur Freed), where the elevation impinges as tongue in cheek (cf. the mock pomposity of ‘heaven’s portal’). In many cases, elevated language is in historical fact linguistic archaism, as with forms like ‘thou hast’ in English, or as with the omission of the article in Greek (n. 6 above), but English examples like ‘mortals’ or noun/adjective inversions (ibid.) show that this need not be so (both examples are Latinisms or Gallicisms, rather than archaisms). ¹³ E. Hipp. 439. βροτός in prose symptomatically in Plato, R. 566d though also ( faute de mieux!) in discussion of human properties at Arist. Top. 133a 31 2, 137a 35 7 (and an unconvincing conjecture by Littré at Hp. Aph. 3.12).

Literary Lexicography: Aims and Principles

301

your life because of love?’),¹⁴ with the same elevating connotations, where the non-elevated Attic norm would be the compounded ἀπόλλυμι.¹⁵ In the midVictorian era, ‘thou hast’ (instead of ‘you have’), as used by the poet Gerard Manley Hopkins, carries comparable connotations—‘Thou hast bound bones and veins in me’¹⁶—except that in context the phrase also evokes the conventional traditions of biblical English (‘Thou’ is God), in addition to those of past poetry.¹⁷ All this is straightforward, not only for a reader, but for a lexicographer, who can readily label such usages as ‘verse word’, ‘verse form’, ‘predominantly verse word’, or whatever.¹⁸ Heightened usage, though also largely found in verse,¹⁹ is different in kind. Under this heading, one subsumes a range of unconventional usages: creative usages whose breach of convention is their precondition. Such usages challenge the lexicographer. From a lexicographical standpoint, there is no difficulty in dealing with established usages like (in English) ‘chair leg’ or ‘trouser leg’ (which it is not helpful—to anticipate—to designate ‘metaphor’ and ‘metonymy’). But how, if at all, should a lexicographer deal with creative innovations like (in English) Shakespeare’s ‘I have immortal longings’, or Keats’s ‘flock in woolly fold’, or Eliot’s ‘the river’s tent’?²⁰ In a living language, like English, there is a practical case for avoiding citation of, or comment on, such usages altogether,²¹ but in the case of ancient Greek the practicalities are different, and the need to deal with equivalent usages unavoidable.

17.2. NORM AND DEVIATION Traditional understanding of heightened usage depends on the model of norm and deviation, and rightly so. There are established usages (as ‘chair leg’ or ‘trouser leg’ or οἱ ἄνθρωποι:people or βροτοί:mortals or ‘thou hast’ or ὄλλυμι: (de)stroy are all established usages), and heightened usage deviates from any of them. Elevated usage, in particular, is a norm—a customary norm in ancient ¹⁴ E. Hipp. 440. ¹⁵ Though (pace LSJ) the simplex verb is not confined to verse usage: p. 328 below. ¹⁶ ‘The Wreck of the Deutschland’, stanza 1. ¹⁷ I.e. the usage evokes (e.g.) ‘Behold, thou hast driven me out’ (Genesis 4.14, in the King James translation: Cain to the Lord God), alongside (e.g.) ‘Thou hast nor youth nor age’ (Shakespeare, Measure for Measure, 3.1). ¹⁸ Ideally, as part of a systematic identification of registers and generic norms. ‘Verse word’ is preferable to ‘poet.’ (LSJ’s formula), because prose can be poetic (n. 19 below). ¹⁹ But heightened usage may also occur in ‘poetic’ prose, like the fragments of the Presocratic Heraclitus, or occasionally in non poetic prose, like the later prose of James Joyce. ²⁰ Respectively: Antony and Cleopatra, 5.2; ‘The Eve of St Agnes’, stanza 1; The Waste Land, III (173). ²¹ Cf. R. Fowler 2004, and n. 50 below.

302

Michael Silk

Greek verse, though a kind of residual norm in modern English religious usage (in the Christian traditions) and a kind of superannuated norm in modern English verse—and, as such, is not in this sense deviant.²² Conversely, innovative tropical usage is, by definition, deviant.²³ The ancient theory of tropes assumes this definitional principle. Quintilian’s formula is representative: ‘tropus est verbi vel sermonis a propria significatione in aliam cum virtute mutatio’—‘a trope is the successful change of a word or sequence from its proper meaning to another’.²⁴ Here, ‘successful’ is open to challenge (tropes can be unsuccessful), and ‘change’ too (tropes assume the ‘proper meaning’ as a point of departure), while the notion of ‘proper’ meaning (especially if taken to mean a single ‘proper’ meaning) is itself infelicitous—but the underlying assumption that a word (or sequence) commonly has one or more established usages, normal usages, from which tropical usage of that word (or sequence) deviates, is sound.²⁵ Pace much loose thinking in recent decades, in particular,²⁶ normal and metaphorical/metonymic usage should be regarded as mutually exclusive categories—and, in some respects, even talk of ‘dead metaphors’ (or ‘dead metonyms’) only serves to confuse the issue.²⁷ ‘Normal usage’, in fact, subsumes any ‘established usage of a word or phrase such as you would expect to find in a compendious dictionary of current and currently available usage’—as (e.g.) ‘mortal/s’ is currently available to a user of English, even if not often actively employed.²⁸ Properly and correctly, then, a Greek lexicon will be expected to house, not only lexemes like ἄνθρωπος, but also lexemes like βροτός; and an English dictionary like OED lexemes like ‘mortal/s’, alongside usages like ‘chair leg’

²² Elevated vocabulary (etc.) in a non elevated context generally creates a perceptible, but only mildly dislocating, switch of tone. Contrast the fundamental, defamiliarizing shock of creative deviance. ²³ Sc. more or less innovative. There are, obviously, degrees of innovativity, and the effectively innovative need not be absolutely innovative though, if reused often enough, a literary/poetic innovation eventually becomes an out and out cliché and in effect ‘normal’ itself. See further n. 44 below. ²⁴ Inst.Or. 8.6.1. ²⁵ In modern times, the norm/deviation model has been challenged by some theorists, from Mary Louise Pratt to W.S. Graham, but the model survives: cf. e.g. Jones 2012, 116 31 and Benjamin 2012, eccentric though the latter discussion in some ways is. The model has been profitably restated across a range of modern theorizing (esp. in literary theory), from Eliot to Bakhtin, from Gerard Manley Hopkins to Paul Valéry, from Mukarovsky to Manfred Bierwisch. Cognitive linguists have tended to confuse the issue by an exclusive focus on implicit analogies in precisely non deviant usage: ‘One of the prime objectives of cognitive linguistics has been to show that metaphor is not . . . deviant’ (Traugott and Dasher 2005, 76) cf. p. 307 with n. 47 below. ²⁶ See nn. 25 (above) and 48 (below). ²⁷ ‘Dead metaphors are no longer metaphors’: Ricoeur 1978, 290. ²⁸ Silk 2003a, 122. Robson 2006, 113 reasonably objects to the use of the phrase ‘normal usage’ for what, in a given context or register, may be ‘unusual language’. For my purposes, the phrase can stand.

Literary Lexicography: Aims and Principles

303

and ‘trouser leg’. And they do. We do not necessarily expect OED to include usages like Shakespeare’s ‘immortal longings’ or Keats’s ‘woolly fold’ or Eliot’s ‘river’s tent’²⁹—but we certainly expect OED to encompass all normal uses of ‘immortal’ and ‘woolly’ and ‘tent’ from which a reader might make sense of these tropical usages in their particular poetic contexts. In this way, a reader would be helped by the dictionary to see that Shakespeare’s ‘immortal longings’ (Cleopatra’s phrase, as she prepares to die by the asp) metaphorically implies boundless, but also impossible, and then metonymically (enforcing that second implication) suggests ‘longings of immortality’—connotatively, then, evoking the queen’s aspiration to defy the human condition (‘I am fire and air; my other elements | I give to baser life’).³⁰ Or that it is Keats’s ‘sheep’ that are literally ‘woolly’ (‘And frozen was the flock in woolly fold’), with woolliness transferred, metonymically, to the fold. Or that Eliot’s ‘tent’ (‘The river’s tent is broken; the last fingers of leaf | Clutch and sink into the wet bank’) refers, metaphorically, to trees, thick with leaves, overhanging a river, with connotations of protective Nature (tents provide shelter). In the case of ancient Greek, and a lexicon like LSJ, we all surely start with corresponding expectations: the presentation of normal usage is paramount. As John Chadwick put it, with LSJ in mind: ‘it is the function of the lexicographer to record how the vocabulary of a language is normally used’.³¹ However, Chadwick’s discussion—representatively?—ignores two crucial issues: first, the implications of the distinctive distribution of lexical evidence for ancient Greek in general and earlier Greek in particular; and, secondly, the problems involved in establishing normality in a dead language in any case. Let us take the points in order.

17.3. THE L EXICAL EVIDENCE FOR ANCIENT GREEK The lexicographer of ancient Greek, bent on recording ‘normal usage’, is faced with the fact that the usages in a range of important sources are intermittently deviant. The relevant facts are easily stated. Quite apart from whatever spoken usages in different periods of Greek, or different parts of the Greek-speaking world, never found their way into writing, our attested, written, samples of ancient Greek constitute only a small fraction of what once existed in written form. Even with the most celebrated literary genres, this is the case. Out of (maybe) three thousand different plays produced at the Athenian dramatic ²⁹ ‘There are no metaphors in dictionaries’: Ricoeur 1978, 170. In fact, OED cites the Shakespeare and the Eliot: n. 50 below. Here and elsewhere ‘OED’ refers to OED² (1989). ³⁰ Antony and Cleopatra, 5.2. ³¹ Chadwick 1996, 16.

304

Michael Silk

festivals between the late sixth and the late fourth centuries, we possess only forty-five more or less complete specimens, along with fragments of others. Most of—even—the known works of Greek literature are effectively lost. But what survives has not, in most cases, survived randomly. Our extant samples of ancient Greek are significantly skewed in favour of esteemed authors and esteemed texts—and in the earlier, pre-Hellenistic period in favour of esteemed verse (as actually with Attic drama!) or (for the fourth century in particular) Attic or quasi-Attic prose (most obviously in the case of oratory, Plato, and Aristotle). In historical perspective, the skewings are no accident. These texts survive, largely because they were esteemed throughout antiquity and were preserved for educational or comparable purposes. One can hardly complain: the extant samples contain a disproportionate number of masterpieces—and it is largely because of these masterpieces that the study of ancient Greek is worthwhile in the first place. For the lexicographer concerned to record ancient norms, however, the situation is far from ideal. But worse: until the mid-fifth century, almost all the extant evidence is verse evidence, simply because prose, as a textual medium, hardly exists before the sixth and only achieves real prominence in the second half of the fifth. This means that a Greek lexicon that attempts to cover the archaic and earlyclassical periods, in particular, has no option but to foreground verse usage. But as it is in verse that the overwhelming proportion of creatively deviant usage occurs, it is inevitable that Greek lexicography will be dealing with deviant, alongside normal, usage as a matter of course.

17.4. DISTINGUISHING NORMAL A ND DE V I ANT US AGE It follows that the lexicographer of ancient Greek must know how to distinguish normal and deviant usage—and then how to present deviant usage itself. These concerns apply to all periods of Greek, but (for the reasons discussed) they apply most pressingly to the early and classical periods that are my special concern. Reflecting on Greek lexicography, Chadwick criticizes LSJ for including citations from ‘many obscure authors . . . which are of little use to students of classical literature’.³² Irrespective of the needs of ‘students of classical literature’, the comment is, to say the least, ill-judged. ‘Obscure’ sources may well be crucial for establishing the norm.

³² Ibid. 3.

Literary Lexicography: Aims and Principles

305

In 1974 I set out principles and procedures for establishing normal usage in post-Homeric, pre-Hellenistic Greek, as distinct from the most common type of creatively deviant usage, innovative (‘live’) metaphor.³³ No subsequent discussions—by myself or others—seem to have offered any consequential challenge to these,³⁴ and (with a few minor modifications or reformulations) I appeal to them now. Mutatis mutandis, they apply to metonymic usage too. The paramount requirement for each usage is to focus on, and make inferences from, the attested distribution (or ‘spread’) of available evidence, with normal usage established positively, by an adequate spread, and deviant usage negatively, by its absence. Two principles (both argued at length in my 1974 discussion) determine the diagnoses: (a) an ‘adequate’ spread will generally include ‘reliable’ prose,³⁵ though if a post-Homeric verse usage (especially of a high-style verse word) is under review, a spread suggestive of normality may sometimes depend significantly, instead, on the evidence of quasi-canonical Homeric usage itself;³⁶ and (b) within the pre-Hellenistic period, later evidence (if the right kind of evidence) from (e.g.) the fifth century suffices to establish the normality of earlier usage, from (e.g.) the seventh or sixth.³⁷ The second principle itself depends on a series of arguments from probability, such that even in the extreme case of a restricted spread consisting of a ³³ Silk 1974, 27 56, 82 4, 211 23, 228 33. ³⁴ The principles and procedures tend to be acknowledged or affirmed en bloc, non discursively, in footnotes, as by Bulloch 1985, 86 n. 1; Pelliccia 1995, 37 n. 58; Conacher 1996, 118 n. 5; Mossman 1996, 63 n. 21; Asper 1997, 15 n. 22; Dover 1997, 124 n. 63; Hordern 2002, 40 n. 115; Catrein 2003, 37 n. 118; Rutherford 2012, 127 n. 20; Wright 2012, 198 n. 10; Brock 2013, xviii n. 13; Maslov 2015, 167 n. 144; Matzner 2016, 19 n. 27. More discursively, Robson 2006 both accepts them (p. 101) and goes on (pp. 101 21) to apply them (with minor modifications) to establishing the tone of (Aristophanic) usage, while Nünlist 1998, 3 7 offers a rare, but uneventful, discussion of some of the issues at stake (on both, see nn. 35 6, 39, 44, below). ³⁵ Where ‘reliable’ prose is summed up as, above all, ‘fully “prosaic” prose, i.e. technical or non literary prose: e.g. prose inscriptions, the Hippocratic corpus [with exceptions], Aristotle, Theophrastus, Aeneas Tacticus’ or else (failing that) ‘ “literary” but generally reliable prose: e.g. the bulk of Herodotus, Thucydides, Xenophon, and certain orators, particularly Lysias and Isaeus’: Silk 1974, 44. On the Hippocratic corpus, see now pp. 328 9 below. The least ‘reliable’ (because most creatively deviant) prose of any extent in pre Hellenistic Greek belongs to Plato: see n. 56 below and cf. Robson 2006, 110 n. 51. Many of the pages cited in n. 33 above are devoted to the rationale for including, or excluding, particular authors on/from one or other of these lists. In Latin studies, similar arguments about the normality of ‘business prose’ have been proposed by Lyne 1989, 7 13. ³⁶ Silk 1974, esp. 35, 39, 43 4 (prose) and 36, 41 (Homer). Nünlist 1998, 4 6 accepts the principle that prose evidence is significantly more indicative of normality (but questions its relevance to earlier verse: n. 39 below). He also (pp. 4 5) follows R.L. Fowler 1987, 3 52 in concluding in effect that ‘ “Homer” has been too easily equated with “the epic” ’ (Fowler 198, 39): cf. Silk 1974, 36: ‘for all [archaic and early classical non epic poetry] . . . early epic usage, especially [but not only] Homeric usage, has a status different in kind from any other poetic usage’. ³⁷ Silk 1974, esp. 36, 40 1: cf. nn. 38 9 below. As in that 1974 discussion, note that in what follows (including my commentaries on LSJ) I ignore evidence from Hellenistic or later sources, unless otherwise indicated.

306

Michael Silk

single usage from considerably earlier verse and a single usage from considerably later, but ‘reliable’, prose, the probability must be that the later evidence implies the normality of the former³⁸—and the more extensive the later spread, the greater the probability. I know of no grounds for challenging these arguments³⁹—but in any case, for present purposes, as indeed in my earlier discussion, I avoid any reliance on the evidence of such ‘extreme cases’.⁴⁰ The following passage from my 1974 discussion illustrates the application of the principles in question. In all cases, the usages under review belong to archaic or early-classical poetry and the ‘evidence’ for (or against) normality to pre-Hellenistic sources:⁴¹ Thus, the spread Herodotus Hippocrates early epic Phocylides Euripides Demades is an eminently trustworthy basis for taking κακὰ ἐπέρχεται to be normal usage for the time of Aeschylus . . . The same applies to the various spreads supporting [the normality for this period of ] ἐπέρχεσθαι of flowing water (inter alia, Herodotus, Hippocrates, Pindar, tragedy); [the normality of ] βένθεα of waters in Stesichorus (Homer, Hesiod, other early epic); [of ] ὀξύς of sharp sounds (e.g. Homer, lyric, comedy, Hippocrates); and [of ] ἐπιρρεῖν of advancing bodies of men (e.g. Homer, Herodotus).⁴²

But regarding Homeric evidence, I offered a proviso: With usages which are found in Homer, but only once, and have a subsequent distribution both limited and inherently suspect [sc. only verse, especially high style verse], the probability is that the usage was [in Homer] and remained [in later centuries] live metaphor [sc. impinges/impinged as deviant, not normal, usage]. Thus σπέρμα, of a spark, occurs only once in Homer . . . and the later

³⁸ See esp. Silk 1974, 40 1. The ‘extreme case’ exemplified is Sappho, 31.10, χρῷ πῦρ ὐπαδεδρόμηκεν, alongside Hp. Fract. 27, ἔρευθος . . . ὑποτρέχει (LSJ s.v. ὑποτρέχω IV): Silk 1974, 40 1 (there is also marginal evidence from other compounds of τρέχειν elsewhere in Hp.: ibid. 42 n. 13). The closest to a direct parallel within the corpus is Hp. Morb. 4.51, ὑποτρέχει ὑπὸ τὴν πληγὴν αἷμα. ³⁹ Robson 2006, 101 2 (and cf. 103 4) plausibly queries the applicability of this principle to (his own concern) the tone of a usage. Nünlist 1998, 5 7 queries the principle as such (‘erheb lichen Zweifeln’: p. 6) but, regrettably, without any counter evidence, counter argument, or discussible alternative. Though himself insisting on the crucial difference between ‘live’ and ‘dead’ metaphor (e.g. pp. 3 5, 228) and though accepting the primacy of plain prose evidence for establishing the norm (pp. 5 6), Nünlist explicitly abandons hope of ‘method’ (p. 228), in favour, as it seems, of intuitive identification, which simply means lumping together, as metaphor, prospective cases of metaphor proper and normal usage (esp. ‘cliché’: cf. n. 44 below), without discrimination (cf. Nünlist 1998, 6 7). ⁴⁰ ‘In practice I shall not be allowing myself anything like so large a margin of error’: Silk 1974, 41. Such spreads of ‘isolated [later] reliable prose plus isolated [earlier] unreliable verse’ are in fact strikingly infrequent: ibid. 48 n. 24. ⁴¹ See ibid. 35 (‘Assumption 5’), 38 9, 49, 82. ⁴² Ibid. 45: q.v. (with notes: ibid. 47) for full evidence and further discussion.

Literary Lexicography: Aims and Principles

307

spread consists of but two passages in Pindar. Those Pindaric passages [and, a fortiori, the Homeric passage itself ] I take to embody live metaphor.⁴³

Regarding that last point, I would now add that establishing normality for Homeric and other early epic usage in its own age (whatever that last phrase might be taken to mean) undoubtedly poses special problems—problems I had no occasion to discuss directly in 1974, when I contented myself with the observation that ‘even on the most generous interpretation, Homer [and other early epic] uses live metaphor (as opposed to cliché)⁴⁴ very infrequently’.⁴⁵ Here too I shall, on the whole, avoid entanglement with such issues.⁴⁶

17.5. NORMAL AND DEVIAN T USAGE: LEXICOGRAPHICAL PROCEDURES First, a restatement: notwithstanding the force of arguments concerning the significance of concealed analogies in ordinary language,⁴⁷ labels suggestive of deviance (like ‘metaphor’) should, in dictionaries, be restricted to apparent cases of deviance itself, and not used for the presumed route of established secondary senses (i.e. regular semantic change).⁴⁸ An English dictionary should not characterize a usage like ‘chair leg’ as metaphorical or a usage like ‘trouser leg’ as metonymic. ‘Leg’, meaning the relevant limb of a human ⁴³ Ibid. 49. ⁴⁴ Where cliché ‘more or less approximates to the status of “dead metaphor”, i.e. normal usage’, but tends to imply ‘usages with a largely poetic . . . currency’: ibid. 2 9. I speak of cliché here, rather than (e.g.) ‘dead metaphor’, on the assumption that routine semantic change by analogy (the change that once produced ‘chair leg’ in English etc.) rarely arises from specifiable (e.g. literary) innovation, but rather develops more or less covertly in a speech community, like phonetic change (the point, and the distinction, seems to have eluded Nünlist 1998, 3, 6 7). Thus Silk 1974, 40 2, 44, 46; cf. now (e.g.) Traugott and Dasher 2005, 34 5, 76 (‘idiosyncratic innovation’ vs ‘diachronic change’); Fritz 2005, 82 3 (‘in der Alltagsprache’); Durkin 2009, 240 (‘largely unconscious processes’). For lexicographical purposes, however, one certainly needs to take special account of the innovativity (albeit short lived) of scientific usages in their first phase: cf. Silk 1974, 44 on ‘electric current’ and McMahon 1994, 191 3 and cf. also p. 329 below. ⁴⁵ Silk 1974, 41: ‘Homeric cliché’ means usages like ποιμένα/ ένι λαῶν (n. 67 below). ⁴⁶ And/but cf. n. 103 below, on the special case of iconyms. ⁴⁷ By Lakoff and other cognitive linguists: see n. 25 above and cf. Silk 2003a, 146. ⁴⁸ OED’s usual practice (with ‘transf.’) provides a model: cf. n. 50 below. Modern linguistics is generally content to follow Aristotle’s pioneering, but muddled, discussion in Poetics 21 (Silk 2003a, 117 18), rather than modern literary theory, in ignoring the distinction between meta phor and analogical semantic change, and between metonymy and semantic change by associ ation: so, e.g. McMahon 1994, 182; Traugott and Dasher 2005, 27; Durkin 2009, 223; Fritz 2005, 81 104. In this respect, linguistics, hugely impressive in many ways, is stuck firmly in the intellectual equivalent of the Dark Ages but, as OED shows, practical lexicography can do better.

308

Michael Silk

being or animal, is normal (and primary) usage, and ‘chair leg’ and ‘trouser leg’ are equally normal (though secondary) usages, and should be marked, ex silentio, as such⁴⁹—qualified (if qualifications are felt to be necessary at all) by labels suggestive of semantic change (‘by analogy’, ‘by association’). Conversely, an English dictionary that opted to cite (e.g.) Shakespeare’s ‘immortal longings’ or Eliot’s ‘river’s tent’ should certainly label them, explicitly, as deviant.⁵⁰ In a lexicon of ancient Greek, the same procedures should be followed, and it is even more important that they should be. A sophisticated native English-speaker who found, under the headword ‘tent’ in an English dictionary, a subdivision suggestive of normal, established usage, ‘overhanging branches . . . tent-like in appearance’ (vel sim.), with citation of the Eliot, would be surprised and rightly scornful.⁵¹ He or she, however, could at least correct the dictionary’s mis-identification by intuitive knowledge of the language. In the case of ancient Greek, there are no native speakers, sophisticated or otherwise, and it is incumbent on the lexicon itself to cultivate a scrupulous differentiation between the normal and the deviant, and to present due evidence indicative of each, as appropriate; and if the usage is presumptively deviant, it should always be marked as such (by ‘metaph.’, ‘trop.’ or whatever), whereas if it is presumptively normal, it should never be marked by ‘metaph.’ (or whatever). And given that it is, almost always, only an adequate spread of evidence that can establish normality, an adequate spread for usage flagged as normal should always be provided.⁵² In its presentation of evidence to establish normality, LSJ is often satisfactory and often at fault. Sometimes it presents an adequate spread, across the genres, with a fair representation of plain-prose usage; sometimes it presents what one can only call randomly selected evidence; but then again, in many instances the criteria for selection are not random at all. This is because the lexicon is inclined to reinforce the skewed survival pattern of extant sources by favouring the evidence of classical Attic literature; it has ‘favourite authors who may not be of much help in reconstructing the norm’,⁵³ privileging, above all, the evidence of

⁴⁹ Conventional and entirely plausible lexicographical practice has long been that the mere presence, under a headword, of a separately demarcated sense or category of usage suffices to indicate ‘normal usage’, unless otherwise qualified: normality is implicit and not spelled out. In the given case, OED duly cites ‘chair leg’ and ‘trouser leg’, as (implicitly) normal, under ‘leg’, II.12 and II.10. ⁵⁰ As OED does, in both cases using its customary formula, ‘transf.’: for immortal, ‘A 1.b., transf. Pertaining to immortal beings or immortality; heavenly, divine’ (a barely adequate gloss on the Shakespeare, but still . . . ); and for tent, ‘2. transf. a. something likened to or resembling a tent . . . ’. ⁵¹ I recycle the example used in Silk 1983a, 309. ⁵² In cases of doubt, doubt should be indicated (‘perh. trop.’, ‘prob. trop. before V ’, etc.). ⁵³ Silk 1974, 83.

Literary Lexicography: Aims and Principles

309

classical Attic drama and classical Attic literary prose.⁵⁴ In historical terms, no doubt, both preferences reflect, at least in part, long-standing, but now outmoded, assumptions about the likely needs of those learning Greek: not least, the need for a resource for ‘reproduction’ activities⁵⁵—prose composition in repro classical Attic and verse composition in repro tragic iambics. Obvious outcomes are the prevalence of citations from (e.g.) Demosthenes and (especially) Plato, neither of whom can be counted on as a straightforward source of normal usage,⁵⁶ and the habit (helpful, as it happens, for identifying stylistic level) of distinguishing some tragic citations as ‘lyr[ics]’ or ‘anap[aests]’, whereby the verse-composer is warned against incorporating the usage in his (usually ‘his’) tragic iambics. Other more sporadic markers point the same way—some of them still helpful (‘poet., Ion., and later prose’),⁵⁷ though a formula like ‘chiefly poet., never in good Attic prose’⁵⁸ makes the reproduction-compositional imperative almost embarrassingly explicit. Both among the instances where its presentation of evidence is adequate, and among those where it is not, LSJ is in any case randomly inconsistent and randomly misleading in its treatment of presumptive normal usage and presumptive metaphorical deviance.⁵⁹ A large majority, no doubt, of the thousands of usages it flags, implicitly, as normal, are indeed presumptively normal, and a bare majority, perhaps, of those marked ‘metaph.’ indeed presumptively deviant—but the occurrence of contrary instances, both ways round, is so common that the user of the lexicon is never able to trust its indications, while the adequacy of evidence provided is always so variable that the user can only be further confused. Among the instances already cited, σπέρμα, of a spark, is correctly marked ‘metaph.’ and ἐπέρχεσθαι, of flowing water, correctly shown as an established usage,⁶⁰ but whereas the (three) relevant citations for σπέρμα:spark are included, ⁵⁴ Implicit in LSJ’s procedures here, as elsewhere, is a schematic understanding of texts and evidence as early, classical, and late, with more attention generally given to the first two than the third. This prejudice, as it happens, is largely congenial to my present concerns. ⁵⁵ I borrow the term from Russell’s felicitous characterization of Second Sophistic neo Attic as ‘reproduction Greek’: Russell 1991, xxi. ⁵⁶ Both authors, esp. Plato, were celebrated in later antiquity for their free use of tropical language: see Silk 1974, 221 (Demosthenes) and 39, 42, 220 1 (Plato). Platonic usage, in particular, is often attested in essentially poetic spreads, as (e.g.) ‘κυμαίνειν of “human passion” has the spread Pindar tragedy (Aeschylus) Plato’: ibid. 46 (cf. LSJ s.v. I.2, where the usage is correctly marked ‘metaph.’) see Silk 1974, 48 n. 24 for other examples, to which add (e.g.) βροτός, n. 13 above. ⁵⁷ See (e.g.) s.v. βαιός. ⁵⁸ As s.v. αὖλαξ (my italics). ⁵⁹ Some of the examples that follow, in this section, are newly chosen, but most are drawn from Silk 1974; further examples can be found by chasing up asterisked headwords listed there on pp. 251 4. ⁶⁰ Though in both cases the relevant subheadings are highly questionable. σπέρμα I.2, headed ‘metaph. germ, origin of anything’, contains not only the metaphorical ‘spark’ passages, but a scatter of diverse prose usages from Anaxagoras to ‘Longinus’, some of which are undoubtedly normal in their time. The loose qualification ‘of anything’, meanwhile, reflects a suspect

310

Michael Silk

the spread of evidence for ἐπέρχεσθαι:water, though indicative of normality, is revealingly partial: ‘Hdt. 2.19, cf. A. Supp. 559 (lyr.), Th. 3.89.’ The usage, as indicated above, is actually attested in a wide range of prose texts, including Hellanicus, fr. 28 J, the Hippocratic περὶ διαίτης (Vict.), 1.27, and Theophrastus’ Historia Plantarum, 4.7.4.⁶¹ Symptomatically, LSJ ignores most of the most decisive evidence here, from three separate plain-prose sources, in favour of the more literary evidence of Herodotus, Aeschylus, and Thucydides. The point, of course, is not that those three authors should be regarded as taboo, but that the overall selection is unhelpfully skewed. Ideally a selection should not be skewed at all: a cross-section of generic habits (and/or of registers) is always desirable. But—if the aim is indeed ‘to record how the vocabulary of a language is normally used’⁶²—of all possible skewings, the one that tends to minimize the evidence of the plainest prose is the most questionable. If LSJ’s treatment of σπέρμα (‘metaph.’) and ἐπέρχεσθαι (not) at least produces what one takes to be the right answer, in many other cases the same cannot be said. Established usages are frequently marked ‘metaph.’, whether without, or indeed with, evidence indicative of normality, though even when such evidence is provided, available plain-prose evidence is still, often, missed. Examples abound. πληγή 6, ‘metaph., blow, stroke of calamity’, is followed by citations from Aeschylus, Hell.Oxy., Aristotle, and Sophocles—but this is a spread indicative of normal usage for the fifth century.⁶³ The same goes for λαμπρός I. 6, ‘metaph. . . . clear, manifest’ (followed by citations from Aeschylus, Sophocles, Xenophon, and Thucydides—while one from the Hippocratic corpus is missing).⁶⁴ With λαμπρός I. 5, ‘metaph. of vigorous action, λ. ἄνεμος a keen wind, Hdt. 2.96, cf. A. Ag. 1180’ (followed by two allusive instantiations in Aristophanes), the spread is in any case suggestive of normal usage—but the given usage, of winds, is far more widely (and prosaically) attested than implied by these citations.⁶⁵ And LSJ is hardly more satisfactory in dealing with cases where the normality is prospectively that of literary cliché. Typical examples are νέϕος II, ‘metaph. . . . a cloud of men, etc.’, for which LSJ adduces understanding of the way secondary senses should be specified: cf. Silk 1974, 33 n. 6 on κύκλος II (‘any circular body’). The ἐπέρχεσθαι usage is subsumed, again loosely, under ἐπέρχομαι III.1 (‘go over or on a space, traverse, mostly of persons, . . . also of water’). ⁶¹ So also Hdt. 8.129.2, Pi. fr. 140c and, e.g. A. Pe. 599 600: Silk 1974, 21, 23 n. 4. Cf. e.g. Hp. Morb.Sacr. 15 and Mul. 1.4 (of flowing blood). ⁶² Chadwick: n. 31 above. ⁶³ Add (e.g.) Heraclit. 11 and Aeschin. 3.147: Silk 1974, 112 n. 14. ⁶⁴ Add (e.g) Hp. Art. 58: Silk 1974, 199 n. 25. ⁶⁵ Ibid. 199 200 n. 28, adding the evidence of a range of sources, from comedy to most pertinently Aristotle’s Meteorologica (361b8) and Theophrastus’ On Winds (Vent. 8, 28). Meanwhile, LSJ’s ‘cf.’ for the Ag. passage represents a commendably distinctive formula for an activated connotation: see pp. 315 17 below.

Literary Lexicography: Aims and Principles

311

evidence from Homer and Herodotus,⁶⁶ and (with a purely verse spread) ποιμήν II. 1, ‘metaph., shepherd of the people’, followed by—perfunctory— citations from Homer, tragedy, and Pindar.⁶⁷ This latter is an unmistakable epic cliché (imitated in later poetry), and should be designated as such. Similarly, ‘poetic cliché’, and not LSJ’s ‘metaph.’, is the appropriate label for the phrase ἄνθος ἥβης,⁶⁸ whereas LSJ’s ἐπιρρέω I. 2, ‘metaph. of large bodies of men, stream on’, is attached to a spread (Homer, Herodotus, Plato) seemingly indicative of (fifth-century?) cliché—but when one adds the missing evidence of plain prose, the usage is seen to be straightforwardly normal.⁶⁹ Conversely, it is not uncommon to find LSJ treating as normal what the available evidence indicates is ‘live’ metaphor: fully deviant usage. A representative instance is ἄγγελος (‘messenger, envoy’), I. 2, ‘generally, one that announces or tells, . . . Μουσῶν ἄγγελος, of a poet, Thgn. 769’. In LSJ, the word ‘generally’ is generally bad news, indicative of fuzzy thinking; and Theognis’ usage is clearly deviant, with the nearest parallel, in early literature, a solitary allusion in Pindar.⁷⁰ A piquant example is associated with Archilochus’ unique use of παλίγκοτος in fr. 176.1–2 West: κεινὸς ὑψηλὸς πάγος | τρηχύς τε καὶ παλίγκοτος—‘that high crag, rough and sullen’. In this sequence, the word παλίγκοτος is again clearly deviant: its—relevant—established usage is of unfriendly people.⁷¹ Outside this bit of Archilochus, the word is not attested of physical objects like rocks—or of any physical object at all, with one significant exception (below). Yet LSJ classifies the usage, implicitly, as normal, even inventing a special sense for it: ‘[παλἰγκοτ]–ος . . . III. steep, rugged, πάγος . . . Archil. 87 [Bergk]’. Procedurally, this is absurd; and hardly less absurd is the item that precedes it in LSJ: ‘II. metaph., of wounds or injuries, growing malignant, festering, Hp. Art. 27 (Sup.).’ Here, a single distinctive instantiation in Archilochus—a highly creative poet, whose work is rich in deviant usage—is taken to suggest normality, whereas a single distinctive instantiation (or what is presented as such) in the sober technical prose of the Hippocratic corpus (Art. is the fifth-century treatise ‘on joints’, περὶ ἄρθρων) supposedly suggests ‘metaphor’. Absurd: even irrespective of the ⁶⁶ See Silk 1974, 123, 125 n. 10, adducing also Archil. 56 (= 106.2 West) and Demad. fr. 1.15 to which Ar. Av. 295, 578, could be added: cliché, certainly, from the fifth century, and arguably from the seventh. ⁶⁷ ποιμένα/ ένι λαῶν is attested over fifty times in the Iliad and Odyssey (Ebeling), as elsewhere in early epic (Asius Sam. 1.3, Iliu Persis 4.2, Minyas 7.8: all Bernabé) and the Hesiodic corpus (Sc. 41, frr. 23(a) 34, 40.1, 193.1, 280.8). ⁶⁸ LSJ s.v. ἄνθος (A) II.1. Silk 1974, 100, 102 n. 16, cites instances (‘e.g.’) from Homer, Hesiod, Tyrtaeus, Mimnermus, Solon, h.Merc., Semonides [= Simon. 20.5 West], Theognis, Pindar, Phryn. Com., along with anon. late archaic verse inscriptions: poetic cliché par excellence. ⁶⁹ Add (e.g.) Aen.Tact. 10.24: Silk 1974, 31 n. 9. ⁷⁰ O. 6.90 1 ἄγγελος ὀρθός, | ἠϋκόμων σκυτάλα Μοισᾶν. ⁷¹ Or harsh human fortunes (αὐτῷ . . . παλιγκότως συνεϕέρετο, Hdt. 4.156). Cf. other usages cited in LSJ s.v. I.1 2.

312

Michael Silk

fact that, on LSJ’s own evidence, this medical usage is attested, repeatedly, in a cluster of derivative forms, elsewhere in this and other early texts of the corpus: παλιγκοταίνω (Fract. 11, Mul. 2.171), παλιγκοτέω (Fract. 11, 25), παλιγκότησις (Fract. 31), παλιγκοτία (Art. 67)⁷²—or the yet more fundamental fact that, far from being a unique attestation, like Archilochus’, the Hippocratic παλίγκοτος itself is actually one of five such attestations in the corpus as a whole.⁷³ LSJ’s treatment of ‘live’ and ‘dead’ metonymy is unsatisfactory on a quite different level. In general, no attempt is made to distinguish deviant usage (metonymy proper) from the established products of semantic change by association (‘trouser leg’), but instead both are presented as normal, indiscriminately. The word αἷμα, along with one of its derivatives, αἱματόεις, provides representative examples. αἷμα, ‘I. blood’, of course is, and is treated as, normal,⁷⁴ as is ‘II. bloodshed, murder’, where the spread given by LSJ is properly indicative of fifth-century normality (tragedy, Demosthenes—and a fifth-century prose inscription).⁷⁵ By contrast, the treatment of the verse adjective αἱματόεις (not actually flagged as ‘poet.’, though it is) is defective. LSJ sets out the word under four subheadings, all, without indication, treated as equally normal. Of these, three (‘1. = αἱματηρός’, sc. ‘blood-stained, bloodied’;⁷⁶ ‘2. blood-red or of blood’; ‘4. bloody, murderous’) are presumptively established verse usages, by virtue of Homeric citations, but the other is visibly deviant: ‘3. suffused with blood, flushed, ῥέθος S. Ant. 528’. More fully, the Sophoclean passage (a choral passage, in anapaests—though here LSJ forgets to say so) runs: νεϕέλη δ’ ὀϕρύων ὑπὲρ αἱματόεν | ῥέθος αἰσχύνει. Jebb translates, ‘a cloud upon her [sc. Ismene’s] brow casts its shadow over her darklyflushing face’, and Griffith (more exactly), ‘a cloud over her brows disfigures her flushed face’. On αἱματόεν itself, Jebb, followed closely by Griffith, comments: ‘“suffused with blood”, darkly flushed. This application of αἱματόεις to the human face seems unparalleled.’ Quite so: this is deviant, metonymic usage, and should be marked as such. Then again, under ‘4. bloody, murderous’, LSJ subsumes three items: ‘πόλεμος, etc., Il. 9.650, etc.; ἔρις A. Ag. 698 (lyr.); βλαχαί Id. Th. 348 (lyr.).’ ⁷² Nine more such instances are in fact attested in the corpus: Kühn Fleischer s.vv. ⁷³ Or seven, if one counts two repeated passages. The list in full is: Art. 19 (= Mochl. 9), 27 (= Mochl. 17); Mochl. 2, 30; Epid. 2.2.24. In several of these passages, the adjective is in the superlative: Art. 19 (but not Art. 27, pace LSJ), Mochl. 2 and 9. ⁷⁴ But NB ‘I.2. of anything like blood, Βακχίου Tim. fr. 7’ (+ post classical prose), where Timotheus (4.4 PMG) uses αἷμα Β. of wine: patently a one off metaphorical use which LSJ contrives to treat as normal (‘anything like’ is another of its unhelpfully loose formulae). ⁷⁵ Add (e.g.) Pl. Lg. 872e. The inscription (SIG 58) is a Milesian legal text of c.450. ⁷⁶ Not actually LSJ’s own translation, under either heading, but appropriate enough for their (sole) citation, under αἱματόεις 1, of Il. 5.82, αἱμ. χείρ. LSJ Revd Suppl. s.v. confuses the issue by replacing ‘= αἱματηρός’ with ‘bloody’ the translation already in use for the labelling of sense 4. For sense 1, leg. ‘blood stained, bloodied, of body parts’ (vel sim.).

Literary Lexicography: Aims and Principles

313

Specifications of ‘etc.’ like the one that follows the Homer passage are opaque. Perversely, LSJ’s own editorial protestation is that ‘etc.’ indicates the existence of additional instances ‘in other authors’—though it would be more helpful if it meant (and sometimes it does seem to mean) ‘and elsewhere in the same author’.⁷⁷ At all events, αἱματόεις of πόλεμος (vel sim.) looks to be a reasonably established usage in Homer and early verse,⁷⁸ which the Agamemnon instance follows, more or less. But it makes no sense, without further comment, to associate these passages with the usage in Aeschylus’ Septem: βλαχαὶ δ’ αἱματόεσσαι | τῶν ἐπιμαστιδίων | ἀρτιτρεϕεῖς—‘bloody/blood-stained bleating, newly nursed/nursing, of at the breast’. αἱματόεσσαι here belongs as much under ‘1. = αἱματηρός’ as under ‘4. bloody, murderous’, but under either heading the adjective is attached to the ‘bleating’ metonymically (by enallage, ‘transferred epithet’)—‘impressionistically’, suggested Headlam⁷⁹— with a series of further metonymic suggestions the outcome. The ‘impressionism’, allied to the open-ended (passive/active) force of the seeming coinage ἀρτιτρεϕεῖς, has the effect of suggesting young mothers, feeding their babies, as well as the babies themselves—all equally ‘blood-stained’ by and amidst the ‘murderous’ slaughter.⁸⁰ A minimally appropriate solution, under sense 4, might be: ‘meton., βλαχαί . . . (with allusion also to signf. 1)’. LSJ’s reluctance to acknowledge metonymic usage as such has strikingly damaging consequences in the case of νηλής. The entry runs: ‘poet. Adj. [as indeed it is], . . . pitiless, ruthless [the sense of the word in Homeric and subsequent poetry] . . . II. Pass., unpitied, ἔκειτο νηλεὲς . . . σῶμα S. Ant. 1197; νηλέα δὲ γένεθλα . . . κεῖται Id. ΟΤ 180 (lyr.)’. If ‘unpitied’ is indeed a plausible implication of these two Sophoclean usages (to which there seems to be no other parallel), it is—only—by virtue of a radical metonymy. The OT passage in particular makes this apparent: νηλέα δὲ γένεθλα πρὸς πέδῳ | θανατηϕόρα κεῖται ἀνοίκτως—‘Pitiless [sic], children on the ground, carrying death, lie without lamentation’. Two usages by a creative poet are not evidence that the word νηλής could mean—as an established meaning—‘unpitied’; and LSJ’s adoption of the supposed sense is as arbitrary as its appeal to the imaginary

⁷⁷ The editorial gloss comes on p. xliv: ‘etc. = et cetera (i.e. in other authors)’. The signal that there are other (in this case) Homeric instances that might be cited is informative; the informa tion that there are other comparable instances in Greek, without a hint of the period or the genre, is close to no information at all. All it can do is assure the user that the word/usage in question is not a hapax, but that, surely, is the case with the overwhelming majority of words/usages in LSJ, and few of them are marked ‘etc.’. The proper procedure would be to flag every word/usage that is a hapax as a hapax (whether in a given period or genre, or indeed in Greek as a whole). ⁷⁸ In Homer, Il. 19.313 and (with additional metonymy) 9.326, along with instances in Archilochus and Tyrtaeus (added in LSJ Supp.) ⁷⁹ Headlam 1902, 435. ⁸⁰ So Tucker ad loc. On the metonymy and its workings here, see further Matzner 2016, 93 4, 108, 119 20.

314

Michael Silk

sense ‘steep’ for Archilochus’ παλίγκοτος.⁸¹ Rather, a suggestion of ‘unpitied’ is implicit in ἀνοίκτως, while νηλέα, attaching itself (‘impressionistically’) to the whole description, points, if anything, to the ‘pitiless’ god of death, himself implicit in θανατηϕόρα.⁸² Deleting ‘II. Pass., unpitied . . . ’ in favour of ‘[I] . . . pitiless . . . meton. . . . ’, though not wholly illuminating, would at least be accurate. By way of a coda to this section, I offer a prime example of LSJ’s hit-andmiss practices, and of the way they impact on the appreciation of poetic usage—but now usage involving connotation play, which will need further consideration in the next section. In Aeschylus, Choephori, 1022–3, the ‘charioteer’ Orestes is out of control, going mad: ἡνιοστροϕῶ δρόμου | ἐξωτέρω, while at Prometheus, 883, the hero finds himself similarly afflicted: ἔξω δὲ δρόμου ϕέρομαι—‘I am . . . charioteering . . . off the track’, ‘I am carried off the track’—but not just ‘off ’ in each case, because ἔξω, as preposition or adverb, is itself idiomatic signifier of madness. Hp. Epid. 5.80, ἔξω ἐγένετο (‘she went mad’), is decisive evidence, alongside a good deal more.⁸³ One would hardly be able to infer this from LSJ, which ignores the adverbial idiom and, for the equivalent prepositional use, cites only ‘ἔ. ϕρενῶν out of one’s senses, Pi. O. 7.47’, followed by comparable usages from Plato (‘ἔ. σαυτοῦ γίγνῃ Pl. Ion 535b’) and Euripides: in itself a spread suggestive of deviance, albeit LSJ—correctly—sets out the usage as non-tropical.⁸⁴ The poetic point here is that in both Aeschylean (or ‘Aeschylean’) passages, ἐξωτέρω/ἔξω, belonging, as it does, to both of the two constituent terminologies (the charioteering and the madness), affects the force or effect of the images in question.⁸⁵ One does not, indeed, expect the lexicon to acknowledge that—any more than one expects OED to write an essay on Cleopatra’s ‘immortal longings’—but one can, reasonably, expect its indication, and documentation, of established senses to point the right way and thus facilitate understanding of the poetic passages in question. And the requisite lexicographical treatment of ἔξω would, likewise, enable understanding of a quite different kind of sequence in Aristophanes. Acharnians, 395–400, features a fraught exchange between Dicaeopolis and Euripides’ house-servant: ἔνδον ἔστ’ Εὐριπίδης; —οὐκ ἔνδον ἔνδον ἐστίν, εἰ γνώμην ἔχεις.

⁸¹ LSJ’s ‘unpitied’ is not its own invention (cf. the Sophocles scholia ad locc.), but the point is unaffected. ⁸² See more fully Silk 2003a, 133 4. ⁸³ See, in full, Silk 1974, 121 n. 117. The Hp. passage, not in LSJ, is now cited in LSJ Revd Suppl. ⁸⁴ For other instances of ἔξω + gen. in this sense (in Menander and Demosthenes), see Silk 1974, 121 n. 117. ⁸⁵ With interactive force (‘explanatory’ or ‘supporting’): ibid. 119, 237.

Literary Lexicography: Aims and Principles

315

—πῶς ἔνδον, εἶτ’ οὐκ ἔνδον; —ὄρθως, ὦ γέρον. ὁ νοῦς μὲν ἔξω ξυλλέγων ἐπύλλια οὐκ ἔνδον, αὐτὸς δ’ ἔνδον ἀναβάδην ποιεῖ τραγῳδίαν.⁸⁶ Approximately: Is Euripides there? —There and not there: d’you understand? —There and not there? —Got it in one, old chap. His mind’s not ‘there’ (it’s gone: out picking phrases), But he’s there, feet up, writing a play. Here again, alongside its overt denotation, ‘outside’, ἔξω implies (connotes) ‘mad’, just as ἔνδον (overtly ‘inside’) also implies (connotes) ‘sane’.⁸⁷ Though the servant is ready to query Dicaeopolis’ mental powers (εἰ γνώμην ἔχεις), his characterization of his master neatly undermines Euripides’ right to be thought compos mentis—but we need some lexicography to help us see this.

17.6. ACTIVATED CONNOTATIONS Heightened language is what Ezra Pound called ‘language charged with meaning to the utmost possible degree’. Such extra ‘meaning’ is created by what T.S. Eliot called ‘words perpetually juxtaposed in new and sudden combinations’⁸⁸—and such combinations frequently involve unexpected (‘new and sudden’) activating of the connotations of words involved.⁸⁹ The ἔξω/ἐξωτέρω/ἔνδον passages discussed in the previous section were adduced to exemplify LSJ’s failure to cite conclusive evidence for established usage. The three passages also involve activated connotations: connotations of ἔξω: madness and ἔνδον:sanity that enrich or inform chariot images in Choephori and Prometheus and a house call to a playwright in Acharnians. All three passages illustrate the unpredictable activating of an established sense of a word, one that adds unpredictably to the linguistic realities of the moment: ‘unpredictably’, then, beyond any of the ordinary unpredictabilities associated with almost any sequence of words in ordinary discourse; and ‘adds’, by virtue ⁸⁶ Text as Sommerstein. ⁸⁷ As LSJ does indicate. On these uses of ἔξω and ἔνδον, see Silk 1974, 121 n. 17 and, further, below, p. 316. Regrettably, comm. on the Aristophanes (including, most recently, Olson) miss the wordplay (contrast Garvie on A. Ch. 233, 1022 3). ⁸⁸ Silk 2014, 436, quoting from Pound’s How to Read (1931) and Eliot’s essay ‘Philip Mas singer’ in The Sacred Wood (1920). ⁸⁹ Here I restate Silk 2014, 436 (where the appeal is to ‘mobilization’, not activation).

316

Michael Silk

of appealing, beyond the apparent denotative significance of a word in context, to connotations which themselves are, very likely, denotations elsewhere (as ‘madness’ is denoted by ἔξω in the Hippocratic text and connoted by ἔξω in the poetic passages cited). What are the implications for lexicography? Not only should any such denotation be registered as an established sense under its due heading; its activation as a connotation is itself part of the evidence (albeit only secondary evidence, and not by itself sufficent evidence) for that usage. ἔξω in the three poetic passages is itself evidence for ἔξω:madness in its own right, and ἔνδον in Acharnians evidence for ἔνδον:sanity. With Aristophanes’ ἔνδον, to its credit, LSJ acknowledges the principle and acts on it: ‘ἔνδον, Adv. within . . . at home . . . 2. b . . . master of oneself, self-possessed . . . ἔ. γενοῦ A. Ch. 233 (οὐκ ἔ. ἔ. ἐστίν with a play on signf. 1, Ar. Ach. 396)’. It might make better sense to put the Aristophanes under ‘signf. 1’ and say, ‘with a play on signf. 2. b’—but more important is the implicit recognition that the Aristophanes belongs under, and is lexicographical evidence for, both senses. In the same vein, LSJ’s response to a passage later in Acharnians is sound: εἰ δέ τις . . . «λιπαρὰς» καλέσειεν Ἀθήνας, | ηὕρετο πᾶν ἂν διὰ τὰς λιπαράς, ἀϕύων τιμὴν περιάψας—‘and if someone called Athens λιπαράς, he’d get anything [out of you he wanted], thanks to that λιπαράς, just by investing you with the dignity of—a sardine’. LSJ s.v. λιπαρός V: ‘λιπαραὶ Ἀθῆναι, favourite epith. with the Athenians, prob. with allusion to the Attic olive, first in Pi. I. 2.20, fr. 76, cf. Ar. Ach. 639, 640 (where he plays on the double sense of brilliant and greasy) . . . ’. Irrespective of the Attic olive, the adjective is indeed a stock epithet of Athens,⁹⁰ as (in verse usage from early epic onwards) of ‘brilliant’ cities (vel sim.) more generally (Χίος . . . νήσων λιπαρωτάτη, h.Ap. 38);⁹¹ and it is indeed standard usage of ‘oily’ and ‘greasy’ things like sardines, fresh or cooked (ἀϕυῆς = τὰ λιπαρά, Ar. fr. 520.1; σιτίοισι . . . λιπαροῖσι . . . θαλασσίοισι, Hp. Morb. 2.64; ἰχθύσι . . . ἑϕθοῖσι . . . λιπαροῖσι, Id. Vict. 3.81). But these Aristophanic wordplays, especially the one on λιπαράς, more or less foreground the connotations in question, and when connotation ‘play’ is less conspicuous (and less ‘play’-ful), LSJ tends to ignore it. In several poetic images already discussed, and many others besides, connotations are operative interactively—and LSJ overlooks them, as it does with ἔξω:madness at Prometheus, 883. A flagrant, but also curious, case involves Cassandra’s λαμπρός at Agamemnon, 1178–81 (καὶ μὴν ὁ χρησμὸς οὐκέτ’ ἐκ καλυμμάτων | ἔσται δεδορκὼς νεογάμου νύμϕης δίκην· | λαμπρὸς δ’ ἔοικεν ἡλίου πρὸς ἀντολὰς | πνέων ἐσᾴξειν)—where four separate senses of the adjective are ‘in play’: ‘“clear”, of the [oracle] . . . “bright”, like the bride’s uncovered face . . . “keen” like the wind . . . “bright” like the sun . . . ’.⁹² ‘Flagrant’, because these are four ⁹⁰ Full citations in Olson ad loc. ⁹¹ LSJ s.v. IV. ⁹² Silk 1974, 197; full discussion and documentation ibid., with nn. 25 9 on pp. 199 200.

Literary Lexicography: Aims and Principles

317

established senses of the word (though not all adequately treated in LSJ); ‘curious’, because one would have expected a one-thing-at-a-time lexicographer to focus on the current denotation, which is surely ‘clear’ (‘my oracular message will no longer be , but clear . . . ’). Instead, our lexicon (s.v. I.5) contents itself with a single citation, but under ‘keen wind’, unwittingly demonstrating the need to acknowledge connotation ‘play’ as it does so.⁹³ A comparable, though superficially quite different, challenge is presented by Pindar’s creative coinage, ἀναξιϕόρμιγγες, in the opening verse of Olympian 2, an epinician ode in honour of Theron, tyrant of Acragas: ἀναξιϕόρμιγγες ὕμνοι, | τίνα θεόν, τίν’ ἥρωα, τίνα δ’ ἄνδρα κελαδήσομεν;— ‘[O you] ἀναξιϕόρμιγγες hymns, which god, which hero, and which man shall we acclaim?’ LSJ opts for a minimalist solution: ‘ἀναξιϕόρμιγξ . . . ruling the lyre, ἀναξιϕόρμιγγες ὕμνοι Pi. O. 2.1.’ I quote the gist of my comment on this usage in a different discussion: [Here] the chief function of ἀναξι is to introduce the poem as a whole and, in particular, the three parallel items of the next verse: god, hero, man. In Greek usage, ἄναξ is a normal honorific term for gods, and then again for heroes but not for ordinary men [cf. LSJ s.v.]. Theron, however, is no ordinary man, but a great man on a par even with . . . ? At the very outset of the poem, then, the metaphorical compound proposes the problematic that underlies Pindar’s epini cian ideology as a whole [sc. the status human? more than human? of the honorand] . . . In addition . . . the lyre is itself ‘lord’ of Greek music it is, after all, the ‘golden’ instrument of Apollo [χρυσέα ϕόρμιγξ, Ἀπόλλωνος . . . κτέανον, Pi. P. 1.1 2] and indeed lyres (ϕόρμιγγες) are themselves right ‘for lords’ (τοῖς ἄναξι). That is: the opening metaphor strictly, the free play of associations of its vehicle creates an explosion of suggestions and connections, one of which, in particular, is momentous for the poem and far outweighs the importance of the notional point of comparison.⁹⁴

What—what more—can we expect the lexicon to do in such a case? ‘Lords of the lyre’, instead of LSJ’s ‘ruling the lyre’, would be a start. Perhaps: ‘lords of the lyre (activating connotations of ἄναξ I, of gods, and II, of heroes)’? As with Cleopatra’s ‘immortal longings’, again, the example raises questions about the limits of lexicography in the face of poetic complexity—but even here, as there, one can still ask the lexicographer to facilitate understanding of the complexity in question.

⁹³ Above, p. 310. LSJ is unforthcoming with dozens of instances discussed in Silk 1974. E.g. Archilochus’ τρηχύς in fr. 176 (above, p. 311) is not only ‘rough’ (of ‘crags’), but also ‘rough, harsh’ of people: Silk 1974, 95, 96 n. 10. LSJ (s.v. τραχύς) fails to cite the Archilochus under either heading, despite the fact that for the latter usage (of people) it is much the earliest attestation (along with Solon, 34.3 West, also not cited). ⁹⁴ Silk 2003a, 130 1.

318

Michael Silk

17.7. ICONYMS For the lexicographer of ancient Greek, verse words present no special problems as such. But verse words are generally archaisms—that is, obsolete in everyday use—and in the Greek poetic experience there is one category of obsolete usage that poses problems of a very special kind: the iconym. I coined the word ‘iconym’, in attempting a first theorization of the phenomenon, in 1983.⁹⁵ Central to the argument is the proposition that, though used in linguistic sequences alongside words with specifiable denotative meanings, iconyms themselves are words that once, no doubt, had denotative meaning, but no longer do: Meaning presupposes communication within a stable speech community . . . A word’s meaning is guaranteed primarily by communal use; in a literate, but not bookish, community before the advent of dictionaries, its meaning is guaranteed only by communal use. If a word becomes obsolete in the speech community, it may survive passively, or as a fossil in set phrases . . . Any appreciable develop ment of literary uses without those restrictions will create an iconym . . . An iconym, then, is a purely literary phenomenon . . . An iconym is a word which has lost its denotations. Its usage is unpredictable and unstable. It has certain properties which ordinary words do not have, but it has less meaning . . . ⁹⁶ [Iconyms have no straightforward reference, no ‘stable centre’,⁹⁷ and only] a few faint scattered connotations: a set of random associations . . . perhaps ran domly overlapping, but largely unrelatable [to each other] . . . The . . . associations will consist partly of earlier literary contexts (from which the knowledge of the word presumably comes), partly, perhaps, of the kind we read as ‘re etymology’. There is a diffuse reference, then, [but] too diffuse to begin to derive a referent [in the real world] from it. If we invoke a referent, we do so via the context.⁹⁸

In the 1983 discussion, I identified the following as iconyms in archaic and/or classical usage: one verb (or verb form), δέδηε; two nouns, ἄωτος/-ον and πέμϕιξ; and nine adjectives, ἁδινός, αἰανής, αἶθοψ, ἀλίαστος, ἀμαιμάκετος, λειριόεις/λείριος, μαλερός, ξουθός, and ὁμοίιος.⁹⁹ All are restricted to verse (or else, as with πέμϕιξ, there is a largely unrelatable prose sense), which usually means elevated genres, while most are words first attested in Homeric epic (πέμϕιξ is again an exception), though usually with a stable (non-iconymic) usage there.¹⁰⁰ More important: as a matter of definition, an iconym must be used, iconymically, by more than one poet (otherwise, we are dealing,

⁹⁵ Inconveniently, the term has since been reinvented/appropriated by the historical linguist Mario Alinei to refer to ‘motivation’ in language change: see Alinei 1997 and 2001. ⁹⁶ Silk 1983, 319. ‘Iconym’ is formally defined ibid. 311. ⁹⁷ Ibid. 313, cf. 309. ⁹⁸ Ibid. 312 (italics added). ⁹⁹ Of these twelve, six begin ἀ or ἁ ; I have no explanation for this. ¹⁰⁰ Cf. Clarke’s objection: n. 103 below.

Literary Lexicography: Aims and Principles

319

prospectively, with the product of idiolectal creativity);¹⁰¹ and, above all, an iconym must be seen to resist classification into ordinary meanings. The symptomatic result of any attempt at such classifying is a dose of guesswork about referents and a ‘spectacularly high proportion of different meanings to occurrences’.¹⁰² Iconyms present, not a multiplicity of different meanings, but a scatter of different (and strictly untranslatable) applications. An iconym’s spread may be much like that of any verse word, but its profile is otherwise distinctive.¹⁰³ Take πέμϕιξ,¹⁰⁴ a word chiefly known from Galen’s discussion of the derivative πεμϕιγώδης in Hp. Epid. 6.1.14. Not counting Galen’s own text, but including solitary attestations of a variant form πεμϕίς and the derivative πεμϕιγώδης itself, the word is attested sixteen times in Greek literature. Of these sixteen occurrences, five are in Hellenistic poetry, two (the Hippocratic πεμϕιγώδης and a use of πέμϕιξ itself, in Euryphon) in classical medical prose, and the remaining nine in classical or archaic verse: four in Sophocles, four in Aeschylus, one in Ibycus. Impressively, all sixteen attestations are cited in LSJ; less impressively, we are not told this; and less impressively still, LSJ seeks to organize these sixteen (or fourteen, for πέμϕιξ itself) into six different meanings, following Galen, more or less—except that Galen (himself following

¹⁰¹ Hence (e.g.) πίτυλος (used variously, but most of the variation is in Euripides) is, on available evidence, probably no iconym: Silk 1983, 325 6. Conversely, αἰανής and ἄωτος are both iconyms and favourite words of (respectively) Aeschylus and Pindar: ibid. 314 (and cf. 304) and 316 17. ¹⁰² Ibid. 313. ¹⁰³ This overall argument, and the concept of the iconym itself, has been effectively accepted by (e.g.) Cairns 1998, 62 3, 68 70; Hummel 1999, 492 5; Tsagalis 2008, 154 as also (beyond the purely classical sphere) Stark 2012, 202, 215 16 while others have acknowledged individual cases: e.g. Asper 1997, 119 n. 45, 120 n. 48, on ἄωτος; Dunbar 1995, 206 on ξουθός; Garvie 1986, 64 on αἰανής; Gerber 2002, 30 1 on μαλερός (contra Olson and Sens 2000, 229); Zimmermann 1992, 88 n. 73 and Knight 1995, 203 n. 253 on λείριος; likewise LfgrE (in those parts of the lexicon published since 1983) s.vv. λειριόεις, μαλερός, ξουθός, ὁμοίιος. The argument has been challenged as such by Clarke, esp. 1995, 12: ‘the only possible evidence for the existence of any iconym will be the scholar’s inability to fit the attestations into a comprehensible pattern’. Not so. Iconyms are always (a) rare verse words/usages (and often the subject of learned debate by later ancient ‘Gramm.’) with (b) a characteristic profile: [i] increasingly diffuse applications (esp. in the post epic age); [ii] visible ‘re etymological’ associations; [iii] often evocation of multiple associations in single instantiations (cf. Cairns 1998, 72); [iv] above all, a wholly disproportionate ratio of apparent ‘senses’ to attestations (in one case, perhaps even to actual uses: see Silk 1983, 308 and p. 320 below). Less comprehensively, without denying the reality of the phenomenon in post Homeric verse, Clarke 1999, 31 questions the possibility of such usage in early/Homeric epic, on the grounds that it assumes fixed texts, which are incompatible with oral compositional culture. That premise seems to me questionable (what is/would be assumed are fossil formulae, of which there are plenty in Homer) but, in any case (pace Clarke 1995, 25 n. 22), Silk 1983 is generally cautious about ascribing iconyms (as opposed to words that later become iconyms) to Homeric poetry itself (see ibid. 313, 321, 325): a position I see no reason to change now. ¹⁰⁴ See the fuller discussion in Silk 1983, esp. 306 8, on which the following summary is based.

320

Michael Silk

earlier ‘grammarians’) largely avoids talk of meanings in favour of apparent applications,¹⁰⁵ and gives every impression, not only that he (and his— evidently Alexandrian—sources) regards the word as a problem, but that he (or his sources) has brought to bear on its elucidation all available evidence. The medical use, as one would expect, entails a real meaning: ‘spots or swellings on the skin’.¹⁰⁶ The other uses, LSJ (s.v. πέμϕιξ) arranges into a series of supposed meanings:¹⁰⁷ 1. ‘breath, blast’ (Sophocles/Aeschylus); 2. ‘ray’ (Sophocles/Aeschylus); 3. ‘drop’ (Aeschylus); 4. ‘cloud’ (Sophocles/ Ibycus + Hellenistic poetry); 5. ‘ghost’ (Hellenistic poetry, with which the πεμϕίς usage belongs)—with the medical use sixth and last. But if, as is likely, we have here all—or even if we only have most—of the literary evidence available to the Alexandrian scholars, it is simply inconceivable that such a thinly attested obsolete word (obsolete outside the non-literary medical use) could mean so many different things: in pre-Hellenistic usage, as it stands, four quite different meanings for nine attestations. The proper conclusion is that, alongside an established (if recherché?) usage in medical prose, the word is current in poetry as an iconym—and perhaps, in the minds of its poetic users, perceived not as ‘the same word’ as the medical term, but as its homonym.¹⁰⁸ Even by the standards of exotic verse words, iconyms are the exception, not the rule. In my 1983 discussion, I identified only twelve, but suggested that in classical Greek there may actually have been ‘between ten and a hundred’.¹⁰⁹ This, I would now suppose, is an overestimate: two dozen is probably nearer the truth—but various candidates are now impossible to be sure about, because, though the words may survive, most of the attestations that would indicate the requisite profile are lost. We seldom have a Galen to help us.

17.8. LEXICOGRAPHICAL PROCEDURES FOR ICONYMS If my argument has been sound, it follows that . . . iconyms are resistant to normal lexicographical procedures . . . I would wish to see LSJ labelling iconyms as such, with the instances arranged into groups defined by contexts or associations, not by meanings, and presented for the most part without reference to meanings, and with sufficient [indication] of when the word seems to have attained its iconymic status.¹¹⁰

In my 1983 discussion, I illustrated the alternative procedure suggested with (among other instances) μαλερός, where LSJ’s entry is: ¹⁰⁵ With formulae like ‘x apparently uses the word of . . . ’ ( . . . δοκεῖ χρῆσθαι ἐπί . . . ): ibid. 307. ¹⁰⁶ Ibid. Cf. p. 322 below. ¹⁰⁷ For LSJ’s full article, see p. 322 below. ¹⁰⁸ Cf. Silk 1983, 313. ¹⁰⁹ Ibid. 315. ¹¹⁰ Ibid. 330.

Literary Lexicography: Aims and Principles

321

μαλερός . . . fierce, raging, in Hom. always epith. of fire, Il. 9.242, 20.316, 21.375, cf. Hes. Sc. 18; πυρὸς μαλερὰ γνάθος A. Ch. 325 (lyr.): metaph., fiery, glowing, ἀοιδαί Pi. Ο. 9.22. 2. fierce, violent, terrible, πόθος A. Pers. 62 (anap.); λέοντες Id. Ag. 141 (lyr.); Ἄρης ὁ μ. S. OT 190 (lyr.); πόνοι Arist. fr. 675.5 (lyr.): neut. pl. as Adv., furiously, E. Tr. 1300 (lyr.). II. μαλεραὶ ϕρένες, = ἀσθενεῖς, subdued, prostrate, Call. fr. anon. 198.¹¹¹

My discussion ran:¹¹² The occurrences, most but not all of which are in LSJ, are best set out in three groups . . . In the first group (a) μαλερός is associated with fire as epithet of the noun πῦρ, and in this restricted context the word is entirely stable throughout the hexameter tradition [six attestations: μαλεροῦ πυρός Il. 9.242, and likewise 20.316, 21.375, Hes. Sc. 18, Hom. Epigr. 4.5, Orac. ap. Hdt. 7.140]. Then in four fifth century passages (b) the word is used with evident allusion to fire, but its immediate applications are quite diversely to ‘songs’, ‘jaw’, ‘Ares’, and [‘roofs’ as standard fifth century metonym for] ‘dwellings’¹¹³

(for all of which we need fuller citations than LSJ gives: πόλιν μαλεραῖς ἐπιϕλέγων ἀοιδαῖς, Pi. O. 9.21–2; πυρὸς μαλερὰ γνάθος, A. Ch. 325; Ἄρεά τε τὸν μαλερόν, ὃς . . . ϕλέγει με, S. OT 190–2; μαλερὰ μέλαθρα πυρὶ κατάδρομα, E. Tro. 1300)— Finally, in four passages from the fifth and fourth centuries (c), its applications have a full iconymic freedom. In the Persians Aeschylus uses the word of πόθος [πόθῳ στένεται μαλερῷ, Pers. 62], in the Agamemnon of ‘lions’ [δρόσοις ἀέπτοις μαλερῶν λεόντων, Ag. 141], Philoxenus in a highly dithyrambic passage applies it to the hands of the diners at a feast [χερσὶ . . . μαλεραῖς, Philox. 836 (e).10 PMG], and Aristotle (poeticus) to great toils [πόνους τλῆναι μαλεροὺς ἀκάμαντας, Arist. 1.5 PMG]. In that last instance, and perhaps in the Aeschylean instances, one might infer some assimilation to μάλα, ‘very’ . . . The word is stable in epic, but fossilized. As an obsolete word, it loses its determinate meaning, and, once taken out of its formulaic strait jacket, it becomes an iconym. The fossilized epic use, nevertheless, survives sufficiently to permit allusion to it. None of this justifies the attribution of ordinary meanings (‘fierce, raging . . . violent, terrible’) by LSJ.

A more appropriate kind of entry, then, might run: μαλερός . . . verse word, attested fourteen times in pre Hell. period. In Hom. always epith. of πῦρ, Il. 9.242 . . . likewise Hes. Sc. 18, Hom.Epigr. 4.5, Orac. ap. Hdt. 7.140. II. From V BC, used as iconym, 1. with allusion to fire, πόλιν μαλεραῖς ἐπιϕλέγων ἀοιδαῖς Pi. O. 9.21 2 . . . 2. variously, πόθῳ στένεται μαλερῷ A. Pers. 62 . . . 3. in Hell. and later verse . . .

¹¹¹ The only change in LSJ Revd Suppl. is an updating of the Call. reference under II. ¹¹² Silk 1983, 322, leaving aside a few qualifications, ibid. nn. 62 4. On μαλερός, cf. also n. 103 above. ¹¹³ E. Tro. 1300, where the notion that μαλερά is adverbial is gratuitous: ibid. 322 n. 63.

322

Michael Silk

It has to be admitted, though, that even this mode of presentation is not always achievable. Take πέμϕιξ again. LSJ’s own entry (along with the entries for the derivative πεμϕιγώδης and the doublet πεμϕίς) runs: πεμϕιγώδης . . . accompanied by vesicular eruption, Hp. Epid. 6.1.14, cf. Gal . . . . πέμϕιξ . . . breath, blast, ἀπῇξε πέμϕιξ Ἰονίου πέλας πόρου prob. in S. fr. 337; κεραυνία π. βροντῆς Id. fr. 538; δυσχείμερος π. A. fr. 195.4. 2. ray, τηλέσκοπον πέμϕιγα χρυσέαν ἰδών S. fr. 338; ἃς οὔτε π. ἡλίου προσδέρκεται οὔτε κτλ. A. fr. 170. 3. drop, Id. fr. 205; μηδ’ αἵματος πέμϕιγα πρὸς πέδῳ βάλῃς Id. fr. 183. 4. cloud, πέμϕιγι . . . ¹¹⁴ ἀγγέλῳ πυρός cloud, harbinger of lightning, S. fr. 539; driving rain or rain cloud, Ibyc. 17; πελιδναὶ ϕλύκταιναι πέμϕιξιν ἐειδόμεναι ὑετοῖο, . . . ¹¹⁵ ἀμυδρήεσσαι ἐς ὠπήν livid pustules like rain clouds (in colour) . . . ¹¹⁶ dim to the sight, Nic. Th. 273 (but = bubbles acc. to Sch.); dub. sens. in Call. fr. 483 (prob. = Oxy.2080.43). 5. ghost, Lyc. 1106, and so prob. in Euph. 134. 6. pustule or part surrounding a pustule, ἐϕίσταται π. οἷον ἐλαίου χλωρῆς ὥσπερ ἀράχνιον Euryphon ap. Gal. . . . πεμϕίς . . . = foreg. 5, Lyc. 686 . . . ¹¹⁷

My 1983 discussion offered no specific template for an improved LSJ entry. For πέμϕιξ itself, and following LSJ’s house style, more or less,¹¹⁸ I now suggest this: πέμϕιξ . . . pustule or blister, or surrounding area of the skin, ἐϕίσταται π. . . . Euryphon . . . cf. πεμϕιγώδης (signf. alluded to in later poetry: see Nic. Th. 273 in section 2. below). II. an iconym (certainly by V ) with varied applications, attested nine times in classical verse: [i] πυκινὰς πέμϕιγας πιόμενοι Ibyc. 31; [ii] ἃς οὔτε π. ἡλίου προσδέρκεται A. fr. 170; [iii] μηδ’ αἵματος πέμϕιγα πρὸς πέδῳ βάλῃς Id. fr. 183; [iv] μή σε προσβάλῃ στόμα πέμϕιξ πικρὰ γάρ Id. fr. 187a; [v] μή σ’ ἀναρπάσῃ [sc. πνοὴ] δυσχειμέρῳ πέμϕιγι Id. fr. 195; [vi] ἀπῇξε π. Ἰονίου πέλας πόρου S. fr. 337; [vii] τηλέσκοπον πέμϕιγα χρυσέαν ἰδών Id. fr. 338; [viii] τάχ’ ἂν κεραυνίαις πέμϕιξι βροντὴ καὶ δυσοσμίᾳ βάλοι Id. fr. 538; [ix] πέμϕιγι πᾶσιν ὄψιν ἀγγελῶ πυρός Id. fr. 539. Mostly used of something threatening or unpleasant (but cf. ii, vii), sometimes to do with bad weather (?i, v, ?vi, viii); usually with significant sensory impact, whether taste (i, iv), sight (ii, vii, ix),¹¹⁹ touch (iv, v), smell (viii), sound (?viii). All nine attestations derive from Galen’s comm. (CMG V 10.2.2.46 54) on πεμϕιγώδης (q.v.), which records suggestions by Gramm. for the referents in question (‘breath, blast’, ‘cloud’, ‘ray’, ‘drop’). 2. later verse usages . . .

¹¹⁵ Ditto. ¹¹⁶ Ditto. ¹¹⁴ LSJ’s marks of omission. ¹¹⁷ See LSJ Revd Suppl. for adjustments to the Hellenistic citations. ¹¹⁸ But (inter alia) with updated references (Ibycus to PMG, Aeschylus and Sophocles to TrGF Radt). ¹¹⁹ Coincidentally or not, the Hp. πεμϕιγώδης usage also foregrounds the visual: πυρετοὶ . . . πεμϕιγώδεες, ἰδεῖν δεινοί. I note in passing that in passages [ii], [iii], and maybe some others, π. could be round (like pustules, though not like blisters): cf. Silk 1983, 312 with n. 34.

Literary Lexicography: Aims and Principles

323

The grammarians’ guesses deserve only a closing glance: they are a distraction from the evidence, and not, in themselves, evidence (unless there is specific reason to think they carry implications for quotations not cited or—more likely—for the lost context of a quotation cited).¹²⁰ Above all, they should not be used to define and structure the entry—in line, indeed, with LSJ’s occasional practice, with rare words, elsewhere.¹²¹ Iconymic status, I would argue, is in principle a black-and-white issue: in a given period, an obsolete word either is, or is not, functioning as an iconym. But determining iconymic status now is a different matter, dependent on the vagaries of extant evidence, and in given cases one may suspect it, without conclusive evidence. In 1983 I discussed the adjective ὄβριμος in such terms.¹²² Another—a new—example is the adjective δαϕοινός, for which LSJ offers this: δαϕοινός . . . epith. of savage animals, tawny (as expld. by most Gramm., though some also give blood reeking), δαϕοινὸν δέρμα λέοντος Il. 10.23; δράκων ἐπὶ νῶτα δαϕοινός 2.308; θῶες δ. 11.474; λαῖϕος δ’ ἐπὶ νῶτα δαϕοινὸν λυγκὸς ἔχει h.Pan. 23; πῆμα δ., of the dragon Python, h.Ap. 304; δ. ἀετός A. Pr. 1022; λεόντων ἁ δ. ἴλα E. Alc. 581 (lyr.); δ. ἄγρα tawny, Pi. N. 3.81. 2. metaph., δ. Κῆρες Hes. Sc. 250; δαλός A. Ch. 607 (lyr.).

There are thirteen occurrences of this word in pre-Hellenistic Greek, ten of them in LSJ. The article begins promisingly, by foregrounding the word’s most usual context, rather than assigning any meaning, before submissively caving in to ‘most Gramm.’, and then, in the end, treating the adjective as

¹²⁰ As perhaps with the earliest citation, the Ibycus. Here the bare phrase cited gives little indication of the context, which the grammarians’ reported comments might be thought to assume. Galen on Hp. Epid. 6.1.29 (CMG V 10.2.2.47): ἐπὶ δὲ τοῦ νέϕους δοκεῖ τετάχθαι [πέμϕιξ] . . . παρ’ Ἰβύκῳ· π. π. π. λέλεκται δὲ οὗτος ὁ λόγος αὐτῷ κατά τινα παραβολὴν ἐπὶ χειμαζομένων εἰρημένην. A ‘proverb’ (παραβολή) concerning those ‘caught in a storm’ could be relevant (though why this should suggest that πέμϕιξ ‘means’ cloud is unclear, not least because Galen goes on to say that ‘most of the grammarians’ take it (here?) to refer to rain drops: ἐπὶ τῶν κατὰ τοὺς ὄμβρους σταγόνων) and in this connection one notes that the word seems to attract alliteration (p p p : see i, iii, iv, vi, ix), which is common in proverbs (Silk 1974, 20 1, 224 7): is it conceivable that the iconymic diffusion of πέμϕιξ depended on a fossilized usage in a proverb? However, no such ‘proverb’ seems to be attested (and perhaps παραβολή here means ‘comparison’?) and the diffusion itself seems remote from anything fossilized, witness the diverse range of epithets (dense/thick [i], bitter/pungent [iv], stormy [v], golden and far seen (or far seeing?) [vii], thundery/lightning y [viii]). In any event, none of these possibilities belongs in a dictionary. ¹²¹ See e.g. the entry for πευκάλιμος: ‘Ep. word used by Hom. only in phrase ϕρεσὶ πευκαλίμῃσι . . . ’, followed by attestations in the Iliad, Hesiod, and elsewhere, with late antique speculations properly confined to the end of the entry. This is no iconym (the word is attested only in a stable phrase), but LSJ’s procedure here is sound and should certainly (a fortiori) be extended to iconyms themselves. ¹²² Silk 1983, 325.

324

Michael Silk

an ordinarily meaningful linguistic commodity by ascribing to it two ‘metaph.’ uses.¹²³ The thirteen attestations fall into four groups: (a) the word is applied to savage animals: δράκων ἐπὶ νῶτα δαϕοινός, Il. 2.308 δαϕοινοὶ θῶες, Il. 11.474 ὄϕις κατὰ νῶτα δα[ϕοινός . . . , Hes. fr. 204.136 M–W¹²⁴ [δράκαιναν] . . . πῆμα δαϕοινόν, h.Ap. 304 πτηνὸς κύων, δαϕοινὸς αἰετός, A. Pr. 1022 (b) to the skins of what were once savage animals: δαϕοινὸν ἑέσσατο δέρμα λέοντος, Il. 10.23 λαῖϕος δ’ ἐπὶ νῶτα δαϕοινὸν | λυγκὸς ἔχει, h.Pan. 23–4 (c) otherwise associated with savage animals: αἰετὸς . . . | . . . ἔλαβεν . . . δαϕοινὸν ἄγραν, Pi. N. 3.80–1 ἔβα . . . λεόντων | ἁ δαϕοινὸς ἴλα, E. Alc. 580–1 (d) with unspecific associations of savagery, without reference to animals: Κῆρες κυάνεαι, λευκοὺς ἀραβεῦσαι ὀδόντας, | δεινωποὶ βλοσυροί τε δαϕοινοί τ’, Hes. Sc. 249–50 παιδὸς δαϕοινὸν | δαλὸν ἥλικ’, A. Ch. 608–9 δαϕοινὸν μάσθλητα δίγονον, S. fr. 129 Radt— along with one fragmentary attestation, ἐπὶ νῶ]τα δαϕοινοί, Choerilus, fr. dub. 13a.13 Bernabé, which presumably belongs to (a) or conceivably (b). If we had only (a) and (b), or indeed only (a) and (c), there would be no reason to consider the case for iconymy. But with a purely verse word,¹²⁵ and one only thinly attested outside the epic traditions, the combination of (b) and (c) might just begin to arouse suspicions (the word is now applied, first to animal-skin clothing, then to the eagle’s ‘prey’ and a ‘platoon’ of lions); and any such suspicions are bound to be aroused more strongly by the range of applications in (d). Here the adjective is applied to the goddesses of death (‘Hesiod’), to the burning log which threatened death to the baby Meleager (Aeschylus), and to a ‘double goad’ (Sophocles). Any attempt to see the word as really meaning ‘tawny’ is stymied by the Hesiodic passage (the goddesses can hardly be both κυάνεαι and tawny), while ‘blood-reeking’ (vel sim.) is close to absurd for the lionskin and the lynx-skin in (b). A single word, no doubt, could mean both (vel sim.), but the profile that emerges overall suggests, rather, an elusive adjective which (as often happens with iconyms) is being re-etymologized¹²⁶ ¹²³ Metaphor assumes prior meaning. If the meaning is uncertain, and especially if it is uncertain that there is a meaning, there can be no talk of metaphor. ¹²⁴ Possibly δα[ϕοινεός (Page). ¹²⁵ Which LSJ, as often, forgets to say. ¹²⁶ Silk 1983, 312, 314.

Literary Lexicography: Aims and Principles

325

by being pulled into the orbit of similar-sounding words—ϕοῖνιξ, ϕοινίσσω, ϕοίνιος (and the rare ϕοινός), and then ϕόνος—with these various associations variously helping to evoke or even create connections with ‘redness’ or ‘blood’ or ‘slaughter’.¹²⁷ So: is δαϕοινός itself an iconym—by the fifth century at least?¹²⁸ There is surely enough in the profile to raise the question, and, all in all, an improved entry might simply translate the evidence and this conclusion into lexicographical form: δαϕοινός . . . verse word, attested thirteen times in classical Greek, most commonly applied to savage animals . . . 2. applied to the skins of what were once savage animals . . . 3. otherwise associated with savage animals . . . 4. with unspecific as sociations of non animal savagery . . . Perh. an iconym, at least from V BC: cf. the range of applications in signf. 3 and 4. Expld. by ancient Gramm. as tawny . . . or blood reeking . . . II. later verse usages . . . .

* * * Although my concern in this chapter is with poetic practice in, and lexicographical response to, pre-Hellenistic Greek, a few words on later usage are appropriate here, in the context of iconyms. In the Hellenistic period, a longstanding antiquarian interest in rare words (γλῶτται) takes on a new intensity and a distinctive significance. When even ‘a leading poet’ like Philitas is prepared to ‘write an extensive work on the subject’,¹²⁹ we find ourselves in an age when ‘the grammarians’ regularly conduct systematic researches into such words, and when poets (sometimes the same people as ‘the grammarians’) are eager, not just to assert their relationship with past poetry by reusing some of its more exotic elements, but specifically to activate and embody the findings of the new scholarship in their own creative work—and, it may be, position themselves vis-à-vis scholarly debate concerning specific interpretations. The Hellenistic reuses of πέμϕιξ are representative.¹³⁰ Given the new preoccupation with γλῶτται, furthermore, one should now also reckon with a poetic inclination both to ‘extend’ the apparent range of an existing iconym¹³¹ and (in effect) to turn rare, but stable, words into new iconyms. One example that I have discussed elsewhere concerns the word ὀπίς.¹³² In classical Greek, the usage of this word is perfectly stable—it refers to divine response to human behaviour, or to men’s religious ‘awe’ in prospect of ¹²⁷ There is also the rare δαϕοινεός (doublet of δαϕοινός?), whose association with ‘blood’ is still stronger: LSJ s.v. ¹²⁸ Cf. Parker on δ. in E. Alc. 581: ‘probably conventional and evocative, rather than descrip tive’. ¹²⁹ Pfeiffer 1968, 90. ¹³⁰ Cf. e.g. Nicander’s use (p. 322 above). On the Hellenistic usages, see further Silk 1983, 306 7. ¹³¹ Cf. Lycophron’s use of πέμϕιξ (p. 322 above). ¹³² Silk 1998, 34 9.

326

Michael Silk

such response—but in Hellenistic verse the word is used, now as if it meant ‘care’, now (seemingly assimilated to ὄπα) as if it meant ‘voice’.¹³³ For the lexicographer, one practical implication is surely that for iconyms, or for any verse word whose profile raises any suspicion of iconymy, Hellenistic and later usage should always be put in a separate heading.¹³⁴

17.9. LITERARY LEXICOGRAPHY: PRACTICAL RECOMMENDATIONS FOR LSJ The foregoing discussion has thrown up a range of criticisms (along with some commendations) of LSJ’s existing practices. It may be helpful to translate these into a set of practical recommendations. Some of the items in the list that follows reflect inconsistencies in LSJ’s current practices; others reflect practices that call for improvement tout court. And while the recommendations mostly concern the treatment of pre-Hellenistic Greek, many can surely claim a more general applicability. 1. LSJ should rigorously and consistently distinguish normal/established and tropical/deviant usage, and should do so on the basis of spreads (in accordance with the principles specified on pp. 304–7 above). 2. Labels suggestive of deviance (like ‘metaph.’) should be restricted to prospectively deviant usage, and not used for the semantic development of established senses. For the latter, no label should be necessary: a separate section (or sense) should always imply normal/established usage.¹³⁵ Conversely, usage presumed to be deviant should always be marked as such (by ‘trop.’, or ‘transf.’, or more specifically by ‘metaph.’ or ‘meton.’), and pseudo-senses should never be invented for such usage (pp. 301, 307–12 above). 3. Like metaphorical deviance, metonymic deviance should be presented as such (however marked) (pp. 307–8, 312–14 above).

¹³³ See respectively Mosch. 4.117 and Maiistas 58 (p. 71 Coll.Alex. Powell). The first of these two innovations reflects learned debate about a problematic (in fact, demonstrably corrupt) passage in Pindar (I. 5.58): Silk (1998) 34 56 (esp. 36 8 with n. 46). This particular reinterpret ation, I would now add, may also imply some ‘re etymological’ assimilation to ὀπηδός, ‘attendant’, as in τέκνων ὀ. (of a παιδαγωγός), E. Med. 53. ¹³⁴ As indicated in my treatments of μαλερός, πέμϕιξ, and δαϕοινός (pp. 321, 322, 325 above); cf. Silk 1983, 313 n. 35, seconded by Cairns 1998, 68. ¹³⁵ If it is felt necessary to specify the presumed mode of semantic development, a separate set of labels should be used: ‘by analogy’, ‘by association’ (pp. 307 8 above).

Literary Lexicography: Aims and Principles

327

4. The spread presented for normal/established usage should be as maximal as possible, privileging the evidence that is the most decisive indication for such usage: plain prose and (in the case of verse words/ usages) Homeric/early epic evidence. But the spread presented should also offer a cross-section of generic norms (and/or registers), and should always include the earliest evidence (pp. 308–12 and n. 93 above). 5. Connotations demonstrably activated in a passage constitute secondary lexical evidence for normal/established usage and can/should be cited, as appropriate (p. 316 above). 6. LSJ’s intermittent practice of using the formula ‘cf.’ to acknowledge an activated connotation as (in effect) secondary evidence is sound, and could/should be extended (n. 65 above). 7. Connotation play may call for cross-referential entries (pp. 313, 316–17 above). 8. Poetic or literary cliché is a species of normal usage and should be labelled, explicitly (as ‘poet. cliché’, or ‘literary cliché’, or specifically ‘epic cliché’ etc.), with a maximal verse spread across genres/authors cited as evidence. As appropriate, LSJ should periodize and indicate doubt (‘perh. trop. until V , then poet. cliché’) (pp. 310–11 above). 9. Verse words/usages should always be indicated as such (and preferably by the phrase ‘verse word/usage’, rather than ‘poet.’). Spreads consisting of verse usage + (e.g.) Plato should be consistently designated as such (‘verse word: also Pl. . . . ’) (nn. 18, 56). 10. A hapax should always be identified as such (with the clear implication that any word/usage not so identified is not a hapax). The practice could profitably be applied to all identifiable uses, not just headwords (n. 77 above). 11. Formulae only relevant to reproduction-compositional activities (‘good Attic prose’) should be adjusted (‘literary Attic prose’) or deleted (p. 309 above). 12. An ‘etc.’ following a citation should be used to mean ‘and elsewhere in the same author’ (not ‘and elsewhere in other [unspecified!] authors’) (pp. 312–13 above). 13. Iconyms and suspected iconyms should be identified as such (in accordance with the principles specified on pp. 318–25 above). 14. For iconyms or suspected iconyms, specification of, and grouping by, ‘meanings’, is not appropriate. Instead, instances should be grouped according to contexts or associations or, failing that, simply listed in approximate chronological sequence (pp. 320–5 above).

328

Michael Silk

15. Given that iconyms are always rare words, the number of actual known attestations should be indicated. This practice could in any case be profitably applied to any rare verse words (pp. 319–24 above). 16. The entry for an iconym should indicate when the word seems to become one. The practice should also be used for doubtful cases (‘perh. an iconym, at least from V ’) (pp. 321–5 above). 17. Iconyms often require fuller citations (wider contexts) than LSJ usually gives (pp. 321–2 above). 18. In the case of iconyms, or suspected iconyms, ancient speculations (by ‘Gramm.’) should not be privileged, but cited s.fin. (p. 323 above). 19. For iconyms, or suspected iconyms, Hellenistic and later verse usage should be cited in a separate section (pp. 325–6). 20. In all cases—normal/established usage, deviant/tropical usage, iconyms/suspected iconyms—loose qualifications of a word’s supposed semantic range should be avoided (‘generally’, ‘of anything’, ‘any’, ‘anything like’) (nn. 60, 74, and p. 311 above).

A P P E N D I X : P O E T IC U S A G E IN TH E HI P P OC R A T I C CORPUS: VERSE WORDS? METAPH OR S? Specialists on the Hippocratic corpus often suggest that the texts are intermittently ‘poetic’. Craik 2015 a recent and informed compendium offers representative examples, close inspection of which is revealing. In what follows, I take the opportun ity to flag up the vagaries of LSJ’s coverage. (a) The term ‘poetic’ is applied generously, from approximations to verse rhythm (e.g. Craik 2015, 5) to especially supposed verse words. In the latter case, many of the attributions turn out to be misleading, because, though perhaps strikingly used in given passages of poetry, the supposed verse words are also used in prose, albeit not necessarily ‘good’ Attic prose: in some cases the spread suggests a literary Ionic usage, in others a more technical affiliation. Thus ibid. 120 on Gland.: ‘an artificial poetic idiom is affected in the use of such terms as παῦρα (2), πῆμα and ἄλλοιος (12), ὄλλυσθαι (14)’. However, ὄλλυσθαι, though indeed alien to Attic prose (cf. p. 301 above), occurs at Melissus, fr. 7.2 D K, and elsewhere in the Hp. corpus itself at Epid. 6.7.2, Mul. 2.120 and 134, Nat. Puer. 30 (none of these in LSJ); παῦρος/ α, likewise, at Thphr. HP 8.7.4, along with twelve occurrences in Hippocratic treatises (Int. and Mul.) elsewhere (Ind. Hipp. s.v.; only the Thphr. is in LSJ); while, even on LSJ’s evidence, ἄλλοιος (Hdt., Thuc., Xen. also Arist. Div.Somn. in LSJ Revd Suppl. not to mention over twenty other instances in the Hippocratic corpus) is plainly no verse word at all. πῆμα, on the other hand, is indeed a verse word (never otherwise attested, outside verse quotations, in classical prose?), and counts as one of the cases of distinctive coincidence between the vocabulary of (esp.) Homer and the corpus (Silk 1983, 326 n. 74). (b) Many of the ‘poeticisms’ cited elsewhere occur in non technical treatises like Flat. (on which cf. Silk 1974, 84) or treatises widely thought to be post classical (like

Literary Lexicography: Aims and Principles

329

Cord., Hebd., Lex, Praec.: cf. Silk (ibid.) and Craik’s own datings of these works). For suggested metaphors in these texts, see Craik 2015, 55, 249, 152, 232. (c) Suggested metaphors in classical age technical treatises are few, and often misleadingly attributed. [i] Some are formal comparisons: thus τὸ χρῶμα τῶν χυμῶν . . . ὥσπερ ἀνθέων, Hum. 1 (Craik 2015, 132), and the earth/belly analogy in Hum. 11 (Craik, ibid.). [ii] Others are not deviant but established usage in technical prose: thus παροχέτευσις, of the ‘diversion’ of moist elements within the body, Hum. 1 (‘meta phorical basis’: Craik, ibid.), for which see παροχετεύεσθαι at Thphr. CP 5.17.4 (of moisture in plants: in LSJ s.v.) and the simplex ὀχετεύειν/ εσθαι at Hp. Oss. 13 and 18 and Arist. PA 668a20 (all of blood flow in the body; only the Arist. in LSJ). Likewise, ἄμπωτις, Hum. 1, of moisture ‘absorbed back’ into the body (‘metaphorical basis’, again: Craik, ibid.), for which see ἀναπίπτειν at Hp. VM 22, Art. 40 and 50, Int. 26 and 40, Nat.Puer. 21 (the first three in LSJ). Likewise, ἄρδεσθαι, of the ‘nourishing’ of body elements, Carn. 13 (‘trope’: Craik 2015, 46), as at Hp. Loc.Hom. 1, Pl. Tim. 82d, cf. Arist. GC 335a14 (both the usage and the instances are missing in LSJ). [iii] One supposed metaphor is not deviant, but established usage (literary cliché?) beyond the technical sphere: ἐρᾷ τοῦ θανάτου, Virg. 1 (‘metaphor’: Craik 2015, 278), for which see attestations of ἐράω/ἔραμαι (pragmatically, ‘the same’ word) at e.g. Il. 9.64 (πολέμου), Archil. 19.3 West, Hdt. 1.96, Pl. Alc. 2.141d (τυραννίδος), Gorg. 11a.15 (πλούτου), A. Pers. 392 (μάχης), Ar. fr. 292 PCG (λοπάδος ἑψητῶν) (different, but broadly adequate, spreads in LSJ s.vv.). [iv] Less clear cut is ἀνθέω, of the ‘humours’, in Hum. 8 (‘metaphor’: Craik 2015, 132). The usage evidently alludes to the comparison in Hum. 1 discussed in [i] above, but might be thought to be subsumable under ἀνθέω, of diseases (LSJ s.v. II.4, ‘to be at the height or pitch’), as at Hp. Epid. 1.25 and S. Tr. 1089 (both in LSJ), as also Hp. Mul. 2.113, cf. Morb.Sacr. 8. [v] An interesting case is πηγή in Morb. 4 (the belly as π. of the body), ‘which is metaphorical . . . [and which] the author seems to claim . . . as his own neologism’ (τὰς δὲ πηγὰς ἃς ὠνόμοσα, Morb. 4.39: Craik 2015, 188). Is this a scientific coinage like ‘electric current’ (n. 44 above)? It may indeed be a distinctive usage at the micro level but that does not necessarily mean that any trope is involved, because (quite apart from the numerous other such uses of π. in Morb. 4. 33 42) there are comparable usages elsewhere in the corpus (καρδίης π. . . . ϕλέψ, Oss. 2) and beyond (π. τοῦ αἵματος [sc. ἡ καρδία], Arist. PA 666a8) (LSJ s.v. II.2, ‘metaph., source, origin’, is inadequate, as well as misleadingly labelled). [vi] In fact, the only one of the various ‘metaphors’ ascribed to these texts in the corpus by Craik that is unquestionably deviant usage occurs at a moment of quasi sophistic flurry at Morb. 1.1: ϕυλάσσοντα δεῖ ἐπιτίθεσθαι ἐν τῇ ἀντιλογίῃ (‘you must stay on guard and attack [your opponent] in your opposing argument’) (‘metaphor ical’: Craik 2015, 172; the usage is not in LSJ). All in all, there is remarkably little here to challenge the principle that the plain, technical, classical prose of the corpus is overwhelmingly a repository of non tropical usage. (d) Notwithstanding the above, it surely is the case that classical Ionic prose, in contrast to its better attested Attic counterpart, is much more liable to maintain the sensuous linguistic concreteness of earlier Ionic, even the Ionic of Homeric epic: Silk 2009, 24 9; cf. e.g. Craik 2015, 132 on Hum. 9, ὄϕις κτλ. But verse words and metaphors are another matter.

18 Lessons Learned During my Time at the Lexikon des frühgriechischen Epos Michael Meier-Brügger

The Lexikon des frühgriechischen Epos (LfgrE) was founded by Bruno Snell after World War II at the University of Hamburg. The focus of LfgrE is specifically on the lexical semantics of the words of early Greek epic and the semantic fields that each one occupies; other considerations are secondary. The publication by Vandenhoeck & Ruprecht of Göttingen started with the first fascicle (Lieferung) in 1955; the final fascicle, no. 25, appeared in 2010. LfgrE is available in four volumes: fascicles 1–9 in vol. 1 (containing all lemmata beginning with A) of 1979, fascicles 10–14 in vol. 2 (Β–Λ) of 1991, fascicles 15–19 in vol. 3 (Μ–Π) of 2004, and fascicles 20–5 in vol. 4 (Π–Ω) of 2010. The work is thus finished. I was responsible for LfgrE from 1984 until its completion in 2010; this short contribution is a reflection on some of the lessons learned in those years of working on the dictionary.¹ To form an estimate of LfgrE, it is important to remember that so-called ‘Archaic’ Greece is not the beginning of Greek linguistic history; rather, the scholar and the lexicographer must also consider the legacy of the Mycenaean period (see Vine, Chapter 6, this volume), when the speech-community of the Greek language was part of a larger Mediterranean world in which many influences were in play, above all those of the poetic traditions and cultural horizons of the Ancient Near East. In spite of this starting point, the special domain of LfgrE is early Greek epic: not only Homer and Hesiod but also later and less-well-known poets up to Panyassis of Halicarnassus in the fifth century BCE. The two Homeric poems, the Iliad and the Odyssey, are known to us as fixed written texts but represent the summit of a centuries-long oral tradition. For

¹ For more information about the development of LfgrE, see http://www1.uni hamburg.de/ Thesaurus/LfgrESch.pdf. Michael Meier-Brügger, Lessons Learned During my Time at the Lexikon des frühgriechischen Epos (LfgrE) In: Liddell and Scott: The History, Methodology, and Languages of the World’s Leading Lexicon of Ancient Greek. Edited by: Christopher Stray, Michael Clarke, and Joshua T. Katz, Oxford University Press (2019). © Oxford University Press. DOI: 10.1093/oso/9780198810803.003.0018

334

Michael Meier-Brügger

this reason, LfgrE is not a general lexicon of the Greek language of the eighth or seventh century BCE but a lexicon of the artificial language (Kunstsprache) of the epic poets of that time. As is now understood, the poets (ἀοιδοί) performed orally, and their language was metrically shaped, with their spoken Ionic dialect combined with Aeolic forms. In the same poem there may co-exist archaisms and neologisms; and some aspects of word-choice and word-form seem to be determined by purely metrical considerations. As indicated in the 1925 Preface (Jones 1940), the scholars originally responsible for the lexicon that is the theme of the present book took as their basis Franz Passow’s admirable Handwörterbuch in its fourth edition (1830).² In addition to Passow they gathered together all lexica available at the time (including specialiazed ones) and created a most comprehensive dictionary. By the time of the new (ninth) edition of 1940, LSJ represented by far the best Greek-English lexicon yet assembled. The aim of this short contribution is to present a sample set of case-studies to check the quality of information presented in LSJ against the results obtained during my life-long research in the field of Greek lexicography, with an emphasis on early epic. A small selection of notices follows.

18.1. ΒΑΡΥΣ Modern Indo-European linguistics has brought many new insights in the reconstruction of the prehistory of the parent language and the steps that lead us from there to the individual attested varieties.³ For instance, the indications of a Proto-Indo-European form in the etymological parenthesis at the end of the lemma for βαρύς are no longer adequate. What is called for, of course, is not *gwr̥ú-, or the same with a ‘long resonant’, but *gwr̥h₂ú- with the confident reconstruction of a laryngeal.⁴

18.2. ΔΟΥ ΛΟΣ LSJ usually presents the best information that was available at the time the entry in question was compiled. The decipherment of Linear B, which marked a new and important step in the knowledge of Ancient Greek, was achieved only in 1952, and it is now essential to consult the modern etymological dictionaries, whether or not Mycenaean Greek has brought new information ² A fifth edition is available: Passow 1841 57. ⁴ See e.g. Mayrhofer 1986 2001, s.v. gurú .

³ See e.g. Meier Brügger 2010.

Lessons Learned During my Time at the LfgrE

335

on the specific word in question (see Frisk 1960–72; Chantraine 1968–80; Beekes 2010; see also Katz, Chapter 5, this volume). Take, for instance, the case of δοῦλος: Mycenaean attests a designation for men with singular /dohelos/ and plural /doheloi/, and for women with singular /dohelā/ and plural /dohelai/. The designation seems to function as an adjective /dohelo-/, /dohelā-/, and its defining concept is a social bond. As Aura Jorro puts it, ‘In terms of its meaning there is no equivalence with the Classical term . . . it seems generally to indicate a relation of dependence, in the majority of cases with regard to another person.’⁵ In contrast, in our earliest alphabetic Greek, δοῦλος designates a slave in the strict sense, an individual who is ‘the permanent private property of another . . . and must work for him’.⁶ Homer employs for ‘chattelslave’ not δοῦλος but a group of words whose use continues in later poetry: δμώς with masculine plural δμῶες, feminine plural δμῳαί, and so on. However, the adjectival derivative δούλιος is indeed attested in Homer, and this shows clearly that the base-term δοῦλος cannot have been foreign to the variety of Greek represented by the Homeric corpus. Despite this, LSJ presents both words separately without mentioning the fact that they belong in a single semantic field.

18 . 3 . ΕΑΦΘΗ This verbal form is attested only twice in the Iliad, in battle scenes in the phrase ἀσπὶς ἑάϕθη (Il. 13.543, 14.419). The meaning is not clear. In a parenthetical etymological note, LSJ mentions that the word has been the subject of lengthy debate since antiquity. Knowing this old crux very well, I asked myself one day: what is the Greek equivalent of German singen? The Proto-Indo-European root *sengwh- was always thought to have but one surviving reflex in Greek, the epic-poetic substantive ὀμϕή ‘divine voice’; the underlying verb was believed lost, presumably replaced by ἀείδω. My proposal is that ἑάϕθη is based on the aorist stem of the verb in question, so that the meaning of the Homeric phrase would be ‘(upon him) resounded the shield’.⁷

⁵ ‘En cuanto a su significado no hay coincidencia con el término clásico . . . parece indicar genéricamente una relación de dependencia respecto a otra persona en la mayoría de los casos’ (Aura Jorro DMic I s.vv. do e ra and do e ro). ⁶ ‘dauernder Privatbesitz eines anderen . . . und [der] für diesen arbeiten muß’. See Schmidt 2006, 118 19, with quotation in n. 3. ⁷ See Meier Brügger 1989.

336

Michael Meier-Brügger

18 . 4 . ΕΠΟΜΑΙ AND Ε ΠΩ The two verbs ἕπω and ἕπομαι go back to different roots but in Greek became homonymous, being differentiated in form only by voice: active in the one case and middle in the other. Consideration must be given to the possibility of secondary contamination between them. LSJ very properly reports the following in summary form: ἕπω (A) ‘to be about, busy oneself with’ < Proto-Indo-European *sep(cf. Sanskrit sápati ‘worship’, ‘tend’ and saparyáti ‘worship’, ‘honor’, Latin sepeliō ‘give funeral honors’; not related to ἕπομαι), simplex only in Il. 6.321, elsewhere compounded with the prepositional preverbs ἀμϕ-, δι-, ἐϕ-, μεθ-, περι-; *ἕπω (B), only in Med. ἕπομαι ‘to be or come after’ < Proto-Indo-European *sekw- (cf. Sanskrit sácate ‘accompany’, ‘follow’, Latin sequor ‘follow’, also in Lithuanian). Other lexica, on the other hand, are either ignorant of the double background or fail to acknowledge it, attempting to explain both verbs as reflexes of a single root and account for the development of the divergent meanings as ‘secondary’.⁸

18.5. Σ Τ Ε Υ Τ Α Ι, ΣΤΕΥΤΟ This Epic verb is only rarely attested, for instance in the description of Tantalus, στεῦτο διψάων (Od. 11, 804). According to LSJ, the meaning is ‘makes as if one would’. In a long parenthesis at the end of the entry, LSJ acknowledges that the meaning of the word has been debated and doubted since antiquity; translation remains difficult. LfgrE takes into account the latest research and is finally in the position to translate the Tantalus phrase in a more adequate manner, offering the paraphrase ‘it was evident that he thirsted’.⁹

1 8. 6 . Σ Τ Ρ Ο Γ Γ Υ Λ Ω ΜΑ The reader may be too slow to look for a mistake in LSJ. The present word is plainly a derivative of στρογγύλοϛ ‘round, circular’, and one is likely to be mystified by LSJ’s entry, which consists of the words ‘pillow or mosquito-net, ⁸ See e.g. Montanari 2015 s.v., where Sanskrit sápati and Latin sequor are erroneously thrown together. ⁹ ‘[E]s war deutlich, dass ihn dürstete’ (V. Langholf s.v.).

Lessons Learned During my Time at the LfgrE

337

τριχῶν’, followed by a single Biblical reference: ‘Al. I Ki. 19.16’. A look at the biblical context explains the mystery. David’s wife, Michal, has put a dummy image of him in his bed in order to deceive his enemies into believing that he is lying there sick. Verse 16 reads, in the words of the King James Bible: ‘And when the messengers were come in, behold, there was an image in the bed, with a pillow of goats’ hair for his bolster.’ It turns out that the abbreviation ‘Al.’ refers to Origen’s collection of variant readings in the biblical text, the Hexapla, keyed to Field’s edition Origenis Hexaplorum quae supersunt (2 vols., Oxford, 1875). As recorded there (vol. 1, 520–1) the mysterious phrase in fact first appears at verse 13, when Michal puts the dummy image in the bed. In the words referring to the object placed at the head of the dummy, the underlying Hebrew is ambiguous between two phonetically similar words, one meaning ‘liver’ and the other ‘net’ or ‘blanket’. Origen’s Greek variants, accordingly, offer στρογγύλωμα τριχῶν τῶν αἰγῶν alongside ἧπαρ τῶν αἰγῶν ‘liver of goats’, with the latter corresponding to the text read by the King James translators. (The entry in the Hexapla at verse 19 simply refers back to this debate a few lines earlier.) What we have been left with in LSJ are no more than the three key words of the discussion of this textual problem, adrift from the context in which they made sense.

1 8. 7 . ΧΕΡΝΗΤΙΣ In this case LSJ follows a sound policy, putting together the whole family of χερνητ- ‘poor, needy’. Usually Homeric philologists take into account only γυνὴ χερνῆτιϛ, and interpret the feminine as a compound ‘spinning with the hands’. From there, χερνητ- would have acquired the general meaning ‘day labourer’ and ‘poor’. However, a move from the specific to the general, rather than the reverse, is an extremely unlikely semantic development. A more credible history is to start from *χερ-νο-as the positive form of χείρων, χερείων ‘worse, weaker’. From this χερν-ητ- and also χερν-ητ-ιδ- are formed.

18.8. CONCLUSION The knowledge base of LSJ is essentially what was current in the fifteen years or so up to the eve of World War II. From that perspective, the general presentation of the data may be said to show skill and good sense.¹⁰ Today, ¹⁰ LSJ certainly includes errors: see in the first instance LSJ Revd Suppl.

338

Michael Meier-Brügger

however, a wealth of new data is available: Mycenaean forms, Proto-IndoEuropean reconstructions, new textual editions, and lexica devoted to particular periods and varieties of the language. An entirely new Greek-English lexicon that includes all this new information in electronic form would be desirable but is not yet in sight.¹¹ LSJ is, therefore, still the best available GreekEnglish lexicon, the one from which we should begin and to which we must all add new insights as they become available.

¹¹ Unfortunately, Montanari 2015 is not the hoped for ‘new LSJ’.

19 Diminishing Returns and New Challenges †

Martin L. West

The story of LSJ in the second half of the twentieth century revolves round the production of two successive Supplements. I was myself implicated in both of them. In around 1960, when E.A. Barber was working on the first Supplement, I saw a notice that he had put in one of the journals inviting people to send in addenda or corrigenda. Having noted several dozen such in the course of reading, I made a list of them and sent them to Barber. He responded by inviting me to go and meet him in the office at the top of the New Bodleian Library which was the centre of his operations. Either then or not long afterwards, he asked if I was willing to do something more for the Supplement. There were recent editions of the Orphic Hymns (by W. Quandt) and of Nonnus’ Dionysiaca (by R. Keydell), and he wanted someone to go through them looking out for new words or meanings. If I would do this, I would be remunerated by the Press at the rate of ten shillings per hour. I agreed to take on the task. There are not many classicists who have read all forty-eight books of Nonnus’ Dionysiaca, and even fewer who have been paid for it. Towards the end of 1963 Barber invited me to assist the project in a more general way. A year and a half later, after some months of ill health, he died, and I found myself left in charge. The work of collecting material for the Supplement was more or less done, but it fell to me to put it together for publication and see it through the Press. It came out in 1968. The office in the New Bodleian was given up, and I took custody of the boxes of slips from which the Supplement had been printed. From then on, if anyone sent in any corrections to the Press, they were forwarded to me and I processed them, and if I happened to come across anything that needed correcting, I made out a new slip and put it in with the rest. My idea was that if the collection of slips was kept up to date, a future supplement could be printed from them at any time. A little over a decade later it seemed to me that it was time to think about getting a new Supplement going to incorporate and expand on the first. In Martin L. West, Diminishing Returns and New Challenges In: Liddell and Scott: The History, Methodology, and Languages of the World’s Leading Lexicon of Ancient Greek. Edited by: Christopher Stray, Michael Clarke, and Joshua T. Katz, Oxford University Press (2019). © Oxford University Press. DOI: 10.1093/oso/9780198810803.003.0019

340

Martin L. West

1979 I made a proposal to the British Academy that it might instigate such a project. The Academy, after two years of leisurely discussions within itself and with the Press, agreed to adopt the LSJ Supplement as one of its Major Projects. A committee had already been set up, with myself as chairman, and an advertisement was now placed for a full-time editor. Several applicants were interviewed, and Peter Glare was appointed. At his interview he was asked whether, after spending so long on the Oxford Latin Dictionary, he could really contemplate spending another ten years on a Greek lexicon. His eyes lit up, and he said fervently, ‘Oh yes!’ So he began work under the committee’s supervision. We used to meet every six months or so in the Clarendon Press Institute in Walton Street, the building that now houses the Centre for Linguistics and Philology. When the Press had it, it was a sort of staff club, with an effective kitchen. We always used to be served a superb lunch before our meeting. So the years went by. At one meeting Robert Browning made us all furiously envious by reporting that in coming from the station he had found in the bargain tray at Waterfield’s bookshop a copy of the Passow–Crönert lexicon for (I think it was) 5p. That work, Wilhelm Crönert’s revision of Passow’s old Wörterbuch der griechischen Sprache, is regarded by many persons of discrimination as the finest Greek lexicon. Dating as it does from 1913, it is of course not up to date with more recent discoveries, and it has the rather more serious deficiency that it never got beyond ἀνά. It breaks off where the fifth Bogen came to an end, in the middle of a quotation from Plutarch, much as Bach’s Art of Fugue ends abruptly in the middle of a bar. That second Supplement eventually came out in 1996. Tranquillity descended once more. But what next? By that time the digital revolution was well under way. It was becoming clear that the future did not lie in the continuing production of supplements in book form every twenty or thirty years. It became natural to think instead in terms of electronic text that could be updated continuously. We older folk cannot really imagine ever managing without the ponderous physical presence of a hard-copy LSJ within reach. It has contributed not a little to the strength and suppleness of our wrists. I know that younger generations see things differently and regard it as highly liberating and convenient to consult the Lexicon online. Of course there are many advantages to that. Perhaps the greatest of them will be the accessibility of a continuously updated version, if that can be made available. But who is to do the updating? Could it be like Wikipedia, with any user able to edit an article to his or her taste, subject to certain controls? No, clearly not, that would be disastrous. It is essential that, as in the past, all contributions be assessed by a qualified editor or a panel, who will decide for themselves in what way, if at all, the lexicon entry is to be modified. But how much can be asked of such an editor? The two Supplements give an idea of what can be achieved in fifteen years by a single editor working full time

Diminishing Returns and New Challenges

341

with some part-time assistance from one or two others. It is not realistic to expect that a larger establishment might come into being in the future. Those editors succeeded reasonably well in their main aim, which was to keep up with new texts. Other improvements were made haphazardly. Obviously it was not possible for them to attempt anything like a systematic revision of the Lexicon. Fortunately, what we have is a work of remarkably high quality. As a onevolume lexicon of ancient Greek covering the whole period from the eighth century BCE to the end of antiquity, LSJ has no serious rival and is unlikely ever to have one. It will never reach a state of absolute completeness or perfection, but we can expect that gradual small improvements will continue to be made as new material is incorporated and as mistaken or inadequate interpretations are replaced with better ones. Yet the gains will be minor. Anyone who has a copy of the ninth edition will still be quite well equipped. But I want to look beyond LSJ, which is after all not the only kind of lexical aid that we use. We have the TLG, which provides something like a complete collection of occurrences of words in literary sources, though it does nothing towards explaining what they mean. Scholars and students nowadays tend to use LSJ and the TLG as complementary resources. You learn from LSJ what a word means and how it is used, and if you want to trace all its occurrences you turn to the TLG. Between these two poles we have an increasing number of indexes, concordances, and special lexica to particular authors or text corpora. The two major examples are the Lexikon des frühgriechischen Epos (recently completed: see Meier-Brügger, Chapter 18, this volume) and the Hippocrates Lexicon. One valuable sort of lexicographical enterprise that we might hope for in the future, besides further refinement of LSJ and the TLG, will be the creation of more works of this kind. There still remains a wide gap between this type of special lexicon and the comprehensive Greek lexicon as represented by LSJ. In this gap, I suggest, it would be useful to think of creating a few works of intermediate character, combining the full and exact detail typical of the special lexica with coverage of a wider band of literature. For instance, a Lexicon of Imperial Prose would be of great value to those studying a particular prose author of the Roman period. It would provide coverage of an area for which LSJ is relatively thin. As a more concrete example of the sort of thing I have in mind, I would like to conjure up the vision of a Poetic Lexicon of Classical Greek, embracing epic, elegy, iambus, lyric, philosophical poetry, drama, and verse inscriptions of the pre-Alexandrian centuries. This makes good sense as a coherent corpus, because although some usages are peculiar to Homer, say, or to tragedy, in general one cannot study lyric or tragic vocabulary in isolation from each other or from epic. If one is investigating a word in Pindar, for example, of course one needs to look at all its occurrences in Pindar, but one also needs to survey its use from Homer on, and its use in tragedy.

342

Martin L. West

I have no ambition to create such a lexicon myself, but to illustrate my conception I have composed a few specimen entries and present them here. I have chosen four words, moderately interesting ones: an interjection, ἆ; an adjective, αἰδοῖος; a verb, ἀΐσσω; and a noun, ἄτη. I have tried to include references to every occurrence of them in pre-Alexandrian verse. The layout of the entries will remind the reader of LSJ, and I have followed the model of LSJ in many respects. One difference is that where there is anything to say about etymologies I put it at the beginning. After giving the necessary information about forms and prosody, it is then logical to comment on the word’s formation (αἰδοῖος – αἰδώς, ἄτη – ἀάω), and to go further back if possible (ἀΐσσω ~ Sanskrit vevijyáte). If there were any Mycenaean evidence relating to any of these words, this would be the place for it, but for these particular items I do not know of any. I consider this to be the proper place for etymologies in a work of this kind, because although a word’s etymology does not necessarily illuminate its meaning in the historical period, it is good if possible to have an idea of where it is coming from, a starting-point to which attested usages might be related. In the case of ἆ there is no ‘etymology’ to be looked for: like many other interjections (ὤ, ἰώ, αἰαῖ, οἰοῖ, ἀτταταῖ, ἐἕ, etc.) it originates as an elementary cry, an involuntary noise wrung from the human frame by a particular physical or emotional stimulus. But such cries tend to become formalized and lexicalized in particular uses, and these we can classify. There is an interesting distinction to be noted between cases where ἆ remains outside the sentence and those where it is part of it. It is legitimate to ask whether all these ἆ’s are the same word that has various uses or whether we should recognize two or three different interjections, perhaps not identical in sound but all represented in texts (faute de mieux) by the letter α. For example, the remonstrating ἆ ἆ, τί δράσεις; in A III 2 may have sounded quite different from Philoctetes’ cries of pain in A III 3, and both may have sounded quite different from what comes out of Silenus’ mouth when he takes a gulp of Odysseus’ excellent wine, ἆ ἆ ἆ, presumably a representation of burping; this at least I have set apart in a section B, together with a couple of other outlandish vociferations that do not seem to belong with the main sections I–III. At the beginning of the other entries I make general comment on registers of use. Thus under αἰδοῖος I say ‘Poetic; not in drama after Aeschylus (except in sense B)’, sense B being αἰδοῖα ‘genitals’. Under ἀΐσσω I say ‘Vivid verb, not common in comedy and classical prose’; under ἄτη, ‘Not in comedy (unless Ar. Pax 605 cj.) or Attic prose’. After this preliminary information on forms, prosody, origins, and register I attempt a general formulation of the word’s essential meaning. Thus for αἰδοῖος I give, ‘Of persons to whom a sense of modesty, shyness, or inhibition attaches, whether (I) intrinsic or (II) felt by others’; this dichotomy then

Diminishing Returns and New Challenges

343

determines the structure of the article. For ἀΐσσω I give, ‘Of quick or unobstructed movement, dart, spring, fly up’; for ἄτη (a difficult one to summarize), ‘harm, ruin, disaster afflicting men’s fortunes, especially because of their own error; often conceived as an active (divine) agent, or as sent by gods’. When it comes to dividing the entry into sections and subsections, the lexicographer has considerable freedom to exercise his or her judgment. There are different ideologies: some favour making as many nice distinctions as possible, with a separate number or letter for each one, while others prefer to run as much as possible together in larger sections in which gentle progressions of meaning may be followed from one passage to another. I have taken a middle way. Under αἰδοῖος A I, is it justified to distinguish 1. of unmarried women, 2. of married women, 3. of goddesses? One may say no, it is all the same sense, but I think it is convenient to make the distinctions. With the goddesses it becomes difficult to say whether the application of the epithet is not also connected with the respect that they inspire in mortals, as in II 6, though it must be significant that the word is used much more of female than of male deities, as I have noted in that entry with a reference back to I 3. At the end of each entry, where appropriate, I have cited passages where the word occurs in a broken context and so cannot be assigned a particular interpretation. Under ἄτη I have also given a list of passages where it either occurs or might be thought to occur but which are too corrupt to make use of. Finally in each entry I list other, related words that occur in the corpus and that would have their own entries in the lexicon. The cross-references are needed because if we are looking for the fullest account of what a particular word means, the usages of related words are sometimes relevant. In the case of ἀΐσσω in particular I have not given full value: I have listed a dozen compounds, and the fact is that the preverbs, ἀνα-, ἀπο-, δια-, and so on, simply add an indication of the direction of the ἀΐσσειν without modifying the actual meaning of the verb, so that for a complete survey of its semantics I ought to have taken the usage of the compounds into account. But this is all I have done, and all I intend to do. My aim is only to give a glimpse of something that someone might aim at sometime in the future.

19.1. SPECIMEN ENTRIES IN AN IMAGINED POETIC LEXICON OF CLASSICAL GREEK ἆ, exclam. (mainly poet.). [Elementary cry, formalized in particular uses.] I. With nom. or voc. adj., commenting on addressee’s fortunes. So always in Hom., and only with δειλός voc.: ἆ δειλέ poor wretch, Λ 441, 452, Π 837, P 201, Ω 518, λ 618; ἆ δ. ξείνων ξ 361, ϕ 288; ἆ δειλώ Ρ 443, ϕ 86;

344

Martin L. West

ἆ δειλοί κ 431, υ 351; ἆ δειλὴ Πενίη O wretched Poverty, Thgn. 351, 649; with nom., ἆ τάλας ἀνήρ Sem. 7.76; ἆ δύσμορος, ἆ τάλαινα Bacch. 16.30. b. Of good fortune: ἆ μάκαρ ὅτ[ις oh happy the man who . . . , Hippon. 117.6, cf. Thgn. 1013, Choeril. Sam. 2.1 B.; ἆ τρισευδαίμ[ων ἀνήρ] Bacch. 3.10. II. In drama for other emotional cries: ἆ ποῖ ποτ᾽ ἤγαγές με; oh, where have you brought me? A. Ag. 1087; ἆ Ζήν O Zeus! A. Su. 162 (text. dub.: αζην cod.); ὀπποποῖ· ἆ μιαρέ S. Ichn. 197 (satyr.). 2. Protesting at another’s action or speech, with neg. injunction: ἆ, μὴ κόλαζε . . . τόνδε oh, don’t chastise him, S. OT 1147; ἆ, μηδαμῶς, μὴ πρὸς θεῶν μεθῆις βέλος id. Ph. 1300; ἆ, μὴ πρόκλαι᾽ ἄκοιτιν E. Alc. 526; cf. eund. Hipp. 503 (cj.), Ion 361, Hel. 445, [E.] Rh. 687, Ar. Pl. 127; so in Men. Sam. 134 with Austin’s reading ἆ μηδαμῶς. Once in prose, Pl. Hi. ma. 295a ἆ, μὴ μέγα ὦ ῾Ιππία λέγε. III. In the preceding uses initial ἆ is part of a sentence. In the follg. it is more typically independent and extra metrum: 1. Noting a new development: ἆ, | οἵδ᾽ οὐκ ἀϕιᾶσι . . . πέπλων ah, they don’t let go of my robe, E. HF 629; on hearing that Aesch. and Eur. are in dispute in Hades, Ar. Ran. 759; marking a sudden new thought: ἆ· | βούληι σϕ᾽ ἐν ὄρεσι συγκαθημένας ἰδεῖν; E. Ba. 810. b. Doubled: ἆ ἆ ἴδου ἴδου oh, oh, look, A. Ag. 1125; ἆ ἆ· | σμοιαὶ γυναῖκες αἵδε, as Orestes sees the Furies, id. Ch. 1048; cf. E. Alc. 28, Hec. 1069, Ba. 586, 596; ἆ ἆ ἔα ἔα [A.] PV 114. 2. Doubled, in remonstrance: ἆ ἆ, τί δράσεις; E. Andr. 1076, Cycl. 565; ἆ ἆ, τί δῆτα μέλλετε; id. Or. 275, cf. Ar. V. 1379; ἆ ἆ· | ποῖ ποῖ σὺ ϕεύγεις; id. Th. 688; ἆ ἆ, διά μ᾽ ὀλεῖτε (by waking Heracles), E. HF 1052, cf. Or. 145; with neg. injunction (as in II 2), ἆ ἆ, | τὴν δᾶιδα μή μοι πρόσϕερε Ar. Pl. 1052; cf. E. Med. [1056], Or. 1598. 3. As a cry of pain: ἆ ἆ ἐἕ [A.] PV 566; ἆ ἆ ἆ ἆ S. Ph. 732, 739, 782, [E.] Rh. 749, 798. B. In the following passages ἆ or ά represents noises made by unseemly characters: ό ό ο ά ά ά A. Su. 825 [cod. M] perhaps signifies the Egyptians’ evil laughter, ho ho ho, ha ha ha; the gramm. tradition (Hsch. α 2, Eust. 855.19, etc.) attests aspirated ἅ ἅ in this sense. ὗ ὗ ὗ ψ ψ ἆ ἆ S. Ichn. 176 are vociferations of excited satyrs. ἆ ἆ ἆ E. Cycl. 157 may represent Silenus’ burping after a gulp of wine; Odysseus enquires, ‘did it tinkle nicely down your throat?’. • Related word: ἄζω (B). αἰδοῖος, -ᾱ/-η, -ον; comp. -ότερος λ 360, sup. -ότατος Pi. P. 5.18 (rest. metri gr. for mss. -έστατον); -έστατος Alcm. 2 (iv) 4, Pi. O. 3.42. Αdv. -ως τ 243.

Diminishing Returns and New Challenges

345

[Deriv. of αἰδώς, q.v.: *aidos-yo-.]:—Of persons to whom a sense of modesty, shyness, or inhibition attaches, whether (I) intrinsic or (II) felt by others; rarely of things (only 5th c.). Poetic; not in drama after Aesch. (exc. in sense B). I. modest, careful of propriety, typical epith. of women. 1. Of unmarried women, παρθένος αἰ. B 514, cf. Hes. Th. 572, Op. 71; κούρην τ᾽ α]ἰδοίην id. fr. 180.13. 2. Of married women, proper, respectable, ἔνθα δὲ γαμβροὶ κοιμῶντο Πριάμοιο παρ᾽ αἰδοίηις ἀλόχοισιν Z 250, cf. Φ 460, κ 11, Hy. 2.148 (v.l.); ὦ γύναι αἰ. Λαερτιάδεω ᾽Οδυσῆος ρ 152 (= τ 165, 262, 336, 583); αἰ. παράκοιτις Νέστορος γ 451, cf. 381, Hy. 2.343, [Hes.] Sc. 14, 46; Λ[αμπι]τὸ αἰ. CEG 66 (epitaph); αἰδοίης ἑκυρῆς ὀπὸς ἔκλυον X 451; μητρὶ πάρ᾽ αἰδοίηι θ 420, cf. Hom. Epigr. 4.2. Of a servant woman, αἰ. ταμίη α 139 (= δ 55, η 175, κ 371, ο 138, ρ 94, 259). 3. Of goddesses: Hera, Διὸς αἰ. παράκοιτις Φ 479, cf. Hy. 5.44, Tyrt. 23.13; Themis, Hes. Th. 16; Hebe, αἰδοίην θέτ᾽ ἄκοιτιν ib. 953; Demeter, id. Op. 301, Hy. 2.374; Demeter and Persephone, σεμναί τ᾽ αἰ. τε ib. 486; Hestia, Hy. 5.21, 29.10; Maia, Hy. 4.5, 18.5; Aphrodite, Hes. Th. 194, Hy. 6.1; Artemis, Hy. 27.2; Athena, Hy. 28.3; Leto, ἤδη γὰρ αἰδοί[ᾱι (v.l. αἰδοῖ[αι) . . . ἐ]βάρυνον ὠ[δ]ῖνες Sim. 102.3; Charis (as one who bestows her own qualities on men, cf. II 7), Pi. O. 6.76. Of nymphs, παῖδες αἰ]δο[ῖ]αι ποταμοῦ (the Asopos), Bacch. 9. 65. 4. Of men, proper in their conduct, αἰδοίως δ᾽ ἀπέπεμπον ἐϋσσέλμου ἐπὶ νηός (my guest Odysseus) τ 243; αἶψα πόδες με ϕέροιεν ἐς αἰδοίων πόλιν ἀνδρῶν Hom. Epigr. 2.2, cf. 6.6; ξείνων αἰδοῖοι λιμένες (you men of Acragas), proper welcomers of visitors, Emp. 112.3; αἰ. πρόξενον respectful (towards suppliants), A. Su. 491. Transf., δέξασθ᾽ ἱκέτην . . . στόλον αἰδοίωι πνεύματι χώρας receive us suppliants with your land’s (duly) respectful breeze (but drive the Egyptians out to sea), ib. 29. 5. bashful, inhibited, diffident, κακὸς δ᾽ αἰ. ἀλήτης a shy beggar is no good, ρ 578. II. inspiring or deserving respect, esp. of persons in certain social categories. 1. Of kings and queens, τὸν δ᾽ οὔ τι προσέϕη . . . αἰδεσθεὶς βασιλῆος ἐνιπὴν αἰδοίοιο abashed by the respectworthy king’s reproach, Δ 402; (Calliope) καὶ βασιλεῦσιν ἅμ᾽ αἰ. ὀπηδεῖ Hes. Th. 80, cf. 434, fr. 361; αἰδοίου βασ[ιλῆος id. fr. 43a.89; ἵν᾽ αἰ. βασίλεια where your respected queen is, σ 314; τὸν ἐμὸν αἰ. πόσιν A. Ag. 600. Of the Areopagus council, id. Eu. 705. 2. Of suppliants, ἀντί τοί εἰμ᾽ ἱκέταο . . . αἰδοίοιο I am in effect your suppliant whom you should respect, Φ 75; (Zeus) ὅς θ᾽ ἱκέτηισιν ἅμ᾽ αἰ. ὀπηδεῖ η 165, 181; before a deity, αἰ. μέν τ᾽ ἐστὶ καὶ ἀθανάτοισι

346

Martin L. West

3.

4.

5.

6.

7.

θεοῖσιν ἀνδρῶν ὅς τις ἵκηται ἀλώμενος . . . ἱκέτης δέ τοι εὔχομαι εἶναι ε 447. Transf., of suppliants’ appeals, πολλῶν ἄκουσον τέρματ᾽ αἰδοίων λόγων A. Su. 455, cf. 194. Of guests, Ζεὺς ξείνιος, ὃς ξείνοισιν ἅμ᾽ αἰδοίοισιν ὀπηδεῖ ι 271; εἵνεκα γὰρ ξείνοιο τάδ᾽ αἰδοίοιο τέτυκται this is done for the sake of our respected guest θ 544, cf. τ 191, 316; ἔϕαγόν τ᾽ ἔπιόν τε καὶ αἰδοίοισιν ἔδωκα I have eaten and drunk and given to the respectworthy (sc. guests and suppliants), o 373 (but cf. below, B). Generally, enjoying public respect, οὔ τις αἰ. μετ᾽ ἀστῶν οὐδὲ περίϕημος θανὼν γίνεται, Arch. 133.1; ἀνδρὸς αἰ. πρόσοψιν θηκάμενος assuming the aspect of a respectworthy man, Pi. P. 4.29; αἰ. μὲν ἦν ἀστοῖς ὁμιλεῖν id. I. 2.37; στείχετ᾽ αἰ. γέροντες A. Ag. 1657. With ref. to interpersonal relationships, (loved and) respected, αἰ. τέ μοί ἐσσι, ϕίλε ἑκυρέ, δεινός τε Γ 172; ϕίλον περ ἐόντα καὶ αἰ. Μενέλαον νεικέσω . . . ὡς εὕδει K 114; αἰ. νεμεσητός, ὅ με προέηκε it is a man worthy of respect and prone to disapproval who sent me, Λ 649; ὥς κεν Φαιήκεσσι ϕίλος . . . γένοιτο δεινός τ᾽ αἰ. τε θ 22, cf. λ 360, ξ 234, τ 254, Sol. 13.6. Cf. below, 6. Of the gods collectively, θεῶν γένος αἰδοῖον Hes. Th. 44; of individuals, mostly of female deities, see I 3; of Zeus as protector of suppliants, A. Su. 192. With ref. to personal relationships among gods, as in 5: (if I could reconcile Oceanus and Tethys,) αἰεί κέ σϕι ϕίλη τε καὶ αἰ. καλεοίμην Ξ 210; Thetis is αἰ. τε ϕίλη τε to Hephaestus and his wife, Σ 386 = 425, cf. 394 (and so is the male Hermes to Calypso in the derivative passage ε 88); the παρθένος Dike is κυδρή τ᾽ αἰ. τε θεοῖς (v.l. θεῶν), Hes. Op. 257. Of things: κτεάνων δὲ χρυσὸς αἰδοιέστατος (v.l. -ον) most respected or awesome, Pi. O. 3.42; δίδοι τέ οἱ αἰδοίαν χάριν καὶ ποτ᾽ ἀστῶν καὶ ποτὶ ξείνων i.e. grant him to be respected both by his fellow-citizens and by others, id. O. 7.89; αἰδοιότατον γέρας that most reverend dignity (of kingship), id. P. 5.18; ἥρωες αἰδοίαν ἐμείγνυντ᾽ ἀμϕὶ τράπεζαν θαμά the solemn festive table, id. fr. 187. B. Neut. pl. as noun, (male) private parts, αἰδοίων τε μεσηγὺ καὶ ὀμϕαλοῦ N 568; αἰ. γονῆι πεπαλαγμένα Hes. Op. 733; αἱματόεντ᾽ αἰ. Tyrt. 10.25 (= αἰδῶ X 75); Ar. Nu. 978, V. 578; also in prose, Hclt. B 15, Hdt., etc.—o 373 (above, II 3) may be adapted from a verse in which αἰδοίοισιν had this sense (Eikasmos 23, 2012, 11–14).

In broken contexts: ]εναιδοιων[ Arch. 145.5; ]α̣ι̣δ̣ο̣ιου[ Sim. 175.3. • Related words: αἰδώς, αἰδόϕρων, ἀναιδής, ἀναίδεια, αἴδομαι/αἰδέομαι, ἐπαιδέομαι, καταιδέομαι.

Diminishing Returns and New Challenges

347

ἀΐσσω, later also contr. ἄισσω, Att. ἄιττω; fut. ἀΐξω, ἄιξω, aor. ἤϊξα, ἦιξα, Lesb./Dor. ἄϊξα; iter. ἀΐξασκε Ψ 369; occasional mid. and pass. forms, ἀΐσσοντο, ἀΐξασθαι, ἠΐχθη, etc. In trag. uncontr. forms occur six times in lyrics, twice in anap., once in trim. (E. Hec. 31). Prosody: ᾱ in hex. verse exc. Φ 126; in 5th c. ᾰ where uncontr. and unaugm., exc. E. Tr. 156, Trag. ad. 668.17, and perh. Pi. N. 8.40. [Prob. *wai-wik-yō, though no trace of initial ϝ; cf. perh. Skt. vevijyáte ‘recoil’; in any case ᾱϊ- will be from αιϊ (Wackernagel, Kl. Schr. 587), and a pap. gives επαιισσει at X 142.]:—Of quick or unobstructed movement, dart, spring, fly up, etc. Vivid verb, not common in com. and class. prose. I. With animate subjects. 1. In battle contexts: rush forward to fight, of one or more warriors, ἐκ δὲ τὼ ἀΐξαντε πυλάων πρόσθε μαχέσθην M 145, cf. O 694, P 734, Mimn. 13a.2 (rest.), Pi. fr. 33a.3, Bacch. 13.144; ἦιξαν δράμημα δεινὸν ἀλλήλοις ἔπι E. Ph. 1379; in a chariot, ἵπποις ἀΐσσων P 460; rush to arms, δμῶες δ᾽ ἰδόντες εὐθὺς ἦιξαν ἐς δόρυ E. El. 844, cf. Ion 997 (cj.), Ph. 1466, IA 80; run after a fleeing opponent, μεταδρομάδην ἔλασ᾽ ὦμον ϕασγάνωι ἀΐξας E 81; spring down from one’s chariot, καθ᾽ ἵππων Λ 423, cf. Ζ 232, Y 401; speed away in flight, ϕόβονδε Ρ 579 [but at A. Pers. 470 ἤϊξ᾽ ἀκόσμωι ξὺν ϕυγῆι is an inferior v.l. for ἵησ᾽]; of dodging a spear, ἀλεύατο κῆρα μέλαιναν λικριϕὶς ἀΐξας darting sideways, Ξ 463; of swift action to cut reins, ἵπποιο παρηορίας ἀπέταμνε ϕασγάνωι ἀΐσσων Θ 88, cf. Π 474; middle, πυλάων . . . ἀντίον ἀΐξασθαι X 195. 2. In other contexts: speed, run, ἐπ᾽ ἄμαξαν ἐΰτροχον ἀΐξασαι, of Andromache and Hecuba running to meet the wagon bearing Hector’s body, Ω 711; of Achilles running from Scamander, ἤϊξεν πεδίοιο ποσὶ κραιπνοῖσι πέτεσθαι Φ 247, cf. 254, 303; of the Boreads pursuing the Harpies round the world, τοὺς πάντα]ς πέρι κύκλωι ἐθύνεον ἀΐσσοντες Hes. fr. 150.20; of girls on an errand, Hy. 2.178; of a stadium runner, dub. in Bacch. 10.23; Κένταυρος ἄϊ[ξ᾽ ἐπὶ νύμϕαν id. fr. 64.16; εὐθέως δ᾽ ἐγὼ κατ᾽ ἴχνος ἄισσω I am hurrying to track him, S. Aj. 32; μάτην ἦιξα I have hurried in vain, E. IA 742; of Io’s frenzied wanderings, [A.] PV 676, 837; of Atalanta, ἄισσουσα δ᾽ ἐξέλαμψεν ἀστραπῆς δίκην Trag. ad. 14a; of wild Bacchic running and dancing, E. Ba. 147, 1090; of Pentheus running about in distraction, ἦισσ᾽ ἐκεῖσε κἆιτ᾽ ἐκεῖσε ib. 625, cf. 631; ἀπ᾽ ἀγχόνας ἀΐξασα Sophr. 8; θεύσει γὰρ ἄιξας εἰς τὸ βουλευτήριον Ar. Eq. 485, cf. V. 120; dub. lect. at Nu. 1299. b. of hurried stage entries and exits, σὺ δ᾽ ἐκτὸς ἦιξας πρὸς τί; why have you come running out? S. El. 1402, cf. OT 1074, Tr. 396, OC 890, 1499, E. Su. 1065, IA 12. c. leap up from sleep (?), ]εν ὕπνον ἀΐξας ἄϕν[ω Trag. ad. 668.16. 3. Of gods speeding down from Olympus, βῆ δὲ κατ᾽ Οὐλύμποιο καρήνων ἀΐξασα B 167 (= Δ 74, H 19, Χ 187, Ω 121), cf. Ξ 225 (= T 114), O 150;

348

Martin L. West

τῶι (a falling star) εἰκυῖ᾽ ἤϊξεν ἐπὶ χθόνα Δ 78; from Onchestus to Pylos, Hy. 4.215; pass., ἐς οὐρανὸν ἀϊχθήτην Ω 97. 4. Of chariot racers, χαλκῆς ὑπαὶ σάλπιγγος ἦιξαν S. El. 711. 5. Of birds and animals: δεξιὸς ἀΐξας ὑπὲρ ἄστεος, of an eagle, Ω 320; of Ceyx as kingfisher, οὔτε . . . παύεται ἀΐσσω[ν . . . ] ἵεται ᾽Αλκυόνη[ς Hes. fr. 10a.95; horses drawing a chariot o 183; a fleeing deer, Λ 118; charging boars, M 148; hunting-dogs, O 580, P 726; so in middle, ἀμϕὶ δέ τ᾽ ἀΐσσονται Λ 417. 6. Of ghosts: τοὶ δὲ σκιαὶ ἀΐσσουσιν but the others dart about as shadows, κ 495; ἢ τὸν ἐς ῞Αιδα μελάγχρωτα πορθμὸν ἄιξω τάλας; am I to fly the dark crossing to Hades? E. Hec. 1106; of the unburied Polydorus, νῦν δ᾽ ὑπὲρ μητρὸς ϕίλης ῾Εκάβης ἀΐσσω, σῶμ᾽ ἐρημώσας ἐμόν, τριταῖον ἤδη ϕέγγος αἰωρούμενος ib. 31. 7. make rapid movement with a weapon or other implement, ϕασγάνωι ἀΐξας K 456; ἦισσον δὲ λόγχαις E. Ph. 1382; ἀΐσσων ὧι ἔγχει, of Odysseus defending himself under pressure, Λ 484; of judges gesturing with sceptres, τοῖσιν ἔπειτ᾽ ἤϊσσον, ἀμοιβηδὶς δὲ δίκαζον Σ 506. 8. Metaph., of eagerly following an impulse, ἐγὼ καὶ διὰ μούσας καὶ μετάρσιος ἦιξα E. Alc. 962; οὐδ᾽ ἦιξας εἰς ἔρευναν ἐξευρεῖν γονάς; did you not hasten to investigate your birth? id. Ion 328; οἷ δ᾽ ἦιξας ὀρθῶς, τοῦτο κἄμ᾽ ἔχει πόθος ib. 572; εἰς δὲ τὴν αὑτοῦ ϕύσιν ἄιξας ἐλεύθερόν τι τολμήσει πονεῖν Men. Epit. 323. b. Of mental turmoil, πλαγκτὰ δ᾽ ὡς εἴ τις νεϕέλα πνευμάτων ὕπο δυσχίμων ἀΐσσω E. Su. 962 (cf. below, ΙΙ 3). II. With inanimate subjects: of spears, fly, θαμέες γὰρ ἄκοντες ἀντίον ἀΐσσουσι θρασειάων ἀπὸ χειρῶν Λ 553 = P 662, cf. E 657, Υ 277; so in pass., ἐκ δέ μοι ἔγχος ἠΐχθη παλάμηϕιν ἐτώσιον Γ 368, cf. E 854; of racing chariots bouncing off the ground, ἅρματα δ᾽ ἄλλοτε μὲν χθονὶ πίλνατο . . . , ἄλλοτε δ᾽ ἀΐξασκε μετήορα Ψ 369; pass., ἐκ δ᾽ ἄρα χειρῶν ἡνία ἠΐχθησαν the reins flew out of his hands, Π 404. 2. Of a speeding ship, ἐμὲ δὲ πόντιον σκάϕος ἀΐσσον πτεροῖσι πορεύσει, E. Tr. 1086. 3. Of winds, ἄτερ στεροπᾶς ἄιξας ὀξὺς Νότος S. Aj. 258; cf. Ibyc. in III 1 below. Of other things seen above the earth: a storm-driven cloud, cf. I 8b above; rising smoke, καπνὸν δ᾽ οἶον ὁρῶμεν ἀπὸ χθονὸς ἀΐσσοντα κ 99; sparks flying up, πυκνοὶ δ᾽ ἄιττουσιν ῾Ηϕαίστου κύνες κούϕως πρὸς αἴθραν Alexis fr. 153.16; a scent, ὀσμὴ δὲ πρὸς μυκτῆρας ἠρεθισμένη ἄισσει Eubul. fr. 75.10; a star, ἐγγὺς τῆς ἑπταπόρου Πλειάδος ἄισσων ἔτι μεσσήρης E. IA 8; the light of beacon fires, ὑψόσε δ᾽ αὐγὴ γίνεται ἀΐσσουσα περικτιόνεσσιν ἰδέσθαι Σ 212. 4. τεμάχη δ᾽ ἄνωθεν αὐτόματα . . . εἰς τὸ στόμ᾽ ἄιττει the meat slices fly down into our mouths by themselves, Metag. fr. 6.10.

Diminishing Returns and New Challenges

349

5. Middle, of a horse’s mane flying up, ἀμϕὶ δὲ χαῖται ὤμοις ἀΐσσονται Z 510 = O 267; the same phrase of girls’ hair, Hy. 2.178; of Oedipus’, κόμη δι᾽ αὔρας ἀκτένιστος ἄισσεται S. OC 1261. b. Of (agile) arms springing from shoulders, τῶν ἑκατὸν μὲν χεῖρες ἀπ᾽ ὤμων ἀΐσσοντο Hes. Th. 150 = 671; οὐ γὰρ ἀπὸ νώτοιο δύο κλάδοι ἀΐσσοντο, οὐ πόδες, οὐ θοὰ γοῦν᾽, οὐ μήδεα λαχνήεντα Emp. 29.1, 134.2. III. With abstract subjects: (ἔρος) ὑπὸ στεροπᾶς ϕλέγων Θρηΐκιος Βορέας ἀΐσσων παρὰ Κύπριδος love rushes on me from Aphrodite like Boreas ablaze with lightning, Ibyc. 286.10; ἀΐσσει δ᾽ ἀρετά, χλωραῖς ἐέρσαις ὡς ὅτε δένδρεον quality springs up like a tree from fresh dews, Pi. N. 8.40 (cj.); ἄ[ϊξον, ὦ] σεμνοδότειρα Φήμα, ἐς Κ[έον Βacch. 2.1 (rest.); μεγάλαν . . . δόμοισι βλάβαν νέων ἀΐσσουσαν γάμων great harm hurtling on the house from the new marriage, S. Tr. 843 (cj. for ἀϊσσόντων); μόχθος δ᾽ ἐκ μόχθων ἄισσει E. IT 191. 2. Of thoughts and feelings: ὡς δ᾽ ὅτ᾽ ἂν ἀΐξηι νόος ἀνέρος, of a man projecting himself in imagination to distant places, O 80; δι᾽ ἐμᾶς ἦιξέν ποτε νηδύος ἅδ᾽ αὔρα E. Hipp. 165; διά μου κεϕαλᾶς ἄισσουσ᾽ ὀδύναι ib. 1351; διὰ δὲ στέρνων ϕόβος ἀΐσσει id. Tr. 156. IV. Transitive: καὶ πρὸς τί δυσλόγιστον ὧδ᾽ ἦιξεν χέρα; why did he speed his hand so irrationally? S. Aj. 40; ἔτυχον . . . αὔραν ῾Ελένας ῾Ελένας εὐπᾶγι κύκλωι πτερίνωι πρὸ παρῆιδος ἀΐσσων I was sending a breeze across Helen’s cheek with a fan, E. Or. 1429. In broken context: ἄϊξεν δ[ Bacch. 24.14. • Compounds: ἀν-, ἀπ-, δι-, εἰσ-, ἐν-, ἐξ-, ἐπ-, κατ-, μετ-, παρ-, προσ-, ὑπ-. Related words: ἀϊκή, κορυθάϊξ, τριχάϊξ (?). ἄτη ( ), Dor. ἄτα, Lesb. ἀυάτα (˘˘ , representing ἀϝάτᾱ); Pi. uses both ἄτα and ἀυάτα. In Hom. and Hes. ἄτη is usu. placed where ἀάτη would scan, but at line-end in T 88 and as v.l. in Γ 100, Z 356, Ω 28. †ἄτη with short ᾰ is unacceptable at Arch. 127. [From ἀ(ϝ)άω; -τᾱ is typical of deverb. abstracts, cf. Buck–Petersen 541.]:—harm, ruin, disaster afflicting men’s fortunes, esp. bec. of their own error; often conceived as an active (divine) agent, or as sent by gods. Not in com. (unless Ar. Pax 605 cj.) or Att. prose. I. In explicit association with ἀάω: Ζεῦ πάτερ, ἦ ῥά τιν᾽ ἤδη . . . βασιλήων τῆιδ᾽ ἄτηι ἆσας you have blighted with this blight, Θ 237; οὐ δύναμαι λελαθέσθ᾽ ἄτης, ἧι πρῶτον ἀάσθην· ἀλλ᾽ ἐπεὶ ἀασάμην καί μοι ϕρένας ἐξέλετο Ζεύς, . . . T 136; Ἄτη ἣ πάντας ἀᾶται T 91, cf. 129. 2. As an active force: τὸν δ᾽ ἄ. ϕρένας εἷλε Π 805, cf. Ω 480; Πέρ]γαμον δ᾽ ἀνέ[β]α ταλαπείριο[ν ἄ]τα Ibyc. S151.8; μή τί σε θυμοπληθὴς δορίμαργος ἄ. ϕερέτω A. Se. 687; αἰανὴς δ᾽ ἄ. διαϕέρει τὸν αἴτιον id. Ch. 68; πρὸς ἄτης θηραθεῖσαι [A.] PV 1072; ὑπ᾽ ἄτης ζεῦγλαν ἀσχάλλει πεσών E. fr. 285.10.

350

Martin L. West

3. Personified, among children of Eris, Δυσνομίην τ᾽ Ἄ. τε συνήθεας ἀλλήληισιν Hes. Th. 230; οὐδέ ποτ᾽ ἰθυδίκηισι μετ᾽ ἀνδράσι Λιμὸς ὀπηδεῖ οὐδ᾽ Ἄ. id. Op. 231; Ate runs ahead harming men, Entreaties follow behind to put things right, I 504–7, cf. T 92–4; but if someone rejects them, they entreat Zeus, τῶι Ἄτην ἅμ᾽ ἕπεσθαι, ἵνα βλαϕθεὶς ἀποτείσηι I 512; T 91, 129, see above, 1; thrown out of heaven after afflicting Zeus, T 126; Ἄτης ἂν λειμῶνα Emp. 121.4, cf. d 17 M.–P. (rest.); σαίνουσα τὸ πρῶτον παράγει βροτὸν εἰς ἀρκύστ Ἄτα A. Pe. 99; οἷον δέδορκεν Ἄ. ib. 1007; ἕστακε δ᾽ Ἄτας τρόπαιον ἐν πύλαις id. Se. 957; Πειθώ, προβούλου παῖς . . . Ἄτας id. Ag. 386, cf. 735, 771, 1433; οὐκ ἀτρίακτος Ἄ. perh. Ate never fails to win, id. Ch. 338; Ἄ. δ᾽ ἀποστατεῖ ϕίλων ib. 824; τότε δ᾽ (after too much drink) Ὕβριος αἶσα καὶ Ἄτης γίνεται ἀργαλέη, κακὰ δ᾽ ἀνθρώποισιν ὀπάζει Pany. 20.8, cf. 22. 4. Laid on men by gods: Ζεύς με . . . ἄτηι ἐνέδησε βαρείηι Β 111 (= Ι 18); Ζeus, Moira, Erinys, οἵ τέ μοι . . . ϕρεσὶν ἔμβαλον ἄγριον ἄ. Τ 88; ἄτην δὲ μετέστενον, ἣν Ἀϕροδίτη δῶκε δ 261; ἐπὶ ϕρεσὶ θῆκε . . . Ἐρινύς o 233; Ζεῦ πάτερ ἠδ᾽ ἄλλοι μάκαρες θεοί . . . ἦ με μάλ᾽ εἰς ἄτην κοιμήσατε μ 372; ἄ. δ᾽ ἐξ αὐτῶν ἀναϕαίνεται, ἣν ὁπότε Ζεὺς πέμψηι τεισομένην, ἄλλοτε ἄλλος ἔχει Sol. 13.75, cf. A. Ch. 383; (some god) δᾶμον μὲν εἰς ἀυάταν ἄγων Alc. 70.12; ὦ πολιοῦχοι θεοί, τοῖσι μὲν ἔξω πύργων . . . ῥίψοπλον ἄταν ἐμβαλόντες i.e. make them lose heart and suffer a shield-abandoning disaster, A. Se. 315; Ἐρινὺν . . . ἄτην ἑτέραν ἐπάγουσαν ἐπ᾽ ἄτηι id. Ch. 403, cf. Eu. 376; οἷς γὰρ ἂν σεισθῆι θεόθεν δόμος, ἄτας οὐδὲν ἐλλείπει γενεᾶς ἐπὶ πλῆθος ἕρπον S. Ant. 584; ὅτωι ϕρένας θεὸς ἄγει πρὸς ἄταν ib. 624, cf. 625; cf. also Crit. 43 F 5.8. 5. Without ref. to gods, condition incurred through impaired judgment, entailing harmful consequences: γνῶι δὲ καὶ Ἀτρείδης . . . ἣν ἄτην that he may recognize his (self-inflicted) harm, A 412 (= Π 274), cf. ψ 223; Ἀλεξάνδρου ἕνεκ᾽ ἄτης Γ 100, Z 356, Ω 28 (v.l. in all three places); ὃ δὲ ϕρεσὶν ἧισιν ἀασθεὶς ἤϊεν ἣν ἄτην ὀχέων ϕ 302, cf. [Hes.] Sc. 93; αὐτὸς ἔτεισε . . . οὐδὲ . . . ἄτην ἐξοπίσω παισὶν ἐπεκρέμασεν he pays for his wrongdoing himself and does not leave ruin hanging over his children, Thgn. 206; ταχέως δ᾽ ἀναμίσγεται ἄτηι Sol. 13.13; Εὐνομίη . . . αὑαίνει δ᾽ ἄτης ἄνθεα ϕυόμενα id. 4.35; τὰν ἀκόρεστον ἀυάταν Lesb. inc. 25b; ἀλλά νιν ὕβρις εἰς ἀυάταν ὑπεράϕανον ὦρσεν Pi. P. 2.28; κόρωι δ᾽ ἕλεν ἄταν ὑπέροπλον id. O. 1.57; ἔσχε τοι ταύταν μεγάλαν ἀυάταν . . . λῆμα Κορωνίδος id. P. 3.24; ὕβρις γὰρ ἐξανθοῦσ᾽ ἐκάρπωσε στάχυν ἄτης A. Pe. 822; ταχεῖα δ᾽ ἄ. πέλει id. Ag. 1124; πρώταρχον ἄτην the original error in the family, ib. 1192; κεκόλληται γένος πρὸς ἄται ib. 1566; ἄτης ἄρουρα θάνατον ἐκκαρπίζεται id. Se. [601]; δαιμονῶντες ἄται ib. 1001; of a deserved doom rightfully inflicted by a man, πέραιν᾽ οὐκ ἐπίμομϕον ἄταν id. Ch. 830, cf. 836; ὤι μοι δύσϕρονος ἄτας S. OC 202; ἄται ἁλοὺς

Diminishing Returns and New Challenges

351

ἐϕόνευσα ib. 547; ἄτηι συγκατέζευκται κακῆι id. Aj. 123; οὐκ ἀλλοτρίαν ἄ. ἀλλ᾽ αὐτὸς ἁμαρτών id. Ant. 1260; ἐμάνην, ἔπεσον δαίμονος ἄτηι (v.l. αἴσηι) E. Hi. 241, cf. 276; μία δ᾽ ἀμϕοτέρους ἄ. πατέρων διέκναισεν id. El. 1307; γνώμας ἐξέσταν, αἰαῖ, πίπτω δ᾽ εἰς ἄ. id. IA 137. II. Generally, disastrous situation, ruin, ἐς μεγάλην ἄτην . . . ἔπεσεν Sol. 13.68; οὐδεὶς ἑκὼν πονηρὸς οὐδ᾽ ἄταν ἔχων no one chooses to suffer toil and calamity, Epich. 66.2, cf. Thgn. 103, A. Ch. 467; σαίνων ποτὶ πάντας ἄταν πάγχυ διαπλέκει Pi. P. 2.82; ϕαινομέναν δ᾽ ἄρ᾽ ἐς ἄταν σπεῦδεν ὅμιλος ἱκέσθαι id. N. 9.21; ἴδε . . . βαθὺν εἰς ὀχετὸν ἄτας ἵζοισαν ἑὰν πόλιν id. O. 10.37; ἄτης δ᾽ ἄβυσσον πέλαγος . . . τόδ᾽ εἰσβέβηκα A. Su. 479, cf. [A.] PV 886; εἰς ἀπέραντον δίκτυον ἄτης ἐμπλεχθήσεσθ᾽ ὑπ᾽ ἀνοίας ib. 1078, cf. A. Ag. 360; ἄτης θύελλαι ζῶσι ib. 819; ἄτης λαθραίου τεύξεται, of Agamemnon, ib. 1230, cf. 1524; ἄλλην τιν᾽ ἄτης ἀντ᾽ ἐμοῦ πλουτίζετε make some other woman affluent in disaster, ib. 1268; ποῖ καταλήξει μετακοιμισθὲν μένος ἄτης id. Ch. 1076; cf. also Ag. 643; ἄτηι πατάξαι θυμόν S. Ant. 1097; στεναγμός, ἄ., θάνατος id. OT 1284; δακέθυμος ἄ. id. Ph. 706, cf. Tr. 1082; cf. also eund. Aj. 195, 307, 363, 643, 908, 976, 1189, El. 936, 1002, 1298, OT 165, Ant. 185, 614, Tr. 850, 1001, 1104, 1274, OC 526; αἱματηρὸν ἄ., of scratched cheeks, E. Or. 962; cf. also eund. Alc. 91, Med. 279, 979, 988, Hi. 1149, 1289, Hec. 688, Hcld. 607, HF 917, Tr. 137, 163, 530, 1314, Ph. 343, fr. 460.1, 752g.31, Trag. ad. 655.47. b. In pl., afflictions, calamities, αἰεὶ δ᾽ ἀμβολιεργὸς ἀνὴρ ἄτηισι παλαίει Hes. Op. 413, cf. 216; ὧιτινι μὴ θυμοῦ κρέσσων νόος, αἰὲν ἐν ἄταις . . . καὶ ἐν μεγάλαις κεῖται ἀμηχανίαις Thgn. 631; πολεμοϕθόροισιν ἄ. A. Pe. 654, cf. Ag. 730; δυσχειμέρους ἄ. ὑϕ᾽ ἧπαρ θερμόν id. Ch. 272; καθαρμοῖσιν ἀτᾶν ἐλατηρίοις ib. 968 (cj.); ἔρωτας ἄταισι συννόμους βροτῶν ib. 598; ϕίλων ἄταισι ποντίαισι id. Pe. 1037; τόνδε . . . δειναὶ κυματοαγεῖς ἆται κλονέουσιν ἀεὶ ξυνοῦσαι S. OC 1244; ἄτας τὰς ἐμάς S. Aj. 848; μὴ τίκτειν σ᾽ ἄταν ἄταις id. El. 235; ἰὼ ματρῶιαι λέκτρων ἆται id. Ant. 863; cf. also A. Ag. 1283, Eu. 983, S. El. 215, 224, OT 1205, E. Med. 129, HF 1284, Tr. 121, IT 148, Or. 988. 2. Concrete, of detested persons, embodiment of disaster, pest, τὰν μελανόζυγ᾽ ἄταν, of the Egyptians, A. Su. 530; τρέϕων δύ᾽ ἄτα κἀπαναστάσεις θρόνων S. Ant. 533, cf. OC 532; Πάρις οὐ γάμον ἀλλά τιν᾽ ἄταν ἀγάγετ᾽ εὐναίαν ἐς θαλάμους Ἑλέναν E. Andr. 103; of the Wooden Horse, Δαρδανίας ἄ. id. Tr. 535; of the Sphinx, δαιμόνων τις ἄ. id. Ph. 1066. 3. In econ. sense, detriment, loss, opp. profit, ἄτης καὶ κέρδεος . . . θεοὶ δώτορες ἀμϕοτέρων Thgn 133; χρημάτων μὲν ἐκ δόμων πορθουμένοις γένοιτ᾽ ἂν ἄλλα κτησίου Διὸς χάριν ἄτης γε μείζω A. Su. 444; opp. κέρδη S. OC 93; Ἐγγύα Ἄτας θυγάτηρ [Epich.] 257, cf. Cratin. Jun. 12.3. b. Ιn pl., κακὰ κέρδεα ἶσ᾽ ἄτηισιν Hes. Op. 352.

352

Martin L. West 4. corruption of gold, silver, χρυσοῦ κιβδήλοιο καὶ ἀργύρου ἀνσχετὸς ἄ. the fault of false gold and silver can be (detected and) kept in check, Thgn. 119.

In broken contexts: Alc. 10.7 (pl.), Ad. iamb. 14.1, S. fr. 221.16, E. fr. 370.48, 822.14. The follg. passages are corrupt: Hom. Epigr. 8.1 (v.l.), Arch. 127, Pi. P. 11.54, A. Su. 110, 164, 850, 886, Ch. 74, S. Ant. 4, E. IT 226, 1117. • Related words: ἀάω, ἀτάομαι, ἀτηρός, ἀτηρία, ἄνατος, ἀνατ(ε)ί.

20 Βάπτω An Illustration of the State of our Ancient Greek Dictionaries Anne Thompson

Words in the corpus of texts making up Thesaurus Linguae Graecae [TLG]¹ have links to headwords in several dictionaries. TLG released a digitized version of Liddell and Scott’s Greek-English Lexicon [LSJ] in 2011², and since then Cunliffe’s Lexicon of the Homeric Dialect has been added,³ also Powell’s Lexicon to Herodotus,⁴ and Trapp’s Lexikon zur byzantinischen Gräzität.⁵ The Lexicographical Resources page also provides links to the site of Diccionario griego-español [DGE],⁶ and to dictionaries of medieval and modern Greek, enabling study of the continuing history of words.⁷ Another archive, ΛΟΓΕΙΟΝ,⁸ developed at the University of Chicago, provides simultaneous lookup of entries in reference works that make up the Perseus Digital Library,⁹ and more resources have been added subsequently. These collections are likely to stimulate questions about how dictionary entries have been written and what the future direction for lexical studies and lexicons should be. ¹ Ed. Maria Pantelia; stephanus.tlg.uci.edu ² There is also a link to Liddell and Scott’s Intermediate (1889), a shortened version founded on LS⁷ [IGL]. ³ Cunliffe 1924. ⁴ Powell 1938. ⁵ Trapp and Hörandner, et al. 1994 2017. The online version is a collaboration between TLG and the Austrian Academy of Sciences: stephanus.tlg.uci.edu/lbg/#eid=1&context=lsj/. The headword equivalent to βάπτω in later Greek is βάϕω. ⁶ Adrados et al. (eds), 1980 . There are also links to Bauer (see Danker 2000 [BDAG]) and Lampe 1961 for definitions only. ⁷ For medieval and modern Greek there are links to Kriaras 2001 and the Triantafyllidis Lexicon 1998. ⁸ logeion.uchicago.edu/ ⁹ G.R. Crane, ed., Perseus Digital Library, www.perseus.tufts.edu/hopper/ Anne Thompson, βάπτω: An Illustration of the State of our Ancient Greek Dictionaries In: Liddell and Scott: The History, Methodology, and Languages of the World’s Leading Lexicon of Ancient Greek. Edited by: Christopher Stray, Michael Clarke, and Joshua T. Katz, Oxford University Press (2019). © Oxford University Press. DOI: 10.1093/oso/9780198810803.003.0020

354

Anne Thompson

For this case study, entries for βάπτω from LSJ, the Revised Supplement [LSJ Revd Suppl.],¹⁰ and DGE are compared. This verb is chosen because LSJ Revd Suppl. replaced almost the entire LSJ entry with a new article. Rewriting of whole entries is rare in LSJ Revd Suppl., so one may wonder what prompted it in this case. DGE presents its own reworked entry. We also have several pages of analysis by John Chadwick in his book Lexicographica Graeca.¹¹ It is instructive to look at the information these editors felt it necessary to provide, and whether everything that needs to be said has been said. The results present a general picture of the state of our dictionaries.

20 . 1 . L S J βάπτω,¹² fut. βάψω (ἐμ ) Ar.Pax959: aor. ἔβαψα S.Aj.95, etc.: Med., fut. βάψομαι Ar.Lys.51: aor. ἐβαψάμην Arat.951, AP9.326 (Leon.): Pass., fut. βᾰϕήσομαι LxxLe.11.32, M.Ant.8.51: aor. ἐβάϕθην AP6.254 (Myrin.), (ἀπ ) Ar.Fr.416; in Att. generally ἐβάϕην [ᾰ] Pl.R.429e, etc.: pf. βέβαμμαι Hdt.7.67, Ar.Pax1176. I. trans., dip, ὡς δ’ ὅτ’ ἀνὴρ χαλκεὺς πέλεκυν . . . εἰν ὕδατι ψυχρῷ βάπτῃ (so as to temper the red hot steel) Od.9.392; β. εἰς ὕδωρ Pl.Ti.73e, cf. Emp.100.11; τἄρια θερμῷ Ar.Ec.216; εἰς μέλι, εἰς κηρόν, Arist.HA 605a29, de An.435a2: Pass., βαπτόμενος σίδηρος iron in process of being tempered, Plu.2.136a; and of coral, become hard, Dsc.5.121 (s.v.l.). b. of slaughter in Trag., ἐν σϕαγαῖσι βάψασα ξίϕος A.Pr. 863; ἔβαψας ἔγχος εὖ πρὸς Ἀργείων στρατῷ; S.Aj.95; ϕάσγανον εἴσω σαρκὸς ἔβαψεν E.Ph.1578 (lyr.); in later Prose, εἰς τὰ πλευρὰ β. τὴν αἰχμήν D.H.5.15; β. τὸν δάκτυλον ἀπὸ τοῦ αἵματος LXXLe.4.17. c. also, dip in poison, ἔβαψεν ἰούς S.Tr.574; χιτῶνα τόνδ’ ἔβαψα ib.580. 2. dye, ἔβαψεν . . . ξίϕος the sword dyed [the robe] red, A.Ch.1011; β. τὰ κάλλη dye the beautiful cloths, Eup.333; β. ἔρια ὥστ’ εἶναι ἁλουργά Pl.R.429d; εἵματα βεβαμμένα Hdt.7.67; τρίχας βάπτειν AP11.68 (Lucill.): abs. in Med., dye the hair, Men.363.4, Nicol.Com.1.33; glaze earthen vessels, Ath.11.480e; of gilding and silvering, Ps. Democr.Alch.p.46 B.: Com., βάπτειν τινὰ βάμμα Σαρδιανικόν dye one in the [red] dye of Sardes, i. e. give him a bloody coxcomb, Ar. Ach.112; but βέβαπται β. Κυζικηνικόν he has been dyed in the dye of Cyzicus, i. e. is an arrant coward, Id.Pax1176 (v. Sch.).

¹⁰ Glare with Thompson 1996. This has been digitized and integrated into the text of LSJ by Logos Bible Software, with some unfortunate results, because it was not written with full integration in mind. ¹¹ Chadwick 1996. ¹² The text is from TLG, where the spacing is generous compared to the cramped printed version.

An Illustration of the State of our Ancient Greek Dictionaries

355

3. draw water by dipping a vessel, ἀνθ’ ὕδατος τᾷ κάλπιδι κηρία βάψαι Theoc. 5.127; ἀρύταιναν . . . ἐκ μέσου βάψασα τοῦ λέβητος ζέοντος ὕδατος draw water by dipping the bucket, Antiph.25, cf. Thphr.Char.9.8; βάψασα ποντίας ἁλός (sc. τὸ τεῦχος) having dipped it so as to draw water from the sea, E.Hec.610. 4. baptize, Arr.Epict.2.9.20 (Pass.). II. intr., ναῦς ἔβαψεν the ship dipped, sank, E.Or.707; β. εἰς ψυχρὸν [αἱ ἐγχέλυς] Arist.HA592a18; εἰ δ’ ὁ μὲν (sc. ἠέλιος) ἀνέϕελος βάπτοι ῥόου ἑσπερίοιο Arat.858 (ῥόον Sch.): c. acc., νῆα . . . βάπτουσαν ἤδη κῦμα κυρτόν dipping into . . . Babr.71.2: also Med., ποταμοῖο ἐβάψατο Arat. 951. 2. βάψας (sc. τὴν κώπην) Ar.Fr.225. (Cf. O Norse kuefia ‘dip’.)

Principal parts are listed at the head of the entry with references to the texts where they occur. The general policy is for the earliest attestation of frequently found forms to be noted, with one or two additional references for those that are infrequent. With the ease of searching now provided by TLG, we can see that the first example of the active future for the simple verb in literature is in Lycophron, Alexandra 1121. There is also an example in LXXLe.4.6. But an earlier fifth century BCE example has been chosen for mention in LSJ, notwithstanding that it is for a compound verb, ἐμβάπτω. The future form is not repeated under the entry for ἐμβάπτω, and the general rule seems to be that the principal parts of compound verbs are to be found under the entry for the simple verb, though there are some exceptions to this policy. Space permitting, it would have been better to give the principal parts for each verb entry and not to rely on cross-referencing. The forms in any case may not inevitably be the same for the simple and compound verbs across the dialects nor from a diachronic perspective. This entry notes in fact that different aorist passives occur (with -βαϕθ- or -βαϕ-), but without full information about their distribution. The aorist passive ἐβάϕην is said to be the ‘usual’ form in Attic, and in fact ἐβάϕθην is extremely rare, even though it is placed first, perhaps because it was felt to be the ‘regular’ form.¹³ An example from the Greek Anthology (AP.6.254 Myrin.) is given, also ἀπεβάϕθην from an Aristophanes fragment. A TLG search finds other examples only in late glossaries and commentaries. The aorist active ἔβαψα is cited from S.Aj.95. LS⁶ had just the playwright’s name (Soph.). Subsequently, in LSJ the work and line number were inserted (though there are two other instances, Tr.574 and 580). Strangely, LS⁷ had Trag., which is more accurate. The aorist also occurs in Aeschylus (Ch.1011, Pr.863) and a few times in Euripides. There is also an example in epic: Cypr.fr.5.2(West).¹⁴ In most verb articles in LSJ, passages cited for the principal parts are picked up in the body of the entry in the classification

¹³ Although both suffixes occur from the time of the earliest texts, the θη suffix is historically of later origin: see Chantraine 1961, 165 70, especially paragr. 192. ¹⁴ M.L. West, Greek Epic Fragments (Cambridge, MA, 2003).

356

Anne Thompson

of senses, but here several do not reappear: Ar.Lys.51, AP9.326, LXXLe.11.32, M.Ant.8.51, Pl.Rp.429e. These could be added respectively to LSJ I 2, I 3 (as middle), I 1a, and I 2 (both as passive), and then the last is a passive beside the active in Pl.Rp.429d in I 2. The article is divided into two major sections I and II, transitive and intransitive. For a verb entry, this is a fundamental distinction that must always be kept in focus: either the categories can be kept apart as here, or one can be placed beneath the other in a subsection for each sense. A disadvantage of the major sections is that examples close in sense are not near each other. In this case, there is not much potential for confusion, since all the examples in II are about dipping or sinking into water, i.e. related to I 1a. Section I is divided into four subsections that relate to sense: firstly dip, intended here to imply ‘in a liquid’. The idea of ‘liquid’ dictates the subsections: b ‘in blood’, c ‘in poison’. Subsection a is mainly about water, though this is not stated explicitly. There are also two Greek quotations about honey and wax. Subsection 2 is about dyeing, a dipping that results in colouring, and 3 about dipping that involves the taking up of water. Subsection 4 is about the dipping of a person in water in a ritual. Different considerations will always compete for focus in an article. For voice, there is a different approach from the transitive and intransitive division. The passives are kept close to the active examples that have the same sense. Passives appear in I 1a and 2, introduced with a label in 1 in a section at the end, but not labelled in 2, where quotations with passives are interleaved with equivalent active examples. In I 4, the translation gloss for the sense is active, but ‘Pass.’ is added at the end as incidental to the particular context. For the middle voice, LSJ uses the abbreviation ‘Med.’, the old–fashioned term ‘Medium’. In I 2, examples from Menander and another comic poet are introduced with ‘Med.’, also at the end of II 1 for an Aratus example. When ‘Med.’ is inserted in I 2, it interrupts a sequence of unlabelled citations with the verb in both active and passive. The reader has to recognize the voice mostly from the quotations. However, the two examples immediately following the ‘Med.’ section have just an indication of context in English: ‘glaze earthen vessels’, and ‘of gilding and silvering’, so the point at which the Med. label ceases to apply is not clear. The punctuation of an em dash in earlier editions before the glaze example made it clear that there was a new start here after the ‘Med.’ label, but this was taken out in LSJ. The various revisions under different hands over several editions often result in a lack of clear overall editorial policy on details. A selection of short quotations is given in each section. Most are not translated, so readers have to rely on their knowledge of Greek or on other books. The general purpose of the quotations in LSJ seems to be mixed, mostly either to indicate different contexts that have a sense in common, or to illustrate syntactic constructions. Occasionally, translations appear when

An Illustration of the State of our Ancient Greek Dictionaries

357

there is an unexpectedly big shift in a sense section, or when there is a particularly difficult phrase. There is also some randomness. In I 1, βαπτόμενος σίδηρος is translated as ‘iron in process of being tempered’, a change in LSJ from ‘tempered iron’ in earlier editions, presumably to give a more accurate translation for the required continuative meaning, but perhaps the quotation was not needed. A translation appears for the first example in I 2, the intention being to bring out the poetic imagery of a sword dyeing a garment red with blood in A.Ch.1011, perhaps deliberately to combine the dye sense with an echo of examples in I 1b where weapons are dipped into blood or flesh. There is a translation for Eup.333, ‘β. τὰ κάλλη dye the beautiful cloths’, even though there is nothing particularly difficult about this phrase.¹⁵ The phrases in Aristophanes at the end of I 2 are translated because there is wordplay that needs to be explained. There is a tendency in LSJ, not consistently followed, to generalize quotations: for Ar.Ach.112 there is ‘βάπτειν τινὰ βάμμα Σαρδιανικόν’, with the general infinitive and τινα someone rather than the wording in the original: ἵνα μή σε βάψω βάμμα Σαρδιανικόν. Sometimes the headwords in quotations are abbreviated to just the initial letter, not perhaps ideal for a strongly inflecting language. It is another way of generalizing, the purpose of which was apparently to suggest that the phrase is a model for what can be found in other contexts, or for what one might write in one’s own prose or verse compositions. In I 3, the translation of Antiph.25 does little more than repeat the definition at the beginning of the section ‘draw water by dipping’. The need for repetition is due to the clumsiness of introducing Theoc.5.127 as the first example, which is about honey rather than water. Furthermore, the container dipped in Theocritus is in the dative case, presenting a different syntax from the following examples where it is in the accusative. There is also ambiguity as to whether to interpret κηρία as accusative object (the honey drawn up) or internal accusative (into honeycomb or honey). A translation is given for E.Hec.610, following again the definition at the beginning. In II both dipped and sank are offered as translations for E.Or.707, apparently to show that the ship not only listed but took on water. dipped connects with other senses already mentioned, but on its own does not make it clear that the result was taking on water. For description of meaning, LSJ employs for the most part translation glosses, rather than true lexicographic definitions. For example, dip is a translation gloss, but it is not precise, because there are several senses of this word in English, some of which do not correspond to the Greek verb (see

¹⁵ See now fr. 363 in S.D. Olson, Fragmenta Comica: Eupolis frr. 326 497 (Heidelberg, 2014), I.C. Storey, ‘Notes on Unassigned Fragments of Eupolis’, Museum Criticum 30 1 (1995 6) 137 57. See also Storey 2003, 382 and 94 8 on the verb βάπτειν. The correct reading might be Βάπταις ‘in the Dyers’ (i.e. in the play of that name). The new fragment number is as in R. Kassel and C. Austin, Poetae Comici Graeci: vol. V Damoxenus Magnes (Berlin/New York, 1986).

358

Anne Thompson

The Oxford English Dictionary [OED]).¹⁶ The superiority of lexicographic definitions has been written about in particular in the field of biblical lexicography.¹⁷ A recurring problem in LSJ for modern readers is nineteenth- and earlier twentieth‐century English (see Williamson, Chapter 2, this volume). Contemporary speakers of English will have difficulty in making all the necessary adjustments. In I 2, for Ar.Ach.112, the translation and explanation are ‘dye one in the [red] dye of Sardes, i.e. give him a bloody coxcomb’. There seems to be an awkward transition from ‘one’ to ‘him’. τινα has been substituted for σε you (enclitic) in the original line and translated as ‘one’, rather than ‘someone’, and then ‘him’ in the explanation looks like it refers to the person in the original text. But ‘one’ is in older English an indefinite pronoun, equivalent of ‘someone’ (see OED C VI 16, 17), and ‘him’ refers back to this without implication of gender, the masculine being the default for common gender. The reader is unlikely to spot this kind of thing in normal consultation of the lexicon. We would now not wish to violate the line and its metre in this way, but would prefer to quote it verbatim with its proper translation. ‘coxcomb’ as a comic appellation for the head is no longer current, and the bloodied head of the victim of the beating and his associated stupidity conjured up by the image of the head and comb of a cockerel is mysterious to us now. For Ar.Pax1176, about the dye of Cyzicus, ‘arrant’ is dated. The Sardes dye in the other passage is a dark or purple red,¹⁸ but the colour associated with Cyzicus and its connection with cowardice are not explained.¹⁹ The coward in battle defecates on his clothes, but this is not stated forthrightly in LSJ. The reader has only ‘(v.Sch.)’ as a direction, a text that requires some proficiency in Greek to read.²⁰ At the end of the article Old Norse kuefia ‘dip’ is given by way of an indication of etymology. This tells the reader little more than that the root

¹⁶ Simpson and Weiner 1989, and online, www.oed.com/ ¹⁷ See Lee 2003 (passim, especially chapters 2 and 3, 15 44). Roberts (2004) shows that even when the method is adopted, things can go wrong. For definitions in classical lexicography and old fashioned ideas about defining, see Chadwick 1996, 20 1 and Glare 1991, 40 1. ¹⁸ S.D. Olson, Aristophanes: Acharnians (Oxford, 2002), 108, on line 112. ¹⁹ A.H. Sommerstein, Aristophanes: Peace (Warminster, 1990), 189, on line 1176, suggests the colour is from the city’s stater coin of a pale gold colour, perhaps with a pun on a derivative with the same stem as χέζω. S. Douglas Olson’s commentary (Oxford, 1998), 292, settles for the colour of the coin, rejecting the possible pun. M. Platnauer’s commentary (Oxford, 1964), 163, accepts the pun as likely, and relates the dye to the pallor of effeminacy and cowardness, one of the explanations in the scholia. ²⁰ LSJ is following good lexicographic practice in not trying to explain too much here, because uncertainty of interpretation is the business of commentators. See D. Holwerda, Scholia in Aristophanem 2.2 (Groningen, 1982), on line 1176, a text that offers several explanations, relating to incontinence associated with sodomy, immorality, cowardice, or effeminacy. The colour of the coin interpretation is not in the scholia, though Eup.fr.247 is quoted there, with the line ἥδε Κύζικος πλέα στατήρων, alluding to the city’s wealth, but with the stater also apparently as a colloquialism for anus or vagina.

An Illustration of the State of our Ancient Greek Dictionaries

359

of βάπτω is thought to be of Indo-European origin. Etymology is the first point on a chronological scale of potentially useful information, but even when certain (and it is not here) it is only a hint, and the range of senses of cognates will usually not be known to the reader. The etymologies in LSJ are now out of date and must always be checked in other reference works.²¹ There seems to be no purpose for the Old Norse here other than to invite further exploration. Etymology, however, is a field that requires expert knowledge.²²

2 0 . 2 . L S J REVD SUPPL. The entry for βάπτω presents rewritten versions of I 1–3 of the LSJ article. I 1a now begins with an additional definition: ‘immerse in a liquid, dip’, indicating without ambiguity which sense of dip in English is meant, and looking ahead to various liquids involved, water, blood, honey, wax. This also looks forward to I 2 where dye is introduced with a new definition: colour by immersion, which picks up the connection with the idea of immerse in I 1, this time with a specific purpose. Without these definitions connecting the entries, the reason the same verb might mean dip and dye is not necessarily obvious. This mixed method of using both lexicographic definitions and translation glosses is a key innovation in LSJ Revd Suppl. LXXLe. 4.17 is moved up from the end of I 1b in LSJ into I 1a here because, although the context involves blood, it is not about slaughter of humans in tragedy. S.Tr. 574 and 580, LSJ’s dip in poison examples in I 1c, are also moved up into I 1a, with quotations, since poison as a liquid is part of the context, and in fact Deianeira at this point does not know the liquid she collects is poison.²³ For S.Tr. 574, however, LSJ Revd Suppl. has failed to adopt the Oxford Classical Text reading, even though this is the specified text in the introductory list.²⁴ In this, μελάγχολος . . . ἰός poison (nominative) is read for μελαγχόλους . . . ἰούς arrows (accusative plural), which alters the meaning of the passage. Also, the insertion of these citations separates Od. 9.392, about ²¹ The standard etymological dictionaries are Chantraine, DELG 1968 80 (new edn, 2009), Frisk 1960 72, and Beekes 2010. Chantraine reports the theory that the verb is a present of the type with a suffix * ye/yo , related to Old Norse kvefja ‘plonger, ėtouffer’, related also to Old Swedish kvaf ‘profondeur’. There is also mention of a theory by Szemerényi (1960, 213 14) that favours a relationship also between βάπτω and βαθύς. Beekes translates the Old Norse as ‘to press down, immerse, choke’. There are perils associated with a policy of including etymology in general lexicons; for New Testament lexicons in this regard, see Lee 2013. ²² See further Katz, Chapter 5, this volume. ²³ See Glare 1987, 13. ²⁴ H. Lloyd Jones and N.G. Wilson, Sophoclis Fabulae (Oxford, 1990). On the problems of interpretation of 572ff., see Halleran 1988. See also the commentaries of R.C. Jebb (Cambridge, 1894); P.E. Easterling (Cambridge, 1982); and M. Davies (Oxford, 1991).

360

Anne Thompson

tempering metal in water, from Pl.Ti. 73e, which is about the creator making bone from a kneaded mixture of earth and marrow, where the language has echoes of metal tempering: καὶ μετὰ τοῦτο εἰς πῦρ αὐτὸ ἐντίθησιν, μετ’ ἐκεῖνο δὲ εἰς ὕδωρ βάπτει, πάλιν δὲ εἰς πῦρ, αὖθίς τε εἰς ὕδωρ then he places it in fire, and after that dips it in water, and from this back into fire, and once again in water. The instruction ‘cf. Emp. 100.11’ is taken over from LSJ, a passage which is not about metalworking but plunging the spout of a clepsydra in a body of water where it may be filled. This is nearer to the passages in LSJ I 3, and it is not clear what the point is of comparing this with the Plato passage. LS⁶ had only the Odyssey, Plato, and Plutarch passive citations, putting all the manufacturing passages together, but subsequently Aristotle’s honey and wax passages were inserted, breaking the connection, now further broken in LSJ Revd Suppl. The passage about coral, originally inserted in LS⁷, and where there might also be something of the idea of tempering, is omitted from LSJ Revd Suppl., presumably because the reading is not sound. I 2 begins with the additional definition colour by immersion, and then moves straight into examples about dyeing textiles (Eup., Pl., and Hdt.). The last is a passive example, but it is not placed at the end of the section as in 1 1a. The section then moves on to the context of dyeing hair, but the absolute middle examples from Men. and Nicol.Com. in LSJ have been omitted. There may have been an intention to place them elsewhere in a dedicated middle section. If so, they have been forgotten. Even if the one example of the active from AP 11.68 (Lucill.) was considered sufficient to illustrate this context, it would still be necessary to mention the middle. Passive uses may be understood from the active, but it is customary for a middle to be specifically recorded. A separate subsection would probably have been better for the dyeing hair context, perhaps with a new definition ‘colour the hair by applying dye’, because colour by immersion does not quite suit. The example at the beginning of LSJ I 2 (A.Ch. 1011), where Aegisthus’s sword ‘dyed [the robe] red’, is now moved to a position later in the dyeing section (LSJ Revd Suppl. I 2). The quotation has been dropped and, rather than a new definition, there is a label ‘poet.’, presumably implying a poetic figure. There is a different sense here, not ‘colour with dye’, but ‘colour as though with dye’, i.e. with blood. As the next lines reveal, it is also a pejorative sense, about staining and spoiling the original dyes and appearance of the fabric. The Aristophanes example about colours associated with Sardes and Cyzicus come next. The LSJ label ‘Com.’ (in the list: ‘Comedy, Comic, in the language of the Comic writers’) is replaced by the description ‘humorously’. Both apply to the genre or audience reaction, not the speaker’s intention, which is serious albeit absurdly hyperbolic. The explanation ‘i.e. give him a bloody head’ replaces LSJ’s coxcomb image. The second Aristophanes example is then introduced by ‘fig.’, to replace LSJ’s rather enigmatic ‘but’. It is hard to see why this is judged figurative as opposed to the first example, or even to A.Ch. 1011. There is

An Illustration of the State of our Ancient Greek Dictionaries

361

still no help with the dye from Cyzicus, nor what it is likened to. But the old-fashioned ‘arrant’ has been changed to ‘thorough’. The translation with a passive ‘he has been dyed’ is questionable, since the coward dyes himself and his clothes by defecating. Interpretation as a middle is preferable. Ps.-Democr.Alch.p. 46 B in LSJ appears correctly in LSJ Revd Suppl. as just Ps.–Democr.p. 46 B, in accordance with the abbreviation given in LSJ’s author list, xxii.²⁵ ‘of gilding and silvering’ is an awkward style however, because it does not clearly indicate if this is from the context, the meaning of the headword, or a blend. LSJ has gilding and silvering in italics, implying these are equivalents of the headword, though it is not legitimate to define using a different part of speech from the headword. TLG identifies the author of this work on alchemy with Bolus of Mendes.²⁶ The verb is actually used twice on p. 46, and with different senses, neither of which is quite gild or silver. Paragr. 12 is about adding a bronze-coloured substance to other substances to create a gold-coloured material: ὁ γὰρ χαλκὸς ἀσκίαστος ξανθὸς [ὢν] γενόμενος πᾶν σῶμα βάπτει for bronze, when it is untarnished and golden, imbues every material.²⁷ Paragr. 14 has ἐὰν οὖν ἐκστραϕῇ, κατάβαψον αὐτὴν εἰς ἔλαιον κίκινον πολλάκις πυρῶν καὶ βάπτων if then it is transmuted, dip it completely into castor oil, repeatedly heating and dipping. A process involving silver is on p. 48, paragr. 18: βάπτε ἄργυρον ἐκ πετάλων ἕως ἀρέσῃ τὸ χρῶμα dip silver leaf (into a mix of saffron and grape juice) until the colour is right. An innovation here in LSJ Revd Suppl. is the introduction of the label ‘transf.’: ‘transferred, in transferred sense’. The meaning is that there is a transfer from the dyeing of textiles to a comparable process in another sphere.²⁸ This label is a feature of Oxford Latin Dictionary [OLD] style, not used anywhere in LSJ. Peter Glare, editor of LSJ Revd Suppl., was previously editor of OLD, and throughout LSJ Revd Suppl. he uses some features that are inconsistent with LSJ, e.g. in labels, punctuation and a preference for the use of lower-case letters. In this article for βάπτω, there is ‘trag.’ for ‘Trag.’, ‘poet.’ for

²⁵ B refers to Berthelot, the editor listed in LSJ for Pseudo Democritus Alchemista: M. Berthelot, Collection des anciens alchimistes grecs (Paris, 1887; date and place not given). The text begins on p. 41. ²⁶ The text was taken over in LSJ Revd Suppl. apparently without being checked. The scale of lexicographic undertaking acts against thoroughness and, even as recently as the compilation of LSJ Revd Suppl., there was not the same ease of consulting texts as is now afforded by TLG and other electronic resources. ²⁷ The paragr. numbers here are as in M. Martelli, The Four Books of Pseudo Democritus (London, 2013), S93 9. His translation is slightly different, as is that of R.R. Steele, Chemical News 61 (1890) 88 125, as included in S.J. Linden (ed.), The Alchemy Reader: From Hermes Tresmegistus to Isaac Newton (Cambridge, 2003), 40 2. ²⁸ On this label, see Thompson 2017, especially 490 3; Clarke 2010, 125; and compare the contrasting approaches to the issue taken by Clarke, Chapter 14, and Silk, Chapter 17, both in the present volume.

362

Anne Thompson

Poet.’, ‘pass.’ for ‘–Pass.’ (without the em dash), also a space between author or work and reference numbers.²⁹ Nonetheless, in most respects, LSJ Revd Suppl. is chained to the methodology and layout of the original work with its dense columns of print that are hard to navigate. There was no possibility of introducing the OLD system of introducing lexicographic definitions alongside translation glosses and highlighting them by keeping them separate from the paragraph of quotations in a smaller font. It seems that the reason for the LSJ Revd Suppl. replacement for most of the βάπτω article was to avoid a complicated set of instructions for corrections to LSJ rather than to present a new entry based on a re-reading of the sources, as was done for all the entries in OLD. For Section I 3, LSJ Revd Suppl. has ‘dip a vessel in order to draw water’, seeing ‘draw’ as part of the context whereas, with its definition ‘draw water by dipping a vessel’, LSJ sees the drawing up of the liquid as part of the meaning of the headword. The focus is on the drawing of water as well as dipping. Otherwise, the rest of the section is the same as in LSJ, except that the first example (Theoc. 5.127), about honey rather than water, is suitably moved to the end. It is introduced with ‘cf.’, a label from LSJ style. Presumably this is intended to mean that the passages have features in common, but the reader has to work out the reason for the comparison. There is no change from LSJ’s policy of occasional translation of quotations. For E.Hec. 610, the part of the translation equivalent to the headword is in italics as in LSJ: ‘having dipped it so as to draw water from the sea’, but this is at odds with the new definition which has only dip in italics. The same applies to ‘draw water by dipping the bucket’ for Antiph. 25, also taken directly from LSJ. It is also questionable whether this can translate the participle βάψασα in the quotation. There is an instruction to compare Thphr.Char. 9.8 with Antiph. 25, with no quotation given, a passage with comparable context and syntax. The quotation for Theoc. 5.127 remains untranslated.

20.3. DGE The third volume of DGE, ἀποκοιτέω—Βασιλεύς was published in 1991. Below is the entry for βάπτω:³⁰

²⁹ As already noticed, LSJ’s ‘Med.’ for middle voice is omitted in this article, but where the label does occur in LSJ Revd Suppl. it is printed ‘med.’, e.g. s.v. ἰθαίνω. ³⁰ The article here with a spacious layout is as presented on the DGE website: http://dge.cchs. csic.es/xdge.

An Illustration of the State of our Ancient Greek Dictionaries

363

I tr. 1 en v. act. c. ac. sumergir, hundir en un líquido εἰς ὕδατος . . . τέρεν δέμας Emp.B 100.11, τἄρια βάπτουσι θερμῷ Ar.Ec.216 • abs. μετ’ ἐκεῖνο εἰς ὕδωρ βάπτει Pl.Ti.73e, ἐς ὀθόνιον βάπτοντα (un trapo), Hp.Mul.2.196, εἰς κηρὸν βάψειέ τις Arist.de An.435a2 • hundir el remo, e.d., navegar βάψας πρὸς ναυτοδίκας Ar.Fr.237 • mojar, empapar σπόγγους βάπτοντα Hp.Epid.5.58, ᾗ . . . ἔβαψεν ἰοὺς θρέμμα Λερναίας ὕδρας donde la criatura de Lerna, la hidra, mojó sus flechas S.Tr.574, τὸν χόρτον εἰς μέλι βάπτοντες Arist.HA 605a29, ἵνα βάψῃ τὸ ἄκρον τοῦ δακτύλου αὐτοῦ ὕδατος Eu.Luc.16.24 • en v. pas. σιδηρίου βαϕέντος ἐς ἔλαιον Hp.Coac.378 • en ceremonias rituales βάψει ὁ ἱερεὺς τὸν δάκτυλον ἀπὸ τοῦ αἵματος LXX Le.4.6, δεσμὴν ὑσσώπου . . . βάψαντες ἀπὸ τοῦ αἵματος LXX Ex.12.22, ἔβαψεν αὐτὸ (τὸ ἄκρον τοῦ σκήπτρου) εἰς τὸ κηρίον τοῦ μέλιτος LXX 1Re.14.27, βάψεις τὸν ψωμόν σου ἐν τῷ ὄξει LXX Ru.2.14 • de un metal sumergir en agua para templarlo templar por inmersión πέλεκυν . . . εἰν ὕδατι βάπτῃ Od.9.392 • en v. pas. βαπτόμενος σίδηρος hierro templado Plu.2.136a, χαλκὸς . . . θερμὸν ὄντα ὑπὸ ὕδατος τούτου βάπτεσθαι λέγουσιν Paus.2.3.3 • por anal. clavar, hundir un objeto punzante ἐν σϕαγαῖσι βάψασα ξίϕος A. Pr.863, ϕάσγανον εἴσω σαρκὸς ἔβαψεν E.Ph.1578, εἰς τὰ πλευρὰ βάψας τὴν αἰχμήν D.H.5.15, ξίϕος βάπτει κατὰ τῆς καρδίας Ach.Tat.1.4.3 • fig. ἔβαψας ἔγχος . . . πρὸς Ἀργείων στρατῷ has hundido la lanza en el ejército de los argivos S.Ai.95 • en v. med. c. ac. βαψάμενα κοίλων ἐντὸς ἄρη λαγόνων AP 7.531 (Antip.). 2 colorear, teñir ϕᾶρος τόδ’ ὡς ἔβαψεν Αἰγίσθου ξίϕος A.Ch.1011, χιτῶνα τόνδ’ ἔβαψα S.Tr.580, βάψαι ἔρια Pl.R.429c, ἐάν τέ τις ἄλλα χρώματα βάπτῃ Pl.R.429d, βάπτετε τὰ κάλλη τὰ περίσεμνα τῷ θεῷ Eup.363, τὰς τρίχας . . . τινὲς βάπτειν σε λέγουσι AP 11.68 (Lucill.), βάπτουσιν Ἀϕροδίτης τὸν πέπλον Ach.Tat.2.11.4 • manchar βάπτει τὸ αἷμα τὴν γένυν Ach.Tat.2.11.5, ὁ χαλκὸς πᾶν σῶμα βάπτει Ps.Democr.p.46, με τὰς χεῖρας ἐς τοῦ παιδὸς αἷμα βάψαντα Philostr. VA 8.7 (p.318), ἐρεῦσαι· ϕοινῖξαι, ἐκ τοῦ ἐρεύθω τὸ βάπτω Sch.Il.18.329 • en v. med. c. ac. hacer teñir κροκωτὸν . . . ἐγὼ βάψομαι Ar.Lys.51 • fig. ἵνα μή σε βάψω βάμμα Σαρδιανικόν para que no tenga que teñirte con tinte de Sardes, e.d., si no quieres que te golpee hasta sangrar Ar.Ach.112 • en v. pas. ser teñido εἵματα βεβαμμένα Hdt.7.67, ἱμάτιον βεβαμμένον αἵματι Apoc.19.13, βεβαμμένα χλανιδία I.BI 4.563, δοραὶ . . . βεβαμμέναι I.AI 3.102, σάγοι Ἀρσινοϊτικοὶ . . . βεβαμμένοι Peripl.M.Rubri 8, τἀκ κόκκου βαϕθέντα . . . θέριστρα AP 6.254 (Myrin.), de cacharros de barro

364

Anne Thompson

βάπτονται εἰς τὸ δοκεῖν εἶναι ἀργυραῖ Ath.480e, del agua de una fuente οὐδαμῶς βαϕήσεται M.Ant.8.51. 3 coger agua hundiendo un objeto en ella λαβοῦσα τεῦχος . . . βάψασ’ ἔνεγκε δεῦρο ποντίας ἁλός E.Hec.610, μὴ βάπτετε πίνετ’ ἀπὸ κρανᾶν Call.Lau. Pall.45, ἁ παῖς ἀνθ’ ὕδατος τᾷ κάλπιδι κηρία βάψαι Theoc.5.127, Ῥυνδακοῦ ποτῶν κρωσσοῖσιν ὀθνείοισι βάψαντας γάνος Lyc.1365. 4 en sent. ritual bautizar en v. act. τὸ μὲν σῶμα ἔβαψεν ὕδατι Cyr.H. Procatech.2 • en v. pas. ser bautizado ὁ βεβαμμένος Arr.Epict.2.9.20, βεβά]μμεθα ἐν ὕδατι ζω[ῆς Eu.Fr.Pap.2.43. II intr. 1 en v. act. sumergirse, mojarse ἐὰν βάπτωσιν εἰς ψυχρόν (αἱ ἐγχέλυς) Arist. HA 592a18, ὁ μὲν ἀνέϕελος (ἠέλιος) βάπτῃ ῥόου ἑσπερίοιο Arat.858 • de barcos hundirse ναῦς . . . ἔβαψεν E.Or.707, ἰδὼν γεωργὸς νῆα . . . βάπτουσαν . . . κῦμα Babr.71.2 • en v. med.-pas. οἱ πόδες τῶν ἱερέων . . . ἐβάϕησαν εἰς μέρος τοῦ ὕδατος LXX Io.3.15, ποταμοῖο ἐβάψατο Arat.951. 2 en v. med. teñirse βαπτόμενος βατραχειοῖς tiñéndose de verde rana Ar. Eq.523, ἐβάπτετο δ’ αἵματι λίμνη Batr.220, βάψομαι (los cabellos), Men. Fr.303.4, τὸ αἷμα ἔρρει πολὺ . . . ὥστε αὐτὰ βάπτεσθαι καὶ ἐρυθρὰ ϕαί νεσθαι Luc.VH 1.17 • mancharse ἐπειδὰν ἐπιστάξῃ ἐπὶ τὰ ἱμάτια, βάπτεται Hp.Mul.2.122 • fig. αὐτὸς βέβαπται βάμμα Κυζικηνικόν se queda teñido con un tinte de Cízico e.d. se caga de miedo Ar.Pax 1176. • Etimología: Quizá de una raíz *g u̯ m̥ bh-/g u̯ embh- ‘profundo’, c. disim. y suf. yod, rel. antiguo nórdico kuefja ‘hundirse en el agua’, cf. βαθύς. DGE offers more supporting passages. Earlier editions of Liddell and Scott were primarily based on authors of the classical era. It was not possible for LSJ to catch up with post-classical literature and all the increasingly available documentary sources.³¹ DGE adds more citations from the Septuagint: LXXEx.12.22 is added to LXXLe.4.6 (which replaces LSJ’s LXXLe.4.17, a passage with similar wording and, as in LSJ Revd Suppl., this context is moved away from the slaughter in tragedy section). Two other citations are given, LXX1Re.14.27 (=LXX1Ki. in LSJ) and LXXRu.2.14. This section is introduced with the words ‘en ceremonias rituales’, a label that presents a key improvement in methodology, that is, paying closer attention to context and semantic domain. However, the last two examples are about food, dipping a stick into honeycomb and bread into liquid, and they are without ritual significance.

³¹ See Aitken 2014, 7.

An Illustration of the State of our Ancient Greek Dictionaries

365

Another Septuagint citation is given in the intransitive section II 1, LXXIo3.15 (=LXXJo. in LSJ), labelled as ‘med.-pas.’, though the form should probably be interpreted as a passive. There is also a policy to include more Christian literature, and two supporting citations are added for the unexpected sense of βάπτω as bautizar/ baptize (βαπτίζω being more usual). Both are taken from Lampe’s Patristic Greek Lexicon (1961), with an updating of one of the references.³² This section is labelled ‘en sent. ritual’, picking up a similar idea to ‘en ceremonias rituales’, but it is not certain if the different wording indicates a deliberate distinction. As noted, the use of this kind of label is an important innovation. Lampe defines, however, as dip, and says for the passages ‘of baptism’. There is a contrast in fact in Cyr.H.Procatech.2, with βαπτίζω: ἐβαπτίσθη, ἀλλ’ οὐκ ἐϕωτίσθη· καὶ τὸ μὲν σῶμα ἔβαψεν ὕδατι, τὴν δὲ καρδίαν οὐκ ἐϕώτισε Πνεύματι he was baptized but was not enlightened; he dipped his body in the water, but did not enlighten his heart with the Spirit. This alleged parallel taken from Lampe does not support the sense baptize: it is a description of the action in the ritual, not the full religious sense. Eu.Fr.Pap.2.43 similarly has a contrast with βαπτίζω, but with that verb in reference to Jewish ritual ablution: ἐγὼ δὲ καὶ οἱ [μαθηταί μου] οὓς λέγεις μὴ βεβα[πτίσθαι βεβά]μμεθα ἐν ὕδασι ζω[ῆς αἰωνίου . . . ] I and my disciples whom you say have not performed ritual ablution have been dipped in the waters of eternal life.³³ It is not made clear in LSJ or DGE that the context for the Arrian quotation from Epictetus is for a proselyte in a Jewish context: ὅταν δ’ ἀναλάβῃ τὸ πάθος τὸ τοῦ βεβαμμένου καὶ ᾑρημένου, τότε καὶ ἔστι τῷ ὄντι καὶ καλεῖται Ἰουδαῖος when he has adopted the attitude of a man who has been ritually dipped in water and made his choice, then truly he is and is called a Jew.’ The definition as baptize is misleading.³⁴ In general, increased coverage for the Hellenistic period onwards is represented by new citations from Callimachus, Lycophron, Batrachomyomachia, Josephus, Lucian, Pausanias, Philostratus, Achilles Tatius, the work Περίπλους τῆς Ἐρυθρᾶς Θαλάσσης (Peripl.M.Rubri), and Scholia on the Iliad. However, the citations Antiph.25 and Thphr.Char.9.8 in LSJ and LSJ Revd Suppl. I 3 do not appear in DGE. A new classical citation, Ar.Eq.523, is introduced in II 2. But others that can be gleaned from TLG, for example E.HF929, X.An.2.2.9, Pl.Ti.83b, and several in Aristotle, do not appear. Ar.Lys.51 and AP6.254 (Myrin.), mentioned in the LSJ principal parts section but not picked up in the body of that article, are in I 2 as middle and passive respectively. Other references in LSJ’s list of principal parts that are not assigned senses, AP9.326 ³² Eu.Fr.Pap. = Euangeliorum incertorum Fragmenta Papyracea; see DGE list of abbreviations for details of these papyri; 2.43 = POxy.840.43 in Lampe 1961. ³³ See B.P. Grenfell and A.S. Hunt, The Oxyrhynchus Papyri vol. V (London, 1898), 7 9. ³⁴ The baptize section sits awkwardly in our dictionaries and should have been questioned. G. Fatouros (Glotta 49 (1971) 87) compares the meaning in Arrian with modern Greek βαμμένος (a form of the perfect passive participle attested from Byzantine times): ‘fanatischer Anhänger einer Partei oder Sekte’. One might compare English ‘dyed in the wool’.

366

Anne Thompson

(Leon.), LXXLe.11.32, and Pl.R.429e, do not appear in DGE either. DGE makes a point of including more citations for medical writers, who use this verb frequently (more than one hundred instances in TLG for the Hippocratic corpus, Galen and Aëtius). Citations from Hippocrates are added in I I and II 2.³⁵ DGE presents a longer entry than LSJ. It adopts the same major transitive (tr.) and intransitive (intr.) distinction as LSJ, and broadly keeps the same sense sections.³⁶ There are no subsections in I 1 corresponding to a, b, and c in LSJ, though the article moves from one kind of context to another, sometimes with new translation glosses. Middle and passive examples are added to the relevant sense sections. A key innovation is that quotations are given for all the citations. There are no abbreviations of Greek words or changes to their form, and the quotations tend to be longer than in LSJ, presenting more of the context and helping the reader with comprehension. Translations are given only for selected quotations, the choices being mostly the same as in LSJ. The senses are defined with translation glosses: in I 1 sumergir, hundir, mojar, empapar, sumergir, clavar, 2 colorear, teñir, manchar, 3 coger agua, 4 bautizar; in II 1 sumergirse, mojarse, hundirse, 2 teñirse, mancharse. There are more than in LSJ, which is in general not necessarily a good thing, because there is a risk of not keeping strictly to the closest equivalents and of encroaching on the meaning of synonyms of the headword.³⁷ I 1 begins with the Emp. citation, presumably because it is felt to be a general example of immersing something in a liquid. But, as has been seen, this context is in fact closer to those found in I 3, about dipping so as to draw water. The second citation Ar.Ec.216 is placed here, apparently accepting the interpretation that it is about dipping wool in water in a cleaning process rather than dyeing.³⁸ Pl.Ti.73e remains with the general section at the beginning, but the passage Od.9.392 about tempering metal, to which Pl.Ti.73e relates in some ways, has been moved much further down in I 1.³⁹ Ar.fr.237 [=Ar.Fr.225] has been moved up from LSJ II 2, because if an object τὴν κώπην ³⁵ There is also a policy for wider coverage of documentary sources, but this verb is very rarely attested in these, and no new senses have been identified. For βάπτω and derivatives of the same root, see Mossakowska 2002, 292 5, a work that explores some of the papyrological sources. ³⁶ See Adrados 1977, for editorial policy and principles of arrangement in DGE. ³⁷ The conventions for author and works abbreviations are mostly as in LSJ, with some changes as appropriate for Spanish. Keeping to universal conventions helps to avoid confusion. LSJ Revd Suppl. in turn adopted many DGE conventions for sources not in LSJ. ³⁸ R.G. Ussher (Aristophanes: Ecclesiazusae (Oxford, 1973), 104 5) thinks the reference is to dyeing, also J. Henderson, in his translation in Aristophanes vol. IV (Cambridge, MA, 2002), 271. The women argue they can run the state because they do things properly, like always using hot water for the wool. See also Monaghan 2001, 42 4. ³⁹ There is a small problem, in that by giving a second definition templar por inmersión, the participle ϕαρμάσσων is ignored (as it is in the quotation). See further, n. 67. Also, in the quotation Paus.2.3.3, for the passive, the first word should read χαλκόν (accusative, as the following adjective shows).

An Illustration of the State of our Ancient Greek Dictionaries

367

is to be understood (and this is far from certain) then it should be taken as transitive or, in traditional terminology, absolute. The reference has been updated to the Kassel and Austin edition.⁴⁰ LSJ Revd Suppl. did not move this citation. As in LSJ Revd Suppl., the LSJ I 1c dip in poison section is abolished, and S.Tr.574 (with the ‘arrows’ reading) is moved up with other dip examples, with poison seen as contextual. But the accompanying S.Tr.580 in LSJ is moved down to I 2 with colorear, teñir. However, it is not a dyeing process, in that Deianeira applies the substance like an ointment with a piece of wool: line 675.⁴¹ The passages about dipping or plunging a weapon in blood or flesh towards the end of I 1 are labelled ‘por anal.’ The introductory list actually gives ‘analóg. = analógico, analógicamente’ that is ‘by analogy, analogically’. LSJ and LSJ Revd Suppl. do not employ such a label for these examples.⁴² Judging from the position of the label, the intention in DGE seems to be to indicate not just an extension from water to blood, but that there is an analogy between a weapon dipped in water for tempering to a weapon dipped in blood or flesh. But S.Ai.95 [=S.Aj.95] is labelled ‘fig.’ It is hard to be sure why this is thought to be different, because there is literal blood. The phrase πρὸς Ἀργείων στρατῷ is as though a collective for all the blood. In I 2 the senses of dye textiles and stain things are distinguished in a clearer manner than in LSJ. The citation Ps.Democr. p. 46 is correctly placed, and Ath.480e [=11.480e in LSJ] is correctly identified as passive not middle as in LSJ. The citations Ar.Ach.112 and Ar.Pax1176 in I 2 and II 2 are placed far apart in the DGE layout, because the second is intransitive. The label ‘fig.’ is used for both, in contrast to ‘Com.’ for both in LSJ, and then ‘humorously’ and ‘fig.’ respectively in LSJ Revd Suppl. Labels are not used consistently by lexicographers with universally agreed meanings.⁴³ DGE correctly takes Ar.Pax1176 as middle not passive, unlike LSJ and LSJ Revd Suppl., and also explains what the reference to colour is about: se caga de miedo. At the end a more detailed etymology is given than in LSJ, with Szemerényi’s theory of a connection with βαθύς also mentioned (see n. 21). Understanding the technicalities of the reconstructed root with vowel gradation, the initial labiovelar, dissimilation, and the yod suffix, requires specialized knowledge. It is not mentioned that the yod suffix results in the form of the present according to a well-established phonetic change: *βαϕ-y-ω becomes βάπτω, ⁴⁰ R. Kassel and C. Austin, Poetae Comici Graeci: vol. III 2 Aristophanes: Testimonia et Fragmenta (Berlin/New York, 1984). Also, see below, p. 379, n. 86. ⁴¹ See above, p. 359, with n. 24. ⁴² Bailly’s Dictionnaire grec français (1950) under βάπτω sees analogy here as well. LSJ uses ‘by analogy, on the analogy of ’ (without abbreviation) to describe processes in phonetic change and word formation, e.g. under κῆρ, ἀνδράποδον, and very occasionally for changes in wording in phrases, e.g. μακρός I 3. ⁴³ On labels and the problems of editorial consistency in general, see Thompson 2017.

368

Anne Thompson

in the same way that there is ἅπτω from a stem ἁϕ- seen in ἁϕάω, or κρύπτω beside κρύϕα.⁴⁴ It may be noted that DGE does not give a list of principal parts.

2 0. 4. MO N T AN AR I GI AND GE Two new Ancient Greek dictionaries edited by Franco Montanari appeared in 1995 and 2015. The original work, Vocabolario della lingua greca, known as GI, was followed by an English version, The Brill Dictionary of Ancient Greek, GE.⁴⁵ The citations on which the article βάπτω depends are nearly all taken from LSJ, though for reasons of space they are fewer in number, since these lexicons are not as long. There are several innovations that make for a more user-friendly layout. Translation glosses are in bold type, which makes them easier for a reader to pick out across the different sections. GI and GE do not give quotations for all the passages cited but, where they do occur, there is a key innovation, which is to give translations in every case. Both glosses and quotation translations are in plain text, with the harder–to-read italics reserved for subsidiary material. Another key innovation is the introduction of indicators of context in Italian (English in GE), a helpful aid to comprehension. For contexts relating to dyeing there is di capelli (of hair), for the coated arrows, con veleno (with poison), for sinking into the sea, del sole (of the sun), di nave (of ships). There is also ferro (iron) for the tempering in Plutarch, and di fonte (of a spring) for the Marcus Aurelius passage. The tradition of including etymology is retained, here simplified and moved to a position after the headword. There is no Indo-European cognate, just Szemerényi’s suggestion with a question mark: ‘[βαθύς?]’. Presumably the reader is invited to make a connection between the depth of a liquid and a dipping action, but this is hardly a scientific exercise and not useful for the interpretation of any passage.⁴⁶ The principal parts are given at the end of the article in GI but moved to the beginning in GE, with others added for the reader’s convenience (for this verb the imperfect in all voices). Reference numbers for the parts are not given, other than where they happen to appear in the body of the article, but a lot of repetition is thus avoided. The forms are written in full in GE (e.g. βάψομαι, as opposed to -ομαι in GI with the stem to

⁴⁴ See M. Lejeune, Phonétique historique du mycénien et du grec ancien (Paris, 1972), 79, paragr. 68. ⁴⁵ For a critical review of GE, see Lee 2017. ⁴⁶ If there is a connection between these two words, it is unlikely a native speaker would have felt it.

An Illustration of the State of our Ancient Greek Dictionaries

369

be understood from βάψω.) The headword is sometimes abbreviated as β. in quotations in both GI and GE, as in LSJ, unlike in DGE.⁴⁷ There are three sections, with 1 for the active divided into subsections, transitive and intransitive. 2 is for the middle voice, with a segment marked ‘ass./abs.’ at the end, and 3 for the passive, separated from passages with similar senses to the active in 1. This is not the policy in LSJ and DGE for this particular verb, although elsewhere, for longer entries, the voices are kept separate. The ordering of active senses in 1 remains similar, moving from dipping or immersing, to dyeing textiles or hair, and then to dipping to draw water. The section begins with PLAT. Tim.73e, in defiance of chronological ordering, perhaps because it is thought to present a general sense. Following is OD. 9.392 [=Od.9.392 in LSJ] εἰν ὕδατι ψυχρῷ β. translated as ‘immergere nell’acqua fredda/to immerse in cold water (per temperare/to temper)’. But there is no mention of metal here, and the sense of to temper may not be recognized without that. The passages about plunging a weapon in flesh are labelled analog. (in GE, anal. in GI), as in DGE I 1 (‘por anal.’), apparently seeing the use as analogical after dipping metal objects in water for tempering. For EUR. Ph. 1577 [=E.Ph.1578] the translation in GE ‘to immerse the sword in flesh’ is not quite idiomatic English. LSJ Revd Suppl. and DGE recognized that VT Lev. 4.17 [=LXXLe.]⁴⁸ was wrongly placed with these citations in LSJ, since it is about dipping a finger in blood, not a weapon, but in Montanari the passage has not been moved. In the next segment, labelled ‘fig.’, there is a mistake in the translation of SOPH. Ai. 95 [=S.Aj.95]: ‘ἔβαψας ἔγχος εὖ πρὸς Ἀργείων στρατῷ having plunged the sword deep in the Argive army’, where the first word is not the aorist participle, and the sentence is a question.⁴⁹ The ⁴⁷ Many style decisions taken are to help learners. For those unfamiliar with classical authors and their works, abbreviations of names are expanded and printed in small capitals to highlight them further, e.g. A. in LSJ and DGE is AESCHL., Pl. is PLAT., Arist. is ARISTOT., MAUR. is for Marcus Aurelius Antoninus, more recognizable than LSJ and DGE’s M.Ant. There are spaces between the author names, the works, and line or paragraph references: PLAT. Tim. 73e (LSJ and DGE: Pl.Ti.73e). Subsection numbers are given for the overly long chapters in the historians: e.g. HDT. 7.67.1 (Hdt.7.67). Plu.2.136a is cited in GI with the essay number in Moralia: PLUT. 11.136a, changed in GE to the Latin title: PLUT. Sanit. 136a (de tuenda sanitate praecepta; in the list of authors and works, xlvii, De tuenda sanitate is given, with praecepta omitted; the Greek title is ὑγιεινὰ παραγγέλματα.) A decision was made in LSJ to use shorter abbreviations for many authors and works than in previous editions, in order to save space for additional material: e.g. S. for Soph. (Sophocles), M.Ant. for M.Anton., Pl.Ti. for Plat.Tim. The original style (without works and reference numbers) can be seen in IGL, where it is conveniently appropriate for a beginner’s dictionary. Usage labels in GI and GE are mostly similar to those in LSJ and DGE, but some abbreviations are longer to make them more recognizable. Section numbers are in white against the background of a dark square in GI, a dark circle in GE. Subsection letters are highlighted by being in small capitals and placed in a small box. Much thought has gone into making the columns readable compared to dense columns of print in other dictionaries. ⁴⁸ VT is VECCHIO TESTAMENTO. ⁴⁹ P.J. Finglass (Sophocles: Ajax (Cambridge, 2011), 165, on line 95) translates ‘Have you dipped your sword thoroughly in the Argive army?’ He argues that πρός is an equivalent of ἐν

370

Anne Thompson

label ‘fig.’ is also used for AESCHL. Pr. 863 [=A.Pr.863] ἐν σϕαγαῖσι βάψασα ξίϕος, but there is actual dipping in blood in both passages, so it is a questionable choice. For ARISTOPH. Ec. 216 [=Ar.Ec.216] Montanari sees the sense as wash (not dye).⁵⁰ The segment relating to dyeing textiles and hair in GE begins with to soak, dye. The first gloss soak helps with explaining the dyeing process for textiles, though it cannot actually be a translation. It does not work at all in the case of dyeing hair. The passage about coating arrows with poison strangely appears at the end of this segment: ‘ἔβαψε ἰούς he poisoned arrows’, SOPH. Tr. 574 [=S.Tr.574]. The sequence ‘soak, dye . . . with poison’, however, does not follow on idiomatically. ‘dip’, ‘coat’ or ‘smear’ would be more natural translations. Also, ‘he’ appears to be wrong, because with this reading the subject is the Hydra, not Herakles. In the next segment, to draw is not long enough for a definition of βάπτω, and it is ambiguous, since there are many senses of the English verb. The middle voice section 2 has for its first quotation ‘ποταμοῖο ἐβάψατο it immersed itself in the river ARAT. 951’ [=Arat.951]. This is the only example under the translation glosses to immerse oneself, dip oneself, which are already somewhat tautologous, and the translation of the quotation adds to this. There is no indication of the subject in Italian or English as is the new policy for other quotations. Any innovation can give rise to unexpected problems. Here, ‘it’ is puzzling because it seems inanimate, but saying that ‘it’ is a crow would also be puzzling, because crows do not do this. The point is that the unusual behaviour is said to presage a storm.⁵¹ LSJ includes βάψομαι (future middle) from Ar.Lys.51 in the list of principal parts, but does not assign it to a sense section. DGE has the citation in I 2, defined as hacer teñir. GI has tingere da sé, followed by a single-word quotation βάψομαι and a translation that hesitates between two alternatives: ‘mi farò tingere o mi tingerò’. In GE this becomes to color oneself, which is not very clear, and in the translation ‘I will dye or stain’, ‘stain’ does not suit the passage. More of the line is needed to make sense of it. DGE has a longer quotation, the full line being κροκωτὸν ἄρα νὴ τὼ θεὼ ’γὼ βάψομαι in that case, by the Two Goddesses, I’ll have a dress dyed saffron.⁵² DGE might be right to here: ‘The regular senses of πρός, “on”, “at”, or “upon” are all inappropriate.’ A.F. Garvie (Warminster, 1998) has ‘did you dye your sword thoroughly in the army of the Argives?’, arguing that for ‘did you dip’ one would expect πρός + accusative. A sense ‘in an engagement with’ may also be possible, see Smyth 1920, paragr. 1695 2. (‘Occupation’); cf. LSJ πρός B II. But the passage is placed under LSJ πρός B I, expressing ‘proximity’. R.C. Jebb (Cambridge, 1896) has ‘have you dyed your sword well in the Greek army?’ and in his note says ‘ “on” them denoting an encounter at close quarters’. H. Lloyd Jones (Cambridge, MA, 1994) opts for ‘have you well stained your sword in the blood of the Argive army?’ This one example reminds that there are good arguments for not translating quotations. ⁵⁰ See above, p. 366, with n. 38. ⁵¹ See D. Kidd, Aratus: Phaenomena (Cambridge, 1997), 503. ⁵² This translation is from J. Henderson, Aristophanes vol. III (Cambridge, MA, 2000).

An Illustration of the State of our Ancient Greek Dictionaries

371

place this example in the transitive section, though to color oneself might be nearer the mark, if κροκωτόν were an internal accusative: I will dress myself colourfully with a saffron dress.⁵³ [DEMOCR.] Phys. 46 B [=Ps.Democr. p. 46 B] is moved to its own segment, divided off with upright bars, under the label ass./abs. (absolute). A separate segment is better than integration with the dyeing of textiles and hair examples, but it has been added in section 2 for the middle voice, and the example is active (see DGE). There may have been misinterpretation of LSJ, where the label ‘Med.’ was inserted in I 2 with no signal as to when the label ceases to apply. LSJ’s ‘of gilding and silvering’ become the gloss translations: dorare, argentare/gild, silver. But the senses on p. 46 B (as noted above, p. 361) are about changing the colour of a substance to gold through mixing (not about applying gold to a surface), and about dipping something in oil.⁵⁴ In 3, the section devoted to the passive, the sense in the third segment is defined as essere tinto/to be colored, with the reference COM. Com. 53 in GI, and COM. 76 in GE. This is a new citation not in LSJ or DGE. In LSJ the reference would be Com.Adesp.53D,⁵⁵ as under εὐρώς, one of the words in the context: εὐρῶτι γήρως τὰς τρίχας βεβαμμένος having hair dyed with the blight of old age. It is hard to see from the introductory lists of authors and abbreviations in GI and GE which editions are used.⁵⁶ The translation glosses will cover this context relating to hair, but it is only figuratively about dye, so not quite the same as in PLAT. Rp. 429e [=Pl.R.429e] which is about dyed textiles, examples of which come in the earlier segment of the section, essere immerso o tuffato/to be immersed or dipped, where neither to be dyed nor to be colored is given as a gloss. The comedy in the line ARISTOPH. Pax 1176 [=Ar.Pax1176] in the first segment is not signalled with a label as in other works. ⁵³ See Chadwick 1996, 14 16 on the difficulty of making choices between different interpret ations, and cases where the lexicographer feels the need to defy all authorities. ⁵⁴ Silver is mentioned elsewhere in the work, see above, p. 361. ⁵⁵ D. signals the work J. Demiańczuk, Supplementum Comicum (Krakow, 1912). [Cracow in LSJ.] ⁵⁶ Four editions are mentioned. Since different readings or reference numbers can cause difficulty, the preferred editions should be specified clearly in the initial lists. For example, Men.363.4 in LSJ is updated to 303.4 from the Oxford Classical Text in DGE and Montanari GI (F. Sandbach, Menandri reliquiae selectae (2nd edn, Oxford, 1990)), and GE has the more recent Kassel and Austin edition number 264.4 (R. Kassel and C. Austin, Poetae Comici Graeci: vol.VI 2 Menander, 1998, Berlin/NY). This is a necessary updating, but several editions are listed in GI and GE without any signal as to the preferred text. Antiph.25 in LSJ is updated to 26.3 but it is hard to determine the edition used because the list gives four. The line number EUR. Ph. 1577 (for E.Ph.1578 in LSJ and DGE) seems to be an error. The editions referred to in the list of authors and works, J. Diggle, Euripidis Fabulae (Vol. III, Oxford, 1994) and D.J. Mastronarde, Euripides: Phoenissae (Leipzig, 1988), both have the word in 1578. For D.H.5.15 εἰς τὰ πλευρὰ β. τὴν αἰχμήν in LSJ I 1b, GI and GE (DION. 5.15) have εἰς τὰς πλευρὰς β. . . . But GI and GE list the same Teubner edition as LSJ, with the Loeb edition added, which also has εἰς τὰ πλευρά: C. Jacoby, Dionysius Halicarnassus: Antiquitates Romanae (Leipzig, 1885 1905); E. Cary (Cambridge, MA, Loeb, 1937 50).

372

Anne Thompson

The quotation βεβαμμένος baptized/battezzato is the last sense in 3, with LSJ’s reference Arr.2.9.20 changed, as in DGE, to reflect that these are discourses of Epictetus as recorded by Arrian (Montanari has ARR. Epict.D; DGE, Arr.Epict.). A general observation is that incorporating material from previous dictionaries, along with presenting new ideas, also moving between English and Italian, is fraught with dangers in terms of accuracy.⁵⁷ An example of the attention to detail required is an epigram of Lucillius that appears in the GI article as LUCIL. AP 11.68.1 [=AP11.68 (Lucill.) in LSJ]. DGE introduced into its lists another author, Gaius Lucilius, a Roman poet who used Greek vocabulary.⁵⁸ He was not included in LSJ’s author list, though he does appear, with name unabbreviated, in LSJ Revd Suppl.’s list. DGE uses the abbreviation Lucil., to differentiate from Lucill. the Greek poet, hence here AP11.68(Lucill.) as LSJ. In GI and GE’s author list, there is LUCIL. for the Latin poet, and LUCIL.¹ for the poet of the Anthology. The superscript numeral is not present under βάπτω in GI, but that is corrected in GE.⁵⁹ The situation is more complicated because, despite his absence in the author list, Lucilius was included as Lucil. in LSJ in a few entries, where he is cited as quoted by other authors, e.g. under ἀγέλαστος, σκληρώδης (in Cicero, Gellius). He also appears at least once, under συμμειρακιώδης, as Lucil.fr.187 (in the Teubner edition,⁶⁰ the edition given in the LSJ Revd Suppl. and DGE lists.) The meaning given in LSJ is altogether childish, but uncertainty about the text is marked with a label ‘(dub.)’. LSJ Revd Suppl. instructs deletion of this entry.⁶¹ GI and GE keep the entry, as LUCIL. fr. 5.187 (without any expression of doubt), though there is no mention of this author under ἀγέλαστος, σκληρώδης, perhaps because the words are attested elsewhere or because the origin of these entries is independent of LSJ. The meaning is given as childish, youthful in GE (words ⁵⁷ Although the English edition was overseen by editors in the United States, there is evidence that the first adaptors were not native English speakers. Apart from obvious unidiomatic phrasing, there are difficulties with small details. For instance, LSJ’s S.Aj. becomes SOPH. Ai. in GI (which could be Greek, Latin, or Italian; cf. S.Ai. in DGE), but S.Ai. is retained in GE, which is not the English spelling Ajax. For LXX (Septuagint) in LSJ, GI has VT (VECCHIO TESTAMENTO), and this is retained in GE, explained in the list in English and Greek as ‘Old Testament / παλαιὰ διαθήκη’ (but not Latin, Vetus Testamentum, which would have given the same abbreviation). ⁵⁸ See H. Petersmann, ‘The Language of Early Roman Satire: its Function and Characteristics’, in J.N. Adams and R.G. Mayer (eds), Aspects of the Language of Latin Poetry (London, 1999), 298 310; also J.N. Adams, Bilingualism and the Latin Language (Cambridge, 2003), 10, 19 20, 77, 326 7, 353 5. ⁵⁹ Nicol.Com. in LSJ becomes Nicol. in GI and GE, distinguished in the introductory list from two other authors Nicol.¹ and Nicol.². ⁶⁰ F. Marx, C. Lucilii: Carminum reliquiae (Leipzig, 1904). ⁶¹ See H. Petersmann, 300 n. 16 (see n. 58). E.H. Warmington, Remains of Old Latin vol. III: Lucilius. The Twelve Tables (Cambridge, MA, Loeb, 1938), 60, has a different reading: quod atechnon et Eisocration lerodesque simul totum ac sit meiraciodes (for ac symmiraciodes in manuscripts). The GI and GE entries for συμμειρακιώδης apparently define meiraciodes rather than symmiraciodes.

An Illustration of the State of our Ancient Greek Dictionaries

373

which are not quite synonymous), translated from infantile, giovane in GI, again indicating independence from LSJ.⁶²

20.5. JOHN CHADWICK, LEXICOGRAPHICA GRAECA Chadwick analysed the senses of βάπτω in this book (59–62).⁶³ He had been a member of the British Academy committee overseeing the LSJ Revd Suppl. project, and his ideas may have been formed when reading a draft of the article there. He says that ‘the danger of using translation words instead of lexicographic definitions is well demonstrated by LSJ’s treatment of this word’, but he does not specifically explain why he thinks this.⁶⁴ He judges the LSJ Revd Suppl. article to be an improvement, but suggests a replacement which he sketches out with a commentary. Although he uses LSJ as a source, he makes clear in the introduction that he is claiming a radical departure from the LSJ tradition by using the very different methodology of the OLD. Chadwick makes the point that the Greek quotations in LSJ are short, and throughout his book he expands them, though he says this would not be possible in a printed lexicon.⁶⁵ On average most are longer than in DGE. However, translations are not provided, which is a pity because, although these would have taken up space, the book would have been opened up to a wider audience.⁶⁶ For the meaning, Chadwick expands on the translation word dip in LSJ: ‘The basic sense is to plunge or dip (in a yielding medium, usually but not necessarily a liquid)’. In English plunge and dip have other senses, so the additional information specifies which is meant. Od.9.392 is placed at the beginning as an illustration of this ‘basic sense’. However, the purpose of dipping the metal implements in cold water is not mentioned specifically, and can only be picked up by the reader from the Greek quotation.⁶⁷ Passages about tempering metal are about a special purpose for dipping in a liquid, from the vocabulary of manufacturing, rather than a general context, so the

⁶² Lucil. does appear under ἀγέλαστος in DGE. On resources used in preparing GE, see Lee 2017. ⁶³ Chadwick 1996. ⁶⁴ For early complaints about the entry in Liddell and Scott, see Scott 1920 1. ⁶⁵ Chadwick 1996, 26; also see Glare 1987, 17 on excessively short quotations. ⁶⁶ The introduction of translations does, however, introduce many further difficulties, with huge potential for inconsistencies across a whole work; see above, n. 49. ⁶⁷ It is debatable whether the participle ϕαρμάσσων in the passage can be translated as ‘tempering’. According to A. Heubeck and A. Hoekstra, Homer: Odyssey vol. II (Oxford, 1989), 34: ‘ϕαρμάσσω (hapax) is a technical term (“treat with a ϕάρμακον [here, liquid to temper the metal]; harden”)’.

374

Anne Thompson

position of this as first is questionable. It would be justified if the policy were for a strictly chronological arrangement, since this is the first attestation in literature, but Chadwick believes in a ‘logical’ ordering of senses.⁶⁸ For use with a different preposition (εἰς + accusative, rather than εἰν + dative in Od.), Pl.Ti.73e is cited next. As already noted, this is about creation of a material (bone) that uses language most likely influenced by manufacturing terminology, so again it is not the most general sense. Arist.HA605a29 is also cited, where the context is honey rather than water. Honey can be thought of either as a liquid or as a yielding medium, and similarly wax in Arist.de An.435a2, as a passage added for comparison. With the honey, however, the purpose is to take up some of it on to animal fodder, a different kind of context. Chadwick says this basic sense ‘leads naturally to the use which emphasizes the result of the immersion, to make wet’. He uses the word ‘immersion’ here, rather than ‘plunging’ or ‘dipping’, and one wonders if immerse could have been another translation gloss in the previous section. In this make wet category he places Ar.Ec.216, from LSJ I 1a, along with LSJ’s S.Tr. poison examples from I 1c. The definition make wet is a kind of metalanguage: ‘make wet (wool in hot water)’ and ‘make wet (arrows with poison)’ are unidiomatic; ‘dip or soak wool’ and ‘smear or coat arrows’ would be more natural. But the passages have somewhat different senses arising from a difference of purpose. In the case of the wool, if this is about cleaning or preparation, none of the liquid is deliberately taken up (that is, if it is not about dyeing, as many take this).⁶⁹ For the arrows, the purpose is the taking up of some of the liquid in which they are dipped. The LXXLe.4.17 passage is transferred here from LSJ I 1b (the ‘slaughter’ section), as in LSJ Revd Suppl. and DGE. Chadwick explains that although LSJ added the citation to the ‘slaughter’ section because it is about a sacrificial victim and blood, the object dipped is a finger, so it does not fit with the weapon examples. He notes also that the preposition ἀπό introduces the instrument. However, the purpose is to take up some of the liquid, as with the arrows, so the preposition ἀπό might not be about the instrument, but the source.⁷⁰ The passive in Dsc.5.121 (s.v.l.) is also added here. Chadwick quotes the variant text as κουράλλιον . . . ἔξαλον γινόμενον καὶ βαπτόμενον ἤτοι πηγνύμενον.⁷¹ LSJ has become hard, but Chadwick is right to

⁶⁸ Chadwick 1996, 18 20. ⁶⁹ See above, n. 38. ⁷⁰ This preposition may rather be a marker of the source from which the finger becomes coated. Sometimes this genitive of source could be classified as means or instrument: e.g. στράτευμα συνέλεξεν ἀπὸ χρημάτων he raised an army from financial resources X.An.1.1.9 could be seen as by means of money. See Smyth 1920, 374, paragr. 1684 c. (4), the ‘Means, Instrument’ category. ⁷¹ The TLG text, with ἁπτόμενον, is quite different, though the general meaning is similar: τὸ δὲ κουράλιον, ὅπερ ἔνιοι λιθόδενδρον ἐκάλεσαν, δοκεῖ μὲν εἶναι ϕυτὸν ἐνάλιον, στερροποιεῖσθαι δέ, ὅταν ἐκ τοῦ βυθοῦ ἑλκυσθῇ τῆς ἁλὸς ἁπτόμενον τοῦ περικεχυμένου ἡμῖν ἀέρος (M. Wellmann, Pedanii Dioscuridis Anazarbei de materia medica libri quinque (Berlin, 1906 14)).

An Illustration of the State of our Ancient Greek Dictionaries

375

say that this is already expressed in the other participle πηγνύμενον. As it stands, the example should perhaps be near Pl.Ti.73e, because it seems to describe a process of creation influenced by manufacturing terminology. Chadwick places baptize at the end of this section as a ‘special sense’. Although there is some similarity to the action of dipping in the other cases, the purpose is different. Chadwick then begins a new section: ‘Both of these usages reappear in a group of examples where the object is a weapon.’ He means that plunge needs to be taken from the first sense (dip is less idiomatic), and make wet is the result of plunging a weapon in blood. One can plunge a weapon in an enemy’s body (E.Ph.1578, D.H.5.15.2), but one can also wet the weapon with blood (for which A.Ch.1011 is mistakenly cited, because the weapon is the subject in this line). In poetic passages there can be an allusion (A.Pr.863, S.Aj.95) to another sense: dip in a colouring medium, dye. The weapon is coloured red with blood as though dyed.⁷² Chadwick continues: ‘Another special use is where the object is a bucket or similar container, which is dipped in a liquid in order to draw it up’ (E.Hec.610, Antiph.25, Thphr.Char.9.8). ‘A different sense, but clearly derived from this is: to draw (a liquid) by dipping’ (Theoc.5.127). Chadwick has not recognized it, but this is somewhat close to LXXLe.4.17, assigned to the make wet section, because the purpose there of dipping the finger in blood is to draw up some of it to sprinkle in a ritual. The next section is also a ‘special’ sense: ‘to dip in a colouring medium, dye’ (Hdt.7.67.1, Pl.R.429d, Ath.11.480e). The Aristophanes citations, where the verb is accompanied by an internal accusative, are added here (Ar.Ach.112, Ar.Pax1176), with the same description ‘humorously’ as in LSJ Revd Suppl. The final section lists citations where the verb is intransitive, ‘to plunge oneself ’, employing the same ones as in LSJ: of eels (Arist.HA592a18),⁷³ of a ship (E.Or.707), of the sun (Arat.858, also middle in Arat.951, of a crow). This last section ends with Babr.71.2 where the construction with internal accusative is pointed out: κῦμα κυρτόν in the swelling sea. Since the publication of Lexicographica Graeca, an in-depth study of βάπτω and words with the same stem has been published by Everett Ferguson.⁷⁴ It is influenced by Chadwick’s work, beginning with a section ‘Literal Usage’ where it is said that ‘the basic meaning of βάπτω is “to plunge”, “to dip” in a yielding medium, usually a liquid’. The lexicographer has a heavy responsibility in that a definition may be repeated verbatim as authoritative. The difficulty of ⁷² See D.J. Mastronarde, Euripides: Phoenissae (Cambridge, 1994), 589, on line 1578. ⁷³ This passage is doubtful; see an alternative interpretation in the analysis of passages below, p. 379. ⁷⁴ See Ferguson 2009. The title ‘Words from the Bapt root’ is not quite accurate in that the root stem is actually baph ; the t suffix, as has been stated above, is a feature of the present tense only.

376

Anne Thompson

finding the right words for translation glosses and lexicographic definitions cannot be underestimated. There is also the challenge of the essential discipline of separating as far as possible the meaning of the headword from the context. One can supplement or replace translation glosses dip and plunge with a phrase describing their precise meaning, though there is an ever-present danger of defining the English word rather than the Greek. For the first transitive sense of dip, the English verb nearest in meaning to βάπτω, OED has: ‘To put down or let down temporarily or partially in or into a liquid, or the like, or the vessel containing it (usually with the notion of wetting, or of taking up a portion of the liquid, etc.); to immerse; to plunge (but with less implication of force and splashing, the sound of the word expressing a light though decided act).’ In framing a definition for Greek, the features here could be assessed for their appropriateness. βάπτω is nearly always about an action that implies removal following the initial action. The range of the action is from a relatively light to a forceful one, and from slow to quick (making both dip and plunge suitable translations in different contexts). It would probably not be worthwhile making these last distinctions in definitions. More important is the result or purpose of the action. Fundamental in Greek is whether any of the liquid is purposefully taken up. When it comes to context, separating this from the meaning of the headword in the wording of definitions is not straightforward. The verb implies the breaking of the surface of a liquid or liquid-like substance, the nature of which may or may not be expressed in accompanying words. Chadwick’s first definition has the liquid as context, but in the case of the sense dye, dip in a colouring medium, he has the liquid as part of the meaning of the headword. In contexts about dye, unless the actual process is being described, the dipping action and liquid tend to become lost, with the focus moving towards the resulting colour. It cannot be said that in Od.9.392 βάπτω in itself contains no idea of a liquid before the cold water is mentioned. Under tingo, OLD has for the first sense (1 a) ‘To wet by plunging into a liquid, dip’, where the liquid is seen as part of the meaning of the headword, though this may be amplified in particular quotations with the kind of liquid, as in Vergil, Aeneid 7.811, when Camilla would speed across the sea: ‘nec tingeret aequore plantas nor would she wet her feet in the waves’. The solution therefore may be to write something like ‘dip or plunge in a liquid (very often with the kind of liquid expressed in the context)’. The vast majority of contexts for this verb are about liquids. The contexts of honey or wax may have prompted Chadwick’s wording ‘or in a yielding medium’, though they could also be thought of as liquid-like, semi-liquids or viscous substances. The move away from liquids takes place in small steps and does not progress far. When there is a move from water to blood, it is only by a small extension that flesh stands in for blood. This probably began in poetry, but became standard in prose. Abstract senses also develop, again particularly in later prose, deriving from the idea of dye.

An Illustration of the State of our Ancient Greek Dictionaries

377

Chadwick has made improvements to the LSJ arrangement along with corrections. The main difference from LSJ Revd Suppl. is the distinction between dip or plunge and make wet. Of key importance is the emphasis on relating senses one to another step by step, setting up a logical framework.⁷⁵ Chadwick does this partly by picking up the same definition words: to plunge or dip are picked up in to plunge oneself, dip in a colouring medium. Relating senses in this way imposes order on the material and aids the reader’s comprehension. To some extent, a good framework positing these developments will reflect historical reality, but it cannot be exact.⁷⁶ The idea of a ‘basic sense’ in particular is questionable. This can only be an artificial construct to express semantic components which all instances of the word have in common, usually expressed in a kind of metalanguage. But, as has been seen with the sense dye, where in many instances there is hardly any idea of dipping or liquid, even this breaks down. All that can be done is to arrange the article with a first sense that is as neutral as possible. This, however, can also break down. For instance, the most neutral sense of a word is often to be found in the arguments of philosophers, yet we know for certain that this does not come first chronologically. A logical ordering can only be just that, with no pronouncements about the actual development of senses.

20.6. BIBLICAL LEXICOGRAPHY Several new dictionaries of biblical Greek have also appeared in recent years. In this field it is now generally acknowledged that lexicographic definitions should be used in larger dictionaries in preference to translation glosses. Two that do not rely solely on glosses are Takamitsu Muraoka’s Lexicon of the Septuagint and Frederick Danker’s revision of the Bauer New Testament dictionary, known as BDAG. But the results are not always convincing.⁷⁷ For βάπτω Muraoka offers: ‘1 to place into liquid, ‘dip, immerse’, also, later in the section ‘ . . . into sth⁷⁸ that contains liquidlike substance’ (in the quotation, ‘into the honeycomb’), then ‘2 to moisten by forming contact with ⁷⁵ See also Chadwick 1986 and 1992b. ⁷⁶ For the third edition, the OED has changed from an arrangement according to plausible semantic development to a chronological one according to date of attestation: see https://public. oed.com/history/oed editions/preface to the third edition/#chronology ‘Chronology and the historical method’. Both systems have advantages and drawbacks. Chadwick’s piece raises questions for us about lexicographic terminology and methodology. Terms such as usage, use, meaning, sense often appear as though they were completely interchangeable. Lexicographers would do well to start out with a clear idea of what each means. ⁷⁷ See the review of Muraoka’s lexicon by Lee 2009, also the review of BDAG by Roberts 2004. ⁷⁸ sth = something

378

Anne Thompson

liquid.’ This offers something very similar to Chadwick’s improvement to LSJ, a division between a dipping action and the resulting wetness.⁷⁹ But in 1, to place is not adequate as a definition on its own. In 2, the phrase by forming contact is rather awkward wording, and tautologous in that moisten contains this on its own.⁸⁰ Danker has ‘to dip somethg. in a liquid, dip, dip in’, with the bold plain type distinguishing the lexicographic definition (or, in Danker’s terminology, extended definition⁸¹) from the italic translation glosses. The object of the verb, which is contextual, is not separated in the wording from the meaning of the headword. Biblical dictionaries tend to offer dip in beside dip.⁸² There is probably no need for dip in. For example, in Ev.Luc.16.24 ἵνα βάψῃ τὸ ἄκρον τοῦ δακτύλου αὐτοῦ ὕδατος so that he may dip the tip of his finger in water, the preposition ‘in’ can be taken as represented by the following genitive case and not part of the headword.

20.7. FUTURE APPROACHES There are several noteworthy improvements in dictionaries published since LSJ. There are longer quotations, sometimes with translations, more contextual and sociolinguistic material, and (most important) a move away from the translation gloss to lexicographic definition. However, revision of material from existing reference works, carried out alongside incorporation of new material and methods, gives rise to inconsistencies and inaccuracies, especially if there is no rereading of the original texts. The complexity of detail is always going to be more challenging than expected. Some anomalies are trivial, but there are also very substantive issues. The sense sections have not moved much away from what is presented in LSJ. For a proper reassessment, rereading of the source texts is going to be essential. Layout and presentation are important but secondary. It is possible to imagine new ways of presenting lexicon entries in the future, especially now in the electronic age when there is the possibility of escape from the linear alphabetic layout of the printed page. Below is a fuller analysis of all the source material for this verb in the dictionaries examined above. This is only a preliminary version, since for any ⁷⁹ The article distinguishes between Le.4.6 and 4.17 [Le. = LXXLe.], dividing them over the two sections: βάψει ὁ ἱερεὺς τὸν δάκτυλον εἰς τὸ αἷμα καὶ προσρανεῖ ἀπὸ τοῦ αἵματος and βάψει ὁ ἱερεὺς τὸν δάκτυλον ἀπὸ τοῦ αἵματος τοῦ μόσχου καὶ ῥανεῖ, but possibly both belong in 2 because, despite the following prepositions, the purpose in both cases is to take up some of the blood. ⁸⁰ There is also a section ‘1 b. mid. “to bathe” . . . ’ a citation that is about ritual washing for healing purposes [4 Kings 5.14], but the text has ἐβαπτίσατο, and the reference also appears under βαπτίζω. ⁸¹ BDAG, Introduction, p. viii. ⁸² Also, Louw and Nida 1988; Trenchard 2003; Lust, Eynikel, and Hauspie 2003.

An Illustration of the State of our Ancient Greek Dictionaries

379

lexicographic project everything needs to be checked and discussed by a team.⁸³ Particular attention is paid to the wider context and also the social context, a methodology that is key to the identification of senses. For a verb, it is also appropriate to look at the subjects and objects.⁸⁴ Persons are the subject of the verb in all sections until the last, where mostly things are the subject. Middle, passive, and intransitive examples are placed with the active transitive examples that are comparable in sense. The perfect passive participle is labelled as an adjective where appropriate. At least one quotation with a translation is given for each section as an illustration. It should be noted that where the idea of a liquid or other medium is part of the meaning of the headword, this may also be specified in the context, as in the first example: ‘in cold water’. 1 DIP or PLUNGE (someone or something) IN WATER Arist.HA592a18 καὶ τὰς μεταβολὰς δ’ οὐχ ὑπομένουσι τὰς ἰσχυράς, οἷον καὶ τοῖς ϕέρουσιν ἐὰν βάπτωσιν εἰς ψυχρόν · ἀπόλλυνται γὰρ ἀθρόαι πολλάκις and they (αἱ ἐγχέλυς eels) cannot bear harsh changes, as is the case for those transporting them if they plunge them in cold water; for often whole batches die (This seems a good starting point, but LSJ and the other dictionaries see the example as intransitive, i.e. the eels plunge themselves.⁸⁵) [Also : Ar.fr.237 K.-A = 225 in LSJ; but the text and interpretation ‘dunk a person’ are uncertain; interpreted in LSJ as ‘dip (the oar’).⁸⁶ Secure examples need be found for this first sense.] b passive: (of persons, things) BE DIPPED or PLUNGED IN WATER LXXJo.3.15 οἱ πόδες τῶν ἱερέων τῶν αἰρόντων τὴν κιβωτὸν τῆς διαθήκης κυρίου ἐβάϕησαν εἰς μέρος τοῦ ὕδατος τοῦ Ιορδάνου the feet of the priests carrying the ark of the Lord’s covenant became wet at the edge of Jordan’s waters c passive perfect participle as adjective: (of persons) WET, SOAKED or DRENCHED

βεβ[αμ]μένου, τρέμοντος drenched, trembling Men.Dysc.657 2 intransitive: DIP or PLUNGE INTO WATER Arat.858 εἰ δ’ ὁ μὲν ἀνέϕελος βάπτῃ ῥόου ἑσπερίοιο if he (the sun) dips without a cloud into the western ocean Babr.71.2 (of a ship) ἰδὼν γεωργὸς νῆα ναυτίλων πλήρη ⁸³ Comparison with words with the same stem or synonyms in particular is likely lead to some revision. ⁸⁴ Other parts of speech will require different approaches. ⁸⁵ For the text and interpretation, see J. Barnes, The Complete Works of Aristotle (Princeton, 1984), 927, translation and note. ⁸⁶ J. Taillardat (Les images d’Aristophane: études de langue et du style (Paris, 1965), paragr. 213) sees this as a synonym of ἰέναι from maritime colloquial vocabulary, French souquer.

380

Anne Thompson

βάπτουσαν ἤδη κῦμα κυρτὸν ἐκ πρῴρης a farmer seeing a ship full of sailors at that moment dipping from the prow under the water in the swelling sea [also : E.Or.707] b middle: DIP or PLUNGE ONESELF INTO WATER Arat.951 ἤ που καὶ ποταμοῖο ἐβάψατο μέχρι παρ’ ἄκρους | ὤμους ἐκ κεϕαλῆς or maybe it (a crow) dipped head and shoulders into the river 3 DIP or PLUNGE (something) INTO A LIQUID (other than water), SEMI-LIQUID or VISCOUS SUBSTANCE Arist.de An.435a2 οἷον εἰ εἰς κηρὸν βάψειέ τις, μέχρι τούτου ἐκινήθη, ἕως ἔβαψεν as for instance if someone dips into wax, the wax is moved up to the point where the dipping is complete 4 DIP or PLUNGE (a weapon) IN BLOOD or FLESH; (sometimes perhaps with allusion to 12 or 14) DYE RED or STAIN (a weapon) IN BLOOD A.Pr.863 γυνὴ γὰρ ἄνδρ’ ἕκαστον αἰῶνος στερεῖ | δίθηκτον ἐν σϕαγαῖσι βάψασα ξίϕος for each woman will deprive her husband of his life, dipping a two-edged sword in his blood S.Aj.95 ἔβαψας ἔγχος εὖ πρὸς Ἀργείων στρατῷ; did you dye your sword well in blood in your encounter with the Argive army? E.Ph.1579 χαλκόκροτον δὲ λαβοῦσα νεκρῶν πάρα ϕάσγανον εἴσω σαρκὸς ἔβαψεν and taking a wrought bronze sword from the dead she plunged it into her flesh D.H.5.15.2 ὁ μὲν εἰς τὰ πλευρὰ βάψας τὴν αἰχμήν, ὁ δ’ εἰς τὰς λαγόνας the one plunging the point of his spear into the ribs, his opponent into the flank [also : Ach.Tat.3.15.4; Ach.Tat.1.4.3 in DGE I 1 is a mistake apparently for 3.15.4; see below 14] b middle AP7.531.2 (Antip.) βαψαμένα κοίλων ἐντὸς ἄρη λαγόνων | μάτηρ your mother plunging the weapon into the hollow of your side 5

DIP or PLUNGE (someone or something) IN WATER IN A CLEANSING or INITIATION RITUAL

Cyr.H.Procatech.2 καὶ τὸ μὲν σῶμα ἔβαψεν ὕδατι, τὴν δὲ καρδίαν οὐκ ἐϕώτισε Πνεύματι he dipped his body in water, but did not enlighten his heart with the Spirit b passive: (of persons, things) BE DIPPED or PLUNGED IN WATER IN A CLEANSING or INITIATION RITUAL LXXLe.11.32 πᾶν σκεῦος, ὃ ἐὰν ποιηθῇ ἔργον ἐν αὐτῷ, εἰς ὕδωρ βαϕήσεται καὶ ἀκάθαρτον ἔσται ἕως ἑσπέρας each utensil in which work may be done will be immersed in water and will be unclean until evening Cyr.H.Procatech.2 ἐβαπτίσθη, ἀλλ’ οὐκ ἐϕωτίσθη he was baptized, but was not enlightened

An Illustration of the State of our Ancient Greek Dictionaries

381

Arr.Epict.2.9.20 ὅταν δ’ ἀναλάβῃ τὸ πάθος τὸ τοῦ βεβαμμένου καὶ ᾑρημένου, τότε καὶ ἔστι τῷ ὄντι καὶ καλεῖται Ἰουδαῖος when he has assumed the condition of a man who has been ritually dipped in water and made his choice,⁸⁷ then he truly is and is called a Jew [also: Eu.Fr.Pap.2.43] 6 DIP or PLUNGE (something, especially metal or other materials) IN WATER or A LIQUID IN A PROCESS OF CREATION, PREPARATION or MANUFACTURE Od.9.392 ὡς δ’ ὅτ’ ἀνὴρ χαλκεὺς πέλεκυν μέγαν ἠὲ σκέπαρνον εἰν ὕδατι ψυχρῷ βάπτῃ μεγάλα ἰάχοντα as when a blacksmith plunges a great axe or adze hissing loudly into cold water (in order to temper the metal) Pl.Ti.73e καὶ μετὰ τοῦτο εἰς πῦρ αὐτὸ ἐντίθησιν, μετ’ ἐκεῖνο δὲ εἰς ὕδωρ βάπτει, πάλιν δὲ εἰς πῦρ, αὖθίς τε εἰς ὕδωρ and after this he (the creator) placed it (a mixture of earth and marrow) in fire, after that dipped it in water, and from this back to fire, and once again in water Ps.-Democr. p. 46 ἐὰν οὖν ἐκστραϕῇ, κατάβαψον αὐτὴν εἰς ἔλαιον κίκινον πολλάκις πυρῶν καὶ βάπτων if then it transmutes, dip it completely into castor oil, heating and dipping repeatedly [also: Ps.-Democr. p. 48] b passive: (of things, especially metal or other materials) BE DIPPED or PLUNGED IN WATER or A LIQUID IN A PROCESS OF CREATION, PREPARATION or MANUFACTURE

Hp.Coac.378 πομϕόλυγος δὲ ὑποπελίου γινομένης ἐπὶ τῆς γλώσσης ἐν ἀρχῇ, οἵη σιδηρίου βαϕέντος ἐς ἔλαιον when at the beginning a somewhat livid blister comes up on the tongue, such as forms with a cautery iron dipped in oil Plu.2.136a ὥσπερ τὸν βαπτόμενον σίδηρον like iron being tempered by dipping Paus.2.3.3 καὶ τὸν Κορίνθιον χαλκὸν διάπυρον καὶ θερμὸν ὄντα ὑπὸ ὕδατος τούτου βάπτεσθαι λέγουσιν and fired and red hot Corinthian bronze is tempered by this water they say [also: Dsc.5.121 s.v.l.] 7

DIP

(a container, instrument, one’s finger or hand) INTO WATER SO AS TO

TAKE UP A CERTAIN QUANTITY

Thphr.Char.9.8 δεινὸς δὲ καὶ πρὸς τὰ χαλκία τὰ ἐν τῷ βαλανείῳ προσελθὼν καὶ βάψας ἀρύταιναν βοῶντος τοῦ βαλανέως αὐτὸς αὑτοῦ καταχέασθαι he is also apt to go up to the hot–water tanks in the baths and, despite the protests of the bath attendant, dip his ladle in and shower himself on his own Ev.Luc.16.24 ἐλέησόν με καὶ πέμψον Λάζαρον ἵνα βάψῃ τὸ ἄκρον τοῦ δακτύλου αὐτοῦ ὕδατος καὶ καταψύξῃ τὴν γλῶσσάν μου, ὅτι ὀδυνῶμαι ἐν τῇ ϕλογὶ ταύτῃ ⁸⁷ Some take this to refer to circumcision, see Moulton and Milligan 1930, 103, under βάπτω.

382

Anne Thompson

take pity on me and send Lazarus so that he may dip the tip of his finger in water and cool my tongue, because I am in agony in this fire [also : Emp.100.11, E.Hec.610, Antiph.25] b middle: DIP (a container) TO DRAW WATER FOR ONESELF AP9.326 (Leon.) χαίρετ’· Ἀριστοκλέης ὅδ’ ὁδοιπόρος, ᾧπερ ἀπῶσα δίψαν βαψάμενος, τοῦτο δίδωμι γέρας Hail! I Aristokles this traveller give you (the Nymphs of the springs) this gift (a cup), with which I drove away my thirst dipping it in your streams [taking this as absolute, rather than intransitive with dative, as c] c intransitive: DIP IN ORDER TO DRAW WATER (with dative: with a container), or DIP A WATER-POT Lyc.1365 ἐπεὶ Πελασγοὺς εἶδε Ῥυνδακοῦ ποτῶν κρωσσοῖσιν ὀθνείοισι βάψαντας γάνος since she saw the Pelasgians dipping into the brightness of the fresh waters of Rhyndacus with their foreign pitchers Call.H.5.45 σάμερον, ὑδροϕόροι, μὴ βάπτετε—σάμερον, Ἄργος, πίνετ’ ἀπὸ κρανᾶν μηδ’ ἀπὸ τῶ ποταμῶ today, water carriers, do not draw water,—today, Argos, drink from the fountains not from the river⁸⁸ 8 DIP (a container or an instrument) INTO A LIQUID (other than water, especially honey) SO AS TO TAKE UP A CERTAIN QUANTITY LXX1Ki.14.27 καὶ ἐξέτεινεν τὸ ἄκρον τοῦ σκήπτρου αὐτοῦ τοῦ ἐν τῇ χειρὶ αὐτοῦ καὶ ἔβαψεν αὐτὸ εἰς τὸ κηρίον τοῦ μέλιτος καὶ ἐπέστρεψεν τὴν χεῖρα αὐτοῦ εἰς τὸ στόμα αὐτοῦ and he stretched out the end of the stick in his hand and dipped it in the honeycomb and turned his hand again towards his mouth (to eat the honey) b intransitive: DIP IN ORDER TO DRAW UP A LIQUID (with dative: with a container) Theoc.5.127 καὶ τὸ πότορθρον ἁ παῖς ἀνθ’ ὕδατος τᾷ κάλπιδι κηρία βάψαι and near dawn may the girl dip with her pitcher into honeycomb instead of water 9

(something) IN A LIQUID or APPLY A LIQUID TO (something) SO AS TO or IMPART PROPERTIES TO THAT THING (especially in contexts relating to food, poison, medicine, perfume) S.Tr.580 χιτῶνα τόνδ’ ἔβαψα I imbued this garment (with the Hydra’s poison) Arist.HA605a29 καὶ τὸν χόρτον εἰς μέλι βάπτοντες διδόασιν ἐσθίειν and dipping the hay into honey they feed it (to a sick elephant) DIP

MOISTEN

⁸⁸ See further A.W. Bulloch, Callimachus: The Fifth Hymn (Cambridge, 1985), 154, on line 45.

An Illustration of the State of our Ancient Greek Dictionaries

383

Hp.Epid.5.58 θαλάσσῃ, ὄξει, θερμοῖσι καταιονῇν, καὶ σπόγγους βάπτοντα πυριῇν make up an application with seawater, vinegar, hot water, and moistening sponges apply as a hot compress Ev.Jo.13.26 ἐκεῖνός ἐστιν ᾧ ἐγὼ βάψω τὸ ψωμίον it is he for whom I shall dip the piece of bread [also : LXXRu.2.14] 10 intransitive : DRAW or SOAK UP A LIQUID (on to something absorbent) Hp.Mul.2.196 ἢν ξηρῆναι δέῃ ῥόον, καλαμίνθην ἐν οἴνῳ μέλανι ἑψεῖν, καὶ ἐς ὀθόνιον βάπτοντα, ἐπιτιθέναι if the flow needs to be dried up, boil mint in red wine, and soaking it up on to a piece of lint, apply it 11

(something) IN BLOOD IN A RITUAL SO AS TO TAKE UP A CERTAIN or COAT THAT THING LXXEx.12.22 λήμψεσθε δὲ δεσμὴν ὑσσώπου καὶ βάψαντες ἀπὸ τοῦ αἵματος τοῦ παρὰ τὴν θύραν καθίξετε τῆς ϕλιᾶς καὶ ἐπ’ ἀμϕοτέρων τῶν σταθμῶν you will take a bunch of hyssop and coating it from the blood that is by the door, you will daub it over the lintel and on both the door jambs LXXLe.4.17 καὶ βάψει ὁ ἱερεὺς τὸν δάκτυλον ἀπὸ τοῦ αἵματος τοῦ μόσχου καὶ ῥανεῖ ἑπτάκις and the priest will coat his finger from the blood of the calf and perform a ritual sprinkling seven times [also : Philostr.VA.8.7] 12 DIP (fibres, textiles, garments) IN LIQUID DYE SO AS TO DRAW UP THE DIP

QUANTITY

COLOUR IN A MANUFACTURING PROCESS

Pl.R.429d ἐπειδὰν βουληθῶσι βάψαι ἔρια ὥστ’ εἶναι ἁλουργά when they wish to dye wool so that it is the best purple Ach.Tat.2.11.4 βάπτουσιν Ἀϕροδίτης τὸν πέπλον they die the peplos of Aphrodite [also : Ar.Ec.216, unless the purpose here is washing rather than dyeing; also : Eup.363K.-A; Pl.R.429d (absolute or intransitive)] b middle: DYE ONE’S CLOTHES, WEAR COLOUR or DRESS COLOURFULLY Ar.Lys.51 κροκωτὸν ἄρα νὴ τὼ θεὼ ’γὼ βάψομαι in that case by the gods I shall get my dress dyed in saffron (or I shall dress colourfully in saffron) Ar.Pax1176 (in a figurative phrase) ἢν δέ που δέῃ μάχεσθ’ ἔχοντα τὴν ϕοινικίδα τηνικαῦτ’ αὐτὸς βέβαπται βάμμα Κυζικηνικόν though I think if he has to fight wearing the scarlet cloak that’s the time he’s dressed in the dye of Cyzicus [also : Ar.Eq.523] c passive: (of fibres, textiles, garments) BE DYED, COLOURED or BRIGHTLY COLOURED

AP.6.254(Myrin.) τἀκ κόκκου βαϕθέντα καὶ ὑσγίνοιο θέριστρα and summer garments dyed from scarlet and crimson dye [also : Pl.R.429e]

384

Anne Thompson

d perfect passive participle as adjective:

DYED, COLOURED

or

BRIGHTLY

COLOURED

Hdt.7.67.1 Σαράγγαι δὲ εἵματα μὲν βεβαμμένα ἔχοντες ἐνέπρεπον and the Saraggai stood out wearing their brightly coloured garments [also : J.BJ.4.563, Peripl.M.Rubr.8] 13 DYE (someone, their hair or person) AP.11.68 (Lucill.) τὰς τρίχας, ὦ Νίκυλλα, τινὲς βάπτειν σε λέγουσιν some say, Nikulla, that you die your hair Ar.Ach.112 (in figurative phrase) ἄγε δὴ σύ, ϕράσον ἐμοὶ σαϕῶς πρὸς τουτονί, | ἵνα μή σε βάψω βάμμα Σαρδιανικόν so come on you, tell me honestly in the face of this (a stick), so that I won’t have to dye you in Sardian crimson b middle: DYE THE HAIR καὶ βάψομαι καὶ παρατιλοῦμαι and I shall dye my hair and have my body hair plucked Men.fr.303.4 (LSJ fr.363) [also : Nicol.Com.1.33] c passive perfect participle as adjective: DYED Com.Adesp.53D (in figurative phrase; of a man) εὐρῶτι γήρως τὰς τρίχας βεβαμμένος having hair dyed with the blight of old age 14 COLOUR (a material or substance, an object) IN A PROCESS OF PREPARATION or MANUFACTURE Ach.Tat.1.4.3 πορϕύραν, εἰς οἵαν τὸν ἐλέϕαντα Λυδία βάπτει γυνή purple, like that in which a Lydian woman colours ivory Ps.-Democr.p. 44 B καὶ ὄπτησον, καὶ ἐπίβαλλε ὕλῃ ξανθὸν γενόμενον, καὶ βάψεις (v.l. βάψῃ) roast it and, when it has become yellow, apply it to the material and you will colour it [also Ps.-Democr.p. 46 B] b passive: (of materials or substances, of objects) BE COLOURED IN A PROCESS OF PREPARATION or MANUFACTURE Ath.11.480e καὶ βάπτονται εἰς τὸ δοκεῖν εἶναι ἀργυραῖ and they (kylikes) are coated so as to look as if made of silver 15 (of a weapon used in murder, of persons or animals bleeding, of blood, other liquids or substances) DISCOLOUR or STAIN (something) A.Ch.1011 ϕᾶρος τόδ’ ὡς ἔβαψεν Αἰγίσθου ξίϕος this garment that Aegisthus’ sword has stained (perhaps with allusion to 12 or 14) S.Tr.574 ἐὰν γὰρ ἀμϕίθρεπτον αἷμα τῶν ἐμῶν | σϕαγῶν ἐνέγκῃ χερσίν, ᾗ μελάγχολος | ἔβαψεν ἰὸς θρέμμα Λερναίας ὕδρας if you take up in your hands the blood clotted all round from my wound, where the black-galled poison, product from the hydra of Lerna, contaminated it

An Illustration of the State of our Ancient Greek Dictionaries

385

Ach.Tat.2.11.5 καὶ τῷ στόματι τοῦ κυνὸς περιρρέει τοῦ ἄνθους τὸ αἷμα, καὶ βάπτει τὸ αἷμα τὴν γένυν καὶ ὑϕαίνει τοῖς χείλεσι τὴν πορϕύραν and the blood (of the murex) with its colour streamed all over the dog’s mouth, and the blood stained its muzzle and on its lips created purple dye [also : Sch.Il.18.329] b passive: (of things) BE DISCOLOURED or STAINED Hp.Mul.122.4 ἐπειδὰν ἐπιστάξῃ ἐπὶ τὰ ἱμάτια, βάπτεται, καὶ δύσπλυτα ἐμμένει when it drips on to clothes, they are stained, and are permanently hard to wash clean Pl.Ti.83b τοτὲ δὲ ἡ πικρότης αὖ βαϕεῖσα αἵματι χρῶμα ἔσχεν ἐρυθρώτερον and then the bitter substance (in diseased flesh) bathed in blood came to have more of a reddish colour Batr.220 ἐβάπτετο δ’ αἵματι λίμνη | πορϕυρέῳ and the marsh was discoloured with dark blood M.Ant.8.51 κἂν πηλὸν ἐμβάλῃ, κἂν κοπρίαν, τάχιστα διασκεδάσει αὐτὰ καὶ ἐκκλύσει καὶ οὐδαμῶς βαϕήσεται and if he throw mud or dung in (into a spring), it will quickly disperse it and wash it away, and will not be sullied [also : Luc.VH.1.17] c perfect passive participle as adjective: DISCOLOURED or STAINED Apoc.19.13 καὶ περιβεβλημένος ἱμάτιον βεβαμμένον αἵματι clothed in a garment stained with blood

2 0.8. T E S T I N G TH E R E S U L TS AG A INS T NEW MATERIAL Contexts for most vocabulary words are not unlimited. With this verb, they frequently relate to water, blood, dye, oil, poison, and food. Passages about tempering metal or preparing medical and chemical recipes feature prominently. When more texts are read, the scheme above can be checked and adjusted. Analysis of the first examples is time‐consuming, but afterwards instances of the word can be assessed relatively quickly. Some additions from classical and Hellenistic literature are as follows: add to 1: AP7.636.4 πικρῇ βάψαι νήοχα πηδάλια ἅλμῃ to dip the ship-guiding rudder in the bitter brine add to 1: AP9.438.3 μηνίσας ὁ πρέσβυς ἐς ὕδατα κρωσσὸν ἔβαψεν ἐνθάδε τοὺς ἀπὸ γῆς οὐ δοκέων πελάσειν the old man in his anger set the jar (of honey) here in water not thinking animals of the land (ants) would be able to come near [not 7, as the water is not to be drawn up]

386

Anne Thompson

add to 3: Pl.Com.198.1 σαπρῷ γάρῳ βάπτοντες ἀποπνίξουσί με dipping me in rotten fish sauce they will suffocate me [and within the parentheses add ‘someone’ to ‘something’, though it is not certain who ‘me’ refers to] add to 3: Orac.Sib.7.147 οὐκέτι τις κόψει βαθὺν αὔλακα γυρῷ ἀρότρῳ· οὐ βόες ἰθυντῆρα κάτω βάψουσι σίδηρον no longer will anyone cut the deep furrow with the curved plough, nor will the oxen dip the guiding iron beneath [‘earth’ an unusual medium in poetic extension, probably on the analogy of contexts where the surface is the sea] add to 4: Hp.Epid.5.1.45 ὁ σκυτεὺς, κάσσυμα κεντῶν, ὀπητίῳ ἐκέντησεν ἑαυτὸν ἐπάνω τοῦ γούνατος ἐς τὸν μηρὸν, καὶ ἔβαψεν ὡς δάκτυλον the cobbler, sewing a leather shoe, pierced himself with a needle in his thigh above the knee, and it went into the flesh to the depth of a finger [within the parentheses ‘or sharp instrument’ needs to be added to ‘a weapon’] add to 4: AP9.240 κάπρος δ’ Ἡράκλειος ἀπορρήξας ἀπὸ δεσμῶν ἐς νηδὺν κριοῦ πᾶσαν ἔβαψε γένυν and the boar of Herakles breaking loose from his tether buried the whole length of his tusks in the ram’s belly add to 5: E.HF.929 μέλλων δὲ δαλὸν χειρὶ δεξιᾶι ϕέρειν, ἐς χέρνιβ’ ὡς βάψειεν he was about to bring a torch in his right hand to dip in the basin of holy water [the context is a sacrifice] add to 5: Sotion fr.24.12 ϕεῦγε δ’ ἐμὴν πηγὴν μισάμπελον, ἔνθα Μελάμπους ῥυσάμενος λύσσης Προίτιδας ἀργαλέης πάντα καθαρμὸν ἔβαψεν ἀπόκρυϕον leave my vine-hating streams where Melampus secretly bathed with every purification ritual the daughters of Proitos delivering them from grievous madness add to 6: D.S.4.36.5 ἡ δὲ κατὰ τὴν γενομένην ὑπὸ τοῦ Νέσσου παραγγελίαν εἰς ἄγγος ἀναλαβοῦσα τὸν γόνον, καὶ τὴν ἀκίδα βάψασα, λάθρᾳ τοῦ Ἡρακλέους ἐϕύλαττεν and she put the semen, as Nessus had told her, into a jar and dipped the barb of the arrow in it and kept it unknown to Heracles [the idea is to add blood from the spear to the jar, not to draw up the liquid; the dipping has the intention of imparting properties to the liquid] add to 7 c intransitive: Arr.Peripl.M.Eux.8.3.2 καὶ ἦν κατὰ μὲν τοῦ ἐπιρρέοντος βάψαντα γλυκὺ τὸ ὕδωρ ἀνιμήσασθαι, εἰ δὲ εἰς βάθος τις καθῆκεν τὴν κάλπιν, ἁλμυρόν and if you dip a water-pot in the surface-flowing water, fresh water is drawn, but if someone lets the pot down into the depths, salty water add to 9: Cypr.5.2 (West) εἵματα μὲν χροῒ ἕστο, τά οἱ Χάριτές τε καὶ Ὧραι | ποίησαν καὶ ἔβαψαν ἐν ἄνθεσιν εἰαρινοῖσιν she was wearing garments on her body, which the Graces and the Seasons made and dipped in spring flowers [A list of the flowers follows. The last line is about perfume: ὧδ’ Ἀϕροδίτη ὥραις παντοίαις τεθυωμένα εἵματα ἕστο thus Aphrodite was dressed in garments scented with all the seasons. This may be about dyeing, or imbuing with scent, or a play on both; add to 12 if about dyes]

An Illustration of the State of our Ancient Greek Dictionaries

387

add to 9: Antid.fr.3 (in Ath.3.109c) λαβόντα θερμοὺς ἐσχαρίτας, πῶς γὰρ οὔ; τούτους ἀνειλίττοντα βάπτειν εἰς γλυκύν taking hot brazier-cakes, why not? rolling them dip them in syrup add to 9: Hp.Mul.126 βάψας ἢ λευκῷ αἰγυπτίῳ ἢ μυρσίνῳ ἢ βακχαρίῳ ἢ ἀμαρακίνῳ having moistened it with white Egyptian ointment, myrtle, hazelwort or marjoram ointment [with dative] add to 10: b middle DIP BREAD: Hp.Morb.2.48 ζωμὸν δὲ μὴ ῥοϕεέτω, μηδὲ βάπτεσθαι let him not drink broth, nor dip into it with bread [but uncertain; ἐμβάπτεσθαι cj.] add to 11: X.An.2.2.9 ταῦτα δ’ ὤμοσαν, σϕάξαντες ταῦρον καὶ κάπρον καὶ κριὸν εἰς ἀσπίδα, οἱ μὲν Ἕλληνες βάπτοντες ξίϕος, οἱ δὲ βάρβαροι λόγχην they made an oath to this effect, slaughtering a bull, boar and ram over a shield, the Greeks dipping a sword in the blood, the barbarians a spear add to 13: AP9.214 τῇ τῶν λόγων σου κογχύλῃ, Πορϕύριε, βάπτεις τὰ χείλη you dye your lips with the purple of your discourse, Porfyrios [in figurative phrase] add to 13 b middle: Arr.Ind.16.4.2 τοὺς δὲ πώγωνας λέγει Νέαρχος ὅτι βάπτονται Ἰνδοί Nearkhos says that Indians dye their beards [with internal accusative: ‘they apply dye on their beards’] add to 13 b middle: Thphr.HP3.13.6 καὶ τὰς χεῖρας τελειούμενοι βάπτονται καὶ τὰς κεϕαλάς they dye themselves on their hands and heads (in elderberry) when being initiated into the mysteries [again, with internal accusative] add new section X (between 14 and 15) (abstract) ADD COLOUR TO, ENHANCE or (pejorative) STAIN, TAINT (someone, their attributes or circumstances; with allusion to 12 or 14) AP11.408.1 τὴν κεϕαλὴν βάπτεις, τὸ δὲ γῆρας οὔποτε βάψεις you colour your hair, but you will never colour your old age [second clause is figurative; add first clause to 13] add to X: AP11.423.1 βάπτων πάντα, βαϕεῦ, καὶ χρωματίοις μεταβάλλων, καὶ πενίην βάψας πλούσιος ἐξεϕάνης dyer you dye all things changing them with your pigments, and you also dyed your poverty and made yourself stand out as a rich man [in figurative phrase] add to X: pejorative: Alcibiades elegiac fragment 1 (West)⁸⁹ †βάπτε με σὺ | ἐν† θυμέληισιν· ἐγὼ δὲ σὲ κύμασι πόντου βαπτίζων ὀλέσω νάμασι πικροτάτοις you dunk me amongst the altars, but I will plunge you in the waves of the sea and destroy you in its cruel waters [uncertain; apparently refers to plunging someone into criticism]⁹⁰ ⁸⁹ M.L. West, Iambi et Elegi Graeci (editio altera), vol. II (Oxford, 1971), 29. ⁹⁰ See also Storey 2003, 97 for the suggestion you dye me, with a figurative and pejorative sense. For the idea that Eupolis had portrayed Alcibiades as ritually dipped in an initiation, and

388

Anne Thompson

add to X: pejorative: AP7.71 σῆμα τόδ’ Ἀρχιλόχου παραπόντιον, ὅς ποτε πικρὴν | μοῦσαν ἐχιδναίῳ πρῶτος ἔβαψε χόλῳ | αἱμάξας Ἑλικῶνα τὸν ἥμερον this tomb by the sea belongs to Archilochus, who in days gone by was first to taint the muse making her bitter with vipers’ gall, bloodying gentle Helikon add to X b middle: pejorative: M.Ant.6.30.1 ὅρα μὴ ἀποκαισαρωθῇς, μὴ βαϕῇς take care not to turn into a Caesar, so you do not taint yourself add to X b passive: M.Ant.5.16.1 οἷα ἂν πολλάκις ϕαντασθῇς, τοιαύτη σοι ἔσται ἡ διάνοια· βάπτεται γὰρ ὑπὸ τῶν ϕαντασιῶν ἡ ψυχή. βάπτε οὖν αὐτὴν τῇ συνεχείᾳ τῶν τοιούτων ϕαντασιῶν as are your habitual thoughts, so will be the character of your mind; for the soul is imbued with your thoughts. Dye it then with a continuous series of such thoughts as these [second instance belongs with active] add to 15: Arist.HA.547a18 θλιβόμενος δὲ βάπτει καὶ ἀνθίζει τὴν χεῖρα and when it (a murex) is squeezed it emits a coloured liquid and colours the hand add to 15 b passive: Arist.Col.791a7 διὰ τὸ τῷ καπνῷ βεβάϕθαι μέλανι ὄντι because they (ashes) are coloured by the smoke which is black add to 15 c perfect passive participle as adjective: Arr.Peripl.M.Eux.8.5.2 ἡ δὲ χρόα τῷ Φάσιδι οἵα ἀπὸ μολίβδου ἢ καττιτέρου βεβαμμένου τοῦ ὕδατος there is a colour to the Phasis that is like water discoloured by lead or tin A concise outline can be extracted from this survey.⁹¹ As it is based on a fresh analysis of source material, it will not be the same as a short dictionary article such as in IGL, because that is based on a long tradition of entries for the large lexicon. 1 dip or plunge (someone or something) into water, a liquid or liquid-like substance; sometimes, in a process of manufacture, or as a ritual; (also intransitive) 2 dip (a container, instrument or material) so as to draw up water or a liquid; (also intransitive) dip a water-pot, draw up water or liquid 3 dip in dye or apply colour to, dye, colour (textiles, artefacts or materials)

that this was the motive for revenge, see R.C.T. Parker, Miasma: Pollution and Purification in Early Greek Religion (2nd edn, Oxford, 1996), 306, n. 125. ⁹¹ The above new citations are not exhaustive. The work on DGE has given rise to a bibliography on Greek vocabulary: P. Boned Colera, J. Rodríguez Somolinos, et al. 1998, now with Suplemento online. Such bibliographies are useful, but the ideal would be not just reference to titles of works, but a summary of the points contained in them. Work on the Montanari lexicons also gave rise to a database: Aristarchus www.aristarchus.unige.net (formerly PAWAG: Poorly Attested Words in Ancient Greek). Invaluable resources are the indices of Supplementum Epigraphicum Graecum [SEG], and the ongoing Bulletin épigraphique and Bulletin papyrologique in Revue des Études Grecques.

An Illustration of the State of our Ancient Greek Dictionaries

389

4 dye (hair or a part of the body); (middle) dye or add colour to oneself (one’s hair, body or clothes) 5 (abstract sense) colour or enhance, or (pejorative) stain or taint (someone or something) 6 (of a liquid or liquid-like substance, or things discharging these) colour, discolour or stain (something) It could be added that the kind of liquid or liquid-like substance may be specified in the context, also that the verb is frequently found with an internal accusative or prepositional phrase giving this information: (dip) into a liquid or source of liquid, (especially) into water, a dye solution, medicament or potion; also, into blood, flesh or a part of the body. There are also examples with a dative or partitive genitive. An ideal would be for every citation from TLG and documentary sources to be linked to the sense categories, as far as these can be safely identified, along with lexicographic definitions. It would then be possible to say which senses occur in each author or text, freeing us from the somewhat unrepresentative overview of lexicons of the past. The methodology above is traditional, similar to that of OED and OLD. But there can be experimentation with other methods and styles of presentation.⁹² The most complicated task is one that for the most part has not been tackled at all, an account of the lexical relations of the whole vocabulary.⁹³ βάπτω belongs to a word family, with related nouns, adjectives, and compound verbs.⁹⁴ The lexicographer needs to work on all these together, not separately as they appear in alphabetic order. When derivatives from the same root are examined, for example, βαπτός, βαϕή, βαπτίζω, senses related to those above will recur. Where they do, they should be defined with similar wording so that the matches are clear. If there are divergences, these will stand out. It is very difficult from a comparison of the LSJ entries for βάπτω and βαπτίζω to say under what circumstances one is used rather than the other. An analysis of the attestations of βαπτίζω along the lines of what has been done above for βάπτω is likely to provide some answers.⁹⁵ In the case of synonyms, such as τέγγω, the precise similarities and differences will also become apparent. ⁹² See for instance the approaches to semantic analysis of Clarke 2010, also Peláez 1996. ⁹³ See Lee 2003, 186 on the challenge of finding a radically different approach to the lexicon with a systematic presentation of the lexical structure or sense relations of the whole vocabulary. ⁹⁴ See Chantraine, DELG and other etymological dictionaries for such groupings, though the study of word formation and word families is not the same as etymology. See also Chadwick 1996, 21 2. ⁹⁵ An analysis for βαπτίζω will be along these lines: ‘1 plunge or immerse (oneself or another person, a part of the body, a thing) in or under water (or other liquid); (intransitive) be in water 2 sink (a ship) 3 (in a ritual) immerse in water (oneself or another, a part of the body, or a thing; occasionally, in blood); (middle) immerse oneself, perform ablutions 4 (Christian usage) baptize (a person); (middle) get oneself baptized 5 plunge (a weapon or instrument, into a part of the

390

Anne Thompson

There are also derivatives with a prepositional prefix within this word family. Very frequently such words are not adequately defined.⁹⁶ The reliance on translation glosses rather than lexicographic definitions contributes to this failure. As an illustration, the compound verb ἀποβάπτω may be set beside the simple verb. For the compound, LSJ offers dip, plunge, and draw (water), as though the presence of the preverb made no difference. DGE has sumergir, mojar which appear also under βάπτω, and a new gloss bañar, which does not in itself indicate why the compound verb might be different. Montanari begins with the same translation glosses as for the simple verb, immergere, tuffare/to immerse, plunge. Only IGL signals a difference by defining as to dip quite or entirely, taken from LS⁷, but quite and entirely were subsequently omitted in the LS⁸ and LSJ. The only help in LSJ is under the entry for the preposition ἀπό, section (D) devoted to the functions as a prefix (IN COMPOS.): 1. asunder, as ἀποκόπτω, ἀπολύω, ἀποτέμνω: and hence, away, off, as ἀποβάλλω, ἀποβαίνω; denoting, removal of an accusation, as ἀπολογέομαι, ἀποψηϕίζομαι. 2. finishing off, completing, ἀπεργάζομαι, ἀπανδρόω, ἀπανθρωπίζω, ἀπογλαυκόω. 3. ceasing from, leaving off, as ἀπαλγέω, ἀποκηδεύω, ἀπολοϕύρομαι, ἀποζέω, ἀπανθίζω, ἀϕυβρίζω. 4. back again, as ἀποδίδωμι, ἀπολαμβάνω, ἀπόπλους: also, in full, or what is one’s own, as ἀπέχω, ἀπολαμβάνω: freq. it only strengthens the sense of the simple. 5. by way of abuse, as in ἀποκαλέω. 6. almost = ἀ- priv.; sts. with Verbs, as ἀπαυδάω, ἀπαγορεύω; more freq. with Adjectives, as ἀποχρήματος, ἀπότιμος, ἀπόσιτος, ἀπόϕονος. IGL seems to be taking ἀποβάπτω as coming under 2., but one might think of 3., or even 4. as a strengthening element. The possibility of the preverb as marking a source, a usual sense of the preposition, comes (not very clearly) under away, off, removal in 1. There will not be enough evidence to determine everything with precision until a number of similar compounds are looked at.⁹⁷ body, one’s own or another’s) 6 (of a quantity of water) overwhelm, flood (a person, a ship, land) 7 (of wine, of a person serving wine) cause (a person) to be inebriated 8 (of powerful persons or conditions) overwhelm (persons, their mind or soul, a city).’ The verb is rare before the fourth century BCE. Although there are crossover points, there is little sense of a temporary action, as with ‘dip’ for βάπτω, nor of taking up any liquid; more usual is an idea of thorough immersion or a resulting permanent state of submergence. See also James (Chapter 10, this volume), pp. 163 4, with n. 84, along with Stray’s remarks (Chapter 1, this volume), p. 14, n. 29, about the history of the LSJ entry. For an analysis of βαπτίζω with a different approach, see Peláez 2014, 266 71. ⁹⁶ On this problem, see Glare 1987, 14 15. ⁹⁷ On understanding the preverb μετα from a general study, see Glare 1991, 38 9.

An Illustration of the State of our Ancient Greek Dictionaries

391

The force is likely to be mostly in some way equivalent to the sense of the preposition, but divergence cannot be ruled out.⁹⁸ In the following, there is a relationship with 7 c in the analysis for βάπτω draw water: LXX2Ma.1.20 ὡς δὲ διεσάϕησαν ἡμῖν μὴ εὑρηκέναι πῦρ, ἀλλὰ ὕδωρ παχύ, ἐκέλευσεν αὐτοὺς ἀποβάψαντας ϕέρειν when they reported to us that they had not found fire, but salty water, he ordered them to draw water from there and bring it

The function of the preverb seems to be to emphasize ‘from the source’, also perhaps with some idea of moving the water from that place to the present one. Wording for a definition should differ somewhat from the simple verb: perhaps ‘draw or draw off water from a source by dipping’. There are several instances of ἀποβάπτω relating to βάπτω 9 above: ‘dip (something) in a liquid so as to moisten that thing’. With the preverb, the second part of the definition is brought into prominence: the liquid as a source, instrument or means. For the definition, more emphasis needs to be placed on the liquid than the dipping action: ‘moisten or wet (something) by dipping in a liquid’. There may also be a sense of change, from the anterior state to a new one. As the liquid is an instrument or means, there will necessarily be a sense of the resulting state after the dipping action. Hp.Int.9.23 τρωγέτω δὲ καὶ τῆς ὀριγάνου τῆς ἁπαλῆς ὡς πλεῖστον, ἐς μέλι ἀποβάπτων and let him eat some tender oregano as much as he likes, after soaking it in honey Arist.Mir.854a1 ϕασὶ τὸ Σκυθικὸν ϕάρμακον, ᾧ ἀποβάπτουσι τοὺς ὀϊστούς, συντίθεσθαι ἐξ ἐχίδνης they say that the Scythian poison, in which they coat arrows, is procured from the viper Ar.fr.416 [426 K.-A.] ὦ κακοδαίμων, ὅστις ἐν ἅλμῃ πρῶτον τριχίδων ἀπεβάϕθη unlucky the anchovy who first got doused in brine (passive) Alexis 62.3 ἀλλὰ τέτταρας, | περιστερὰς ἀϕῆκεν ἀποβεβαμμένας, | εἰς οὐχὶ ταὐτὸν μὰ Δία τὴν αὐτὴν μύρον, | ἰδίῳ δ’ ἑκάστην but he sent him four doves that are perfused, by god, not in the same perfume, but in a different one for each (perfect passive participle as adjective) In the following passage, the purpose of dipping in liquid is not to take up but to infuse that liquid with properties from the object dipped, relating to

⁹⁸ Bortone (2010, 120) emphasizes this. The forms can diverge, as the Latin preposition cum is con as a prefix, and in Modern Greek γιά is the preposition, whereas δια remains as the prefix, though with more limited potential for creating new compounds than in the ancient language. It is to be expected that the senses will diverge as well.

392

Anne Thompson

βάπτω 6 above. This, along with D.S. 4.36.5 (above, p. 386), prompts the addition of a subsection, with a definition ‘dip (something) in a liquid to impart properties of the thing dipped’. With the preverb, the focus again shifts from the dipping action to the liquid, and the subsequent removal: Arist.HA607a25 γίνεται δὲ καὶ ἐν τῷ σιλϕίῳ τι ὀϕείδιον, οὗ καὶ λέγεται ἄκος εἶναι λίθος τις, ὃν λαμβάνουσιν ἀπὸ τάϕου βασιλέως τῶν ἀρχαίων καὶ ἐν οἴνῳ ἀποβάψαντες πίνουσιν amongst silphium plants there is a certain snake, the antidote for which, it is said, is a certain stone which they take from the tomb of a king of the ancients, and after immersing it in wine they drink the wine

The following are related to section 5, ‘dip, plunge or immerse (something) in water in a ritual’. The first may be somewhat more of a medical cure than a ritual, but it is more than just ‘plunge in water’. The purpose of the dipping is to profit from qualities that the water contains, the preverb emphasizing the liquid as their source. dip or plunge may be the translation gloss, but the definition might be along the lines of ‘treat in water in a procedure or ritual (someone) by dipping or plunging (them)’. Arist.Mir.832b6 περὶ Κιλικίαν δέ ϕασιν εἶναι ὕδατος συστρεμμάτιον, εἰς ὃ τὰ πεπνιγμένα τῶν ὀρνέων καὶ τῶν λοιπῶν ζῴων ὅταν ἀποβαϕῇ, πάλιν ἀναβιοῖ and they say in Cilicia there is an eddy of water, in which birds and other animals who have been suffocated come to life again when they are dipped in it (passive) Arist.Pol.1336a16 συμϕέρει δ’ εὐθὺς καὶ πρὸς τὰ ψύχη συνεθίζειν ἐκ μικρῶν παίδων τοῦτο γὰρ καὶ πρὸς ὑγίειαν καὶ πρὸς πολεμικὰς πράξεις εὐχρηστότατον. διὸ παρὰ πολλοῖς ἐστι τῶν βαρβάρων ἔθος τοῖς μὲν εἰς ποταμὸν ἀποβάπτειν τὰ γιγνόμενα ψυχρόν, τοῖς δὲ σκέπασμα μικρὸν ἀμπίσχειν, οἷον Κελτοῖς and it is also advantageous to accustom them at once from early childhood to cold, for this is most useful both for health and with a view to military service. Hence there is a custom among many of the barbarian peoples of plunging newborn children into a cold river, and among others, like the Celts, of clothing them with little protection

To these may be added an example with tmesis from Herodotus, where the water is a source of qualities that ritually wash away pollution: Hdt.2.47.1 (with tmesis) ἤν τις ψαύσῃ αὐτῶν παριὼν ὑός, αὐτοῖσι τοῖσι ἱματίοισι ἀπ’ ὦν ἔβαψε ἑωυτὸν βὰς ἐς τὸν ποταμόν if any one of them when passing brushes against a pig, going to the river he immerses himself in it along with his clothes

Another example in Herodotus has a ritual dipping of weapons in blood and wine in an oath ceremony, relating to βάπτω 11. The weapons will be coated but there may also be a focus, as in Arist.HA607a25 above, on the qualities the weapons will have bestowed on the liquid to be drunk, implying the threat of violence in defence of the oath: Hdt.4.70.1 ὅρκια δὲ ποιεῦνται Σκύθαι ὧδε πρὸς τοὺς ἂν ποιέωνται. ἐς κύλικα μεγάλην κεραμίνην οἶνον ἐγχέαντες αἷμα συμμίσγουσι τῶν τὸ ὅρκιον ταμνομένων, τύψαντες ὑπέατι ἢ ἐπιταμόντες μαχαίρῃ σμικρὸν τοῦ σώματος καὶ ἔπειτα

An Illustration of the State of our Ancient Greek Dictionaries

393

ἀποβάψαντες ἐς τὴν κύλικα ἀκινάκην καὶ ὀϊστοὺς καὶ σάγαριν καὶ ἀκόντιον ἐπεὰν δὲ ταῦτα ποιήσωσι, κατεύχονται πολλὰ καὶ ἔπειτα ἀποπίνουσι the Scythians conduct oaths for those taking part in this way: pouring wine into a big earthen ware bowl they mix it together with blood from those performing the oath ceremony, pricking with a large needle or making a small cut on the body’s skin with a knife, and then ritually immerse an oriental sword, arrows, an oriental battleaxe and a javelin into the bowl; when they do this, they utter many solemn promises over it and then drink from it

The wording of definitions for similar instances should be consistent. These are just a few preliminary observations on such compounds, and much remains to be done.

20.9. CONCLUSIONS An ancient Greek lexicon along these lines would require a huge amount of work, but perhaps no more than has been invested in other linguistic studies, phonology, morphology, etymology, and syntax. Vocabulary has been very neglected by comparison, and there is no scientific foundation on which to build reference works. It is clear that producing new dictionaries by revising old entries and inserting extra material can produce a useful reference tool which is modernized in some respects, but it is far from a wholly successful and accurate process. The level of detail and checking required is always underestimated. Study of this one word shows how dictionaries do not very accurately reflect what is there to be known. The analysis of senses for longer entries is rarely completely correct, and translation glosses in old-fashioned English can mislead. The most comprehensive work is the ongoing DGE, and it can be judged to be an excellent replacement for LSJ. Any new lexicographic work in English would require different methodologies.⁹⁹ Whatever the style decided upon, there needs to be a completely new start and, as happened with OLD, the source texts need to be read again. The whole corpus of literary and documentary texts needs to be drawn upon, and material from later Greek should also be there for reference. In reconstructing the history of a word any one piece of evidence may be relevant.¹⁰⁰ There is, however, no need initially for exhaustive collections of source material and bibliography. A relatively small sample of texts can be used to set up reasonably accurate preliminary models that can ⁹⁹ See Chadwick 1994 and Lee 1997. Reasons for revision of a dictionary remain similar to those in the nineteenth century. Our own language is changing, as well as our knowledge of Greek; see Imholtz 2007. ¹⁰⁰ On Modern Greek in LSJ and LSJ Revd Suppl., see Thompson 2015.

394

Anne Thompson

be added to, corrected, and modified relatively quickly once the first time-consuming analysis has been done.¹⁰¹ The setting of words in their wider linguistic and real world contexts will sharpen our understanding of ancient texts and provide an abundance of new insights into ancient society. As Hugh Lloyd-Jones wrote in a review of LSJ Revd Suppl. and Chadwick’s Lexicographica Graeca in the Times Literary Supplement (1 November 1996, 9): ‘All this shows that what is really needed is a permanent body of wellqualified scholars engaged not only in scrutinizing new material but in a thorough revision of the entire body of the Lexicon. Modern technology would very considerably improve the prospects of such a plan. To put it into effect would require a considerable, but not vast sum of money. It is not hard to think of projects a good deal less deserving on which far greater sums are spent. To finance such an operation would be to make a priceless contribution to humanistic scholarship.’

¹⁰¹ Chadwick 1996, 28: ‘A continuously progressive lexicon should be created, probably at one location with online facilities for consultation at a distance . . . ’ Chadwick founded a project for a revision of IGL as a first stage. After his death, the direction changed and this is now the Cambridge Greek Lexicon [CGL], an expanded new work, with a re reading of many of the source texts, but with considerable departure from the methodology he envisaged.

21 Liddell and Scott and the Oxford English Dictionary John Considine

The OED depends for its lexicography entirely on Liddell and Scott’s Greek English Lexicon, first published in 1843 and itself merely an English version of a great work by the German classicist Franz Passow, who had clearly formulated the principles of a historical dictionary as early as 1812 in a seventy page essay called Über Zweck, Anlage und Ergänzung Griechischer Wörterbücher. (Aarsleff 1962, 419)

The claim made here by the historian of the language sciences Hans Aarsleff is supported by contemporary evidence. For instance, Herbert Coleridge, the first editor of the English dictionary of the Philological Society of London, which was to be published as A New English Dictionary on Historical Principles (NED) and then came to be known from the 1890s onwards as the Oxford English Dictionary (OED), wrote in 1860 that ‘the theory of lexicography we profess is that which Passow was the first to enunciate clearly and put in practice successfully—viz., “that every word should be made to tell its own story”’ (Coleridge 1860, 72).¹ This ‘theory of lexicography’, the ‘historical principles’ of the full title of the New English Dictionary, was indeed to be fundamental in the making of the NED/OED, as it has continued to be fundamental in its revision to this day.² The story of the relationship between Liddell and Scott and the NED/OED falls into three parts. The first concerns the place of Liddell and Scott in the thought of the founders of the NED/OED project in the late 1850s and early 1860s. The second concerns the Lexicon’s place in the thought and practice of the founding editor of the printed ¹ For Coleridge and his context, see Gilliver 2016, 11 40. ² The historical principles of lexicography and their origins are treated in Considine 2014a, b and 2016, on which the first paragraphs of this chapter draw. John Considine, Liddell and Scott and the Oxford English Dictionary In: Liddell and Scott: The History, Methodology, and Languages of the World’s Leading Lexicon of Ancient Greek. Edited by: Christopher Stray, Michael Clarke, and Joshua T. Katz, Oxford University Press (2019). © Oxford University Press. DOI: 10.1093/oso/9780198810803.003.0021

396

John Considine

dictionary, James Murray, from the late 1870s onwards. The third concerns the relationship between editions of Liddell and Scott and of the OED in the twentieth and twenty-first centuries. *

* *

Let us begin with Herbert Coleridge’s reference to Franz Passow. Coleridge and his associates in the Philological Society could have learned about the historical principle as conceived and practiced by Passow in any or all of three complementary ways. First, they could have encountered his writings directly. The early programmatic pamphlet of 1812 mentioned by Aarsleff does not appear to have circulated widely in England, and it only begins to enunciate the vital point that a dictionary entry tells a chronologically ordered story.³ However, the preface of the second edition of Passow’s Handwörterbuch der griechischen Sprache is much clearer: The dictionary should . . . set out, as it were, the life story [Lebensgeschichte] of each single word in a conveniently ordered overview: it should state where and when each one was (as far as we know, of course) first hit upon, in which directions it developed . . . and finally, at what period it disappeared from use.⁴

Passow’s Handwörterbuch was readily available in England, and Coleridge’s German was good enough to give him access to its preface and its main text, while his colleague Richard Chenevix Trench was also reading it in the 1850s and indeed thereafter.⁵ Trench’s acquaintance with the Handwörterbuch makes it highly likely that when he wrote in his pamphlet On Some Deficiencies in our English Dictionaries—the manifesto of 1857 to the second edition of which Coleridge’s words of 1860 were appended—that a word’s birth and death should be documented in a dictionary just as those of a person would be in a biography (Trench 1857, 33), he was taking up Passow’s image of the Lebensgeschichte of a word.⁶ The Handwörterbuch was appreciatively discussed in two unsigned articles in the Quarterly Review, of 1834 and 1845, by J.R. Fishlake, the translator from ³ The only treatment of chronological ordering in Passow 1812 is at 32, ‘nicht der erste, der beste; sondern der älteste als erste Auctorität für das Wort, das zur Sprache kommt, angeführt werden muss’ (‘the first authority adduced for a word which comes into the language should not be the first in quality, the best, but rather the earliest’). It is referred to briefly in LS¹, v. ⁴ Passow (2nd ed.) xvi xvii, ‘Das Wörterbuch soll . . . gleichsam die Lebensgeschichte jedes einzelnen Wortes in bequem geordneter Ueberschaulichkeit entwerfen: es soll Auskunft geben, wo und wann ein jedes (natürlich immer soviel wir wissen) zuerst gefunden werde, in welchen Richtungen es sich fortbildete . . . endlich um welche Zeit es etwa aus dem Gebrauch versch winde.’ ⁵ For Coleridge’s German, see his ability to read the preface to the Grimms’ Deutsches Wörterbuch (Coleridge 1860, 74); for Trench and Passow, see Trench 1854, 65 and the later comparison of editions of Passow in Trench 1876, xix. ⁶ For the image, with references to its appearance in Trench 1853, see Mugglestone 2013, 14 15.

Liddell and Scott and the Oxford English Dictionary

397

German of Philipp Buttmann’s Lexilogus.⁷ These articles represent the second route by which Passow’s ideas could have reached members of the Philological Society. Fishlake gave special attention to the structure of Passow’s entries, noting in his first article that ‘His leading principle was to draw out, wherever it was possible, a kind of biographical history of each word’ ([Fishlake] 1834, 150), and referring in his second to ‘the great leading principle which Passow adopted and exemplified in his Lexicon, and which we may call “the historical principle”’ ([Fishlake] 1845, 306). This latter passage appears to be the place in which the English phrase ‘historical principle’ was first applied to lexicography; it was taken up by the members of the Philological Society in 1859, when they wrote of their projected dictionary that ‘In the treatment of words the historical principle will be uniformly adopted’ (Philological Society 1859, 4). Fishlake referred to Passow’s ‘historical principle’ in order to say that this principle had been the basis of the first edition of Liddell and Scott’s Lexicon, which was the third route by which Passow’s ideas could have travelled to the Philological Society. The Lexicon was of course known to all the members of the society by the end of the 1850s, and Coleridge’s phrase ‘that every word should be made to tell its own story’, which he encloses in quotation marks, rather than being his translation from a work by Passow, appears to be his paraphrase of a passage in Liddell and Scott’s preface to their first edition, ‘In most cases the word will tell its own story’ (LS¹, viii). So, the fundamental principles which underlie the Lexicon and the OED are those of Passow, and the founders of the OED are not only likely to have known Passow’s work directly, but can also be shown by verbal echoes in their programmatic writings to have known his principles by way of their description by Fishlake and their description and use by Liddell and Scott. Liddell and Scott’s Lexicon was a model of a second fundamental principle in the planning of the dictionary of the Philological Society: that of comprehensiveness. When Richard Chenevix Trench considered the possibility of making a new dictionary of English which would have a normative function like the Dictionnaire de l’Académie Française, and would therefore exclude words which did not belong to the preferred standard vocabulary of the language, he rejected this possibility sternly: What sort of completeness, or what value, would a Greek lexicon possess, a Scott and Liddell, from whose pages all the words condemned by Phrynichus and the other Greek purists, and, so far as style is concerned, many of them justly condemned, had been dismissed? (Trench 1857, 5) ⁷ Fishlake’s authorship is confirmed from the records of John Murray, the publisher of the Quarterly Review, in Brightfield 1944, 498, 507. Liddell and Scott knew that the article of 1834 was his: LS¹, v note c.

398

John Considine

Liddell and Scott had themselves drawn attention to their documentation of ‘the unusual words introduced by the learned Epic school of [Alexandria], Callimachus, Apollonius, etc., or by that wholesale coiner Lycophron’ (LS¹, vii). Trench was aware that the corpus of English literature was vast, and that anything like a comprehensive collection of its vocabulary would have to be gathered by a collaborative effort, which he visualized by means of Herodotus’ account of the Persian tactic of bringing soldiers together in a human chain to sweep through a conquered territory. ‘This σαγηνεύειν, this drawing as with a sweep-net over the whole extent of English literature, is that which we would fain see’, he wrote in the final paragraph of On Some Deficiencies; thus ‘we can hope that the innumerable words which have escaped us hitherto will ever be brought within our net, that an English Dictionary will prove that allembracing πάναγρον which, indeed, it should be’ (Trench 1857, 57). His use of the word σαγηνεύειν is unsurprising, because it is in the passage of Herodotus to which he refers, but his use of πάναγρον has a strong lexicographical flavour. It is a rare word: Liddell and Scott cited its sense ‘fishing or hunting net’ only from Oppian. They noted there that it is formed from πάναγρος, which they defined as ‘catching or grasping all’, hence Trench’s ‘all-embracing’. There is a reason why Trench might have looked at Liddell and Scott’s entry for πάναγρον as he thought about dictionaries: the previous entry in the Lexicon was πανάγριος ‘quite wild or rude’, also cited only from Oppian, and the range of senses of ἄγριος is discussed at some length by J.R. Fishlake, just before his account of Liddell and Scott ([Fishlake] 1845, 302).⁸ There, Fishlake asks whether ‘troublesome’ is a strong enough equivalent for ἄγριος in Iliad 19:30. Had Trench read Fishlake’s article, reflected that the semantic range of ἄγριος might be illuminated by consulting an entry for πανάγριος, turned to the entry in Liddell and Scott, and then read the adjacent entries πάναγρον and πάναγρος and stored them in his mind? Without forcing the point, it must certainly be said that as Trench concluded his manifesto for an English dictionary, his mind was running on some out-of-the-ordinary Greek. Thinking about ‘Scott and Liddell’ as a model for the desirable comprehensiveness of an English dictionary did lead to conceptual problems of which Trench was not, perhaps, fully aware. He was ready to admit all English words, ‘let their claim to belong to our book-language be the humblest’ (1857, 6), and all the Greek in the Lexicon was book-language in that sense. So, for instance, all obsolete literary words were to be welcomed, and Trench did not need to draw the obvious analogy with the place in a Greek dictionary of words attested only from Homer and Hesiod (8–12). But ‘provincial or local words’ were to be excluded (12), as were words by ‘the rabble of scribblers who hang on the skirts of literature’ (23), and as were ‘purely technical words’ ⁸ Later editions of Liddell and Scott introduced the entry παναγρίς between πανάγριος and πάναγρον.

Liddell and Scott and the Oxford English Dictionary

399

(46). As far as a lexicographer of Greek was concerned, the literary dialects were obviously to be included, as were scientific and technological texts, and ephemeral texts had by definition been lost. Working out in practice where to draw the line in each of these cases in the lexicography of English was to be a very difficult task for the successors of Trench and Coleridge.⁹ Herbert Coleridge died prematurely, but he was able to draft a few specimen dictionary entries, which were published posthumously by his successor, F.J. Furnivall.¹⁰ Each of these begins with a headword in capitals, and a part of speech, for instance ‘ABACK, adv.’. An etymology follows, in square brackets: aback is derived from Old English and Old Norse forms. Then numbered senses are given: sense 1 of aback is ‘(denoting motion) Backwards, towards that which lies behind’ and sense 2 is ‘(denoting rest) On or in that which is behind’. These senses are followed by the phrases and glosses take aback ‘to surprise by something novel and unexpected’ and put aback ‘to rebuke, check suddenly’, the pair being followed by the observation that ‘The leading idea is that of a forward motion in a given course being suddenly arrested.’ Each sense is followed by numbers which correspond to those of a separate section of numbered quotations, each of which is given a date and a bibliographical reference, arranged in a single chronological sequence. The two ways in which this entry structure differs from that of, for instance, Johnson’s Dictionary of 1755 both suggest awareness of developments in Greek lexicography. Firstly, the provision of what Ladislav Zgusta called a ‘semantic bridge’ (see Zgusta 2006, 31–8, 73–4) between take aback and put aback is strongly Passowian, and is to be found in LS¹ (1843): so, for instance, the first sense of πάλιν is given in the Lexicon as ‘back, backwards’, and then the second sense begins ‘Connected herewith is 2. the notion of opposition’, and similarly, once παλίντονος has been defined as being used ‘1. of the strung or bent bow . . . 2. of the unstrung bow’, an explanation follows, beginning ‘all the passages may be reduced to one sign[i]f[ication], bending backwards and forwards’. Secondly, the chronological ordering of the quotations in a single sequence is one way of observing the historical principle. It does not derive from Liddell and Scott, who, like Passow, place each quotation after the sense which it illustrates. The system of numbering which tied definitions to quotations appears to build on an observation by Fishlake (1834, 161–2), who had criticized the first edition of James Donnegan’s New Greek and English Lexicon (1826) for its policy of placing all the quotations together in an undivided block, with translations to distinguish their senses, and had recommended a system like that of Robert Ainsworth’s Latin dictionary, in which the block of

⁹ See Mugglestone 2005, esp. 70 109, and Gilliver 2016, ad indicem, s.v. ‘inclusion, limits of ’. ¹⁰ Furnivall 1862 [125 6], reproduced in part in Gilliver 2016, 39.

400

John Considine

quotations was subdivided by numbers corresponding to the numbers of the definitions which preceded the quotations.¹¹ A significant innovation in Coleridge’s draft entries was the consistent provision of dates for the quotations. This had not been done in earlier English and Scots dictionaries such as Johnson’s of 1755, Jamieson’s of 1808, and Richardson’s of 1836. Nor, of course, was it done in Passow’s Handwörterbuch or Liddell and Scott: the former relied on its readers’ ability to remember the dates of the authors it cited, or their willingness to consult the list in Passow’s Grundzüge der griechischen und römischen Litteraturgeschichte of 1816, while Liddell and Scott provided a list of authors and editions in the front matter of the dictionary, giving each a floruit date, arrived at ‘chiefly by the aid of Mr. Fynes Clinton’s Fasti Hellenici’ of 1824–34, which included authors in its chronological tables.¹² Herbert Coleridge must have seen that in this respect, the precise dating which was made possible by the printed publication of so many of his primary texts could best be conveyed by dating each quotation individually. An advantage of listing dated quotations in a single sequence separate from the definitions was that it resolved a serious practical problem which arose from any attempt at lexicography on historical principles. The life-story of a word was prima facie likely to have a rational, coherent shape, so that, for instance, it would be used first in a literal sense and then metaphorically. But the imperfections of the available evidence might mean that the first written attestation of a word was not an attestation of the sense which was likely to have developed first: as Liddell and Scott put it, ‘deviations from the strict Historical order must occur. Homer sometimes uses a word in a metaphorical sense only, the literal sense of which first occurs (perhaps) in Plato. In such instances, of course, we give the literal and actual sense the preference’ (L&S 1843: viii).¹³ Coleridge had worked out a way to preserve the chronological order of his quotations while assigning logical order—that is, the order in which the senses were likeliest to have developed (cf. Hiorth 1954–5)—to his definitions. *

* *

Between Coleridge’s draft entries of 1862 and the publication of the first fascicle of the New English Dictionary under the editorship of Furnivall’s

¹¹ For Donnegan, see Stray 2010b, 103 4. The advantage of translating Greek quotations is discussed anew in Lee 2010b, 129 30. ¹² Passow 1816 is discussed briefly by Considine 2014, 269; Clinton 1824 34, cited LS¹, x, is discussed by Murray 1997, 523 4. ¹³ Liddell and Scott cite [Fishlake] 1834, 172, where there is a scathing criticism of Charles Richardson’s adherence to chronological order in his English dictionary, and a conclusion that ‘Passow’s whole lexicon is a striking and beautiful illustration’ of the right way to take account of the logic of sense development in the application of historical principles. Cf. Müller 1899, 16, ‘The Dean’s [Liddell’s] good sense has generally kept him on the via media between a purely chronological and purely logical arrangement of meanings.’

Liddell and Scott and the Oxford English Dictionary

401

successor, James Murray, in 1884, much progress was made by the editors of the English dictionary, some of it attributable to the influence of Liddell and Scott, and some of it a matter of the development of new and independent procedures. The increasing independence of the project from the influence of Liddell and Scott was to be expected. Neither Furnivall nor Murray was a classicist in the way that Herbert Coleridge and Richard Trench had been: Coleridge had won the Newcastle Scholarship at Eton and had earned a double First in classics and mathematics at Oxford, and Trench had published on New Testament Greek, while Furnivall and Murray were both polymaths with a particular interest in, respectively, the English and Scots languages. So, it was only in its very first years, the late 1850s and the beginning of the 1860s, that the NED project was planned by men for whom studying Greek had ever been a central intellectual activity. On the other hand, from 1877 onwards, Henry Liddell was taking a direct personal interest in the NED project in his capacity as one of the Delegates of Oxford University Press, to which it was offered in that year, with Murray proposed as editor.¹⁴ Liddell had some prior knowledge of Murray, and thought well of his work, but when he was shown a draft dictionary specimen prepared by Murray, he was not particularly pleased with it. His objections, and those of other Delegates, concerned ‘the extent to which proper names, transparent formations on suffixes, and items of regional vocabulary were to be included . . . , and the acceptability of newspaper quotations as evidence’ (Gilliver 2016, 93). All three of these were predictable grey areas. The latter two, together with a slightly later suggestion from the Delegates that the new dictionary should not indicate the pronunciation of words and should leave etymologies to W.W. Skeat’s forthcoming etymological dictionary of English, do suggest that the scope of the NED project was seen as particularly problematic when it went beyond that of Johnson’s Dictionary and classical dictionaries such as Liddell and Scott, but they do not suggest particularly strong input from Liddell, or particularly close modelling of the Delegates’ vision of the NED on the Lexicon. Liddell was a member of a committee struck at Murray’s request in 1879 ‘to confer with him from time to time on literary questions’, as was his fellow delegate Max Müller (the committee achieved little, but both men would be thanked in a list of people who had contributed to the first volume of the NED); he supported Murray’s view that quotations in the NED would have to be long enough to show the sense of the words illustrated; and he annotated early proofs of the dictionary critically, catching, inter alia, an error in heraldic terminology.¹⁵ So he did make real ¹⁴ For Liddell as Delegate, see Curthoys 2013, 62 3. For Liddell and OED, see Gilliver 2016, ad indicem, s.n. ¹⁵ For these contributions, see Gilliver 2016, 107, 133 4, and 163 n. 190; for the acknow ledgements, see NED 1, xiii.

402

John Considine

contributions to the dictionary, but Murray’s principal adviser on the Greek language was not Liddell but a third Delegate, the Aristotelian scholar and book collector Ingram Bywater (NED 1, xii; 3.1, vi; 5, vii; etc.; Gilliver 2016, 154, 164). The English dictionary was bound, if it was to do anything like justice to the enormous corpus of written English, to be much larger than Liddell and Scott. Furnivall made the comparison explicitly in 1862, distinguishing ‘the large Dictionary’ from a smaller parallel project which he had in mind at the time (it came to nothing), ‘one of the size of Liddell & Scott, 1600 pages 4to’ (quoted Gilliver 2016, 47). Its longest entries were also bound to be much longer than those of Liddell and Scott, and indeed much longer than Coleridge’s draft entries. This meant that the definition of each sense and the quotations for that sense would really have to stay together, as they do in the Lexicon, if their relationship was to be apparent to any reader. It also meant that the structure of entries would have to be indicated by very careful use of typography. Murray’s printed page, as realized in the first fascicle of 1884 and thereafter throughout the dictionary, resembled that of Liddell and Scott in so far as his headwords were in bold type and slightly indented, and the use of bold type in headwords does appear to have been pioneered in the Lexicon (see Stray 2010b, 94–5). But he had to go beyond their example to make typographical sense of some of his very long entries, in which numerous quotations are given in full, so the distinctive beauty of an OED page, in which each sense starts on a new line, and the quotation block following each definition is again on a new line and in smaller type, is largely Murray’s achievement. As for the tension between the probable order of the development of senses of a given word and the dates of actual attestations, this called for careful thought. Furnivall would, in the 1860s, have liked to arrange the senses in logical order within an entry, whatever the dating of the available evidence (Gilliver 2016, 72–4). However, Murray’s preferred practice, and therefore the practice of the published dictionary, was to be guided by the dating of the evidence. In this respect, Murray could afford to be more rigorously historical than Furnivall or Liddell and Scott, because he had fewer gaps in his evidence: much had been gathered since Furnivall began to edit the English dictionary project, and much more was available to him than could be available to a lexicographer of ancient Greek. (The same historical rigour was applied by his colleagues: hence the first sense of the adjective soft, edited by William Craigie, is a metaphorical one, attested a couple of centuries before the literal one; John Chadwick regarded the treatment of μαλακός in LSJ, where the literal sense is given before the metaphorical one, as ‘more intelligent’.¹⁶) A dispute over the ¹⁶ Chadwick 1996, 19; since both the literal and the metaphorical sense of μαλακός are in Homer, it was possible for Liddell and Scott to put the literal one first without doing violence to chronology.

Liddell and Scott and the Oxford English Dictionary

403

wording of the title page of the first fascicle of the NED suggests the distance between Murray and Liddell. The working title of the dictionary until the end of September 1883 had been ‘A New English Dictionary on a Historical Basis’. Benjamin Jowett, who as Vice-Chancellor of the University of Oxford (1882–6) was ex officio a Delegate of the Press, disliked ‘on a Historical Basis’ and proposed ‘showing the History of the Language from the Earliest Times’; when this was rightly objected to on the grounds that there is more to the history of the language than the history of words, Liddell suggested ‘arranged so as to show the continuous history of the Words’.¹⁷ This emphasis on continuous history, which would have tended to commit the dictionary to a logical arrangement of senses in cases where the chronological order of attestations would have made the sense-development appear logically discontinuous, was rejected by Murray, and the final form of words, ‘on Historical Principles’ (cf. Philological Society 1859, 4), was adopted in November 1883 (Gilliver 2016, 172). However, Murray did impose a logical structure on historically ordered senses in the case of words like bad. There, he not only distinguished the uses of the word in three grammatical functions followed by a short section of compounds, all marked by letters—(A) adjective, (B) noun, (C) adverb, (D) compounds—but also distinguished the uses of the adjective in two semantic branches, marked by Roman numerals, (I) ‘In a privative sense: Not good’, followed by senses 1 to 4, the first three being in chronological order of first attestation, and (II) ‘In a positive sense: Evil, ill, noxious’, followed by senses 5 to 8, all four being in chronological order of first attestation. A remarkably similar structure is in fact found in the treatment of κακός in the first edition of L&S: the adjective is followed by (B) τὸ κακόν and τὰ κακά, (C) κακῶς, (D) a note on comparative and superlative forms (treated by Murray in the etymology of bad), and (E) a note on compounds, and the adjective itself is separated into semantic branches (I) ‘bad in its kind, bad, worthless, useless’, followed by senses 1 to 3, and (II) ‘bad, evil, mischievous’. Whether or not Murray actually turned to Liddell and Scott for guidance as he started to analyse the semantic range of the English word (it is worth noting that by the time he was compiling his entry, the most recent editions of the Lexicon had rearranged the semantic branches into ‘of persons’ and ‘of outward things’), he was at least applying Liddell and Scott’s principles. We do, finally, know that Murray and Liddell talked about lexicography together—which does not go without saying, because Liddell was not always an easy man to talk to. In 1902, after the Dean’s death, Murray passed on advice which Liddell had given him to his colleague William Craigie: ‘Everybody can make distinctions: it is the lexicographer’s business to make broad ¹⁷ Gilliver 2016, 167, 172; a proof with Jowett’s form of words corrected by Liddell is reproduced in Mugglestone 2005, 156.

404

John Considine

definitions which embrace them; the synthetic power is far above the analytic’ (quoted Gilliver 2016, 276). Here we see Liddell’s interest in the unified entry expressed once more, in a way which Murray took as a guide for his own practice. We have noticed Henry Liddell’s initial reluctance to approve the policy of including extensive etymologies in NED. The relationship of the etymologies in Liddell and Scott to those drafted for the NED project from Herbert Coleridge’s day onwards also calls for notice. The etymological material in Coleridge’s draft entries was in fact very limited. The only one to have an elaborate etymology was for the interjection ah, where Coleridge noted ‘The root is AK or AH, denoting sorrow or pain, a root which reappears constantly, as Gr. ἄχος, Eng. ache, A[nglo] S[axon] ace, etc.’ (in Furnivall 1862, [126]). As an etymology of ah, this is fanciful: Johnson had sensibly left the word unetymologized in 1755, and even the notoriously feeble eighteenth-century etymologist George William Lemon had confined himself to deriving it from the Latin interjection a (1783, first alphabetical sequence, s.v.). But it is paralleled in the entry for ἄχος in the first edition of the Lexicon, where the Greek headword and the English equivalent ache are given in capital letters to indicate that they are both roots, the implication being that they are both reflexes of the same root, and a cross-reference leads to an etymological note at ἄχω that ‘The root is the ejaculation of pain Ach! Ah! ἀχ–, cf. ἄχθομαι’. The relationship between Coleridge’s etymology and Liddell and Scott’s is not entirely clear: either or both could have been influenced by an article of 1836 by the early philologist Richard Garnett which brought ache and ἄχος together with many words for sharp things, pain, and fear, and was quoted with approval as late as the 1860s, but this article does not adduce ah and ach.¹⁸ At this early stage, Liddell and Scott had already ‘introduced a little Comparative Etymology, by quoting kindred Roots from Sanscrit, and other of the great family of Indo-European Tongues’ (LS¹, ix), their main named source being Pott’s Etymologische Forschungen of 1833–6 (in which ἄχος appears not to be discussed). By the sixth edition (1869), they had cleaned up the etymological treatment of ἄχος: it was still given in capitals with ache as a capitalized equivalent, but beyond this implicit suggestion that the two words were cognate, the only etymological statement in the entry was a reference to material in the entry for ἄγχω taken from Georg Curtius’ Grundzüge der griechischen Etymologie of 1858–62. In this latter entry, they provided Sanskrit, Latin, Gothic, and Old High German cognates such as Latin angustus and anxius. Half a century later, in the first fascicle of LSJ, ἄχος was no

¹⁸ Garnett 1836, 375, cited with strong approval in Farrar 1865, 94.

Liddell and Scott and the Oxford English Dictionary

405

longer in capitals, ache was no longer offered as an equivalent, and a terser cross-reference led to the etymological material s.v. ἄγχω. Meanwhile, the etymology of the English word had also been cleared up by Murray: the first fascicle of the NED pointed out that ah is not attested in Old English, and may therefore appear in Middle English under the influence of a similar form in Old French; it made no connection between ah and ache; and it pointed out in the etymology of the verb ache that it is not cognate with ἄχος (the verb had been spelt ake until its connection with the Greek word started to be supposed, so it was worth noting that the connection was not valid). As for English words which are indeed cognate with ἄχος, the NED noted at angina that Latin angina may be compared with Latin angere and Greek ἀγχόνη, and at anguish that Latin angustus is cognate with Greek ἄγχω, but at neither entry did it point out the wider range of cognates from Curtius which Liddell and Scott had provided, and it did not comment on the cognates of Latin anxius s.v. anxious. So, Liddell and Scott in the 1840s and Coleridge around 1861 were presenting the same mistaken etymology; in both cases, the etymology was improved in the light of nineteenth-century European comparative philology; but by the 1880s, Liddell and Scott were presenting somewhat more etymological information about the family represented by ἄχος than the NED. This was not because James Murray did not know the work of Curtius, which he corrected in the etymology of burn ‘spring, small stream’ in a fascicle published in 1888, and it was not because he was averse to giving surveys of cognates in etymologies, but rather because he was cautious of elaborate treatment of the etymologies of learned borrowings: space was always an issue in his dictionary work. * *

*

In the twentieth century, while Liddell and Scott and the OED co-existed as two highly visible Oxford dictionaries, the latter came to be more widely known, and James Murray came for many to be the model lexicographer. As John Chadwick put it in the introduction to his Lexicographica Graeca, ‘In rare cases the author of an outstanding dictionary is treated with the utmost marks of distinction, as having achieved something beyond the powers of ordinary mortals. I need only instance Sir James Murray . . . ’ He added less resoundingly that ‘To this class we may perhaps assign Henry George Liddell and Robert Scott’ (1996, 2). Chadwick went on to refer to the 1968 Supplement to LSJ as ‘unworthy both of Liddell and Scott and of the Oxford tradition of lexicography’ (1996, 8). Whether or not this harsh criticism was justified, it shows Chadwick’s sense that there was such a thing as an Oxford tradition of lexicography, and not just of Greek lexicography: he evidently saw OED and Liddell and Scott as having common standards because they were members of the same tradition. He had himself contributed to this tradition at more than one point, because he had served on the editorial staff of the Oxford Latin

406

John Considine

Dictionary and on the committee which oversaw the production of the 1996 Supplement to LSJ; since he had been recruited to OLD by James Wyllie, whose lexicographical training had been on the 1933 Supplement to the OED, he could also see that the tradition embraced OED and its classical cousins.¹⁹ The contributions of Oxford men such as Falconer Madan and R.W. Chapman to both Liddell and Scott and the OED illustrate the unity of the tradition, as does the progress of the Malayanist and former Craven scholar Richard Greentree from work on the OED to work on the Lexicon in 1905, and Chapman’s suggestion around 1910 that F.G. Fowler, who had recently co-edited the Concise Oxford Dictionary, might move on to work on a new edition of Liddell and Scott.²⁰ When the histories of the Lexicon and of OED are seen side by side within a single tradition, the two dictionaries can be seen as diverging in some respects and as converging in others, the divergences having been more marked. The most obvious divergence, increasing what was always a significant dissimilarity, has been the increasing scale of the OED, made desirable by the vast increase in the size of the corpus of English as well as by the increasing accessibility of earlier English texts, and made possible by the willingness of Oxford University Press to continue funding a very expensive, but very visible, scholarly project, and by the availability of electronic publication, which makes the production and distribution of an extremely large dictionary feasible. Henry Stuart Jones’s preface of 1925 had already compared the scale of Liddell and Scott’s compact original work to that of the sprawling reworkings of Henri Estienne’s Thesaurus graecae linguae of 1819–28 and 1831–65 (LSJ, iii–iv), and the scale of his own work to that of the ambitious projects of the first decade of the twentieth century for a new dictionary of ancient Greek (v–vi), and to that of the extant fascicles of Wilhelm Crönert’s reworking of Passow (x). Jones also noted in his preface that when the possibility of a new Greek dictionary on the comprehensive scale of the Thesaurus linguae latinae was being discussed, the medievalist Paul Meyer, who was a consultant to the NED, had suggested that ‘a more promising plan’ for a Greek dictionary would be ‘that of the New English Dictionary’ (quoted LSJ, vi), in other words a multi-volume dictionary founded on a large but by no means comprehensive body of quotations. This divergence of total length has been partly caused by a convergence of ambition. In 1948, at a point when the OED files were being maintained by James Wyllie but no new edition or supplement was being actively prepared,

¹⁹ For the sad story of Wyllie, see Brewer 2007, esp. 80 94, and Gilliver 2016, ad indicem, s.n.; cf. Stray 2012. ²⁰ For Madan, see Madan 1920, 130 1 (Liddell and Scott) and Gilliver 2000, 242 (OED); for Chapman, see Chapman 1948, 32 (Liddell and Scott) and 8 9 (OED); for Greentree, see Stray 2010b, 110; for Fowler, see McMorris 2001, 99.

Liddell and Scott and the Oxford English Dictionary

407

R.W. Chapman had reflected that ‘in Liddell & Scott you will find all the words known to have been written down . . . Nothing was for them too temporary’— he gave the Aristophanic ἀρχαιομελισιδωνοφρυνιχήρατα as an example—‘or too esoteric or too vulgar’ (the latter claim might be disputed, but one sees Chapman’s point). The lexicographer of English cannot, on the other hand, hope to register every recorded form from English literature, let alone nonliterary texts: ‘He must, however reluctantly, omit drear-nighted and skies of couple-colour’ (1948, 11). Keats’s drear-nighted had in fact been recorded in the first edition of OED, and Hopkins’s couple-colour is still not recorded in the dictionary—but some of the distinctive language of Hopkins had appeared in the Supplement of 1933, and more was to appear in that of 1972–86, its inclusion being a matter of deliberate policy on the part of the editor of the latter supplement, Robert Burchfield.²¹ The more nearly OED approached Liddell and Scott’s degree of comprehensiveness, the further it would expand beyond the size of the Lexicon. An argument in Chadwick’s Lexicographica Graeca adds a twist to the story of the relative sizes of the two dictionaries. Information about many specialized words which may be of interest to an epigraphist or a papyrologist ‘is useless for the average reader of Greek texts, and ought to be consigned to a special lexicon of record’, according to Chadwick, and this lexicon should be an online publication (Chadwick 1996, 19, 28). A new printed Liddell and Scott could then be produced, with a carefully selected vocabulary ‘slanted to give prominence to the authors most often studied in schools and university courses’ (28). Chadwick’s online ‘lexicon of record’ can be compared to the online OED, and his printed dictionary to the printed Shorter Oxford English Dictionary: the development which he imagined for Liddell and Scott is just the one which has taken place for OED. Nor is this a new comparison: as we have seen, when F.J. Furnivall imagined a concise companion to the NED in 1862, he suggested that it would be the size of Liddell and Scott, and when the Shorter Oxford English Dictionary was first conceived in 1899, it was referred to in one document as ‘the English Liddell & Scott’ (quoted Gilliver 2016, 285). The existence of a whole set of smaller English dictionaries, more or less direct offspring of the OED, running down from the Shorter through the Concise to a Pocket Oxford Dictionary, echoed the set of Liddell and Scott in its full version, the abridgement ‘chiefly for the use of schools’ which was published in the same year, and the intermediate version of 1889.²² One more point can be made about the relative size and comprehensiveness of the two dictionaries. In 1919, one of James Murray’s co-editors, William ²¹ Brewer 2007, 165, 186; see also ibid. 187 for Geoffrey Hill’s complaint in a review of the second edition of OED in 1989 that the language of Hopkins was still inadequately documented. ²² For the three sizes of Liddell and Scott, see Stray 2013a, 455 6; for the English dictionaries, see Brewer 2007, 76 80.

408

John Considine

Craigie, set out an ambitious plan for the compilation of a number of dictionaries on historical principles which would overlap in coverage with the OED, but would treat particular periods or local varieties in more detail than was possible for the more comprehensive dictionary (Adams 2009). Most of these ‘period dictionaries’ have now been produced, among them Craigie’s own Dictionary of American English, which R.W. Chapman described as providing the nearest parallel in English to Liddell and Scott’s ability to give ‘not merely good illustrations of every sense, but the best illustrations . . . all the most famous, all the most significant, and almost all the most entertaining examples’ (1948, 6). Although the provision of a set of Greek dictionaries has been imagined in the past—for instance by Hermann Diels (see LSJ, vi)—the closest thing to a ‘period dictionary’ to be produced as a companion to Liddell and Scott has been the Patristic Greek Lexicon edited by Geoffrey Lampe and completed in 1969 (see Bell 2013, 364–7). The typography of Liddell and Scott and of OED has undergone convergence in one respect and divergence in another. OED pages came to look slightly more like those of the Lexicon from 1972 onwards, when its headwords began to be printed with a lower-case initial letter (unless the word would normally be capitalized), as Liddell and Scott headwords had always been. Conversely, LSJ pages came to look slightly less like those of OED from 1925 onwards, when more entries than before were run together in blocks as a space-saving measure to ensure that the dictionary could still be sold as a single volume. In this respect, OLD has followed OED, setting not only each headword but each new sense and each quotation block on a new line, the result being that an opening of OLD is more elegant and easier to navigate than an opening of LSJ—and that the single-volume first edition of OLD was physically so unwieldy that the revised edition is issued in two volumes. An important internal divergence between the two dictionaries over the last century, and particularly over the last twenty years, has been the treatment of etymology. Whereas, as we have seen, nineteenth-century editions of Liddell and Scott sporadically provided more elaborate etymological information than early fascicles of the NED, etymology was always an important element of the latter, as its place among the opening elements of every entry made clear, and as one fascicle succeeded another in the decades-long publication of the NED, the etymologies tended to become more elaborate. Meanwhile, those in Liddell and Scott remained terse. By the end of the nineteenth century, Max Müller looked back on the ‘old objectionable etymologies’ of the earliest editions of Liddell and Scott, remarked on the superiority of the etymologies based on Curtius which had been added since then, but added that ‘a careful revision by a young scholar . . . would be very useful even now, and would be highly appreciated by classical scholars, who rightly recognize in every true etymology the pre-historic development of Greek words and Greek ideas’

Liddell and Scott and the Oxford English Dictionary

409

(1899, 16).²³ However, Jones reported in 1925 that ‘After careful consideration it was decided that etymological information should be reduced to a minimum’ (LSJ, x), although work was done on the etymologies of the ninth edition by Roderick McKenzie. Thereafter, the preface to the 1968 Supplement was silent on the subject, and the preface to the 1996 Revised Supplement acknowledged frankly that ‘the occasional etymological notes in LSJ are frequently out of date’ and that ‘No attempt has been made to bring them up to date’, referring readers to Frisk and Chantraine instead. By contrast, in the course of the revision of OED, the rewriting of etymologies has been undertaken by a substantial team, the results being often considerably more elaborate than NED etymologies or those in any complete etymological dictionary of English (see Durkin 1999 and 2009): judicious and elaborate etymologies for the interjections a and ah and the verb ache have been published online since 2009. Revision is now the most important element in the pattern of convergence and divergence in the story of the two dictionaries. Passow’s Handwörterbuch was itself a dynamic product of revision: a stage-by-stage reworking of Johann Gottlob Schneider’s Kritisches griechisch–deutsches Handwörterbuch of 1797–8, paying close attention to the vocabulary of Homer and Hesiod in the first edition, and taking that of Herodotus into account in the fourth, after which Passow had hoped to proceed with further editions in which more and more recent vocabulary would be revised.²⁴ Liddell and Scott revised all the classical vocabulary in Passow in their first edition rather than proceeding with his plan of incremental chronological progress, but their Lexicon was then revised from edition to edition, so it too progressed incrementally, though on a broader front. A former pupil of Liddell’s remembered that ‘In the interval between one Lecture and another he would be found standing at his desk over his interleaved copy of the first edition, correcting and amending it’—this must have been between 1843 and 1845—and a colleague told the story of hearing Liddell solemnly announcing that he was ‘writing the last sheet of the last edition of the Lexicon which I shall undertake’—this must have been the seventh edition, of which the preface is dated October 1882— and then, a year or so later, finding him ‘busy on one of the sheets of the lately issued edition, preparing already for the next’.²⁵ These reminiscences suggest

²³ Cf. Müller 1899, 14, recalling conversations with Liddell at some time shortly after the latter’s return to Oxford in 1855: ‘The Dean, though not a professed student of Comparative Philology himself, had read enough to know that the whole etymological portion would have either to be left out altogether or to be written again.’ ²⁴ The plan is set out in a letter of 1819 in Passow 1839, 261; its execution is assessed in [Fishlake] 1834, 150 3, and more briefly in LS¹, v. ²⁵ Thompson 1899: 48 and 80 1; the first reminiscence is by the Revd Henry Harvey, who matriculated at Christ Church in 1842 and received his BA in 1846, and would indeed therefore have been able to see Liddell at work preparing the second edition, and the second is by the Revd

410

John Considine

one of the two reasons why the Handwörterbuch and then Liddell and Scott could undergo nearly constant revision through most of the nineteenth century: on the one hand, there is a steady demand for a good textbook, and so new editions were a sound commercial proposition (Liddell and Scott received a couple of hundred pounds each in annual payments from the press, suggesting very healthy steady sales), and on the other, editorial costs were low, since the revision work was done by editors who were not employees of the dictionary publishers.²⁶ By contrast, the years over which the first edition of the OED was produced, from 1884 to 1928, were of course dedicated to the making of one edition of one work, and the only practicable way to add new material once the first edition was completed was in the form of a supplement, which was issued in 1933, and followed by a second supplement in four volumes from 1972 to 1986. The first edition and the supplements were keyboarded and merged into a single alphabetical sequence in electronic form, from which a second print edition was published in 1989; the dictionary in its electronic form was released on CD-ROM and then became the basis of an online edition which is being revised incrementally by a large editorial staff, with a quarterly schedule for the online publication of revised entries. Meanwhile, the deaths of Scott in 1887 and of Liddell in 1898 left the Lexicon orphaned. Falconer Madan remembered that Liddell had offered him the editorship before leaving Christ Church in 1891, no doubt in order to avoid the hiatus in editorship which in fact took place at the end of the nineteenth century and the beginning of the twentieth, followed by the ineffective editorship of Arthur Sidgwick, who had been talked into the position at a convivial dinner.²⁷ The preparation of the ninth edition of 1925–40 called for the co-ordination of expert work by many consultants, as Jones’s preface of 1925 made clear. Further revision was now bound to be a slow process: no longer could a single editor stand at his desk over an interleaved copy of the most recent edition of the dictionary, correcting it between his other duties. Since then, Liddell and Scott, like OED, has had two twentieth-century supplements which have presented numerous updatings and much new material, but the sort of revision which tweaks an existing set of entries in one small respect, or which rewrites all the definitions of a certain class or in a certain alphabetical range, has scarcely been possible. Like OED,

George Kitchin, who held various posts at Christ Church until 1883. See also LSJ, iv for Liddell’s collection of epigraphical material in an interleaved copy of the abridged Lexicon ‘for some years after the publication of the seventh edition’. ²⁶ For payments to Liddell and Scott, see Feather 2013, 347; see also Stray 2010b, 100 and Stray 2013b, 456 n. 89. ²⁷ Madan 1920, 131; Stray 2010b, 109, and Stray 2013b, 428; cf. Stray, Chapter 1, this volume.

Liddell and Scott and the Oxford English Dictionary

411

the Lexicon has been made available online, but unlike that of OED, the online texts of Liddell and Scott are static. Now that the shorter printed successor to LSJ imagined by John Chadwick is forthcoming in the form of the Cambridge Greek Lexicon (Diggle et al. 2019), it is possible to imagine how the online ‘lexicon of record’ which he also imagined might be created by incremental revision and augmentation of an online text of the lexicon. John A.L. Lee has set out a more detailed account than Chadwick’s of how this might be done, identifying ‘the electronic OED’ as a model for the on-screen presentation of entries (Lee 2010b, 128–35, esp. 130). Incremental online revision has certainly given a vitality to the OED which the release of printed supplements, however excellent, could not have continued to give, and the same, if the funding were available, would surely be true of Liddell and Scott. The long-term preservation of online scholarly resources should be a matter of grave concern to those who make and use them, and it would be sad to see Liddell and Scott or the OED cease altogether to be released in printed form, but the ability to make steady improvements which Henry Liddell and Robert Scott enjoyed in the nineteenth century is likely to be enjoyed by their twenty-first-century successors, as it is by those of James Murray, as part of an online programme of revision.

Bibliography Aarsleff, H. (1962) ‘The early history of the Oxford English Dictionary’, Bulletin of the New York Public Library 66: 417 39. Aarsleff, H. (1983) The Study of Language in England 1780 1860 (London). Accorinti, D. (ed.) (2016) Brill’s Companion to Nonnus of Panopolis (Leiden and Boston). Adams, J.N. (1982) The Latin Sexual Vocabulary (Baltimore). Adams, M. (2009) ‘The period dictionaries’, in A.P. Cowie (ed.), The Oxford History of English Lexicography (Oxford), 1.326 52. Adorno, F. et al. (2015) Corpus dei papiri filosofici greci e latini. Testi e lessico nei papiri di cultura greca e latina. Parte II 2: Sentenze di Autori noti e ‘Chreiai’ (Florence). Adrados, F.R. (1977) ‘Organización de los artículos. Criterios a seguir’, in id., E. Gangutia, J. López Facal, and C. Serrano Aybar (eds), Introducción a la lexico grafía griega (Madrid), 259 80. Adrados, F.R. et al. (1980 ) (eds) Diccionario griego español [DGE] (Madrid). Online at http://dge.cchs.csic.es/ Aijmer, K. and Simon Vandenbergen, A. M. (2011) ‘Pragmatic markers’, in J. Zienkowski, J. O. Östman, and J. Verschueren (eds), Discursive Pragmatics (Amsterdam), 223 47. Aitken, J.K. (1999) ‘The language of the Septuagint: Recent theories, future prospects’, Bulletin of Judaeo Greek Studies 24: 24 33. Aitken, J.K. (2014a) No Stone Unturned: Greek Inscriptions and Septuagint Vocabulary (Winona Lake, IN). Aitken, J.K. (2014b) ‘The language of the Septuagint and Jewish Greek identity’, in Aitken and Carleton Paget 2014, 120 34. Aitken, J.K. and Carleton Paget, J. (eds) (2014) The Jewish Greek Tradition in Antiquity and the Byzantine Empire (Cambridge). Aland, B., Aland, K., Karavidopoulos, J., Martini, C.M., and Metzger, B.M. (2000) The Greek New Testament, 4th rev. edn (Stuttgart). Aland, K., Black, M., Martini, C.M., Metzger, B.M., and Wikgren, A. (1968) The Greek New Testament, 2nd edn (Stuttgart). Alinei, M. (1997) ‘Principi di teoria motivazionale (iconomia) e di lessicologia moti vazionale (iconomastica)’, in L. Mucciante and T. Telmon (eds), Lessicologia e lessicografia (Rome), 9 36. Alinei, M. (2001) ‘Aspects of a theory of motivation (iconymy)’, VS/Quaderni di Studi Semiotici 88: 89 97. Anastassiou, A. and Irmer, D. (1997 2012) Testimonien zum Corpus Hippocraticum (Göttingen). Andersen, Ø. and Haug, D.T. (eds) (2012) Relative Chronology in Early Greek Epic Poetry (Cambridge). Andriotis, N.P. (1948) Το γλωσσικό ιδίωμα των Φαράσων (Athens). Andriotis, N.P. (1974) Lexikon der Archaismen in neugriechischen Dialekten (Vienna).

414

Bibliography

Anon. (1837) ‘A Greek fragment lately discovered at Derrynane’, Fraser’s Magazine 15: 578 80. Anon. (1891) ‘The Dean of Christchurch’, Saturday Review 72, 24 Oct.: 464 5. Anon. (1898) Review of F. Max Müller, Auld Lang Syne, The Speaker, 7 May: 582 3. Anon. (1899a) ‘Dean Liddell’, The Speaker 20, 2 Sept.: 240 2. Anon. (1899b) Review of Thompson 1899, Athenaeum 3744, 20 May: 620. Anon. (1937) ‘Oxford professor’s death: An Edinburgh man’, The Scotsman, 26 June: 17. Anon. (1940) ‘Postscript 1940’, in LSJ, xiii xiv. Asper, M. (1997) Onomata Allotria: Zur Genese, Struktur und Funktion poetologischer Metaphern bei Kallimachos (Stuttgart). Atherstone, A. (2003) ‘Benjamin Jowett’s Pauline commentary: An atonement con troversy’, Journal of Theological Studies 54: 139 53. Atkins, B.T.S. and Rundell, M. (2008) The Oxford Guide to Practical Lexicography (Oxford). Aune, D.E. (2003) ‘Distinct lexical meanings of ἀπαρχή in Hellenism, Judaism and early Christianity’, in Fitzgerald, Olbricht, and White 2003, 103 29. Aura Jorro, F. (1985 93) Diccionario micénico, 2 vols (Madrid). Babiniotis, G. (2011) Ετυμολογικό Λεξικό της Νέας Ελληνικής, 2nd edn (Athens). Bach, K. (1999) ‘The myth of conventional implicature’, Linguistics and Philosophy 22: 262 83. Bailly, A. (1950) Dictionnaire grec français, rev. L. Séchan and P. Chantraine (Paris). Bain, D.M. (1991) ‘Six Greek verbs of sexual congress (βινῶ, κινῶ, πυγίζω, ληκῶ, οἴϕω, λαικάζω)’, CQ 41: 51 77. Bain, D. and Coker, A. (2014) ‘Praefanda: The lexicography of ancient Greek aischro logia’, Eikasmos 25: 391 416. Bakker, E.J. (1988) Linguistics and Formulas in Homer: Scalarity and the Description of the Particle per (Amsterdam). Bakker, E.J. (ed.) (2010) A Companion to the Ancient Greek Language (Chichester). Balz, H.R. (1971) Heilsvertrauen und Welterfahrung: Strukturen der paulinischen Escatologie nach Römer 8, 18 39 (Munich). Barrett, C.K. (1957) A Commentary on the Epistle to the Romans (London). Batey, M. (2006) ‘Knox (Alfred) Dillwyn (1884 1943)’, in the Oxford Dictionary of National Biography (Oxford) (http://www.oxforddnb.com/view/article/37641, accessed 26 June 2017). Bauckham, R. (1983) Jude and 2 Peter: Word Biblical Commentary (Waco, TX). Bauer, W., Aland, K., and Aland, B. (1988) Griechisch deutsches Wörterbuch zu den Schriften des Neuen Testaments und der frühchristlichen Literatur, 6th edn (Berlin). Bauer, W., Arndt, W.F., and Gingrich, F.W. (1957) A Greek English Lexicon of the New Testament and Other Early Christian Literature: A Translation and Adaptation of Walter Bauer’s Griechisch Deutsches Wörterbuch zu den Schriften des Neuen Testa ments und der übrigen urchristlichen Literatur, fourth revised and augmented edition 1952 (Chicago). Bauer, W., Danker, F.W., Arndt, W.F., and Gingrich F.W. (2000) A Greek English Lexicon of the New Testament and Other Early Christian Literature, 3rd edn (Chicago). Beck, S. and Gergel, R. (2015) Contrasting English and German Grammar: An Intro duction to Syntax and Semantics (Berlin).

Bibliography

415

Beekes, R. (2010) Etymological Dictionary of Greek (Leiden and Boston). Bell, A. (2013) ‘Scholarly and reference publishing’, in Eliot 2013, 3.325 88. Bell, H.I. and Skeat, T.C. (1935) Fragments of an Unknown Gospel and Other Early Christian Papyri (London). Belson, M. (2003) On the Press: Through the Eyes of the Craftsmen of Oxford University Press (Witney). Benjamin, S. (2012) ‘On the distinctiveness of poetic language’, NLH 43: 89 111. Bentein, K., Janse, M., and Soltic, J. (eds) (2017) Variation and Change in Ancient Greek Tense, Aspect and Modality (Leiden). Benveniste, E. (1937) ‘Expression indo européenne de l’éternité’, Bulletin de la Société de Linguistique 38: 103 12. Benveniste, E. (1949) ‘Noms d’animaux en indo européen’, Bulletin de la Société de Linguistique de Paris 45: 74 103 Benveniste, E. (1966) ‘Problèmes sémantiques de la reconstruction’, in id., Problèmes de linguistique générale (Paris), 289 307. Benveniste, É. (1969) Le Vocabulaire des institutions indo européennes, 2 vols (Paris). Berenguer Sánchez, J.A. (2000) Estudio sobre las partίculas indoeuropeas con base consonántica y laringal (Madrid). Bertrand, N. (2015) ‘Le pronom anaphorique ὅ γε chez Homère’, Gaïa 18: 275 291. Biber, D. (2005) Dimensions of Register Variation: A Cross linguistic Comparison (Cambridge). Blackie, J.S. (1861) ‘Homer and his translators’, Macmillan’s Magazine 4: 268 80. Blakemore, D. (2002) Relevance and Linguistic Meaning: The Semantics and Pragmat ics of Discourse Markers (Cambridge). Blakemore, D. (2004) ‘Discourse markers’, in L.R. Horn and G. Ward (eds), Handbook of Pragmatics (Oxford), 221 40. Blakiston, H.E.D. (2004) ‘Jones, Sir Henry Stuart (1867 1939)’, rev. by D. Gill, in H.C. G. Matthew and B. Harrison (eds), Oxford Dictionary of National Biography, in Association with the British Academy: From the Earliest Times to the Year 2000 (Oxford), 30.521 4. Blanshard, A.J.L. (2010) Sex: Vice and Love from Antiquity to Modernity (Chichester). Blass, F., Debrunner, A., and Funk, R.W. (1961) A Greek Grammar of the New Testament and Other Early Christian Literature: A Translation and Revision of the Ninth Tenth German Edition incorporating Supplementary Notes of A. Debrunner (Chicago). Boisacq, É. (1916, 4th edn 1950) Dictionnaire étymologique de la langue grecque étudiée dans ses rapports avec les autres langues indo européennes (Heidelberg). Boned Colera, P., Rodríguez Somolinos, J., et al. (1998) Repertorio bibliográfico de la lexicografía griega. DGE Anejo III (Madrid). Boned Colera, P., Rodríguez Somolinos, J. and Elías, M. (2008) RBLG Suplemento 1, 2, and 3. Online at dge.cchs.csic.es/blg/blg s.htm Bonifazi, A. (2009a) ‘Discourse cohesion through third person pronouns: The case of κεῖνος and αὐτός in Homer’, in G.C. Wakker and S.J. Bakker (eds), Discourse Cohesion in Ancient Greek (Leiden), 1 19. Bonifazi, A. (2009b) ‘The pragmatic meanings of some discourse markers in Homer’, in E. Rieken and P. Widmer (eds), Pragmatische Kategorien: Form, Funktion und

416

Bibliography

Diachronie. Akten der Arbeitstagung der Indogermanischen Gesellschaft vom 24. 26.09.2007 in Marburg (Wiesbaden), 29 36. Bonifazi, A. (2012) Homer’s Versicolored Fabric: The Evocative Power of Ancient Greek Epic Word making (Cambridge, MA). Bonifazi, A, Drummen, A., and de Kreij, M. (eds) (2016) Particles in Ancient Greek Discourse: Five Volumes Exploring Particle Use across Genres (Washington, DC). Bortone, P. (2010) Greek Prepositions: From Antiquity to the Present (Oxford). Boys Stones, G.R. (ed.) (2003) Metaphor, Allegory, and the Classical Tradition: Ancient Thought and Modern Revisions (Oxford). Boys Stones, G.R. and Haubold, J.H. (eds) (2010) Plato and Hesiod (Oxford). Bracke, E. (2009) ‘Of Μῆτις and magic: The conceptual transformations of Circe and Medea in ancient Greek poetry’. PhD thesis, National University of Ireland, Maynooth. Bradley, M. (2013) ‘Colour as synaesthetic experience in antiquity’, in S. Butler and A. Purves (eds), Synaesthesia and the Ancient Senses (London), 127 40. Brent, R. (1983) Liberal Anglican Politics: Whiggery, Reaction, and Reform (Oxford). Brewer, C. (2007) Treasure house of the Language: The Living OED (New Haven and London). Brightfield, M.F. (1944) ‘Lockhart’s Quarterly Contributors’, PMLA 59: 491 512. Broadhead, H.D. (ed.) (1960) The Persae of Aeschylus (Cambridge). Brock, N. van (1961) Recherches sur le vocabulaire médical du grec ancien: soins et guérison (Paris). Brock, R. (2013) Greek Political Imagery from Homer to Aristotle (London). Brown, C.H. (2000) ‘Folk classification: An introduction’, in P. Minnis (ed.), Ethno botany: A Reader (Norman, OK), 65 8. Browning, R. (1983) Medieval and Modern Greek (Cambridge). Bruce, F.F. (1985) The Letter of Paul to the Romans: An Introduction and Commentary (Leicester). Buck, C.D. and Petersen, W. (1945) A Reverse index of Greek Nouns and Adjectives (Chicago). Bulloch, A.W. (1985) Callimachus: The Fifth Hymn (Cambridge). Büring, D. (1997) The Meaning of Topic and Focus: The 59th Street Bridge Accent (London). Burnett, A. (2007) The Letters of A.E. Housman, 2 vols (Oxford). Burrow, J.W. (1981) A Liberal Descent: Victorian Historians and the English Past (Cambridge). Bussemaker, U.C. (1862) Poetae bucolici et didactici (Paris). Butcher, S.H. and Lang, A. (trs) (1878) The Odyssey of Homer (London). Cadell, H. (1973) ‘Papyrologica: A propos de πυρός et de σῖτος’, Chronique d’Égypte 48: 329 38. Caird, G.B. (1968) ‘Towards a Lexicon of the Septuagint. I’, Journal of Theological Studies 19: 453 75. Caird, G.B. (1969) ‘Towards a Lexicon of the Septuagint. II’, Journal of Theological Studies 20: 21 40. Cairns, D.L. (1998) ‘Ἄωτος, ἄνθος and the death of Archemorus in Bacchylides’ Ninth Ode’, Papers of the Leeds International Latin Seminar 10: 57 73.

Bibliography

417

Cairns, D.L. (2014) ‘Psuche, thumos and metaphor in Homer and Plato’, Étudies platoniciennes 11, http://etudesplatoniciennes.revues.org/566 Calder, W.M. III (1989) ‘In response to Kenneth F. Kitchell, Jr., “How the wrong parts wrote Scott and the right parts wrote Liddell”: The Westminster epigram on Dean Liddell’, CJ 84: 265 6. Camden, W. 1595. Institutio Graecae grammatices compendiaria (London). (Consulted via Early English Books Online: https://eebo.chadwyck.com/home.) Carastro, M. (2006) La Cité des mages: penser la magie en Grèce ancienne (Paris). Carlyle, T. (2013) On Heroes, Hero worship, and the Heroic in History (New Haven and London). Cassio, A. C. (2009) ‘The language of Hesiod and the Corpus Hesiodeum’, in F. Montanari, A. Rengakos, and C. Tsagalis (eds), Brill’s Companion to Hesiod (Leiden and Boston), 179 201. Catenacci, C. (1996) ‘Ὁμοίϊοϛ (γελοίϊοϛ, ὀλοίϊοϛ): evidenti aggiustamenti metrici nell’esametro’, Quaderni Urbinati di Cultura Classica 52.1: 133 44. Catrein, C. (2003) Vertauschte Sinne: Untersuchungen zur Synästhesie in der römischen Dichtung (Munich). Cavallo, G. and Maehler, H. (1987) Greek Bookhands of the Early Byzantine Period, .. 300 800 (London). Chadwick, H. (1982) ‘The Patristic Greek Lexicon’ in Moule 1982, 66 72. Chadwick, J. (1986) ‘The semantic history of Greek ἐσχάρα’, in A. Etter (ed.), o o pe ro si: Festschrift für Ernst Risch zum 75. Geburtstag (Berlin), 515 23. Chadwick, J. (1992a) The Decipherment of Linear B, 4th edn (first published 1958) (Cambridge). Chadwick, J. (1992b) ‘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), 281 8. Chadwick, J. (1994) ‘The case for replacing Liddell and Scott’, BICS 39: 1 11. Chadwick, J. (1996) Lexicographica Graeca: Contributions to the Lexicography of Ancient Greek (Oxford). Chadwick, J. and Baumbach, L. (1963) ‘The Mycenaean Greek vocabulary’, Glotta 41: 157 271. Chadwick, O. (1962) ‘Patristic Greek’, CR 12: 222 4. Chadwick, O. (1970) The Victorian Church, Part 2: 1860 1901 (London). Chamberlain, A. (1903) ‘Primitive theories of knowledge: A study in linguistic psych ology’, The Monist 13: 295 302. Chantler, T. and DeArmitt, P. (2008) Sarah Kofman’s Corpus (New York). Chantraine, P. (1961) Morphologie historique du grec, 2nd edn (Paris). Chantraine, P. (1968 80) Dictionnaire étymologique de la langue grecque: histoire des mots (Paris). Chantraine, P. (1999) Dictionnaire étymologique de la langue grecque, augmented edn (Paris). Chantraine, P. (2009) Dictionnaire étymologique de la langue grecque, new edn (Paris). Chapman, M.D. (2016) ‘New Testament revision company (act. 1870 1881)’, Oxford Dictionary of National Biography (Oxford) (http://www.oxforddnb.com/view/ theme/96401, accessed 20 June 2017). Chapman, R.W. (1948) Lexicography (London).

418

Bibliography

Christidis, A. Ph. (ed.) (2007), A History of Ancient Greek: From the Beginnings to Late Antiquity (Cambridge). Cinque, G. (1999) Adverbs and Functional Heads: A Cross linguistic Perspective (Oxford). Clackson, J. (2007) Indo European Linguistics: An Introduction (Cambridge). Clark, G. (1944) ‘A reference to Plotinus in Liddell and Scott’, AJPh 65: 244 5. Clarke, M.C. (ed.) (1860) Shakespeare’s Works (New York). Clarke, M.J. (1995) ‘Aeschylus on mud and dust’, Hermathena 158: 7 26. Clarke, M.J. (1999) Flesh and Spirit in the Songs of Homer (Oxford). Clarke, M.J. (2004) ‘The semantics of colour in the early Greek word hoard’, in L. Cleland and K. Stears (eds), Colours in the Ancient Mediterranean World, BAR International Series 1267 (Oxford), 131 9. Clarke, M.J. (2005) ‘Etymology in the semantic reconstruction of early Greek words: The case of anthos’, Hermathena 179: 13 38. Clarke, M.J. (2010) ‘Semantics and vocabulary’, in Bakker 2010, 120 33. Clarke, M.L. (1945) Greek Studies in England 1700 1830 (Cambridge). Clay, J.S. (2003) Hesiod’s Cosmos (Cambridge). Clayton, B. (2004) A Penelopean Poetics: Reweaving the Feminine in Homer’s Odyssey (Lanham, MD). Clinton, H.F. (1824 34) Fasti Hellenici (Oxford). Coker, A. (forthcoming 2019) ‘Looking for dysphemistic language in ancient Greek (or: how filthy was Cleopatra?)’, Journal of Historical Pragmatics 20.2 (Special issue, ed. K. Ridealgh). Coleridge, H. (1860) ‘A letter to the Very Rev. the Dean of Westminster’, in Trench 1860, 71 8. Collinge, N.E. (1985) The Laws of Indo European (Amsterdam and Philadelphia). Colvin, S. (2007) A Historical Greek Reader: Mycenaean to the Koiné (Oxford and New York). Colvin, S. (2014) A Brief History of Ancient Greek (Chichester). Conacher, D.J. (1996) Aeschylus: The Earlier Plays and Related Studies (Toronto). Considine, J. (2009) ‘ “Les voyelles ne font rien, et les consonnes fort peu de chose”: On the history of Voltaire’s supposed comment on etymology’, Historiographia Lin guistica 36: 181 9. Considine, J. (2010) Current Projects in Historical Lexicography (Newcastle upon Tyne). Considine, J. (2014a) Academy Dictionaries 1600 1800 (Cambridge). Considine, J. (2014b) ‘John Jamieson, Franz Passow, and the double invention of lexicography on historical principles’, Journal of the History of Ideas 75: 261 81. Considine, J. (2016) ‘Historical dictionaries: History and development; current issues’, in Durkin 2016a, 163 75. Considine, J. (2017) Small Dictionaries and Curiosity from the Late Middle Ages to the Nineteenth Century (Oxford). Consani, C. (1991) Διάλεκτος: Contributo alla storia del concetto di ‘dialetto’ (Pisa). Cornford, F.M. (1920) ‘Liddell and Scott’, CR 34: 79. Cornford, F.M. (1937) Plato’s Cosmology (London). Cox, D. (2012) A Certain Share of Low Cunning: A History of the Bow Street Runners, 1792 1839 (Oxford and New York).

Bibliography

419

Craik, E.M. (2015) The ‘Hippocratic’ Corpus: Content and Context (Abingdon). Craik, H. and Smail, R. (2004) ‘Scott, Robert (1811 1887)’, in the Oxford Dictionary of National Biography (Oxford) (http://www.oxforddnb.com/view/article/24907, accessed 21 June 2017). Cranfield, C.E.B. (1975) A Critical and Exegetical Commentary on the Epistle to the Romans: Volume I, Introduction and Commentary on Romans I VIII (Edinburgh). Crystal, D. (1999) A Dictionary of Language, 2nd edn (Chicago). Crystal, D. (2014) Words in Time and Place: Exploring Language Through the Histor ical Thesaurus (Oxford). Cumberland, R. (tr.) (1812) The Clouds, in C. Dunster, H. Fielding, and R. Cumberland (trs), Comedies of Aristophanes (London). Cunliffe, R.J. (1924) A Lexicon of the Homeric Dialect (London). Currie, B. (2007) ‘Heroes and holy men in early Greece: Hesiod’s theios aner’, in A. Coppola (ed.), Eroi, eroismi, eroizzazioni. Dalla Grecia antica a Padova e Venezia (Padua), 163 203. Curthoys, M. (2013) ‘The Press and the University’, in Eliot 2013, 2.27 75. Curtius, G. (1858; 5th edn rev. E. Windisch 1879) Grundzüge der griechischen Etymo logie (Leipzig). Curtius, G. (1886) Principles of Greek Etymology, trans. A.S. Wilkins and E.B. England, 5th edn (London). Cuyckens, H. and Geeraerts, D. (eds) (2007) The Oxford Handbook of Cognitive Linguistics (Oxford). Dalby, A. (2003) Food in the Ancient World from A to Z (London). Danker, F.W. (2000) A Greek English Lexicon of the New Testament and other Early Christian Literature (BDAG). Based on Walter Bauer’s Griechisch deutsches Wörterbuch zu den Schriften des Neuen Testaments und der frühchristlichen Literatur, 6th edn, ed. K. Aland and B. Aland with V. Reichmann, and on previous English editions by W.F. Arndt, F.W. Gingrich, and F.W. Danker (Chicago). Danker, F.W. (2004) ‘Lexical evolution and linguistic hazard’, in Taylor et al. 2004, 1 31. Danker, F.W. and Krug, K. (2009) The Concise Greek English Lexicon of the New Testament (Chicago). Dawkins, R.M. (1916) Modern Greek in Asia Minor: A Study of the Dialects of Sílli, Cappadocia and Phárasa, with Grammar, Texts, Translations and Glossary (Cambridge). de Blécourt, W. (1994) ‘Witch doctors, soothsayers and priests: On cunning folk in European historiography and tradition’, Social History 19: 285 303. Denniston, J.D. (1954) The Greek Particles, 2nd edn (Oxford). Derksen, R. (2008) Etymological Dictionary of the Slavic Inherited Lexicon (Leiden). Detienne, M. and Vernant, J. P. (1974) Les Ruses de l’intelligence: la mètis des grecs (Paris). Devine, A.M. and Stephens, L.D. (2013) Semantics for Latin: An Introduction (Oxford). Dickey, E. (2007) Ancient Greek Scholarship (Oxford). Diggle, J. et al. (forthcoming 2019) Cambridge Greek Lexicon (Cambridge).

420

Bibliography

Dimitrakos, D.V. (1936 50) Μέγα λεξικὸν της Ελληνικής γλώσσης (Athens). Dimitrakos, D.V. (1964) Nέον λεξικόν ορθογραϕικόν και ερμηνευτικόν όλης της Ελληνικής γλώσσης: αρχαίας, μεσαιωνικής, καθαρεύουσης, δημοτικής (Athens). Dines, J.M. (2004) The Septuagint (London). Dodd, C.H. (1932) The Epistle of Paul to the Romans (London). Donnegan, J. (1826) A New Greek and English Lexicon; Principally on the Plan of the Greek and German Lexicon of Schneider (London). Donnegan, J. (1842) A New Greek and English Lexicon, Fourth Edition, Considerably Enlarged, Carefully Revised, and Materially Improved Throughout (London). Dover, K.J. (ed.) (1968) Aristophanes, Clouds (Oxford). Dover, K.J. (1981) ‘The expurgation of Greek literature’, Entr. Hardt 26, 55 89 [reprinted in K.J. Dover (ed.) (1988), The Greeks and Their Legacy (Oxford), 270 91]. Dover, K.J. (1997) The Evolution of Greek Prose Style (Oxford). [Draco], De metris poeticis = G. Hermann (ed.) (1812) Δράκωντος Στρατονικέως Περὶ μέτρων ποιητικῶν, καὶ πρῶτον περὶ χρόνων = Draconis Stratonicensis liber De metris poeticis (Leipzig). Dubois, P. (1988) Sowing the Body: Psychoanalysis and Ancient Representations of Women (Chicago). Duhoux, Y. (2008) ‘Mycenaean anthology’, in Y. Duhoux and A. Morpurgo Davies (eds), A Companion to Linear B: Mycenaean Greek Texts and Their World (Louvain la Neuve and Walpole, MA), 1.243 393. Dunbabin, R. (1946) ‘Notes on the new Liddell and Scott’, CR 60: 8 11. Dunbar, G. (1844) A New Greek and English and Greek and English Lexicon, 2nd edn (Edinburgh). Dunbar, G. and Barker, E.H. (1831) A Greek and English Lexicon (Edinburgh). Dunbar, N. (1995) Aristophanes, Birds (Oxford). Dunkel, G.E. (2014) Lexikon der indogermanischen Partikeln und Pronominalstämme (Heidelberg). Dunn, J.D.G. (1988) Romans 1 8 (Dallas, TX). Dunkle, R. (1987) ‘Nestor, Odysseus, and the μῆτιϛ : βίη antithesis: The funeral games, Iliad 23’, CW 81: 1 17. Durkin, P. (1999) ‘Root and branch: Revising the etymological component of the OED’, Transactions of the Philological Society 97: 1 49. Durkin, P. (2009) The Oxford Guide to Etymology (Oxford). Durkin, P. (ed.) (2016a) The Oxford Handbook of Lexicography (Oxford). Durkin, P. (2016b) ‘Etymology, word history, and the grouping and division of material in historical dictionaries’, in id. 2016a, 236 52. Ebeling, H. (1913) Griechisch deutsches Wörterbuch zum Neuen Testamente mit Nachweiss der Abweichungen des neutestamentlichen Sprachegebrauchs vom Attischen und mit Hinweis auf seine Übereinstimmung mit dem hellenistischen Griechisch (Hanover and Leipzig). Eberhard, B. and Joosten, J. (eds) (2016) Die Sprache der Septuaginta/The Language of the Septuagint (Gütersloh). Edwards, G.P. (1971) The Language of Hesiod in Its Traditional Context (Oxford). Edwards, M.W. (1991) The Iliad: A Commentary, Vol. V: Books 17 20 (Cambridge). Eliot, S. (ed.) (2013) The History of Oxford University Press, 3 vols (Oxford).

Bibliography

421

Eliot, S. and Rose, J. (eds) (2007) A Companion to the History of the Book (Oxford). Ellis, I. (1980) Seven Against Christ: A Study of ‘Essays and Reviews’ (Leiden). Erbse, H. (1953) ‘Bemerkungen zu Homer und zu seinen Interpreten’, Glotta 32: 236 47. Erlemann, K. (2001) ‘Anfänge, die das Ganze verbürgen: Überlegungen zu einer frühchristlichen Metaphergruppe’, Theologische Zeitschrift 57: 60 87. Eton Greek Grammar (1819) Graecae grammatices rudimenta in usum Regiae Scholae Etonensis, editio nova (Eton). Eton Greek Grammar, tr. H.J. Tayler (1843) = The Rudiments of Greek Grammar as used in the Royal College at Eton (London). Euler, W. (1979) Indoiranisch griechische Gemeinsamkeiten der Nominalbildung und deren indogermanische Grundlagen (Innsbruck). Fabricius, C. (1967) ‘A Patristic Greek Lexicon. Edited by G.W.H. Lampe. Fasc. 3 (εὐσυναλλάκτως μετεωρίζω)’, Journal of Theological Studies 18: 213 17. Fabricius, C. (1968) ‘A Patristic Greek Lexicon. Edited by G.W.H. Lampe. Fasc. 4 (μετεώρισμα προκατατίθημι)’, Journal of Theological Studies 19: 311 21. Faraone, C.A. and Teeter, E. (2004) ‘Egyptian Maat and Hesiodic Metis’, Mnemosyne 57: 177 208. Farrar, F.W. (1865) Chapters on Language (London). Faulkner, A. (2008) The Homeric Hymn to Aphrodite: Introduction, Text, and Com mentary (Oxford). Feather, J. (2013) ‘Authors and publishers’, in Eliot 2013, 2.321 53. Febvre, L. and Martin, H. J. (1997) The Coming of the Book: The Impact of Printing, 1450 1800 (London). Ferguson, E. (2009) ‘Words from the Bapt root in Classical and Hellenistic Greek’, in id., Baptism in the Early Church: History, Theology, and Liturgy in the First Five Centuries (Grand Rapids, MI), 38 59. Fillmore C.J. and Atkins, B.T.S. (2000) ‘Describing polysemy: the case of “crawl” ’, in Y. Ravin and C. Leacock (eds), Polysemy: Theoretical and Computational Approaches (Oxford), 91 110. Fillmore, C.J., Kay, P., and O’Connor, M.C. (1988) ‘Regularity and idiomaticity in grammatical constructions: The case of let alone’, Language 64: 501 38. Filos, P. (2018) Review of Montanari 2015, Bryn Mawr Classical Review 2018.03.46 (= http://bmcr.brynmawr.edu/2018/2018 03 46.html). Fischer, J. (2017) Food in Mycenaean Greece (Cracow). Fischer, R. (2010) Die Aḫhi̮ jawa Frage, mit einer kommentierten Bibliographie (Wiesbaden). Fisher, N.R.E. (1980) Review of M. Detienne and J. P. Vernant, Cunning Intelligence in Greek Culture and Society (1978), CR 30: 228 9. [Fishlake, J.R.] (1834) ‘Greek and German and Greek and English lexicography’, Quarterly Review 51, no. 101: 144 77. [Fishlake, J.R.] (1845) ‘Greek and English lexicography’, Quarterly Review 75, no. 150: 293 324. Fitzgerald, G.M. (1931) ‘Excavations at Beth Shan in 1930’, PEQ 63: 59 70. Fitzgerald, J.T., Olbricht, Th. H., and White, L.M. (eds) (2003) Early Christianity and Classical Culture: Comparative Studies in Honour of Abraham J. Malherbe (Leiden and Boston).

422

Bibliography

Fletcher, G.B.A. (1947) ‘Remarks on the new Liddell and Scott’, CR 61: 82 4. Flow, C. (2015) ‘Thesaurus matters? Frames for the study of Latin lexicography’, in C. Stray and G. Whitaker (eds), Classics in Practice: Studies in the History of Scholarship (BICS Supplement 128, London), 33 73. Foesius, A. (1588) Oeconomia Hippocratis (Frankfurt). Fowler, R. (2004) ‘Text and meaning in Richardson’s Dictionary’, in J. Coleman and A. McDermott (eds), Historical Dictionaries and Historical Dictionary Research (Tübingen), 53 62. Fowler, R.L. (1987) The Nature of Early Greek Lyric (Toronto). Franchi, R. (2016) ‘Approaching the “Spiritual Gospel”: Nonnus as an interpreter of John’, in Accorinti 2016, 240 66. Frantzen, A.J. (1990) Desire for Origins: New Language, Old English, and Teaching the Tradition (New Brunswick, NJ). Fraser, B. (1996) ‘Pragmatic markers’, Pragmatics 6: 167 90. Fraser, B. (2009) ‘An account of discourse markers’, International Review of Pragmat ics 1: 293 320. Frere, J.H. (1911 [1820]) ‘Introduction’, in The Plays of Aristophanes, vol. 2 (London). Frisk, H. (1960 72) Griechisches etymologisches Wörterbuch, 3 vols (Heidelberg). Fritz, G. (2005) Einführung in die historische Semantik (Tübingen). Fryer, P. (1963) Mrs Grundy (London). Furnivall, F. J. (ed.) (1862) An Alphabetical List of English Words Occurring in the Literature of the Eighteenth and Nineteenth Centuries, and Forming a Basis of Comparison for the Use of Contributors to the New Dictionary of the Philological Society, part III, M to Z (Hertford). García Ramón, J. L. (1990) ‘Mykenisch e qi ti wo e/(h)ekwhthiwohe(s) “umgekommen, tot”, homerisch ἔϕθιται, ἔϕθιεν, (o)ἔϕθιτο und das Perfekt von idg. *dhgwhei im Griechischen’, Münchener Studien zur Sprachwissenschaft 51: 7 20. García Ramón, J.L. (1992) ‘Mycénien ke sa do ro /Kessandros/, ke ti ro /Kestilos/, ke to /Kestōr/: grec alphabétique Αἰνησιμβρότα, Αἰνησιλάος, Αἰνήτωρ et le nom de Cassandra’, in J. P. Olivier (ed.), Mykenaïka (Actes du IXe Colloque international sur les textes mycéniens et égéens . . . , Athènes, 2 6 octobre 1990) (Athens and Paris), 239 55. García Ramón, J.L. (2011) ‘Mycenaean onomastics’, in Y. Duhoux and A. Morpurgo Davies (eds), A Companion to Linear B: Mycenaean Greek Texts and their World (Louvain la Neuve and Walpole, MA), 2.213 51. García Ramón, J.L. (2012) ‘En travaillant à une grammaire du mycénien: 1. a pi e qe /amphihenkwe/ “(on) mentionna, (on) énuméra”. 2. Absence d’augment et mode injonctif. 3. di ri mi jo: Drimios, fils de Zeus’, in P. Carlier, C. de Lamberterie, M. Egetmeyer, N. Guilleux, F. Rougement, and J. Zurbach (eds), Études mycéniennes 2010 (Actes du XIIIe Colloque international sur les textes égéens, Sèvres, Paris, Nanterre, 20 23 septembre 2010) (Pisa and Rome), 437 56. Garnett, R. (1836) ‘English dialects’, Quarterly Review 55: 354 87. Garnett, R. (1899) Review of Liddell and Scott, 8th edn, CR 13: 291 3. Garvie, A.F. (1986) Aeschylus, Choephori (Oxford). Garvie, A.F. (ed.) (2009) Aeschylus’ Persae (Oxford). Gascoigne, G. (1575) The Noble Art of Venerie, excerpted in The Green Knight, ed. R. Pooley, 1982 (Manchester).

Bibliography

423

Geeraerts, D. (1997) Diachronic Prototype Semantics (Oxford). Geeraerts, D. (2007) ‘Lexicography’, in Cuyckens and Geeraerts 2007, 1160 74. Geeraerts, D. (2010) Theories of Lexical Semantics (Oxford). George, C., McCullagh, M., Nielsen, B., Ruppel, A., and Tribulato, O. (eds) (2007) Greek and Latin from an Indo European Perspective (Cambridge). George, C. (2010) ‘Jewish and Christian Greek’, in Bakker 2010: 267 80. Gerber, D.E. (2002) A Commentary on Pindar Olympian Nine (Stuttgart). Giannakis, G.K. (ed.) (2014), Encyclopedia of Ancient Greek Language and Linguistics (Brill). Gibson, R.K. (2002) ‘ “Cf. e.g.”: A typology of “parallels” and the function of commen taries on Latin poetry’, in id. and C.S. Kraus (eds), The Classical Commentary: Histories, Practices, Theory (Leiden), 331 57. Gildersleeve, B.L. (1882) ‘Brief mention’, AJPh 3: 515. Gildersleeve, B.L. (1885) Pindar. The Olympian and Pythian Odes (London). Gildersleeve, B.L. (1898) ‘Brief mention’, AJPh 19: 233. Giles, J.A. (1839) A Lexicon of the Greek Language, for the Use of Colleges and Schools (London). Gilliver, P. (2000) ‘OED personalia’, in L. Mugglestone (ed.), Lexicography and the OED: Pioneers in the Untrodden Forest (Oxford), 232 52. Gilliver, P. (2016) The Making of the Oxford English Dictionary (Oxford). Glare, P.G.W. (1987) ‘Liddell & Scott: Its background and present state’, in R. Burch field (ed.), Studies in Lexicography (Oxford) 1987, 1 18. Glare, P.G.W. (1991) ‘Starting from the wrong end’, in G. Cigman and D. Howlett (eds) Birthday Celebration for Naky Doniach (Oxford), 35 41. Glare, P.G.W., with the assistance of A.A. Thompson (1996) Greek English Lexicon: Revised Supplement (Oxford). Glare, P.G.W. (1997) ‘Liddell and Scott: Then and now’, Hyperboreus 3: 205 17. Glare, P.G.W. et al. (1968 82) The Oxford Latin Dictionary [OLD] (Oxford). Goldstein, D.M. (2010) ‘Wackernagel’s law in fifth century Greek’, PhD diss., Univer sity of California, Berkeley. Goldstein, D.M. (2013a) Review of Devine and Stephens 2013, Bryn Mawr Classical Review 2013.07.04. Goldstein, D.M. (2013b) ‘The synchrony and diachrony of a scalar coordinator: Latin nēdum “let alone” ’, Journal of Indo European Linguistics 1: 68 106. Goldstein, D.M. (2014) ‘Wackernagel’s law I’, in Giannakis 2014, 3.508 13. Goldstein, D.M. (2016a) Classical Greek Syntax: Wackernagel’s Law in Herodotus (Leiden). Goldstein, D.M. (2016b) Review of Devine and Stephens 2013, Kratylos 61: 155 63. Gonis, N., Maltomini, F., Henry, W.B., and Slattery, S. (2016) The Oxyrhynchus Papyri, Volume LXXXII (London). Grant, E. 1575. Της Ελληνικης Γλωσσης σταχυολογία: Græcæ Linguæ Spicilegium. London: Ex officina Henrici Binemani pro Francisco Coldock. (Consulted via Early English Books Online: https://eebo.chadwyck.com/home.) Greenhill, W.A. (1872) Contributions to Medical Literary History: Adversaria Medico Philologica (London).

424

Bibliography

Greenlee, H.J. (1960) ‘A suggestion with regard to Romans 8:23’, The Bible Translator 11: 91 2. Greentree, R. and Nicholson, E.W.B. (1910) Catalogue of Malay Manuscripts . . . in the Bodleian Library (Oxford). Greetham, D.C. (1999) Theories of the Text (Oxford). Grenfell, B.P. and Hunt, A.S. (1908) The Oxyrhynchus Papyri V (London). Grenfell, B.P., Hunt, A.S., and Goodspeed, E.J. (1907) The Tebtunis Papyri II (London). Grice, H.P. (1975) ‘Logic and conversation’, in P. Cole and J. L. Morgan (eds), Syntax and Semantics 3: Speech Acts (New York), 41 58. Griffith, R.D. (2010) ‘ΤΡΕΦΕΙΝ ΓΑΛΑ (Odyssey 9.246)’, CPh 105: 301 8. Grigg, G. and Kirshner, D. (2015) Biology and Evolution of Crocodylians (Ithaca). Groenendijk, J.A.G. and Stokhof, M.J.B. (1984) ‘Studies on the semantics of questions and the pragmatics of answers’, PhD diss., University of Amsterdam. Groves, J. (1839) A Greek and English Dictionary (Boston). Gurr, A. (ed.) (2003) King Richard II, 2nd edn (Cambridge). Gutzmann, D. (2015) Use conditional Meaning (Oxford). Hajnal, I. (2006) ‘Die Tafeln aus Theben und ihre Bedeutung für die griechische Dialektologie’, in S. Deger Jalkotzy and O. Panagl (eds), Die neuen Linear B Texte aus Theben: Ihr Aufschlusswert für die mykenische Sprache und Kultur (Vienna), 53 69. Halleran, M.R. (1988) ‘Repetition and irony at Sophocles Trachiniae 574 81’, CPh 83: 129 31. Halliwell Phillipps, J.O. (1852) A Dictionary of Archaic and Provincial Words (London). Hamblin, C.L. (1958) ‘Questions’, Australian Journal of Philosophy 36: 159 68. Hamblin, C.L. (1973) ‘Questions in Montague English’, Foundations of Language 10: 41 53. Hanks, P. (2013) Lexical Analysis: Norms and Exploitations (Cambridge, MA). Hare, A.J.C. (1896) The Story of My Life II (London). Hare, J. (1647) St. Edwards Ghost, or Anti Normanisme, being a Patheticall Complaint and Motion in the Behalf of our English Nation against her Grand (Neglected) Grievance, Normanisme (London). Harris, J.R. (1886) Fragments of Philo Judaeus (Cambridge). Harrison, E. (1935) ‘More of the new Liddell and Scott’, CR 49: 226 7. Harrison, E. (1941) ‘The completion of the new Liddell and Scott’, CR 55: 28 30. Harrison, S.J. and Pelling, C.B.R. (eds) (forthcoming 2020) Classics and Classicists: Essays on the History of Scholarship in Honour of Christopher Stray (Oxford). Harrison, S.J. and Stray, C.A. (eds) (2012a) Expurgating the Classics: Editing Out in Greek & Latin (London). Harrison, S.J. and Stray, C.A. (2012b) ‘Introduction’, in eid. 2012a, 1 7. Hartung, J.A. (1832 3) Lehre von den Partikeln der griechischen Sprache (Erlangen). Hass, U. (ed.) (2012) Grosse Lexika und Wörterbücher Europas: Europäische Enzyklo pädien und Wörterbücher in historischen Porträts (Berlin). Hauspie, K. (2004) ‘The LXX quotations in the LSJ supplements of 1968 and 1996’, in Taylor et al. 2004, 108 25. Hawhee, D. (2004) Bodily Arts: Rhetoric and Athletics in Ancient Greece (Austin).

Bibliography

425

Hawke, A. (2016) ‘Quotation evidence and definitions’, in Durkin 2016a, 176 203. Headlam, W. (1902) ‘Metaphor, with a note on transference of epithets’, CR 16: 434 42. Heath, D. (2007) ‘Obscenity, censorship & modernity’, in Eliot and Rose 2007, 508 19. Henderson, J. (1991) The Maculate Muse: Obscene Language in Attic Comedy, 2nd edn (Oxford). Henderson, J.G.W. (1998) Juvenal’s Mayor: The Professor who Lived on 2d a Day (Cambridge). Herbert, A.P. (1970) A. P. H.: His Life and Times (London). Hermann, G. (1806) Homeri Hymni et Epigrammata (Berlin). Heubeck, A. and Hoekstra, A. (1989) A Commentary on Homer’s Odyssey: Volume II, Books IX XVI (Oxford). Heussler, R. (1981) British Rule in Malaya: The Malayan Civil Service and its Prede cessors, 1867 1942 (Westport, CT). Hickie, W.J. (1853) The Comedies of Aristophanes: A New and Literal Translation, vol. 1 (London). Hill, C. (1958) ‘ The Norman yoke’, in id., Puritanism and Revolution: Studies in Interpretation of the English Revolution of the 17th Century (London), 50 122. Hinchliff, P. (1982) ‘Benjamin Jowett and the Church of England: or “Why really great men are never Clergymen” ’, in J. Prest (ed.), Balliol Studies (London), 125 58. Hiorth, F. (1954 5) ‘Arrangement of meanings in lexicography: Purpose, disposition, and general remarks’, Lingua 4: 413 24. Hire, P. (2005) ‘The Cambridge new Greek lexicon project’, CW 98: 179 85. Hiscock, M. (2010) ‘The politics of classical allusion at the end of the eighteenth century: Radical or redundant?’, CCJ 56: 95 139. Holmberg, I. (1997) ‘The Sign of MHTIS’, Arethusa 30: 1 33. Holmberg, I. (1998) ‘Μῆτις and Gender in Apollonius Rhodius’ Argonautica’, TAPhA 128: 135 59. Holton, D., Mackridge, P., and Philippaki Warburton, I. (1997) Greek: A Comprehensive Grammar of the Modern Language (London). Holton, D. and Manolessou, I. (2010), ‘Mediaeval and Early Modern Greek’, in Bakker 2010, 539 63. Holton, D., Mackridge, P., and Philippaki Warburton, I. (2012) Greek: A Comprehensive Grammar 2nd edn rev. V. Spyropoulos (London). Hoogeveen, H. (1788) Doctrina particularum Graecarum, ed. Chr. Schütz (Leipzig). Hoogeveen, H. (1829) Hoogeveen’s Greek Particles, ed. J. Seager (London). Hordern, J.H. (2004) Sophron’s Mimes: Text, Translation, and Commentary (Oxford). Horrocks, G. (1997) Greek: A History of the Language and Its Speakers (London). Horrocks, G. (2010) Greek: A History of the Language and Its Speakers, 2nd edn (Chichester and Malden, MA). Horrocks, G. (2017) ‘High and low in Medieval Greek’, in Bentein, Janse, and Soltic 2017, 219 41. Horsley, G.H.R. (1989) New Documents illustrating Early Christianity: Vol. 5, Linguis tic Essays (North Ryde, NSW). Howsam, L. (2009) Past into Print: The Publishing of History in Britain, 1850 1950 (London).

426

Bibliography

Hughes, G. (1998) Swearing: A Social History of Foul Language, Oaths & Profanity in English (London). Hummel, P. (1999) L’Épithète pindarique (Bern). Hunter, R. (2014) Hesiodic Voices: Studies in the Ancient Reception of Hesiod’s Works and Days (Cambridge). Ilgen, K.D. (1796) Hymni Homerici cum reliquis carminibus minoribus Homero tribui solitis et Batrachomyomachia. Addita est Demetrii Zeni versio Batrachomyomachiae dialecto vulgari et Theodori Prodromi Galeomyomachia. Textum recensuit et anim adversionibus criticis illustravi (Halle). Imholtz, A. (2007) ‘Liddell and Scott: Precursors, nineteenth century editions, and the American contributions’, in Stray 2007b, 117 34. Israel, M. (2011) The Grammar of Polarity: Pragmatics, Sensitivity, and the Logic of Scales (Cambridge). Jackendoff, R. (1972) Semantic Interpretation in Generative Grammar (Cambridge, MA). Janko, R. (1982) Homer, Hesiod and the Hymns: Diachronic Development in Epic Diction (Cambridge). Janko, R. (ed.) (1992) The Iliad: A Commentary, Volume IV: Books 13 16 (Cambridge). Janko, R. (2012) ‘πρῶτόν τε καὶ ὕστατον αἰὲν ἀείδειν. Relative chronology and the literary history of the early Greek epos’, in Andersen and Haug 2012, 20 43. Janse, M. (2003) ‘The metrical schemes of the hexameter’, Mnemosyne 56: 343 8. Janse, M. and Joseph, B.D. (2014) ‘A new historical grammar of early Demotic Greek: Reflections on the Κοινή Ελληνική in the 19th and 20th centuries as seen through Thumb’s handbook of the Modern Greek vernacular’, CHS Research Bulletin 2.2: 1 10. Jay, J. (1940) ‘Review of LSJ’, Classical Weekly 34: 86 7. Jeffreys, E. (1998) Digenis Akritis: The Grottaferrata and Escorial Versions (Cambridge). Jocelyn, H.D. (1980) ‘A Greek indecency and its students: λαικάζειν’, PCPhS 26: 12 66. Johnson, S. (1755) Dictionary of the English Language (London). Jones, H.S. (1922) ‘ΣΠΙΛΑΣ. ΑΠΑΡΧΗ ΠΝΕΥΜΑΤΟΣ ’, Journal of Theological Studies 23: 282 3. Jones, H.S. (1925) ‘Preface 1925’, in LSJ, iii xii. Jones, H.S. (1940) ‘Aids to the reader’, in LSJ, xv xlv. Jones, H.S. (1941) ‘The making of a lexicon’, CR 55: 1 13. Jones, H.S., Alford, M., and Maas, P. (1940) ‘Addenda and corrigenda’, in LSJ, 2043 111. Jones, H.S. and McKenzie, R. (1940) A Greek English Lexicon Compiled by H.G. Liddell and R. Scott. A New Edition, Revised and Augmented Throughout [LSJ] (Oxford). Jones, J. (1823) A Greek and English Lexicon (London). Jones, J. (1825) The Tyro’s Greek English Lexicon (London). Jongkind, D., Williams, P., Head, P., and James, P. (2017) The Greek New Testament produced at Tyndale House, Cambridge (Cambridge and Wheaton, IL). Joosten, J. (2005) ‘The ingredients of New Testament Greek’, Analecta Bruxellensia 10: 56 69. Joosten, J. (2014) ‘Varieties of Greek in the Septuagint and New Testament’, in Aitken and Carleton Paget 2014, 22 45.

Bibliography

427

Joosten, J. (2016) ‘Septuagint Greek and the Jewish sociolect in Egypt’, in Eberhard and Joosten 2016, 246 56. Joseph, B.D. (2009) ‘Greek’, in B. Comrie (ed.), The World’s Major Languages, 2nd edn (London), 347 72. Kaplan, D. (1999) ‘The meaning of ouch and oops: Explorations in the theory of meaning as use’, MS., University of California, Los Angeles. Käsemann, E. (1973) An die Römer (Tübingen). Katz, J.T. (2007) ‘The epic adventures of an unknown particle’, in George et al. 2007, 65 79. Katz, J.T. (2010a) ‘Etymology’, in A. Grafton, G.W. Most, and S. Settis (eds), The Classical Tradition (Cambridge, MA), 342 5. Katz, J.T. (2010b) ‘Inherited poetics’, in Bakker 2010, 357 69. Katz, J.T. (2010c) ‘Nonne lexica etymologica multiplicanda sunt?’, in Stray 2010a, 25 48. Katz, J.T. (2013a) ‘Gods and vowels’, in J.V. García and A. Ruiz (eds), Poetic Language and Religion in Greece and Rome (Newcastle upon Tyne), 2 28. Katz, J.T. (2013b) ‘The hymnic long alpha: Μούσαϛ ἀείδω and related incipits in Archaic Greek poetry’, in S.W. Jamison, H.C. Melchert, and B. Vine (eds), Proceed ings of the 24th Annual UCLA Indo European Conference, Los Angeles, October 26th and 27th, 2012 (Bremen), 87 101. Katz, J.T. (2016) ‘Etymological “alterity”: Depths and heights’, in S. Butler (ed.), Deep Classics: Rethinking Classical Reception (London), 107 26. Katz, J.T. (2018) ‘Μῆνιν ἄειδε, θεά and the form of the Homeric word for “goddess” ’, in D. Gunkel and O. Hackstein (eds), Language and Meter (Leiden), 54 76. Katz, J.T. (2019) ‘Gods and vowels’, in S. Butler and S. Nooter (eds), Sound and the Ancient Senses (Abingdon), 153 70. (Revision of Katz 2013a.) Katz, P. (1942) ‘Some detached remarks concerning Liddell Scott. Oxford, 10.4.1942’, PW/4/1: 1 11, Archive of the Library of the Faculty of Asian and Middle Eastern Studies, Cambridge. Kay, C., Roberts, J., Samuels, M., and Wotherspoon, I. (eds) (2009) Historical The saurus of the Oxford English Dictionary (Oxford). Kay, P. (1992) ‘At least: New essays in semantic and lexical organization’, in A. Lehrer and E.F. Kittay (eds), Frames, Fields, and Contrasts (Hillsdale, NJ), 309 31. Kazazis, I.N. and Karanastasis, T.A. (2001 ), Επιτομή του Λεξικού της Μεσαιωνικής Ελληνικής Δημώδους Γραμματείας (1100 1669) του Εμμανουήλ Κριαρά (Thessaloniki). Kemble, J.M. (1849) The Saxons in England (London). Keene, C. (1893) ‘Notes on Liddell and Scott’s Lexicon’, CR 7: 328 9. Kendal, G.H. (1985) In Praise of Baldness: Synesius of Cyrene. Translated from the Greek with Introduction and Notes, illus. J. Kamins (Vancouver). Kenney, E.J. (1974) The Classical Text: Aspects of Classical Scholarship in the Age of the Printed Book (Cambridge). Kilgarriff, A. (1997) ‘I don’t believe in word senses’, Computers and the Humanities 31: 91 113. Kirk, G.S. (1962) ‘The structure and aim of the Theogony’, in O. Reverdin (ed.), Hésiode et son influence: six exposés et discussions (Geneva), 63 107.

428

Bibliography

Kirk, G.S., Raven, J.E., and Schofield, M. (1983) The Presocratic Philosophers, 2nd edn (Cambridge). Kitchell, K.F., Jr. (1988) ‘How the wrong parts wrote Scott and the right parts wrote Liddell’, CJ 84: 47 52. Kitchell, K.F., Jr. (1989) ‘The “Liddell Riddle”: Some further thoughts’, CJ 84: 266 8. Klein, R. (1986) ‘The Mētis of centaurs’, Diacritics 16.2: 2 13. Knapp, C. (1921) ‘Revised edition of Liddell and Scott’s Greek English Lexicon’, CW 14: 136. Knight, G.W. III (1992) The Pastoral Epistles: A Commentary on the Greek Text (Grand Rapids, MI and Cambridge). Knight, V. (1995) The Renewal of Epic: Responses to Homer in the Argonautica of Apollonius (Leiden). Knox, A.D. (1913) ‘ΣΠΙΛΑΔΕΣ ’, Journal of Theological Studies 14: 547 9. Knox, A.D. (1915) ‘Apocalypse xviii 15, 17; σπιλάς’, Journal of Theological Studies 16: 77 8. Koning, H. (2010) Hesiod: The Other Poet (Leiden). Körner, O. (1917) Das homerische Tiersystem und seine Bedeutung für die zoologische Systematik von Aristoteles (Wiesbaden). Kövecses, Z. (2006) Language, Mind, and Culture: A Practical Introduction (Oxford). Kövecses, Z. (2010) Metaphor: A Practical Introduction, 2nd edn (Oxford). Kretschmer, P. (1926) ‘Brot und Wein im Neugriechischen’, Glotta 15: 60 5. Kriaras, E. (1968 ) Λεξικό της Μεσαιωνικής Ελληνικής Δημώδους γραμματείας (1100 1669) (Thessaloniki). Kriaras, E. (2001) Επιτομή του λεξικού της Μεσαιωνικής Ελληνικής Δημώδους γραμματείας (1100 1669) (Thessaloniki). Krifka, M. (1993) Review of E. König, The Meaning of Focus Particles: A Comparative Perspective (1991), Language 69: 593 7. Krifka, M. (2008) ‘Basic notions of information structure’, Acta Linguistica Hungarica 55: 243 76. Kroesch, S. (1929) ‘The semantic development of OE “cræft” ’, Modern Philology 26: 433 43. Kroonen, G. (2013) Etymological Dictionary of Proto Germanic (Leiden). Kühn, C.G. (1821 30) Claudii Galeni Opera Omnia (Leipzig). Kvı́čala, J. (1870) ‘Untersuchungen auf dem Gebiete der Pronomina, besonders der lateinischen’, Sitzungsberichte der Österreichischen Akademie der Wissenschaften. Philosophisch historische Klasse 65: 77 155. Lakoff, G. (1987) Women, Fire, and Dangerous Things: What Categories Reveal about the Mind (Chicago). Lakoff, G. (1999) ‘Cognitive models and prototype theory’, in E. Margolis and S. Laurence (eds), Concepts: Core Readings (Cambridge, MA), 391 421. Lakoff, G. and Johnson, M. (1980) Metaphors We Live by (Chicago). Lampe, G.W.H. (ed.) (1961 8) A Patristic Greek Lexicon (Oxford). Langslow, D.R. (2000) Medical Latin in the Roman Empire (Oxford). Lang, A., Leaf, W., and Myers, E. (1883) The Iliad of Homer, Done into English Prose (London). Lardinois, A.P. (1998) ‘How the days fit the works in Hesiod’s Works and Days’, AJPh 119: 319 36.

Bibliography

429

Latham, D.C. (1910) Letters on Church and State of W.E. Gladstone (London). Latham, R.G. (1841) The English Language (London). Laurent, P.E. (tr.) (1846) The Nine Books of the History of Herodotus (Oxford). Ledger Lomas, M. (2013) ‘Theology, divinity, and sermons’, in Eliot 2013, 2.415 24. Lee, J.A.L. (1969) ‘A note on Septuagint material in the Supplement to Liddell and Scott’, Glotta 47: 234 42. Lee, J.A.L. (1997) ‘Hebrews 5:14 and ἕξις: A history of misunderstanding’, Novum Testamentum 39: 151 76. Lee, J.A.L. (2003) A History of New Testament Lexicography (New York). Lee, J.A.L. (2009) Review of Muraoka, 2009, Bulletin of the International Organization of Septuagint and Cognate Studies 43: 115 25. Lee, J.A.L. (2010a) ‘Auxiliary θέλω’, in T.V. Evans and D.D. Obbink (eds), The Language of the Papyri (Oxford), 15 34. Lee, J.A.L. (2010b) ‘Releasing Liddell Scott Jones from its past’, in Stray 2010a, 119 38. Lee, J.A.L. (2013) ‘Etymological follies: Three recent lexicons of the New Testament’, Novum Testamentum 55: 383 403. Lee, J.A.L. (2017) ‘Review article: The Brill Dictionary of Ancient Greek. New clothes for an old emperor’, Novum Testamentum 59: 415 35. Lee, M.M. (2004) ‘ “Evil wealth of raiment”: Deadly πέπλοι in Greek tragedy’, CJ 99: 253 79. Leeper, A. (1893) ‘Notes on Liddell and Scott’s Lexicon’, CR 7: 71 3. Leerssen, J. (2006) National Thought in Europe (Amsterdam). Lejeune, M. (1987) ‘Elles avaient nom a qi ti ta’, Revue de Philologie 61: 181 4. Lemon, G.W. (1783) English Etymology, or, A Derivative Dictionary of the English Language (London). Leumann, M. (1945) ‘Unregelmäßige griechische Steigerungsformen’, Museum Helve ticum 2: 1 14. Leumann, M. (1950) Homerische Wörter (Basel). Lewandowska Tomaszczyk, B. (2007) ‘Polysemy, prototypes and radial categories’, in Cuyckens and Geeraerts 2007, 139 69. Lewis, D.K. (1970) ‘General semantics’, Synthese 22: 18 67. Lewis, T. (2003) ‘Legislating morality: Victorian and modern legal responses to porn ography’, in J. Rowbotham and K. Stevenson (eds) (2003) Behaving Badly: Social Panic & Moral Outrage Victorian & Modern Parallels (Aldershot), 143 58. Liddell, H.G. (1889) An Intermediate Greek English Lexicon (Oxford). Liddell, H.G., Scott, R., and Jones, H.S., (1968) Greek English Lexicon: A Supplement, ed. E.A. Barber, with the assistance of P. Maas, M. Scheller, and M.L. West [LSJ Suppl.] (Oxford). Liddell, H.G., Scott, R., Jones, H.S., and McKenzie, R. (1996) Greek English Lexicon: Revised Supplement, ed. P.G.W. Glare, with the assistance of A.A. Thompson [LSJ Revd Suppl.] (Oxford). Littlebury, I. (tr.) (1818) The History of Herodotus (Oxford). Littlefield, G.E. (1904) Early Schools and School books of New England (Boston). Littré, E. (1839 61) Œuvres complètes d’Hippocrate (Paris). Lloyd, G.E.R. (1987) The Revolutions of Wisdom: Studies in the Claims and Practice of Ancient Greek Science (Berkeley).

430

Bibliography

Lloyd, G.E.R. (2003) ‘The problem of metaphor: Chinese reflections’, in Boys Stones 2003, 101 14. Lloyd Jones, H. (1979) Mention of M. Detienne and J. P. Vernant, Cunning Intelligence in Greek Culture and Society (1978), JHS 99: front matter. Longobardi, G. (2001) ‘Formal syntax, diachronic minimalism, and etymology: The history of French chez’, Linguistic Inquiry 32: 275 302. Louw, J.P. and Nida, E.A. (1988) Greek English Lexicon of the New Testament based on Semantic Domains (New York). Lucy, J. (1992) Language Diversity and Thought: A Reformulation of the Linguistic Relativity Hypothesis (Cambridge). Lust, J., Eynikel, E., and Hauspie, K. (2003) Greek English Lexicon of the Septuagint, rev. edn (Stuttgart). Lyne, R.O.A.M. (1989) Words and the Poet: Characteristic Techniques of Style in Vergil’s Aeneid (Oxford). McKenzie, R. (1925) ‘Etymologies’, CQ 19: 208 10. McKenzie, R. (1928) ‘Οὖλον κεκλήγοντεϛ’, CQ 22: 206. McKenzie, R. (1932a) ‘Bibliography: Graeco Roman Egypt. A: Papyri (1930 1931) 8. Lexicography and grammar’, JEA 18: 101. McKenzie, R. (1932b) ‘A lexicon to Josephus’, CR 46: 76 7. McKenzie, R. (1936) ‘Bibliography: Graeco Roman Egypt Part I: Papyrology (1935) §8. Lexicography and grammar’, JEA 22: 91. McKenzie, R. (1937) ‘Contributions to ‘Bibliography: Graeco Roman Egypt Part I: Papyrology (1936) §8. Lexicography and grammar’, JEA 23: 103 4. McKitterick, D.J. (2003) Print, Manuscript and the Search for Order, 1450 1830 (Cambridge). McMahon, A.M.S. (1994) Understanding Language Change (Cambridge). McMorris, J. (2001) The Warden of English: The Life of H.W. Fowler, Author of Fowler’s Modern English Usage (Oxford). McReady, E.S. (2005) ‘The dynamics of particles’, PhD diss., University of Texas, Austin. McReady, E.S. (2012) ‘Formal approaches to particle meaning’, Language and Lin guistics Compass 6: 777 95. McWhorter, J.H. (2014) The Language Hoax: Why the World Looks the Same in any Language (New York). Mackridge, P. (2009) Language and National Identity in Greece, 1766 1976 (Oxford). Madan, F. (1920) ‘Some experiences of a bibliographer’, The Library 4th ser., 1: 129 40. Madden, D.H. (1897) The Diary of Master William Silence: A Study of Shakespeare & of Elizabethan Sport (London). Maehler, H. (ed.) (2004) Bacchylides: A Selection (Cambridge). Mallory, J. P. and Adams, D.Q. (1997) Encyclopedia of Indo European Culture (London). Marquardt, P. (1993) ‘Penelope as weaver of words’, in M. DeForest (ed.), Woman’s Power, Man’s Game: Essays on Classical Antiquity in Honor of Joy K. King (Wauconda, IL), 149 58. [Marshall, George] (1843) A Lexicon abridged from Liddell and Scott’s Greek English Lexicon (Oxford).

Bibliography

431

[Marshall, George] (1850) A Copious Greek English Vocabulary, Compiled from the Best Authorities (Oxford). [Marshall, George] (1871) A Lexicon abridged from Liddell and Scott’s Greek English Lexicon (Oxford). Martín Zorraquino, M.A. and Portolés Lázaro, J. (1999) ‘Los marcadores del discurso’, in I. Bosque and V. Demonte (eds), Gramática descriptiva de la lengua española (Madrid), 4051 207. Maslov, B. (2015) Pindar and the Emergence of Literature (Cambridge). Mason, M. (1995) The Making of Victorian Sexuality (Oxford). Mathews, C. (1843) The Various Writings of Cornelius Mathews (New York). Matos Amaral, P., Roberts, C., and Smith, E.A. (2007) Review of Potts 2005, Linguistics and Philosophy 30: 707 49. Matthews, D. (2000) The Invention of Middle English: An Anthology of Primary Sources (University Park, PA). Matzner, S. (2016) Rethinking Metonymy: Literary Theory and Poetic Practice from Pindar to Jakobson (Oxford). Mavrogordato, J. (1956) Digenes Akrites (Oxford). Mayrhofer, M. (1986 2001) Etymologisches Wörterbuch des Altindoarischen (Heidelberg). Meier Brügger, M. (1989) ‘Griechisch ὄμφη, ἑάφθη, ἄαπτοϛ und ἑψία’, Münchener Studien zur Sprachwissenschaft 50: 91 6. Meier Brügger, M. (2006) ‘Zur Bildung von griechisch θεόϛ’, Incontri Linguistici 29: 119 25. Meier Brügger, M. (2010) Indogermanische Sprachwissenschaft, 9th edn (Berlin and New York). Meißner, T. (2006) S stem Nouns and Adjectives in Greek and Proto Indo European: A Diachronic Study in Word Formation (Oxford). Melchert, H.C. (1997) ‘Luvian /tāna / “sanctified, inviolable” ’, Historische Sprach forschung 110: 47 51. Melena, J.L. (2014) ‘Mycenaean writing’, in Y. Duhoux and A. Morpurgo Davies (eds), A Companion to Linear B: Mycenaean Greek Texts and their World (Louvain la Neuve and Walpole, MA), 3.1 186. Metzger, B. M. (2000) A Textual Commentary on the Greek New Testament (Stuttgart). Milligan, G. (1922) Here and There among the Papyri (London). Mitteis, L. and Wilcken, U. (1912) Grundzüge und Chrestomathie der Papyruskunde, II. Hälfte: Chrestomathie (Leipzig and Berlin). Momma, H. (2013) From Philology to English Studies: Language and Culture in the Nineteenth Century (Cambridge). Monaghan, M.D. (2001) ‘Coats of many colours: Dyeing and dyeworks in Classical and Hellenistic Greece’, PhD diss., University of Leicester. Montanari, F. (1995, 2nd edn 2004, 3rd edn 2013) Vocabolario della lingua greca [GE] (Turin). Montanari, F. (2015) The Brill Dictionary of Ancient Greek, English edn by M. Goh and C. Schroeder, advisory eds G. Nagy and L. Muellner [GI] (Leiden). Moo, D. (1996) The Epistle to the Romans (Grand Rapids, MI and Cambridge). Moran, M. (2004) ‘ “Light no Smithfield fires”: Some Victorian attitudes to witchcraft’, Journal of Popular Culture 33: 123 51.

432

Bibliography

Moritz, L.A. (1955) ‘Corn’, CQ 5: 135 41. Morpurgo Davies, A. (1999) ‘The morphology of personal names in Mycenaean and Greek: Some observations’, in S. Deger Jalkotzy, S. Hiller, and O. Panagl (eds), Floreant Studia Mycenaea: Akten des X. Internationalen Mykenologischen Colloqui ums in Salzburg vom 1. 5. Mai 1995 (Vienna), 2.389 405. Morris, J. (2004) ‘Text and sacrament: Victorian Broad Church philology’, Studies in Church History 38: 365 74. Mossakowska, M. (2002) ‘Tissus colorés et décorés exportés d’Égypte au premier siècle ap. J. C. (d’après le Periplus Maris Erythraei)’, Topoi 10: 289 314. Mossman, J.M. (1996) ‘Chains of imagery in Prometheus Bound’, CQ 46: 58 67. Moule, C.F.D. (1981) ‘Geoffrey William Hugo Lampe, 1912 1980’, PBA 67: 399 409. Moule, C.F.D. (ed.) (1982) G.W.H. Lampe: Christian, Scholar, Churchman. A Memoir by Friends (London). Moulton, J.H. (1901) ‘Notes from the papyri’, The Expositor 6th ser., 3: 271 82. Moulton, J.H. (1908) A Grammar of New Testament Greek: Volume I, Prolegomena, 3rd edn, with corrections and additions (Edinburgh). Moulton, J.H. and Milligan, G. (1914) The Vocabulary of the Greek Testament illus trated from the Papyri and other Non literary Sources: Part I (London, New York, and Toronto). Moulton, J.H. and Milligan, G. (1915) The Vocabulary of the Greek Testament illus trated from the Papyri and other Non literary Sources: Part II Βάαλ to δωροϕορία (London, New York, and Toronto). Moulton, J.H. and Milligan, G. (1919) The Vocabulary of the Greek Testament illus trated from the Papyri and other Non literary Sources: Part III ἐάν to θώραξ (London, New York, and Toronto). Moulton, J.H. and Milligan, G. (1920) The Vocabulary of the Greek Testament illus trated from the Papyri and other Non literary Sources: Part IV Ἰάειρος to λύω (London, New York, and Toronto). Moulton, J.H. and Milligan, G. (1924) The Vocabulary of the Greek Testament illus trated from the Papyri and other Non literary Sources: Part V μαγεύω to ὀψώνιον (London, New York, and Toronto). Moulton, J.H. and Milligan, G. (1926) The Vocabulary of the Greek Testament illus trated from the Papyri and other Non literary Sources: Part VΙ παγιδεύω to ῥώννυμι (London, New York, and Toronto). Moulton, J.H. and Milligan, G. (1928) The Vocabulary of the Greek Testament illus trated from the Papyri and other Non literary Sources: Part VΙΙ σαβαώθ to Τύχικος (London, New York, and Toronto). Moulton, J.H. and Milligan, G. (1929) The Vocabulary of the Greek Testament illus trated from the Papyri and other Non literary Sources: Part VΙΙΙ ὑακίνθινος to ὠϕέλιμος (London, New York, and Toronto). Moulton, J.H. and Milligan, G. (1930) The Vocabulary of the Greek Testament illus trated from the Papyri and other Non literary Sources (London). Moulton, W.F. and Geden, A.S. (1897) A Concordance to the Greek New Testament according to the Texts of Westcott and Hort, Tischendorf, and the English Revisers (Edinburgh).

Bibliography

433

Mugglestone, L. (2000) ‘ “Pioneers in the untrodden forest”: The New English Dictionary’, in ead. (ed.), Lexicography and the OED (Oxford), 1 21. Mugglestone, L. (2005) Lost for Words: The Hidden History of the Oxford English Dictionary (London). Mugglestone, L. (2007) ‘ “The indefinable something”. Representing rudeness in the English Dictionary’, in M. Gorji (ed.), Rude Britannia (London), 23 34. Mugglestone, L. (2013) ‘ “Life writing”: The lexicographer as biographer in the Oxford English Dictionary’, in R.W. McConchie et al. (eds), Selected Proceedings of the 2012 Symposium on New Approaches in English Historical Lexis (Somerville, MA), 14 26. Mühlestein, H. (1968) ‘Deutung einiger Linear B Wörter’, in A. Bartoněk (ed.), Studia Mycenaea: Proceedings of the Mycenaean Symposium, Brno, April 1966 (Brno), 113 16. Mühlestein, H. (1987) Homerische Namenstudien (Frankfurt am Main). Müller, F.M. (1899) ‘Dean Liddell: As I knew him’, Fortnightly Review new ser., 65: 10 24. Muraoka, T. (2009) A Greek English Lexicon of the Septuagint (Louvain). Murray, G. (1940) Aeschylus the Creator of Tragedy (Oxford). Murray, J. (1967) The Epistle to the Romans: The English Text with Introduction, Exposition and Notes (London and Edinburgh). Murray, O. (1997) ‘The beginnings of Greats, 1800 1872: Ancient History’, in M.G. Brock and M.C. Curthoys (eds), The History of the University of Oxford, vol. 6: Nineteenth Century Oxford, Part I (Oxford), 520 42. Nagy, G. (1990) Greek Mythology and Poetics (Ithaca, NY). Nead, L. (2000) Victorian Babylon (London). Neil, R.A. (1901) The Knights of Aristophanes (Cambridge). Nestle, E. (1898) Novum Testamentum Graece cum apparatu critico (Stuttgart). Nestle, E. and Aland, K. (1963) Novum Testamentum Graece (Stuttgart). Nestle, E., Aland, K., Aland, B., Karavidopoulos, J., Martini, C.M., and Metzger, B.M. (1993) Novum Testamentum Graece (Stuttgart). Nestle, E., Aland, K., Aland, B., Karavidopoulos, J., Martini, C.M., and Metzger, B.M. (2012) Novum Testamentum Graece (Stuttgart). Newman, F.W. (1856) The Iliad of Homer Faithfully Translated into Unrhymed English Metre (London). Nida, E.A. (1964) Towards a Science of Translating (Leiden). Nida, E.A. (2003) The Theory and Practice of Translation (Leiden). Niederwimmer, Κ. (1993) Die Didache, 2nd edn (Göttingen). Nikolaev, A. (2007) ‘The name of Achilles’, in George et al. 2007, 162 73. Nisbet, G. (2004) ‘Hesiod, “Works and Days”: A didaxis of deconstruction?’ Greece & Rome 51: 147 63. Nisetich, F. (1975) ‘Olympian 1.8 11: An epinician metaphor’, HSCPh 79: 55 68. Nockles, P.B. (1997) ‘ “Lost causes and . . . impossible loyalties”: The Oxford Move ment and the University’, in M.G. Brock and M.C. Curthoys (eds) The History of the University of Oxford, vol. 6: Nineteenth century Oxford, Part I (Oxford), 195 267. Nünlist, R. (1998) Poetologische Bildersprache in der frühgriechischen Dichtung (Stuttgart). Oke, C.C. (1957) ‘A suggestion with regard to Romans 8:23’, Interpretation 11: 455 60.

434

Bibliography

Olson, S.D. (2012) The Homeric Hymn to Aphrodite and Related Texts: Text, Trans lation and Commentary (Berlin). Olson, S.D. and Sens, A. (2000) Archestratus of Gela (Oxford). Orrells, D. (2015) Sex: Antiquity and Its Legacy (London). Páez, M. (2012) ‘Bibliography about Greek particles’, Syntaktika 42: 3 43. Paley, F.A. (1883) The Epics of Hesiod (London). Palmer, L.R. (1963) The Interpretation of Mycenaean Greek Texts (Oxford). Pantelia, M.C. (ed.) Thesaurus Linguae Graecae: A Digital Library of Greek Literature [TLG] (Irvine). See online at: http://stephanus.tlg.uci.edu/index.php Panther, K. U. and Thornburg, L.L. (2007) ‘Metonymy’, in Cuyckens and Geeraerts 2007, 236 63. Parry, M. (1971) The Making of Homeric Verse: The Collected Papers of Milman Parry, ed. A. Parry (Oxford). Passow, F. (1812) Über Zweck, Anlage und Ergänzung griechischer Wörterbücher (Berlin). Passow, F. (1816) Grundzüge der griechischen und römischen Litteraturgeschichte (Berlin). Passow, F. (1819) Handwörterbuch der griechischen Sprache, 2nd edn (Leipzig). Passow, F. (1830) Handwörterbuch der griechischen Sprache, 4th edn (Leipzig). Passow, F. (1839) Leben und Briefe, ed. A. Wachler (Breslau). Passow, F. (1841 57) Handwörterbuch der griechischen Sprache, new 5th edn, ed. V. Chr. Fr. Rost, Fr. Palm, et al., 4 vols (Leipzig). Paul, H. (1920) Prinzipien der Sprachgeschichte, 5th edn (Tübingen). Peek, W. (ed.) (1968 75) Lexikon zu den Dionysiaka des Nonnos (Hildesheim). Peláez, J. (1996) Metodología del Diccionario griego español del Nuevo Testamento (Córdoba). Peláez, J. (2014) ‘Contextual factors in the Greek Spanish Dictionary of the New Testament (DGENT)’, in R.A. Taylor and C.E. Morrison (eds), Reflections on Lexicography (Piscataway, NJ), 265 75. Pelliccia, H. (1995) Mind, Body, and Speech in Homer and Pindar (Göttingen). Pellizer, E. (1979) ‘Il fodero e la spada. Metis amorosa e ginecofobia nell’episodio di Circe, Od. X 133 ss.’, Quaderni Urbinati di Cultura Classica 1: 67 82. Peters, Martin (1986) ‘Indogermanische Chronik 32b, No. 727’, Die Sprache 32: 547. Petitmengin, P. (1983) ‘Deux têtes de pont de la philologie allemande en France: le Thesaurus Linguae Graecae et la Bibliothèque des auteurs grecs (1830 1867)’, in M. Bollack and H. Wismann (eds), Philologie und Hermeneutik im 19. Jahrhundert, II, (Göttingen), 76 107. Pfeiffer, R. (1968) History of Classical Scholarship: From the Beginnings to the End of the Hellenistic Age (Oxford). Philological Society (1859) Proposal for the Publication of a New English Dictionary by the Philological Society (London). Pinkerton, J. (1786) Ancient Scottish Poems (London and Edinburgh). Pott, A.F. (1833 6) Etymologische Forschungen auf dem Gebiete der indogermanischen Sprachen (Berlin). Potts, C. (2005) The Logic of Conventional Implicatures (Oxford).

Bibliography

435

Powell, J.E. (1938) A Lexicon to Herodotus, 2nd edn (Cambridge). Prato, C. (1988) ‘Liddell, Scott, Jones s.v. νοσοτυϕέω’, Quaderni Urbinati di Cultura Classica 29.2: 125. Preisigke, F. (1911) Griechische Urkunden des Aegyptischen Museums zu Kairo (Strasbourg). Preisigke, F. (1925 31) Wörterbuch der griechischen Papyrusurkunden (Berlin). Prellwitz, W. (1905) Etymologische Wörterbuch der griechischen Sprache, 2nd edn (Göttingen). Prest, J.M. (1966) Robert Scott and Benjamin Jowett (Oxford). Pucci, P. (1986) ‘Les Figures de la métis dans l’Odyssée’, Mètis 1: 7 28. Race, W.M. (ed. and tr.) (1997) Pindar, Nemean Odes, Isthmian Odes, Fragments (Cambridge, MA). Rahlfs, A. (1914) Verzeichnis der griechischen Handschriften des Alten Testaments (Berlin). Rahlfs, A. (ed.) (1935a) Septuaginta: id est Vetus Testamentum graece iuxta LXX interpretes, vol. 1: Leges et historiae (Stuttgart). Rahlfs, A. (ed.) (1935b) Septuaginta: id est Vetus Testamentum graece iuxta LXX interpretes, vol. 2: Libri poetici et prophetici (Stuttgart). Reece, S. (2009) Homer’s Winged Words: The Evolution of Early Greek Epic Diction in the Light of Oral Theory (Leiden). Revuelta Puigdollers, A.R. (2009) ‘Particles and discourse cohesion in ancient Greek’, in G.K. Giannakis, M. Baltazani, G.I. Xidopoulos, and T. Tsagalidi (eds), Proceedings of the 8th International Conference on Greek Linguistics, University of Ioannina, Greece, August 30th September 2nd 2007 (Ioannina), 394 405. Reynolds, L. and Wilson, N.G. (2013) Scribes & Scholars: A Guide to the Transmission of Greek & Latin Literature, 4th edn (Oxford). Richards, G.C. (1925) ‘Review: A Greek English Lexicon’, Journal of Theological Studies 27: 76 7. Ricoeur, P. (1978) The Rule of Metaphor, tr. R. Czerny (London). Rijksbaron, A. (ed.) (1997) New Approaches to Greek Particles: Proceedings of the Colloquium Held in Amsterdam, January 4 6, 1996, to Honour C.J. Ruijgh on the Occasion of His Retirement (Amsterdam). Riplinger, G.A. (2008) Hazardous Materials: Greek and Hebrew Study Dangers (Ararat, VA). Risch E. (1974) Wortbildung der homerischen Sprache, 2nd edn (Berlin). Risch, E. (1987) ‘Die älteste Zeugnisse für κλέος ἄϕθιτον’, Zeitschrift für Vergleichende Sprachforschung 100: 3 11. Roberts, C. (2012) ‘Information structure: Towards an integrated formal theory of pragmatics’, Semantics & Pragmatics 5: 1 69. Roberts, T. (2004) ‘A Review of BDAG’, in Taylor et al. 2004, 53 65. Robinson, M.A. and Pierpont, W.G. (2005) The New Testament in the Original Greek: Byzantine Textform, 2nd edn (Southborough, MA). Robson, J. (2006) Humour, Obscenity and Aristophanes (Tübingen). Rolfe, J.C. (tr.) (1913) Suetonius, Lives of the Caesars, 3 vols (Cambridge, MA). Rooth, M.E. (1985) ‘Association with focus’, PhD diss., University of Massachusetts, Amherst.

436

Bibliography

Rooth, M.E. (1992) ‘A theory of focus interpretation’, Natural Language Semantics 1: 75 116. Rooth, M.E. (1996) ‘Focus’, in S. Lappin (ed.), The Handbook of Contemporary Semantic Theory (Oxford), 271 97. Rowe, C. (2007) Plato and the Art of Philosophical Writing (Cambridge). Ruffell, I. (2012) ‘ “Seeing the meat for what it is”: Aristophanic expurgation and its phallacies’, in Harrison and Stray 2012a, 25 51. Ruijgh, C.J. (2004) ‘The stative value of the PIE verbal suffix * éh₁ ’, in J.H.W. Penney (ed.), Indo European Perspectives: Studies in Honour of Anna Morpurgo Davies (Oxford), 48 64. Russell, D.A. (1991) An Anthology of Greek Prose (Oxford). Rutherford, R.B. (2012) Greek Tragic Style (Cambridge). St. John Stevas, N. (1956) Obscenity & the Law (London). Sanders, H.A. (1912) The Washington Manuscript of the Four Gospels (New York). Schmid, H.J. (2007) ‘Entrenchment, salience and basic levels’, in Cuyckens and Geeraerts 2007, 117 38. Schmidt, M. (2006) ‘Die Welt des Eumaios’, in A. Luther (ed.), Geschichte und Fiktion in der homerischen Odyssee (Munich), 117 38. Schoch, R.W. (1998) Shakespeare’s Victorian Stage: Performing History in the Theatre of Charles Kean (Cambridge). Schubart, W. (1919) Aegyptische Urkunden aus den Staatlichen Museen zu Berlin. Griechische Urkunden: V, Der Gnomon des Idios Logos. Erster Teil: Der Text (Berlin). Schwyzer, E. (1988) Griechische Grammatik: Syntax und syntaktische Stilistik, 5th edn, vol. 2 (Munich). Scott, J.A. (1920 1) ‘The meaning of the verb βάπτω, βαπτίζω’, CJ 16: 53 4. Scott, R. (1839) A Sermon Preached on Sunday, Jan. 20, 1839, in the Parish Church of Holywell, Oxford, in behalf of the Society for the Propagation of the Gospel in Foreign Parts (Oxford). Scott, R. (1841) Christian Quietness: A Sermon (London and Oxford). Scott, R. (1851) Twelve Sermons (London). Scott, W. (1903) Ivanhoe; edited for School Use by W.E. Simonds (Chicago). Scott, W. (2000 [1820]) Ivanhoe, ed. G. Tulloch (London). Scully, S. (2015) Hesiod’s Theogony: From Near Eastern Creation Myths to Paradise Lost (New York and Oxford). Sherry, L.F. (1996) ‘The Paraphrase of St. John attributed to Nonnus’, Byzantion 66: 409 30. Shipp, G.P. (1979) Modern Greek Evidence for the Ancient Greek Vocabulary (Sydney). Sicking, C.M.J. and van Ophuijsen, J.M. (1993) Two Studies in Attic Particle Usage: Lysias and Plato (Leiden). Silk, M.S. (1974) Interaction in Poetic Imagery with Special Reference to Early Greek Poetry (Cambridge). Silk, M.S. (1983) ‘LSJ and the problem of poetic archaism: From meanings to iconyms’, CQ 33: 303 30. Silk, M.S. (1998) ‘Pindar’s poetry and the obligatory crux: Isthmian 5.56 63, text and interpretation’, TAPhA 128: 25 88.

Bibliography

437

Silk, M.S. (2003a) ‘Metaphor and metonymy: Aristotle, Jakobson, Ricoeur, and others’, in Boys Stones 2003, 115 47. Silk, M.S. (2003b) ‘Metaphor and simile’, in S. Hornblower and A. Spawforth (eds), The Oxford Classical Dictionary, 3rd edn (Oxford), 966 8. Silk, M.S. (2009) ‘The invention of Greek’, in A. Georgakopoulou and M.S. Silk (eds), Standard Languages and Language Standards: Greek, Past and Present (Farnham), 3 31. Silk, M.S. (2014) ‘The language of Greek lyric poetry’, in E.J. Bakker (ed.), A Companion to the Ancient Greek Language, rev. edn of Bakker 2010 (Chichester), 424 40. Simons, M., Tonhauser, J., Beaver, D., and Roberts, C. (2011) ‘What projects and why’, in N. Li and D. Lutz (eds), Proceedings of the 20th Semantics and Linguistic Theory Conference, April 29 May 1, 2010 in Vancouver, British Colombia (Ithaca, NY), 309 27. Simpson, J.A. and Weiner, E.S.C. (1989) The Oxford English Dictionary, 2nd edn [OED] (Oxford). Online (with some 3rd edn entries) at www.oed.com/ Skeat, T.C. (1938) ‘The lilies of the field’, Zeitschrift für die Neutestamentliche Wissenschaft 37: 211 14. Skoda, F. (1988) Médecine ancienne et métaphore. Le vocabulaire de l’anatomie et de la pathologie en grec ancien (Paris). Slater, W.J. (1969) Lexicon to Pindar (Berlin). Slobin, D.I. (2003) ‘Language and thought online: Cognitive consequences of linguistic relativity’, in D. Gentner and S. Goldin Meadow (eds), Language in Mind: Advances in the Study of Language and Thought (Cambridge, MA), 157 91. Smyth, H.W. (1920) Greek Grammar (Cambridge, MA). Smyth, H.W. (1956) Greek Grammar, rev. G.M. Messing (Cambridge, MA). Snell, B. et al. (1955 2010) (eds), Lexikon des frühgriechischen Epos [LfrgE] (Göttingen). Solmsen, F. (1990) Hesiod, Theogonia, Opera et Dies, Scutum, Fragmenta Selecta, 3rd edn (Oxford). Soltic, J. (2014) ‘The particle γάρ: From ancient Greek sentence connector to blatant line filler? A case study on the Late Medieval Greek Chronicle of Morea’, Symbolae Osloenses 88: 136 47. Sommerstein, A. (ed. and tr.) (2008) Aeschylus, Persians, Seven Against Thebes, Suppliants, Prometheus Bound (Cambridge, MA). Sommerstein, A. (2010) ‘Aeschylus’ epitaph’, in id., The Tangled Ways of Zeus and Other Studies in and around Greek Tragedy (Oxford), 195 9. Sommerstein, A. (2012) ‘Notes on Aristophanes’ Frogs’, in C.W. Marshall and G. Kovacs (eds), No Laughing Matter: Studies in Athenian Comedy (London), 115 25. Sophocles, E.A. (1870) A Greek Lexicon of the Roman and Byzantine Periods (from B.C. 146 to A.D. 1100) (Boston). Sophocles, E.A. (1900) A Greek Lexicon of the Roman and Byzantine Periods (from B.C. 146 to A.D. 1100), ed. J.H. Thayer and H. Drisler (New York). Souter, A. (1910) Novum Testamentum Graece (Oxford). Souter, A. (1916) A Pocket Lexicon to the Greek New Testament (Oxford). Souter, A. (1926 7) ‘ἀγαπητός’, Journal of Theological Studies 28: 59 60. Stafford, E. (2000) Worshipping Virtues: Personification and the Divine in Ancient Greece (London).

438

Bibliography

Stanford, W.B. (1958) The Odyssey of Homer: Books I XII (London and New York). Stark, R. (2012) Ezra Pound’s Early Verse and Lyric Tradition (Edinburgh). Starr, J. (1937) ‘The Byzantine inscriptions of Bethshan Scythopolis’, AJPh 58: 83 9. Stolk, J.V. (2015) ‘Dative by genitive replacement in the Greek language of the papyri: A diachronic account of case semantics’, Journal of Greek Linguistics 15: 9 121. Stolk, J.V. (2017a) ‘Dative and genitive case interchange in Greek papyri from Roman Byzantine Egypt’, Glotta 93: 182 212. Stolk, J.V. (2017b) ‘Dative alternation and dative case syncretism in Greek: The use of dative, accusative and prepositional phrases in documentary papyri’, TPhS 115: 212 38. Storey, I.C. (2003) Eupolis: Poet of Old Comedy (Oxford). Stray, C.A. (ed.) (1996) The Mushri English Pronouncing Dictionary: A Chapter in 19th century English Public School Lexicography (Reading). Stray, C.A. (2007a) Classical Books: Scholarship and Publishing in Britain since 1800 (BICS Supplement 101, London). Stray, C.A. (ed.) (2007b) Oxford Classics: Teaching and Learning 1800 2000 (London). Stray, C.A. (2007c) ‘Sir William Smith and his dictionaries: A study in scarlet and black’, in Stray 2007a, 35 54. Stray, C.A. (ed.) (2010a) Classical Dictionaries: Past, Present and Future (London). Stray, C.A. (2010b) ‘Liddell & Scott: Myths & markets’, in Stray 2010a, 94 118. Stray, C.A. (2010 11) ‘Reading silence: Books that never were’, Hyperboreus 16 17: 527 38. Stray, C.A. (ed.) (2013a) Sophocles’ Jebb: A Life in Letters (CCJ Supplement 38, Cambridge). Stray, C.A. (2013b) ‘Classics’, in Eliot 2013, 2.435 70. Stray, C.A. (2016) ‘A semi sacred monster: Charles Wordsworth’s Graecae Gramma ticae Rudimenta’, Philologia Classica 11: 98 115. Stray, C.A. (2018) Classics in Britain 1800 2000: Scholarship, Education, and Publish ing (Oxford). Stray, C.A. (forthcoming 2020) ‘Classical communities’, in Harrison and Pelling forthcoming 2020. Stuart Jones, T. (2012) Poetic Language (Edinburgh). Sutcliffe, P. (1978) Oxford University Press: An Informal History (Oxford). Szabolcsi, A. (2015) ‘What do quantifier particles do?’, Linguistics and Philosophy 38: 159 204. Szemerényi, O. (1960) ‘The origin of Greek βαθύς and βόθυνος’, Glotta 38: 211 16. Taubenschlag, R. (1959) Opera minora (Warsaw). Taylor, B.A., Lee, J.A.L., Burton, P.R., and Whitaker, R.E. (eds) (2004) Biblical Greek Language and Lexicography: Essays in Honour of Frederick W. Danker (Grand Rapids, MI and Cambridge). Taylor, G. (1990) Reinventing Shakespeare: A Cultural History from the Restoration to the Present (London). Taylor, J.R. (1995) Linguistic Categorization: Prototypes in Linguistic Theory, 2nd edn (Oxford). Taylor, J.R. (2002) Cognitive Grammar (Cambridge). Tcherikover, V. and Fuks, A. (1960) Corpus Papyrorum Judaicarum, vol. 2 (Cambridge, MA).

Bibliography

439

Thackeray, H. St J. (1909) A Grammar of the Old Testament in Greek according to the Septuagint, Volume I: Introduction, Orthography and Accidence (Cambridge). Thiselton, A.C. (2000) The First Epistle to the Corinthians: A Commentary on the Greek Text (Grand Rapids, MI and Cambridge). Thompson, A.A. (2015) ‘Modern Greek in LSJ and the Revised Supplement’, in J.K. Aitken and T.V. Evans (eds), Biblical Greek in Context: Essays in Honour of John A.L. Lee (Leuven), 207 39. Thompson, A.A. (2017) ‘The lexicographic editor and the problem of consistency’, in T. Li. and K. Dyer (eds), From Ancient Manuscripts to Modern Dictionaries: Select Studies in Aramaic, Hebrew and Greek (Piscataway, NJ), 481 507. Thompson, H.L. (1899) Henry George Liddell, Dean of Christ Church, Oxford: A Memoir (London). Thumb, A. (1910) Handbuch der neugriechischen Volkssprache, 2nd edn (Strasbourg). Tischendorf, C. von (1843) Codex Ephraemi Syri rescriptus, sive, Fragmenta Novi Testamenti e codice Graeco Parisiensi celeberrimo quinti ut videtur post Christum seculi (Leipzig). Tischendorf, C. von (1849) Novum Testamentum Graece (Leipzig). Tischendorf, C. von (1869 72) Novum Testamentum Graece: Editio octava critica maior (Leipzig). Tischendorf, C. von, Gregory, C.R., and Abbot, E. (1884, 1890, 1894) Novum Testa mentum Graece: Editio octava critica maior. Prolegomena (Leipzig). Tonhauser, J. (2012) ‘Diagnosing (not )at issue content’, in E. Bogal Allbritten (ed.), Proceedings of Semantics of Under represented Languages of the Americas (SULA) 6 (Amherst, MA), 239 54. Toosarvandani, M. (2010) ‘Scalar reasoning and the semantics of let alone’, in M. Bane, J. B. Holle, T. Grano, A.L. Grotberg, and Y. McNabb (eds), Papers from the Forty Fourth Regional Meeting of the Chicago Linguistic Society (Chicago), 51 64. Trapp, E. and Hörandner, W., et al. (1994 2017) (eds) Lexikon zur byzantinischen Gräzität (Vienna). Online at http://stephanus.tlg.uci.edu/lbg/#eid=1&context=lsj Trapp, J.B. (1971) ‘The conformity of Greek and the vernacular: The history of a Renaissance theory of languages’, in R.R. Bolgar (ed.), Classical Influences on European Culture AD 500 1500 (Cambridge), 239 44. Traugott, E.C. and Dasher, R.B. (2005) Regularity in Semantic Change (Cambridge). Trédé, M. (1992) Kairos: l’à propos et l’occasion, le mot et la notion, d’Homère à la fin du IVe siècle avant J. C. (Paris). Tregelles, S.P. (1857 72) The Greek New Testament edited from the Ancient Authorities, with their Various Readings in Full, and the Latin Version of Jerome (London). Trench, R.C. (1853) On the Study of Words (London). Trench, R C. (1854) Synonyms of the New Testament, 2nd edn (Cambridge). Trench, R.C. (1857) On Some Deficiencies in Our English Dictionaries (London). Trench, R.C. (1860) On Some Deficiencies in Our English Dictionaries, 2nd edn (London). Trench, R.C. (1876) Synonyms of the New Testament, 8th edn (London). Trenchard, W.C. (2003) A Concise Dictionary of New Testament Greek (Cambridge).

440

Bibliography

Triantafyllidis Lexicon (1998) Λεξικό της κοινής Νεοελληνικής (το Ινστιτούτο Νεοελληνικών Σπουδών. Ίδρυμα Μανόλη Τριανταϕυλλίδη) (Thessaloniki). Online at http://www.greek language.gr/greekLang/modern greek/tools/lexica/triantafyllides/ Tsagalis, C. (2008) The Oral Palimpsest (Washington, DC). Turcan, R. (1976) Review of Detienne and Vernant 1974, Revue de l’Histoire des Religions 189: 223 5. Turner, N. (1976) A Grammar of New Testament Greek (J.H. Moulton). Volume IV: Style (Edinburgh). Turner, S. (1799 1805) The History of the Anglo Saxons (London). Van Noorden, H. (2014) Playing Hesiod: The ‘Myth of the Races’ in Classical Antiquity (Cambridge). Van Windekens, A.J. (1986) Dictionnaire étymologique complémentaire de la langue grecque: nouvelles contributions à l’interprétation historique et comparée du vocabu laire (Louvain). Veitch, W. (1866) Greek Verbs, Irregular and Defective: Their Forms, Meaning, and Quantity; Embracing all the Tenses used by the Greek Writers, with References to the Passages in which they are Found (Oxford). Ventris, M. and Chadwick, J. (1956) Documents in Mycenaean Greek (Cambridge). Ventris, M. and Chadwick, J. (1973) Documents in Mycenaean Greek, 2nd edn (Cambridge). Verdenius, W.J. (1985) A Commentary on Hesiod: Works and Days, Vv. 1 382 (Leiden). Vergados, A. (2012) The ‘Homeric Hymn to Hermes’: Introduction, Text and Com mentary (Berlin). Verhoogt, A.M.F.W. (1998) Menches, Komogrammateus of Kerkeosiris: The Doings and Dealings of a Village Scribe in the Late Ptolemaic Period (120 110 B.C.) (Leiden). Vernant, J. P. (1990) ‘The myth of Prometheus in Hesiod’, in id., Myth and Society in Ancient Greece (New York), 183 201 (first published 1974). Vilborg, E. (1960) A Tentative Grammar of Mycenaean Greek (Göteborg and Stockholm). Vine, B. (2014) ‘Grassmann’s Law’, in Giannakis 2014, 1.394 5. Vittiglio, N.A. (2017) Il lessico miceneo riferito ai cereali (Florence). von Bäumlein, W. (1861) Untersuchungen über griechische Partikeln (Stuttgart). Vorster, W.S. (1998) ‘In what sense is the Louw/Nida dictionary authoritative?’, in id. and J.E. Botha (ed.), Speaking of Jesus: Essays on Biblical Language, Gospel Narrative, and the Historical Jesus (Leiden), 37 51. Wagner, G. (1987) Les Oasis d’Égypte à l’époque grecque, romaine et byzantine d’après les documents grecs (Paris). Wakker, G.C. (1997) ‘Emphasis and affirmation: Some aspects of μήν in tragedy’, in Rijksbaron 1997, 209 31. Walcot, P. (1976) Review of Detienne and Vernant 1974, JHS 96: 212 13. Walker Gore, C. (2015) ‘Crafty women and artful misses: Femininity, feminism and female crafts in the nineteenth century novel’, https://victorianist.wordpress.com/ 2015/06/01/crafty women and artful misses femininity feminism and female crafts in the nineteenth century novel/.

Bibliography

441

Watkins, C. (1979) ‘NAM.RA GUD UDU in Hittite: Indo European poetic language and the folk taxonomy of wealth’, in W. Meid and E. Neu (eds), Hethitisch und Indogermanisch: Vergleichende Studien zur historischen Grammatik und zur dia lektgeographischen Stellung der indogermanischen Sprachgruppe Altkleinasiens (Innsbruck), 269 87. Watkins, C. (1995) How to Kill a Dragon: Aspects of Indo European Poetics (New York and Oxford). Watson, F. (1908) The English Grammar Schools to 1660: Their Curriculum and Practice (Cambridge). Wayte, W. (1894) ‘Corrections for Liddell and Scott’s Lexicon’, CR 8: 146. Webster, N. (1833) A Dictionary for Primary Schools, 11th edn (New York). Weiss, M. (1996) ‘Greek μυρίοϛ “countless”, Hittite mūri “bunch of fruit” ’, Historische Sprachforschung 109: 199 214. West, M.L. (1966) Hesiod, Theogony (Oxford). West, M.L. (1978) Hesiod, Works and Days (Oxford). West, M.L. (2001) Studies in the Text and Transmission of the Iliad (Berlin). West, M.L. (2008) ‘The Hesiod papyri and archaic epic language’, in G. Bastianini and A. Casanova (eds), Esiodo: cent’anni di papiri. Atti del convegno internazionale di studi (Florence), 29 42. West, M.L. (2012) ‘Towards a chronology of early Greek epic’, in Andersen and Haug 2012, 224 41. Westcott, B.F. and Hort, F.J.A. (1881) The New Testament in the Original Greek (Cambridge). Wevers, J.W. (1990) Notes on the Greek Text of Exodus (Atlanta). Wevers, J.W. (1991) Exodus: Vetus Testamentum graecum auctoritate Academiae Scientiarum Gottingensis editum (Göttingen). Wharton, E.R. (1890) Etyma Graeca (London). Whatmough, J. (1937) ‘Review of A Greek English Lexicon, Part VIII’, CPh 32: 168 70. Whatmough, J. (1938) ‘Review of A Greek English Lexicon, Part IX’, CPh 33: 233 6. Whatmough, J. (1942) ‘Review of LSJ 1940’, CR 37: 96 8. Wheeler, C.S. (ed.) (1842) Herodotus, from the Text of Schweighäuser, with English Notes (Boston). Widmer, P. (2009) ‘Funktion und Syntax informationsstrukturierender Partikeln im Hethitischen und Britannischen’, Habilitationsschrift, Philipps Universität Marburg. Wierzbicka, A. (1997) Understanding Cultures through Their Key Words: English, Russian, Polish, German, and Japanese (Oxford). Wifstrand, A. (1963) ‘The Lexicon of Patristic Greek (α βαραθρόω)’, Journal of Theo logical Studies 14: 400 6. Wifstrand, A. (1964) ‘A Patristic Greek Lexicon. Edited by G. W. H. Lampe. Fasc. 2 (βαρβαρεύω εὐσυμπαθήτως)’, Journal of Theological Studies 15: 397 402. Wilkins, J.M. and Hill, S. (2006) Food in the Ancient World (Oxford). Williams, P.J. (2012) ‘The Bible, the Septuagint, and the Apocrypha: A consideration of their singularity’, in G. Khan and D. Lipton (eds), Studies on the Text and Versions of the Hebrew Bible in Honour of Robert Gordon (Leiden), 169 80. Williams, R.F. (1838) Shakespeare and his Friends; or, The Golden Age of Merrie England (Paris).

442

Bibliography

Willinsky, J. (1994) Empire of Words: The Reign of the OED (Princeton). Winkler, J. (1990) ‘Penelope’s cunning and Homer’s’, in id.,The Constraints of Desire (New York), 129 61. Winter, B.W. (1998) ‘Puberty or passion? The referent of ΥΠΕΡΑΚΜΟΣ in 1 Corinthians 7:36’, Tyndale Bulletin 49: 71 89. Wodtko, D.W., Irslinger, B., and Schneider, C. (2008) Nomina im indogermanischen Lexikon (Heidelberg). Wolf, F.A. (1795) Prolegomena ad Homerum, sive de Operum Homericorum prisca et genuina forma variisque mutationibus et probabili ratione emendandi (Berlin). Wolf, F.A. (1986) Prolegomena to Homer, 1795, tr. A. Grafton, G. Most, and J.E.G. Zetzel (Princeton). Wright, F.A. (1932) A History of Later Greek Literature from the Death of Alexander in 323 B.C. to the Death of Justinian in 565 A.D. (London). Wright, M. (2012) The Comedian as Critic: Greek Old Comedy and Poetics (London). Wright, N.T. (2005) Paul: Fresh Perspectives (London). Zgusta, L. (2006) Lexicography Then and Now: Selected Essays, ed. F. Dolezal and T. Creamer (Tübingen). Zimmermann, B. (1992) Dithyrambos: Geschichte einer Gattung (Göttingen). Zimmermann, M. (2011) ‘Discourse particles’, in K. von Heusinger, C. Maienborn, and P. Portner (eds), Semantics: An International Handbook of Natural Language Meaning (Berlin), 2.2011 38. Zobel, S. (2015) ‘On the distribution of German discourse particles across types of questions’, Handout, talk presented at the University of Vienna, March. Zorell, F. (1911) Lexicon Graecum Novi Testamenti (Paris). Zucker, A. (2005) Aristote et les classifications zoologiques (Louvain la Neuve).

Greek Index Alphabetic Greek ἆ 342, 343 4 ἀάω 349, 352 ἀβοατί 202, 203 Ἀβοριγῖνεϛ 49 ἀγαθόϛ 126 9, 139 ἀγάλακτοϛ 60 ἀγαπητόϛ 170 ἄγγελοϛ 311 ἄγγοϛ 143 ἄγκαθεν 55 ἁγνευτήριον 154 ἄγνοϛ 55 ἄγριοϛ 398 ἀγχόνη 405 ἀγχυλομήτηϛ 227, 234 ἄγχω 404 5 ἄγω 92, 102 ἁδινόϛ 318 ἀδολέσχηϛ 129 30 ἀδολεσχία 129 30 ἄεθλοϛ 213, 215, 217 ἀείδω 89, 335 ἄζω 344 ἆθλοϛ 213, 215, 217 αἰαῖ 342 αἰανήϛ 318 αἰδέομαι 346 αἰδοιολείκτηϛ 57, 70 αἰδοῖοϛ 342 3, 344 6 αἴδομαι 346 αἰδόφρων 346 αἰδώϛ 249, 342, 345, 346 αἶθοψ 318 ἀϊκή 349 αἷμα 312 αἱματόειϛ 312 13 αἰολόμητιϛ 227, 234 αἰόλοϛ 238 ἀΐσσω 342 3, 347 9 αἰσχροποιΐα 56, 70 αἰσχρότηϛ 70 αἰσχρουργέω 57 αἰών 249 52, 258 ἄκρατον 196 ἄλγοϛ 90 ἀλέξανδροϛ 100 ἀλίαστοϛ 113, 318 ἀλλοῖοϛ 328

ἄλοχοϛ 23 ἄλσοϛ 258 60 ἀλφηστήϛ 110 12 ἅμα 94 ἀμαιμάκετοϛ 318 ἀμαυρόω 121 22 ἀμέιβω 55, 59 ἄμπωτιϛ 329 ἀμφίπολιϛ 54 5 ἄν 53 ἀν(ά) 201 2, 340 ἀναγράφω 201 ἀναίδεια 346 ἀναιδήϛ 346 ἀνάλωμα 202 ἄναξ 102, 317 ἀναξιφόρμιγγεϛ 317 ἀναπίπτω 329 ἀνατ(ε)ί 352 ἄνατοϛ 352 ἀνατρέπω 201 ἀναφλασμόϛ 70 ἀναψύχω 201, 203 ἀνδρεία 75 ἀνδρεῖοϛ 206, 207 ἀνδροβατέω 56 ἀνδράποδα/ ον 291 2 ἀνέλεοϛ 174 ἀνθέω 329 ἄνθοϛ 311 ἄνθρωποϛ 288, 295, 300, 301 2 ἀνίλεωϛ 174 ἀνοίκτωϛ 313 14 ἀνταγείρω 154 ἀνταποπέρδω 56 ἀντιμετρέω 173 ἀντιοχεύομαι 57 ἄνω 147 ἀορτήρ 33 4 ἀπαρχή 164 6 ἄπειμι 102 ἀπεκρίθην 177 ἀπεχθάνομαι 216 ἀπηλιώτηϛ 50 ἀπηύρα 117 18 ἄπλαστοϛ 109 10 ἀπό 390 ἀποβάπτω 355, 390 3 ἀπόλλυμι 301

444 ἄποροϛ 238 ἀργαλέοϛ 40 Ἀργεϊφόντηϛ 100 ἄρδομαι 329 ἄρθρον 145 ἄρκτοϛ 48 ἁρμόϛ 145 ἀρραβών 165 ἄρρην/ἄρσην 177, 203 5 ἀρτέμων 50 ἄρτοϛ 183, 187, 192 9 ἀρχάγγελοϛ 156 ἀρχαιομελισιδωνοφρυνιχήρατα 407 ἀρχιεπίσκοποϛ 154 ἀρχιέρεια 52 ἀρχιεταῖροϛ 171 2 ἄσπροϛ 183 ᾄσσω/ᾄττω 347 ἀσφόδελοϛ 115 16 ἀτάομαι 352 ἅτε 94 ἄτη 342 3, 349 52 ἀτηρία 352 ἀτηρόϛ 352 ἀτταταῖ 342 ἀυάτα 349 αὐτόϛ 131 2, 139 αὐξάνω 177 αὔξω 177 αὐχήν 179 ἄφθιτοϛ 100 ἀφίημι 102 ἄωτον/ϛ 318 Ἀχαιόϛ 89 90 ἄχθομαι 404 Ἀχιλλεύϛ 89 ἄχοϛ 404 5 ἄχω 404 βαθύϛ 364, 367, 368 βαίνω 95 βάλανοϛ 78 βαπτίζω 162, 365, 389 βαπτόϛ 389 βάπτω 353 94 βαρύϛ 148, 334 βαφή 389 βένθοϛ 306 βίη 232 3, 235, 237, 239 βινέω 57, 72 βλοσυρόϛ 130 1 βόσκω 297 βοτόν 297 βούκα 196 7 βουκάκρατον 195 6 βουκελλάριοϛ 196 7 βουκία 197

Greek Index βουκί(ο)ν 197 βουλή 242 βούλομαι 95, 289 βοῦϛ 291 2 βροτόϛ 300, 301 2 γάρ 278 γε 268 87 γένοϛ 132, 137, 138, 139 γλοιόϛ 40 1 γλωσσόκομοϛ 76 γλῶττα 325 γλωττοδεψέω 56 γονή 148 γοῦν 278 γυναικεῖα 148 δαιδάλλω 38 9 δαφοινόϛ 323 5 δέ 271 δέδηε 318 δέμω 52 δεσπότηϛ 101 δευτερόπρωτον 173 4 δέφω 71, 72, 74, 80 δηλονότι 178 διαίρεσιϛ 132, 137, 139 δίαιτα 147 διαλέγομαι 132 3 διαλογισμόϛ 171 διαπτύω 52 δίκη/Δίκη 119 20 δίπουϛ 292 δίψιοϛ 55 δμῳή 335 δμώϛ 335 δοκεύω 237 δοκέω 263 δολόμητιϛ 234 δόλοϛ 227, 234, 237, 240 δόξα 262 3 δούλιοϛ 335 δοῦλοϛ 104, 334 5 δούξ 49 δρῖλοϛ 74 δρόσοϛ 148 δρῶπαξ 50 δυσμενήϛ 254 ἔασσα 102 ἑάφθη 335 ἐγκυκλόω 238 ἐγχέζω 57 ἐγώ 92 ἔγωγε 278 9 ἔδομαι/ἔδω 92 ἐἕ 342 ἐθέλω 177, 289 εἶδοϛ 133, 137 8, 139

Greek Index εἰλεόϛ 146 εἰμί 102 εἰρωνεία 134 εἰϛ 93 ἐκμιαίνω 70 ἐκτρίβω 52 ἐκχέζω 57 ἔλαφοϛ 293 ἐλαφρόϛ 148 ἐλπίϛ 118 19 ἐμβάπτω 355 ἔμβλημα 50 ἐν 93 ἐναίσιμοϛ 58 9 ἔνδον 315 16 ἕννυμι 207 8 ἐντονία 57 ἔξω 314 16 ἕξω 101 ἐξωτέρω 314 ἐπαιδέομαι 346 ἐπαυρέω 117 18 ἐπέρχομαι 306, 309 10 ἐπερώτημα 155, 160 1, 172 ἐπίκλοποϛ 38 9 ἐπιλοιβή 216 ἐπιούσιοϛ 155 ἐπιρρέω 306, 311 ἐπίσκοποϛ 162 ἐπιχείρησιϛ 175 ἕπομαι 336 ἕπω 336 ἐπωμίϛ 145 ἔραμαι 329 ἐράω 329 ἐρευτάϛ 102 ἔριϛ 119 ἑρπετόν 295, 296 ἕρπω 296 ἐρυθρόϛ 249 ἔρχομαι 85 Ἐρχομενόϛ 100 ἐϋκτίμενοϛ 220 1 εὐμενήϛ 254 εὐρύπρωκτοϛ 56 ευϛ 101 εὐσχεθήϛ 109 εὐφροσύνη/ἐϋφροσύνη 213, 215, 217 εὔχομαι 94 ἐχθάνομαι 216 ἐχθέϛ 177 ἔχω 101 ζῷον 288 9, 294 5 ζώϛ 295 ἦμαρ 101 ἥρωϛ 249

θάλασσα/θάλαττα 203 θέλεοϛ 52 θεά 89, 91 θεάομαι 93 θέλω 177 θεόϛ 89, 91, 103 θερμοπώλιον 51 θήρ 293 4 θηρίον 293 4 θρέμμα 297 θυμόω 59 θώρηξ 33 4 ἰδέα 134 5, 137 8, 139 ἴδιοϛ 209 ἵημι 102 Ἰησοῦϛ 156 60 ἰκμάϛ 146 ἰλεόϛ 146 ἵνα 177 ἴον 50 ἵππαρχοϛ 52 ἵπποϛ 291 2 ἰσαίων 109 ἵστημι 48 ἰσχυρόϛ 148 ἰώ 342 καίριοϛ 149 καιρόϛ 149, 238 καίω 147 κακκάω 57 κακοστόματοϛ 56 κανονίζω 31 κάπηλοϛ 25, 30 1, 41 2 κάπνοϛ 31 καρταίποδα 290, 292 κατά 95 καταιδέομαι 346 κατάληψιϛ 50 κατάρροοϛ 146 κατασπιλάζω 167 κάτω 147 καυλόϛ 81 κέρδοϛ 242 κεφαλαιόω 174 κλάσμα 195 κλέοϛ ἄφθιτον 100 κνάω 297 κνώδαλον 296 7 κνώδαξ 296 κνώδων 296 κοινόϛ 209 25 κομήειϛ 102 κονδύλωμα 147 κονδύλωσιϛ 147 κόποϛ 147 κορυθάϊξ 349

445

446 κράβακτοϛ 175 κράββατοϛ 175 κρασί 183 4 κράτοϛ 239 κρημνόϛ 148 κριθή 188 κριθόπυρον 190 κτάομαι 297 κτῆνοϛ 297 κτίζω 297 κτίσιϛ 297 κτίσμα 297 κτίτηϛ 101 κτοίνα 101 κυνῶπιϛ 262 κύσθοϛ 72, 74, 80 κύων 249, 262 3 Λαέρτηϛ 100 λαμπρόϛ 310, 316 17 Λαόδοκοϛ 100 λαόϛ/λεώϛ 100, 203 4 λείκτηϛ 70 λειριόειϛ 318 λείριοϛ 318 Λεσβιάζω 56 λέων 293 λῆμμα 50 λιμόϛ 121 λιπαρόϛ 316 λοιβή 216 λουτρόν 163 4 μάλα 321 μαλακόϛ 402 μαλερόϛ 318, 320 1 μαυρόω 121 22 μέθεξιϛ 135, 137, 139 μελίκρατον 195 μέν 178, 271 μένοϛ 252 5 μενοῦν(γε) 178 μετά( ) 94, 201 μεταδιωκόμενοϛ 201 μεταλεύομαι/μεταλεύω 201 μεταμελόμενοϛ 201 μεταχρόνιαι 114 15 μετεωρολόγοϛ 129 30 μέτωπον 101 μῆδοϛ 234 μηθείϛ 177 μῆνιϛ 89 μητίετα 239, 240 μῆτιϛ 226 43 μήτρα 148 μιαρόϛ 41 μισέω 265 6 μυελόϛ 145, 250 1

Greek Index μῦθοϛ 242 μύκηϛ 147 μύξα 145 μυρίοϛ 89 μύρμηξ 48 μυρτόχειλα 74 μύχιοϛ 52 ναόϛ/νεώϛ 203 5 ναῦϛ 178 9 νερό 183 4 νέφοϛ 310 11 Νέσσανδροϛ 100 Νέστωρ 100 νέω 48 νηλήϛ 313 14 νηνία 51 νόοϛ 177 νότιοϛ 55 νύμφη 78 νῦν 94 ξενεῖον 220 1 ξουθόϛ 318 ξύν( ) 203, 208 9 ξυσταθεύω 209 ό ό ο ά ά ά 344 ὁ, ἡ, τό 94 ὄβριμοϛ 323 ὅγε 278 οἶνοϛ 183, 195 οἰοῖ 342 οἰωνιστήϛ 52 ὄλισβοϛ 74 ὄλλυμαι/ὄλλυμι 89, 300 1, 328 ὁμοίϊοϛ 113, 318 ὁμοῖοϛ 113 ὀμφή 335 ὀμψύχω 201, 203 ὀν(α) 201 2 ὀνάλα 201 2 ὄνοϛ 53 ὄνυξ 59 ὀξύϛ 149 50, 306 ὄπα 326 ὀπηδέω 121 ὄπιϛ 325 6 ὅπου 53 ὀργάω 149 Ὀρέστηϛ 100 ὁρίζω 135 ὀρνιθοβοσκεῖον 51 Ὀρχομενόϛ 100 ὅϛ, ἥ, ὅ 88, 90 ὅταν 53 οὐά/οὐᾶ 49 οὐδείϛ 178 οὐθείϛ 177 8

Greek Index οὐλόμενοϛ 89 ὄφιϛ 210 11, 329 ὀφρύκνηστον 52 ὀχετεύομαι/ὀχετεύω 329 ὀψάριον 199 ὄψον 193, 199 ὀψοφάγοϛ 193 παθικόϛ 51, 56 παλιγγενεσία 163 παλιγκοταίνω 312 παλίγκοτοϛ 311 12, 314 πάλιν 399 πανάγριοϛ 398 πάναγρον 398 πάναγροϛ 398 πάνυ γε 278 9 παράδειγμα 138 9 παραμηχανάομαι 154 παρθένια 148 παρθένοϛ 148, 265 παροχετεύομαι 329 παροχέτευσιϛ 329 πᾶϛ 102 πατήρ 48 παῦροϛ 328 πεδα 201 2 πεδαίρω 202 πεδαλευόμενοϛ 201 πεδίον 260 πείρατα γαίηϛ 112 13 πελώρη 112 πέμμα 188 πεμμάτιον 188 πεμματόλογοϛ 188 πεμφιγώδηϛ 319, 322 πέμφιξ 318, 319 20, 322 3, 325 πέοϛ 72, 75 Περσεφόνη 206, 207, 213 πηγή 329 Πηλεύϛ 91 πῆμα 328 πλακουντάριον 187 πλακοῦϛ 187 πλάστηϛ 155 πλατύσημοϛ 52 πλευρῖτιϛ 146 πληγή 310 πλοῖον 178 πνεῦμα 176 ποιέω 94 ποικιλόμητιϛ 234 ποικίλοϛ 238 ποιμήν 311 ποιότηϛ 139 πολιτικόϛ 288 πολυγηθήϛ 112

πολύμητιϛ 240 πολύτροποϛ 237 πόνοϛ 147 πόντιοϛ 260 πόσθη 72, 73, 75 ποτιμάστιοϛ 208 πρεσβύτεροϛ 162 πρίγκιπεϛ 49 προβαίνω 292 πρόβασιϛ 291 2 πρόβατα 290, 292 πρόρρηγμα 143 πρόϛ( ) 94, 208 προστίηθμι 177 πρόσωπον 101 προτιόσσομαι 220 1 πρόφασιϛ 150 πρωκτίζω 56 πρῶτοϛ 94 πυγίζω 57, 72, 74, 80 πυκινόϛ 237 πυρόϛ 188 90 πυτίζω 50 ῥᾴδιοϛ 148 ῥαχίτηϛ 145 Ῥέα 206 7 ῥωχμόϛ 143 σαγηνεύω 398 σάθη 72, 75, 80 σαλαγέω 56 σαρδίνη 55 σάρξ 176, 265 σίαλον 145 σιτάρι((ο)ν) 190 1 σιτία 191 3 σιτίον 186, 188, 192 σιτόκριθον 190 σῖτοϛ 188 91 σκότοϛ 177 Σοῦχοϛ 186 σοφία 227, 234, 240, 242 σπαργάω 149 σπείρω 263 σπεκουλάτωρ 49 σπέρμα 148, 263 4, 306 7, 309 10 σπιλάϛ 166 8 σταμαγορίϛ 202 στέργω 266 7 στεῦται/ το 336 στεφανοσταύριον 154 στεφανόω 109 στρογγύλοϛ 336 σρογγύλωμα 336 7 στρόφιϛ 41 στύω 72, 80 συκοφάντηϛ 23

447

448 σύμπαϛ 103 σύν( ) 178, 203, 208 9 συναγωγή 137 σύρβη 220 1 συσταθεύω 208 9 σφριγάω 149 σωφροσύνη 206 7, 213 ταῦροϛ 48, 76 τε 94, 178 τέγγω 389 τείνω 149 τέμνω 147 τέτανοϛ 74, 78 τετράποδα 290 τετράπο(υ)ϛ 290 2 τεῦχοϛ 101 τεύχω 101 τέχνη 229, 234, 237, 240, 242 τηλία 220 1 τίθημι 88, 90 τιμή 249 τίϛ 95 τιτίϛ 76 *τοιχοδόμοϛ 100 τόποϛ 76 του 179 τράχηλοϛ 179 τρέφω 257 τριχάϊξ 349 τριχοπλάστηϛ 155 6 τύρβη 220 1 τύχη 120 ὗ ὗ ὗ ψ ψ ἆ ἆ 344 ὕδωρ 183, 199 υἱόϛ 102, 265 ὑπέρακμοϛ 160 1 ὑπό 177 ὕσσωποϛ 170 1 φάσηλοϛ 50 φέρω 102 φθίω 102 φιλοπυτιστήϛ 56 φιλόσοφοϛ 136 7 φλάω 57 φλέγμα 53, 146 φλεγμαίνω 149 φλέψ 144, 149 φοῖνιξ 325 φοίνιοϛ 325 φοινίσσω 325 φοίτηϛ 100 φόνοϛ 325 φόντηϛ 100 φύσιϛ 297 φωτίζω 163 φώτισμα 163

Greek Index χαιρέφυλλον 51 χέζω 73, 75 χείρων 337 χερείων 337 χερνῆτιϛ 337 χέω 146 χθέϛ 94, 177 χλωρόϛ 150, 257 χοῖροϛ 81 χολή 146 χυλόϛ 146 χυμόϛ 146 ψάμμη 207, 214 ψάμμοϛ 207 ψάρι 199 ψυχή 135 6, 294 ψωλή 56 7, 62, 72, 73, 75, 80 ψωλόϛ 73 ψωμί 183 4, 185 ψωμί(ο)ν 186, 188, 193 5, 197 9 ψωμόϛ 193 4, 198 ὤ 342 ὡϛ/ὥϛ 53 4, 59 60

Mycenaean Greek a ke 102 a mo ra ma 101 a pe a sa 102 a pe e ke 102 a pe e si 102 a pe o( te) 102 a qi ti ta 100 a re ka sa da ra 100 a re ke se u 100 di ri mi jo 104 di wo i je we 102 do e ra 335 do e ro 104, 335 do po ta 101 e keo 101 e ke 102 e ke e 102 e ko me no 100 e qi ti wo e 102 e re u te re 102 e ti ra wo 100 i je we 102 i ju 102 ki ri ta 188 ki ti ta 101 ko ma we( to/ te) 102 ko ma we te ja 102 ko to( i) na 101 ko to no o ko 101

Greek Index ku su pa 103 me ta ki ti ta 101 ne ti ja no 100 o o ko 101 o ko me ne u 100 o pi te u ke e we 101 o po qo 101 o re ta 100 pa si 102 pa si te o i 103 pa te 102 pe re 102

qe to ro po pi 104, 290 1 qo i ta 100 qo ta 100 ra wo 100 si to 188 te tu ko wo a₂ 101 te u ke pi 101 to ko do mo 100 to so pa 102 wa na ka( te) 102 zo wi jo 294 zo wo 294

449