Form and Function in Greek Grammar

Amsterdam Studies in
Classical Philology

Series Editors

Irene J.F. de Jong

Caroline H.M. Kroon

Editorial Board

Rutger J. Allan
Mark A.J. Heerink

volume 30

The titles published in this series are listed at brill.com/ascp

Form and Function
in Greek Grammar
Linguistic Contributions
to the Study of Greek Literature

By

Albert Rijksbaron

Edited by

Rutger J. Allan
Evert van Emde Boas
Luuk Huitink

LEIDEN | BOSTON
The Library of Congress Cataloging-in-Publication Data

Names: Rijksbaron, Albert, author. | Allan, Rutger J., editor. | Emde Boas, Evert van,
1982- editor. | Huitink, Luuk, 1981- editor.
Title: Form and function in Greek grammar : linguistic contributions to the study
of Greek literature / by Albert Rijksbaron ; edited by Rutger J. Allan, Evert van
Emde Boas, Luuk Huitink.
Description: Leiden ; Boston : Brill, 2018. | Series: Amsterdam studies in classical
philology ; volume 30 | Includes bibliographical references and indexes.
Identifiers: LCCN 2018047457 (print) | LCCN 2018051232 (ebook) |
ISBN 9789004386129 (ebook) | ISBN 9789004385771 (hardback : alk. paper)
Subjects: LCSH: Greek language–Grammar. | Greek literature–History and
criticism.
Classification: LCC PA251 (ebook) | LCC PA251 .R55 2018 (print) | DDC 485–dc23
LC record available at https://lccn.loc.gov/2018047457

Typeface for the Latin, Greek, and Cyrillic scripts: “Brill”. See and download: brill.com/brill‑typeface.

ISSN 1380-6068
ISBN 978-90-04-38577-1 (hardback)
ISBN 978-90-04-38612-9 (e-book)

Copyright 2019 by Koninklijke Brill NV, Leiden, The Netherlands.

Koninklijke Brill NV incorporates the imprints Brill, Brill Hes & De Graaf, Brill Nijhoff, Brill Rodopi,
Brill Sense, Hotei Publishing, mentis Verlag, Verlag Ferdinand Schöningh and Wilhelm Fink Verlag.
All rights reserved. No part of this publication may be reproduced, translated, stored in a retrieval system,
or transmitted in any form or by any means, electronic, mechanical, photocopying, recording or otherwise,
without prior written permission from the publisher.
Authorization to photocopy items for internal or personal use is granted by Koninklijke Brill NV provided
that the appropriate fees are paid directly to The Copyright Clearance Center, 222 Rosewood Drive,
Suite 910, Danvers, MA 01923, USA. Fees are subject to change.

This book is printed on acid-free paper and produced in a sustainable manner.

Discipulis

∵
Contents

Acknowledgements ix
Albert Rijksbaron—CV and Publications xi

Introduction 1

1 A Review of: H. Hettrich, Kontext und Aspekt in der altgriechischen

Prosa Herodots 5

2 The Greek Perfect: Subject versus Object 39

3 The Discourse Function of the Imperfect 60

4 Sur les emplois de λέγε et εἰπέ chez Platon 80

5 On False Historic Presents in Sophocles (and Euripides) 99

6 ἀκροᾶσθαι or ἀκροάσασθαι (Plato, Ion 530d9)? 124

7 The Imperfect as the Tense of Substitutionary Perception 133

8 How Does a Messenger Begin His Speech? Some Observations on the

Opening Lines of Euripidean Messenger Speeches 170

9 Discourse Cohesion in the Proem of Hesiod’s Theogony 185

10 On the Syntax and Pragmatics of inquit Formulae in Plato’s Narrated

Dialogues 210

11 Sur quelques différences entre οὗτος ὁ + substantif, οὗτος δὲ ὁ +

substantif, ὁ δὲ + substantif + οὗτος chez Hérodote 223

12 Sur les emplois de ἐάν et ἐπεάν (à propos d’ Euripide, Bacchae

50–51) 237
viii contents

13 The Syntax and Semantics of Expressions of Sorrow and Related

Concepts in Homer 254

14 The Meaning and Word Class of πρότερον and τὸ πρότερον 279

15 Adverb or Connector? The Case of καὶ … δέ 293

16 Sur l’article avec nom propre 319

17 Does Ancient Greek Have a Word for ‘no’? The Evidence from οὐκοῦν
(…) οὐ Questions 335

18 The Treatment of the Greek Middle Voice by the Ancient

Grammarians 357

19 A Question of Questions: peusis, erôtêsis and [Longinus] Περὶ ὕψους

18.1 370

20 The Xenophon Factory: One Hundred and Fifty Years of School

Editions of Xenophon’s Anabasis 376

Index Rerum 407

Index Graecitatis 411
Index Locorum 413
Acknowledgements

The author and editors wish to acknowledge the kind permission of the rele-
vant publishers to reprint the chapters collected in this volume:
1. Review of H. Hettrich, Kontext und Aspekt in der altgriechischen Prosa
Herodots, in Lingua 48 (1979) 223–257
2. ‘The Greek Perfect: Subject versus Object’, translated by A. Rijksbaron
from ‘Het Griekse perfectum: subject contra object’, Lampas 17 (1984)
403–420
3. ‘The Discourse Function of the Imperfect’, in A. Rijksbaron, H.A. Mul-
der, G.C. Wakker (eds), In the Footsteps of Raphael Kühner: Proceedings
of the International Colloquium in Commemoration of the 150th Anniver-
sary of the Publication of Raphael Kühner’s Ausführliche Grammatik der
griechischen Sprache, II. Theil: Syntaxe (Amsterdam: Gieben, 1988) 237–
254
4. ‘Sur les emplois de λέγε et εἰπέ chez Platon’, in B. Jacquinod et J. Lallot
(eds), Études sur l’aspect verbal chez Platon (Saint-Etienne: Publications
de l’Université, 2000) 151–170
5. ‘On False Historic Presents in Sophocles (and Euripides)’, in I.J.F. de Jong,
A. Rijksbaron (eds), Sophocles and the Greek Language (Leiden: Brill,
2006) 127–149
6. ‘ἀκροᾶσθαι or ἀκροάσασθαι? (Plato Ion 530d9)’, Appendix III in Plato: Ion.
Or: On the Iliad, edited with introduction and commentary (Leiden: Brill,
2007) 261–269
7. ‘The Imperfect as the Tense of Substitutionary Perception’, in P. da Cunha
Corrêa, M. Martinho, A. Pinheiro Hasegawa (eds), Hyperboreans: Essays
in Greek and Latin Poetry, Philosophy, Rhetoric and Linguistics (São Paulo:
Humanitas, 2012) 331–377
8. ‘How Does a Messenger Begin His Speech? Some Observations on the
Opening Lines of Euripidean Messenger Speeches’, in J.M. Bremer,
S.L. Radt, C.J. Ruijgh (eds), Miscellanea tragica in honorem J.C. Kamerbeek
(Amsterdam: Hakkert, 1976) 293–309
9. ‘Discourse Cohesion in the Proem of Hesiod’s Theogony’, in S.J. Bakker,
G.C. Wakker (eds), Discourse Cohesion in Ancient Greek (Leiden: Brill,
2009) 241–267
10. ‘On the Syntax and Semantics of inquit Formulae in Plato’s Narrated Dia-
logues’, in J. de la Villa Polo et al. (eds), Ianua Classicorum: Temas y formas
del Mundo Clásico, vol. 1 (Madrid: Sociedad Española de Estudios Clásicos,
2015) 397–410
x acknowledgements

11. ‘Sur quelques différences entre οὗτος ὁ + substantif, οὗτος δὲ ὁ + substantif

et ὁ δὲ + substantif + οὗτος chez Hérodote’, Lalies 12 (1993) 119–130
12. ‘Sur les emplois de ἐάν et ἐπεάν’, Lalies 12 (1993) 131–144
13. ‘The Syntax and Semantics of Expressions of Sorrow and Related Con-
cepts in Homer’, adapted by A. Rijksbaron from ‘D’ où viennent les ἄλγεα?
Quelques observations à propos d’ ἄλγε᾽ ἔχειν chez Homère’, in F. Létou-
blon (ed.), La langue et les textes en grec ancien: Actes du Colloque P. Chan-
traine, Grenoble, 5–8 septembre 1989 (Amsterdam: Gieben, 1991) 181–193,
and from ‘Further Observations on Expressions of Sorrow and Related
Expressions in Homer’, in E. Banfi (ed.), Atti del Secondo Incontro inter-
nazionale di Linguistica greca (Trento: Editrice Universita degli Studi,
1997) 215–224
14. ‘The Meaning and Word Class of πρότερον and τὸ πρότερον’ in E. Crespo
et al. (eds), Word Classes and Related Topics in Ancient Greek (Louvain-la-
Neuve: Peeters, 2006) 405–441
16 ‘Sur l’article avec nom propre’, in J.-L. Breuil et al. (eds), Ἐν κοινωνίᾳ πᾶσα
φιλία: Mélanges pour Bernard Jacquinod (Saint-Etienne: Presses universi-
taires, 2006) 243–257
17. ‘Does Ancient Greek Have a Word for ‘No’? The Evidence from οὐκοῦν …
οὐ Questions’, Journal of Greek Linguistics 12 (2012) 140–160
18. ‘The Treatment of the Greek Middle Coice by the Ancient Grammarians’,
in H. Joly (ed.), Actes du Colloque International ‘Philosophies du langage et
théories linguistiques dans l’Antiquité’, Grenoble 1985 (Bruxelles: Editions
Ousia-Université des Sciences Sociales de Grenoble, 1986) 427–444
19. ‘A Question of Questions: peusis, erôtêsis and [Longinus] Περὶ ὕψους 18.1’
Mnemosyne 56 (2003) 733–736
20. ‘The Xenophon Factory: One Hundred and Fifty Years of School Editions
of Xenophon’s Anabasis’, in R. Gibson, C.S. Kraus (eds), The Classical Com-
mentary: Histories, Practices, Theory (Leiden: Brill, 2002) 235–267
Albert Rijksbaron—CV and Publications

1 Curriculum Vitae

1968 doctoraalexamen (MA), cum laude, University of Amsterdam

1969 appointment as wetenschappelijk medewerker (PhD candidate /
Lecturer) in Ancient Greek, University of Amsterdam
1976 PhD (dissertation: Temporal and Causal Conjunctions in Ancient
Greek: With Special Reference to the Use of ἐπεί and ὡς in Herodotus),
cum laude, University of Amsterdam
1981–1984 member of the Faculty board; initiator of new degree course Euro-
pean Studies
1985 appointment as Universitair Hoofddocent (Associate Professor) of
Ancient Greek and European Studies (after 1998: Ancient Greek
only)
2000 appointment as Hoogleraar (full Professor) of Ancient Greek
2005 retirement

Visiting studentships/scholarships at École Normale Supérieure Paris (1968–

1969), Harvard University (1987 and 1996). Member of the research group
‘Aspect verbal’ (1992–2010).

Rijksbaron served as editor of the journals Lampas (1972–2005) and Mnemos-

yne (2001–2012) and editor of the Amsterdam Studies in Classical Philology
series (1991–2014).

2 Complete List of Publications

2.1 Books, Articles, Chapters

Items with an asterisk are reprinted in this volume; relevant chapter numbers
are given in square brackets. Volumes (co-)edited by Rijksbaron are not listed
separately if he appears as contributor in the volume in question.

1966
Dutch translation of Lucretius, De rerum natura 1.921–950, Hermeneus 37: 173–174.

1972
‘De Griekse genitivus’, Lampas 5: 68–77.
‘ἐπεί en ὡς’, Lampas 5: 465–477.
xii albert rijksbaron—cv and publications

1974
Review of C.M.J. Sicking, Hoofdstukken uit de Griekse ‘syntaxis’, Lampas 7: 207–218
‘Het praesens historicum in het Oudgrieks’, Handelingen van het XXXIIe Nederlands
Filologencongres (Amsterdam) 146–148.

1976
Temporal and Causal Conjunctions in Ancient Greek: With Special Reference to the Use of
ἐπεί and ὡς in Herodotus (Amsterdam).
*‘How Does a Messenger Begin His Speech? Some Observations on the Opening Lines
of Euripidean Messenger Speeches’, in J.M. Bremer, S.L. Radt, C.J. Ruijgh (eds), Mis-
cellanea tragica in honorem J.C. Kamerbeek (Amsterdam) 293–309. [ch. 8]

1979
*Review of H. Hettrich, Kontext und Aspekt in der altgriechischen Prosa Herodots, Lin-
gua 48: 223–257. [ch. 1]

1980
‘Ancient Greek Relative Clauses and Functional Grammar’, in S. Daalder, M. Gerritsen
(eds), Linguistics in the Netherlands: 1980 (Amsterdam) 121–127.
‘De semantiek van Griekse hypothetische bijzinnen’, Lampas 13: 130–146.

1981
‘Relative Clause Formation in Ancient Greek’, in A.M. Bolkestein et al., Predication and
Expression in Functional Grammar (London) 235–259.
‘Deux notes: Esch. Pers. 598; Eur. Or. 1281’, Mnemosyne 34: 388–390.
Dutch translation of ‘De Eed van Hippocrates’, in H.M. Beliën et al. (eds) Een geschiede-
nis van de Oude Wereld: Bronnen (Haarlem) 123.

1984
The Syntax and Semantics of the Verb in Classical Greek: An Introduction (Amsterdam).
‘Chariton 8.1.4 und Aristot. Poet. 1449b28’, Philologus 128: 306–307.
*‘Het Griekse perfectum: subject contra object’, Lampas 17: 403–420 [ch. 2, translated
into English].

1986
‘The Pragmatics and Semantics of Conditional and Temporal Clauses: Some evidence
from Dutch and Classical Greek’, Working Papers in Functional Grammar 13.
*‘The Treatment of the Greek Middle Voice by the Ancient Grammarians’, in H. Joly
(ed.), Actes du Colloque International ‘Philosophies du langage et théories linguis-
tiques dans l’Antiquité’, Grenoble 1985 (Bruxelles) 427–444. [ch. 18]
albert rijksbaron—cv and publications xiii

‘Infinitivus en participium als complement in het Oudgrieks: het probleem van ἄρχομαι
en πειρῶμαι’, Lampas 19: 175–192.
‘Taalkunde en de structuur van Herodotus’ Historiën’, Lampas 19: 220–231.

1987
‘Hebben Herodotus’ Historiën een strekking?’ Bulletin VCN 12.

1988
*‘The Discourse Function of the Imperfect’, in A. Rijksbaron, H.A. Mulder, G.C. Wakker
(eds), In the Footsteps of Raphael Kühner: Proceedings of the International Colloquium
in Commemoration of the 150th Anniversary of the Publication of Raphael Kühner’s
Ausführliche Grammatik der griechischen Sprache, II. Theil: Syntaxe (Amsterdam)
237–254. [ch. 3]
‘“Waar ben ik nu weer verzeild?!” Odysseus en Nausikaa’, in A.M. van Erp Taalman Kip
et al. (eds) Propemptikon: Afscheidsbundel W.J.H.F. Kegel (Amsterdam) 36–43.

1989
Aristotle, Verb Meaning and Functional Grammar: Towards a New Typology of States of
Affairs (Amsterdam).
‘Wat was Europa, en waarom?’, in J. Leerssen, J.H. Reestman, A. Rijksbaron (eds),
Tussen wetenschap en werkelijkheid: Europese opstellen aangeboden aan Max Weis-
glas (Amsterdam) 122–130.
‘De zaak van de draaibank. Eur. Ba. 1066 vv.’, Lampas 23: 343–347.

1991
Grammatical Observations on Euripides’ Bacchae (Amsterdam).
*‘D’où viennent les ἄλγεα? Quelques observations à propos d’ ἄλγε᾽ ἔχειν chez Homère’,
in F. Létoublon (ed.), La langue et les textes en grec ancien: Actes du Colloque P. Chan-
traine, Grenoble, 5–8 septembre 1989 (Amsterdam) 181–193. [ch. 13, adapted]

1992
(with I.J.F. de Jong) ‘“Zo lag de held Odysseus …”: Enige opmerkingen bij de Odyssee-
vertaling van I. Dros’, Lampas 25: 198–213.
Dutch translation of Homer, Odyssey 6.1–40, in D. den Hengst (ed.), Van Homerus tot
Van Lennep: Griekse en Latijnse literatuur in Nederlandse vertaling (Muiderberg) 25–
27.

1993
*‘Sur quelques différences entre οὗτος ὁ + substantif, οὗτος δὲ ὁ + substantif et ὁ δὲ +
substantif + οὗτος chez Hérodote’, Lalies 12: 119–130. [ch. 11]
xiv albert rijksbaron—cv and publications

*‘Sur les emplois de ἐάν et ἐπεάν’, Lalies 12: 131–144. [ch. 12]
‘Why is the Incident on Thrinacia Mentioned in Od. 1.7–9?’, Mnemosyne 46: 528–529.
‘Energie’, in T. Eijsbouts et. al. (eds), De onrust van Europa (Amsterdam) 61–71.

1994
The Syntax and Semantics of the Verb in Classical Greek: An Introduction, second, revised
edition (Amsterdam).
‘“Héél kilo aardbeien 4 gulden”. Of: heeft FG retorische functies nodig?’, in F.G. van
Werkgem (ed.), Dubbel Nederlands: 23 opstellen voor Simon C. Dik (Amsterdam) 49–
52.
‘Nadrukkelijke en onnadrukkelijke verwijzingen in Herodotus: Over de positie en func-
tie van οὗτος’, Lampas 27: 230–241.
‘Herodotus, de vader van de geschiedvervalsing?’, Lampas 27: 242–248.

1995
‘Euripides, Bacchae 35–36’, Mnemosyne 48: 198–200.
‘Les valeurs aspectuelles selon A. Poutsma’, Syntaktika 8: 3–15.
‘Van “doortrapt” tot “losbandig” en alles wat daar tussen ligt: De schurk in de Griekse
roman’, in A.M. van Erp Taalman Kip, I.J.F. de Jong (eds), Schurken en schelmen: Cul-
tuurhistorische verkenningen rond de Middellandse Zee (Amsterdam) 97–107.

1997
(with K.A. Worp) The Kellis Isocrates Codex, The Dakhleh Oasis Project: Monograph 5
(Oxford).
*‘Further Observations on Expressions of Sorrow and Related Expressions in Homer’, in
E. Banfi (ed.), Atti del Secondo Incontro internazionale di Linguistica greca (Trento)
215–242. [ch. 13, adapted]
‘Introduction’ and *‘Adverb or connector? The case of καὶ … δέ’, in A. Rijksbaron (ed.),
New Approaches to Greek Particles (Amsterdam) 1–14; 187–208. [ch. 15]
‘Van steen tot mens: Een verkenning van de Homerische wereld’, Lampas 30: 198–212.

1998
(with K.A. Worp) ‘Isocrates bilinguis Berolinensis’, Mnemosyne 51: 718–723.
‘Euripides, Hippolytus 141’, Mnemosyne 51: 712–715.

1999
‘Lethe for the lathe? Euripides, Bacchae 1066–67 again’, Mnemosyne 52: 705–710.

2000
*‘Sur les emplois de λέγε et εἰπέ chez Platon’, in B. Jacquinod et J. Lallot (eds), Études
sur l’aspect verbal chez Platon (Saint-Etienne) 151–170. [ch. 4]
albert rijksbaron—cv and publications xv

Review of E.-C. Gerö, Negatives and Noun Phrases in Classical Greek, Kratylos 45: 203–
206.
‘De toekomst in de geschiedenis’, in M. Spiering et al. (eds), De weerspannigheid van de
feiten: Opstellen aangeboden aan W.H. Roobol (Hilversum) 189–200.
(with S.R. Slings, P. Stork, G.C. Wakker) Beknopte syntaxis van het klassiek Grieks (Lun-
teren).

2001
‘De Xenophon-fabriek. Honderdvijfig jaar schoolcommentaren op de Anabasis.’ Lam-
pas 34: 114–140.
Over bepaalde personen (Amsterdam) (inaugural lecture).

2002
*‘The Xenophon Factory: One Hundred and Fifty Years of School Editions of Xeno-
phon’s Anabasis’, in R. Gibson, C.S. Kraus (eds), The Classical Commentary: Histories,
Practices, Theory (Leiden) 235–267. [ch. 20]
The Syntax and Semantics of the Verb in Classical Greek: An Introduction, third, revised
edition (Amsterdam).
‘Karl Lachmann en zijn methode: het voorbeeld van de klassieke filologie’, in J. Leerssen,
M. Mathijsen (eds), Oerteksten: Nationalisme, edities en canonvorming (Amsterdam)
13–21.

2003
*‘A Question of Questions: peusis, erôtêsis and [Longinus] Περὶ ὕψους 18.1’ Mnemosyne
56: 733–736. [ch. 19]

2004
(with J.-F. Courtine) ‘τὸ τί ἦν εἶναι’, in B. Cassin (ed.), Vocabulaire européen des philoso-
phies (Paris) 1298–1304.
(with S. de Voguë and four other authors) ‘Aspect’, in B. Cassin (ed.), Vocabulaire
européen des philosophies (Paris) 116–144.
‘Een dramatisch moment van tweetaligheid’, in A. van Heerikhuizen et al. (eds), Het
Babylonische Europa: Opstellen over veeltaligheid (Amsterdam) 13–19.
‘Over et tu, Brute en ander niet bestaand Latijn’, Lampas 37: 229–234.

2005
Review of G. Cooper III, Greek Syntax, Classical Review 55: 479–482.

2006
*‘On False Historic Presents in Sophocles (and Euripides)’, in I.J.F. de Jong, A. Rijksbaron
(eds), Sophocles and the Greek Language (Leiden) 127–149. [ch. 5]
xvi albert rijksbaron—cv and publications

*‘The Meaning and Word Class of πρότερον and τὸ πρότερον’ in E. Crespo et al. (eds),
Word Classes and Related Topics in Ancient Greek (Louvain-la-Neuve) 405–441.
[ch. 14]
*‘Sur l’article avec nom propre’, in J.-L. Breuil et al. (eds), Ἐν κοινωνίᾳ πᾶσα φιλία: Mé-
langes pour Bernard Jacquinod (Saint-Etienne) 243–257. [ch. 16]
Over punten en komma’s, in het bijzonder bij Plato (Amsterdam) (valedictory lecture).

2007
The Syntax and Semantics of the Verb in Classical Greek, third edition, American reprint
(Chicago).
Plato: Ion. Or: On the Iliad, edited with introduction and commentary (Leiden).
Review of E. Crespo, L. Conti, H. Maquieira, Sintaxis del Griego Clásico, Kratylos 52: 65–
67.

2008
Review of M. Napoli, Aspect and Actionality in Homeric Greek: A Contrastive Analysis,
Journal of Hellenic Studies 128: 268–269.

2009
*‘Discourse Cohesion in the Proem of Hesiod’s Theogony’, in S.J. Bakker, G.C. Wakker
(eds), Discourse Cohesion in Ancient Greek (Leiden) 241–267. [ch. 9]
‘Negatives in Questions (and Answers): The Case of οὐκοῦν … οὐ. Or: Does Ancient
Greek Have a Word for ‘No’?’, in K. Loudová, M. Záková (eds), Early European Lan-
guages in the Eyes of Modern Linguistics (Brno) 261–271.

2010
‘Voorwerpen als personages: De revolver en de pijp van Maigret’, in A. van Heerikhui-
zen, I.J.F. de Jong, M. van Montfrans (eds), Tweede levens: Over personen en person-
ages in de geschiedschrijving en de literatuur (Amsterdam) 167–185.

2011
‘Introduction’ and ‘The Profanation of the Mysteries and the Mutilation of the Hermae:
Two Variations on Two Themes’, in J. Lallot, A, Rijksbaron, B. Jacquinod, M. Buijs
(eds), The Historical Present in Thucydides: Semantics and Narrative Function / Le
présent historique chez Thucydide: Sémantique et fonction narrative (Leiden) 1–19;
177–195.

2012
*‘Does Ancient Greek Have a Word for ‘No’? The Evidence from οὐκοῦν … οὐ Questions’,
Journal of Greek Linguistics 12: 140–160. [ch. 17]
albert rijksbaron—cv and publications xvii

*‘The Imperfect as the Tense of Substitutionary Perception’, in P. da Cunha Corrêa et

al. (eds), Hyperboreans: Essays in Greek and Latin Poetry, Philosophy, Rhetoric and
Linguistics (São Paulo) 331–377. [ch. 7]

2013
Σύνταξη και σημασιολογία του ρήματος στην κλασική ελληνική: Εισαγωγή (Θεσσαλονίκη)
(Modern Greek translation of The Syntax and Semantics of the Verb in Classical Greek,
third edition).
Review of D. Riaño Rufilanchas, El complemento directo en griego antiguo, Gnomon 85,
108–112.

2015
*‘On the Syntax and Semantics of inquit Formulae in Plato’s Narrated Dialogues’, in J.
de la Villa Polo et al. (eds), Ianua Classicorum: Temas y formas del Mundo Clásico,
vol. 1 (Madrid) 397–410. [ch. 10]
‘Stative Historical Presents in Greek Tragedy: Are They Real?’ Philologus 159: 224–250.

2019
(with E. van Emde Boas, L. Huitink, M. de Bakker) Cambridge Grammar of Classical
Greek (Cambridge).

2.2 Publications for Secondary Schools

1977
(as co-author) Basis: Griekse leergang voor het VWO (Den Haag).

1982
(as co-editor) Metabasis: Een bloemlezing Griekse teksten voor het VWO (Leeuwarden).

1987
(with E. Jans and P. Stork) Basis: Leergang Grieks, completely revised edition, 2 vols
(Leeuwarden).

1989
(as co-author) Herodotos: Literator en historicus (Leeuwarden).

1992
(with A. Jansen, C. Hupperts) Stephanos: Een bloemlezing uit de Anthologia Graeca
(Leeuwarden).
xviii albert rijksbaron—cv and publications

1994
(as co-author) Xerxes, een koning die zijn grenzen niet kende (Leeuwarden).
(as co-author) Pallas: Leergang Grieks voor het VWO: Deel 1, 2 vols (Leeuwarden).

1995
(as co-author) Pallas: Leergang Grieks voor het VWO: Deel 2, 2 vols (Leeuwarden).
Introduction

[T]ake languages seriously. Whenever there is some overt difference be-

tween two constructions X and Y, start out on the assumption that this
difference has some kind of functionality in the linguistic system.
– Simon C. Dik, The Theory of Functional Grammar, Part 1: The Structure of the
Clause

∵
Simon Dik’s statement strikes an appropriate opening chord to this collection
of selected papers by Albert Rijksbaron, for at least two reasons. First, Dik’s
statement well captures an important principle of Rijksbaron’s methodologi-
cal approach to Greek linguistics, an approach in which the texts provide the
natural starting point of linguistic research. Rijksbaron’s publications often
depart from the observation that two (or more) morphological forms or syn-
tactic constructions co-occur in a particular literary text, apparently without
any clear difference in meaning or function between them. A thorough exami-
nation of the occurrences of these linguistic alternatives in their (literary) con-
text reveals, however, that there are in fact important semantic or pragmatic
differences between them. Rijksbaron’s contributions, in other words, exhibit
a strongly inductive ‘bottom-up’ approach: linguistic theorizing is used as an
important tool in his research but it is rarely pursued as an end in itself. The
ultimate goal of linguistics à la Rijksbaron is to come to a better understanding
of the texts.
A second reason why the quotation is a suitable motto is the fact that its
author is Simon Dik, the founder of the theory of Functional Grammar. Func-
tional Grammar has been a constant and important inspiration to Rijksbaron’s
work. He knew Simon Dik well personally and he belongs to the first gener-
ation of linguists (together with, for example, his close colleagues, the Latin-
ists Machtelt Bolkestein and Harm Pinkster) who applied the theory and con-
tributed to its further development.
Rijksbaron’s work occupies a prominent place in the field of Greek linguis-
tics. His Syntax and Semantics of the Verb in Classical Greek has become a stan-
dard reference work for Greek linguists; unusually for an academic monograph
in this field, the book has gone through three editions and an American reis-
sue and has been translated into Modern Greek. The forthcoming Cambridge

© koninklijke brill nv, leiden, 2019 | doi:10.1163/9789004386129_002

2 introduction

Grammar of Classical Greek, a new reference grammar which he wrote together

with a number of his students (two of whom are also editors of the present vol-
ume), will, it may be hoped, have a similar influence on the study of Greek
language and literature. Rijksbaron has also regularly employed his linguistic
expertise in the editing of texts and in writing commentaries. Examples of this
line of scholarly work are the edition, together with Klaas Worp, of the Kellis
Isocrates codex, his linguistic commentary on Euripides’ Bacchae, and his edi-
tion with commentary of Plato’s Ion.
Rijksbaron is no doubt the most recognizable voice of ‘Dutch pragmatic crit-
icism’ (the term is Victor Bers’),1 which is now commonly acknowledged as a
distinctive and productive strand of classical scholarship. ‘Dutch’ though it may
be (or precisely because it is), Rijksbaron’s work is the opposite of parochial.
Over the course of a career that spans more than four decades, Rijksbaron has
been committed to stimulating a dialogue between different national tradi-
tions of classical scholarship. After spending time as a student in Paris, he has
retained a special affinity with French intellectual life, often publishing in that
language (four French papers are included in the present volume). He also was
for a long time an active member of the ‘Groupe Aspect’, a bi-annual meeting of
French and Dutch Greek linguists, and co-editor of several collaborative pub-
lications by that group.
A perhaps lesser-known aspect of Rijksbaron’s work is the contribution he
has made to the teaching of Greek language and literature in schools. The lines
of influence here are both direct (Rijksbaron authored a number of course-
books, anthologies, and exam syllabuses used in Dutch schools) and indirect
(many school teachers were trained by Rijksbaron or use(d) his work to come
to grips with Greek grammar). In his scholarly output, too, Rijksbaron has often
paid attention to how work in Greek linguistics filters down (or ought to do so)
into teaching practice; the final chapter in this volume, on school editions of
Xenophon’s Anabasis, has been included to reflect this aspect of his work.
The initiative for the present volume, which offers a selection of Rijksbaron’s
linguistic papers, was taken originally by Irene de Jong; the editorial work was
entrusted to us, three of Rijksbaron’s pupils. The papers have been selected
to represent Rijksbaron’s rich and varied œuvre: he has published on a wide
range of different topics in Greek linguistics, including conjunctions, cases, syn-
tax, tense and aspect, mood and modality, negation, discourse analysis, and

1 See Bers, V., Review of I.J.F. de Jong, A. Rijksbaron (eds), Sophocles and the Greek Language:
Aspects of Diction, Syntax and Pragmatics (Leiden 2006), Bryn Mawr Classical Review
2006.08.46, http://bmcr.brynmawr.edu/2006/2006‑08‑46.html.
introduction 3

particles. Still, two main areas of interest may perhaps be identified: the Greek
tense-aspect system and the discourse analysis of Greek literary texts. In these
two fields, Rijksbaron’s work has exerted a strong and lasting influence on the
development of the scholarly debate. Most of the contributions in this volume,
then, focus in one way or another on these two areas. This is most obviously true
of the first ten chapters in the selection (chs. 1–10), which deal explicitly with
questions of tense-aspect or of discourse structuring. In the following seven
chapters (chs. 11–17) the semantics of individual lexical items and the functions
of individual Greek particles and the article occupy a more central position.
Yet while those chapters focus on the use and meaning of individual words or
phrases, Rijksbaron never loses sight of how such words or phrases fit into the
wider context. Indeed, contextual factors often prove to be the key to the cor-
rect interpretation of the individual items. The final three chapters (chs. 18–20),
while still tackling linguistic questions, are also good examples of Rijksbaron’s
interest in two other, related fields: ancient grammatical theory and the com-
mentary tradition (from antiquity to the present).
Another way in which the selection is representative is that it presents a
good overview of Rijksbaron’s working methods. We have already mentioned
his typical bottom-up approach: readers will quickly become familiar with
the systematic manner in which Rijksbaron collects and analyses his material.
Related to this is the commentary-like structure of several of the chapters (or
sections in them): Rijksbaron’s writing typically operates on the basis of close
reading of individual passages. Other recurring characteristics of his work are
also on abundant display: his keen and vigilant eye for textual problems, as well
as his soberingly realistic attitude to their solvability; his firm hand in deal-
ing with the published commentaries and, especially, translations of others;
his broad knowledge of other languages and texts from other literary canons,
which often serve as suitable parallels. Finally, a constant thread in Rijksbaron’s
work is that it is sensitive to the literary implications of linguistic minutiae. It
is in this regard that Rijksbaron’s work has had arguably the most significant
impact (certainly on us, his students): it demonstrates how much linguistics
can contribute to the interpretation of literary texts.

R.J.A.
E.v.E.B
L.H.
4 introduction

Note on Editorial Practice

The text of the original chapters and articles is reproduced here largely un-
changed. One paper has been translated from the original Dutch (ch. 2, ‘The
Greek Perfect: Subject versus Object’), and one is a completely new adapta-
tion in English of two articles, one of which was originally published in French
(ch. 13, ‘The Syntax and Semantics of Expressions of Sorrow and Related Con-
cepts in Homer’). Other chapters in French have been left untranslated, in line
with Rijksbaron’s deep engagement with Greek linguistic scholarship in France.
Original page numbers have been added in the margins, with page divisions
marked by the sign |. Internal cross-references copied from the originals main-
tain the original pagination, using square brackets (e.g. ‘see p. [123]’). Footnote
numbering has also been left unchanged, as has the numbering of examples
(even in the few instances, in chs. 3 and 11, where a number was acciden-
tally skipped). Square brackets are also used to indicate newly added cross-
references within the volume (e.g. ‘[See also ch. 7 in this volume]’), and, in
chapter 2, to mark off a few addenda by Rijksbaron in the footnotes.
We have introduced a uniform style for numbered Greek examples. In some
cases missing translations of examples (by Rijksbaron’s hand) have been added,
but we have not sought to do this exhaustively, not least because of Rijksbaron’s
emphasis on the principle that translation implies interpretation.
The formatting of references and bibliographies has been made uniform and
brought in line with the standard of the Amsterdam Studies in Classical Philol-
ogy series. We have added bibliographies to chapters that originally did not
have one of their own. It is Rijksbaron’s practice not to list standard text edi-
tions and commentaries in his bibliographies, and we have not deviated from
that practice here.
We have added abstracts to chapters originally without one (chs. 1, 6, 7, 8, 9,
13, 15, 16, 18, 19, 20).
Finally, we have silently corrected a few minor imperfections in the original
texts.
 chapter 1

A Review of: H. Hettrich, Kontext und Aspekt in der

altgriechischen Prosa Herodots

Abstract

This chapter is a long review article of Heinrich Hettrich’s Kontext und Aspekt in der
altgriechischen Prosa Herodots (Göttingen: Vandenhoeck & Ruprecht, 1976; 128 pp.).
Endorsing the main thrust of Hettrich’s argument, it counters the common view that
a speaker is in principle free to choose between imperfective and perfective (aoris-
tic) forms; the notion of a ‘moment donné’, established by Ruijgh and elaborated by
Hettrich, often appears to be a decisive factor: when one reads in linear order, a tense-
form T1 often serves as the reference-point for following tense-forms (T2, T3, etc.). Thus,
in subordinate clauses that precede their main clause, an imperfect signals that the
actions to be presented in the main clause stand in a relationship of simultaneity to
the action of the subordinate clause, while an aorist signals a relationship of anteri-
ority. It is argued that an extension of this rule to other, larger linguistic units requires
precision and care. In the case of larger chunks of discourse, the imperfect often creates
a temporal framework within which other actions occur.

1 Introduction

1.1 Aspect as a Universal Category

Verbal aspect has now for more than a century attracted scholarly attention.*
Ever since, roughly, the mid-19th century, when Curtius’ book on tenses and
moods in Greek and Latin appeared (Curtius 1846), an increasing number of
books and articles have been published on this subject. Initially, they were
mostly concerned with Slavonic languages and Ancient Greek (indeed, Cur-
tius himself seems to have been influenced by Slavonic studies, cf. S-D 2, 251,
note 1) but progressively the verbal systems of more and more other languages
were taken into consideration, notably Latin, French and English. A number
of non-IE languages have also been studied with respect to aspect and other

* I am indebted to Elseline Vester for her criticism on a number of points.

© koninklijke brill nv, leiden, 2019 | doi:10.1163/9789004386129_003

 6 chapter 1

verbal categories. For a—non-exhaustive—survey one may consult the bibli-

ography in Comrie (1976); for English also Schopf (1974).
It must be admitted, however, that all this work has not led to a communis
opinio on even such basic questions as what aspect is, what semantic phenom-
ena are connected with it, what factors determine the choice of one rather
than another verb-form, etc. Briefly, there is no general aspect theory.1 It must
equally be admitted, however, that this situation is perhaps inevitable: there
is no a priori reason why in every language possessing a verbal category that
could be called ‘aspect’, the same semantic functions should be expressed
through that category. It could as well be that such a category has a num-
224 ber of language-specific realisations.2 Nor are there | any a priori reasons for
the often applied method of assigning, within one language, the same aspec-
tual value to all forms formed upon a certain verbal stem, since in actual
usage these forms may occur in quite different syntactic and semantic envi-
ronments.

1.2 Aspect versus Tense

The interrelationship of aspect and tense has been, and still is, one of the most
vexing problems of the study of the verb. Very often, the one is explicitly sep-
arated from the other, which mostly means that aspect is given a much more
prominent place than tense.3 As an example, take the remarks of Schwyzer and
Debrunner in their Greek grammar (2, 248): ‘Lat. cantabat und cantavit, frz.
il chantait und il chanta, griech. ἐδίδου ἔδωκεν δέδωκεν ἐδεδώκει gehen alle auf
die Vergangenheit; ihr Unterschied kann also nicht im rein Zeitlichen liegen’.4

1 Cf. the pertinent remarks by Szemerényi (1965: 161 ff.) about the definitions of aspect in
Robins, General Linguistics: An Introductory Survey and Hockett, Course in General Linguis-
tics. What is perhaps more remarkable is that not even for a much-studied language like Greek
there exists agreement among scholars; cf. the remarks in Humbert (1960: 144).—For a history
of the difficult term ‘aspect’ cf. Mazon (1913).
2 This position is taken by those who adhere to the so-called ‘contrastive aspectology’, e.g.
Maslov (1962) and cf. Maslov’s lecture at the XIth International Congress of Linguists (Vienna
1977). Cf. also Hettrich, pp. 10–11. Observe that some scholars would deny the usefulness of
setting up a universal category aspect altogether, cf. Szemerényi (1965: 166): ‘Within the IE
family there is no justification for using this term for any language—except Slavic’.
3 In such a case, it would seem natural to study aspect primarily in those constructions where
tense would seem to play much more clearly a subordinate role, e.g., in Greek, infinitives after
βούλομαι; final clauses, etc. In fact, in a note on p. 256 Schwyzer and Debrunner announce that
for precisely this reason they will take, as far as possible, non-indicative forms as a starting-
point for their discussion of aspect. In actual practice, however, they are far from consistent
on this point.
4 They connect this rejection of ‘das Zeitliche’ with a theory about the original situation in
a review of: h. hettrich 7

A very outspoken modern representative of this direction is Comrie, whose

statement on this point (in his recent general introduction to aspect, Comrie
1976: 3) hardly differs from the opinion quoted above: ‘The difference in French
between il lisait and il lut, or in English between he was reading and he read, is
not one of tense, since in both cases we have absolute past tense’.5 This is, of
course, where aspect comes in. As to the question which values these aspectual
forms have, there is no agreement, neither for the individual languages nor in a
more general way (see 1.1). I will refrain from enumerating all suggestions and
signal only one point. Whatever aspectual value one wants to assign to forms
as he was reading and he read, it will not do to base these values upon sen-
tences consisting of main clause and temporal clause, as Comrie does; see, for
example, p. 3: John was reading when I entered,6 and p. 4 (example with while). It
would seem a rather perverse method to eliminate tense by using constructions
that explicitly signal that temporal relationships are involved. By this method,
Comrie is forced to use rather elaborate paraphrases to explain the effect of
simultaneity that is present in a sentence like John was reading when I entered
(see his p. 5).

the Ursprache. Their remark is too curious and too indicative of a certain view of language
and its study to be only referred to, so I give the text in full: ‘Eine ausgebildete Kategorie
Tempus gab es im frühen Indogermanischen noch nicht. (–) Ein Volk und eine Sprache die
in ihrer Gegenwart aufgehen, brauchen diese nicht durch besondere Verbalformen auszu-
drücken, und die Zukunft wird ihnen wichtiger sein als die Vergangenheit. Aber statt des rein
temporalen Ausdrucks des Künftigen steht ihnen der triebhafte reichlich zu Gebote (durch
Imperativ, Optativ, Desiderativ, Konjunktiv)’ (pp. 253–254).
5 Even if this is so, why could not tense, or rather, time, also be involved in another way, viz.
as expressing relative time? There is no principal objection against taking this view, and in
actual practice an analysis that makes use of this notion in many cases recommends itself,
e.g. when temporal sentences are concerned. See the text and below p. [227]ff.
6 While discussing this sentence, Comrie introduces the notion ‘internal temporal constitu-
ency’ to explain the difference between entered and was reading: entered ‘presents the totality
of the situation referred to … without reference to its internal temporal constituency: the
whole of the situation is presented as a single, unanalysable whole … no attempt is made
to divide this situation up into the various individual phases that make up the action of my
entry’ (nor, I should add, would this have been easy, since John was reading when I was entering
is highly doubtful). On the other hand, was reading ‘make(s) explicit reference to the inter-
nal temporal constituency of the situation … in particular, reference is made to an internal
portion of John’s reading (–).’ These features, he adds, are responsible for the effect of simul-
taneity. Notice that he does not speak at all about the presence of the temporal conjunction
when.
 8 chapter 1

1.3 Subjectivity; Freedom of Choice; Sentence and Context

It is often said that in a language having aspectual distinctions, the language-
user may make a free, subjective choice out of the aspectually differing verb-
forms. Or, to put it otherwise, the choice of one form rather than another is
usually taken for granted; one seldom asks how and why the language-user has
225 come to choose form x rather than form y. This is | connected with another phe-
nomenon. Most scholars studying aspect concentrate on the sentence, or even
on the verb-forms alone; thereby, they tend to ignore that the choice of a par-
ticular verb-form may have been influenced by, and may itself exert influence
upon, other constituents in the context. As an example of this attitude, take,
again, Comrie’s remarks (1976: 4): ‘In discussing aspect it is important to grasp
that the difference between perfectivity and imperfectivity is not necessarily
an objective difference between situations. … It is quite possible for the same
speaker to refer to the same situation once with a perfective form, then with an
imperfective, without in any way being self-contradictory’. He illustrates this
point with, among others, the following sentences:

(1) John read that book yesterday; while he was reading it, the postman came

(2) Jean lut ce livre hier; pendant qu’il le lisait le facteur vint

It may well be, indeed, that read and was reading, lut and lisait refer to the
same situation of reading. It may be doubted, however, whether the language-
user had any other possibilities; in (1) it would seem that, given the presence
of John read that book yesterday, he had no alternative, in beginning a sentence
with while, but to use the progressive.7 Similarly, in the French sentence (2),
the presence of pendant que automatically triggers the imparfait, not only here
but always. The point, then, is that whereas the speaker may be referring to the
same situation in different ways, he did not have, given the specific form of the
sentences concerned, any other option. In other words, the choice is, at least in
these cases, not free. The moral is that the presence of contextual elements,
e.g. conjunctions like while and pendant que, should be taken into account.

7 In other cases, however, while + simple past is perfectly admissible, e.g. in: ‘While the mes-
sage concerning Sardis went up to the king (–), matters fell out as I will show’ (from Godley’s
translation of Hdt. 5.108.1). Observe that, since in Comrie’s theory there is no ‘internal tempo-
ral constituency’ involved with went up, the above sentence will present some difficulties to
this theory: went up does not make ‘explicit reference to the internal temporal constituency of
the situation’, this situation is, thus, not presented as sub-divisible, and yet another situation
is presented as occurring within the former situation.
a review of: h. hettrich 9

I fully agree with Hettrich’s remark on p. 17: ‘Mann kann die Bedeutung der Tem-
pusstämme nicht aus sich heraus erkennen, sondern muss sie aus dem Kontext
ermitteln’.8
To conclude this section I add a more general objection against the tenet
of freedom of choice: an aspect-theory that emphasizes this freedom could,
in the end, force us to see the verb-forms concerned as being in free varia-
tion, which would make it pointless to set up any semantic differences between
these forms.

2 Hettrich’s Analysis

As to Hettrich’s book, the author is well aware of some of the above points, and
acts accordingly. As the title of his book indicates, he explicitly | wants to study 226
aspect in larger units than single sentences. And, as he says in the introduction
(p. 9), he will try to furnish some contributions to the solution of a ‘Teilkom-
plex der Aspektproblematik’. Also, he states (pp. 10–11) that, since there is as
yet no general ‘übereinzelsprachliche’ (cross-linguistic) definition of aspect, his
objective will be to establish the functioning of present and aorist forms by
interpreting them in their context, and within the framework of the language
of one particular writer, rather than to apply the principles of any one linguistic
theory. H.’s general attitude is, then, refreshingly unpretentious.
My critical remarks will follow the general structure of the book.
The Einleitung is preceded by a survey of abbreviations that calls for some
comment. One reads: ‘PSt = Praesensstamm ausser Indikativ Praesens’. H. does
not justify his omitting the present indicative. There are, indeed, strong reasons
for its omission; H. might have referred to the discussion in Ruipérez (1954:
72). As Ruipérez remarks, the present indicative is of no use for establishing
the aspectual value of the present stem, since it is neutral as to the distinc-
tion between present and aorist: there is no opposition between the present
indicative and some other form (the opposition is ‘neutralized’, in Ruipérez’
terminology).9

8 ‘Context’, in my opinion, has to be taken in a wide sense, comprising supra-sentential units; cf.
below for Greek, pp. [231ff.]. Engl. was reading and read may also have different functions on a
supra-sentential level; cf. the interesting article by Nickel (1968) on contextual links between
sentences in English.
9 Since the pres. indic. does not partake of the opposition impf. : aorist indic., it has a position
of its own, too, when it is used as a historical present. Depending on the context a historical
present may have various, rather diverging, functions. Cf. Rijksbaron (1974a: 146–148).
 10 chapter 1

Chapter 1 is a short but useful Geschichte der Forschung. On p. 12 the author

enumerates on what points there exists general agreement about the functions
of the tense-stems; this list is correct, as far as I can see, be it somewhat sim-
plified. I give some additional remarks. It should be noted that the distinctions
made do relate only once to formal contextual features, viz. when the presence
of the negation is said to favour the choice of the present stem. In all other
cases the respective distinctions (e.g., the present stem is used to express an
enduring, as yet unfinished action, the aorist stem to express a finished action,
etc.) are mostly based upon the formal opposition between present stem and
aorist stem alone,10 without any attention being paid to possible other factors.11
It should also be noted that usually these distinctions are considered to hold
throughout the verbal paradigm, although they are established primarily for
imperfect and aorist indicative. There are, however, no a priori reasons for such
a ‘holistic’ approach: after all, the different tense stem forms serve quite differ-
ent functions; cf. also 1.1 and note 3.
On pp. 12–13 H. mentions the problem of aspectual versus temporal cate-
gories; he notes that most discussions take the primary importance of aspect
for granted, and he rightly stresses that the treatment of questions of tense/
227 time show a remarkable vagueness and lack of agreement. | As we will see, H.
himself operates extensively with temporal notions.
On pp. 13–14 H. briefly discusses the much-debated issue of the nature of the
opposition between present and aorist. This opposition has often been viewed
as privative, sometimes as equipollent. H. does not express his own attitude
towards the question whether the present or the aorist is the marked term of
the opposition (both views have been defended), nor does he make any further
use of it, and rightly so. I refer to my remarks in Lampas (1974b).
Another tricky topic is discussed on pp. 15–17: the question whether ‘sub-
jectivity’ is a useful notion in the study of aspect. H. is very suspicious on this
point, and I agree with him, see section 1.3 above.
On p. 18 Hettrich comes to his subject proper: his endeavour to find the rules
governing the use of present and aorist takes the context as a starting-point. He
bases his analysis on the theory put forward in Ruijgh (1971), viz. that aspect is

10 Observe that sometimes impf. and aor. are formally identical, e.g. in the third person sing.,
with a number of verbs that are formed from a stem ending on a nasal and the suffix-
y: πλύνω, κτείνω, ὀτρύνω, etc. Forms like ἔπλυνε, ἔκτεινε, ὤτρυνε, etc. thus would seem to
provide useful checking-material for an aspect-theory: such a theory should be able to
indicate whether such a form represents an impf. or an aor. As far as I know they never
have been used as such.
11 Cf. Hettrich’s remark cited on p. [225] above.
a review of: h. hettrich 11

connected with time, and that in the last resort it is a category of a temporal
nature (‘d’ordre temporel’). In this connection, Ruijgh introduces the notion
‘moment donné par le contexte ou la situation’; according to him the present
stem expresses a process (this term is used in a general way, and comprises also
actions and states) that is not yet finished or closed at the ‘moment donné’. The
aorist, on the other hand, expresses a process that is finished at the ‘moment
donné’. In constructions involving temporal clauses, the ‘moment donné’ is
given by the main clause: the present stem in the subordinate clause indicates
that the process expressed by the subordinate clause is simultaneous with that
of the main clause, the aorist stem that this process is anterior to that of the
main clause.12 It is clear, then, that Ruijgh, and H. after him, explicitly connects
the use of the present and aorist stem with the expression of relative time. H.’s
main objective is to check this theory with the help of a very extensive number
of temporal clauses in Herodotus,13 totalling 2719 cases.14 They consist both of

12 To demonstrate this, both H. and Ruijgh use clauses introduced by ὅτε, and translate by
‘pendant que’ and ‘während’ for ὅτε + present stem. However, ὅτε is seldom, if ever, used
with the meaning ‘while’, its primary meaning being ‘(at the time) when’, ‘while’ being
expressed by ἕως and ἐν ᾧ; cf. Rijksbaron (1976: 131ff. (ὅτε) and 139ff. (ἕως, ἐν ᾧ)). Observe
that the latter conjunctions are not discussed by H.; nor are πρίν ‘before’ and μέχρι, ἕως, ἐς
ὅ etc. ‘until’; H. does not justify these omissions.
13 On p. 22 H. states that he will take into account all clauses introduced by ἐπεί, etc., even
though sometimes they express causality rather than temporality. H. doubts whether an
independent class of causal clauses exists at all in Greek, since there is no unequivocal
‘signifiant’. The latter, however, does not entail there being no class of causal clauses at
all. Actually, several types of causal clauses or clauses involving causality have to be reck-
oned with, among which there are some that, in fact, also express temporality (e.g. ἐπεί
‘now that’). But there is also one group of ἐπεί-clauses where temporal meaning is cer-
tainly lacking, viz. those where ἐπεί has the meaning ‘(I say this) because’; H. nevertheless
groups these cases together with the other instances of ἐπεί (p. 21); see also below, p. [237]f.
For the whole subject see Rijksbaron (1976).
14 H.’s survey of numbers on p. 21 has to be handled carefully. E.g. for ἐπεί+ aorist stem he
gives a total of 59; this is, indeed, the number one arrives at by adding the examples of
Powell’s sections I.1.a and II.3 (Powell 1938). However, this raises a problem: the exam-
ples of this group differ semantically (the first group = roughly after, the second now that,
since) and it may a priori be questioned whether they should be treated as one group. I
name only one important point of difference: whereas the first group consists of exam-
ples occurring in narrative passages, the second group comprises examples occurring in
direct speech and followed, mostly, by an imperative in the main clause; semantically and
syntactically they show very much resemblance to cases of ἐπεί + present indicative in
direct speech (cf. Rijksbaron 1976: 70–71). The latter, however, are not taken into consider-
ation by H., simply because no examples whatsoever of pres. ind. are discussed (cf. p. 4).
Briefly, H. has been guided too much by morphology and too little by syntax and seman-
tics.
 12 chapter 1

temporal clauses sensu stricto, involving the conjunctions ἐπεάν, ἐπεί, ἐπειδή,
ἐπείτε, εἶτε, ὁκότε, ὅκως, ὅταν, ὅτε and ὡς, and all participial clauses formed from
present stem and aorist stem in book 7.15
In detail, H. proceeds in the following way (p. 22). First and foremost he
examines for all examples whether the use of the tenses can be explained
with the notions ‘simultaneity’ and ‘anteriority’. He at once gives away the
result: in the great majority of cases this is, indeed, possible. He adds that,
since lack of space forbids treating every example separately, he will discuss
228 a representative sample of 29 cases (ch. 2). The chapters 3–8 are | devoted to a
discussion of the examples that do not behave in conformity with the general
rules.
In reviewing chapter 2 (Beispiele aus Herodot für Ruijghs These) I will con-
centrate on the use Hettrich makes of the context. As stated above, in sentences
containing temporal clauses the temporal reference point is given by the main
clause: subordinate clauses with present stem forms are simultaneous with the
moment expressed in the main clause, those with aorist forms anterior, irre-
spective of the tense of the main verb. Generally speaking, this is attractive,
although, as we will see, another approach, that takes the subordinate clause as
a starting-point, has some advantages (see p. [230]). Notice that, ultimately, this
does not exclude these phenomena being connected with aspectual notions.16
To support his analysis, H. might have pointed to the anaphorical temporal ele-

15 H. calls the subordinate clauses and participles ‘B-Glied’ (B-member) and the main clause
‘A-Glied’ (A-member).—Of course, the theory set forth in the text in principle only applies
to sentences involving temporal clauses and participles. It is an open question whether it
also holds for other constructions. Both Ruijgh and H. are aware of its limitations (Ruijgh
1971: 252, H.: 19). See also below pp. [229 ff.].
16 Thus, simultaneity might be connected with the present stem expressing non-closedness,
anteriority with the aorist stem expressing closedness. But, anyhow, temporal notions
have to be reckoned with, since the closedness or non-closedness of an action may only
be ascertained with regard to a certain reference-point: an action is not closed as such but
at a certain time. Cf. Ruijgh (1971: 229) and below, p. [230].
There is, furthermore, a more general consideration to justify the claim that these
clauses express relative time. Herodotus writes history and is, thus, confronted par excel-
lence with events that in one way or another are situated temporally vis-à-vis each other.
The structure of his work simply presupposes the expression of such temporal relation-
ships. And, of course, this is done not only by temporal clauses, but also by other means,
like temporal prepositional groups and adverbs. It would seem to me an entirely artificial
procedure to acknowledge the existence of such expressions of relative time as: ἐν τούτῳ
τῷ χρόνῳ ‘in that time, at that moment’, μετὰ ταῦτα ‘thereafter’, πρίν ‘before’, ὕστερον ‘later’,
πρότεροv ‘earlier’, etc., but to deny this possibility to conjunctional clauses. Cf. also Ruijgh
(1971: 228).
a review of: h. hettrich 13

ments that sometimes are present in the main clause (cf. Rijksbaron 1976: 74
and 109): we find ἐπεί etc. + impf. taken up by ἐν τούτῳ (τῷ χρόνῳ) (‘in/at/dur-
ing that time’), ἐπεί etc. + aor. by μετὰ ταῦτα, (τὸ) ἐνθεῦτεν (‘thereafter’) but not
vice versa.
For Ruijgh, the context is simply the reference point in the main clause. H.,
however, is not content with this. His intention is, rather, (cf. p. 20) to take
the ‘nähere und weitere Kontext’ into consideration, which in actual prac-
tice means (cf. pp. 29–30) that he makes use of a procedure, by which he
abstracts from the linguistic data, viz. the tense forms in the temporal clauses
under discussion. He states that in the examples discussed by him on pp. 25–
29 the temporal relationship between subordinate clause and main clause
appears also without taking the tense forms into account, viz. from the natural
order of the actions described (‘aus dem natürlichen Ablauf der beschriebenen
Vorgänge’). According to him, then, the situation ‘described’ would give suffi-
cient information on the temporal relationship independently from linguistic
means, specifically, the tense forms in the subordinate clause. Thus, he remarks
on:

(3) ἐπείτε … ἦλθον [aor.] ὀπίσω παρὰ τὰς Ἀμαζόνας, ἔλεξαν αἱ γυναῖκες πρὸς
αὐτοὺς τάδε (4.115.1)

After they [the young Scythians] came back to the Amazons, the women
said to them17

that the speaking of the Amazons to the Scythians could only have occurred
after the return of the latter to them. This is essentially correct, but since | H.’s 229
formulations are rather concise, I shall try to elaborate somewhat the idea put
forward by him.
Given two events that are put into a temporal relationship, by means of
a subordinate conjunctional clause and a main clause, one will investigate
whether there are any features that indicate, independently from the tense of
the verb in the subordinate clause, whether this temporal relationship involves
simultaneity or anteriority. For an illustration cf. the English sentence:

(4) When John had bought a ticket he went to platform II

17 The translations are either borrowed from, or based upon, the Loeb-edition by A.D. God-
ley. The Greek text is that of the Oxford-edition by C. Hude and J.E. Powell.
 14 chapter 1

Abstracting from the tense in the when-clause we may note the result in the
following way:

(4a) When [John buy ticket], John went to platform II

Now, the going to the platform can, in fact, have taken place only after the buy-
ing of the ticket has been completed. Thus, when such events are put into a
temporal relationship of a specific form, in which the ‘ticket-buying’ appears in
a when-clause and the ‘going to the platform’ in the main clause, our knowledge
of the world tells us what kind of temporal relationship this is. The specific rela-
tionship involved in our example may only find formal expression in a clause
with a pluperfect, cf.:

(5) *When John bought a ticket, he went to platform II

It is important to realise, however, that it is not ‘bought’ as such that is not

suitable for expressing anteriority; this only holds for ‘bought’ in the particular
construction we are dealing with here. It is, after all, perfectly possible to say

(6) John bought a ticket and went to platform II

where the same conditions are present as to the possible relationships between
the ‘going’ and the ‘buying’.18
To return to Hettrich: having established the values of aorist and imperfect
in the way sketched above, he proceeds by transferring the results to those con-
texts (called ‘neutral’ by H.) where there are no situational indications as to the
230 kind of temporal relationship between the | events involved. In those, then, it
is exclusively by the tense of the verb of the subordinate clause that this infor-
mation is conveyed. One example:

(7) ὃς ἐπειδὴ ἡ νηῦς ἡλίσκετο [impf.] ἐς τοῦτο ἀντεῖχε μαχόμενος ἐς ὃ κατεκρεορ-

γήθη ἅπας (7.181.1)

When his ship was (being) taken, he would not give over fighting till he
was all hacked about with wounds

18 Also, bought may appear in a clause introduced by after which unequivocally expresses
anteriority.
a review of: h. hettrich 15

Here, it is only the imperfect that makes clear at what time the fighting of the
soldier occurred. An aorist would have been possible too, conveying that the
ἀντέχειν took place after the enemy got hold of the ship.19
Two further points deserve some attention in connection with the ‘moment
donné’ and the context. As to the former. H., following Ruijgh, situates the tem-
poral reference point in the main clause; thus, the actions (processes etc.) of the
subordinate clause are simultaneous with, or anterior to, that point. Of course,
this has a prima facie plausibility, but actually it is too limited, in my opinion,
for the following reasons.
In general, the subordinate clause precedes the main clause and, thus, comes
first, in the linear structure of the text. It would seem counterintuitive not to
assign a proper value to the tense in the subordinate clause, but to make this
value depend upon the presence of an element that still has to come. Also,
because usually the information of the subordinate clause refers back to the
preceding context, whereas the information of the main clause is completely
new (cf. Rijksbaron 1976: 70 and notes), it would seem natural to take the for-
mer as a reference-point for the latter, and not the other way round. Now this
may seem somewhat far-fetched for the sentences under discussion, since the
first clause is formally subordinate to the second. However, using the ‘moment
donné’-criterion for other imperfects and aorists, viz. those that are not subor-
dinate to another clause, one runs into difficulties. One would have to indicate,
for a given non-subordinate imperfect or aorist, that it is simultaneous with,
or anterior to, a particular ‘moment donné’. The latter, however, may not be
present in the immediate context; in fact, it may be given much later (see (11)
below), whereby our first verb-form for a long time would remain temporally
unspecified. More important, the ‘moment donné’ may not be present at all,
as happens sometimes with aorists: with these, the notion ‘anterior to a given
moment’ is irrelevant (see (15) and (16) below).
For these reasons, then, I think it preferable to start, within a given text, from
a tense-form T1, to assign a certain value to this form, and to | take it as the 231
reference-point for any tense-forms T2, T3 etc. that may follow T1 (instead of
taking, e.g., T3 as the reference-point for T1).
Returning to our subordinate clauses, I suggest the following definition for
the value of the imperfect (presently I confine myself to this): an imperfect in
an ἐπεί etc. clause, referring to an unfinished action, signals that the actions
to be presented in the main clause stand in a relationship of simultaneity to

19 Notice, as is also done by H., p. 31, that by the explanation put forward here nothing is said
about the verb of the main clause.
 16 chapter 1

the action of the subordinate clause. My point may be briefly illustrated by the
following English clause:

(8) When John bought a ticket …

When confronted with such a clause, we are informed that the action to be
mentioned the main clause is simultaneous with that of the ‘buying’, for exam-
ple: he saw a 10-pound note on the floor. To put it otherwise, the imperfect fur-
nishes a framework within which other actions may occur.20
As to the second point, although H. claims that he makes use of the con-
text in explaining the values of imperfect and aorist, he actually is taking into
account the semantic properties of the situations described in subordinate
clause and main clause, which I would view as ‘knowledge of the world’ rather
than as ‘context’. Also, H. is concerned primarily with sentences; as he remarks
on p. 23, he does so on purpose, and he warns the reader not to draw rash con-
clusions as to the uses of the tenses in other constructions.
H. himself applies the results arrived at for sentences consisting of tem-
poral clause and main clause, to suprasentential units in four different pas-
sages, viz. Hdt. 7.184 and 2.121β–γ.1 (on pp. 40–43 and 89–92, respectively), and
Homer Iliad 16.101–111 and 5.297–310 (on pp. 84–89), the objective being to show
that temporal relationships are also involved when the various verb-forms are
not formally interconnected, as they are in subordinate clause/main clause-
constructions, viz. by the conjunction. Generally speaking, his analysis of these
texts is convincing, but the choice of Hdt. 7.184 is not a very fortunate one, since,
contrary to what H. states on p. 40, we are not dealing here with ‘events’ and
‘series of actions’ but, rather, with a state, viz. a description of the composition
and the size of the Persian army.
Be that as it may, in my opinion H. still makes use of the context in a too lim-
ited way. I now come back to my remarks on p. [228]. There, I suggested that
an imperfect in an ἐπεί etc. clause signals that the action to be mentioned in
the main clause occurs simultaneously with that of the subordinate clause; or,
232 in other words, that the imperfect furnishes a | framework within which other
actions may occur. I suggest that this value of the imperfect is also present in
other constructions.21 The aorist, on the other hand, does not have this value;
or, in positive terms, the aorist, referring to a finished action, signals that any

20 To be sure, within certain limits, cf. pp. 6–7.

21 My claim concerns strictly the imperfect, not the present stem as such. My ‘framework’
shows some resemblance to the ‘Inzidenzschema’ of e.g. Strunk (1971: 201ff.).
a review of: h. hettrich 17

actions to be mentioned thereafter are posterior to that referred to by the aorist.

Some illustrations now:

(9) … ἐπὶ δὲ Αἴγυπτον ἐποιέετο [impf.] στρατηλασίην (2.1.2)

… and he made an expedition against Egypt

According to the principles laid down above, the imperfect, referring to an

unfinished action, signals that other actions may be expected to be mentioned
that are simultaneous with the action referred to by the imperfect. In this case,
this expectation is thwarted: these actions do not follow immediately, and for a
very specific reason: Herodotus at this moment cuts off the narrative proper
and starts the lengthy description of Egypt. It is only in the first chapter of
book 3 that he returns to the narrative: there we again find an imperfect, which,
once again, indicates that other, simultaneous actions are to follow. Again, how-
ever, Herodotus interrupts himself; ἐστρατεύετο ‘he marched’ is taken up in 3.4.1
by ὁρμημένῳ δὲ στρατεύεσθαι ‘when he started to march’. Only now are the other
actions mentioned: φράζει, etc.
Some instances of στρατιὴν ποιέεσθαι and similar verb phrases corroborate, I
think, the functioning of the imperfect as sketched above. Cf.:

(10) Κροῖσος δὲ … ἐποιέετο [impf.] στρατηίην ἐς Καππαδοκίην (–). παρασκευαζο-

μένου δὲ Κροίσου στρατεύεσθαι ἐπὶ Πέρσας, τῶν τις Λυδῶν … συνεβούλευσε …
(1.71.1)

Croesus prepared an expedition against Cappadocia (–). But while he was

preparing to march against the Persians, a certain Lydian … advised him
…22

The giving of the advice falls within, and is thus simultaneous with, the prepar-
ing of the expedition. Notice that the actual marching against the Persians is
not mentioned before ch. 75.2 ἐστρατεύετο ἐς τὴν Περσέων μοῖραν where the
action referred to by ἐστρατεύετο serves as the framework for the arrival at the
Halys, mentioned in the next sentence.

22 Godley translates: ‘Croesus invaded Cappadocia’, wrongly. In itself στρατηίην ἐποιέετο may
mean ‘(he) invaded’, cf. (11), but the actual invasion of Cappadocia is not mentioned until
75.3.
 18 chapter 1

(11) Ἅρπαγος … ἐποιέετο [impf.] στρατηίην ἐπὶ Κᾶρας καὶ Καυνίους καὶ Λυκίoυς
[long interruption] [174.1] οἱ μέν νυν Κᾶρες ἐδουλώθησαν ὑπὸ Ἁρπάγου
233 (1.171.1) |

Harpagus made an expedition against the Carians, the Caunians and the
Lycians (–). The Carians then were enslaved by Harpagus

Notice that in this case ἐποιέετο στρατηίην refers to the expedition itself, of
which the enslaving of the Carians forms part; the other stages are mentioned
in chapter 176. Forms like ἐποιέετο στρατηίην are thus ambiguous; the context
has to decide whether the ‘preparation’ interpretation suits best, or the ‘expe-
dition’ interpretation.
Two more examples:

(12) Κῦρος … ἐποιέετο [impf.] ἐκ τοῦ ἐμφανέος ἐπὶ τοὺς Μασσαγέτας στρα-
τηίην (–). ἔχοντι δέ οἱ τοῦτον τὸν πόνον πέμψασα ἡ Τόμυρις κήρυκα ἔλεγε τάδε
(1.205.2–206.1)

Cyrus … openly prepared to attack the Massagetae. (–) But while he was
at his work Tomyris sent a herald to him with this message

(13) Ἀρισταγόρης δὲ (–) ἐποιέετο [impf.] στρατηίην ἐς Σάρδις. αὐτὸς μὲν δὴ οὐκ
ἐστρατεύετο ἀλλ᾽ ἔμενε ἐν Μιλήτῳ. (–). ἀπικόμενοι δὲ τῷ στόλῳ τούτῳ Ἲωνες
ἐς Ἔφεσον πλοῖα μὲν κατέλιπον … (5.99.1–100.1)

Aristagoras planned a march against Sardis. He himself went not with the
army, but stayed still at Miletus (–). The Ionians, having come in this for-
mation to Ephesus, left their ships …

Of course, the use of the imperfect has to be compared with that of the aorist.
The latter entirely lacks the features found for the imperfect; the aorist, refer-
ring to a finished action, does not furnish a framework within which other
actions may occur. Any action to be mentioned following the aorist form occurs
after the action referred to by the aorist. Since the aorist refers to a finished
action, and accordingly does not form a framework for other actions, it some-
times refers to an action that stands on its own, having no temporal rela-
tionship whatsoever with other actions, cf. examples (15) and (16) below; this,
I think, cannot be conveyed by the imperfect. Some examples of the aorist
are:
a review of: h. hettrich 19

(14) Καμβύσεω δὲ ἐπ᾽ Αἴγυπτον στρατευομένου ἑποιήσαντο [aor.] καὶ Λακεδαι-

μόνιοι στρατηίην ἐπὶ Σάμον τε καὶ Πολυκράτεα … ὃς ἔσχε Σάμον ἐπαναστάς.
[Long story about Polycrates.] [3.54.1] Λακεδαιμόνιοι δὲ στόλῳ μεγάλῳ ὡς
ἀπίκοντο, ἐπολιόρκεον Σάμον (3.39.1) | 234

While Cambyses was attacking Egypt, the Lacedaemonians too made war
upon Samos and Polycrates; he had revolted and won Samos … (–). Now
after the Lacedaemonians came with a great host, they laid siege to Samos

Ἐποιήσαντο, referring to a finished action, unambiguously indicates that the

Spartan expedition reached Samos, and that no further actions are to be
expected that might fall within the ‘marching’, and, thus, be simultaneous with
that action. In fact, ἀπίκοντο in 3.54.1 simply takes up the action referred to by
ἐποιήσαντο στρατηίην ἐς Σάμον; in 39.1 no intervening actions occur.

(15) ταύτην πρώτην στρατιὴν ἐς τὴν Ἀσίην Λακεδαιμόνιοι Δωριέες ἐποιήσαντο

[aor.] (3.56.2)

This was the first expedition to Asia made by Dorians of Lacedaemon

This sentence concludes Herodotus’ story about the Spartan expedition to

Samos. No further actions connected with this expedition are mentioned there-
after.

(16) [When still young, Cambyses says to his mother that, once he has grown
up, he will turn Egypt upside down] … ἐπείτε ἀνδρώθη καὶ ἔσχε τὴν βασι-
ληίην, ποιήσασθαι [aor.] τὴν ἐπ᾽ Αἴγυπτον στρατηίην (3.3.3)

when he grew up and became king, he made the expedition against Egypt

Cambyses’ expedition to Egypt is, at his moment, only succinctly mentioned,

no details are given.
It will have become clear, I hope, that at least with verbs or verbal groups
like στρατηίην ποιέεσθαι the use of imperfect and aorist has something to do
with the presentation of actions as finished or not finished, and thereby, more
importantly, with the possibility to furnish (impf.) or not furnish (aor.) a frame-
work for other actions. By these values, then, imperfect and aorist belong to the
inherent means of the language to structure narrative.
Coming back to H.’s book, I pass over to chapter 3 Zur genaueren Bestimmung
des A-Gliedes (for this term cf. note 15). In this chapter, H. discusses a number of
 20 chapter 1

examples where the theory put forward by him as to the relationship between
235 subordinate clause and main clause (cf. p. [227f.] | above) gives some problems;
on p. 34 he proposes a solution to these problems. Whereas this proposal may
hold for the examples discussed by him on pp. 33–36, it certainly does not for
the examples on pp. 32–33. I will comment upon one of his examples.

(17) ὁ γὰρ Ἱστιαῖος … ἐτύγχανε [impf.] τοῦτον τὸν χρόνον ἐὼν ἐν Σούσοισι, ὅτε οἱ
Νάξιοι ἦλθον (5.30.2)

For Histiaeus … happened to be at Susa when the Naxians came

H. remarks:

Es handelt sich hier um ein Inzidenzschema, ἐτύγχανε ἐών stellt eine ver-
laufende Handlung dar, in die das ἦλθον hineinstösst. Die Hauptsatzhand-
lung ἐτύγχανε ἐών ist hier im Widerspruch zu dem in 2.7 Festgestellten das
B-Glied [i.e. the ‘subordinate’ clause], dessen Tempusstamm vom Zeitver-
hältnis zur untergeordneten A-Handlung ἦλθον abhängt.

According to H., then, from the point of view of the information structure the
main clause is formed by ἦλθον and the subordinate clause by ἐτύγχανε ἐών.
He translates: ‘Denn Histiaios war zu dieser Zeit gerade in Susa, als die Naxier
ankamen.’ His analysis, however, is not correct, since the arrival of the Naxians
is not an action that ‘hineinstösst’ in the state (rather than ‘verlaufende Hand-
lung’, as H. has it) referred to by ἐτύγχανε ἐών. Their arrival has been mentioned
before (ἀπίκοντο ἐς Μίλητον 30.1) and more or less as a parenthesis Herodotus
adds where Histiaeus was at that time, which explains why they did not find
Histiaeus in Miletus, of which city he was tyrant. The translation, then, should
be: ‘For Histiaeus happened to be in Susa at the time of the arrival of the Nax-
ians [viz. in Miletus]’. Cf. for this and similar examples Rijksbaron (1976: ch. 7,
note 1, p. 198).
H. proposes to call these ὅτε-clauses ‘invers’, following the terminology in
Latin grammars (e.g. Kühner and Stegmann 1912: 2, 338); but this will not do.
Although syntactically (17) resembles Latin ‘cum-inversum’–clauses, in that,
contrary to normal usage, the subordinate clause follows the main clause,
its semantic characteristics are different. In Latin, the different order of con-
stituents reflects a difference in information structure: the information that
refers back to the preceding context (the ‘old’ information), which usually is
given by the subordinate clause (cf. for Greek Rijksbaron 1976: 67ff.), in this
case is presented by the main clause, whereas the new information is given
a review of: h. hettrich 21

in the subordinate clause. Or, in the words of Kühner and Stegmann (loc. cit.):
‘… der Hauptgedanke | hingegen, nämlich das neue Moment der Erzählung, in 236
einem Nebensatz mit cum nachfolgt’. For an English example cf.:

(18) Having entered, Peter decided to go straight to the apartment. He walked

through the corridor and was about to enter the elevator, when he saw
John

The Greek examples adduced by H. do not have these semantic features.23

In chapter 4 (Kriterien für den T(empus)st(amm)-Gebrauch in negierten B-
Gliedern) H. discusses the important subject of negation and aspect/tense. It
has often been noticed that there is a clear relationship between the presence
of the negation and the choice of the tense. In Herodotus this is particularly
conspicuous, at least with subordinate temporal clause and participles, as the
following numbers show: in non-negated clauses the present stem and the
aorist stem are distributed at a ratio of ca. 47 : 53, in negated ones, however,
at a ratio of 83 : 17. An example is:

(19) ἐπεὶ … οὐδεὶς ἐφαίνετό [impf.] σφι ἐπαναγόμενος … ἐνθαῦτα Λευτυχίδης ὑπὸ
κήρυκος προηγόρευε τοῖσι Ἴωσι λέγων (9.98.2)

When they found nobody was putting out to meet them … Leotychidas
made the following proclamation to the Ionians by the voice of a herald

In the ἐπεί-clause reference is made to a situation in which, properly speak-

ing, no action occurs; there is, thus, a situation of ‘not being found, not being
seen’, which still continues when the action of the main clause sets in, and is,
thus, simultaneous with the latter (cf. H. p. 47); so far I agree with H.,24 see my

23 In fact, such ‘inverse’ constructions seem to have been extremely rare in Greek. Ruijgh
(1971: 485 ff.) mentions Hom. Il. 16.11; cf. also Rijksbaron (1976: ch. 7, note 8 on p. 200).
24 H. remarks (p. 47): ‘Wären sie nicht negiert, so bestände wohl überall Vorzeitigkeit’. On
this point H. seems to have been misled by his material. His remark does seem to apply, in
fact, to the examples discussed by him on pp. 46–47, but there only punctual or termina-
tive verbs are involved, which, when not negated, have a preference for the aorist (i.e. in
subordinate temporal clauses). It does not apply, however, to other verb-types, e.g. ἐβού-
λετο in 1.165.1, ἐπειρᾶτο in 1.77.1, τολμώντων in 7.10.1 ἐούσης in 7.99.1, ἀρχόμενοι in 7.103.3. In
such cases the negation would not seem to have favoured the choice of the impf.: with-
out it, there would have been simultaneity, too. This is especially clear in 7.103.3, where an
anterior aorist participle instead of ἀρχόμενοι would have been very strange, if possible at
all.
 22 chapter 1

remarks in Rijksbaron (1976: 112ff.), where also some specific semantic features
of these clauses are discussed.
Now there are also some examples where the aorist occurs alongside the
negation. These are discussed by H. on pp. 47–49; he suggests some explana-
tions, but rejects these, and avows (p. 48): ‘Mann wird schliesslich nicht umhin
können, hier eine gewisse Freiheit im Tempusstamm-Gebrauch anzuerkennen’.
In my opinion, however, the use of the aorist can be explained in the following
way.
In ἐπεὶ οὐ + imperfect clauses often an iterative nuance is present, especially
with punctual and terminative verbs, like, for example, φαίνομαι ‘appear, turn
up’ and πείθω ‘persuade’. In (19), too, this may appear: the ascertainment that
237 ‘nobody was putting out against them’ may have | been made at several, other-
wise unspecified, moments; for some other cases see 2.118.4 (φαίνομαι), 5.104.2
(πείθω) and 9.20 (καταβαίνω ‘descend’). Now the use of the aorist of such verbs
(I confine myself to these) in ἐπεὶ οὐ clauses seems to indicate that the action
referred to occurred only once; it is, thus, semelfactivity we are dealing with
here. In view of the information presented in the preceding context this often
is not surprising. Cf.:

(20) Μαζάρης … ὡς οὐκ εὗρε [aor.] ἔτι ἐόντας τοὺς ἀμφὶ Πακτύην ἐv Σάρδισι,
πρῶτα μὲν … ἠνάγκασε (1.157.2)

When Mazares found Pactyas’ followers no longer in Sardis, he first of all

compelled …25

In 1.157.1 we have been told that Pactyas has left for Cyme. The aorist in (20)
indicates that Mazares at his arrival in Sardis at once noticed Pactyas’ absence,
which, when seen from the point of view of the reader, is not unnatural, given
the information of the preceding section. Or, to put it otherwise, the aorist sig-
nals that Mazares did not have to look for Pactyas; the latter would have been
conveyed by the imperfect, as in:

(21) [The Persians are for the second time unsuccessful in battle] ὡς δὲ οὐδὲν
εὕρισκον [impf.] ἀλλοιότερον οἱ Πέρσαι ἢ τῇ προτεραίῃ ἐνώρων, ἀπήλαυνον
(7.212.2)

25 The other instances are: 1.214.1; 2.134.4; 5.98.4.

a review of: h. hettrich 23

When the Persians found the Greeks in no way different from what the
day before had shown them to be, they drew off from the fight

εὕρισκον signals that the Persians noticing their ill luck occurred at several occa-
sions. Here, too, we have an indication in the preceding context for this: the
Greeks fought ἐν μέρει ἕκαστοι ‘each in turn’.26
To conclude this section, I present some remarks about H.’s material. The
only criterion for listing an item seems to have been the presence of the nega-
tion. However, this group is not as homogeneous as H. would have us believe.
It contains, for instance, many clauses that are not strictly temporal, e.g. (par-
ticiples) 7.16.1 (οἷα οὐκ ἐξιεύμενος, causal), 7.86.2 (οὐ λειπομένας, attributive),
7.89.1 (οὐκ ἐχούσας, attributive), 7.211.2 (ἅτε οὐκ ἔχοντες, causal), 7.211.3 (ἐν οὐκ
ἐπισταμένοισι μάχεσθαι, participle used as a substantive), 7.222 (οὐ βουλόμενοι,
predicative, not strictly temporal, but | adjective-like; cf. ἀέκοντες which pre- 238
cedes), 7.10.η.1 (οὐκ ἐόντων, attributive), 7.166 (οὔτε ζώντα, predicative) etc. Also
ἐπεί-clauses that follow the main clause and have to be interpreted as clauses
of motive, viz. 1.141.2; 3.36.3; 4.43.1; 4.117; 3.88.1; 9.80.2 (cf. Rijksbaron 1976: 78ff.)
Whereas with some of the participial clauses temporal relationships may be
involved (e.g. with ἅτε-clauses), this certainly does not hold for the ἐπεί-clauses
listed above: in these there is no direct temporal relationship between main
clause and ἐπεί-clause; in fact, the latter does not function as a subordinate
clause at all (cf. Rijksbaron 1976: 71–72, 78ff.). Thus, in, for example, 9.80.2 ἐσκύ-
λευον in no way is the main verb to ἐγίνετο in the ἐπεί-clause.
In chapter 6 (Von der Verbalsemantik abhängige Besonderheiten im Gebrauch
des PSt) H. discusses a heterogeneous group of examples where the present
stem is used, although at first sight it would seem that anteriority is involved. A
relatively large subgroup of those is formed by verba dicendi (in a wide sense,
also verb like ἀποπέμω ‘send away’ are included), or, in the words of Blass (1889:
410): ‘Verba, Handlungen bezeichnend, die ihr Ziel und ihre Vollendung in dem
Thun eines anderen haben’.27 Following a suggestion by Svensson (1930), H.
proposes the following values for the present and aorist stems of these verbs
(p. 61): the present stem expresses:

26 H. too proposes a semelfactive interpretation for the aorist, but in a rather curious way
(p. 48). He suggests that the aorist perhaps occurs in those negated clauses where the
verbform is in the singular. This not being the case, he rejects a semelfactive value for the
aorist altogether.
27 Of course, it may a priori be questioned whether the verbs brought together under the
heading ‘verba dicendi’ really form a homogeneous group. There are, after all, with regard
to the reaction demanded, quite some differences between verbs like ἀγγέλλω ‘report’ and
αἰτέω ‘ask’ or καλέω ‘call’.
 24 chapter 1

(a) the ‘Akt des Sagens’ (the ‘action of speaking’), including the enduring sit-
uation that is created by this action, up to the reaction of the addressee;
(b) the action of speaking alone in its extension.
The aorist stem, on the other hand, expresses the action of speaking as such,
without emphasizing its duration.
It will be noticed that H. does not speak about the reaction of the hearer
in connection with the aorist. In his additional remarks on p. 62, however, he
opposes Present stem (a) above to the aorist and says: the speaker may use
either the present stem in sense (a), or the aorist stem, the difference being,
in his opinion, that the present stem expresses the going on (‘Weiterwirken’,
a rather impressionistic term) of the speaking, the calling etc. at the time the
main action begins, whereas the aorist expresses that the action of speaking,
calling etc. is at that moment finished.
In my opinion the above definitions are rather rough; the following distinc-
tions, at least, have to be reckoned with:
(1) When the present stem of a verbum dicendi that is directed towards elicit-
239 ing a reaction from the hearer is used in a subordinate temporal | clause,
the demand, calling, etc. is presented as ‘open’, ‘unfinished’. This has the
following consequences in actual usage:
(a) when a positive reaction occurs, the narrow temporal relationship
between the two actions is emphasized. That is, the actual utter-
ing of the demand etc. is anterior to the reaction, but by the use of
the present stem the two actions are linked as closely, as ‘simultane-
ously’, as possible;
(b) when a negative reaction occurs, the factors causing the failure of
the demand etc. are mentioned.
(2) When the aorist stem of a verbum dicendi that is directed towards elicit-
ing a reaction from the hearer is used in a subordinate temporal clause,
the demand, calling, etc. is presented as ‘closed’, ‘finished’. This has the
following consequences in actual usage:
(a) often the ‘closedness’, of the demand etc., has the specific implica-
tion that the hearer reacts positively;
(b) when a positive reaction is mentioned explicity, the two actions
are presented without special emphasis on the narrow relationship
between them; this occurs, for example, when the reaction is indi-
rect, i.e. not directed to the person asking;
(c) sometimes no reaction occurs; in that case the demand, etc. is pre-
sented as such: possible reactions are considered irrelevant.28

28 Of course, when the subject of the subordinate clause and the main clause are co-
a review of: h. hettrich 25

These caracteristics may be schematized as follows:

reaction

positive no reaction
negative
expl. impl.

imperfect + – + –

aorist + + – +

From this schema it appears that both the imperfect and the aorist are used
when a reaction is explicitly expressed. Above I have tried to assess the differ-
ence between them, cf. (1a) and (2b). This analysis starts from the assumption
that in Greek similar phenomena are involved with certain verba dicendi as in,
for example, Dutch and English. Consider the following sentences (some exam-
ples illustrating (1b), (2a) and (2c) will be given below):

(22) Toen Jan me vroeg of ik zou komen, zei ik ja

(23) When John asked me whether I would come, I said yes

(24) ?Toen/nadat Jan me gevraagd had of ik zou komen, zei ik ja | 240

(25) ?When/after John had asked me whether I would come, I said yes

(26) Toen ik Jan riep, schrok hij op

(27) When I called John, he gave a start

(28) ?Toen/nadat ik Jan geroepen had, schrok hij op

(29) ?When/after I had called John, he gave a start

referential, the notion ‘reaction’ usually will be irrelevant; cf. examples (30)–(31) below,
also p. [243] on Greek. This point seems to be missed by H. when he classifies, e.g., ἀπε-
πέμπετο in 3.50.2 or φράζοντες in 7.168,1 as “Handlung ohne Fortwirken” (p. 107).
26 chapter 1

(30) Toen ik Jan geroepen had, liep ik verder

(31) When/after I had called John, I walked on

(32) *Toen ik Jan riep, liep ik verder

(33) *When I called John, I walked on

The imperfect in (22) and (23), and in (26) and (27), suggests that the subject of
the main clause reacted promptly to the request. Apparently, such a close rela-
tionship between actions like ‘asking’ and ‘answering’ is a normal feature, for
sentences like (24)–(25) and (28)–(29), where it is indicated that we are dealing
with two separate, strictly successive, actions, are rather strange, if possible at
all. On the other hand, when no reaction is involved, as in (30)–(31), we are deal-
ing with a strict sequence of actions, in which case the pluperfect is normal. In
this case the use of the imperfect would seem to be impossible, cf. (32)–(33). Cf.
also p. [229]f. above on conditions on the possible relationships between two
actions.
I think, then, that some such semantic differences play a role, too, in distin-
guishing present stem and aorist stem of certain verba dicendi in subordinate
temporal clauses in Greek.29 Of course, it will be difficult to find examples
where these rather subtle differences can be shown to be operative, the more so
because formal contextual features—other than the verb-forms—that might
distinguish present stem and aorist stem, would seem to be lacking. The values
discussed above may be present in, for example:

(34) οἱ δὲ Πεισιστρατίδαι … ἐπεκαλέοντο [impf.] ἐκ Θεσσαλίης ἐπικουρίην, (–)

Θεσσαλοὶ δέ σφι δεομένοισι [ptc. praes.] ἀπέπεμψαν … χιλίην τε ἵππον …
(5.63.3)

But the sons of Pisistratus … sent to ask help from Thessaly (–). The Thes-
salians at their entreaty sent a thousand horsemen …

(35) οἱ δὲ Κορίνθιοι (–) Ἀθηναίοισι διδοῦσι δεομένοισι [ptc. praes.] εἴκοσι νέας
(6.89.1)

29 I avow that in some cases it is very difficult to detect any significant semantic and syntac-
tic differences, cf. e.g. κελευόντων τῶν Περσέων in 6.13.1 and κελευσάντων Περσέων in 6.25.1;
τὰ ἀγγελλόμενα in 3.126.2 and τὰ ἀγγελθέντα in 8.76.1.
a review of: h. hettrich 27

The Corinthians (–) at the entreaty of the Athenians sent twenty ships to
them | 241

(36) Αἰγινῆται δὲ (–) Θηβαίων δεηθέντων [ptc. aor.] πόλεμον ἀκήρυκτον Ἀθηναί-
οισι ἐπέφερον (5.81.2)

But the Aeginetans (–) at the entreaty of the Thebans made war on the
Athenians without sending of herald

(37) οὐκ ἔχω δ᾽ ἀτρεκέως εἰπεῖν οὔτε εἰ ἦλθον μὲν ἀπολέοντες τοὺς Φωκέας δεηθέν-
των [ptc. aor.] Θεσσαλῶν … (9.18.2)

Now I cannot with exactness say if they came at the Thessalians’ desire to
slay the Phocians …

(38) ταῦτα Σκυθέων ἐπαγγελλομένων [ptc. praes.] ἐβουλεύοντο οἱ βασιλέες (4.119.1)

Such being the message of the Scythians the kings took counsel

(39) ταῦτα ἐπαγγειλαμένου [ptc. aor.] μετὰ ταῦτα οὐδὲν ἐγίνετο πλοίων σπανιώτε-
ρον (8.25.1)

After this proclamation, there was nought so hard to get as a boat30

Notice that in (34) and (35) there is a more direct link between the request and
the reaction than in (36) and (37), in that in the former two the sending of help
directly concerns the persons asking, whereas in the latter the actions of the
main clause are directed towards other persons. As to (38) and (39), there may
be a difference of situation, too. In (38) the taking of the counsel directly con-
cerns the Scythian messengers: they are waiting for a reaction. In (39), on the
other hand, the reaction to the proclamation is not directed towards Xerxes’
herald, but to Xerxes himself, who is the eventual person asking.
I will now give some examples of the other values assigned above to present
stem and aorist stem.

30 Anaphoric temporal modifiers like μετὰ ταῦτα are not found following a clause contain-
ing a present stem form, and it may be asked whether they would be acceptable there. Of
course, if the value assigned by me to the present stem is valid, the occurrence of such
expressions of anteriority is not to be expected.
 28 chapter 1

(1b) present stem, no positive reaction to the request occurs and the factors
causing the failure of the request are mentioned:

(40) ὡς δὲ … τὸν Ἄμασιν ἐκάλεε [impf.] ὁ Πατάρβημις, ὁ Ἄμασις (ἔτυχε γὰρ ἐπ᾽
ἵππου κατήμενος) ἐπάρας ἀπεματάϊσε καὶ τοῦτό μιν ἐκέλευε Ἀπρίῃ ἀπάγειν
(2.162.3)

When Patarbemis … summoned Amasis, he, being then on horseback,

lifted his leg and made an unseemly gesture, and bade the messenger take
242 that token back to Apries |

(41) ὡς δὲ … Λευτυχίδης … ἀπαίτεε [impf.] τὴν παραθήκην οἱ [δ᾽] Ἀθηναῖοι προ-

φάσις εἶλκον οὐ βουλόμενοι ἀποδοῦναι (6.86.1)

So when Leutychides … demanded that what had been entrusted be

restored, the Athenians, being loath to restore it, made excuses

(2a) aorist stem, positive reaction implied:

(42) Ξέρξης δὲ καλέσας [ptc. aor.] Δημάρητον εἰρώτα … (7.234.1)

Xerxes sent for Demaratus and questioned him …

(43) καλέσας [ptc. aor.] Περσέων ἄνδρας δοκίμους πεντεκαίδεκα ἐνετέλλετό σφι …
(3.135.1)

Having called to him fifteen notable Persians he bade them …

(44) … Εὔρυτον … αἰτήσαντά [ptc. aor.] τε τὰ ὅπλα καὶ ἐνδύντα … (7.229.1)

(It is told) … that Eurytus called for his armour and put it on …

(45) ὁ Δωριεὺς … αἰτήσας [ptc. aor.] λεὼν Σπαρτιήτας ἤγε ἐς ἀποικίην (5.42.2)

Dorieus … asked the Spartans for a company of folk, whom he took away
as colonists

(2c) aorist stem; possible reactions are considered irrelevant:

a review of: h. hettrich 29

(46) ἐπαγγείλας [ptc. aor.] δὲ καὶ Λακεδαιμονίοισι παρεῖναι ἐς χρόνον ῥητόν (1.77.2)

Having summoned the Lacedaemonians, too, to join him at a fixed time31

(47) Ἀθηναῖοι δέ σφεας ἐπὶ ῥητοῖσι ἐδέξαντο σφέων αὐτῶν εἶναι πολιήτας, οὐ πολλῶν
τεων καὶ οὐκ ἀξιαπηγήτων ἐπιτάξαντες [ptc. aor.] ἔργεσθαι (5.57.2)

The Athenians received them as citizens of their own on set terms, debar-
ring them from many practices not here deserving mention

(48) τοῖσι Ἴωσι παραγγείλας [ptc. aor.] πλέειν ἐς τὸν Πόντον … (4.89.1)

Darius … having bidden the Ionians to sail into the Pontus …

I conclude this section with some remarks on H.’s material, since this is open
to some criticism. | 243
(1) H. would seem to underestimate the possibilities of explaining the use
of the present stem otherwise than by using notions like ‘Fortwirken’
and ‘Reaktion’. This holds especially for iterative interpretations; these are
at least possible in a number of cases; I mention 7.1.2 περιαγγελλομένων
(notice κατὰ πόλις in the main clause); 8.70.1 παρήγγελλον; 7.136.1 κελευόν-
των.
(2) Sometimes no aorist form of the verbs presented by H. in the present stem
occurs in the same constructions, viz. subordinate temporal clause and
participle. This holds e.g. for ὑποτείνω, φράζω, χρηστηριάζομαι, ἐπισκήπτω.
On principal grounds, such verbs provide doubtful illustrations of specific
values of the present stem.
(3) Sometimes H. lists verbs that are no verba dicendi: ἐκλεγόμενος in 7.6.4
(‘pick out, choose’); συλλεγομένων in 7.145.1 and 7.213.2 (‘assemble’); ἐπιλε-
γόμενος in 7.49.5 and 7.220.4 (‘consider’).
(4) On p. 107 H. groups a number of examples under the heading ‘Handlung
ohne Fortwirken’; for most of them this is correct, not, however, for ἐπι-
λέγομαι (cf. the preceding remark), nor for καλέοντα in 7.15.1, which is a
future form. He does not mention, however, the feature that is essential
for this interpretation, viz. that the subject of the subordinate clause, or

31 Notice that ἐπαγγείλας does not refer to an actual summoning, but to a planned one: it
refers to an action that is anterior to a future action ἐνένωτο … ἅμα τῷ ἦρι στρατεύειν. There
could not possibly have been a reaction.
 30 chapter 1

the participle, is co-referential with the subject of the main clause; of

course, in these cases a notion like ‘reaction of the addressee’ is irrele-
vant.32 Cf. also note 28.
On pp. 63–69 H. discusses a number of other verbs that, according to him,
exhibit the semantic characteristics of verba dicendi, notably γίνομαι, μίσγω (or
rather, συμμίσγω), εἶμι, ἵζομαι, πυνθάνομαι, νικῶ, καταλαμβάνω, ἑσσοῦμαι, δέκομαι,
αἱρέω, ἥκω, εὑρίσκω, οἴχομαι. With these verbs too, then, the present stem, while
referring to a finished action, often would express the ‘Fortwirken’ of the ver-
bal action. H. connects this phenomenon with these verbs being ‘non-durative’
(or, as I would prefer, terminative or punctual), which makes that, for exam-
ple, imperfects expressing simultaneity tout court occur but seldom with these
verbs.
I agree that for some of these verbs (not for all, see below) we have to reckon,
indeed, with a notion like ‘Fortwirken’. However, I add immediately that this
does not, or not solely, seem to be conveyed by the present stem, but, rather, by
the semantic aspect of the verbal lexeme (the ‘Aktionsart’). Consider the fol-
lowing sentences:

(49) ὡς δὲ ἐγίνετο [impf.] ἐν τῷ πελάγει, ἀποσιεύμενος τὴν ἐξόρκωσιν τοῦ Ἐτεάρχου

244 σχοινίοισι αὐτὴν διαδήσας κατῆκε ἐς τὸ πέλαγος (4.154.4) |

When he was on the high seas he bound the girl about with ropes and let
her down into the sea …, thus duly fulfilling the oath that he had sworn
to Etearchus

(50) ὡς δὲ ἀπὸ τῆς νήσου ἑκὰς ἐγένετο [aor.], περιελόμενος τὴν σφρηγίδα … ῥίπτει
ἐς τὸ πέλαγος (3.41.2)

And when he was far from the island, he took off the seal-ring … and cast
it into sea

H., who discusses (49) on p. 64, translates ‘als er … angelangt war (sc. und sich
dort befand)’; the latter, however, also would seem to hold for (50), which has
an aorist. I think, moreover, that H.’s translation makes ἐγίνετο ‘too anterior’; in
German, as in other languages, the imperfect of a verb like anlangen, ankom-

32 The subjects are not co-referential in 7.152.1, but this example has a structure of its own,
since λέγοντα refers back to the words mentioned in 150.2; its function is that of an attribu-
tive modifier of κήρυκα and ἀπέπεμψε is thus, strictly speaking, not the main verb to
λέγοντα.
a review of: h. hettrich 31

men (cf. Du. aankomen, komen, Engl. arrive, come), when it occurs in a temporal
clause, may receive an anterior interpretation, which is due to the punctual
semantic aspect of the verb. For some details cf. Rijksbaron (1976: 109 ff.).33 This
feature is especially noticeable with πυνθάνομαι: two examples (also mentioned
by H., p. 66):

(51) ὡς ἐπυνθάνοντο [impf.] δὲ λεγόντων αὐτῶν τῶν ἄγχιστα δέεσθαι, εἶπαν οἱ

Θηβαῖοι ἀκούσαντες [ptc. aor.] τούτων (5.79.2)

And when the Thebans learnt the message “that they must entreat their
nearest”, they said when they heard it …

(52) οἱ δὲ Παίονες ὡς ἐπύθοντο [aor.] ἐχομένας τὰς πόλιας, αὐτίκα διασκεδασθέντες

κατ᾽ ἑωυτοὺς ἕκαστοι ἐτράποντο (5.15.3)

When the Paeonians learnt that their towns were taken, they straightaway
broke and each went his own way

In (51) H. translates: ‘als die Thebaner erfuhren’ and adds: ‘(sc. und damit
wussten)’, but the latter implication is also present in (52). Notice also, in
(51), the presence of the aorist ἀκούσαντες τούτων, which takes up ἐπυνθάνοντο,
apparently without any difference of meaning.
All in all, I think we have to acknowledge a certain freedom of choice, as
regards present stem and aorist stem with verbs having a punctual semantic
aspect.34

33 Such an anterior interpretation is, however, not compulsory; it depends also on the type
of information in the main clause. For a probable ease of simultaneous ὡς ἐγίνετο + loca-
tive expression see: 3.86.1; 8.37.2; 8.94.2; 9.69.2.—For the difficulties connected with the
assignment of specific values to verbal lexemes, see Sicking (1971: 36ff.).
34 Although, then, in my opinion the notion ‘Fortwirken’ is a lexical feature of the verb, I
draw attention to an interesting phenomenon that in an unexpected way may give some
support to H.’s hypothesis about the value of the present stem. As far as I can see, in all
examples of ἐγίνετο + local expression in a subordinate clause, the action referred to in the
main clause never occurs on a spot other than that referred to in the ὡς- or ἐπεί-clause.
With the aorist, on the other hand, the action of the main clause may occur both on the
same spot as that referred to in the subordinate clause, or on another. In the latter case
the implicit result of arriving somewhere, viz. to be there, does not apply. An example of
the former relationship is (50), discussed on p. [244]; an example of the latter, καί μιν, ὡς
ἐγένετο ἐν τῷ Αἰγαίῳ, ἐξῶσται ἄνεμοι ἐκβάλλουσι ἐς τὸ Αἰγύπτιον πέλαγος (‘And when he got
in the Aegean, contrary winds drove him in the Egyptian sea’, 2.113.1). Alexander did, thus,
not stay in the Aegean sea. Possibly, then, in the case of ὡς ἐγίνετο + local expression the
impf. explicitly brings out the ‘Fortwirken’-feature of the verbal meaning, which in the
case of the aorist is implicit or absent.
 32 chapter 1

As I remarked above, I think that H.’s group of verbs showing ‘Fortwirken’ in

the present stem is too large, if this feature is connected with verbs having a ter-
245 minative or punctual ‘Aktionsart’. At least four verbs | of his group do not seem
to have such an ‘Aktionsart’, viz. νικῶ, ἑσσοῦμαι, ἥκω, οἴχομαι. These verbs are,
rather, stative, νικῶ having approximately the meaning ‘to be conqueror’, ἥκω ‘to
have come’, ἑσσοῦμαι ‘to be inferior to’, οἴχομαι ‘be gone’. In these cases it makes
little sense to say that the present stem, when used in temporal clauses, has the
combined value of anteriority + ‘Fortwirken’; such clauses express, rather, sim-
ple simultaneity.35 This also would seem to hold for some other verbs of H.’s
group, being durative rather than punctual or terminative. Thus, in the follow-
ing example προσίῃ may express simultaneity.

(53) ἐπεὰν δὲ σαπῇ [aor.] καὶ προσίῃ [praes.] ὁ τεταγμένος χρόνος, ἀπικνέεται …
(2.41.4)

When the carcase is decomposed, and the time appointed is at hand, a

boat comes …

There is no need to view this as a case of anteriority + ‘Fortwirken’, as H. does

(p. 65), who translates: ‘wenn … herangekommen (sc. und dann da) ist’; ‘wenn
… herankommt’, Engl. ‘when … comes near, approaches’ or ‘when … is at hand’,
as Godley has it, suit very well.
To summarize the main results of the discussion on chapter 6: for a num-
ber of verbs, having a punctual or terminative lexical semantic aspect, it often
is difficult, when they are used in subordinate temporal clauses, to view them
as expressing simple simultaneity. H.’s solution of this problem, viz. to assume
that the present stem expresses the ‘Fortwirken’ of an, otherwise finished,
action, does not seem to apply: the ‘Fortwirken’ is not so much connected with
the present stem but is, rather, an inherent lexical feature of these verbs. In my
opinion we have to reckon, in these cases, with freedom of choice between the
present and the aorist stem.36
In chapter 7 (Ingressiver Aorist und Vorzeitigkeit) H. discusses another case
where there is interference between the lexical semantic aspect of a verb and

35 H. himself makes a remark to this effect on p. 69 in connection with νικῶ.

36 For a possible ‘Fortwirken’-like effect of the present stem of γίνομαι see note 34.—In other
languages, verbs like hear, arrive, find out, discover, etc. would seem to behave in a similar
way with respect to the choice of verb-forms in temporal clauses. Cf. Du. Toen ik hoorde dat
de voorstelling afgelopen was, ben ik weggegaan; Engl. When I heard that the show was over,
I went away; Germ. Als ich erfuhr, dass die Vorstellung zu Ende war, bin ich weggegangen.
a review of: h. hettrich 33

the verb-forms, viz. the ingressive aorist. H.’s discussion starts from Ruipérez’
hypothesis (Ruipérez 1954) that durative verbs have aorists with two differ-
ent values, a ‘valor puntual’ and a ‘valor neutro’; the latter will not concern us
here. As to the ‘valor puntual’, this has two different realisations: with durative-
transformative verbs (‘transformative’ equals more or less ‘terminative’) as
‘aoristo finitivo’ (others use terms like ‘confective’, ‘complexive’), with durative-
non-transformative verbs as ‘aoristo initivo’ (= ingressive). In the first case, the
end of the action | or process is taken into account; examples are πίπτω ‘fall’, 246
ἄγω ‘lead, bring’, ποιέω ‘make’. In the second case, the aorist signals the starting-
point of the action or process; examples are ἄρχω ‘rule’, βασιλεύω ‘be king’, χαίρω
‘be pleased’.
In H.’s opinion, Ruipérez’ hypothesis as to the two realisations of the ‘valor
puntual’ is confirmed by the material from Herodotus, that is, a durative verb
either has an ingressive aorist or a confective (complexive) one, dependent
upon the ‘non-transformative’ or the ‘transformative’ character of the verb
(pp. 72–73). He admits only one exception to this rule: ‘non-transformative’
verbs may have a complexive aorist when the context provides a temporal mod-
ifier that indicates the extension of the action or process, as in:

(54) συνήνεικε αὐτὸν Δαρεῖον, βασιλεύσαντα [ptc. aor.] τὰ πάντα ἕξ τε καὶ τριή-
κοντα ἔτεα, ἀποθανεῖν (7.4)

Death came upon Darius, after a reign of thirty-six years in all

However, such aorists are not confined to these contexts; there are at least
three cases of complexive aorists of βασιλεύω, without there being any tem-
poral modifiers. Cf.:

(55) οἱ Πέρσαι ὕπαρχον ἐπιστᾶσι Λυκάρητον τὸν Μαιανδρίον τοῦ βασιλεύσαντος

[ptc. aor.] Σάμου ἀδελφεόν (5.27.1)

The Persians appointed as a governor Lycaretus, the brother of Maean-

drius who had been king of Samos

The other examples are: 1.7.3 and 2.147.2. In this connection another objec-
tion may be made against H.’s analysis of verbs like βασιλεύω. He states (p. 73
and p. 76) that the speaker/author with these verbs had the freedom to choose
either the (ingressive) aorist or the imperfect; in both cases the ‘B-Handlung’,
i.e. the action of the subordinate clause, would be parallel to that of the main
clause. Thus, by way of an illustration, both ἐπεὶ Δαρεῖος ἐβασίλευσε and ἐπεὶ
 34 chapter 1

Δαρεῖος ἐβασίλευε would convey that the action of the main clause occurred
during Darius’ reign. When viewed in isolation this may be right, but I do not
think for all that that ἐβασίλευσε and ἐβασίλευε are interchangeable: the ingres-
sive aorist only seems to occur in contexts where a statement precedes to the
247 effect that some other king has died. |
I conclude this review with some short remarks on the first section of chap-
ter 8 (Partizip und Nebensatz-Konstruktionen ausserhalb der A-B-Beziehung).37
In this section, H. discusses constructions involving participles and subordi-
nate clauses with ὅτε that are not temporal.38 An example of such a participle
is:

(56) ὑμεῖς δ᾽ ἄν μοι τάδε ποιέοντες [ptc. praes.] χαρίζοισθε (7.8δ1)

As for you, you will please me best by acting as follows

Here we are confronted with what is called by H. ‘Identitätsfall’, that is, the par-
ticiple does not refer to an independent action standing in a temporal relation-
ship to the main verb. Rather, the participle refers to an action that specifies the
way in which the main action is performed (H.’s formulation: ‘Die hinter bei-
den Verben stehenden Handlung ist also dieselbe’ is somewhat unfortunate).
The aorist participle also occurs:

(57) ἡ μὲν δὴ Πυθίη ὑπομνήσασα [ptc. aor.] ταῦτα ἔσχε βουλομένους τιμωρέειν
τοῖσι Ἕλλησι (7.171.2)

By reminding them of these things, the Pythia kept them from aiding the
Greeks as they would have done

37 The other—short—sections of this chapter deal with (1) τυγχάνω and other verbs getting a
participle as their complement; (2) so-called periphrastic constructions; however, H. does
not discuss ἔχω + aorist participle. For the whole subject cf. Dietrich (1973), Kahn (1973)
and Aerts (1965). The latter two authors are not in H.’s bibliography.
38 From H.’s discussion one would get the impression that all temporal clauses may be used
in this non-temporal—or at least not strictly temporal—way, but actually this only holds
for ὅτε; cf. Rijksbaron (1976: 135–136 and the footnotes there). To be sure, on p. 117, H. men-
tions two instances of ὡς-clauses, but this is not correct: the content of ἐργάσατο … ἀγαθά
is expressed by what follows, not by the ὡς-clauses. H. also treats 6.19.3 as a case of this
‘explicative’ ὅτε, as it may be called; on p. 134 I proposed some other interpretations, but
H. may be right. 2.121ε.4 should not be listed under ‘main verb in the present stem’ (H.,
p. 116), since the main verb is εἴη ἐργασμένος.
a review of: h. hettrich 35

H. rightly notes that these cases formally are indistinguishable from tem-
poral cases and that the context has to decide. Thus, H. remarks that in (57)
in principle two interpretations are possible, a temporal one (‘als … erinnert
hatte’) and an interpretation involving ‘Identität’ (‘indem … erinnerte’). The
correct interpretation, the second one, is only established by looking back to
7.169, where manner and effect of the Pythian answer are mentioned in detail
(H., p. 79). I add that general semantic considerations as to the structure of
sentence (57) may play a role, too: if we were not to take ὑπομνήσασα as an
instrumental modifier of ἔσχε,39 we would be left with the question how the
checking (σχεῖν) of the Cretans was brought about.
I notice one further point. Whereas in the case of the present participle one
still could speak of simultaneity (after all, ποιέοντες in (56) is simultaneous with
χαρίζοισθε) it is virtually impossible to say that the aorist participle is anterior to
the main action; this would mean that we are dealing with two strictly succes-
sive actions. For this reason several scholars have been tempted to say that the
aorist participle in principle may express simultaneity as well as anteriority (cf.
the references in Hettrich, | p. 79). H. rejects this; as he says on p. 79, the func- 248
tioning of present stem and aorist stem can, at least partially, be described in
terms of simultaneity and anteriority, respectively, and the usefulness of these
notions is not disproved because in other ‘Textkonstellationen’ present stem
and aorist stem are used according to other criteria. His own solution for exam-
ples like (56) and (57) is that the choice of the tense stem of the participle is
conditioned to a high degree by the tense of the main verb: if the main verb
is in the present stem, there is a preference for a present participle, if it is in
the aorist stem, there is a preference for an aorist participle. In support of his
analysis he adduces some statistical data, that suggest, indeed, that a ‘tense
attraction’, as it may be called, may be involved. For all that, I still think that
the aorist participle has its proper value: while not expressing anteriority, in
the strict, temporal, sense, it does refer to a completed action. This holds, for
instance, for ὑπομνήσασα in (57): the Pythia has finished her speech. Now the
important point is that the coming about of the action referred to by the main
verb logically depends upon the completion of the first action: it is only when

39 Such participles may, indeed, be called instrumental modifiers, cf. p. 26 and also Vester
(forthcoming) on some syntactic problems connected with Latin instrumental modifiers.
Of course, still other interpretations than those discussed in the text may be envisaged, e.g.
causal ones, (H. takes παραχρεώμενοι and αἰτέοντες in 7.223.4 as ‘Identitäts’–examples, cf.
his list on p. 110, but they are, rather, causal modifiers.) For the whole subject of the various
interpretations of the participle one may consult Oguse (1962); instrumental participles
are discussed on pp. 146–160 (this book is not in H.’s bibliography).
 36 chapter 1

the speech is finished that the σχεῖν is brought about. We might say, then, that
the action of the participle is logically prior to that of the main verb, not tempo-
rally. That we, in last analysis, interpret the participial clause as specifying the
manner in which the σχεῖν is brought about, and, thereby, as occurring simul-
taneously with the σχεῖν, is due to syntactic factors, the ‘reminding’ appearing
as a modifier of the ‘keeping from’.
To conclude, I give a very clear example of the dual possibilities of such a par-
ticiple, and of the importance of the context, that is, the nature of the actions
referred to by participle and main verb in interpreting them, taken from Homer:

(58) ὣς εἰπὼν ὤτρυνε μένος καὶ θυμὸν ἑκάστου (Hom. Il. 5.470 and passim)

So he spoke, and stirred the spirit and strength in each man

(59) ὣς εἰπὼν ἔνδυνε περὶ στήθεσσι χιτῶνα (Hom. Il. 10.131)

After this speech, he slipped the tunic upon his body

In both cases we are dealing with a participle that refers to a completed action:
the respective speeches are finished. In (59) this action is followed by an action
that is entirely independent of the former, which leads to our interpreting the
first action as anterior to the second: we are, thus, dealing with two successive
249 actions. In (58), however, the action referred | to by the main verb does not
exist independently of that of the participle: the coming about of the ‘stirring
up’ depends upon the completion of the speech, or, to put it otherwise, the
completion of the speech is a necessary condition for the coming about of the
‘stirring up’. It may be asked, of course, how we are to decide, in a given case,
which of these relationships applies. As far as I can see, no general semantic
rules can be given for this, in terms of, for example, the semantic features of the
verbs of participial clause and main clause; a rather practical rule-of-thumb is
that a non-instrumental interpretation of participles like that of (58) generally
leads to an unacceptable sentence.

3 Conclusion

Summarizing the above discussion, the main results appear to be as follows.

Hettrich’s assumption, following Ruijgh (1971), that the use of the present
stem and the aorist stem in temporal clauses and participles can be explained
in terms of simultaneity and anteriority, respectively, is highly convincing. The
a review of: h. hettrich 37

merit of H.’s book lies in the forceful way in which the relevance of tempo-
ral notions for the functioning of the above constructions is emphasized. H.’s
suggestion that these notions are also operative in other structures involving
present stem and aorist stem, is not implausible, but needs to be studied in
more detail. As to H.’s claim that he will make use extensively of the context,
this appears to be not wholly warranted; in my opinion, the role of the con-
text can be given a much more prominent place. H.’s treatment of a number
of examples that do not satisfy his general hypothesis, is of uneven quality. His
interpretations often show a lack of precision, and not seldom he is rather care-
less in choosing his material.
I finally note that H.’s book compares favourably with many linguistic stud-
ies, in that he presents full lists of examples, which enables the reader to imme-
diately check H.’s analyses.

Bibliography

Aerts, W., Periphrastica (Amsterdam 1965).

Blass, F., ‘Demosthenische Studien III (Aorist und Imperfekt)’, RhM 44 (1889) 406–430.
Comrie, B., Aspect (London 1976).
Curtius, G., Die Bildung der Tempora und Modi im Griechischen und Lateinischen (Berlin
1846).
Dietrich, W., ‘Der periphrastische Verbalaspekt im Griechischen und Lateinischen’,
Glotta 51 (1973) 188–228.
Humbert, J., Syntaxe grecque, 3e éd. revue et augmentée (Paris 1960).
Kahn, Ch.H., The Verb ‘Be’ in Ancient Greek (Dordrecht 1973).
Kühner, R., Stegmann, C., Ausführliche Grammatik der lateinischen Sprache. II: Sat-
zlehre. 2 vols. (Hannover 1912, repr.).
Maslov, Ju.D., ‘Вопросы глагольного вида в современом зарубешном языкозна-
нии/ Voprosy glagol’nogo vida v sovremennom zarubežnom jazykoznanii [Ques-
tions of verbal aspect in modern foreign linguistics]’, in Ju.D. Maslov (ed.), Вопросы
глагольного вида: сборник/ Voprosy glagol’nogo vida: sbornik (Moscow 1962) 7–
32.
Mazon, A., ‘La notion morphologique de l’aspect des verbes chez les grammariens
russes’, in Mélanges offerts à E. Picot (Paris 1913) I, 343–367.
Nickel, G., ‘Kontextuelle Beziehungen zwischen Sätzen im Englischen’, Praxis der neu-
sprachlichen Unterrichts 15 (1968) 15–25. Repr. in W. Dressier (ed.), Textlinguistik
(Darmstadt 1978) 147–166.
Oguse, A., Recherches sur le participe circonstantiel en grec ancien (Paris 1962).
Powell, J.E., A Lexicon to Herodotus (London 1938)
38 chapter 1

Rijksbaron, A., ‘Het praesens historicum in het Oudgrieks’, Handelingen van het 32e Ned-
erlandse Filologencongres (Amsterdam 1974a) 146–148.
Rijksbaron, A., Review of Sicking 1971, Lampas 7 (1974b) 207–215.
Rijksbaron, A., Temporal and Causal Conjunctions in Ancient Greek: With Special Refer-
ence to the use of ἐπεί and ὡς in Herodotus (Amsterdam 1976).
Ruipérez, M.S., Estructura del sistema de aspectos y tiempos del verbo griego antiguo
(Salamanca 1954)
Ruijgh, C.J., Autour de ‘τε épique’: Études sur la syntaxe grecque (Amsterdam 1971).
Schopf, A. (ed.), Der englische Aspekt (Darmstadt 1974).
Sicking, C.M.J., Hoofdstukken uit de Griekse syntaxis (Amsterdam 1971).
Strunk, K., ‘Historische und deskriptive Linguistik bei der Textinterpretation’, Glotta 49
(1971) 191–216.
Svensson, A., Zum Gebrauch der erzählenden Tempora im Griechischen (Lund 1930).
Szemerényi, O., ‘Unorthodox Views of Tense and Aspect’, Arch. Linguist. 17 (1965) 161–
171.
Vester, E. (forthcoming), ‘Review of: H. Happ, Grundfragen einer Dependenz-Grammatik
des Lateinischen (1976)’ [since published, Mnemosyne 34 (1981) 169–174].
 chapter 2

The Greek Perfect: Subject versus Object

Abstract

This paper argues, following McKay (1965, 1980), that the views of Wackernagel and
Chantraine on the development of the perfect in Ancient Greek are untenable; cp.
also Magnien (1929). The active perfect, that is, did not change from a subject-oriented
into an object-oriented tense. In other words, the active transitive perfect both in
Homer and in later Greek expresses the state of the Agent-subject rather than that
of the Patient-object. More specifically, the perfect stresses the lasting involvement
of the Agent in the state he has brought about; often his responsibility for this state
is emphasized. To express the state of the Patient (the ‘semantic object’) the passive
perfect was used. It is argued that failure to distinguish between semantic and syn-
tactic functions has led Wackernagel and Chantraine astray; also, that Dik’s Functional
Grammar (cp. Dik 1978) offers a suitable framework for a more adequate analysis of the
Greek perfect. The active perfect kept its subject-oriented value well into Hellenistic
times, as in shown by an analysis of the use of perfect and aorist in the Rosetta Stone.
It did not, therefore, become a kind of doublet of the aorist. Finally, a hypothesis is
put forward concerning the rapid increase in the use of the active transitive perfect
in the fifth century. This increase is argued to be connected with the growing aware-
ness of, and interest in, the responsibility of the individual, as opposed to the society
described by Homer, where there is much less room for notions like individual respon-
sibility.

1 Introduction

To situate the discussion of the Greek perfect in a wider framework I first of all
give a brief survey of the five tense stems (or aspect stems)1 which characterize
the verbal system of Ancient Greek:

1 The nature of the relationship between ‘tense’ and ‘aspect’ in Ancient Greek is not without
problems but these will be ignored here. Generally speaking, I share the view of those schol-
ars (e.g. Ruijgh 1970) who consider the importance of the category ‘aspect’ for the functioning
of the verb in Ancient Greek overrated.

© koninklijke brill nv, leiden, 2019 | doi:10.1163/9789004386129_004

 40 chapter 2

– the present stem (e.g. βουλεύε-σθαι)

– the aorist stem (e.g. βουλεύσα-σθαι)
– the perfect stem (e.g. βεβουλεῦ-σθαι)
– the future stem (e.g. βουλεύσε-σθαι)
– the future perfect stem (e.g. βεβουλεύσε-σθαι)2

The semantic values of these stems can be defined as follows:

– the present stem signifies that the action3 expressed by the verb stem con-
tinues through a moment of time which is given in the context or situation,
and is, therefore, not-completed (imperfective value);
– the aorist stem signifies that at a moment of time given in the context or
situation the action expressed by the verb stem is completed (confective
value);
– the perfect stem signifies that at a moment of time given in the context or
404 situation | a state exists which is the result of the completion of the action
expressed by the verb stem (stative-confective value);
– the future stem signifies that the action expressed by the verb stem is located
after a moment of time given in the context or situation; the notion ‘com-
pletion’ is not relevant for the future stem;
– the future perfect stem signifies that after a moment of time given in the
context or situation a state will exist which is the result of the completion of
the action expressed by the verb stem.
This survey shows that the notions ‘(not-)completed’ and ‘stat(iv)e’ are of prime
relevance for the Greek verbal system. The following table illustrates the vari-
ous values on the basis of the active indicative of παιδεύω:

2 Middle rather than active forms are given, since these show the various morphological fea-
tures of the stems more clearly than their active counterparts.
3 ‘Action’ is used here in a general sense, to comprise processes and states as well. The term
‘state of affairs’ is also often used for this general sense (Dik 1978: 25).
the greek perfect: subject versus object 41

action is

completed
not-completed
situated in + resulting state – resulting state

present πεπαίδευκα παιδεύω4

past ἐπεπαιδεύκη ἐπαίδευσα ἐπαίδευον

future πεπαιδευκὼς ἔσομαι παιδεύσω

Starting with Georg Curtius’ influential and controversial book Die Bildung der
Tempora und Modi im Griechischen und Lateinischen (1846) the focus for Greek
has been on the present and aorist stems. Actually, these stems are still the
object of debate, and every now and then new proposals are put forward regard-
ing the exact semantic values of these two stems (although the importance of
the notion ‘completion’ would seem to be widely acknowledged).
For the perfect stem the situation is quite different; its position in the Greek
verbal system would seem to be more or less established. To a certain extent this
is due to the low frequency of the forms of the perfect stem, when compared
with those of the present and aorist stems. A far more important factor, how-
ever, is the influence of two authorative monographs on the perfect, by Jakob
Wackernagel (1904) and Pierre Chantraine (1927), respectively. Their analysis
has been almost universally accepted.5 There have, to be sure, been dissenting
voices, like those of Hartmann (1930), Magnien (1929) and especially McKay
(1965, 1980), but their critical observations have had no or little impact. In what
follows I will argue, on many points inspired by the papers of McKay, that the
calm surrounding the perfect is by no means justified. I will concentrate | on 405
the primary perfect indicative (the ‘perfect’, for short) of so-called ‘terminative’

4 The present indicative is, of course, also used generically, i.e in timeless statements and in the
descriptions of habits and properties, e.g. ‘The sun rises in the east’ or ‘The Dutch eat a lot of
potatoes’.
5 See also Wackernagel (1920: 1, 166 ff.) For a recent instance of approval see Armstrong (1981:
3): ‘The peculiar history of the Greek perfect … is brilliantly described in J. Wackernagel (1920:
1, 166 ff.) and in more detail in his earlier work (1904)’.
42 chapter 2

verbs.6 First, I present a survey of the views of Wackernagel and Chantraine.

This will be followed by a discussion of some instances from Homer and clas-
sical Greek, and the way in which these perfects are analysed by Wackernagel
and especially Chantraine. Next the relationship between perfect and aorist is
discussed; a prominent role in this discussion is given to the use of the perfect
in the Rosetta Stone. In the conclusion the main results are summarized, and
the perfect in Homer returns one more time.

2 The Views of Wackernagel and Chantraine

On p. [403] above the value of the perfect stem was described as: signifies that
at a moment of time given in the context or situation a state exists which is the
result of the completion of the action expressed by the verb stem. The indica-
tive locates this state at the moment of speaking, e.g. in πεπαίδευκα τοὺς παῖδας
‘I have educated the children’. Now the question is: whose state is involved
here? According to Wackernagel and Chantraine it is the state of the object,
e.g. τοὺς παῖδας in the example sentence given above. To bring this out more
clearly this sentence might be paraphrased as follows: ‘I have educated the
children ⟨so they are now in a state of being (well-)educated⟩’. Wackernagel
and Chantraine further believe that this transitive use of the active perfect is
found mainly in classical Greek; in earlier Greek, especially Homer, this use
would be exceptional. In Homer, the perfect expresses rather the state of the
subject. To support these views they point to the large number of intransi-
tive perfects in Homer, as opposed to the very frequent perfect of transitive
verbs in later Greek, especially those formed with the suffix -κ-, which is rare
in Homer. They further claim that from the fourth century BC onward, and
possibly already in the fifth century, the value ‘completion’ ousted the value
‘state’. As a result the perfect (i.e. the perfect indicative) gradually became a
more or less expressive doublet of the aorist (indicative). This process is sum-
marized by Schwyzer-Debrunner (2, 263) as follows: ‘… für die spätere Entwick-
lung wichtiger (viz. than the ‘intensive’ perfect, cp. note 6) war das zunächst …
intransitive Perfekt des erreichten Zustandes (–). Zuerst wirkt dieser Vorgang
(viz. that expressed by the perfect) nur am Subjekt nach (–). Ferner kann das
Perfekt des erreichten Zustandes zu … transitiven Verben gebildet werden. Aus

6 I.e. verbs expressing actions with an inherent endpoint, like ‘educate’, ‘conquer’, ‘build’ etc.
Not discussed, therefore, are so-called ‘intensive’ perfects, i.e. perfects of stative verbs, e.g.
γέγηθα ‘be full of joy’ alongside γηθῶ ‘rejoice’ (cp. S-D 2, 263), nor perfect-like periphrastic
constructions, e.g. those formed by means of ἔχω + aorist participle.
the greek perfect: subject versus object 43

diesem transitivem Gebrauch ging, sicher seit dem 5. Jahrhundert, das … Resul-
tativperfekt hervor, das nicht mehr ein Zustand des Subjektes, sondern die in
der Gegenwart fortdauernde Nachwirkung einer abgeschlossener Handlung in
oder am Objekt | bezeichnet (–). Indem beim Resultativperfekt … der Begriff 406
der Abgeschlossenheit zur Hauptsache wird, entsteht das erzählende Vergan-
genheitsperfekt (–).
What is not immediately clear from this quotation (for it is nowhere men-
tioned explicitly) is that this argument concerns only the perfect active. Below
it will become clear that the analysis of Wackernagel and Chantraine is flawed
by its ignoring the passive perfect, and as a corollary, the differences between
syntactic and semantic functions. To show this I will first discuss some Homeric
examples and Chantraine’s analysis of these.7

3 The Perfect in Homer

The majority of Homeric perfects are either active perfects of intransitive

verbs8 or passive perfects, that is, morphologically passive perfects with pas-
sive meaning. Examples are (1)–(3) and (4)–(6), respectively:

(1) ἀνδρὶ δὲ κεκμηῶτι μένος μέγα οἶνος ἀέξει,

ὡς τύνη κέκμηκας ἀμύνων τοῖσι ἔτῃσι (Il. 6.261–262)

In a tired man, wine will bring back his strength to its bigness, in a man
tired as you are tired, defending your neighbours. (Transl. here and below
Lattimore)

(2) Ἥρη, τίπτε βέβηκας; (Il. 15.90)

Hera, why have you come?

(3) καὶ δὴ δοῦρα σέσηπε νεῶν … (Il. 2.135)

and the timbers of our ships have rotted away

7 I will focus on Chantraine as his analyses are more explicit and detailed than those of Wack-
ernagel.
8 Middle verbs often have a morphologically active perfect, e.g. γίγνομαι : γέγονα, σήπομαι :
σέσηπα, etc.
 44 chapter 2

(4) ἅρμα δέ οἱ χρυσῷ τε καὶ ἀργύρῳ εὖ ἤσκηται (Il. 10.438)

his chariot is fairly ornate with gold and with silver

(5) τίπτε τὰρ ὧδ᾽ Ἀχιλεὺς ὀλοφύρεται υἷας Ἀχαιῶν,

ὅσσοι δὴ βέλεσιν βεβλήαται; (Il. 11.657)

Now why is Achilleus being so sorry for the sons of the Achaians, who
have been wounded with spears thrown?

(6) … θηοῖό κεν αὐτὸς ἐπελθών,

οἷον ἐερσήεις κεῖται περὶ δ᾽ αἷμα νένιπται (Il. 24.419)

… you yourself can see when you go there how fresh with dew he lies, and
the blood is all washed from him

In all cases the perfect forms express a state of the subject: the addressee is
tired (1), or present (2), the timbers are rotten (3), the chariot is well equipped
(4), the Greeks are covered by wounds (5), and the blood is washed from Hector
(6). However, the subject does not in all cases have the same semantic function,
something ignored by Chantraine: in (1) and (2) it has the semantic function
Agent,9 in (3) that of Internally Affected,10 but in the passive instances (4)–
(6) the subject is rather the ‘semantic object’, since it refers to the entity that
has undergone the action. In more technical terms: the syntactic subject has
here the semantic function Patiens, and the perfect expresses the state of this
Patiens. By attaching so much weight to the syntactic function subject Wacker-
nagel, and especially Chantraine, pass by an essential semantic difference, with
407 far-reaching | consequences.11
Active perfects of transitive verbs are much rarer.12 On the basis of the cases
discussed above one may expect that in these cases, too, it is the state of the
subject that is expressed, and this is indeed what we find, e.g. in the following
examples:

9 Or rather, in more precise terms: the Agent is involved in the action of which the state
expressed by the perfect is the result.
10 This corresponds to Dik’s ‘Processed’ (Dik 1978: 37 ff.).
11 To be sure, Wackernagel acknowledges the function of the passive perfect (1904: 7), but he
does not pursue this point any further.
12 I confine myself here to discussing active perfects of active present indicatives. However,
my analysis also holds good for active, i.e. semantically non-passive, perfects of middle
verbs, like ἔκτημαι : κτῶμαι.
the greek perfect: subject versus object 45

(7) ναὶ μὰ τόδε σκῆπτρον, τὸ μὲν οὔποτε φύλλα καὶ ὄζους

φύσει, ἐπεὶ δὴ πρῶτα τομὴν ἐν ὄρεσσι λέλοιπεν. (Il. 1.234–235)

in the name of this sceptre, which never again will bear leaf nor branch,
now that it has left behind the cut stump in the mountains

(8) ὢ πόποι, ἦ δὴ μύρι᾽ Ὀδυσσεὺς ἐσθλὰ ἔοργε (–)

νῦν δὲ τόδε μέγ᾽ ἄριστον ἐν Ἀργείοισιν ἔρεξεν (Il. 2.272, 274)

Come now: Odysseus has done excellent things by thousands, (–) but now
this is far the best thing he ever has accomplished

The branch upon which Achilles swears his oath (ex. (7)) is in a state of being
separated for good from its tree. In (8), the anonymous speaker expresses
judgement on Odysseus’ record of service: the perfect expresses the idea that
Odysseus finds himself in a state of having carried out many feats; the per-
fect could be paraphrased by: Odysseus has many brilliant deeds to his name.
When individual deeds are concerned the aorist is used, see τόδ(ε) ἔρεξεν in (8).
Now interestingly, the above line of reasoning is also followed by Chantraine,
but thereby he gets into trouble. For what we have here are active perfects of
transitive verbs, so common in later Greek, and we might therefore expect—
in Chantraine’s view, that is!—that we are dealing here with the state of the
object. And yet this is, according to Chantraine not the case, although he does
not explain why it isn’t. On (7) he simply remarks (1927: 11): ‘Τομὴν détermine
λέλοιπεν, pourtant il ne s’agit pas d’un résultat’ (i.e., a result that relates to
τομή).13 In his view the first clear example of the latter use is to be found in
Pindar.

13 According to Chantraine there is one perfect that to some extent anticipates the ‘object-
oriented’ use of the perfect, viz. τοῖον ἄχος βεβίηκεν Ἀχαίους ‘such grief has fallen upon the
Achaians’ (e.g. at Il. 10.145). Why this value is present here but not in (7) and (8) he does
not explain. Be that as it may, this example can very well be taken as a ‘subject-oriented’
perfect: ‘grief weighs heavy on’. This interpretation is already found in the commentary of
Ameis-Hentze-Cauer ad locum: ‘hat überwaltigt, d.i. liegt schwer auf.’
 46 chapter 2

4 The Perfect after Homer, and the Rise of the -κ- perfect

The example referred to by Chantraine is:

(9) … ἴστε μάν

Αἴαντος ἀλκάν, (–)
ἀλλ᾽ Ὁμηρός τοι τετίμακεν δι᾽ ἀνθρώπων, … (Pi. I. 4.35–37)

Surely you know of Ajax’s valor, (–) But Homer, to be sure, has made him
honored … (Transl. here and below Race)

He observes (1927: 22): ‘Le parfait exprime bien encore un état; mais ce n’est
408 plus l’état du sujet, c’est celui de l’objet.’ In other words, the perfect | here
indicates that Ajax’ valour is forever honoured. However, as McKay (1965: 10)
has convincingly argued, this is only half of the truth. Naturally, it is also the
permanent fame of Ajax which is involved here, but if Pindar had wanted to
emphasize this point he could have done this more effectively by using a pas-
sive perfect, i.e. by making Αἴαντος ἀλκά the subject, with the semantic function
Patiens. The use of the active perfect, with its explicit Agent-subject, Homer,
indicates that the state in question relates primarily to Homer. And the ensu-
ing lines show this to be in fact the case, for these lines emphasize the crucial
role played by Homer in the transmission of the glorious past:

(10) … δι᾽ ἀνθρώπων, ὃς αὐτοῦ

θεσπεσίων ἐπέων λοιποῖς ἀθύρειν (Pi. I. 4.37–39)

… among mankind, who set straight his entire achievement and declared
it with his staff of divine verses for future men to enjoy.

In the course of the 5th century the use of the active perfect of transitive verbs
quickly increases, especially that with the suffix -κ-. Yet, in spite of the fact that
these perfects are construed with an object, it is not the state of the object
that is primarily involved here. Just as in the examples from Homer and Pin-
dar discussed above the perfect has a ‘subject-oriented’ value. This also applies
to the following example from Herodotus, which, like the example from Pindar,
clearly illustrates the importance of taking account of contextual information
for the interpretation of the perfect.

(11) Ἀχελῴου, ὃς … τῶν Ἐχινάδων νήσων τὰς ἡμισέας ἤδη ἤπειρον πεποίηκε (Hdt.
2.10.3)
the greek perfect: subject versus object 47

… Achelous, which … has already made half of the Echinades islands to

be mainland. (Transl. here and below Godley)

At first sight this may seem the example par excellence of the ‘object-oriented’
use of the active perfect: the river Achelous has caused half of the Echinades
islands to be part of the mainland, and so they are now in a permanent state
of ‘being mainland’. Unsurprisingly, this is also the view of Chantraine.14 How-
ever, although the state of the islands is of course expressed here, this is not
the real issue. What this sentence emphasizes is the state of the Achelous,
i.e. its importance as a river, as can be inferred from the preceding lines, not
taken into account by Chantraine. In these lines the Achelous is described
as one of the great rivers of Greece, rivers that have great deeds to their
name:

(12) εἰσὶ δὲ καὶ ἄλλοι ποταμοί, οὐ κατὰ τὸν Νεῖλον ἐόντες μεγάθεα, οἵτινες ἔργα | 409
ἀποδεξάμενοι μεγάλα εἰσί· τῶν ἐγὼ φράσαι ἔχω οὐνόματα καὶ ἄλλων καὶ οὐκ
ἥκιστα Ἀχελῴου, ὃς …

There are also other rivers, not so great as the Nile, that have wrought
great effects; I could declare their names, but chief among them is Ache-
lous, which …

Here too, then, we may observe that the perfect expresses first and foremost the
state of the subject (the Agent). And here, too, to express the state of the object,
i.e. the constituent with the semantic function Patiens, the passive perfect was
available.
The ‘subject-oriented’ value of the perfect active could be demonstrated
from numerous other examples. I confine myself here to discussing a few illus-
trative cases.

(13) γέγραφε δὲ καὶ ταῦτα ὁ αὐτὸς Θουκυδίδης Ἀθηναῖος (Th. 5.26.1)

These events too the same Thucydides from Athens has put on record

In cases like (13) the perfect has the specific nuance that it emphasizes the
responsibility of the Agent for his deeds, in this case Thucydides’ writing his
account of the Peloponnesian war. Note the elaborate way in which the iden-

14 He is emphatically followed by Keil (1963: 32).

 48 chapter 2

tity of the author is highlighted. The perfect passive, on the other hand, is used
to express the state of the Patient-subject, as in

(14) ταῦτα δὲ τὰ δέκα ἔτη ὁ πρῶτος πόλεμος ξυνεχῶς γενόμενος γέγραπται (Th.
5.24.2)

The first war, which had gone on without interruption during these ten
years, has now been put on record.

Here, the emphasis is on the completion on the first of part of the work; the
author is of secondary importance: he is not even mentioned.

(15) τίς οὖν … ἐπιτηδειότερος Νικομάχου δοῦναι δίκην; τίς ἐλάττω τὴν πόλιν ἀγαθὰ
πεποίηκεν ἢ πλείω ἠδίκηκεν; ὂς … ἡμάρτηκεν … (Lys. 30.24)

And from whom … could (requital) be more suitably exacted than from
Nicomachus? Who has rendered less service or done more wrong to the
city? He has offended …

The perfects active emphasize the responsibility of Nicomachus for his

(mis)deeds. This nuance is a frequent phenomenon in Lysias and the other ora-
tors, not surprisingly perhaps; after all, they pre-eminently deal with matters of
guilt and responsibility. I note in passing that the increasing use of the transi-
tive perfect active may be connected with the rise of certain genres and with
social factors. I will return to this point in the conclusion.

5 Perfect and Aorist

In this section another element of Chantraine’s view will be scrutinized,

namely the supposed rapprochement between perfect and aorist (indicative)
410 (cp. above p. [405]). | According to Chantraine the perfect gradually lost its sta-
tive value, the value ‘completion’ becoming more and more important. Thus
the perfect hardly differed any longer from the aorist semantically, and so it
developed into a narrative tense. The first signs of this development are already
visible, according to Chantraine, in the fifth century, e.g. in the following pas-
sages from Sophocles:

(16) ἀπόλωλα τλήμων, προδέδομαι. τί μ᾽, ὦ ξένε,

δέδρακας; ἀπόδος ὡς τάχος τὰ τόξα μοι. (S. Ph. 923–924)
the greek perfect: subject versus object 49

Then I am lost, I am betrayed. Stranger, what have you done to me? Give
back my bow. (Transl. Greene, combined with Torrance)

(17) ὦ λιμένες, ὦ προβλῆτες, ὦ ξυνουσίαι

θηρῶν ὀρείων, ὦ καταρρῶγες πέτραι,
(–)
ἀνακλαίομαι …,
οἷ᾽ ἔργ᾽ ὁ παῖς μ’ ἔδρασεν οὑξ Ἀχιλλέως (S. Ph. 936–937; 939–940)

Harbors and promontories—fellow-creatures who roam the Mountain-

side, steep-rising cliffs (to you) I lament … the wrongs Achilles’ son has
done to me (Transl. Torrance)

Chantraine comments (1927: 184): ‘Il arrive souvent que le parfait soit employé
dans un passage où notre analyse grammaticale a peine à le justifier. Les textes
permettent de voir combien les deux temps sont parfois voisins.’ At first sight
it looks as if δέδρακας and ἔδρασεν are indeed ‘voisins’, the more so because
the two forms refer to the same action, viz. Neoptolemus’ snatching away
Philoctetes’ bow. On closer inspection, however, it appears that the contexts
of these passages are widely divergent, and this fact explains the use of the
perfect δέδρακας in (16), which has its normal subject-oriented value. In (16)
Philoctetes, who is utterly desperate and considers himself lost (cp. ἀπόλωλα),
makes an appeal to Neoptolemus to give him back his bow. By using δέδρακας he
stresses Neoptolemus’ responsibility for, and involvement in, the crime he has
committed, which in turn enhances the force of the appeal. In (17), however,
Philoctetes, who is at this point alone, bursts out into a soliloquy to the nature
that surrounds him. Now he refers to Neoptolemus’ behaviour by means of the
aorist, expressing that in the situation at hand it is this behaviour itself that is
of prime importance. Next, in lines 941–948 Neoptolemus’ misdeeds are men-
tioned one by one.15 In the situation at hand it would have been rather pointless
to stress Neoptolemus’ responsibility, as he is not present. The other examples
adduced by Chantraine can be explained in a similar way.16

15 The aorist is used, then, with so-called ‘countable’ actions; cp. Armstrong (1981).
16 E.g. S. OC 825 and 854; Ant. 442 and 443. See Chantraine (1927: 164ff.). [Addendum 2018.
The lines from Antigone are particularly revealing. To Creon’s question φής, ἢ καταρνῇ μὴ
δεδρακέναι τάδε; (‘Do you admit or deny that you are responsible for these deeds?’, tr. AR)
Antigone responds: καὶ φημὶ δρᾶσαι κοὐκ ἀπαρνοῦμαι τὸ μή (‘I admit that I have done them’).
By using the perfect infinitive Creon indcates that he wants to hear from Antigone: δέδρακα
‘I am responsible for them.’ Antigone, however, by using the aorist infinitive, representing
 50 chapter 2

In my view the above examples by no means suggest that the perfect must
be taken as a kind of ‘pseudo-aorist’. I should add that even if the perfect had
become aorist-like this does not mean that it functioned as a narrative tense.
For in instances like (16) and (17)—and this applies to virtually all other perfects
that are considered aoristic by Chantraine—the aorists that are presented to
show the similarity between perfect and aorist are not narrative aorists, since
they do not occur in a narrative text, but rather so-called constative aorists,
411 which typically, as here, are found in direct speech.17 |
That the perfect active of transitive verbs, rather than becoming a kind of
doublet of the aorist indicative, kept its ‘subject-oriented’ value well into Hel-
lenistic times can be illustrated from its use in the famous Rosetta Stone. This
stone contains a decree, drawn up in 196BC by Egyptian priests, to celebrate the
first commemoration of the coronation of Ptolemaeus V Epiphanes. The decree
consists of an introduction specifying place and time (lines 1–6), a preamble,
that opens with the usual ἐπειδή ‘considering that’, ‘since’ (lines 6–36), and the
decree proper: ἔδοξεν τοῖς ἱερεῦσι ‘the priests have decided’ (lines 36–54). In
what follows I discuss the use of perfect and aorist in the preamble (see the
accompanying text with translation at the end of this chapter).18
The preamble opens with a series of perfects active, predominantly of tran-
sitive verbs, which all depend on ἐπειδή: εὐεργέτηκεν, ἀνατέθεικεν, ὑπομεμέ-
νηκεν, intransitive πεφιλανθρώπηκε, one puzzling aorist ἀφῆκεν,19 κεκούφικεν:

constative ἔδρασα, merely admits that she has done them. See also Rijksbaron 2006: 36,
note 2].
17 The constative aorist can be modified by the temporal adverb νῦν, just like the present
indicative and the perfect indicative, but unlike the narrative aorist, which occurs in alter-
nation with the imperfect. For the constative aorist see Rijksbaron (1984/2006: 28). The
real competitor of the narrative aorist is rather the pluperfect, but this matter is not fur-
ther discussed here.—To be complete I may add that the difference between perfect and
aorist mentioned here naturally is also found in the passive. Contrast Φείδων αἱρεθεὶς ὑμᾶς
διαλλάξαι (‘Pheidon, after being chosen to reconcile you …’, Lys. 12.58) with στρατηγὸς ὑφ᾽
ὑμῶν ᾑρημένος (‘having been elected general by you’, Lys. 18.4): the aorist participle refers
to one single assignment, the perfect participle to a permanent function.
18 The text is taken from the PHI Greek Inscriptions database (https://epigraphy.packhum
.org/text/219002); the translation is that of Wallis Budge (1904: 104ff.). The perfects are
marked by a straight, the aorists by a dotted line. Other Hellenistic texts are extensively
discussed in McKay 1980. The Rosetta Stone decree is written in two languages, Egyptian
and Greek, and three scripts: hieroglyphic, demotic and Greek. I discuss only the Greek
text.
19 Perhaps a hearing/writing error for ἀφεῖκεν? In general, however, the writing is very care-
ful. Or perhaps the combination with εἰς τέλος ‘completely’ (rather than Wallis Budge’s
‘finally’, cp. LSJ s.v. II 2.b) has lent the aorist a perfect-like meaning? [Addendum 2018.
The Η of ΑΦΗΚΕΝ, which is found in all modern texts of the Rosetta Stone, may not be
the greek perfect: subject versus object 51

‘Considering that king Ptolemy has given many benefactions …, has dedicated
…, has incurred great expenses …, has shown magnanimity …, some (taxes) he
has finally suppressed …, others he has diminished.’ These perfects are followed
by a long series of aorists, which refer to individual deeds of Ptolemy, illustrat-
ing and justifying the general qualifications expressed by the perfects. In other
words, his benefactions etc. consist of the various actions that start with ἀφῆκεν
in 13 and end with προσδιωρθώσατο in 35. This is series is once interrupted by
a perfect, viz. διατετήρηκεν in 33, probably with good reason, for we are deal-
ing here with priviliges of the priests that were maintained by Ptolemy and so
with what must have counted for them as a major service.20 Finally, there is just
before the decision one last perfect, δεδώκασιν in 35, with the gods as subject:
they are the ones responsible for Ptolemy’s health, victory, power and all other
good things.
All in all the employment of the perfect here is not different from that dis-
cussed above (p. [410]). The transitive perfect active expresses a state that con-
cerns certainly also the Patient-object but is primarily relevant for Ptolemy as
Agent-subject: it is his responsibility for the actions concerned, and thus his
merits, that for the priests is the main point.21 The aorists, just as in ex. (17),
simply express the individual deeds performed by him.
Now remarkably Mayser, in his monumental grammar of the Greek papyri,
writes (1926: II, 1, 211): ‘Noch bunter (viz. than in some other honorary decrees)
ist der Wechsel (of perfect and aorist) … im Rosettastein.’ In the light of what I
argued above this view seems untenable. On the contrary, I believe the Rosetta
Stone shows a keen, one might almost say ‘classical’, awareness of the respec-
tive values of perfect and aorist. Observe, finally, that here, too, the perfect is
not used as a narrative tense. | 412

there. This is suggested by the text in the edition of R. Parkinson (1999), in which the stone
has been cleansed of the white chalk with which the incised signs had been filled ‘early in
its museum history, a procedure that was designed to make the text more legible’ Parkin-
son (1999: 23; the cleansed text is on p. 18)). To all appearances, the letters after ΑΦ are ΚΚ,
which must be an error for either ΕΙΚ or ΗΚ. Since this form is preceded by ΑΝΑΤΕΘΕΙ-
ΚΕΝ, ΥΠΟΜΕΜΕΝΗΚΕΝ, ΠΕΦΙΛΑΝΘΡΩΠΗΚΕ and followed by ΚΕΚΟΥΦΙΚΕΝ I propose
to read ΑΦΕΙΚΕΝ. There are a few examples of similar errors, e.g. ΠΤΟΜΑΙΟΥ in line 38,
ΧΡΥΣΙΟ 33, ΙΡΩΝ 35, ΧΟΝΟΥ 23. At the end of line 13 we find an indubitable instance of
ΑΦΗΚΕΝ, which is the first of a series of aorists, that ends with ΠΡΟΣΔΙΩΡΘΩΣΑΤΟ in
line 34; in line 28 a second example of ΑΦΗΚΕΝ occurs.]
20 There is also one imperfect, ἀνανεοῦτο in 35 (Wallis Budge 1907 wrongly has ἀνενεοῦτο, the
correct reading in Wallis Budge 1929; curiously, in this book there is no reference to the
earlier edition), which may have iterative value.
21 By using the perf. δεδώκασιν for the compensation provided by the gods the priests also
stress the responsibility and merits of the gods.
52 chapter 2

6 Conclusion

I hope to have shown that the views of Wackernagel and Chantraine concern-
ing the perfect transitive active cannot stand. The state expressed by this per-
fect primarily is the state of the Agent-subject, rather than that of the Patient-
object. The latter was expressed by the perfect passive.22 The aorist merely
expresses that the action in question has been completed. It has also been
shown that the subject-oriented value of the perfect was still fully operative in
the Hellenistic Greek of the Rosetta Stone. The perfect did not develop, then,
into some kind of double of the aorist indicative. A second conclusion that can
be drawn is of a methodological nature: it is not well possible to analyse the
use of the perfect without taking into account differences between the perfect
active and the perfect passive, and between syntactic functions like subject and
object, and semantic functions like Agent and Patient. Their ignoring these dif-
ferences to all appearances has put Wackernagel and Chantraine on the wrong
track.23 To put it differently, and in more positive terms: the perfect can only be

22 Cp. McKay’s remark: ‘Of course when the sort of emphasis postulated for the “resultative”
(i.e. object-oriented—AR) perfect is needed it can be expressed by the passive’, and Com-
rie (1976: 86): ‘The perfect passive is precisely that form which predicates a change of state
to the (semantic) object of an action.’ Cp. also next note.
23 Occasionally Chantraine seems to doubt the correctness of his own analysis, in the light
of his subject-oriented translations of βέβληκα ‘je suis dans l’état de celui qui a frappé
(quelqu’ un)’ (1927: 121) and φὴς δεδρακέναι τάδε; ‘avoues tu être coupable?’ (S. Ant. 442)
(1927: 167). Cp. also Magnien’s question (1929: 461): ‘Est-il bien sûr qu’en attique λέλυκα
τὰς σπονδάς signifies “j’ ai rompu le traité et il reste rompu” et non “j’ai rompu le traité et
je reste dans l’ état qui en est la conséquence”?’
Incidentally, from a cognitive point of view the reasoning of Wackernagel and Chan-
traine is quite understandable. For both in the case of the passive and that of the active
perfect the result of the action, i.e. its effect, will have primarily been observable on the
Patient, as with John has hit Bill. I may refer in this connection to the interesting obser-
vations of Comrie (1981: 70): ‘With some sentences with perfect meaning either A(gent)-
orientation or P(atient)-orientation is equally likely; an example is John has hit Bill, where
the point at issue may be either Bill’s state after having been hit, or John’s state (e.g. the
fact that he has broken his thumb) after hitting Bill. With most two-place (i.e. roughly
transitive—AR) predicates, however, in particular those that, lexically, describe a change
in state (e.g. ‘kill’, ‘melt’ (transitive), ‘move’ (transitive)) the state is primarily attributed to
P(atient).’ A few lines further he adds: ‘it is quite likely that there would be no ascertain-
able change’ in the Agent. Therefore ‘the resultant change of state is attributed primarily
to P., rather than A., in the perfect. Given this, one might expect the perfect to favor P.-
orientation, that is, that the perfect would be more likely than other aspects to correlate
quite highly with passive (–).’ All this sounds plausible enough; unfortunately, however,
Comrie gives no justification for the existence of the perfect active, neither here nor else-
where in his paper.
the greek perfect: subject versus object 53

analysed in a satisfactory way within the framework of a linguistic theory that

explicitly and unambiguously distinguishes between syntactic and semantic
functions. One of the attractive features of Simon Dik’s Functional Grammar is
that such distinctions belong to the core ‘apparatus’ of this theory.24
One intriguing question still remains. Why is it that the perfect active of tran-
sitive verbs is so rare in Homer, whereas intransitive and passive perfects are
quite common? As an explanation I venture the following, somewhat specula-
tive, hypothesis, which has a social, perhaps even, ethical foundation.25
In Homer, the intransitive perfect active expressing the state of the sub-
ject is a common phenomenon. Now it is of course an inherent property of
intransitive verbs that the action involves no other entitities than the Agent
(I restrict myself here to agentive verbs). This means in turn that the result of
this action concerns only the Agent; in other words, the Agent is only respon-
sible for actions that affect his own person. On the other hand, in those cases
in Homer where other entities than the Agent are involved in the action, more
specifically with transitive verbs, the state of the Patient is expressed by the
perfect passive. There is thus a certain asymmetry in the perfect system in
Homer: Agent-subject oriented perfects of intransitive verbs alternate with
Patient-subject-oriented perfects passive of transitive verbs. To be sure, the
Agent-subject-oriented perfect active is also used, as in examples (7) and (8),
but on a very limited scale. Now the latter perfect is very common in classical
Greek, starting with Sophocles. So there must have been a factor (Wackernagel’s
‘something’, see n. 25) which stimulated | the use of this perfect. In my view, 413
this factor was the growing awareness of, and interest in, the role of the indi-
vidual, especially his responsibilities vis-à-vis other individuals. In the society
described by Homer there is but little room for individual responsibility. There,
human actions are basically determined by the gods, and on the human side the
most important social framework is that of the collective.26 It is only in the 5th

24 Cp. Dik (1978: ch. 5).

25 Wackernagel confines himself (1904: 3) to noticing that there must have been ‘something’
(‘etwas’), either in the form or in the meaning of e.g. δίδωμι, which in the time of Homer
blocked the formation of δέδωκα. This is not much of an explanation, of course. This also
applies to his remark (p. 15): ‘Es ist natürlich (emphasis added), dass sich Wunsch und
Bedürfnis nach entsprechendem (viz. with the perfect passive—AR) aktivem Ausdruck
regten.’
26 In Homer, we are dealing with ‘une croyance en une reponsabilité collective. Les dieux
comme les hommes font en effet souvent retomber le châtiment d’une faute sur la famille
ou la cité toute entière du coupable. Une telle conception reflète l’existence d’un groupe
social très fermé, où l’ homme n’est considéré que comme membre d’une communauté
déterminée et n’ existe pas en tant qu’ individu autonome’ (Saïd 1978: 150).
54 chapter 2

century that the importance of the collective begins to diminish. The begin-
nings of this development can be found in Aeschylus,27 but it becomes fully
visible in Sophocles, who is the first author to use the Agent-subject-oriented
perfect active of transitive verbs with great effect. It need cause no wonder that
Homer uses this perfect on such a small scale: with him an ethical-social notion
like ‘individual responsibility’ is of minor importance. Nor is it surprising that
such perfects are so frequent in the genre that is pre-eminently concerned with
individual responsibility, viz. that of judicial oratory. In my view, then, there is
direct link between the rise of the Agent-oriented perfect and the rise of the
concept of the individual.

Rosetta Stone Preamble, Lines 9–3628

(9) ἐπειδὴ βασιλεὺς Πτολεμαῖος αἰωνόβιος, ἠγαπημένος ὑπὸ τοῦ Φθᾶ, θεὸς Ἐπιφανὴς
Εὐχάριστος, ὁ ἐγ βασιλέως Πτολεμαίου καὶ βασιλίσσης Ἀρσινόης, θεῶν Φιλοπατό-
ρων, κατὰ πολλὰ εὐεργέτηκεν τά θ᾽ ἱερὰ καὶ (10) τοὺς ἐν αὐτοῖς ὄντας καὶ τοὺς ὑπὸ
τὴν ἑαυτοῦ βασιλείαν τασσομένους ἅπαντας, ὑπάρχων θεὸς ἐκ θεοῦ καὶ θεᾶς καθά-
περ Ὧρος ὁ τῆς Ἴσιος καὶ Ὀσίριος υἱός, ὁ ἐπαμύνας τῶι πατρὶ αὐτοῦ Ὀσίρει, τὰ
πρὸς θεοὺς (11) εὐεργετικῶς διακείμενος ἀνατέθεικεν εἰς τὰ ἱερὰ ἀργυρικάς τε καὶ
σιτι⟨κ⟩ὰς προσόδους, καὶ δαπάνας πολλὰς ὑπομεμένηκεν ἕνεκα τοῦ τὴν Αἴγυπτον εἰς
εὐδίαν ἀγαγεῖν καὶ τὰ ἱερὰ καταστήσασθαι (12) ταῖς τε ἑαυτοῦ δυνάμεσιν πεφιλαν-
θρώπηκε πάσαις καὶ ἀπὸ τῶν ὑπαρχουσῶν ἐν Αἰγύπτωι προσόδων καὶ φορολογιῶν
τινὰς μὲν εἰς τέλος ἀφῆκεν,............. ἄλλας δὲ κεκούφικεν, ὅπως ὅ τε λαὸς καὶ οἱ ἄλλοι πάν-
τες ἐν (13) εὐθηνίαι ὦσιν ἐπὶ τῆς ἑαυτοῦ βασιλείας, τά τε βασιλικὰ ὀφειλήματα, ἃ
προσώφειλον οἱ ἐν Αἰγύπτωι καὶ οἱ ἐν τῆι λοιπῆι βασιλείαι αὐτοῦ, ὄντα πολλὰ τῶι
............. καὶ τοὺς ἐν ταῖς φυλακαῖς (14) ἀπηγμένους καὶ τοὺς ἐν αἰτίαις ὄντας
πλήθει, ἀφῆκεν,
............... τῶν ἐνκεκλ⟨η⟩μένων· προσέταξε
ἐκ πολλοῦ χρόνου ἀπέλυσε .................. δὲ καὶ τὰς προσόδους τῶν
ἱερῶν καὶ τὰς διδομένας εἰς αὐτὰ κατ᾽ ἐνιαυτὸν συντάξεις σιτικάς (15) τε καὶ ἀργυ-
ρικάς, ὁμοίως δὲ καὶ τὰς καθηκούσας ἀπομοίρας τοῖς θεοῖς ἀπό τε τῆς ἀμπελίτιδος
γῆς καὶ τῶν παραδείσων καὶ τῶν ἄλλων τῶν ὑπαρξάντων τοῖς θεοῖς ἐπὶ τοῦ πατρὸς
αὐτοῦ (16) μένειν ἐπὶ χώρας προσέταξεν .................... δὲ καὶ περὶ τῶν ἱερέων ὅπως μηθὲν πλεῖον
διδῶσιν εἰς τὸ τελεστικὸν οὗ ἐτάσσοντο ἕως τοῦ πρώτου ἔτους ἐπὶ τοῦ πατρὸς αὐτοῦ·

27 It may be no coincidence that Aeschylus, who in many respects is a representative of

the old ‘collective’ world, has very few transitive perfects active, with one exception:
the Prometheus, a tragedy about individual resistance to divine power (Prometheus 6
instances, the other six plays: 4).—Outside Attic authors the same tendency is found in
Pindar (see above, example (9)) and Herodotus (ex. (11)).
28 See footnote 18 above.
the greek perfect: subject versus object 55

ἀπέλυσεν
................ δὲ καὶ τοὺς ἐκ τῶν (17) ἱερῶν ἐθνῶν τοῦ κατ᾽ ἐνιαυτὸν εἰς Ἀλεξάνδρειαν
κατάπλου· προσέταξεν.................... δὲ καὶ τὴν σύλληψιν τῶν εἰς τὴν ναυτείαν μὴ ποιεῖσθαι, τῶν
τ᾽ εἰς τὸ βασιλικὸν συντελουμένων ἐν τοῖς ἱεροῖς βυσσίνων (18) ὀθονίων ἀπέλυσεν ................
τὰ δύο μέρη, τά τε ἐγλελειμμένα πάντα ἐν τοῖς πρότερον χρόνοις ἀποκατέστησεν ........................... εἰς
τὴν καθήκουσαν τάξιν, φροντίζων ὅπως τὰ εἰθισμένα συντελῆται τοῖ⟨ς⟩ θεοῖς κατὰ τὸ
(19) προσῆκον· ὁμοίως δὲ καὶ τὸ δίκαιον πάσιν ἀπένειμεν, ................. καθάπερ Ἑρμῆς ὁ μέγας
καὶ μέγας· πρ⟨ο⟩σέταξεν
.... . . . ............. δὲ καὶ τοὺς καταπορευομένους ἔκ τε τῶν μαχίμων καὶ τῶν
ἄλλων τῶν ἀλλότρια (20) φρονησάντων ἐν τοῖς κατὰ τὴν ταραχὴν καιροῖς κατελθόν-
τας μένειν ἐπὶ τῶν ἰδίων κτήσεων· προενοήθη .................. δὲ καὶ ὅπως ἐξαποσταλῶσιν δυνάμεις
ἱππικαί τε καὶ πεζικαὶ καὶ νῆες ἐπὶ τοὺς ἐπελθόντας (21) ἐπὶ τὴν Αἴγυπτον κατά τε
τὴν θάλασσαν καὶ τὴν ἤπειρον, ὑπομείνας δαπάνας ἀργυρικάς τε καὶ σιτικὰς μεγάλας,
ὅπως τά θ᾽ ἱερὰ καὶ οἱ ἐν αὐτῆι πάντ[ε]ς ἐν ἀσφαλείαι ὦσιν· παραγινόμενος (22) δὲ
καὶ εἰς Λύκων πόλιν τὴν ἐν τῶι Βουσιρίτηι, ἣ ἦν κατειλημμένη καὶ ὠχυρωμένη πρὸς
πολιορκίαν ὅπλων τε παραθέσει δαψιλεστέραι καὶ τῆι ἄλληι χορη⟨γί⟩αι πάσηι, ὡς
ἂν ἐκ πολλοῦ (23) χ⟨ρ⟩όνου συνεστηκυίας τῆς ἀλλοτριότητος τοῖς ἐπισυναχθεῖσιν εἰς
αὐτὴν ἀσεβέσιν, οἳ ἦσαν εἴς τε τὰ ἱερὰ καὶ τοὺς ἐν Αἰγύπτωι κατοικοῦντας πολλὰ κακὰ
συντετελεσμένοι, καὶ ἀντικαθίσας (24) χώμασίν τε καὶ τάφροις καὶ τείχεσιν αὐτὴν ἀξι-
ολόγοις περιέλαβεν,
.................... τοῦ τε Νείλου τὴν ἀνάβασιν μεγάλην ποιησαμένου ἐν τῶι ὀγδόωι
ἔτει καὶ εἰθισμένου κατακλύζειν τὰ (25) πεδία κατέσχεν ................ ἐκ πολλῶν τόπων ὀχυρώ-
σας τὰ στόματα τῶν ποταμῶν, χορηγήσας εἰς αὐτὰ χρημάτων πλῆθος οὐκ ὀλίγον καὶ
καταστήσας ἱππεῖς τε καὶ πεζοὺς πρὸς τῆι φυλακῆι (26) αὐτῶν, ἐν ὀλίγωι χρόνωι
τήν τε πόλιν κατὰ κράτος εἷλεν ........ καὶ τοὺς ἐν αὐτῆι ἀσεβεῖς πάντας διέφθειρεν, καθά-
περ [Ἑρμ]ῆς καὶ Ὧρος ὁ τῆς Ἴσιος καὶ Ὀσίριος υἱὸς ἐχειρώσαντο τοὺς ἐν τοῖς αὐτοῖς
(27) τόποις ἀποστάντας πρότερον. τοὺς ⟨δὲ⟩ ἀφηγησαμένους τῶν ἀποστάντων ἐπὶ
τοῦ ἑαυτοῦ πατρὸς καὶ τὴν χώραν ἐ[νοχλήσ]αντας καὶ τὰ ἱερὰ ἀδικήσαντας παρα-
γενόμενος εἰς Μέμφιν, ἐπαμύνων (28) τῶι πατρὶ καὶ τῆι ἑαυτοῦ βασιλείαι, πάντας
ἐκόλασεν
................ καθηκόντως καθ᾽ ὃν καιρὸν παρεγενήθη πρὸς τὸ συντελεσθῆ[ναι αὐτῶι τὰ]
προσήκοντα νόμιμα τῆι παραλήψει τῆς βασιλείας. ἀφῆκεν ............. δὲ καὶ τὰ ἐ[ν] (29) τοῖς
ἱεροῖς ὀφειλόμενα εἰς τὸ βασιλικὸν ἕως τοῦ ὀγδόου ἔτους, ὄντα εἰς σίτου τε καὶ ἀργυ-
ρίου πλῆθος οὐκ ὀλίγον· ὡσαύ[τως δὲ] καὶ τὰς τιμὰς τῶν μὴ συντετελεσμένων εἰς τὸ
βασιλικὸν βυσσίνων ὀθ[ονί]ων (30) καὶ τῶν συντετελεσμένων τὰ πρὸς τὸν δειγματι-
σμὸν διάφορα ἕως τῶν αὐτῶν χρόνων· ἀπέλυσεν ................ δὲ τὰ ἱερὰ καὶ τῆς ἀ[ποτεταγ]μένης
ἀρτάβης τ⟨ῆ⟩ι ἀρούραι τῆς ἱερᾶς γῆς, καὶ τῆς ἀμπελίτιδος ὁμοίω[ς] (31) τὸ κεράμιον
τῆι ἀρούραι, τῶι τε Ἄπει καὶ τῶι Μνεύει πολλὰ ἐδωρήσατο ................... καὶ τοῖς ἄλλοις ἱεροῖς ζώι-
οις τοῖς ἐν Αἰγύπτωι, πολὺ κρεῖσσον τῶν πρὸ αὐτοῦ βασιλείων φροντίζων ὑπὲρ τῶν
ἀνηκόν[των εἰς] (32) αὐτὰ διὰ παντός, τά τ᾽ εἰς τὰς ταφὰς αὐτῶν καθήκοντα διδοὺς
δαψιλῶς καὶ ἐνδόξως καὶ τὰ τελισκόμενα εἰς τὰ ἴδια ἱερὰ μετὰ θυσιῶν καὶ πανηγύ-
ρεων καὶ τῶν ἄλλων τῶν νομι[ζομένων], (33) τά τε τίμια τῶν ἱερῶν καὶ τῆς Αἰγύπτου
διατετήρηκεν ἐπὶ χώρας ἀκολούθως τοῖς νόμοις, καὶ τὸ Ἀπιεῖον ἔργοις πολυτελέσιν
......................... χορηγήσας εἰς αὐτὸ χρυσίο⟨υ⟩ τε κ[αὶ ἀργυρί]ου (34) καὶ λίθων πολυ-
κατεσκεύασεν
56 chapter 2

τελῶν πλῆθος οὐκ ὀλίγον, καὶ ἱερὰ καὶ ναοὺς καὶ βωμοὺς ἱδρύσατο ............... τά τε προσδεόμενα
ἐπισκευῆς προσδιωρθώσατο
.............................. ἔχων θεοῦ εὐεργετικοῦ ἐν τοῖς ἀνήκου[σιν εἰς τὸ] (35)
θεῖον διάνοιαν· προσπυνθανόμενός τε τὰ τῶν ἱ[ε]ρῶν τιμιώτατα ἀνανεοῦτο ἐπὶ τῆς
ἑαυτοῦ βασιλείας ὡς καθήκει· ἀνθ᾽ ὧν δεδώκασιν αὐτῶι οἱ θεοὶ ὑγίειαν, νίκην, κράτος
καὶ τἄλλ᾽ ἀγαθ[ὰ πάντα], (36) τῆς βασιλείας διαμενούσης αὐτῶι καὶ τοῖς τέκνοις εἰς
τὸν ἅπαντα χρόνον·

(9) Inasmuch as King Ptolemy, the beloved of Ptah, the God Epiphanes Eucha-
ristos, the Offspring of King Ptolemy and Queen Arsinoë, the Gods Philopa-
tores, hath given many benefactions, both to the temples, and (10) to those that
dwell therein, and unto all those who are subject to his dominion, being a God,
[and] born of a God and Goddess, even like Horus, the son of Isis and Osiris,
who avenged his father Osiris; and towards the gods (11) being full of benevo-
lent piety, hath dedicated to the temple revenues in money and in provisions;
and hath incurred great expenses in order that he might bring Egypt into a state
of prosperity, and might establish the temples; (12) and hath given away freely
of all the moneys which were his own; and of the taxes and dues which come to
him in Egypt, some he hath finally suppressed, and others he hath diminished,
so that the people (i.e., natives) and all the others (i.e., the foreigners and for-
eign settlers) (13) might be prosperous during his reign; and hath remitted to
the inhabitants of Egypt, and to those who were in the rest of his kingdom, the
debts which were due to the royal treasury, and were very many indeed in num-
ber; and those who were in the prisons, (14) and who were there because of the
[non-settlement of the cases] in which they were parties, and which were of
long standing, he hath set free from the charges which had been made against
them; and hath ordered that the revenues of the temples and the grant which is
made to them each year, both in corn (15) and money, and also the proper por-
tion [which is assigned] to the gods from the vineyards, and from the gardens
and the other things which belonged to the gods, should, as in the reign of his
father, (16) remain the same; and, in respect of the priests also, he hath com-
manded that they should pay no more as their fee for consecration than what
they had been assessed up to the first year of the reign of his father; and he hath
further released (17) those who belong to the priestly class from sailing down
[the Nile] each year to Alexandria; and he hath likewise commanded that men
shall no longer be seized [and compelled to serve] in the navy; and of the tax
upon byssus cloth which is paid to the royal treasury by the temples (18) he hath
remitted two-thirds; and whatsoever things had been neglected in times past
he hath restored and set in the order in which they should be; and he hath taken
care that the things which ought to be performed for the gods should be (19)
rightly carried out; and moreover, he hath administered justice unto every man,
the greek perfect: subject versus object 57

even like Hermes, the great and great; and he hath further ordered that those
of the soldiers who had returned, and of the others (20) who had held rebel-
lious opinions in the troubled times, should, having come back, be allowed to
keep possession of their own property; and he made provision that forces of
cavalry and infantry should be despatched, and ships also, against those who
were about to invade (21) Egypt, both by sea and by land, incurring [in the work]
great expenses in money and in corn, so that the temples and all that were in
the country might be in a state of security; and having gone (22) to Lycopolis,
that [city] which is in the Busirite nome, which had been taken and fortified
against a siege with an arsenal which was full of weapons of war and supplies
of every kind, of long standing being the disaffection of the impious men who
were gathered together into it, who had wrought much evil on the temples and
on all those who dwelt in Egypt, and having encamped (24) against them, he
surrounded it with mounds, and trenches, and marvellous walls; and when the
Nile made a great rise in the eighth year, and being abolut, as usual, to flood out
(25) the plains, he held in check [the river], having dammed up in many places
the mouths of the canals [which flow from it], and in the carrying out of this
work, spent no small sum of money; and having established both cavalry and
infantry to keep guard over [the canals], (26) he took by storm the city in a very
short time, and destroyed all the impious men who were therein, even as Her-
mes and Horus, the son of Isis and Osiris, brought into subjection in these very
same places (27) those who had rebelled; and the men who had led astray the
rebels in the time of his father, who had made trouble in the country, and had
treated the temples sacrilegiously, having come into Memphis for the purpose
of avenging (28) his father and his own sovereignty, he punished according to
their deserts at the time when hee duly appointed ceremonies for his receiving
the crown; and moreover he hath remitted in (29) the temples that which was
due to the royal treasury up to the viiith year [of his reign], which was no small
amount of corn and money; and moreover, he hath remitted the dues upon
the byssus cloth[s] which had not been paid into the royal treasury, (30) and
also the charges made for the examination of these which had been sent in for
the same period; and he hath freed the temples from [the tax of] one artaba for
each aroura of [temple-] land, and also the [tax of one] jar of wine (31) for each
aroura of vineyard; and to [the Bull] Apis and to [the Bull] Mnevis, and to the
other sacred animals in Egypt, he dedicated many gifts, far more indeed than
the kings who were before him, and he was careful in respect of what belonged
to them in (32) every particular; and for their burials he gave all that was needed
with magnificent generosity, and that which was necessary for private shrines
and for sacrifices, and for commemorative feasts, and for the ordinances by law
prescribed; (33) and the honourable estate of the temples and of Egypt he hath
58 chapter 2

maintained in a proper manner, according to traditional custom; and he hath

decorated the Temple of Apis with fine work, expending upon it gold, and silver,
and (34) precious stones, in no small quantities; and he hath founded temples,
and shrines, and altars; and he hath restored those which needed repairs, hav-
ing the zeal of a beneficent god in the matters which relate to (35) the divine
service, and having discovered which of the temples were most held in hon-
our, he restored them during his reign, as was meet.—In return for all these
things the gods have given him health, and victory, and power, and all other
good things, and may his royal (36) dominion remain with him, and with his
children for all time.

Bibliography

Armstrong, D., ‘The Ancient Greek Aorist as the Aspect of Countable Action’ in Tede-
schi and Zaenen (1981) 1–12.
Chantraine, P., Histoire du parfait grec (Paris 1927).
Comrie, B., Aspect (Cambridge 1976).
Comrie, B., ‘Aspect and Voice: Some Reflections on Perfect and Passive’, in Tedeschi and
Zaenen (1981) 65–78.
Curtius, G., Die Bildung der Tempora und Modi im Griechischen und Lateinischen (Berlin
1846).
Dik, S.C., Functional Grammar (Amsterdam 1978).
Hartmann, F., Review of Chantraine 1927, Gnomon 6 (1930) 177–191.
Keil, F., ‘Untersuchungen zum Perfektgebrauch Herodots’, Glotta 41 (1963) 10–51.
Magnien, V., Review of Chantraine 1927, Revue des Études Grecques 42 (1929) 460–461.
Mayser, E., Grammatik der griechischen Papyri aus der Ptolemäerzeit, Band II: Satzlehre,
vol. 1 (Berlin/Leipzig 1926).
McKay, K.L., ‘The Use of the Perfect Down to the Second Century A.D.’, Bulletin of the
Institute of Classical Studies 12 (1965) 1–21.
McKay, K.L., ‘On the Perfect and Other Aspects in Greek Non-literary Papyri.’Bulletin of
the Institute of Classical Studies 27 (1980) 23–50.
Parkinson, R., Cracking Codes: The Rosetta Stone and Decipherment (London 1999).
Rijksbaron, A., The Syntax and Semantics of the Verb in Classical Greek: An Introduction
(Amsterdam 1984; third ed. Chicago 2006)
Ruijgh, C.J., Autour de ‘τε épique’: Études sur la syntaxe grecque (Amsterdam 1970).
Saïd, S., La faute tragique (Paris 1978).
Tedeschi, Ph.J., Zaenen, A. (eds), Tense and Aspect (New York 1981).
Wackernagel, J., Studien zum griechischen Perfektum (Göttingen 1904; repr. in Kleine
Schriften (Göttingen 1953) 1000–1021).
the greek perfect: subject versus object 59

Wackernagel, J., Vorlesungen über Syntax, Erste Reihe (Basel 1920).

Wallis Budge, E.A., The Decrees of Memphis and Canopus: The Rosetta Stone (London
1904).
Wallis Budge, E.A., The Rosetta Stone in the British Museum and the Decipherment of
Aegyptian Hieroglyphs (London 1929).
 chapter 3

The Discourse Function of the Imperfect

Abstract

In this paper the functioning of the imperfect is examined, both within the overall
Greek verbal system and on discourse level. To this end first the temporal properties
of the various indicatives are discussed, the main question being: how much tem-
poral information does a given indicative provide us with? It can be shown that the
imperfect is the past tense par excellence of Greek, and that the aorist indicative is
not inherently a past tense. If it were, a number of non-past uses of the aorist would
be very difficult to explain. The common value of all aorists, it is argued, is that they
express ‘closedness’. Its use as a past tense may, in its turn, be explained as being due
to the general tendency to associate ‘perfectivity’ with ‘past’. The imperfect, then, is,
and the aorist is not, inherently connected with narrative, i.e. the reporting of past
events.
The main issue of the second part is the functioning of the imperfect on the dis-
course level. Apart from locating a state of affairs as non-closed, i.e. as ongoing. Thereby
an imperfect creates a ‘framework’ or ‘time anchor’ for other states of affairs, and raises
expectations as to what is going to happen next. Mostly, these expectations are fulfilled
in the immediately following context, but, as an analysis of a large part of Herodotus’
Histories makes clear, imperfects also occur regularly at a break in the narrative, e.g.
when the author puts in a long ethnographic description. The fulfilment of our ‘nar-
rative expectations’ is, in that case, temporarily suspended, and the imperfect signals
that the author will continue his original story elsewhere. By this ‘cataphoric’ function
the imperfect serves as a powerful means to establish textual cohesion.

1 Introduction

It is not immediately clear, perhaps, why only the imperfect is mentioned in

the title of this paper, and not also, to name the most plausible other candi-
date, the aorist indicative. The reasons for giving prominence to the imperfect
will become clear, I hope, in what follows.
In the first part of this paper I will make some observations on the place of
the imperfect within the overall system of the Greek verb; in the second part
some aspects of the functioning of the imperfect on discourse level will be dis-
cussed.

© koninklijke brill nv, leiden, 2019 | doi:10.1163/9789004386129_005

the discourse function of the imperfect 61

Thanks to the treatment of the verbal system of Ancient Greek both in

school grammars and in large-scale scholarly grammars, we are wont, more
or less unconsciously, I think, to associate one particular verb form with one
particular meaning. This habit, while being perfectly understandable, if only
for didactic reasons, may nevertheless result in a grossly simplified picture of
what is, in reality, a far more complex | whole. To mention only one instance 238
of this risk, which I noticed rather often during grammar courses for first year
students: there is a general reluctance to assign to the present indicative its
habitual or omnitemporal value, although this value is at least as important as
the ‘actual’ value given in the verbal paradigm of the schoolbooks. In this case,
the traditional name ‘present’ tense probably reinforces the tendency to view
this tense primarily as referring to the ‘present’.
To counterbalance, as it were, the traditional treatment of the Greek verb,
I have asked myself the following question: how much temporal information
does a given indicative on its own provide us with, i.e. to what extent are we
informed by a given indicative about the temporal location of the state of
affairs expressed by that indicative? The search for an answer to this question
produces, I think, some unexpected results. While searching, we may use as
a motto Wackernagel’s remark: ‘Schon in den Indikativformen der Tempora
stimmt vieles nicht’ (1926: 151). The investigation has the form of a discussion
of a number of representative examples of the various indicatives, with some
additional comments on further features in the notes.

2 The Temporal Properties of the Indicatives

2.1 Present Indicative

A state of affairs expressed by the present indicative may have various temporal
locations. In (1) and (2) we are dealing with a ‘habitual’ and an ‘omnitemporal’
present, respectively:

(1) συσσιτοῦμεν … ἐγώ τε καὶ Μελησίας ὅδε καὶ ἡμῖν τὰ μειράκια παρασιτεῖ (Pl.
La. 179b)

My friend Melesias and I take our meals together, and our boys share our
table

(2) ἁ χεὶρ τὰν χεῖρα νίζει (proverb)

one hand washes the other

 62 chapter 3

Both (1) and (2) express non-single states of affairs which, although they are,
strictly speaking, not going on at the speech moment, may nonetheless, with
Comrie (1985: 39), be considered instances of location at the speech moment,
because ‘the habit does hold at the present moment’ (= speech moment - AR).
They differ in that in the habitual use the subject constituent, as well as any
other constituent, typically refers to definite entities, whereas in the omnitem-
poral use the nominal constituents typically have generic reference. Also, habit-
ual states of affairs may in principle come to an end, becoming past habitual
239 states of affairs; | omnitemporal states are, rather, universally valid or ‘tempo-
rally unrestricted’ (Lyons 1977: 680).1
Unlike (1) and (2), which are time-bound, i.e. linked to the world as we know
it, (3) is timeless, i.e. the state of affairs that it describes ‘is outside time alto-
gether’ (Lyons 1977: 680):

(3) τὰ δὶς πέντε δέκα ἐστίν

Two times five makes ten

Of course, much more could be said about the similarities and discrepancies
between habitual, omnitemporal and timeless statements, e.g. concerning the
precise referential properties of the nominal constituents involved and the
types of verbs occurring in these statements, but this would fall outside the
scope of this paper.2
Next, consider (4):

(4) ὁ δ᾽ ὦμος οὑτοσί πιέζεται (Ar. Ra. 30)

My shoulder here is stuck

This is, of course, an example of a state of affairs that is strictly located at the
speech moment and is, moreover, presented as ongoing, as not having come to
completion.
With the following examples we are moving out of the sphere of the present.
(5) and (6) illustrate the use of the present to express a past state of affairs (the
‘historic present’), which may occur both in main (5) and in subordinate (6)
clauses:

1 Among omnitemporal statements one may also reckon expressions like Ὅμηρος λέγει.
2 Some observations on these points may be found in Lyons (1977: 679f.) and Dahl (1985: 99f.).
the discourse function of the imperfect 63

(5) … ἐχώρεε ἔξω. καὶ ἡ γυνὴ ἐπορᾷ μιν ἐξιόντα (Hdt. 1.10.2)

he slipped from the room. And the woman saw (‘sees’) him as he went out

(6) ἐπειδὴ … ἀφίεται τοῦ δόρατος, τότ᾽ ἤδη καὶ … οὐκέτι οἷοί τ᾽ ἦσαν τὸν γέλωτα
κατέχειν (Pl. La. 184a)

When he let (‘lets’) go the spear, then (the others) could not restrain their
laughter anymore either

Finally, a present indicative state of affairs may also be located in the future:

(7) καὶ εἰ αὕτη ἡ πόλις ληφθήσεται, ἔχεται καὶ ἡ πᾶσα Σικελία (Th. 6.91.3)

And if that town be taken, all of Sicily will be (‘is’) captured | 240

(8) τί οὖν … οὐκ ἐρωτᾷς; :: ἀλλ᾽ ἐρήσομαι (Pl. Ly. 211d)

Why not ask him then? :: But I am going to3

A state of affairs which is characterized by the present indicative may be

located, then, in the past, the present—in an extended sense—and in the
future. This could be taken to mean that the verb form as such does not provide
specific temporal information, and, consequently, that the present indicative
is not an absolute tense, i.e. a tense ‘which includes as part of its meaning
the present moment as deictic center’ (Comrie 1985: 36), but, rather, a rela-
tive tense, i.e. a tense ‘where the reference point for location of a situation
is some point of time given by the context’ (Comrie 1985: 56). On the other
hand, it would seem more plausible, both on language-internal and on cross-
linguistic grounds, to consider the habitual/omnitemporal and present values
of the present indicative as basic, and the past and future uses as derivative
or secondary. First, the past or future location of a present indicative depends
much more clearly on contextual information (cf. ἐχώρεε in (5), ληφθήσεται in
(7), τί οὖν οὐκ in (8)) than the location in the present. Second, as will become
clear presently, there are two absolute tenses to locate a state of affairs in the

3 Questions introduced by τί οὖν οὐκ are, from a pragmatic point of view, urgent variants of the
imperative. They refer, therefore, to the future. This is borne out by the fact that, in principle,
the answer contains an imperative or a future indicative (as in (8)), not a present indicative.
Cf. also (21).
 64 chapter 3

past or in the future, imperfect and future indicative, respectively. Third, the
habitual/omnitemporal and the present values of the present indicative are
attested as basic in a wide variety of other languages.4 If, then, the past and
future values are, indeed, secondary, they may be explained as pragmatically
motivated. Being ‘pseudo-presents’, they function as marked variants of the
241 more neutral ‘normal’ tenses in contexts like those of (5) and (7).5 |

2.2 Perfect Indicative

As regards the temporal properties of the perfect indicative more or less the
same can be said as with respect to the present indicative. (Note, though, that
there is no ‘historic’ perfect.) It is basically opposed to the present indicative by
its aspectual value, since it expresses both that a state of affairs is closed and
that a state exists which is the result of the completion of that state of affairs.
This state may be habitual/omnitemporal (9), it may be located at the speech
moment (10), or it may be located in the future (11).

(9) εἰ γὰρ καὶ ἐναντιωθῆναί τι θέλει, βεβούλευται μὲν οὐδέν ἧσσον εὖ (Hdt. 7.10δ2)

For even though it be thwarted later, yet nonetheless has the plan been
good

(10) γῆν … ἐκτήμεθα ὀλίγην (Hdt. 9.122.2)

We have acquired (i.e. possess) little land6

4 And yet some doubts remain. Strictly speaking, it could be defended that a pres.ind. does not
‘include as part of its meaning the present moment as deictic centre’. Thus, to interpret πιέζε-
ται in (4) as located at the speech moment, it is crucial to know that this utterance occurs in
direct speech; again, to interpret συσσιτοῦμεν in (1) as unequivocally habitual, one must know
that the state of affairs in question is not occurring at the speech moment. In both cases, then,
contextual support is needed, after all. All this is completely different in the case of e.g. the
imperfect, see below.
5 See S-D (2, 271ff.) for a brief discussion of some of the effects of the use of the historic present,
and (273) for the ‘praesens pro futuro’. An interesting discussion of the latter phenomenon in
Russian may be found in Barentsen (1985: 199–323).
6 In this use the perfect has ‘current relevance’. (Cf. Quirk et al. 1972: 91 for this value of the
English perfect; also Comrie 1976: 56ff.). It is perhaps appropriate to note here that the per-
fect tense in many languages is often considered a past tense, whereas in reality it is not;
cf. e.g. Dutch ik zal blijven zoeken tot ik het gevonden heb, in which the perfect gevonden heb
refers to a future state of affairs (the same holds for the equivalent sentences in e.g. French
and German).
the discourse function of the imperfect 65

(11) εἰ … πτῶμα θανάσιμον πεσῇ | τέθνηκα κἀγώ (E. El. 686–687)

If you fall to death, I too am dead

Here, too, the habitual/omnitemporal and the present values are probably
basic, whereas the use of the perfect in a future situation (11) would seem to
be derivative, serving pragmatic purposes.7

2.3 Future Indicative

Generally speaking, the future indicative may be considered an absolute tense,
since it locates the state of affairs after the speech moment. Compare ἐρήσομαι
in: | 242

(12) = (8) τί οὖν οὐκ ἐρωτᾷς; :: ἀλλ᾽ ἐρήσομαι

Why not ask him then? :: But I am going to

Occasionally, however, a habitual/omnitemporal future occurs, as in:

(13) ὄρθρου δὲ γενομένου λοῦνται καὶ ἀμφότερον ἄγγεος γὰρ οὐδένος ἅψονται πρὶν
ἂν λούσωνται (Hdt. 1.198)

and at dawn they both take a bath; for they will touch no vessel before
they have bathed8

Note, finally, that a ‘future in the past’, as e.g. would in John left for the West-
Indies. He would never come back seems to be lacking.9

7 E.g. by presenting, as in (11), a future state—which could also have been expressed by the
future perfect, see 2.4.—as ‘being present’, whereby the consequences of the fulfilment of
the condition are presented in a more forceful way than by τεθνήξω (fut. perf.: ‘I will be dead’)
or θανοῦμαι (future: ‘I will die’). Cf. also K-G (1, 150). It should be noticed, finally, that such a
‘futural’ perfect indicative may only occur if the context makes it sufficiently clear that the
future is involved (as is done by the conditional clause in (11)). Otherwise a future state can
only be expressed by the future perfect, see. 2.4.
8 Possibly, the future indicates that the state of affairs ‘touching a vessel’ follows upon that of
‘taking a bath’. Observe also the presence of the anterior πρίν-clause (‘not before’, ‘only after’).
Cf. K-G (1, 172 Anm. 1) for some further examples of habitual futures, and Dahl (1985: 100) on
the rarity of examples combining ‘HAB and FUT’.
9 Such a ‘narrative’ future is expressed by ἔμελλε + future infinitive. Cf. Basset (1979: 60ff.) and
Ruijgh (1985: 323).
 66 chapter 3

2.4 Future Perfect

To all appearances the future perfect is an absolute tense: it always locates the
state that is the result of the completion of the state of affairs concerned after
the speech moment (in contradiction to e.g. the English future perfect, cf. Com-
rie 1985: 70). An example is:

(14) καί με ἐὰν ἐξελέγχῃς, οὐκ ἀχθεσθήσομαί σοι ὥσπερ σὺ ἐμοί, ἀλλὰ μέγιστος
εὐεργέτης παρ᾽ ἐμοὶ ἀναγεγράψῃ (Pl. Grg. 506c)

And if you refute me, I shall not be vexed with you as you were with me;
you will only be recorded in my mind as my greatest benefactor10

2.5 Imperfect
One example will suffice, since the picture presented by the imperfect is very
243 simple: |

(15) Κροῖσος ἦν Λυδὸς μὲν γένος … (Hdt. 1.6.1)

Croesus was a Lydian by birth

The imperfect always has the speech moment as its reference point, or, to put
it in Comrie’s terms, the deictic centre for the imperfect is the speech moment.
This means that the imperfect locates a state of affairs unambiguously in the
past: it is, therefore, an absolute tense.11 It has, of course, also an aspectual
value:12 it presents a state of affairs as ‘not closed’, as ‘on-going’, at a given
moment in the past. Some effects of this will be discussed in the second part of
this paper.

10 Note the opposition between the future ἀχθεσθήσομαι, which, being formed from the aorist
stem (cf. ἠχθέσθην), may have a slightly ingressive value (‘I will not (even) begin to be
vexed’, although this is a matter of dispute, cf. S-D 2, 265–266), and the state expressed
by ἀναγεγράψῃ (‘Du wirst in der Liste meiner Wohlthäter stehen’, K-G 1, 180).—The ‘future
state’ is absent from Dahl’s categories of the future tense (Dahl 1985: 103–112).
11 With all that, the state of affairs is, of course, as yet not located on the time axis, i.e. it
is not specified in terms of calendar time. It goes without saying that a Greek historian
was confronted by far more serious problems of chronological ordering than his modern
colleagues. For some issues connected with calendar time in Herodotus see the masterly
article by Strasburger (1965).
12 I use ‘aspectual’ in a rather informal way (and neglect voice differences): it roughly cov-
ers that part of the meaning of a given verb form that does not concern the location of
the state of affairs on the time axis. This does not exclude, of course, that ‘aspect’ may
the discourse function of the imperfect 67

2.6 Pluperfect
This is an interesting tense because, while far less frequent than the imperfect,
it shares with the latter the unambiguous temporal location in the past. Two
examples are:

(16) καὶ Βαβυλὼν μὲν οὕτω τότε πρῶτον ἀραίρητο (Hdt. 1.191.6)

Thus had Babylon then for the first time been taken

(17) … ἔπεμπε Ξέρξης κατάσκοπον … ἀκηκόεε δ᾽ ἔτι ἐὼν ἐν Θεσσαλίῃ ὡς … (Hdt.

1.208.1)

… Xerxes sent a spy; while still in Thessaly he had learned that …

The pluperfect too is, then, an absolute tense. It is opposed to the imperfect
(and to the aorist, see below) by its aspectual value: whereas the imperfect
presents a state of affairs as on-going in the past, the pluperfect expresses both
that a state of affairs is closed and that a state exists which is the result of the
completion of that state of affairs. Compare (16), where ἀραίρητο comes at the
end of the long narrative concerning the capture of Babylon. It expresses, thus,
‘current relevance’ in the past, just as the perfect expresses ‘current relevance’
in the | present.13 Observe that the pluperfect does not inherently express that 244
the completion of the state of affairs is prior to another reference point, i.e. it is
not inherently a past-in-the-past. Whether this applies depends from contex-
tual factors (as in (17)). The past-in-the-past use may be termed, with Comrie
(1985: 65), ‘absolute-relative’: it signals both that the state is located in the past
and that it is located before a reference point in the past.

to a large extent cover phenomena of a temporal nature. Thus, ‘ongoingness’ may imply
‘simultaneity’, just as ‘closedness’ may imply ‘anteriority’.
13 Compare Comrie (1976: 56), who calls a comparable English construction a ‘perfect-in-
the-past’; also Dahl (1985: 146–147).—Note that in a way the pluperfect, too, may be said
to express ‘on-goingness’. A state is, after all, something which is not-closed and it may,
therefore, come to an end (just as with states of affairs in the imperfect). This explains, in
fact, its use in (16): after some time the Persians lose Babylon again to the Babylonians. Cf.
also the presence of πρῶτον, which explicitly signals that the state in question will come
to an end and will be followed by another capture. The latter is described in 3.151ff., where
an aorist is used (αἱρέθη in 3.159.1): there is no explicit mention of a state, and Herodotus
does not, in fact, come back to Babylon. This implies that Babylon remained in the hands
of the Persians.
 68 chapter 3

2.7 Aorist Indicative

The aorist indicative is an excellent example of the tendency to view a tense as a
whole in terms of one of its uses (cf. the present tense, section 2.1.): it is primar-
ily considered a past tense. On the level of its actual use this is, of course, not
incorrect, but it is another matter whether the past value is an inherent feature,
just as in the case of the imperfect, i.e. whether the aorist is an absolute tense. I
think it is not. If it were, it would be very difficult indeed, and sometimes down-
right impossible, to account for its other uses. Consider the following examples:

(18) στρατὸς πολλὸς ὑπὸ ὀλίγου διαφθείρεται κατὰ τοιόνδε· ἐπεάν σφι ὁ θεὸς …
φόβον ἐμβάλῃ …, δι᾽ ὦν ἐφθάρησαν ἀναξίως ἑωυτῶν (Hdt. 7.10ε)

(Thus) a numerous host is destroyed by one that is lesser in the following

way: when a god has sent panic among them, they perish unworthily

(19) εἰ μέν κ᾽ αὖθι μένων Τρώων πόλιν ἀμφιμάχωμαι, | ὤλετο μέν μοι νόστος ἀτάρ
κλέος ἄφθιτον ἔσται (Il. 9.412)

If I stay here and fight beside the city of the Trojans, my return home is
gone, but my glory shall be everlasting (transl. Lattimore)

(20) φὴς … ἐλᾶν στρατόν … ἐς τὴν Ἑλλάδα. καὶ δὴ καὶ συνήνεικε … ἑσσωθῆναι (–).
οὐκ ὦν ἀμφοτέρῃ σφι ἐχώρησε (Hdt. 7.10β1–2)

245 You say that | you will march your army to Hellas. Now ⟨I will suppose that⟩
matters have so fallen out that you are worsted (–). Well, ⟨be it granted
that⟩ they win not success by land and by sea

(21) τί οὖν … οὐκ … ἔλεξας …; :: ἄκουε δή … (X. Cyr. 2.1.4)

Why then not tell us? :: All right, listen

(22) διὸ ὑμέας νῦν ἐγώ συνέλεξα (Hdt. 7.8a.2)

For this reason I have now summoned you together

(23) ἔπαιζε δὲ μετ᾽ ἄλλων … καὶ οἱ παῖδες παίζοντες εἵλοντο τοῦτον … τὸν … παῖδα
(Hdt. 1.114.1)

He was playing with the others … The boys in their play chose … that boy
(for their king)
the discourse function of the imperfect 69

(24) τὴν δὲ μεταλλαγὴν ταύτην … προηγόρευσε ἔσεσθαι (Hdt. 1.74.2)

(Thales) had foretold that this change would happen

As appears from these examples, the aorist may express a habitual/omnitem-

poral (18),14 a future ((19), (20)15 and (21)16), a past, both an | immediate (22) 246
and an undetermined (23), past,17 and a past-in-the-past (24) state of affairs.
In none of these examples, except (23), an imperfect could have been used;
the aorist is, then, crucially different from the imperfect.18 What all aorist
indicatives have in common is not that they refer to the past, but, rather, that
they, just like e.g. the aorist participle, express a closed state of affairs, i.e. closed,
vis-à-vis a reference point given in the context, this reference point not being

14 The differences between habitual present and aorist indicatives are a matter of dispute.
Possibly the aorist in (18) expresses a concrete instance of the state of affairs involved
(in other words, it singles out a ‘token’ of the ‘type’ expressed by the present indicative),
this instance being of prime importance in that particular context. That the aorist also
expresses a habitual state of affairs is in that case of secondary importance. Alternatively,
the aorist in (18) may have a nuance of (sudden) completion, as against a more gradual
value of the present indicative. (A similar example is Hdt. 3.82.3.) In other contexts, how-
ever, especially when a series of states of affairs is involved, the aorist apparently expresses
a closed, i.e. anterior, state of affairs vis-à-vis another state of affairs, as in e.g. Hdt. 1.194.4.
See further Ruijgh (1971: 255 ff.) and Péristérakis (1962: passim). Observe, finally, that, if the
aorist is not intrinsically a past tense, its use in habitual statements may not be as big a
problem as it is usually thought to be.
15 Another instance of a future indicative coordinated with an aorist indicative is Hom.
Il. 4.161; for further examples of aorists expressing a supposition see K-G (1, 202–203). Cf.
also Wackernagel (1926: 176–177): ‘Der Aorist kann etwa auch von einer zukünftigen Hand-
lung gebraucht werden.’ Wackernagel would probably consider the aorists in (20) as cases
of ‘etwas das in einem künftigem Moment als Vergangenheit betrachtet … wird’, see his
discussion of Hdt. 8.102.2 on p. 177. Interestingly, Russian has a similar ‘future’ use of the
so-called perfective preterite, see Barentsen (1985: 224–226). With some hesitation, Bar-
entsen explains this use as being due to a shift in the orientation of the speaker. This, too,
would seem to imply that this ‘preterite’ is no more inherently a past tense than the Greek
aorist.
16 For my taking this aorist as having future reference see note 3, on a similar use of the
present.
17 These two differ in that the former typically occurs in direct speech and the latter in nar-
rative discourse. See also Rijksbaron (1984: 28 ff.).
18 This basic difference has been duly noted by Wackernagel (1926: 176): ‘Zunächst von ein
paar Funktionen, wo die Konkurrenz des Imperfekts nicht in Betracht kommt’. Cf. also
Kuryłowicz (1964: 104): ‘From the synchronic point of view the indicative of the Greek
aorist has therefore the primary function of a past perfective and a secondary function of
a general perfective’.
 70 chapter 3

temporally fixed. This means, that, in principle, the aorist indicative is charac-
terized by an aspectual value only, and not by a specific temporal one. The fact
that the aorist is used predominantly as a past tense may be explained as due
to the phenomenon that, all things being equal, the presentation of a state of
affairs as closed is ‘likely to be a matter of record, in retrospect’, as Givón notes,
in a discussion of the general tendency of perfectivity to be associated with
the past (1984: 277–278). This past value would seem to be crucially connected
with the speech moment serving as the reference point: the feature ‘closed
before the speech moment’, i.e. anterior to the speech moment, makes us rein-
terpret the aorist as a past tense. Otherwise than with the imperfect, then, the
value [past] is not part of the meaning of the aorist indicative; it is, rather, in
Gricean terms, a conversational implicature, be it a very strong one.19 As to its
functioning in a narrative, i.e. in a report of past events, in alternation with
the imperfect, this rests solely upon the replacement, as it were, of the speech
moment as a reference point by a reference point provided by a given imper-
fect. As soon as there is no such imperfect, the reference point is—in historical
discourse—the speech moment. When this happens there is, strictly speaking,
no narrative; rather, we are dealing with events that have as their common fea-
ture that they all simply occurred before the speech moment (or the ‘writing’
247 moment), the so-called constative use of the aorist.20 |
A clear case in point is the ‘real’ beginning of Thucydides’ account of the
Peloponnesian War (ex. (25)):

(25) Ἐπίδαμνός ἐστι πόλις (–). Ταύτην ἀπῴκισαν … οἰκιστὴς δ᾽ ἐγένετο … ξυνῴκι-
σαν … ἐγένετο … ἐφθάρησαν … ἐστερήθησαν. Τὰ δὲ τελευταῖα πρὸ τοῦδε τοῦ
πολέμου ὁ δήμος αὐτῶν ἐξεδίωξε τοὺς δυνατούς, οἱ δὲ ἀπελθόντες μετὰ τῶν
βαρβάρων ἐλῄζοντο τοὺς ἐν τῇ πόλει … (Th. 1.24)

The city of Epidamnos is … The city was colonized by [Legrand, better: ‘a

été fondée par’] … the leader was … … joined in establishing the colony …

19 See Levinson (1983: ch. 3) for an illuminative discussion of this difficult notion.—I should
add that in some cases the aorist indic. always has past reference, viz. (a) when it occurs
in conditional and temporal clauses, (b) when it is accompanied by the so-called ‘ἄν iter-
ativum’, (c) when the counterfactual ἄν is present, and (d), when it occurs in unrealizable
wishes. I cannot go into this matter any further, here.
20 For this reason Dahl is, in my opinion, mistaken when he considers Caesar’s famous words:
veni, vidi, vici (Greek: ἦλθον, εἶδον, ἐνίκησα) as an example of an ‘ideal narrative’ (1985: 112).
It may be doubted whether this is formally characterized as a narrative; in Greek, at least,
it is not. [See also ch. 7 in this volume, pp. [341]–[352].]
the discourse function of the imperfect 71

(the city) became (great) … they were crippled … and deprived (of most
of their power). Finally, just before the Peloponnesian War, the populace
expelled the aristocrats, and they, making common cause with the bar-
barians, came back and plundered those who were in the city

Thucydides first presents some background information by way of a number of

aorist states of affairs, which are past vis-à-vis the speech moment; the narra-
tive proper starts with the imperfect ἐλῄζοντο: this provides the reference point
for the states of affairs presented subsequently.21 The imperfect creates, then,
to use a recent term, a ‘time anchor’, and that in two ways: it unequivocally
brings us to the past and it provides, by its aspectual value of ‘non-closedness’,
a framework for other states of affairs (see below).22
If the above is correct remarks like that by Hirt (1928: 173) are clearly beside
the point: ‘Wenn nun aber das Augment nicht notwendig zur Bezeichnung
der Vergangenheit war, was diente denn dazu? Zunächst finden wir in dieser
Verwendung den Indikativ Aoristi. (–) Das Griechische und Indische besitzen
ferner ein Imperfektum …’ The point is not that Greek ‘also possesses’ an imper-
fect to express ‘Vergangenheit’, but rather the other way round: there is also an
aorist to express past actions.23 The essential difference between imperfect and
aorist indicative can be seen by comparing their distributional properties in the
various discourse types: | 248

(27) Frequency of finite verb forms in Hdt. 7.8–10 (direct speech; ca. 6 pp. OCT)
Present indicative 73
Perfect indicative 8
Future indicative 17
Future perfect –
Imperfect (non-modal) 5
Pluperfect (non-modal) 1

21 To put it somewhat differently: the imperfect is the sure sign that we are dealing with a
narrative (cf. Longacre and Levinsohn 1978: 107).
22 ‘Time anchor’ is used, be it in a slightly different way, by e.g. Bertinetto (1986). See also
note 25.
23 It should be noticed, incidentally, that the aorist is conspicuously absent from Plato’s dis-
cussion of past, future and present verb forms in Prm. 141d–e.—Observe also that the fact
that the aorist indicative for some persons has so-called secondary endings—just like the
imperfect—does not tell against a non-past analysis, since the temporal and aspectual
information provided by these endings is virtually redundant. The aor. indic. is distin-
guished from the present indic, and imperf. primarily by its being formed from a different
stem. See Rix (1976: 239).
 72 chapter 3

Aorist indicative (non-modal) 24

Subjunctive 17
Optative 7
Imperative 10

(28) Frequency of verb forms in Hdt. 5.92β.-η4 (narrative; ca. 4 pp. OCT)
Present indicative 10
Imperfect 34
Pluperfect 3
Aorist indicative 31
[Optative 1]

In ‘real’ direct speech, i.e. interactive speech (or conversation), cf. (27), the
(non-modal) imperfect is practically absent, whereas the aorist is quite com-
mon, even more so than the future indicative. This can be readily explained,
for what counts, in a debate, are the results, the facts of the past, rather than
the circumstances under which these results came about. In the other passage,
however, a narrative passage (cf. (28)), the imperfect is the most common tense.
The latter passage is particularly interesting, because it occurs in direct speech,
viz. in a speech delivered by the Corinthian Soclees, in Sparta. What we have
here, then, is an embedded story (the famous Cypselus-story); the signal that
we are leaving interactive speech and are passing over to the narrative is pro-
vided by the two imperfects of ex. (29):

(29) Κορινθίοισι … ἦν πόλιος κατάστασις τοιήδε· ἦν ὀλιγαρχίη … (Hdt. 5.92β.1)

249 The Corinthian state was ordered in the following way. The Few ruled …24 |

2.8 Conclusion of Sections 2.1.–2.7.

From the above overview it may be concluded that, when viewed in terms of
the verb system, the imperfect is of prime importance: it is the past tense par
excellence. This is, of course, not the whole story. As has already been noticed
in passing, the imperfect not only locates a state of affairs in the past, it also
presents this state of affairs as not-closed, as on-going. Thus, it typically creates

24 Up to these imperfects, no imperfect occurs (apart from a modal one).—Observe also that
in the, by their very nature ‘non-narrative’, Mycenean texts no imperfects occur, whereas
there are some instances of the aorist indicative. Cf. Duhoux (1988). [See also ch. 7 in this
volume, p. [351].]
the discourse function of the imperfect 73

a temporal framework for other states of affairs, serving, as their ‘time anchor’.25
The latter are, thus, simultaneous with the state of affairs expressed by the
imperfect.26 Aorist states of affairs, on the other hand, expressing closed events,
typically fall within the framework created by the imperfect. Some examples
are (23) and (30):

(30) διέβαλλον … ἔτι τούτων ταῦτα λεγόντων … ἐνέβαλε νηὶ Ἀττικῇ Σαμοθρηικίη
νηῦς (Hdt. 8.90.1)

They accused (the Ionians) … While they yet spoke as aforesaid, a Samoth-
racian ship charged an Attic one

Note that in (30) ἐνέβαλε may be interpreted as putting an end to the διαβάλλειν.
In fact, the aorist often has the function of bringing an open-ended imperfect
state of affairs to completion, as in ἐπορεύετο … ἀπικόμενος δέ … (‘he marched …
Having arrived …’, e.g. Hdt. 4.89.3). This means that the aorist gives direction to
the narrative; a narrative consisting solely of imperfects would show virtually
no progress.27
So far the discussion has dealt with the importance of the imperfect on
what may be called the micro-level, i.e. the level of a small scale narrative | 250
unit, e.g. a battle, where the various states of affairs making up that battle are
grouped together, to put it somewhat impressionistically. However, the imper-
fect can also be shown to be of major importance on what may be called the

25 I do not think, therefore, that the following remark by Bertinetto (1986: 68) is valid for
Ancient Greek (and I strongly doubt whether it is valid for the Romance imperfect or
the English past progressive): ‘The event described by the Imperfect tense (and by the
Past Progressive) evokes some other event, which is viewed as (partially) overlapping with
the first one. Thus, the Romance Imperfect and the English Past Progressive call … for a
‘time anchor’.’ ‘Evokes’—yes; but the idea that the imperfect needs rather than provides
temporal support is mistaken. Nor do I agree with Vet and Molendijk (1986), who hold
that the French imparfait has to be temporally ‘bound’ by a verb form in the preceding
text.
26 Rather than the other way round. For the simultaneity function of the Greek imperfect cf.
especially Hettrich (1976). Incidentally, it is worth noticing that Kühner, in the 1835 edition
of his grammar, gives ‘Gleichzeitigkeit’ as the essential feature of the imperfect (1835: 62;
67). This has disappeared in the third, and last, edition.
27 In Hopper and Thompson’s terminology (1980: 285ff.), aorist states of affairs might be
called ‘foregrounded’ and imperfect ones ‘backgrounded’, provided that these notions are
not interpreted as ‘more’ and ‘less important’, respectively. See, however, also Foley and
Van Valin (1984: 370 f.) for some criticism of Hopper and Thompson’s views. [See also ch. 7
in this volume, n. 28.]
 74 chapter 3

macro-level, i.e. the level of large scale narrative units. It can be argued, in fact,
that, just as the imperfect creates a temporal framework on the micro-level
by establishing temporal cohesion between the various events that make up
a given narrative unit, it also creates a narrative framework on the macro-level,
by establishing discourse cohesion between parts of the narrative that may be
at a considerable distance from each other.28
To illustrate this latter cohesion function I shall discuss a significant part of
Herodotus’ narrative, viz. the passage that covers book 4 and the beginning of
book 5, roughly speaking the Scythian, Libyan and Thracian campaigns of Dar-
ius and his generals.

3 The Discourse Function of the Imperfect in Herodotus

The fourth book opens with a reference to the—second!—capture of Babylon

(cf. note 13); thereby the following events are located approximatively on the
time axis. Observe that αὐτοῦ implies that also campaigns by other generals
will be reported. The first finite verb is ἐγένετο, which in this particular context
has a so-called complexive function: it summarizes beforehand an important
part (viz. up to 5.11) of the narrative that is to follow, presenting it as a well-
defined, single whole.29 In the next passage some background information is
given-about Darius’ action. The narrative proper gets going with συνήγειρε ἐπ᾽
αὐτοὺς στράτευμα; it may be called the first stage of the narrative (cf. for this
term Longacre and Levinsohn 1977: 104). συνήγειρε, which can formally also be
an aor. indic. must be viewed as an imperfect, for reasons that will become clear
251 presently.30 The imperfect συνήγειρε |

28 Cf. Longacre and Levinsohn’s remark (1978: 107): ‘Significant chunks of the whole verb sys-
tem of a language can be classified in terms of functions in discourse and paragraph’. For a
succinct but insightful discussion of the notion ‘cohesion’ compare Brown and Yule (1983:
191–199).
29 Possibly all of the narrative up to 5.23 is summarized, depending on whether Megabazus’
activities are or are not considered part of the Scythian campaign. Observe that the first
vertical arrow is interrupted; this is because the sentence with ἐγένετο ἔλασις as such does
not point forward to 4.4. In fact, in other contexts, specifically when no narrative fol-
lows, the whole report could have consisted of just this sentence. There is, thus, only an
implicit link between ἐγένετο ἔλασις and the actual beginning of the expedition with συν-
ήγειρε.
30 A simultaneous imperfect συνήγειρε οn the micro-level occurs in: ὁ δὲ Ἀρκεσίλεως … συν-
ήγειρε πάντα ἄνδρα … συλλεγομένου δὲ στρατοῦ πολλοῦ ἐστάλη ἐς Δελφοὺς ὁ Ἀρκεσίλεως …
(‘Arcesilaus gathered all the men that he could, … and while a great army thus mustered,
he made a journey to Delphi’, Hdt. 4.163.1).
the discourse function of the imperfect 75

The structure of Herodotus, book 4 and the beginning of book 5

4.1 μετὰ δὲ τήν Βαβυλῶνος αἵρεσιν ἐγένετο ἐπὶ Σκύθας αὐτοῦ Δαρείου ἔλασις
4.1–4.4 background of Darius’ decision to campaign against the Scythians
4.4 τῶν δὲ εἵνεκα ὁ Δαρεῖος τείσασθαι βουλόμενος συνήγειρε ἐπ᾽ αὐτοὺς στράτευμα
4.5–4.82 logos on the Scythians, concluded by the transitional formula τοῦτο μέν νυν
τοιοῦτόν ἐστι, ἀναβήσομαι δὲ ἐς τὸν κατ᾽ ἀρχὰς ἤια λέξων λόγον
4.83 παρασκευαζομένου Δαρείου ἐπὶ τοὺς Σκύθας καὶ περιπέμποντος ἀγγέλους … ἐχρήιζε. (-)
ἐπειδή οἱ τὰ πάντα παρεσκεύαστο, ἐξήλαυνε
4.83–4.143 Scythian campaign
4.143.1 Δαρεῖος δὲ διὰ τῆς Θρηίκης πορευόμενος ἀπίκετο ἐς Σηστὸν τὴν Χερσονήσου· ἐνθεῦτεν δὲ
αὐτὸς μὲν διέβη τῇσι νηυσὶ ἐς τὴν Ἀσίην, λείπει δὲ στρατηγὸν ἐν τῇ Εὐρώπῃ Μεγάβαζον …
4.143–4.144 anecdote about Megabazus
4.144.3 οὗτος δὴ ὦν τότε ὁ Μεγάβαζος στρατηγὸς λειφθεὶς … τοὺς μὴ μηδίζοντας κατεστρέ-
φετο
4.145.1 οὗτος μέν νυν ταῦτα ἔπρησσε, τὸν αὐτὸν δὲ τοῦτον χρόνον ἐγίνετο ἐπὶ Λιβύην ἄλλος
στρατιῆς μέγας στόλος, διὰ πρόφασιν τὴν ἐγὼ ἐπηγήσομαι προδιηγησάμενος τάδε·
4.145–4.167 background of Libyan expedition
4.167.3 αὕτη μέν νυν αἰτίη πρόσχημα τοῦ λόγου ἐγίνετο, ἐπέμπετο δὲ ἡ στρατιή,
ὡς ἐμοὶ δοκέειν, ἐπὶ Λιβύων καταστροφῇ. Λιβύων γὰρ ἔθνεα πολλὰ …
4.168–199 Libyan logos; closure: ταῦτα μέν νῦν ἐπὶ τοσοῦτον εἰρήσθω.
4.200 οἱ δὲ Φερετίμης τιμωροὶ Πέρσαι ἐπείτε ἐκ τῆς Αἰγήπτου σταλέντες ὑπὸ
Ἀρυάνδεω ἀπίκατο ἐς τὴν Βάρκην, ἐπολιόρκεον …
4.200–205 results of the Libyan expedition
4.205 ἡ μὲν δὴ Φερετίμης τῆς Βάττου τοιαύτη τε καὶ τοσαύτη τιμωρίη ἐγένετο
ἐς Βαρκαίους.
5.1 οἱ δὲ ἐν τῇ Εὐρώπῃ τῶν Περσέων καταλειφθέντες ὑπὸ Δαρείου, τῶν ὁ Μεγάβαζος
ἦρχε, πρώτους μὲν Περινθίους κατεστρέψαντο …
5.2 ὡς δὲ ἐχειρώθη ἡ Πέρινθος, ἤλαυνε τὸν στρατὸν ὁ Μεγάβαζος διὰ τῆς Θρηίκης. (-)
Θρηίκων δὲ ἔθνος …
5.3–5.10 digression about the Thracians; closure: ταῦτα μέν νυν τῆς χώρης ταύ-
της πέρι λέγεται,
5.10 τὰ παραθαλάσσια δ᾽ ὦν αὐτῆς Μεγάβαζος Περσέων κατήκοα ἐποίεε
5.11 Δαρεῖος δὲ ὡς διαβὰς τάχιστα τὸν Ἑλλήσποντον ἀπίκετο ἐς Σάρδις, ἐμνήσθη …
5.11 digression about Histiaeus and Coës; closure: οὗτοι μὲν … ἐτράποντο
5.12 Δαρεῖον δὲ συνήνεικε … ἐπιθυμῆσαι ἐντείλασθαι Μεγαβάζῳ Παίονας ἑλόντα ἀναστάτους
ποιῆσαι ἐκ τῆς Εὐρώπης ἐς τὴν Ἀσίην
5.12–13 background of this decision
5.14 ἐνθαῦτα Δαρεῖος γράφει γράμματα Μεγαβάζῳ, τὸν ἔλιπε ἐν τῇ Θρηίκῃ στρατηγόν, ἐντελ-
λόμενος …
5.15.3 οἱ δὲ Παίονες … παρεδίδοσαν σφέας αὐτοὺς τοῖσι Πέρσῃσι. οὕτω δὴ Παιόνων Σιριο-
παίονές τε καὶ … ἐξαναστάντες ἤγοντο ἐς τὴν Ἀσίην.
5.16 part of the Paeonians not subjugated
5.17 Παιόνων μὲν δὴ οἱ χειρωθέντες ἤγοντο ἐς τὴν Ἀσίην, Μεγάβαζος δὲ ὡς ἐχειρώσατο
τοὺς Παίονας, πέμπει ἀγγέλους ἐς Μακεδονίην
5.17–5.22 vicissitudes of the envoys in Macedonia
5.23.1 Μεγάβαζος δὲ ἄγων τοὺς Παίονας ἀπίκετο ἐπὶ τὸν Ἑλλήσποντον, ἐνθεῦτεν δὲ διαπε-
ραιωθεὶς ἀπίκετο ἐς τὰς Σάρδις. | 252
 76 chapter 3

creates a framework for other actions, it makes us expect that some further
information will be given concerning states of affairs that are connected with
the συναγείρειν. This expectation, however, is initially not fulfilled, because
Herodotus interrupts his narrative for the long excursus about the habits of the
Scythians and the topography of their country. The information provided by the
chapters 5–82 is, of course, not temporally related to συνήγειρε, i.e. it does not
fall with the framework created by συνήγειρε, but forms what Genette (1972: 128)
calls a ‘descriptive pause’. Thereby this imperfect acquires a cataphoric func-
tion: it indicates that the author will come back to the συναγείρειν at some other
point in the narrative.
In ch. 82 the description of the Scythians comes to an end; the closure of
this topic and the transition to the former one is marked by the transitional
particle combination μέν νυν. The present participles παρασκευαζομένου and
περιπέμποντος refer back, in a somewhat expanded form, to the συνήγειρε that
we lost track of in ch. 4. It is within the framework created by these states
of affairs that Artabanus’ warning, unheeded by Darius, occurs. The comple-
tion of the preparations, expressed in a pluperfect subordinate clause (ἐπειδή οἱ
πάντα παρεσκεύαστο), is followed by the second stage of the narrative, the actual
launching of the campaign, expressed by ἐξήλαυνε. The report of this campaign
runs, in the first resort, up to ch. 143, διέβη ἐς τὴν Ἀσίην, but comes to a close in
ch. 5.11, when Darius arrives in Sardes. Meanwhile Herodotus has introduced,
with the historic present λείπει (4.143.1), an important new topic, viz. the activ-
ities of Darius’ general Megabazus. After a short interruption the first stage of
Megabazus’ actions is, again, expressed by the imperfect, κατεστρέφετο. Just like
συνήγειρε, this imperfect makes us expect other actions to follow, located within
the framework provided by κατεστρέφετο. Again, however, this expectation is
thwarted, since Herodotus indicates, by the transitional formula οὗτος μέν νυν
ταῦτα ἔπρησσε, that he is embarking on a new topic, viz. the expedition to Libya.
To be sure, this expedition is simultaneous with Megabazus’ doings, which is
indicated explicitly by τὸν αὐτὸν τοῦτον χρόνον; properly speaking, however, it
does not fall within the framework created by κατεστρέφετο since it is totally
unconnected with the earlier expedition. In other words, the new expedition
does not fulfil the expectations raised by κατεστρέφετο, since, for the time being,
we get no further information concerning this expedition. It should be noted,
in this connection, that an imperfect is not a necessary condition for another
state of affairs to occur simultaneously.31 Herodotus could also have used κατε-
253 στρέψατο/ἔπρηξε, but in that case the story would have | been different, for the

31 Compare e.g. Hdt. 6.27.2, where an aorist ἀπήνεικε is followed by τὸν αὐτὸν τοῦτον χρόνον …
the discourse function of the imperfect 77

report on Megabazus would have come to an end in 4.144. As it is, the imperfect
is, again, cataphoric: the narrative about Megabazus will be continued later.
In the instances discussed above, the imperfect essentially has a discourse
function: the narrative is interrupted, and the imperfect carries it over, as it
were, to another part of the text. Or, to put it differently, the imperfect enables
the author to dissociate the narrative moment from the action moment of the
states of affairs that are evoked by the imperfect. Thus, the narrative moment
of ἐχρήιζε (4.82) is, in the text, at a considerable distance from συνήγειρε; the
action moment in the historical reality, as it is described by Herodotus, is, of
course, not separated from that of συνήγειρε.32 This dissociation allowed the
author to insert, at a given point in the narrative, some piece of information
that he considered relevant at that juncture, be it information that falls out-
side the story proper—like the description of the Scythians—or information
concerning parallel actions, like the account of the Libyan expedition.
This account is structured in the same way. The first stage is expressed by ἐγί-
νετο, which, once again, however, signals that the narrative will be continued
elsewhere, since it is interrupted for some background information concerning
the expedition. The continuation of the narrative follows at 4.167.3, ἐπέμπετο,
but again Herodotus breaks off his story, to insert a description of the Libyans
and their country. After a transitional formula (cp. μέν νυν) he returns, in 4.200,
to the account of the expedition proper: ἐπέμπετο is brought to an end by ἐπείτε
… ἀπίκατο, which is followed, in turn, by a new state of affairs: ἐπολιόρκεον. This
imperfect provides the framework for some further acts of war. The whole story
is summarized in the transitional formula in 4.205, where ἐγένετο indicates that
the account of the Libyan expedition is closed.
In book 5, ch. 1 Herodotus returns to Megabazus. Both items of the infor-
mation in 4.144.3 on Megabazus are present here as well: στρατηγὸς λειφθείς
is taken up by oἱ … ἦρχε, and κατεστρέφετο is now further elaborated into a
number of specific subjections, that of the Perinthians being the first stage; it
is presented as closed (cf. the aorist κατεστρέψατο). Note that πρώτους points to
other subjections. Both ἤλαυνε and κατήκοα ἐποίεε continue κατεστρέφετο.
In 5.11 the narrative returns to Darius, whom we had left in 4.143.1. | διαβάς 254
refers back to διέβη in 4.143.1, and Darius’ campaign is closed by ἀπίκετο. For
the remainder of the narrative it is important that Herodotus in 5.14 brings
about a change in Megabazus’ actions, expressed by the historic present γρά-

32 Using Genette’s terms (1972: 72): the ‘histoire’ goes on, while the ‘récit’ is interrupted. In
a sense the imperfect always has a cataphoric function. On the micro-level, however, it
has no discourse effects, since the states of affairs evoked by the imperfect are usually
described immediately, and narrative moment and action moment coincide.
78 chapter 3

φει. Megabazus stops the κατήκοα ποιέειν and turns to the execution of Darius’
orders. This new action is told in three different stages, by the two imperfects
ἤγοντο and the participle ἄγων, and brought to an end by the aorist ἀπίκετο. At
this point the stories about Darius and Megabazus merge, and the account of
Darius’ actions in Europe, which started in 4.1, has come to an end.

4 Some conclusions

The following conclusions may be drawn from the above discussion. First, the
imperfect could be shown to occupy a crucial position, both in terms of the verb
system, especially when confronted with the aorist indicative, and in terms of
the actual uses of the tenses. Unlike the aor. indic. it unequivocally locates a
state of affairs in the past, and it has an important function on the discourse
level: it is crucially connected with historical narrative. As to the latter, two
main features may be distinguished. On the level of small-scale narrative units
it serves as the time anchor for other states of affairs; on the level of large-scale
narrative units it establishes cohesion between different and, more specifically,
distant parts of a given narrative, if, for some reason or other, this is split up.
This latter function is, I think, insufficiently acknowledged in the literature
on tense and aspect in Greek. Furthermore, the functioning of the imperfect,
and, for that matter, of the other tenses, is conspicuously absent from literary
studies about, e.g., the structure of Herodotus’ Histories. A glaring example is
provided by the recent book by Lang (1984). It can be shown, last but not least,
that, contrary to what is often stated, Herodotus has taken considerable care in
tying together the parts of his, at times rather discontinuous, narrative.

Bibliography

Barentsen, A.A. ‘Tijd’, ‘Aspect’ en de conjunctie ‘poka’ (Amsterdam 1985).

Basset, L., Les emplois périphrastiques du verbe grec μέλλειν (Lyon 1979).
Bertinetto, P.M., ‘Intrinsic and Extrinsic Temporal References: On Restricting the
Notion of “Reference Time”’, in Lo Cascio & Vet (1986) 41–79.
Brown, G., Yule, G., Discourse Analysis (Cambridge 1983).
Lo Cascio, V., Vet, C. (eds), Temporal Structure in Sentence and Discourse (Dordrecht
1986) 133–161.
Comrie, B., Aspect (Cambridge 1976).
Comrie, B., Tense (Cambridge 1985).
Dahl, Ö., Tense and Aspect Systems (Oxford 1985)
the discourse function of the imperfect 79

Duhoux, Y., ‘Le système verbal grec: l’état mycénien’, in A. Rijksbaron, H.A. Mulder,
G.C. Wakker (eds), In the Footsteps of Raphael Kühner (Amsterdam 1988) 121–134.
Foley, W.A., Van Valin jr., R.D., Functional Syntax and Universal Grammar (Cambridge
1984).
Genette, G., Figures III (Paris 1972).
Givón, T., Syntax: A Functional-Typological Introduction, vol. 1 (Amsterdam/Philadel-
phia 1984).
Hettrich, H., Kontext und Aspekt in der altgriechischen Prosa Herodots (Göttingen 1976).
Hirt, H., Indogermanische Grammatik 4 (Heidelberg 1928).
Hopper, P.J., Thompson, S.A., ‘Transitivity in Grammar and Discourse’, Language 56
(1980) 251–300.
Kuryłowicz, J., The Inflectional Categories of Indo-European (Heidelberg 1964).
Levinson, S.C., Pragmatics (Cambridge 1983).
Lyons, J., Semantics (Cambridge 1977).
Longacre, R., Levinsohn, S., ‘Field Analysis of Discourse’, in W.U. Dressler (ed.), Current
Trends in Textlinguistics (Berlin 1978) 103–122.
Péristérakis, A.G. Essai sur l’aoriste intemporel en grec (Athens 1962).
Quirk, R. et al., A Grammar of Contemporary English, 2nd ed. (London 1973).
Rijksbaron, A. The Syntax and Semantics of the Verb in Classical Greek: An Introduction.
(Amsterdam 1984)
Rix, H., Historische Grammatik des Griechischen (Darmstadt 1976).
Ruijgh, C.J. Autour de “τε épique”: Études sur la syntaxe grecque (Amsterdam 1971).
Ruijgh, C.J., Review of Basset (1979), Lingua 65 (1985) 323–333.
Strasburger, H., ‘Herodots Zeitrechnung’, in W. Marg (ed.), Herodot: Eine Auswahl aus
der Neueren Forschung (Darmstadt 1965) 688–737.
Vet, C., Molendijk, A., ‘The Discourse Function of the Past Tenses of French’, in Lo Cas-
cio & Vet (1986) 133–161.
Wackernagel, J., Vorlesungen über Syntax, Erste Reihe, 2nd ed. (Basel 1926).
 chapter 4

Sur les emplois de λέγε et εἰπέ chez Platon

Résumé

Une analyse de λέγε et εἰπέ dans le Gorgias, le Philèbe et le Ménexène, trois textes de
nature assez différente, mène à la conclusion suivante.
Dans le Gorgias et le Philèbe, λέγε s’emploie de deux façons différentes, qui sont
liées au caractère du texte où λέγε se trouve et au rôle joué par celui qui parle; εἰπέ par
contre s’emploie de la même façon dans les deux dialogues.
Dans le Gorgias, un dialogue élenctique, λέγε est surtout dit par Socrate, qui l’em-
ploie pour obtenir de son interlocuteur ou bien une réponse à une seule question
spécifique, avec une valeur ‘immédiative’ (‘tu n’as qu’à le dire’, ‘vas-y’), ou bien des
réponses à une série de questions, avec une valeur itérative-protocolaire. Εἰπέ invite
toujours l’interlocuteur à établir un point spécifique; il marque donc une démarche
importante dans l’interrogation (valeur d’accomplissement).
Dans le Philèbe, en revanche, un dialogue qui pour la plus grande part est un rai-
sonnement plutôt qu’une réfutation, λέγε (27 exemples) est surtout employé par les
interlocuteurs de Socrate, afin de lui permettre de poursuivre son raisonnement, ou
bien avec une valeur ‘immédiative’ (‘tu n’as qu’à parler’) ou bien avec une valeur conti-
nuative, ou encore avec une combinaison de ces valeurs (‘tu n’as qu’à continuer’). Il
n’ y a qu’un seul cas de εἰπέ, qui est employé de la même façon que εἰπέ dans le Gorgias.
Dans le Ménexène, un texte qui n’est pas un dialogue, l’emploi de εἰπέ diffère net-
tement de son emploi dans un texte dialogué: εἰπέ a le sens idiomatique de ‘prononce
une oraison funèbre’. Bien que λέγε y manque, il y a bien une opposition entre εἰπέ
et les formes de λεγ-: εἰπέ dénote une oraison quelconque in abstracto, λεγ- dénote
l’exécution, la réalisation d’une oraison concrète.

Abstract

An analysis of λέγε and εἰπέ in Gorgias, Philebus and Menexenus, three texts of a quite
different nature, leads to the following conclusion.
In the Gorgias and Philebus, λέγε is used in two different ways, which are linked with
the nature of these texts and with the role played by the person speaking; εἰπέ, however,
is used in the same way in both dialogues.
In the Gorgias, an elenctic dialogue, λέγε is used almost exclusively by Socrates, who
uses it to obtain from his interlocutor either an answer to a single, specific question,

© koninklijke brill nv, leiden, 2019 | doi:10.1163/9789004386129_006

sur les emplois de λεγε et ειπε chez platon 81

with an ‘immediative’ value (‘fire away’), or a series of answers to a series of questions,

with an iterative value. Εἰπέ always invites the interlocutor to establish a specific point;
thereby, it serves to mark an important step in the discussion (accomplishment-value). | 152
In the Philebus, on the other hand, a dialogue that for the greater part is an argu-
mentative rather than an elenctic text, λέγε (27 instances) is mostly used by Socrates’
interlocutors, so as to enable him to proceed with his argument, either with an ‘imme-
diative’ (‘fire away’) or with a continuative value, or with a combination of these values
(‘just go on’). There is only one instance of εἰπέ, which is used in the same way as in the
Gorgias.
Finally, in the Menexenus, which is not a dialogue, the use of εἰπε differs conspicu-
ously from that in a dialogic text: εἰπέ has the idiomatic meaning of ‘deliver a speech
at a funeral’. Although λέγε does not occur, there is an opposition between εἰπέ and the
forms of λεγ-: εἰπέ denotes a speech in abstracto, λεγ- denotes the production of an
actual speech.

1 Introduction

Avant que je n’aborde le sujet propre de mon exposé, il n’est pas superflu peut-
être d’indiquer en quelques grandes lignes quelle est ma position dans le débat
quasi-permanent sur les valeurs sémantiques des thèmes verbaux en grec, ou
du moins des thèmes de PR et d’AO, ne serait-ce que pour éviter le plus possible
d’attribuer à λέγε et à εἰπέ des valeurs ad hoc.*
La valeur sémantique fondamentale du thème de PR est qu’ il dénote une
action qui peut être interrompue. Ceci implique que l’ accent est mis sur la
réalisation partielle de l’action, sur son non-accomplissement, et donc sur
le processus de sa réalisation per se : l’action est en cours de réalisation1. Il
y a notamment deux phénomènes qui m’amènent à considérer cette valeur
comme fondamentale:
(i) Si les verbes παύεσθαι et ἀπαγορεύειν/ἀπειπεῖν sont construits avec un par-
ticipe, c’est toujours un participe PR.

* Je tiens à remercier Gerard Boter, Jean Lallot, Frédéric Lambert, Odile Mortier-Waldschmidt,
Kees Ruijgh, Gerry Wakker et les membres du ‘Amsterdamse Hellenistenclub’ pour leurs
remarques et suggestions.
1 L’ importance de la notion d’ ‘interruptibilité’ pour analyser l’aspect verbal est reconnue expli-
citement par Thelin, dans son introduction à Thelin (1990), p. ex. aux pp. 31 et 41. La notion
de ‘réalisation partielle’ est probablement déjà présente dans Apollonius Dyscole, cf. ses
remarques à propos de ἔγραφον, Synt 3.21 (= p. 286 Uhlig, p. 205 Bekker), le terme utilisé étant
ἀπὸ μέρους γεγονότα.
 82 chapter 4

(ii) Dans son emploi temporel l’adverbe μεταξύ ‘en plein milieu de’ est tou-
jours accompagné d’un participe PR ou d’ une autre forme du thème de
153 PR2. |
De cette valeur générale découlent plusieurs valeurs sémantiques dérivées, qui
sont co-déterminées par l’Aktionsart du lexème verbal et par le contexte. Dans
les emplois conatifs, itératifs et duratifs3, c’est la valeur du ‘non-accomplisse-
ment’ qui prédomine. Dans l’emploi dit ‘inceptif’ ou ‘inchoatif’, c’ est la valeur
d’‘action en cours de réalisation’ qui est dominante: l’ action est présentée
comme en voie d’être réalisée, et plus spécifiquement, dès qu’ une autre action
a été accomplie4. À ceci s’ajoute que tous les emplois du PR ont ceci en com-
mun qu’ils peuvent créer, en tant qu’action ‘en cours’, un cadre pour d’ autres
actions; celles-ci sont dans ce cas-là simultanées avec la première.

2 Voir p. ex. Ap. 40b2 : καίτοι ἐν ἄλλοις λόγοις πολλαχοῦ δή με ἐπέσχε λέγονταPR μεταξύ ‘Bien sou-
vent pourtant, en d’ autres circonstances, elle m’a fait taire, au beau milieu de mon proposPR’,
Ly. 207b1 : ἔπειτα Μενέξενος ἐκ τῆς αὐλῆς μεταξὺ παίζωνPR εἰσἐρχεται ‘Puis Ménexène entra en
quittant la cour au beau milieu de son jeuPR’ [trad. AR].
3 Conatif : πείθειν, ἀναγκάζειν, μανθάνειν, πυνθάνεσθαι… ; itératif: κόπτειν, ἐρωτᾶν, βοᾶν…; duratif:
γελᾶν, βασιλεύειν, περιπατεῖν… ; dans certains contextes, l’emploi duratif s’est spécialisé dans
un emploi (dis)continuatif, cf. ἄρχε en S. Ai. 1107 : ὧνπερ ἄρχεις ἄρχεPR ‘CommandePR à ceux
qui sont sous ton commandement’ (= ‘Continue à commander’) [trad. Mazon], μὴ πατέριζε
en Ar. V. 652 ; ΒΔ. ἀτάρ, ὦ πάτερ ἡμέτερε, Κρονίδη – ΦΙ. παῦσαι καί μή πατέριζεPR – (Bdélycléon)
‘Cependant, ô notre père, fils de Cronos…’ – (Philocléon) ‘Assez; ne disPR plus “notre père”’
[trad. Van Daele, modifiée]. À noter que les verbes par excellence pour exprimer l’itérativité,
φιλεῖν et εἰωθέναι, sont toujours suivis d’ un infin. PR.
4 P. ex. X. An. 4.8.18 : οἱ δέ… πελτασταί… ἀνακραγόντες ἔθεονPR, où l’on pourrait traduire par ‘Aus-
sitôt qu’ ils avaient poussé un cri ils accouraientPR’ (= ‘ils étaient en train d’accourir’), Th. 8.1.4:
καὶ ὡς ἔδοξεν αὐτοῖς καὶ ἐποίουνPR ταῦτα ‘Et ces résolutions qu’ils arrêtèrent entraientPR effec-
tivement en vigueur’ [trad. Weil]. Souvent cette valeur est rendue explicite par la présence de
εὐθύς ou εὐθέως, cf. Th. 2.93.4: ὡς δὲ ἔδοξεν αύτοῖς καὶ ἐχώρουνPR εὐθύς (litt.) ‘Sitôt la chose déci-
dée, ils étaientPR en route’, ou par le moyen d’ une expression comme μὴ διάτριβε, cf. R. 472b2:
ἀλλὰ λέγεPR καὶ μὴ διάτριβεPR ‘Parle donc, sans nous remettrePR davantage’ [trad. Chambry].
Les termes ‘inceptif’ et ‘inchoatif’ (pour le dernier cf. par exemple Classen-Steup ad Th.
1.26.5 ἐπολιόρκουν τὴν πόλιν (‘obsidere coeperunt’, avec une douzaine d’autres exemples de
cet ‘inkohatives Impf.’), L. Hillesum (1908: 26ss., 77ss.), Poutsma (1928: §5), pour le premier
Ruijgh (1985: 1ss. = 1996: 535ss.)) sont un peu incommodes, parce qu’il s’agit plutôt de la réali-
sation immédiate que du début de la réalisation. Pour l’ imparfait, Moorhouse (1982: 191) se sert
du terme plus exact ‘imperfect of consecutive action’. Voir aussi Bakker (1966: 50: l’impératif
PR est utilisé quand la situation ‘requires immediate action’; le cadre théorique de Bakker dif-
fère toutefois du mien), Rijksbaron (1994: 17s., 43ss.). Dans ce qui suit je parlerai de la valeur
‘immédiative’. On trouve la même valeur dans l’ emploi de μέλλω avec infinitif PR et celui de
ὥρα + inf. PR pour exprimer une action imminente. Comparer aussi l’emploi de οὐκ ἂν φθάνοις,
toujours + participe PR, pour inciter quelqu’ un à réaliser l’action du participe sur-le-champ,
p. ex. οὐκ ἂν φθάνοις λέγων Pl. Smp. 185e4 (‘Rede sofort’, K-G 2, 65, Anm. 12).
sur les emplois de λεγε et ειπε chez platon 83

De l’impossiblité d’employer les formes de l’ ΑΟ tant avec μεταξύ qu’ avec

παύεσθαι et ἀπαγορεύειν/ἀπειπεῖν, en revanche, on peut conclure que l’ ΑΟ
dénote essentiellement une action qui ne peut pas être interrompue; ceci
implique que l’accent est mis sur l’accomplissement | de l’ action5 et non pas 154
sur le processus de réalisation. L’AO présente donc l’ action comme un tout
indivisible. Les emplois de l’AO s’opposent à ceux du PR comme suit : l’ AO
peut exprimer la réussite (<-> conativité), la semelfactivité (<-> itérativité),
l’ accomplissement pur et simple (<-> durativité)6. Tous ces emplois ont en
commun que l’ ΑΟ ne peut jamais servir de cadre pour une autre action ; par
là, l’AO exprime souvent l’antériorité.
Dans son emploi ‘dynamique’ (non référentiel), une forme de l’ ΑΟ (notam-
ment l’infinitif dynamique et l’impératif) est souvent suivie d’ une forme du
PR du même verbe. Dans ces cas-là l’ ΑΟ dénote la réalisation tout court d’ une
certaine action, in abstracto, pour ainsi dire, tandis que le PR met l’ accent sur
sa réalisation concrète. Deux exemples nets:

(1) ΛΥ. ἀλλ᾽ εἴ τι καὶ σύ, ὦ παῖ Σωφρονίσκου, ἔχεις τῷδε τῷ σαυτοῦ δημότῃ ἀγαθὸν
συμβουλεῦσαιAO, χρή συμβουλεύεινPR (La. 180d6–8)

Lysimaque: Mais si tu as quelque bon conseil à me donnerAO, à moi qui

suis de ton dème, ô fils de Sophronisque, tu dois me le donnerPR (trad.
Croiset)

5 J’ emploie ‘accomplissement’ tant pour l’ accomplissement au sens strict, p. ex. dans le cas
d’ une action télique (οἰκοδομῆσαι, παιδεῦσαι), que, un peu καταχρηστικῶς, pour l’arrivée
à terme (un terme arbitraire, non inhérent) d’ une action atélique, soit statique (γελάσαι,
βασιλεῦσαι, περιπατῆσαι) soit momentanée (λαβεῖν, εὑρεῖν, ἀφικέσθαι). Voir aussi la note sui-
vante.
6 Réussite : πεῖσαι, ἀναγκάσαι, μαθεῖν, πυθέσθαι… ; semelfactivité: κόψαι, ἐρέσθαι, βοῆσαι…;
accomplissement pur et simple : γελάσαι, βασιλεῦσαι, περιπατῆσαι…
L’ aoriste des verbes atéliques statiques a deux valeurs différentes, qui dépendent du
contexte. Ou bien l’ ΑΟ dénote un état qui s’ est terminé, donc un état qui est présenté comme
un tout, et qui, le cas échéant, est antérieur à une autre action (cf. Hdt. 2.147.2: μετὰ τὸν ἱερέα…
βασιλεύσανταAO ἐστήσαντο δυώδεκα βασιλέας ‘… après le règne du prêtre d’Héphaistos, les
Égyptiens créèrent douze rois…’ [trad. Legrand]), ou bien il dénote la réalisation de l’action
qui constitue la phase initiale de l’ état (emploi ingressif). Dans le dernier cas il s’agit de ce
que l’ on pourrait appeler ‘accomplissement’ ou ‘antériorité interne’: ainsi, le point de réfé-
rence pour βασιλεύσας, dans Hdt. 2.2.1 (ἐπειδὴ δὲ Ψαμμήτιχος βασιλεύσαςAO ἠθέλησε εἰδέναι…
‘Mais, depuis que Psammétique, devenuAO roi, voulut savoir…’ [trad. Legrand]), n’est pas une
action externe, mais le ‘ἐβασίλευε’ impliqué par l’ aoriste.
 84 chapter 4

Pour faire ressortir les différences on pourrait paraphraser par : ‘Si tu peux
réaliser x, passe à la réalisation’ (emploi ‘immédiatif’). En d’ autres termes,
155 l’ infinitif PR invite Socrate à suivre le chemin qui a été frayé par l’ infinitif AO. |

(2) a. καί μοι… μὴ θορυβήσητεAO, μηδ᾽ ἐὰν δόξω τι ὑμῖν μέγα λέγειν (Ap. 20e3–5)

Maintenant, n’allez pas faire de bruitAO, Athéniens, si je vous parais

présomptueux (trad. Croiset, modifiée)

Un peu plus bas Socrate répète sa demande, mais cette fois-ci il emploie
l’ impératif PR:

(2) b. καί, ὅπερ λέγω, μὴ θορυβεῖτεPR, ὦ ἄνδρες… (Ap. 21a5)

et comme je disais, ne faites pas de bruitPR, juges… (trad. AR)

Noter dans (2a) la présence de la conditionnelle, qui précise le cas où l’ ordre/

demande de Socrate est pertinent, sans qu’il ait dit encore quelque chose de μέγα,
tandis qu’en 21a5 les mots cités sont suivis immédiatement par le ‘gros mot’, à
savoir la question de Chéréphon à Delphes, demandant s’ il y a quelqu’ un qui
est plus σοφός que Socrate. La valeur de μὴ θορυβεῖτε est donc ‘ne faites pas de
bruit dès que vous aurez entendu ce que je vais dire’7. Pour éviter tout malen-

7 La traduction de Croiset (‘de grâce, juges, ne vous récriez pas en l’entendant’) me semble assez
inexacte. Noter que la dénotation de μὴ θορυβεῖτε ne diffère pas de celle de μὴ θορυβήσητε;
sinon, Socrate n’aurait pas pu ajouter à l’ impératif PR ὅπερ λέγω pour renvoyer au subjonc-
tif AO. La différence est donc une différence d’ ordre pragmatique: l’impératif PR s’emploie
dans une autre situation d’ énonciation que l’ impératif AO. On trouve un cas similaire en Plt.
257c8–10 : ΞΕ. διαναπαύσωμεν αὐτὸν μεταλαβόντεςAO αὐτοῦ τὸν συγγυμναστὴν τόνδε Σωκράτη ;
ἤ πῶς συμβουλεύεις; ΘΕΟ. καθάπερ εἶπες, μεταλάμβανεPR·… (‘L’Étranger: Lui donnerons-nous
quelque répit, en le remplaçantAO par son compagnon d’exercices, le Socrate que voici? Ou
bien que conseilles-tu ?’ – Théodore: Fais ce que tu proposes, remplacePR-le’; trad. Diès). Par
l’ emploi de l’ impératif PR, Théodore insiste sur l’ exécution immédiate de l’action. À noter
également que dans les trois cas cités les formes du PR sont employées absolument; ainsi
l’ accent est mis sur l’ action en soi, les conditions ou l’ objet de l’action étant spécifiés avec la
forme de l’ ΑΟ qui précède. Cf. aussi R. 608d : σοῦ δ᾽ ἂν ἡδέως ἀκούσαιμιΑΟ τὸ οὐ χαλεπόν τοῦτο. –
ΆκούοιςPR ἄν, ἦν δ᾽ ἐγώ. – ΑέγεPR μόνον. (‘(Glaucon)… mais j’aurais plaisir à t’ entendreAO faire
cette démonstration facile. – (Socrate) ÉcoutePR, dis-je. – (Glaucon) Tu n’as qu’à parlerPR’;
trad. Chambry). (Voir aussi Basset (2000 : 237, ex. (1)), et Oréal (2000: 295–300), avec une
explication différente.) La même chose vaut pour les séquences ἀνάγνωθι… ἀναγίγνωσκε et
ἀνάγνωθι… λέγε chez les orateurs, cf. les exemples présentés par Amigues (1977: 233ss.). Bien
qu’ il opère dans un cadre théorique différent, Sicking (1991: 156ss.) analyse de telles séquences
sur les emplois de λεγε et ειπε chez platon 85

tendu, je souligne que la valeur ‘abstraite’ de θορυβήσητε en (2a) n’est pas une
valeur inhérente de l’AO mais un effet contextuel. Si la forme de l’ AO | n’est 156
pas suivie d’une forme de PR, un impératif à l’AO dénote une action qui doit
être réalisée sans plus8.
C’est dans le cadre esquissé ci-dessus que je voudrais analyser l’ emploi de
λέγε et εἰπέ chez Platon.

2 λέγε et εἰπέ chez Platon

Tout d’abord, il est intéressant de noter que la répartition de ces deux impéra-
tifs n’est pas du tout homogène, comme il ressort de la table suivante (où les
dialogues apparaissent dans leur ordre traditionnel) : | 157

λέγε εἰπέ

Euthphr. 4 7
Ap. 2 6
Cri. 3 2
Phd. 7 1
Cra. 7 5
Tht. 14 5
Sph. 16 4
Plt. 8 1
Prm. – 3
Phlb. 27 1
Smp. 4 6
Phdr. 6 9
Alc. 1 6 5
Alc. 2 – –
Chrm. 6 6
La. 3 2
Ly. 2 5
Euthd. – 9
Prt. 1 7

d’ impératifs AO et PR d’ une façon comparable. À noter qu’on ne trouve jamais, chez les ora-
teurs, εἰπέ avec le sens de ‘lis’, cf. Fournier (1946 : 67, 79), Amigues (1977: 232).
8 Le ‘plus’ du PR étant qu’ il ajoute une nuance d’ urgence.
 86 chapter 4

(cont.)

λέγε εἰπέ

Grg. 9 21
Men. 4 16
Hp.Ma. 6 6
Hp.Mi. 1 4
Ion – 3
Mx. – 3
R. 27 17
Ti. 2 –
Criti. – –
Lg. 29 3
Totaux 194 167

Tandis qu’il y a un certain nombre de dialogues où les deux formes sont repré-
sentées plus ou moins en proportion égale, il y en a d’ autres où les différences
158 sont très nettes, voire spectaculaires, tels le Philèbe, | l’Euthydème, le Gorgias, le
Ménexène et les Lois9. Ma démarche sera d’analyser d’ abord, de manière assez
circonstanciée, le Gorgias, ensuite de dire quelques mots sur le Philèbe, pour
finir avec le Ménexène.

2.1 Le Gorgias
Je commence par une des discussions du Gorgias, à savoir le passage où Socrate
et Polos sont les principaux participants (461b–481b). Il s’ agit ici d’ une discus-
sion formelle, où λέγε et εἰπέ alternent avec des verbes ‘de discussion’, notam-
ment ἐλέγχειν/ἐλέγξαι et ἀποκρίνεσθαι/ἀποκρίνασθαι ; aussi tiendrai-je égale-
ment compte de ces verbes-ci, en premier lieu des impératifs. J’ en traiterai dans
l’ ordre de leur apparition.

9 Les impératifs du pluriel et du duel, et l’ impér. de la troisième personne, ne sont pas repré-
sentés dans la table. – Noter que dans les dialogues considérés comme tardifs, comme le
Philèbe, λέγε l’ emporte en général nettement sur εἰπέ. On peut faire l’hypothèse que ce fait
est lié au caractère non élenctique de ces dialogues. Voir plus bas, §2.2. et notes 20 et 23. Selon
Dodds, le Gorgias et le Ménexène appartiennent à la même période (environ 387–385); voir
son édition du Gorgias, p. 24.
sur les emplois de λεγε et ειπε chez platon 87

2.1.1 L’organisation de la discussion: la distribution des rôles

(461b3–462b2)
Après que Socrate a conclu qu’un traitement satisfaisant du sujet de l’ emploi
injuste de la rhétorique ‘n’est pas l’affaire d’un moment’ (461b1), Polos inter-
vient, en protestant contre le comportement impertinent de Socrate ; alors
Socrate lui donne formellement la parole par l’ injonction ἐπανόρθου (461d1),
qui a une valeur itérative ‘fais le correcteur’: Polos est invité à corriger tous les
points contestables de la discussion avec Gorgias, que Socrate à cette fin est
prêt à réintroduire10.

462a4 ἐρωτῶν τε καὶ ἐρωτώμενος… ἔλεγχέ τε καὶ ἐλέγχου.

Tout comme ἐπανόρθου ces impératifs ont une valeur itérative: ‘tiens le rôle de
réfuteur et de réfuté’. Noter l’absence d’un complément d’ objet direct, ce qui
favorise probablement une interprétation itérative. L’ on a affaire ici à l’ emploi
que Jean Lallot a nommé ‘protocolaire’ (voir Lallot 2000).
En 462b1, Socrate donne à Polos (qui, tout comme Gorgias lui-même, se
déclare prêt à être interrogé sur tout ce qu’on veut) le choix : ἐρώτα ἢ ἀποκρίνου;
nous sommes encore dans l’itératif-protocolaire: Polos peut ou bien poser les
questions ou bien donner les réponses.
À la différence de Gorgias, Polos choisit le rôle de questionneur, donc le rôle
habituel de Socrate – un mauvais choix, comme il ne tardera | pas à s’ en aper- 159
cevoir, car Socrate est souvent obligé de lui suggérer une question.

2.1.2 Polos (et Gorgias) questionneur, première partie (462b3–467c1)

La discussion proprement dite commence en 462b3. Dans cette discussion,
les impératifs AO marquent les démarches principales de l’ argumentation : ils
concernent toujours des points qui doivent être établis (< valeur d’ accom-
plissement), afin qu’ils puissent servir de base pour d’ autres démarches. La
première démarche est marquée par ἀπόκριναι, qui concerne une question
concrète et spécifique (σὺ αὐτὴν τίνα φῂς εἶναι ; 462b5). La réponse de Socrate
(ἐμπειρίαν ἔγωγέ τινα, 462c3), jugée satisfaisante par Polos (cf. 462c6–d4), est
ensuite invalidée par Socrate lui-même, quand il se livre à une comparaison
avec une autre pseudo-τέχνη (462d8: Ἐροῦ νῦν με, ὀψοποιία ἥτις μοι δοκεῖ τέχνη
εἶναι, suivi de ἐρωτῶ δή, dit par Polos)11. Ἀπόκριναι est alors repris par εἰπέ en

10 Dans εἰ… ἐπανορθώσασθαι αὐτὸν βούλει (462a2), l’ aoriste exprime que Socrate vise ici à la
correction effective des points contestables.
11 Dans ce passage il y a deux cas de φάθι (462d10 et 11). J’avoue que les conditions de
l’ emploi de cet impératif, tout comme l’ emploi de φάναι en général dans la discus-
 88 chapter 4

463a5 (dit par Gorgias), et ce n’est qu’après cet AO que Socrate donne une
réponse plus concrète et élaborée: la rhétorique est une forme de κολακεία
(463a9ss.).
Bien que ce point soit établi, Socrate ne considère nullement l’ affaire comme
réglée (cf. 463c1ss.: οὐ γάρ πω πέπυσται (sc. Polos) ὁποῖόν φημι ἐγὼ τῆς κολα-
κείας μόριον εἶναι τὴν ῥητορικήν) et il ouvre la voie pour une deuxième série
de démarches, qui est préparée dans un premier temps par πυνθανέσθω en
463b8 (= ‘qu’il continue à tenir le rôle de questionneur’), et ensuite par ἐρώτα
en 463c6. Noter qu’ici ἐρώτα est suivi, à la différence de ἐρώτα en 462b1, mais
tout comme ἐροῦ en 462d8, d’une question concrète et spécifique (ὁποῖον μόριον
etc.). Par la présence de ce complément d’objet direct, cet impératif PR com-
bine la valeur protocolaire et celle de l’AO: ‘que tu continues ton interrogation
en me posant la question suivante’ (cf. Wakker 2000 : 223 (‘Protocole II’)). Les
démarches de cette deuxième série sont, elles aussi, marquées par des impé-
ratifs et d’autres formes d’AO: ἀπόκριναι (463c8), εἰπέ (463e3, dit par Gorgias),
(πειράσομαι) φράσαι (463e5), ἐπιδείξω (464b2). À la fin de son exposé, Socrate
signale que pour sa part il considère le sujet en question comme épuisé (ὃ μὲν
160 οὖν ἐγώ φημι | τὴν ῥητορικὴν εἶναι, ἀκήκοας, 465d7, voir aussi συχνὸν λόγον ἀποτέ-
τακα en 465e3, ταύτῃ τῇ ἀποκρίσει en 466a2). Ceci est suivi d’ une remarque de
nature protocolaire: comme il ressort de σοῦ ἀποκρινομένου en 466a1, Socrate
présuppose que dès lors c’est Polos qui va tenir le rôle de ‘répondeur’. Il ajoute
que, tout comme lui-même, Polos a le droit de développer son propos pour éla-
borer un point, chaque fois que sa réponse ne suffit pas à Socrate (ἐὰν μὲν οὖν
καὶ ἐγὼ σοῦ ἀποκρινομένου μὴ ἔχω ὅ τι χρήσωμαι, ἀπότεινε καὶ σὺ λόγον) ; ἀπότεινε
s’ inscrit dans le rôle protocolaire de Polos et a donc une valeur itérative, noter
λόγον ‘un discours’; la même chose vaut pour ἔα et χρῶ un peu plus loin. Mais
Polos ignore la consigne donnée par Socrate et continue à faire le question-
neur; il répète même simplement, en forme mutilée, une question que Socrate
croyait traitée à fond, au grand étonnement de ce dernier (466a7–8 : ἀλλ᾽ οὐ
μνημονεύεις τηλικοῦτος ὤν, ὦ Πῶλε; τί τάχα δράσεις ;). Pendant un certain temps
le rôle de Polos reste ambigu (466c3: Νὴ τὸν κύνα, ἀμφιγνοῶ μέντοι, ὦ Πῶλε, ἐφ᾽
ἑκάστου ὧν λέγεις πότερον αὐτὸς ταῦτα λέγεις καὶ γνώμην σαυτοῦ ἀποφαίνῃ, ἣ ἐμέ
ἐρωτᾷς), mais peu à peu les rôles sont renversés, informellement à partir de
466e9, et formellement en 467c1–4: ΣΩ. ἀλλ᾽ εἰ μὲν ἔχεις ἐμὲ ἐρωτᾶν, ἐπίδειξον ὃτι

sion, m’échappent. Il semble toutefois que souvent le sens soit ‘croire, penser’ (sens non
reconnu, d’ ailleurs, par Des Places dans son Lexique) plutôt que ‘dire’. Comparer un pas-
sage comme 462b6–10, où φάναι alterne avec δοκεῖν (ἆρα ἐρωτᾷς ἣντινα τέχνην φημὶ εἶναι ; –
ἔγωγε. – οὐδεμία ἔμοιγε δοκεῖ… – ἀλλὰ τί σοι δοκεῖ ἡ ῥητορικὴ εἶναι ;); pour ces alternances cf.
Fournier (1946 : 17). Peut-être φάθι = ‘donne ton opinion’?
sur les emplois de λεγε et ειπε chez platon 89

ψεύδομαι, εἰ δὲ μή, αὐτός ἀποκρίνου. ΠΩΛ. ἀλλ᾽ ἐθέλω ἀποκρίνεσθαι, ἵνα καὶ εἰδῶ ὅ
τι λέγεις. Noter l’AO ἐπίδειξον : Socrate demande à Polos de prouver ‘complète-
ment’ qu’il se trompe.

2.1.3 Socrate questionneur, première partie (467c5–470c7)

L’ interrogatoire de Socrate commence par un échange rapide de questions et
de réponses (467c5–468e5). Dans tout ce passage les verbes de discussion font
défaut, mis à part deux cas de la question impatiente τί οὐκ ἀποκρίνῃ; (468c7–
8 et 468d6). La même chose vaut pour le passage suivant (468e6–469e8), qui
n’est pas, en fait, une vraie discussion mais une tentative de la part de Polos
pour pousser Socrate à clarifier son attitude personnelle à propos de ce qu’ il
vient de dire12. Socrate reprend l’interrogatoire proprement dit en 470a1, ce qui
est formellement marqué par la réapparition d’ un verbe de discussion (Ἔχεις
οὖν εἰπεῖν δι᾽ ὅ τι…;), où εἰπεῖν indique que Socrate veut mener au bout le proces-
sus de ‘dire’, c’est-à-dire établir un point de base. Après la réponse affirmative
de Polos à cette question préparatoire, Socrate invite Polos à transformer sa
connaissance abstraite en un acte de parole concret : λέγε : ‘tu n’as qu’ à le dire’,
‘parle’ (valeur ‘immédiative’; cf. le cas cité plus haut du | Lachès). Cette pre- 161
mière démarche est donc réalisée en deux temps13. La deuxième démarche est
marquée par εἰπὲ τίνα ὅρον ὁρίζῃ, mais cette demande est renvoyée par Polos
à Socrate (σὺ μὲν οὖν… ἀπόκριναι τοῦτο). Après la réponse positive de Socrate
(470c1), Polos remarque ironiquement qu’il est difficile de réfuter (ἐλέγξαι)
Socrate: un enfant pourrait le réfuter (ἐλέγξαι). De son côté, Socrate fait remar-
quer qu’il saurait gré à cet enfant et à Polos lui-même, ἐάν με ἐλέγξῃς. Ces trois
AO expriment que Polos et Socrate pensent tous les deux à mener à bien la
réfutation (Socrate seulement en apparence, bien sûr).

2.1.4 Polos questionneur, deuxième partie (470c7–474c3)

De la phrase de Polos ‘un enfant pourrait te réfuter’, Socrate a apparemment
inféré que Polos juge cette réfutation si facile qu’ il est prêt à l’ éluder, puisqu’ il

12 ‘Polus, defeated, falls back on an argumentum ad hominem’, Dodds dans son commentaire
ad loc.
13 Noter que plus bas, 476b1–2, l’ on trouve ἔχεις οὖν λέγειν ὡς…; καὶ διασκεψάμενος εἰπέ. Ici,
l’ inf. PR est peut-être employé parce que la question fait partie d’une interrogation qui
est déjà en cours (‘peux tu maintenir que… ?’ ; cf. aussi καλεῖς en 476a8), tandis que l’inf.
AO est plutôt employé pour établir un point préliminaire. Quant à la nuance apportée
par εἰπέ, je ne vois pas trop bien comment l’ expliquer. Faut-il inférer de la présence de
διασκεψάμενος que Socrate souligne après tout que Polos doit donner une réponse très
précise, et qu’ il s’ agit donc d’ un point de base ? Pour les emplois de εἰπεῖν et λέγειν après
ἔχω – assez difficiles à saisir – je renvoie à l’ article de Basset (2000: 305–316).
 90 chapter 4

ajoute (470c7–8): ἀλλὰ μὴ κάμῃς φίλον ἄνδρα εὐεργετῶν, ἀλλ᾽ ἔλεγχε. L’impératif
PR a la valeur, me semble-t-il, de ‘soumets-moi à une réfutation’ (valeur itéra-
tive-protocolaire, avec peut-être une nuance immédiative). La réaction de
Polos indique qu’il ne pense point à une réfutation de nature itérative: οὐδέν
γέ σε δεῖ παλαιοῖς πράγμασιν ἐλέγχειν. Des événements récents suffiront large-
ment à réfuter Socrate, ἐξελέγξαι. Noter l’apparition de l’ inf. AO: non seule-
ment Polos va passer à la réfutation, mais cette réfutation sera faite une fois
pour toutes. Néanmoins, après leur discussion, Socrate explique par une ques-
tion rhétorique qu’il ne se considère nullement comme réfuté pour de bon
(471d5ss.: … οὗτός ἐστιν ὁ λόγος ᾧ με καὶ ἄν παῖς ἐξελέγξειε, καὶ ἐγὼ ὑπὸ σοῦ νῦν, ὡς
σὺ οἴει, ἐξελήλεγμαι τούτῳ τῷ λόγῳ…;). Noter ici que Socrate renchérit sur l’ AO
de Polos en le remplaçant par un indicatif PFT.
Le caractère de l’ ἔλεγχος continue à jouer un rôle important dans ce qui suit,
voir 471e7, et notamment 472c4–5, où Socrate propose de comparer le τρόπος
ἐλέγχου de Polos avec son τρόπος à lui: παραβαλόντες… σκεψώμεθα εἴ τι διοίσου-
σιν ἀλλήλων. L’AO exprime que Socrate veut en arriver à examiner à fond et
162 ainsi établir les points de différence. (Cf. aussi Vassilaki 2000.) Tandis que | l’ ΑΟ
vise aux résultats, l’examen au sens propre, donc comme procès, ne commence
qu’en 473a4: σκόπει δὲ καὶ σύ.
Le prochain verbe de discussion est διανοώμεθα dans ἄλλο τι ὡς οὕτω σοῦ νομί-
ζοντος διανοώμεθα ; (472d4). Tant σοῦ νομίζοντος, qui reprend ἡγῇ (472d3), que
διανοώμεθα ont une valeur durative: devons-nous considérer que cette opinion
est la tienne?
En 473a3ss., Socrate commence à faire le bilan des différences, et il invite
Polos à donner suite à son invitation préliminaire exprimée par σκεψώμεθα :
σκόπει δὲ καὶ σύ. Son sommaire va dans un premier temps jusqu’ à 473b7. Noter,
dans ce sommaire, la présence de l’aoriste ἐξηλέγχθην ὑπὸ σοῦ (473a10), qui
pourrait faire croire à Polos que Socrate est battu. En effet, Polos tombe les
yeux ouverts dans le piège tendu par Socrate, puisqu’ il répond: ναὶ μὰ Δία. Alors
Socrate ajoute sèchement: ὡς σύ γε οἴει. En 473b7 Socrate introduit un point qui
n’a pas encore donné lieu à une réfutation: βούλει καὶ τοῦτο ἐλέγχειν; Ici encore
(cf. ἔλεγχε en 470c8) le PR a probablement la valeur de ‘soumettre à une réfuta-
tion’, mais une interprétation conative est également possible. Polos réagit de
la même façon ironique (473b9: Ἀλλ’ ἔτι… χαλεπώτερόν ἐστιν ἐξελέγξαι), mais
après la réaction de Socrate (qui le prend au sérieux : Οὐ δῆτα, ὦ Πῶλε, ἀλλ᾽ ἀδύ-
νατον· τὸ γὰρ ἀληθὲς οὐδέποτε ἐλέγχεται), il est tellement étonné qu’ il produit
une longue question rhétorique, dont la force perlocutoire est que l’ homme
injuste qui échappe à une punition est évidemment plus heureux que celui qui
est arrêté. Et il finit par une deuxième question rhétorique (ταῦτα λέγεις ἀδύνα-
τον εἶναι ἐξελέγχειν;). À noter qu’ici Polos emploie l’ inf. PR, et non pas l’ inf. AO,
sur les emplois de λεγε et ειπε chez platon 91

comme en 473b9. Cet inf. PR a peut-être été choisi pour mettre l’ accent sur le
procès de la réfutation que Polos vient de donner : ‘je viens de soumettre tout
cela à une réfutation’. Il va de soi que Socrate rejette cette ‘réfutation’ (473d3:
οὐκ ἐλέγχεις). Ensuite Socrate veut que Polos lui rappelle un point de détail,
ὑπόμνησόν με σμικρόν (cf. Lallot 2000: 42, n. 10), où l’ ΑΟ a une valeur sémelfac-
tive. Après la réponse de Polos, Socrate précise ce point, mais cette précision
amène Polos à éclater de rire (473e1: τί τοῦτο, Πῶλε; γελᾷς ;). Socrate réagit en
demandant à Polos si se moquer de quelqu’un au lieu de le réfuter (ἐλέγχειν,
avec valeur protocolaire) est encore une forme de réfutation. Une fois de plus
Polos a recours à une question rhétorique (473e6: Οὐκ οἴει ἐξεληλέγχθαι…), où le
PFT exprime que Polos, lui, considère la réfutation comme une chose acquise.
En fin de compte Socrate propose à Polos de changer de rôle, 474b1–2: ὅρα
οὖν εἰ ἐθελήσεις ἐν τῷ μέρει διδόναι ἔλεγχον ἀποκρινόμενος τὰ ἐρωτώμενα. Après
que Polos, plein de scepticisme, y a consenti, Socrate commence son question-
nement, sinon son interrogatoire, en 474c4: Λέγε δή μοι… | 163

2.1.5 Socrate questionneur, deuxième partie (474c4–479c6)

Le changement de procédure ne pourrait être plus radical. Premier indice :
la présence de λέγε en tête de l’interrogation. C’ est le signal univoque que
Socrate, lui, a en vue un vrai ἔλεγχος, et non pas la mécanique rhétorique que
Polos a mise en œuvre. Λέγε a une valeur itérative nette14, qui est encore renfor-
cée par la subordonnée ὥσπερ ἂν εἰ ἐξ ἀρχῆς σε ἠρώτων : ‘réponds-moi successi-
vement à chaque point que je soulève’. La portée de ce λέγε va jusqu’ à 475d4–5,
où l’interrogation aboutit à une question à l’ΑΟ sur un point tout à fait fon-
damental: Δέξαιο ἂν οὖν σὺ μᾶλλον τὸ κάκιον καὶ τὸ αἴσχιον ἀντὶ τοῦ ἧττον; Après
la réponse de Polos (Ἀλλ᾽ οὐκ ἂν δεξαίμην) Socrate conclut que οὐδεὶς ἀνθρώ-
πων δέξαιτ᾽ ἂν μᾶλλον ἀδικεῖν ἢ ἀδικεῖσθαι. Ce point établi – ce qui est encore
souligné par la formule sommaire en 476a2 καὶ τοῦτο μὲν ἡμῖν οὕτως ἐχέτω –,
Socrate invite Polos à passer à la seconde démarche de l’ ἔλεγχος : μετὰ τοῦτο δὲ
περὶ οὗ τὸ δεύτερον ἠμφεσβητήσαμεν σκεψώμεθα. L’AO a la même valeur qu’ en
472c5, voir plus haut: Socrate veut établir si la punition d’ un homme injuste
est oui ou non le plus grand des maux. Ici encore le procès concret de l’ examen
est marqué par le subjonctif PR: Σκοπώμεθα δὲ τῇδε (476a7). D’ emblée, Socrate
veut établir un point de base: διασκεψάμενος εἰπέ (476b2 ; mais cf. aussi note 13).
Ensuite, σκόπει invite Polos à considérer une nouvelle série de questions. On
trouve le même emploi de σκόπει en 477b1. (Voir aussi Vassilaki 2000.)

14 Il semble que λέγε, à la différence de ἀποκρίνου, n’a jamais une valeur protocolaire au sens
strict, et n’est donc pas utilisé quand les rôles de questionneur et de répondeur sont éta-
blis.
 92 chapter 4

La fin de l’interrogatoire est indiquée par ὦ Πῶλε, ἆρ᾽ αἰσθάνῃ τὰ συμβαίνοντα

ἐκ τοῦ λόγου ; ἢ βούλει συλλογισώμεθα αὐτά; (479c4). Parce que Polos opte pour
la seconde proposition, Socrate fait brièvement le bilan des résultats15. À partir
de 479d7, Socrate tire les conséquences des résultats obtenus pour la position
originelle de Polos, à savoir que la non-punition d’ un homme injuste est un
bien16.

2.2 Le Philèbe
Je passe maintenant au Philèbe, où il y a une majorité écrasante de λέγε (voir
164 la table). Le Philèbe, qui est généralement considéré comme | tardif, diffère des
autres dialogues tardifs en ceci que c’est Socrate qui conduit la discussion, et
non pas un Étranger, comme dans le Sophiste ou le Politique. Mais est-ce bien le
même Socrate que celui du Ménon, du Gorgias…17 ? En effet, les opinions sur le
rôle exact joué par Socrate dans le Philèbe, et donc sur le caractère général du
dialogue, diffèrent considérablement. Ainsi Hackforth écrit, dans sa traduction
commentée (1958: 8): ‘It has been urged that in the Philebus Socrates is unlike
himself in that he expounds rather than argues or persuades’18. Hackforth s’ en
prend à cette opinion (ibid.); il ne croit pas que dans le Philèbe Socrate soit
devenu un ‘ex cathedra lecturer. (…) Plato could still write a Socratic dialogue’.
De son côté, Diès écrit, dans l’introduction à son édition : ‘Quand on passe du
Politique au Philèbe, on a l’impression de rentrer brusquement dans le cadre
normal des dialogues platoniciens… (…) Socrate a retrouvé son rôle habituel : il
pose le problème, interroge, discute et conclut; ou plutôt, comme d’ ordinaire,
à peine la conclusion formulée, promet… de reprendre bientôt le débat pour
éclaircir quelques détails laissés dans l’ombre’. Diès ajoute toutefois : ‘Mais ce
n’est pas, cette fois, dans un avenir plus ou moins proche, c’ est tout de suite, ici
même, que la discussion est supposée se prolonger’.

15 Ici et ailleurs (p. ex. Grg. 498e10) συλλογίζομαι = ‘énumérer ensemble’, ‘faire le bilan en-
semble’, plutôt que simplement ‘reckon up’, ‘recapitulate’ (Dodds).
16 Dans cette partie finale le seul verbe de discussion est φημί: 480b6 (Pol.) τί γὰρ δή φῶμεν ;
et 480d6–7 (Socr.) φῶμεν οὕτως ἢ μὴ φῶμεν ; Peut-être (voir note 11) le sens est-il ‘Que croire,
en dehors de cela ?’ et ‘Faut-il penser ainsi ou non ?’
17 Cf. la question de Frede (1996: 216) : ‘In what sense does the Socrates of the Philebus
resemble the Socrates of the earlier dialogues ?’ Elle annonce que son analyse sera sur-
tout orientée vers la forme littéraire du dialogue (p. 217: ‘The Dramatis Personae and their
Roles’), mais elle ne donne aucune analyse formelle de ces rôles. Il en résulte qu’une
réponse claire à sa question manque.
18 À la p. 7, n. 4, il cite l’ opinion de Raeder (1920: 354) : ‘Der Sokrates, der hier auftritt, hat mit
dem Sokrates, der sonst in den platonischen Dialogen als Leiter des Gesprächs erscheint,
nur den Namen gemeinsam’.
sur les emplois de λεγε et ειπε chez platon 93

Pour ma part, je crois que l’opinion de Hackforth est fausse, ou du moins trop
restrictive. Le Philèbe diffère indéniablement d’un dialogue comme le Gorgias.
Un indice clair: tandis que dans le Gorgias il y a 41 exemples du verbe (ἐξ)ἐλέγ-
χειν, soit en moyenne un sur deux pages d’Estienne, il n’ y a, dans le Philèbe,
que cinq exemples19, soit un sur dix pages d’Estienne ; de même pour ἀποκρί-
νεσθαι: 55 exemples dans le Gorgias, 18 dans le Philèbe. En plus, les exemples
de (ἐξ)ἐλέγχειν dans le Philèbe sont tous au thème de PR ; dans le Gorgias, en
revanche, comme l’on a vu, l’AO est bien attesté. La conclusion s’ impose que
le Philèbe n’est nullement un dialogue au sens usuel, où Socrate combat et, le
cas échéant, réfute (AO) les opinions d’autrui. Le but du Philèbe est, en effet,
autre, à savoir de démontrer, par le moyen d’une méthode dialectique et non
pas élenctique, que le plaisir a une valeur moindre que l’ activité intellectuelle.
En apparence il s’agit d’une discussion, en réalité c’ est bien, pace Hackforth,
en grande partie une exposition, un | raisonnement20. Or, l’ emploi de λέγε four- 165
nit une contribution importante à ce caractère expositionnel du dialogue. Sans
entrer dans les détails, p. ex. le rôle d’autres verbes de discussion, je signale les
points suivants.
Des 27 exemples de λέγε dans le Philèbe, 21 sont dits par Protarque, cinq par
Socrate et un par Philèbe21. En revanche, des neuf exemples de λέγε dans le
Gorgias, huit sont dits par Socrate et seulement un par un autre participant,
Calliclès22.
Les λέγε de Protarque (tout comme celui de Philèbe) ont pour la plupart une
forme et une fonction tout à fait formulaires. Leur forme : λέγε, λέγε μόνον, λέγε
πῶς/ὅπῃ/τίνος, λέγε σαφέστερον etc. Leur fonction primaire: ils invitent Socrate
ou bien à commencer (une partie de) son raisonnement, son λόγος, ou bien à
le continuer. Voici deux cas typiques:

19 14b3, 14e2, 15a4, 23a7, 52d10.

20 Peut-être Diès a-t-il le caractère ‘expositionnel’ en vue quand il parle d’une ‘discussion pro-
longée’. Cf. aussi Thesleff (1967: 149) : ‘Phlb. is more like a discursive tract than any other
Platonic dialogue’. L’opinion de Waterfield (1980: 271) que le Philèbe est ‘a bad example of
the earlier dialogue form’ est complètement erronée. Hackforth (1958: 8) compare le rôle
de Socrate dans le Philèbe avec son rôle dans la République. C’est exact, mais il faut préci-
ser que c’ est notamment le Socrate de la discussion avec Glaucon et Adimante; Diès, dans
son Introduction à l’ édition des Belles Lettres de la République, parle à juste titre (p. xxiii)
d’ ‘une discussion qui deviendra progressivement un exposé presque continu’. Voir aussi
la note 23.
21 Protarque : 13el, 14c6, 16c4, 21a7, 23d6, 25c4, 25e6, 28d1, 29a8, 30a2, 31d3, 33a7, 41c10, 43a9,
48c5, 49c9, 53d2, 53d8, 53e2, (54b2 coni. Badham), 61d6, 64b9; Socrate: 36el, 45d5 (après
ἀπόκριναι : un bel exemple de l’ emploi immédiatif de λέγε), 46d7, 50d6, 65el; Philèbe :
27e4.
22 Socrate : 451d5, 470a4, 474c4, 492d5, 494b7, 495a2, 515a7, 516d6; Calliclès; 506c4.
 94 chapter 4

(3) ΣΩ. Σὺ δέ μοι συνακολούθησον τῷ λόγῳ.

ΠΡΩ. ΛέγεPR μόνον. (Phlb. 25c2–4)

Socrate: Suis mon raisonnement jusqu’à son terme.

Protarque: Vas-y, parle. (trad. Diès, modifiée)

(4) ΣΩ. Ἴθι δή, τὸν ἐπιόντα περὶ τούτων νῦν ἡμῖν λόγον ἄθρει.
ΠΡΩ. ΛέγεPR μόνον. (Phlb. 29a6–8)

Socrate: Eh bien, considère donc, à ce propos, le raisonnement qui nous

survient.
Protarque: Tu n’as qu’à continuer (trad. Diès, modifiée)

Socrate, en revanche, ne dit jamais λέγε μόνον, λέγε ou λέγε σαφέστερον. Quand
166 c’est lui qui emploie cet impératif, c’est le signe qu’ il ne raisonne | plus, mais
interroge. Voir 36e1 λέγε δή μοι, usage tout à fait comparable à celui de λέγε δή
μοι dans le Gorgias 474c4, voir plus haut23. De son côté, Protarque ne dit jamais
λέγε δή μοι24.
Je ne puis terminer, évidemment, sans dire quelques mots sur le seul cas de
εἰπέ, qui se trouve en 32e4–5:

(5) ΣΩ. σφόδρα δὲ προσέχων τὸν νοῦν εἰπέAO· ἆρα οὐ πᾶσα ἀνάγκη πᾶν ἐν τῷ τότε
χρόνῳ ζῷον μήτε τι λυπεῖσθαι μήτε ἥδεσθαι μήτε μέγα μήτε σμικρόν ; (Phlb.
32e4–5)

23 Il est intéressant de noter (voir note 20) que l’ emploi de λέγε dans la partie de la Répu-
blique où Socrate discute avec Glaucon (427e–445e, 450b–487a) est tout à fait similaire à
son emploi dans le Philèbe : des dix exemples de λέγε, neuf sont dits par Glaucon, un par
Socrate. Or, les λέγε de Glaucon sont tous du type ‘protarquien’, voir R. 436c4, 445c3, 445d2,
451b5, 451b8, 457c6 (λέγε δή ; noter toutefois les conjectures φέρε δή de Cobet et ἄγε δή de
Richards), 471e3, 472b2, 473c10. Le λέγε de Socrate (459a1) ressemble, lui aussi, aux λέγε
‘socratiques’ du Philèbe. L’emploi de λέγε est du même type dans le Sophiste et le Politique :
dans le Sophiste les λέγε de Théétète, et dans le Politique ceux du Jeune Socrate, sont du
type ‘protarquien’, tandis que les λέγε de l’ Étranger dans ces deux dialogues sont du type
‘socratique’. Quant au Parménide, ce dialogue a une position à part parmi les dialogues
tardifs par l’ absence totale de λέγε (voir la table). Le Parménide est en effet sui generis,
tant parce qu’ il est composé essentiellement d’ une série d’hypothèses et d’antinomies
que parce que les interlocuteurs principaux sont Parménide et Aristote.
24 En effet, il semble que, dans de telles demandes, μοι soit uniquement employé par celui
qui conduit la discussion, en tant que personne qui y a le plus d’intérêt, soit Socrate (p.
ex. dans le Gorgias, le Ménon, la République, le Philèbe), soit une autre personne, comme
Polos (Grg. 462b3: καί μοι ἀπόκριναι) ou l’ Étranger (Sophiste, Politique).
sur les emplois de λεγε et ειπε chez platon 95

Socrate: Prête-moi toute ton attention et disAO-moi : n’est-il pas totale-

ment inévitable que tout vivant, dans une telle circonstance, ne sente ni
peu ni prou de douleur ni de plaisir? (trad. Diès)

Le choix des mots emphatiques (σφόδρα προσέχων τὸν νοῦν, πᾶσα ἀνάγκη, πᾶν
ζῷον) indique que Socrate considère ce point comme étant d’ une grande
importance, et il prépare, en effet, l’introduction, en 33a8–9, de l’ homme qui a
choisi la vie de sagesse. L’emploi de εἰπέ s’explique donc bien : l’ aoriste indique
qu’il s’agit ici d’établir un point important (voir plus haut).

2.3 Le Ménexène
Dans le Ménexène, où λέγε manque, il y a trois cas de εἰπέ, très voisins les uns
des autres: | 167

(6) ΜΕΝ. Τί οὖν οὐ διῆλθες ;

ΣΩ. Ἀλλ᾽ ὅπως μή μοι χαλεπανεῖ ἡ διδάσκαλος, ἂν ἐξενέγκω αὐτῆς τὸν λόγον.
ΜΕΝ. Μηδαμῶς, ὦ Σώκρατες, ἀλλ’ εἰπέAO, καὶ πάνυ μοι χαριῇ, εἴτε Ἀσπα-
σίας βούλει λέγεινPR, εἴτε ὁτουοῦν· ἀλλὰ μόνον εἰπέAO.
ΣΩ. Ἀλλ᾽ ἴσως μου καταγελάσῃ, ἄν σοι δόξω πρεσβύτης ὢν ἔτι παίζειν.
ΜΕΝ. Οὐδαμῶς, ὦ Σώκρατες, ἀλλ’ εἰπέAO παντὶ τρόπῳ. (Mx. 236c1–10)

Ménexène: Qu’attends-tu donc pour l’exposer ?

Socrate: Prends garde que mon maître ne se fâche contre moi, si je
divulgue son discours!
Ménexène: Ne crains rien, Socrate, et parleAO. Tu me feras le plus grand
plaisir, que ce soit d’Aspasie ou de tout autre que tu veuilles rappor-
terPR les propos. ParleAO seulement.
Socrate: Mais peut-être vas-tu rire de moi, si je te parais, vieux comme je
suis, m’adonner encore au badinage.
Ménexène: Point du tout, Socrate. ParleAO, de toute façon. (trad. Méri-
dier)

Ces trois εἰπέ ont une position tout à fait à part: (1) ce sont les seuls cas de ἀλλ᾽
εἰπέ où εἰπέ est employé absolument; (2) le deuxième est le seul cas où εἰπέ
est modifié par μόνον25. En effet, l’emploi de εἰπέ ici s’ inscrit dans un contexte
qui, lui aussi, est différent des contextes ‘normaux’ de εἰπέ (et λέγε), parce qu’ il

25 Tandis que μόνον dans λέγε μόνον ou μόνον λέγε renforce la valeur ‘immédiative’ de l’impé-
ratif (‘tu n’as qu’ à parler’), sa valeur dans ἀλλὰ μόνον εἰπέ est plutôt ‘il n’y a qu’une chose
qui importe : prononce une oraison funèbre’.
 96 chapter 4

s’ agit non pas d’un dialogue mais d’une brève discussion entre Ménexène et
Socrate sur les capacités de celui-ci de parler ἐπὶ τοῖς ἀποθανοῦσιν (234b6), cf.

(7) ΜΕΝ. Ἦ οἴει οἷός τ᾽ ἄν εἶναι αὐτὸς εἰπεῖνAO, εἰ δέοι καὶ ἕλοιτό σε ἡ βουλή ;
ΣΩ. Καὶ ἐμοὶ μέν γε, ὦ Μενέξενε, οὐδὲν θαυμαστὸν οἵῳ τ᾽ εἶναι εἰπεῖνAO, ᾧ…
(Mx. 235e1)

Ménexène: Te croirais-tu capable de prononcer toi-même une oraison,

s’il le fallait et que tu fusses choisi par le Conseil ?
Socrate: Moi aussi, bien sûr, Ménexène, il ne serait point surprenant que
je fusse en état de prononcer une oraison, moi qui… (trad. Méridier,
168 modifiée) |

Dans tout ce passage les formes de εἰπ-, pour autant qu’ elles soient employées
absolument, ont le sens spécialisé et idiomatique de ‘prononcer une orai-
son (funèbre)’26. Et quoique λέγε manque, il y a bien une opposition avec
λεγ-. Voir notamment 236c5–7, cité ci-dessus: ἀλλ’ εἰπέ, καὶ πάνυ μοι χαριῇ, εἴτε
Ἀσπασίας βούλει λέγειν εἴτε ὁτουοῦν· ἀλλὰ μόνον εἰπέ27. Tandis que εἰπέ dénote
l’ oraison tout court in abstracto (‘prononce une oraison funèbre d’ une façon
ou d’une autre’, cf. παντὶ τρόπῳ en 236c10), λέγειν se réfère plutôt à l’ exécution,
à la réalisation d’une oraison concrète. Noter la présence d’ un objet avec
λέγειν (Ἀσπασίας, sc. λόγον), tout comme en 236b4 (οἷα) et 236b7 (ἅ)28. Lorsque
Socrate cède finalement aux supplications de Ménexène et prononce son orai-
son (qui est celle d’Aspasie), il emploie naturellement λεγ- (ἔλεγε… οὑτωσί,
236d2–3).

26 Noter que le sens de ‘prononcer une oraison’ peut être rattaché à la valeur d’accomplisse-
ment. Cf. aussi l’ emploi absolu de εἰπεῖν (‘tenir un discours’) dans Smp. 194e4ss.: ΑΓ.
ἐγὼ δὲ δὴ βούλομαι πρῶτον μὲν εἰπεῖν ὡς χρή με εἰπεῖν, ἔπειτα εἰπεῖν (‘Agathon: Eh bien!
moi, je veux d’ abord préciser comment il faut que je tienne un discours, et après cela
tenir un discours’ ; trad. AR). On notera le jeu avec εἰπεῖν + complément = ‘établir, préci-
ser’.
27 Pour λεγ- voir aussi 236b4, 236b7 et 236d2–3.
28 En 236a7, c’ est λέγειν qui est employé absolument ; ici, cet infinitif a probablement le sens
spécialisé de ‘paraître en orateur, être ὁ λέγων’, cf. 235c9 ὁ λέγων.
sur les emplois de λεγε et ειπε chez platon 97

3 Conclusion

Les principaux résultats de l’analyse présentée ci-dessus peuvent être résumés

comme suit.

Λέγε et εἰπέ dans le Gorgias et dans le Philèbe

Λέγε s’emploie de deux façons différentes, qui sont liées au caractère du texte
où λέγε se trouve et au rôle joué par celui qui parle ; εἰπέ par contre s’ emploie
de la même façon dans les deux dialogues.
(i) Dans le Gorgias, un dialogue élenctique, λέγε est surtout dit par Socrate,
qui l’ emploie pour obtenir de son interlocuteur ou bien une réponse à
une question spécifique, avec une valeur ‘immédiative’ (‘tu n’as qu’ à le
dire’, ‘vas-y’), ou bien des réponses à une série de questions, avec une
valeur itérative-protocolaire. Quant à εἰπέ, cet impératif invite toujours
l’interlocuteur à établir un point spécifique ; il marque donc une dé-
marche importante dans l’interrogation (valeur d’ accomplissement).
(ii) Dans le Philèbe, en revanche, un dialogue qui pour la plus grande part est
un raisonnement plutôt qu’une réfutation, λέγε est surtout employé par
l’interlocuteur de Socrate, afin de lui permettre de | poursuivre son raison- 169
nement, ou bien avec une valeur ‘immédiative’ (‘tu n’as qu’ à parler’) ou
bien avec une valeur continuative, ou encore avec une combinaison de
ces valeurs (‘tu n’as qu’à continuer’). L’unique cas de εἰπέ est employé de
la même façon que εἰπέ dans le Gorgias.

Εἰπέ dans le Ménexène

Dans le Ménexène, un texte qui n’est pas un dialogue, l’ emploi de εἰπέ, tout
comme celui des autres formes de εἰπ-, diffère nettement de son emploi dans un
texte dialogué: εἰπ- a le sens idiomatique de ‘prononcer une oraison funèbre’.
Bien que λέγε y manque, il y a bien une opposition entre εἰπ- et les formes de
λεγ-: εἰπ- dénote une oraison in abstracto, λεγ- dénote l’ exécution, la réalisa-
tion d’une oraison concrète.

Références

Amigues, S., ‘Les temps de l’impératif dans les ordres de l’orateur au greffier’, REG 90
(1977) 223–238.
Bakker, W.F., The Greek Imperative (Amsterdam 1966).
Basset, L., ‘Corrélations aspectuelles PR/AO entre verbes régissants et verbes régis’, in
Jacquinod (2000) 233–245.
98 chapter 4

Fournier, H., Les verbes “dire” en grec ancien (Paris 1946).

Frede, D. 1996. ‘The Hedonist’s Conversion: The Role of Socrates in the Philebus’, in
C. Gill and M.M. McCabe (eds), Form and Argument in late Plato (Oxford 1996), 213–
248.
Hackforth, R., Plato’s Examination of Pleasure: A Translation of the Philebus with Intro-
duction and Commentary (Cambridge 1958).
Hillesum, L. (diss.), De imperfecti et aoristi usu Thucydideo, pars prior (Amsterdam
1908).
Jacquinod, B. (ed.), Études sur l’aspect verbal chez Platon (Saint-Etienne 2000).
Lallot, J., ‘Essai d’interprétation de l’opposition PR-AO à l’impératif de ἀποκρίνεσθαι
dans l’oeuvre de Platon’, in Jacquinod (2000) 29–74.
Moorhouse, A.C., 1982. The Syntax of Sophocles (Leiden 1982).
Oréal, E., ‘Détermination et indétermination: un paramètre du fonctionnement de
l’aspect en grec ancien’, in Jacquinod (2000) 285–302.
Poutsma, A., Over de tempora van de imperativus en de conjunctivus hortativus-prohibi-
tivus in het Grieks (Amsterdam 1928).
Raeder, H., Platons philosophische Entwicklung, 2nd ed. (Leipzig 1920).
Rijksbaron. A., The Syntax and Semantics of the Verb in Classical Greek, 2nd ed. (Amster-
dam 1994).
Ruijgh, C.J., ‘L’emploi ‘inceptif’ du thème du présent du verbe grec’, Mnemosyne 38
(1985) 1–61 (= Scripta minora II (Amsterdam 1996), 535–595).
Sicking, C.M.J., ‘The Distribution of Aorist and Present Stem Forms in Greek, Especially
in the Imperative’, Glotta 69 (1991) 14–43, 154–170.
Thelin, N.B. (ed.), Verbal Aspect in Discourse (Amsterdam/Philadelphia 1990).
Thesleff, H., Studies in the Styles of Plato (Helsinki 1976).
Vassilaki, S., ‘A propos des valeurs des impératifs σκόπει (PR) – σκέψαι (AO) chez Platon’,
in Jacquinod (2000) 171–201.
Wakker, G., ‘Infinitif PR et infinitif AO: l’opposition ἀποκρίνεσθαι – ἀποκρίνασθαι chez
Platon’, in Jacquinod (2000) 203–230.
Waterfield, R.A.H., ‘The Place of the Philebus in Plato’s Dialogues’, Phronesis 25 (1980)
270–305.
 chapter 5

On False Historic Presents in Sophocles (and

Euripides)

Abstract

Many messenger speeches in our current Sophocles and Euripides texts contain forms
that may be interpreted either as a historic present or as an—augmentless—imperfect,
depending on the accent. Examples are κυρεῖ/κύρει, κυνεῖ/κύνει, φοιτᾷ/φοίτα. In such
cases our texts most often have the ‘present’ accentuation, with or without support
of the manuscripts. Since I am about to transform several of such historic presents
into imperfects, I should perhaps first provide some linguistic background to my argu-
ment, both with respect to the historic present and to the augmentless imperfect. I will
start by briefly setting forth what is in my view the function of the historic present.
I will then mention a number of semantic and syntactic features of uncontroversial
historic presents both in drama and in two historical works, Herodotus’ Histories and
Xenophon’s Anabasis, the idea being that these features may guide us in determin-
ing whether a given tragic candidate for this function qualifies as a historic present or
not. This will be followed by some remarks on augmentless past tenses in Sophocles
and Euripides. Finally, I will discuss a number of dubious historic presents in the two
dramatists.

1 The Function of the Historic Present

The preferred habitat of the historic present is, of course, the narrative, i.e.
a report of past events, which are predominantly narrated in imperfects and
aorist indicatives;1 there is, in fact, no narrative without these two tenses. The
aorist presents the main events of the story, i.e. the events that warrant that
the narrative moves on, while the imperfects | provide the background to these 128
events. At varying intervals the flow of these past tenses may be interrupted
by the occurrence of a present indicative form, traditionally called the his-
toric present, sometimes also narrative present. An important function of these

1 Occasionally, the historic present occurs in what may be called mixed narrative-argumenta-
tive passages: an example is μισθοῦται in πῶς οὖν ταῦτ᾽ ἐποίησεν; μισθοῦται τουτονί (D. 18.148).

© koninklijke brill nv, leiden, 2019 | doi:10.1163/9789004386129_007

 100 chapter 5

presents is to present events that the narrator considers crucial or decisive

for the development of the plot.2 An oft-cited example of this use is ἐπορᾷ in
Herodotus 1.10.2:

(1) ὡς δὲ κατὰ νώτου ἐγένετο ἰούσης τῆς γυναικὸς ἐς τὴν κοίτην, ὑπεκδὺς ἐχώρεε
ἔξω. καὶ ἡ γυνὴ ἐπορᾷ μιν ἐξιόντα … (Hdt. 1.10.2)

And when she turned her back upon him, going to her bed, he crouched
and slipped from the room. And the woman saw him as he went out …

where the fact that Candaules’ wife sees Gyges leave the bedroom completely
overthrows the course of events as planned by Candaules. Needless to say, it
is not just the perception verb ἐπορᾷ that is marked here as crucial by the
present tense but this verb plus the dependent state of affairs (ἐξιόντα). I
realize that in assigning this ‘decisive’ function to the historic present I run
a risk of falling victim to circular reasoning, there being no other indica-
tions of ‘decisiveness’ than the historic present. Below, I will try to neutral-
ize this risk by pointing to a number of features of the historic present that
can satisfactorily be accounted for if expressing ‘decisiveness’ is considered
the basic function of the historic present and that otherwise remain unex-
plained.
There is, not surprisingly, much controversy about the question as to what
is and is not a decisive event. Thus, Irene de Jong (1991: 40) observes à propos
129 Medea 1141: |

(2) κυνεῖ δ᾽ ὁ μέν τις χεῖρ᾽, ὁ δὲ ξανθὸν κάρα παίδων· (E. Med. 1141)

2 This view is already found in K-G (1, 132): ‘Oft neben Aoristen oder Imperfekten zur Hervorhe-
bung einzelner besonders bemerkenswerter und für die Folge wichtiger Momente’. Cf. further
Eriksson (1943: 9; ‘Hauptpunkte’), S-D 2, 271 (‘entscheidende und neue Momente’), also Kells
(1973: on S. El. 35; ‘… reporting a critical act or occurrence’), Sicking & Stork (1997: 147–156).—
Ultimately, and at a more theoretical level, the ‘decisive’ function of the historic present can be
accounted for along the lines sketched by Benveniste in his article ‘Le langage et l’expérience
humaine’ (Benveniste 1974). By interrupting the flow of past tenses the present tense creates
the illusion that the event concerned does not belong to the past, but is ‘un moment neuf,
non encore vécu’, in the words of Benveniste (1974: 74). Being presented as new and ‘having
not yet been gone through’, the event has, in a context of events ‘that have been gone through’,
another information status than these other events. By suddenly confronting the reader with
‘un moment non encore vécu’, the narrator wants him to know that this is not an ordinary, but
an extraordinary event of the past, an event that may be expected to have more far-reaching
effects than the surrounding events.
on false historic presents in sophocles (and euripides) 101

And someone kissed the hands and another the blond heads of the chil-
dren. (Trans. here and below Kovacs.)

‘[…] in the case of Med. 1141 (servants kiss—κυνεῖ—the hands and heads of
Medea’s children) it would be difficult to maintain that this is a decisive event’.
I could not agree more. Fortunately, Irene de Jong’s uneasy feelings can, in this
case at least, be dissipated quite easily. In fact, this quotation brings us straight-
away to the heart of the matter. For it so happens that κυνεῖ is a conjecture of
Brunck’s, the reading of the MSS being the imperfect, κύνει. At the end of my
paper I will come back to this case.
While in drama, too, the ‘decisive’ function is an important one, the narrative
situation there differs from that in historical texts, in as much as the messengers
and other narrators report about events of which they themselves have been a
witness; to quote Irene de Jong once more (1991: 8): ‘[…] being an eyewitness is
the messenger’s very theatrical raison d’être’. The historic present is the means
par excellence to give expression to his eyewitness status: the narrator speaks as
if he is again on the spot and is experiencing a ‘moment non encore vécu’. This,
in turn, may lend a certain vividness to his report. However, ‘vividness’ is to my
mind not a central notion in connection with the historic present.3 To show
its doubtful status it suffices perhaps to replace ‘decisive’ in the above quota-
tion from Narrative in Drama with ‘vivid’. The resulting statement is devoid of
meaning.4

2 Some Syntactic and Semantic Features of the Historic Present

The use of the historic present is far more restricted than that of imperfect and
aorist, as appears from the following list of features. | 130

I. The fact that it is (almost) confined to narrative discourse has already been
mentioned.

3 Supporters of the ‘vividness’ theory sometimes appeal to [Long.] Subl. 25, where it is said
that ‘when you introduce the past things as happening and being present, οὐ διήγησιν ἔτι τὸν
λόγον ἀλλ᾽ ἐναγώνιον πρᾶγμα ποιήσεις’. Pace LSJ and others I do not think that ἐναγώνιος here
means ‘energetic, vivid’, but rather, as elsewhere, ‘argumentative, suited for forensic oratory’.
In fact, the example from Demosthenes’ De corona mentioned in note 1 probably belongs to
this category.
4 I have discussed the problems connected with the notion ‘vivid’ elsewhere, see Rijksbaron
(2002: 257 and 261 f.). See also Sicking & Stork (1997: 131–134).
 102 chapter 5

II. A consequence of I is that the historic present does not occur in excla-
mations or questions,5 but only in declarative sentences. The—remarkable—
exception that proves the rule (for Sophocles and Euripides) is Creon’s ques-
tion

(3) Φυ. ταύτην γ᾽ ἰδὼν θάπτουσαν ὃν σὺ τὸν νεκρὸν

ἀπεῖπας. ἆρ᾽ ἔνδηλα καὶ σαφῆ λέγω;
Κρ. καὶ πῶς ὁρᾶται κἀπίληπτος ᾑρέθη;
Φυ. τοιοῦτον ἦν τὸ πρᾶγμ᾽· ὅπως γὰρ ἥκομεν, … (S. Ant. 404–407)

Phy. I saw her burying the corpse that thou hadst forbidden to bury. Is
that plain and clear?
Cr. And how was she seen? How taken in the act?
Phy. It befell on this wise. When we had come to the place … (Trans. here
and below Jebb.)

which comes just before the full report of the Guard, who had already given
away, as messengers always do, the main event (l. 404). With πῶς ὁρᾶται, Creon
asks the Guard to elaborate upon that ‘seeing’. The use of the historic present
leaves no doubt that for Creon a crucial event is involved, and that he asks the
messenger to transfer him mentally to the flagrant délit.

III. A further consequence of I is that the historic present does not occur in the
second person, but only in the third and first persons.

IV. Not all verbs have historic presents. They are, in fact, confined to telic,
or terminative, and momentaneous verbs (so-called ‘accomplishments’ and
‘achievements’, respectively), and do not occur, then, with durative-stative
verbs. This means that forms like ἐστί, κεῖται, μένει, ἔχει, νομίζει, εὕδει are never
131 used as a historic present.6 As we | will see presently, this has important con-

5 With the exception of rhetorical questions, i.e. questions that are pragmatically equivalent to
a statement, e.g. τίς οὐ σίδηρον προσφέρει, τίς οὐ πέτρον; E. Andr. 1153.
6 ‘Nicht von Zustände’, S-D 2, 271; ‘des verbes comme εἰμί, κεῖμαι n’apparaissent jamais au
présent historique’, Ruipérez (1982: 182). Moorhouse (1982: 185) claims that a number of his-
toric presents in the report of Orestes’ ‘death’ in Electra have ‘imperfective’ meaning, but
this is unlikely. (βίᾳ) φέρουσιν at 725 is said ‘de rectore auferendo’ (Ellendt-Gendte); ἀνα-
κωχεύει (732) = ‘suspend sa marche un moment’ (Mazon); διώκει (738) = ‘se lance’ (Mazon);
ἑλίσσεται (746) must be connected with the preceding σύν = ‘he was rolled up in a ball’
(Kells).
on false historic presents in sophocles (and euripides) 103

sequences for our material. The activity involved is normally displayed by a

human being,7 but occasionally by natural phenomena, as at X. An. 4.1.15 εἰς
δὲ τὴν ὑστεραίαν γίγνεται χειμὼν πολύς.8

V. Historic presents are rare in the passive voice; there are four examples in
Sophocles: Ant. 406 and 423 ὁρᾶται, already mentioned above, OT 812 (ὁ πρέ-
σβυς) ἐκκυλίνδεται, El. 746 (Orestes) σὺν δ᾽ ἑλίσσεται τμητοῖς ἱμᾶσι, and two in
Euripides: Alc. 184 (δέμνιον) δεύεται (unless this has middle meaning, see also
note 27) and Hipp. 1236–1237 αὐτὸς δ᾽ ὁ τλήμων … | … ἕλκεται δεθείς.

VI. Historic presents are rare in subordinate clauses. An example from Sopho-
cles is

(4) ὁ δ᾽ ὡς ὁρᾷ σφε, στυγνὸν οἰμώξας ἔσω

χωρεῖ πρὸς αὐτὸν … (S. Ant. 1226–1227; messenger speaking)

But when Creon saw him, with a dreadful groan he came inside towards
him …

VII. Historic presents cannot be combined with the negative, except under spe-
cial conditions, as in:

(5) ὁ δ᾽ οὖν Τισσαφέρνης ὡς μεῖον ἔχων ἀπηλλάγη, πάλιν μὲν οὐκ ἀναστρέφει, εἰς
δὲ τὸ στρατόπεδον ἀφικόμενος τὸ τῶν Ἑλλήνων ἐκεῖ συντυγχάνει βασιλεῖ (X.
An. 1.10.8)

At any rate, after Tissaphernes had thus come off with the worst of it, he
did not wheel round again, but went on to the camp of the Greeks and
there fell in with the King

which is, in the narrative parts of Books 1 and 2 of the Anabasis, one of the
three instances of οὐ + historic present, alongside 44 cases of οὐ with imperfect
or aorist. The negated historic present οὐκ ἀναστρέφει is only acceptable, I think,
because it is balanced by a positive statement (συντυγχάνει).9 In Sophocles

7 Explicitly or implicitly, the latter at e.g. X. HG 3.2.31 τούτων δὲ συγχωρηθέντων εἰρήνη τε γίγνεται
καὶ συμμαχία Ἠλείων πρὸς Λακεδαιμονίους.
8 In Euripides, there is just one example of a non-human subject of a historic present, Hipp.
1212 χωρεῖ, subject a gulf, in Sophocles none (for Tr. 767 see below, ex. (14)).
9 The other two instances are 1.10.1 οἱ μὲν μετὰ Ἀριαίου οὐκέτι ἵστανται, ἀλλὰ φεύγουσι, and 2.6.3
 104 chapter 5

132 and Euripides there | are no such exceptions: οὐ never modifies a historic
present; cf. also Moorhouse (1982: 184).

VIII. As appears from the following table, the sentences in which historic
presents appear are normally introduced by δέ or καί:

Herodotus 1.1–70 Sophocles Euripides

δέ 23 24 84
καί 3 19 18
τε – 1 1
ἀλλά – 2 2
asyndeton 9 7 5
γάρ – 5 2

One of the striking features of this survey is the total absence of γάρ from
Herodotus, and its small numbers in Sophocles and Euripides. The marginal
position of γάρ in connection with the historic present also appears from the
fact that of the 241 instances of γάρ in the first four Books of Xenophon’s Anaba-
sis just one is combined with a historic present. This involves, moreover, a
special use of γάρ. The example runs:

(6) ὁποῖόν τι μὲν δὴ ἐστὶ τὸ τοιοῦτον ὄναρ ἰδεῖν ἔξεστι σκοπεῖν ἐκ τῶν συμβάντων
μετὰ τὸ ὄναρ. γίγνεται γὰρ τάδε. εὐθὺς ἐπειδὴ … (X. An. 3.1.13)

Now what it really means to have such a dream one may learn from the
events which followed the dream—and this is what happened. Immedi-
ately after …

In this example γάρ introduces an embedded narrative, explaining what the

συμβάντα mentioned in the preceding sentence consisted of.10 This function of

ἐνταῦθα οὐκέτι πείθεται, ἀλλ᾽ ᾤχετο πλέων εἰς Ἑλλήσποντον. Observe that here, too, the
negated historic present is balanced by a positive statement.
10 Our passage is exceptional in that the narrative is introduced in two stages, so to speak. It
might also have started straightaway with εὐθὺς γὰρ ἐπειδή …, but as it is, it is announced
by the historic present + γάρ, whereby the importance of the events following the dream is
heavily emphasized. And with good reason, for Xenophon is speaking here about a dream
on false historic presents in sophocles (and euripides) 105

γάρ has recently been discussed by Sicking and Van Ophuijsen (1993: 20–21) and
De Jong (1997).11 In Sophocles | and Euripides, too, if the historic present occurs 133
in a γάρ-clause, this is only in γάρ-clauses introducing an embedded narrative
of the type exemplified by (6). A clear case in point is:

(7) ἠγόμην δ᾽ ἀνὴρ

ἀστῶν μέγιστος τῶν ἐκεῖ, πρίν μοι τύχη
τοιάδ᾽ ἐπέστη, … |
ἀνὴρ γὰρ ἐν δείπνοις μ᾽ ὑπερπλησθείς μέθῃ
καλεῖ παρ᾽ οἴνῳ πλαστὸς ὡς εἴην πατρί. (S. OT 775–780; Oedipus speaking)12

and I was held the first of all the folk in that town, until the following
chance befell me. At a banquet a man full of wine cast it at me in his cups
that I was not the true son of my sire.

Here, the sentence introduced by γάρ explains the nature of τύχη τοιάδε; καλεῖ
marks the event as decisive. I should add that in such sentences the past tenses
are also, and indeed more frequently, found.13 For an example from drama with
a pluperfect see below, example (18).
In regular explanatory γάρ-clauses, on the other hand, γάρ is never found
with the historic present. In such clauses (that is, in narrative) only the imper-
fect and—far less often—aorist occur; three examples with an imperfect are:

(8) πολλὰ τῶν ὑποζυγίων ἀπώλετο ὑπὸ λιμοῦ· οὐ γὰρ ἦν χόρτος (X. An. 1.5.5)

… many of the baggage animals died of hunger, for there was no fodder …

he had himself, a dream that eventually led to his appointment as the commander-in-chief
of the Greeks.
11 Kroon (1995: 148 ff.) observes, on a similar use of Latin nam introducing narratives: ‘[…]
a general or summarizing statement […] is followed by a particular instance (not seldom
in the form of an extensive narratio) which clarifies or fills in the detail of the general
statement’.
12 The other examples in S. are: Aj. 764, Ant. 1001, El. 35. The examples from Euripides are Ph.
1410 and 1458; IT 1395 does not belong here, see below, note 27.
13 Cf. e.g. ἐπικατερρίπτουν at X. An. 4.7.13, where the embedded narrative is formally an-
nounced by δεινὸν … θέαμα: ἐνταῦθα δὴ δεινὸν ἦν θέαμα. αἱ γὰρ γυναῖκες ῥίπτουσαι τὰ παιδία
εἶτα ἑαυτὰς ἐπικατερρίπτουν, καὶ οἱ ἄνδρες ὡσαύτως. In Sophocles and Euripides, too, this is
common, cf. e.g. OT 1268 ff., Ba. 760.
 106 chapter 5

(9) καὶ οἱ μὲν πλεῖστοι ἐξέφυγον· πλησίον γὰρ ἦν τὸ ὄρος· (X. An. 7.4.6)

Now most of the villagers made their escape, for the mountain was close
at hand

(10) κοὐδεὶς ἐτόλμα τἀνδρὸς ἀντίον μολεῖν.

ἐσπᾶτο γὰρ πέδονδε καὶ μετάρσιος (S. Tr. 785–786)

and no one dared to to come before the man. For the pain dragged him to
134 earth, or made him leap into the air |

In fact, the basic function of γάρ is to provide this type of background infor-
mation.14 The two uses can in principle be distinguished by the following rule-
of-thumb: regular γάρ presupposes an implicit question like ‘how was this pos-
sible?’ (cf. (9)) or ‘what was the cause/reason of this?’ (cf. (8) and (10)), while
narrative γάρ presupposes rather a question like ‘what did x consist of?’ (cf. (6)
and (7)).
The importance of the distinction between these two uses of γάρ will
become clear below.

3 Some Conclusions

The features mentioned at IV through VIII should, of course, be accounted for,

and this can be done in a satisfactory way if signalling ‘decisive events’ is taken
as the basic function of the historic present:
– (IV. non-occurrence of stative historic presents) the semantics of stative
verbs are inherently at odds with the notion ‘decisive event’; verbs like
‘be’, ‘think’, ‘have’, ‘sleep’ are rather used to provide background informa-
tion
– (V. near-absence of historic presents in the passive voice) decisive events are
typically caused by actively operating persons, not by persons undergoing an
action
– (VI. rareness in subordinate clauses) since subordinate clauses typically
provide information that is on a lower hierarchical level than that pro-

14 In what is called, in the pragmatic model known as ‘the Geneva model’, a ‘subsidiary
discourse act’, to be distinguished from ‘central acts’, like, in our examples, πολλὰ τῶν ὑποζυ-
γίων ὰπώλετο ὑπὸ λιμοῦ·and οἱ μὲν πλεῖστοι ἐξέφυγον, respectively. Cf. for these distinctions
Kroon (1995: 65–66).
on false historic presents in sophocles (and euripides) 107

vided by independent sentences, one may expect decisive events to occur at

the highest level
– (VII. near-absence of negated historic presents) non-events may be expected
to qualify less easily as decisive events than real events15
– (VIII. near-absence of historic presents with γάρ) decisive events may be
expected to avoid collocation with a particle whose main function is to intro-
duce background information. | 135
Of these semantic features, it is especially the features mentioned at IV and VIII
that can help us determine whether κυρει and the other forms under discussion
should be interpreted as historic presents or as imperfects.

4 The Unaugmented Imperfect (and Aorist)

Having given an idea of my theoretical framework concerning the historic

present, I now turn to the unaugmented imperfect, or to be more precise, the
imperfect without the syllabic augment, for the absence of the temporal aug-
ment plays no role in my corpus.16
The existence in Greek drama of unaugmented past tenses was for a long
time denied by scholars, but is nowadays more or less grudgingly acknowl-
edged. Rather more than less—for generally speaking the situation is still thus,
that if an editor sees a way to get rid of an unaugmented past tense he will not
hesitate to do so. This may involve rather innocent measures like adding a coro-
nis in cases where aphaeresis is believed to be involved, as in (11), where most
editors follow Brunck in reading ᾽θώυξεν:

(11) παίσας κάρα ᾽θώυξεν (S. Aj. 308)

he struck his head and uttered a loud sound

θώυξεν codd. : ᾽θώυξεν Brunck

15 The special information status of negative statements may be illustrated from the fol-
lowing example from Givón (1984: 323): (a. Context)—‘What’s new?’ (b. AFF(IRMATIVE)-
reply)—‘Oh, my wife is pregnant’; (c. NEG-reply)—‘Oh, my wife is not pregnant’. Sentence
(c.) is ‘distinctly odd’, as Givón puts it.
16 Although verbs beginning with a long vowel or diphthong pose similar problems of accen-
tuation. An example is E. IT 1395 ὠθεῖ Kirchhoff: ὤθει L, see also notes 18 and 27.
 108 chapter 5

but also slightly less innocent measures like changing the accent, or the verb
ending, in the cases I am going to discuss below.
The most extensive modern treatments are, in chronological order: Lauten-
sach (1899: 165–174; ‘Fehlen des syllabischen Augments’), Page’s note in his com-
mentary from 1938 on Medea 1141 (strangely enough, Page does not mention
Lautensach), and Bergson (1953: 121–128; ‘The omitted augment in the messen-
gers’ speeches of Greek tragedy’). Of these, Bergson’s article is of no use for my
purpose, for his objective is to show that the omission of the augment is not
an epicism, but a matter of metrical convenience. He does not question the
136 reliability of the lists of instances used by Lautensach and | Page. In fact, he
mentions forms like κυρει only in passing on p. 124. So I can confine myself to
Page and Lautensach.
Page writes: ‘The exx. in Messengers’ speeches are: A. Pers. 313 πέσον, 376 τρο-
ποῦτο, 416 παίοντο, 458 κυκλοῦντο, 506 πῖπτον : S. El. 715 φορεῖτο, 716 φείδοντο : OT.
1249 γοᾶτο : Tr. 904 βρυχᾶτο, 915 φρούρουν : OK. 1606 κτύπησε, 1607 ῥίγησαν, 1624
θώυξεν : Eur. Ba. 767 νίψαντο, 1066 κυκλοῦτο, 1084 σίγησε, 1134 γυμνοῦτο.’ Lauten-
sach has the same list, plus A. Ch. 738 θέτο and E. fr. 495.36 δῦνον. These lists,
however, present a fundamental problem, for they are based upon what may
be called a modern vulgate (incidentally, Lautensach does not mention which
text he used), where a number of unaugmented imperfects either have been
removed per coniecturam or remain hidden in the apparatus criticus. In fact,
Page himself (still on Med. 1141) mentions a number of such cases, which he
calls ‘dubious or corrupt exx.’, viz. S. Ph. 371; E. Med. 1141 (+1207); Hec. 580 and
1153; Ion 1205, and admits that in the case of S. Ph. 371 (κυρει), E. Med. 1141 (κυνει)
and 1207 (κυνει) ‘it remains doubtful whether the verbs should be accented as
imperfects or as “historic presents.”’ The last three will reappear in my own,
corrected, list below. A more comprehensive list should contain at least the fol-
lowing possible instances,17 where the imperfect has some, overwhelming or
137 even exclusive manuscript support.18 For Sophocles these are: |

17 I confine myself to forms where a change of accent or verb ending changes the interpre-
tation. Forms like λέγον (E. Hec. 580) and θάκουν (Hec. 1153) are therefore not included.
Incidentally, once the idea has been given up that unaugmented past tenses should be
removed as much as possible from our texts, both λέγον and θάκουν, that are virtually
always corrected, are perfectly acceptable forms.
18 And this is not all, for it is worth noticing that cases like S. Ant. 407 ἥκομεν, 432 ἱέμεσθα
and 433 θηρώμεθ᾽; E. Andr. 1159 κομίζομεν; IT 330 χειρούμεθα, 334 κομίζομεν, that all occur in
messenger(-like) speeches and in principle may be taken (without any change of accent)
either as (unaugmented) imperfects or as present tense forms are virtually never men-
tioned as possible instances of imperfects, wrongly, I think. An exception is Kamerbeek,
who discusses the problem involved in his extensive note on S. Ant. 432 (the problem is
on false historic presents in sophocles (and euripides) 109

5 A List of Historic Presents in Sophocles That May Actually Be

Imperfects; Text and Apparatus Lloyd-Jones & Wilson, with
Additions

(12) πύλας δ᾽, ὅπως εἰσῆλθ᾽, ἐπιρράξασ᾽ ἔσω,

καλεῖ τὸν ἢδη Λάϊον πάλαι νεκρόν,
μνήμην παλαιῶν σπερμάτων ἔχουσ᾽, …
γοᾶτο δ᾽ εὐνάς, … (S. OT 1245–1248; messenger speaking)

once within the chamber, she dashed the doors together at her back; then
called on the name of Laius, long since a corpse, mindful of that son …,
and she bewailed the wedlock …

1245 καλεῖ Ambros. G 56 sup., coni. Erfurdt [etiam Pearson, Jebb, Dawe (apud
quem nihil in app. invenitur), Ellendt-Genthe (‘recte sic Erf. : libri male κάλει’)]
: κάλει codd. [Brunck (᾽κάλει), Wolf-Bellermann, Campbell, Kuiper, Dain-Mazon,
Kamerbeek]

(13) βοῶν γὰρ εἰσέπαισεν Οἰδίπους, ὑφ᾽ οὗ

οὐκ ἦν τὸ κείνης ἐκθεάσασθαι κακόν,
ἀλλ᾽ εἰς ἐκεῖνον περιπολοῦντ᾽ ἐλεύσσομεν.
φοιτᾷ γὰρ ἡμᾶς ἔγχος ἐξαιτῶν πορεῖν (S. OT 1252–1255; messenger speaking)

For crying out loud Oedipus burst in, and suffered us not to watch her woe
unto the end; on him, as he rushed around, our eyes were set. To and fro
he went, asking us to give him a sword … And, in his frenzy, a power above
man was his guide

largely ignored by Jebb, and wholly by Campbell and Griffith). But I will not pursue this
issue further here; the whole subject of unaugmented forms in drama deserves a fresh
study à la Lautensach.—As for ‘manuscript support’ a word of caution is perhaps in order
here. While our classical texts go back, in one form or another, to their respective authors,
this is not so with the accents, for these are due to the editorial activity of Byzantine
scholars and copyists. Cf. S-D 1, 373: ‘Systematisch ist die Akzentuation erst seit dem 9.
und 10. Jahrhundert durchgeführt’; cf. also Reynolds & Wilson (1991: 4). So if all MSS read,
say, κύρει, this means that the Byzantine copyists agreed upon this accentuation. Perhaps
these copyists occasionally continued the editorial practice of scholars from antiquity, but
in any case the accents cannot go back further than to the time of Alexandrinian schol-
arship. All this entails that the accents in our MSS have less authority than the words in
these MSS. They should not be ignored, of course, but ultimately the putting of accents
on ambiguous forms is a matter of interpretation.
 110 chapter 5

1255 φοιτᾷ Lrpat [etiam Brunck, Campbell, Jebb, Kuiper, Pearson, Dawe, Ellendt-
Genthe, Kamerbeek, Moorhouse] : φοίτα p, coni. Blaydes [Roussel, Dain-Ma-
zon]

(14) ὅπως δὲ σεμνῶν ὀργίων ἐδαίετο

φλὸξ αἱματηρὰ κἀπὸ πιείρας δρυός
ἱδρὼς ἀνῄει χρωτὶ καὶ προσπτύσσεται
πλευραῖσιν ἀρτίκολλος, ὥστε τέκτονος,
χιτὼν … (S. Tr. 765–769; Hyllus speaking)

But when the blood-fed flame began to blaze from the holy offerings and
from the resinous pine, a sweat broke forth upon his flesh and the tunic
138 clung to his sides … close-glued, as if by a craftsman’s hand |

767 προσπτύσσεται Musgrave [etiam Ellendt-Genthe (‘libri προσπτύσσετο contra

tragicorum usum’) Jebb, Pearson, Dawe] : -ετο codd. [Hermann, Campbell, Dain-
Mazon, Kamerbeek]

(15) ἐπεὶ δ᾽ ἀπεῖπε, …

τότ᾽ ἐκ προσέδρου λιγνύος διάστροφον
ὀφθαλμὸν ἄρας εἶδέ μ᾽ ἐν πολλῷ στρατῷ
δακρυρροῦντα, καί με προσβλέψας καλεῖ·
“ὦ παῖ, πρόσελθε, …”, (S. Tr. 789, 794–797; Hyllus speaking)

But when he gave over … then, from out of the shrouding altar-smoke,
he lifted up his wildly-rolling eyes, and saw me in the in the great crowd,
weeping. He turned his gaze on me, and called me: “O son, draw near …”.
That command sufficed …

796 καλεῖ H. Stephanus : κάλει codd.

(16) κἀγὼ ᾽κδακρύσας εὐθὺς ἐξανίσταμαι

ὀργῇ βαρείᾳ, καὶ καταλγήσας λέγω·
“ὦ σχέτλι᾽, …” |
ὁ δ᾽ εἶπ᾽ Ὀδυσσεύς, πλησίον γὰρ ὢν κυρεῖ· (S. Ph. 368–371; Neoptolemus
speaking)

The tears came into my eyes,—I sprang up in passionate anger,—and said

in my bitterness: “Wretch …”. Then said Odysseus, for he chanced to be
near
on false historic presents in sophocles (and euripides) 111

371 κυρεῖ Porson [etiam Ellendt-Genthe, Jebb, Pearson, Dain-Mazon, Webster,

Dawe, Kamerbeek, Ussher] : κύρει codd. [Campbell : ἦν κυρῶν Brunck, Her-
mann]

(17) ἦν μὲν σιωπή, φθέγμα δ᾽ ἐξαίφνης τινὸς

θώϋξεν αὐτόν, ὥστε πάντας ὀρθίας
στῆσαι φόβῳ δείσαντας ἐξαίφνης τρίχας.
καλεῖ γὰρ αὐτὸν πολλὰ πολλαχῇ θεός.
“ὦ οὗτος οὗτος, Οἰδίπους, …”
ὁ δ᾽ ὡς ἐπήσθετ᾽ ἐκ θεοῦ καλούμενος,
αὐδᾷ … (S. OC 1624–1630; messenger speaking)

There was a stillness; and suddenly a voice of someone cried aloud to him,
so that the hair of all stood up on their heads for sudden fear, and they
were afraid. For the god called him with many callings and manifold: “You
there, Oedipus, …”. But when he perceived that he was called of the god,
he craved that …

1626 καλεῖ] κάλει t (= Triclinius) | 139

5.1 Discussion
All in all this gives us a maximum of six new instances of unaugmented imper-
fects from Sophocles. I will argue that of these instances at least four should
indeed be analysed in this way: (12), (13), (14), and (16).

Example (12), Oedipus Tyrannus 1245

Campbell, who reads κάλει, apparently felt a need to defend his choice, for he
writes: ‘The imperfect κάλει without augment […] is retained from the MSS,
and agrees with ἵετο and γοᾶτο better than καλεῖ […] For καλεῖ, which may be
thought more vivid, cf. O.C. 1630 [= 1655 AR].’ I will turn to the latter example
shortly. The commentators who prefer καλεῖ do so without Campbell’s scru-
ples. Possibly they considered it, in fact, more vivid, but I will not fall into
the trap of speculating about the arguments they may have had. But what
about decisiveness here? This interpretation should, I think, be rejected, unless
one is prepared to accept that calling upon the dead Laius may be a decisive
event in Jocasta’s life just before her own death.19 I have little doubt that the
manuscripts’ κάλει should be retained, along the lines set out by Campbell, and

19 If καλεῖ is considered ‘vivid’ I am of course defenceless.

 112 chapter 5

that it has iterative meaning, just like γοᾶτο.20 In fact, it adds to the pathos of
the scene if Jocasta is described as shouting Laius’ name repeatedly.

(13), Oedipus Tyrannus 1255

With some hesitation I prefer the imperfect, not because of γάρ, for this could
be considered a case of γάρ introducing an embedded narrative, which in that
case would indicate what Oedipus’ περιπολεῖν consisted of, but because of the
semantics of φοιτᾶν, which has frequentative meaning and is therefore an atelic
verb. If the historic present expresses ‘decisiveness’, as I argued above, φοιτᾷ
would express a frequentative decisive event, which is almost a contradiction
in terms. Be that as it may, I have found no parallels for such a use of the present
tense of a frequentative verb.

(14), Trachiniae 767

Here, we are not dealing with a change of accent, but of verb ending. With the
140 majority of the editors I fail to see why προσπτύσσετο | should be changed to
προσπτύσσεται. As appears from the note which accompanies his correction
(‘[προσπτύσσετο] offensioni est ob augmenti absentiam’), Musgrave has found
fault with the imperfect simply because of the omitted augment, but this is
hardly a sufficient reason. The imperfect should, then, be retained, and is in
line with the other imperfects here: ἐδαίετο in 765 and ἀνῄει in 767.

(15), Trachiniae 796

The fact that the ‘calling’ is followed by direct speech may favour the historic
present reading. A parallel is Ant. 1227, and probably OC 1626, although in the
latter case there is a variant κάλει (see ex. (17)). In fact, speeches occurring in
messenger speeches are frequently introduced by verbs of saying in the historic
present, e.g. λέγει, βοᾷ, αὐδᾷ, αἰτεῖ, ἀντιφωνεῖ21 and notably ἐννέπει (of which,
however, no imperfect or aorist exists). Sometimes a decisive event is involved,
but more generally the combination of historic present and quoted speech may
add to the persuasiveness of the messenger’s report: the messenger speaks not
only as an eyewitness but also as an ‘earwitness’. There are no instances where
the imperfect of καλέω or the other verbs mentioned above is followed by direct
speech, although with still other verbs this is not uncommon, cf. E. Hel. 1584
ηὔχετο, Heracl. 825 παρήγγελλε. On balance, Stephanus’ conjecture is to be pre-
ferred.

20 But unlike ἵετο, which rather has immediative meaning.

21 S. Aj. 773; ἀντιφωνεῖ Lrpat : ἀντεφώνει p.
on false historic presents in sophocles (and euripides) 113

(16), Philoctetes 371

I will discuss this example, mentioned by Page among his ‘dubious or cor-
rupt exx.’, in more detail, because on a small scale it gives us a fascinating
insight into the way our profession works, or perhaps I should say, sometimes
works.
Let me say for starters that Porson’s (if it is really Porson’s, see below) κυρεῖ is
definitely impossible, for the following reasons: (i) the sentence contains regu-
lar explanatory or ‘background’ γάρ; in this use γάρ does not combine with the
historic present, see above on examples (8)–(10); (ii) with κυρεῖ we get a stative
historic present, viz. κυρεῖ ὤν—a use for which there are no parallels, cf. item
IV above, p. [130]. | 141
Now interestingly and reassuringly, but also slightly depressingly, Gottfried
Hermann followed already roughly the same line of reasoning, for he rejects,
in his edition of the Philoctetes of 1824 (18332), the present tense with the fol-
lowing remark: ‘Nam praesentis in narrando usus quum ab alacritate quadam
loquentis repetendus sit, convenit primariis partibus narrationis, non etiam iis,
quae obiter et explicandi caussa inseruntur. Itaque nec Latine dicas, tum dicit
Vlysses, praesens est enim, nec, si alia lingua utare: ut hoc non iniuria etiam
a Graeca lingua alienum putari oporteat.’22 With this remark, except for the
deplorable use of ‘alacritas’, Hermann hits the nail on the head. For as I argued
in connection with γάρ, the historic present does, indeed, not befit ‘quae obiter
et explicandi caussa inseruntur’. As for the second half of Hermann’s remark, if
κυρεῖ ὤν is replaced by ἐστι, it is immediately clear that the resulting sentence
is, in fact, impossible: ὁ δ᾽ εἶπ᾽ Ὁδυσσεύς, πλησίον γάρ ἐστιν.23
Jebb, however, was not impressed: ‘Hermann objected to the historic present
as unsuitable to a parenthetic remark; but without cause’. To support this note
Jebb refers to his own note on Ant. 253 (‘historic pres. combined with past tense;
cf. Lys. Or. 1 §6 […]’), and to E. Hec. 963. But all three passages are very different
from Philoctetes 371. The Antigone passage runs:

22 ‘As the use of the present in narrating must derive from a certain liveliness of the speaker, it
befits the main parts of the narrative rather than those passages that are inserted obliquely
and for the sake of explaining. Thus, you would not say in Latin “[…]” nor in other lan-
guages. So we may safely assume that Greek, too, did not know this use.’
23 Before Hermann, Brunck had already observed: ‘Sane res ipsa manifesto tempus imperfec-
tum postulat: verum Attici poëtae rarissime augmentum omittunt, nec erat hic licentiae
locus’. Unfortunately Brunck does not tell us why in this case the ‘licentia’ does not apply.
Anyhow, he replaced ὢν κύρει, by ἦν κυρῶν which was adopted, alas, by Hermann. Appar-
ently, Hermann shared Brunck’s view of the augment here.
 114 chapter 5

(18) ὅπως δ᾽ ὁ πρῶτος ἡμὶν ἡμεποροσκόπος

δείκνυσι, πᾶσι θαῦμα δυσχερὲς παρῆν·
ὁ μὲν γὰρ ἡφάνιστο … (S. Ant. 253–255)

And when the first day-watchman showed it to us, sore wonder fell on all.
The dead man was veiled from us …

And that in Lysias:

(19) ἐπειδὴ δὲ μοι παιδίον γίγνεται, ἐπίστευον ἤδη … (Lys. 1.6)

142 But when a child was born to me, thenceforward I began to trust her … |

There are simply no parentheses here, only temporal clauses.

As for the Hecuba, this is not convincing either, if only because there is a
variant τύγχανον:

(20) σὺ δ᾽, εἴ τι μέμφῃ τῆς ἐμῆς ἀπουσίας,

σχές· τυγχάνω γὰρ ἐν μέσοις Θρῄκης ὅροις
ἀπών, ὅτ᾽ ἦλθες δεῦρ᾽· ἐπεὶ δ᾽ ἀφικόμην,
… | ἐς ταὐτὸν ἥδε συμπίτνει δμωὶς σέθεν (E. Hec. 962–965)

As for you, if you find fault with my absence, check the thought. It hap-
pens that I was away in the inland regions of Thrace when you arrived here

963 τυγχάνω Ω̣ ξΖcΖm et GSΤZC : τύγχανον vel ᾽τ- GPPaZ et B3Rγρ : ἐτύγχανον FVaTZ
et B3RSSaS

But even if the present tense is correct, this sentence is not a parenthesis either.
Nor are we dealing with a narrative or a historic present. If correct, τυγχάνω …
ἀπών should be taken as an instance of the use discussed by K-G (1, 200 Anm.
9): ‘So von vergangenen Handlungen [the present participle] auch nach präsen-
tischem Hauptverbum’. Kovacs correctly translates ‘It happens that I was away
[…]’. This is the same use of the participle which we find in e.g.

(21) φαίνομαι οὖν τρισκαιδεκέτης ὢν ὅτε ὁ πατὴρ ὑπὸ τῶν τριάκοντα ἀπέθνῃσκε.
(Lys. 10.4)

It is clear, therefore, that I was thirteen when my father was put to death
by the Thirty.
on false historic presents in sophocles (and euripides) 115

So much for Jebb’s rejoinder to Hermann’s objections. I should also add that
one dearly misses a positive appraisal on Jebb’s part of the present tense form.
What is its effect? What do we gain by reading the present? Why is it better than
the reading of the MSS? As for the other commentators, Campbell—who is vir-
tually the only one who resisted Porson’s (and Elmsley’s, see below) authority,
for he prints the imperfect—notes, in his Paralipomena Sophoclea (1907): ‘The
historic present would not be amiss, but there hardly seems sufficient cause
for departing from the traditional text’. While the first part of this statement is
not correct, the second is, of course, entirely to the point. Again, others were
not impressed. Webster (1970) writes: ‘the present has to be explained as a his-
toric or vivid present; the MSS κύρει is the imperfect without augment; this is
rare in tragedy and usually the word begins the line, but it cannot be ruled
out here.’ | Not only Jebb, Kamerbeek, too, I am afraid, had an off-day when 143
he wrote his note. While he accepts κυρεῖ and takes it as a historic present, he
adds: ‘If we follow the MSS in reading κύρει, the form without augment may be
defended on the ground that it occurs in a narrative comparable to a messen-
ger’s story, where such an epicism is not rare’. So far so good. But he then goes on
with: ‘Besides he could have used κύρει (present), for he has ἔκυρον (imperfect)
O.C. 1159’. Unfortunately this is metrically impossible, for κύρω (a very rare by-
form of κυρέω) has a long -υ-. Perhaps he was misled by Ellendt-Genthe, who
did not have their day either, for they print κῠ́ρω. Ussher (1990) has no note at
all.
Finally, a few words about the origin of κυρεῖ—we have not yet reached the
end of the off-days. In all apparatuses this form is ascribed to Porson. But where
did Porson propose this reading? The first part of Hermann’s note sends us
to Porson’s note on Medea 1141, for Hermann writes: ‘Porson ad Med. 1138 [=
1141] κυρεῖ quem sequitur Elmsleius ad Med. 1110’. Naturally, from this remark
we must infer that κυρεῖ was a conjecture by Porson. Porson himself, however,
believed that κυρεῖ was the reading of the manuscripts, for he writes24 at Medea
1138 [= 1141]: ‘Hic et 1204. [= 1207] κύνει habent Lasc. Ald.25 solenni errore, quem
correxit Brunckius [viz. into κυρεῖ, in his edition of the Medea of 1793, see above
example (2)—AR]; quocirca magis eum miror in locis similibus tantas turbas
interdum ciere. Sophocl. Philoctet. 371. πλησίον γὰρ ὢν κυρεῖ [sic—AR] sine ulla
causa in πλησίον γὰρ ἦν κυρῶν mutavit.’26 In other words, Porson here castigates

24 Ed. of 1812, originally published in 1801, and also in those of 1821 and 1825.
25 The edition of Lascaris (1494) and the Aldina (1503).
26 ‘Here and at 1204 Lasc. Ald. have κύνει, by a common error, which Brunck has corrected; I
am therefore amazed that in similar cases he sometimes makes such a fuss.’ Brunck had
published his Sophocles edition in 1786.
 116 chapter 5

Brunck for rejecting a present tense form of the MSS, while elsewhere he intro-
duced such present tense forms himself in our texts.
As for Elmsley, mentioned in Hermann’s note, he writes (1818, on Medea 1110
= 1141): ‘Hic et v. 1176 [= 1210] κυρεῖ reposuit Brunckius’, followed by a quotation
of Porson’s words ‘quocirca […]’ etc. Elmsley then makes things worse, for he
rephrases Porson’s words as follows: ‘Apud Sophoclem scilicet Phil. 371 πλησίον
144 γὰρ ὢν κύρει | [sic] in πλησίον γὰρ ἦν κυρῶν mutare maluit (viz. Brunck), quam
praesens κυρεῖ [sic] praeteriti imperfecti sensu admittere’. This is of course
quite a mess, and apart from everything else one wonders whether Elmsley
really believed that the present could have the meaning of an imperfect.
All in all, it looks as if everybody believed that Porson proposed the conjec-
ture κυρεῖ for Philoctetes 371, while Porson himself thought that κυρεῖ was in the
manuscripts. How should this comedy of errors be represented in an accurate
apparatus? Perhaps as follows: ‘κυρεῖ Porsono attribuunt edd.; is autem falso
hanc formam pro lectione libr. habuit’. In the last resort we must conclude that
κυρεῖ is a ghost-form, since nobody actually proposed to read it.

5.2 Discussion, Continued

(17), Oedipus Coloneus 1626
For καλεῖ and κάλει (which may well be a conjecture, of course) followed by
direct speech see above, on Tr. 796, example (15). What about the effect here
of γάρ and πολλὰ πολλαχῇ? As for γάρ, this is, I think, an instance of γάρ intro-
ducing an embedded narrative, for it explains what the φθέγμα mentioned in
line 1623 consisted of. On this count, then, both καλεῖ and κάλει are possible.
As for πολλὰ πολλαχῇ, in his note Jebb slightly varies on his translation, for he
writes: ‘with repeated and manifold calling’; Lloyd-Jones, too, has an iterative
translation: ‘For the god called him often and from many places’. In any event,
if this is the meaning of these two adverbs, this strongly suggests that indeed
a repeated event is meant, which would plead in favour of the imperfect. On
the other hand, the fact that καλει is followed by direct speech and takes up
ἐξαίφνης θώϋξεν suggests rather that there is no repetition involved. All in all,
I believe that the latter is correct. Πολλά should then be taken as a manner
adverb, and πολλαχῇ as a locative adverb: ‘loudly and by many ways’ (i.e. ‘from
145 many sides’). |
on false historic presents in sophocles (and euripides) 117

6 Historic Presents in Euripides That May Be Imperfects

For reasons of space I confine myself to discussing Medea 1141, which got
already a brief mention above (ex. (2)), and Bacchae 728:27

(22) … δι᾽ ὤτων δ᾽ εὐθὺς ἦν πολὺς λόγος

σὲ καὶ πόσιν σὸν νεῖκος ἐσπεῖσθαι τὸ πρίν.
κυνεῖ δ᾽ ὁ μέν τις χεῖρ᾽, ὁ δὲ ξανθὸν κάρα
παίδων· (E. Med. 1139–1142; messenger speaking)

For our ears buzzed with the loud report that you and your husband had
brought your former quarrel to and end. And someone kissed the hands
and another the blond heads of the children.

1141 κυνεῖ Brunck : κύνει codd. (‘fortasse recte. Cf. v. 1207’—Murray, Page)

(23) αἳ δὲ τὴν τεταγμένην

ὥραν ἐκίνουν θύρσον ἐς βακχεύματα,

27 Other cases in Euripides where edd. prefer a historic present to an imperfect are:
(i) Alc. 183–184 κυνεῖ and δεύεται (κυνεῖ BOV : κύνει LP; δεύεται BOV et gB; δεύετο P : aut
-αι in -ο aut -ο in -αι mutavit L1e);
(ii) Med. 1207 κυνεῖ (κυνεῖ V et iΣh : κύνει Ω̣ LP);
(iii) IT 1395 ὠθεῖ (ὠθεῖ Kirchhoff : ὤθει L).
I add a few remarks. Med. 1207 is like 1141, and there is no reason to treat it differently,
so the imperfect should be preferred. At IT 1395 Kirchhoff’s conjecture should be rejected,
for it occurs in a regular ‘background’ γάρ-clause, where the historic present is impossible,
see above p. [133]. ὤθει is simultaneous with the preceding imperfect ἠπείγετο: the blow-
ing of the wind and the speed of the ship coincide. Cf. also Platnauer ad loc.: ‘I see little
reason to suspect the imperf.’. As for Alc. 183–184, δεύετο is generally rejected for metrical
reasons, cf. e.g Dale ad loc.: ‘Since δεύετο here is prosodically impossible we can justifiably
remove δάκρυσε 176 and κύνει 183’. ‘Impossible’ is perhaps just a bit too strong, for there
are some cases where muta cum liquida in initial position makes position, cf. especially
Barrett on E. Hipp. 760. Moreover, the situation here is similar to that of the unaugmented
imperfects: some instances may be hidden in the apparatus, or have been removed per
coniecturam. To the possible examples at least one should be added, IA 1579 ἵνᾱ πλήξειεν
ἄν. Also, the alternative forms κυνεῖ and δεύεται at Alc. 183–184 are not without problems
either, at least if marking decisive events is considered the basic function of the historic
present (the proponents of ‘vividness’ will of course haven no qualms). That δεύετο may
be the correct reading finds some support in line 185 ἐπεὶ δὲ πολλῶν δακρύων εἶχεν κόρον:
this clause seems more appropriate after an imperfect than after a historic present. For
κυνεῖ see also below on Med. 1141.
 118 chapter 5

… πᾶν δὲ συνεβάκχευ᾽ ὄρος

καὶ θῆρες, οὐδὲν δ᾽ ἦν ἀκίνητον δρόμῳ.
κυρεῖ δ᾽ ᾽Αγαύη πλησίον θρῴσκουσά μου·
κἀγὼ ᾽ξεπήδησ᾽ ὡς συναρπάσαι θέλων (E. Ba. 723–729; messenger speak-
146 ing) |

And the women at the appointed time of day began to wave their thyr-
soi and to worship Dionysus … the whole mountain with its beasts was as
possessed as they were, and everything was set in rapid motion. Agave’s
leaping happened to bring her near me, and I left my hiding-place and
jumped up to seize her

728 κυρεῖ P2 : κύρει LP

6.1 Discussion
Example (22), Medea 1141
To explain his text, that, as far as I could see, has been adopted by all subse-
quent editors, Brunck provides just the following remark:28 ‘κυνεῖ hic et v. 1207
praesens, non imperfectum. Sic in narrationibus variari tempora sciunt, qui
paullo attentius Tragicos legerunt.’29 To be sure, in the second sentence Brunck
ingeniously combines an appeal to variatio as a critical concept with an argu-
mentum ex auctoritate, but if this is all that can be put forward against the
reading of the manuscripts, that reading should perhaps not be given up too
easily, the less so because κυνεῖ introduces a very inappropriate historic present
(see above on ex. (2)). In fact, κύνει makes perfect sense, and has iterative mean-
ing: it is not simply ‘they kissed’, but ‘they covered with kisses’, simultaneously
with ἦν πολὺς λόγος.

Example (23), Bacchae 728

This example much resembles Sophocles, Philoctetes 371, discussed above,
example (16), since it combines a form of κυρέω with a present participle and
the same locative adverb, πλησίον. As in the example from Medea (ex. (22)), all
editors seem to prefer the present indicative accentuation, κυρεῖ, which is due
to the corrector of P. In Kovacs’ translation, κυρεῖ δ᾽ ᾽Αγαύη πλησίον θρῴσκουσά
μου expresses the idea that Agaue, while jumping, ended up in the immedi-

28 In the Leipzig edition of 1794.

29 ‘That the tenses show such variation in narratives is well-known to people who have read
the tragedians just a bit more attentively.’
on false historic presents in sophocles (and euripides) 119

ate vicinity of the messenger. Likewise the translations of Roux (‘Agavé passe à
ma portée, bondissante’) and Grégoire (‘or voici qu’ Agavé bondit à ma portée’),
and probably also Kirk (‘Agaue chances to jump close by me’). Compare also the
translations of Willink and Kovacs in another case of πλησίον θρῴσκειν, viz. at
Orestes 257: αὗται γὰρ αὗται πλησίον θρῴσκουσί | μου: ‘[…] here they come galop- 147
ping near me’ (Willink), ‘Here they come leaping toward me!’ (Kovacs). But I do
not think these translations are correct, since they all treat θρῴσκω as some verb
of ‘moving toward’. In reality, ‘bring’ or ‘come’ is not part of the meaning of the
present θρῴσκειν.30 Both at Bacchae 728 and Orestes 257 the meaning is rather
‘jump to and fro’, ‘move, jump around’. In other words, θρῴσκω does not have
telic, but durative-stative meaning (cf. also φοιτᾶν above, example (13)). Just as
in the case of Sophocles, Philoctetes 371, the historic present would clash with a
durative-stative meaning,31 so the imperfect must be preferred: ‘she happened
to be jumping near me’. Κύρει … θρῴσκουσα continues the imperfects ἐκίνουν …
συνεβάκχευ᾽ … οὐδὲν ἦν ἀκίνητον, which seems unobjectionable.

7 General Conclusions

The results of the above discussion are summarized in the following table,
where the text of the OCT editions of Sophocles and Euripides should be taken
as representative of most modern editions:32

Manuscripts OCT This article

S. OT 780 (ex. (7)) καλεῖ καλεῖ καλεῖ

OT 1245 (ex. (12)) κάλει καλεῖ κάλει
OT 1255 (ex. (13)) φοιτᾷ majority of the φοιτᾷ φοίτα
MSS; φοίτα minority
Tr. 767 (ex. (14)) προσπτύσσετο προσπτύσσεται προσπτύσσετο
Tr. 796 (ex. (15)) κάλει καλεῖ καλεῖ

30 Θρῴσκω may have this meaning all right, but only in the aorist, see LSJ s.v. (who wrongly
classify Or. 257 and Ba. 728 under the heading ‘leap upon, attack’).
31 There is one other instance of κυρεῖ in a Euripidean messenger-speech, viz. at El. 777 κυρεῖ
… βεβώς. Since here, too, there is a clash between the stative value of this phrase and the
historic present I am tempted to read κύρει.
32 Observe, however, that Dain-Mazon, in their Budé-edition, read an imperfect at OT 1245,
1255 and Tr. 767.
 120 chapter 5

(cont.)

Manuscripts OCT This article

Ph. 371 (ex. (16)) κύρει κυρεῖ κύρει

OC 1626 (ex. (17)) καλεῖ καλεῖ καλεῖ
E. Med. 1141 (ex. (22)) κύνει κυνεῖ κύνει
Med. 1207 (note 27, ex. (ii)) κύνει κυνεῖ κύνει
IT 1395 (note 27, ex. (iii)) ὤθει ὠθεῖ ὤθει
Ba. 728 (ex. (23)) κύρει κυρεῖ κύρει |

148 What other conclusions can be drawn from the above observations? First, that
a linguistic framework may be of some help in interpreting the elusive verb
forms of our Greek texts. Second, that the number of historic presents in Sopho-
cles and Euripides goes slightly down while that of imperfects goes slightly up.
Third, that the phenomenon of unaugmented past tenses is more widespread
than is generally assumed, and should no longer be considered some kind of
aberration. Fourth, that an apparatus criticus should be as detailed as possible.
Fifth, that consulting our predecessors of at least the last two centuries is indis-
pensable to form a balanced opinion, although, sixth, one should always reckon
with one or more, or even collective, off-days. Seventh, that the consultation of
these predecessors should perhaps start with Gottfried Hermann.

Appendix: Historic and Actual Presents

There comes a moment, of course, when the messenger has to put an end to his
report. This is in general indicated by a reference to the actual situation, which
in turn may be followed by a concluding evaluation; actuality and evaluation
may also coincide. Not unexpectedly, the reference to the actuality is expressed
by present tense forms. But how can we be certain that these presents are not
historic presents? Let us look paullo attentius at example (a) and at some trans-
lations of line 1240 and especially of κεῖται:

(a) ἐς δ᾽ ὑγρὸν
ἀγκῶν᾽ ἔτ᾽ ἔμφρων παρθένῳ προσπτύσσεται
καὶ φυσιῶν ὀξεῖαν ἐκβάλλει ῥοὴν
λευκῇ παρειᾷ φοινίου σταλάγματος.
κεῖται δὲ νεκρὸς περὶ νεκρῷ, … (S. Ant. 1236–1240; messenger speaking)
on false historic presents in sophocles (and euripides) 121

Lloyd-Jones: ‘Still living, he clasped the maiden in the bend of his feeble
arm, and shooting forth a sharp jet of blood, he stained her white cheek.
He lay, a corpse holding a corpse […]’

Jebb: ‘he lies’

Dain-Mazon: ‘Il est là […]’.

Observe that Lloyd-Jones translates κεῖται with a past tense, and we must there-
fore assume that he takes κεῖται as another instance of a historic present, after
προσπτύσσεται and ἐκβάλλει. But this is impossible, if only because, as I noticed
earlier, stative κεῖμαι does not | qualify for this use. Also, with this translation 149
there is properly speaking no formal transition to the actual situation. That
Lloyd-Jones’ choice of a past tense is not an incident appears from example
(b), and from the similar example (c):

(b) τοιόνδε κεῖται προπετές (S. Tr. 701)

Lloyd-Jones: ‘So there it lay’ (sc. the piece of wool)

Contrast Jebb: ‘In such a state it lies as it fell’ and Dain-Mazon: ‘Il est là
[…]’

(c) καί νιν πυρᾷ κέαντες εὐθὺς ἐν βραχεῖ

χαλκῷ μέγιστον σῶμα δειλαίας σποδοῦ
φέρουσιν ἄνδρες Φωκέων τεταγμένοι,
ὅπως πατρῴας τύμβον ἐκλάχῃ (-οι cett.) χθονός.
τοιαῦτά σοι ταῦτ᾽ ἐστίν, (S. El. 757–761)

Lloyd-Jones: ‘at once carried […] his mighty form, so that he should be
accorded burial […]’

Contrast again Jebb: ‘are bringing him’ and Dain-Mazon: ‘les Phocidiens
ont été délégués pour vous l’apporter’

Why Lloyd-Jones in all three cases has chosen to translate in this way is quite a
riddle. Even more puzzling is that he translates ἐστίν at Electra 761 (ex. (c)) with
a past tense: ‘Such was this event’, instead of ‘Such is the situation’. Observe that
these choices are not simply a matter of stylistic preference; in (c), for instance,
the translation with ‘carried’ brings about a radical change in the dramatic situ-
122 chapter 5

ation, for the use of this past tense implies that the ashes of ‘Orestes’ are already
on the soil of his fathers, i.e. Mycenae (cf. l. 760), which they are not.33 Jebb,
Mazon and others correctly take κεῖται etc. as actual presents.34

References

Benveniste, E., ‘Le langage et l’expérience humaine’, in Problèmes de linguistique gén-

érale, 2, (Paris 1974) 67–78.
Bergson, L., ‘The Omitted Augment in the Messengers’ Speeches of Greek Tragedy’, Era-
nos 51 (1953) 121–128.
Campbell, L., Paralipomena Sophoclea (London 1907).
Eriksson, K.V., Das Präsens historicum in der nachklassischen griechischen Historiogra-
phie (Lund 1943).
Givón, T., Syntax, 2 vols. (Amsterdam/Philadelphia 1984)
de Jong, I.J.F., Narrative in Drama: The Art of the Euripidean Messenger-Speech (Leiden
1991).
de Jong, I.J.F., ‘Γάρ Introducing Embedded Narratives,’ in A. Rijksbaron (ed.), New
Approaches to Greek Particles (Amsterdam 1997) 175–186.
Kells, J.H. (ed.), Sophocles: Electra (Cambridge 1973).
Kroon, C.H.M., Discourse Particles in Latin: A Study of nam, enim, autem, vero, and at
(Amsterdam 1995).
Lautensach, O., Grammatische Studien zu den griechischen Tragikern und Komikern:
Augment und Reduplikation (Hannover 1899).
Moorhouse, A.C., The Syntax of Sophocles (Leiden 1982).
Page, D.L. (ed.), Euripides: Medea (Oxford 1938).
Reynolds, L.D., Wilson, N.G., Scribes and Scholars, third ed. (Oxford 1991).
Rijksbaron, A., ‘The Xenophon Factory: One Hundred and Fifty Years of School Editions
of Xenophon’s Anabasis’, in Gibson, R.K., Kraus, C.S. (eds), The Classical Commen-
tary: Histories, Practices, Theory (Leiden 2002) 235–267 [ch. 20 in this volume].
Ruipèrez, M.S., Structure du système des aspects et des temps du verbe en grec ancien:
analyse fonctionnelle synchronique (Paris 1982).
Sicking, C.M.J., van Ophuijsen, J.M., Two Studies in Attic Particle Usage Lysias and Plato
(Leiden 1993).

33 The chosen ‘Phocians’ (i.e. Orestes and Pylades) only arrive at the palace at l. 1098; at 1113
we learn that they bring the remains of ‘Orestes’, φέροντες αὐτοῦ σμικρὰ λαίψαν᾽ ἐν βρα-
χεῖ/τεύχει, where φέροντες of course echoes φέρουσιν at 759.
34 For similar actual presents in Euripides see e.g. κεῖνται at Med. 1220 and κομίζομεν at Andr.
1159.
on false historic presents in sophocles (and euripides) 123

Sicking, C.M.J., Stork, P., ‘The Grammar of the So-Called Historical Present in Ancient
Greek,’ in E.J. Bakker (ed.), Grammar as Interpretation: Greek Literature in its Linguis-
tic Contexts (Leiden 1997) 131–168.
Ussher, R.G. (ed., tr.), Sophocles: Philoctetes (Warminster 1990).
Webster, T.B.L. (ed.), Sophocles: Philoctetes (Cambridge 1970).
 chapter 6

Ἀκροᾶσθαι or ἀκροάσασθαι (Plato, Ion 530d9)?

Abstract

This chapter, which was originally published as Appendix III to Rijksbaron’s commen-
tary on Plato’s Ion (2007), poses the question whether, at Ion 530d9, σχολὴν ἀκροάσασθαι
(the reading of S and F, with aorist infinitive) or σχολὴν ἀκροᾶσθαι (the reading of T,
with present infinitive) is the better reading. After a number of general observations
on the dynamic infinitive and the manuscripts of Plato, an overview of all occurrences
of ἀκροᾶσθαι and ἀκροάσασθαι in the Platonic corpus is given, from which it appears
that the former, presenting the ‘listening’ as open-ended, often has an iterative mean-
ing or denotes a general line of conduct (‘lend one’s ear’), whereas the latter presents
the ‘listening’ as an action in its entirety, from beginning to end. It is next shown that
σχολή, both in Plato and elsewhere, is more often followed by a present infinitive than
by an aorist one, and that in the latter case it is always negated. It is concluded that,
although most editors prefer the aorist, the present infinitive is the better reading in
the Ion passage.

The choice between the present infinitive and the aorist infinitive in the
dynamic use, i.e. after verbs of volition, commanding, having time, ability, etc.,
belongs to the most intriguing and elusive parts of Greek syntax, which poses
special problems when the MSS present both forms, as here, which both yield
good Greek.1 In such cases Greek scholars as a rule stand helpless. The problem
is either ignored—massively—or it leads to puzzling remarks like the following
one by Dodds (1959), on Grg. 448a5, where F reads λαβεῖν, and BTW f λαμβάνειν:
‘F’s λαβεῖν is more appropriate than λαμβάνειν, and is confirmed by Olympi-
odorus (18.12 Norvin)’. Why λαβεῖν is more appropriate Dodds does not say.
Burnet preferred λαμβάνειν. And this is understandable enough, for two rea-
sons. First, whether we read λαβεῖν or λαμβάνειν at Grg. 448a5, or ἀκροάσασθαι
or ἀκροᾶσθαι in our case, and in countless similar cases, the interpretation in
terms of denotation is, or at least seems to be, the same. Related to this point is
the fact that in our translations present and aorist infinitives are usually trans-

1 The same holds, incidentally, for the choice between present and aorist imperative, and
present and aorist subjunctive and optative in purpose clauses.

© koninklijke brill nv, leiden, 2019 | doi:10.1163/9789004386129_008

ακροασθαι or ακροασασθαι (plato, ion 530d9)? 125

lated in the same way.2 Second, a generally accepted theory of such differences
was not available in Dodds’ days nor is it now. Yet in recent times some new
light, at least, has been shed on this aspectual distinction in two extensive
and at the same time in-depth studies of this subject, the dissertation by the
Dutch scholar Peter Stork (1982), on the dynamic infinitive in Herodotus, and
Jacquinod (2000), a collection of papers written by a French-Dutch group of
scholars, that for the greater part deal with the use of this infinitive in Plato. In
the ‘Présentation’, the introduction to this book (17), the editors aptly speak of a
‘projet de capturer Protée’. Anyone interested in the actual state of research on
this matter I may refer to this introduction, and to the book as a whole, where
the notions used below and | elsewhere in the present book in connection with 262
aspectual oppositions are all discussed in detail.
Before I return to our passage, I must mention one further complicating
factor, now of a codicological nature. The fact is that if there is manuscript vari-
ation involving the dynamic infinitive, MSS S and F virtually always have the
aorist infinitive, and T always and W mostly the present infinitive. This is not
only the case in the restricted corpus of the Ion, but also in the Gorgias, which
I checked by way of comparison, using Dodds’ edition with its full apparatus
(where, however, of the S F family, for reasons proper to this dialogue, only F is
mentioned, and TW are accompanied by B).3 Consider the following facts:

Ion

530d9 ἀκροάσασθαι S F : ἀκροᾶσθαι T W

533c2–3 συμβαλέσθαι W S F (all three with aorist accent) : συμβάλλεσθαι T
538b8 πιεῖν S F : πίνειν T W

Gorgias

448a5 λαβεῖν F (Olymp.) : λαμβάνειν B T W f

448d7 ἀποκρίνασθαι F : ἀποκρίνεσθαι cett.

2 Nor is this surprising, since after verbs of volition, etc., modern European languages (but also
Latin, for that matter) simply do not have two infinitives that would correspond semantically
to the two Greek infinitives.
3 This phenomenon is also found with non-infinitive forms, e.g. Ion 530c2 συνείη S F Prisc. :
συνιείη W f : συνίη T, 540e1 ἀπεκρίνω S F : ἀπεκρίνου T W, Grg. 511d5 διαπραξαμένη F (Olymp.)
: διαπραττομένη cett. Cp. further e.g. Hp.Mi. 363c5 ἀποκρίνεσθαι T W : ἀποκρίνασθαι S F, Men.
84a6 ἀπεκρίνετο B T W f : ἀπεκρίνατο F, 85d6 ἀναλαμβάνειν] ἀναλαβεῖν F, 87e4 ἀναλαμβάνοντες]
ἀναλαβόντες F. For the, less frequent, reverse situation cp. e.g. Hp.Mi. 367a7 ψεύσαιτο T W :
ψεύδοιτο S F, Men. 72c7 ἀποκρινόμενον W F : ἀποκρινάμενον B T.
 126 chapter 6

493d1 μεταθέσθαι F : μετατίθεσθαι B T W f

513a8 ἀπεργάσασθαι F : ἀπεργάζεσθαι cett.
505d1 καταλιπεῖν P F : καταλείπειν B T W f

Once the reverse is found: Grg. 475d6 ἀποκρίνεσθαι F : ἀποκρίνασθαι cett.

I must confess that I fail to see how this bias towards either the aorist or
the present infinitive, depending on one’s starting point, should be explained.
If ever, it is of course impossible here to establish an ‘original’ reading. There
were apparently different traditions, which for all we know may both go back
263 to ‘Plato’. Or then again one tradition may | have been dominant, the problem,
however, being that we are not in the position to decide which one. I should
mention at least one other phenomenon, however, viz. that in Byzantine Greek
from late antiquity onwards the forms of the aorist stem seem gradually to have
encroached ever more upon those of the present stem.4 As a consequence, the
copyists of the late Byzantine MSS S and F (or those of their immediate prede-
cessors) may have tended unconsciously to replace present stem forms with
aorist forms (although the dynamic infinitives in question had by that time
almost disappeared).
Be that as it may, we still have to decide whether ἀκροάσασθαι or ἀκροᾶσθαι
should be preferred. A fairly thorough investigation both of the dynamic use of
ἀκροάσασθαι and ἀκροᾶσθαι and of that of σχολή + infinitive, in Plato and else-
where, leads to the following picture.5

Ἀκροᾶσθαι

Ἀκροᾶσθαι presents the ‘listening’ as an open-ended (unbounded, atelic)

action, i.e. as an action ‘in course’ (in technical terms: an activity);6 one might
also say that it is ‘process-oriented’. It has frequently iterative (habitual) mean-
ing; in that case the infinitive denotes a general line of conduct. Typical exam-
ples are:

4 See Lallot (2000: 261) on the preponderance of aorist subjunctives as replacers in Modern
Greek of present stem dynamic infinitives in Ancient Greek.
5 ‘Elsewhere’ = Aeschylus, Sophocles, Euripides, Aristophanes, Herodotus, Thucydides, Xen-
ophon, Andocides, Lysias, Isocrates, Demosthenes, Isaeus, Aeschines. Ἀκροάομαι does not
occur in Aeschylus, Sophocles, Euripides, Herodotus and Xenophon. It does occur in Aristo-
phanes, but not in the infinitive.
6 For activity and accomplishment, used below, see Rijksbaron (1989: 17ff.).
ακροασθαι or ακροασασθαι (plato, ion 530d9)? 127

Ly. 205d4 ταῦτ᾽ ἐστὶν ἃ οὗτος λέγων τε καὶ ᾄδων ἀναγκάζει καὶ ἡμᾶς ἀκροᾶσ-
θαι—repeatedly; cp. the generic presents ᾄδει … ποιεῖ τε καὶ λέγει
at 205c2–6.
Grg. 488c2 Πότερον δὲ τὸν αὐτὸν βελτίω καλεῖς σὺ καὶ κρείττω; οὐδὲ γάρ τοι τότε
οἷός τ᾽ ἦ μαθεῖν σου τί ποτε λέγοις, πότερον τούς ἰσχυροτέρους κρείτ-
τους καλεῖς καὶ δεῖ ἀκροᾶσθαι τοῦ ἰσχυροτέρου τοὺς ἀσθενεστέρους—
in general; note δεῖ and the generic articles τούς, τοῦ and τούς. | 264
Th. 6.17.3 καὶ οὐκ εἰκὸς τὸν τοιοῦτον ὅμιλον οὔτε λόγου μιᾷ γνώμῃ ἀκροᾶσθαι οὔτε
ἐς τὰ ἔργα κοινῶς τρέπεσθαι—note generic τὸν τοιοῦτον ὅμιλον.
And. 1.69 Οὑτωσὶ δὲ ἔχει, ὦ ἄνδρες· μέχρι τούτου ἀναβήσονται καὶ λέξουσιν ὑμῖν,
ἕως ἂν ἀκροᾶσθαι βούλησθε, ἔπειτα δ᾽ ἐγὼ περὶ τῶν ἄλλων ἀπολογή-
σομαι—unbounded (durative) ‘listening’ tout court.
Lys. 13.79 ἀνάγκη δὲ ἦν στρατηγοῦ ἀνδρὸς ἀκροᾶσθαι, εἴπερ ἔμελλον σωθήσεσ-
θαι—the necessity to listen (obey) existed in general; note indefi-
nite, generic, στρατηγοῦ ἀνδρός.
Isoc. 15.20 Ὧν χρὴ μεμνημένους μὴ προπετῶς πιστεύειν τοῖς τῶν κατηγόρων λό-
γοις, μηδὲ μετὰ θορύβου καὶ χαλεπότητος ἀκροᾶσθαι τῶν ἀπολογου-
μένων—in general, with implications for the actual situation; note
the generic article in τῶν κατηγόρων and τῶν ἀπολογουμένων.

Also with the lexical variant ποιεῖσθαι τὴν ἀκρόασιν:

Isoc. 15.12 Χρὴ δὲ τοὺς διεξιόντας αὐτὸν πρῶτον μὲν ὡς ὄντος μικτοῦ τοῦ λόγου καὶ
πρὸς ἁπάσας τὰς ὑποθέσεις ταύτας γεγραμμένου ποιεῖσθαι τὴν ἀκρό-
ασιν, ἔπειτα προσέχειν τὸν νοῦν ἔτι μᾶλλον τοῖς λέγεσθαι μέλλουσιν ἢ
τοις ἤδη προειρημένοις, πρός δὲ τούτοις μὴ ζητεῖν εὐθὺς ἐπελθόντας ὅλον
αὐτὸν διελθεῖν, ἀλλὰ τοσοῦτον μέρος ὅσον μὴ λυπήσει τοὺς παρόντας.

Other examples are: Th. 2.21.3, Lys. 12.55, D. 8.23, 9.55, Aeschin. Epp. 7.4.

Ἀκροάσασθαι

Ἀκροάσασθαι, on the other hand, presents the ‘listening’ as a completed

(bounded, telic) action, i.e. as an action in its entirety, from beginning to end
(as an accomplishment). One might also say that it is ‘result-oriented’. It is fre-
quently used by the orators at the end of the exordium, in appeals to the jury
to listen to the speaker to the very end. There may also be another nuance
involved, viz. that the verbal action is considered without any thought of it
being carried out (unlike the present infinitive), and thus in abstracto. (Cp. also
 128 chapter 6

on ἐξηγήσασθαι at 531b7–9.) (See also on 531b8–9.) The aorist infinitive does

not occur in the genuine works of Plato, with the possible exception of our
265 passage. |
Typical examples are:

Lys. 16.9 δέομαι οὖν ὑμῶν μετ᾽ εὐνοίας ἀκροάσασθαί μου (i.e. my defence). ποιή-
σομαι δὲ τὴν ἀπολογίαν ὡς ἂν δύνωμαι διὰ βραχυτάτων.
D. 18.2 Πρῶτον μέν … τοῖς θεοῖς εὔχομαι … τοσαύτην (sc. εὔνοιαν) ὑπάρξαι μοι
παρ᾽ ὑμῶν …, ἔπειθ᾽ …, τοῦτο παραστῆσαι τούς θεούς ὑμῖν, μὴ τὸν ἀντίδι-
κον σύμβουλον ποιήσασθαι … ἀλλὰ τοὺς νόμους καὶ τὸν ὅρκον, ἐν ᾧ … καὶ
τούτο γέγραπται, τὸ ὁμοίως ἀμφοῖν ἀκροάσασθαι. τοῦτο δ᾽ ἐστὶν οὐ μόνον
τὸ μὴ προκατεγνωκέναι μηδέν, οὐδὲ τὸ τὴν εὔνοιαν ἴσην ἀποδοῦναι, ἀλλὰ
τὸ καὶ τῇ τάξει καὶ τῇ ἀπολογίᾳ, ὡς βεβούληται καὶ προῄρηται τῶν ἀγω-
νιζομένων ἕκαστος, οὕτως ἐᾶσαι χρήσασθαι.—Here, τὸ … ἀκροάσασθαι,
just like the other aorist infinitives, denotes the action in abstracto;7
notice that it is the content of a legal provision. The ‘from beginning
to end’ nuance is also present, however; cp. μὴ προκατεγνωκέναι in
the next sentence.
[D.] 47.3 δέομαι δὲ ὑμῶν καὶ ἐγὼ μετ᾽ εὐνοίας μου ἀκροάσασθαι περὶ τοῦ πρά-
γματος ἐξ ἀρχῆς ἅπαντα, ἵνα ἐκ τούτων εἰδῆτε ὅσα ἐγώ τε ἠδικήθην καὶ
ἐξηπατήθησαν οἱ δικασταὶ καὶ οὗτοι τὰ ψευδῆ ἐμαρτύρησαν.

Also with the lexical variant τὴν ἀκρόασιν ποιήσασθαι:

And. 1.9 τάδε δὲ ὑμῶν δέομαι, μετ᾽ εὐνοίας μου τὴν ἀκρόασιν τῆς ἀπολογίας ποιή-
σασθαι, καὶ μήτε μοι ἀντιδίκους καταστῆναι μήτε ὑπονοεῖν τὰ λεγόμενα,
μήτε ῥήματα θηρεύειν, ἀκροασαμένους δὲ διὰ τέλους (‘from beginning to
end’) τῆς ἀπολογίας τότε ἤδη ψηφίζεσθαι τοῦτο ὅ τι ἂν ὑμῖν αὐτοῖς ἄρι-
στον καὶ εὐορκότατον νομίζητε εἶναι.

Also relevant is

Lys. 19.11 ὅμως δὲ καὶ τούτων ὑπαρχόντων ῥᾳδίως γνώσεσθε ὅτι οὐκ ἀληθῆ ἐστι τὰ
κατηγορημένα. δέομαι δ᾽ ὑμῶν πάσῃ τέχνῃ καὶ μηχανῇ μετ᾽ εὐνοίας ἀκρο-
ασαμένους ἡμῶν διὰ τέλους, ὅ τι ἂν ὑμῖν ἄριστον καὶ εὐορκότατον νομίζητε
266 εἶναι, τοῦτο ψηφίσασθαι. |

7 Butcher, and Dilts in the new Demosthenes OCT, therefore rightly prefer ἀκροάσασθαι to the
variant ἀκροᾶσθαι.
ακροασθαι or ακροασασθαι (plato, ion 530d9)? 129

Note that at [D.] 47.3 ἀκροάσασθαι is construed with two restrictive (‘bind-
ing’) constituents, by which the implicit ‘from beginning to end’ value of the
aorist is made explicit: the object ἅπαντα, and the temporal modifier ἐξ ἀρχῆς;
for such restrictive constituents cp. also 533c2–3. In a different way, this implicit
‘from beginning to end’ value of the infinitive ποιήσασθαι is made explicit in the
context by ἀκροασαμένους διὰ τέλους at And. 1.9, and by τὸ μὴ προκατεγνωκέναι
μηδέν (‘do not have made up your mind before the end’) at D. 18.2.
Other examples of ἀκροάσασθαι are: [Pl.] Demod. 383b3, Lys. 30.1, Isoc. 14.6,
15.28, Aeschin. De falsa leg. 62, In Ctes. 59–60, [D.] 43.2, Is. De phil. 2.
Next, the constructions of σχολή will be discussed.

Σχολή + Present Infinitive

Σχολή is mostly followed by a present infinitive; it is often negated. Gener-

ally speaking, the present infinitive has the same value as that found above
for ἀκροᾶσθαι: it presents the action denoted by the infinitive as open-ended
(unbounded, atelic), i.e. as an action ‘in course’ (an activity), which has fre-
quently iterative (habitual) meaning; in the latter case the infinitive denotes
a general line of conduct. The examples from Plato are:

Phdr. 227b8 ΣΩ. Τίς οὖν δὴ ἦν ἡ διατριβή; ἢ δῆλον ὅτι τῶν λόγων ὑμᾶς ὁ Λυσίας
εἱστία; ΦΑΙ. Πεύσῃ, εἴ σοι σχολὴ προϊόντι ἀκούειν.—ΣΩ. Τί δέ; οὐκ
ἂν οἴει με κατὰ Πίνδαρον “καὶ ἀσχολίας ὑπέρτερον” πρᾶγμα ποιήσα-
σθαι τὸ τεήν τε καὶ Λυσίου διατριβὴν ἀκοῦσαι;
ποιήσασθαι Par. 1811 : ποιήσεσθαι BT
R. 406c5 καὶ οὐδενὶ σχολὴ διά βίου κάμνειν ἰατρευομένῳ.

Observe that at Phdr. 227b8 Socrates reacts to Phaedrus’ ἀκούειν with the aorist
infinitive ἀκοῦσαι. While Phaedrus wants to know from Socrates whether he has
the time to listen in the situation at hand, to lend him his ear, so to speak, as
they continue their walk, Socrates assures him that he is interested in hearing
the whole conversation; notice, again (cp. above on ἀκροάσασθαι), the presence
of an object with ἀκοῦσαι.
The other examples from Plato are: R. 406d4 and 500b8. | 267
From other authors:8

8 Σχολή + infinitive is not found in Euripides, Lysias, Herodotus, Thucydides, Andocides, Lysias,
Demosthenes, Isaeus, Aeschines.
 130 chapter 6

A. A. 1055 οὔτοι θυριαίᾳ τῇδ᾽ ἐμοὶ σχολὴ πάρα | τρίβειν·

S. Aj. 816 … εἴ τῳ καὶ λογίζεσθαι σχολή
Ar. Ach. 409 καταβαίνειν δ᾽ οὐ σγολή.
X. An. 5.1.9 σχολὴ τοῖς πολεμίοις λήζεσθαι—right now, in the situation at
hand
X. Mem. 1.6.9 ἐὰν δὲ … δέῃ, ποτέρῳ ἡ πλείων σχολὴ τούτων ἐπιμελεῖσθαι—
general line of conduct
X. Cyr. 2.1.16 εἴ τι χείρους ἡμῶν ταῦτα ποιεῖν ἦτε, οὐδὲν θαυμαστόν οὐ γὰρ ἦν ὑμῖν
σχολὴ τούτων (sc. certain weapons) ἐπιμέλεσθαι

See further X. Cyr. 1.6.17, 4.3.12 bis, 7.5.50, 8.3.48, Hier. 10.5.

Σχολή + Aorist Infinitive

Σχολή is rarely followed by an aorist infinitive; it is always negated, with the pos-
sible exception of ἀκροάσασθαι in our passage. The aorist presents the action
denoted by the infinitive as completed (bounded, telic), i.e. as an action in its
entirety (an accomplishment). In actual practice, i.e. in the context of οὐ/μὴ
σχολή, the aorist expresses the idea that full realization of the infinitive action
did not or should not come about.9 The examples are:

Ap. 23b9 οὔτε τι τῶν τῆς πόλεως πρᾶξαί μοι σχολὴ γέγονεν ἄξιον λόγου οὔτε
τῶν οἰκείων
X. An. 4.1.17 σχολὴ δ᾽ οὐκ ἦν ἰδεῖν παρελθόντι τὸ αἴτιον τῆς σπουδῆς (‘but there
was no time to go forward and find out (rather: ‘establish’—AR)
the reason for his haste’—Brownson, Loeb)
X. Cyr. 4.2.22 μὴ δῶμεν αὐτοῖς σχολὴν μήτε βουλεύσασθαι μήτε παρασκευάσασθαι
268 ἀγαθὸν αὑτοῖς μηδέν … |

Also relevant is:

Isoc. 4.112 … ἐπαυσάμεθ᾽ ἀλλήλους ἐλεοῦντες. οὐδενὶ γὰρ τοσαύτην σχολὴν παρέ-
λιπον ὥσθ᾽ ἑτέρῳ συναχθεσθῆναι (‘… since there was no man to

9 Why is positive σχολή + aorist infinitive non-existent or rare? Perhaps because it is slightly
odd to say, upon completion of some action, that you have had the time to bring about this
action. Having completed an action entails that you have had the time to do this. Interestingly,
negated σχολή + infinitive has the reverse entailment; thus, οὔτε τι τῶν τῆς πόλεως πρᾶξαί μοι
σχολὴ γέγονεν … οὔτε τῶν οἰκείων (Ap. 23b9) entails (ἀσχολίᾳ) οὔτε τι τῶν τῆς πόλεως ἔπραξα …
οὔτε τῶν οικείων.
ακροασθαι or ακροασασθαι (plato, ion 530d9)? 131

whom they allowed enough of respite so that he could share an-

other’s burdens’—Norlin, Loeb)

Conclusion

The main characteristics of, on the one hand, ἀκροᾶσθαι and ἀκροάσασθαι, and,
on the other, σχολή + infinitive can be summarized as follows.
– ἀκροᾶσθαι presents the ‘listening’ as an open-ended (unbounded, atelic)
action, i.e. as an action ‘in course’ (in technical terms: an activity)
– ἀκροάσασθαι presents the ‘listening’ as a completed (bounded, telic) action,
i.e. as an action in its entirety, from beginning to end (an accomplishment).
– σχολή is mostly followed by a present infinitive; it is often negated. The action
denoted by the infinitive is open-ended (unbounded, atelic), i.e. it is an
action ‘in course’ (an activity)
– σχολή is rarely followed by an aorist infinitive; it is always negated, with the
possible exception of ἀκροάσασθαι in our passage. The aorist presents the
action denoted by the infinitive as completed (bounded, telic), i.e. as an
action in its entirety (an accomplishment).
What conclusions can be drawn from these features for the text at Ion 530d9?
Reading ἀκροάσασθαι, the listening is presented as a bounded action, ‘from
beginning to end’. Does this make sense in our passage? Perhaps it does, if ἀκρο-
άσασθαί σου refers to Ion’s words ἄξιόν γε ἀκοῦσαι ὡς εὖ κεκόσμηκα τὸν Ὅμηρον
at 530d6–7, where ἀκοῦσαι denotes a bounded action. A similar correlation
exists below, at 536d6ff. οἶμαι δὲ οὐδ᾽ ἂν σοὶ δόξαιμι, εἴ μου ἀκούσαις λέγοντός
τι περὶ Ὁμήρου.—ΣΩ. Καὶ μὴν έθέλω γε ἀκοῦσαι, οὐ μέντοι πρότερον πρὶν ἄν μοι
ἀποκρίνῃ τόδε. The fact, however, that ἀκροάσασθαι does not occur elsewhere in
Plato, and that σχολή is elsewhere only followed by an aorist infinitive when it
is negated seems to plead against the aorist infinitive here. I prefer therefore
the present infinitive. Also, | one wonders why Socrates would use ἀκροάσα- 269
σθαι here, instead of picking up Ion’s ἀκοῦσαι verbatim, as at 536d8. Ἀκροᾶσθαι
presents the listening as an unbounded action, indicating that Socrates in a
general way is willing to lend Ion his ear, and is not interested in anything in
particular.
132 chapter 6

Bibliography

Dodds, E.R. (ed.), Plato: Gorgias (Oxford 1959).

Jacquinod, B. (ed.), Etudes sur l’aspect verbal chez Platon (Saint-Etienne 2000).
Lallot, J., ‘Aspects contrastés: l’Apologie de Socrate en grec ancien et en grec moderne’,
in Jacquinod 2000, 247–265.
Rijksbaron, A., Aristotle, Verb Meaning and Functional Grammar: Towards a New Typol-
ogy of States of Affairs (Amsterdam 1989).
Rijksbaron, A., Plato: Ion. Or: On the Iliad, edited with introduction and commentary
(Leiden 2007).
Stork, P., The Aspectual Usage of the Dynamic Infinitive in Herodotus (Groningen 1982).
 chapter 7

The Imperfect as the Tense of Substitutionary

Perception

Abstract

This chapter focuses on Substitutionary Perception. This is a phenomenon of narra-

tive texts, by which the narrator substitutes the perception of a character for his own
perception. It is suitable for highligthing the cognitive involvement of the character
in the events (s)he is experiencing. The chapter adduces a wide range of examples
from French, English and German to establish its main characteristics, viz. a special
use of tense (the French imparfait and the English past progressive) and the frequent
presence of a Window Opener (a verb or other lexemes indicating a character’s percep-
tion). After a survey of the tenses of Greek and a general exposition on the connection
between the imperfect, perception and narrativity, it discusses a wide range of Greek
examples of Substitutionary Perception, as well as an example from Virgil’s Aeneid. It
also argues that the recognition of the category of Substitutionary Perception in Greek
solves the otherwise mystifying presence of imperfect forms in the proem of Hesiod’s
Theogony (10) and Plato’s Phaedrus (253e5–255a1).

…
Aus diesen den neueren Sprachen gewidmeten Arbeiten ist auch für
Griechisch und Latein viel zu lernen
– Wackernagel (1928: 249)

∵
Consider the following sentences and text fragments:*

* This is a much expanded and rewritten version of a paper which I presented at the Confer-
ence on Classical Studies, that was organised by the Universidade de São Paulo in November
2010. I am indebted to members of the audience at that conference, and to the members of
the ‘Amsterdamse Hellenistenclub’, especially Jan Ebele van der Veen and Gerry Wakker, for
their valuable remarks on earlier versions of this paper.

© koninklijke brill nv, leiden, 2019 | doi:10.1163/9789004386129_009

 134 chapter 7

(1) ‘Here comes Jack,’ said Fred

(2) On turning round Fred saw Jack coming across the street towards him

332 (3) ‘Look!’ Fred turned round. Jack was coming across the street towards him |

(4) Look! And he nodded his head towards the moving mass of creatures before
them. Birds and animals were now swaying together, closely encircling
Mary Poppins who was rocking lightly from side to side. Even the trees
were bending and lifting gently and the moon seemed to be rocking in the
sky as a ship rocks on the sea. (emphasis Fehr)

(5) Everybody listened. A bell was ringing and a deep gruff voice could be
heard coming nearer and nearer. (emphasis Fehr)

These sentences are at the basis of an essay by the Swiss anglicist Bernhard Fehr
(Fehr 1938). Although the essay is called ‘classic’ by Monika Fludernik (1993:
308), and has indeed been used fruitfully by a number of English literary schol-
ars,1 it is rarely mentioned in English linguistic studies, and as far as I could
see absent from current general linguistic and literary studies. Thus, it is not
mentioned in Bal’s influential Narratology (20093).
Sentences (1)–(3) were made by Fehr for the occasion, fragments (4) and
(5) were taken by him from P.L. Travers’ novel Mary Poppins. Sentence (1) is
a report about something said by Fred,2 plus a quotation of Fred’s words in
direct speech. These words imply (Fehr uses ‘connote’) a Perception—Fred see-
ing Jack coming across the street towards him. Sentence (2) is a report about
Fred’s seeing Jack coming towards him. It is the reporter who tells us so, he is not
quoting Fred, but apparently had access to Fred’s mind.3 The third sentence,
finally, is not a report, but ‘is intended to express Fred’s vision’, in Fehr’s words.
‘It is the reporter running in on the actor’s vision and it is the actor lending his
sight to the articulate reporter.’ Fehr calls this, appropriately, ‘vision by proxy’,
‘Erlebte Wahrnehmung’ or ‘Substitutionary Perception’. As the word ‘erlebte’

1 An instructive example is chapter 4 ‘Progressive verbal forms in the narrative’ in Michael

Toolan’s book about Faulkner’s novel Go down, Moses (Toolan 1990).
2 Actually, Fehr simply writes about (1): ‘No. 1 is Direct Speech connoting a Perception’, omitting
the role of the reporter. See further below, note 10.
3 For a narrator’s ‘access to the mind of others’ see Fludernik (2009: 38, and ch. 8 ‘Thoughts,
Feelings and the Unconscious’).
the imperfect as the tense of substitutionary perception 135

suggests, there is a strong resemblance with erlebte Rede, or Free Indirect

Speech, or Substitutionary Speech, as in | 333

(6) What next? He had his plan; it should be easy enough. He’d get the gold.
And afterwards, when the kid was given back, they would hunt the coun-
try for him. (From E.L. Grant Watson, The Nun and the Bandit, Albatross,
p. 169; Fehr p. 79)4

I should perhaps add that Fehr is silent about some important questions raised
by his deceptively simple sentences, notably about the role of the reporter,
which differs crucially in these fragments, see examples (1′)–(3′).

(1′) ‘Here comes Jack,’ said Fred. But he was mistaken, for it was Peter.

(2′) On turning round Fred saw Jack coming across the street towards him.
*But he was mistaken, for it was Peter.

(3′) ‘Look!’ Fred turned round. Jack was coming across the street towards him.
*But he was mistaken, for it was Peter.

While, as appears from (1′) the reporter can correct Fred’s words about his
(Fred’s) seeing Jack, it appears from (2′) that he cannot correct his own state-
ment about Fred’s seeing Jack. The reporter apparently not only had access to
Fred’s perception but also shared it. In fact, ‘see’ presupposes the truth of the
(proposition of the) complement, unlike ‘say’. As for (3′), this in turn shows that,
once he has said ‘Fred turned round’, as in (3), the reporter cannot correct a fol-
lowing statement like ‘Jack was coming across the street towards him’ either.
In (3), too, then, the reporter shares Fred’s perception, or, as Fehr put it, ‘the
reporter [is] running in on the actor’s vision’. Although both in (2) and in (3),
then, the reporter and the character share their vision, there is a subtle differ-
ence. In (2), Fred’s seeing stems from the reporter, as indicated by saw, in (3)
from Fred himself, it being his consciousness that we get to know, not that of
the reporter. | 334
Sentence (4) illustrates the phenomenon of (3) in ‘real’ English. Finally, sen-
tence (5) shows that Substitutionary Perception may also involve Substitution-

4 Free Indirect Speech and Thought, and Substitutionary Perception can be subsumed under
the heading Substitutionary Cognition, cp. Toolan (1990: 112) or Character Consciousness; for
the latter term cp. Verdonk (2001: 49).
 136 chapter 7

ary Audition.5 In more modern, narratological, terms one might of course say
that in (3) Jack’s crossing the street is focalised by Fred, and is thus an instance
of (implicit) embedded focalisation. As a matter of fact, I have profited much
from narratological work, notably Irene de Jong’s Commentary on the Odyssey
(2001). I prefer using, however, Fehr’s concept of Substitutionary Perception,
because it makes use of explicit linguistic markers, notably the imperfect tense,
is combined with the very useful concept ‘Window Opener’ (see below) and
does not suffer from the inflation suffered by the concept ‘focalisation’.6
How can we tell whether we are dealing with SP? There are in (3)–(5) three
Perception Indicators, as Fehr calls them (p. 99):
(a) ‘Look’, ‘he nodded his head’ and ‘listened’, which are like ‘window openers’
(b) the progressive forms ‘was coming’, ‘were swaying’ etc.
(c) what Fehr calls ataxis, i.e. the absence of a formal connection between the
Indicator and the words expressing SP, as illustrated by Jack was coming
in (3) as against saw Jack coming in (2).
To (a) Fehr might have added now in the third sentence of fragment (4), and
nearer and nearer in (5), seemed in (4) and could be heard in (5). About the
335 progressives of (b) Fehr notes: ‘The reason [of its use] is obvious. A window is |
opened abruptly into a scene where things are going on for the eye to see, for the
ear to hear.’7 In fact, it will have been noticed that in all five examples ‘things are
going on’, i.e. we are dealing with durative states of affairs.8 I will return to this
point below. It is especially this second characteristic that will be called upon

5 Some other terms for the same phenomenon are ‘represented perception’ (Brinton 1980), ‘nar-
rated perception’ (Fludernik 1993: 305; but this term should rather be reserved for saw in (2))
and ‘inferred perception’ (Smith 2009: 383). For still other terms see Brinton (1980: 370). For
the perception expressed by saw in saw coming towards him Smith uses the term ‘indirect
perception.’
6 Thus, in connection with instances like (2) it is often said (cp. e.g. Bal (2009: 147)) that Jack’s
coming across the street is focalised by Fred, because of the presence of saw, wrongly, to my
mind, for it is the narrator who tells us that Fred saw something and what that something
was. For a similar critique see Bronzwaer (1981). As a matter of fact, Bal’s narratological model
does not recognize the perception implied by (3) as a separate type, to be distinguished from
that of (2). For both the term ‘(embedded) focalisation’ is used. For a critique of the way the
concept ‘focalisation’ has become applied to ‘un’area di significato troppo ampia’ see Dorati
(2008: 143 ff.).
7 For further discussion see sections 3.1 (‘The linguistic markers of represented perception’)
and 3.2 (‘Unique linguistic markers of represented perception’) of Brinton’s paper (Brinton
1980: 372–375; 375–376).
8 To be more precise: while the verbs in these instances belong to different semantic classes,
since come towards (1)–(3), bend and lift (4) are so-called telic (bounded) verbs, and sway and
rock (4) as well as ring (5) are atelic (unbounded) verbs, the progressive form causes them all
to have durative meaning.
the imperfect as the tense of substitutionary perception 137

to interpret and explain certain phenomena in Greek involving the imperfect.

To familiarize the reader a bit more with the concept of Substitutionary Per-
ception I will first give a sample of three further text fragments from English,
French and German novels where SP is found. The first I borrowed from sec-
ondary literature, the other two were noted by myself.

(7) She stopped at a window in the passage and held back the curtain.
Beneath was the garden, bathed in sun. The grass was sleek and shining.
Three white pigeons were flirting and tiptoeing as ornate as ladies in ball
dresses. Their elegant bodies swayed as they minced with tiny steps on
their little pink feet upon the grass. Suddenly, up they rose in a flutter, cir-
cled, and flew away. (V. Woolf, Between the acts, p. 54, quoted by Fludernik
(1993: 306); bold type and italics added by Fludernik)

(8) (a) Quand les trois hommes quittèrent l’immeuble, il n’y avait plus que
Mme Josselin et sa fille dans l’appartement. (b) Le bébé de la concierge,
après une nuit agitée, avait dû s’endormir, car la loge était obscure, et le
doigt de Maigret avait hésité un instant sur le bouton de sonnerie. (Brief
discussion about a suggestion by Maigret to have a drink together. This is
accepted, and then the doctor, one of the three men, says: ‘Il doit y avoir
un bar encore ouvert au carrefour Montparnasse’). | 336
(New paragraph) (c) Le jour ne pointait pas encore. (d) Les rues étaient
presque vides. (e) Maigret leva la tête vers le troisième étage et vit la
lumière s’éteindre dans le salon où une des fenêtres restait ouverte.
(New paragraph) (f) Est-ce que Véronique Fabre allait enfin se dévêtir et
s’étendre dans son ancienne chambre? (–)
(New paragraph) (g) Il n’y avait que la rue Vavin à parcourir. (h) Larue et
Maigret marchaient le long du trottoir. (i) Le médecin était un homme
assez petit … (and some other details about Larue).
(New paragraph) (j) C’était l’heure fraîche, un peu humide. (k) Un léger
halo entourait les becs de gaz. (l) Plusieurs voitures stationnaient devant
un cabaret encore ouvert au coin du boulevard Raspail; (m) le portier
en uniforme, debout à l’entrée, prit les deux hommes pour des clients
éventuels et, poussant la porte, fit jaillir des bouffées de musique.
(G. Simenon, Maigret et les braves gens. Paris 1962, Editions Presses de la
Cité, pp. 32–34; bold type and italics added)

(9) ‘… Wohin gehen wir?’ Esch hatte kein Ziel, und daher gingen sie zu Mut-
ter Hentjens Wirtschaft hinauf. In der Kastellgasse blieb Martin stehen:
‘Haben sie dir ein anständiges Zeugnis gegeben?’ (–), sie traten ein. Es
 138 chapter 7

war ein mäßig großer, düsterer Raum, … [Long description of the space
as seen by the men; or description by the narrator?]. Schließlich war
noch zwischen dem Büfett und dem Glasschrank ein Zinkblechbecken
mit Wasserhahn diskret eingebaut.
[New paragraph] Das Lokal war ungeheizt und seine Kälte stank. Die bei-
den Männer rieben sich die Hände, … (H. Broch, Esch oder die Anarchie.
Frankfurt/M., Fischer Bücherei 1954, pp. 6–7 (origin. 1931–1932))

All three fragments illustrate SP to varying degrees, sometimes with and some-
337 times without an introductory Window Opener. With this Opener: (7) | She
stopped at a window in the passage and held back the curtain, (9) sie traten ein;
note, in this fragment, the Substitutionary Feeling expressed by ‘war ungeheizt’:
it is the two men who felt the coldness. Therefore they frotted their hands,
as told by the narrator. Fragment (7) is discussed at some length by Monika
Fludernik in her impressive study of 1993, on pp. 306 ff. She mentions the past
progressives, the presence of little pink feet and the front position of up as signs
of ‘narrated perception’, as she calls it. Curiously, she does not mention the pres-
ence of the introductory sentence, nor the presence of the simple pasts swayed,
minced and rose, that must also be part of the ‘narrated perception’. In fact, Fehr
had already observed (1938: 105) that ‘the illusion of S.P. can be kept up as long
as simple preterites alternate with functional progressives.’9
On p. 308 Fludernik writes: ‘The most important single feature of narrated
perception … is its correlation with the German (preterital) indicative …, the
English past (progressive) and the French imparfait’. Fludernik’s observation
is borne out by the fragment from the Maigret novel (8), which has no clear
introductory Window Opener (but I will come back to this below). Here, it is
only the imparfait that tells us that perception by a character, or rather charac-
ters, is involved, characters whom we have followed when they went out into
the street, being what may be called ‘potential perceivers’, and who from that
moment onwards have been present implicitly or explicitly in the text, alone,
as a pair, or as a group of three. Interestingly, we have also been informed in the
discussion on p. 32 that once they were in the street, the three decided to look
for a bar, as can be inferred from Il doit y avoir un bar encore ouvert au carrefour
Montparnasse; so there is, after all, a Window Opener, a rather inconspicuous
one, that will dominate, however, the whole scene that follows. The fragment

9 This view may be too narrow. Thus, the iterative state of affairs denoted by ‘ticked’ in: John
entered the president’s office. The clock ticked loudly can probably also get an SP interpretation
in this particular context. For the (artificial) fragment see Smith (2009: 383).
the imperfect as the tense of substitutionary perception 139

starts with narrator perception (and narration, of course),10 which with avait | 338
dû s’endormir passes into Free Indirect Thought and next, with était, into Sub-
stitutionary Perception (note also the evaluative adjective obscure), but passes
then again into exclusive narrator perception with le doigt de Maigret avait hés-
ité. The presence of the proper name Maigret makes it clear that this comes
from an omniscient narrator, as also further on in this fragment.11 Sentences
(c) and (d) express the vision of the three characters introduced in (a).12 Sen-
tence (e) shows that the imparfait need not entail SP; here, it cannot possibly
be SP, because of the proper name in the sentence, which must stem from the
narrator. Observe that sentence (f) expresses Maigret’s thoughts as a question
in free indirect speech, which is continued for some time. In this context, the
proper name (Véronique Fabre) naturally stems from Maigret. With sentence (g)
we return to the narrative, which is continued in (h); note the proper names.
The physical description of the doctor in sentence (i) must stem from the
narrator. In other words, imparfaits like marchaient in (h) and était in (i) are hit
by the gaze of the narrator only. Sentences (j), (k) and (l), however, express the
perception of the two characters introduced in (h), Maigret and Larue, a feeling
in (j) and a vision in (k) and (l). Note also the presence of un peu in (j): this | 339
evaluative adverb points to an assessment by a human perceiver rather than by
an impersonal, external, narrator. The humidity, in turn, explains why Maigret
and Larue saw a halo around the streetlamps. We follow these two characters

10 Both features are often ignored in discussions of narrators and characters. Thus, concern-
ing his example (1), see above, Fehr simply writes: ‘No. 1 is Direct Speech connoting a
perception—Fred seeing Jack coming across the street towards him,’ and does not men-
tion that the whole fragment is due to a narrator or reporter, who ‘has overheard’ Fred
and reports to us, the readers, what he has perceived. As Toolan puts it (1988: 7): ‘a mini-
malist definition of narrative might be: a perceived sequence of non-randomly connected
events.’ (On ‘sequence’ see further below). Fludernik, too (1993: 449), considers the percep-
tion part especially important, arguing that narration is structured around ‘the script of
experiencing or viewing, rather than telling events’ (emphasis original).
11 In a narrative, the narrator is the default source of proper names. See also below, on sen-
tence (f).
12 Alternatively, it could be argued that the three characters have receded into the back-
ground to such an extent that these two sentences describe the general setting and atmo-
sphere surrounding the characters, as ‘couleur locale.’—Lips (1926: 57), one of the earliest
monographs on Free Indirect Speech, also discusses some cases of SP as instances of style
indirect libre (she does not have a separate category of SP). She rightly observes that there
are cases where one may hesitate. About the fragment ‘Yves discerna … quelque chose qui
bougeait … Il se pencha un peu pou regarder de plus près: c’était un crapaud.’ she writes:
‘On hésite entre deux interprétations; ou bien: ce qu’il vit était un crapaud; on a une énon-
ciation (i.e. by the narrator—AR); ou bien: il se pencha et constata: c’était un crapaud, alors
on a le style indirect libre.’
 140 chapter 7

as they ‘parcourent la rue Vavin’ until they reach the corner of the Boulevard
Raspail, where the narrator, with prit les deux hommes and fit jaillir, takes over;
note les deux hommes, which cannot stem from the characters.
Much more could be said about this fragment, e.g. about the passés simples
leva and vit in sentence (e), but this would keep me too long from discussing
the Greek material. Let me conclude this brief excursion into Simenon’s work
by adding that his work is a real goldmine for research in the use of the tenses
in narrative, for he is a very sophisticated and shrewd manipulator of point of
view and tense use.13
Ultimately, the effect of the device of Substitutionary Perception may be as
described by McHale (1978: 278): ‘By locating descriptive details within the per-
ceptual apparatus of a character, the reader makes them serve no longer simply
as residuals of the “real”, but as marks and measures of human consciousness.
The notion of “Substitutionary Perception”, in other words, is a principal strat-
egy for organizing a text according to limited points of view.’ To this I may add
that SP would seem to highlight the mental, especially cognitive, involvement
of a character in the events he is experiencing, the purpose being, we must
assume, to enhance the reader’s interest in, and understanding of, the char-
acter, just as free indirect speech and thought ‘can be utilized to great effect
in the detailed portrayal of characters’ sentiments and feelings and thoughts’
(Fludernik 1993: 79).14 Thus, the atmosphere in Paris just before dawn described
in (8) makes us see through the eyes of the characters themselves, without an
intervening narrator, the circumstances in which Maigret and his collaborators
work. In short, Substitutionary Perception is a means to show the reader what
a character sees of the outer world, Free Indirect Discourse shows him what a
340 character thinks of that world. |
How can this effect of Substitutionary Perception created by the imperfect
be accounted for in theoretical terms? Or, to put it differently, what feature of
the imperfect makes this, rather than e.g. the passé simple, the tense of Substi-
tutionary Perception? Fehr, who was not a linguist, is silent on this point. But it
has attracted the attention of linguists, whose work is summarized by Fludernik
(1993: 308–309) as ‘the traditional approach.’ She rejects this approach, which
‘has been to explain perception as a punctual directedness of the perceiver’s
gaze in relation to which perceived activities appear to be simultaneous. The
tense is then taken to reflect its customary “simultaneity” meaning, and the
perception itself identified with the pattern of incidence: while something is

13 Some illuminating observations about these matters may be found in Alavoine (1999).
14 Also, by choosing the point of view of a character the speaker may show empathy with
that character. For this notion see Croft and Cruse (2004: 61–62).
the imperfect as the tense of substitutionary perception 141

going on, PERCEIVER’s gaze impinges on it’. She argues rather that ‘the per-
ceptual implication of the imparfait (past progressive/German preterite), its
consciousness factor, is primary and that … its function in the pattern of inci-
dence carries with it the implication of a perceptional schema … since the
incidence can be identified with an awareness of interruption of ongoing being,
state or activity.’ I take this sentence to mean that the imparfait always implies a
perceiver (a consciousness, in Fludernik’s terms). If no contextual indications
to the contrary are present this perceiver will be the narrator.15 Thus, in sen-
tence (g) of fragment (8) marchaient is perceived by the narrator. However,
since the state of affairs is presented as ongoing or unbounded it is there also
for other perceivers to be perceived, if present (which is not so in this case).
The passé simple, on the other hand, and its equivalents, which denote a ‘sin-
gle, complete whole’, in Comrie’s words (1976: 23), can only be perceived by one
perceiver, the narrator, and not by other persons. Thus, she stopped at a window
in the passage and held back the curtain in (f) is only perceived by the (exter-
nal) narrator and told post factum,16 while what follows is perceived in fiendo
by the woman introduced in the first sentence.17 As long as unbounded | states 341
of affairs are going on they may be ‘hit’ by any perceiver’s gaze, in this case that
of the woman introduced in the first sentence.

Narrativity and the Imperfect

Before I turn to Greek, I would like to add a few words about the imperfect as
such, and its importance for narrative texts, starting from the following quo-
tation from Østen Dahl’s well-known book about tense and aspect (Dahl 1985:
112):

(10) ‘I define a narrative discourse as one where the speaker relates a series of
real or fictive events in the order they are supposed to have taken place.
As an example of a maximally short narrative discourse, Julius Caesar’s
famous statement (4.6) may be quoted:

15 In connection with the narrative of Thucydides’ Histories Egbert Bakker has proposed to
view the Greek imperfect, in fact, as the tense of ‘the observer’, with some rather far-
reaching implications (Bakker 1997). See further below, fn. 65.
16 This is not to say that e.g. she held back the curtain cannot have a simultaneous reading,
as in while she held back the curtain John slipped out of the room. But both these states of
affairs are still perceived and told by one single narrator.
17 Stretching a bit the meaning of fio, for ‘gardens’ do not happen, of course.
 142 chapter 7

(4.6) Veni, vidi, vici ‘I came, I saw, I conquered’

In actual texts, such ideal or pure narrative discourses are of course rela-
tively seldom found. Normally, the main story-line is continuously inter-
rupted by various kinds of flashbacks and points of background informa-
tion.’

Such pure narrative discourses are indeed exceptional, for it is the ‘interrup-
tions’ mentioned in the last sentence of this quotation, rather than the main
story-line, that turn such terse, list-like, narratives (‘… and then, and then’)
into something that is (potentially) interesting for an addressee. E.M. Forster
famously wrote: ‘“The king died and then the queen died” is a story’, a story
having been defined by him as ‘a narrative of events being arranged in their
time sequence’ (p. 29), while ‘“The king died, and then the queen died of grief”
is a plot. The time-sequence is preserved, but the sense of causality overshad-
ows it’ (Forster 1974: 60).18 Upon this Fludernik comments (2009: 79): ‘There is
342 a very good reason why E.M. Forster’s example of a plot—The king | died, and
then the queen died of grief—is a sentence in which the key element that trans-
forms the story (sequence of events) into a narrative or plot consists in pointing
out the connection between the events;19 it can be gathered from the queen’s
emotional response to what happens. Our sympathy for the queen in her grief
makes the story worth reading; we can understand the how and why of what
happened.’20 Likewise Dahl’s example (4.6) is a story (in Forster’s sense) rather
than a narrative. How, why, where and when, these events occurred remains
completely unclear.21 The message conveyed by these three words is: ‘This is
all there is to tell. I had barely seen the enemy when he was already defeated’,
which in the original context and situation of utterance is perfectly under-
standable. Here is this context, to be found in Plutarch’s Life of Caesar:

(11) εὐθὺς οὖν ἐπὶ τὸν ἄνδρα (: Pharnaces, the king of the Bosporianum regnum)
τρισὶν ἤλαυνε τάγμασι, καὶ περὶ πόλιν Ζῆλαν μάχην μεγάλην συνάψας αὐτὸν
μὲν ἐξέβαλε τοῦ Πόντου φεύγοντα, τὴν δὲ στρατιὰν ἄρδην ἀνεῖλε· καὶ τῆς μάχης

18 ‘… is a plot’ would nowadays rather be formulated as ‘has a plot’.

19 Observe that Fludernik equates ‘narrative’ with ‘plot’. I would rather prefer the equation
‘story + plot = narrative.’
20 Cp. also her remark on p. 5 (speaking about novels): ‘… the primary concern in narratives
is not actually chains of events but the fictional worlds in which the characters in the story
live, act, think and feel.’
21 It is probably the shortest example not of a narrative but of a log-book or chronicle.
the imperfect as the tense of substitutionary perception 143

ταύτης τὴν ὀξύτητα καὶ τὸ τάχος ἀναγγέλλων εἰς Ῥώμην πρός τινα τῶν φίλων
Μάτιον ἔγραψε τρεῖς λέξεις· ἦλθον εἶδον ἐνίκησα. Ῥωμαïστὶ δ᾽ αἱ λέξεις, εἰς
ὅμοιον ἀπολήγουσαι σχῆμα ῥήματος, οὐκ ἀπίθανον τὴν βραχυλογίαν ἔχουσιν.
(Plut. Caes. 50.2–3)22

The addressee, C. Matius, a friend of Caesar’s, was naturally familiar with Cae-
sar’s expedition, and he knew what these words referred to, and so do we,
for | Plutarch has written out for us what they referred to. This ultra-short 343
list reflects, then, the intensity and swiftness of the battle and its outcome; it
may be considered a rather bold attempt to make the erzählte Zeit as brief as
the Erzählzeit. The Latin text, with its rhetorically effective isosyllabicity and
homoioteleuton was, according to Suetonius, on display during Caesar’s Ponti-
cus triumphus:

(12) Pontico triumpho inter pompae fercula trium uerborum praetulit titulum
veni vidi vici non acta belli significantem sicut ceteris, sed celeriter con-
fecti notam. (Suet. Caes. 37.2)23

Apparently Suetonius, for one, was well aware that the events of this ‘titu-
lus’ did not qualify as a narrative, and not even as acta belli, the official war
records.24 Linguistically, the fact that the three verb forms of (4.6) are aorists,
and in the Latin variant three perfects, is of course of crucial importance for
the discourse status of this fragment: a mere succession of aorists or perfects
does not yield a narrative. What would be needed to turn veni vidi vici or ἦλθον,
εἶδον, ἐνίκησα into a narrative is the presence of imperfect forms, preferably
quite a lot of them, imperfects that flesh out the bare bones provided by the
perfects and aorists. At least one major European novelist would have agreed,

22 ‘Caesar immediately marched against him with three legions, and having joined him in a
major battle near the town of Zela, he drove him out of Pontus, and totally defeated his
army. When he gave Matius, a friend of his at Rome, an account of the promptness and
rapidity of this battle, he used three words, ‘Came saw won’. In Latin the words, which end
in the same form, have a very persuasive brevity.’
23 ‘In his Pontic triumph he displayed among the show-pieces of the procession an inscrip-
tion of but three words …, not indicating the events of the war as the others did, but the
speed with which it was finished’ (transl. Rolfe).
24 Interestingly Quintilian, in the part of his Institutio Oratoria where he discusses the nar-
ratio of forensic speeches, observes (4.2.12) on the statement Non occidi hominem in a
speech: nulla narratio est, followed by elements that would turn this into a narrative, e.g.
the presentation of the defendant’s past life, and of the reasons which have brought him
into jeopardy.
 144 chapter 7

for writing about Der Zauberberg Weinrich observes (1972: 24): ‘Thomas Mann
hat mehrfach seine Überzeugung ausgedrückt, daß die Tempus-Form Imper-
fekt die für eine Erzählung richtige Tempus-Form ist.’
Dahl is not alone in entertaining a rather one-sided idea about narratives.
Thus, Carlota Smith (2009: 168–169) concludes her discussion of the advance-
344 ment of narrative time with: ‘I take it that the basic case of narrative | is
sequence.’25 One of her longer sample texts is a series of simple pasts denot-
ing perfective events. But this, too, is rather a story, in Forster’s sense, than a
narrative, see ex. (13):

(13) She put on her apron, took a lump of clay from the bin and weighed off
enough for a small vase. The clay was wet. Frowning, she cut the lump in
half with a cheese-wire to check for air bubbles, then slammed the pieces
together much harder than usual. A fleck of clay spun off and hit her fore-
head, just above her right eye. (From Peter Robinson, A Necessary End,
New York 1989 p. 182)

Interestingly, it is the only sentence which denotes a state rather than an event
(‘The clay was wet’) that lends this fragment a more or less narrative character,
for this sentence explains why the ensuing events occur at all.
Such lists as in (13) are never long, and with good reason, for they become
quickly unnatural and boring, or they achieve an effect of caricature or parody,
as in the following fragment from Voltaire’s Candide:

(14) Elle (: Cunégonde) rencontra Candide en revenant au château, et rougit:

Candide rougit aussi. Elle lui dit bonjour d’ une voix entrecoupée; et Can-
dide lui parla sans savoir ce qu’il disait. Le lendemain, après le dîner,
comme on sortait de table, Cunégonde et Candide se trouvèrent der-
rière un paravent; Cunégonde laissa tomber son mouchoir, Candide le
ramassa; elle lui prit innocement la main, le jeune homme baissa inno-
cement la main de la jeune demoiselle avec une vivacité, une sensibilité,
une grâce toute particulière; leurs bouches se rencontrèrent, leurs yeux
s’enflammèrent, leurs genoux tremblèrent, leurs mains s’ égarèrent. M.
le baron de Thunder-ten-tronckh passa auprès du paravent, et, voyant
345 cette cause et cet effet, | chassa Candide du château à grands coups

25 Dowty holds a similar view. He considers (Dowty 1982: 38) the sequence of past tenses in
the following—artificial—fragment a narrative discourse:
John came to visit me yesterday. He was depressed when he arrived. He told me about
his problems. Gradually, he became happier. He left in a good mood.
the imperfect as the tense of substitutionary perception 145

de pied dans le derrière; Cunégonde s’évanouit; elle fut souffletée par

madame la baronne dès qu’elle fut revenue à elle-même; et tout fut
consterné dans le plus beau et le plus agréable des châteaux possibles.
(quoted by Weinrich p. 96; this quotation taken from pp. 4–6 of the edi-
tion of 1928, Paris)26

Notice, incidentally, the parody in the reference to Leibniz’ le meilleur des mon-
des possibles in the last sentence. In fact, the whole tale ridicules Leibniz’
ideas.27
‘The basis case of narrative’, then, is to my mind not simply ‘sequence’ but
sequence in a context of circumstances, sequence as the result of plans, prepa-
rations, hesitations, efforts, deliberations, decisions, fears etc., briefly sequence
as part of a plot. It is the elements just mentioned which transform what would
be a bare enumeration of events into a text that is (potentially) interesting.28
According to Arthur Palmer in his provocative book Fictional Minds, ‘Narra-
tive is in essence the presentation of fictional mental functionings’ (Palmer
2004: 177). Though this presentation is undoubtedly very important, Palmer’s
statement is perhaps a bit one-sided, too. In my view narrative is about the
behaviour of complex characters in complex situations, rather than about mere
sequences or just mental functionings.29 | 346
Also, and not unimportantly, events plus circumstances must be situated in
the past, as signalled by imparfait, Präteritum, simple past and other so-called
past tenses.30 Much more could be said about this characteristic, e.g. exactly

26 For ‘das ganz ungewöhnliches Übergewicht des Passé simple’ in this satire see the perti-
nent remarks by Weinrich (1971: 96–97).
27 That were published in his Théodicée (1710); Candide is from 1759.
28 Interestingly, while a novel consisting of just simple pasts or passés simples is hardly con-
ceivable, novels that have virtually no such tenses but predominantly imperfects do exist.
An example is Simenon’s Le chat, a grim novel about an elderly couple who hate each
other so much that they, although they live under one roof, lead two completely separate
lifes, each preparing his or her own food etc. The overwhelming presence in this novel
of the imperfect is meant to convey (and conveys) a gruesome lack of plot development,
there is virtually no advancement of narrative time, only ratiocinations and emotions. The
man and his wife are trapped in a completely immobile universe. Other French novelists
who favour the imperfect (but not to this extent) are Balzac and Flaubert. Cp. also Wein-
rich (1971: 97): ‘Flauberts Romane wie überhaupt die erzählende Literatur des Realismus
und Naturalismus sind … durch ein auffälliges Überwiegen des Imparfait in der Erzählung
gekennzeichnet.’
29 An interesting recent study of narrative, which strongly focuses on the importance of char-
acters in analysing narratives, is Altman (2008).
30 Although present tense narratives are also possible, not only as an eyewitness account
 146 chapter 7

which past, i.e. past with respect to which point of orientation, and whose past
is involved, as well as the relationship between the past tenses in novels and
in historiography, but this would lead me too far away from my main subject.
Such questions have led to an enormous amount of discussion, and I confine
myself to referring to two authors whose work I found particularly illuminating
and thought-provoking, Harald Weinrich and Paul Ricœur, see Weinrich (1971)
and Ricœur (1983–1985).

Greek Narrative and the Imperfect

Following a chiastic ordering I will now first make a number of observations

about the Greek imperfect, as well as the aorist, in general, to provide a frame-
work for what follows, and then proceed to its use in Substitutionary Percep-
tion. For this use is, indeed, found in Greek (and in Latin).
In my contribution to In the Footsteps of Raphael Kühner (Rijksbaron 1988
[ch. 3 in this volume]) I argued, following Wackernagel and others, that ancient
Greek has only one inherently past tense, the imperfect, leaving aside the rather
marginal pluperfect. Only the imperfect locates a state of affairs unequivocally
in the past, with respect to a given point of orientation.31 In my view, then,
the aorist indicative is not inherently a past tense but only by conventional
implicature, i.e. the interpretation of the aorist as a past tense is a default inter-
pretation. The point is that the aorist can be shown to behave in ways that
347 would be inappropriate for a real past tense, as in the following examples: |

(15) εἰ μέν κ᾽ αὖθι μένων Τρώων πόλιν ἀμφιμάχωμαι,

ὤλετο μέν μοί νόστος, ἀτὰρ κλέος ἄφθιτον ἔσται·
εἰ δέ κεν οἴκαδ᾽ ἵκωμι φίλην ἐς πατρίδα γαῖαν,
ὤλετό μοι κλέος ἐσθλόν, ἑπὶ δηρὸν δέ μοι αἰὼν
ἔσσεται, οὐδέ κέ μ᾽ ὦκα τέλος θανάτοιο κιχείη. (Hom. Il. 9.412–416)32

of, for instance, a football match, but also of past events, especially in first person narra-
tives. Cp. Fludernik (2009: 51–52) for ‘narrative in the present tense.’ For a witty critique of
this way of story telling see the article by Ph. Pullmann in The Guardian of 18 September
2010.
31 More exactly: this is the default interpretation. The words used by De Mulder and Vetters
(2002: 126) in connection with the French imparfait apply also to the Greek imperfect: ‘it
normally signals reference to a past event, unless there are indications to the contrary.’
One may think here e.g. of wishes and counterfactuals. See also below, on the aorist.
32 ‘If I stay here and fight beside the city of the Trojans, | my return home is gone, but my glory
shall be everlasting; but if I return home to the beloved land of my fathers, | the excellence
the imperfect as the tense of substitutionary perception 147

(16) οὕτω δὲ καὶ στρατὸς πολλὸς ὑπὸ ὀλίγου διαφθείρεται κατὰ τοιόνδε· ἐπεάν σφι
ὁ θεὸς φθονήσας φόβον ἐμβάλῃ ἢ βροντήν, δι᾽ ὦν ἐφθάρησαν ἀναξίως ἑωυτῶν.
(Hdt. 7.10ε)33

(17) ζεύξας φὴς τὸν Ἑλλήσποντον ἑλᾶν στρατὸν διὰ τῆς Εὐρώπης ἐς τὴν Ἑλλάδα.
Καὶ δὴ καὶ συνήνεικε ἤτοι κατὰ γῆν ἢ [καὶ] κατὰ θάλασσαν ἑσσωθῆναι, ἢ καὶ
κατ᾽ ἀμφότερα· (–). οὐκ ὦν ἀμφοτέρῃ σφι ἐχώρησε· (Hdt. 7.10β.1–2)34

(18) Διὸ ὑμέας νῦν ἐγὼ συνέλεξα, ἵνα τὸ νοέω πρήσσειν ὑπερθέωμαι ὑμῖν· (Hdt.
7.8α.2)35 | 348

In (15) we find two aorists coordinated with a future indicative, in (16) a gnomic
aorist, which temporally does not differ from the generic present of the first
sentence (note also the subjunctive + ἄν), in (17) two aorists are located in the
future expressed by ἐλᾶν and in (18), finally, an aorist is modified by νῦν, refer-
ring to the speech situation. In all these cases the aorist indicative could not be
replaced by an imperfect. For details I may refer to the article just mentioned.36
On the assumption that narratives concern the past we may infer that to be
a narrative a text must first of all have one or more imperfects: this is the signal
that past states of affairs are involved, without it the reader is not transferred
to the past. This is confirmed by phenomena like the following.

(I) Many forensic speeches of Lysias have an exordium which ends in a sentence
like ἐγὼ τοίνυν ἐξ ἀρχῆς ὑμῖν ἅπαντα ἐπιδείξω τὰ ἐμαυτοῦ πράγματα, or ἐντεῦθεν
ὑμῖν ἄρξομαι διηγεῖσθαι. In such cases the first or second independent verb of the
ensuing narratio is always an imperfect. Διηγεῖσθαι/διηγήσασθαι, then, means
using the imperfect.37

of my glory is gone, but there will be a long life for me, | and my end in death will not come
to me quickly.’ (transl. Lattimore).
33 ‘Thus a numerous host is destroyed by one that is lesser, for instance in this way: after
the god, having become jealous, has cast panic or lightning among them, they at once are
gone, unworthily.’ (transl. AR).
34 ‘You will bridge the Hellespont (so you say) and march … And guess what? It so happens
that you have been defeated … So they have not been successful by sea and by land’ (transl.
Godley + AR).
35 ‘For this cause I have now summoned you together, that I may impart to you my purpose’
(transl. Godley).
36 As for English and French, it seems highly unlikely that e.g. the passé simple can have a
comparable non-past meaning. I should add, however, that novels situated in a utopian
future may have past tenses, an example being Orwell’s 1984.
37 At least, for Lysias; for διηγήσασθαι, which occurs outside the exordium, see 1.22. Out-
 148 chapter 7

Some examples:

(19) … ἐγὼ τοίνυν ἐξ ἀρχῆς ὑμῖν ἅπαντα ἐπιδείξω τὰ ἐμαυτοῦ πράγματα …

1.6 Ἐγὼ γάρ, ὦ Ἀθηναῖοι, ἐπειδὴ ἔδοξέ μοι γῆμαι καὶ γυναῖκα ἠγαγόμην εἰς τὴν
349 οἰκίαν, τὸν μὲν ἄλλον χρόνον οὕτω διεκείμην | ὥστε μήτε λύπεῖν μήτε λίαν ἐπ᾽
ἐκείνῃ εἶναι ὅ τι ἂν ἐθέλῃ ποιεῖν, ἔφύλαττόν τε ὡς οἷόν τε ἦν, καὶ προσεῖχον τὸν
νοῦν ὥσπερ εἰκὸς ἦν.38

(20) … πειράσομαι … διδάξαι.

7.4 Ἦν μὲν γὰρ τοῦτο Πεισάνδρου τὸ χωρίον, …39

(21) … ἐντεῦθεν ὑμῖν ἄρξομαι διηγεῖσθαι.

13.5 Ἐπειδὴ γὰρ αἱ νῆες αἱ ὑμέτεραι διεφθάρησαν καὶ τὰ πράγματα ⟨τὰ⟩ ἐν τῇ
πόλει ἀσθενέστερα ἐγεγένητο, οὐ πολλῷ χρόνῳ ὕστερον αἵ τε νῆες αἱ Λακεδαι-
μονίων ἐπὶ τὸν Πειραιᾶ ἀφικνοῦνται, καὶ ἅμα λόγοι πρὸς Λακεδαιμονίους περὶ
τῆς εἰρήνης ἐγίγνοντο. ἐν δὲ τῷ χρόνῳ τούτῳ οἱ βουλόμενοι νεώτερα πράγματα
ἐν τῇ πόλει γίγνεσθαι ἐπεβούλευον, …40

In (19) and (21) the narrative opens with a temporal clause which briefly sum-
marizes the prehistory of the narrative sensu stricto that will follow. Notice, in

side Lysias one may compare e.g. Aeschin. In Tim. 43 and Isoc. 18.4 and 21.1. However,
Isocrates also has a series of aorists after διηγεῖσθαι, see 19.4. As for Demosthenes, he (i)
only uses βούλομαι, ἀναγκαῖον etc. διηγήσασθαι, not διηγεῖσθαι, and (ii) διηγήσασθαι is usu-
ally followed by aorists. A typical example is 34.5 ff. MacDowell observes (1990: 32) that
the first piece of narrative in Meidias (or. 21; §§ 13–18) is ‘a fairly straightforward cata-
logue of incidents.’ Note that the predominant tense here is the aorist. In fact, §§13–18
are a story rather than a narrative (cp. above, pp. [341–343].) MacDowell’s observation
seems to hold good for many other speeches as well, but further research is needed
here.
38 ‘I shall therefore set forth the whole of my story from the beginning. (–) When I decided
to marry and brought a wife into my house, for some time I was disposed …, and I kept a
watch on her … and observed her …’ (transl. here and below Lamb).
39 ‘Nevertheless I will try to explain the affair to you from the beginning. This plot of ground
belonged to Peisander …’ Jan-Ebele van der Veen (per litt.) suggests influence from tran-
sitions like that at Hdt. 1.5.4–6.1: τὸν δὲ οἶδα αὐτὸς πρῶτον ὑπάρξαντα ἀδίκων ἔργων ἐς τοὺς
Ἔλληνας, τοῦτον σημήνας προβήσομαι … (6.1) Κροῖσος ἦν Λύδὸς μὲν γένος … (‘But I do know
who was the first man to begin unjust acts against the Hellenes. I shall describe him and
then proceed. … Croesus was of Lydian ancestry …’).
40 ‘I shall therefore start my relation at a point where … When your ships had been destroyed
and …, the ships of the Lacedaemonians arrived soon after …, and negotiations for peace
were made at once. … At this moment those who desired to have a revolution in the State
were busy with their plot …’
the imperfect as the tense of substitutionary perception 149

(21), the historical present ἀφικνοῦνται, denoting a turning-point, after which,

with ἐγίγνοντο, the narrative begins. | 350
For further examples see 3.5, 10.4 (with a somewhat different introductory
sentence), 12.4, 16.4 (probably) and 32.4.

(II) In a similar way the introductory part of messenger speeches in tragedy

mostly ends in an appeal to the messenger to tell his story, e.g. Med. 1134 λέξον,
Hipp. 1171 εἰπέ, Andr. 1084 σήμαινε, El. 773 βούλομαι μαθεῖν, Hec. 517 εἰπέ, Supp.
649 λέξον, Or. 863 λέγε, etc. The messenger speech proper usually opens with
an imperfect, sometimes after a temporal clause.41 The following two examples
are representative of this procedure:

(22) Ηλ. λέγ᾽, ὦ γεραιέ 863

(–)
Αγ. ἐτύγχανον μὲν ἀγρόθεν πυλῶν ἔσω
βαίνων, πυθέσθαι δεόμενος τά τ᾽ ἀμφὶ σοῦ
τά τ᾽ ἀμφ᾽ Ὀρέστου· σῷ γὰρ εὔνοιαν πατρὶ
ἀεί ποτ᾽ εἶχον, καί μ᾽ ἔφερβε σὸς δόμος (E. Or. 866 ff.)42

(23) Πη. Σήμαιν᾽· 1084

Αγ. ἐπεὶ τὸ κλεινὸν ἤλθομεν Φοίβου πέδον,
τρεῖς μὲν φαεννὰς ἡλίου διεξόδους
θέᾳ διδόντες ὄμματ᾽ ἐξεπίμπλαμεν.
καὶ τοῦθ᾽ ὕποπτον ἦν ἄρ᾽· εἰς δὲ συστάσεις
κύκλους τ᾽ ἐχώρει λαὸς οἰκήτωρ θεοῦ. (E. Andr. 1085 ff.)43

(III) That narratives and imperfects are intimately connected can also be
shown from a quite different angle. If the presence of imperfects signals narra-
tivity one may expect that the absence of imperfects signals non-narrativity.
This is | borne out by the total or near-total absence of the imperfect from 351
documentary texts, non-narrative texts par excellence, such as contracts, inven-
tories, decrees. Thus, in the Linear B tablets there are no certain instances of the

41 For a comprehensive study of the messenger speeches in Euripides see De Jong (1991).
42 ‘Tell me old sir …’—‘I was making my way from the country to the citadel … I always felt
good will toward your father: your house gave me sustenance …’ (transl. Kovacs).
43 ‘Explain …’—‘When we had come to …, we spent three shining circuits of the sun in look-
ing around and … This caused suspicion …: the people … gathered in knots and circles’
(transl. Kovacs).
 150 chapter 7

imperfect, nor of the pluperfect, for that matter. There is a maximum of 2 pos-
sible instances, on a total of 60 finite verb forms. The majority of these forms
consist of present indicatives (between 21–23 instances) and aorist indicatives
(between 21–27 instances); there are also a few cases of future and perfect
indicative. Ambiguous forms like pe-re (φέρει or φέρε) are generally taken as
present indicatives, no doubt rightly. In view of the cataloguing character of
these texts it seems highly unlikely that the two possible instances of the imper-
fect are really imperfects. See for this whole subject the recent and very thor-
ough treatment in Bartoněk (2003: 324ff.).
So far I only discussed the temporal properties of the Greek imperfect, but
this is only part of the picture, of course, for I ignored the aspectual part. Ulti-
mately, it is the combination of the past meaning of the secondary indicative,
the aspectual meaning of the present stem, as well as the lexical meaning of
the various verbs, or rather verb stems, which makes the imperfect the narra-
tive tense par excellence. Specifically, this combination is the source of all the
uses traditionally assigned to the imperfect, which together lend a narrative
text its narrative character, e.g.:
– the durative use, with atelic (unbounded) verbs like παίζειν: ἔπαιζε μετ᾽ ἄλλων
ἡλίκων ἐν ὁδῷ (Hdt. 1.114.1)44
– the iterative use, frequently with momentaneous verbs like εἰσβάλλω ‘in-
vade’: ἐπολιόρκεε τὴν Μίλητον τρόπῳ τοιῷδε. ὅκως μὲν εἴη ἐν τῇ γῇ καρπὸς ἁδρός,
352 τηνικαῦτα ἐσέβαλλε τὴν στρατιήν· (Hdt. 1.17.1)45 |
– the conative use, with telic (bounded) verbs like πείθω: ἦσαν μέν τινες οἳ τὸν
Ἀγησίλαον ἔπειθον χωρὶς τούτοις ἐπιθέσθαι· ὁ δὲ … ἔγνω κράτιστον εἶναι … (X.
HG 6.5.16)46
Perhaps its most important function in narrative texts, however, is of a more
rhetorical-stylistic nature, in that it creates a framework within which other
states of affairs may occur. It in principle makes the reader wonder ‘what else
happened?’, as, for instance, in the fragments from Herodotus and Xenophon
above. This function makes it the means par excellence for creating discourse
cohesion. For these and other uses see e.g. the discussion in Rijksbaron (2006:
11–20).
The common aspectual meaning in all these uses is that a given state of
affairs is presented as not-completed. For like many others I take it that the

44 ‘there he was playing in the road with others of his age.’

45 ‘This was the manner in which he laid siege to Miletus. Whenever the crops on its land
were ripe, at that time he would thrust his army upon it.’ (transl. Godley + Purvis)
46 ‘And there were some who tried to persuade Agesilaus to attack these troops separately;
he, however, … judged it best …’ (transl. Brownson).
the imperfect as the tense of substitutionary perception 151

basic aspectual meaning of the present stem is to denote ‘eine nicht vollendete
Tätigkeit’, in the words of Mayser (1926: 135). That we are entitled to take this as
the basic meaning is clear from the fact that only present stem states of affairs
can be interrupted or broken off, as in these instances of the present participle:

(24) παῦσαι σπεύδων τὰ σπεύδεις (Hdt. 1.206.1)47

(25) ἔπειτα ὁ Μενέξενος ἐκ τῆς αὐλῆς μεταξὺ παίζων εἰσέρχεται (Pl. Ly. 207b)48

*Παῦσαι σπεύσας and *μεταξὺ παίσας are ungrammatical.

To the narrative uses of the imperfect I would now like to add that of Substi-
tutionary Perception. | 353

Substitutionary Perception in Greek

Consider the following fragments, where

– imperfects and pluperfects of Substitutionary Perception, as well as some
other elements implying Substitutionary Perception, are in italics and bold
type
– narrator imperfects, aorists and historical presents are in bold type
– generic (omnitemporal) presents and gnomic aorists are underlined
– a diplê (>) after a verb form indicates that this verb form functions as a Win-
dow Opener.49

(26) αὐτὰρ Ὀδυσσεὺς

Ἀλκινόου πρὸς δώματ᾽ ἴε > κλυτά· πολλὰ δέ οἱ κῆρ
ὅρμαιν᾽ ἱσταμένῳ >, πρὶν χάλκεον οὐδὸν ἱκέσθαι.
ὥς τε γὰρ ἠελίου αἴγλη πέλεν ἠὲ σελήνης
85 δῶμα καθ᾽ ὑψερεφὲς μεγαλήτορος Ἀλκινόοιο.
χάλκεοι μὲν γὰρ τοῖχοι ἐληλέδατ᾽ ἔνθα καὶ ἔνθα,
ἐς μυχὸν ἐξ οὐδοῦ, περὶ δὲ θριγκὸς κυάνοιο·
χρύσειαι δὲ θύραι πυκινὸν δόμον ἐντὸς ἔεργον·
ἀργύρεοι δὲ σταθμοὶ ἐν χαλκέῳ ἕστασαν οὐδῷ,
90 ἀργύρεον δ᾽ ἐφ᾽ ὑπερθύριον, χρυσέη δὲ κορώνη.
χρύσειοι δ᾽ ἑκάτερθε καὶ ἀργύρεοι κύνες ἦσαν,

47 ‘stop striving after the things that you are striving after.’
48 ‘next Menexenus steps in from the court, in the middle of his game.’
49 For the narrative technique displayed in this passage see also De Jong (2001).
 152 chapter 7

οὓς Ἥφαιστος ἔτευξεν ἰδυίῃσι πραπίδεσσι

δῶμα φυλασσέμεναι μεγαλήτορος Ἀλκινόοιο,
ἀθανάτους ὂντας καὶ ἀγήρως ἤματα πάντα.
95 ἐν δὲ θρόνοι περὶ τοῖχον ἐρηρέδάτ᾽ ἔνθα καὶ ἔνθα,
ἐς μυχὸν ἐξ οὐδοῖο διαμπερές, ἔνθ᾽ ἐνὶ πέπλοι
354 λεπτοὶ ἐΰννητοι βεβλήατο, ἔργα γυναικῶν. |
ἔνθα δὲ Φαιήκων ἡγήτορες ἑδριόωντο
πίνοντες καὶ ἔδοντες· ἐπηετανὸν γὰρ ἔχεσκον.
100 χρύσειοι δ᾽ ἄρα κοῦροι ἐϋδμήτων ἐπὶ βωμῶν
ἕστασαν αἰθομένας δαΐδας μετὰ χερσὶν ἔχοντες,
φαίνοντες νύκτας κατὰ δώματα δαιτυμόνεσσι.
πεντήκοντα δέ οἱ δμῳαὶ κατὰ δῶμα γυναῖκες
αἱ μὲν ἀλετρεύουσι μύλῃσ᾽ ἔπι μήλοπα καρπόν,
105 αἱ δ᾽ ἱστοὺς ὑφόωσι καὶ ἠλάκατα στρωφῶσιν
ἥμεναι, οἷά τε φύλλα μακεδνῆς αἰγείροιο·
καιρουσσέων δ᾽ ὀθονέων ἀπολείβεται ὑγρὸν ἔλαιον.
ὅσσον Φαίηκες περὶ πάντων ἴδριες ἀνδρῶν
νῆα θοὴν ἐνὶ πόντῳ ἐλαυνέμεν, ὣς δὲ γυναῖκες
110 ἱστὸν τεχνῆσσαι· περὶ γάρ σφισι δῶκεν Ἀθήνη
ἔργα τ᾽ ἐπίστασθαι περικαλλέα καὶ φρένας ἐσθλάς.
ἔκτοσθεν δ᾽ αὐλῆς μέγας ὄρχατος ἄγχι θυράων
τετράγυος· περὶ δ᾽ ἕρκος ἐλήλαται ἀμφοτέρωθεν.
ἔνθα δὲ δένδρεα μακρὰ πεφύκασι τηλεθάοντα,
115 ὄγχναι καὶ ῥοιαὶ καὶ μηλέαι ἀγλαόκαρποι
συκέαι τε γλυκεραὶ καὶ ἐλαῖαι τηλεθόωσαι.
τάων οὔ ποτε καρπὸς ἀπόλλυται οὐδ᾽ ἀπολείπει
χείματος οὐδὲ θέρευς, ἐπετήσιος· ἀλλὰ μάλ᾽ αἰεὶ
ζεφυρίη πνείουσα τὰ μὲν φύει, ἄλλα δὲ πέσσει.
120 ὄγχνη ἐπ᾽ ὄγχνῃ γηράσκει, μῆλον δ᾽ ἐπὶ μήλῳ,
αὐτὰρ ἐπὶ σταφυλῇ σταφυλή, σῦκον δ᾽ ἐπὶ σύκῳ.
ἔνθα δέ οἱ πολύκαρπος ἀλωὴ ἐρρίζωται,
τῆς ἕτερον μέν θ᾽ εἱλόπεδον λευρῷ ἐνὶ χώρῳ
τέρσεται ἠελίῳ, ἑτέρας δ᾽ ἄρα τε τρυγόωσιν,
125 ἄλλας δὲ τραπέουσι· πάροιθε δέ τ᾽ ὄμφακές εἰσιν
355 ἄνθος ἀφιεῖσαι, ἕτεραι δ᾽ ὑποπερκάζουσιν. |
ἔνθα δὲ κοσμηταὶ πρασιαὶ παρὰ νείατον ὄρχον
παντοῖαι πεφύασιν, ἐπηετανὸν γανόωσαι.
ἐν δὲ δύω κρῆναι ἡ μέν τ᾽ ἀνὰ κῆπον ἅπαντα
130 σκίδναται, ἡ δ᾽ ἑτέρωθεν ὑπ᾽ αὐλῆς οὐδὸν ἵησι
πρὸς δόμον ὑψηλόν, ὅθεν ὑδρεύοντο πολῖται.
the imperfect as the tense of substitutionary perception 153

τȏι’ ἄρ᾽ ἐν Ἀλκινόοιο θεῶν ἔσαν ἀγλαὰ δῶρα.

ἔνθα στὰς θηεῖτο πολύτλας δῖος Ὀδυσσεύς.
αὐτὰρ ἐπεὶ δὴ πάντα ἑῷ θηήσατο θυμῷ,
135 καρπαλίμως ὑπὲρ οὐδὸν ἐβήσετο δώματος εἴσω.
εὗρε δὲ Φαιήκων ἡγήτορας ἠδὲ μέδοντας
σπένδοντας δεπάεσσιν … (Hom. Od. 7.81–137)50 | 356

In the brilliantly structured passage 81–132—called ‘inorganic’ by Heubeck et

al. in their commentary, that elsewhere, too, shows a certain lack of sensitivity
to the literary sophistication of the Odyssey—the phenomenon of Substitu-
tionary Perception makes a quite spectacular entrance into European litera-
ture.51 There is a gradual movement of a camera (which is Odysseus) from the
outside to the inside of the palace of Alcinous. After we had been told about
the arrival of Athena and Odysseus at Alcinous’ palace (46) we now learn that
Odysseus was actually approaching the palace (πρὸς δώματ᾽ ἴε, 82, perceived
and narrated by the narrator), following Athena’s order at l. 50 σὺ δ᾽ ἔσω κίε (the

50 ‘But now Odysseus was approaching > the famous house of Alkinoôs. He pondered >
many things …, while he was busy finding a place, until he arrived at the threshold. For a
shining splendour was all over the house. Brazen walls had been laid from the threshold
to the inner room, … golden doors were guarding …, silver pillars stood on the thresh-
old … On each side there were golden and silver dogs, which Hephaistos had fashioned
[narratorial intrusion] … (95) And within thrones had been placed along the walls from
the threshold to the door …, where cloths had been spread. (98) There the leaders were
sitting drinking and eating [rather than: ‘held their sessions’ (Lattimore)], for they held
these forever [narratorial intrusion] … there were young men of gold standing on strong-
compounded bases, shedding light through the house by night … (98) Of the fifty serving
women in his house some grind the grain, some weave … and turn the distaff. (107) And
from the cloths oozes the limpid olive oil. As much as … so their women ⟨are⟩ skilled … For
Athena had bestowed on them … (112) On the outside of the courtyard ⟨is⟩ a great orchard,
a fence has been driven around it … There tall fruit trees grow … (117) Never their fruit is
spoiled, never does it give out, but always the West Wind makes some grow and ripens
others. Pear matures on pear …, and apple upon apple … (122) There also a vineyard with
many grapes has been planted …, some of which dry in the sun, others they gather and still
others they trample. In front of these are unripe grapes, while others are darkening. (127)
And there grow all kinds of greens … And there two springs distribute water, one through
…, and one jets out by the threshold of the courtyard, where the townspeople were fetch-
ing their water [not: come for their water (Lattimore)!]. Such were [not: are (Lattimore)]
the glorious gifts of the gods … (133). Standing there Odysseus was full of admiration. But
when his mind was done with all admiration, he stepped over the threshold and went on
into the palace, and there he found the leaders of the Phaiakians … pouring libation in
cups …’ (transl. Lattimore, with adaptations).
51 A much shorter example, without omnitemporal presents, occurs already before this pas-
sage, viz. at 5.55–74, for which see De Jong (2001) on 63–75.
 154 chapter 7

actual going in occurs at l. 135), and that he pondered (ὅρμαινε) many things
while he came to a standstill (ἱσταμένῳ, 82), before he arrived (πρὶν … ἱκέσθαι,
83) at the threshold. Ὅρμαινε, together with ἱσταμένῳ, functions as Window
Opener for the scene that is described from l. 86 onwards,52 while πρὶν χάλκεον
οὐδὸν ἱκέσθαι functions as what might be called a ‘Window Closer’. The next two
lines (84–85), however, are probably still due to the narrator, since they men-
tion the light which explains how Odysseus was able to perceive and ponder the
things mentioned in the next lines. Starting at line 86 the narrator makes way
for Odysseus’ perceptions, for lines 86–102 describe, in imperfects and pluper-
fects,53 the things Odysseus pondered, things, we may infer, seen by him during
the time of his ἵστασθαι: first the walls ἐς μυχὸν ἐξ οὐδοῦ and elements belong-
ing to the walls (86–94, with a temporary return to the narrator in lines 92–94),
and then the space enclosed by the walls as far as he could see it, i.e. the court-
yard (95–102; notice 95 ἐν δὲ θρόνοι, and cp. ὑπ᾽ αὐλῆς οὐδόν at 130), with the
357 Phaeacian leaders being in session while | drinking and eating (98 ἑδριόωντο |
πίνοντες καὶ ἔδοντες). In fact, Athene had told Odysseus at l. 49 δήεις δὲ διοτρε-
φέας βασιλῆας | δαίτην δαινυμένους. Then what is going on inside the palace itself,
and what is situated outside the αὐλή (112 ἔκτοσθεν αὐλῆς), viz. the orchard, nei-
ther of which can be seen by Odysseus, is told by the omniscient narrator in
omnitemporal present (and perfect) indicatives (103–131 ὑψηλόν).54 With ὑδρεύ-
οντο at 131 we return to the perception by Odysseus: he sees the townsmen come
for the water. Τοῖ(α) … ἔσαν still continues Odysseus’ perception; note evalua-
tive τοῖα. Incidentally, Odysseus himself cannot be seen, for Athena has covered
him with a cloud (ll. 41–42). Finally, at 133 the narrative is resumed; ἔνθα refers
back to the οὐδόν introduced in l. 83, while the aorist participle στάς picks up,
and closes off, the present participle ὅρμαιν᾽ ἱσταμένῳ at l. 83: Odysseus has
come to a standstill, and the window opened by ἱσταμένῳ is closed. Next θηεῖτο,
which is due to the narrator, expresses the idea that once Odysseus stands on
the threshold he is no longer pondering but full of admiration. Having had his
fill of admiration he goes into the palace, where the leaders of the Phaeacians,
whom we have left while they were in the middle of their dinner (98), are just
finishing their meal with libations.

52 Interestingly, Ameis-Hentze’s note on ἱσταμένῳ suggests that they, too, considered it a kind
of Window Opener, since they write: ‘ἵσταμένῳ indem er stehen blieb, womit die folgende
Beschreibung vorbereitet wird.’ For other present stem forms of ἵσταμαι meaning ‘be busy
taking a position’ see Od. 8.263, 9.381 and 402.
53 De Jong (2001: on Od. 5.63–75) had already observed that ‘[w]hen characters focalize
scenery, we usually find past tenses’. Rather, I argue, imperfects and pluperfects.
54 Cp. De Jong (2001: 176): ‘… the omniscient narrator … provid[es] all kinds of information
which exceeds the perception or knowledge of the hero.’
the imperfect as the tense of substitutionary perception 155

The similarities with the examples from modern novels discussed above
should be noticed: the presence of Window Openers at the beginning (πρὸς …
ἴε 82, ὅρμαιν᾽ ἱσταμένῳ 83), and the abundant presence of imperfects, and plu-
perfects, which describe what Odysseus saw while taking position. Notice that
Odysseus’ vision is restricted, for it runs from the threshold (83, 87, 89, 96) to
the inner room (87, 96). Just as in the examples from modern novels above a
durative state of affairs in the imperfect may, as long as it is going on, be ‘hit’ by
any perceiver’s gaze, in this case that of Odysseus. Notice also the very different
role of the omnitemporal presents and perfects in which the narrator describes
the interior of the palace and its garden, a diference that returns in examples
below, e.g. (30). | 358
The next examples, from prose, show the working of Substitutionary Percep-
tion in a more unobtrusive way.

(27) οἱ δ᾽ ἀπὸ τῆς Ὀλύνθου τοῖς Ποτειδαίοις βοηθοί (ἀπεῖχε δὲ ἑξήκοντα μάλιστα
σταδίους καὶ ἔστι καταφανές) ὡς ἡ μάχη ἐγίγνετο καὶ τὰ σημεῖα ἤρθη, βραχὺ
μέν τι προῆλθον … (Th. 1.63.2)55

ἀπεῖχε A B E F M: ἀπέχει C, Hude

(28) ἔπλευσαν εἰς Αἰγὸς ποταμοὺς ἀντίον τῆς Λαμψάκου· διεῖχε δ᾽ ὁ Ἐλλήσποντος
ταύτῃ σταδίους ὡς πεντεκαίδεκα. (X. HG 2.1.21)56

(29) τὸ δὲ στράτευμα ὁ σῖτος ἐπέλιπε, καὶ πρίασθαι οὐκ ἦν εἰ μὴ ἐν τῇ Λύδίᾳ ἀγορᾷ

ἐν τῷ Κύρου βαρβαρικῷ, τὴν καπίθην ἀλεύρων ἢ ἀλφίτων τεττάρων σίγλων. ὁ
δὲ σίγλος δύναται ἑπτὰ ὀβολοὺς καὶ ἡμιωβέλιον Ἀττικούς· ἡ δὲ καπίθη δύο χοί-
νικας Ἀττικὰς ἐχώρει. κρέα οὖν ἐσθίοντες οἱ στρατιῶται διεγίγνοντο. (X. An.
1.5.6)57 | 359

55 ‘Meanwhile the auxiliaries of the Potidaeans from Olynthus, which was about seven
miles off, and is in sight of Potidaea, when the battle began and the signals were raised,
advanced a little way …’ (transl. Crawley-Strassler in: The Landmark Thucydides, New York
2008—although they have ‘is about seven miles off’, perhaps reading ἀπέχει.) Classen-
Steup ad loc. already observed: ‘Impf., indem die lokalen Umstände in die Vergangenheit
der Erzählung hineingezogen werden.’
56 ‘… they sailed to Aegospotami, which is opposite Lampsacus, the Hellespont at this
point being about fifteen stadia wide’ (transl. Brownson, wrongly rendering διεῖχε by
‘being’).
57 ‘As for the troops, their supply of grain gave out, and it was not possible to buy any except
in the Lydian market …, at the price of of four sigli for a capithê of wheat flour or barley
meal. The siglus is worth seven and one-half Attic obols, and the capithê had the capacity
 156 chapter 7

In (27) and (28) the distance between two points is presented as perceived
by the βοηθοί and the shipsmen (28), respectively. In (27) the imperfect is com-
bined with omnitemporal ἔστι. Why? Perhaps because being clearly visible is
not something judged by estimation?
For (29) I have consulted a number of commentaries, to see whether they
notice the change from δύναται to ἐχώρει. Some do, and one hits the nail on the
head. Here are their remarks:
– Vollbrecht 1877 (1907): ‘Wechsel des tempus; das Imperf. steht oft von Din-
gen, welche in der Gegenwart noch fortbestehen, um den ehemaligen Zus-
tand zu berücksichtigen.’
– Krüger 18897 refers the reader to his note at 1.4.9 ἐνόμιζον: ‘ἐνόμιζον für νομί-
ζουσι. Der lebhafte Grieche drückt öfter als wir nur aus, wie etwas zur Zeit
der erzählten Vorgänge gefunden wurde, auch wenn es sich dauernd so ver-
hält.’
– Couvreur 1900: ‘Remarquer l’ imparfait ἐχώρει, pour exprimer un fait qui est
toujours vrai, mais que Xénophon constatait alors; il emploie souvent cet
imparfait dans les descriptions géographiques.’
Krüger’s ‘wie etwas gefunden wurde’ is not bad, but Couvreur’s ‘Xénophon con-
statait’ is definitely better, since it makes the SP by a specific perceiver explicit.
The imperfect may have been chosen to convey the idea that Xenophon found
the size of the καπίθη very small, in fact, too small.58 After all, this is why the
soldiers had to live on meat.
Incidentally, example (29) raises an interesting problem, which for reasons
of space I can only mention here. As is well known, the Anabasis has the form
360 of a third-person, hetero-diegetic, narrative, with an anonymous, | covert, nar-
rator. On the other hand, there is the name ‘Xenophon’, which occurs dozens of
times in the text, referring to a participant in the expedition. And, thirdly, there
is also an overt first-person narrator, for on four occasions (1.2.5, 1.9.28, 2.3.1 and
2.6.6) a first person pops in, e.g. 1.2.5 Κῦρος δὲ ἔχων οὓς εἴρηκα ὡρμᾶτο …, a nar-

of two Attic choenices. The soldiers therefore managed to subsist by eating meat’ (transl.
Brownson).
58 With ‘constatait’ compare ‘erkannte’ in K-G 1, 145–146: ‘Das Imperfekt scheint bisweilen
statt des Präsens zu stehen, indem die durch dasselbe ausgedrückte Handlung in der
Gegenwart fortbesteht. Der Redende nimmt alsdann keine Rücksicht auf das Fortbeste-
hen der Handlung in der Gegenwart, sondern verstetzt sich in den Zeitpunkt der Vergan-
genheit zurück, in welchem er dieselbe erkannte oder von ihr die Rede war’. However,
while ‘der Redende erkannte’ is correct in our case, since narrator and perceiver coincide,
their remark does not mean that they recognize something like Substitutionary Percep-
tion. See also the main text, next paragraph. For ‘constatait’ cp. fn. 12.
the imperfect as the tense of substitutionary perception 157

rator, incidentally, who is as nameless as the third-person narrator.59 In all four

cases the ‘I’ is an external, non-participating, narrator. Yet, this ‘I’ must be the
perceiver implied by ἐχώρει. But how do we know that both the ‘I’ of εἴρηκα and
the third person narrator is Xenophon?60

(30) (a) ἐντεῦθεν ἐξελαύνει > διὰ τῆς Ἀραβίας τὸν Εὐφράτην ποταμὸν ἐν δεξιᾷ ἔχων
σταθμοὺς ἐρήμους πέντε παρασάγγας τριάκοντα καὶ πέντε. (b) ἐν τούτῳ δὲ τῷ
τόπῳ ἦν μὲν ἡ γῆ πεδίον ἅπαν ὁμαλὲς ὥσπερ θάλαττα, ἀψινθίου δὲ πλῆρες· (c)
εἰ δέ τι καὶ ἄλλο ἐνῆν ὕλης ἢ καλάμου, ἅπαντα ἦσαν εὐώδη ὥσπερ ἀρώματα· (d)
δένδρον δ᾽ οὐδὲν ἐνῆν, θηρία δὲ παντοῖα, πλεῖστοι ὄνοι ἄγριοι, πολλαὶ δὲ στρου-
θοὶ αἱ μεγάλαι· (e) ἐνῆσαν δὲ καὶ ὠτίδες καὶ δορκάδες· (f) ταῦτα δὲ τὰ θηρία οἱ
ἱππεῖς ἐνίοτε ἐδίωκον. (g) καὶ οἱ μὲν ὄνοι, ἐπεί τις διώκοι, προδραμόντες ἕστα-
σαν· (h) πολὺ γὰρ τῶν ἵππων ἔτρεχον θᾶττον (i) καὶ πάλιν, ἐπεὶ πλησιάζοιεν
οἱ ἵπποι, ταὐτὸν ἐποίουν, καὶ οὐκ ἦν λαβεῖν, εἰ μὴ διαστάντες οἱ ἱππεῖς θηρῷεν
διαδεχόμενοι. (j) τὰ δὲ κρέα τῶν ἁλισκομένων ἦν παραπλήσια τοῖς ἐλαφείοις,
ἁπαλώτερα δέ. (k) στρουθὸν δὲ οὐδεὶς ἔλαβεν· (l) οἱ δὲ διώξαντες τῶν ἱππέων
ταχὺ ἐπαύοντο· (m) πολὺ γὰρ ἀπέσπα φεύγουσα, τοῖς μὲν ποσὶ | δρόμῳ, ταῖς δὲ 361
πτέρυξιν αἴρουσα, ὥσπερ ἱστίῳ χρωμένη. (n) τὰς δὲ ὠτίδας ἄν τις ταχὺ ἀνιστῇ
ἔστι λαμβάνειν· πέτονται γὰρ βραχὺ ὥσπερ πέρδικες καὶ ταχὺ ἀπαγορεύουσι.
(o) τὰ δὲ κρέα αὐτῶν ἥδιστα ἦν. (p) πορευόμενοι δὲ διὰ ταύτης τῆς χώρας ἀφι-
κνοῦνται > ἐπὶ τὸν Μάσκαν ποταμόν, τὸ εὖρος πλεθριαῖον. ἐνταῦθα ἦν πόλις
ἐρήμη … (X. An. 1.5.1–4)61

59 For the distinction between ‘covert’ and ‘overt’ narrators see e.g. Fludernik (2009: 21–22).
60 For the problems involved in distinguishing narrator from participating character in
Xenophon (and Thucydides) see Dorati (2008: 140–141). Ultimately, it may perhaps be
questioned whether the Anabasis, with its mixture of covert third-person and overt first-
person narrative, really exhibits Substitutionary Perception. Thus, it is worth noting that
in (29) there is no other (potential) perceiver around than the narrator himself. So nar-
rator and character coincide. On the other hand, ἐχώρει might be seen as the signal that
the narrator is no longer speaking as an external narrator but as a participating charac-
ter. See also below, on Hesiod’s Theogony. De Jong (1991: ch. 1 ‘The Messenger-Speech as
a First-Person Narrative’) provides a succinct but fundamental discussion of some impor-
tant characteristics of first-person narratives.
61 ‘Thence they marched > through Arabia, … In this region the ground was an unbroken
plain … and whatever else there was … was always fragrant like spices; trees there were
none, but wild animals …, besides bustards and gazelles. These animals were sometimes
chased by the horsemen. As for the asses … they would run on ahead and stop—for they
ran much faster than the horses—and … they would do the same thing again, and it was
impossible to catch them unless … The flesh of those that were captured was like venison,
but more tender. But no ostrich was captured by anyone, and any horseman who chased
one speedily desisted; for it would distance him … The bustards, on the other hand, can be
 158 chapter 7

This fragment exhibits basically the same uses of imperfect and omni-tem-
poral present as in (26), but on a much larger scale. Observe that in (30) some
imperfects of Substitutionary Perception are accompanied by Perception Indi-
cators: ὁμαλές and ὥσπερ in sentence (b), εὐώδη in (c), παραπλήσια and ἁπαλώ-
τερα in (j).62 Observe also that, after the hunting of ὄνοι ἄγριοι and στρουθοί has
been described in imperfects (sentences (g)–(m)), that of the ὠτίδες (bustards)
is described in omnitemporal presents, which changes back again, however, to
an imperfect of Substitutionary Perception in (n). Again, why? Might this imply
that Xenophon has not seen the hunting of the bustards with his own eyes, but
362 did taste their meat? |
So far in the majority of my examples the imperfect of SP occurred in a con-
text where also omnitemporal presents are found. To avoid the impression that
such a context is necessary for an imperfect to get an SP interpretation, I now
present two fragments of a strictly narrative nature, without any omnitempo-
ral present indicatives, taken from Thucydides’History of the Peloponnesian War
and Longus’ Daphnis and Chloe, respectively.

(31) (a) oἱ δ᾽ Ἀθηναῖοι κατὰ μίαν ναῦν τεταγμένοι περιέπλεον αὐτοὺς κύκλῳ καὶ
ξυνῆγον ἐς ὀλίγον, ἐν χρῷ αἰεὶ παραπλεόντες > καὶ δόκησιν παρέχοντες αὐτίκα
ἐμβαλεῖν· (b) προείρητο δ᾽ αὐτοῖς ὑπὸ Φορμίωνος μὴ ἐπιχειρεῖν πρὶν ἂν αὐτὸς
σημήνῃ. (c) ἤλπιζε γὰρ αὐτῶν οὐ μενεῖν τὴν τάξιν, ὥσπερ ἐν γῇ πεζήν, ἀλλὰ
ξυμπεσεῖσθαι πρὸς ἀλλήλας τὰς ναῦς καὶ τὰ πλοῖα ταραχὴν παρέξειν, εἴ τ᾽
ἐκπνεύσειεν ἐκ τοῦ κόλπου τὸ πνεῦμα, ὅπερ ἀναμένων τε περιέπλει καὶ εἰώθει
γίγνεσθαι ἐπὶ τὴν ἕω, οὐδένα χρόνον ἡσυχάσειν αὐτούς· (d) καὶ τὴν ἐπιχείρησιν
ἐφ᾽ ἑαυτῷ τε ἐνόμιζεν εἶναι, ὁπόταν βούληται, τῶν νεῶν ἄμεινον πλεουσῶν, καὶ
τότε καλλίστην γίγνεσθαι. (e) ὡς δὲ τό τε πνεῦμα κατῄει καὶ αἱ νῆες ἐν ὀλίγῳ
ἤδη οὖσαι ὑπ᾽ ἀμφοτέρων, τοῦ τε ἀνέμου τῶν τε πλοίων, ἅμα προσκειμένων
ἐταράσσοντο, καὶ ναῦς τε νηὶ προσέπιπτε καὶ τοῖς κοντοῖς διεωθοῦντο, βοῇ τε
χρώμενοι καὶ πρὸς ἀλλήλους ἀντιφυλακῇ τε καὶ λοιδορίᾳ οὐδὲν κάτηκουον οὔτε
τῶν παραγγελλομένων οὔτε τῶν κελευστῶν, καὶ τὰς κώπας ἀδύνατοι ὄντες ἐν
κλύδωνι ἀναφέρειν ἄνθρωποι ἄπειροι τοῖς κυβερνήταις ἀπειθεστέρας τὰς ναῦς

caught if …; and their flesh was delicious. Marching on through this region they arrived at
the Mascas river, which is a plethrum in width. There, in the desert, was a … city …’ (transl.
Brownson). Note, incidentally, that τὸ εὖρος πλεθριαῖον is basically ambiguous, Brownson
opts for ‘is a plethrum in width’, but its width may also have been perceived by the troops:
‘it was a plethrum in width.’
62 Words like ὥσπερ and παραπλήσια, and morphological elements like the comparative suf-
fix –τερος, will normally, I think, point to Substitutionary Perception being involved: it is
the character rather than the narrator who compares.
the imperfect as the tense of substitutionary perception 159

παρεῖχον, (f) τότε δὴ κατὰ τὸν καιρὸν τοῦτον σημαίνει, καὶ οἱ Ἀθηναῖοι προσπε-
σόντες πρῶτον μὲν καταδύουσι τῶν στρατηγίδων νεῶν μίαν, ἔπειτα δὲ καὶ τὰς
ἄλλας ᾗ χώρήσειαν διέφθειρον, καὶ … (Th. 2.84.1–3)63 | 363

The first part of this passage (sentences (a)–(d)) is due to the narrator,64 whose
gaze hits the περιπλεῖν of the Athenians,65 but contains also a Window | Opener, 364

63 ‘The Athenians, formed in line, sailed round and round them, and forced them to con-
tract their circle, by continually brushing past > and making as though they would attack
at once, having been previously cautioned by Phormio not to do so till he gave the sig-
nal. [2] His hope was that the Peloponnesians would not retain their order like a force on
shore, but that the ships would fall foul of one another and the small craft cause confusion;
and if the wind which usually rose toward morning should blow from the gulf (in expec-
tation of which he kept sailing around them), he felt sure they would not remain steady
an instant. He also thought that it rested with him to attack when he pleased, as his ships
were better sailers, and that an attack timed by the coming of the wind would tell best. [3]
When the wind came up, the enemy’s ships were now in a narrow space, and what with the
wind and the small craft dashing against them, at once fell into confusion: ship fell foul
of ship, while the crews were pushing them off with poles, and by their shouting, swear-
ing, and struggling with one another, made captains’ orders and boatswains’ cries alike
inaudible, and through being unable for want of practice to clear their oars in the rough
water, prevented the vessels to obey their helmsmen properly. At this moment Phormio
gave the signal, and the Athenians attacked. Sinking first one of the commanders’ ships
they …’ (transl. Crawley-Strassler).
64 Note the presence of the proper names Ἀθηναῖοι in (a) and Φορμίων in (b), as well as the
oblique optative ἐκπνεύσειεν in (d), indicating that Phormio’s thoughts are presented from
the perspective of the narrator; for the optative cp. Rijksbaron (2006: 53).
65 According to Bakker (1997: 37 ff.) the Greek imperfect, rather than simply referring to an
event in the past, expresses ‘the displacement of its observation to the past.’ In his view all
imperfects in Thucydides would in principle be used to create the illusion that ‘events are
observed on the spot’ (37). Bakker calls this way of presentation ‘the mimetic mode’. Thus,
when Thucydides writes about the plague he ‘is in the past’ (38); he does not ‘know’ but
‘observes’ in ‘mimetic discourse, presented from the internal standpoint of the observer,
and largely conducted with imperfect verbs.’ So far so good. But then, on p. 39, he con-
cludes: ‘The description of the plague … is a case where Thucydides’ role as an “internal
observer” is not in doubt. Other cases are less clear, and often we may doubt whether
Thucydides was on the spot when things happened’. We may indeed, the narrative of the
battle of Naupactus in book 2 being a case in point. Bakker seems to fail here to distinguish
between narrator and author. Narrators are free to observe whatever they like, and can
indeed create the illusion that they are present on the spot. But ‘illusion’ is the key word
here, reality has nothing to do with it. If the narrator describes the plague in imperfects
and if his gaze hits what is going on, to use the terminology adopted above, and if, further-
more, this narrator in this particular case happens to coincide with the author Thucydides
(but how can we know this for sure?) the description may stem from Thucydides the man.
But it is a giant step to claim that Thucydides must have been on the spot every time he
uses an imperfect, for that is what Bakker’s claim amounts to. This positivist view of the
 160 chapter 7

viz. ἐν χρῷ αἰεί παραπλέοντες: the Spartans are within seeing and hearing dis-
tance of the Athenians (and vice versa, but this is irrelevant here, although
with δόκησιν παρέχοντες αὐτίκα ἐμβαλεῖν we for a moment pass over to percep-
tion by the Spartans; note αὐτίκα). Note that the narrator knows what is going
on in Phormio’s mind: ἤλπιζε (sentence (c)) and ἐνόμιζεν (sentence (d)). But
in the long temporal-circumstantial clause (e)66 the narrator makes way for
Phormio’s Substitutionary Perception, who, ἐν χρῷ αἰεὶ παραπλέων, can see and
hear what is happening on the enemy’s side, note the presence of the Percep-
tion Indicators βοῇ, λοιδορίᾳ, τῶν παραγγελλομένων and τῶν κελευστῶν, and the
imperfects, of course:67 Phormio’s perception hits the turmoil that is develop-
ing on the other side, and reacts accordingly, as we learn in (f), for here the
narrator takes over again: it is he who is in a position to call the moment at
hand ὁ καιρὸς οὗτος and to let the reader know, by using an historical present,
that Phormio’s σημαίνειν was of crucial importance for the development of the
battle.68

(32) (a) Ἐν τῷδε τῷ ἀγρῷ νέμων αἰπόλος, Λάμων τοὔνομα, παιδίον εὗρεν > ὑπὸ μιᾶς
τῶν αἰγῶν τρεφόμενον. (b) δρυμὸς ἦν καὶ λόχμη βάτων καὶ κιττὸς ἐπιπλανώ-
365 μενος καὶ πόα μαλθακή, | καθ᾽ ἧς ἔκειτο τὸ παιδὶον. (c) ἐνταῦθα ἡ αἲξ θέουσα
συνεχὲς ἀφανὴς ἐγίνετο πολλάκις καὶ τὸν ἔριφον ἀπολιποῦσα τῷ βρέφει παρέ-
μενε. (d) φυλάττει > τὰς διαδρομὰς ὁ Λάμων οἰκτείρας ἀμελούμενον τὸν ἔριφον,
(e) καὶ μεσημβρίας ἀκμαζούσης κατ᾽ ἴχνος ἐλθὼν ὁρᾷ τὴν μὲν αἶγα πεφυλαγμέ-
νως περιβεβηκυῖαν, μὴ ταῖς χηλαῖς βλάπτοι πατοῦσα, τὸ δὲ ὥσπερ ἐκ μητρῴας

relation between language and reality must be rejected, and not only for methodological
reasons. For tradition has it that Thucydides during the greater part of the Peloponnesian
was absent from Athens, having been banned in 424, so he cannot have been ‘on the spot’,
for instance, during the preparations of the Sicilian expedition.
66 For a discussion of circumstantial clauses I may refer to Rijksbaron (1976: 112–117), where
I argued that with such clauses ‘the event of the main clause … forms a reaction to it (i.e.
to the event of the ὡς-clause); this reaction, in turn, is based upon an observation on the
part of the subject of the main clause.’
67 The present participles (ἤδη) οὖσαι and προσκειμένων belong also to Phormio’s Substitu-
tionary Perception. Interestingly, ἤδη with οὖσαι may reinforce the SP effect, like ‘now’ in
the English translation, for this is Phormio’s ‘now’: he is the deictic center of the activities
described.
68 To appreciate the difference between a narrative passage and a list-of-events (see above
pp. [7]ff.) one may compare Thucydides’ account of the battles of the Gulf of Corinth
(2.83–92), with that by Diodorus Siculus at 12.48 of his universal history. The 16 finite verbs
in this passage are all aorist indicatives: no imperfects, futures, perfects, pluperfects, sub-
junctives, optatives, or imperatives. No preparations, no projected states of affairs (plans,
orders, desire, will, request, obligation …), no direct speech, no indirect speech or thought.
In sum: a simple inventory of what Diodorus saw as the events that mattered in this battle.
the imperfect as the tense of substitutionary perception 161

θηλῆς τὴν ἐπιρροὴν ἕλκον τοῦ γάλακτος. (f) καὶ θαυμάσας, ὥσπερ εἰκὸς ἦν,
πρόσεισιν > ἐγγὺς καὶ εὑρίσκει > παιδίον ἄρρεν, μέγα καὶ καλὸν καὶ τῆς κατὰ
τὴν ἔκθεσιν τύχης ἐν σπαργάνοις κρέιττοσι· (g) χλανίδιόν τε γὰρ ἦν ἁλουργὲς
καὶ πόρπη χρυσῆ καὶ ξιφίδιον ἐλεφαντόκωπον. (Longus 1.2.1–3)69

In the first sentence of this fragment a narrator is speaking, notice the proper
name Λάμων, which is normally due to an external narrator. Παιδίον εὗρεν sum-
marizes beforehand what is thereafter narrated; εὗρεν is a so-called complexive
aorist, and functions at the same time as a Window Opener.70 Notice bare παι-
δίον: the narrator does not know the gender of the infant. In (b) the narrator
makes way for the perception of Lamon; notice the imperfects ἦν and ἔκειτο, as
well as evaluative μαλθακή. The perception by Lamon is continued with ἀφανὴς
ἐγίνετο and παρέμενε in (c). This is subsequently confirmed by | the narrator, 366
see φυλάττει in (d)71 and ὁρᾷ in (e). In sentence (f) it is still the narrator who is
speaking, note especially his comment clause ὥσπερ εἰκὸς ἦν. But with πρόσει-
σιν ἐγγύς and εὑρίσκει the narrator introduces a Window Opener, since it is of
course Lamon who is the finder, and it is Lamon who perceives that the baby
is a boy, something the narrator did not know, as we saw. This effect is ruined
by Thornley-Edmonds in the old Loeb translation, which is otherwise, too, very
bad, for they translate in sentence (a): ‘found one of his goats suckling an infant-
boy.’ Henderson, in the new Loeb translation, correctly has: ‘found a baby.’ The
description of the boy in sentences (f) and (g), which is full of evaluative ele-
ments, stems also from Lamon, although τὴν ἔκθεσιν may be due to the narrator
(unless a shepherd may have concluded immediately, upon finding the apparel
of a boy, that he was dealing with a case of child abandonment.)

69 ‘While pasturing his flock on this estate a goatherd named Lamo found > a baby being
suckled by a she-goat. It was a copse … where the baby lay. The goat kept running out of
sight in this direction … and stayed with the newborn. Lamo kept an eye on these com-
ings and goings and … he tracked her and saw the goat standing over the child carefully
while it sucked her flowing milk as if from a mother’s breast. … Naturally he was amazed
and coming > closer he discovered > a male child, large and handsome, and in swaddling
clothes too fine for a foundling’s condition: for there was …’ (transl. Henderson).
70 This use of εὗρεν is already found in Homer, see e.g. Od. 1.103–112: στῆ δ᾽ … | οὐδοῦ ἐπ᾽
αὐλείου· … | … | εὗρε > δ᾽ ἄρα μνηστῆρας ἀγήνορας· οἱ μὲν ἔπειτα | … ἔτερπον, … | κήρυκες δ᾽
… | οἱ μὲν ἄρ᾽ οἶνον ἔμισγον … | οἱ δ᾽ αὖτε … | νίζον καὶ πρότιθεν, τοὶ δὲ κρέα πολλὰ δατεῦντο.
Ameis-Hentze’s note on οἱ μὲν ἔπειτα is worth quoting again (cp. fn. 52): ‘die alsdann, von
dem weiter Wahrgenommenen, nach dem allgemeinen εὗρεν κτἑ.’ See also Od. 24.226ff.
and, with τέτμεν, 5.58 ff.
71 Φυλάττει is remarkable, for in principle the present indicative of durative-stative verbs
cannot be used as a historical present. For details see Rijksbaron (2006: 130).
 162 chapter 7

I conclude this part with two texts where commentators and translators have
seriously misunderstood the text by not recognizing the Substitutionary Per-
ception function of the imperfect.
First, the proem of Hesiod’s Theogony.

(33) Μουσάων Ἑλικωνιάδων ἀρχώμεθ᾽ ἀείδειν,

αἵ θ᾽ Ἑλικῶνος ἔχουσιν ὄρος μέγα τε ζάθεόν τε,
καί τε περὶ κρήνην ἰοειδέα πόσσ᾽ ἁπαλοῖσιν
ὀρχεῦνται καὶ βωμὸν ἐρισθενέος Κρονίωνος
5 καί τε λοεσσάμεναι τέρενα χρόα Περμησσοῖο
ἠ᾽ Ἳππου κρήνης ἠ᾽ Ὀλμειοῦ ζαθέοιο
ἀκροτάτῳ Ἑλικῶνι χοροὺς ἐνεποιήσαντο,
καλοὺς ἱμερόεντας, ἐπερρώσαντο δὲ ποσσίν.
10 ἔνθεν ἀπορνύμεναι κεκαλύμμέναι ἠέρι πολλῷ
367 ἐννύχιαι στεῖχον περικαλλέα ὄσσαν ἱεῖσαι (Hes. Th. 1–10)72 |

After omnitemporal ἔχουσιν in line 2, ὀρχεῦνται in 4, and gnomic (= omnitempo-

ral) ἐνεποιήσαντο and ἐπερρώσαντο in lines 7 and 8, which are due to the omni-
scient singer referred to in l. 1 and must represent impersonal, conventional
knowledge, the next finite verb is the imperfect στεῖχον. The change from the
omnitemporal tenses of lines 1–8 to this imperfect has baffled many scholars,
has sparked a great deal of controversy and led to positively desperate solu-
tions. To bring the imperfect in line with the preceding verbs some scholars,
e.g. West in his commentary ad loc., did not hesitate to postulate the existence
of a ‘timeless’ imperfect, which temporally would not differ from the present
and aorist indicatives of ll. 2–8.73 I will refrain from discussing this bizarre view,
and I can do this the more easily because West himself, in a much later publi-
cation (West 1989), without discussing his earlier view, abandoned it, only to
exchange it for another untenable view. In West (1989) he considers στεῖχον an
instance of a non-past, so-called ‘injunctive’, use of augmentless forms with sec-
ondary endings. But this injunctive use does not exist. Cp. Beekes (1995: 245):

72 ‘Let us begin to sing from the Heliconian Muses, who possess the great and holy moun-
tain of Helicon and dance on their soft feet around the violet-dark fountain and the altar
of Cronos’ mighty son. And after they have washed their tender skin in Permessus or Hip-
pocrene or holy Olmeius, they perform choral dances on highest Helicon, beautiful, lovely
ones, and move nimbly with their feet. Starting out from there, shrouded in thick invisi-
bility, by night they walked, sending forth their very beautiful voice …’ (Translation Most,
except for στεῖχον in l. 10).
73 Likewise e.g. Evelyn-White, Loeb (1914): ‘Thence they arise and go abroad by night …’ and
Most, Loeb (2006): ‘… by night they walk …’
the imperfect as the tense of substitutionary perception 163

‘In Greek the injunctive forms are completely equal to the indicative forms …;
it thus has no sense to speak here of an injunctive (because there is no meaning
component). Only in Sanskrit can one speak of an injunctive’.
In a recent article (Rijksbaron 2009 [ch. 9 in this volume]) I have proposed
to take στεῖχον as an imperfect of Substitutionary Perception (or focalisation,
the term used in that article), arguing that this imperfect is both the sign that
a narrative begins, and that the στείχειν is presented from the point of view
of a character rather than that of the narrator,74 since the Muses could be
heard while they were moving; note the Perception Indicator ὄσσαν ἱεῖσαι. The
question now of course is: Who is this character? At lines 22–23 the narrator
mentions Hesiod (αἵ νύ ποθ᾽ Ἡσίοδον καλὴν ἐδίδαξαν ἀοιδήν, | ἄρνας ποιμαίνονθ᾽
Ἑλικῶνος | ὕπο ζαθέοιο),75 from which we must infer that he was the charac- 368
ter perceiving the στείχειν of the Muses. Next, this perceiving character turns
out to be none other than the narrator himself: με … θεαὶ πρὸς μῦθον ἔειπον, 24:
here, the third-person narrative of lines 10–23 becomes a first-person narra-
tive and thus the report of a personal experience. Eventually, then, narrator,
perceiving character and ‘I’ coincide, as in text (29) from Xenophon, discussed
above.
The perception by the ‘I’, Hesiod, returns later in the proem, viz. at lines 68–
70:

(34) αἳ τότ᾽ ἴσαν πρὸς Ὄλυμπον, ἀγαλλόμεναι ὀπὶ καλῇ,

ἀμβροσίῃ μολπῇ· περὶ δ᾽ ἴαχε γαῖα μέλαινα
ὑμνεύσαις, ἐρατὸς δὲ ποδῶν ὕπο δοῦπος ὀρώρει (Hes. Th. 68–70)76

Notice the abundance of Perception Indicators: ὀπὶ καλῇ, ἴαχε, ὑμνεύσαις, δοῦπος
ὀρώρει.

Next, the myth from Plato’s Phaedrus.

In the myth of Plato’s Phaedrus, more specifically in the passage 253e5–
255a1, the omnitemporal, habitual behaviour of the horses of the soul is de-

74 Or perhaps rather the singer of lines 1–8, for strictly speaking there has so far been no
narrative.
75 This must mean that on some occasion during their descents the Muses gave Hesiod
instruction; indeed ‘descents’, for στεῖχον must have iterative meaning. For details I refer
to Rijksbaron (2009).
76 ‘They then went towards Olympus, exulting in their beautiful voice, and around them the
black earth resounded as they sang, and from under their feet a lovely din rose up’ (transl.
Most, adapted).
 164 chapter 7

scribed by means of present and (gnomic) aorist indicatives.77 But in between

369 there is, at 254d6, just one striking imperfect, ἦσαν, in a temporal clause. |

(35) (253e5) ὅταν δ᾽ οὖν ὁ ἡνίοχος … ὑποπλησθῇ, ὁ μὲν εὐπειθὴς … τῶν ἵππων, ἀεί τε
καὶ τότε αἰδοῖ βιαζόμενος, ἑαυτὸν κατέχει … ἐντρέπεται, … φέρεται, … ἀναγκά-
ζει … ἀντιτείνετον (254b1) … πορεύεσθον …· καὶ (254b4) πρὸς αὐτῷ τ᾽ ἐγένοντο
καὶ εἶδον … ἠνέχθη, καὶ πάλιν εἶδεν … ἔδεισε … ἀνέπεσεν … ἠναγκάσθη …
(254c5) ἔβρεξε … ἐλοιδόρησεν … (254d2) μόγις συνεχώρησεν δεομένων εἰς
αὖθις ὑπερβαλέσθαι. ἐλθόντος δὲ τοῦ συντεθέντος χρόνου … βιαζόμενος, χρεμε-
τίζων, ἕλκων (254d5) ἠνάγκασεν αὖ προσελθεῖν τοῖς παιδικοῖς ἐπὶ τοὺς αὐτοὺς
λόγους, καὶ ἐπειδὴ ἐγγὺς ἦσαv, ἐγκύψας καὶ ἐκτείνας τὴν κέρκον, ἐνδακὼν τὸν
χαλινόν, μετ᾽ ἀναιδείας ἕλκει· ὁ δ᾽ (254e1) ἡνίοχος … τήν τε κακηγόρον γλῶτταν
καὶ τὰς γνάθούς καθῄμαξεν καὶ τὰ σκέλη … ὀδύναις ἔδωκεν. ὅταν δὲ … λήξῃ,
ταπεινωθεὶς ἕπεται ἤδη τῇ τοῦ ἡνιόχου προνοίᾳ, καὶ ὅταν ἴδῃ τὸν καλόν, φόβῳ
διόλλυται· etc. (Pl. Phdr. 253e5–255a1)78

Like στεῖχον in (32), ἦσαν has often been misunderstood, and been translated
as if it were a present indicative or a gnomic aorist. Here follow some faulty
370 English translations: |

77 Although this is a description of non-past, omnitemporal and therefore iterative (mythi-

cal) events rather than a narrative sensu stricto the internal make-up of the passage shares
important features with ‘real’ narratives, notably the presentation of the behaviour of the
horses of the soul as a series of successive events in a context of circumstances (see above,
p. [345]); notice especially the presence of temporal clauses at 253e5 and 254e7, and the
combination of ἀεί and τότε at 253e5, where the omnitemporal (ἀεί) events are also par-
ticularized (τότε), as in a ‘real’ narrative.
78 ‘Now when the charioteer … is full of …, the horse that is obedient, constrained then as
always by modesty, controls himself … but the other no longer heeds … but springs for-
ward … and forces … they pull back … (254b1) they go forward … and (254b1) they come to
him and behold … his (the charioteer’s) memory is borne back and he sees it again … he
is afraid … and falls backward … he is forced … (254c5) one horse wets its soul with sweat
… breaks forth into angry reproaches … (254d2) he hardly yields to their prayer that he
postpone the matter to another time. Then when the time comes which they have agreed
upon … struggling, and neighing, and pulling (254d5) he forces them again with the same
purpose to approach the beloved one, and now that they were near him, he lowers his
head, raises his tail, takes the bit in his teeth, and pulls shamelessly. As for the charioteer
(254e1) … he covers his scurrilous tongue and jaws with blood, and … causes much pain
to his legs. Now when the (bad horse) has ceased from his unruliness, he is humbled and
follows henceforth the wisdom of the charioteer, and when he sees the beautiful one, he
is overwhelmed with fear.’ (transl. Fowler, with adaptations).
the imperfect as the tense of substitutionary perception 165

(36) 254d5f.; emphasis added

Jowett (18923): (the unruly horse) ‘… forces them [i.e. the other horse and
the charioteer] to draw near again. And when they are near he stoops his
head …, and … pulls shamelessly’

Fowler (Loeb, 1914): ‘… he forces them again … to approach the beloved

one, and when they are near him, he lowers his head … and pulls shame-
lessly’,

Hackforth (1952): ‘and when they have come close …’

Rowe (1986; he has no note on ἦσαν): ‘… it forces them to approach the

beloved again …, and when they are nearby … it pulls shamelessly.’

‘When … are’ or ‘have come’ are of course impossible: to have ‘when’ in English,
the Greek should have had ὅταν or ἐπειδάν + subjunctive, as indeed earlier
and later in the myth. Robin, however, correctly—and brilliantly—translates
(Budé, 1961), avoiding the use of a verb in his translation and using voilà instead:
‘… une fois de plus il les a contraints d’approcher du bien-aimé … Enfin, main-
tenant que les voilà à proximité, il se penche en avant sur lui … il tire sans ver-
gogne.’ The imperfect all of a sudden, dramatically, transfers the reader from the
general description of the behaviour of the ‘good’ horse and the charioteer to
the point of view of the ‘bad’ horse. Having forced the other two to approach the
beloved one (ἠνάγκασε προσελθεῖν: they have come close (aorist) to the beloved
one) the ‘bad’ horse sees that they are there, and reacts accordingly: ‘now/see-
ing that they were close, it lowers its head and … pulls shamelessly.’ Note also,
incidentally, the presence of the Perception Indicator ἐγγύς.

Epilogue: an Example from Latin

Perhaps inspired by the Homeric passage on Odysseus’ seeing and taking in

the palace of Alcinous, the Vergilian narrator tells us about a similar visual
experience of Aeneas in Carthage, in a similar mixture of narrative and Sub-
stitutionary Perception; moreover, Aeneas, too, is covered by a cloud. | 371

(37) ‘o fortunati, quorum iam moenia surgunt!’

Aeneas ait et fastigia suspicit urbis.
infert > se saeptus nebula (mirabile dictu)
 166 chapter 7

440 per medios, miscetque > uiris neque cernitur ulli.

Lucus in urbe fuit media, laetissimus umbrae,
quo primum iactati undis et turbine Poeni
effodere loco signum, quod regia Iuno
monstrarat, caput acris equi; sic nam fore bello
445 egregiam et facilem uictu per saecula gentem.
hic templum Iunoni ingens Sidonia Dido
condebat, donis opulentum et numine diuae,
aerea cui gradibus surgebant limina nexaeque
aere trabes, foribus cardo stridebat aënis.
450 hoc primum in luco nova res oblata timorem
leniit, hic primum Aeneas sperare salutem
ausus et adflictis melius confidere rebus.
namque sub ingenti lustrat dum singula templo,
372 reginam opperiens, … (Verg. Aen. 1.437–454)79 |

After Aeneas, climbing the hill qui plurimus urbi | imminet (419–420), has
admired (miratur … Aeneas, 420: narrated perception) the buildings and the
building activities of the Carthaginians, and has exclaimed how fortunate they
are quorum iam moenia surgunt (437, direct speech of Aeneas, implying a per-
ception and ‘overheard’ by the narrator), he mingles with the locals (439–440);
both infert and miscet function as Window Openers. The whole of lines 418–
440 is due to the narrator, note the presence of Aeneas both at the beginning
and the end of this passage. The indentation in line 441 suggests that there is a
narrative break at this point, but in reality Aeneas continues to be present as a
perceiver, for with the imperfects condebat (447), surgebant (448) and stridebat
(449) the narrator makes way for Aeneas’ Substitutionary Perception, which is
made possible by the Window Openers of ll. 439–440.80 Aeneas, who is hid

79 ‘Veiled in a cloud, he enters >—wondrous to tell—through their midst, and mingles >
with the people, seen by none!
Amid the city was a grove, luxuriant in shade, the spot where first the Phoenicians,
tossed by waves and whirlwind, dug up the token which queenly Juno had pointed out, a
head of the spirited horse, for thus was the race to be famous in war and rich in substance
through the ages. Here Sidonian Dido was founding to Juno a mighty temple, rich in gifts
and the presence of the goddess. Brazen was its threshold uprising on steps; bronze plates
were its lintel beams, on doors of bronze creaked the hinges. In this grove first did a strange
sight appear to him and alloy his fears; here first did Aeneas dare to hope for safety and
put surer trust in his shattered fortunes. For while beneath the mighty temple, awaiting
the queen, he scans each object, …’ (transl. Fairclough-Goold, Loeb).
80 On condebat Conway and Austin confine themselves to noting: ‘“was building”; the work
the imperfect as the tense of substitutionary perception 167

by a cloud and must not be detected, is of course no longer in a position to

comment directly, in exclamations, on what he sees. Note, incidentally, the
correspondence surgunt (437): direct perception, with surgebant (448): Sub-
stitutionary Perception. That Aeneas is indeed still present is confirmed by the
narrator, who takes over again in lines 450–454 (note Aeneas) and tells us that
Aeneas ‘beneath the mighty temple, awaiting the queen, scans each object’.
Note, finally, that Aeneas and the narrator share their view of the size of the
temple: see templum … ingens in 446 and ingenti … templo in 453.

Conclusion

With this fragment from the Aeneid we have come full circle. Examples (1)–(3)
showed a report of direct speech implying a perception (ex. 1), narrated per-
ception (ex. 2) and Substitutionary Perception (ex. 3). We find these modes of
narrating in fragment (37) combined in one passage: report of direct speech
implying a perception at l. 437, narrated perception at ll. 418–440 and 450 ff.,
and Substitutionary Perception at ll. 447–449. This and the other texts | dis- 373
cussed in this paper show that the stylistic device known as Substitutionary
Perception, which resembles the focalisation of narratology but has a more
powerful explanatory potential, and is a frequent feature of modern narra-
tive texts, was also known to Greek (and Latin) narrative texts. It enables the
narrator to substitute the perceptions of one of his characters for his own per-
ceptions, thereby giving this character the opportunity to show himself what he
sees of the world around him, as ‘marks and measures of human consciousness’
(McHale), rather than to ‘rely’ on the narrator to tell us what he, the charac-
ter, saw. In this way SP, like free indirect speech and thought, would seem to
highlight the cognitive involvement of a character in the events he is experi-
encing, the purpose being, we must assume, to enhance the reader’s interest in,
and understanding of, the character. Both in Greek and in modern languages
its function is crucially linked with the use of imperfective, unbounded tenses
like the imperfect in Greek, the French imparfait and the (past) progressive in
English. Since these tenses denote durative, ongoing states of affairs they can be
‘hit’ by the gaze of some character, if he happens to be present in the narrative at
that particular point; if not, it is the narrator who perceives (and narrates) what
is going on in the world of the narrative. Bounded, complete states of affairs,

was not yet complete, though far advanced’ (Conway, confusing reality and literature), and
‘work was still in progress. But the goddess’ presence was already in the place, glorifying
it.’ (Austin).
168 chapter 7

on the other hand, expressed by perfective tenses like the aorist in Greek, the
passé simple in French and the simple past in English, can only be perceived by
the narrator, as a complete whole: they do not permit the intrusion of the gaze
of a character (nor that of the narrator, for that matter). The perception by a
character is often announced by what may be called a Window Opener, e.g. a
verb of entering, looking or finding, or a phrase that points to spatial closeness
(e.g. near and equivalents).
Also, and not unimportantly, in some cases where an imperfect has caused
serious interpretive difficulties (Hesiod, proem of the Theogony; Plato, the myth
of the Phaedrus) an analysis which makes use of the concept of SP can provide
a satisfactory solution to these difficulties.

References

Alavoine, B., Les enquêtes de Maigret de Georges Simenon: Lecture des textes (Amiens
1999).
Altman, R., A Theory of Narrative (New York 2008).
Bakker, E.J., ‘Verbal Aspect and Mimetic Description in Thucydides’, in E.J. Bakker (ed.),
Grammar as Interpretation: Greek Literature in its Linguistic Contexts (Leiden 1997)
1–51.
Bal, M., Narratology, 3rd ed. (Toronto 2009).
Bartoněk, A., Handbuch des mykenischen Griechisch (Heidelberg 2003).
Beekes, R.S.P., Comparative Indo-European Linguistics (Amsterdam/Philadelphia 1995).
Brinton, L., ‘“Represented Perception”: A Study in Narrative Style’, Poetics 9 (1980) 363–
381.
Bronzwaer, W.J.M., ‘Mieke Bal’s Concept of Focalisation: A Critical Note’, Poetics Today
2 (1981) 193–201.
Comrie, B., Aspect (Cambridge 1976).
Dahl, Ø., Tense and Aspect Systems (Oxford 1985).
Dorati, M., ‘Considerazioni sulla focalizzazione e sul narratore omnisciente nel rac-
conto storico’, Studi italiani di filologia classica 101 (2008) 133–193.
Dowty, D., ‘Tenses, Time Adverbials, and Compositional Semantic Theory’, Linguistics
and Philosophy 5 (1982) 23–55.
Fehr, B., ‘Substitutionary Narration and Description: A Chapter in Stylistics’, English
Studies 20 (1938) 97–107.
Fludernik, M., The Fictions of Language and the Languages of Fiction (London/New York
1993).
Fludernik, M., An Introduction to Narratology (London/New York 2009).
Forster, E.M., Aspects of the Novel (London 1974).
the imperfect as the tense of substitutionary perception 169

Jong, I.J.F. de, Narrative in Drama (Leiden 1991).

Jong, I.J.F. de, A Narratological Commentary on the Odyssey (Cambridge 2001).
Lips, M., Le style indirect libre (Paris 1926).
Mayser, E., Grammatik der griechischen Papyri aus der Ptolemäerzeit, II: Satzlehre (Ber-
lin/Leipzig 1926).
McHale, B., ‘Free Indirect Discourse: A Survey of Recent Accounts’, Poetics and Theory
of Literature 3 (1978) 249–287.
Mulder, W. de, Vetters, C., ‘The French Imparfait, Determiners and Grounding’, in F. Bris-
ard (ed.), Grounding: The Epistemic Footing of Deixis and Reference (Berlin/New York
2002) 113–150.
Palmer, A., Fictional Minds (Lincoln/London 2004).
Ricœur, P., Temps et récit I–III. (Paris 1983–1985).
Rijksbaron, A., Temporal and Causal Conjunctions in Ancient Greek: With Special Refer-
ence to the Use of ἐπεί and ὡς in Herodotus (Amsterdam 1976).
Rijksbaron, A., ‘The Discourse Function of the Imperfect’, in A. Rijksbaron, H.A. Mulder
and G.C. Wakker (eds), In the Footsteps of Raphael Kühner (Amsterdam 1988) 237–
254 [ch. 3 in this volume].
Rijksbaron, A., The Syntax and Semantics of the Verb in Classical Greek, 3rd ed. (Chicago/
London 2006).
Rijksbaron, A., ‘Discourse Cohesion in the Proem of Hesiod’s Theogony’, in S.J. Bakker,
G.C. Wakker (eds), Discourse Cohesion in Ancient Greek (Leiden 2009) 241–267 [ch. 9
in this volume].
Smith, C.S., ‘Accounting for Subjectivity (Point of View)’, in R.P. Meier et al. (eds), Text,
Time and Context: Selected Papers of Carlota S. Smith (Dordrecht etc. 2009) 371–395
(origin. publ. in 2002).
Toolan, M.J., Narrative: A Critical Linguistic Introduction (London 1988).
Toolan, M.J., The Stylistics of Fiction: A Literary-Linguistic Approach (London/New York
1990).
Weinrich, H., Tempus: besprochene und erzählte Welt, 2nd ed. (Mainz 1971).
West, M.L. (ed.), Hesiod: Theogony (Oxford 1966).
West, M.L., ‘An unrecognized injunctive usage in Greek’, Glotta 67 (1989) 135–138.
 chapter 8

How Does a Messenger Begin His Speech? Some

Observations on the Opening Lines of Euripidean
Messenger Speeches

Abstract

This chapter investigates the opening lines of Euripidean messenger speeches. A sub-
stantial portion of them start with an ἐπεί-clause. It is demonstrated that, being pre-
supposition-triggers, these clauses, with few exceptions, contain information that is
already familiar to both the speaker and the addressee, referring back to something that
was said before or that belongs to the general background knowledge presupposed by
the play. By contrast, those messenger speeches which start with a simple decalarative
clause do not pick up information already presented earlier, but either elaborate on
the setting (time and place) of the ensuing narrative or immediately present a piece of
news. The question remains how a messenger knows that he can take certain informa-
tion for granted and start his speech with an ἐπεί-clause. It is argued that in a number
of cases it is certain or highly likely that both the messenger and the addressee were on-
stage and present at the event to which the ἐπεί-clause refers, though in a number of
other cases it must be assumed that the messenger is by convention allowed to suppose
that the addressee is acquainted with the information he presents.

1 Introduction

All Euripidean tragedies except the Troades have one or more so-called ‘mes-
senger speeches’.*,1 In these a character who is generally called an ἄγγελος ‘mes-
senger’ in the mss, but may also be a θεράπων or some other person, gives some
kind of report of off-stage events.2

* I am indebted to J.M. Bremer for his criticism on a number of points.

1 I have not discussed the Cyclops, a satyr-play, and the Rhesus, since it is probably not by
Euripides, although both plays do have a messenger-speech, vv. 382ff. and 762ff., respec-
tively.—Talthybius’ role in the Troades is primarily that of a herald, who transmits messages
from the Greek camp to the Trojan women; cf. 235–277; 709–739; 1260–1271. Vv. 1123–1155 are
closer to a messenger-speech, in that they form a report of off-stage events.
2 Barrett aptly remarks, in his commentary ad 1151 of the Hippolytus: ‘The conventional trans-
lation ‘messenger’ is unfortunate: the ἄγγελος brings not a message but news’.

© koninklijke brill nv, leiden, 2019 | doi:10.1163/9789004386129_010

how does a messenger begin his speech? 171

These scenes may be analysed into the following scheme: (1) a messenger
enters the stage; his arrival has sometimes been announced by one of the per-
sons already present on the stage with some appropriate expression, ‘there I
see someone coming who will bring us news on such-and-such a topic’; (2)
he speaks one or more introductory lines that arouse the attention of the
hearer(s), who ask him to go on; (3) the messenger proceeds by giving the main
point of his news whereupon he is asked to tell his story in detail; (4) he takes
a deep breath and tells the story proper. The way in which he begins this story
will form the subject of my contribution. | 294

2 Survey of Content

Of a total of 24 messenger speeches in the tragedies,3 15 open, either imme-

diately or after a number of lines which do not form part of the story, with a
temporal subordinate clause introduced by ἐπεί;4 the other nine start with a
simple sentence. This difference in form reflects, as I will try to show, a differ-
ence in content. My argument proceeds in the following manner: in section 3 I
describe three characteristics of ἐπεί-clauses at the beginning of a messenger-
speech, and discuss two of them more fully. Section 4 is devoted to speeches
without an ἐπεί-clause, and in section 5 I return at greater length to the third
characteristic of ἐπεί-clauses.5

3 Messenger Speeches That Open with an ἐπεί-clause

3.1 List of Plays

The following plays have one or more messenger speeches that open with
an ἐπεί-clause: Alcestis (158ff.),6 Medea (1136ff.), Heraclidae (800 ff.), Andro-
mache (1085ff.), Heracles (923ff.), Ion (1122ff.), Electra (774 ff.), Iphigenia in Tau-

3 These include IA 414 ff. and 1543 ff.; the latter speech is probably spurious from v. 1568 onwards,
but I include the beginning in my discussion. On IA 414 ff., see below; I think there are 4 mes-
senger speeches in Phoenissae: see below, note 13.
4 Kamerbeek, in his commentary ad S. OT 1241, draws attention to similar openings in Sopho-
cles with temporal ὅπως.
5 Erdmann (1964) does not discuss the opening lines of the messenger speeches. He says (1964:
95 n. 5) that 13 speeches open with ἐπεί, omitting Ba. 1043 and HF 923; on the latter cf. my
note 11 below).
6 Curiously enough, Grube (1961: 25) says that the Alc. has no messenger-speech, but for Her-
akles’ one-line report in 1140.
 172 chapter 8

ris (261ff. and 1327ff.), Helen (1526ff.), Phoenissae (1090 ff., 1359 ff. and 1428 ff.),
Bacchae (1043ff.), Iphigenia at Aulis (1543ff.). Now it is a characteristic of ἐπεί-
clauses, and of temporal subordinate clauses in general, that when they occur
in narrative passages, they refer back to information that has been given
before.7 This may be briefly illustrated by the following English example. When
someone says (or writes), while telling a story: After John came home …, this
will normally follow an earlier statement about John leaving some place to go
295 home, or the like.8 Naturally, for such an utterance to | be meaningful, it has
to satisfy another requirement: the person addressed (henceforth called the
addressee) must be acquainted with the event the temporal clause refers to.
Since the ἐπεί-clauses under discussion occur, typically, in narrative texts three
questions arise: (1) do these clauses refer back and, if so, to what? (2) Is the
addressee in fact acquainted with the event referred to? In the case of the mes-
senger speeches there is a further and important requirement that has to be
satisfied: (3) the messsenger should, in principle, know where to start his story;
or, to put it differently, how does the messenger know when he begins his story
that his addressee is acquainted with the event he is referring to? To this point
I will come back in section 5. I will now discuss our ἐπεί-clauses with regard to
the first two points.

3.2 What Do the ἐπεί-clauses Refer to? Is the Addressee Acquainted with
the Event Referred to?
Alcestis 158ff.
The servant speaks to the chorus: ‘For when she understood the fatal day was
come …’9 Now it is said several times, in the introductory scene that precedes,
that this day will be Alcestis’ last (e.g. 19ff. (Apollo to the audience), 73–74
(Thanatos to Apollo)), but it has nowhere been stated that Alcestis herself
was acquainted with this fact. Thus, the reference of the ἐπεί-clause is not to
some explicit mention of Alcestis realizing that her time has come. As, how-
ever, the entire play is based on the assumption that the present day is Alcestis’
last, everyone, Alcestis included, is presumed to know this. In fact the cho-

7 See my thesis (Rijksbaron 1976).

8 Or, alternatively, it may be a more or less literal repetition of something said before, e.g. At six
o’clock John came home. After he came home … The first type of clause may be called continu-
ative, in that it gives the sequel to an earlier event, the latter resumptive: no new information
is added. See Rijksbaron 1976.
9 The English translations are those of The Complete Greek Tragedies, ed. by D. Grene and
R. Lattimore (third ed., Chicago-London 1969), vol. III–IV, by abbreviation: ‘the Chicago-
translation’. At Ion 1122 and Ba. 1043 where the Chicago-translation has no temporal conjunc-
tion, I have added when.
how does a messenger begin his speech? 173

rus also knows it, as appears from e.g. 79ff., where they express their fear that
Alcestis may be already dying. Miss Dale has noticed the difficulty, witness her
commentary on line 158: ‘we are not told how’ (scil. Alcestis ᾔσθετο … ἥκου-
σαν).

Medea 1136ff.
The messenger, to Medea: ‘When those two children, born of you, had entered
…’ The reference is to 974ff., where Medea sends away her children to the bride’s
house. Here, then, the addressee is better acquainted with the event the mes-
senger is referring to than anyone else, | since it is Medea herself who sent the 296
children away.10

Heraclidae 800ff.
The servant, to Alcmene: ‘When we had drawn our own troops up …’ The ἐπεί-
clause refers to 709ff. where Iolaus, in the presence of Alcmene, sets out to fight
the Argives (cf. also n. 25).

Andromache 1085ff.
The messenger, apparently a servant of Neoptolemus (cf. 1071, 1085 (ἤλθομεν),
1100 (ἡμεῖς)), to Peleus: ‘When we arrived in Apollo’s famous territory …’. As in
the Alcestis, the reference is here to an event that lies outside the play, the jour-
ney of Neoptolemus to Delphi. Neither his departure for Delphi nor his arrival
there are enacted on the stage, but everybody present is supposed to know of
his mission; cf. e.g. 50ff., where Andromache mentions it. Peleus, whom the
messenger is addressing, also knows it, as appears from e.g. 558.

Heracles 923ff.
Messenger to the chorus: ‘Offerings to Zeus were set before the hearth to purify
the house, after Heracles had cast the body of the king outside’.11 The reference
of the ἐπεί-clause, and especially of κτανών, is to 754ff., where Lycus’ death-cries
from within the palace are overheard by the chorus.

10 Medea already knows that her children had arrived at the bride’s house, from the words
of the Paidagogos (1002 ff.). But, of course, the messenger does not know that she knows.
11 After is mine, the Chicago-translation has for. This is not implausible in itself, but I think
that the ἐπεί-clause has precisely the same function here as it has elsewhere, when it pre-
cedes the main clause. This ‘dislocation’ of the ἐπεί-clause may be due to the contrastive
μὲν … δέ structure of the passage as a whole.
 174 chapter 8

Ion 1122ff.
Servant to the chorus: ‘When Creusa’s husband came out from the shrine’. The
reference is to the second episode: the chorus was present when Xuthus came
out of the temple (517) and left the sanctuary together with Ion in order to pre-
pare the banquet (651–668).

Electra 774ff.
Messenger to Electra: ‘When we rose from your cottage and walked down the
hill’. The reference is to 620ff., especially to 634 where the Old Man and Elec-
tra urge Orestes to set off. After v. 698 Orestes does in fact leave in search of
Aegisthus. Electra has been present throughout, so she knows that Orestes
297 went. |

Iphigenia in Tauris 261ff. and 1327ff.

First speech: A cowherd to Iphigenia: ‘Just as we drove our cattle from the
woods to that long hollow where the curving tide has cut away the cliff …’
This speech differs from the others, in that the reference of the ἐπεί-clause is
to some information that has been given shortly before by the herdsman him-
self, viz. to 252–255, where he tells Iphigenia that he found some strangers on
the shore. The ἐπεί-clause, then, does not continue a story (an event) that has
been broken off earlier.12 Second speech: Messenger to Thoas: ‘It was just when
we had reached the bend of shore …’, which refers to 1192ff., where Iphigenia
announces to Thoas that she will go to the sea to perform purificatory rites.
After v. 1233 she leaves the stage.

Helen 1526ff.
Messenger to Theoclymenus: ‘After Zeus’ daughter left the palace here, and you,
and was escorted to the sea …’. This refers to 1399ff., where Helen, in the pres-
ence of Theoclymenus, says she will go to the sea. After line 1450 she does go,
with Menelaus.

Phoenissae 1090ff., 1359ff., and 1428ff.

First speech. Messenger to Jocasta: ‘When Creon’s son, who died to save the
city, on the highest tower standing had thrust his sword through his own throat
…’. The ἐπεί-clause refers to 994ff., where Menoeceus says that he will sacrifice
himself to save his country. However, he is addressing the chorus, and Jocasta

12 This would have been virtually impossible, since the messenger-scene occurs so early in
the play that it has not been preceded by any action to which the herdsman could possibly
have referred.
how does a messenger begin his speech? 175

is not present at all. How does Jocasta know about Menoeceus’ death? We may
return to this problem later (see p. [305] below).
Third speech:13 Messenger to Creon: ‘When they were ordered in their shin-
ing arms, the two young sons of Oedipus the old …’. The reference is probably to
the messenger’s own words in 1242–1243: ἔκρυπτον σῶμα παγχάλκοις ὅπλοις etc.
(cf. p. [305]). These words, however, are spoken to Jocasta, not to Creon, who
does not enter the stage until vv. 1308–1309. That I think is why vv. 1356–1357
are added; see the further discussion on p. [305]. | 298
Fourth speech:14 Messenger to the chorus: ‘Just as her fallen children left this
life …’. The reference is to the immediately preceding 1424, from the messenger’s
own narrative: γαῖαν δ’ ὀδὰξ ἑλόντες … πίπτουσιν ἄμφω.

Bacchae 1043ff.
Messenger to the chorus: ‘When, leaving behind us the last outlaying farms of
Thebes, we forded the Asopus …’. This refers to the departure of Pentheus and
the Stranger, after lines 972 and 976, respectively, in the presence of the chorus.

Iphigenia at Aulis 1543ff.

Messenger to Clytaemnestra: ‘When we came to Artemis’ grove …’, The refer-
ence is to the departure of Iphigenia after v. 1508, when of course Clytaemnestra
was present.15

3.3 Conclusion
A positive answer can thus be given to both questions in section 3.1, viz. (1)
do ἐπεί-clauses at the opening of a messenger-speech refer back? and (2) is
the addressee acquainted with the event referred to? As to the first question,
there are only three exceptions: in Alcestis and Andromache the ἐπεί-clause
does not refer back to some specific earlier passage; rather, the reference con-
cerns something that lies outside the play, something that is one of the play’s
presuppositions. Again, in the first speech of Iphigenia in Tauris the reference is
not to an earlier event, but to the herdsman’s own recent words. As for the sec-
ond question, there are two exceptions: in Phoenissae I and III, the addressee

13 I consider that the second speech is at v. 1207 ff. Although the setting of what is told in
vv. 1217–1267 does not differ from that of vv. 1067ff., and their subjects are related, this
speech does tell a new story. Erdmann (1964: 114 ff.) also distinguishes four speeches in the
Phoenissae.
14 This is properly speaking the sequel to lines 1356–1424. But again the subject changes.
15 The messenger’s failure to make known in the preceding lines the main burden of his
news, that Iphigenia is safe, is exceptional.
 176 chapter 8

was not present at the event the messenger is referring to. So she must have
arrived at the knowledge of that event in a different way. Since this point is
closely linked with the question of how the messenger knows where to begin
his speech, I refer to the discussion of that question on pp. [302–307] below. I
299 now turn to the speeches that do not open with an ἐπεί-clause. |

4 Messenger Speeches That Open with a Simple Sentence

4.1 List of Plays

The speeches involved occur in the following plays: Hippolytus (1173 ff.), Hecuba
(521ff.), Supplices (650ff.), Helen (605ff.), Phoenissae (1219 ff.), Orestes (866 ff.
and 1400ff.), Bacchae (677ff.), Iphigenia at Aulis (414 ff.).

Hippolytus 1173ff.
Messenger to Theseus: ‘We were combing our horses’ coats beside the sea,
where the waves came crashing to the shore …’. There is no ἐπεί-clause here,
merely a simple sentence. And its function is quite different from the cases we
have discussed: no reference whatsoever is made to information given earlier.
What the messenger says in 1173ff. is, on the contrary, completely new. Hippoly-
tus’ final words on stage, προπέμψατε χθονός (1099), are not picked up in the
opening lines of the messenger’s report.16 In other words, the event reported is
not viewed in the same manner as in speeches with an ἐπεί-clause, but more
‘externally’. This feature is apparent here. The information presented in 1175ff.
(ἦλθε γάρ τις etc.) is not new to Theseus, nor to the audience: we all know from
vv. 1090ff. that Hippolytus has been banished. But the messenger apparently
did not, since he briefly explains how he knows, before describing Hippolytus’
arrival. We find the same feature in four other speeches:

Hecuba 521ff.
Talthybius to Hecuba: ‘The whole army of the Greeks, drawn up in ranks, was
present at the execution …’. This is quite similar to the speeches which open
with an ἐπεί-clause, in that some of what Talthybius says is known to Hecuba,
namely that her daughter would be slain at Achilles’ tomb. On the other hand,
the mention of the assembly of the Achaean army is new, and does not refer
to an earlier scene, in which the desirability of such a gathering was stressed.

16 Of course, the speech as a whole picks up earlier information, in that it presents what
happened after Hippolytus’ exit after v. 1101. The same holds for Hec., Supp., Or. I and II,
but not for Ba., see note 18.
how does a messenger begin his speech? 177

In other words the scene that ends with Odysseus and Polyxena leaving the
stage (after v. 437), is not directly continued in the opening-words of Talthy-
bius.17 | 300

Supplices 650ff.
The messenger to the chorus: ‘A brilliant shaft of sunlight, straight and clear,
lit up the field as I stood at Electra gate, where a tower gave a sweeping view’.
These are the opening lines of the description of the battle between Theseus
and the Thebans: they indicate the time and place of that battle, that is, new
information. There is no link with any earlier scene.

Orestes 866ff.
Messenger to Electra: ‘By strange coincidence, ma’am, I’d just now come into
town from the country …’. The event described is new and there is no reference
to anything earlier. Note also the occurrence of ἐτύγχανον, which emphasizes
that the messenger is an outsider.

Bacchae 677ff.
Messenger to Pentheus: ‘About that hour when the sun lets loose its light to
warm the earth, our grazing herds of cows had just begun to climb the path
…’. New information, except that these opening words continue the messen-
ger’s own words of 661: ἥκω Κιθαιρῶν᾽ ἐκλιπών … But there is no reference to an
earlier scene.18

There remain four speeches. Those of Hel. and IA are in some ways alike, while
the second speech in Or. and that of Ph. are rather different.

Helen 605ff.
Messenger to Menelaus: ‘Your wife is gone, swept up and away and out of sight
into the hollows of the high air’. This speech, it should be noted, does not give
details of some event that has been briefly described before, but describes the

17 In fact, the arrival of Polyxena at the Greek camp is not mentioned, but of course, from
what Talthybius says e.g. in v. 508 it is evident that she has arrived. It would have been
irrelevant, not to say pedantic, for Talthybius to begin his speech with an ἐπεί-clause, such
as ‘after Polyxena had come to the Greek camp …’
18 This would have been virtually impossible given the nature of the herdsman’s report: there
has been no previous scene in which Autonoe, Agave and the other women have left the
stage, in order to go to Mount Cithaeron. This is not the case in the Hippolytus for example,
see above, p. [299].
 178 chapter 8

event itself (up to v. 616);19 it is, thus, real news. Notice that it forms the answer
to Menelaus’ words (604): ἀγγέλλεις δὲ τί;

Iphigenia at Aulis 414ff.

Messenger to Agamemnon: ‘O commander of all the armies of Greece, King
Agamemnon, I am here to bring your daughter, Iphigenia’. This speech resem-
bles the preceding example in that it gives real news, and does not resume the
important aspects of an event already known. At the same time, however, the
301 speech has some further peculiarities: (1) it begins | not with a report of a past
event, but of something which is still happening as the messenger speaks;20
(2) there are no introductory lines, indicating, e.g., where the messenger comes
from; his entrance is so abrupt that he begins to speak in the middle of the line.
On both points this speech is without parallel.21

Orestes 1400ff.
Phrygian to the chorus: ‘Into palace came a pride of lions, Greekers, twins’. Both
the form and the content of this speech stand apart: (1) it is not in iambic trime-
ters, but in lyrics; (2) after lamenting his own fate (1395–1399) the Phrygian
begins the narrative proper with a sentence that would seem to refer back to
the scene that ends with Orestes and Pylades entering the palace (after v. 1245).
It could be argued, then, that there should have been an ἐπεί-clause at the
beginning of his speech. Note, however, that the Phrygian speaks about Orestes
and Pylades as persons unknown to the chorus whom he is addressing; this is
suggested by the absence of the definite article with λέοντες, and the lengthy
description of the character of Orestes and Pylades. As far as the Phrygian
is concerned, his story has no prehistory—it begins simply with Orestes and
Pylades entering the palace; in other words, ἦλθον is not meant to refer back.

Phoenissae 1219ff.
Messenger to Iocasta: ‘Your sons intend—Oh shamefulness of daring—a single
combat, severed from the host’. There follows the report on the debate between

19 So also Erdmann (1964: 173–174).

20 But which does not occur on stage.
21 Cf. also Erdmann (1964: 176). Some 19th-century scholars considered this speech to be the
work of an interpolator, not because of the two points mentioned in the text, but because
of ‘its confused, insipid and stupid’ (Dindorf) content. It has been defended by Hermann
and others. Anyhow, as Paley ad loc. notes, if it is spurious, it must have replaced another
messenger-speech, with approximately the same content, since Agamemnon in vv. 455ff.
mentions the imminent arrival of his wife and daughter.
how does a messenger begin his speech? 179

Eteocles and Polyneices. Since this report froms the sequel to the first speech
of the messenger, it naturally begins without special introduction.

4.2 Conclusion
In general, the messenger speeches which begin not with a temporal ἐπεί-
clause but with a simple declarative sentence, can be shown to open with
completely new information, which does not pick up information | that has 302
been presented earlier in the play. Two main groups can be discerned: in the
first (Hipp., Hec., Suppl, Or., Ba.) the opening lines generally give the time and
place of the subsequent narratives, the main points of which have already been
made known in the first lines spoken by the messenger. In the second (Hel., IA),
the content of the speeches forms news of a direct kind. The second speeches
of Ph. and Or. are untypical.

5 How Does the Messenger Know Where to Begin His Speech?

As we have seen, most speeches22 that open with an ἐπεί-clause refer back to
some information that has been presented earlier in the play.23 Further, the per-
son addressed by the messenger is usually acquainted with the event referred to
in the ἐπεί-clause, since he was present when that event occurred. We may now
return to the third feature of ἐπεί-clauses mentioned on p. [295] above. How
does the messenger know that his addressee is acquainted with that event? We
may sketch an ideal scenario for Euripides.
At some point in the play one of the principal characters leaves the stage
with the explicit remark that he is not alone, but is leaving accompanied by
e.g. a number of servants. Everyone present on the stage will thus know the cir-
cumstances under which this character made his exit. At a later point, e.g. after
a stasimon by the chorus, the stage is occupied by the persons, for instance the
chorus, who have witnessed the departure of the principal character and his
escort. Then someone dashes up, and after a quick dialogue with the person(s)
present, begins his story: ‘After we left this place …’. Clearly he is a member of
the escort: ‘we’ = ‘the principal character and I’. So he was on stage before, viz.
at the departure-scene of the principal character, and knows that his addressee
was also present then. Accordingly, he can begin his speech with a reference to
the earlier scene. The crucial point, then, is that there should somewhere be an

22 Not the first speech of the IT, see p. [297].

23 Of course this can be seen as a continuity-device.
 180 chapter 8

indication that the principal character went away with one or more servants,
including the future messenger. We do indeed sometimes find such an indica-
tion in the text, for example of the Helen in the second speech. The messenger
begins: ‘After Zeus’ daughter left the palace here, and you, and was escorted to
303 the sea …’ (1526ff.) and continues | (v. 1530): ‘Now as we came to your shipyards
…’. He (1) is speaking to a person (Theoclymenus) who actually was present
when Helen left the stage; (2) indicates that he was with Helen (ὡς ἤλθομεν,
v. 1530). Now Theoclymenus in v. 1412 orders a servant to go away and to procure
a ship and a crew for Helen and Menelaus: χώρει σὺ καὶ … δός … If the messenger
of 1512ff. is the σύ of 1412, he can refer to the departure without implausibility,
since it was Theoclymenus himself who ordered him to go to the sea.24
A similar pattern is perhaps also found in Heraclidae and in Iphigenia in
Tauris (second speech). In Heraclidae 748–867 a θεράπων gives to Alcmene an
account of the battle between Iolaus and the Argives. The reference of 800–801
is to 709ff. (cf. above p. [296]), where Iolaus tells Alcmene that he will join Hyl-
lus in the battle. A θεράπων is also present, witness e.g. vv. 720ff. If we are dealing
304 with the same θεράπων on both occasions, the reference to 709 ff. is easy.25 |

24 A slight inconsistency remains: ἤλθομεν in 1530 indicates that the messenger accompa-
nied Helen and Menelaus, whereas the χώρει of 1412, on a natural interpretation, implies:
go before Helen and Menelaus. Cf. also Kannicht’s long note ad 1410ff. He rightly remarks
that the palace-servant of 1412–1413 has changed into a ναύτης at vv. 1512ff. (cf. e.g. v. 1522,
1525).
25 Unfortunately, there is considerable disagreement over the name and the identity of the
θεράπων of vv. 784ff., and, in this connection, of the person who appears at v. 928 and is
designated as ἄγγελος in LP. Leaving the latter problem aside, the following remarks may
be made: (1) the θεράπων of vv. 630ff. is called a servant of Hyllus: 639 Ὕλλου πενέστης; (2)
he explicitly states that he will accompany Iolaus to battle, cf. e.g. 724–725: νῦν μὲν πορεύου
γυμνός, ἐν δὲ τάξεσιν | κόσμῳ πυκάζου τῷδ’· ἐγὼ δ᾽ οἴσω τέως; 739: εἰ δή ποθ’ ἥξομεν γε (that
is, to the battlefield); (3) the indication θεράπων at 784 occurs, according to Garzya in his
new Teubner-edition (see his app. crit.), only in Mr, a late copy of L, LP having nothing but
dashes in the margin (Murray’s note in his app. crit. ‘Θερ. LP’ is incorrect in that case); (4)
the ‘θεράπων’ of 784 addresses Alcmene as δέσποινα, this may, but need not, entail that he
is her servant (cf. LSJ s.v. δέσποινα); further, Hyllus, though mentioned several times (802,
818, 843) is never referred to in this speech as ‘master’; (5) in vv. 788–789 Alcmene says
something like ‘this good news sets you free’ (διήλασεν has been suspected); this would
seem to imply that she is the slave’s mistress; cf. also 888–891. The facts adduced under (4)
and (5), and some other considerations, led Rassow (1883: 7–9) to distinguish the person
who speaks 784–891 from the slave of Hyllus of vv. 630ff. All things considered this is not
implausible. Murray, on the other hand, thinks they are the same person; regrettably he
only says (ad Dram. Pers.): ‘una eademque mihi videntur esse persona’. Garzya agrees with
Rassow.
how does a messenger begin his speech? 181

The Iphigenia in Tauris presents an interesting problem. In v. 1284 the mss all
have the indication ἄγγελος. Wecklein however (see Murray’s apparatus) pro-
posed to read θεράπων ‘cl. 1205 et 1319’. Indeed, whether or not we actually read
θεράπων, it seems plausible to suppose that the messenger has been on stage
earlier, having been a member of the group that is referred to in 1208 (rather
than the group referred to in 1205), οἵδε ὁμαρτήσουσί σοι.26 Thus, the messen-
ger knows that Thoas knows that Iphigenia intended to go to the sea, which
enables him to start with: ‘When we reached the shore …’. There is no example
apart from these three plays of an explicit indication that the messenger was
on stage earlier. Now it should be noted that in the three plays discussed above
the messenger opens with a first person plural pronoun, ‘we’ meaning ‘the prin-
cipal character and I’. Using this as a criterion we may postulate that wherever
the speech begins with such a pronoun the principal character left the stage in
the company of the future messenger. These plays are: Andromache (1085),27
Electra (774), Bacchae II (1043), Iphigenia at Αulis II (1543).
The ἑπεί-clause in the rest does not contain an expression for ‘we’. From this
it may be inferred, I think, that the messenger was not present at the event
he is referring to in the ἑπεί-clause; and in fact it would sometimes have been
dramatically rather odd if he had. In most of these plays we must assume that
the messenger is tacitly, by convention, as it were, provided with all the knowl-
edge that is necessary to make his opening lines intelligible to the person he
is addressing. The speeches we are concerned with here are Alcestis, Medea,
Heracles, Phoenissae I, III and IV, and Ion.
Alcestis: There is no need to amplify here the discussion on p. [295].
Medea (1136): The messenger apparently was in the bridal appartment when
the children arrived (cf. also vv. 1118–1119 where Medea says that she sees an ὁπα-
δὸς Ἰάσονος coming). Obviously therefore he could not begin with something
like: ‘when we arrived at the bridal appartment’. By the same token he cannot
have been present when Medea sent her children | on their errand. In other 305
words, he tacitly assumes that it was in fact Medea who sent the children.28
Heracles (923): Given the form of the opening, it would have been rather
strange to find an expression for ‘we’ in the ἐπεί-clause, since this would have

26 Cf. also 1329 ἡμᾶς μέν, οὓς σὺ … συμπέμπεις.

27 The event referred to in the ἐπεί-clause here lies outside the play, cf. p. [296].
28 This appears also from φαρμάκων τῶν σῶν ὕπο (1126): σῶν indicates that the messenger
simply takes for granted that the φάρμακα were Medea’s. Such assumptions are, of course,
inevitable and, indeed, necessary if the action is to be developed intelligibly. Note once
more (cf. note 23) that it does not matter to the audience who know, of course, that it was
Medea’s φάρμακα that are responsible.
 182 chapter 8

meant that the messenger had suddenly become a participant in the doings
of the principal character (the killing of Lycus). Secondly, when the messenger
refers to Lycus’ death he is taking for granted that the chorus are acquainted
with this event; strictly speaking he cannot know this, since he has come from
the palace and so has not been on stage to overhear Lycus’ death with his
addressee.
Phoenissae I (1090): Here a good deal of information is given only implic-
itly. (1) The ἐπεί-clause refers to Menoeceus’ suicide, which we are to sup-
pose occurred after v. 1018; (2) the messenger speaks to Jocasta, but how does
she know about Menoeceus’ death? She was not present when Menoeceus
announced he will die to save his country. The only thing Menoeceus does
say (986) is that he will go to Jocasta, not to tell her of his intended suicide,
but of his so-called ‘flight’ to save his life, thereby hoodwinking Creon; (3) it
would have been difficult for the messenger to begin his speech with some-
thing like: ‘when we had seen Menoeceus die’; or ‘when we (= Men. and I) had
gone away’: Menoeceus clearly goes off alone to execute his plan. All this means
that the messenger, in referring to Menoeceus’ death, simply takes for granted
that Jocasta knows about it. This is confirmed by another point: in the phrase
ὁ γῆς ὑπερθανών, the definite article indicates that he supposes Menoeceus is
known to have died for his country.
Phoenissae III (1359): The messenger is speaking to Creon. The reference
of the ἐπεί-clause is to 1242–1243; given the beginning, he could not possibly
have started with an expression for ‘we’, since this would have meant that he
was actively involved in the combat between Eteocles and Polyneices, which
is clearly impossible. Apart from this, the opening lines (1356–1360) are rather
problematic: (1) it is generally believed that the messenger who speaks here is
306 not the same as the messenger of 1067ff., | the latter being Eteocles’ armour-
bearer (cf. 673–674).29 Yet this is not indisputable: one indication to the con-
trary might be that the second messenger in v. 1461 speaks of Eteocles as δεσπό-
την ἐμόν. If he is not the same it is certainly remarkable that he picks up the
words of the other messenger almost verbatim; (2) if we are dealing with one
and the same person, the messenger knows that Creon is not acquainted with
the event he is referring to in the ἐπεί-clause, since Creon was not present when
1242–1243 were spoken: he does not enter the stage until v. 1308 (Creon has been
informed about the duel in the meantime (1322ff.), but the messenger is not

29 They are distinguished in the Dram. Pers.: ἄγγελος and ἕτερος ἄγγελος. To be sure, Pearson,
in the Introduction to his commentary (p. XV, note 3) remarks: ‘He is of course a different
person to the armour-bearer’. But this begs the question.
how does a messenger begin his speech? 183

present then); (3) of course, if the messenger is not the same, he cannot possi-
bly know whether Creon knows about the event referred to in the ἐπεί-clause.
In either case we have the same result: an opening with simply the ἐπεί-clause
would have suggested too much.30 The messenger would have had tacitly to
assume that Creon knows about the preparations for the duel, which is, I think,
too particular an event for such an assumption, unlike e.g. Menoeceus’ death
in the first speech of Phoenissae or Lycus’ death in Heracles. To remedy this dif-
ficulty, I suggest, Euripides wrote the rather odd opening lines 1356–1358; the
messenger simply states that Creon is acquainted with the earlier stages of the
(his) story. The result is hardly felicitous: the transition from the εὐτυχήματα to
the preparations for the duel jars somewhat.
Phoenissae IV (1428): The messenger could hardly have started with an ex-
pression for ‘we’, for the same reasons as are suggested above. But there can
be no question as to the referential correctness of the ἐπεί-clause, since it is
the messenger himself who spoke the line to which the ἐπεί-clause refers, each
time to the same person.
Finally, the Ion (1122): The messenger is called a servant of Creusa in the
Dramatis personae (cf. also 1107). The ἐπεί-clause refers to the departure of
Xuthus and Ion after v. 675. From the absence of an expression for ‘we’ in the
ἐπεί-clause, we may infer that the servant was not in the company of Xuthus
and Ion when they left. (Indeed, this would have been difficult, if he really
is Creusa’s slave). Secondly, the addressee, the chorus, is acquainted with the
event referred to in the ἐπεί-clause, since it was present | at the departure of 307
Xuthus and Ion. But how does the servant know that the chorus knows? Appar-
ently he takes the chorus’ knowledge for granted. We may also note παῖδα τὸν
καινόν in 1123, where the definite article indicates that the speaker supposes the
person mentioned to be known to the addressee; and the mention of δεῖπνα
and θυσίαι in 1124, which refers to 652–653, where the chorus is also present.31

6 Conclusion

In half of the 1432 speeches which begin with an ἐπεί-clause, our third require-
ment, that the messenger has to know where to begin, is satisfied. In three

30 Of course, this does not hold for the audience, since for them the story broke off at just
this point.
31 Note that until v. 1194 (ἐπίμπλαμεν) there is no explicit indication that the slave has wit-
nessed the events he is describing.
32 Fourteen, not fifteen, since the first speech of the IT has a position of its own.
184 chapter 8

instances33 the messenger is explicitly said to have been on the stage before,
perhaps as a servant, together with the person whom he is addressing, both as
witnesses of the event he is referring to in the ἐπεί-clause. In four other cases we
may infer, from the presence in the ἐπεί-clause of an expression for ‘we’, mean-
ing the principal character and the messenger, that he was present at the event
he is referring to (again, in company of his addressee). In the above cases, then,
the messenger knows that his addressee knows what he is talking about.
Such explicit authorial guidance is lacking in the other seven plays. In all but
one, Phoenissae 1359ff. (see pp. [305–306]) we have to assume that the messen-
ger is allowed by convention to suppose that his addressee is acquainted with
the event he is referring to. In most cases such ‘conventional’ knowledge was
the only possibility, since the messenger could not possibly have been present
at the event he is referring to.

References

Erdmann, G. (diss.), Der Botenbericht bei Euripides (Kiel 1964).

Grube, G.M.A., The Drama of Euripides, second edition (London 1961).
Rassow, J., Quaestiones selectae de Euripideorum nuntiorum narrationibus (Greifswald
1883).
Rijksbaron, A., Temporal and Causal Conjunctions in Ancient Greek: With Special Refer-
ence to the Use of ἐπεί and ὡς in Herodotus (Amsterdam 1967).

33 Or two, if Heracl. 748–867 are omitted.

 chapter 9

Discourse Cohesion in the Proem of Hesiod’s

Theogony

Abstract

This chapter takes up the vexed question of the coherence of the proem of Hesiod’s
Theogony (1–115). It makes more precise, and at the same time modifies, Verdenius’
view that the unity of the proem lies more in the continuity of its progress than in the
interdependency of its parts. In a line by line commentary on relevant items it estab-
lishes the sequence of actions, started off by the imperfect στεῖχον (10). It is also shown,
however, that unity is created by the fact that the imperfect verb forms, which refer to
sights and sounds, presuppose a human observer, and that this observer can be identi-
fied with the poet himself, who experienced these events in the past. Further unity is
created by the recurrent references to Hesiod as a servant (θεράπων) of the Muses.

Few pieces of Greek literature have sparked so much controversy as the proem
of Hesiod’s Theogony (lines 1–115).* Among the controversial issues are: the
identity of the Muses, the strangely selective catalogue of gods at lines 11–21,
and the temporal and spatial coordinates, so to speak, of Hesiod’s encounter
with the Muses and especially those of the movements of the Muses during the
proem. It is the temporal and spatial side of the proem that will be the subject
of this paper. In 1972, W.J. Verdenius published a series of notes on the proem of
the Theogony, which he concluded with the following verdict: ‘The unity of the
proem does not lie in the interdependency of its parts, but in the continuity of
its progress’ (Verdenius 1972: 260). In what follows I will argue that Verdenius
rightly viewed the continuity of the progress of the proem as a unifying princi-
ple, but underestimated the unifying force of the interdependency of its parts.1

* I am indebted to the members of the Amsterdamse Hellenistenclub, especially Niek van der
Ben, Jan Maarten Bremer and Irene de Jong, and to Gerry Wakker and Stephanie Bakker, for
their critical comments.
1 As is also shown by Thalmann in his, on the whole convincing, analysis of the proem (1984:
134–150). For the role of devices like ring composition I may refer to these pages. At the end of
this paper the full text of the proem (lines 1–115) is presented. [Hes. Th. 1–10 is also discussed
in chapter 7 (ex. 33) of this volume.]

© koninklijke brill nv, leiden, 2019 | doi:10.1163/9789004386129_011

 186 chapter 9

The Theogony sets off with what is often regarded as a quaestio maior, whose
status is formulated by Thalmann as follows:

Explanations of the different verb tenses in Theog. 1–10 have been sug-
gested that rival in their complexity those given for the similar mixture in
h. Apol. 1–13. This elaborateness by itself makes them suspect. Any such
explanation, to be convincing, would have to account for similar incon-
sistencies in other passages cited by West (1966: 155), on Theog. 6 (i.e. h.
19.10–15, 27–29; h. Aphr. 260–261). His solution to the problem, that the
verbs in all these passages are timeless, is more economical and more
plausible than its convoluted rivals. Verdenius takes issue with West on
242 this point; but his own explanation, (1972: 227 on Theog. 5)—that the |
narrative of a particular event begins at line 5 and continues through
line 35—ignores ποτε (l. 22) and drives him to an untenable view (p. 249)
of τότε in line 68.
Thalmann 1984: 227, fn. 9

I will now review the relevant tense forms and adverbs in the form of a run-
ning commentary, following Verdenius in this respect, too. In the process I will
also discuss some other points that perhaps deserve further or new comments.
Present indicatives are printed in italics, imperfects in bold type, and aorist
indicatives in bold italic type.

Μουσάων Ἑλικωνιάδων ἀρχώμεθ᾽ ἀείδειν,

αἵ θ᾽ Ἑλικῶνος ἔχουσιν ὄρος μέγα τε ζάθεόν τε,
καί τε περὶ κρήνην ἰοειδέα πόσσ᾽ ἁπαλοῖσιν
ὀρχεῦνται καὶ βωμὸν ἐρισθενέος Κρονίωνος
5 καί τε λοεσσάμεναι τέρενα χρόα Περμησσοῖο
ἠ᾽ Ἵππου κρήνης ἠ᾽ Ὀλμειοῦ ζαθέοιο
ἀκροτάτῳ Ἑλικῶνι χοροὺς ἐνεποιήσαντο,
καλοὺς ἱμερόεντας, ἐπερρώσαντο δὲ ποσσίν.
ἔνθεν ἀπορνύμεναι κεκαλυμμέναι ἠέρι πολλῷ
10 ἐννύχιαι στεῖχον περικαλλέα ὄσσαν ἱεῖσαι2

2 ‘Let us begin to sing from the Heliconian Muses, who possess the great and holy mountain of
Helicon and dance on thier soft feet around the violet-dark fountain and the altar of Cronos’
mighty son. And after they have washed their tender skin in Permessus or Hippocrene or
holy Olmeius, they perform choral dances on highest Helicon, beautiful, lovely ones, and
move nimbly with their feet. Starting out from there, shrouded in thich invisibility, by walked,
sending forth their very beautiful voice (…)’ (translation Most, except for στεῖχον in line 10, see
below).
discourse cohesion in the proem of hesiod’s theogony 187

1 Μουσάων (…) ἀρχώμεθ᾽ ἀείδειν

The opening line of the Theogony makes us expect that the poet is about to start
singing,3 making the Muses the starting-point of his song. This expectation,
however, is only partially fulfilled, for after seven lines on some of the character-
istic activities of the Muses, the poet first explains how he is able to undertake
such an enterprise at all, and why the Muses are his starting-point (lines 9–34).
It will turn out that the self-hortatory ἀρχώμεθα is the consequence of an order
given to Hesiod by the Muses themselves (ἐκέλοντο, line 33).4 After he has given
the explanation he makes a new start at line 36, where his Muses-song really
begins. | 243

ἀρχώμεθα
The poet may have preferred the plural to the singular to involve the audience
right from the start in his enterprise.5 For this use of the plural cp. Chantraine
(1953: 2, 34) on Od. 1.10 (to the Muse) (…) εἰπὲ καὶ ἡμῖν: ‘(…) le poète parle de

3 Temporally, the hortative subjunctive is uttered at a fictional moment of speaking, which cre-
ates an implicit ‘now’. For a similar use of the first person future indicative see Pfeijffer (1999:
19–20). See also Furley & Bremer (2001: 51–52) on self-exhortations in hymns.
4 Verdenius wrongly says that ἀρχώμεθα ‘is a traditional formula used at the beginning of a
hymn.’ On the contrary, it is anything but traditional. He refers to h. Dem. 1, but there we find
ἄρχομ᾽ ἀείδειν, which is, indeed, a formula, cp. h. 9.8, 11.1, 13.1, 16.1, 22.1, 26.1, 28.1. The horta-
tive subjunctive makes Th. 1 crucially different. (Μουσάων ἄρχωμαι, at h. 25.1, is, just like the
rest of this short hymn, a clear imitation of Th. 1). According to Pucci (2007: 33), ἀρχώμεθα
expresses the idea that ‘il cantore … si mette in prima linea con una insolita enfasi sul proprio
“io” precedendo le Muse.’ This ignores the reappearance of ἀρχώμεθα at line 36; see there.
5 For the role of the audience, and for other characteristics of hymns see Danielewicz (1976: 119),
as quoted by Furley & Bremer (2001: 59): ‘The specific character of the hymn [as a type—
AR] is to be seen in the simultaneous existence of two communicative settings, the first of
which (the author/performer → the formal addressee, viz. the god) is supplemented by a new
one: the author performer → the real recipient, viz. the listeners. The latter setting enables
the poet to deal with the theme of the utterance from a distance: to begin and conclude
with a meta-textual formula, to treat the addressed god, until the moment of salutation, as
an object of description or narration, to emphasize the presence of the performer, and to
imply the virtual audience.’ All these elements are present, at various places, in the proem
of Th., and yet the proem is entirely sui generis, since not only the god(s) but also the poet
himself is ‘an object of narration’, viz. in the narrative about his encounter with the Muses.
The elements are: line 1 meta-textual formula, emphasizing the presence of a performer and
implying an audience; lines 2–8 description passing into narrative (see below); lines 9–34
personal narrative: Hesiod’s encounter with the Muses, perception of their songs, Dichter-
weihe; lines 35–36 meta-textual formulas; lines 37–67 description of the Muses’ activities on
Olympus, interrupted by a narrative digression at lines 53–60: Mousogony; lines 68–71 and
lines 75–79 personal narrative, continued: Hesiod and the Muses; lines 71–74 description of
Zeus; lines 80–103 description of the Muses’ activities, through their servants, the singers,
among mortals; lines 104–115 salutation and prayer.
 188 chapter 9

lui-même en s’associant à ses auditeurs.’ Compare a speaker starting a lecture

with: ‘Ladies and gentlemen, let’s talk about the future of the classics,’ where
the talking of course will be done just by the speaker. But see also below, on
ἀρχώμεθα at line 36.

2 ἔχουσιν, 4 ὀρχεῦνται
‘Omnitemporal’ (Faulkner 2005: 66), rather than ‘timeless’ (e.g. Thalmann fol-
lowing West, see above), ‘eternal’ (Stoddard 2004: 130) states of affairs. ‘Time-
less’ states of affairs are ‘outside time altogether’ (Lyons 1977: 680), and are
confined to mathematical truths like ‘two times five equals ten’, or, in Greek, τὰ
δὶς πέντε δέκα ἐστίν. As for ‘omnitemporal’, Lyons uses this term for situations
which are ‘(…) time-bound but temporally unrestricted’ (Lyons 1977: 680).6 This
244 applies in our | case: as long as there are Muses—this may change, of course—
they are the masters of Mount Helicon and dance there.7 Note that these two
descriptive omnitemporal statements occur in a relative clause, which means
that we are dealing with a so-called ‘attributive section’, i.e. a section in which
one or more fixed attributes of a god are mentioned.8 They are due to an omni-
scient narrator and must represent impersonal, conventional knowledge.

ἔχουσιν
= ‘possess’ (Lombardo, Most), not ‘bewohnen’ (Marg). Compare lines 108 and
113 below: εἴπατε δ᾽ ὡς (…) τὰ πρῶτα πολύπτυχον ἔσχον Ὄλυμπον (‘took posses-
sion of’, Most). ‘Live’ is rather expressed by οἰκία or δώματ᾽ ἔχειν and similar
expressions, like ἔνθα σφιν λιπαροί τε χοροὶ καὶ δώματα καλά (line 63). The Muses
did indeed not live on Mount Helicon, but on Olympus (lines 62–63): Ὀλύμπου·
| ἔνθα σφιν λιπαροί τε χοροὶ καὶ δώματα καλά.

5–9 καί τε (…) χοροὺς ἐνεποιήσαντο, ἐπερρώσαντο δὲ ποσσίν

Perhaps I may quote Ruijgh (1971: 900) here: ‘L’emploi de τε adverbial invite
à considérer le fait exprimé comme permanent (aoriste gnomique). Comme à

6 Omnitemporal states of affairs always involve iteration or habituality. There are also tem-
porally restricted habitual states of affairs, like ‘The Royal Ballet dances at the Royal Opera
House’. If the habitual dancing of the Muses lasts longer, so to speak, than that of mortals,
this is due to the characteristics of the subject and of the locative phrase; it is not formally
marked on the verb (although in the epic language the omnitemporal interpretation of a
present indicative may be indicated by the presence of epic τε, see Ruijgh (1971: chapter 1)).
7 Clay (1988: 324) writes: ‘Hesiod does not tell us what the Muses sing on the peak of Helicon.’
But they just dance (ὀρχεῦνται) there, there is no verb of singing in these lines. The singing
occurs during their descents from Helicon, see below.
8 For attributive sections see Bakker (2002: 66), Depew (2004: 129), and Faulkner (2005: 60–61).
discourse cohesion in the proem of hesiod’s theogony 189

priori, l’indicatif II de l’aoriste peut également exprimer un fait du passé, il y

a une transition graduelle vers la phrase suivante (Th. 9 sqq.), qui exprime net-
tement un fait du passé (στεῖχον). Ainsi, le poète prépare le récit de sa propre
expérience, c’est-à-dire de sa rencontre avec les Muses (Th. 22sqq.).’ The aorists
start, then, as elements of an attributive section but must eventually be rein-
terpreted as elements of the narrative that is initiated by στεῖχον (see further
below). For the sequence (omnitemporal) present indicative: (gnomic) aorist
indicative cp. e.g. Hdt. 7.10ε: φιλέει γὰρ ὁ θεὸς τὰ ὑπερέχοντα πάντα κολούειν. οὕτω
δὲ καὶ στρατὸς πολλὸς ὑπὸ ὀλίγου διαφθείρεται κατὰ τοιόνδε· ἐπεάν σφι ὁ θεὸς φθο-
νήσας φόβον ἐμβάλη ἢ βροντήν, δι’ ὦν ἐφθάρησαν ἀναξίως ἑωυτῶν. | 245

9–10 ἔνθεν ἀπορνύμεναι κεκαλυμμέναι ἠέρι πολλῷ | ἐννύχιαι στεῖχον περικαλλέα

ὄσσαν ἱεῖσαι

9 ἔνθεν
Not ‘puis’ (Mazon) but ‘from there’ (Evelyn-White, Most, Marg and others), i.e.
from the top, hence their being covered by ἀήρ.

10 στεῖχον
After the omnitemporal present indicatives of lines 2–4, and the ‘gnomic-
cum-past’ aorist indicatives of lines 7–8, στεῖχον expresses ‘nettement un fait
du passé’ (see Ruijgh as quoted above at lines 5–9). More specifically, στεῖχον
should be taken as a ‘focalising’ imperfect, that is, an imperfect which presents
a certain state of affairs from the point of view of a character rather than that of
the narrator.9 Who is this character? At lines 22–23 the narrator mentions Hes-
iod (αἵ νύ ποθ᾽ Ἡσίοδον καλὴν ἐδίδαξαν ἀοιδήν, ἄρνας ποιμαίνονθ᾽ Ἑλικῶνος ὕπο
ζαθέοιο), from which we must infer that he was the character perceiving the
στείχειν of the Muses. Observe that the Muses moved περικαλλέα ὄσσαν ἱεῖσαι:
they could be heard while they were moving. Next, this perceiving character
turns out to be none other than the narrator himself: με (…) θεαὶ πρὸς μῦθον
ἔειπον, line 24: here, the third person narrative of lines 10–23 becomes a first
person narrative and thus the report of a personal experience.
The focalising use of the imperfect after a general description is not always
recognized by scholars. An illuminating example from Homer occurs at the end
of the passage Od. 7.104–132: (δμῳαί) (…) ἀλετρεύουσι (…) ὑφόωσι καὶ (…) στρωφῶ-

9 Smith (2002: 71) uses the term ‘inferred perception’ for this phenomenon, and calls the sen-
tences involved ‘perspectival’. See also Bakker (1997: 40) on the remarkable combination, at
Thuc. 7.70.7, of the imperfect ἐγίγνετο with νῦν, expressing ‘the perception and experience of
the crews on the ships.’
 190 chapter 9

σιν (…) ἀπολείβεται (…) (line 113) ἐλήλαται (…) πεφύκασι (…)· οὔ ποτ᾽ ἀπόλλυται οὐδ᾽
ἀπολείπει (…) τὰ μὲν φύει, ἄλλα δὲ πέσσει. (line 120) (…) γηράσκει, (…) ἐρρίζωται, (…)
τέρσεται (…) τρυγόωσιν, (line 125) (…) τραπέουσι· (…) εἰσιν (…) ὑποπερκάζουσιν. (…)
πεφύασιν, which is followed and concluded by

ἐν δὲ δύω κρῆναι ἡ μέν τ᾽ ἀνὰ κῆπον ἅπαντα

130 σκίδναται, ἡ δ᾽ ἑτέρωθεν ὑπ᾽ αὐλῆς οὐδὸν ἵησι
πρὸς δόμον ὑψηλόν, ὅθεν ὑδρεύοντο πολῖται.
246 τοῖ᾽ ἄρ᾽ έν Ἀλκινόοιο θεῶν ἔσαν ἀγλαὰ δῶρα10 |

Here ὑδρεύοντο and τοῖ(α) ἔσαν are presented from the standpoint of Odysseus
there and then on the treshold of the palace.11 Some prose examples are: X.
An. 6.4.1–6: ηὐλίζοντο ἐπὶ τοῦ αἰγιαλοῦ πρὸς τῷ λιμένι. τὸ δὲ χωρίον τοῦτο (…) ἔστι
(…) καλεῖται (…) ἔστι (…) ἐστὶν (…) ἐστιν (…) λέγονται (…) κεῖται (…) ἔστι δ᾽ (…)
τὸ δὲ ὄρος (…) ἀνήκει (…) φέρει γὰρ ἡ γῆ (…) ἡ μὲν χώρα ἦν τοιαύτη. ἐσκήνουν
δ᾽ (…). Here, a geographical description presenting conventional knowledge
in omnitemporal present indicatives is closed by an imperfect plus an evalua-
tive adjective (cp. τοῖα at Od. 7.132), whereby the whole preceding description
is turned into something that was perceived by the soldiers camping there; X.
An. 1.5.6 (…) ὁ δὲ σίγλος δύναται ἑπτὰ ὀβολοὺς καὶ ἡμιωβέλιον Ἀττικούς· ἡ δὲ καπίθη
δύο χοίνικας Ἀττικάς ἐχώρει (the soldiers found out that this was the contents of
the kapithê); X. An. 1.5.3 τὰς δὲ ὠτίδας ἄν τις ταχὺ ἀνιστῇ ἔστι λαμβάνειν· πέτονται
γὰρ βραχὺ ὥσπερ πέρδικες καὶ ταχὺ ἀπαγορεύουσι. This omnitemporal, habitual,
description of the behaviour of the bustards whenever someone hunts them is

10 Cp. De Jong (2001: 176–177) on Od. 7.81–135.

11 In the brilliantly structured passage 81–132—called ‘inorganic’ by Heubeck et al.—the
focalising imperfect is already present before. There is a gradual movement of the camera
(which is Odysseus) from the outside to the inside of the palace of Alcinous. Odysseus was
approaching (ἴε, line 82) the palace, and he pondered (ὅρμαινε) many things while he was
busy finding a place to stand (ἱσταμένῳ, line 82), until he arrived (ἱκέσθαι, line 83) at the
treshold. Next, lines 84–102 describe, in imperfects and pluperfects, the things Odysseus
pondered, things, we may infer, seen by him during the time of his ἵστασθαι: first the walls
ἐς μυχὸν ἐξ οὐδοῦ and elements belonging to the walls (lines 84–94), and then the space
enclosed by the walls as far as he could see it, i.e. the courtyard (lines 95–102; notice
line 95 ἐν δὲ θρόνοι, and cp. ὑπ᾽ αὐλῆς οὐδόν at line 130). Then what is going on inside
the palace itself, and what is situated outside the αὐλή (line 112 ἔκτοσθεν αὐλῆς), viz. the
orchard, neither of which can be seen by Odysseus, are told by the omniscient narrator in
omnitemporal present (and perfect) indicatives (lines 103–131 ὑψηλόν). Finally, at line 133
the narrative is resumed; ἔνθα (στάς) refers back to the οὐδὸν introduced in l. 83, while
θηεῖτο expresses the idea that once Odysseus stands on the threshold he is no longer pon-
dering but full of admiration. (De Jong (2001: 176) wrongly thinks that Odysseus steps over
the threshold in line 83.)
discourse cohesion in the proem of hesiod’s theogony 191

followed by τὰ δὲ κρέα αὐτῶν ἥδιστα ἦν—for the soldiers there and then.12 Also
in a narrative, after an aorist indicative: X. HG 2.1.21 ἔπλευσαν εἰς Αἰγὸς ποταμοὺς
ἀντίον τῆς Λαμψάκου. διεῖχε δ᾽ ὁ Ἑλλήσποντος ταύτῃ σταδίους ὡς πεντεκαίδεκα. I
have discussed this use of the imperfect and further examples in Rijksbaron
(1995), with literature. For a rather spectacular instance of such an imperfect
(not | discussed in the article just mentioned), which illustrates its function 247
very well, see the Appendix.
Στεῖχον at Theogony line 10 should be taken, then, in the same way, the
human perceiver being presupposed by the presence of περικαλλέα ὄσσαν ἱεῖσαι,
| ὑμνεῦσαι Δία etc. The words sung by the Muses in lines 11–21 are therefore heard
by the poet.13 Note that the non-verbal sound implied by the movement of the
feet of the Muses (ἐπερρώσαντο (…) ποσίν in line 8) may also have been heard
by the poet, and may therefore prepare the ground for the meaningful sounds
mentioned in lines 10–21. I think, moreover, that στεῖχον has iterative meaning:
after the impersonal, conventional information on the omnitemporal activi-
ties of the Muses in lines 2–8 we are now dealing with a habitual activity of the
Muses in the past, in Hesiod’s past, that is, as witnessed (aurally) by Hesiod.14
In other words, with στεῖχον we enter a narrative.15 See the next note, on ποτε. | 248

12 The imperfect may also appear before the omnitemporal present, as in Od. 3.291–293 ἔνθα
διατμήξας τὰς μὲν Κρήτῃ ἐπέλασσεν, | ἧχι Κύδωνες ἔναιον Ἰαρδάνου ἀμφὶ ῥέεθρα. | ἔστι δέ τις
λισσὴ αἰπεῖά τε εἰς ἅλα πέτρη—ἔναιον, because perceived by Odysseus.
13 For the idea cp. A.R. 4.1381 Μουσάων ὅδε μῦθος, ἐγὼ δ᾽ ὑπακουὸς ἀείδω | Πιερίδων.
14 West, in his commentary ad locum, considered στεῖχον a ‘typifying’, that is to say a time-
less, imperfect, which temporally would not differ from the present and aorist indicatives
of lines 2–8 (likewise already e.g. Evelyn-White, Loeb (1914): ‘Thence they arise and go
abroad by night …’). I will refrain from discussing this bizarre view, and I can do this the
more easily because West himself, in a much later publication (West 1989), without dis-
cussing his earlier view, abandoned it, only to exchange it for another untenable view.
In West (1989) he considers στεῖχον an instance of a non-past, so-called ‘injunctive’, use
of augmentless forms with secondary endings. But this injunctive use does not exist. Cp.
Beekes (1995: 245): ‘In Greek the injunctive forms are completely equal to the indicative
forms (…); it thus has no sense to speak here of an injunctive (because there is no mean-
ing component). Only in Sanskrit can one speak of an injunctive.’ Remarkably, West never
tells us what is wrong with taking the imperfect as an imperfect.—It will be recalled that
according to Thalmann the use of the tenses in Th. 1–10 should be explained in connection
with ‘similar inconsistencies’ in other passages, notably h. 3.1–13, 19.10–15, 27–29; h. Aphr.
260–261. Why? The phrase ‘similar inconsistencies’ begs the question, of course; above I
argued (with others) that the tenses in Th. 1–10 are not used inconsistently at all but in
accordance with the normal rules of Greek. Which means that any inconsistencies that
might be found in the hymns—one may think notably of the imperfect μίμνε at h. Ap. 5,
which is definitely unlike στεῖχον at Th. 10—may, as a matter of principle, be sui generis,
and demand a different explanation from that of the tenses in Th. 1–10. For a recent, inge-
nious but not quite convincing, analysis of the beginning of h. Ap. see Bakker (2002).
15 Cp. Clay (2003: 54): ‘In effect, the imperfect στεῖχον conveys precisely and vividly the
192 chapter 9

22 αἵ νύ ποθ᾽ Ἡσίοδον καλὴν ἐδίδαξαν ἀοιδήν

νυ
Νυ stresses the idea that the διδάξαι of the Muses did actually happen as it is
described in this sentence. Cp. Ruijgh (1971: 842): ‘(…) elle [the particle νυ] sig-
nale que le fait exprimé par la phrase a la même importance qu’ un fait actuel.’

ποθ᾽
Cp. Thalmann (1984: 136): ‘The Muses’ gift of song to Hesiod was a single
instance, which occurred at a particular moment (ποτε, ‘once,’ line 22), of their
often-repeated activities on the heights of Helikon (lines 1–21).’ This is correct,
but for the reference to lines 1–21 as a whole; Thalmann wrongly followed West
in assuming that all verbs in lines 1–21 are timeless. Incidentally, how could the
Muses possibly have met Hesiod in the human world if their often-repeated
activities are timeless, i.e. outside time altogether? For this encounter to hap-
pen it is necessary that they performed an activity in (and not outside) the
human world, and this activity is provided for by στεῖχον. The encounter of the
Muses with Hesiod occurred, indeed, at a particular moment, namely during
one of their descents from Mount Helicon.16

ἐδίδαξαν
A complexive aorist, which denotes a state of affairs (‘ “bestow an ability” rather
than “teach”’—Verdenius) that occurs within the framework created by στεῖχον
and comprises ἔειπον (line 24), ἔδον (line 30) and ἐνέπνευσαν (line 31).17 For the

moment of transition from the eternal time of the gods to the temporality of mankind.’ We
are not dealing with ‘a moment’, however. That στεῖχον rather expresses a habitual activ-
ity, as claimed above, is an effect of ποτε in l. 22: ‘once, on one occasion’ presupposes the
existence of a series of activities during which another activity occurred. For this effect of
ποτε cp. the sequence ἐφύλασσε (…) κοτε (…) διέφθειραν at Hdt. 9.93.2: ἔνθα δὴ τότε ὁ Εὐή-
νιος οὗτος ἀραιρημένος ἐφύλασσε· καί κοτε αὐτοῦ κατακοιμίσαντος τὴν φυλακὴν παρελθόντες
λύκοι ἐς τὸ ἄντρον διέφθειραν τῶν προβάτων ὡς ἑξήκοντα. The first verb may also be explicitly
marked as habitual, as at Hes. Op. 633–635 ὥς περ ἐμός τε πατὴρ καὶ σός, μέγα νήπιε Πέρση,
| πλωίζεσκ᾽ ἐν νηυσί, βίου κεχρημένος ἐσθλοῦ· | ὅς ποτε καὶ τεῖδ᾽ ἦλθε πολὺν διὰ πόντον ἀνύσ-
σας. Pucci (2007: 43–44) confines himself to enumerating the various interpretations of
lines 9–10.
16 In spring or early summer, cp. ἄρνας ποιμαίνοντ(α), line 23 (lambs are typically born in the
spring; not ‘sheep’ (West and others), for this is ποιμαίνοντ᾽ ἐπ᾽ ὄεσσι, Il. 11.106), and proba-
bly at dawn, for Hesiod must have seen the ‘marvellous branch’ given to him—as he tells
us—at lines 30–31. We are not told that he actually saw the Muses, and must assume that
they remained covered by mist, as indicated by the perfect participle κεκαλυμμέναι (line 9).
17 For the complexive aorist see K-G (1, 155) and Rijksbaron (2007: 11–12).
discourse cohesion in the proem of hesiod’s theogony 193

idea cp. Od. 8.480–481 πᾶσι γὰρ | ἀνθρώποισιν ἐπιχθονίοισιν ἀοιδοὶ | (480) τιμῆς 249
ἔμμοροί εἰσι καὶ αἰδοῦς, οὕνεκ᾽ ἄρα σφέας | οἴμας Μοῦσ᾽ ἐδίδαξε, φίλησε δὲ φῦλον
ἀοιδῶν, also 8.488.

24 τόνδε δέ με πρώτιστα θεαὶ πρὸς μῦθον ἔειπον

τόνδε should be connected with μῦθον, not with με (Waltz 1914: 232, Clay 2003:
52, Bakker 1999: 10, Pucci 2007: 58–59 and others): deictic pronouns are never
combined with unemphatic clitic pronouns. This, it is true, could be remedied
by printing δ᾽ ἐμέ, but this, too, should be rejected: the deictic use of ὅδε with
first and, rarely, second person pronouns is not found in narratives but only in
direct speech: they presuppose the presence of an addressee in a face to face
situation. Cp. Il. 19.40 ἐγὼν ὅδε, Od. 1.76 ἡμεῖς οἵδε, 16.205 ὅδ᾽ ἐγώ, 21.207 ὅδ᾽ αὐτὸς
ἐγώ, 2.45–46 ἐν ὑμῖν | τοίσδεσσιν, Th. 1.53.2 ἡμᾶς τούσδε. For further examples, also
of other combinations with ὅδε, see K-G (1, 641–643). For τόνδε (…) μῦθον cp. Il.
15.202 τόνδε φέρω Διὶ μῦθον (…) | ἦ τι μεταστρέψεις (…).

25 Μοῦσαι Ὀλυμπιάδες, κοῦραι Διὸς αἰγιόχοιο

Probably the Muses are now called ‘Olympian’ because the hymn that is to fol-
low (lines 37–67) deals with their activities on the Olympus.18

33–34 καί μ᾽ ἐκέλονθ᾽ ὑμνεῖν μακάρων γένος αἰὲν ἐόντων, σφᾶς δ᾽ αὐτὰς πρῶτόν τε
καὶ ὕστατον αἰεν ἀείδειν.

ἐκέλοντο
After the Muses have provided him with a purpose for his ‘divine voice’ (lines
31–32), they urge Hesiod to apply his newly acquired art to sing a hymn to the
gods collectively (γένος) and especially to the Muses themselves. By ordering
him to ὑμνεῖν μακάρων γένος αἰὲν ἐόντων they ask him to do among mortals what
they themselves do on Olympus: line 37 ὑμνεῦσαι τέρπουσι (…) ἐντὸς Ὀλύμπου,
further specified at 44 as θεῶν γένος αἰδοῖον πρῶτον κλείουσιν ἀοιδῇ.
As for the imperfect, this creates a certain expectation on the part of the
reader/listener as to the reaction of the addressee, a use of verbs of saying that
was first recognized and discussed by Blass (1889). Thus, at | Il. 1.386 αὐτίκ᾽ ἐγὼ 250
πρῶτος κελόμην θεὸν ἱλάσκεσθαι· | Ἀτρεΐωνα δ᾽ ἔπειτα χόλος λάβεν (…), χόλος λάβεν
describes Agamemnon’s reaction to Achilles’ κελόμην. The aorist lacks this fea-
ture, as at Il. 6.110, 116: Ἔκτωρ δὲ Τρώεσσιν ἐκέκλετο μακρὸν ἀΰσας· ‘(…)’. | ὡς ἄρα

18 ‘There is (…) a shift from one mountain to the other as the focus of the Muses’ activities’
(Thalmann 1984: 135).
 194 chapter 9

φωνήσας ἀπέβη κορυθαίολος Ἔκτωρ. There is no mention of the reaction of the

Trojans. See further Rijksbaron (2007: 18–19).

35 ἀλλὰ τίη μοι ταῦτα περὶ δρῦν ἢ περὶ πέτρην

With this line Hesiod breaks off the narrative about his encounter with the
Muses: the time has come to turn to the matter at hand.19 It will turn out, how-
ever, that the narrative is not yet finished, for the poet resumes it at line 68, quo
vide.

36 τύνη (…) ἀρχώμεθα

This remarkable combination has not got the attention it deserves. West has
only a note on the form τύνη. Verdenius observes: ‘Hesiod addresses himself,
not the audience’, rightly, but this is not the only point to be made.20 In fact,
the use of a second person pronoun with a first person plural hortative sub-
junctive is, as far as I could see, unique, and understandably so. For what we
have here is a speaker who is giving an order to himself as if he is somebody
else: ‘You ⟨and me⟩, let us start singing (…).’21 This very daring construction,
which is the more remarkable because in the previous line μοι is present, must
suggest that Hesiod is addressing himself both as the Hesiod of lines 5–34, so
to speak, the Hesiod who was favoured by, and got orders from, the Muses on
Helicon, and as the accomplished singer he is now: the Hesiod of the earlier
lines should stop talking about how the Muses inspired him, and finally start
251 singing his hymn to them. |

τύνη, Μουσάων ἀρχώμεθα, ταί (…)

Having explained how he is able to sing, the poet now restarts, so to speak,
the hymn to the Muses and the gods, which was announced by ἀρχώμεθα in
line 1:22 ‘au vers 36, ce “commencement” constitue clairement l’ application

19 Whatever the exact meaning of ταῦτα περὶ δρῦν ἢ περὶ πέτρην, it can hardly stand for speak-
ing about ‘irrelevant matters’ (Most, in a note on p. 5), since it must refer to the actions
of the Muses described in lines 30–34, and may even refer to Hesiod’s encounter with the
Muses as a whole. Perhaps περὶ δρῦν ἢ περὶ πέτρην = ‘things belonging to the past,’ cp. one
of the scholia ad loc. (τί μοι χρεία ἐστὶν ἀρχαιολογεῖν) and West (1966: 168).
20 Pucci (2007: 76) mainly discusses its use in the Iliad. Most translates ‘come then’, but τύνη
is not a hortative particle like ἄγε.
21 To be sure, the ‘we’ of the hortative subjunctive may be subdivided into ‘you’ and ‘me’ but
then two different persons are involved, as at Pl. Phdr. 237c6 ἐγὼ οὖν καὶ σὺ μὴ πάθωμεν ὃ
ἄλλοις ἐπιτιμῶμεν.
22 By the presence of τύνη, ἀρχώμεθα here does not simply pick up the form of line 1, where
I argued that it is used to involve the audience in the activity of the poet.
discourse cohesion in the proem of hesiod’s theogony 195

directe des recommandations formulées par les Muses au vers 34’ (Leclerc 1993:
170–171).23 This does not mean that we have to imagine the poet starting to sing
while the Muses are still on Mount Helicon: the hymn to the Muses is executed
within, not outside, the framework of the proem. To have the latter interpre-
tation the poet should have continued with ‘and so I began to sing from the
Muses (…).’ Temporally, we are here still at the implicit ‘now’ of line 1; the poet
really resumes his activity, be it in different circumstances, for the audience has
by now been fully informed about the legitimity of his undertaking.

36–37 ταὶ Διὶ πατρὶ | ὑμνεῦσαι τέρπουσι

With these words we leave the narrative and enter the omnitemporal world
of the gods on Olympus, in another attributive section, which describes the
hymnic activities of the Muses on Olympus and their effect: ὑμνεῦσαι τέρπουσι
(line 37), ῥέει αὐδή (line 39), γελᾷ (…) δώματα (line 40), ἠχεῖ (…) κάρη (line 42),
θεῶν γένος (…) κλείουσιν (line 44), ὑμνεῦσαι τέρπουσι (line 55), μέλπονται (line 66)
and κλείουσιν (line 67). We have already been informed that knowledge of the
world of the gods has been granted to Hesiod by the Muses themselves, viz.
when he overheard them singing of the gods of lines 11–21. The hymn first runs
up to and including θεῶν γένος κλείουσιν (line 44) plus the genealogical rela-
tive clauses (οὓς (…) ἔτικτεν, line 45, οἵ᾽ τ᾽ (…) ἐγένοντο, line 46), thereafter up
to and including line 52. This is followed by a ‘Mousogony’, a narrative digres-
sion within the attributive section, where both ἐν Πιερίῃ and Κρονίδῃ (…) πατρί
are to be connected with μιγεῖσα:24 Mnemosyne was visited by Zeus in her own
abode, where she mingled in love with him as a father. Cp. lines 56–57: ἐννέα
γάρ οἱ νύκτας ἐμίσγετο μητίετα Ζεὺς | νόσφιν ἀπ᾽ ἀθανάτων | ἱερὸν λέχος εἰσανα- 252
βαίνων.25 Here, ἐννέα (…) νύκτας ἐμίσγετο specifies the nature of the encounter
of Zeus and Mnemosyne, which was first presented by a complexive aorist
(μιγεῖσα), and explains the number of the Muses who will be born. Lines 60
and 62 conclude the Mousogony; although the time of the birth of the Muses
is not mentioned, the place is specified: at line 60 ἔτεκ᾽ picks up the τέκε of
line 53, while line 62 adds the place: ἡ δ᾽ ἔτεκ᾽ ἐννέα κούρας, ὁμόφρονας (…) | ,

23 Strangely enough Minton seems to take Μουσάων as the addressees of these words, since
he writes (1970: 358) that line 36 is an ‘exhortation to the Muse to celebrate the god.’
24 Often, ἐν Πιερίῃ is taken with both μιγεῖσα and τέκε, or with τέκε only, but this is less likely
in view of line 62. See below, note 26.
25 Zeus visited Maia in the same way, cp. h. 4.1–7 (Hermes) ὃν τέκε Μαῖα | νύμφη ἐϋπλόκαμος
Διὸς ἐν φιλότητι μιγεῖσα | αἰδοίη· μακάρων δὲ θεῶν ἠλεύαθ᾽ ὅμιλον | ἄντρον ἔσω ναίουσα παλί-
σκιον, ἔνθα Κρονίων | νύμφῃ ἐϋπλοκάμῳ μισγέσκετο νυκτὸς ἀμολγῷ, | ὄφρα κατὰ γλυκὺς ὕπνος
ἔχοι λευκώλενον Ἥρην.
 196 chapter 9

(…) | τυτθὸν ἀπ᾽ ἀκροτάτης κορυφῆς νιφόεντος Ὀλύμπου.26 In between there is

a brief attributive section at lines 60–61 (ᾗσιν ἀοιδὴ | μέμβλεται (…)), which is
continued at lines 63–67 (ἔνθα σφιν (…)).27

68–70 αἳ τότ᾽ ἴσαν πρὸς Ὄλυμπον, ἀγαλλόμεναι ὀπὶ καλῇ,

ἀμβροσίῃ μολπῇ· περὶ δ᾽ ἴαχε γαῖα μέλαινα
ὑμνεύσαις, ἐρατὸς δὲ ποδῶν ὕπο δοῦπος ὀρώρει28
What does τότε refer to? Most scholars take it to refer to the birth of the Muses
mentioned in line 60. Thus e.g. West on line 68: ‘The first thing a newborn god
does—even if he is born practically on the summit of Olympus!—is to go and
join the other gods’, and likewise Minton (1970: 363): ‘Their dwelling place is
just below the peak of snowy Olympus, and there they join with the Graces and
Desire in choral song praising the gods. Then (emphasis Minton) they proceed
253 up to the halls of | Zeus on the peak of Olympus (…)’,29 Thalmann (1984: 137):
‘68–79: Muses’ procession to Olympos at the time of their birth,’ Clay (2003: 68):
‘Immediately after their birth, the goddesses, singing and dancing, all the while,
made their ascent to their father’ and Pucci (2007: 97): ‘le Muse andarono a pre-
sentarsi al padre loro, Zeus’. Cp. also the translation of e.g. Bonnafé: ‘Sitôt nées,
les voilà qui partaient vers l’Olympe.’ This must be rejected, however, since it
yields an impossible situation. First of all, the Muses, whether their birth is

26 Can this be reconciled with the information of line 53 if this is taken as ‘Mnemosyne bore
her children in Pieria’? Or, to put it differently, can Pieria be said to be located ‘at a small
distance from the highest top of the Olympus’? I think not. ‘Pieria is the region immedi-
ately to the north of Olympus, between it and the Haliacmon’ (West). In fact, the nearest
it gets to Olympus seems to be as a slope of that mountain, cp. E. Ba. 409–411 ἁ καλλιστευ-
ομένα Πιερία | μούσειος ἕδρα, | σεμνὰ κλιτὺς Ὀλύμπου. Cp. also Eustath. Ad Iliadem 3.623.23,
where Pieria is called an ἀκρώρεια (‘ridge’) Ὀλύμπου. Nowhere is Pieria mentioned as a
region near the summit. I take it, then, that the Muses were born not in Pieria but near
the summit of Olympus, and that they were begot in Pieria (see on line 36). Another god
born on Olympus is Hermes, at least according to a tradition that is found in Philostr. Im.
1.26.1: (Hermes) τίκτεται μὲν ἐν κορυφαῖς τοῦ Ὀλύμπου, κατ᾽ αὐτὸ ἄνω τὸ ἕδος τῶν θεῶν, and
that probably goes back to Alc. fr. 2 D. χαῖρε, Κυλλάνας ὀ μέδεις, σὲ γάρ μοι | θῦμος ὔμνην, τὸν
κορύφαισιν † αὐγαῖς † | Μαῖα γέννατο Κρονίδαι μίγεισα | παμβασίληϊ. In the hymn to Hermes,
the god is born in Maia’s cave, cp. line 23, where Hermes, whose birth has been mentioned
in line 13, steps over οὐδὸν (…) ἄντροιο.
27 Line 62 can hardly belong to the attributive section, for in that case τυτθὸν (…) Ὀλύμπου
would modify μέμβλεται. But why would the Muses’ care for song be locally restricted?
28 ‘They then went towards Olympus, exulting in their beautiful voice, and around them the
black earth resounded as they sang’ (translation Most, adapted).
29 Minton’s idea that line 68 belongs to the same time sphere as lines 60–67 (see his ‘Then
they proceed’), as if there is no imperfect (or did he take this as timeless?), is not supported
by the text.
discourse cohesion in the proem of hesiod’s theogony 197

located in Pieria as part of Olympus (cp. note 26), or, as I prefer, near the sum-
mit of Olympus (cp. again note 26), are already on Olympus, so they cannot go
there. Note that in West’s explanatory note πρὸς Ὄλυμπον is omitted, while his
‘to join the gods’ is not in the Greek. In fact, the basic flaw of the above interpre-
tations is that they ignore line 62. Second, if a newborn god does go to Olympus,
he/she, being an infant, is brought there, as in two of the parallels adduced by
West for his claim that ‘[t]he first thing a newborn god does (…) is to go and
join the other gods’, viz.

h. 6.14–16 (subject: the Horai, object: Aphrodite)

αὐτάρ ἐπεὶ δὴ πάντα περὶ χροΐ κόσμον ἔθηκαν

45 ἦγον ἐς ἀθανάτους· οἱ δ᾽ ἠσπάζοντο ἰδόντες
χερσί τ᾽ ἐδεξιόωντο καὶ (…).30

After Zephyros has brought Aphrodite to the shores of Cyprus (lines 3–4) and
the Horai have taken her over (δέξαντ᾽, line 6), the latter start clothing and
adorning her (lines 6–13), which is followed by the lines quoted above. There is
no activity of Aphrodite herself: it is the Horai who bring her to the gods.

h. 19.40–45 (τόν = Pan)

40 τὸν δ᾽ αἶψ᾽ Ἑρμείας ἐριούνιος εἰς χέρα θῆκε

δεξάμενος, χαῖρεν δὲ νόῳ περιώσια δαίμων.
ῥίμφα δ᾽ ἐς ἀθανάτων ἕδρας κίε παῖδα καλύψας
δέρμασιν ἐν πυκινοῖσιν ὀρεσκῴοιο λαγωοῦ.
πὰρ δὲ Ζηνὶ καθῖζε καὶ ἄλλοις ἀθανάτοισιν,
45 δεῖξε δὲ κοῦρον ἑόν·31

In this case, too, the newborn god is not active himself, cp. lines 42 and 45: it is 254
Hermes who brings Pan to the assembly of the gods.

30 ‘When they had wholly arrayed the goddess in splendor, | They led her to the immortals,
who, seeing her, welcomed her kindly | and offered their right hands in greeting …’ (trans-
lation Sargent).
31 ‘But luck-bringing Hermes, receiving his son, at once | Took him into his arms, his heart
filled with joy beyond measure, | And swiftly he went to the seat of the undying gods |
With the child warmly wrapped in the skin of mountain-bred hares, | And he sat down
beside Zeus and the other immortals, | And presented his son;’ (translation Sargent, with
corrections in lines 44–45).
 198 chapter 9

If, therefore, Theogony 60–68 would be about the Muses joining the gods
immediately after their birth, one would expect them to be brought there as
well, rather than to go (ἴσαν, 68) there all by themselves, let alone ‘singing and
dancing, all the while’.32
Now to return to τότ᾽ ἴσαν, what does τότε refer to, then? I take it that the
imperfect ἴσαν, the first imperfect having the Muses as its subject after στεῖχον
at line 10, is the sign that the poet resumes the narrative about his encounter
with the Muses, which he had broken off for a hymn to them, and that τότε
refers back to the form which concluded the first part of the narrative, i.e. ἐδί-
δαξαν (line 22).33 After they have bestowed ἀοιδή on Hesiod their mission on
Helicon is completed, and, there being no more need for their στείχειν on its
slopes, they now go (= return) to the Olympus, where their δώματα are (line 63).
For in principle νίσομαι (νισομένων 72), a variant of νέομαι, does not mean ‘go’
but ‘return’. Cp. Chantraine 1968–1980: s.v. νέομαι: ‘«revenir, retourner» (notam-
ment chez soi), rarement aus sens de «venir»’, Lexikon des frühgr. Epos s.v. νίσο-
255 μαι: ‘“zurückkehren”, s.v. νέομαι: “heimkehren.”’34 (LSJ: ‘go, come’, wrongly). |
A similar use of τότε is found at Odyssey 15.256–257 τοῦ μὲν ἄρ᾽ υἱὸς ἐπῆλθε,
Θεοκλύμενος δ᾽ ὄνομ᾽ ἦεν, | ὃς τότε Τηλεμάχου πέλας ἵστατο, which comes after
a long genealogical digression about the descent of Theoclymenus. Here, τότε
ultimately refers back, via ἐπῆλθε at 256, to ἤλυθεν at 223, a distance of 34 lines.35

32 Only newborn Hermes does not behave like a baby, see the elaborate description of the
first day of his life at h. 4.20–102, which was explicitly announced in lines 13–18: καὶ τότ᾽
ἐγείνατο παῖδα πολύτροπον, αἱμυλομήτην, | ληϊστῆρ᾽, ἐλατῆρα βοῶν, ἡγήτορ᾽ ὀνείρων, | νυκτὸς
ὀπωπητῆρα, πυληδόκον, ὃς τάχ᾽ ἔμελλεν | ἀμφανέειν κλυτὰ ἔργα μετ᾽ ἀθανάτοισι θεοῖσιν. | ἠῷος
γεγονὼς μέσῳ ἤματι ἐγκιθάριζεν, | ἑσπέριος βοῦς κλέψεν ἑκηβόλου Ἀπόλλωνος, | τετράδι τῇ
προτέρῃ τῇ μιν τέκε πότνια Μαῖα. Nothing of the kind is found in our passage from the
Theogony. Incidentally, ‘the first thing a newborn god does’ is in Hermes’ case definitely
not to go to Olympus. As for the other passage mentioned by West (Th. 201–202), this is
not about a first voyage to Olympus either: τῇ δ᾽ Ἔρος ὡμάρτησε καὶ Ἵμερος ἔσπετο καλὸς |
γεινομένῃ τὰ πρῶτα θεῶν τ᾽ ἐς φῦλον ἰούσῃ. Here we are just told that Eros accompanied, and
Himeros followed, Aphrodite ‘as soon as she was born and when she went to the gods’, not
necessarily, then, immediately after her birth; τὰ πρῶτα must be connected with γεινομένῃ
only, cp. Il. 6.489, Od. 8.553 ἐπὴν τὰ πρῶτα γένηται. In fact, we had already been told that
the first thing Aphrodite did after she had grown in the foam of the sea (lines 191–192) was
to go to Cythera (lines 192–193), whereafter she went to Cyprus (line 193).
33 According to Verdenius τότε refers to στεῖχον (line 10), but this view is less likely, since it
ignores the presence of ἐδίδαξαν.
34 The lemma in the LfgrE lacks precision, for it is not clear whether νισομένων at Th. 71 is
viewed as an instance of zurückkehren, or of hin(und her)gehen/fahren. Also, the lemma
wrongly claims (top of col. 407) that the Muses go ‘durch die Luft’.
35 ἐπῆλθε at line 256 merely repeats (σχεδόθεν (…)) ἤλυθεν (line 223), there is no temporal pro-
gression. This entails that to know the precise reference point of τότε the hearer/reader
discourse cohesion in the proem of hesiod’s theogony 199

Naturally, our case differs from Od. 15.256–257 in that in the Theogony, if τότε is
meant to refer back to ἐδίδαξαν at line 22, it has to cross the barrier, so to speak,
of line 35 ἀλλὰ τίη μοι ταῦτα περὶ δρῦν ἢ περὶ πέτρην. Irene de Jong (p.c.) considers
this barrier unsurmountable. She points out that, in the Iliad, the comparable
formula ἀλλὰ τίη μοι ταῦτα φίλος διελέξατο θυμός (…) is used five times to put
an end to some line of thought for good (11.407, 17.97, 21.562, 22.122, 22.385),
and that, in Pindar, too, after a so-called Abbruchsformel (e.g. N. 3.26) the poet
never returns to the matter presented before this formula. This may be true,
but ignores the fact that the proem of the Theogony, presenting as it does a
mixture of hymn and personal narrative, is quite different from both the solil-
oquies and speeches in the Iliad and the Pindaric odes. As a matter of fact, as I
argued already above, the proem is entirely sui generis.
The movements of the Muses at Th. 68–70 are still witnessed by the poet, cp.
the presence of ὀπί (line 68), ἴαχε (line 69), ὑμνεύσαις and ἐρατὸς δὲ ποδῶν ὕπο
δοῦπος ὀρώρει (70), which presuppose the presence of a human perceiver, and
continue the situation of lines 10–11 στεῖχον περικαλλέα ὄσσαν ἱεῖσαι, | ὑμνεῦσαι
Δία τ᾽ αἰγίοχον καὶ πότνιαν Ἥρην etc. In the same vein ἄειδον (line 75) refers to
an action that is perceived by the poet as well. This entails that ταῦτ᾽ (…) ἄει-
δον turns lines 71–74 into songs heard by Hesiod, while ultimately ταῦτα also
comprises the song of the Muses at lines 11–21.36 | 256

71 ἐμβασιλεύει
The singing of the Muses about the kingship of Zeus on Olympus mentioned
in this brief attributive section prepares us for the special bond of the Muses
with earthly kings, who come to the fore in lines 80–92.

83–84 τῷ μὲν ἐπὶ γλώσσῃ γλυκερὴν χείουσιν ἐέρσην,

τοῦ δ᾽ ἔπε’ ἐκ στόματος ῥεῖ μείλιχα·
These words echo lines 39–40: (τῶν = the Muses) τῶν δ᾽ ἀκάματος ῥέει αὐδὴ |
ἐκ στομάτων ἡδεῖα, naturally without ἀκάματος: mortal kings (and singers, see
line 97) are not infatigable.

has to go back 34 lines. One may also compare Od. 7.133 (quoted in note 11), where ἔνθα
resumes the narrative after 49 lines of focalised and omnitemporal descriptions. For a still
greater distance spanned by τότε see αἷς ἐχειμάσθην τότε, spoken by the Guard, at S. Ant.
391, which spans a distance of some 80 lines, since it refers back to lines 305–314, where
Creon speaks threateningly to the Guard.
36 For a similar wide-ranging ταῦτα with a form of ἀείδειν see Od. 8.367, where ταῦτ᾽ ἄρ᾽ ἀοιδὸς
ἄειδε summarizes the whole of lines 267–366, that were introduced by αὐτὰρ ὁ φορμίζων
ἀνεβάλλετο καλὸν ἀείδειν at 266.
 200 chapter 9

93 τοίη Μουσάων ἱερὴ δόσις ἀνθρώποισιν

Here τοίη (…) δόσις summarizes the preceding attributive section (lines 79–92)
on Calliope as a whole, while δόσις refers back specifically to line 83 τῷ μὲν
ἐπὶ γλώσσῃ γλυκερὴν χείουσιν ἐέρσην—and ultimately echoes ἔδον, line 30—and
generic ἀνθρώποισιν expresses the idea that eventually the gift of the Muses was
not just meant for the kings, but for all mankind, through the services of the
kings as well as of the singers. The latter are introduced through a paratactic
analogon (lines 94–96): while the kings are the offspring of Zeus, as was implied
by lines 81–82, the singers are the offspring of the Muses and Apollo. For this
rhetorical figure, which mostly involves the use of μέν (…) δέ, see the discus-
sion in K-G (2, 232–233), and Denniston (1954: 370–371). A typical example is
Il. 1.163–164, mentioned by Kühner-Gerth: (ὡς ἔμ᾽ ἀφαιρεῖται Χρυσηΐδα Φοῖβος
Ἀπόλλων, |) τὴν μὲν ἐγὼ σὺν νηΐ τ᾽ ἐμῇ καὶ ἐμοῖς ἑτάροισι | πέμψω, ἐγὼ δέ κ᾽ ἄγω
Βρισηΐδα καλλιπάρῃον ‘so werde ich, während ich diese (…) schicken werde, die
Bris. in das Zelt führen.’37 At Theogony 94–96 we are dealing with the rather rare
variant in which it is the first clause that ‘carries the weight’ (Denniston 1954:
370); Denniston mentions as an example E. Cyc. 199, where while μέν is lacking
τοι is present.

94 ἐκ γάρ τοι (…)

τοι appeals to Hesiod’s addressee in the human world, who was already implied
by ἀνθρώποισιν (line 93). Cp. Denniston (1954: 537): ‘(…) τοι implies (…) an audi-
257 ence (…).’ See also note 4. |

96–97 (…) ὁ δ᾽ ὄλβιος, ὅντινα Μοῦσαι

φίλωνται· γλυκερή οἱ ἀπὸ στόματος ῥέει αὐδή.
The position of this sentence after ἐκ δὲ Διὸς βασιλῆες might suggest that the
poet is speaking about the kings. Actually, however, he must be speaking about
the singers (or rather singer, see below), who were the most important ele-
ment of the preceding sentence. And, in fact, the ‘rulers [have] already been
treated’ (Thalmann 228–229, fn. 16), viz. at lines 81–92. Observe that the phras-
ing here differs from that at line 81: there we read, in connection with kings:
ὅντινα τιμήσουσι Διὸς κοῦραι μεγάλοιο, but here ὅντινα Μοῦσαι φίλωνται, where
Μοῦσαι instead of Διὸς κοῦραι points to a different group, viz. the singers, being
meant. Compare Od. 8.480 Μοῦσ᾽ (…) φίλησε δὲ φῦλον ἀοιδῶν. Observe also that
at line 81, in the context of διοτρεφέων βασιλήων, the Muses are referred to as Διὸς

37 Μέν may also be omitted, for which see K-G (2, 233) and Denniston (1954: 165): ‘[μέν is
sometimes omitted] (…) even when the close connexion between coordinated clauses
implies logical subordination.’
discourse cohesion in the proem of hesiod’s theogony 201

κοῦραι, but here as Μοῦσαι. Finally, it is worth observing that the singer who is
loved by the Muses is called ὄλβιος (line 96), while the king honoured by the
Διὸς κοῦραι (line 81) is not provided with a similar laudatory predicative adjec-
tive: singers are just a bit more valued than kings.

97 γλυκερή οἱ ἀπὸ στόματος ῥέει αὐδή

These words echo lines 83–84 (quo vide), and through them lines 39–40.

At lines 98–103 the singers at last get their due:

εἰ γάρ τις καὶ πένθος ἔχων νεοκηδέι θυμῷ

ἄζηται κραδίην ἀκαχήμενος, αὐτὰρ ἀοιδὸς
100 Μουσάων θεράπων κλεῖα προτέρων ἀνθρώπων
ὑμνήσει μάκαράς τε θεοὺς οἳ Ὄλυμπον ἔχουσιν·
αἶψ᾽ ὅ γε δυσφροσυνέων ἐπιλήθεται οὐδέ τι κηδέων
μέμνηται· ταχέως δὲ παρέτραπε δῶρα θεάων.

With Mazon,38 I take it that αὐτάρ introduces the main clause to the condi-
tional clause εἰ (…) ἅζηται; ὑμνήσει is a future indicative and the main verb.
‘Even if someone, feeling sorrow in a newly anguished spirit, is parched in his
heart with grieving, yet a singer will sing of the | glorious deeds (…)’, i.e., how- 258
ever sad the circumstances, there will always be a singer to oppose the glorious,
comforting, deeds of earlier people, and the blessed gods, to the sorrow of the
grieving man. For apodotic αὐτάρ + future indicative after εἰ see Il. 3.288–291
εἰ δ᾽ ἂν ἐμοὶ τιμὴν Πρίαμος Πριάμοιό τε παῖδες | τίνειν οὐκ ἐθέλωσιν Ἀλεξάνδροιο
πεσόντος, | αὐτὰρ ἐγὼ καὶ ἔπειτα μαχήσομαι (…); Il. 22.389–390 εἰ δὲ θανόντων περ
καταλήθοντ᾽ εἰν Ἀΐδαο | αὐτὰρ ἐγὼ καὶ κεῖθι φίλου μεμνήσομ᾽ ἑταίρου, and cp. Den-
niston (1954: 55). For the use of the future indicative in a habitual context cp.
K-G (1, 171) and especially Pfeijffer (1999: 45–52 ‘Futures expressing a general
thought’). Furthermore, the change from the plural ἀοιδοί at line 95 to the sin-
gular ἀοιδός here should be seen as a sign that Hesiod is implicitly referring to
himself. Finally, note that I have put a high dot after ἔχουσιν; the last two lines

38 ‘Un homme porte-t-il le deuil dans son cœur (…)? qu’un chanteur (…) célèbre les hauts
faits des hommes d’ autrefois ou les dieux bienheureux, habitants de l’ Olympe: vite, il
oublie ses déplaisirs (…).’ Mazon prefers the variant reading ὑμνήσῃ, which he apparently
takes as a hortative subjunctive. For this—extremely rare—use in the third person cp.
Chantraine (1953: 207). I should add that Mazon’s punctuation of the Greek text—he has
a comma before αἶψ(α), line 102—does not match that of his translation, which has a colon
before ‘vite’. For a different view of this sentence see below.
 202 chapter 9

follow therefore asyndetically: they describe the effect of ὑμνήσει. For similar
effect-describing clauses introduced by αἶψα cp. Hes. Op. 43–46 ῥηιδίως γάρ κεν
καὶ ἐπ᾽ ἤματι ἐργάσσαιο, | ὥστε σε κεἰς ἐνιαυτὸν ἔχειν καὶ ἀεργὸν ἐόντα· | αἶψά κε
πηδάλιον μὲν ὑπὲρ καπνοῦ καταθεῖο, Od. 1.392 οὐ μὲν γάρ τι κακὸν βασιλευέμεν·
αἶψά τέ οἱ δῶ | ἀφνειὸν πέλεται καὶ τιμηέστερος αὐτός, 15.313–317 καί κ᾽ ἐλθὼν (…) |
ἀγγελίην εἴποιμι (…) | καί κε μνηστήρεσσιν (…) μιγείην, | εἴ μοι δεῖπνον δοῖεν ὀνείατα
μυρί᾽ ἔχοντες. | αἶψά κεν εὖ δρώοιμι μετὰ σφίσιν, ὅττι θέλοιεν.
Others take a different view of the syntax of our sentence. In their view, αὐτάρ
connects ἅζηται and ὑμνήσει, which is taken as a subjunctive, and ἐπιλήθεται at
line 101 is taken as the main verb. This is less likely, for although αὐτάρ can,
indeed, connect two verbs in a conditional clause (Od. 18.376–377 εἰ δ᾽ αὖ (…)
ὁρμήσειε Κρονίων | σήμερον, αὐτὰρ ἐμοὶ σάκος εἴη καὶ δύο δοῦρε | (…) | τῶ κέ μ᾽ ἴδοις
(…)) the result at Th. 99 is very lame. If the ἀοιδός is the subject of the condi-
tional clause, on a par with τις, he makes a most inconspicuous entrance, while
we might expect him to be the most important element of the passage, and to
contrast with the grieving man.39 Observe that at Il. 3.288–291 and 22.389–390,
259 quoted above, αὐτάρ is followed by strongly contrastive ἐγώ. |

100–101 κλεῖα προτέρων ἀνθρώπων

ὑμνήσει μάκαράς τε θεοὺς οἳ Ὄλυμπον ἔχουσιν·
Just as the Muses Διὶ πατρὶ | ὑμνεῦσαι τέρπουσι μέγαν νόον ἐντός Ὀλύμπου (lines
36–37) by glorifying in their song the venerated race of the gods (θεῶν γένος
αἰδοῖον, line 44) and Zeus himself (Ζῆνα, line 47), and by singing of the race
of human beings and the Gigants (ἀνθρώπων τε γένος κρατερῶν τε Γιγάντων,
line 50), so the mortal singer, i.e. Hesiod, will comfort the human beings by
singing about the κλεῖα προτέρων ἀνδρῶν (…) μάκαράς τε θεοὺς οἳ Ὄλυμπον ἔχου-
σιν, lines 100–101. See also at line 114.

102 αἶψα
This echoes αἶψα at line 87: both a king and a singer have success quickly.

39 Stretching the meaning of αὐτάρ West notes: ‘we would say “and then”.’ But ‘then’ would
not seem to be part of the meaning of αὐτάρ. Observe that at Od. 18.376–377, quoted in the
main text, αὐτάρ cannot possibly be = ‘and then’. Stretching the meaning of both εἰ and
αὐτάρ, Most translates ‘even if (…), yet when (…)’. This will not do for ‘when’ is not in the
Greek: if ὑμνήσει, like ἅζηται, is taken as a subjunctive and still depends on εἰ γὰρ καί, the
translation should be ‘even if someone is parched with grieving, and a singer sings about
the deeds of (…)’, which gives an impossible meaning.
discourse cohesion in the proem of hesiod’s theogony 203

102–103 οὐδέ τι κηδέων | μέμνηται·

Echoes lines 60–61: ᾗσιν ἀοιδὴ | μέμβλεται, ἐν στήθεσσιν ἀκηδέα θυμόν ἐχούσαις.
Just as ἀοιδή for the Muses is connected with the absence of κήδεα, so the ἀοι-
δός (line 99) is connected with dissipating the κήδεα of human beings.

103 ἐπιλήθεται οὐδέ τι κηδέων

μέμνηται· ταχέως δὲ παρέτραπε δῶρα θεάων.
(Gnomic) παρέτραπε at line 104 probably expresses an activity that is anterior
to (omnitemporal) ἐπιλήθεται and μέμνηται.
Παρ- in παρέτραπε echoes παρ- in παραιφάμενοι (line 90)—both kings and
singers ‘divert’ their audience, from bad plans and sorrows, respectively—,
while δῶρα echoes δόσις (line 93) as well as ἔδον (line 30) and announces δότε
(line 104).

104–106 χαίρετε τέκνα Διός, δότε δ᾽ ἱμερόεσσαν ἀοιδήν·

κλείετε δ᾽ ἀθανάτων ἱερὸν γένος αἰὲν ἐόντων,
οἳ Γῆς ἐξεγένοντο (…)
Since Hesiod has now sung extensively of the Muses, and has thereby executed
the second part of the order given to him at lines 33–34, he apparently feels free,
being truly their θεράπων, to salute the Muses and ask them to give him ἀοιδή,
so that he can execute the first part of their order, viz. to sing of the race of the
gods (ὑμνεῖν μακάρων γένος αἰὲν ἐόντων, line 33). Note that this is the first time
he has to ask for their assistance: at line 31 the Muses had of their own accord
provided him with the power to sing. His request basically amounts to ‘breath | 260
again a divine voice into me, so as to enable me now to glorify the history of the
gods’, as part of τὰ πρὸ ἐόντα, cp. line 32 ἵνα κλείοιμι τά τ᾽ ἐσσόμενα πρό τ᾽ ἐόντα.

114 ταῦτά μοι ἔσπετε Μοῦσαι Ὀλύμπια δώματ᾽ ἔχουσαι

ἐξ ἀρχῆς, καὶ εἴπαθ᾽, ὅτι πρῶτον γένετ᾽ αὐτῶν.
Like the Muses among the gods (cp. lines 43–46 αἱ δ᾽ ἄμβροτον ὄσσαν ἱεῖσαι |
θεῶν γένος αἰδοῖον πρῶτον κλείουσιν ἀοιδῇ | ἐξ ἀρχῆς, οὓς Γαῖα καὶ Οὐρανὸς εὐρὺς
ἔτικτεν, οἵ τ᾽ ἐκ τῶν ἐγένοντο), Hesiod is going to sing among mortals of the gods
ἐξ ἀρχῆς. He in fact appeals to the Muses to transfer their activities on Olympus
to an audience of mortals, using Hesiod as their θεράπων. Compare Thalmann
(1984: 139): ‘The Muses’ gift to the mortal poet, the Theogony itself, is the human
realization of the divine song.’
 204 chapter 9

Conclusion

Verdenius’ view, that the ‘(…) unity of the proem [lies] in the continuity of
its progress’ could be shown to be correct. The progress is established by the
sequence στεῖχον (line 10), ποτ᾽ ἐδίδαξαν (line 22), viz. during the στείχειν of
the Muses, and αἳ τότ᾽ ἴσαν (line 68) (…) νισομένων πατέρ᾽ εἰς ὅν (line 71), τότε
referring back to ἐδίδαξαν. This course of the actions strongly suggests that the
Muses seized the occasion of one of their regular visits to Helicon to dance
there (ὀρχεῦνται, line 4), to bestow the gift of poetry on Hesiod. But, contrary to
what Verdenius believed, the unity is also brought about by a number of inter-
dependencies. Thus, the imperfect verb forms which express a sound (ἄειδον,
line 75), or are accompanied by verb forms expressing a sound (στεῖχον περικαλ-
λέα ὄσσαν ἱεῖσαι, | ὑμνεῦσαι, line 10; αἳ τότ᾽ ἴσαν (…), ἀγαλλόμεναι ὀπὶ καλῇ, | (…)
μολπῇ· (…) ἴαχε | ὑμνεύσαις, (…) δοῦπος, lines 68–70) presupppose the presence
of a human perceiver: the poet himself, who is present, so to speak, during the
whole proem; they all refer to the ‘same’ past of the poet. Also, there is a recur-
rent reference to Hesiod acting as the servant (θεράπων, line 100) of the Muses,
to do among mortals what they themselves do among the gods, to sing both of
the gods and of mankind:
– line 33: the Muses to Hesiod: καί μ᾽ ἐκέλονθ᾽ ὑμνεῖν μακάρων γένος αἰὲν ἐόν-
261 των; |
– lines 43–45: the Muses on Olympus: αἱ δ᾽ ἄμβροτον ὄσσαν ἱεῖσαι | θεῶν γένος
αἰδοῖον πρῶτον κλείουσιν ἀοιδῇ | ἐξ ἀρχῆς, οὓς Γαῖα καὶ Οὐρανὸς εὐρὺς ἔτικτεν, οἵ
τ᾽ ἐκ τῶν ἐγένοντο;
– lines 50–51: the Muses on Olympus: (…) ἀνθρώπων τε γένος κρατερῶν τε Γιγάν-
των | ὑμνεῦσαι τέρπουσι Διὸς νόον ἐντὸς Ὀλύμπου
– lines 99–101: a mortal singer among mortals: ἀοιδὸς | Μουσάων θεράπων κλεῖα
προτέρων ἀνθρώπων | ὑμνήσει μάκαράς τε θεοὺς οἳ Ὄλυμπον ἔχουσιν;
– lines 104–105: Hesiod among mortals, to the Muses: δότε δ᾽ ἱμερόεσσαν ἀοι-
δήν· | κλείετε δ᾽ ἀθανάτων ἱερὸν γένος αἰὲν ἐόντων, 108 εἴπατε δ᾽ ὡς τὰ πρῶτα θεοὶ
καὶ γαῖα γένοντο, 114–115 ταῦτά μοι ἔσπετε Μοῦσαι Ὀλύμπια δώματ᾽ ἔχουσαι | ἐξ
ἀρχῆς, καὶ εἴπαθ᾽, ὅτι πρῶτον γένετ᾽ αὐτῶν.

Appendix to στεῖχον, line 10: a Focalising Imperfect in the Myth of

Plato’s Phaedrus

In the myth of Plato’s Phaedrus, more specifically in the passage 253e5–255a1,

the omnitemporal, habitual behaviour of the horses of the soul is described
by means of present and (gnomic) aorist indicatives. But in between there
 discourse cohesion in the proem of hesiod’s theogony 205

is, at 254d6, just one imperfect form, ἦσαν, in a temporal clause: (…) κατέχει
(…) ἐντρέπεται, (…) φέρεται, (…) ἀναγκάζει (…) ἀντιτείνετον (254b1) (…) πορεύε-
σθον (…), καὶ (254b4) πρὸς αὐτῷ τ᾽ ἐγένοντο καὶ εἶδον (…) ἠνέχθη, καὶ πάλιν εἶδεν
(…) ἔδεισε (…). ἀνέπεσεν (…) ἠναγκάσθη (…) (254c5) ἔβρεξε (…) ἐλοιδόρησεν (…)
(254d2) μόγις σννεχώρησεν δεομένων εἰς αὖθις ὑπερβαλέσθαι. ἐλθόντος δέ τοῦ συντε-
θέντος χρόνου (…) βιαζόμενος, χρεμετίζων, ἕλκων (254d5) ἠνάγκασεν αὖ προσελθεῖν
τοῖς παιδικοῖς ἐπὶ τοὺς αὐτοὺς λόγους, καὶ ἐπειδὴ ἐγγὺς ἦσαν, ἐγκύψας καὶ ἐκτείνας
τὴν κέρκον, ἐνδακὼν τὸν χαλινόν, μετ᾽ ἀναιδείας ἕλκει· ὁ δ᾽ (254e1) ἡνίοχος (…) τήν
τε κακηγόρον γλῶτταν καὶ τὰς γνάθους καθῇμαξεν καὶ τὰ σκέλη (…) ὀδύναις ἔδωκεν.
ὅταν δὲ (…) λήξῃ, ταπεινωθεὶς ἕπεται ἤδη τῇ τοῦ ἡνιόχου προνοίᾳ, καὶ ὅταν ἴδῃ τὸν
καλόν, φόβῳ διόλλυται· etc.
This imperfect is often misunderstood, and has been translated as if it were
a present indicative or a gnomic | aorist. Here follow some faulty English trans- 262
lations: Jowett (18923): (the unruly horse) ‘(…) forces them [i.e. the other horse
and the charioteer] to draw near again. And when they are near he stoops
his head (…), and (…) pulls shamelessly’, Fowler (Loeb, 1914): ‘(…) he forces
them again (…) to approach the beloved one, and when they are near him,
he lowers his head (…) and pulls shamelessly’, Rowe (1986; he has no note on
ἦσαν): ‘(…) it forces them to approach the beloved again (…), and when they
are nearby (…) it pulls shamelessly.’ ‘When (…) are’ is of course impossible: to
have ‘when’ in English, the Greek should have had ὅταν or ἐπειδάν + subjunc-
tive, as indeed later in the myth. Robin, however, correctly—and brilliantly—
translates (Budé, 1961), avoiding the use of a verb in his translation: ‘(…) une
fois de plus il les a contraints d’approcher du bien-aimé (…). Enfin, maintenant
que les voilà à proximité, il se penche en avant sur lui (…) il tire sans vergogne.’
The imperfect all of a sudden, dramatically, transfers the reader from the gen-
eral description of the behaviour of the ‘good’ horse and the charioteer to the
point of view of the ‘bad’ horse. Having forced the other two to approach the
beloved one (ἠνάγκασε προσελθεῖν: they have come close (aorist) to the beloved
one) the ‘bad’ horse sees that they are there: ‘now/seeing that they were close,
it lowers its head and (…) pulls shamelessly.’ | 263

The Structure of the Proem of Hesiod’s Theogony

Μουσάων Ἑλικωνιάδων ἀρχώμεθ᾽ ἀείδειν, 1 Metatextual formula

αἵ θ᾽ Ἑλικῶνος ἔχουσιν ὄρος μέγα τε ζάθεόν τε, 2–8 Hymn to the Muses part I attributive
καί τε περὶ κρήνην ἰοειδέα πόσσ᾽ ἁπαλοῖσιν section: description of some characteristic
ὀρχεῦνται καὶ βωμὸν ἐρισθενέος Κρονίωνος activities of the Muses on Mount Helicon
5 καί τε λοεσσάμεναι τέρενα χρόα Περμησσοῖο (omnitemporal presents: 2 ἔχουσιν, 4 ὀρχεῦν-
 206 chapter 9

ἀκροτάτῳ Ἑλικῶνι χοροὺς ἐνεποιήσαντο, ται, gnomic aorists: 7 ἐνεποιήσαντο, 8

καλοὺς ἱμερόεντας, ἐπερρώσαντο δὲ ποσσίν. ἐπερρώσαντο) passing into a narra-
ἠ᾽ Ἵππου κρήνης ἠ᾽ Ὀλμειοῦ ζαθέοιο tive
ἔνθεν ἀπορνύμεναι κεκαλυμμέναι ἠέρι πολλῷ 9–34 Personal narrative: Hesiod hears the
10 ἐννύχιαι στεῖχον περικαλλέα ὄσσαν ἱεῖσαι, Muses singing on the slopes of Mount Helicon
ὑμνεῦσαι Δία τ᾽ αἰγίοχον καὶ πότνιαν Ἥρην (imperfect + verb of sound: 10 στεῖχον (…)
Ἀργείην, χρυσέοισι πεδίλοις ἐμβεβαυῖαν, ὑμνεῦσαι)
κούρην τ᾽ αἰγιόχοιο Διὸς γλαυκῶπιν Ἀθήνην
Φοῖβόν τ᾽ Ἀπόλλωνα καὶ Ἄρτεμιν ἰοχέαιραν
15 ἠδὲ Ποσειδάωνα γαιήοχον ἐννοσίγαιον
καὶ Θέμιν αἰδοίην ἑλικοβλέφαρόν τ᾽ Ἀφροδίτην
Ἥβην τε χρυσοστέφανον καλήν τε Διώνην
Λητώ τ᾽ Ἰαπετόν τε ἰδὲ Κρόνον ἀγκυλομήτην
Ἠῶ τ᾽ Ἠέλιόν τε μέγαν λαμπράν τε Σελήνην
20 Γαῖάν τ᾽ Ὠκεανόν τε μέγαν καὶ Νύκτα μέλαιναν
ἄλλων τ᾽ ἀθανάτων ἱερὸν γένος αἰὲν ἐόντων.
αἵ νύ ποθ᾽ Ἡσίοδον καλὴν ἐδίδαξαν ἀοιδήν, (…) they bestow on him the ability to sing
ἄρνας ποιμαίνονθ᾽ Ἑλικῶνος ὕπο ζαθέοιο. (Dichterweihe) (aorist: 22 ἐδίδαξαν ἀοιδήν) …
τόνδε δέ με πρώτιστα θεαὶ πρὸς μῦθον ἔειπον,
25 Μοῦσαι Ὀλυμπιάδες, κοῦραι Διὸς αἰγιόχοιο·
‘ποιμένες ἄγραυλοι, κάκ᾽ ἐλέγχεα, γαστέρες οἶον,
ἴδμεν ψεύδεα πολλὰ λέγειν ἐτύμοισιν ὁμοῖα,
ἴδμεν δ᾽ εὖτ᾽ ἐθέλωμεν ἀληθέα γηρύσασθαι.’
ὣς ἔφασαν κοῦραι μεγάλου Διὸς ἀρτιέπειαι,
30 καί μοι σκῆπτρον ἔδον δάφνης ἐριθηλέος ὄζον
δρέψασαι, θηητόν· ἐνέπνευσαν δέ μοι αὐδὴν
θέσπιν, ἵνα κλείοιμι τά τ᾽ ἐσσόμενα πρό τ᾽ ἐόντα,
καί μ᾽ ἐκέλονθ᾽ ὑμνεῖν μακάρων γένος αἱὲν ἐόντων, 33 (…) and urge him to use his newly acquired
σφᾶς δ᾽ αὐτὰς πρῶτόν τε καὶ ὕστατον αἰὲν ἀείδειν. ability to sing of the gods and first of all of
35 ἀλλὰ τίη μοι ταῦτα περὶ δρῦν ἢ περὶ πέτρην; themselves (imperfect: 33 ἐκέλοντο)
τύνη, Μουσάων ἀρχώμεθα, ταὶ Διὶ πατρὶ 35–36 Metatextual formulas
ὑμνεῦσαι τέρπουσι μέγαν νόον ἐντὸς Ὀλύμπου, 36 ταί—67 Hymn to the Muses part II attribu-
εἴρουσαι τά τ᾽ ἐόντα τά τ᾽ ἐσσόμενα πρό τ᾽ ἐόντα, tive section: description of the hymnic activi-
φωνῇ ὁμηρεῦσαι, τῶν δ᾽ ἀκάματος ῥέει αὐδὴ ties of the Muses on Olympus, who sing of the
40 ἐκ στομάτων ἡδεῖα· γελᾷ δέ τε δώματα πατρὸς gods and of mankind
Ζηνὸς ἐριγδούποιο θεᾶν ὀπὶ λειριοέσσῃ (omnitemporal presents: 37 τέρπουσι, 39
σκιδναμένῃ, ἠχεῖ δὲ κάρη νιφόεντος Ὀλύμπου ῥέει, 40 γελᾷ, 42 ἠχεῖ, 44 κλείουσιν, 51 τέρ-
δώματά τ᾽ ἀθανάτων· αἱ δ᾽ ἄμβροτον ὄσσαν ἰεῖσαι πουσι, 66 μέλπονται, 67 κλείουσιν) [264]
θεῶν γένος αἰδοῖον πρῶτον κλείουσιν ἀοιδῇ
45 ἐξ ἀρχῆς, οὓς Γαῖα καὶ Οὐρανὸς εὐρὺς ἔτικτεν,
οἵ τ᾽ ἐκ τῶν ἐγένοντο, θεοὶ δωτῆρες ἐάων·
δεύτερον αὖτε Ζῆνα θεῶν πατέρ᾽ ἠδὲ καὶ ἀνδρῶν,
[ἀρχόμεναί θ᾽ ὑμνεῦσι θεαὶ † λήγουσαί τ᾽ ἀοιδῆς,]
ὅσσον φέρτατός ἐστι θεῶν κάρτει τε μέγιστος·
50 αὖτις δ᾽ ἀνθρώπων τε γένος κρατερῶν τε Γιγάντων
ὑμνεῦσαι τέρπουσι Διὸς νόον ἐντὸς Ὀλύμπου
Μοῦσαι Ὀλυμπιάδες, κοῦραι Διὸς αἰγιόχοιο.
 discourse cohesion in the proem of hesiod’s theogony 207

τὰς ἐν Πιερίῃ Κρονίδῃ τέκε πατρὶ μιγεῖσα 53–62 Mousogony: narrative digression
Μνημοσύνη, γουνοῖσιν Ἐλευθῆρος μεδέουσα, within the attributive section
55 λησμοσύνην τε κακῶν ἄμπαυμά τε μερμηράων. (non-gnomic aorists: 53 τέκε, 60 ἔτεκε, itera-
ἐννέα γάρ οἱ νύκτας ἐμίσγετο μητίετα Ζεὺς tive imperfect: 56 μίσγετο)
νόσφιν ἀπ᾽ ἀθανάτων ἰερὸν λέχος εἰσαναβαίνων·
ἀλλ᾽ ὅτε δή ῥ᾽ ἐνιαυτὸς ἔην, περὶ δ᾽ ἔτραπον ὧραι
μηνῶν φθινόντων, περί δ᾽ ἤματα πόλλ᾽ ἐτελέσθη,
60 ἡ δ᾽ ἔτεκ᾽ ἐννέα κούρας, ὀμόφρονας, ᾗσιν ἀοιδὴ
μέμβλεται ἐν στήθεσσιν, ἀκηδέα θυμόν ἐχούσαις,
τυτθὸν ἀπ᾽ ἀκροτάτης κορυφῆς νιφόεντος Ὀλύμπου·
ἔνθα σφιν λιπαροί τε χοροὶ καὶ δώματα καλά,
πὰρ δ᾽ αὐτῇς Χάριτές τε καὶ Ἵμερος οἰκί᾽ ἔχουσιν
65 ἐν θαλίῃς· ἐρατὴν δὲ διὰ στόμα ὄσσαν ἰεῖσαι
μέλπονται, πάντων τε νόμους καὶ ἤθεα κεδνὰ
ἀθανάτων κλείουσιν, ἐπήρατον ὄσσαν ἱεῖσαι. 67 End of attributive section on the Muses
αἳ τότ᾽ ἴσαν πρὸς Ὄλυμπον, ἀγαλλόμεναι ὀπὶ καλῇ, 68–79 Narrative about Hesiod and the Muses,
ἀμβροσίῃ μολπῇ· περὶ δ᾽ ἴαχε γαῖα μέλαινα continued: τότε refers back to 22 ἐδίδαξαν,
70 ὑμνεύσαις, ἐρατὸς δὲ ποδῶν ὕπο δοῦπος ὀρώρει return of the Muses to Olympus (their abode,
νισομένων πατέρ᾽ εἰς ὅν· ὁ δ᾽ οὐρανῷ ἐμβασιλεύει, 63) as perceived by the poet (reappearance
αὐτὸς ἔχων βροντὴν ἠδ᾽ αἰθαλόεντα κεραυνόν, (cp. 10) of imperfects and verbs of sound:
κάρτει νικήσας πατέρα Κρόνον· εὖ δὲ ἔκαστα 68 ἴσαν (…) ἀγαλλόμεναι ὀπὶ καλῇ, 69–70 ἴαχε
ἀθανάτοις διέταξε νόμους καὶ ἐπέφραδε τιμάς. γαῖα (…) ὑμνεύσαις, 70 δοῦπος ὀρώρει, 75 ἄει-
75 ταῦτ᾽ ἄρα Μοῦσαι ἄειδον Ὀλύμπια δώματ᾽ ἔχουσαι, δον)
ἐννέα θυγατέρες μεγάλου Διὸς ἐκγεγαυῖαι, 71–74 Interrupted by brief attributive section
Κλειώ τ᾽ Εὐτέρπη τε Θάλειά τε Μελπομένη τε on Zeus (omnitemporal present: ἐμβασι-
Τερψιχόρη τ᾽ Ἐρατώ τε Πολύμνιά τ᾽ Οὐρανίη τε λεύει) + brief narrative digression (non-
Καλλιόπη θ᾽· ἡ δὲ προφερεστάτη ἐστὶν ἁπασέων. gnomic aorists: διέταξε, ἐπέφραδε) [265]
80 ἡ γὰρ καὶ βασιλεῦσιν ἅμ᾽ αἰδοίοισιν ὀπηδεῖ. 79–80 Brief attributive section on Calliope,
ὅντινα τιμήσουσι Διὸς κοῦραι μεγάλοιο the most prominent of the Muses (omnitem-
γεινόμενόν τε ἴδωσι διοτρεφέων βασιλήων, poral presents: 79 ἐστίν, 80 ὀπηδεῖ), which
τῷ μὲν ἐπὶ γλώσσῃ γλυκερὴν χείουσιν ἐέρσην, triggers an attributive section-like descrip-
τοῦ δ᾽ ἔπε᾽ ἐκ στόματος ῥεῖ μείλιχα· οἱ δέ νυ λαοὶ tion of the Muses’ activities among mortals
85 πάντες ἐς αὐτὸν ὁρῶσι διακρίνοντα θέμιστας through kings (81–92) and singers (94–103);
ἰθείῃσι δίκῃσιν· ὁ δ᾽ ἀσφαλέως ἀγορεύων like the Muses, the latter sing of mankind
αἶψά τι καὶ μέγα νεῖκος ἐπισταμένως κατέπαυσε· and the gods (verbless clauses: 88 ἐχέφρονες,
τούνεκα γὰρ βασιλῆες ἐχέφρονες, οὕνεκα λαοῖς 93 τοίη, 96 ὄλβιος; omnitemporal presents:
βλαπτομένοις ἀγορῆφι μετάτροπα ἔργα τελεῦσι 83 χείουσιν, 84 ῥεῖ, 85 ὁρῶσι, 89 τελεῦσι, 91
90 ῥηιδίως, μαλακοῖσι παραιφάμενοι ἐπέεσσιν· ἱλάσκονται, 92 πρέπει, 95 ἔασιν, 97 ῥέει, 102
ἐρχόμενον δ᾽ ἀν᾽ ἀγῶνα θεὸν ὣς ἱλάσκονται ἐπιλήθεται, 103 μέμνηται; gnomic aorists: 87
αἰδοῖ μειλιχίῃ, μετὰ δὲ πρέπει ἀγρομένοισι. κατέπαυσε, 103 παρέτραπε; generic future:
τοίη Μουσάων ἱερὴ δόσις ἀνθρώποισιν. 101 ὑμνήσει)
ἐκ γάρ τοι Μουσέων καὶ ἑκηβόλου Ἀπόλλωνος
95 ἄνδρες ἀοιδοὶ ἔασιν ἐπὶ χθόνα καὶ κιθαρισταί,
ἐκ δὲ Διὸς βασιλῆες· ὁ δ᾽ ὄλβιος, ὄντινα Μοῦσαι
φίλωνται· γλυκερή οἱ ἀπὸ στόματος ῥέει αὐδή.
εἰ γάρ τις καὶ πένθος ἔχων νεοκηδέι θυμῷ
ἄζηται κραδίην ἀκαχήμενος, αὐτὰρ ἀοιδὸς
 208 chapter 9

100 Μουσάων θεράπων κλεῖα προτέρων ἀνθρώπων

ὑμνήσει μάκαράς τε θεοὺς οἳ Ὄλυμπον ἔχουσιν·
αἶψ᾽ ὅ γε δυσφροσυνέων ἐπιλήθεται οὐδέ τι κηδέων
μέμνηται· ταχέως δὲ παρέτραπε δῶρα θεάων.
χαίρετε τέκνα Διός, δότε δ᾽ ἱμερόεσσαν ἀοιδήν· 104–115 Salutation and prayer: having exe-
105 κλείετε δ᾽ ἀθανάτων ἱερὸν γένος αἰὲν ἐόντων, cuted the second part of the order of lines 33–
οἳ Γῆς ἐξεγένοντο καὶ Οὐρανοῦ ἀστερόεντος, 34, viz. to sing of the Muses, Hesiod now
Νυκτός τε δνοφερῆς, οὕς θ᾽ ἁλμυρὸς ἔτρεφε Πόντος. salutes the Muses and asks them to give him
εἴπατε δ᾽ ὡς τὰ πρῶτα θεοὶ καὶ γαῖα γένοντο ἀοιδή so that he can execute the first part of
καὶ ποταμοὶ καὶ πόντος ἀπείριτος οἴδματι θυίων their order, viz. to sing of the gods and their
110 ἄστρά τε λαμπετόωντα καὶ οὐρανὸς εὐρὺς ὕπερθεν· birth
[οἵ τ᾽ ἐκ τῶν ἐγένοντο, θεοὶ δωτῆρες ἐάων·]
ὥς τ᾽ ἄφενος δάσσαντο καὶ ὡς τιμὰς διέλοντο,
ἠδὲ καὶ ὡς τὰ πρῶτα πολύπτυχον ἔσχον Ὄλυμπον.
ταῦτά μοι ἔσπετε Μοῦσαι Ὀλύμπια δώματ᾽ ἔχουσαι
115 ἐξ ἀρχῆς, καὶ εἴπαθ᾽, ὅτι πρῶτον γένετ᾽ αὐτῶν.

Bibliography

Bakker, E.J., ‘Verbal Aspect and Mimetic Description in Thucydides’, in E.J. Bakker (ed.),
Grammar as Interpretation: Greek Literature in its Linguistic Contexts (Leiden 1997)
7–54.
Bakker, E.J., ‘Homeric οὗτος and the Poetics of Deixis’, Classical Philology 94 (1999) 1–19.
Bakker, E.J., ‘Remembering the God’s Arrival’, Arethusa 35 (2002) 63–81.
Beekes, R.S.P., Comparative Indo-European Linguistics (Amsterdam 1995).
Blass, F., ‘Demosthenische Studien. Aorist und Imperfekt’, RhM 44 (1889) 406–430.
Bonnafé, A. (ed.), Hésiode, Théogonie: Texte et traduction (Paris 1993).
Chantraine, P., Grammaire homérique, 2 vols. (Paris 1953).
Chantraine, P., Dictionnaire étymologique de la langue grecque: Histoire des mots (Paris
1968–1980).
Clay, J.S., ‘What the Muses Sang: Theogony 1–115’, GRBS 29 (1988) 323–333.
Clay, J.S., The Politics of Olympus (Princeton 1989).
Clay, J.S., Hesiod’s Cosmos (Cambridge 2003).
Danielewicz, J., Morfologia hymnu antyznego (Poznán 1976).
Denniston, J.D., The Greek Particles, 2nd ed. (Oxford 1954).
Depew, M., ‘Gender, Power, and Poetics in Callimachus’ Book of Hymns’, in M.A. Harder,
R.F. Regtuit, G.C. Wakker (eds), Callimachus II (Leuven 2004) 117–138.
Evelyn-White, H.G. (ed.), Hesiod: Homeric Hymns, Homerica (Cambridge, MA. 1914).
Faulkner, A., ‘Aphrodite’s Aorists: Attributive Sections in the Homeric Hymns’, Glotta 81
(2005) 60–79.
Furley, W.D., Bremer, J.M. (eds), Greek Hymns, 2 vols. (Tübingen 2001).
discourse cohesion in the proem of hesiod’s theogony 209

de Jong, I.J.F., A Narratological Ccommentary on the Odyssey (Cambridge 2001).

Leclerc, M.-C., La parole chez Hésiode (Paris 1993).
Lombardo, S., Hesiod: Works and Days and Theogony (Indianapolis/Cambridge 1993).
Lyons, J., Semantics (Cambridge 1977).
Marg, W. (ed.), Hesiod: Sämtliche Gedichte (Zürich/Stuttgart 1970).
Mazon, P. (ed.), Hésiode (Paris 1928).
Minton, W.W., ‘The Proem-Hymn of Hesiod’s Theogony’, TAPA 101 (1970) 357–377.
Most, G.W. (ed.), Hesiod: Theogony, Works and Days, Testimonia (Cambridge, MA 2006).
Pfeijffer, I.L., First Person Futures in Pindar (Stuttgart 1999).
Pucci, P. (ed.), Inno alle Muse (Esiodo, Teogonia, 1–115), Testo, Introduzione, Traduzione e
Commento (Pisa-Roma 2007).
Rijksbaron, A., ‘Euripides Bacchae 35–36’, Mnemosyne 48 (1995) 198–200.
Rijksbaron, A., The Syntax and Semantics of the Verb in Classical Greek: An Introduction,
3rd ed. (Chicago 2007).
Ruijgh, C.J., Autour de ‘τε épique’: Études sur la syntaxe grecque (Amsterdam 1971).
Sargent, T. (ed.), The Homeric Hymns (New York 1973).
Smith, C.S., ‘Perspective and Point of View: Accounting for Subjectivity’, in H. Hassel-
gård, S. Johansson, B. Behrens, C. Fabricius-Hansen (eds), Information Structure in a
Cross-linguistic Perspective (Amsterdam 2002) 63–79.
Stoddard, K., The Narrative Voice in the Theogony of Hesiod (Leiden 2004).
Thalmann, W.G., Conventions of Form and Thought in Early Greek Epic Poetry (Balti-
more/London 1984).
Verdenius, W.J., ‘Notes on Hesiod’s Theogony’, Mnemosyne 25 (1972) 225–260.
Waltz, P., ‘Note sur la Théogonie’, REG 27 (1914) 229–235.
West, M.L. (ed.), Hesiod: Theogony (Oxford 1966).
West, M.L., ‘An Unrecognized Injunctive Usage in Greek’, Glotta 67 (1989) 135–138.
 chapter 10

On the Syntax and Pragmatics of inquit Formulae in

Plato’s Narrated Dialogues

Abstract

One of the most conspicuous features of Plato’s narrated dialogues1 is the seemingly
arbitrary alternation of the inquit formulae ἔφη/ἔφην and ἦ δ᾽ ὅς/ἦν δ᾽ ἐγώ, accompa-
nied by a stretch of direct speech. In this paper I will try to shed some light on the
syntactic and pragmatic rules that may lie behind this variation.

1 Frequency and Distribution of the inquit Formulae ἔφη/ἔφην and ἦ

δ᾽ ὅς/ἦν δ᾽ ἐγώ

framed ἔφη ἔφην ἦ δ’ ὅς/ἥ2 ἦν δ’ ἐγώ

Phd. 1st person narrative (Phaedo) 244 3 63 8

Smp. 3d person narrative (Aristodemus) 93 11 10 17
embedded in 1st person narrative (Apol- (φάναι: 57)
lodorus)
Euthd. 1st person narrative (Socrates) 201 52 39 66
Prt. 1st person narrative (Socrates) 128 40 8 43
total 666 106 120 134

1 These consist of dialogues with an introductory ‘dramatic’ conversation, plus a narrated

philosophical discussion, so-called ‘framed’ dialogues (Phaedo, Symposium, Euthydemus, Pro-
tagoras), and dialogues with an introduction that is part of the narrative: Parmenides, Char-
mides, Lysis, Respublica. Much has been written on the various forms of the Platonic dialogue,
for instance on the possible philosophical implications of the differences, and much more on
the preliminary question why he wrote dialogues at all. It is not the purpose of my paper to
enter in any detail into the discussion of these questions. For the latter question I may refer,
for instance, to several of the papers collected in Platonic Writings, Platonic Readings (Gris-
wold 1988), and to Kahn’s Plato and the Socratic Dialogue (Kahn 1996), and for the former to
Clay (1992), Johnson (1998) and Morgan (2004). Of the older literature Andrieu (1951) is still
valuable.
2 Diotima in Smp.

© koninklijke brill nv, leiden, 2019 | doi:10.1163/9789004386129_012

on the syntax and pragmatics of inquit formulae 211

(cont.)

narrative throughout ἔφη ἔφην ἦ δ᾽ ὅς/ἥ ἦν δ᾽ ἐγώ

Prm. 3d person narrative (Pythodorus), that 24 – – –

is embedded in another 3d person nar- (φάναι: 43)
rative (Antiphon), which is embedded
in a 1st person narrative (Cephalus)
Chrm. 1st person narrative (Socrates) 88 19 21 60
Ly. 1st person narrative (Socrates) 82 2 13 54
R. 1st person narrative (Socrates) 1064 76 188 601

total 1258 97 222 715

grand total 1924 2033 342 849

Of the third person verb forms ἔφη is clearly much more frequent than ἦ δ᾽ ὅς:
there is a chance of about 7 to 1 that the third person verb of an inquit formula
is ἔφη. For the first person forms, however, the opposite is true: ἦν δ᾽ ἐγώ is far
more frequent than ἔφην. Both these facts will turn out to be relevant when it
comes to analysing their respective functions.

2 ἔφη/ἦ δ᾽ ὅς and ἔφην/ἦν δ᾽ ἐγώ in Context

Above I only presented some raw numbers about the frequency and | distribu- 399
tion of the inquit formulae. But it is of course the way they function in con-
text, i.e. their pragmatic function, which is the interesting issue. In analysing
this function I will also give my view of the semantics of the verb forms used
in the inquit formulae, and of the particle δέ in the ἦ δ᾽ ὅς/ἦν δ᾽ ἐγώ for-
mula.
To appreciate their function in context I have chosen two passages, the brief
opening section from Lysis, and a longer passage from Phaedo.

(1) Ἐπορευόμην μὲν ἐξ Ἀκαδημείας εὐθὺ Λυκείου τὴν ἔξω τείχους ὑπ᾽ αὐτὸ τὸ
τεῖχος· ἐπειδὴ δ᾽ ἐγενόμην κατὰ τὴν πυλίδα ᾗ ἡ Πάνοπος κρήνη, ἐνταῦθα συν-

3 ἔφην: 104; ἔφην ἐγώ: 92; ἐγὼ ἔφην: 7.

 212 chapter 10

έτυχον Ἱπποθάλει … καί με προσιόντα ὁ Ἱπποθάλης ἰδών, Ὦ Σώκρατες, ἔφη,

ποῖ δὴ πορεύῃ και πόθεν;
Ἐξ Ἀκαδημείας, ἦν δ᾽ ἐγώ, πορεύομαι εὐθὺ Λυκείου.
Δεῦρο δή, ἦ δ᾽ ὅς, εὐθύ ἡμῶν. οὐ παραβάλλεις; ἄξιον μεντοι.
Ποῖ, ἔφην ἐγώ, λέγεις, καὶ παρὰ τίνας τοὺς ὑμᾶς;
Δεῦρο, ἔφη, δείξας μοι ἐν τῷ καταντικρὺ τοῦ τείχους περίβολόν τέ τινα καὶ
θύραν ἀνεῳγμένην. διατρίβομεν δέ, ἦ δ᾽ὅς, αὐτόθι ἡμεῖς τε αὐτοὶ καὶ ἄλλοι πάνυ
πολλοὶ και καλοί. (Ly. 203a1–b2)

I was making my way from the Academy straight to the Lyceum, by the
road outside the town wall,—just under the wall; and when I reached the
little gate that leads to the spring of Panops, I chanced there upon Hip-
pothales … Then Hippothales, as he saw me approaching, said: Socrates,
whither away, and whence?
From the Academy, I replied, on my way straight to the Lyceum.
Come over here, he said, straight to us. You will not put in here? But
you may as well.
Where do you mean? I asked; and what is your company?
Here, he said, showing me there, just opposite the wall, a sort of enclo-
sure and a door standing open. We pass our time there, he went on; not
only we ourselves, but others besides,—a great many, and handsome.
(transl. Lamb, Loeb ed.)

How should the alternation of ἔφην and ἦν δ᾽ ἐγώ be explained? I will try to
formulate an answer to this question after I have provided myself with a num-
ber of explanatory tools, which relate to preferences of ἔφη/ἔφην and ἦ δ᾽ ὅς/ἦν
δ᾽ ἐγώ for combining with the participial modifier ὑπολαβών and with certain
400 types of direct speech. |

3 Combinatory Preferences of ἔφη/ἦ δ᾽ ὅς and ἔφην/ἦν δ᾽ ἐγώ

3.1 ἔφη ὑπολαβών and Variants ‘he said, taking up this point, taking over’

(2) Πάνυ γε, ἔφη ὑπολαβὼν ὁ Σιμμίας, (Phd. 64c3)

(3) Ὑπολαβὼν οὖν ὁ Ἀλκιβιάδης, Οὐ καλῶς λέγεις, ἔφη, ὦ Καλλία· (Prt. 336b7)

Examples (2) and (3) represent the 17 cases where ἔφη is modified by ὑπολαβών.
This combination has the meaning ‘he said, taking up this point,’ or ‘reacting to
 on the syntax and pragmatics of inquit formulae 213

this, he said …’4 ‘He’ is never Socrates, nor is ἔφην ‘I said’, i.e. Socrates, modified
by ὑπολαβών:5 Socrates does not take up points made by his interlocutors, he is
rather the one who introduces such points into the discussion. The combina-
tion of ὑπολαβών with ἦ δ᾽ ὅς or ἦν δ᾽ ἐγώ does not occur, which is important for
establishing the pragmatic value of ἦ δ᾽ ὅς, as will soon become clear.
Only ἔφη, then, combines with ὑπολαβών, a so-called coincident aorist par-
ticiple, which specifies the manner in which the action of the main verb is
executed: ‘he said, taking op this point’, or ‘reacting to this, he said …’ From
this I infer that the narrator uses ἔφη when the quoted speaker continues
the line of questioning in which he is involved. And conversely, from the
absence of the combination of ἦ δ᾽ ὅς and ἦν δ᾽ ἐγώ with ὑπολαβών it may
be inferred that these inquit formulae were not suited to continue the dis-
cussion. This, in turn, suggests that they are rather used to close off a certain
part of the discussion. These two inferences served as my hypothesis in look-
ing for further tests to establish a difference between the two inquit formu-
lae. | 401

3.2 ἀληθῆ, ἀληθέστατα (λέγεις), inquit

ἔφη ἔφην ἦ δ᾽ ὅς ἦν δ᾽ ἐγώ BARE

ἀληθῆ (λέγεις)

Phd. 12 1 (Phaedo) – – 3
Prm. 1 (φάναι 2) – – – 26
Smp. – (φάναι 1) 3 (Socrates to – 1 (Socrates to 2
Diotima) Diotima)
Chrm. – – – – 4
Ly. 2 – – – 9
Euthd. 2 – – – 1
Prt. 3 1 (Socrates to – – 1
Protagoras)

4 Rather than ‘answering’ or ‘interrupting.’ Thus, at Phd. 60c8 Cebes reacts to the presence of
Aesopus in Socrates’ words, which causes him to bring up a point connected with this poet.
He in no way answers a question or interrupts Socrates. True, ‘reacting to this’ may also occur
in answers (e.g. at Euthd. 289c9), but this does not entail that ὑπολαβών itself must have the
meaning ‘answering’.
5 With one exception, Ep. 7.349b2, where Socrates picks up the thread of his own argument.
 214 chapter 10

(cont.)

ἔφη ἔφην ἦ δ᾽ ὅς ἦν δ᾽ ἐγώ BARE

ἀληθῆ (λέγεις)

R. 54 – 5 332a6, 395b7, 3 372c4, 430c1 4 334d2, 459c7

444a3, 463e7, 467b7 475b10, 601c14
495b7
TOTAL 74 5 5 4 50
(without φάναι)

It appears form the above table that there is a rather striking preference of
ἀληθῆ, ἀληθέστατα (λέγεις), an agreement formula that makes it possible for
the interlocutor to go on,6 for combining with ἔφη rather than with ἦ δ᾽ ὅς (a
chance of about 15 to 1), if an inquit formula is expressed at all.7 From this I
infer that the pragmatic force of ἔφη is like that of ἀληθῆ (λέγεις): while the
quoted words (ἀληθῆ, ἀληθέστατα) as such enable the interlocutor to go on, the
use of ἔφη signals at another level, viz. that of the narrative, that the speaker
goes along with what his interlocutor had said earlier in the discussion, rather
than breaking the discussion off or switching to another subject.8 Semantically,
402 this use of ἔφη is in accordance | with its being an imperfect,9 for the imperfect

6 That interlocutor is virtually always Socrates, for Socrates himself says ἀληθῆ λέγεις only rarely,
and only to interlocutors who are socially and intellectually at least equal to himself (three
times to Diotima, once to Protagoras, and three times to Glaucon, Plato’s brother).
7 Other agreement formulae show the same characteristics. Thus, in the Republic ὀρθῶς is the
content of ἔφη in 48 and of ἔφην in 2 cases, but of ἦ δ᾽ ὅς in just 4 and of ἦν δ᾽ ἐγώ in 2 cases.
8 Naturally, the first speaker may still change the subject after his interlocutor has said ‘You’re
right.’ Thus, after Thrasymachus’ words ἀληθέστατα λέγεις at R. 336a8, which is the content of
ἔφη, Socrates quotes himself (ἦν δ᾽ ἐγώ) as saying: Εἰἑν, ἦν δ᾽ ἐγώ· ἐπειδὴ δὲ οὐδὲ τοῦτο ἐφάνη
ἡ δικαιοσύνη ὂν οὐδὲ τό δίκαιον, τί ἂν ἄλλο τις αὐτὸ φαίη εἶναι; (‘That is most true,’ he replied.
‘Very well,’ said I, ‘since it has been made clear that this too is not justice and the just, what
else is there that we might say justice to be?’), thereby both closing off this particular part of
the discussion and introducing a new subject. For this function of ἦ δ᾽ ὅς and ἦν δ᾽ ἐγώ see
below.
9 According to Fournier, who obviously did not know Blass’ explanation of such verbs (see fn.
10), in Homer ‘ἔφη/φάτο, imparfaits d’ origine, jouent le rôle d’aoristes. C’est surtout après
discours direct que la formule se rapproche le plus de l’aoriste: le discours fini, quelle utilité
d’ en signaler la durée et le déroulement?’ (1946: 20), and he claims that the same holds true
of Attic ἔφη(ν) (p. 21). Remarkably, however, he considers (p. 46) ἦ ‘après discours direct’ an
imperfect, without further comment. Svensson (1930: 60–66) presents a detailed critique of
on the syntax and pragmatics of inquit formulae 215

of verbs of saying may be used to present a speech act that as such is com-
pleted, as open-ended, having its ultimate completion in the reaction of the
addressee.10 Syntactically, too, ἔφη continues the narrative, since in principle it
has no formal link with the preceding context, i.e. it follows asyndetically after
the preceding utterance, which itself may or may not be accompanied by ἔφη, ἦ
δ᾽ ὅς, or a semantically non-empty verb of saying. Ἔφη does not, therefore, for-
mally indicate that another speaker takes the turn; if this applies, it is entirely
determined by the context.11 And, finally, if a connector is present this is con-
tinuative καί rather than distinctive δέ (cp. fn. 12). I note that the continuative
force of ἔφη in combination with ἀληθῆ (λέγεις) is fully in accordance with what
we found above for its use in the combination ἔφη ὑπολαβών.
Speaking about continuity: I take it that ἦ δ᾽ ὅς does all the things that ἔφη
does not. (1) (syntax) While asyndetic ἔφη has continuative force and does not
formally mark a change of speaker, ἦ δ᾽ ὅς does mark such a change, | by δέ and 403
the overt new subject ὅς.12 I think, in fact, that the narrator, by using ἦ δ᾽ ὅς (and
ἦν δ᾽ ἐγώ), formally signals that the present turn in the discussion is distinct
from the preceding turn, and has therefore passed from speaker A to speaker
B; (2) (semantics) although ἦ and ἦν are usually considered imperfects, they do
not belong to a present stem paradigm, and I am assuming, in fact, that they,
in competition with the imperfects ἔφη and ἔφην, have acquired aorist mean-
ing and denote a complete speech act;13 (3) (pragmatics) by the combination

the aorist view of ἔφη(ν). For the meaning development of φημί (orig. ‘proférer des paroles
magiques ou sacrées’) see Fournier (1946: 8 ff.).
10 A use discussed and explained by Friedrich Blass, in an article on ‘Verba … die ihr Ziel
und Vollendung in dem Thun eines anderen haben’ (Blass 1889: 410). On this Jacobsthal
(see Svensson 1930: 3) rightly observed: ‘mann kan hinzufügen “in der Zustimmung eines
anderen”.’ See further Svensson (1930: 3–5), Rijksbaron (2006: 18–19). Svensson (who does
not discuss Plato) observes on speeches in Xenophon that ‘ἔφη häufig in eine Rede hier
und da hineingeschoben wird, um die Kontinuität derselben hervorzuheben.’
11 As a rule of thumb one may assume that if two ἔφη’s follow each other they are not spoken
by the same speaker. Overt, explicit, subjects of ἔφη are rare; they are present in only 198
cases out of 1924 (= 9,7 %; 173 instances of ἔφη ὁ Χ, of which 65 in the Phaedo; 25 instances
of the anaphoric pronoun ὅς | ἥ, which is always preceded by καί).
12 The general pragmatic function of δέ is to express ‘discreteness’, i.e. it expresses the idea
that a certain piece of information is pragmatically distinct from the information that pre-
cedes (Kroon 1995: 226, on Lat. autem; for Greek cp. Sicking & van Ophuijsen 1993: 11–12). I
am assuming, then, that δέ in ἦ δ᾽ ὅς and ἦν δ᾽ ἐγώ is a meaningful element, without having
full connective force. It is rather an adverb, as in the combination καὶ … δέ. Note, in this
connection, that ἦ δ᾽ ὅς and ἦν δ᾽ ἐγώ can be preceded by coordinating καί, as at R. 392c9
Καὶ ὁ Ἀδείμαντος, Τοῦτο, ἦ δ ὅς, οὐ μανθάνω ὅτι λέγεις.—For details on ὅς see Chantraine
1968–1980: s.v. ὁ.
13 Ameis-Hentze-Cauer, in their commentary on Iliad 1.219, take ἦ in Homer, which occurs
 216 chapter 10

of distinctive δέ and aoristic ἦ and ἦν the pragmatic function of ἦ δ᾽ ὅς and ἦν

δ᾽ ἐγώ is different from that of ἔφη and ἔφην. While the open-ended imperfects
continue a certain line of questioning, aoristic ἦ δ᾽ ὅς and ἦν δ᾽ ἐγώ indicate that
the current (part of the) discussion either comes to an end (with answers)14 or
takes a new turn (with questions).15

404 Fortunately my claims do not just depend on the behaviour of ἀληθῆ | (λέγεις)
and ἔφη. There are several other formal indications to support them. Consider
the following tables.

3.3 Εἰἑν, inquit

Εἰἑν, ἔφη Εἰἑν, φάναι Εἰἑν, ἔφην Εἰἑν, ἦ δ᾽ ὅς Εἰἑν, ἦ δ᾽ ἐγώ BARE Εἰἑν

Phd. 3 – – 1 – 2
Prm. – 1 – – – 1
Smp. – 1 – – 2 –
Chrm. – – – – 2 1
Ly. – – – – 6 –
Euthd. – – – – 5 –
Prt. – – – – 1 3
R. I16 — – – 10 6
TOTAL 4 2 – 1 26 13

always after direct speech, as an ‘Abschluß- und Übergangsformel’. I note for the record
that in a scholium on that line Zenodotus paraphrases ἦ as ὦς εἰπών, and that Elias, in
Aristot. Cat. comm. 146, writes ‘παρὰ Πλάτωνι « ἦ δ᾽ ος » ἀντὶ τοῦ « εἶπεν οὗτος »’.
14 One may compare the use of Ὡμολόγησεν at Euthd. 277b2, with which the narrator closes
off a discussion between Clinias and Euthydemus. Earlier in this discussion he had used
open-ended ὡμολόγει three times: at 277a3, a4, and a6.
15 One wonders, of course, how the instances of ἀληθῆ/ἀληθέστατα with ἦ δ᾽ ὅς and ἦν δ᾽ ἐγώ
might be explained. I confine myself here to suggesting an explanation for two cases. At R.
332a6 the line after Ἀληθῆ, ἦ δ᾽ ὅς does not continue the line of questioning but introduces
a new point: Ἄλλο δή τι ἢ τὸ τοιοῦτον … λέγει ὁ Σιμωνίδης (‘True,’ said he. ‘It is, then some-
thing other than this that Simonides must mean.’). Likewise, at R. 372c5 (Ἀληθῆ, ἦν δ᾽ ἐγὼ,
λέγεις. ἐπελαθόμην ὅτι …), the line after ἦν δ᾽ ἐγώ (λέγεις) does not continue the question-
ing but provides a comment on the way of questioning. In other words: the questioning
itself comes temporarily to a stand-still.
16 337c2, subject Thrasymachus.
on the syntax and pragmatics of inquit formulae 217

It appears that the transition formula εἰἑν,17 which conveys that the speaker
considers some point settled and ‘is ready to proceed to the next point or the
next step’ (Barrett on E. Hipp. 297), has a very strong preference for combining
with ἦν δ᾽ ἐγώ (I confine myself here to this inquit formula).18 This can be read-
ily understood if the latter has the pragmatic value discussed above: after the
speaker, with εἰἑν, has put an end to the current discussion (‘Right’, ‘Okay’, ‘Well,
then’; ‘Eh bien,’ ‘Soit’ (Chantraine)) and thus has cleared the way for himself to
take his turn and to proceed to a new point, the speaker-as-narrator rounds
off, too, by means | of ἦν δ᾽ ἐγώ: the turn shifts, and so does the subject of the 405
discussion, for non-continuative ἦν δέ clears the way for introducing a new sub-
ject into the discussion. Naturally, it belongs to the leader of the discussion
(Socrates, of course) to use such a formula as εἶἑν. In fact, it is never said by
one of his interlocutors. A typical example is

(4) Ὁ δὴ Θρασύμαχος ὡμολόγησε μὲν πάντα ταῦτα, … ἐπειδὴ δὲ οὖν διωμολογη-

σάμεθα τὴν δικαιοσύνην … Εἶεν, ἦν δ᾽ ἐγω, τοῦτο μὲν ἡμῖν οὕτω κείσθω, ἔφαμεν
δὲ δὴ καὶ ἰσχυρὸν εἶναι τὴν ἀδικίαν … (R. 350c10 ff.)

Thrasymachus made all these admissions … But when we did reach our
conclusion that justice … ‘Good,’ said I, ‘let this be taken as established.
But we were also affirming that injustice is a strong and potent thing …’
(transl. Shorey)

Note the completed states of affairs expressed by the aorists ὡμολόγησε and διω-
μολογησάμεθα; note also that the rounding-off function of εἰἑν is here reinforced
by τοῦτο … οὕτω κείσθω.

17 For the aspirated ε cp. Chantraine 1968–1980: s.v.

18 As for the combination with ἔφη, this would seem to express incomplete, partial, rounding
off at Phd. 105e8 Εἶἑν, ἔφη· τοῦτο μὲν δὴ ἀποδεδεῖχθαι φῶμεν; ἢ πῶς δοκεῖ; subject Socrates;
note the presence of μέν after τοῦτο: there is more to come.—115b1 Ταῦτα δὴ εἰπόντος αὐτοὗ
ὁ Κρίτων, Εἶἑν, ἔφη, ὦ Σώκρατες· τί δὲ τούτοις ἢ ἐμοὶ ἐπιστέλλεις ἢ περὶ …;—this falls outside
the discussion proper, it does not react to a participant, and there is no new turn. Likewise
at 117a8 ἰδὼν δὲ ὁ Σωκράτης τὸν ἄνθρωπον, Εἶἑν, ἔφη, ὦ βέλτιστε, σὺ γὰρ τούτων ἐπιστήμων, τί
χρὴ ποιεῖν;—outside the discussion, does not react to a participant, no new turn.
 218 chapter 10

3.4 τοίνυν, inquit

τοίνυν, ἔφη τοίνυν, ἔφην τοίνυν, ἦ δ’ ὅς τοίνυν, ἦνδ’ ἐγώ τοίνυν

Phd. 9 1 5 – 5
Prm. – – – –
Smp. 2 – – – 5
Chrm. 4 2 2 7 2
Ly. – – – 1 4
Euthd. – 2 1 2 1
Prt. 1 2 – 1 4
R. 11 3 504c9, 564b4, 584b1 – 53 44
TOTAL 27 10 8 64 65

Note again the preponderance of ἦν δ᾽ ἐγώ. As a matter of fact, τοίνυν, ἦν δ᾽ ἐγώ,

406 too, is predominantly used to express the transition from some | point in the
discussion that has been settled to a new point. For this function of τοίνυν cp.
Denniston (1954: 574–575).19 A typical example is

(5) Σύμψηφός σοί εἰμι, ἔφη, τούτου τοῦ νόμου, καί μοι ἀρέσκει. Οὗτος μὲν τοίνυν,
ἦν δ᾽ ἐγώ, εἷς ἂν εἴη … (R. 380c6)

Observe that here, too, the speaker text is in accordance with the rounding-off
function of the narrator text.

3.5 Results of the Above Investigation

The results of the above investigation can be summarized as follows: by ἔφη and
ἔφην the narrator presents the quoted utterance as an argumentative move20
that continues the line of questioning in which the speaker and his interlocutor
were involved, while by ἦ δ᾽ ὅς and ἦν δ᾽ ἐγώ the narrator presents the quoted
utterance as an argumentative move that disrupts the line of questioning of
speaker and interlocutor, either as a closure or to mark a change of speaker
and subject matter.

19 Transitional ἆρ᾽ οὖν behaves in the same way: in R. there is one instance of ἆρ᾽ οὖν (…),
ἔφην, as against 17 of ἆρ᾽ οὖν (…), ἦν δ᾽ ἐγώ (and 77 of bare ἆρ᾽ οὖν; there are no instances
of ἆρ᾽ οὖν, ἔφη or ἆρ᾽ οὖν, ἦ δ᾽ ὅς).
20 For the use of the concept ‘move’ in discourse analysis see Kroon (1995: 65–66).
on the syntax and pragmatics of inquit formulae 219

4 The Opening Section of Lysis

Back now to the opening section of Lysis, example (1). ἔφη, the first verb of
speaking, is the normal verb form to open the narrative of a discussion. In fact,
ἦ δ᾽ ὅς is never used in this way, another indication of the pragmatic value of
this formula: it contrasts with what precedes, but here there is nothing with
what it might contrast. By its imperfective aspectual value ἔφη suggests that
there is more to come. The reaction of Socrates-the-participant is described by
Socrates-the-narrator as ἦν δ᾽ ἐγώ, conveying that he wants his answer to be
definitive (effect of closure). For Hippothales’ reaction the narrator uses ἦ δ᾽
ὅς, conveying that the reaction does not, indeed, continue the words spoken
by Socrates, e.g. by asking him about the purpose of his walk to the Lyceum,
but constitutes a new move: he ignores Socrates’ εὐθὺ Λυκείου, and proposes
another destination. This | move is acknowledged and picked up by Socrates, 407
as Socrates the narrator indicates by using continuative ἔφην. The fact that the
narrator also uses continuative ἔφη for Hippothales’ answer conveys that Δεῦρο
is not meant to close the conversation, it is not, so to speak, the final word on
the destination. Remarkably, ἔφη is then followed by the other formula, spo-
ken by the same speaker. Note that this formula is preceded by διατρίβομεν δέ,
where δέ conveys that this piece of information is pragmatically distinct from
the information that precedes (cp. fn. 12): the narrator switches from the phys-
ical aspect of the destination as pointed at by Hippothales, to what is going on
at the destination. I submit that the δέ in the speaker text is reflected by the δέ
of ἦ δ᾽ ὅς: the speech act, too, is distinct from the previous one. In an English
translation this nuance could perhaps be achieved by rendering ἦ δ᾽ ὅς as ‘he
added’ rather than as ‘he went on’; the latter would rather correspond to ἔφη.

5 Application to an Extended Passage: Plato, Phaedo 60d8–62a1

(6) 60d8 (subj. Socrates) Λέγε τοίνυν, ἔφη, αὐτῷ, ὦ Κέβης, τἀληθῆ, ὅτι οὐκ …
61b8 ταῦτα οὖν, ὦ Κέβης, Εὐήνῳ φράζε, και ἐρρῶσθαι καί, ἂν σωφρονῇ, ἐμὲ
διώκειν ὡς τάχιστα. ἄπειμι δέ, ὡς ἔοικε, τήμερον κελεύουσι γὰρ Ἀθηναῖοι.
61c1 Καὶ ὁ Σιμμίας, Οἷον παρακελεύῃ, ἔφη, τοῦτο, ὦ Σώκρατες, Εὐήνῳ. πολλὰ
γὰρ ἤδη ἐντετύχηκα τῷ ἀνδρί· σχεδὸν οὖν ἐξ ὧν ἐγώ ᾔσθημαι οὐδ᾽ ὁπωστι-
οῦν σοι ἑκὼν εἶναι πείσεται.
61c6 Τί δέ; ἦ δ’ ὅς, οὐ φιλόσοφος Εὔηνος;
61c7 Ἔμοιγε δοκεῖ, ἔφη ὁ Σιμμίας.
61c8 Ἐθελήσει τοίνυν καὶ Εὔηνος καὶ πᾶς ὅτῳ ἀξίως τούτου τοῦ πράγματος
μέτεστιν. οὐ μέντοι ἴσως βιάσεται αὑτόν· οὐ γάρ φασι θεμιτὸν εἷναι. Καὶ
 220 chapter 10

ἅμα λέγων ταῦτα καθῆκε τα σκέλη ἐπὶ τὴν γῆν, καὶ καθεζόμενος οὕτως
ἤδη τὰ λοιπὰ διελέγετο.

c8 τοίνυν ἔφη καὶ δ

61d3 Ἤρετο οὖν αὐτον ὁ Κέβης· Πῶς τοῦτο λέγεις, ὦ Σωκρατες, τὸ μὴ θεμι-
τὸν εἶναι ἑαυτὸν βιάζεσθαι, ἐθέλειν δ᾽ ἂν τῷ ἀποθνῄσκοντι τὸν φιλόσοφον
408 ἕπεσθαι; |
61d6 (no inquit) Τί δέ, ὦ Κέβης; οὐκ ἀκηκόατε σύ τε καὶ Σιμμίας περὶ τῶν
τοιούτων Φιλολάῳ συγγεγονότες;
61d8 (no inquit) Οὐδέν γε σαφές, ὦ Σώκρατες.
61d9 (no inquit) Ἀλλὰ μὴν καὶ ἐγὼ ἐξ ἀκοῆς περὶ αὐτῶν λέγω·
61e5 (no inquit) Κατὰ τί δὴ οὖν ποτε οὔ φασι θεμιτὸν εἶναι αὐτὸν ἑαυτὸν ἀπο-
κτεινύναι, ὦ Σώκρατες; …
62a1 Ἀλλὰ προθυμεῖσθαι χρή, ἔφη· τάχα γὰρ ἂν καὶ ἀκούσαις.
(Pl. Phd. 60d8–62a1)

60d8 Then tell him, Cebes, said he, the truth, that …
61b8 So tell Evenus that, Cebes, and bid him farewell, and tell him, if he’s
wise, to come after me as quickly as he can. I, it seems, am going to-
day; for that is the order of the Athenians.
61b2 And Simmias said, What a message that is, Socrates, for Evenus! I
have met him often, and from what I have seen of him, I should say
that he will not take your advice in the least if he can help it.
61c6 Why so? said he. Is not Evenus a philosopher?
61c7 I think so, said Simmias.
61c8 Then Evenus will take my advice, and so will every man who has
any worthy interest in philosophy. Perhaps, however, he will not take
his own life, for they say that it is not permitted. And as he spoke
he put his feet down on the ground and remained sitting this way
through the rest of the conversation.
61d3 Then Cebes asked him, What do you mean by this, Socrates, that
it is not permitted to take one’s life, but that the philosopher would
desire to follow after the dying?
61d6 (no inquit) How is this, Cebes? Have you and Simmias, who are
pupils of Philolaus, not heard about such things?
61d8 (no inquit) Nothing definite, Socrates.
61d9 (no inquit) I myself speak of them only from hearsay.
61e5 (no inquit) Why in the world do they say that it is not permitted to
kill oneself, Socrates? …
on the syntax and pragmatics of inquit formulae 221

62a1 You must have courage, said he, and perhaps you might hear some-
thing.
(transl. Fowler, Loeb ed.)

The use of ἔφη and ἦ δ᾽ ὅς in the larger passage from Phaedo can be accounted
for in the same way. Observe that although Socrates speaks to Cebes it is Sim-
mias who reacts. ἔφη indicates that his reaction stays with Socrates’ last words
and with Evenus. Then Socrates, provoked by Simmias’ ideas about Evenus,
gives another turn to the discussion, as appears, first, from the introductory
exclamation Τί δέ, in his own text (which is | ‘How is that?’ rather than ‘Why 409
so?’), and, second, from Phaedo’s use of ἦ δ᾽ ὅς. (Recall that it is Phaedo who
is the narrator.) The discussion is no longer about Socrates’ advice to Evenus,
but about the latter’s status as a philosopher. Continuative ἔφη indicates that
Simmias acknowledges this new turn, and goes along.
61c8 presents an interesting problem. The text printed here is that of the
new OCT. It appears from the apparatus that part of the MSS add ἔφη after τοί-
νυν (not mentioned, incidentally, by Burnet). This reading may well be right, for
the omission of ἔφη results first of all in a certain imbalance, since from 60b3
onward all utterances were accompanied by inquit formulae, 6 times ἔφη and
once ἦ δ᾽ ὅς. But more importantly, without ἔφη the words ἐθελήσει etc. might
still be spoken by Simmias, for in scriptio continua, as practised in Plato’s time
and for a long time after him, there would be nothing to suggest that they are
not spoken by Simmias.21
The next passage begins with ἤρετο. Why this, rather than an inquit formula?
Probably because so far there had been no ‘philosophical’ questions. In fact,
the only question so far, at 61c5, was about Evenus as a person. ἤρετο must have
been chosen to signal explicitly that here the philosophical part of the dialogue
begins.
Note that in the next passage there are all of a sudden four questions and
answers, or two adjacency pairs, as they are called in conversation analysis,
without inquit formulae. This can be explained as being due to their deviant
pragmatic status. As a matter of fact, not only does Socrates not answer Cebes’
question about the permissibility of suicide, he completely ignores it, and
instead starts asking about Cebes’ and Simmias’ philosophical education with
Philolaus. The two adjacency pairs form a block that is outside the discussion

21 In fact, ἔφη may elsewhere, too, have this punctuation-like function. Being a post-positive
element it signals that there is a light syntactic boundary before the word preceding ἔφη.
For details cp. Dik (1995: 31–37) and Scheppers (2011: 8, 199, and the Greek Index s.v. ἔφη).
For similar phenomena in medieval romances see Brandsma (1998).
 222 chapter 10

proper, and this, I suggest, is why they do not have inquit formulae. Observe
410 that as soon as the philosophical discussion is resumed we find ἔφη, 62a1. |

Works cited

Andrieu, J., Le dialogue antique: structure et présentation (Paris 1954).

Blass, F., ‘Demosthenische Studien. Aorist und Imperfekt’, RhM 44 (1889) 406–430.
Brandsma, F., ‘Knights’ Talk: Direct Discourse in Arthurian Romance’, Neophilologus 82
(1998) 513–525.
Chantraine, P., Dictionnaire étymologique de la langue grecque: Histoire des mots (Paris
1968–1980).
Clay, D., ‘Plato’s First Words’, in F.M. Dunn, Th. Cole (eds), Beginnings in Classical Liter-
ature (Cambridge 1992 (= Yale Class. Studies, vol. 29)) 113–129.
Dik, H.J.M., Word Order in Ancient Greek: A Pragmatic Account of Word Order Variation
in Herodotus (Amsterdam 1995).
Fournier, H., Les verbes “dire” en grec ancien (Paris 1946).
Griswold, Ch.L. (ed.), Platonic Writings, Platonic Readings (University Park, PA 1988).
Johnson, W.A. ‘Dramatic Frame and Philosophic Idea in Plato’, AJP 119 (1998) 577–598.
Kahn, Ch.H., Plato and the Socratic Dialogue: The Philosophical Use of a Literary Form
(Cambridge 1996).
Kroon, C.H.M., Discourse Particles in Latin: A study of nam, enim, autem, vero and at
(Amsterdam 1995).
Morgan, K., ‘Plato’, in I.J.F. de Jong, R. Nünlist, A. Bowie (eds), Narrators, Narratees and
Narratives in Ancient Greek Literature (Leiden/Boston 2004) 359–376.
Rijksbaron, A., The Syntax and Semantics of the Verb in Classical Greek, 3rd ed. (Chicago
2006).
Scheppers, F., The Colon Hypothesis: Word Order, Discourse Segmentation and Discourse
Coherence in Ancient Greek (Brussel 2011).
Sicking, C.M.J., Ophuijsen, J.M. van, Two Studies in Attic Particle Usage: Lysias and Plato
(Leiden 1993).
Svensson, A., Zum Gebrauch der erzählenden Tempora im Griechischen (Lund 1930).
 chapter 11

Sur quelques différences entre οὗτος ὁ + substantif,

οὗτος δὲ ὁ + substantif, ὁ δὲ + substantif + οὗτος chez
Hérodote

Résumé

Chez Hérodote, le pronom οὗτος, dans son emploi de déterminant dans un groupe
nominal défini, se rencontre essentiellement dans les trois types de syntagme suivants:
(1) οὗτος ὁ Κροῖσος βαρβάρων πρῶτος τῶν ἡμεῖς ἴδμεν τοὺς μὲν κατεστρέψατο Ἑλλήνων…
(1.6.2)
(2) οὗτος δὲ ὁ Γύγης πρῶτος βαρβάρων τῶν ἡμεῖς ἴδμεν ἐς Δελφοὺς ἀνέθηκε ἀναθήματα
μετὰ Μἱδην τὸν Γορδἱεω… (1.14.2)
(3) οἱ δὲ Ἀλευάδαι οὗτοι ἦσαν Θεσσαλἱης βασιλέες. (7.62)
Dans cet article l’auteur se propose de démontrer que cette variation n’est pas arbi-
traire, mais qu’elle exprime des différences sémantico-pragmatiques spécifiques. Ces
différences portent sur la ‘saillance’ du (référent du) nom aussi bien que des informa-
tions données sur ce nom. Dans les trois types de syntagme, on a affaire à un Topique
Repris. Dans le type (1), il s’agit d’un groupe nominal avec οὗτος prépositif, qui se trouve
dans une phrase qui n’est pas liée formellement au contexte (phrase asyndétique).
Dans de tels cas, tant le Topique que le contenu de la phrase sont marqués comme
saillants. Ces groupes nominaux avec οὗτος prépositif peuvent être considérés comme
des expressions d’‘anaphore forte’. Souvent, le Topique et ses actions ont une impor-
tance primordiale pour l’histoire. Dans le type (2) en revanche, qui contient un GN
avec οὗτος prépositif (‘anaphore forte’) et δέ connectif, la situation est différente. Tandis
que le Topique est marqué comme saillant, la particule δέ indique que les informations
données sur ce Topique sont des informations additionnelles, et donc non saillantes.
Dans (3) enfin, on trouve οὗτος postpositif (‘anaphore faible’) et δέ connectif. Dans ces
cas, tant le Topique que le contenu de la phrase sont marqués comme non saillants.

Abstract

In the Histories of Herodotus the pronominal determiner οὗτος occurs in three differ-
ent constructions involving a definite noun phrase:
(1) οὗτος ὁ Κροῖσος βαρβάρων πρῶτος τῶν ἡμεῖς ἴδμεν τοὺς μὲν κατεστρέψατο Ἑλλήνων…
(1.6.2)

© koninklijke brill nv, leiden, 2019 | doi:10.1163/9789004386129_013

 224 chapter 11

(2) οὗτος δὲ ὁ Γύγης πρῶτος βαρβάρων τῶν ἡμεῖς ἴδμεν ἐς Δελφοὺς ἀνέθηκε ἀναθήματα
μετὰ Μἱδην τὸν Γορδίεω… (1.14.2)
(3) οἱ δὲ Ἀλευάδαι οὗτοι ἦσαν Θεσσαλἱης βασιλέες. (7.62)
In this paper it is argued that this variation is not arbitrary, but expresses clear seman-
tic-pragmatic differences. These differences relate to the saliency both of the (referent
of the) noun and of the information given about that noun.
In all three cases we are dealing with Resumed Topics. In example (1) we find a noun
phrase with prepositive οὗτος in a sentence which is not formally linked to the previ-
ous sentence; there is, thus, asyndeton. In such cases both the Topic and the contents
120 of the | sentence in which the Topic occurs are marked as salient. Noun phrases with
prepositive οὗτος may be considered instances of ‘strong anaphora’. Often, the Topic
and his activities are of prime importance for the story. In (2), on the other hand, where
we have an NP with prepositive οὗτος (‘strong anaphora’) and connective δέ, the situa-
tion is different. While the Topic is marked as salient, the information about that Topic
is rather, by δέ, characterized as additional information, and thus as non-salient. In (3),
finally, we find postpositive οὗτος (‘weak anaphora’) and connective δέ. In these cases
both the Topic and the contents of the sentence are marked as non-salient.

Chez Hérodote, le pronom démonstratif-anaphorique οὗτος se rencontre, dans

son emploi comme déterminant dans un groupe comprenant par ailleurs
article et substantif, essentiellement1 dans trois syntagmes différents: ou bien
οὗτος précède le groupe article – substantif, dans une phrase qui est liée asyn-
détiquement au contexte (οὗτος ὁ + substantif), ou bien le pronom précède un
tel groupe, dans une phrase qui est liée par δέ au contexte (οὗτος δὲ ὁ + substan-
tif), ou bien il suit le groupe article – substantif, dans une phrase qui est, elle
aussi, liée par δέ au contexte (ὁ δὲ + substantif + οὗτος)2.
Dans cet exposé je voudrais démontrer que cette variation n’est pas arbi-
traire mais exprime des différences sémantico-pragmatiques bien spécifiques.
Mis à part quelques observations générales sur la syntaxe de ces groupes ana-
phoriques, je me limiterai à leur emploi en début absolu de phrase; je laisserai

1 Voir Powell. Je fais remarquer toutefois que la classification de Powell, bien qu’elle soit assez
utile en tant qu’ elle se fonde sur la position de οὗτος, reste très globale et ne donne pas
d’ informations supplémentaires, p. ex. sur la position du groupe à οὗτος dans la phrase, ou
(dans la rubrique ‘prec(eding) n(oun)’ (321 exemples)) sur la nature de la particule qui se
trouve dans les syntagmes de ce type. Ainsi, cette rubrique contient des exemples avec δέ
aussi bien qu’ avec δή, ὦν, etc. En outre, les exemples sont parfois mal placés: ainsi 1.6.2 doit
être rangé sous A I 2 (‘w. asyndeton’), et non pas sous A I 1.
2 En plus, il y a deux exemples du groupe ὁ + substantif + οὗτος; voir note 20.
différences entre ουτοσ ο / ουτοσ δε ο / ο δε ουτοσ 225

donc de côté leur emploi à l’intérieur d’une phrase3. Deuxième remarque:

je passerai également sous silence les syntagmes où l’ on trouve des (combi-
naisons de) particules comme ὦν, δή, δὴ ὦν etc. à côté de οὗτος4. Troisième
remarque: l’analyse qui suit est fondée sur l’emploi de οὗτος dans les livres 1
et 7 d’Hérodote, et seulement dans les parties narratives, les discours directs
n’étant pas pris en compte5. Les résultats ont donc | forcément un caractère 121
provisoire et demandent à être confirmés par une analyse plus approfondie.
Les grandes grammaires gardent le silence sur les différences éventuelles
entre les groupes contenant οὗτος. Plus généralement, on constate qu’ elles
concentrent leur attention sur l’emploi déictique des pronoms démonstratifs,
et ne traitent guère leur emploi anaphorique. Ainsi, Kühner-Gerth consacrent,
aux pages 1, 641ss., six paragraphes à l’emploi déictique de ὅδε, οὗτος et ἐκεῖνος,
tandis que l’emploi anaphorique est traité dans deux paragraphes seulement. À
part les remarques très sommaires des grammaires, il n’ y a, à ma connaissance,
qu’une étude qui analyse l’emploi anaphorique d’ une façon plus détaillée : la
petite monographie de l’helléniste suédois Palm (Palm 1960). Toutefois, dans
cette monographie l’emploi asyndétique de οὗτος est ignoré.
Comme point de départ de mon analyse je prends les exemples suivants :

(1) οὗτος ὁ Κροῖσος βαρβάρων πρῶτος τῶν ἡμεῖς ἴδμεν τοὺς μὲν κατεστρέψατο
Ἕλλήνων… (1.6.2)

Ce Crésus, le premier des Barbares à notre connaissance, asservit certains

Grecs… (traduction Legrand, comme dans la suite)

(2) οὗτος δὲ ὁ Γύγης πρῶτος βαρβάρων τῶν ἡμεῖς ἴδμεν ἐς Δελφοὺς ἀνέθηκε ἀνα-
θήματα μετὰ Μίδην τὸν Γορδίεω… (1.14.2)

3 Comme dans παραλαβὼν δὲ τοῦτο τὸ ἔπος… (1.126.4), ἦσαν δὲ οἵδε Ἑλλήνων οἱ ὑπομένοντες τὸν
Πέρσην ἐν τούτῳ τῷ χώρῳ (7.202), ou, avec οὗτος postpositif, παρεὼν δὲ καὶ μεμφόμενος τὴν γνώ-
μην ταύτην Κροῖσος… (1.207.1)
4 Comme dans οὗτος δὴ ὦν ὁ Κανδαύλης ἠράσθη… (1.8.1), τούτων ὦν τῶν ἀνδρῶν Λίχης ἀνεῦρε…
(1.68.1), ou, avec οὗτος postpositif, πρὸς ὦν τὴν ὄψιν ταύτην… (1.38.2)
5 Notre corpus est donc le suivant :
– οὖτος ὁ + substantif (22 exemples): 1.6.2, 13.2, 24.1, 50.3, 65.1, 60.4, 70.2, 96.2, 144.3, 162.2,
183.3, 187.3 ; 7.3.2, 27.1, 59.3, 87, 145.1, 142.3, 143.1, 190, 225.3;
– οὗτος δὲ ὁ + substantif (15 exemples): 1.14.2, 31.5, 62.1, 142.4, 146.3, 149.2, 188.2; 7.70.2, 80,
89.2, 109.2, 117.2, 144.2, 155.2, 198.1 ;
– ὁ δὲ + substantif + οὗτος (22 exemples): 1.66.4, 74.2, 143.3, 157.3, 160.4, 163.1, 176.3, 184, 201;
7.6.2, 59.2, 72.1, 83.1, 91, 109.2, 111.2, 113.2, 153.1, 164.1, 180, 186.1, 213.3. Pour 1.85.4, voir note 16.
 226 chapter 11

Ce Gygès est le premier Barbare à notre connaissance qui ait consacré des
offrandes à Delphes, après Midas fils de Gordias…

(3) οἱ δὲ Ἀλευάδαι οὗτοι ἦσαν Θεσσαλίης βασιλέες (7.6.2)

Ces Aleuades régnaient en Thessalie

À en croire Kühner-Gerth, il n’y aurait pas de différence entre (1) et (2), puis-
qu’ils écrivent, 2, 343 γ: ‘Sehr häufig aber (wird das Asyndeton angewendet)
auch bei einem Demonstrative, das auf einen vorhergehenden Begriff zurü-
ckweist, wo keine Folgerung, sondern bloss eine Fortsetzung der Rede ausge-
drückt wird, also δέ stehen könnte’. Dans la suite, j’ espère démontrer que cette
opinion est partiellement fausse, mais d’abord je voudrais signaler quelques
différences d’ordre syntaxique entre les groupes οὗτος ὁ + substantif et ὁ + sub-
stantif + οὗτος comme tels, qui nous donnent des indications précieuses sur les
différences pragmatiques.
Comme le montrent les exemples (4) et (5), le groupe οὗτος ὁ + substantif
peut être précédé de l’adverbe additif καί, qui indique que le constituant qui
suit est le Focus de la phrase:

(4) ὡς δὲ τῷ Κύρῳ καὶ τοῦτο τὸ ἔθνος κατέργαστο, … (1.201)

122 Quand Cyrus eut réduit aussi cette nation, … |

(5) παραπλέων δὲ καὶ ταύτην τὴν χώρην… (7.123.2)

Longeant à son tour la côte de ce pays…

Par contre, l’on ne trouve pas de syntagmes comme (6), où καί précéderait οὗτος
dans un groupe ὁ + substantif + οὗτος:

(6) * τὸ ἔθνος καὶ τοῦτο…

* τὴν χώρην καὶ ταύτην…

Du fait que οὗτος postpositif ne peut pas être modifié par un adverbe intensif,
on peut déduire que dans ces syntagmes οὗτος a une fonction anaphorique que
l’ on pourrait appeler ‘faible’. En revanche, οὗτος prépositif exprime l’ anaphore
‘forte’. Ce que j’entends par ces termes deviendra plus clair dans ce qui suit6.

6 L’ on trouve, bien sûr, des syntagmes comme… ποταμὸν διαβάς, ἐπ᾽ οὗ καὶ ὁ κόλπος οὗτος τὴν
différences entre ουτοσ ο / ουτοσ δε ο / ο δε ουτοσ 227

(7) et (8) illustrent le même phénomène. Dans (7), la particule intensive δή

indique que ταύτην est, ici encore, le Focus de la phrase; l’ on pourrait appeler
ce type de Focus ‘Focus identifiant’: ‘précisément ce sentier’, ‘ce sentier et nul
autre’7 :

(7) κατὰ ταύτην δὴ τὴν ἀτραπὸν καὶ οὕτως ἔχουσαν… (7.217.1)

par ce sentier, tel que nous venons de le dire…

Or, le fait qu’il n’y a pas un syntagme correspondant avec οὗτος postpositif,
donc:

(8) * κατὰ τὴν ἀτραπὸν ταύτην δή…

m’amène une fois de plus à penser que οὗτος postpositif a une fonction ana-
phorique faible, tandis que οὗτος prépositif a plutôt une fonction forte8.
Pour revenir maintenant aux exemples (1)–(3), je vais présenter tout d’ abord
ces phrases dans leur contexte. Premièrement: exemple (1), οὗτος ὁ + substan-
tif. | 123

1 οὗτος ὁ + substantif

(9) Κροῖσος ἦν Λυδὸς μὲν γένος, παῖς δὲ Ἀλυάττεω, τύραννος δὲ ἐθνέων τῶν ἐντὸς
Ἅλυος ποταμοῦ, … οὗτος ὁ Kροῖσος βαρβάρων πρῶτος τῶν ἡμεῖς ἴδμεν τοὺς μὲν
κατεστρέψατο Ἑλλήνων… (1.6.1)

Crésus était de race lydienne, fils d’Alyatte, souverain des peuples habi-
tant en deçà du fleuve Halys… Ce Crésus, le premier des Barbares à notre
connaissance, asservit certains Grecs…

ἐπωνυμίην ἔχει… (7.58.3) (‘ayant franchi ce fleuve dont le golfe Mélas tient son nom…’), mais
dans ce cas-là c’ est plutôt le groupe article + substantif qui est le Focus (‘ce golfe s’appelle
également Mélas’, tout comme le fleuve). Dans (4) et (5), en revanche, le Focus est sur τοῦτο
et ταύτην : ‘ce peuple aussi’, tout comme d’ autres peuples, ‘cepays aussi’, tout comme d’autres
pays. Dans le premier cas, l’ entité modifiée par καί est mise en rapport avec des entités d’une
autre classe, dans le second cas, elle est mise en rapport avec des entités de la même classe.
7 Pour les différents types de Focus cf. Dik (1989 : 282ss.).
8 À noter que, dans un exemple comme (ἀπικομένους δὲ τοὺς Φοίνικας) ἐς δὴ τὸ Ἄργος τοῦτο
(1.1.2), le domaine syntaxique de δή est le groupe entier τὸ Ἄργος τοῦτο, et non pas le seul
τοῦτο.
 228 chapter 11

Le groupe οὗτος ὁ Κροῖσος vient immédiatement après l’ introduction de Cré-

sus comme personnage dans le récit, dans une phrase où il est pourvu d’ un
certain nombre de traits personnels. Après avoir été introduit comme Topique
Nouveau9, Crésus est expressément repris comme Topique. La reprise de Cré-
sus par un groupe pronominal avec οὗτος préposé et répétition du nom propre
met un accent fort sur son identité (anaphore forte): ‘Ce Crésus et nul autre’10.
Cette mise en relief de son identité s’explique bien, car un tel participant ainsi
que ses actions ont une importance primordiale pour l’ histoire.
L’absence de δέ, à son tour, signale qu’Hérodote commence en quelque
sorte à nouveau, ce qui s’explique très bien car ce n’est qu’ avec la phrase qui
commence par οὗτος ὁ Κροῖσος qu’Hérodote commence son discours narratif
à proprement parler, et que le personnage ‘statique’ devient un participant
‘dynamique’. Il me semble, en effet, que, contrairement à ce que suggèrent
Kühner-Gerth (voir supra), on ne peut pas ajouter δέ, puisqu’ il n’y a pas de
vraie ‘Fortsetzung’.
Cette réintroduction emphatique en phrase asyndétique, comme partici-
pant, d’un personnage qui vient d’être introduit se trouve régulièrement chez
Hérodote. En voici deux autres exemples:

(10) ἐν τῷ δήμῳ τῷ Παιανιέï ἦν γυνή, τῇ οὔνομα ἦν Φύη, μέγεθος… ταύτην τὴν

γυναῖκα σκευάσαντες… (1.60.4)

Dans le dème de Paiania, il y avait une femme nommée Phyè, d’ une

124 taille… Ils revêtirent cette femme… |

9 Pour les différentes façons dont un Topique Nouveau peut être introduit voir Dik (1989:
268ss.).
10 Cf. Grimes (1975: 92ss.), qui, lui aussi, classe les différents types d’anaphore sur une échelle
qui va de l’ anaphore plus forte (plus explicite) jusqu’à l’anaphore plus faible (moins expli-
cite). Il se sert de cette échelle pour définir la notion de ‘champ d’identification’ (‘iden-
tification span’): ‘… a series of identifications of the same participant (…) in which no
identification is stronger than the one before it’. Avec Dik (1989: 271) on pourrait dire qu’un
Topique Nouveau qui est repris par une expression anaphorique forte (comme dans Hdt.
1.6.1) ouvre un ‘nouveau champ d’ identification’ (‘new identification span’). Je fais remar-
quer que l’ anaphore forte d’ Hérodote 1.6.1. est exceptionnelle, car plus souvent elle vient
après (une série) d’ expressions anaphoriques faibles, comme dans (14). Cf. aussi l’exemple
de Dik : ‘Yesterday I got a phone call from the tax inspector (NewTop). He wanted me to
come to his office, and he gave me the impression that I was in for some trouble. Now, this
inspector happens to be a friend of my sister’s…’. Ici, le premier ‘champ d’identification’
va de ‘He’ jusqu’ à ‘trouble’ inclusivement. En revanche, ‘this inspector’ ouvre un nouveau
champ d’ identification.
différences entre ουτοσ ο / ουτοσ δε ο / ο δε ουτοσ 229

(11) ἀνὴρ ἐν τοῖσι Μήδοισι ἐγένετο σοφὸς τῷ οὔνομα ἦν Δηιόκης, παῖς δὲ ἦν Φραόρ-
τεω. οὗτος ὁ Δηιόκης ἐρασθεὶς τυραννίδος ἐποίεε τοιάδε. (1.96.2)

Il y avait chez les Mèdes un homme avisé, appelé Dèiokès, qui était fils de
Phraorte. Épris du souverain pouvoir, ce Dèiokès agit de la façon suivante:
…

Tandis que, dans des cas comme (9)–(11), où le groupe avec οὗτος vient après
l’ introduction d’un personnage comme Topique Nouveau dans le récit, l’ asyn-
dète s’explique aisément, il y a d’autres cas, au milieu d’ un récit, où la présence
de δέ ne paraît nullement exclue. Voir (12) et (13) :

(12) τοσόνδε μέντοι εἶπε ἡ Πυθίη, ὡς Ἡρακλείδῃσι τίσις ἥξει ἐς τὸν πέμπτον ἀπό-
γονον Γύγεω. τούτου τοῦ ἔπεος Λυδοί τε καὶ οἱ βασιλέες αὐτῶν λόγον οὐδένα
ἐποιεῦντο, … (1.13.2)

Toutefois, la Pythie ajouta que les Héraclides seraient vengés sur le qua-
trième descendant de Gygès. Mais ni les Lydiens ni leurs rois ne tinrent
aucun compte de cette prédiction avant qu’ elle fût accomplie.

(13) Τόμυρις δέ… συνέβαλε Κύρῳ. ταύτην τὴν μάχην… κρίνω ἰσχυροτάτην γενέσθαι,
καὶ δὴ καὶ πυνθάνομαι οὕτω τοῦτο γενόμενον. (1.214.1)

Tomyris… en vint aux mains avec lui. Ce combat… fut à mon avis le plus
violent. D’après mes informations, il se passa de la manière suivante : …

Je voudrais avancer l’hypothèse que, tout comme dans (9)–(11), la position

prépositive du pronom met l’accent sur l’identité de l’ entité en question11,

11 Dans ces cas aussi le groupe οὗτος ὁ + substantif reprend une entité qui vient d’être intro-
duite comme Topique Nouveau, à savoir τοσόνδε (εἶπε) ὡς…, et μάχη, qui est implicite
dans συνέβαλε. Je fais observer cependant que dans ces cas-là l’entité en question n’est
pas caractérisée de façon explicite comme Topique Nouveau, à la différence de (9)–(11).
Les différences entre (9)–(11) d’ une part et (12)–(13) de l’autre, peuvent être mises en
lumière également de la façon suivante. Après l’ introduction de Crésus par Κροῖσος ἦν
Λυδὸς μὲν γένος κτλ., le narrateur ne peut guère s’ arrêter, car le lecteur/auditeur s’attendra
à ce qu’Hérodote se serve, pour ainsi dire, de ce Topique, en faisant de lui un participant.
Par contre, les phrases τοσόνδε εἶπε ἡ Πυθίη ὡς dans (12) et Τόμυρις δέ… συνέβαλε Κύρῳ dans
(13) ne contiennent pas une obligation inhérente d’en dire plus sur ces événements. Je
crois donc que l’ opinion de Dik (1989: 268) : ‘… the object or second argument position is
often used for the introduction of NewTops’, est trop forte. Son exemple est: ‘In the circus
 230 chapter 11

tandis que l’absence de δέ signale à nouveau qu’ on n’ a pas affaire à une ‘Fort-
setzung’, et qu’Hérodote recommence donc en quelque sorte son récit. Dans
ces contextes, l’effet produit est que l’entité en question (ἔπος et μάχη, res-
pectivement), ainsi que toutes les informations données sur cette entité, sont
125 détachées du contexte et obtiennent une certaine | indépendance. Par là, le nar-
rateur appelle notre attention sur cette entité et sur son rôle dans l’ histoire12.
Ici encore, ces effets pragmatiques s’expliquent bien, car dans les deux cas il
s’ agit d’une entité qui a une importance capitale pour l’ histoire: la négligence
des Lydiens est la cause ultime du sort de Crésus (ex. (12)), et la bataille entre
Tomyris et Cyrus se termine par la mort du roi.
Dans tous ces cas il s’agit donc d’une entité et d’ une information sur cette
entité qui sont toutes les deux marquées comme ‘saillantes’13.
Bien entendu, l’emploi de οὗτος ὁ + substantif ne peut être compris qu’ en
faisant une comparaison avec l’emploi de οὗτος δὲ ὁ + substantif.

2 οὗτος δὲ ὁ + substantif

Ici encore je commence par présenter l’exemple (2) dans son contexte :

(14) Γύγης δὲ τυραννεύσας ἀπέπεμψε ἀναθήματα ἐς Δελφοὺς οὐκ ὀλίγα, … οὗτος

δὲ ὁ Γύγης πρῶτος βαρβάρων τῶν ἡμεῖς ἴδμεν ἐς Δελφοὺς ἀνέθηκε ἀναθήματα
μετὰ Μίδην τὸν Γορδίεω… (1.14.2)

we saw an elephant called Jumbo’. Ici, ‘an elephant’ n’a pas le même statut pragmatique
que ‘an elephant’ dans l’ exemple suivant: ‘Once upon a time there was an elephant called
Jumbo’. À mon avis, il vaut mieux appeler ‘an elephant’ dans le premier cas un Topique
Nouveau potentiel. Bien entendu, une telle entité fonctionnera le plus souvent comme
un vrai Topique Nouveau, car normalement un narrateur n’introduit pas une entité ‘pour
rien’.
12 Cf. Grimes (1975: 96) : ‘In a system that makes extensive use of linkages it is the absence of
a linking clause [ou, dans notre cas, d’ une particule de liaison] that catches the hearer’s
attention’.
13 Pour être complet, j’ ajoute que l’ asyndète se trouve parfois dans des phrases qui com-
mencent avec un groupe οὗτος ὁ + substantif, sans qu’il s’agisse d’une entité qui est
importante pour l’ histoire, cf. 1.50.3 : ἐποιέετο δὲ καὶ λέοντος εἰκόνα χρυσοῦ ἀπέφθου, ἕλκου-
σαν σταθμὸν τάλαντα δέκα. οὗτος ὁ λέων, ἐπείτε κατεκαίετο ὁ ἐν Δελφοῖσι νηός, κατέπεσε…
Noter toutefois qu’ il s’ agit d’ un monument spectaculaire, qui apparemment mérite d’être
mis en relief; cf. à nouveau Grimes (1975: 96) : ‘… ASYNDETON or break in the sequence
may be used to signal… a point of special emphasis’. Voir aussi note 15.
différences entre ουτοσ ο / ουτοσ δε ο / ο δε ουτοσ 231

Gygès, quand il fut souverain, envoya en hommage à Delphes des off-

randes abondantes; … Ce Gygès est le premier Barbare à notre connais-
sance qui ait consacré des offrandes à Delphes, après Midas fils de Gor-
dias…

Dans ce cas-là, le groupe οὗτος (δὲ) ὁ Γύγης n’est pas précédé d’ une phrase où
Gygès est introduit comme personnage dans le récit. Bien au contraire, il est
le Topique primaire depuis le chapitre 12.2, et nous est connu comme parti-
cipant depuis 8.1. En outre, l’information sur les offrandes envoyées par lui à
Delphes est, elle aussi, déjà connue. Si Hérodote veut tout de même identifier ce
Topique de cette façon explicite, c’est qu’il le reprend après une digression14. Ici
encore, il s’agit d’anaphore forte: ‘ce Gygès et nul autre’. Quant à la conjonction
δέ, elle indique bien ici qu’Hérodote enchaîne, et qu’ il y a donc ‘Fortsetzung’.
Plus précisement, δέ indique que l’information donnée dans cette phrase est de
l’ information additionnelle sur l’action de Gygès qu’Hérodote vient de décrire.
Cette information est donc marquée comme ‘non | saillante’, ou peut-être, pour 126
employer une formule plus prudente, comme ‘pas plus saillante que les infor-
mations précédentes sur ce Topique’15.
Les exemples (16) et (17) peuvent être analysés de la même façon.

(16) διὰ τοῦτον δὲ τὸν φόνον αἱ γυναῖκες αὗται νόμον θέμεναι σφίσι αὐτῇσι ὅρκους
ἐπήλασαν… (1.146.3)

C’est à cause de ce meurtre que ces femmes s’ imposèrent cette loi,

qu’elles firent serment d’observer…

(17) αὗται δὲ αἱ πόλιες τῇσι πρότερον λεχθείσῃσι ὁμολογέουσι κατὰ γλῶσσαν οὐδέν,
σφίσι δὲ ὁμοφονέουσι. (1.142.4)

14 οὗτος ὁ Γύγης marque le début d’ un nouveau champ d’identification (voir note 10). Je fais
remarquer cependant qu’ il n’est pas nécessaire qu’ un groupe οὗτος ὁ + substantif vienne
après une digression, cf. par exemple (16) ci-dessous.
15 La différence entre οὗτος ὁ λέων… κατέπεσε dans 1.50.3 (voir note 13), et οὗτος δὲ ὁ Γύγης…
ἀνέθηκε dans (14) est, me semble-t-il, que l’ information sur le lion est, dans le contexte, plus
intéressante que l’ information sur Gygès dans son contexte. On notera que dans (14) on
trouve πρῶτος βαρβάρων (τῶν ἡμεῖς ἴδμεν), tandis que dans (9) l’ordre est βαρβάρων πρῶ-
τος (τῶν ἡμεῖς ἴδμεν). Il est difficile d’ expliquer la raison d’être de cette variation. Peut-être
l’ hyperbate de βαρβάρων dans (9) indique-t-il avec unce certaine emphase qu’Hérodote
va tout d’ abord se concentrer sur le rôle des non-grecs dans le conflit, comme il l’avait
annoncé au ch. 5.3.
 232 chapter 11

celles-là, au point de vue de la langue, diffèrent tout à fait des précédentes

et s’accordent entre elles

Dans (16), Hérodote donne un détail sur le meurtre qu’ il a mentionné dans la
phrase précédente. Δέ signale le caractère additionnel de ce détail, tandis que,
par l’antéposition de οὗτος, le φόνος est identifié de façon emphatique avec le
φόνος impliqué dans ἐφόνευσαν de la phrase précédente. Pour ce qui est de αἱ
γυναῖκες αὗται, en revanche, il s’agit d’anaphore faible, voir plus bas. Quant à
(17), l’identification emphatique est d’autant plus marquée que le constituant
en question est opposé à un autre constituant (τῇσι πρότερον λεχθείσῃσι).
Pour résumer, on pourrait dire que dans les exemples du deuxième type
il y a simultanément deux effets présents: premièrement, tout comme dans
les exemples du premier type, οὗτος a une fonction anaphorique ‘forte’: par sa
position prépositive il identifie d’emblée le référent du nom (propre) avec le
référent du même nom mentionné (ou impliqué) dans le contexte précédent.
À la différence du premier type, toutefois, la présence de δέ indique que le nar-
rateur présente des informations additionnelles sur le constituant en question.
L’ intérêt du contenu d’une telle phrase est donc moins grand que celui d’ une
phrase asyndétique. Bref, la formule de type (2) introduit une entité marquée
comme saillante et, sur cette entité, une information non marquée comme
telle.
127 Troisièmement, il y a le groupe ὁ δὲ + substantif + οὗτος. |

3 ὁ δὲ + substantif + οὗτος

(18) τοῦτο μὲν ἀπὸ τῆς Θεσσαλίης παρὰ τῶν Ἀλευαδέων ἀπιγμένοι ἄγγελοι ἐπεκα-
λέοντο βασιλέα πᾶσαν προθυμίην παρεχόμενοι ἐπὶ τὴν Ἑλλάδα (οἱ δὲ Ἀλευάδαι
οὗτοι ἦσαν Θεσσαλίης βασιλέες), … (7.6.2)

de Thessalie, de chez les Aleuades, étaient venus des messagers, qui

déployaient tout leur zèle pour inciter le Roi à attaquer la Grèce (ces
Aleuades régnaient en Thessalie)

Comme dans le type précédent, δέ indique que les informations présentées

dans cette phrase sont des informations additionnelles, qui ont donc relative-
ment peu d’intérêt. Quant à οὗτοι, il a ici, en raison de sa postposition, une fonc-
tion anaphorique faible. Autrement dit, l’anaphore est établie en premier lieu
par le groupe οἱ δὲ Ἀλευάδαι, et on pourrait même dire, avec Palm (p. 28), que le
pronom a ici une position quasi-enclitique. L’effet combiné résultant de la pré-
différences entre ουτοσ ο / ουτοσ δε ο / ο δε ουτοσ 233

sence de δέ et de l’anaphore faible est que les phrases introduites par le groupe
ὁ δὲ + substantif + οὗτος ont souvent le caractère d’ une (brève) digression, qui
ressemble beaucoup à nos notes de bas de page: tant l’ entité que l’ information
sur cette entité sont marquées comme ‘non saillantes’. Quand elles se trouvent
dans un contexte narratif ces phrases présentent essentiellement des informa-
tions ‘de second plan’ (‘background information’). À ce propos je fais observer
que le plus souvent le verbe de ces phrases est un verbe d’ état, notamment une
forme de εἶναι. Voir (18), et des exemples comme (19) et (20)16 ; dans (19), les
parenthèses dans la traduction de Legrand mettent bien en évidence le carac-
tère digressif de ὁ δὲ χῶρος οὗτος κτλ. Dans les phrases qui commencent par
οὗτος ὁ + substantif ou οὗτος δὲ ὁ + substantif, en revanche, il n’y a aucune forme
de εἶναι17.

(19) τὰς μὲν δὴ νέας… οἱ ναύαρχοι… ἐς τὸν αἰγιαλὸν τὸν προσεχέα Δορίσκῳ ἐκόμι-
σαν, ἐν τῷ… ὁ δὲ χῶρος οὗτος τὸ παλαιὸν ἦν Κικόνων. (7.59.2)

… les capitaines conduisirent les vaisseaux à la plage voisine du château,

où… (ce territoire, jadis, était aux Ciconiens) | 128

(20) ἑσσωθέντες δὲ τῇ συμβολῇ, ὅσοι αὐτῶν ἐζωγρήθησαν, πέδας τε ἔχοντες· τὰς ἐφέ-
ροντο αὐτοὶ καὶ… τὸ πεδίον τὸ Τεγεητέων ἐργάζοντο. αἱ δὲ πέδαι αὗται ἐν τῇσι
ἐδεδέατο ἔτι καὶ ἐς ἐμὲ ἦσαν σόαι ἐν Τεγέῃ. (1.66.4)

Mais ils eurent le dessous dans la rencontre; et tous ceux d’ entre eux qui
furent pris vivants, chargés des entraves qu’ eux-mêmes avaient appor-
tées, mesurèrent au cordeau la plaine tégéatique en y travaillant. Les
entraves qui avaient servi à les enchaîner subsistaient encore à mon
temps à Tégée.

À propos de (20), je signale un phénomène important d’ ordre syntaxique. Le

groupe αἱ δὲ πέδαι αὗται est suivi d’une phrase relative, plus précisement d’ une

16 Autres exemples de εἶναι : 1.157.3, 160.4, 201 ; 7.6.2, 59.2, 91, 109.2, 111.2, 153.1, 180. Dans 1.74.2 la
phrase à ὁ δὲ + substantif + οὗτος est également marquée comme faisant partie de l’arrière-
plan par προ- dans προηγόρευσε. Il est intéressant de noter que dans 1.82.2, où un groupe ὁ
+ substantif + οὗτος se trouve dans une phrase qui décrit bel et bien un événement qui fait
partie du récit principal, la conjonction est γάρ et non pas δέ. Dans 1.85.4 (ὁ δὲ παῖς οὗτος
ὁ ἄφωνος), je crois qu’ il faut combiner οὗτος avec ὁ ἄφωνος, et non pas avec ὁ παῖς, contra
Powell. Par son contenu non digressif, cette phrase diffère fondamentalement des phrases
régulieres avec οὗτος postpositif.
17 Dans le corpus examiné. Dans 1.187.3 ἦν ἀκίνητος (= οὐκ ἐκινέετο) n’exprime pas un état.
234 chapter 11

relative digressive (la relative n’est nullement nécessaire pour l’ identification

de l’antécédent et pourrait être omise). Or, il s’avère que, si un groupe du type ὁ
δὲ + substantif + οὗτος est suivi d’une relative, (a) il s’ agit toujours d’ une digres-
sive, et (b) οὗτος est toujours anaphorique. Si, par contre, un groupe οὗτος ὁ +
substantif est suivi d’une relative, il s’agit toujours d’ une relative restrictive.
En outre, dans ces cas-là le pronom a une fonction cataphorique et non pas
anaphorique18. Il faut toutefois que j’ajoute que le comportement syntaxique
de ces syntagmes-ci n’est pas tout à fait identique, car ils ne semblent pas se
trouver au début absolu d’une phrase mais seulement à l’ intérieur. Voir par
exemple:

(21) … καὶ οἰκήσαντας τοῦτον τὸν χῶρον τὸν καὶ νῦν οἰκέουσι (1.1.1)

… après avoir établi leur demeure dans le territoire qu’ ils habitent encore
aujourd’hui

(22) … ὡς κατὰ τοῦτο τὸ χωρίον ἐγίνοντο ἵνα τῆς παροιχομένης νυκτὸς κατεδέδετο
ἡ θήλεα ἵππος (3.86.1)

… quand ils furent près du lieu où, la nuit précédente, avait été attachée
la jument

J’ en viens à ma conclusion, qui sera présentée dans l’ ordre inverse de l’ exposé

qui précède.

4 Conclusion

Dans tous les cas, les syntagmes traités ci-dessus, qui se trouvent au début
absolu d’une phrase, ont une fonction anaphorique : ils reprennent comme
Topique une entité qui est présente dans le contexte précédent. Dans les
phrases qui commencent par ὁ δὲ + substantif + οὗτος, le pronom a une fonc-
tion anaphorique faible; l’identification du Topique se fait en premier lieu par

18 C’ est donc à tort que Powell classe les deux types sous la rubrique IV ‘Antecedent’. En
réalité, ce n’ est que οὗτος préposé qui sert d’ antécédent. Le fait que οὗτος postposé ne
s’ emploie apparemment pas comme un pronom cataphorique confirme, me semble-t-il,
le caractère non emphatique de l’ anaphore établie par le syntagme ὁ δὲ + substantif +
οὗτος en général, ainsi que le statut enclitique du pronom. Comparer aussi les remarques
ci-dessus sur (7) et (8).
différences entre ουτοσ ο / ουτοσ δε ο / ο δε ουτοσ 235

le substantif. La particule de liaison δέ indique que le narrateur présente des

informations additionnelles sur ce | Topique ‘non saillant’. Tant l’ entité que 129
l’ information sur cette entité sont donc marquées comme ‘non saillantes’.
Dans les phrases qui commencent par οὗτος δὲ ὁ + substantif, en revanche,
il s’agit d’anaphore forte: dans l’identification du Topique le pronom joue un
rôle essentiel. Le narrateur veut apparemment mettre en relief qu’ il s’ agit tou-
jours du même Topique: ‘cette entité et nulle autre’. Quant à la particule δέ,
tout comme dans le type précédent, elle indique que le narrateur présente des
informations additionnelles, et donc ‘non saillantes’, sur le Topique, qui, lui, est
‘saillant’.
Enfin, le groupe οὗτος ὁ + substantif, sans δέ. Ici encore, il s’ agit d’ anaphore
forte. Tout comme dans le cas précédent le narrateur veut apparament mettre
en relief qu’il s’agit toujours du même Topique: ‘cette entité et nulle autre’, ‘pré-
cisément cette entité’. C’est donc un Topique ‘saillant’. Quant à l’ absence de
δέ, il faut distinguer deux emplois. Le premier est représenté par des phrases
où l’absence de δέ semble obligatoire, notamment les phrases qui viennent
après l’introduction d’un personnage dans le récit, du type : ‘X était le fils
de…’; ‘Il y avait là un X…’. Dans de tels cas, la phrase qui commence par
οὗτος ὁ + substantif marque le vrai début du récit. Deuxième type d’ emploi :
l’ absence de δέ se trouve en plein milieu du récit. Dans ces cas-là l’ asyndète
donne une certaine indépendance à la phrase en question. Dans les deux cas
les informations données dans la phrase asyndétique sont marquées comme
‘saillantes’.
Les résultats peuvent être résumés d’une façon plus schématique comme
suit:
οὗτος ὁ + substantif: Topique saillant, informations saillantes ;
οὗτος δὲ ὁ + substantif: Topique saillant, informations non saillantes ;
ὁ δὲ + substantif + οὗτος: Topique non saillant, informations non saillantes19.

19 Pour être complet, j’ ajoute qu’ il y a un exemple chez Hérodote de la combinaison ὁ + sub-
stantif + οὗτος : ὁ νόμος οὗτος διατελέει ἐὼν ὅμοιος τὸ μέχρι ἐμεῦ ἀπ᾽ ἀρχῆς (2.113.2; v.l. οὕτως),
ainsi qu’ un exemple d’ un tel groupe avec μέν : τὰ μὲν χρηστήρια ταῦτά σφι ἐχρήσθη, οἱ δὲ…
(9.94.1), où il y a probablement asyndète. Ici on a donc affaire à un Topique non saillant
avec des informations qui sont marquées comme saillantes, ce qui paraît assez étrange.
Mais faute d’ autres exemples il est difficile d’ en dire plus. Il se peut que l’occurrence de
tels syntagmes chez d’ autres auteurs permette de mieux discerner leur valeur.
236 chapter 11

Références

Dik, S.C., The Theory of Functional Grammar (Dordrecht 1989).

Grimes, J.E., The Thread of Discourse (La Haye/Paris 1975).
Palm, J., Zur Funktion und Stellung des attributiven Demonstrativums im Griechischen
(Lund 1960).
Powell, J.E., A Lexicon to Herodotus (Cambridge 1938; repr. Hildesheim 1966).
 chapter 12

Sur les emplois de ἐάν et ἐπεάν (à propos d’ Euripide,

Bacchae 50–51)

Résumé

On étudie ici les emplois des propositions conditionnelles du type ἐάν + subjonctif,
ainsi que ceux des propositions temporelles avec ἐπεάν + subj. On essaie de démon-
trer qu’il y a une différence fondamentale, pour les subordonnées introduites par ἐάν,
entre celles qui ont valeur de futur et celles qui ont valeur ‘générique’. Dans les proposi-
tions qui réfèrent au futur, ἐάν + subjonctif exprime qu’il est fort possible que l’état de
choses en question se réalise. (Tandis que εἰ + indicatif futur ne présente aucune indi-
cation quant au degré de probabilité, et que εἰ + optatif indique que la réalisation de la
condition est considérée comme simplement possible). Les propositions avec ἐπεάν +
subj., par contre, s’ils réfèrent au futur, expriment que la réalisation de l’état de choses
est pratiquement certaine.
Ces deux types se trouvent également dans des phrases génériques, c’est-à-dire des
phrases qui font partie des descriptions d’habitudes. Comme il s’agit dans ce cas-là de
faits (présentés comme) réels, il est impossible de dire que ἐάν + subjonctif exprime
que la réalisation de l’état de choses est ‘fort possible’, ἐάν + subjonctif exprime plutôt
que l’état de choses n’est pas réalisé en toute occasion possible, et donc seulement de
temps à autre dans la situation en question, ἐπεάν + subj., par contre, présuppose la
réalisation de l’état de choses, ce qui veut dire que l’état de choses est toujours réalisé
dans la situation en question.

Abstract

In this article the uses of conditional clauses of the type έάν + subjunctive, and those of
temporal clauses involving ἐπεάν + subj., are examined. It is argued that there is a funda-
mental difference between ‘future’ ἐάν-clauses and ‘generic’ ἐάν-clauses. In clauses hav-
ing future reference, ἐάν + subjunctive expresses that it is quite possible that the state
of affairs concerned will be realised (as against εἰ + future indicative, which present no
indication as to the likelihood of the realisation of the state of affairs, and εἰ + optative,
which expresses that it is just possible that the state of affairs will be realised). ἐπεάν +
subj. clauses, on the other hand, if they have future reference, express that realisation
of the state of affairs is virtually certain.

© koninklijke brill nv, leiden, 2019 | doi:10.1163/9789004386129_014

 238 chapter 12

Both types are also used in generic clauses, i.e. clauses occurring in descriptions of
habits. Since in this case facts are described, it makes no sense to say that ἐάν + sub-
junctive expresses that realisation of the state of affairs is ‘quite possible’. Rather, ἐάν +
subjunctive expresses that the state of affairs is realised sometimes, i.e. not on all pos-
sible occasions in the given situation, ἐπεάν + subj., on the other hand, presupposes the
realisation of the state of affairs, which means specifically that the state of affairs is
132 always realised in the given situation. |

1 Subordonée en ἐάν à valeur future: ‘éventualité’ ou ‘attente’ ?

Comme point de départ pour mon analyse de la proposition conditionnelle

introduite par ἐάν je prends le passage suivant:

(1) ἢν δὲ Θηβαίων πόλις

ὀργῇ σὺν ὅπλοις ἐξ ὄρους βάκχας ἄγειν
ζητῇ, ξυνάψω μαινάσι στρατηλατῶν. (Ε. Ba. 50–51)

Et si Thèbes en courroux veut ramener les bacchantes de la montagne

par la force des armes, je prendrai le commandement des ménades et
j’engagerai la bataille (trad. Roux)

Les commentaires gardent le silence sur l’interprétation de la conditionnelle.

C’est d’ailleurs le comportement habituel des commentateurs, qui ne s’ arrê-
tent que rarement à l’emploi des modes. Ici, Verdenius (1980) est le seul à
faire une remarque; il dit, à tort il me semble, que ‘the possibility of an armed
conflict is mentioned as a mere eventuality’. Dans cet exposé je voudrais trai-
ter d’abord brièvement la valeur sémantique du subjonctif + ἄν dans des cas
comme Ba. 50–51. Ensuite, j’aborderai de façon plus détaillée la question de
savoir si cette valeur est également présente dans les conditionnelles au sub-
jonctif + ἄν que l’on trouve dans des descriptions. Dans les deux cas je ferai une
comparaison avec l’emploi du subjonctif dans les temporelles apparentées, à
savoir celles introduites par ὅταν, ἐπεάν, etc.
Tout d’abord il faut dire que le silence quasi-universel des commentaires
dans des cas comme Ba. 50 se comprend très bien. Après tout, l’ interprétation
globale d’une conditionnelle comme ἢν ζητῇ ne pose normalement aucun pro-
blème. On risque toutefois de perdre une nuance de quelque importance en
traduisant par un simple ‘si’, ‘if’, etc., à savoir la nuance apportée par le sub-
jonctif + ἄν. À mon avis on peut expliquer l’emploi du subjonctif + ἄν dans
les conditionnelles qui portent sur l’avenir en attribuant à ce mode la valeur
sur les emplois de εαν et επεαν 239

sémantique de ‘forte possibilité’. Par l’emploi du subjonctif le locuteur exprime

qu’il considère la réalisation de l’état de choses en question comme fortement
possible, ou même probable1. Ici donc Dionysos indique qu’ il s’ attend à ce que
les Thébains essaient de chasser les Bacchantes. Avec Kühner-Gerth (2, 473) on
pourrait appeler ce subjonctif le mode de l’attente: ‘Ἐάν c. conj. wird gebraucht,
wenn die Bedingung als eine solche hergestellt wird, deren Verwirklichung zu
erwarten steht’.
Malheureusement les savants allemands ne s’ en tiennent pas là, puisqu’ ils
ajoutent ‘je nach Lage der Umstände’, donc ‘éventuellement’, apparemment
pour faire ressortir la valeur de ἄν; comparer aussi leur traduction de εἰ ἄν
= ‘wenn vorkommendenfalls’. À la page suivante ‘l’ éventualité’ l’ a nettement
emporté sur | l’‘attente’ (Erwartung); là, ils ne parlent que de ‘eventuelle Ver- 133
wirklichung’. Or ce terme d’‘éventuel’ me paraît assez mal choisi ; il est trop
vague, et ne fait pas ressortir suffisamment la différence avec les autres modes
qu’on peut trouver dans les conditionnelles portant sur l’ avenir. Par son vague,
il donne à penser qu’il s’agit d’un état de choses qui pourrait en effet se réa-
liser éventuellement, donc à la rigueur: voir la remarque de Verdenius que je
viens de citer. (Je fais observer entre parenthèses que pour l’ attique cela n’a
guère de sens de vouloir attribuer une valeur spécifique à ἄν, puisque dans les
subordonnées en question la particule est quasi-obligatoire)2. Quand K-G en
viennent à faire une comparaison de ce subjonctif avec d’ autres modes, on voit
clairement les effets indésirables de leur préférence pour une ‘eventuelle Ver-
wirklichung’, témoin la citation suivante (K-G 2, 474): ‘Der Unterschied (à savoir
entre εἰ + ind. fut. et ἐάν + subj.) liegt nur darin, dass der Redende, wenn er εἰ c.
ind. Fut. gebraucht, von der Wirklichkeit geflissentlich absieht, den Satz als ein
rein logische Schlussfolgerung ausspricht, während er mit ἐάν c. conj. auf die
eventuelle Verwirklichung hindeutet. Oft kann also ebensowohl die eine, wie
die andere Form gewählt werden. Daher wechseln zuweilen die beiden Aus-
drücksweisen ohne wesentlichen Unterschied ab’. Pour illustrer cette remarque
ils mentionnent le passage Hérodote 1.71.3:

1 Comparer aussi Amigues (1977: 117): ‘Le subjonctif est affecté à l’expression d’un procès dont
la réalisation est tenue pour probable mais non assurée’; elle appelle toutefois ce subj. un
subjonctif éventuel. En outre, elle croit que ce subj. s’ emploie tant dans les conditionnelles
que dans les temporelles, à tort selon moi. J’ y reviens plus bas. Bien que la notion de ‘pro-
babilité’ soit souvent pertinente, j’ ai préfére employer la notion de ‘forte possibilité’ comme
terme englobant, parce qu’ il y a des contextes où la notion de ‘probabilité’ ne convient mani-
festement pas ; voir plus bas, ex. 9.
2 Pour ce qui est d’ Homère, si ἄν a une valeur spécifique, c’est plutôt ‘auf Lage der Umstände’
que ‘je nach Lage der Umstände’; ἄν relie la réalisation de la condition à la situation où le locu-
teur se trouve : c’ est cette situation qui l’ amène à s’ attendre à la réalisation de la condition.
 240 chapter 12

(2) τοῦτο μὲν δή, εἰ νικήσεις, τί σφεας ἀπαιρήσεαι, τοῖσί γε μὴ ἔστι μηδέν; τοῦτο
δέ, ἤν νικηθῇς, μάθε ὅσα ἀγαθὰ ἀποβαλέεις. (Hdt. 1.71.3)

Alors, si tu les vaincs, que leur prendras-tu, à des gens qui n’ont rien ? Si
au contraire tu es vaincu, rends-toi compte de tous les avantages que tu
vas perdre (trad. Legrand)

Ici il n’y a donc pas, à en croire K-G, de différence réelle entre l’ indicatif futur
et le subjonctif. Or cela me paraît très douteux. Il me semble au contraire
qu’il y a une distinction nette. Dans le chapitre 71, le locuteur, Sandanis, essaie
de persuader Crésus qu’il ferait mieux d’abandonner son projet de monter
une expédition contre les Perses. Dans ce contexte, l’ emploi des modes est
tout à fait significatif. L’emploi de l’indicatif futur indique que le locuteur
s’ abstient de tout jugement sur la probabilité de réalisation de la condition : il
se déclare complètement neutre à cet égard; comparer le ‘rein logische Schluss-
folgerung’ de K-G3. Peut-être Crésus remportera-t-il la victoire, peut-être non.
Comme souvent, l’effet pragmatique de l’emploi de l’ indicatif est que le locu-
teur exprime un certain scepticisme en ce qui concerne la réalisation de la
condition. Le subjonctif par contre indique que le locuteur considère le second
terme de l’alternative comme probable; comme Dionysos dans Ba. 50, Sanda-
nis s’attend à le voir se réaliser – tout naturellement, puisqu’ il veut dissuader
Crésus de cette expédition qu’il tient pour néfaste.
J. Humbert partage l’opinion de K-G, comme il ressort de la citation suivante
(p. 220): ‘On conçoit aisément qu’il soit impossible d’ établir une distinction
rigoureuse entre un indicatif futur – qui se donne l’ avenir comme une réalité
134 proche – et un | subjonctif – qui attend une éventualité dont la réalisation peut
être immédiate: le tenter serait d’autant plus vain que le subjonctif et le futur
possèdent des affinités profondes. En principe, le futur comporte quelque chose
de plus affirmé, de plus assuré que le subjonctif et apparaît bien comme tel dans
de nombreux exemples; mais aussi simplement le souci de varier l’ expression
peut les faire employer côte à côte dans la même phrase’4.
Quant au troisième type de conditionnelle portant sur l’ avenir, celui à l’ opta-
tif, le mode y exprime que la réalisation de la condition est présentée comme
une simple possibilité. Si l’on voulait employer le terme d’ éventuel, il serait

3 Il s’ ensuit que cet indicatif n’ a qu’ une valeur temporelle: il situe l’état de choses à un moment
postérieur au moment d’ énonciation.
4 À noter que la valeur attribuée par Humbert à l’ indicatif du futur est fausse: dans les condi-
tionnelles l’ indicatif n’ exprime jamais la réalité. Voir supra.
sur les emplois de εαν et επεαν 241

mieux de mise pour caractériser l’optatif que le subjonctif. Souvent, en effet, la

réalisation de la condition, tout en étant possible, n’a qu’ une pertinence limi-
tée dans le contexte. Voir par exemple:

(3) ἐγὼ δὲ οἷόν τε ἐγίγνετο λέξω, καὶ ἀφ᾽ ὧν ἄν τις σκοπῶν, εἴ ποτε καὶ αὖθις ἐπι-
πέσοι, μάλιστ᾽ ἂν ἔχοι τι προειδὼς μὴ ἀγνοεῖν, ταῦτα δηλώσω (Thuc. 2.48.3)

Pour moi, je dirai comment cette maladie se présentait; les signes à obser-
ver pour pouvoir le mieux, si jamais elle se reproduisait, profiter d’ un
savoir préalable et ne pas être devant l’inconnu : voilà ce que j’ exposerai
(trad. de Romilly)

L’ optatif indique que Thucydide considère la répétition de la peste comme une

possibilité lointaine, mais tout de même comme une possibilité. Par l’ emploi
d’un subjonctif par contre il aurait suggéré qu’il s’ attendait à ce que la peste
réapparaisse, dépassant peut-être ainsi les limites de ses connaissances (viola-
tion des maximes gricéennes de la qualité (‘do not say that for which you lack
adequate evidence’) et de la quantité (‘do not make your contribution more
informative than is required’, cf. Levinson 1983: 101)). L’indicatif futur à son
tour aurait suggéré un certain scepticisme à cet égard, ce qui est rhétorique-
ment faible.
Dans beaucoup de cas le locuteur, par une sorte de jeu d’ esprit, envisage
un certain état de choses simplement pour épuiser les possibilités. Dans de tels
cas il peut y avoir un assez grand écart entre la situation actuelle et celle décrite
dans la conditionnelle. Du coup, la valeur de la conditionnelle à l’ optatif peut
être voisine de l’irréel, comme dans l’exemple suivant :

(4) τὸ μὲν αὐτῶν (scil. ὀΐων) ἕτερον ἔχει τὰς οὐρὰς μακράς, τριῶν πήχεων οὐκ ἐλάσ-
σονας, τὰς εἴ τις ἐπείη σφι ἐπέλκειν, ἕλκεα ἂν ἔχοιεν ἀνατριβομένων πρὸς τῇ γῇ
τῶν οὐρέων· νῦν δ’… (Hdt. 3.113.1)

Ceux de la première espèce ont une longue queue ne mesurant pas moins
de trois coudées; si on les laissait les traîner, le frottement de cette queue
contre le sol leur donnerait des ulcères; mais en fait… (trad. Legrand)

Noter la présence de νῦν δέ, que l’on trouve également souvent après une
irréelle. Par l’emploi de l’optatif, toutefois, Hérodote fait croire qu’ il serait à
la rigueur possible de laisser traîner les queues sur le sol. | 135
Un exemple net des différences entre le subjonctif et l’ optatif est :
242 chapter 12

(5) ΣΩ. Σχολὴ μὲν δή, ὡς ἔοικε· καὶ ἅμα μοι δοκοῦσιν ὡς ἐν τῷ πνίγει ὑπὲρ κεφα-
λῆς ἡμῶν οἱ τέττιγες ᾄδοντες καὶ ἀλλήλοις διαλεγόμενοι καθορᾶν καὶ ἡμᾶς. εἰ
οὖν ἴδοιεν καὶ νὼ καθάπερ τοὺς πολλοὺς ἐv μεσημβρίᾳ μὴ διαλεγομένους ἀλλὰ
νυστάζοντας καὶ κηλουμένους ὑφ᾽ αὑτῶν δι᾽ ἀργίαν τῆς διανοίας, δικαίως ἂν
καταγελῷεν, … ἐὰν δὲ ὁρῶσι διαλεγομένους καὶ παραπλέοντάς σφας ὥσπερ
Σειρῆνας ἀκηλήτους, ὃ γέρας παρὰ θεῶν ἔχουσιν ἀνθρώποις διδόναι, τάχ᾽ ἂν
δοῖεν ἀγασθέντες. (Pl. Phdr. 259a)

En tout cas nous avons le temps, à ce qu’il paraît ! Et tout ensemble, j’ en

ai l’idée, les cigales qui, selon l’usage au fort de la chaleur, chantent et
conversent entre elles au-dessus de nos têtes, ont l’ œil sur nous. Si donc
elles nous voyaient, même nous deux, à l’heure de midi imiter les gens du
commun et ne point converser, mais au contraire laisser choir notre tête et
céder à leurs enchantements par inertie intellectuelle, à juste titre elles se
riraient de nous (…). Si, au contraire, elles nous voient converser, et notre
esquif les côtoyer comme des Sirènes, sans céder à leurs enchantements,
alors ce privilège dont les dieux leur ont accordé de faire aux hommes
le présent, peut-être nous en feraient-elles présent dans leur satisfaction !
(trad. Robin)

Socrate s’attend donc à ce que les cigales voient Phèdre converser avec lui. Prag-
matiquement ceci implique, bien entendu, que Socrate espère que la seconde
option sera réalisée. Je signale en passant la différence aspectuelle entre ἴδοιεν,
un état de choses sémelfactif, et ὁρῶσι, un état de choses duratif: un regard suf-
fit pour que les cigales se mettent à rire, si Socrate et Phèdre ne conversent
pas, mais elles continueront à les regarder s’ils conversent bel et bien, tout
comme elles. Il faudrait donc traduire ἴδοιεν par ‘remarquer’ plutôt que par
‘voir’.
Le sujet parlant grec avait donc, en somme, trois modes à sa disposition
pour exprimer son point de vue sur le degré de probabilité de réalisation d’ une
condition portant sur l’avenir:
– l’emploi de l’indicatif futur indique que le locuteur ne s’ exprime pas sur le
degré de probabilité;
– l’emploi du subjonctif + ἄν indique que le locuteur s’ attend à ce que la condi-
tion se réalise;
– l’emploi de l’optatif indique que le locuteur considère la réalisation de la
condition comme simplement possible.
Je fais observer finalement que si les conditionnelles se distinguent de la façon
décrite elles ont cependant un trait en commun, à savoir qu’ elles présentent
toutes une ‘situation disjonctive’, pour utiliser le terme introduit par Chr. Leh-
sur les emplois de εαν et επεαν 243

mann. Toute conditionnelle dénote de par sa nature un état de choses dont la

réalisation est une possibilité et non pas une nécessité.
En dehors des conditionnelles portant sur l’avenir, les modes qui viennent
d’être mentionnés ont. bien entendu, divers autres emplois. Dans le cadre de
cet exposé, où je me borne au subjonctif, je n’en relève que trois: (a) les tem-
porelles portant sur l’avenir; (b) les conditionelles ‘génériques’, c’ est à dire
les conditionelles qui se trouvent dans des descriptions d’ habitudes ; et (c)
les temporelles ‘génériques’, qui se trouvent également dans des descriptions
d’habitudes. | 136

2 Subordonnée en ἐπεάν à valeur future: la ‘certitude’

Quelle est l’interprétation de ἐπεάν dans un exemple comme Hdt. 1.9.3 ?

(6) ἐπεὰν δὲ ἀπὸ τοῦ θρόνου στείχῃ ἐπὶ τὴν εὐνὴν κατὰ νώτου τε αὐτῆς γένῃ, σοὶ
μελέτω τὸ ἐνθεῦθεν ὅκως μή σε ὄψεται ἰόντα διὰ θυρέων. (Hdt. 1.9.3)

Puis, quand du siège elle se dirigera vers le lit et que tu te trouveras dans
son dos, à toi de t’arranger alors pour qu’elle ne te voie pas pendant que
tu franchiras la porte (trad. Legrand)

La temporelle ressemble à une conditionnelle en ce qu’ elle porte, elle aussi,

sur l’avenir. Il en résulte que le locuteur ne peut pas être absolument cer-
tain que l’état de choses en question se réalisera vraiment. Mais par l’ emploi
de la conjonction temporelle avec le mode de la ‘forte possibilité’, il indique
qu’il considère la réalisation comme quasi assurée. La subordonnée exprime
donc le plus haut degré de probabilité en ce qui concerne la réalisation de
l’ état de choses en question. Dans (6) l’emploi de ἐπεάν s’ explique en effet très
bien. Le locuteur, Kandaulès, décrit au destinataire, Gygès, le futur comporte-
ment de sa femme. Or comme il doit avoir été témoin de ce comportement
en d’innombrables occasions précédemment, il peut présenter l’ arrivée de sa
femme comme certaine. Les conjonctions temporelles diffèrent évidemment
de la conjonction conditionnelle en ce qu’elles n’expriment pas une condition,
mais un lien temporel; elles ne présentent donc pas une situation disjonctive.
En termes positifs, elles présupposent la réalisation de l’ état de choses en ques-
tion. Les conjonctions temporelles appartiennent, en effet, à la catégorie des
‘déclencheurs de présupposition’, des ‘presupposition-triggers’, pour utiliser le
terme de Levinson (1983: 179). Cette différence fondamentale entre les condi-
tionnelles et les temporelles est souvent ignorée, voir la remarque suivante de
 244 chapter 12

K-G à propos des conditionnelles (2, 474): ‘Diese hypothetische Form (sc. ἐάν +
subj.) kann, …, geradezu als temporale Form bezeichnet werden. Wie nahe sie
sich mit den Temporalsätzen berührt, zeigen Beispiele wo ὅταν, ὁπόταν und ἐάν
mit einander wechseln…’.
Si l’on a l’impression que les deux types sont assez voisins, c’ est qu’ ils
portent tous deux sur l’avenir, où il n’y a pas, à proprement parler, de place
pour une notion comme ‘présupposition’. Il faut dire toutefois qu’ il y a beau-
coup de situations où il est tout à fait naturel de considérer la réalisation de tel
ou tel état de choses comme quasi certaine; Hdt. 1.9.3, mentionné plus haut,
en est un exemple clair. Il en va de même en E. Ba. 940 (7), Hdt. 3.3.3 (8), et
8.80.2.(9):

(7) ἦ πού με τῶν σῶν πρῶτον ἡγήσῃ φίλων, ὅταν παρὰ λόγον σώφρονας βάκχας
ἴδῃς (E. Ba. 940)

Tu me tiendras pour le premier de tes amis, n’est-ce pas, lorsque – contre

ton attente – tu verras la tenue décente des bacchantes (trad. Roux)

(8) τοιγάρ τοι, ὦ μῆτερ, ἐπεὰν ἐγὼ γένωμαι ἀνήρ, Αἰγύπτου τὰ μὲν ἄνω κάτω θήσω,
τὰ δὲ κάτω ἄνω. (Hdt. 3.3.3)

Eh bien, mère, quand moi je serai grand, je mettrai l’ Égypte sens dessus
137 dessous! (trad. Legrand) |

(9) σὺ δὲ ἐπεί περ ἥκεις χρηστὰ ἀπαγγέλλων, αὐτός σφι ἄγγειλον. ἢν γὰρ ἐγὼ αὐτὰ
λέγω, δόξω πλάσας λέγειν καὶ οὐ πείσω ὡς οὐ ποιεύντων τῶν βαρβάρων ταῦτα.
ἀλλά σφι σήμηνον αὐτὸς παρελθὼν ὡς ἔχει. ἐπεὰν δὲ σημήνῃς, ἢν μὲν πείθων-
ται, ταῦτα δὴ τὰ κάλλιστα, ἢν δὲ αὐτοῖσι μὴ πιστὰ γένηται, ὅμοιον ἡμῖν ἔσται.
(Hdt. 8.80.2)

Mais toi, puisque tu arrives porteur d’excellentes nouvelles, annonce-les

leur toi-même; si c’est moi qui le dis, ils ne me croiront pas, convain-
cus que les barbares ne font rien de cela. Mais toi, présente-toi devant le
conseil, annonce-leur ce qu’il en est. Lorsque tu l’ auras fait, s’ ils le croient,
ce sera pour le mieux; si tes paroles les laissent incrédule, notre situation
sera la même (trad. Legrand)

Dans (7) l’emploi de ὅταν se comprend très bien, car Dionysos est bien entendu
tout à fait sûr du futur comportement des bacchantes. En fait, l’ emploi de ἐάν
aurait été assez étonnant, puisque cette conjonction aurait suggéré que Dio-
sur les emplois de εαν et επεαν 245

nysos n’était pas vraiment sûr des bacchantes, ce qui, à son tour, aurait eu
des effets indésirables sur Penthée. Dans (8), l’ emploi du tour temporel ne
surprend pas: il appartient en effet aux conventions socio-culturelles de consi-
dérer l’état adulte comme un état que l’on atteindra, et ἐάν aurait produit ici
un effet assez bizarre, vu que dans tout ce passage Cambyse parle d’ un ton
très assuré. Pour un autre phénomène naturel dans une proposition temporelle
comparer:

(10) τάφος δὲ ποῖος δέξεταί μ᾽, ὅταν θάνω; (E. IT 625)

quel tombeau me recevra après ma mort?

Finalement, des conventions socio-culturelles jouent également un rôle dans

(9), où le locuteur, Thémistocle, après avoir donné l’ ordre σήμηνον à Aris-
tide, considère tout naturellement l’exécution de son ordre comme allant de
soi. Une fois de plus ἤν serait étrange. L’incertitude exprimée par ἢν ‘disjonc-
tif’ serait entrée en conflit avec l’autorité exprimée par l’ impératif. ‘Fais cela.
Quand tu l’auras fait…’ est une séquence plus naturelle que ‘Fais cela. Si tu le
fais…’. À propos des conditionnelles ἢν… ἐγὼ αὐτὰ λέγω, et ἢν δὲ αὐτοῖσι μὴ πιστὰ
γένηται, je fais remarquer qu’elles montrent clairement qu’ il y a des cas où la
notion de ‘probabilité’ ne s’applique pas (cf. note 1). Dans le premier cas, où
ἢν… ἐγὼ αὐτὰ λέγω suit αὐτός σφι ἄγγειλον, Thémistocle ne tient évidemment
pas la réalisation de l’état de choses ‘(si) je les annonce moi-même’ pour pro-
bable, puisqu’il vient de demander à Aristide d’annoncer lui-même les bonnes
nouvelles. Les conditionnelles ἢν μὲν πείθωνται… ἢν δὲ αὐτοῖσι μὴ πιστὰ γένη-
ται n’expriment pas non plus des conditions dont la réalisation est tenue pour
probable, puisqu’elles s’excluent mutuellement. Dans de tels cas, la notion
de ‘probabilité’ fait place à celle de ‘forte possibilité’, sans plus. Par ἢν… ἐγὼ
αὐτὰ λέγω, Thémistocle indique qu’il pourrait en principe très bien parler lui-
même; l’effet pragmatique du tour qu’il emploie est qu’ il présente le rejet de la
demande d’Aristide (ἀλλ᾽ ἐσελθών σφι ταῦτα σήμηνον, fin du chap. 79) sous une
forme assez polie. Pour ce qui est de ἢν μὲν πείθωνται… ἢν δὲ αὐτοῖσι μὴ πιστὰ
γένηται, par l’emploi du subjonctif tant dans l’une que dans l’ autre condition-
nelle, Thémistocle indique qu’il se peut fort bien que les conseillers croient
Aristide, mais qu’il est également fort possible qu’ ils ne le croient pas. Pour un
cas semblable, voir Hdt. 1.13.1. | 138
246 chapter 12

3 Ἐάν et ἐπεάν ‘génériques’: ‘réalisation occasionnelle’ et ‘réalisation

nécessaire’

Dans l’analyse que je viens de donner de l’emploi du subjonctif dans les condi-
tionnelles et temporelles portant sur l’avenir, j’ai attribué à ce mode une valeur
sémantique spécifique. Ceci se justifie tant par le fait que le subjonctif est
concurrencé par d’autres modes (ceci vaut notamment pour les condition-
nelles)5 que par le fait que l’on a affaire à l’avenir, où il est pertinent d’ opérer
avec des notions comme: ‘le locuteur exprime son point de vue sur le degré de
probabilité de réalisation de l’état de choses’.
Pour ce qui est de l’emploi du subjonctif dans des descriptions – je renvoie
ici aux exemples (11) et (12) – la situation est tout à fait différente.

(11) ἀπὸ Ἐλεφαντίνης πόλιος ἄνω ἰόντι ἄναντές ἐστι χωρίον· ταύτῃ ὦν δεῖ τὸ πλοῖον
διαδήσαντας ἀμφοτέρωθεν κατά περ βοῦν πορεύεσθαι· ἢν δὲ ἀπορραγῇ, τὸ
πλοῖον οἴχεται φερόμενον ὑπὸ ἰσχύος τοῦ ῥόου. (Hdt. 2.29.2)

A partir d’Éléphantine, en remontant, on trouve un pays escarpé ; pour

avancer, il faut attacher le bateau des deux côtés, comme un bœuf ; s’ il
échappe à ses liens, il est emporté par la violence du courant (trad.
Legrand)

(12) τὰ πλοῖα αὐτοῖσί ἐστι τὰ κατὰ τὸν ποταμὸν πορευόμενα ἐς τὴν Βαβυλῶνα ἐόντα
κυκλοτερέα πάντα σκύτινα· (…) ἐπεὰν ὦν ἀπίκωνται πλέοντες ἐς τὴν Βαβυ-

5 En ce qui concerne la valeur du subjonctif dans les temporelles, la situation n’est cepen-
dant pas tout à fait la même que dans les propositions introduites par ἐάν. Tandis que ἐάν
est concurrencé par εἰ + optatif et par εἰ + indicatif, ἐπεάν et ὅταν sont, dans le domaine des
temporelles, quasi-seuls. Comme l’ on sait, l’ indicatif futur n’est pas admis du tout après une
conjonction temporelle; noter que dans un exemple comme Hom. Od. 18.272: νὺξ δ᾽ ἔσται ὅτε
δὴ στυγερὸς γάμος ἀντιβολήσει | οὐλομένης ἐμέθεν, il ne s’ agit pas d’une vraie temporelle, mais
d’ une temporelle explicative, qui spécifie la nature de la nuit. Quant à l’optatif, bien qu’il
ne soit pas exclu il est toutefois très rare. Un exemple net est X. An.7.7.17: ἡμεῖς τοίνυν ἀπίοι-
μεν ἄν ὁπότε τὸν μισθὸν ἔχοιεν οἱ ταῦτα ὑμῖν πράξαντες. À propos de cet exemple K-G disent (2,
452) que la conjonction exprime ‘neben der Bedingung zugleich die Zeit’; pour ma part je
pense plutôt que s’ il y a une nuance conditionnelle ici, c’est que l’optatif présente la réali-
sation de l’ état de choses comme une simple possibilité. Quoiqu’en principe les temporelles
à l’ optatif fassent donc concurrence à ἐπεάν et ὅταν + subjonctif, il semble que dans la pra-
tique il n’existe pas de vraie compétition. Dans la catégorie ‘la réalisation de l’état de choses
est une simple possibilité’, εἰ + optatif était apparemment nettement préféré aux conjonc-
tions temporelles. Je fais d’ ailleurs observer à ce propos que l’on ne trouve jamais ἐπεί +
opt.
sur les emplois de εαν et επεαν 247

λῶνα καὶ διαθέωνται τὸν φόρτον, νομέας μὲν τοῦ πλοίου καὶ τὴν καλάμην πᾶσαν
ἀπ᾽ ὦν ἐκήρυξαν, τὰς δὲ διφθέρας ἐπισάξαντες ἐπὶ τοὺς ὄνους ἀπελαύνουσι ἐς
τοὺς Ἀρμενίους. (Hdt. 1.194.1–4)

Les bateaux des gens du pays, qui descendent le cours du fleuve pour
aller à Babylone, sont de forme circulaire et tout en cuir. (…) Lors donc
que les bateliers sont arrivés à Babylone et qu’ ils ont placé leur cargai-
son, ils vendent à la criée la carcasse du bateau et toute la paille ; puis
ils empilent les peaux sur leurs ânes et repartent pour l’ Arménie (trad.
Legrand) | 139

Tout d’abord, il ne s’agit pas, bien entendu, d’ assertions sur l’ avenir, ou du

moins pas sur un avenir défini, mais d’assertions génériques, qui réfèrent à
des états de choses qui ne sont pas limités dans le temps et qui ont donc une
valeur habituelle, ou, le cas échéant, omnitemporelle. Il faut souligner qu’ il
s’ agit donc d’états de choses qui sont réels ou du moins présentés comme
tels. Si les phrases génériques semblent porter sur l’ avenir, ce n’est que par
implication. Cet ‘avenir impliqué’ est bien sûr un autre avenir que celui qui
est envisagé dans les cas traités plus haut, où il s’ agit d’ un avenir défini par
référence à un moment d’énonciation concret, et qui peut être l’ objet de spé-
culations diverses. A priori il me semble donc que, pour l’ analyse des modes
dans les descriptions, cela n’a pas de sens d’opérer avec une notion comme
‘l’attente du locuteur’. Les positions de K-G et d’ Humbert sur les condition-
nelles génériques – voir les citations suivantes – me paraissent donc intenables
d’un point de vue méthodologique: ‘(ἐάν + subj.) steht… verallgemeinernd von
Bedingungen, deren Verwirklichung jederzeit erwartet werden kann : jedesmal
wenn’ (K-G 2, 473); ‘La protase suppose… un fait général, dont la réalisation
permanente est l’objet d’une attente’ (Humbert, p. 220). On verra d’ ailleurs
que les notions de ‘jederzeit’ et de ‘réalisation permanente’ ne s’ appliquent nul-
lement aux conditionnelles. Deuxième point, qui est lié au point précédent:
dans les descriptions on ne trouve que le subjonctif, il n’y pas de concurrence
d’autres modes. En d’autres termes, dans ce type de discours le subjonctif
apparaît automatiquement. Il n’y a donc aucune raison d’ attribuer au sub-
jonctif la même valeur sémantique que dans son emploi futurel. Je voudrais
encore aller plus loin et dire que le subjonctif n’ a aucune valeur spécifique
(sauf, bien entendu, celle du ‘non-passé’). On pourrait toutefois m’objecter que
le subjonctif exprime la répétition. J’en doute; la notion de ‘répétition’ est une
caractéristique générale du contexte, qui le plus souvent y est établie d’ emblée
par un présent générique. Ainsi, pour prendre un cas typique, le présent ποιεῦσι
δὲ ὧδε dans (13) réfère à un état de choses répété.
 248 chapter 12

(13) ποιεῦσι δὲ ὧδε (scil. une sorte d’huile)· παρὰ τὰ χείλεα τῶν τε ποταμῶν και
τῶν λιμνέων σπείρουσι τὰ σιλλικύπρια ταῦτα, … ταῦτα… καρπὸν φέρει πολλὸν
μέν, δυσώδεα δέ· τοῦτον ἐπεὰν συλλέξωνται, οἱ μὲν κόψαντες ἀπιποῦσι, οἱ δὲ καὶ
φρύξαντες ἀπέψουσι καὶ τὸ ἀπορρέον ἀπ᾽ αὐτοῦ συγκομίζονται (Hdt. 2.94.1–
2)

… la préparent de la façon suivante. Sur les bords des cours d’ eau et des
marais, ils sèment ces ricins… cette plante produit des fruits abondants
mais qui sentent mauvais; les Égyptiens les ramassent, les concassent et
les pressent, ou bien les font bouillir après les avoir torréfiés, et recueillent
ce qui en découle (trad. Legrand)

Du coup, tout ce qui est dit après le prédicat générique ποιεῦσι, que l’ on pour-
rait appeler un prédicat hyperonyme, doit être pris comme faisant partie de
l’ exécution de cet état de choses, y compris tous les états de choses enchâssés.
Ceci vaut non seulement pour les subordonnées à verbe fini, comme ἐπεάν συλ-
λέξωνται, mais également pour les participes, p. ex. le participe aoriste κόψαν-
τες. Or il serait absurde de dire que ces participes expriment la répétition.
C’est le contexte qui nous invite à prendre les participes comme référant à
un état de choses répété. En fait les participes aoristes n’ont qu’ une valeur
sémantique, à savoir qu’ils expriment l’antériorité. En appliquant ce raison-
nement aux subordonnées à verbe fini, on pourrait dire fort bien qu’ elles aussi
n’expriment que l’antériorité, ou, le cas échéant, la simultanéité ou la posté-
riorité.
Reste à savoir sous quel rapport les conditionnelles se distinguent des tem-
porelles dans ce type de discours. Tout d’abord on constate que ces subordon-
140 nées ne | diffèrent que par la conjonction. La valeur de répétition – s’ il y en a
une – est bien sûr présente dans l’une et l’autre subordonnée. Or à mon avis la
différence est que la conditionnelle exprime que l’ état de choses en question ne
se réalise pas en toute occasion possible, et donc seulement de temps à autre,
pendant la durée d’un état de choses hyperonyme. La temporelle par contre
exprime que l’état de choses se réalise en toute occasion possible pendant la
durée d’un état de choses hyperonyme. Je voudrais illustrer cette remarque par
une analyse des deux exemples déjà mentionnés sous (11) et (12). (11) fait partie
d’un passage où Hérodote décrit comment on voyage sur le Nil. À un certain
endroit difficile, nous dit-il, le voyageur doit quitter son bateau, et attacher des
câbles aux deux côtés du bateau pour pouvoir le haler. Dans la phrase citée
sous (11), ἤν indique qu’il s’agit d’une situation disjonctive. Comme dans les
autres emplois de la conjonction conditionnelle, ἤν exprime donc une possi-
bilité. Or, dans le contexte d’une description, ‘possibilité’ implique ‘réalisation
sur les emplois de εαν et επεαν 249

occasionnelle’. L’on pourrait paraphraser la conditionnelle par ‘il arrive que les
câbles se rompent; dans ce cas le bateau est emporté par le courant’. Noter qu’ il
n’y a aucune indication sur la fréquence des ruptures, ce qui en principe aurait
été possible, puisque Hérodote pourrait avoir ajouté une expression comme ‘ce
qui arrive parfois/régulièrement/souvent’. Il y a toutefois une chose qui est cer-
taine, c’est que les câbles ne se rompent pas tout le temps (ce qui, pris en soi,
aurait été un peu bizarre dans ce contexte). Ceci aurait été exprimé non pas par
une conditionnelle mais par une temporelle, comme dans (12). Ici, ἐπεάν fonc-
tionne de nouveau comme un ‘déclencheur de présupposition’, comparer (6),
mais cette fois-ci dans le plein sens du mot, puisqu’ il s’ agit d’ un état de choses
habituel et réel, et non pas d’un état de choses futur. La conjonction temporelle
indique donc que l’état de choses dénoté par ἀπίκωνται se produit nécessaire-
ment, c’est-à-dire en toute occasion possible pendant l’ état de choses dénoté
par πορευόμενα et πλέοντες. En d’autres termes: chaque voyage des bateaux se
termine par une arrivée. La présence de ἤν, par contre, aurait indiqué qu’ il y a
des voyages qui ne se terminent point par une arrivée.
J’ajoute que l’on pourrait analyser les différences signalées de deux façons
plus techniques, en opérant dans un cadre logique. D’ abord, on pourrait dire
que ἤν exprime que la réalisation de l’état de choses de la subordonnée est un
trait contingent de l’état de choses dénoté par le prédicat hyperonyme, tandis
que ἐπεάν exprime qu’il s’agit d’un trait nécessaire de ce prédicat. La notion de
contingence est employée par van der Auwera (1983: 301) dans son analyse de
la phrase (14):

(14) It can be cold in Berlin

qui contient l’implication ‘il fait parfois froid à Berlin’. ‘Le froid’ est donc un
trait contingent du climat berlinois. Van der Auwera définit la notion de contin-
gence comme ‘covers the middle ground between possibility and necessity’.
Pareillement, c’est un trait contingent d’un voyage sur le Nil que les câbles
des bateaux se rompent parfois6. Dans l’autre analyse, on pourrait opérer
avec la notion de quantification. Ainsi, ἤν implique ce qu’ on pourrait appe-
ler une ‘quantification non exhaustive’, ou bien une ‘quantification relative
non proportionnelle’, ou encore une ‘quantification partielle’, tandis que ἐπεάν
implique une ‘quantification exhaustive’ ou ‘universelle’. La notion de ‘quanti-
fication relative non proportionnelle’ est employée par Dik (1989: 153) dans son
analyse des cas comme: | 141

6 Il est intéressant de noter que Powell, dans son Lexicon s.v. ἤν 2, parle de ‘repeated contingen-
cies in the present’.
250 chapter 12

(15) I saw many/some/few birds

sur lequel il fait observer: ‘Each of these quantifiers presupposes some implicit
norm with which the number of members of the set is compared. This norm
cannot be absolutely fixed and is co-determined by extra-linguistic properties
of, and expectations about the referent set’. En appliquant cette analyse à notre
cas, on pourrait dire que la norme y est également implicite, le nombre des
voyages sur le Nil n’étant pas spécifié. Et une telle spécification est bien sûr
exclue, car c’est un des traits pertinents des descriptions d’ un état de choses
habituel que le nombre des répétitions est illimité. La notion de ‘quantification
exhaustive’ est employée dans un contexte un peu différent par Brown (1985:
139).
Les différences entre ἐάν et ἐπεάν dans les descriptions peuvent être illus-
trées par deux autres phénomènes. Comme ἐάν réfère à une situation disjonc-
tive, et que l’état de choses en question peut donc ne pas se réaliser, cet état de
choses peut être opposé à un autre état de choses, comme dans (16) :

(16) μεθυσκόμενοι δὲ ἐώθασι βουλεύεσθαι τὰ σπουδαιέστατα τῶν πρηγμάτων. τὸ δ᾽

ἂν ἅδῃ σφι βουλευομένοισι, τοῦτο τῇ ὑστεραίῃ νήφουσι προτιθεῖ ὁ στέγαρχος,
ἐν τοῦ ἂν ἐόντες βουλεύωνται· καὶ ἢν μὲν ἅδῃ καὶ νήφουσι, χρέωνται αὐτῷ, ἢν
δὲ μὴ ἅδῃ, μετιεῖσι· (Hdt. 1.133.4)

D’autre part ils ont la coutume de discuter en état d’ ivresse les affaires les
plus importantes. Ce qu’ils ont trouvé bon dans leur discussion leur est
soumis le lendemain, alors qu’ils sont à jeun, par le maître de la maison
où ils se trouvent pour discuter; s’ils le trouvent bon aussi étant à jeun, ils
s’y tiennent; s’ils ne le trouvent pas bon, ils y renoncent (trad. Legrand)

Ici, il est impossible de remplacer ἐάν par ἐπεάν, puisque cette conjonction ne
laisse pas de place à la non-réalisation, pour ainsi dire. Deuxièmement, tout
comme les conjonctions temporelles dans les récits, ἐπεάν peut reprendre un
état de choses mentionné avant, comme dans (17) :

(17) αἱ δὲ πλέονες ποιεῦσι ὧδε· ἐν τεμένεϊ Ἀφροδίτης κατέαται… πολλαὶ γυναῖκες.

(…) ἔνθα ἐπεὰν ἵζηται γυνή, οὐ πρότερον ἀπαλλάσσεται ἐς τὰ οἰκία ἤ τίς οἱ
ξείνων ἀργύριoν ἐμβαλὼν ἐς τὰ γούνατα μειχθῇ (…) ἐπεὰν δὲ μειχθῇ, ἀποσι-
ωσαμένη τῇ θεῷ ἀπαλλάσσεται ἐς τὰ οἰκία, … (Hdt. 1.199.3–4)

La plupart agissent de cette façon: elles se tiennent assises dans l’ enceinte

sacrée d’Aphrodite… Lorsqu’une femme a pris place en ce lieu, elle ne
sur les emplois de εαν et επεαν 251

retourne pas chez elle avant qu’un étranger lui ait jeté de l’ argent sur les
genoux et qu’elle se soit unie à lui [plutôt: ‘qu’il se soit uni à elle’]… Après
s’être unie à lui, quitte envers la déesse de ses obligations religieuses, elle
retourne chez elle (trad. Legrand)

Ici c’est le remplacement par ἐάν (ἤν) qui est exclu, car l’ information d’ une
conditionnelle serait contraire à celle donnée par κατέαται: il n’y a pas de
femmes qui ne s’assoient pas à l’endroit indiqué. De même, ἢν μειχθῇ au lieu de
ἐπεὰν μειχθῇ s’est pas possible non plus, car la phrase précédente présuppose
que la femme fait l’amour. Cette présupposition aurait été invalidée par ἤν. | 142
Je signale encore trois points d’un autre ordre. Premièrement, dans des
cas comme (11) la notion de ‘condition’ a un tout autre contenu que dans les
conditionnelles qui portent sur l’avenir. Après tout, la condition est bel et bien
réalisée, bien que non nécessairement. Pour ne pas brouiller les choses il vau-
drait donc mieux employer un autre terme, par exemple ‘temporelle indéfinie’,
comme l’a préconisé l’helléniste belge van Pottelbergh, dans une monographie
parue en 1939, qui malheureusement n’est guère connue. Autant que je sache,
van Pottelbergh est le seul à signaler qu’il y a une différence entre les deux
emplois de ἐάν. Deuxième point: jusqu’ici j’ai évité le terme d’ ‘hypothèse’ dans
mon analyse. En principe, l’analyse des conditionnelles portant sur l’ avenir
pourrait se fonder sur cette notion. Ainsi, l’on pourrait dire à propos du sub-
jonctif qu’il exprime que l’hypothèse a toute chance de se réaliser. Dans les
descriptions cette notion n’a toutefois aucune raison d’ être, puisqu’ il s’ agit
d’états de choses qui sont présentés comme réels. Dans (11), on n’a pas affaire
à un état de choses hypothétique, mais à une réalité : Hérodote présente la rup-
ture des câbles comme un fait. Troisièmement, il me semble que le si français
a la même valeur dans des descriptions. Comparer la traduction de Legrand de
(11) ‘s’il échappe à ses liens, il est emporté…’. À ce propos je fais observer que
les grammaires des langues modernes, tout comme celles du grec, ne font pas
état de cet emploi de la conditionnelle. | 143
 252 chapter 12

4 Conclusion

En guise de conclusion, je présente le schéma suivant :

type de subordonnée

conditionnelles: temporelles :
trait commun : décrivent une situation traits communs:
disjonctive : la condition peut ou non se réaliser – établissent un lien tem-
porel;
– présupposent en prin-
cipe l’état de choses de la
subordonnée

type de discours εἰ + ind. fut. εἰ + opt. ἐάν + subj. ἐπεάν, ὅταν + subj.

discours cond. neutre ; la réalisation la réalisation réalisation de l’état de

(in)direct : le locuteur ne de la condition de la condition choses quasi-certaine (effet
la subord. porte s’ exprime pas est simplement relève d’une combiné de la valeur semi-
sur l’avenir sur la probabilité possible forte possibilité présuppositionnelle de la
de réalisation de conjonction et de la valeur
la condition de ‘forte possibilité’ du sub-
jonctif.
N.B. Puisqu’il s’agit de
l’avenir, la conjonction ne
peut avoir ici qu’une valeur
semi-présuppositionnelle)

descriptions : manque manque réalisation pos- l’état de choses se produit

l’état de choses sible implique: à toute occasion possible
de la subordon- l’ état de choses pendant la durée de l’état
née fait partie de la subordon- de choses hyperonyme
d’un état de née se réalise (‘quantification exhaus-
choses hyper- parfois pen- tive’)
onyme dant la durée de
l’ état de choses
hyperonyme
(‘quantification
non exhaustive’)
sur les emplois de εαν et επεαν 253

Références

Amigues, S., Les subordonnées finales par ὅπως (Paris 1977).

Auwera, J. van der, ‘Conditionals and Antecedent Possibilities’, Journal of Pragmatics 7
(1983) 297–309.
Brown, D.R., ‘Term Operators’, in A.M. Bolkestein et al. (eds), Predicates and Terms in
Functional Grammar (Dordrecht 1985) 127–145.
Dik, S.C., The Theory of Functional Grammar, Part I: The Structure of the Clause (Dor-
drecht 1989).
Goodwin, W.W., Syntax of the Moods and Tenses of the Greek Verb (London 1889).
Humbert, J. Syntaxe grecque, 3e éd. (Paris 1960).
Levinson, S.C., Pragmatics (Cambridge 1983).
Pottelbergh, R. van, Over de geschiedenis en de betekenis van den εἰ-zin in het Grieksch
(Gent 1939).
Verdenius, W.J., ‘Notes on the Prologue of Euripides’ Bacchae’. Mnemosyne 33 (1980) 1–
16.
 chapter 13

The Syntax and Semantics of Expressions of Sorrow

and Related Concepts in Homer

Abstract

This chapter presents a full-scale inquiry of the fine-grained semantic differences

between Homeric expressions for ‘sorrow’ and related concepts, viz. ἄλγος, ἄχος, κακόν,
κῆδος, ὀϊζύς, πένθος, πῆμα and πόνος. It closely considers the syntactic behaviour of these
words (especially the kinds of verbs with which they occur as subject or object), as well
as their occurrence in the singular and/or plural. The conclusions are presented in the
form of a small Homeric ‘lexicon of sorrow’.

Introduction

By way of introduction to the present paper I shall briefly summarize the main
findings of Rijksbaron (1991).* In that article I addressed the question in what
sense ἄλγεα, in the expression ἄλγε᾽ ἔχειν, could be considered a ‘possession’ of
a human being, as suggested by the verb ἔχειν, and how human beings ‘acquire’
these possessions. I argued that the ἄλγεα sometimes originate with the gods, in
which case the verb used is διδόναι, and sometimes with human beings; in the
latter case the verb used is τιθέναι: unlike the gods, human beings do not have
ἄλγεα at their disposal and as a consequence they cannot give them. The con-
struction τιθέναι ἄλγεα τινι should semantically be analysed as ‘to create ἄλγεα
for somebody’, i.e. ‘cause somebody to receive/have/feel ἄλγεα’.1 Some typical
examples of these uses are:

(1) ἥ ῥ᾽ Ὀδυσῆ᾽ ἐλέησεν ἀλώμενον, ἄλγε᾽ ἔχοντα (Od. 5.336)

She felt pity for Odysseus, as he wandered and had sorrows

* I am indebted to the members of the ‘Amsterdamse Hellenistenclub’ for their valuable com-
ments.
1 Besides τιθέναι the verb τεύχειν is used in a similar construction, both with a god and a human
being as agent, cp. Rijksbaron (1991: 186). I am not taking into account this verb in the present
paper.

© koninklijke brill nv, leiden, 2019 | doi:10.1163/9789004386129_015

the syntax and semantics of expressions of sorrow 255

(2) ἀλλά μοι αἰγίοχος Κρονίδης Ζεὺς ἄλγε᾽ ἔδωκεν (Il. 2.375)

Zeus of the aegis, son of Kronos, has given me bitterness

(3) μάλιστα δ᾽ ἐμοὶ περὶ πάντων ἄλγε᾽ ἔθηκε (Il. 22.422; subject: Achilles)

He has given us most sorrow, beyond all others

Human beings ‘acquire’ and ‘possess’ ἄλγεα either because of their own (mis)-
conduct, as appears from (4), or as an inevitable constitutive element of human
life, in accordance with the principle formulated in (5), cp. also μόρος in (4):

(4) ὢ πόποι, οἷον δή νυ θεοὺς βροτοὶ αἰτιόωνται.

ἐξ ἡμέων γάρ φασι κάκ᾽ ἔμμεναι, οἱ δὲ καὶ αὐτοὶ
σφῇσιν ἀτασθαλίῃσιν ὑπὲρ μόρον ἄλγε᾽ ἔχουσιν. (Od. 1.32–34, Zeus speaking)

Humph! How mortals now blame gods, for they say that evils are from us.
Yet they themselves have woes beyond their lot by their own recklessness

(5) ὡς γὰρ ἐπεκλώσαντο θεοὶ δειλοῖσι βροτοῖσι

ζώειν ἀχνυμένοις· αὐτοὶ δέ τ᾽ ἀκηδέες εἰσί.
δοιοὶ γάρ τε πίθοι κατακείαται ἐν Διὸς οὔδει
δώρων οἷα δίδωσι κακῶν, ἕτερος δὲ ἑάων. (Il. 24.525–528, Achilles speaking)

Such is the way the gods spun life for unfortunate mortals, that we live
in unhappiness, but the gods themselves have no sorrows. There are two
urns that stand on the door-sill of Zeus. They are unlike for the gifts they
bestow: an urn of evils, an urn of blessings

As to the relevance of the notion ‘possession’ in these expressions, I argued that

this notion should probably be taken metaphorically. Thus, human beings are
not capable of manipulating ἄλγεα, e.g. to get rid of them, unlike real posses-
sions. Ἄλγεα do not form part, then, of the ‘belongings’ of a human being; they
are rather phenomena that burden him.
Besides the expression ἄλγε᾽ ἔχειν there exists another possessive expression,
viz. of the singular ἄλγος as subject of ἔσσεται or of a verbless clause. In this use
we are dealing with a single or a collective ἄλγος, which is caused by an event
rather than by a god or a human being.2 Compare:

2 Presumably, the difference between the ‘possession’ of these phenomena as expressed by

256 chapter 13

(6) εἰ δ᾽ ἤδη τεθνᾶσι καὶ εἰν Ἀΐδαο δόμοισιν,

ἄλγος ἐμῷ θυμῷ καὶ μητέρι τοὶ τεκόμεσθα (Il. 22.52–53; Priam speaking, the
event being the possible death of his sons)

But if they are dead already and gone down to the house of Hades, it is
sorrow to our hearts, who bore them, myself and their mother

Furthermore, it could be shown that the claim of many dictionaries that ἄλγος/
ἄλγεα also occurs as the subject of ἔχειν is unwarranted. Finally, I argued that
the primary meaning of both ἄλγος and ἄλγεα is not ‘pain of body or mind’,
‘souffrance’, ‘Schmerz’, as suggested by the standard dictionaries, but rather
‘affliction, ordeal’.
Below I shall investigate the semantic and syntactic behaviour of a num-
ber of words that belong to the same semantic sphere: ἄχος/ἄχεα, κῆδος/κήδεα,
ὀϊζύς, πένθος, πόνος, πῆμα/πήματα, κακόν/κακά.3 The format of my treatment is
the same as that with ἄλγος/ἄλγεα. I start, therefore, with a survey of expres-
sions involving ἔχειν, διδόναι and τιθέναι, and related verbs; this will be followed
by a number of particulars of the affect-words involved. Then follows a discus-
sion of the constructions where the singular of these words is found, notably
as the subject of εἶναι, of verbs of seizing and of verbs of reaching.4 As in the
case of ἄλγος/ἄλγεα I believe that in principle a distinction should be made
between the uses of these words as object and as subject, and between the plu-
ral (if present) and the singular. The ultimate goal of taking into account all
these factors is to arrive at more precise meanings than those given in the stan-
dard dictionaries; some suggestions to that effect are given at the end of my
paper.5

ἔχειν and as expressed by εἶναι + dative, is that ἔχειν presupposes the activity of an explicit
‘giver’ (διδόναι) or ‘causer’ (τιθέναι), while εἶναι + dative does not. This means that, in (6), a con-
struction like ἔχομεν ἄλγος/ἄλγεα would seem to be excluded. Regrettably, the monograph of
Kahn (1973) is of no help here, because his treatment is flawed by the idea that all possessive
constructions involving εἶναι and the dative are ‘the result of a transformational derivation’
(p. 270). Kahn does not discuss possible (dis)similarities with ἔχειν.
3 In this order, because πῆμα/πήματα and κακόν/κακά must be contrasted with the other ones,
as will become clear presently.
4 The discussion of verbs of seizing and verbs of reaching includes ἄλγος, since these verb types
were not discussed in Rijksbaron (1991).
5 The words to be discussed are not confined, of course, to the expressions considered in this
paper. Thus, ἄλγεα may be the object of μογήσας, and ἄλγος the subject of μέλει, κήδεα the
subject of ἐφῆπται etc. In other words, the present paper is by no means exhaustive.
the syntax and semantics of expressions of sorrow 257

1 Ἄχεα etc. as Object

1.1 Constructions with ἔχειν

(7) ἔχω δ᾽ ἄχε᾽ ἄκριτα θυμῷ (Il. 3.412, Helen speaking; 24.91, Thetis speaking)

and my heart even now is confused with sorrows

(8) ἀλλ᾽ ἕνεκ᾽ οὐλομένης γαστρὸς κακὰ κήδε᾽ ἔχουσιν

ἀνέρες, ὅν τιν᾽ ἵκηται ἄλη καὶ πῆμα καὶ ἄλγος (Od. 15.344–345, Odysseus
addressing Eumaios)6

yet for their accursed belly’s sake men have evil troubles, anyone to whom
wandering and pain and sorrow come

(9) ἤλυθες Οὔλυμπον δὲ θεὰ Θέτι κηδομένη περ,

πένθος ἄλαστον ἔχουσα μετὰ φρεσίν· (Il. 24.105, Zeus speaking)7

You have come to Olympos, divine Thetis, for all your sorrow, with an
unforgotten grief in your heart

(10) ἔχων ἀλάλημαι ὀϊζύν (Od. 11.167, Odysseus speaking; cp. ὀχεῖν ὀϊζύν, Od. 7.211)

I’ve wandered always and had sorrow

(11) τοῖον γὰρ ἔχον πόνον ἀμφιέποντες (Il. 5.667; subject: the Trojans)

such hard work did they have attending him

(12) οὐ γάρ πω σχεδὸν ἦλθον Ἀχαιΐδος, οὐδέ πω ἁμῆς

γῆς ἐπέβην, ἀλλ᾽ αἰὲν ἔχω κακά. (Od. 11.481–482, Odysseus speaking)

For I never came near Achaean land, or ever set foot on my own, but
always had sorrow

6 Line 345 is omitted by Von der Mühll and Heubeck-Hoekstra.

7 Other instances are Od. 7.218, 219, 10.376, 18.324, 24.233, in all cases except 10.376 with a loca-
tive modifier of the form (ἐνὶ, μετὰ) φρεσί.
258 chapter 13

From this survey it appears that as object of ἔχειν, besides ἄλγεα, are found:
ἄχεα, κήδεα, ὀϊζύς, πένθος, πόνος, κακά. The exclusive use of the plural, of those
words that are marked for number, indicates that habitual states of affairs are
involved; apparently, one cannot ‘have’ a single ἄλγος etc.8
Observe also that of the words under discussion only πῆμα and πήματα are
never construed with ἔχειν, which suggests that they cannot be ‘possessed’.
As for κήδε᾽ ἔχειν in example (8)—the only example with ἔχειν—this is
slightly peculiar, since these κήδεα are caused by hunger; compare Od. 17.287
(note 9), from which we learn that the stomach ‘gives’ κακά. Also, unlike ἄλγεα,
which cannot be ended by a human being but only by a god, we must assume
that such κήδεα can be dispelled rather simply, viz. by eating.

1.2 Constructions with διδόναι and Other Verbs of Giving

(13) ἀργαλέον, βασίλεια, διηνεκέως ἀγορεῦσαι

κήδε᾽, ἐπεί μοι πολλὰ δόσαν θεοὶ οὐρανίωνες (Od. 7.241–242, Odysseus speak-
ing. Note that ἄλγε’ was metrically acceptable)

It’s hard, my queen, to recount at length my troubles, since the heavenly

gods have given me many

(14) ὄφρ᾽ ἱλασόμεσθα ἄνακτα,

ὃς νῦν Ἀργείοισι πολύστονα κήδε᾽ ἐφῆκεν. (Il. 1.444–445, Odysseus speaking;
ἄνακτα = Apollo)

that we may propitiate the lord who has heaped unhappiness and tears
on the Argives

(15) ἀλλὰ θεοὶ δυόωσι πολυπλάγκτους ἀνθρώπους,

ὁππότε καὶ βασιλεῦσιν ἐπικλώσωνται ὀϊζύν (Od. 20.195–196, Philoitios speak-
ing about Odysseus)

But the gods plunge into misery far-wandering men, when, even for kings,
they spin a thread of woe

8 Πόνος occurs a few times in the plural (Il. 10.89, 245, 279, Od. 13.301), and may be used, then as
a count noun, but is essentially a mass noun, like ὀϊζύς and πένθος.
the syntax and semantics of expressions of sorrow 259

(16) θεοὶ δ᾽ ὤπαζον ὀϊζύν (Od. 23.210, Penelope addressing Odysseus)

The gods gave us misery

(17) αὐτὰρ ἐμοὶ καὶ πένθος ἀμέτρητον πόρε δαίμων (Od. 19.512, Penelope speak-
ing)

But a divinity gave even me immeasurable sadness

(18) ἀτὰρ θεὸς ἄλλοτε ἄλλῳ

Ζεὺς ἀγαθόν τε κακόν τε διδοῖ (Od. 4.236–237, Helen speaking)9

god Zeus at one time or another gives both good and bad

From this survey it appears that, besides ἄλγεα, the following words occur as the
object of διδόναι etc., with a god as subject: κήδεα, ὀϊζύς, πένθος, κακόν. Presum-
ably, then, ἄχεα, πόνος and πῆμα/πήματα are not given by the gods. As for ἄχεα,
I should immediately add that the plural nominative/accusative occurs only
three times, as against 55 instances of the singular ἄχος. Ἄχεα is only found as
the object of ἔχειν (twice, cp. ex. (7)) and once as subject of ἔσται (Il. 6.413, spo-
ken by Andromache, referring to herself). These possessive constructions are
rather remarkable, because, as we will see later, sg. ἄχος usually is a ‘schmer-
zlicher Schreck, selten anhaltende Empfindung’, in the words of the Lexikon
des frühgriechischen Epos. The scantiness of the material makes it impossible
to ascertain where the ἄχεα come from.10
Note that the violent form of giving expressed in (14) obviously refers to the
sending of Apollo’s arrows, mentioned at lines 48 ff.
Example (18) may be considered a variant of the principle formulated in Il.
24.525–528, quoted above as (5).

9 The plural κακά, too, occurs as object of διδόναι, however not with a god but with the stom-
ach as subject: γαστέρα … | … ἣ πολλὰ κάκ᾽ ἀνθρώποισι δίδωσι (Od. 17.286–287: ‘a belly, a
damned thing that gives men many evils’, Odysseus speaking, cp. also 17.472–473).
10 A unique expression is found at Od. 19.167–168: ἦ μέν μ᾽ ἀχέεσσί γε δώσεις | πλείοσιν ἢ ἔχο-
μαι (Odysseus addressing Penelope: ‘Yes, you give me to more sorrows than I’m held by’).
Here, both the subject and the object of διδόναι are human beings. Also, the ἄχεα are repre-
sented as active forces, which have Odysseus in their grip. Although sg. ἄχος often occurs
as the subject of verbs of violence (see below), the active counterpart of ἀχέεσσι ἔχομαι is
not attested.
260 chapter 13

1.3 Constructions with τιθέναι

(19) αὐτὰρ ὁ διογενὴς Ὀδυσεὺς ὅσα κήδε᾽ ἔθηκεν

ἀνθρώποις ὅσα τ᾽ αὐτὸς ὀϊζύσας ἐμόγησε,
πάντ᾽ ἔλεγ᾽ (Od. 23.306–308)

Then Zeus-born Odysseus, all the troubles he’d caused for men, and all
he’d suffered in his misery. He told it all.

(20) ἀρρητὸν δὲ τοκεῦσι γόον καὶ πένθος ἔθηκας (Il. 17.37, subject Menelaos;
Patroklos speaking)

and left to his parents the curse of lamentation and sorrow

(21) ὡς δ᾽ ὅτε καπνὸς ἰὼν εἰς οὐρανὸν εὐρὺν ἵκηται

ἄστεος αἰθομένοιο, θεῶν δέ ἑ μῆνις ἀνῆκε,
πᾶσι δ᾽ ἔθηκε πόνον, πολλοῖσι δὲ κήδε᾽ ἐφῆκεν,
ὣς Ἀχιλεὺς Τρώεσσι πόνον καὶ κήδε᾽ ἔθηκεν (Il. 21.522–525; on the subject of
ἔθηκε at 524 see below)11

and as when smoke ascending goes up into the wide sky from a burning
city, with the anger of the gods let loose upon it which inflicted labour
upon them all, and sorrow on many, so Achilleus inflicted labour and sor-
row upon the Trojans.

(22) νῆας …, αἳ δεῦρο θεῶν ἀέκητι μολοῦσαι

ἡμῖν πήματα πολλὰ θέσαν (Il. 15.720–721, Hektor speaking; only example of
τιθέναι with πήματα)12

the ships that came here in spite of the gods’ will and have visited much
pain on us

11 At Il. 17.157–158 we find οἳ περὶ πάτρης | ἀνδράσι δυσμενέεσσι πόνον καὶ δῆριν ἔθεντο (‘men
who, for the sake of their country, have made the hard hateful work come between them
and their enemies’), a clear example of the indirect-reflexive meaning of the middle voice
(‘to their own profit’).
12 An example like Od. 11.555 τὰ δὲ πῆμα θεοὶ θέσαν Ἀργείοισι (‘the gods gave this as a misery
to the Argives’) is, of course, different, for here πῆμα is a predicative complement.
the syntax and semantics of expressions of sorrow 261

The object of τιθέναι, with a human being as subject is, then (besides ἄλγεα):
κήδεα, πένθος, πόνος, and also πήματα, on the assumption that ‘the ships’ is used
metonymically for ‘the crew of the ships’. Ἄχεα (but see above), ὀϊζύς, πῆμα13
and κακόν14 are absent from this list.
Example (21) is a very controversial passage, the main question being who or
what is the subject of ἔθηκε (and ἐφῆκεν) at 524. Leaf hesitates between καπνός
and ‘the conflagration’, or more generally ‘the state of things’, but in the end he
inclines to omit the line, following Bothe. Ameis-Hentze think the subject may
be ‘the fire’, to be supplied from καπνός, while Richardson, in the new large-
scale commentary on the Iliad, is completely silent on this point. Be that as it
may, with all options the construction is unparallelled: in view of the construc-
tions presented in this section one would expect the subject to be a human
being or a concrete entity like ‘ships’, used metonymically. Observe that for the
third alternative, viz. of taking θεῶν μῆνις as subject of ἔθηκε, there is no parallel
either. If the line is retained perhaps ἄστυ αἰθόμενον should be taken as subject:
‘the burning of the city’, with αἰθόμενον as a so-called ‘dominant’ participle.15

1.4 Constructions with πῆμα/πήματα as Object

Above we saw that πῆμα/πήματα are not given by the gods. This does not mean,
however, that they cannot originate with them, but this is expressed by an alto-
gether different construction, involving verbs of bringing. Compare:

(23) ἤδη μὲν σὲ καὶ αὐτὸν ὀΐομαι εἰσορόωντα

γιγνώσκειν ὅτι πῆμα θεὸς Δαναοῖσι κυλίνδει (Il. 17.687–688, Menelaos speak-
ing, referring to the attack of the Trojans, following Patroklos’ death)

You can see for yourself, I think, already, from watching, how the god is
wheeling disaster against the Danaans

(24) πῆμα also object of ἤγαγον (Il. 24.547; subject: the gods), προσήγαγε (Od.
17.446; subject: δαίμων), φέρε (Od. 12.231; subject: Skylla)

13 At Il. 11.413 we find: ἔλσαν δ᾽ ἐν μέσσοισι, μετὰ σφίσι πῆμα τιθέντες (‘(they) penned him in
their midst, but made thereby a wound in their ranks’; subject: the Trojans, object of ἔλσαν:
Odysseus), where πῆμα is either the object of μετὰ … τιθέντες, or a predicative complement
with Odysseus; on both analyses the dative is governed by μετά, not by τιθέντες.
14 In ξυνὸν δὲ κακὸν πολέεσσι τιθεῖσι (Il. 16.262: ‘they do something that hurts many people’) the
subject is: the wasps. There are, then, no examples of τιθέναι κακόν/κακά τινι, with a human
being as subject-agent. Instead Homeric Greek uses ῥέζειν κακόν/κακά τινα (Il. 2.195, 3.351,
354 etc.).
15 This analysis may find some support in the comparison of Achilles with a fire at Il. 21.12ff.
262 chapter 13

From these examples it may be inferred that πῆμα often refers to something
concrete, cp. the presence of εἰσορόωντα in (23).16 This is confirmed by two other
constructions of πῆμα, compare:

(25) εὕροι δ᾽ ἐν πήματα οἴκῳ (Od. 9.535, the Cyclops speaking; subject: Odys-
seus)

and may he find trouble in his house

(26) αὐτὰρ ἐμοὶ Ζεὺς αὐτός, ἔχοντί περ ἄλγεα θυμῷ,

ἱστὸν …
ἐν χείρεσσιν ἔθηκεν, ὅπως ἔτι πῆμα φύγοιμι. (Od. 14.310–312, Odysseus speak-
ing)

But Zeus himself, though I had sorrows in my heart, put the mast … in my
hands, so I’d yet escape misery

Πῆμα is, then something that may be found, cp. (25),17 and fled (26).18 Ἄλγεα
etc. are never the object of verbs like εὑρεῖν.
On the other hand, there are a number of instances where to all appearances
πήματα comes very close to ἄλγεα, especially in connection with πάσχειν, as in
example (27):

(27) δήμῳ ἔνι Τρώων, ὅθι πάσχετε πήματ᾽ Ἀχαιοί (Od. 3.100; ten other instances
with a form of πάσχειν)

in the Trojan kingdom, where you Achaeans suffered sorrows

(28) μηδέ τι μεσσηγύς γε κακὸν καὶ πῆμα πάθῃσι (Od. 7.195, Alkinoos speaking;
subject: Odysseus)

and not suffer woe or evil in the meantime

16 Also Il. 9.229: πῆμα … εἰσορόωντες (‘a trouble that we look on’). Note the presence, in Il.
24.547 πῆμα τόδ᾽ ἤγαγον οὐρανίωνες (‘now the Uranian gods brought us, an affliction upon
you’) of τόδε, referring to the situation (see below section 1.5) at hand. Cp. also Il. 11.347.
Ἄλγεα and the words discussed above are never modified by deictic pronouns.
17 Cp. Od. 11.115: δήεις δ᾽ ἐν πήματα οἴκῳ (‘In your house you’ll find misery’).
18 Mawet’s (1979: 100) remark: ‘Πῆμα figure encore comme complément de verbes transi-
tifs divers qui, par leur caractère isolé, ne permettent aucune conclusion’ (viz. about its
the syntax and semantics of expressions of sorrow 263

In fact, the frequent combination of πήματα with a form of πάσχειν19 is best

considered a formulaic variant of ἄλγεα πάσχειν (cp. Od. 3.220: δήμῳ ἔνι Τρώων,
ὅθι πάσχομεν ἄλγε᾽ Ἀχαιοί, as Mawet (1979: 98ff.) argues, conditioned by met-
rical factors). As for (28), which has an aorist verb form, this is the singular
(semelfactive) variant of (27). Cp. the singular ἄλγος in ἔνθα κεν … πάθεν ἄλγος:
Od. 14.32. This formulaic variation need not entail, of course, that ἄλγεα and
πήματα are synonymous; it could well be the case that a certain phenomenon
is sometimes described in terms of ἄλγεα (‘afflictions’) and sometimes in terms
of πήματα (‘disasters’, see end of paper).

1.5 Some Particulars

1.5.1 ἄχεα
From the fact that it is Thetis who uses ἄχεα at Il. 24.91 (above, ex. (7)) we may
perhaps infer that referentially the word does not differ from ἄλγεα, since the
latter is also used by her to express her feelings (at Il. 18.431). Possibly ἄχεα adds
a nuance of anger (for this meaning see below) to her distress.

1.5.2 κήδεα
Consider the following passages:

(29) εἴ γε μὲν εἰδείης σῇσι φρεσίν, ὅσσα τοι αἶσα

κήδε᾽ ἀναπλῆσαι, πρὶν πατρίδα γαῖαν ἱκέσθαι,
ἐνθάδε κ᾽ αὖθι μένων κτλ. (Od. 5.206–208, Calypso speaking to Odysseus;
note that ἄλγε’ was metrically acceptable)

If you only knew in your mind how many sorrows make up your destiny
before you reach your fatherland

(30) ἥ μ᾽ ἔφατ᾽ ἐν πόντῳ, πρὶν πατρίδα γαῖαν ἱκέσθαι,

ἄλγε᾽ ἀναπλήσειν· (Od. 5.301–302, Odysseus, addressing himself; note that
κήδε᾽ was metrically acceptable)

who said that on the sea, before I reached my fatherland, I’d have my fill
of sorrows

nature, examples being εὑρίσκειν and δήειν), shows that she does not realise that these
uses strongly point to πῆμα having a concrete nature.
19 It would seem that by folk etymology πῆμα, which probably is not connected with πάσχειν,
came to be considered a cognate object of that verb. Note that neither πάθος nor πάθημα
are attested in Homer.
264 chapter 13

Passage (29) is spoken by Calypso to Odysseus before he sails off, (30) by

Odysseus himself as he is in the middle of great difficulties. From the fact that
these utterances are identical but for the replacement of κήδεα by ἄλγεα, it
may be inferred that these words are referentially similar but conceptually dif-
ferent: ἄλγεα refers to phenomena that are actually afflicting a person, while
κήδεα rather refers to the same phenomena in abstracto. Mawet (1979: 356)
already observed: ‘… κῆδος (this should be: κήδεα) semble être une expres-
sion plus abstraite, plus intellectuelle de la souffrance qu’ ἄλγος [read: ἄλγεα]’.
Refining her observation, by making use of the semantic-ontological terms dis-
cussed by Lyons (1977: 442ff.), we could say that ἄλγεα are second-order entities,
while κήδεα are rather third-order entities. ‘Whereas second-order entities are
observable and … have a temporal duration, third-order entities are unobserv-
able and cannot be said to occur or to be located either in space or time. Third-
order entities are such that ‘true’, rather than ‘real’, is more naturally predicated
of them; they can be asserted or denied, remembered or forgotten’ (ibid. 445).
I think, then, that κήδεα, in (29), are the third-order variant of second-order
ἄλγεα, and, more specifically, that they are mental representations of ἄλγεα.20
Other examples of this use can be found in two passages from Iliad 18, and in a
passage from Odyssey 15. First, Iliad 18:

(31) κλῦτε κασίγνηται Νηρηΐδες, ὄφρ᾽ ἐῢ πᾶσαι

εἴδετ᾽ ἀκούουσαι ὅσ᾽ ἐμῷ ἔνι κήδεα θυμῷ. (Il. 18.52–53, Thetis speaking)

Hear me, Nereids, my sisters; so you may all know well all the sorrows that
are in my heart, when you hear of them from me

(32) Ἥφαιστ᾽, ἦ ἄρα δή τις, ὅσαι θεαί εἰσ᾽ ἐν Ὀλύμπῳ,

τοσσάδ᾽ ἐνὶ φρεσὶν ᾗσιν ἀνέσχετο κήδεα λυγρὰ
ὅσσ᾽ ἐμοὶ ἐκ πασέων Κρονίδης Ζεὺς ἄλγε᾽ ἔδωκεν; (Il. 18.429–431, Thetis
speaking)

Hephaistos, is there among all the goddesses on Olympos one who in her
heart has endured so many grim sorrows as the griefs Zeus, son of Kronos,
has given me beyond others?

20 Cp. Il. 18.8 μὴ δή μοι τελέσωσι θεοὶ κακὰ κήδεα θυμῷ (Achilles speaking: ‘May the gods not
accomplish vile sorrows upon the heart in me’). Achilles fears that the gods will fulfill his
κήδεα, i.e., one might say, will turn his κήδεα into ἄλγεα.
the syntax and semantics of expressions of sorrow 265

In (31), Thetis’ words are followed by a reference to the ‘future fact’ that she
will not see back her son, Achilles, alive, and by some lines indicating her uncer-
tainty about the reason of his staying out of the war. At that time she has not yet
seen Achilles nor spoken with him; at line 63 she announces that she is going
to find out what πένθος has reached him. Lines 18.429–431 ((32)), on the other
hand, are spoken after she has met Achilles (ll. 18.79ff.), and now her κήδεα have
become ἄλγεα. She now knows that Patroklos is dead, and that the time has
come for Achilles to go back into the war, thereby provoking his own death, cp.
lines 95–96. As for κήδεα in l. 430, this, too, might refer to afflictions in abstracto,
indicating that the goddesses are not actually afflicted at the time of Thetis’
speech, which is also suggested, of course, by ἀνέσχετο.
In Odyssey 15 Eumaios says:

(33) κήδεσιν ἀλλήλων τερπώμεθα λευγαλέοισι

μνωομένω· μετὰ γάρ τε καὶ ἄλγεσι τέρπεται ἀνήρ,
ὅς τις δὴ μάλα πολλὰ πάθῃ καὶ πόλλ᾽ ἐπαληθῇ. (Od. 15.399–401)

delight in the dismal troubles of each other as we recall them. For a man
can delight even in sorrows, whoever’s suffered much and wandered far.

Here, ἄλγεσι refers to the afflictions as they originally occurred. When they are
over, they may be the object of recollection, in the form of κήδεα.
The differences between ἄλγεα and κήδεα in terms of concrete, actual (sec-
ond-order) vs. abstract, representational (third-order) are also apparent from
the following features.
(a) Quite often κήδεα is the object of a verbum dicendi, cp. ἀγορεῦσαι in (13)
and ἔλεγε in (19); other instances are: Od. 5.5, 9.12, 11.369, 376, 382, 542,
14.47, 185, 197,21 and, with ἀκούειν, Il. 18.53. This occurs once with ἄλγεα
(Od. 15.487), but there ἄλγεα is not directly governed by λέγων. Ἄλγεα
are phenomena that are experienced, κήδεα are phenomena that can be
related.22

21 Mostly those of Odysseus or his comrades (at Od. 11.382), the exceptions being Od. 11.542,
where the κήδεα of ψυχαί are involved (‘Each ghost asks after his κήδεα, whatever is most
important to him’, Heubeck-Hoekstra), and 23.306–308, see above (19). Observe that the
ghosts may also have ἄλγεα, referring to afflictions that are present in the Hades there and
then, e.g. Tantalos (Od. 11.582–584) and Sisyphos (Od. 11.593–594).
22 Observe that, when Odysseus’ actual vicissitudes are concerned both the narrator and
Odysseus himself use ἄλγεα (examples (1) and (26)), while at Alkinoos’ court κήδεα is the
favourite word (Od. 9.12, 11.369, 376, 382; also νόστον πολυκηδέα at Od. 9.15 and 37).
266 chapter 13

(b) A further characteristic of κήδεα as object of a verbum dicendi is that the

word may be modified by ‘possessive’ genitives and possessive pronouns,
e.g.: λέγε κήδεα πόλλ᾽ Ὀδυσῆος (Od. 5.5), τὰ σ᾽ αὐτοῦ κήδε᾽ ἐνίσπες (Od.
14.185), λέγων ἐμὰ κήδεα θυμοῦ, (Od. 14.197). Interestingly, ἄλγεα is never
modified by such elements. Κήδεα, being third-order entities, may ‘belong
to’ a person, in an abstract way, i.e. as a characteristic of him outside ‘space
and time’ (Lyons, as quoted above), while ἄλγεα, as second-order entities,
rather ‘occur’ and do not ‘belong to’ a person. Compare, in English, the
unacceptability of *‘his disasters’ (second-order; ‘disasters’ occur, and do
not belong to somebody) as against ‘his sorrows’ (third-order; ‘sorrows’ do
not occur, and do belong to somebody).
(c) Unlike ἄλγεα, κήδεα is never the object of πάσχειν or παθεῖν: κήδεα are not
experienced. The word is found, to be sure, with πεπαθυίῃ (Od. 17.555), but
here the use of the perfect stem may be significant: Penelope is under the
influence of κήδεα, but not actually experiencing them.
(d) Again unlike ἄλγεα, κήδεα can be dispelled; cp. Lyons’ remark above about
the possiblity to forget third-order entities:

(34) ἀλλ᾽ ἄγε πείρησαι, σκέδασον δ᾽ ἀπὸ κήδεα θυμοῦ (Od. 8.149, Laodamas,
addressing Odysseus)

But come, make a try, and scatter troubles from your heart

1.5.3 ὀϊζύς
Ὀϊζύς is never located in an organ (θυμός, φρένες etc.) and is therefore not a
feeling. Rather it refers to circumstances that accompany a human being. This
appears from the use of ὤπαζον ‘gave as a companion’ in (16),23 and also from:

(35) γήθησε δέ μοι φίλον ἦτορ,

δυσμόρῳ· ἦ γὰρ μέλλον ἔτι ξυνέσεσθαι ὀϊζυῖ (Od. 7.269–270, Odysseus speak-
ing about his vicissitudes at sea)24

and my dear heart rejoiced, ill-fated me, for I would still meet with much
hardship.

23 Cp. e.g. Il. 13.416: οἱ ὤπασα πομπόν (‘I have sent him an escort’).
24 This feature is also relevant for ex. (16) above, where Penelope continues with (Od. 23.211–
212): οἳ νῶϊν ἀγάσαντο παρ᾽ ἀλλήλοισι μένοντε | ἥβης ταρπῆναι καὶ γήραος οὐδὸν ἱκέσθαι. (‘who
begrudged that we two stay beside each other, to enjoy our youth and reach the threshold
of old age.’). Penelope and Odysseus were not allowed to remain each other’s companion,
they both got ὀϊζύς instead.
the syntax and semantics of expressions of sorrow 267

As Mawet (1979: 190) notes, the word is especially connected with ‘roaming’,
cp. ἀλάλημαι in example (10) and πολυπλάγκτους in (15), also (35) and ἄλη in (36)
below. Mawet also remarks that the word ‘semble désigner le sort misérable de
l’ homme’, pointing to the association of ὠκύμορος with ὀϊζυρός at Il. 1.417; one
may also note the presence of δυσμόρῳ in (35). Nevertheless it is possible for a
human being to get rid of ὀϊζύς, by the intervention of another human being, cp.:

(36) μ᾽ ἔπαυσας ἄλης καὶ ὀϊζύος αἰνῆς (Od. 15.342, Odysseus, addressing Eu-
maios)25

you’ve stopped me from wandering and grim misery

Recall (end of section 1.3) that the origin of ὀϊζύς is never a human being.

1.5.4 πένθος
Πένθος is the word par excellence for sorrow, grief as a lasting feeling, as appears
e.g. from its being modified by the adjectives ἄλαστον (‘not to be forgotten’) and
ἀμέτρητον (‘immeasurable’), and often has the specialised meaning ‘mourn-
ing’.26 It may be connected with physical symptoms like weeping, cp. γόος in
Il. 17.37 and 24.741, and with bodily neglect, as in Od. 24.226ff., where Laertes’
condition is described; that his condition is caused by πένθος may be inferred
from the occurrence of the word at 231 and 233.
We are not told whether it is possible for a human being or a god to put an
end to πένθος; possibly it just wears off. Be that as it may, the feeling of πένθος
may temporarily be ousted by the demands of the stomach, as Odysseus grudg-
ingly acknowledges at Od. 7.218–219, in a heavily rhetorical passage, where he
uses not less than six different words to bring home to the Phaeacians the
seriousness of his plight: ὀϊζύς (211), ἄλγεα (212), κακά (213), κηδόμενος (215), τει-
ρόμενος (218), πένθος (218 and 219).

1.5.5 πόνος
As noted above, πόνος is never given by the gods. The gods can, to be sure,
increase (ὀφέλλειν, Il. 2.42) or stir up (ἐγείρειν, Il. 16.651) πόνος, but human beings
are, we may submit, very well capable on their own to make life full of it, and
do not need the gods to have πόνος.

25 And also by that of a god, Od. 4.35. Cp. also Od. 4.812 καί με κέλεαι παύσασθαι ὀϊζύος ἠδ᾽ ὀδυ-
νάων (‘and you bid me cease from my grief and the many sorrows’), spoken by Penelope
to Athena-Iphthime.
26 Mawet (1979: 259) puts too much emphasis on this latter meaning.
268 chapter 13

Some forms of πόνος can, like κήδεα and ὀϊζύς, be ended by a human being:

(37) ἐπεὶ παύσαντο πόνου (Il. 1.467; viz. of preparing a meal)

after they had finished the work

The πόνος of war, however, can only be ended by a god, cp.:

(38) ὅρμηνεν δ᾽ ἀνὰ θυμὸν ὅπως παύσειε πόνοιο

δῖον Ἀχιλλῆα (Il. 21.137; subject: the Skamandros)

and he pondered in his heart as to how he could stop the labour of bril-
liant Achilleus

Cp. also Il. 7.36, 15.235, 21.249.

1.5.6 πῆμα
Πῆμα has many features of a first-order entity, cp. again Lyons (1977: 443): ‘…
it is a characteristic of all first-order entities … that they are located, at any
point in time, in what is, psychologically at least, a three-dimensional space;
and that they are publicly observable’. Recall that a πῆμα can, in fact, be pointed
at (note 16), be found (ex. (25)) and be seen (ex. (23)). Interestingly, πῆμα is often
equated with a full-fledged first-order entity, viz. a person, cp. e.g. τόδε πῆμα, …
Ἕκτωρ (Il. 11.347), σὺ … πῆμα μέγιστον (Il. 22.288), Il. 6.282, 22.421.27
Not surprisingly, in view of the characteristics mentioned above, πῆμα is
never located in an organ (θυμός, φρένες etc.).

1.5.7 κακόν/κακά
Above we saw that κακά, like ἄλγεα, may be the object of ἔχειν and of διδόναι,
and κακόν, like κήδεα, the object both of διδόναι and of τιθέναι. Another feature
it shares with κήδεα is that it may be the object of a verbum dicendi, e.g. μυθήσα-
σθαι at Od. 7.213. On the other hand, κακόν also shares some features with πῆμα,
since it may be the object of a verb of finding or fleeing, cp.:

27 On the other hand, if πήματα, in πήματα πάσχειν, is more like ἄλγεα, the plural may
have second-order characteristics. Again, the presence of ἀρχή, in κυλίνδετο πήματος ἀρχή
(Od. 8.81) indicates that πῆμα, too, may be used in a more abstract way. Heubeck-West-
Hainsworth note: ‘a traditional metaphor’, quite mistakenly, for this expression is unique.
We do find more often πῆμα κυλίνδεται (Il. 11.347, Od. 2.163) or πῆμα κυλίσθη (Il. 17.99), but
here there may be no metaphor at all.
the syntax and semantics of expressions of sorrow 269

(39) μὴ ἴομεν, μή πού τις ἐπίσπαστον κακὸν εὕρῃ (Od. 24.462, Halitherses address-
ing the Ithacians)

Let’s not go, lest anyone by chance finds evil he’s brought on himself.

(40) οὐ γάρ τις νέμεσις φυγέειν κακόν (Il. 14.80)

There is no shame in running, even by night, from disaster

Also, like πῆμα, it is never located in an organ (θυμός, φρένες etc.).

All these features suggest that κακόν/κακά, being hyperonyms and there-
fore semantically rather vague (‘something bad, evil’), could serve, with some
restrictions (non-occurrence of * ἔχειν κακὰ θυμῷ etc.), as replacements of the
more specific terms.

2 The Same Words as Subject, but Mostly in the Singular

2.1 Of εἶναι and γενέσθαι, + Dative; Also of Verbless Sentences

For ἄλγος see above, ex. (6).

(41) ἀλλά μοι αἰνὸν ἄχος σέθεν ἔσσεται, ὦ Μενέλαε,

αἴ κε θάνῃς … (Il. 4.169, Agamemnon speaking. Also Il. 9.249, 16.55, once in
plural, 6.413. In a verbless clause 5.759, 20.293)

But I shall suffer a terrible grief for you, Menelaos, if you die

(42) Πατρόκλῳ δ᾽ ἄρ’ ἄχος γένετο φθιμένου ἑτάροιο,

ἴθυσεν δὲ διὰ προμάχων (Il. 16.581; cp. Il. 1.188, 12.392, 13.417, 14.458, 486,
16.508, Od. 11.208)28

And the sorrow took hold of Patroklos for his fallen companion. He
steered his way through the ranks of the front fighters

28 The imperfect is found once: καί σφιν ἄχος κατὰ θυμὸν ἐγίγνετο δερκομένοισι | Τρῶας (Il.
13.86–87: ‘discouragement of the heart came over them, as they watched’), probably
expressing a gradual process, simultaneously with the seeing of the Trojans.
270 chapter 13

(43) καί οἱ κατέλεξεν ἅπαντα

κήδε’, ὅσ’ ἀνθρώποισι πέλει τῶν ἄστυ ἁλώῃ (Il. 9.591–592, Phoinix speaking,
subject of κατέλεξεν the wife of Meleagros)

and rehearsed in their numbers before him all the sorrows that come to
men when their city is taken

(44) νῦν δ᾽ ἵνα καὶ σοὶ πένθος ἐνὶ φρεσὶ μυρίον εἴη
παιδὸς ἀποφθιμένοιο, τὸν οὐχ ὑποδέξεαι αὖτις (Il. 18.88–89, Achilles speak-
ing; σοί = Thetis; also Od. 17.477)

As it is, there must be on your heart a numberless sorrow for your son’s
death, since you can never again receive him

(45) οὔ θην οἴοισίν γε πόνος τ᾽ ἔσεται καὶ ὀϊζὺς

ἡμῖν (Il. 14.480–481, Akamas speaking. Only example of ὀϊζύς with εἶναι.
Πόνος also at Il. 22.488)29

we shall not be the only ones to be given hard work and sorrow

(46) νῦν δὲ δὴ ἐνθάδ’ ἐμοὶ κακὸν ἔσσεται· οὐ γὰρ ὀΐω

σὰς χεῖρας φεύξεσθαι … (Il. 21.92–93, Hektor, addressing Achilles. Also
plural in a similar construction: περὶ … κακὰ πάντοθεν ἔστη (Od. 14.270,
17.439))

the evil shall be mine in this place, since I do not think I shall escape your
hands

As subject of εἶναι etc. + dative occur: ἄλγος, ἄχος/ἄχεα, κήδεα, ὀϊζύς, πένθος,
πόνος, κακόν/κακά. In these expressions ἄχος/ἄχεα, κήδεα, and πένθος are used,
like ἄλγος (see Introduction), for phenomena that are caused by an event, rather
than by a god or a human being: the taking of a city ((43)), the (possible) death
of a warrior ((41) and (42)). The construction ἄχος (ἐ)γένετο is typically used
to express a sudden emotional reaction to an important event, e.g. the loss
of a comrade, as in (42), and is usually immediately followed by an action

29 In these two cases, πόνος probably is a second-order entity. It may also be a first-order
entity, e.g. in: ὣς τῶν μισγομένων γένετο ἰαχή τε πόνος τε (Il. 4.456: ‘such, from the coming
together of men, was the shock and the shouting’). Cp. ἔβη θεὸς ἂμ πόνον ἀνδρῶν (Il. 13.239:
‘(he) strode on, a god, through the mortals’ struggle’), etc.
the syntax and semantics of expressions of sorrow 271

in response to that event, see again (42).30 This type of ἄχος, one might say,
does not develop into a more permanent ἄχος, precisely because the person
‘seized’ by ἄχος immediately reacts. The more permanent form is expressed,
not surprisingly, by a form of εἶναι. Here, the presence of ἄχος does not prompt
a reaction.
There are no examples of πῆμα with εἶναι + dative.

2.2 Of Momentaneous Verb Forms of Seizing: λαβεῖν, ἑλεῖν

(47) βεβλήκει· τότε καί μιν ἀνήκεστον λάβεν ἄλγος (Il. 5.394; object: Hera)

struck (her) so that the pain he gave her could not be quieted.

(48) τὴν δ᾽ ἅμα χάρμα καὶ ἄλγος ἕλε φρένα (Od. 19.471; object Eurykleia, upon
recognizing Odysseus)

Pain and joy seized her mind at the same time

(49) δούπησεν δὲ πεσών· πυκινὸν δ’ ἄχος ἔλλαβ’ Ἀχαιούς,

ὡς ἔπεσ᾽ ἐσθλὸς ἀνήρ· (…)
(…)· οὐδ᾽ ἄρ᾽ Ἀχαιοὶ
ἀλκῆς ἐξελάθοντο, μένος δ᾽ ἰθὺς φέρον αὐτῶν. (Il. 16.599–602; cp. Il. 14.475.
Also with εἷλε (Il. 13.581), τύψε (Il. 19.125), βεβίηκεν (Il. 10.145, 16.22))

He fell, thunderously, and the closing sorrow came over the Achaians as
the great man went down … nor did the Achaians let go of their fighting
strength, but steered their fury straight at them.

(50) ὣς ἔφατο, Τρῶας δὲ κατὰ κρῆθεν λάβε πένθος

ἄσχετον, οὐκ ἐπιεικτόν
(…)
βὰν δ᾽ ἰθὺς Δαναῶν … (Il. 16.548–549, 552; after death of Sarpedon; only
example)

30 Cp. Mawet (1979: 300). A person may actively contribute to loosening the grip of ἄχος, itself
caused by the death of a comrade, by killing an opponent: ὀλίγον γε Μενοιτιάδαο θανόντος |
κῆρ ἄχεος μεθέηκα χερείονά περ καταπεφνών. (Il. 17.538–539: ‘I have put a little sorrow from
my heart for Patroklos death, although the man I killed was not great as he was’).
272 chapter 13

He spoke, and the Trojans were taken head to heel with a sorrow untake-
able, not to be endured (…) They went straight for the Danaans …

As subject of λαβεῖν, ἑλεῖν occur: ἄλγος, ἄχος, πένθος. These constructions ex-
press a single momentaneous feeling, as a reaction to some event. As in the
case of ἄχος (ἐ)γένετο, the construction ἄχος ἔλλαβε is typically used to express
a sudden emotional reaction to an important event, e.g. the loss of a comrade,
as in (49), and is usually immediately followed by an action in response to that
event, see again (49). Example (50) is unique, in that elsewhere the occurrence
of πένθος is not followed by an action. Usually, πένθος is used in situations where
nobody is acting at all.
There are no examples with κῆδος or ὀϊζύς: κῆδος expresses a stative (lasting),
not a momentaneous (i.e. violent) feeling (cp. below 2.3), while ὀϊζύς does not
express a feeling at all, but rather a more or less permanent characteristic of
man’s life; cp. the remarks made above in section 1.5.
There are no examples with πόνος, κακόν or πῆμα either: all three are, in fact,
only very rarely represented as active forces; for κακόν and πῆμα see below.31

2.3 With Verbs of Reaching32

(51) μάλιστα γὰρ ἄλγος ἱκάνει

θυμὸν ἐμόν (Il. 3.97–98, Menelaos speaking. Cp. Od. 2.41, without θυμόν;
Telemachos speaking)

Listen now to me also; since beyond all others this sorrow comes closest
to my heart

(52) ἥ μ᾽ ἐσάωσ᾽ ὅτε μ’ ἄλγος ἀφίκετο τῆλε πεσόντα (Il. 18.395, Hephaistos speak-
ing; acute physical pain)

She saved me when I suffered much at the time of my great fall

31 Od. 2.45 (… μοι κακὸν ἔμπεσεν οἴκῳ: ‘evil has befallen my house’) is a rare example of κακόν
as subject of a momentaneous verb form expressing a violent action.
32 It is a heavily debated question whether these expressions should be taken metaphorically
or involve personifications; in the first case ἄλγος etc. would operate as second-order enti-
ties, in the second case as (semi-) first-order entities. De Boel (1988: 89ff.) argues, convinc-
ingly, I think, that they are metaphors. (The same would seem to apply to the expressions
of section 2.2). This may be inferred e.g. from the fact that the expressions of this section
are never modified by source-expressions, indicating where the phenomenon concerned
came from. This points to these phenomena being, in fact, second-order entities. But a
detailed account of the problems involved falls outside the framework of this paper.
the syntax and semantics of expressions of sorrow 273

(53) ἐπεί μ’ ἄχος αἰνὸν ἱκάνει (Il. 19.307, Achilles speaking. Cp. 312: πυκινῶς ἀκα-
χήμενον. Also with ἐν … δῦνε (Il. 19.367), δύη, ἀμφεχύθη, ἀμφιβέβηκεν, κὰδ
… χύτο (Il. 20.282))33

since this strong sorrow has come upon me

(54) δύνασαι δὲ σὺ πάντοσ᾽ ἀκούειν

ἀνέρι κηδομένῳ, ὡς νῦν ἐμὲ κῆδος ἱκάνει.
ἕλκος μὲν γὰρ ἔχω τόδε κτλ.
(…)
ἔγχος δ᾽ οὐ δύναμαι σχεῖν ἔμπεδον, οὐδὲ μάχεσθαι (Il. 16.515–520, Glaukos
speaking. Cp. Il. 13.464, 15.245)

and wherever you are you can listen to a man in pain, as now this pain
has descended upon me. For see, I have this strong wound on me, I can-
not hold my spear up steady, I cannot … fight

(55) … ὀϊζύος, ἥ μιν (: Odysseus) ἱκάνει (Od. 5.289, Poseidon speaking; only
example)34

(to escape the great bond of) misery that’s come to him

(56) ὦ πόποι ἦ μέγα πένθος Ἀχαιΐδα γαῖαν ἱκάνει (Il. 1.254, Nestor speaking, about
the quarrel between Agamemnon and Achilles. Also with ἱκάνει: Il. 7.124,
Od. 6.169. With ἵκετο: Il. 1.362, 18.64, 73, 24.708, Od. 23.224 (in all cases not
necessarily a lasting feeling), Od. 1.342 (καθίκετο + ἄλαστον = lasting feel-
ing))

Oh, for shame. Great sorrow comes on the land of Achaia.

33 Of the phenomena discussed in sections 2.1–2.3, ἄχος is the most versatile when it comes
to manifesting itself to a person. It may also go away, a unique feature:
ἦ κέ μοι αἰνὸν ἀπὸ πραπίδων ἄχος ἔλθοι
(Il. 22.43, Priam speaking: ‘bitter sorrow (would) so be taken from my heart’)
As appears from the context the departure of Priam’s ἄχος, which is caused by Achilles
having killed many of his sons (lines 44–45), depends on Achilles’ death. Both the coming
and the going of ἄχος are, then, typically caused by an event.
34 Although πόνος is not construed with ἱκάνειν/ἱκέσθαι, it is once found with a related verb,
in: ἐπεί σε μάλιστα πόνος φρένας ἀμφιβέβηκεν (Il. 6.355, Helen addressing Hektor: ‘since it
is on your heart beyond all that the hard work has fallen’). This is a unique example, else-
where πόνος is not connected with an organ.
274 chapter 13

(57) = (8) ἀλλ᾽ ἕνεκ᾽ οὐλομένης γαστρὸς κακὰ κήδε᾽ ἔχουσιν

ἀνέρες, ὅν τιν᾽ ἵκηται ἄλη καὶ πῆμα καὶ ἄλγος (Od. 15.344–345, Odysseus
addressing Eumaios)

yet for their accursed belly’s sake men have evil troubles

Subjects of verbs of reaching are: ἄλγος, ἄχος, κῆδος, ὀϊζύς, πένθος, (πῆμα).35
When a form of ἱκάνειν is involved they express a lasting feeling or situation
(ὀϊζύς);36 this is always the case with κῆδος; cp. also the presence of κηδομένῳ.
Note, in (53), the presence of the perfect participle ἀκαχημένον after ἄχος ἱκά-
νει. Unlike the constructions of sections 2.1 and 2.2, the expressions with ἱκάνειν
typically are not accompanied by expressions indicating the cause of the phe-
nomenon. They describe, then, the emotional state of the object in a general
way. An exception is (54), where the κῆδος is the result of a wound.

2.4 With Verbs of Approaching

(58) τοῖσιν γὰρ μέγα πῆμα κυλίνδεται· οὐ γὰρ Ὀδυσσεὺς

δὴν ἀπάνευθε φίλων ὧν ἔσσεται (Od. 2.163–164, Halitherses speaking; τοῖσιν
= the suitors. Cp. Il. 11.347 (… πῆμα κυλίνδεται ὄβριμος Ἕκτωρ), 17.99)

since great trouble rolls toward them, for Odysseus will not long be far
from his loved ones

(59) ἐξ αὖ νῦν ἔφυγες θάνατον κύον· ἦ τέ τοι ἄγχι

ἦλθε κακόν (Il. 11.362–363, Diomedes addressing Hektor. Cp. μέγα γὰρ
κακὸν ἐγγύθεν ἦεν (Od. 9.423 etc.))

Once again now you escaped death, dog. And yet the evil came near you

Example (58) is the middle-passive variant of (23). The subjects of verbs of

approaching are, then: πῆμα and κακόν. The use of these verbs is consistent with
the first-order features of πῆμα and κακόν mentioned above. In view of these
features I think that πῆμα κυλίνδεται etc. are not to be taken metaphorically.

35 I have put πῆμα between parentheses, because this is a dubious example. See note 6.
36 The full clause runs: ἔνθα οἱ αἶσα | ἐκφυγέειν μέγα πεῖραρ ὀϊζύος, ἥ μιν ἱκάνει (‘where it’s his
destiny to escape the great bond of misery that’s come to him’), where the meaning of πεῖ-
ραρ ὀϊζύος is not clear (see the note in Heubeck et al.). The presence of ἐκφυγέειν suggests
that in this case ὀϊζύς may have first-order features.
the syntax and semantics of expressions of sorrow 275

3 Main Conclusions

1. a. Ἄλγεα, ἄχεα, κακά, κήδεα, ὀϊζύς, πένθος, πόνος, may be the object of
ἔχειν and thus be presented as the ‘possession’ of a human being.
The exclusive use of the plural, of those words that are marked for
number, indicates that habitual states of affairs are involved.
b. The ‘possession’ of ἄλγεα, κακά, κήδεα, ὀϊζύς, πένθος may originate
with the gods. In that case a verb of giving is used, e.g. διδόναι, πορεῖν,
ὀπάζειν. Πόνος does not originate with the gods: it belongs par excel-
lence to the human world.
c. The ‘possession’ of ἄλγεα, κήδεα, πένθος may also originate with a
human being; this is always the case with πόνος. In that case the verb
used is τιθέναι, meaning ‘to create ἄλγεα etc. for somebody’, i.e. ‘cause
somebody (dative) to receive/have/feel ἄλγεα etc.’
d. The origin of ἄχεα cannot be established, by lack of material.
2. Πῆμα is not a feeling but rather a concrete threatening phenomenon or
person: ‘disaster, calamity’; it may originate with the gods, in which case a
verb of bringing is used. Πῆμα is never construed as the object created by
a human being for another human being. On the other hand, the plural
πήματα, in combination with a form of πάσχειν, is usually an alternative
for ἄλγεα, conditioned by metrical factors.
3. Ἄλγος/ἄλγεα, ἄχος/ἄχεα, κακόν, κήδεα, ὀϊζύς, πένθος, πόνος may occur in
an alternative possessive construction, with forms of εἶναι + dative. Ἄχος
is also frequently the subject of (ἐ)γένετο + dative, expressing a sudden
feeling of anger, which prompts a reaction. In these cases the feeling con-
cerned is typically caused by an event. Again, πῆμα is not used in this way.
4. Ἄλγος, ἄχος, πένθος (once) may occur as the subject of a verb of seizing,
always in the aorist, expressing a single momentaneous feeling, caused by
an event and prompting a reaction.
5. Ἄλγος, κῆδος, ἄχος, πένθος, ὀϊζύς, πόνος may occur as the subject of a verb
of reaching, mostly expressing a lasting feeling.
6. Πῆμα and κακόν may be the subject of a verb of approaching.
7. Κακόν/κακά are used as replacers for the more specific terms.
8. Ἄλγος/ἄλγεα, ἄχος/ἄχεα, κῆδος, ὀϊζύς, πένθος, πήματα and πόνος are so-
called second-order entities (abstract action nouns). Κήδεα are usually
third-order entities (abstract ‘non-action’ noun); they are typically men-
tal representations of ἄλγεα. Πῆμα comes close to being a first-order entity
(concrete noun). Κακόν/κακά may be all three types of entities.
9. a. The words involved occur in a number of different constructions;
also, the syntactic behaviour of the singular ἄλγος, κῆδος, ἄχος, πῆμα
differs in important respects from that of the plural.
276 chapter 13

b. The syntactic differences often involve striking semantic differ-

ences, that are not sufficiently acknowledged in the standard dictio-
naries etc. These differences suggest that dictionary entries should
be arranged according to the syntactic and semantic features of the
constructions in which the words concerned occur. E.g.:

ἄλγος (subj.)
1. With a form of εἶναι + dat.: ‘affliction, ordeal’; also ‘mental pain’: ‘sorrow’,
esp. when a word like θυμῷ is present
2. With verb of seizing, always in aorist: ‘pain’, physical or mental
3. With verb of reaching: ‘pain’, physical or mental; stative, i.e. a lasting feel-
ing, with a form of ἱκάνειν

ἄλγεα (obj.)
With ἔχειν, διδόναι (subject: a god), τιθέναι (subject: a human being): ‘afflictions’;
when θυμῷ is present, ‘feelings of sorrow’.
– LSJ: ‘pain of body …’; ‘pain of mind’
– Chantraine: ‘souffrance physique, souffrance’
– Lex. frühgr. Epos: ‘Schmerz’
– Mawet (1979: 224): ‘souffrance infligée par un phénomène, un événement ou
une volonté surnaturels …’, ignoring its use with τιθέναι.

ἄχος (subj.)
1 a. With form of εἶναι + dat.: ‘anger mixed with grief’; single feeling, pos-
sibly of a lasting nature
b. With (ἐ)γένετο + dat.: ‘anger’; not a lasting feeling, prompts an action
2. With verb of seizing, and other verbs of violence, always in aorist or per-
fect: ‘anger’, prompts an action
3. With verb of reaching: ‘anger mixed with grief’; when a form of ἱκαν- is
involved it expresses a lasting feeling; does not prompt an action

ἄχεα
(subj.) With form of εἶναι + dat.: ‘afflictions’?
(obj.) With ἔχειν + θυμῷ: ‘sharp feelings of sorrow’
– LSJ simply gives: ‘pain, distress’.
– Mawet (1979: e.g. p. 297, 313) correctly observes that ἄχος frequently means
‘colère’, but nevertheless she discusses its uses under the general heading
(p. 296): ‘Ἄχος dans les contextes du deuil’.
– Lex. frühgr. Epos: ‘schmerzlicher Schreck, selten anhaltende Empfindung’.
‘Schreck’ is perhaps too ‘gentle’.
the syntax and semantics of expressions of sorrow 277

κῆδος
(subj.) With verb of reaching, always a form of ἱκάνειν: ‘vexing paralysing, feel-
ing’, of a lasting nature

κήδεα
(subj.) With form of εἶναι + dat.: ‘vexing feelings’, ‘tourments de l’ âme’
(Mawet 1979: 362)
(obj.) With ἔχειν, διδόναι (subject: a god), τιθέναι (subject: a human being): ‘vex-
ing feelings’, ‘tourments de l’âme’
– LSJ: (‘care about’ + genitive, referring to Od. 22.254, a unique example; not
discussed in the present paper);37 2. ‘anxiety, grief’ … mostly in pl. ‘troubles’
… b. esp. for the dead, ‘funeral rites, mourning’. The latter meanings are far
too concrete.
– Chantraine: ‘soin, souci (Hom.)’; but there are no instances of the former
meaning.

πῆμα
(subj.) With verb of reaching (1×): ‘disaster’ (if authentic)
(obj.) 1. With verbs of bringing or fleeing: ‘disaster’. 2. With a form of παθεῖν:
‘disaster’

πήματα (obj.)
1. With verb of finding: ‘disasters’
2. With a form of πάσχειν: ‘afflictions’?; or possibly again ‘disasters’
– LSJ: ‘misery, calamity’; no mention of the use of the plural with πάσχειν.
– Chantraine: ‘souffrance, malheur’; no information about possible dif-
ferences between sing. and plur. The first meaning is very unlikely for
the singular.
– Mawet’s treatment of πῆμα/πήματα is very complicated, and flawed
by an etymological bias. Following Benveniste and others she believes
that πῆμα, being a noun in -μα, has a ‘middle’ meaning, expressing ‘une
activité à laquelle participe le sujet ou dans laquelle il est engagé’ (1979:
86). Since this is patently at variance with the use of the singular, she
is at great pains to explain how e.g. the meaning ‘calamity’, which she
duly recognizes, should be combined with the alleged meaning of the
suffix -μα.

37 By giving this as the first meaning LSJ suggests that ‘care’ is the basic meaning of κῆδος,
which may strongly be doubted, in view of the other uses.
278 chapter 13

Select bibliography

Mawet (1979) contains a full bibliography of the older literature.

Boel, G. de, Goal Accusative and Object Accusative in Homer. A Contribution to the Theory
of Transitivity (Brussels 1988)
Kahn, Ch., The Verb ‘Be’ in Ancient Greek (Dordrecht/Boston 1973)
Lyons, J., Semantics (Cambridge 1977)
Mawet, F., Recherches sur les oppositons fonctionnelles dans le vocabulaire homérique de
la douleur (Brussels 1979)
Rijksbaron, A., ‘D’où viennent les ἄλγεα? Quelques observations à propos d’ ἄλγε᾽ ἔχειν
chez Homère’, in F. Létoublon (ed.), La langue et les textes en grec ancien. Actes du
Colloque P. Chantraine (Grenoble, 5–8 septembre 1989) (Amsterdam 1991) 181–193
Rijksbaron, A. ‘Further Observations on Expressions of Sorrow and Related Expressions
in Homer’, in E. Banfi (ed.), Atti del Secondo Incontro internazionale di Linguistica
greca (Trento 1997) 215–242
 chapter 14

The Meaning and Word Class of πρότερον and τὸ

πρότερον

Abstract

If we are to believe the grammars and dictionaries of Ancient Greek there is no differ-
ence between temporal adverbs like νῦν, ἔπειτα and πρότερον, and the variants where
these adverbs are combined with what looks like the article τό or τά. See e.g. LSJ s.v.
πρότερον: ‘after Hom. neut. πρότερον freq. as Adv., before, earlier … Also used with the
Art. τὸ π.’ In this paper it is argued that this view, at least for τὸ πρότερον, is untenable.
Whatever its precise morpho-syntactic status (an issue that is discussed at the end of
the paper), τό modifies the meaning of πρότερον in an article-like way, since it turns an
adverb that expresses an indefinite time-position into a definite expression. Τὸ πρότε-
ρον is used in two, rather different, ways. If τὸ πρότερον modifies an aorist-stem form its
meaning is ‘the previous time, the last time’; if, on the other hand, it modifies a present-
stem form its meaning is ‘during the preceding period’. The meaning of single πρότερον
does not depend on the form of the verb, it is always ‘before, earlier’.

1 Introduction

As is well known, the Greek article, apart from its functioning as a determiner
of nouns, may be used to form various kinds of other noun phrases, either fully
developed noun phrases, i.e. phrases that are inflected for all three morpholog-
ical categories (case, gender, number), or semi-developed noun phrases, | hav- 442
ing a reduced number of nominal properties.* To the first class belong phrases
with adjectives (ὁ σοφός, ἡ καλή, τὸ ἀγαθόν, οἱ καλοὶ κἀγαθοί, τὸ ἡμέτερον), par-
ticiples (ἡ τίκτουσα, ὁ οὐκ εἰδώς, τὰ νομιζόμενα) and with the particles μὲν … δέ
in fixed expressions (ὁ/οἱ μὲν … ὁ/οἱ δέ, ὁ δέ alone). The second class consists
of the infinitive (τὸ λέγειν), certain prepositional phrases (oἱ περὶ Κῦρον, οἱ ἐν
ἄστει, τὰ ἐν πολέμῳ), adverbs (οἱ νῦν, τὸ αὐτίκα, τὰ ἐκεῖ) and numerals (τὸ ἕν, οἱ
ἕνδεκα),1 all of which to varying degrees lack one or more of the full nominal

* I am indebted to Dr G.C. Wakker for her comments and suggestions.

1 I will ignore in this paper the attributive use of the adverb, as in οἱ ἐκεῖ ἄνθρωποι, οἱ νῦν τύραννοι.

© koninklijke brill nv, leiden, 2019 | doi:10.1163/9789004386129_016

 280 chapter 14

properties.2 And thirdly there is the use where we find what looks like the defi-
nite neuter article, but may in reality be something different. Examples of this
use are expressions like τὸ ἐπ᾽ ἐμοί, τὸ κατ᾽ ἐμέ, τὸ λεγόμενον and also the subject
of the present paper, the article with temporal adverbs.
By way of orientation, I start with giving a (non exhaustive) list of such
phrases, using the works of Plato as my main corpus.3
– τὸ νῦν, τὰ νῦν
– τὸ αὐτίκα, τὸ παραυτίκα
– τὸ παραχρῆμα
– τὸ τήμερον
– τὸ πάλαι, τὸ παλαιόν
– τὸ ἀρχαῖον
– τὸ πρίν
– τὸ μετὰ ταῦτα, τὰ μετὰ ταῦτα
– τὸ ἐντεῦθεν
– τὸ ἔπειτα
– τὸ λοιπόν4
– τὸ πρῶτον, τὰ πρῶτα
– τὸ πρότερον
– τὸ ὕστερον
443 – τὸ δεύτερον etc. |

Immediately a qualification is in order. Many of these combinations may also

function as noun phrases, i.e. they may have all the functions that normal noun
phrases have, e.g. that of a subject, as τὰ νῦν in:5

(1) καὶ γάρ, ὦ ξένε, ἡμῖν καὶ τὰ πρότερον ὀρθῶς σοι παιδείας πέρι καὶ τὰ νῦν εἰρῆ-
σθαι δοκεῖ (Pl. Lg. 653c–d)

It seems to us that both what you said before and what you said now about
education is quite right

or as an object, as τὸ αὐτίκα in:

2 E.g. for τὸ λέγειν: no gender alternatives, no plural; οἱ περὶ Κῦρον: no singular; τὸ ἕν: no gender
alternatives, no plural; οἱ ἕνδεκα: no singular.
3 I notice here in passing that not all temporal adverbs may be modified by the article, since
e.g. the following combinations are not found: *τὸ ἔτι, *τὸ ἢδη, *τὸ ἄπαξ, *τὸ ἑκάστοτε. For an
explanation see note 18.
4 Τὸ λοιπόν is a special case, since single λοιπόν would not seem to be used adverbially.
5 The translations are either borrowed from, or based upon, those in the Loeb series.
the meaning and word class of προτερον and το προτερον 281

(2) ὅταν … τὸ αὐτίκα περισκοπῶν ἐνδοιάζῃ (Th. 1.36.1)

When … he hesitates through cautious consideration of the immediate

chances

Observe that in the above cases the article cannot be omitted without radically
changing the meaning of the sentence. Although these uses are interesting in
themselves I will leave them aside in the present paper.
It is high time now to turn to my real subject, viz. the so-called adverbial uses
of these phrases,6 that will be illustrated from a phrase that occurs with some
frequency in Plato (and elsewhere): τὸ πρότερον.7

2 Some (Morpho-)syntactic Observations

Above, I briefly mentioned the nominal functions these articular groups may
fulfil. In fact, the first question to be answered is how we are able to distinguish
the (pro)nominal use from the adverbial use. As a rule of thumb, and not sur-
prisingly, of course, τὸ πρότερον must be considered adverbial if the argument | 444
positions of the verb are taken by other constituents. The following examples
may illustrate this point:

(3) τῷ ἀδελφῷ ὑμῶν τῷ ὁμομητρίῳ τί ἦν ὄνομα; οὐ γὰρ μέμνημαι. παῖς δέ που ἦν,
ὅτε τὸ πρότερον ἐπεδήμησα δεῦρο ἐκ Κλαζομενῶν· πολὺς δὲ ἤδη χρόνος ἐξ ἐκεί-
νου. (Pl. Prm. 126b)

What was your half-brother’s name? For I don’t remember. He was only a
boy when I came here from Clazomenae before (so Fowler, wrongly, see
below); and that is now a long time ago.

6 The question whether e.g. πρότερον should, indeed, be taken as an adverb will be addressed at
the end of this paper. For the other items of the list I confine myself here to giving a few refer-
ences; some of them will turn up again in the discussion below. Τὸ αὐτίκα, τὸ παραυτίκα (not
in Plato): Hdt. 6.63.2; τὸ παράχρημα: Pl. Lg. 670d; τὸ τήμερον: Smp. 176e; τὸ πάλαι: Phdr. 251b;
τὸ παλαιόν: Cra. 401c; τὸ ἀρχαῖον: Prt. 316e; τὸ πρίν: Criti. 112a; τὸ μετὰ τοῦτο: Sph. 218a; τὰ μετὰ
ταῦτα: Lg. 629e; τὸ ἐντεῦθεν: Prt. 321b; τὸ ἔπειτα (very rare): Th. 5.115.4; τὸ λοιπόν: Pl. Phdr. 256c;
τὸ ὕστερον: Th. 3.87.2; τὸ πρῶτον: Pl. Phd. 75a; τὸ δεύτερον: Grg. 476a; Hdt. 3.159.1; τὸ τελευταῖον:
Th. 8.8.2.
7 There are some 20 examples in the works of Plato (this does not include the doubtful works
and the spuria).
 282 chapter 14

(4) ἦσαν δὲ οἱ Σύριοι οὗτοι τὸ μὲν πρότερον ἢ Πέρσας ἄρξαι Μήδων κατήκοοι, τότε
δὲ Κύρου (Hdt. 1.72.1)

and these Syrians before the Persian rule were subjects of the Medes, and,
at this time, of Cyrus

In (3), τὸ πρότερον occurs as an optional modifier with two-place, or possibly

three-place, ἐπιδημέω, δεῦρο and ἐκ Κλαζομενῶν being the second and third
argument, respectively; in (4) with two-place κατήκοός εἰμι, Μήδων being the
second argument. In this respect, τὸ πρότερον is exactly like single πρότερον, cp,
e.g. Ly. 212d, where we find πρότερον combined with one-place δοκέω:

(5) ἀλλοίως ἄρα νῦν ἡμῖν δοκεῖ ἢ πρότερον ἔδοξεν (Pl. Ly. 212d)

So you see, we now hold a different view from what we held before

I will come back to the issue of the word class of τὸ πρότερον and πρότερον at
the end of this paper.

3 Τὸ πρότερον and πρότερον

Above (ex. (3)), I added ‘wrongly’ to Fowler’s translation ‘before’ of τὸ πρό-

τερον, the reason being that he ignores the article, And he is definitely not
the only one;8 on the contrary, there is a strong tendency in dictionaries and
grammars, as well as among translators and commentators, to make no dis-
tinction between τὸ πρότερον and πρότερον, nor, for that matter, between the
other adverbs of the list on p. [442] and their non-articular counterparts. The
description of πρότερον in LSJ may be taken as representative of this tendency:
445 ‘after Hom. neut. | πρότερον freq. as Adv., before, earlier … Also used with the
Art. τὸ π.’9 Although this view tallies with the widespread belief that in several
types of noun phrases the article may be added or omitted at will, without any
discernable semantic effect,10 it is of course, as a matter of principle, highly

8 There are exceptions, though. Thus, Diès, in his Budé-edition, translates ‘lors de mon pre-
mier voyage’, which, by ‘mon’, is a definite expression. And if τὸ πρότερον is just opposed to
Parmenides’ actual trip it is of course correct. See also example 17 and discussion.
9 Cp. also Bailly s.v. πρότερος ‘neutre adv. πρότερον … ou τὸ πρότερον … auparavant’.
10 A typical example of this belief is Goodwin (1894: 206): ‘Proper names may take the arti-
cle; as ὁ Σωκράτης or Σωκράτης Socrates … Abstract nouns often take the article; as ἡ ἀρετή,
the meaning and word class of προτερον and το προτερον 283

questionable. Whatever the precise semantic characteristics of the Greek arti-

cle, it seems a sound procedure to start from the assumption that its func-
tion in τὸ πρότερον is no different from that in, say, ὁ πρότερος πόλεμος,11 and
that it functions therefore as a definiteness marker. Consequently, τὸ πρότε-
ρον should be taken as a definite temporal modifier. It can be shown to have
two completely different functions, which depend on the tense stem of the
verb:
– if the verb is from the aorist stem, τὸ πρότερον indicates that the present
occurrence of the state of affairs was preceded by another occurrence of the
same state of affairs at some definite moment in time,12 i.e. a moment which
the speaker considers identifiable for the hearer. Its literal meaning is ‘the
previous time’, Fr. ‘la fois passée’,13 Ge. ‘das vorige Mal’, Sp. ‘la vez pasada’, Du.
‘de vorige keer’,14 or possibly, if τὸ πρότερον is contrasted with just one later
occasion, ‘the first time’, ‘la première fois’ etc.
– if, on the other hand, the verb is a form of the present stem, τὸ πρότερον gets
a durative reading, indicating that during some definite stretch of time a
state of affairs existed which preceded a reference point given by the con-
text. See example (4). In this use the literal meaning of τὸ πρότερον is: ‘during
the | preceding (time)’. I should add that in this use τὸ πρότερον sometimes 446
comes close to the function of single πρότερον. I will come back to this
below.
Finally, single, non-articular, πρότερον. Being indefinite, its basic function is
to indicate that the state of affairs concerned is simply prior with respect
to some other state of affairs or to the speech moment. Its meaning is ‘at
some earlier moment, previously, before’; the nature of the tense stem of the
accompanying verb is of no consequence for the interpretation. Two examples
are:

virtue, … But ἀρετή etc. are also used in the same sense.’ In Rijksbaron (2002) [chapter 16 in
this volume] I have tried to show that there are clear semantic differences between proper
names with and without the article.
11 Cp. e.g. Mx. 242d, Th. 7.18.2.
12 This use well illustrates the ‘countability’ of states of affairs in the aorist, cp. Armstrong
(1981).
13 Cp. the definition of ‘fois’ in Le Petit Robert: ‘moment du temps où un événement, conçu
comme identique à d’ autres événements, se produit.’
14 In English and the Romance languages these literal meanings seem to be rather strange,
if acceptable at all. One says rather ‘(the) last time’ (as in ‘the last time I saw him was …’),
‘la dernière fois’, ‘la última vez’. Nonetheless, these translations are rarely if ever found for
τὸ πρότερον. The reason may be that ‘last’, ‘dernière’ etc. intuitively clash with ‘earlier’, as
expressed by πρότερον.
284 chapter 14

(6) ἔπειτα τὰ μὲν ἐκ τοῦ παραχρῆμά μοι διῄει, οἷα δέοι λέγειν, τὰ δὲ πρότερον
ἐσκεμμένη, ὅτε μοι δοκεῖ συνετίθει … (Pl. Mx. 236b)

and thereupon she rehearsed to me the speech in the form it should take,
extemporizing in part, while other parts of it she had previously prepared,
as I imagine, at the time when …

(7) ΣΩ. … τί τὸ κεφάλαιόν ἐστι τῆς ἐργασίας;

ΕΥΘ. Καὶ ὀλίγον σοι πρότερον εἶπον, ὦ Σώκρατες, ὅτι πλείονος ἔργου ἐστὶ …
(Pl. Euthphr. 14a)

Soc. … what is the chief result of their work?

Euth. I told you a while ago, Socrates, that it is a long task to …

Πρότερον is frequently modified by adverbs like ὀλίγον ‘a little’, as in (7). In fact,

by such adverbs πρότερον, which in itself is not very specific of course, becomes
much more informative. Example (7) is a case in point, for by the presence
of ὀλίγον the moment referred to must have been easily retrievable for the
hearer.15 Observe also that the inherently vague πρότερον is specified in (6) by
ὅτε … συνετίθει.

4 Some Syntactic-Semantic Differences

Consider the following examples:

(8) … περὶ ἐκείνων τῶν ἀνδρῶν ὧν ὀλίγῳ πρότερον ἔλεγες (Pl. Grg. 515c)

… those men whom you mentioned a little while ago

(9) μνήμην, ὡς ἔοικεν, ὅτι ποτ᾽ ἔστιν πρότερον ἀναληπτέον, καί κινδυνεύει πάλιν
ἔτι πρότερον αἴσθησιν μνήμης … (Pl. Phlb. 33c)

We must, apparently, first take up memory, and perception even before

memory …

(10) τῶν μὲν ἐρώντων πολλοὶ πρότερον τοῦ σώματος ἐπεθύμησαν ἢ τὸν τρόπον ἔγνω-
σαν (Pl. Phdr. 232e)

15 The reference is to 9b4.

the meaning and word class of προτερον and το προτερον 285

… many lovers are moved by physical passion before they get to know the
character … | 447

(11) οὐ πρότερον αὐτὸν ἀφήσει Σωκράτης, πρὶν ἂν βασανίσῃ ταῦτα εὖ (Pl. La. 188a)

Socrates will never let him go until he has thoroughly and properly put all
his ways to the test

From these examples it appears that πρότερον

– can be modified by intensifiers like ὀλίγῳ, cp. (8), cp. also ὀλίγον in (7), and
ἔτι, cp. (9)
– can be construed with ἤ or a genitivus comparationis, cp. (10)
– can be used with respect to a point of reference in the future, i.e. after the
speech moment, cp. (9) and (11), a frequent construction.
I should add that at least the second of these constructions is also found with τὸ
πρότερον in its durative reading, cp. example (4) above, but I will not go into this
phenomenon here. However, τὸ πρότερον is never used in these constructions
when it is combined with an aorist form, i.e. in its ‘the previous time’ reading.
What can we infer from these facts? First, that single πρότερον, in conformity
with its formal characteristics, functions essentially as a comparative.16 Like all
comparatives, it can be modified by intensifiers, and it is truly indefinite, since
it does not refer to a fixed moment in time. Second, that in the construction of
τὸ πρότερον + aorist form the comparative nature of πρότερον is no longer oper-
ative. The reason, we may add, is that the article binds πρότερον to a definite
point in time.17 Below I will come back in more detail to some of the theoreti-
cal issues involved here.18 | 448

16 That is, synchronically. The original function of the suffix -τερος was to express opposition,
cp. e.g. Risch (1974: 91 ff.). See also below, example (17).
17 Just as in the nominal domain ὁ binds a comparative to a definite entity in oppositional
pairs like ὁ μὲν πρεσβύτερος … ὁ δὲ νεώτερος (X. An. 1.1.1), and also in ὁ πρότερος πόλεμος for
that matter.
18 In fn. 3 I mentioned the phenomenon that adverbs like ἔτι, ἤδη, ἅπαξ are not found with the
article, unlike adverbs like νῦν, αὐτίκα, πρότερον etc. I submit that this is due to an impor-
tant semantic difference between these two groups. The group of νῦν etc. has referential
meaning, i.e. they locate the state of affairs at some identifiable point of time, either with
respect to the speech moment or with respect to another state of affairs. In short, they can
serve as an answer to the question πότε ‘when?’. As with other lexical items having refer-
ential meaning, the point of time involved may be definite or indefinite, as was illustrated
by (τὸ) πρότερον. Adverbs like ἔτι, ἤδη, ἅπαξ, ἐνίοτε, ἑκάστοτε, on the other hand, although
they are semantically heterogeneous, have in common that they do not locate their state
of affairs in time: ἔτι and ἤδη say something about the state of affairs in terms of some con-
286 chapter 14

I conclude this discussion with some other examples illustrating the differ-
ences.

5 Further Examples

5.1 Of τὸ πρότερον with a Form of the Aorist Stem; = ‘the previous time’

(12) ὅτε … τὸ πρότερον ἐπεδημησάτην (Pl. Euthd. 273e)

not ‘… when you visited our city before’ (Lamb) but ‘… the last time you
visited our city’19

(13) ἔτι γὰρ παῖς ἦ ὅτε τὸ πρότερον ἐπεδήμησε (Pl. Prt. 310e)

‘I was still a child when he paid us his previous visit’, (Lamb); for Croiset’s
‘la première fois qu’il est venu ici’ see below

(14) νυνὶ δὲ ὅπερ ἄρτι σε ἠρόμην πειρῶ σαφέστερον εἰπεῖν. οὐ γάρ με, ὦ ἑταῖρε, τὸ
πρότερον ἱκανῶς ἐδίδαξας ἐρωτήσαντα τὸ ὅσιον ὅτι ποτ᾽ εἴη, … (Pl. Euthphr.
6d)

At present try to tell me more clearly what I asked you just now. For, my
friend, you did not give me sufficient information before, when I asked
what holiness was … (Fowler; this should be ‘the previous time’ or ‘the
first time’)

(15) ὅθεν … τὸ πρότερον ἀπέλιπον … (Pl. Euthd. 288c)

‘from where I left off before’, Lamb; a most unlikely translation

(16) τὸ γὰρ πρότερον ἡμᾶς ἐπηγάγεσθε (Th. 6.86.1)

not ‘for you brought us over before’ (Smith), but ‘the previous time you
brought us over …’,

ventional measure of time, while the frequency adverbs say something about the number
of times the state of affairs was realised. In both cases, the adverbs cannot serve as an
answer to πότε ‘when?’ Consequently, notions like ‘definiteness’ are irrelevant for them.
19 ‘Lors de votre premier séjour’—Meridier, which is only correct if there were just two visits.
the meaning and word class of προτερον and το προτερον 287

Above, I mentioned the possiblity that τὸ πρότερον should be interpreted as

‘the first time’, viz. if it is contrasted with just one later occurrence of the same
state of affairs. This reflects the original function of the suffix -τερος, which was
to express opposition, cp. e.g. Risch (1974: 91ff.; see also note 16). An example
where this opposition or contrast is overtly present is:

(17) τοῦ δ᾽ ἐπιγιγνομένου χειμῶνος ἡ νόσος τὸ δεύτερον ἐπέπεσε τοῖς Ἀθηναίοις …

παρέμεινε δὲ τὸ μὲν ὕστερον οὐκ ἔλασσον ἐνιαυτοῦ, τὸ δὲ πρότερον καὶ δύο ἔτη
(Th. 3.87.2)

In the course of the following winter the plague again fell upon the | Athe- 449
nians … And it continued the second time no less than a year, having run
for two full years on the previous occasion

Observe that the translator, Smith, has ‘the second time’ for τὸ ὕστερον but ‘the
previous occasion’ for τὸ πρότερον, perhaps striving after variatio; observe also
that he renders τὸ δεύτερον ‘for the second time’ as ‘again’, for no obvious reason.
Although it cannot be excluded that τὸ πρότερον = ‘the first time’ also in cases
where there is no explicit contrast (cp. Croiset’s translation of τὸ πρότερον in
(13)), I have as a rule preferred the ‘the previous time’ interpretation.
If there is no opposition involved ‘the first time’ is rather expressed by τὸ
πρῶτον, as in:

(18) πρὸ ἐκείνου τοῦ χρόνου ὅτε τὸ πρῶτον ἰδόντες τὰ ἴσα ἐνενοήσαμεν ὅτι … (Pl.
Phd. 75a)

… before the time when we first saw equal things and came to realize that
… (transl. Fowler)20

5.2 Of τὸ πρότερον with a Form of the Present Stem; = ‘during the

preceding time’

(19) ὃ δὲ τὸ πρότερον ἐλέγομεν φίλον εἶναι, ὕθλος τις ἦν (Pl. Ly. 221d)

… while what we kept saying previously about friends was all mere drivel
(rather than ‘… while our earlier statement about friends …’, Lamb)

20 Τὸ πρῶτον far more often means ‘at first, in the first instance’.
 288 chapter 14

(20) οὐκοῦν καὶ ἰατρῶν ἐν χρείαις ἐσόμεθα πολὺ μᾶλλον οὕτω διαιτώμενοι ἢ ὡς τὸ
πρότερον (Pl. R. 373d)

‘Doctors, too, are something whose services we shall be much more likely
to require if we live thus than as before’ (Shorey); but rather (liter.): ‘than
as we lived during the preceding time’

(21) κἀγὼ διενοήθην ὅτι ᾠηθήτην ἄρα ἡμᾶς τὸ πρότερον παίζειν, ἡνίκα ἐκελεύομεν
… (Pl. Euthd. 283b)

At this I reflected that they apparently thought that previously we were

jesting, when we urged them …

(22) οὗτοι δὲ διαβάντες μὲν ἐς τὴν Ἀσίην ἐκλήθησαν Βιθυνοί, τὸ δὲ πρότερον ἐκαλέ-
οντο … Στρυμόνιοι (Hdt. 7.75)

These took the name of Bithynians after they crossed over to Asia; before
450 that they were called … Strymonians |

5.3 Of πρότερον = ‘before, earlier’

(23) πολλὰ γὰρ καὶ καλὰ Γοργίας ἡμῖν ὀλίγον πρότερον ἐπεδείξατο (Pl. Grg. 447a)

‘for Gorgias gave us a fine and varied display but a moment ago’ (Lamb)

(24) οἱ δὲ πρότερον Ἄγρωνος βασιλεύσαντες ταύτης τῆς χώρης ἦσαν ἀπόγονοι Λυ-
δοῦ τοῦ Ἄτυος, ἀπ᾽ ὅτεο ὁ δῆμος Λύδιος ἐκλήθη ὁ πᾶς οὗτος, πρότερον Μηίων
καλεόμενος (Hdt. 1.7.3)

The kings of this country before Agron were descendants of Lydus, … from
whom all this Lydian district took its name; before that it was called the
land of the Meii

It will be noticed that example (22) of τὸ πρότερον is rather similar to example

(24) of πρότερον. So the question arises whether there are any differences. In
fact, at the risk of some over-interpretation, I believe there is a difference. In
(22), in conformity with the use discussed above, τὸ πρότερον should be under-
stood as referring to a definite stretch of time, viz. the one up to the crossing
of the Thracians to Asia; the translation ‘before that’ brings out this nuance
nicely. In (24), on the other hand, where we find single πρότερον, it is only indi-
the meaning and word class of προτερον and το προτερον 289

cated that the name ‘Bithynians’ during some period of time was preceded by
the name ‘Strymonians’. Naturally, this may imply that the name ‘Bithynians’
immediately succeded that of ‘Strymonians’, but this is not expressed formally.
So in this case ‘before that’ seems less correct than, say, ‘in earlier times’.

6 The Word Class of πρότερον and τὸ πρότερον

Finally, I will say a few words about the possible word class of πρότερον and
τὸ πρότερον, both from a traditional lexical and morpho-syntactic, and from a
modern, semantic-functional, point of view.

6.1 From a Traditional Point of View

First of all, I assume that πρότερον, ὕστερον etc., just like πρότερος and ὕστερος,
should be reckoned among the ordinals, and more particularly among the ‘gen-
eral ordinals’. In this I follow the treatment in Quirk et al. of comparable English
adverbs (1985: 262). They reckon next and last among the ‘general ordinals’, as
against first, second etc., which are rather ‘ordinal numerals’.21 | 451
Second, there is the morphological status of the ending -ον. Naturally, estab-
lishing the morphological properties of adjective-like words like πρότερον and
τὸ πρότερον depends to a large extent on one’s definitions. Above, I observed
(see examples (3)–(4) and (6)) that in their use as temporal modifier πρότερον
and τὸ πρότερον are in no way dependent on the predicate, which suggests that
they are not accusatives (nor nominatives), in short, that they are not nominal
elements but rather adverbs.22 But this is not the whole story, for they still might
be considered accusatives, on the analogy of, for instance, τοῦτον τὸν χρόνον in:

(25) τοῦτόν γε τὸν χρόνον αὐτὸν τὸν πρὸ τοῦ θανάτου ἧττον τοῖς παροῦσιν ἀηδὴς
ἔσομαι … (Pl. Phd. 91b)

I shall not be burdensome to my friends … in these last moments

where τοῦτον τὸν χρόνον is not governed by the predicate either (but see below),
but still has the properties of an accusative noun phrase, as can be inferred from

21 In their discussion of the ‘Ordnungszahlwörter’, Kühner-Blass mention πρότερος, but only

in its use as ‘the first of two’ (1890: 1, 622).
22 Whatever its other properties, in an inflected language like Greek a lexical item qualifies
as an adverb if its form is invariable. In my dissertation I discussed the same issue in con-
nection with causal τί; ‘why?’, see Rijksbaron (1976: 10–18).
 290 chapter 14

the fact that the phrase is modified by attributive τὸν πρὸ τοῦ θανάτου. One might
even be tempted to assume that there is a relationship of dependency, after all,
between τοῦτον … θανάτου and ἀηδὴς ἔσομαι, if the temporal phrase is taken as
modifying the inherent stative features of the predicate. Be that as it may, such
attributive modifiers are not found with (τὸ) πρότερον; in fact, πρότερον and τὸ
πρότερον are, in the use under discussion, never modified like ordinary nomi-
nals, so on this—important—count they do not qualify as accusative forms.23
Actually, πρότερον and τὸ πρότερον are exactly like ἧττον in example (24), which
in no meaningful way can be said to depend on the predicate. So if ἧττον is
taken as an adverb rather than as an adjective in the accusative, πρότερον, at
least, should also be taken in this way. Τὸ πρότερον is more difficult, of course,
because of the presence of τὸ. This phrase can perhaps best be analysed as a
combination of the article and adverbial πρότερον, on the analogy of temporal
modifiers like τὸ ἔπειτα and τὸ αὐτίκα. Here we find an undisputable adverb that
is modified by what is to all intents and purposes an article in the accusative.24
The latter can be inferred from the fact that these combinations can be gov-
erned by the preposition εἰς;25 for εἰς τὸ ἔπειτα see e.g. Smp. 193d, for ἑς τὸ αὐτίκα
452 Th. 2.64.6. It is true | that the combination with εἰς is not found with τὸ πρότε-
ρον, but this may be just a coincidence, for εἰς τὸ ὕστερον does occur, e.g. at Prt.
353d.26

6.2 From a Functional Point of View

Thirdly, as for the semantic-functional approach, I confine myself to a discus-
sion of τὸ πρότερον in its ‘the last (= previous) time’ reading, since this is the
most interesting use from a theoretical point of view. I take it that τὸ πρότερον,

23 The discussion of the classification criteria in Pinkster (1972: 18–33) is still relevant for this
matter.
24 The meaning of these phrases is another matter, which will be ignored here.
25 Interestingly, εἰς can also govern pure adverbs, as in εἰς νῦν (Ti. 20c), εἰς πότε; (S. Aj. 1185), εἰς
αὖθις (Euthphr. 6c and some 36 other instances in Plato), εἰς/ἐς αἰεί (Hdt., Th., X.). In fact,
it is because of the existence of such combinations that I would take single ὕστερον in εἰς
ὕστερον (e.g. Pl. Sph. 247e) as an adverb. Prepositons may also govern adverbs of place, cp.
e.g. μέχρι ἐνταῦθα in Smp. 210e. Observe that the same constructions occur in other lan-
guages, e.g. English (till now, near here), French (à bientôt, jusqu’ici), and Dutch (tot nu, tot
hier). For English cp. Quirk et al. (1985: 454).
26 Or is there a semantic reason, εἰς being oriented towards the future?—Theoretically, ὕστε-
ρον in εἰς τὸ ὕστερον could also be taken as an accusative, like τό. However, this is inherently
unlikely, because one would expect there to be a parallel phrase with a noun, just as ἡ προ-
τεραία may be taken as a variant of ἡ προτεραία ἡμέρα. But such a full noun phrase is lacking
in the case of εἰς (τὸ) ὕστερον that is, with a neuter noun. Phrases with a form of χρόνος are
quite common, e.g. εἰς τὸν ὕστερον χρόνον.
the meaning and word class of προτερον and το προτερον 291

just like ‘(the) last time’, ‘de vorige keer’, etc., belongs to the class of ‘quantifi-
cational predication operators’,27 and more specifically to what may be called
‘predication ordinators’, on the analogy of the ordinators that function as term
operators. In his discussion of term operators, Dik (1997: 178ff.) gives the follow-
ing definition of ‘ordinators’: ‘By “ordinators” we understand such elements | as 453
first, second, third, … last. These elements presuppose that the intended refer-
ent belongs to a linearly ordered series of similar entities, and they indicate the
position of the intended referent in the sequence.’28 As examples Dik gives It is
the third house on your left hand, where ‘third’ is an ‘absolute ordinator’, and The
next problem, where ‘next’ is a ‘relative ordinator’. To Dik’s analysis we should
add that in this use the ordinator is typically definite, since it is preceded by the
definite article. If we apply this analysis to ‘the third time’ and ‘the previous/last
time’, as well as to τὸ πρότερον, we can say that ‘the third time’ functions as an
‘absolute temporal ordinator’, while τὸ πρότερον and ‘the last/previous time’, or
‘the next time’, for that matter, function as ‘definite relative ordinators’, since
their position in the sequence is fixed.

References

Armstrong, D., ‘The Ancient Greek Aorist as the Aspect of Countable Action’, in Ph. Te-
deschi and A. Zaenen (eds), Tense and Aspect (New York 1981) 1–13.
Dik, S.C., The Theory of Functional Grammar, part I (Berlin/New York 1997).
Goodwin, W.W., Greek Grammar (London 1894; repr. London 1997).
Kühner, R. and Blass, F., Ausführliche Grammatik der griechischen Sprache, 1. Teil (Han-
nover 1890).
Pinkster, H., On Latin Adverbs (Amsterdam 1972).

27 It is true that in Functional Grammar predication operators are confined to ‘the grammat-
ical (my italics) means through which the SoA designated by the core predication can be
quantified …’ (Dik 1997: 236), notably aspect and tense operators, but this seems rather
arbitrary, in view of the fact that first, second etc. are lexical rather than grammatical ele-
ments, but are nevertheless treated as term operators. Another option would be to take
the last time, the next time etc. as predication satellites. In fact, as Dik (1997: 243) observes,
‘Most of the distinctions which have been discussed as potentially expressable by π2 (=
predication) operators can in principle also be expressed in the form of σ2 satellites.’ All
in all, on the analogy of their use in term operators, I prefer to take quantifying elements
like the first and the last in the first time, the last time, as well as τὸ πρότερον in the use under
discussion, as operators. Conversely, since τὸ πρότερον in the ‘during the preceding time’
reading, and πρότερον ‘before’, have no quantifying function, these should rather be taken
as satellites.
28 Instead of ‘similar entities’ I would prefer ‘identical entities’. See also note 12.
292 chapter 14

Quirk, R. et al., A Comprehensive Grammar of the English Language (London 1985).

Rijksbaron, A., Temporal and Causal Conjunctions in Ancient Greek: With Special Refer-
ence to the Use of ἐπεί and ὡς in Herodotus (Amsterdam 1976).
Rijksbaron, A., Over bepaalde personen (On Definite Persons) (Amsterdam 2002).
Risch, E., Wortbildung der homerischen Sprache (Berlin 1974).
 chapter 15

Adverb or Connector? The Case of καὶ … δέ

Abstract

This chapter examines the syntactic function and semantic value of the particle com-
bination καὶ … δέ. It defends the view of Denniston and Kühner-Gerth that the com-
bination consists of connective καί and adverbial δέ (meaning something like ‘and, on
the other hand’) against the more common view that it is the other way round (yielding
a meaning ‘and also’). It is shown that καὶ … δέ either connects single nouns, adjectives
or verbs, or clauses and sentences. In the former case καί cannot be omitted, while in
the latter case the existence of correlative constructions (e.g. τε … καί …) demonstrates
that here, too, καί is the connector. It is argued that the function of adverbial δέ is a
pragmatic one; it sets the word or phrase which it modifies apart as an item to be con-
sidered in its own right.

1 Introduction

In Rijksbaron 1997 I made mention of Denniston’s carelessness in syntactic

matters, using his discussion of ἀλλὰ γάρ as an example. Another particle
combination illustrating this attitude is καὶ … δέ. In fact, the treatment of
this combination shows us Denniston at his most easy-going. On p. 199 he
says: ‘This is a natural enough combination, the former particle denoting that
something is added, the latter that what is added is distinct from what pre-
cedes.’1 As appears from a footnote this analysis was inspired by Jebb: ‘Jebb,
on S. Ph. 1362, argues for the view that, in καὶ … δέ, καί is the conjunction,
while δέ means “on the other had”, “also”’, on which Denniston comments:
‘This is, I think, the right explanation of most of the passages. But there are
others (see (2) below) in which δέ seems to be the conjunction, while καί
means “also”.’2 One immediately wonders, first, how Denniston can advocate

1 Incidentally, in his discussion of δέ Denniston uses the terms ‘contrast’, ‘adversative’ (p. 162)
and ‘opposed’ (p. 165), but not ‘distinct’; these terms are considered, moreover, values of the
connective use only (p. 162).
2 Jebb is less categorical than Denniston suggests, for his remark runs: ‘The usual account of it
is that the καί = ‘also’, while δέ = ‘and’. This suits those instances in which, as here, καὶ … δέ is
preceded by a full stop, or by a pause’ (i will come back to the Philoctetes passage below—

© koninklijke brill nv, leiden, 2019 | doi:10.1163/9789004386129_017

 294 chapter 15

a translation of καὶ … δέ in which δέ, which should express ‘distinctness’,

means ‘also’, while at the same time admitting that there are cases where it
is καί that means ‘also’. Rather surprisingly, the net result of his analysis is
that in either case the combination καὶ … δέ means ‘and also’. Second, one
also wonders in what way the combination of καί and δέ differs from that
of δέ and καί. Most importantly, Denniston’s remarks on δέ amount to say-
ing that in most cases of καὶ … δέ it is δέ which is the adverb, although the
term as such is avoided. But thereby he flatly contradicts his own rather apo-
dictic statement on p. 162: ‘Except in the apodotic use, δέ is always a con-
188 nective.’ Matters are further complicated by Denniston’s treatment of | οὐδὲ
… δέ, the negative counterpart of καὶ … δέ. On p. 203 he remarks: ‘Here δέ
is clearly the connective, and οὐδέ is adverbial.’ Why this is so remains unex-
plained.
Denniston might also have turned to Kühner-Gerth for support. They men-
tion our particles on page 2, 253, significantly not in the section on ‘καί, etiam,
als Adverb’, and say that epic καὶ δέ (with no intervening word) = ‘und ander-
erseits, hinwiederum, ferner, und doch’, implying that δέ is adverbial. Rather
confusingly, however, they speak some lines further about: ‘Die Verbindung
der … Konjunktionen (plural) καὶ … δέ’, as if both particles perform the same
function.3 But further on, in the section on δέ, they emphatically state, page 2,
261, that δέ originally was an adverb ‘mit der Bedeutung andererseits, dage-
gen’, and that this meaning is ‘ganz deutlich’ preserved in Homeric καὶ δέ
and post-Homeric καὶ … δέ, as well as in their negative counterpart οὐδὲ …
δέ.
To conclude these preliminary remarks I should add that while Kühner-
Gerth categorically, and Denniston with some reservations, defend the view
that, of καὶ … δέ, it is δέ which is the adverb, many commentators, as we will
see, categorically take the opposite position.
For reasons that will become clear shortly I have arranged my examples into
two groups. First, I shall discuss the use of καὶ … δέ as a connective device, to
use a neutral term, between single words (adjectives, nouns, and verbs). For
practical reasons I have further divided the group of nominal constituents into

AR) ‘but it is less natural where καὶ … δέ links a new clause to a preceding one in the same
sentence; as in Thuc. 4.24: id. 9.71 (read: 6.71).’ (On 6.71 see further below—AR.) Jebb’s appeal
to punctuation to decide this matter involves, of course, a circular argument, there being no
way to punctuate independently from our syntactic analysis.
3 LSJ s.v. δέ II 2 b note, not very helpfully: ‘In the combination καὶ δέ (-), καὶ … δέ (-), each particle
retains its force.’ Schwyzer-Debrunner do not discuss καὶ (…) δέ.
adverb or connector? the case of και … δε 295

two sub-classes, according to the number of the words involved. In the second
part of my paper the use of καὶ … δέ to connect clauses and sentences will be
discussed.4

2 The Use of καὶ … δέ Connecting Single Words5

2.1 καὶ … δέ Connects Two Items | 189

(1) Ἄρξομαι δὲ ἀπὸ τῶν προγόνων πρῶτον· δίκαιον γὰρ αὐτοῖς καὶ πρέπον δὲ ἅμα
… τὴν τιμὴν ταύτην τῆς μνήμης δίδοσθαι (Th. 2.36.1)

I shall speak first of our ancestors for it is right and at the same time fit-
ting … to give them this place of honour. (Unless indicated otherwise, the
translations are those of the Loeb-series, sometimes with adaptations to
bring out the value of δέ.)

The words connected by καὶ … δέ are two predicative neuter adjectives.

Classen-Steup ad loc. render: ‘aber auch zugleich … geziemend’, referring to
1.132.4. There they had, in fact, observed: ‘In der Verbindung der Partikeln καὶ
… δέ (2, 36, 1; 4, 24, 2; 6, 71, 2; 7, 56, 3 … u. 8, 67, 3) ist καί nicht Kopula, son-
dern auch.’ But this is very unlikely. The point is that if καί in (1) is taken as an
additive adverb, and thus as optional,6 δέ must be capable of connecting the

4 I concentrate on Attic Greek and Herodotus, but will refer to some Homeric instances in
fn. 30. Καὶ … δέ is distributed rather unevenly; also, not seldom there is ms. variation or δέ
has been removed per coniecturam. The numbers for some Attic writers are (fragments not
included): Aeschylus: 9; Sophocles: 1 or 2 (at Ant. 432, a’s χἠμεῖς ἰδόντες δ᾽ may conceal χἠμεῖς δ᾽
ἰδόντες); Euripides: 4; Aristophanes: 3, all in Pax; Thucydides: 6; Plato: 35; Lysias: 2; Xenophon:
approx. 265; Demosthenes: 22; Theophrastus’ Characters approx. 60. For Herodotus the num-
bers are: 6; for Homer: 33. Homer has only the combination καὶ δέ, while in Herodotus and
Attic Greek only καὶ … δέ occurs. There is nothing exceptional about this. Similar discrep-
ancies are found with particle combinations like καὶ γάρ (the only combination occurring
in Homer), as against Attic καὶ γάρ and καὶ … γάρ. It is a priori not clear whether these for-
mal differences reflect a possible semantic or pragmatic difference. The uses of particles that
are known to both Homer and Attic Greek may, of course, vary considerably, some examples
being περ, καὶ μήν, μέντοι and οὖν. In principle, then, the fact that Attic Greek uses καὶ … δέ in
one way does not exclude that Homeric καὶ δέ may have been used in another way. But let me
add immediately that I do not think this is the case. See fn. 30.
5 This use is not known to Homer, who uses καὶ δέ only at clause and sentence level.
6 Cf. Engl. John, Peter and (also) Richard/and Richard (too). For optionality as a defining char-
acteristic of most adverbs see Pinkster (1972: 75–76).
 296 chapter 15

two adjectives all on its own. This, however, is not the case, as appears from the
unacceptability of ex. (2), both with δέ as ‘and’ and as ‘but’:7

(2) *δίκαιον πρέπον δέ

On the other hand, καί in (1) is, of course, capable of connecting the two adjec-
tives on its own, which strongly suggests that it is δέ that is an (optional) adverb
here. To be sure, δέ may connect two adjectives, but only as a ‘balancing adver-
sative’ (Denniston p. 165), viz. when the second adjective contrasts with an
implication of the first adjective, as in (3) and (4).8

(3) πονηρά γ᾽, ὦ Νεφέλαι, δίκαια δέ (Ar. Nu. 1462)

Ah Clouds, hard words, but they’re fair … (transl. Sommerstein)

(4) οὕτω δὴ τὰ μὲν στοιχεῖα ἄλογα καὶ ἄγνωστα εἶναι, αἰσθητὰ δέ (Pl. Tht. 202b)

Thus the elements are not objects of reason or of knowledge, but only of
perception

190 Otherwise, the connection must be established by καί, as in (5): |

(5) ταῦτα γὰρ δίκαια καὶ πρέποντα ἅμα τῇδε τῇ πόλει ἀποκρίνασθαι (Th. 1.144.2)

This answer is just and at the same time consistent with the dignity of the
city9

7 Strangely enough, this fundamental difference between καί and δέ, viz. that the latter in prin-
ciple is not suited to connect single words, is mentioned more or less in passing by Denniston,
in one sentence and a footnote on p. 162.
8 Such implications are mostly a matter of convention. They may be conventionalized to such
a degree that they are virtually an inherent semantic feature of the adjective involved. Thus,
poor always implies ‘not-happy’, as may be inferred from the fact that poor but happy is a
well-formed expression, while *poor but unhappy is not. However, they may also be of a more
informal, individualized, nature, as when someone says This wall is blue but (and yet) beauti-
ful. Apparently for this particular speaker ‘blueness’ implies ‘ugliness’.
9 Scholars often mention cases where καί seemingly connects items that are semantically
opposed, e.g. E. Med. 1243 τὰ δεινὰ κἀναγκαῖα … κακά. In a small but (and?) insightful mono-
graph, Blomqvist (1979) convincingly argues that this fact should not induce us to assume
the existence of a ‘καί adversativum’. An adversative particle is ‘eine Partikel, die nicht nur
Gegensätze verknüpft, sondern diese auch als Gegensätze bezeichnet’ (p. 56); cf. also Slings
(1980: 122) ‘… an adversative interpretation (of καί) can be suggested by the semantic rela-
adverb or connector? the case of και … δε 297

Now it might be objected that it is precisely the presence of καί which makes
the otherwise unacceptable δέ in (2) acceptable, on the analogy of e.g. English
and Dutch. In these languages combinations like that of (2) are equally unac-
ceptable: *justified but fitting, *gerechtvaardigd maar gepast; interestingly, how-
ever, the presence of also and ook turns them into fully acceptable ones: it is
justified, but also fitting; het is gerechtvaardigd, maar ook gepast. These combi-
nations probably counter a possible objection like Only justified?, Alleen maar
gerechtvaardigd?; but and maar may be considered ‘abbreviations’, so to speak,
of not only …, but also, niet alleen …, maar ook. Likewise the addition of adver-
bial καί might make the use of δέ to connect non-contrastive items acceptable.
Some commentators explain καὶ … δέ, in fact, in this way. Thus, Frohberger
remarks, in his Anhang to Lys. 19.5: ‘Die Verbindung … καὶ … δέ dient stets der
gradatio und betont den eingeschobenen Begriff (hier = οὐ μόνον ἐγὼ … ἀλλὰ
καὶ ὐμᾶς).’ However, we need not, and indeed cannot, follow this way out. First,
if the not only … but also-approach were correct (i.e. not just as an explana-
tory but as a grammatical solution) one would expect ἀλλὰ καί rather than καὶ
… δέ. Second, there sometimes may be gradatio involved, but this is definitely
not ‘stets’ the case.10 Third, and most important, there are examples of καὶ … δέ
where καί without any doubt is the connector, like:

(6) ἀλλ᾽ αἰνίττεται … καὶ οὗτος καὶ ἇλλοι δὲ ποιηταὶ σχεδόν τι πάντες ([Pl.] Alc. II
147b8)

But he, …, like almost every other poet, speaks in riddles

ἇλλοι ci. Burnet, ἄλλοι B : οἱ ἄλλοι T

Of course, the presence of preparative καί here is of prime importance for our
analysis: we must be dealing with corresponsive καὶ … καί, which means that,

tions of two members coordinated by καί’. Regrettably, Blomqvist does not discuss καὶ (…)
δέ. Incidentally, I do not think that in a case like Med. 1243 there is a semantic opposition.
By the use of καί the speaker/author deliberately presents the two qualifications as being
valid at the same time: for the speaker there is no contrast between them. Cf. an example
like Engl. This food is very unhealthy and very tasteful, where the standard implications of
unhealthy make us expect but very tasteful. See also fn. 8.
10 Thus, there is clearly no gradatio in the very example used by Frohberger, which runs:
ἀκούω γὰρ ἔγωγε, καὶ ὑμῶν δὲ τοὺς πολλοὺς οἶμαι εἰδέναι, ὅτι κτλ. The καί-clause (which
belongs to my second type, see below § 3) is a parenthesis rather than a highlighting clause.
Also, the presence of -γε with ἐγώ indicates that the speaker is primarily concerned with
himself.
 298 chapter 15

of καὶ … δέ, it is καί which is the connector. Another example of two nominal
elements connected by καὶ … καὶ … δέ is X. Smp. 2.9 (ἐν πολλοῖς μὲν … καὶ ἄλλοις
δῆλον καὶ ἐν οἷς δ᾽ ἡ παῖς ποιεῖ κτλ). Often, the two nouns are part of a series of
191 nouns | connected by καί (see §2.2 below), e.g. πελταστὰς πολλοὺς καὶ τοξότας
καὶ σφενδονήτας καὶ ἱππέας δὲ … in ex. (15).
But what, then, is the function of δέ in καὶ … δέ? Like Denniston, and like
K-G, for that matter, I think that this particle presents the item concerned as
distinct from the first item, not semantically, however, as we have seen, but
pragmatically.11 That is, while καί expresses the idea that the two items seman-
tically belong together (cf. fn. 11), δέ indicates that the (referent of the) second
item should be considered in its own right, and is, thus, at least as important
as the first item. I may perhaps quote here the words of Caroline Kroon in her
admirable study of a number of Latin particles, on the very similar Latin com-
bination et … et … autem: ‘The expression et … et … autem is used to emphasize
on the one hand the coordination of corresponding units of information, and
on the other the individuality and mutual discreteness of the coordinated ele-
ments’ (Kroon 1995: 231).12 On p. 229 she had argued that by the presence of
autem the ‘focus of attention shifts from A to B’. The eventual effect of δέ in a
καὶ … δέ group may be rendered in various ways, e.g. by ‘and on the other hand’,

11 Expressing ‘distinctness’ in the pragmatic sense may be the common function of all uses
of δέ: in καὶ … δέ, in apodosi, in combinations like ὁ μὲν … ὁ δέ, and at clause and sentence
level. In the latter use δέ is (syntactically) also a connector, with the semantic function
‘adds y to x’. By setting off the unit of information it introduces from the previous infor-
mation unit, δέ often not just adds information but implies that the two units contrast
qua content. This, in turn, often correlates with a shift in Topic or Theme. The difference
with καί. is perhaps that καί only connects and adds, and does not set off: the two units
should be taken as forming together one non-contrastive information unit. Asyndeton, on
the other hand, does not connect, but only sets off: the unit of information is presented as
having no formal links at all with what precedes. Thereby this information gains a certain
independence vis-à-vis the information that precedes; the addressee is invited not to look
backward, and to concentrate on the unit of information at hand. In practice, this means,
especially in a narrative, that asyndeton signals a point of special interest, often marking
the beginning of a new piece of narrative. Cf. Rijksbaron (1993). For the function of δέ see
also Bakker (1993) and Sicking & Van Ophuijsen (1993: 11f., 47).
12 Kroon calls this ‘mixture of (formal) conjunction and (semantic or pragmatic) disjunc-
tion’ ‘remarkable’; for καὶ … δέ, too, this qualification would seem more appropriate than
Denniston’s ‘this is a natural enough combination’. In formal pragmatic terms one might
say that in et … autem the two particles operate at different levels of discourse, et at the
representational, and autem at the presentational level, cf. Kroon (1995: 69 and 226ff.; for
these distinctions cf. also Wakker 1997: 211). The same analysis applies to καὶ … δέ. I should
perhaps mention here the possibility that καὶ … δέ, with adverbial δέ, fills the gap, so to
speak, created by the impossiblility of combining καί ‘and’ and καί ‘also’: *καὶ καί.
adverb or connector? the case of και … δε 299

‘and apart from that’, ‘and again’, ‘and further’. Often the effect would seem to be
the same as the effect conveyed by English ‘for that matter’, Dutch ‘trouwens,
overigens’ (= ‘apart from’): it anticipates a possible question on the part of the
addressee ‘what about item X?’, and indicates that the speaker is aware that that
item is relevant as well.
On the basis of this discussion I conclude that in (1), too, it is καί which is the
connector. As suggested above, the function of δέ is to individualize the second
item; it is perhaps best rendered by ‘for that matter’ or ‘on the other hand’.13 | 192
Another example of καὶ … δέ connecting predicative adjectives is:

(7) τὸ μὲν οὖν μὴ οὐχὶ ἡδέα εἶναι τὰ ήδέα λόγος οὐδεὶς ἀμφισβητεῖ· κακὰ δ᾽ ὄντα
αὐτῶν τὰ πολλὰ καὶ ἀγαθὰ δέ, ὡς ἡμεῖς φαμέν, ὅμως πάντα σὺ προσαγορεύεις
ἀγαθὰ αὐτά, … (Pl. Phlb. 13b)

Now of course nobody attempts to maintain the thesis that pleasant

things are not pleasant; but though they are in some cases (indeed in
most) bad and in others good … nevertheless you designate them all as
good, … (transl. Hackforth)

Again, δέ cannot possibly be taken as the connector, for the same reasons as
those set out above. I am not sure about the eventual effect of δέ in καὶ ἀγαθὰ
δέ; perhaps the addition serves, after the information that most pleasant things
are bad, as a kind of afterthought: for the sake of the argument Socrates con-
cedes explicitly that pleasant things are ‘on the other hand’ sometimes good,
‘sometimes’ being an effect of the presence of τὰ πολλά.

The next three examples all conform to the pattern set out above: καί cannot
be omitted, and δέ has an ‘on the other hand’ or ‘for that matter’ value.

(8) αὐτίκα γὰρ ἐδόκουν οἱ Ἕλληνες καὶ πάντες δὲ ἀτάκτοις σφίσιν ἐπιπεσεῖσθαι (X.
An. 1.8.2)

for the thought of the Greeks, and of all the rest in fact, was that he would
fall upon them immediately, while they were in disorder

13 This pragmatic function correlates with the presence of a syntactic boundary before the
item followed by δέ (in (1): πρέπον), a boundary which is due to δέ being a postpositive. For
the function of δέ and other postpositives as boundary markers see Ruijgh (1996: 629ff.),
Dik (1995: 35 ff.). As for the relationship between this use of δέ, and that of αὖ ‘on the other
hand’, ‘further’, see below p. [195].
 300 chapter 15

(9) εἰ μέντοι ῥήτωρ ἐστὶν οἵους ἐνίους τῶν λεγόντων ἐγὼ καὶ ὑμεῖς δ᾽ ὁρᾶτε … οὐκ
ἂν εἴην οὗτος ἐγώ (D. 21.189)

But if by orator he means one of those speakers such as I and you, for that
matter, often see, … I cannot be one (‘for that matter’ added)14

(10) ᾤμην … προσήκειν Εὐβουλίδῃ, καὶ πᾶσιν δ᾽ ὅσοι … κατηγοροῦσιν, … (D. 57.4)

I should have thought … that it was fitting for Eubulides and for all
those, for that matter, who are … making accusations … (‘for that matter’
added)15

193 καὶ … δέ may also connect two predicates, cf.: |

(11) ἐγὼ νὴ Δί᾽ ἐρῶ, καὶ γράψω δέ, ὥστ᾽ ἂν βούλησθε χειροτονήσετε. (D. 9.70)

I solemnly promise that I will answer it and will also move a resolution,
for which you can vote, if so disposed

Again, connection by just δέ is not possible: *ἐρῶ γράψω δέ, while the sequence
ἐρῶ καὶ γράψω is perfectly viable. This means once more that δέ is optional and
should be taken as an adverb. Both ‘for that matter’ and ‘moreover’ would seem
suitable ‘translations’ to make clear its pragmatic function.

2.2 καὶ … δέ Connects Two Items in a Series of Three or More Items

The examples of this section do not differ basically from those discussed above,
but the connective function of καί is even clearer, since the other items are usu-
ally preceded by καί as well. The καὶ … δέ-phrase is always the last item of the
series. Let me repeat that in none of these cases καὶ can be omitted. In all cases
the function of δέ is to set off the item it modifies from the previous items.

(12) … περὶ τὰ ὑγιεινὰ καὶ νοσώδη ἐθελῆσαι ἂν φάναι μὴ πᾶν γύναιον καὶ παιδίον,
καὶ θηρίον δέ, ἱκανὸν εἶναι ἰᾶσθαι … (Pl. Tht. 171e)

14 In their index to Neun Philippische Reden, Leipzig 1886, s.v. καὶ … δέ, Rehdantz-Blass
have collected a great number of instances from Demosthenes. Like Frohberger (see
above), they take them all as ‘aber auch’, ‘mit Betonung des dazwischen stehenden Wortes’.
Strangely enough, MacDowell in his commentary on the In Meidiam not even mentions
the difficulties involved in καὶ … δέ.
15 Other examples are: ἐκπίπτει ὁ Ἀβραδάτας καὶ ἄλλοι δὲ … (‘Abradatas and others were
adverb or connector? the case of και … δε 301

… he might be willing to say that in matters of health and disease not every
woman or child—or beast, for that matter—is able to cure itself (‘for that
matter’ is present in the Loeb-translation)

Since we are dealing here with a series of ever more unlikely items, the καὶ
θηρίον δέ phrase being the last in the series, the status of this phrase can perhaps
also be indicated by rendering ‘let alone a beast’, the ‘let alone’-effect being due
to the presence of the negative with the first item.

(13) πρῶτος δὲ εἴσεισιν ὁ τὸ σταδίον ἁμιλλησόμενος σὺν τοὶς ὅπλοις, δεύτερος δὲ ὁ

τὸν δίαυλον, καὶ τρίτος ὁ τὸν ἐφίππιον, καὶ δὴ καὶ τέταρτος ὁ τὸν δόλιχον, καὶ
πέμπτος δὲ ὃν ἀφήσομεν πρῶτον ὡπλισμένον … (Pl. Lg. 833a–b)

First, then, there shall enter the man who, with his arms, is to run the
furlong,—second, the runner of the quarter-mile,—third, the half-
miler,—fourth, the runner of the three-quarters,—and fifth, the runner
in the long-distance race whom we shall despatch first, fully armed

In this example the δέ after δεύτερος sets off all the items that are to follow from
the first one, the σταδιοδρόμος, which had already been given a separate treat-
ment in the preceding sentence. Within the group set off by δέ, the connection
is brought about by καί, while δέ individualizes the last item. Obviously, there
is no question here of a ‘for that matter’ effect, since the last item is the most
important for the description that follows. | 194

(14) παρεῖχε δὲ ἡ θεὸς ἄλφιτα, …, καὶ τῶν θυομένων … λάχος, καὶ θηρευομένων δέ.
καὶ γὰρ θήραν ἐποιοῦντο … (X. An. 5.3.9)

And the goddess would provide for the banqueters barley meal … and a
portion of the sacrificial victims … as well as of victims taken in the chase.
For (they) used to have a hunting expedition …

Both τῶν θυομένων and θηρευομένων are partitive genitives with λάχος, con-
nected by καί.

thrown to the ground …’, X. Cyr. 7.1.32), ὦ Ὑστάσπα, καὶ οἱ ἄλλοι δὲ οἱ παρόντες (‘Hystaspes
and you others here’, X. Cyr. 8.4.17); Pl. Lg. 862c.
 302 chapter 15

(15) … ὁρῶντι δὲ πελταστὰς πολλοὺς καὶ τοξότας καὶ σφενδονήτας καὶ ἱππέας δὲ …
(X. An. 5.6.15)

… (his) eyes rested upon … and likewise upon a great body of peltasts,
bowmen, slingers, and horsemen also …

Here, too, the last item is set off from the rest; and naturally so, for the first three
groups all operate on foot.

(16) … ποιήσειν … καὶ ἀνέμους καὶ ὕδατα καὶ ὥρας καὶ ὅτου δ᾽ ἂν ἄλλου δέωνται τῶν
τοιούτων. (X. Mem. 1.1.15)

(do they imagine) that they will create winds, waters, seasons and such
things to their need …

‘Setting off’ would also seem to be the function of δέ in the next examples.

(17) εἰ μὴ … σαφῶς διώρισεν, τί πρῶτον δεῖ ποιεῖν τοὺς ἁντιδεδωκότας καὶ τί δεύτε-
ρον καὶ τἄλλα δ᾽ ἐφεξῆς, οὐκ οἶδ᾽ … ([D.] 42.1)

For if he had not clearly defined for us what is the first thing to be done
by those who have offered an exchange, and what the second, and so on
in due order, I do not know …

(18) εἰπεῖν μὲν οὖν μοι ταῦτα πρὸς τὸν Ἀρχεβιάδην καὶ τὸν Ἀριστόνουν καὶ πρὸς
αὐτὸν δὲ τὸν Κηφεσιάδην οὐδὲν διαφέρει ([D.] 52.11)

to suggest what you propose to Archebiades and Aristonoüs and to Ceph-

esiades himself, for that matter, can cause me no trouble (‘for that matter’
added)

Now syntactically, an example like (18) does not differ from an example like
(19), which exhibits the combination καὶ … δή:

(19) σκεψώμεθα δὴ … ποῖά ἐστιν ἃ ἡμᾶς ὠφελεῖ. ὑγίεια, φάμεν, καὶ ἰσχὺς καὶ κάλλος
καὶ πλοῦτος δή (Pl. Men. 87e)

Then let us see … what sort of things they are that profit us. Health, let us
195 say, and strength, and beauty, and wealth … |
adverb or connector? the case of και … δε 303

or from (20), featuring καὶ … γε:

(20) ἕτεραι δέ γέ εἰσι τῶν τεχνῶν αἳ διὰ λόγου πᾶν περαίνουσι, καὶ ἔργου … ἢ οὐδένος
προσδέονται ἢ βραχέος πάνυ, οἶον ἡ ἀριθμητικὴ καὶ λογιστικὴ καὶ γεωμετρικὴ
καὶ πεττευτική γε καὶ ἄλλαι πολλαὶ τέχναι (Pl. Grg. 450d)

But there is another class of arts which achieve their whole purpose
through speech and … require either no action to aid them, or very lit-
tle; for example, numeration, calculation, geometry, draught-playing, and
many other arts …

or, again, from καὶ … αὖ in:

(21) ἀλλὰ μὲν δὴ ἥ γε ἀρετὴ ἑκάστου, καὶ σκεύους καὶ σώματος καὶ ψυχῆς αὖ καὶ
ζῴου πάντος, … (Pl. Grg. 506d)

But surely the virtue of each thing, whether of an implement or of a body,

or again of a soul or of any live creature, …

On καὶ … δή Denniston remarks, p. 254: ‘Καί … δή joins sentences, clauses, and

single words’, e.g. ‘at the end of a catalogue’, the passage from Meno being an
example. And on p. 253 we read: ‘These combinations (viz. καὶ … δή and καὶ δὴ
καί) signify that the addition made by καί is an important one. They thus dif-
fer from καὶ … γε, which merely stresses the fact that an addition is made.’ One
wonders why Denniston did not include καὶ … δέ in his discussion. I believe,
incidentally, that the values of δή and γε are different. In a case like (19), δή
would seem to express the idea that the addition of the item it modifies will
not come as a surprise to the addressee.16 Again, γε in καί … γε rather indicates
that to the list which exemplifies the statement of the main clause the item
concerned certainly belongs, an effect of the limitative value: ‘if anything else,
then certainly this item’: πεττεία being a game, there is no ἔργον whatsoever
involved in ἡ πεττευτική.17

16 Cf. Van Ophuijsen’s adaptation of Brugmann-Thumb’s definition (1993: 141, footnote 2): ‘δή
wies auf dasjenige was von dem Sprechenden als dem Hörenden gegenwärtig Vorliegen-
des … vorgestellt wird hin’. At Men. 87e δή ‘no doubt contains sarcasm’, as Bluck ad loc.
observes.
17 Dodds explains γε as: ‘ “yes and …”, adding an unexpected item to the enumeration’, but
this seems rather ad hoc.
 304 chapter 15

As for καὶ … αὖ (and καὶ αὖ), which contains the adverb that expresses par
excellence ‘on the other hand, in turn’ and related nuances, one may well ask in
what respect it differs from καὶ … δέ. In general, it would seem that αὖ expresses
a more marked contrast than καὶ … δέ. Thus, in (21) αὖ contrasts two sets, each
consisting of two items, that are not just opposed pragmatically but semanti-
cally. In fact, καὶ (…) αὖ, unlike καὶ … δέ, often connects antonyms, e.g. φιλόλογος
… καὶ αὖ μισόλογος (Pl. La. 188c), χρήσιμόν τε καὶ ὠφέλιμον καὶ ἄχρηστον αὖ καὶ
196 βλαβερόν (R. 518e).18 See also p. [206], καὶ αὖ … ὡσαύτως. |

I have discussed this first group of examples together because they share one
crucial syntactic characteristic: of the two particles present δέ can be left
out, just as δή and γε in the combinations καὶ … δή and καὶ … γε.19 In other
words, again just as in the latter combinations, it is καί which is the connec-
tor.
With respect to my second group the situation is altogether different, since
both καί and δέ may, in principle at least, each on their own operate as con-
nectors at clause and sentence level. A further, pragmatic, problem is that,
whether δέ is taken adverbially or as a connector, in both cases it would mark
a Topic shift; cf. fn. 11 and see further below. Yet the syntactic similarity with
καὶ … δέ of the first group strongly suggests that καί is the connector. Also, in
many cases the context simply excludes the meaning ‘and/but also’. And let me
add that on the basis of the results achieved so far I considered it a matter of
methodological soundness to try and analyse these cases of καὶ … δέ in the same
way.

3 καὶ … δέ Connects Clauses or Sentences

3.1 Preceded by τε or καί

My first examples contain the correlative construction τε … καὶ … δέ. As in the
case of the combination καὶ … καὶ … δέ discussed above, the presence of τε indi-
cates that καί should be taken as the connector. Consider (22) and (23):

18 In neither case could αὗ be replaced by δέ: apart from semantic considerations καὶ δέ does
not occur in Attic, and καὶ … δέ never contrasts sets, but only single items.
19 In (21) αὖ cannot be left out as easily, probably because, as I argued above, the items are
semantically opposed. Without αὖ there would be a suggestion that ψυχή belongs to the
same category as σκεῦος and σῶμα. This effect is even clearer in φιλόλογος … καὶ αὖ μισόλο-
γος, or χρήσιμόν τε καὶ ὠφέλιμον καὶ ἄχρηστον αὖ καὶ βλαβερόν, quoted above. Unlike δέ (cf.
fn. 12), αὖ probably operates at the representational level.
adverb or connector? the case of και … δε 305

(22) καὶ τὸν πόλεμον αὐτόθεν ποιεῖσθαι οὔπω ἐδόκει δυνατὸν εἶναι, πρὶν ἂν ἱππέας
τε μεταπέμψωσιν ἐκ τῶν Ἀθηνῶν καὶ ἐκ τῶν αὐτόθεν ξυμμάχων ἀγέιρωσιν …,
καὶ χρήματα δὲ ἅμα αὐτόθεν τε ξυλλέξωνται καὶ παρ᾽ Ἀθηναίων ἔλθῃ, τῶν
τε πόλεων τινας προσαγάγωνται …, τά τε ἄλλα … παρασκευάσωνται … (Th.
6.71.2)

… and it seemed as yet impossible to carry on the war from this base until
they should send to Athens for horsemen, besides collecting them from
their allies in Sicily … And they wanted at the same time to collect money
from the island itself and to have a supply come from Athens; also to bring
over some of the cities …; and to prepare other things

Classen-Steup remark: ‘ἱππέας τε steht erst zu καὶ χρήματα δέ in Beziehung;

wegen der entfernteren Verbindung wird das καί zu καὶ … δέ verstärkt …’20 | 197
This is a particularly revealing example, because of the balanced structure
of the sentence, a balance brought about by the subdivision of the temporal
πρίν-clause into four members, marked by τε …, καὶ … δὲ ἅμα, … τε and again
τε (printed in bold type). Observe that Classen-Steup’s analysis of this exam-
ple is quite different from what they suggest in their general note on καὶ … δέ
at 1.132.4, also quoted above: ‘In der Verbindung der Partikeln καὶ … δέ … ist
καί nicht Kopula, sondern auch.’ The change of opinion can be easily under-
stood, for the coordination with the three τε’s makes it virtually impossible
not to consider καί here the connector.21 Their view that δέ has been added
to ‘strengthen’ καί because of its distance from τε is entirely ad hoc. Rather,

20 Two other examples of τε … καὶ … δέ are: ὁ δὲ Κυαξάρης … αὐτός τε ἐμεθύσκετο …, καὶ τοὺς
ἄλλους δὲ Μήδους ᾤετο παρεῖναι ἐν τῷ στρατοπέδῳ … (‘Cyaxares … got drunk; and he sup-
posed that the rest of the Medes were all in camp …’, X. Cyr. 4.5.8), and: καὶ οἵ τε ἄλλοι
προθύμως τῷ Τελευτίᾳ ὑπηρέτουν … καὶ ἡ τῶν Θηβαίων δὲ πόλις … προθύμως συνέπεμψε καὶ
ὁπλίτας καὶ ἱππέας (‘And all the states gave their hearty support to Teleutias … while the
Theban state in particular … eagerly sent with him both hoplites and horsemen’, X. HG
5.2.37). In Pl. Cri. 48b: οὗτός τε ὁ λόγος … ἔμοιγε δοκεῖ ἔτι ὅμοιος εἶναι καὶ πρότερον· (om. βδ)
καὶ τόνδ᾽ αὗ σκόπει, there is a ‘slight anacoluthon’, as Burnet ad loc. observes, ‘since the καί
(b4) introduces a changed construction’. Burnet is silent on δέ. Note the presence of αὗ
reinforcing δέ, also found in e.g. καὶ ἄλλο δέ τι αὗ, X. Cyr. 4.5.46.
21 It might be argued that, in (22), (χρήματα) δέ corresponds with (ἱππέας) τε. However, this
is unlikely in view of the position of δέ. For δέ to correspond with τε one would rather
expect the order ἅμα δὲ καί, as in: κατά τε τὸ δίκαιον ὑπεδέξαντο τὴν τιμωρίαν, …, ἅμα δὲ καὶ
μίσει τῶν Κερκυραίων (‘The Corinthians undertook the task, partly on the ground of right
… partly also through hatred of the Corcyraeans’), Th. 1.25.3. On τε … δέ see further K-G (2,
244 Anm. 3), Denniston p. 513.
 306 chapter 15

δέ should be given the ‘on the other hand’ or ‘for that matter’ meaning dis-
cussed above; it explictly sets off χρήματα from ἱππέας, and marks, thus, a Topic
shift.22
The same effect is present in (23), an example of καὶ … καὶ … δέ:

(23) καὶ ὁ πατὴρ ἀεὶ λέγει καὶ σὺ φῂς καὶ οἱ ἄλλοι δὲ πάντες ὁμολογοῦσιν, ὡς … (X.
Cyr. 3.3.19)

And my father always says, and so do you, and all the rest agree, that …

Note also that καί cannot be omitted, while δέ can.

3.2 Other Cases

The example from Sophocles (without preceding τε) that played such an impor-
tant role in Denniston’s discussion of καὶ … δέ, should be analysed in the same
198 way: |

(24) οἷς γὰρ ἡ γνώμη κακῶν

μήτηρ γένηται, πάντα παιδεύει κακούς.
καὶ σοῦ δ᾽ ἔγωγε θαυμάσας ἔχω τόδε.
χρῆν γάρ σ᾽ κτλ. (S. Ph. 1360ff.)

For when the mind hath once become a parent of evil, it teaches men
to be evil henceforth. And in thee, too, this conduct moves my wonder
(Philoctetes addressing Neoptolemus; transl. Jebb)

We have seen above that Jebb takes καί here as ‘also’, and δέ as ‘and’; he adds that
‘there is an emphasis on the intervening word.’ But this interpretation does not
suit the passage, for it entails that Philoctetes has voiced amazement in the pre-
ceding lines, which is not the case. In those lines he is tormented by the thought
of having to deal with his enemies, the Atreidae and Odysseus. I think that καὶ

22 Besides καὶ … δὲ ἅμα Thucydides also uses ἅμα δὲ καί to connect clauses and sentences,
as in: κατεσκευάσαντο δὲ … πολλοὶ … ἅμα δὲ καὶ τῶν πρὸς τὸν πόλεμον ἥπτοντο (‘Many also
established themselves … And while all this was going on, (the Athenians) applied them-
selves to the war …’, Th. 2.17.4). Here, δέ must be taken as the connector, since καί in this
position cannot possibly function as such; naturally, therefore, this combination is not
found between single words, as in ex. (1). The difference with καὶ … δὲ ἅμα at clause level,
as in (22), probably is that in 2.17.4 καί has the whole clause that follows in its scope, while
in (22) the emphasis is primarily on χρήματα. Similar differences are found with καὶ … δὲ
ὡσαύτως, see the Appendix.
adverb or connector? the case of και … δε 307

… δέ signals that Philoctetes now turns to Neoptolemus himself: ‘And as for you,
I am amazed at your present behaviour’, viz. of trying to persuade him to go to
Troy. The ‘emphasis on the intervening word’ is not due, then, to καί but to δέ.
As in examples (22) and (23), δέ is here a marker of Topic shift.23
The same effect of καὶ … δέ is present in:

(25) εὖ οὖν ἴσθι ὅτι ἐγὼ ταῦτα ἀκούων χαίρω ὅτι εὐδοκιμεῖς, καὶ σὺ δὲ ἡγοῦ με ἐν
τοῖς εὐνούστατόν σοι εἶναι (Pl. La. 181b)

So let me tell you that I rejoice to hear this and to know you have such a
good reputation; and you in return must count me as one of your warmest
well-wishers (Lysimachus addressing Socrates)

Again, καὶ σὺ δέ cannot possibly mean ‘and you too (must count me as one of
your warmest well-wishers)’, because there has been no mention, in the pre-
ceding clause, of someone else who considered the speaker (Lysimachus) well-
wishing. Rather, the construction is as follows. By its position ἐγώ approaches
ἐγὼ μέν, and we expect a second verb governed by ὅτι. But then the construc-
tion changes, and there starts an independent clause, introduced by καί, which
connects ἡγοῦ with ἴσθι. At the same time δέ causes σύ to contrast with ἐγώ in
the ὅτι-clause. Lamb’s ‘in return’ captures the effect pretty well. | 199
The non-connective status of δέ may seem less apparent in a case like (26):

(26) ΣΩ.Ὅταν ἄρα δυοῖν καλοῖν θάτερον κάλλιον ᾖ, ἢ τῷ ἑτέρῳ τούτοιν ἢ ἀμφοτέ-
ροις ὑπερβάλλον κάλλιόν ἐστιν, ἤτοι ἡδονῇ ἢ ὠφελίᾳ ἢ ἀμφοτέροις.
ΠΩΛ. Πάνυ γε.
ΣΩ. Καὶ ὅταν δὲ δὴ δυοῖν αἰσχροῖν τὸ ἕτερον αἴσχιον ᾖ, ἤτοι λύπῃ ἢ κακῷ
ὑπερβάλλον αἴσχιον ἔσται·

23 A similar Topic shift marked by καὶ … δέ occurs e.g. at E. El. 1117: ΗΛ. τί δ᾽ αὗ πόσιν σὸν
ἄγριον εἰς ἡμᾶς ἔχεις; | ΚΛ. τρόποι τοιοῦτοι· καὶ σὺ δ᾽ αὐθάδης ἔφυς (‘And me—why let your
husband treat me so fiercely? Kl. That is his way; you too are self-willed, you know’, transl.
Cropp). Denniston ad loc. notes: ‘καὶ … δέ is rare in tragedy. Here, and in some other places,
the combination is only an apparent one, δέ being the connective and καί meaning “also”:
“and you too” (Particles, p. 200).’ But this is impossible, for ‘too’ only would make sense if
ἄγριος and αὐθάδης were synonymous. For the same reason Cropp’s translation should be
rejected: Aegisthus is not αὐθάδης. We should rather interpret: ‘and you, in turn, …’ Cf. also
exx. (25) and (28) below.—In Theophrastus’ Characters καὶ … δέ is the means par excel-
lence to achieve a Topic shift, and thus to structure the description of a given type, cf. e.g.
23 Ἀλαζονεία (3: καὶ συνοδοιπόρου δέ, 4: καὶ γράμματα δέ, 6: καὶ ἀγνώτων δέ, 7: καὶ προσελθὼν
δ᾽). Note that Diels, in the Index verborum to his edition, takes καί in καὶ … δέ as = et, not
as = etiam.
 308 chapter 15

So.: Thus when of two fair things one is fairer, the cause is that it sur-
passes in either one or both of these effects, either in pleasure, or in
benefit, or in both.
Pol.: Certainly.
So.: And again, when one of two foul things is fouler, this will be due to
an excess either of pain or of evil (Pl. Grg. 475a)

‘And again’ is Lamb’s translation, which suggests that he takes καί as the con-
nector. I think, in fact, that καί expresses the idea that Socrates continues the
same line of reasoning, while δέ signals that this procedure is now applied to a
different item; the ὅταν-clause serves to introduce the new Topic.24 One might
be tempted, perhaps, to take δέ as the connector, with καί as = ‘also’: ‘and/but
also when’. Thereby, however, δέ would be given the position of a sentence con-
nector, although it occurs in a subordinate clause. There are, to be sure, some
instances where γάρ is used in this way (e.g. κεἰ μὴ γὰρ ἔστίν ὁ θεὸς οὗτος, … |
παρὰ σοί λεγέσθω (E. Ba. 333)), but this construction is exceptional, and proba-
bly confined to fixed combinations like καὶ εἰ (μή); cf. E. Pho. 946, Ar. Eccl. 888.25
Be that as it may, there are no examples of καὶ Subordinator δέ where δέ must be
taken as the connector; see Hdt. 4.68.4 (καὶ ἢν δέ (καί om. d, Hude), correspond-
ing with καὶ ἢν μέν), X. Cyr. 8.2.25 (καὶ ὁπότε δέ; actually καί = ‘also’ excluded), Pl.
R. 436e (καὶ ὅταν δέ (καί Galenus, om. AFDM)), [Pl.] Erx. 400e (καὶ ὅσα δέ; καί =
‘also’ excluded). The status of δέ in e.g. καὶ ὅσα δέ is not different from that of αὖ
in καὶ ὅσα αὖ, e.g. Th. 1.91.5. A particularly clear example of the adverbial use of
δέ in a dependent construction is:

(27) ταῦτα γὰρ οὐχ ὥσπερ εἰ …, καὶ εἰ νόσου δὲ πολεμίοις ἐμπεσούσης κρατήσειεν,
200 εὐτυχέστερος μὲν ἂν εἴη, στρατηγικώτερος δὲ οὐδὲν ἄν· | (X. Ages. 10.1)

These are the marks that distinguish him (: Agesilaus) from … or from the
man who wins victory through an outbreak of sickness among the enemy,
and adds to his success but not to his knowledge of strategy

24 Observe that while καί connects the sentence as a whole with the preceding sentence, δέ
operates at the level of the temporal clause only, contrasting this clause with the previous
ὅταν-clause. Bakker (1993: 287ff.) believes that also in cases like ὡς δὲ ὥρα … ἑωυτὸν ἐμακά-
ρισε (Hdt. 7.45), δέ just belongs to the temporal clause. But this is impossible; in accordance
with Wackernagel’s Law δέ appears in second position, but it is, of course, still a sentence
connector. To put it otherwise: in (26) δέ appears in the second position of the clause, in
ὡς δὲ ὥρα … ἑωυτὸν ἐμακάρισε in the second position of the sentence.
25 More common are καὶ γὰρ εἰ (e.g. thrice in Plato, 5× in Isocrates, 20× in Demosthenes), εἰ
γὰρ καί (e.g. once in Herodotus, 4× in Plato, 3× in Demosthenes). I have found no exam-
adverb or connector? the case of και … δε 309

Here δέ exceptionally modifies a genitive absolute within a καὶ εἰ-clause. Of

course, δέ cannot operate as a connector here, since this would amount to
ignoring two syntactic boundaries, viz. those effected by εἰ and by the genitive
absolute.
The position between καί and δέ may also be taken by a participium coni-
unctum, cf.:

(28) καὶ θέλων δὲ τούτων πέρι σαφές τι εἰδέναι …, ἔπλευσα καὶ ἐς Τύρον … (Hdt.
2.44.1)

Moreover, wishing to get clear knowledge of this matter …, I took ship to

Tyre …

Abicht ad loc.: ‘Auf dem zwischen καί und δέ eingeschobenen Worte liegt
… der Ton.’ Stein ad loc.: ‘Nicht verschieden von θέλων δὲ καί. Das im Attis-
che geläufige καὶ—δέ zur Betonung des Parentheton findet sich bei H.
noch nicht.’

This passage is preceded by two sentences introduced by so-called ‘progressive’

καὶ μέν, and followed by a sentence introduced by καὶ εἶδον. This definitely sug-
gests that in our example καί is the connector. From Abicht’s remark it does
not appear whether he takes καί as = ‘also’. Be that as it may, this meaning is
excluded for the same reasons as those mentioned above in the discussion of
examples (24) and (25). To arrive at ‘also’, one would expect Herodotus to have
spoken about willing other things in the preceding context, but this is not the
case.26 Here, too, καὶ … δέ marks a shift, in this case of the ‘setting’ or ‘attendant

ples of καὶ ὅταν/ἐπειδὰν (ὅτε/ἐπειδὴ) γάρ. Instead, one finds καὶ γὰρ ὅταν (Pl. R. 343e), etc.
I should perhaps add that it is far from certain whether καί in καὶ γὰρ εἰ, καὶ γὰρ ὅταν etc.
always means ‘also’; καὶ (…) γάρ is probably the most Protean of all particle combinations.
26 It might be argued that in (26) (and elsewhere) the focus of καί is not the word that follows
but the clause as a whole: ‘and also’ = ‘and further’. But such an approach to καὶ … δέ runs
counter to the syntactic considerations mentioned above, which indicate that it is καί that
is the connector. Nor are there parallels for such a use of καί in other combinations. To be
sure, καί in καὶ γάρ may be used in this way (e.g. X. An. 1.1.8), but this does not occur with
καὶ … γάρ separated, where in general the focus is on the word following καί. For all that,
‘and further’ for καὶ … δέ is often not inappropriate, but in such cases the ‘further’ nuance
is conveyed by δέ, not by καί. Notice, in this connection, that Denniston renders both καὶ
γάρ (p. 109, in some uses) and καὶ τοίνυν (p. 565) by ‘and further’, in both cases with καί
as the connector. Incidentally, part of the confusion created by Denniston’s rendering of
καὶ … δέ by ‘and also’, both when he takes καί and when he takes δέ as the connector (see
section 1), may be due to his not distinguishing the different ‘also’s’ of English, for which
cf. Quirk et al. (1985: 604 ff.).
 310 chapter 15

circumstances’: after the speculative reasoning of the two καί-clauses in §§ 2–3,

καὶ … δέ marks the transition to a more practical step in Herodotus’ argument
concerning the name ‘Herakles’. As for Stein, I wonder why Herodotus did not
201 simply use δὲ καί, if it all amounted to the same.27 A more difficult example is: |

(29) Ἰὼ Πρασιαὶ … | ὡς ἀπολεῖσθε τήμερον.

…
Ἰὼ Μέγαρα Μέγαρ᾽, ὡς ἐπιτρίψεσθ᾽ αὐτίκα
…
Ἰὼ Σικελία, καὶ σὺ δ᾽ ὡς ἀπόλλυσαι (Ar. Pax 243–244, 246, 250)

Oho Prasiae … how you’ll be murdered today!

Ho Megara, Megara, how you will be crushed presently …
Oho Sicily, what a doom is yours too! (transl. Sommerstein)28

In view of the preceding lines, it is tempting, of course, to take this καί adver-
bially and δέ as the connector. However, on the analogy of καὶ σοῦ δέ in (24) and
καὶ σὺ δέ in (25) (also E. El. 1117 (note 23)), as well as καὶ ὑμῖν δέ, καὶ ὑμᾶς δέ in (37)
and (38) below, I think that, once again, καί should be taken as the connector,
just as καί is the connector in καὶ … γε in similar greeting and farewell formulas,
e.g.:

(30) Καὶ χαῖρε πόλλ᾽, ὦδελφέ. :: ΔΙ. Νὴ Δία καὶ σύ γε ὑγίαινε. (Ar. Ra. 164)

And fare thee well, my brother. :: Di. And to you good cheer. (transl.
Rogers)

(31) ὁ δὲ Κροῖσος ὡς εἶδε τὸν Κῦρον, Χαῖρε, ὦ δέσποτα, ἔφη. (–) Καὶ σύ γε, ἔφη, ὦ
Κροῖσε (X. Cyr. 7.2.10)

And when Croesus saw Cyrus, he said: “I salute you, my sovereign lord.
(–)” “And I you, Croesus”

27 Observe that in Ξέρξης δὲ … δρησμὸν ἐβούλευε· θέλων δὲ μὴ ἐπίδηλος εἶναι … χῶμα ἐπειρᾶτο
διαχοῦν … (Hdt. 8.97.1) the δέ-sentence contrasts not just pragmatically, but semantically
with the preceding sentence. Here, καὶ … δέ would seem excluded.
28 Van Leeuwen on Pax 250 only comments that καὶ … δέ, which, according to him, often has
dubious ms authority in Aristophanes, ‘hoc certe loco … suspicione vacat’. Platnauer has
no remark at all.
adverb or connector? the case of και … δε 311

Cf. also Ar. Lys. 6, Pax 718, Eccl. 477. It should also be observed that in (29), as
in many other cases as we have seen, καί cannot be omitted, while δέ can. This
again strongly points to δέ not being a sentence connector, but an adverb: ‘And
you, on the other hand/for that matter’.29

3.3 καὶ … δέ = ‘and also/too’?

With regard to example (25) I observed that ‘and you too’ for καὶ σὺ δέ is
excluded, given the lack of someone else who considered Lysimachus a well-
wisher; a similar reasoning applies to e.g. καὶ σοῦ δ᾽ in (24) and to examples (37)
and (38) below. In more general terms one might say that for καί = ‘also’ to be
acceptable the | states of affairs of the two clauses or sentences must be iden- 202
tical or similar, while the entities involved are different. Indeed, this explains
why in example (29) ‘and you too’ is in principle fully acceptable: ὡς ἀπόλλυ-
σαι repeates, so to speak, ὡς ἐπιτρίψεσθ’ of line 246, and καὶ … δέ expresses the
idea that the state of affairs mentioned in the preceding sentence applies to
a second item as well.30 There are many comparable passages where our first
inclination will be, I think, to take καί as = ‘also’. Consider:

(32) ἀπιόντα δὲ τὸν Κῦρον προὔπεμπον ἅπαντες … καὶ οὐδένα ἔφασαν ὅντιν᾽ οὐ
δακρύοντ᾽ ἀποστρέφεσθαι. καὶ Κῦρον δὲ αὐτὸν λέγεται σὺν πολλοῖς δακρύοις
ἀποχωρῆσαι (X. Cyr. 1.4.26)

29 In view of the syntactic similarity I am inclined to follow the same reasoning at Pax 523 (ὦ
χαῖρ᾽ Ὀπὠρα, καὶ σὺ δ᾽ ὧ Θεωρία). Another possibility would be to take both καί and δέ as
adverbs, e.g. as in Engl.: ‘you too, for that matter’. But this would leave the clause/sentence
with no connecting particle at all.
30 Note that the wording of the καὶ … δέ-sentence is not an exact duplicate of the preceding
sentence. In fact, this is very rare, mostly there is some degree of variation, cf. e.g. (32) and
(34) below. There are several instances of this type in Homer, e.g. οὗτος γὰρ δὴ ὀνήσει ἐϋκνή-
μιδας Ἀχαίους, | καὶ δ᾽ αὐτὸς ὃν θυμὸν ὀνήσεται, αἴ κε φύγῃσι (‘He shall be the one to gladden
the strong-greaved Achaians, and to be glad within his own heart, if he can come off’, Il.
7.173; transl. Lattimore). Some other examples are Il. 12.272, 13.73, Od. 6.60, and, without
αὐτός or ἄλλος, Il. 7.113, 375, Od. 13.302. Whenever Ameis-Hentze comment on such cases
they always interpret καί as = ‘also’. Kühner-Gerth, on the other hand, always take δέ as
adverbial and καί as the connector. I think K-G are right. Thus, in Il. 7.173 καὶ δ᾽ means ‘and
apart from that, and moreover’; ‘and, for that matter’ is also possible. The point is that,
just as in Attic Greek, there are many cases where καὶ δέ cannot possibly be taken as ‘and
also’, with the emphasis on the following word, e.g. Il. 17.351, 19.178, 20.28, Od. 4.391, 13.302.
In the last three cases A-H were aware of the problem, for there they take καί at sentence
level, e.g. Il. 20.28 καὶ δέ τέ μιν καὶ πρόσθεν ‘das erste καί (bezieht) sich auf dem ganzen Satz’
(similarly Edwards ad loc.), but this is highly unlikely, cf. footnote 26. It should be noted,
finally, that καί in the combination καί τε is taken also as = ‘also’ by A-H, whereas in all
probability it is the connector, τε being adverbial; cf. Ruijgh (1971: 764ff.).
 312 chapter 15

And everybody … escorted him …, and they say that there was no one
who turned back without tears. And Cyrus also, it is said, departed very
tearfully

(33) Δοκεῖ γοῦν ἔφη ὁ Κριτόβουλος, οἰκονόμου ἀγαθοῦ εἶναι εὖ οἰκεῖν τὸν ἑαυτοῦ
οἶκον. Ἦ καὶ τὸν ἄλλου δὲ οἶκον, ἔφη ὁ Σωκράτης, εἰ ἐπιτρέποι τις αὐτῷ, οὐκ ἂν
δύναιτο, εἰ βούλοιτο, εὖ οἰκεῖν, ὥσπερ καὶ τὸν ἑαυτοῦ; (X. Oec. 1.3)

“Well, I suppose that the business of a good estate manager is to manage

his own estate well.” “Yes, and in case he were put in charge of another
man’s estate, could he not, if he chose, manage it as well as he manages
his own?”

(34) καὶ σπείσας καὶ εὐξάμενος ἔπιε, καὶ οἱ ἄλλοι δὲ οἱ περὶ αὐτὸν οὕτως ἐποίουν (X.
Cyr. 7.1.1)

And when he had poured a libation and prayed, he drank; and the rest,
his staff-officers, followed his example

Observe that in (33) the predicate and the object of the second clause (τὸν οἶκον
εὖ οἰκεῖν) literally repeat those of the first clause (εὖ οἰκεῖν τὸν οἶκον), while in (32)
203 there is lexical variation (σὺν πολλοῖς δακρύοις ἀποχωρῆσαι as against οὐδένα |
ὅντιν᾽ οὐ δακρύοντ᾽ ἀποστρέφεσθαι). In (34), finally, the second clause has the
pro-forms οὕτως ἐποίουν, representing the three states of affairs of the preced-
ing clause. Observe also that the examples contain a form of αὐτός ‘self’ or
ἄλλος ‘other’. In fact, many passages where an ‘and also’ interpretation seems
to impose itself contain such forms, not surprisingly perhaps, since ‘self’ and
‘other’ are especially at home in clauses or sentences that are semantically
(near-)repetitions of the preceding sentence or clause.31
But is this ‘and also’ interpretation inevitable? I think not. Assigning an over-
all ‘and on the other hand’-interpretation to καὶ … δέ is quite possible, perhaps
most clearly in (33), where καὶ … δέ might be rendered more specifically by ‘and
again’, while both in (32) and in (34) ‘and in turn’ is more appropriate. As in the
other cases, then, δέ marks a Topic shift. I prefer this analysis to distinguishing
a separate use of καὶ … δέ where καί = ‘also’.32

31 That the presence of αὐτός alone is not sufficient for this effect appears e.g. from X. Cyr.
2.4.17.
32 In itself such a double use is not impossible, cf. καὶ (…) γάρ, where καί may or may not be
‘also’.
adverb or connector? the case of και … δε 313

Another obvious example of ‘and also’ would seem to be:

(35) ἐπυνθάνοντο δέ καὶ ἐς τοὺς Εἵλωτας πράσσειν τι αὐτόν, καὶ ἦν δὲ οὕτως· ἐλευ-
θέρωσίν τε γὰρ ὑπισχνεῖτο αὐτοῖς … ἀλλ᾽ οὐδ᾽ ὣς … ἠξίωσαν … (Th. 1.132.4)

They were informed also that he was intriguing with the Helots; and it
was even so, for he was promising them freedom … But not even then …
did they think it best …

where οὕτως takes up πράσσειν τι. Indeed, as in the other cases of καὶ … δέ in
Thucydides, Classen-Steup and others take καί as = ‘also’. We have already seen,
however, that this is very unlikely at 2.36.1 (ex. (3)) and 6.71.2 (ex. (22)), and here,
too, there is no need to take it in this way. In the preceding context we have
been told how the Spartans distrusted Pausanias because of his lifestyle, and
discovered all kinds of incriminating evidence. The ἐπυνθάνοντο-clause men-
tions another potentially damaging fact. Now καὶ ἦν δὲ οὕτως plus the following
sentence give Thucydides’ comments on this clause, by way of a parenthesis;
δέ emphasizes ἦν, and marks the shift from ‘rumours’ to ‘reality’: ‘and, as a mat-
ter of fact, this was the case; for …’ On this analysis, the ἐπυνθάνοντο-clause is
answered by ἀλλ᾽ οὐδ᾽ ὣς etc.33 | 204

3.4 οὐδὲ … δέ
The next example to be discussed contains οὐδέ … δέ. Recall that according to
Denniston (p. 203) ‘… δέ is clearly the connective, and οὐδέ is adverbial.’

(36) καὶ οὐδὲν μέντοι οὐδὲ τοῦτον παθεῖν ἔφασαν, οὐδ᾽ ἄλλος δὲ τῶν Ἑλλήνων …
ἔπαθεν οὐδεὶς οὐδέν (X. An. 1.8.20)

yet it was said that even he was not hurt in the least, nor, for that matter,
did any other single man among the Greeks get any hurt whatever … (‘for
that matter’ is present in the Loeb-transl.)

I do not think Denniston is right. Instead of continuing, after οὐδὲ τοῦτον παθεῖν
ἔφασαν, with οὐδ᾽ ἄλλον (δέ), Xenophon starts a new sentence, which no longer

33 To bring this out the whole of καὶ ἦν δὲ οὕτως … ξυγκατεργάσωνται should, in fact, be put
between parentheses.—The remaining examples from Thucydides can be analysed in a
similar way, 4.24.2: ‘and they themselves, for that matter’; 7.56.3: ‘and, as a matter of fact,
…’, Du. ‘En trouwens, de strijd was memorabel’; 8.67.3: ‘and, on the other hand’, ‘and, in
turn’.
 314 chapter 15

represents other people’s words, but voices his own comments. By the tran-
sition to a new main verb there is a slight anacoluthon, but οὐδ᾽ (ἄλλος) still
corresponds with οὐδὲ (τοῦτον). In other words, οὐδέ is the connector and δέ
should be taken adverbially. And it should be noted that, just as in the case of
καὶ … δέ, δέ can be omitted, while οὐδέ must be present.34

3.5 μὲν … καὶ … δέ

To conclude I present two cases where καὶ … δέ combines with μέν, a combi-
nation also mentioned by Denniston (p. 203), who does not tell us, however,
whether he takes καί or δέ as the connector.

(37) ἐγὼ μὲν τοίνυν εὔχομαι πρὶν ταῦτα ἐπιδεῖν ὑφ᾽ ὑμῶν γενόμενα μυρίας ἐμέ γε
κατὰ τῆς γῆς ὀργυίας γενέσθαι, καὶ ὑμῖν δὲ συμβουλεύω … πειρᾶσθαι τῶν
δικαίων τυγχάνειν (X. An. 7.1.30)

For my part, therefore, I pray that sooner than live to behold this deed
wrought by you, I may be laid ten thousand fathoms under the ground.
And to you my advice is that … you endeavour to obtain your just rights
…

(38) Ἐγὼ μὲν οὖν … κατέστησά τε τουτουσὶ εἰς ἀγῶνα καὶ ὑπὸ τὴν ὑμετέραν ψῆφον
ἤγαγον. καὶ ὑμᾶς δὲ χρὴ … ψηφίσασθαι τὰ δίκαια ([D.] 59.126)

I therefore … have brought them to trial and submitted them to be judged

by you. It is now your duty to render the verdict which justice demands
205 …|

How should these passages be analysed?35 (Observe that the translations sug-
gest that καί is not taken adverbially, rightly, as we will see). First, there can be
little doubt that δέ is meant to contrast ὑμῖν, and ὑμᾶς, with ἐγὼ μέν. Next, it
would seem that καί can be easily omitted, which would point to δέ being the
connector. And a sequence like ἐγὼ μὲν τοίνυν εὔχομαι … ὑμῖν δὲ συμβουλεύω is, of
course, perfectly acceptable. Syntactically, then, δέ might be the connector. But
for semantic reasons καί must be taken as such. The point is that if καί is taken

34 This passage is mentioned explicitly by K-G (2, 261) as an instance of adverbial δέ.—
Strangely enough, Denniston gives Aristot. EN 1120a31 as another instance of adverbial
οὐδέ, although the presence in the preceding clauses of οὐ and οὐδέ definitely points to
οὐδέ being the connector.
35 Another example of μὲν … καὶ … δέ is Pl. Prt. 331b.
adverb or connector? the case of και … δε 315

adverbially, i.e. as = ‘also’, with the emphasis on the pronoun, there is a semantic
clash between the two sentences, a phenomenon we have already met in cases
like (24) and (25). For to interpret (37) as ‘and I advise you, too, to try and get …’,
the first sentence must necessarily have mentioned other persons who got this
advice. But this is not the case. Moreover, to express, in the appropriate context,
the Greek equivalent of the sentence just mentioned, the order should rather
be συμβουλεύω δὲ καὶ ὑμῖν. The same analysis applies to (38): ‘you, too, must vote
…’, is, in the context, nonsense. In both cases δέ has its by now familiar ‘on the
other hand, in turn’ value.

4 Conclusion

In the introduction I noted that Denniston to a large extent, and Kühner-Gerth

wholly, analyse the particle combination καὶ … δέ as consisting of connective
καί and adverbial δέ, roughly meaning: ‘and on the other hand’. Neither Kühner-
Gerth nor Denniston, however, provided any arguments for their view. Many
commentators, on the other hand, if they comment at all, take καί as = ‘also’
and δέ as the connective. The upshot of the present paper is that Kühner-Gerth
and Denniston were right. Its main points are as follows.
The combination καὶ … δέ occurs in two different syntactic surroundings.
Firstly, it may connect single nouns, adjectives and verbs, type: δίκαιον καὶ πρέ-
πον δέ. In these phrases καὶ cannot be omitted. This fact, as well as the existence
of correlative καὶ … καὶ … δέ in phrases like καὶ οὗτος καὶ ἇλλοι δὲ ποιηταί and καὶ
τοξότας καὶ σφενδονήτας καὶ ἱππέας δέ, shows that καί, rather than δέ, should be
taken as the connector. The function of δέ is a pragmatic one: it presents the
item concerned as distinct from the preceding item(s), and indicates that the
καὶ … δέ item should be considered in its own right.
Secondly, καὶ … δέ is used to connect clauses and sentences. Here, in prin-
ciple both καὶ and δέ are suitable connectors. Again, however, the existence of
correlative constructions, e.g. πρὶν ἂν ἱππέας τε μεταπέμψωσιν …, καὶ χρήματα δὲ
ἅμα … ξυλλέξωνται …, and καὶ ὁ πατὴρ ἀεὶ λέγει καὶ σὺ φῄς καὶ οἱ ἄλλοι δὲ πάντες
ὁμολογοῦσιν indicates that it is καὶ that should be taken as the connector. The
general value is the same as that mentioned above: ‘and on the other hand’.
Often, καὶ … δέ marks a Topic shift. | 206
I further argued that those cases where at first sight an ‘and also’ interpreta-
tion for καὶ … δέ seems plausible can very well be analysed in conformity with
the ‘and on the other hand’ approach. Since there are no cases where καί must
be taken as ‘also’, while there are many cases where καί cannot possibly be taken
in this way, I conclude that, in the combination καὶ … δέ, καί is the connector
and δέ an adverb.
316 chapter 15

Appendix: καὶ … δὲ ὡσαύτως, and Other Constructions Involving

ὡσαύτως

In Plato, we find a number of combinations involving ὡσαύτως that prima facie

are rather similar: καὶ … ὡσαύτως, καὶ … δὲ ὡσαύτως, καὶ αὖ … ὡσαύτως and ὡσαύ-
τως δὲ καί, as in the following examples:

(i) δοκεῖ οὖν σοι … ἀπολείπειν ἄν τι ὁ τοιοῦτος ἀρετῆς, εἴπερ εἰδείη τά τε ἀγαθὰ
πάντα καὶ παντάπασιν ὡς γίγνεται καὶ γενήσεται καὶ γέγονε, καὶ τὰ κακὰ ὡσαύ-
τως (Pl. La. 199d)

Now do you think … there could be anything wanting to the virtue of

a man who knew all good things, and all about their production in the
present, the future, and the past, and all about evil things likewise?

(ii) Εἰσὶν δὴ κατὰ τοὺς νῦν λόγους ψευδεῖς ἐν ταῖς των ἀνθρώπων ψυχαῖς ἡδοναί,
μεμιμημέναι μέντοι τὰς ἀληθεῖς ἐπὶ τὰ γελοιότερα, καὶ λῦπαι δὲ ὡσαύτως (Pl.
Phlb. 40c6)

Hence we reach the result that false pleasures do exist in men’s souls,
being really a rather ridiculous imitation of true pleasures; and the same
applies to pains, for that matter (transl. Hackforth; ‘for that matter’ added)

(iii) ἐν δὲ ταύτῃ οὔσῃ τοιαύτῃ ἀνὰ λόγον τὰ φυόμενα φύεσθαι, δένδρα τε καὶ ἄνθη
καὶ τούς καρπούς· καὶ αὖ τὰ ὄρη ὡσαύτως καὶ τοὺς λίθους ἔχειν … τήν τε λειό-
τητα … (Pl. Phd. 110d)

And in this fair earth the things that grow, the trees, and flowers and fruits,
are correspondingly beautiful; and so too the mountains and the stones
are smoother …

(iv) ἥ τε οὖν ἰατρική … πᾶσα διὰ τοῦ θεοῦ τούτου κυβερνᾶται, ὡσαύτως δὲ καὶ
γυμναστικὴ καὶ γεωργία (Pl. Smp. 186e)

And so not merely is all medicine governed … through the influence of

this god, but likewise athletics and agriculture

Actually, however, the differences between these four expressions are, I think,
roughly as follows.
In examples (i)–(iii) καί functions as a connector. Example (i) is pragmat-
ically unmarked; καί corresponds with τε, and simply adds an item to which
adverb or connector? the case of και … δε 317

εἴπερ εἰδείη etc. applies as well, the latter being summarized by the pro-form
ὡσαύτως. In example (ii) the speaker, by adding λῦπαι by means of καὶ … δέ,
sets them off from the preceding item, and indicates that he is fully aware that
this item is relevant as well. Observe that in (ii) δέ can and καί cannot be omit-
ted; conversely, in the passage from Laches δέ might be added without affecting
the syntax of the καί-clause. Example (ii) is, then, pragmatically marked. In (iii),
finally, the contrast between the two members is the most marked, since it is
semantic rather than merely pragmatic: while καί indicates that the line of rea-
soning followed in the first member will be continued, αὖ emphatically signals
that this reasoning will now be applied to a different set, that of non-growing
items.36 | 207
As to example (iv), this differs syntactically, of course, from the former three,
since in this case δέ must be taken as the connector; for δέ corresponding with
τε cf. fn. 20. Consequently, καί must be taken adverbially. The difference with
the examples where ὡσαύτως occurs later in the clause may be that in (iv)
ὡσαύτως, being in front-position, functions more emphatically as a pro-form:
it stresses that the speaker’s main Topic still is διὰ τοῦ θεοῦ τούτου κυβερνᾶσθαι,
a Topic introduced at 186b: … ὡς … ἐπὶ πᾶν ὁ θεο͂ς τείνει.

References

Bakker, E.J., ‘Boundaries, Topic and the Structure of Discourse: An Investigation of the
Ancient Greek Particle dé’, Studies in Language 17 (1993) 275–311.
Blomqvist, J., Das sogenannte καί adversativum: Zur Semantik einer griechischen Partikel
(Uppsala 1979).
Denniston, J.D. The Greek Particles, 2nd ed. (Oxford 1954).
Dik, H.J.M., Word Order in Ancient Greek: A Pragmatic Account of Word Order Variation
in Herodotus (Amsterdam 1995).
Kroon, C.H.M., Discourse Particles in Latin: A Study of nam, enim, autem, vero and at
(Amsterdam 1994).
Pinkster, H., On Latin Adverbs (Amsterdam 1972).
Quirk, R., Greenbaum, S., Leech, G., Svartvik, J., A Comprehensive Grammar of the
English Language (London 1985).

36 As in the use of αὖ at word-level (see ex. (21)), αὖ contrasts sets that are semantic oppo-
sites. Of course the ἡδόναι and λῦπαι of (ii) are also semantically opposed. But καὶ … δέ
does not contrast them qua opposites. The pragmatic and semantic markedness of (iii)
also appears from the fact that in this example καί is provided with a full-blown clause of
its own, where ὡσαύτως is not just a pro-form, as in (i) and (ii).
318 chapter 15

Rijksbaron, A., ‘Sur quelques différences entre οὗτος ὁ + substantif, οὗτος δὲ ὁ + substan-
tif et ὁ δέ + substantif + οὗτος chez Hérodote’, Lalies 12 (1993) 119–130 [ch. 11 in this
volume].
Rijksbaron, A., ‘Introduction’, in A. Rijksbaron (ed.), New Approaches to Greek Particles
(Amsterdam 1997) 1–14.
Ruijgh, C.J., Autour de ‘τε épique’: Études sur la syntaxe grecque (Amsterdam 1971)
Ruijgh, C.J., ‘La place des enclitiques dans l’ordre des mots chez Homère d’après la loi
de Wackernagel’. in Scripta minora II (Amsterdam 1996) 627–647 [orig. published in
H. Eichner, H. Rix (eds.), Sprachwissenschaft und Philologie: Jakob Wackernagel und
die Indogermanistik heute (Wiesbaden 1990) 213–233].
Sicking, C.M.J., van Ophuijsen, J.M., Two Studies in Attic Particle Usage: Lysias & Plato
(Leiden 1993).
Slings, S.R., ‘καί adversativum: Some Thoughts on the Semantics of Coordination’, in
D.J. van Alkemade et al. (eds), Linguistic Studies Offered to Berthe Siertsema (Ams-
terdam 1980) 101–125.
Wakker, G.C., ‘Emphasis and Affirmation: Some Aspects of μήν in Tragedy’, in A. Rijks-
baron (ed.), New Approaches to Greek Particles (Amsterdam 1997) 209–231.
 chapter 16

Sur l’article avec nom propre

Abstract

It is well known that the article can be used with proper names in Greek, but the factors
which determine its presence or absence remain largely unknown. Through an inves-
tigation of the presence or absence of the article with the name Κῦρος (in all cases) in
the first book of both Xenophon’s Anabasis and Cyropaedia and through an additional
analysis of Plato’s Phaedo, it is argued that, in an extension of its normal anaphoric
function, the article is used to underline the prominence of a character in narrative
discourse, either to contrast that character with another character (especially to accen-
tuate changes of speaker in reported dialogues) or to highlight their role as a principal
actor in the story.

1 Article et nom de personne en grec ancien, en général…

En grec, les noms propres peuvent être accompagnés de l’ article. Mais com-
ment l’article fonctionne-t-il dans cet emploi, si fonction il y a ? Voilà la ques-
tion qui sera le sujet de ma contribution à ce recueil en l’ honneur de mon
collègue stéphanois Bernard Jacquinod, depuis des années co-animateur et co-
hôte des séances à Paris du groupe sur l’aspect verbal. Cela distrait un peu
l’ esprit, après toute cette agitation sur le présent et l’ aoriste !
En guise d’orientation je présente d’abord les opinions de quelques gram-
maires grecques sur l’emploi de l’article avec nom propre1, qui sont typiques,
je crois, de la manière dont on aborde ce sujet. Je commence par les opi-
nions de Basil Gildersleeve, que l’on retrouve, sous une forme ou une autre,
dans les autres manuels. Gildersleeve, qui dans sa grammaire de 1900–1911, au
chapitre intitulé «The doctrine of the article», traite de l’ article d’ une façon
assez détaillée2, fait remarquer tout d’abord: « Proper names being in their
nature particular do not require the explicit article »3. Si l’ article est néanmoins

1 Ne sont discutés que les noms de personnes – et non ceux de dieux, de villes, de fleuves, etc. –,
et parmi les noms de personnes, seulement les groupes Article + Nom Propre – et non ceux
des groupes comme Ἀρίστιππος ὁ Θετταλός, Πίγρητα τὸν ἑρμηνέα, etc.
2 Le chapitre sur l’ article est d’ ailleurs de la main d’ un élève de Gildersleeve, Charles Miller.
3 Gildersleeve (1900–1911 : 215).

© koninklijke brill nv, leiden, 2019 | doi:10.1163/9789004386129_018

 320 chapter 16

présent avec un nom propre c’est une caractéristique qui « belongs to popular
244 diction»4. De son coté, Jean Humbert affirme: |

En principe, l’article n’est jamais nécessaire avec un nom propre, puisque

celui-ci, en raison de son caractère singulier, ne peut être confondu avec
un autre. Cependant, l’article est fréquemment employé, pour des rai-
sons subjectives fort variées. D’une façon générale, la présence de l’ article
indique que la personne (ou la chose) désignée par le nom propre est
familière à celui qui parle ou l’intéresse particulièrement5.

Il fait suivre cette remarque par une série de remarques très diverses sur la pré-
sence ou l’absence de l’article dans toute sorte de genres. Maurice Bizos, en
245 revanche, écrit, dans une remarque qui vise clairement la syntaxe d’ Humbert : |

4 Gilderleeve (1900–1911 : 229).

5 Humbert (1960: 46–47) ; c’ est Humbert qui souligne. Les opinions de Gildersleeve et Humbert
sur la « particularité» ou la « singularité» inhérente du nom propre sont erronées. Un exemple
clair qui en prouve la fausseté est dans Platon, La. 180e (c’est Lysimaque qui parle): τὰ μει-
ράκια… τάδε πρὸς ἀλλήλους οἴκοι διαλεγόμενοι θαμὰ ἐπιμέμνηνται Σωκράτους… οὐ μέντοι πώποτε
αὐτούς ἀνηρώτησα εἰ τὸν Σωφρονίσκου λέγοιεν. Pour ses contemporains, «notre» Socrate n’était
qu’ un Socrate parmi d’ autres Socrate, qui pouvait très bien être confondu avec ses homo-
nymes. – La question de l’ unicité ou non-unicité de la référence d’un nom propre a occupé
les esprits de beaucoup de linguistes, et surtout de philosophes. Pour cette question, ainsi
que pour la question qui y est liée de la signification éventuelle d’un nom propre, je renvoie à
l’ étude intéressante et détaillée de M.-N. Gary-Prieur, Grammaire du nom propre, PUF, Paris,
1994. En combinant les théories de J.S. Mill (System of Logic, Londres, 1843) et du philosophe
américain S.A. Kripke (Naming and Necessity, Blackwell, Oxford, 19802) elle arrive à la conclu-
sion suivante: « L’intérêt majeur, pour les linguistes, des théories de Mill et Kripke, c’est de
montrer la nécessité d’ aborder l’ étude du nom propre dans le cadre de l’énoncé, rompant
ainsi avec la tentation constante de confronter nom propre et nom commun dans une pers-
pective lexicale» (p. 25). Comparer aussi Lyons (1977: 655): «Uniqueness of reference… is
always context-dependent ». Le problème de la référence du nom propre et de son unicité
éventuelle est également soulevé, dans une perspective différente, dans S. Ferret, Le bateau
de Thésée : le problème de l’ identité à travers le temps, Editions de Minuit, Paris, 1996, en parti-
culier au ch. 1 : « Identité et changement». L’essai brillant d’A. Gardiner, The theory of proper
names. A controversial essay, Oxford University Press, Londres, 19542, n’a rien perdu de son
intérêt : Gardiner s’ oppose vigoureusement à toute théorie qui veut attribuer une significa-
tion au nom propre, notamment celle de B. Russell, « The philosophy of logical atomism»,
The Monist, 28 (1918), p. 479 sqq. : « Assuredly the most fantastic theory of proper names that
has ever come to birth is that propounded by Bertrand Russell» (p. 57). F. Lambert, «Le
nom propre entre signifié et référence: l’ apport d’ Apollonios Dyscole», dans A. Mettouchi
et H. Quintin (éds.), La réference 2. Statut et processus, Presses Universitaires de Rennes, 1999,
p. 31 sqq., traite des idées stoïciennes sur le nom propre, et notamment de l’idée assez remar-
quable selon laquelle les noms propres n’ont pas de fonction référentielle, mais seulement
une signification, à savoir « être la personne qui est appelée /N/».
sur l’ article avec nom propre 321

Les noms propres prennent ou ne prennent pas l’ article, indifféremment.

Les distinctions souvent établies à ce sujet semblent factices6.

L’ impression générale qui subsiste après lecture des remarques d’ Humbert et

de celles de Gildersleeve est qu’il y a bien quelques tendances générales dans
l’ emploi ou le non-emploi de l’article (ainsi, non-emploi dans l’ épopée et la
lyrique, rareté dans la poésie dramatique, etc.), mais que par ailleurs la pré-
sence ou l’absence de l’article avec un nom propre est une chose qui n’obéit à
aucune règle fixe et que l’arbitraire y règne donc, notamment chez les prosa-
teurs, comme Platon ou Xénophon7. Le lecteur ne peut que conclure qu’ il est
impossible d’en dire plus8. Dans ce qui suit j’essaierai de démontrer que cette
conclusion serait prématurée.

2 … et dans l’Anabase de Xénophon en particulier

Le point de départ de mon analyse est la présence du nom propre « Cyrus » dans
le premier livre de l’Anabase de Xénophon. Comme l’ on sait, c’ est dans ce livre
que l’on trouve le récit de l’anabase au sens strict, la « montée» de Cyrus et de
ses troupes mercénaires à la rencontre de son frère Artaxerxès, le roi des Perses.
Il va de soi que Cyrus, qui est tant l’organisateur que le commandant en chef
de l’expédition contre Artaxerxès, joue un rôle important dans ce livre, et qu’ il
est mentionné avec une grande régularité. Des 223 exemples du nom de Κῦρος
que j’ai comptés moi-même dans l’Anabase9, il y en a 144 qui se trouvent dans
les chapitres 1–8 du premier livre, c’est-à-dire dans les chapitres qui traitent de
l’ expédition au sens propre. À la fin du ch. 8 Xénophon décrit comment Cyrus
tombe pendant la bataille de Counaxa. Néanmoins son nom revient | encore 246
souvent après sa mort, au total 79 fois. C’est qu’ il joue un rôle dans des flash-

6 Bizos (1995: 6). Comparer Goodwin (18942 : § 943) : « Proper names may take the article; as ὁ
Σωκράτης or Σωκράτης, Socrates».
7 Chez Platon « the articular form is used with the utmost freedom» (Gildersleeve 1900–1911:
229) ; et à propos de Xénophon: « According to Joost, Sprachgebrauch Xenophons in der Ana-
basis, [Berlin 1892] Κῦρος occurs 196 times in the Anabasis, ὁ Κῦρος 28 times» (Ibid.).
8 Il est vrai que Gildersleeve (1900–1911: 215) écrit aussi que «when the article is used with
them (sc. les noms propres), it retains much of its original demonstrative force», mais il ne
développe pas cette idée, qui est pourtant précieuse; voir ci-dessous, §4. Pour être complet
j’ ajoute que la notion d’ « intérêt», soulevée par Humbert et que l’on retrouve par exemple
dans K-G 1, 598, semble en effet pertinente pour une partie des emplois de l’article au moins;
voir infra, § 5.
9 Ne sont pas inclus les trois cas du vocatif, qui ne peut pas être modifié par l’article.
322 chapter 16

backs, des considérations rétrospectives et des anecdotes10. Or qu’ en est-il de

l’ article? De la remarque que je viens de citer (supra, note 7) il est clair que
les formes de Κῦρος avec l’article sont beaucoup moins nombreuses que celles
sans l’article: il y a 28 cas de ὁ Κῦρος (τοῦ Κύρου, etc.) contre 196 cas de Κῦρος
(Κύρου, etc.), soit environ 12,5% avec, et 87,5% sans l’ article. Mais ces chiffres
ne prennent de l’intérêt que si l’on considère le cas (la forme casuelle) auquel
le nom de «Cyrus» se présente. La répartition des formes casuelles est la sui-
vante:

nominatif génitif

Κῦρος11 ὁ Κῦρος Κύρου τοῦ Κύρου

68 21 54 1

datif accusatif

Κύρῳ τῷ Κύρῳ Κῦρον12 τὸν Κῦρον

39 3 35 3

De ce tableau il ressort donc qu’il y a 89 formes au nominatif, dont 21 (=

23%) ont l’article, et que l’on rencontre 55 cas au génitif, 42 au datif, et 38
à l’accusatif, soit 135 formes aux cas obliques, dont seulement 7 (= 5 %) ont
l’ article.
Somme toute, le nominatif est le cas préféré, ce qui n’a rien de surprenant,
bien entendu, puisque Cyrus est la personne qui agit par excellence dans ces
chapitres. Pourtant, il est frappant qu’au nominatif l’ article est beaucoup plus
fréquent qu’aux cas obliques13. Comment expliquer ce phénomène ? Avant que
je n’aborde cette question je mentionne un autre phénomène qui donne à

10 J’ ai laissé de côté les résumés (non-authentiques) au début des livres 2, 3, 4, 5 et 7.

11 Y compris 1.3.1, où Hude avec une partie des manuscrits lit Κῦρος; d’autres éditeurs pré-
fèrent la variante ὁ Κῦρος. Le texte sans l’ article est probablement préférable. Voir aussi
infra, n. 27.
12 Sans Κῦρον dans 1.9.1, car ce Cyrus est un autre Cyrus, à savoir Cyrus le Grand, roi des Perses
de 559–529 avant J. C.
13 Pour les cas obliques voir infra, n. 25.
sur l’ article avec nom propre 323

réfléchir: le nom «Cyrus» n’est jamais accompagné de l’ article quand il appa-

raît dans un discours direct, c’est-à-dire dans le texte d’ un personnage par-
lant. Il faut que j’ajoute tout de suite que ceci vaut également pour tout autre
personnage engagé dans l’expédition, et, chose non moins importante, pour
beaucoup de noms propres chez un auteur comme Platon – j’ y reviendrai.
L’ explication de ce phénomène est probablement que les personnes mention-
nées dans un discours direct sont | normalement présentes dans la situation 247
communicative; leur identité est donc connue14. Quoi qu’ il en soit, on peut
conclure de ce fait que l’emploi de l’article avec nom propre est exclusivement
lié aux parties narratives de l’Anabase15.

3 Κῦρος contre ὁ Κῦρος

Passons donc au terrain où l’article est le plus présent: les parties narratives de
l’Anabase, en l’occurrence les huit premiers chapitres du livre 1. Dans ces cha-
pitres, il y a 19 cas de ὁ Κῦρος contre 43 de Κῦρος (soit 30 % contre 70%). Or il
est intéressant de noter que ces 19 cas sont répartis de façon assez inégale dans
notre texte; on obtient la liste suivante où les lignes sont comptées consécuti-
vement suivant l’édition de Hude (Leipzig 19722) ; chiffres gras = ὁ Κῦρος:

3, 9, 36, 46, 51, 62, 104, 130, 136, 143, 154, 170, 187, 194, 196, 200, 212, 236, 242,
247, 286, 364, 396, 399, 412, 416, 431, 441, 444, 476, 488, 533, 547, 572, 590,
620, 639, 642, 646, 649, 651, 654, 673, 681, 704, 721, 740, 756, 759, 763, 769,
783, 794, 797, 819, 822, 836, 846, 865, 879, 893, 895.

Il est évident que la façon normale de renvoyer à la personne de Cyrus dans ces
chapitres est par la forme sans article (je me borne ici donc au nominatif)16. A

14 Il se peut toutefois qu’ il y ait aussi un effet de genre, ou de position sociale, car les
noms propres avec article ne sont pas rares dans la comédie, notamment avec les noms
d’ esclaves, ainsi ὁ Ξανθίας (Ar., Grenouilles, 271), ὁ Διτύλας χὠ Σκεβλύας χὠ Παρδόκας,
χωρεῖτε δευρὶ… (Ibid., 608). Ces formes s’ emploient surtout dans des appels. Cf. Svennung
(1958 : 217 sqq.), qui appelle (p. 222) l’ emploi de l’ article + nominatif «schroff» («rude»).
15 Quant aux 79 cas du nom de « Cyrus » que l’ on trouve dans le récit après sa mort (voir ci-
dessus), seulement deux d’ entre eux sont au nominatif et ont l’article. Ces cas se trouvent,
eux aussi, dans des passages narratifs, 3.1.5 et 3.1.9 ; dans ces passages le lecteur est reporté
à une situation d’ avant le début de l’ expédition, où Cyrus, alive and kicking, est active-
ment engagé dans la préparation de son expédition. Ici donc Cyrus est exceptionellement
réintroduit dans le récit comme personne vivante.
16 L’ emploi fréquent du nom « Cyrus », dans un texte où il est de loin le Topique du discours
 324 chapter 16

248 cet égard ce nom propre diffère des noms | communs, qui normalement, après
qu’ils ont été introduits dans un texte, ont l’article lorsqu’ ils sont repris dans
la suite. La première phrase de l’Anabase en fournit un exemple :

Δαρείου καὶ Παρυσάτιδος γίγνονται παῖδες δύο, πρεσβύτερος μὲν Ἀρταξέρξης,

νεώτερος δὲ Κῦρος· ἐπεὶ δὲ ἠσθένει Δαρεῖος καὶ ὑπώπτευε τελευτὴν τοῦ βίου,
ἐβούλετο τὼ παῖδε ἀμφοτέρω παρεῖναι.

Mais pour la reprise d’un nom propre l’article n’est donc nullement néces-
saire. Si l’article est néanmois présent, quelle est alors sa fonction ? Je crois que
l’ article garde après tout sa fonction anaphorique et que l’ ensemble de l’ article
et du nom propre établit ce que l’on pourrait appeler une anaphore forte, qui
sert à souligner l’identité du référent, et ainsi à le mettre en relief17. Dans la
pratique, cette «mise en relief» s’opère de deux façons différentes.
Il est frappant qu’une partie considérable des cas de ὁ Κῦρος se trouve regrou-
pée dans deux passages, au début (lignes 9–62 = 1.1.2–10) et au trois-quarts du
récit (lignes 639–651 = 1.6.7–8). C’est notamment la haute concentration de ὁ
Κῦρος dans 639–651 qui demande notre attention. C’ est pourquoi je traiterai
tout d’abord de ce passage, ainsi que du passage qui le précède (1.6.1–5).
Dans ces passages Xénophon décrit la trahison et le procès d’ un officier
perse, un certain Orontas. Orontas fait son entrée dans le récit au ch. 1.6.1,
comme comploteur (Ὀρόντας δέ, Πέρσης ἀνήρ, … ἐπιβουλεύει Κύρῳ…). C’ est un
personnage nouveau, donc sans l’article, et de réputation douteuse, car il avait
déjà eu un conflit avec Cyrus à une occasion antérieure (καὶ πρόσθεν πολεμήσας).
À un moment non spécifié cet Orontas fait une proposition à Cyrus (οὗτος Κύρῳ
εἶπε, ligne 607 = 1.6.2). τῷ δὲ Κύρῳ… ταῦτα ἐδόκει ὠφέλιμα εἶναι, après quoi Oron-
tas procède à son act perfide (ὁ δ᾽ Ὀρόντας… γράφει ἐπιστολὴν παρὰ βασιλέα ὅτι

le plus important, peut étonner et n’est pas facile à expliquer. Quoi qu’il en soit, cet emploi
contredit l’ opinion assez répandue selon laquelle les noms de personne apparaissent
de préférence, ou même exclusivement, lorsque le référent a une «accessibilité basse»
(« low accessability », Givón 1990: 741), ou est « momentanément non-actif» («currently-
inactive », ibid. : 937), c’ est-à-dire lorsque il y a une distance relativement grande entre
l’ expression anaphorique et son référent. Un exemple clair de cette présence de Cyrus est
le passage An. 1.4.9–11, où il y a trois cas de Κῦρος assez proches les uns des autres sans
qu’ il n’existe aucune opacité référentielle si le nom est absent. Pour quelques suggestions
intéressantes sur les effets pragmatiques de ce type d’anaphore («overspecified referen-
tial expressions»), où l’ identification du référent ne pose aucun problème, je renvoie à
Maes (1996 : 9–11).
17 Pour la notion d’ « anaphore forte», voir Grimes (1975: 92 sqq.) et Rijksbaron (1993: 123–
130).
sur l’ article avec nom propre 325

ἥξοι, 1.6.3). Mais il est trahi à son tour, et la lettre est transmise à Cyrus, qui le fait
arrêter (ἀναγνοὺς δὲ αὐτὴν ὁ Κῦρος συλλαμβάνει Ὀρόνταν). Suit le procès, qu’ un
des officiers grecs, Cléarque, rapporte à Xénophon. Cyrus interroge Orontas;
il est toujours cité en discours direct, tandis que les réactions d’ Orontas sont
tantôt résumées très brièvement, sans qu’il parle lui-même, tantôt | citées en 249
discours direct (chapitre 6.7–8; les noms propres et les verbes dicendi sont en
caractères gras):
Après une brève introduction Cyrus en vient aux faits : Ὀρόντα, ἔστιν ὅ τι σε
ἠδίκησα ; – ἀπεκρίνατο ὅτι οὔ.

1 πάλιν δὲ ὁ Κῦρος ἠρώτα· Οὐκοῦν ὕστερον, ὡς αὐτὸς σὺ ὁμολογεῖς, οὐδὲν ὑπ’

ἐμοῦ ἀδικούμενος ἀποστὰς εἰς Μυσοὺς κακῶς ἐποίεις τὴν ἐμήν χώραν ὅ τι
ἐδύνω;
1′ ἔφη ὁ Ὀρόντας.
2 Οὐκοῦν, ἔφη ὁ Κῦρος, ὁπότ’ αὖ ἔγνως τὴν σαυτοῦ δύναμιν, ἐλθὼν ἐπὶ τὸν τῆς
Ἀρτέμιδος βωμὸν μεταμέλειν τέ σοι ἔφησθα καὶ πείσας ἐμὲ πιστὰ πάλιν ἔδω-
κάς μοι καὶ ἔλαβες παρ’ ἐμοῦ;
2′ καὶ ταῦθ’ ὡμολόγει ὁ Ὀρόντας.
3 τί οὖν, ἔφη ὁ Κῦρος, ἀδικηθεὶς ὑπ’ ἐμοῦ νῦν τὸ τρίτον ἐπιβουλεύων μοι φανερὸς
γέγονας ;
3′ εἰπόντος δὲ τοῦ Ὀρόντα ὅτι οὐδὲν ἀδικηθείς,
4 ἠρώτησεν ὁ Κῦρος αὐτόν· Ὁμολογεῖς οὖν περὶ ἐμὲ ἄδικος γεγενῆσθαι ;
4′ Ἦ γὰρ ἀνάγκη, ἔφη ὁ Ὀρόντας.
5 ἐκ τούτου πάλιν ἠρώτησεν ὁ Κῦρος· Ἔτι οὖν ἂν γένοιο τῷ ἐμῷ ἀδελφῷ πολέμιος,
ἐμοὶ δὲ φίλος καὶ πιστός ;
5′ ὁ δὲ ἀπεκρίνατο ὅτι οὐδ’ εἰ γενοίμην, ὦ Κῦρε, σοί γ’ ἄν ποτε ἔτι δόξαιμι.

πρὸς ταῦτα Κῦρος εἶπε τοῖς παροῦσιν…

4 Les alternances des prises de parole

L’ on a affaire ici à une scène avec deux rôles qui sont formellement établis,
celui d’un interrogateur et celui d’un répondeur. Les deux rôles sont tour à tour
mis en relief, chaque rôle au total cinq fois. Or, c’ est précisement ce « tour à
tour» auquel s’attache la présence de l’article ὁ. Chaque fois que le tour passe
à l’interlocuteur, ceci est signalé par l’article, dans certains cas en combinaison
avec la particule δέ. Il y a donc un effet de «l’un… l’ autre». Pour faire ressor-
tir cet effet on pourrait même, dans une traduction un peu chargée, rendre
l’ article par «il/lui», et prendre le nom propre comme une apposition : « De son
 326 chapter 16

coté lui, Cyrus, demandait…», «lui, Orontas, en convint», « dit-il, lui, Cyrus »,
etc.; comparer l’emploi pronominal de ὁ dans l’ expression figée ὁ μέν… ὁ δέ. Il
s’ agit ici en d’autres termes de ce qui est appelé, dans l’ analyse moderne de la
250 conversation, «turn-taking», en français «alternance des prises de parole»18. |
Lorsque la partie d’Orontas est finie et que l’alternance des prises de parole
se termine donc, son nom n’apparaît plus (voir 5′) tandis que Κῦρος apparaît
sans article, comme le montre la dernière phrase citée ci-dessus, § 6.9, où il
adresse la parole non plus à Orontas mais aux témoins19. De plus, il est intér-
essant de noter que dans ce passage le nom propre vient normalement après
le verbe, dans quatre cas dans la combinaison ἔφη ὁ + nom propre, une combi-
naison dont on verra plus loin qu’elle est très souvent utilisée pour exprimer
l’ alternance des prises de parole20. À noter aussi que la répartition des rôles
entre Cyrus et Orontas a déjà été mise en œuvre antérieurement, à savoir au
chapitre 6.2–4; aussitôt qu’Orontas a pris l’initiative, ses propres actions et
celles de Cyrus se relayent, ce qui est marqué, ici aussi, par l’ article : τῷ δὲ Κύρῳ,
ὁ δ᾽ Ὀρόντας, ὁ Κῦρος21.

18 Cf. Brown et Yule (1983: 104–105); Levinson (1983: 296 sqq.). Je fais remarquer, entre
parenthèses, que la recherche de l’ alternance des prises de parole ne concerne que la
conversation « réelle» ; autant que je sache, il n’existe pas d’études sur la signalisation
de l’ alternance des prises de parole dans la langue écrite.
19 Il y a toutefois une variante ὁ Κῦρος. Pour τὸν Ὀρόνταν dans 1.6.10 voir infra, n. 25.
20 Bien que cet ordre des mots semble avoir un caractère formulaire, je pense que pragma-
tiquement le verbe est dans la position de Focus, tandis que les noms de personne sont
des Topiques contrastifs. L’information nouvelle, saillante, est après tout fournie par le
verbe (ἔφη, εἰπόντος, ἠρώτησεν ou ὡμολόγει), ou le cas échéant par le verbe plus le dis-
cours (in)direct qui en dépend. Autrement dit, tandis qu’il est établi que Cyrus et Orontas
parlent à tour de rôle, ce qu’ ils disent ne l’ est pas. Cf. Dik (1995: ch. 4), bien qu’elle ne
traite pas de ἔφη. À noter toutefois que dans le passage cité sous 1, le nom de personne
précède le verbe. Cet ordre sert peut-être, après ἀπεκρίνατο sans sujet, à introduire Cyrus
de façon explicite comme Topique principal du discours; noter aussi la présence excep-
tionelle de πάλιν δέ (« de son côté»). A côté de scènes où alternent les prises de parole
détaillées comme celle avec Orontas, on en rencontre également d’ autres beaucoup plus
rudimentaires, qui ne deviennent pas de vraies scènes; ainsi en 1.7.9, où Cyrus est le seul
à avoir l’ article : ἐν δὲ τῷ καιρῷ τούτῳ Κλεάρχος ὧδέ πως ἤρετο τὸν Κῦρον· Οἴει γάρ σοι μαχεῖ-
σθαι, ὦ Κῦρε, τὸν ἀδελφόν; Νὴ Δί᾽, ἔφη ὁ Κῦρος. Pour τὸν Κῦρον, voir infra, n. 25. Remarquer
toutefois que les mss D et F ont Κῦρον.
21 Le texte que j’ ai présenté ci-dessus sous 1–5 est celui de l’édition de Hude, qui lit toujours,
avec une partie des manuscrits, ὁ Ὀρόντας. D’ autres éditeurs préfèrent la variante Ὀρόν-
τας, donc sans l’ article (paléographiquement, la présence comme l’absence de ο devant
οροντας serait naturellement facile à expliquer). À mon sens, le texte de Hude est préfé-
rable, vu d’ abord la présence de ἔφη ὁ Κῦρος, et la fréquence générale de la combinaison
ἔφη ὁ + nom propre que je viens de mentionner, mais aussi en raison de la présence de τοῦ
devant Ὀρόντα au passage 3′, et de ὁ δ᾽ Ὀρόντας dans 1.6.3. Voir aussi infra, n. 25.
sur l’ article avec nom propre 327

Dans le passage cité ci-dessus il s’agit de personnes qui réagissent l’ un à

l’ autre. La présence de l’article est en effet étroitement liée à cette situation
communicative, car s’il s’agit de personnes qui parlent sans qu’ ils réagissent
l’ un à l’autre, l’article manque. Cette situation se présente par exemple lorsque
des orateurs | se relayent dans un conseil de guerre, comme dans 3.1.34 : Ἱερώ- 251
νυμος Ἠλεῖος ἤρχετο λέγειν ὧδε – ἐκ τούτου λέγει τάδε Ξενοφῶν. – μετὰ δὲ τοῦτον
εἶπε Χειρίσοφος.
Voici venu le bon moment d’interrompre mon argumentation. Dans un
article méconnu, paru dans un des innombrables Schulprogramme allemands
du XIXe siècle, Adolf Zucker avait déjà en 1899 mis en lumière le rôle de l’ article
dans le «lebendige Wechselgespräch»22. Cet article, intitulé « Beobachtungen
über den Gebrauch des Artikels bei Personennamen in Xenophons Anaba-
sis», est une étude détaillée et pleine d’observations précieuses sur l’ emploi
de l’article avec des noms de personne dans l’Anabase. Dans plus d’ un cas j’ ai
pu développer ses idées. Le titre est mentionné par Gildersleeve, mais je doute
qu’il l’ait vraiment utilisé23. Car c’est en effet un article assez inaccessible : 67
pages d’un allemand sévère, qui contient des séries interminables d’ exemples,
sans aucune sous-division, et imprimées en alphabet gothique – Zucker s’ est
assurément donné peu de peine d’enthousiasmer ses lecteurs. Ceci n’empêche
cependant pas que l’article de Zucker aurait mérité un traitement plus appro-
fondi.

5 La position spéciale de Cyrus

Jusque-là les choses sont relativement claires. Dans les passages mentionnés
ci-dessus ὁ Κῦρος contraste avec un autre nom de personne pourvu de l’ article,
dans une situation où l’on trouve une alternation régulière d’ action et de
réaction des deux personnes; au total il y a neuf exemples de cet emploi (=
43%). Mais que penser de l’autre passage où un certain nombre de cas de ὁ
Κῦρος se trouvent plus ou moins proches l’un de l’ autre, au chapitre premier
du livre 1? Ici l’emploi de l’article est d’un autre ordre. Ce premier chapitre
consiste en une série de brefs épisodes, qui ont tous une importance plus ou
moins grande pour la préhistoire de l’expédition, ou pour l’ organisation de
l’ expédition même, et dans lesquelles plusieurs autres personnages, aussi bien

22 Avant Zucker, Schmidt (1890 : 54) avait déjà écrit : «Nomen proprium quod coniungitur
cum ἔφη in nullo scriptore sine articulo inveni ».
23 Ceci vaut également pour S-D 2, 25. Bien qu’ ils mentionnent Zucker, ils n’ont pas utilisé
son article dans leur traitement de l’ article avec nom de personne.
 328 chapter 16

ennemis qu’amis de Cyrus, nous sont présentés. Aucun de ces personnages

252 n’est pourvu de l’article. Mon | hypothèse est que dans les cinq cas de ὁ Κῦρος
l’ article sert de nouveau à souligner l’identité de Cyrus et ainsi à mettre en relief
son rôle dans ces épisodes. Comme dans le cas traité ci-dessus, on pourrait faire
ressortir cet effet en (sur)traduisant par «lui, Cyrus » – je rappelle que l’ article
n’est nullement nécessaire pour renvoyer à Cyrus dans le contexte précédent.
Il s’agit de:
(a) Κῦρον δὲ μεταπέμπεται (Darius) ἀπὸ τῆς ἀρχῆς ἧς αὐτὸν σατράπην ἐποίησε·
καὶ στρατηγὸν δὲ αὐτὸν ἀπέδειξε πάντων ὅσοι… ἁθροίζονται· ἀναβαίνει οὖν ὁ
Κῦρος… (1.1.2)
(b) ἐν Μιλήτῳ δὲ Τισσαφέρνης προαισθόμενος τὰ αὐτὰ ταῦτα βουλευομένους ἀπο-
στῆναι πρὸς Κῦρον, τοὺς μὲν αὐτῶν ἀπέκτεινε τοὺς δ᾽ ἐξέβαλεν. ὁ δὲ Κῦρος
ὑπολαβὼν τοὺς φεύγοντας συλλέξας στράτευμα ἐπολιόρκει Μίλητον καὶ κατὰ
γῆν καὶ κατὰ θάλατταν. (1.1.7)
(c) οὐδὲν ἤχθετο (le roi) αὐτῶν (Cyrus et Tissaphernès) πολεμούντων. καὶ γὰρ ὁ
Κῦρος ἀπέπεμπε τοὺς γιγνομένους δασμοὺς βασιλεῖ ἐκ τῶν πόλεων ὧν Τισσα-
φέρνους ἐτύγχανεν ἔχων. (1.1.8)
(d) Κλέαρχος Λακεδαιμόνιος φυγὰς ἦν· τούτῳ συγγενόμενος ὁ Κῦρος ἠγάσθη τε
αὐτὸν καὶ δίδωσιν αὐτῷ μυρίους δαρεικούς. (1.1.9)24
(e) Ἀρίστιππος δὲ ὁ Θετταλὸς ξένος ὢν ἐτύγχανεν αὐτῷ, καὶ πιεζόμενος ὑπὸ τῶν
οἴκοι ἀντιστασιωτῶν ἔρχεται πρὸς τὸν Κῦρον καὶ αἰτεῖ αὐτὸν εἰς δισχιλίους
ξένους καὶ τριῶν μηνῶν μισθόν, ὡς οὕτως περιγενόμενος ἂν τῶν ἀντιστασιωτῶν,
ὁ δὲ Κῦρος δίδωσιν αὐτῷ εἰς τετρακισχιλίους καὶ ἓξ μηνῶν μισθόν. (1.1.10)25

24 Il faut remarquer que lorsque ce même épisode est rapporté dans la nécrologie de Clé-
arque (2.6.5), Κῦρος n’ a pas l’ article : δίδωσι δὲ αὺτῷ Κῦρος μυρίους δαρεικούς. C’est que le
rôle central de Cyrus n’ a plus d’ intérêt ici.
25 À noter aussi qu’ ici la présence de ὁ Κῦρος a été préparée exceptionellement par la com-
binaison τὸν Κῦρον, avec peut-être un effet de focalisation: pour Aristippe, tout comme
pour le narrateur, Cyrus joue un rôle crucial. Quant aux autres exemples des cas obliques
(τοῦ Κύρου 1.4.12 ; τῷ Κύρῳ 1.1.4, 1.6.2, 1.8.13; τὸν Κῦρον 1.1.3, 1.7.9), ils s’expliquent à peu
près comme le nominatif : τοῦ Κύρου dans 1.4.12 est peut-être un exemple de l’emploi «en
relief », mais il y a une variante sans l’ article et Cobet et Méridier omettent les mots. Pour
τὸν Κῦρον dans 1.7.9, voir supra, n. 20. Les autres se trouvent en opposition à d’autres noms
(propres) qui ont pour la plupart l’ article (τῷ Κύρῳ dans tous les trois cas; τὸν Κῦρον dans
1.1.3), dans un emploi que l’ on pourrait considérer comme une extension de l’emploi des
alternances des prises de parole. Àpart ces formes obliques de ὁ Κῦρος il y a aussi deux
exemples de τοῦ Ὀρόντα et un de τὸν Ὀρόνταν. L’un des deux cas de τοῦ Ὀρόντα fait par-
tie de l’ alternance des prises de parole (voir supra, p. [249]). L’autre se trouve hors cette
alternance, dans 1.6.5 ; peut-être l’ article met-il l’ adversaire de Cyrus en relief, juste avant
le début de son procès. D’ une façon identique τὸν Ορόνταν, qui se trouve dans 1.6.10, pour-
rait le mettre en relief au moment crucial où il est mené à son exécution.
sur l’ article avec nom propre 329

Dans les quatre derniers cas, il s’agit d’actions de Cyrus qui sont essentielles
pour la guerre qu’il est en train d’organiser. Tout en | trompant le roi, il lève 253
toujours de nouvelles troupes; or, chaque fois qu’ un nouveau contingent est
nommé on trouve la combinaison ὁ Κῦρος, jamais Κῦρος seul. Cyrus est présenté
expressément comme la personne «en tête», comme le « héros» de l’ histoire.
Quant au premier cas, bien que Cyrus ne soit pas encore actif comme orga-
nisateur, sa position et son importance ont été mentionnées expressément
dans la phrase précédente. Lorsque les préparations sont terminées, et que
l’ expédition au sens propre commence (à partir du ch. 2), la combinaison ὁ
Κῦρος disparaît pendant longtemps et fait place à Κῦρος sans article. Quand
ὁ Κῦρος resurgit, à la ligne 170 (1.2.14), c’est à nouveau à propos de son rôle
d’organisateur et de commandant en chef, dans une situation spéciale, à savoir
une revue des troupes: ἐξέτασιν ποιεῖται ἐν τῷ πεδίῳ τῶν Ἑλλήνων καὶ τῶν βαρβά-
ρων. ἐκέλευσε δὲ τοὺς Ἕλληνας… ταχθῆναι καὶ στῆναι, … ἐτάχθησαν οὖν ἐπὶ τεττά-
ρων·… ἐθεώρει οὖν ὁ Κῦρος πρῶτον μὲν τοὺς βαρβάρους…· εἶτα δὲ τοὺς Ἕλληνας26.
Il est à remarquer qu’en 1.2.10 se trouve une phrase semblable, à ceci près que
l’ article est absent: ἐνταῦθ’ ἔμεινεν ἡμέρας τρεῖς· ἐν αἷς Ξενίας ὁ Ἀρκὰς… ἀγῶνα
ἔθηκε·… ἐθεώρει δὲ τὸν ἀγῶνα καὶ Κῦρος. Or, il y a une différence nette entre les
deux phrases. La course mentionnée dans la seconde est une affaire exclusive-
ment grecque – on notera la présence de Ξενίας… ἀγῶνα ἔθηκε. Cyrus n’a donc
joué aucun rôle dans l’organisation de la course, il est présent plus ou moins
par hasard: Cyrus «aussi» était parmi les spectateurs. Dans le premier exemple,
en revanche, la phrase ἐθεώρει κτλ. mentionne le résultat des mesures prépa-
ratives de Cyrus lui-même, comme le signale οὖν. On traduira donc, non pas
« Cyrus aussi était un spectateur», mais «Ainsi se fit-il que lui, Cyrus, inspec-
tait ses troupes», dans une traduction qui vise de nouveau à mettre en lumière
l’ effet de ὁ (et celui de οὖν). Plus généralement on peut dire que, lorsque le texte
précise ὁ Κῦρος, la présence de l’article est le signal que Cyrus est au centre d’ un
épisode qui met en relief sa position particulière27.
Il va de soi que pour faire un tableau complet il me faudrait traiter égale-
ment des autres noms propres dans l’Anabase, mais en | faisant cela je dépas- 254
serais le cadre de cet article. Je me borne donc à présenter trois observations
générales, qui peuvent en même temps servir de conclusion provisoire. (1) Les

26 D’ après Zucker (1899 : 59), qui discute, lui aussi, les deux phrases, la fonction de l’article
est de signaler que cette épisode pour l’ auteur « besonderes Interesse hat», en raison des
activités glorieuses des Grecs, mais cette explication me paraît faite sur mesure.
27 C’ est pour cette raison que je partage l’ opinion de Hude (cf. n. 11), selon laquelle, dans
1.3.1, la variante Κῦρος est meilleure que ὁ Κῦρος: l’ épisode en question n’a rien de spécial;
au contraire, il s’ agit d’ une action banale (ἐνταῦθα ἔμεινε…).
 330 chapter 16

scènes d’alternance des prises de parole sont responsables de la majorité des

cas d’articles avec nom propre. (2) Dans les autres cas, la fonction de l’ article est
de mettre en relief, tout comme dans le cas de Cyrus, la position spéciale d’ un
personnage dans un épisode qui a une importance particulière. (3) Lorsque le
narrateur-auteur, Xénophon, parle de lui-même il emploie l’ article beaucoup
plus souvent qu’avec les autres noms propres. Il y a, dans les parties narratives,
154 cas de Ξενοφῶν au nominatif, dont 73, soit 47 %, avec l’ article28. Pour Κῦρος
ce pourcentage était de 30%. Xénophon se présente lui-même donc comme un
personnage qui a une position encore un peu plus centrale que Cyrus…

6 Les noms de personne dans la Cyropédie

On doit se demander, bien sûr, si les résultats obtenus pour l’Anabase sont
également valables pour les autres œuvres de Xénophon29. Comme œuvre
de contrôle, j’ai retenu le premier livre de la Cyropédie, la biographie fictive
d’un autre Cyrus, à savoir Cyrus le Grand, roi des Perses de 559 à 529. Le nom
« Cyrus» y est naturellement très fréquent. Dans les parties narratives de ce
livre il y a au total 53 mentions du nom au nominatif, dont 44 fois avec l’ article.
Les proportions sont donc radicalement différentes de celles rencontrées dans
l’Anabase. Des 44 cas de ὁ Κῦρος, 21 sont à mettre au compte de l’ alternance
des prises de parole (environ 48%), souvent du type ἔφη ὁ Κῦρος, et 23 à celui
de l’autre emploi de ὁ Κῦρος (52%). Puisque pour l’Anabase ces pourcentages
étaient de 43% et 57%, la différence n’est pas choquante. La question est donc
surtout de savoir: pourquoi Κῦρος sans article est-il si peu fréquent dans la Cyro-
pédie et si fréquent dans l’Anabase?
Cette différence de fréquence est reliée à la forme des œuvres. Le premier
livre de la Cyropédie consiste pour l’essentiel en des scènes plus ou moins déta-
chées, dans lesquelles Cyrus ou bien entre en débat avec d’ autres personnes, ou
bien est présenté comme la personne centrale de situations nouvelles; le pre-
mier livre de l’Anabase, en revanche, est le récit ininterrompu d’ une expédition
de Cyrus dont il est la figure centrale dès le début, et où sa position centrale
255 n’est mise en relief que dans des cas particuliers. L’emploi n’est donc pas |
fondamentalement différent, mais doit être attribué à la différence du type du
discours: dans l’amalgame de scènes qu’est la Cyropédie, c’ est ὁ Κῦρος qui pré-
domine, dans le récit continu de l’Anabase, c’est Κῦρος seul.

28 Ὁ Ξενοφῶν se trouve en outre trois fois dans un discours indirect: 6.4.14, 7.7.55, 7.8.21. J’ai
ignoré ces cas.
29 Et pour d’ autres écrivains, bien entendu. Pour quelques observations sur Platon voir §7.
sur l’ article avec nom propre 331

7 Article et nom de personne chez Platon

Je conclue cette contribution avec quelques mots sur l’ emploi de l’ article avec
nom de personne chez Platon, l’auteur par excellence de discussions et donc
d’alternances des prises de parole.
Comme l’on sait, les discussions platoniciennes sont présentées de manières
diverses, entre autres sous la forme d’un discours rapporté par un narrateur.
Ce narrateur peut être Socrate, comme dans le Charmide et le Lysis, mais dans
d’autres dialogues il y a un narrateur externe, donc une personne qui ne parti-
cipe pas à la dicussion, tels le Banquet et le Phédon30. Je me borne ici au second
type, prenant le Phédon comme exemple. Dans le Phédon, le « titulaire» du dia-
logue, Phédon, rapporte une discussion entre Socrate et quelques-uns de ses
élèves, notamment Cébès et Simmias. Le fait qu’ il s’ agit d’ un récit, d’ une διή-
γησις (voir 59c8: πάντα πειράσομαι διηγήσασθαι), est, tout comme dans le cas de
l’Anabase, d’une importance capitale.
Comme point de départ pour mon argumentation je prends une remarque
brève mais pertinente de John Burnet dans son commentaire sur le Phédon, à
63a231. La plupart des manuscrits ont ici le texte suivant: Ἀκούσας οὖν ὁ Σωκρά-
της ἡσθῆναί τέ μοι ἔδοξε τῇ τοῦ Κέβητος πραγματείᾳ, καὶ ἐπιβλέψας εἰς ἡμᾶς, Ἀεί
τοι, ἔφη, ὁ Κέβης λόγους τινὰς ἀνερευνᾷ, mais le manuscrit T présente Κέβης, sans
l’ article. Cette variante est préférée par Burnet et par la plupart des éditeurs
de Platon; Burnet ajoute la note explicative suivante: « it is Plato’s almost uni-
form practice to insert the article with proper names in the narrative…, and
to omit it in the dialogue when directly reported…»32. Avec cette remarque il
a failli mettre le doigt dessus – failli, car la présence de l’ article n’est pas tant
une caractéristique «of the narrative» que des scènes d’ alternance de prises
de parole, tout comme dans la scène entre Cyrus et Orontas dans l’Anabase33.
Ainsi, | les 38 cas de ὁ Κέβης dans le Phédon apparaissent tous dans des alter- 256
nances des prises de parole34. Quant à Κέβης sans article, il y en a huit exemples,

30 À vrai dire la situation dans le Banquet est plus complexe, parce que le narrateur, Apollo-
dore, raconte à son tour ce qu’ il a entendu d’ Aristodème.
31 À son tour Burnet a utilisé un article de J.I. Beare dans Hermathena 9 (1896), p. 197 sqq.
32 En 86d5, dans une situation tout à fait semblable, T est de nouveau le seul MS qui a Σιμ-
μίας, mais curieusement Burnet, tout comme le nouveau Oxford Classical Text, préfère ici
la variante ὁ Σιμμίας, sans aucune explication dans son commentaire.
33 La présence ici de ὁ s’ explique bien comme un effet de la formule ἔφη ὁ Κέβης, très fré-
quente dans le Phédon, voir ci-dessous.
34 Voir ainsi en 92a2 : ἔφη ὁ Κέβης… ; en 92a4 : ἔφη ὁ Σιμμίας…; en 92a6: Καὶ ὁ Σωκράτης… ἔφη.
Il y a, bien sûr, des variantes sans nom propre (Πάνυ μὲν οὖν, ἔφη, en 65b8, 67d3, …), et sans
nom propre ni verbe (Πάνυ μὲν οὖν, en 68e1), et d’ autres formules avec la même fonction,
 332 chapter 16

dont sept dans les mots d’un participant à la discussion (y compris 63a2, men-
tionné ci-dessus). Une seule fois, en 84c3, Κέβης se trouve donc, dans la partie
narrée, en combinaison avec le nom de son compagnon, Simmias : Κέβης δὲ καὶ
Σιμμίας σμικρὸν πρὸς ἀλλήλω διελεγέσθην. On remarquera qu’ il n’est pas ques-
tion ici d’une alternance de prises de parole. Pour être complet, il faut ajouter
que le second type d’emploi de l’article, celui qui met en relief la position spé-
ciale d’une personne, se trouve également dans le Phédon; ainsi en 117d1: ὁ δὲ
Κρίτων ἔτι πρότερος ἐμοῦ… ἐξανέστη. Ἀπολλόδωρος δὲ… À la fin du Phédon, Cri-
ton est la personne qui assiste Socrate plus que les autres pendant ses derniers
moments35. En revanche, Apollodore – sans article – joue un rôle tout à fait
accessoire; il était, en effet, depuis un certain temps accablé de chagrin.

8 Conclusion

Au début de cette contribution j’ai fait remarquer que selon les grammaires
grecques l’emploi de l’article avec un nom de personne chez les prosateurs est
une chose qui appartient essentiellement au domaine de l’ arbitraire, et que
Platon, par exemple, se sert ou ne sert pas de l’article « avec une liberté totale».
Or, il n’y a rien de moins vrai. J’espère avoir démontré qu’ il y a bien un système
dans ce désordre apparent, au moins – soyons prudents – dans les oeuvres ana-
lysées ci-dessus.
Chez Xénophon et Platon, la fonction de l’article avec un nom de personne
257 est de mettre en relief (le rôle de) cette personne, ou bien | d’ accentuer le
contraste entre celle-ci et une autre personne, dans une situation d’ alternance
des prises de parole (ἔφη ὁ Ὀρόντας… ἔφη ὁ Κῦρος (Xénophon, Anabase) ; ἔφη
ὁ Κέβης… ἔφη ὁ Σιμμίας… Καὶ ὁ Σωκράτης… ἔφη (Platon, Phédon)), ou bien

notamment ἦ δ᾽ ὅς. Les conditions d’ emploi de toutes ces formules sont tout à fait obs-
cures. Cette structure se trouve aussi dans le discours indirect du Banquet, cf. 189a1: ἔφη
εἰπεῖν τὸν Ἀριστοφάνη… ; 189a7 : Καὶ τὸν Ἐρυξίμαχον… φάναι…; 189b3: Καὶ τὸν Ἀριστοφάνη…
εἰπεῖν…, etc. [See further ch. 10 in this volume.]
35 On a une extension de cet emploi « de relief» dans un cas comme Phdr. 269e: Κινδυ-
νεύει… εἰκότως ὁ Περικλῆς πάντων τελεώτατος εἰς τὴν ῥητορικὴν γενέσθαι. Ici, une personne
bien connue, qui n’appartient pas aux personnages du dialogue, est pourvue de l’article
pour mettre en relief sa notoriété: « le fameux Périclès». Noter le caractère solennel
du contexte. Normalement Περικλῆς s’ emploie sans l’article. L’emploi emphatique de
l’ article défini en français dans « Vous êtes bien Billy the Kid? LE Billy the Kid? » (pareille-
ment en d’ autres langues, comme l’ allemand, l’ anglais, l’italien, le néerlandais…) res-
semble à cet emploi de ὁ/ἡ. Voir Gary-Prieur (1994: 100 sqq.); elle explique cet emploi
comme « une anaphore qui renvoie à un présupposé extérieur au discours» (p. 101).
sur l’ article avec nom propre 333

de mettre une personne en relief sans plus, dans certaines situations d’ un

intérêt spécial (ἀναβαίνει οὖν ὁ Κῦρος; ἐθεώρει οὖν ὁ Κῦρος [Xénophon, Anabase] ;
ὁ δὲ Κρίτων… ἐξανέστη [Platon, Phédon]). Il s’agit bien d’ un emploi anapho-
rique, mais par la combinaison avec un nom propre, qui est, lui aussi, employé
anaphoriquement, il y a un effet pragmatique spécial, qu’ on pourrait appe-
ler «anaphore forte». Autre conclusion: la présence de l’ article avec un nom
propre est une caractéristique de textes narratifs (dialogue rapporté, récit his-
torique)36.

Références

Bizos, M., Syntaxe grecque (Paris 1955).

Brown, G., Yule, G., Discourse Analysis (Cambridge 1983)
Dik, H.J.M., Word Order in Ancient Greek: A Pragmatic Account of Word Order Variation
in Herodotus (Amsterdam 1995).
Gardiner, A., The Theory of Proper Names: A Controversial Essay. 2nd ed. (Oxford 1954)
Gildersleeve, B.L., Syntax of Classical Greek (New York 1900–1911).
Goodwin, W.W., Greek Grammar. 2nd ed. (London 1894).
Givón, T., Syntax (Amsterdam/Philadelphia 1990).
Grimes, J.E., The Thread of Discourse (The Hague/Paris 1975).

36 Une question se pose encore, bien sûr, de savoir quelle est la relation entre l’emploi de
l’ article avec nom propre et celui avec nom commun. Puisque le traitement de cette ques-
tion dépasserait largement le cadre de cet article, je me borne ici à signaler l’opinion de
Sansone (1993) sur l’ emploi de l’ article dans un cas comme ἔχεις μοι εἰπεῖν, ὦ Σώκρατες, ἆρα
διδακτὸν ἡ ἀρετή (Pl. Men. 70a). Sansone fait remarquer tout d’abord (p. 201), et à juste titre,
que « Since this is the beginning of a conversation the article obviously cannot be ana-
phoric ». Selon lui, l’ article a ici une fonction « topiqualisante»: «It is used with the noun
because that which the noun denotes is the topic, not only of this sentence, but of the
entire converstion that follows» Mais il ignore ainsi la fonction «identifiante» de l’article,
qui semble tout aussi importante. En effet, on pourrait paraphraser la question de Ménon
par « Si j’ emploie l’ article, Socrate, c’ est que je présuppose que tu saches, tout comme
moi, à quoi ἀρετή se réfère». Or il est intéressant de noter que c’est précisément cette pré-
supposition qui n’est pas acceptée par Socrate, car il dit quelques lignes plus bas (71a) ἐγὼ
δὲ τοσοῦτον δέω εἴτε διδακτὸν εἴτε μὴ διδακτὸν εἰδέναι, ὥστ᾽ οὐδὲ αὐτὸ ὅτι ποτ᾽ ἐστὶ τὸ παράπαν
τυγχάνω εἰδώς. Socrate ne sait donc nullement ce qu’est l’ ἀρετή. À noter que dans la suite
de la conversation on trouve à plusieurs reprises la question τί ἐστιν ἀρετή, jamais avec
l’ article : τί ἐστιν ἀρετή (79c4) ; ἀπόκριναι…· τἰ φῂς ἀρετὴν εἶναι (79e); ζητεῖν ἀρετὴ ὅτι ἐστίν
(81e) ; ζητεῖν τί ποτ᾽ ἐστὶν ἀρετή (86c). Dans tous ces cas ἀρετή semble aussi bien le Topique
que dans la première phrase du Ménon. Voir aussi dans le Lachès à propos d’ ἀνδρεία: ἐπι-
χειρήσωμεν εἰπεῖν ἀνδρεία τί ποτ᾽ ἐστίν (190d) ; πειρῶ εἰπεῖν ὃ λέγω, τί ἐστιν ἀνδρεία (190e),
etc.
334 chapter 16

Humbert, J., Syntaxe grecque (Paris 1960).

Lyons, J., Semantics 2 Vols (Cambridge 1977).
Maes, A., Nominal Anaphors and the Coherence of Discourse (Louvain 1996).
Rijksbaron, A. ‘Sur quelques différences entre οὗτος ὁ + substantif, οὗτος δὲ ὁ + substan-
tif et ὁ δὲ + substantif + οὗτος chez Hérodote’, Lalies 12 (1993) 119–130 [ch. 11 in this
volume].
Sansone, D., ‘Towards a New Doctrine of the Article in Greek: Some Observations on
the Definite Article in Plato’, CPh 88 (1993) 191–205.
Schmidt, C., De articulo in nominibus propriis apud Atticos scriptores pedestres (Kiel
1890).
Svennung, J., Anredeformen: Vergleichende Forschungen zur indirekten Anrede in der
dritten Person und zum Nominativ für den Vokativ (Uppsala 1958).
Zucker, A., Beobachtungen über den Gebrauch des Artikels bei Personennamen in Xeno-
phons Anabasis (Nuremberg 1899).
 chapter 17

Does Ancient Greek Have a Word for ‘No’? The

Evidence from οὐκοῦν (…) οὐ Questions

Abstract

Among the uses of οὐκοῦν Denniston (1954: 235) mentions ‘οὐκοῦν οὐ, οὐκοῦν … οὐ,
expecting a negative answer.’ This paper argues that Denniston’s view, which is shared
by most (all?) other grammars and dictionaries of Ancient Greek (e.g. K-G 2, 164: ‘Wenn
nach οὐκοῦν eine verneinende Antwort erwartet wird, so wird demselben die Negation
οὐ nachgesetzt’, Smyth §2651 a: ‘οὐκοῦν οὐ expects the answer no’, LSJ s.v. οὐκοῦν) should
be rejected. Actually, the answer is never no. As always, οὐκοῦν expects an affirmative
answer, in this case to a negated question: ‘Is it not true, then, that not X?’ = ‘Surely, then,
not X?’ To be sure, οὐ does occur as an answer, but this can be shown to be a proposition
(or sentence) negative (= not), rather than an answering particle like no. The situation
in Greek is compared with negatives in several other languages, notably Latin and Old
French. Finally, Modern Greek is briefly discussed, which, unlike Ancient Greek, does
have a negative answering particle, viz. όχι, alongside a proposition negative, viz. δε(ν).

…
… negation is to the linguist and linguistic philosopher as fruit to Tanta-
lus: waving seductively, alluringly palpable, yet just out of reach, within
the grasp only to escape once more
– Horn (1989: xiv)

…
Aus diesen den neueren Sprachen gewidmeten Arbeiten ist auch für
Griechisch und Latein viel zu lernen
– Wackernagel (1928: 249)

© koninklijke brill nv, leiden, 2019 | doi:10.1163/9789004386129_019

336 chapter 17

1 Preliminaries

We may readily agree with Horn that negatives are among the most elusive
parts of any natural language.* One may think of phenomena like the double
negation in English (‘I don’t know nothing about it’) and other languages, the
presence in French of so-called ‘expletive’ ne after verbs like empêcher, and in
plus souple qu’on ne le pensait, the existence of opaque negative conjunctions
like lest, and of negative questions with positive meaning like Isn’t that a Rem-
brandt?, to which I will return in more detail below. Greek is of course a gold-
mine of such phenomena.1 One needs only think of the combinations οὐδεὶς …
οὐ nobody … not and οὐκ … οὐδείς not … nobody, and of constructions like ἀπα-
γορεύω μὴ διαλέγεσθαι I forbid (you) to have a discussion and οὐκ ἀπαγορεύω μὴ
οὐ διαλέγεσθαι I don’t forbid you to have a discussion, or οὐ μή + subjunctive in
declarative utterances, as in οὐ μὴ πίθηται he won’t obey. To these phenomena
also belongs the use of οὐ in questions which are already negative, type οὐκοῦν
… οὐ;, which will be the subject of the first half of my paper. They are treated in
some detail, because the answers to these questions will enable me to propose
an answer the question of the title: Does Ancient Greek have a word for ‘No’?
Grammars and monographs alike regard the combination οὐκοῦν (…) οὐ as
an expression which expects a negative answer. Here follow the views of Den-
niston and others:

(1) – ‘Wenn nach οὐκοῦν eine verneinende Antwort erwartet wird, so wird
demselben die Negation οὐ nachgesetzt’ (K-G 2, 164)
– ‘οὐκοῦν οὐ, οὐκοῦν … οὐ, expecting a negative answer’ (Denniston 1954:
435)
– ‘οὐκοῦν οὐ expects the answer no’ (Smyth 1956: § 2651)
– ‘(οὐκοῦν) folld. by οὐ when a negative answer is invited’ (LSJ s.v. οὐκοῦν
I)2

* For their valuable comments and suggestions I would like to thank the two anonymous refer-
ees for the Journal of Greek Linguistics and the participants at the 2010 Nijmegen conference
‘Ancient Greek and Semantic Theory’, especially Maribel Romero, who commented on a pre-
liminary version of the present paper. That version was published, incidentally, as Rijksbaron
2009. Hero Hokwerda provided useful information on Medieval Greek.
1 For readers who are not familiar with the language I note that Ancient Greek has two negative
particles, οὐ and μή. Basically, οὐ is used in assertions and in a certain type of questions (more
on these below), while μή is rather used in directives and wishes, and in a different type of
questions than οὐ, as well as in most dependent clauses and with the infinitive after volitional
and related verbs. See also n. 6.
2 Note that Smyth simply gives Engl. no as an answer, while Kühner-Gerth, Denniston and
does ancient greek have a word for ‘no’? 337

Since according to the same authors bare οὐκοῦν expects or invites an affir-
mative answer one might expect to find some explanation as to how οὐκοῦν … οὐ
comes to have the opposite meaning, but this is not given. Perhaps this should
not surprise us, for their view cannot be defended. To show this I start with a
discussion of bare οὐκοῦν,3 which will also involve the rather thorny question
as to what exactly is the value of the negative in such questions.

2 Questions with Bare οὐκοῦν and the Answers to οὐκοῦν4

In answers to questions with bare οὐκοῦν we find inter alia the following items:5

Veridical ἐστί and Related Expressions: ἔστι ταῦτα ‘that is the Case’, ἔστιν οὕτω
‘(that) is so’, ἀληθῆ ‘true’, ὀρθῶς ‘correct’

(2) ΣΩ. Οὐκοῦν καὶ περὶ τοῦ μείζονος καὶ ἐλάττονος εἰ διαφεροίμεθα, ἐπὶ τὸ
μετρεῖν ἐλθόντες ταχὺ παυσαίμεθ᾽ ἂν τῆς διαφορᾶς;
ΕΥ. Ἔστι ταῦτα. (Pl. Euthphr. 7c)

So.: Then, too, if we were to disagree about the relative size of things, we
should quickly put an end to the disagreement by measuring?
Eu. Yes.6

Liddell-Scott-Jones do not specify what they mean by ‘negative answer’. They must have con-
sidered it self-evident that this was οὐ, and that this was the equivalent of German nein and
Engl. no, as is also implied by Smyth. Wrongly, as I will argue in this paper.
3 My argument is based on the current texts of the authors mentioned, which sometimes do
not print a question mark after οὐκοῦν, although they clearly elicit an answer. ‘I should be
inclined,’ writes Denniston (1954: 433), ‘to … insert the question-mark everywhere.’—I am
ignoring the difficult problem of the accentuation of ουκουν, for which see Denniston (1954:
430 f.).
4 Unless indicated otherwise the examples are from the very rich corpus of interrogative dis-
course provided by Plato.
5 I give notably those answers that are relevant for the subsequent discussion of οὐκοῦν … οὐ.—
Question words like ἆρα, bare οὐ, ἆρ᾽ οὐ, μή, ἆρα μή, μῶν are not discussed in this paper, since
they turned out to be less relevant for the purpose of this investigation than οὐκοῦν … οὐ.
Where this seemed relevant, however, I have added references to the use(s) of these question
words.
6 Unless indicated otherwise the translations are taken from the translations in the Loeb series,
without adaptations, although these might often be considered. Thus, in (2) the conducive
force of the question, i.e. the fact that ‘[it [indicate[s] that the speaker is predisposed to the
kind of answer he has wanted or expected’ (Quirk et al. 1985: 808) is not expressed. This could
be remedied e.g. by adding a tag question after ‘measuring’: isn’t it?. The answer, too, is unsat-
338 chapter 17

(3) Οὐκοῦν τοιοῦτον (sc. μουσικόν) μὲν εὑρόντες ἐπειθόμεθ᾽ ἄν, ἄμουσον δέ, ἠπι-
στοῦμεν; :: Ἀληθῆ. (Pl. Tht. 144e)

Then if we found that he was a musician, we should believe him, but if

not, we should refuse to take his word? :: Yes.

Verbs Expressing Agreement: φημί ‘I say so’, ὁμολογῶ ‘I agree’ and Compounds

(4) Οὐκοῦν … ὅταν … εἷς τις ἄριστα λέγͅη, γνώσεται δήπου τις τὸν εὖ λέγοντα; ::
Φημί. (Pl. Ion 531e)

Well now …; when one of them speaks better than the rest, I suppose there
is some one who will distinguish the good speaker? :: I agree.

Also ‘narrated agreement’ in a past tense, in dialogues having a narrator, e.g.7

(5) Οὐκοῦν, ἦν δ᾽ ἐγώ, ἑνὶ ἑκάστῳ τῶν ἐναντίων ἓν μόνον ἐστὶν ἐναντίον καὶ οὐ
πολλά; :: Συνωμολόγει. (Pl. Prt. 332c)

Now, I went on, each single opposite has but one opposite, not many? ::
He admitted this.

Words Expressing Various Degrees of Likelihood: ἀνάγκη ‘(that’s) inevitable’,

φαίνεται ‘(so) it seems’, εἰκός (γε) ‘(that’s) likely’, δοκεῖ μοι ‘(so) it seems to me’,
κινδυνεύει ‘(that’s) probable’

(6) Οὐκοῦν καὶ ἐνταῦθα ἡ τοῦ φιλοσόφου ψυχὴ μάλιστα ἀτιμάζει τὸ σῶμα …; ::
Φαίνεται. (Pl. Phd. 65c–d)

In this matter also, then, the soul of the philosopher greatly despises the
body …? :: Evidently.

isfactory, for ἔστι ταῦτα obviously does not mean ‘Yes’. There is, in fact, a tendency in these
translations to abandon the variety of the Greek answers by rendering affirmative answers
simply as ‘Yes’.
7 This includes ἐπινεύω ‘nod (yes)’, which naturally occurs only in narrated dialogues, e.g. Prt.
360c6–7 Οὐκοῦν ἡ τῶν δεινῶν καὶ μὴ δεινῶν ἀμαθία δειλία ἂν εἴη; Ἐπένευσε. (‘Then ignorance of
what is dreadful and not dreadful will be cowardice? He nodded assent’).
does ancient greek have a word for ‘no’? 339

Affirmative Adverbs and Adverbial Phrases, the Greek Equivalents of ‘cer-

tainly/yes/by all means/of course’: καὶ μάλα, μάλα γε, μάλιστα, πῶς δ’ οὐ, πῶς
γὰρ οὐ, πάνυ γε, παντάπασί γε, παντάπασι μὲν οὖν, ναί

(7) Οὐκοῦν ἐντεῦθέν ποθεν πρῶτον ἐπέρχεται ἐρέσθαι ἡμῖν τί οὖν ποτ᾽ ἐστὶ τὸ μέγα
αὖ καὶ τὸ σμικρόν; :: Παντάπασι μὲν οὖν. (Pl. R. 524c)

And is it not in some such experience as this that the question first occurs
to us, what in the world, then, is the great and the small? :: By all means.

(8) Οὐκοῦν νῦν φαμεν περὶ μαθήματος σκοπεῖν τῆς ψυχῆς ἕνεκα τῆς τῶν νεανί-
σκων; :: Ναί. (Pl. La. 185d–e)

And we say that our present subject is an accomplishment studied for the
sake of the young men’s souls? :: Yes.

Affirmative Particle Attached to a Constituent to be Supplied from the Ques-

tion

(9) Οὐκοῦν τόνδε τοῦτο ἐρωτᾷς; :: Ἔγωγε. (Pl. La. 194e)

Well, will you put that question to your friend here? :: I do.

Repetition of the Word Which Was the Focus of the Question, Sometimes
with Adaptations, Often Strengthened by an Additional Affirmative Expres-
sion, or by an Emphatic Adverb

(10) Οὐκοῦν αὖ ὡμολογήσαμεν …; :: Ὡμολογήσαμεν, ἔφη· πῶς γὰρ οὔ; (Pl. Phd. 94c)

Did we not agree …? :: Yes, he replied, we did, of course.8

(11) Οὐκοῦν οἰκτρὸν ἂν εἴη τὸ πάθος, εἰ …; :: Νὴ τὸν Δία … οἰκτρὸν δῆτα. (Pl. Phd.
90c–d)

Wouldn’t it be a sad experience, if …? :: By Zeus, definitely sad.

8 Compare, in a bare yes/no question And. 1.14: ἦσθα ζητητής, ὦ Διόγνητε, ὅτε …; :: Ἦν. :: Οἶσθα
οὖν μηνύσαντα Ἁνδρόμαχον …; :: Οἶδα. (‘You were on the commission of enquiry, Diogne-
tus, when …? :: I was. :: So you know that Andromachus laid an information as to …? :: I
340 chapter 17

It is clear from this survey that οὐκοῦν expects or invites, in fact, an affirma-
tive answer, i.e. the speaker invites the interlocutor to confirm the truth of the
proposition underlying the question.9 The sequence of questions and answers
can be paraphrased as follows:

Q οὐκοῦν X; = ‘Isn’t it true, then, that X?’= ‘Surely, then, X?’

A ‘Indeed, X’10

3 The Semantics of οὐ in οὐκοῦν Questions

Before I turn to οὐκοῦν … οὐ questions perhaps a few words about the seman-
tics of the negative in οὐκοῦν questions are in order, since this may not be
immediately clear. In this combination the meaning of οὐ, just like that of bare
οὐ and ἆρ᾽ οὐ introducing questions, can be paraphrased as ‘Isn’t it true that
X …?’, which pragmatically, and in positive terms, amounts to ‘Surely X …?’11
How can it be that a negative particle is meant to elicit an affirmative answer?
Observe that the same problem exists for English not, in simple conducive
questions, like Hasn’t he told you what to do? = ‘Surely he has told you what
to do?’, and especially in questions with checking tags12 (He has told you what
to do, hasn’t he?), which express ‘maximum conduciveness’ (Quirk et al. 1985:

know.’)—In some languages this type of answer is obligatory, e.g. Old Welsh: A eisteddi di?
:: Eisteddaf. ‘Will you sit down? :: I will sit’ (Ball 1993: 337).
9 ‘Proposition’ is used as in Dik’s Functional Grammar, i.e. a ‘possible fact’, ‘something which
can be known, believed, mentioned, remembered, etc.’ (Dik 1997: 1, 65–66). Traditionally,
proposition negatives are called, more loosely, ‘sentence negatives.’
10 To avoid a possible misunderstanding I should add that the affirmative items mentioned
above are by no means confined to answers to οὐκοῦν questions. They may also appear, for
instance, after (neutral) ἆρα questions, as in Pl. Grg. 462b Ἆρα ἐρωτᾷς ἥντινα τέχνην φημὶ
εἶναι; :: Ἔγωγε (‘Are you asking what kind of art I call it? :: Yes’ (liter.: ‘Yes I (am)’)); cp. ex.
(9); R. 566a ἆρα ἀνάγκη …; :: Πολλὴ ἀνάγκη (‘Is it then inevitable …? :: Quite inevitable’); cp.
ex. (10). See also below.
11 As for -οῦν in οὐκοῦν, this presents the question at hand as the consequence of a point
raised and discussed in the preceding discourse. In fact, οὐκοῦν questions never come first
in a series of questions. First questions are rather introduced by ἆρα, negative ἆρ᾽ οὐ or
bare οὐ. Also, with a different meaning, ἆρα μή and bare μή, and μῶν. See also n. 36.
12 For the term see Lyons (1977: 764). The positive conducive force of negative questions
is generally known as ‘positive bias’. For a discussion of positively and negatively biased
questions in English (the latter for instance in Do you really want to leave now? [‘Surely
you don’t want to’]; Quirk et al.’s example) see e.g. Quirk et al. (1985: 808–809) and Lyons
(1977: 764–766), and, in a formal semantic framework, Romero & Han (2004), and Asher
& Reese (2005). See also n. 13.
does ancient greek have a word for ‘no’? 341

810). Of course Dutch niet, French ne … pas, German nicht, and no doubt the
negatives in many other languages pose similar problems.
According to Quirk et al. (1985: 808) a question like Hasn’t he told you what to
do? implies ‘that the speaker had originally hoped for a positive response, but
new evidence now suggests that the response will be negative.13 Thus, Hasn’t he
told you what to do? means “Surely he has told you what to do, hasn’t he? I would
have thought that he had told you.” Here there is a combining of positive and
negative attitude, which one may distinguish as the old expectation (pos-
itive) and new expectation (negative).’ They do not explain, however, how
such a sentence can elicit a positive response. Perhaps the simplest explanation
of this phenomenon is that of Moore (1934: 165): ‘… the interrogative form is a
form of doubt or denial and cancels the οὐ negation, so that the two negatives
make a positive.’ A more elaborate version of this view may be found in Hud-
son (1975). While discussing the sentences (Hudson’s numbers) (37a) Do you
enjoy making me miserable? and (37b) Don’t you enjoy making me miserable?,
Hudson notes (1975: 17–18): ‘In both these sentences the speaker is calling into
question the truth of the proposition, which in (37a) is positive and in (37b) is
negative. From (37a–b) respectively, then, the hearer can deduce that the cor-
responding negative proposition is true (cf. Surely you don’t enjoy making me
miserable?), and that the corresponding positive proposition is true (cf. Surely
you enjoy making me miserable?).’14 To this we should add that the hearer not
only can make deductions, but by the interrogative form is explicitly invited
to confirm that the corresponding negative or positive proposition is, in fact,
true, in the first case by saying No(, indeed I don’t), in the second by saying Yes(,
indeed I do).
All in all, negative questions with positive bias would seem to function as
a pragmatic-rhetorical device which is used to ensure the cooperation of the
interlocutor in the communicative setting at hand.15 Much more could be, and

13 In the same vein Lyons, in a discussion of the questions (Lyons’ numbers) (21) Is the door
open? and (22) Isn’t the door open?, argues (1977: 765) that ‘[t]he speaker utters (22) rather
than (21) because there is some conflict between his prior belief that p is true and present
evidence which would tend to suggest that ~p is true. He questions ~p because it is the
negative proposition that occasions his doubt or surprise.’
14 For negative questions this analysis might be rephrased in slightly more formal, and
speaker-oriented, terms as: by asking a question of the form ~p? the speaker calls into
question the truth of ~p, thus implying that he believes that p rather than ~p is true, and
asks the addressee to confirm the correctness of this belief. Alternatively, the combined
presence of question and negation could also be viewed as involving an entailment rela-
tionship: if the speaker questions the truth of ~p, he must consider p true.
15 The invitation to react affirmatively may of course also be rejected, but with οὐκοῦν
342 chapter 17

indeed has been, said about negative questions,16 but for the purpose of this
paper the above observations may perhaps suffice.
I now turn to οὐκοῦν … οὐ questions.

4 Questions with οὐκοῦν … οὐ and the Answers to οὐκοῦν … οὐ

In answers to questions with οὐκοῦν … οὐ we find inter alia the following items
(for their meaning see above, at bare οὐκοῦν):

Veridical ἐστί and Related Expressions

(12) Οὐκοῦν ὅστις ἂν μὴ ἔχῃ τινὰ τέχνην, ταύτης τῆς τέχνης τὰ λεγόμενα ἢ πραττό-
μενα καλῶς γιγνώσκειν οὐχ οἷός τ᾽ ἔσται; :: Ἀληθῆ λέγεις. (Pl. Ion 538a)

Then he who has not a particular art will be incapable of knowing aright
the words or works of that art? :: True.

Verbs Expressing Agreement

Lacking

Words Expressing Various Degrees of Likelihood: ἀνάγκη, φαίνεται, εἰκός (γε),

δοκεῖ μοι

(13) Οὐκοῦν ἢ ἄρτι οὐ καλῶς ὡμολογοῦμεν, ἢ εἰ τότε καλῶς, νῦν οὐκ ὀρθῶς τιθέμεθα;
:: Ἔοικεν. (Pl. Euthphr. 15c)

Then either our agreement a while ago was wrong, or if that was right, we
are wrong now? :: So it seems.

this seems to have been extremely rare, in Plato, at least. Normally, the interlocutor co-
operates. (‘[T]he answers given in the (Platonic) dialogues are usually the expected ones’
(Denniston 1954: 275)). An example of a non-co-operative interlocutor is the rhapsode Ion
at Ion 541a5: οὐκοῦν καὶ ὅστις ἀγαθὸς στρατηγὸς τυγχάνει ὤν, ἀγαθὸς καὶ ῥαψῳδός ἐστιν; :: Οὐκ
αὖ μοι δοκεῖ τοῦτο. (‘Is he who happens to be a good general not also a good rhapsode? ::
This time I don’t agree.’)
16 Recently, negative questions have attracted the interest of several formal semanticists. See
e.g. Romero and Han (2004), which is a thorough study of our questions, also in other lan-
guages than English, with further literature, and Romero (2005) for some modifications
to the earlier paper. To account for the positive bias of negative yes/no questions Romero
and Han make use, among other things, of an epistemic operator VERUM, that formally
captures the informal paraphrase ‘Isn’t it true that …?’ used above.
does ancient greek have a word for ‘no’? 343

Affirmative Adverbs and Adverbial Phrases: καὶ μάλα, μάλα γε, μάλιστα, πῶς
γάρ, πάνυ γε, παντάπασί γε, παντάπασι μὲν οὖν

(14) Οὐκοῦν οὐκ ἂν εἴη τὸ μὴ λυπεῖσθαί ποτε ταὐτὸν τῷ χαίρειν; :: Πῶς γὰρ ἄν (sc.
εἴη);17 (Pl. Phlb. 43d)

Then freedom from pain would not be identical with pleasure? :: Certainly
not.

(15) Οὐκοῦν καὶ εἰκόνας γραμμάτων, εἴ που ἢ ἐν ὕδασιν ἢ ἐν κατόπτροις ἐμφαίνοιντο,

οὐ πρότερον γνωσόμεθα, πρὶν ἂν αὐτὰ γνῶμεν, …; :: Παντάπασι μὲν οὖν (Pl. R.
402b.)

And is it not also true that if there are any likenesses of letters reflected in
water or mirrors, we shall never know them until we know the originals
…? :: By all means.

Affirmative Particle Attached to a Constituent to be Supplied from the Ques-

tion

(16) Οὐκοῦν εἴ τι μὴ ἐπίστασαι, οὐκ ἐπιστήμων εἶ; :: Ἐκείνου γε, ὦ φίλε ἦν δ᾽ ἐγώ
(Pl. Euthd. 293c)

Then if you do not know something, you are not knowing? :: Not in that
thing, my dear sir, I replied.

Repetition of the Word Which was the Focus of the Question (= οὐ), Often
Strenghtened by an Additional Affirmative Expression

(17) Οὐκοῦν ψυχὴ οὐ δέχεται θάνατον; :: Οὔ. :: Ἀθάνατον ἄρα ψυχή. :: Ἀθάνατον. (Pl.
Phd. 105e)

17 Observe that οὐκοῦν questions may be answered by πῶς γὰρ οὔ; (e.g. (10) above), while
οὐκοῦν … οὐ questions may be answered by πῶς γάρ; ((14) above). There is thus a parallelism
οὐκοῦν … ; – πῶς γὰρ οὔ; and οὐκοῦν … οὐ; – πῶς γάρ;. Both these rhetorical answer-questions
emphatically confirm the truth of the proposition underlying the preceding question, πῶς
γὰρ οὐ by expressing the idea that in no way it cannot be true (πῶς γὰρ οὐ sc. ὡμολογήσαμεν;
‘how/in what way did we not agree?’ = πάντως ὡμολογήσαμεν ‘in all ways we agreed’), πῶς
γὰρ ἄν by expressing that in no way it can be true (πῶς γὰρ ἄν, sc. εἴη ταὐτόν; ‘how could it
be the same?’ = οὐδαμῶς ‘in no way’, sc. ἄν εἴη ταὐτόν).
344 chapter 17

And the soul does not admit death? :: No. :: Then the soul is immortal. ::
Yes.

(18) Οὐκοῦν εἰ ταῦτα οὕτως ἔχει, οὐκ ἂν εἶεν φύσει οἱ ἀγαθοί. :: Οὔ μοι δοκεῖ. (Pl.
Men. 89a)

Then if this is so, good men cannot be good by nature. :: I think not.

(19) Οὐκοῦν ἐπειδὴ οὐ φύσει, οὐδὲ οἱ ἀγαθοὶ φύσει εἶεν ἄν. :: Οὐ δῆτα. (Pl. Men.
98d)

Then since they are not natural, good people cannot be good by nature
either. :: Of course not.

(20) Οὐκοῦν οὕτω γε οὐ δεῖ παιδοποιεῖσθαι; :: Οὐ γὰρ οὖν, ἔφη. (X. Mem. 4.4.23)

In that case then, they ought not to have children? :: Most certainly not.

It is clear from this survey that, just like bare οὐκοῦν, οὐκοῦν … οὐ, too, expects
or invites an affirmative answer:

Q οὐκοῦν (…) οὐ X; = ‘Isn’t it true, then, that not X?’ = ‘Surely not X?’
A ‘Indeed, not X’

To be sure, this would seem to be contradicted by the final category, where we

find οὐ as an answer. Actually, however, the form of the answer differs in no
way from the answers in the final category of οὐκοῦν, see examples (10) and
(11). This is seen most clearly in (17); here, the οὐ of the answer repeats the οὐ
of the question, thereby confirming its correctness, just as in the answer to the
next question Ἀθάνατον affirmatively repeats the Ἀθάνατον of the question.18
In all such cases, then, οὐ should not be taken as No, i.e. an answering par-
ticle, but as not, i.e. a proposition negative, which in answers functions as an
elliptic substitute for the entire underlying negative proposition of the ques-
tion: (ex. 17) Οὐκοῦν (ψυχὴ) οὐ (δέχεται θάνατον); :: Οὔ = ψυχὴ οὐ δέχεται θάνατον

18 Cp. also above, nn. 8 and 10. As in the case of οὐκοῦν (see n. 6), the translations of the
answers sometimes have a negative where there is no negative in the Greek, see (14) and
(16). These negatives must have been viewed by the translators as the most adequate
means to render the meaning of the answer. But this is a matter of English rather than
Greek syntax.
does ancient greek have a word for ‘no’? 345

(‘Isn’t it true, then, that (the soul) does not (admit death)?’—‘Not’ (= the soul
does not admit death)), (ex. 20) Οὐκοῦν (οὕτω γε) οὐ (δεῖ παιδοποιεῖσθαι); :: Οὐ
γὰρ οὖν = οὕτω γε οὐ δεῖ παιδοποιεῖσθαι. (‘Isn’t it true, then, that (in that case) not
(one must make children)?’—‘Most certainly not’ (= In that case one must not
make children)). Note, in the latter example, the presence of strongly assever-
ative γὰρ οὖν.
If the above analysis is correct it follows that Ancient Greek does not have a
separate negative answering particle like Eng. No, Du. Nee, Fr. Non, Ge. Nein, It.
No, Russian njet, alongside the proposition negatives not, niet, ne … pas, nicht,
non, ne, respectively.19 Note also that in (19) οὐ is emphasized by the particle
δῆτα certainly, surely. On the assumption that δῆτα is like certainly and similar
adverbs in other languages, this, too, points to οὐ being a proposition negative
rather than an answering particle. Compare, in English, answers like Certainly
not, Perhaps not, where *Certainly no and *Perhaps no are ungrammatical. Fur-
ther support for this claim can be derived from the combinations Τί οὐ; and Διὰ
τί οὐ; = Why not? Cp. indeed Engl. Why not?, not *Why no? Likewise for Dutch
Absoluut niet/*Absoluut nee, Waarom niet?/*Waarom nee?; German Bestimmt
nicht/*Bestimmt nein, Warum nicht?/*Warum nein?20 Observe that the answers
in the English Loeb translations of (18)–(20) have not, and rightly so: in answers
not, too, functions as an elliptic substitute for the entire negative proposition
of the question.
It might be objected that the above claim is too strong, since it is only based
on the behaviour of οὐκοῦν … οὐ. In fact, it is this combination, plus the answers,
that allows us to make inferences about the word class of οὐ. Naturally, this has
to be checked against the behaviour of οὐ in other environments, e.g., impor-
tantly, in answers after non-biased (neutral) questions. These have either no
introductory question word, or they are introduced by ἆρα or ἦ. Here follow
some examples with οὐ as an answer:

19 For the Romance languages see further n. 20 and section 5.

20 Cf. also, in Greek, Pl. Chrm. 164b: ἴσως οὔ (‘perhaps not’). In French both Pourquoi pas?
and Pourquoi non? are in use, and so are Certainement pas and Certainement non. In these
uses non may continue its original use as a propositon negative, for which cp. section 5,
and eventually Latin Quor non? as in Pl. Pseud. 318: Tibi ego credam? :: Quor non? ‘Should
I believe you? :: Why not?’ and Cic. de domo sua 51: Quid enim diceres? Damnatum? Certe
non (‘For what could you say? That he has been condemned? Certainly not’). Remarkably,
in Italian, that has both the proposition negative non and the answering particle No, it is
no, not non that can be modified by perche and certamente: perche no?, not *perche non?
and certamente no, not *certamente non. In view of the situation in French and Spanish,
Italian no, too, should probably be considered a proposition negative in this use.
346 chapter 17

(21) ἔχομεν φράζειν; :: Οὐ δῆτα, ὦ ξένε, σαφῶς γε. (Pl. Lg. 962c; no question word)

Can we answer? (viz. the preceding question) :: Certainly not clearly21

(22) Ἆρ᾽ οὖν οἷόντε εὖ διοικεῖν … πόλιν …, μὴ σωφρόνως … διοικοῦντα; :: Οὐ δῆτα.

(Pl. Men. 73a)

And is it possible to manage a state well … if you do not manage it tem-

perately …? :: Surely not

From the comedies of Aristophanes:

(23) ἔδωκας ἤδη τουτῳί κάττυμα παρὰ σεαυτοῦ ταῖς ἐμβάσιν …; :: οὐ δῆτα μὰ τὸν
Ἀπόλλω. (Ar. Eq. 870; no question word)

Have you ever given him a free patch for his shoes …? :: By Apollo, he never
has (said by a third person)

(24) ἀλλ᾽ ἦ παραφρονεῖς ἔτεον …; :: οὔκ, ἀλλ᾽ ὕπνος μ᾽ ἔχει τις ἐκ Σαβαζίου (Ar. V. 8)

Whoa there, are you losing your mind …? :: No, Sabazius has put me under
a sleepy spell22

In all four cases οὐ can be analysed as an elliptic negative, which denies the
truth of the proposition underlying the preceding question.23 Notice especially
οὐ δῆτα in (21), (22) and (23), comparing (19) plus discussion. (23) is particularly
interesting, since the answer does not stem from the addressee of the question
but from a bystander. This makes ‘No’ virtually impossible, for this can only be
said by the addressee: Q Are you ill? A No. Q Are you ill? A by third person: *No.
This is only acceptable in the form No, he isn’t, or, without No: He isn’t.24

21 Translation mine. The Loeb translation has: ‘No Stranger, at least, not clearly’, where δῆτα
is not translated, while οὐ is translated twice, first by No and then by not, apparently to
avoid the impossible *at least, no clearly.
22 In the analysis of οὐ advocated here, the No of Henderson’s Loeb translation should rather
be I’m not.
23 Note that in these cases the οὐ of the answer denies the truth of the underlying positive
proposition (e.g. in (24): Q παραφρονεῖς; ‘Are you losing your mind?’, underlying propo-
sition: παραφρονεῖς ‘You are losing your mind’; A οὔκ = οὐ παραφρονῶ ‘I’m not losing my
mind’), while in the οὐκοῦν questions of (17)–(20) it confirms the truth of the underlying
negative proposition.
24 Οὐ is also found in reactions to statements, e.g. S. Tr. 413–414: ἄπειμι. μῶρος δ᾽ ἦ πάλαι κλύων
does ancient greek have a word for ‘no’? 347

In sum, since I have found no examples where οὐ must be taken as a negative

answering particle, i.e. as No, but several in which οὐ cannot be taken as such a
particle (notably in οὐ δῆτα and οὐ γὰρ οὖν), I conclude, applying Occam’s razor,
that οὐ in answers everywhere functions as an elliptic negative proposition par-
ticle.

5 Latin and Other Parallels

Ancient Greek is by no means alone among Indo-European languages in hav-

ing only one negative particle (apart from prohibitive μή of course). Latin, too,
has only one negative (apart from prohibitive nē),25 which may very well be
taken as a proposition negative only, and which goes back to a neuter adverbial
accusative *nē + oinom ‘in not one thing, in no respect’. Thus, I would take Non
in Pl. Pseud. 1067: Venitne homo ad te? :: Non (‘Has the man come to you? :: Neg-
Ptcle’) not as the equivalent of No, as proposed by Thesleff (1960: 56), but just
like Greek οὐ as an elliptic substitute for the full proposition: ‘Non’ = non venit
ad me (= he has not come to me). A fuller phrase, with repetition of the verb
of the question, occurs e.g. at Ter. Ad. 569 Estne frater intus? :: Non est. (‘Is your
brother inside? :: He is not’).26 For more examples see Thesleff (1960: 56) and
cp. also Kühner-Stegmann 2, 531.27
In the Romance languages, the Latin situation is still present in Spanish,
Portuguese and Romanian, where there is just one negative, no, não and nu,
respectively.28 Other I-E languages with only one negative are Czech and other
Slavonic languages (ne) (but apparently not Russian, which has the proposi-
tion negative ne and the answering particle njet) and Albanian ( jo). Outside

σέθεν. | :: οὔκ, πρίν γ᾽ ἂν εἴπͅ ης … (‘I shall be off; I was a fool to listen to you for so long. :: No,
not before you’ve answered …’). The fact that οὔκ is modified here by a temporal clause to
my mind proves that it does not function as an answering particle (*‘No, before …’) but as
a negative proposition particle (‘Not before …’); the lexicon of Ellendt-Genthe correctly
explains: οὔκ, sc. ἄπει (‘You will be off’). Note that in the Loeb translation the initial choice
for ‘No’ made it necessary to continue with ‘not’.
25 And haud, which has a far more restricted use than non. See Kühner-Stegmann 1, 813–816.
Haud is not found in answers.
26 Interestingly, Latin does not have an exclusive word for ‘Yes’ either. The common ita ‘so
(is it)’ in answers should probably also be taken as an elliptic substitute for a proposition,
which confirms the correctness of the proposition of the preceding question. See e.g. Pl.
Capt. 262 captus est? :: ita.
27 Orlandini (2001) does not discuss the problems involved in establishing the function of
non.
28 Cp. Sp. ¿porqué no?, with Zeijlstra (2004: 154). For French and Italian see above n. 20.
348 chapter 17

the I-E family one may compare e.g. Hungarian (nem) and Arabic (lâ). In all
these cases the negative should be considered a proposition negative.29 As for
French, in Oldest French the Latin system, too, was still extant, there being only
one negative, non (Cantilène de sainte Eulalie ca. 880). Gradually, a differentia-
tion occurred, leading on the one hand to ne, an elided form of non, which was
used proclitically and was strengthened ‘dès son apparition’ (Buridant 2000:
707) by words like pas, point, mie (from Latin mica) and on the other to non,
which on its own could be used as an answer. See the detailed discussion in
Buridant (2000: 697–726). Next, due to the process known as Jespersen’s Cycle
(see Jespersen 1917: 4), proclitic ne disappeared, at least in spoken and written
informal French, transferring its negative value to pas, point, etc.30 Remnants of
the old situation, where non was used as a proposition negative, are still present
in expressions like je pense que non, qu’il le veuille ou non.

6 Modern Greek

Unlike Ancient Greek, the modern language does have two negatives, όχι in
answers, and δέν/δε(ν) with propositions. The latter form goes back to οὐδέν,
itself originally a negative neuter adverbial accusative of οὐδείς (‘no(t) one’),
so ‘in no respect at all’ (cp. Latin non < *ne +¯oinom), which after the Classical
period gradually ousted οὐ as a negative, or to be more precise as a proposi-
tion negative. Cp. the entry in Hesychius’ lexicon (5th–6th c.): οὐδὲν ἀντὶ τοῦ
οὐ. The subsequent development of οὐδέν into δέν can be explained along the
lines described above for French ne … pas.31 Incidentally, according to Cowgill
(1960), who was followed by Chantraine (1968–1980; hesitatingly: ‘plus plau-

29 The fact that these languages have only one—proposition—negative pleads against the
view that generally in child language the acquisition of ‘no’ precedes that of ‘not’; see for a
discussion Horn (1989: 154–168). If this occurs it must be a language-specific phenomenon.
30 Jespersen’s Cycle is still fully alive in French, see Du tout for Pas du tout in Êtes-vous orig-
inaire de Biarritz? :: Du tout. Je suis de Bordeaux. Perhaps the obligatory final accent of
French words and word-groups plays a role here?—The process is also known from Jes-
persen’s native Danish and the other Scandinavian languages. Cp. Danish ikke ‘not’ (<
*eit-gi, < *ne eit-gi ‘not ever a thing’; for ei(t)- cp. Dutch eeuw etc. (below, n. 32)). Note
that Danish has also a separate answering particle, Nej.
31 The working of Jespersen’s Cycle in Modern Greek can also be detected in the use of the
indefinite pronouns κανείς/κανένας ‘anyone’, τίποτα(-ε) ‘anything’ in answers, where they
have the meaning ‘Nobody’, ‘Nothing’. Cp. Δε θέλω τίποτα ‘I don’t want anything’, with the
negative, alongside Τι θέλεις; :: Τίποτα ‘What do you want?’ :: ‘Nothing’, without the nega-
tive, and Ποιος ήρθε σήμερα; :: Κανένας (‘Who came today?’ :: ‘No one’, without the negative).
See Holton et al. (1997: 321–324 and 420–422), with further examples. For the conditions
under which κανένας etc. mean ‘nobody’ etc. see Giannakidou (1999). For the working of
does ancient greek have a word for ‘no’? 349

sible’, viz. than other etymologies) and Beekes (1995; unhesitatingly; see now
also his his etymological dictionary (Beekes 2010)), Ancient Greek οὐ, too, was
the result of this process. In Beekes’ words (1995: 144, 122): ‘Greek ou from ne
h2oiu ‘not ever’ from which ne disappeared, as ne in spoken French ne pas.’32
For some time both οὐδέν and δέν were in use, but eventually οὐδέν became a
proclitic particle, which in turn led to the weakening and disappearence of οὐ,
so that δέν/δε(ν) became the only form. See Jannaris (1897: §§ 1796–1800). The
first occurrence of δέν is to be found in a 6th c. documentary papyrus, and is
thus much earlier than is usually assumed. Cp. Gignac (1976: 186): ‘This Mod-
ern Greek form [i.e. δέν] is anticipated in ὅμος δὲν ἑ ἁμαρτίε ὑμο͂ν ἐσίν (for ὅμως
δὲν αἱ ἁμαρτίαι ὑμῶν εἰσίν POxy. 1874.13).’ I should add that the replacement of οὐ
by δέν involved also οὐ in questions: Δέν σου τό ’λέγα ’γώ; ‘Did I not tell you so?’
(Jannaris 1897: §2049), with positive bias. In answers, however, Modern Greek
uses όχι, which goes back to Ancient Greek οὐχί (or οὔχι), an emphatic variant of
οὐ (Jannaris §2061).33 There was, then, a functional split of the Ancient Greek
proposition negative οὐ: while the function as proposition negative was taken
over by δέν, οὐ itself acquired, in a more emphatic form, a new function as an
answering particle.34

7 Semantic and Pragmatic Features of ‘no’ and ‘not’ in Answers

As for the semantics and pragmatics of όχι and the other answering parti-
cles meaning ‘No’, it is sometimes claimed that these particles, just like those
expressing ‘not’, are substitutes for the preceding sentence, cp. e.g. Sweet (1892:

Jespersen’s Cycle in French see further Rowlett (1998: ch. 6 Jespersen’s Generalization),
and in general Zeijlstra (2004: ch. 3.2.3 The Jespersen Cycle).
32 This must be very old for already in Mycenaean there are no traces of an original *ne before
ou, the negative being just ou: o-u-di-do-si = ou-didonsi ‘they do not deliver’. For h2oiu cp.
Greek αἰών, αἰέν (< *aiw-), Lat. aevum, Du. eeuw, Eng. (n)ever, Ge. ewig.—A survey of I-E
languages where Jespersen’s Cycle is or was active is given by Bernini (1987: 64), who, does
not, however, refer to Jespersen to explain this phenomenon. For parallels outside Indo-
European see Croft (1991).
33 For the etymology see Chantraine s.v. οὐ. Οὐχί is in classical Greek by and large used as οὐ,
but 1) *οὐκοῦν … οὐχί does not occur; 2) it is not used in answers, with the exception of the
pseudo-Platonic dialogue Hipparchus (date uncertain; 4th c. BC?), at 226b (Οὐκ ἄρα οἴεται
… :: Οὐχί. ‘So he does not think that … :: (He does) not’. Or ‘No’? see below); 3) unlike οὐ it
is used in tag questions with positive bias, e.g. Plato Meno 87e πάντα … τὰ ἀγαθὰ ὠφέλιμα.
οὐχί; :: Ναί. (‘All good things are profitable, [are they] not? :: Yes.’). See further the Appendix.
34 But just like οὐ in Ancient Greek, όχι is also used as a contrastive negative: όχι ο Κώστας, αλλά
ο Γιάννης βγήκε πρώτος ‘Not Costas but Gianni went out first’, θα πας είτε το θέλεις είτε όχι
‘you’ll go whether you like it or not.’ It is also used in negative complements: νομίζω/ελπίω
350 chapter 17

§ 368): ‘no in the above example [‘The answer to the question is he here? can be
either the affirmative yes or the negative no’] is equivalent to he is not here; it
is, therefore, at the same time the absolute form corresponding to the conjoint
not.’ I do not think this is correct. If it were, one would expect that ‘No’-words
could be modified by adverbs like certainly, which they cannot (see above). In
fact, they cannot be modified in any way, unless in mention (a qualified ‘no’,
etc.). They are rather like interjections. I believe that No basically expresses the
idea that the speaker rejects the truth of the proposition that underlies the pre-
ceding question,35 while not denying it. Often, the two are of course combined:
Is John here? :: No, he is not (here)/he isn’t, where the underlying proposition
‘John is here’ is first rejected by No, and then denied by not. Straightforward
rejection being rather impolite, there is a tendency for such answers to be more
complete than is strictly needed: rather No … not, than just No.

8 Summary

Let me sum up. Contrary to what is suggested by grammars and dictionaries,

οὐκοῦν … οὐ questions do not expect the answer ‘No’.36 From the actual usage
of οὐκοῦν … οὐ questions it is clear that these, too, expect an affirmative answer,
just like οὐκοῦν questions without οὐ: they ask the addressee to confirm the cor-
rectness of the negative. This confirmation can be given in various ways, also
by repeating the negative οὐ. In that case the negative is not an answering parti-
cle like ‘No’, but an elliptic substitute for the entire proposition of the preceding
question. In later Greek οὐ made way for οὐδέν, which in turn led to the modern
negative δεν. This is only used as a proposition negative. The development of οὐ
> οὐδέν > δέν (δεν) was accompanied by a parallel development οὐ > οὐχί/οὔχι >
όχι, which is used predominantly as an answering particle, ‘No’.37 In a similar

πως όχι ‘I think/hope (that) not’. Likewise, όχι continues AGr οὐ in γιατί όχι; (‘Why not?’);
*γιατί δε(ν); does not occur. Sometimes Ancient Greek οὐ has been preserved unchanged,
as in ουδόλως ‘absolutely not’, and ούτε … ούτε ‘neither … nor.’
35 One might also say that the speaker refuses to accept the proposition, just as No expresses
refusal rather than rejection in cases like Come on, hurry up! :: No.
36 To be complete I should add that besides οὐκοῦν (…) οὐ the combination μῶν (contracted
from μὴ οὖν) οὐ, too, exists. In principle μῶν (‘it is not true, is it?’—negative bias) οὐ expects
a negative answer to a negative question: ‘It isn’t true, then, is it, that not X …?’. But here
the two particles cancel each other, and in actual practice this combination expects an
affirmative answer: ‘Surely, then, X …?’. E.g. Plato Leges 657c μῶν οὐχ οὕτως; :: Οὕτω μὲν οὖν
(‘Is not that so? :: Yes, that is so’).
37 As for affirmative answers, ναί, too, may not be simply an answering particle like ‘yes’. At
does ancient greek have a word for ‘no’? 351

way Latin non, itself from *ne + ōinom ‘not one thing’, developed into French
(ne …) pas, a proposition negative, on the one hand, and non, which is mainly
an answering particle, on the other.
Schematically the two developments can be presented as follows:

From Classical Greek to Modern Greek

(*ne h2oiu ‘not ever’, which through Jespersen’s Cycle >)

Classical Greek:

οὐ negative proposition particle: ‘not’, not ‘No’

Post-classical/Byzantine Greek:

οὐ → 1) οὐδὲν ἀντὶ τοῦ οὐ (Hesychius 5th–6th c.) which through Jespersen’s

Cycle > δέν ‘not’, a proposition negative; as early as 6th c. AD (a
papyrus from Oxyrhynchus; the Vita S. Marinae) > Modern Greek:
δε(ν))
→ 2) οὐχί/οὔχι > όχι negative answering particle: ‘No’; hypothesis: this
development occurred when δέν became more frequent as a re-
placer of οὐ as a proposition negative. (NB Just as οὐ in ancient
Greek, όχι is also used as a contrastive negative: όχι ο Κώστας, αλλά
ο Γιάννης βγήκε πρώτος ‘Not Costas but Gianni went out first’)

From Classical Latin to Modern French

(*ne + ōinom ‘not one thing’ >)

Classical Latin:

nōn negative proposition particle: ‘not’, not ‘No’

least this is suggested by its use in sentences like Aristot. Metaph. 1034a17: ὅσων οὖν τοιαύτη
ἡ ὕλη, οἷον οἱ λίθοι, ἀδύνατον ὡδὶ κινηθῆναι εἰ μὴ ὑπ᾽ ἄλλου, ὡδὶ μέντοι ναί· (‘It is impossible,
then, for any things whose matter is of this kind (e.g. stones) to be moved in this par-
ticular way except by something else; but in that particular way it is possible’). Here ναί
is obviously not an affirmative answering particle meaning ‘yes’. It is rather an emphatic
affirmative adverb that modifies a proposition and contrasts with the negative prefix of
ἀδύνατον: ‘It is not possible’ as opposed to ‘It is possible.’ Unlike English, that in such cases
puts contrastive emphasis on the verb of the second clause, Dutch, like Ancient Greek,
uses an emphatic particle, wel: ‘op deze manier niet, op die manier wel’.
352 chapter 17

Oldest French (Cantilène de sainte Eulalie, around 880):

non negative proposition particle

Old French:

1) > nen > ne > ne … pas negative proposition particle (La vie de saint
Alexis, end of 11th c.); modern spoken French: pas (through Jes-
persen’s Cycle); ne retained in expressions like je ne saurais le dire
non 2) non: negative proposition particle in certain fixed collocations
((Voillent o nun ‘qu’ils le veuillent ou non’, La vie de saint Alexis);
modern French qu’ils le veuillent ou non, je crois que non,38 proba-
bly also in non pas (x mais y)
3) non: answering particle (Nel feras? :: Non. ‘Tu ne le feras pas? ::
Non’, Le jeu d’Adam, 12th c.)

Appendix: οὐχί and οὔχι/ὄχι/όχι (Continuation of Footnote 33)

The etymology deriving όχι from οὐχί is less evident than might seem, for the
shift of the accent, and the change of ου to ο are both problematic, and have
been, and still are, the subject of discussion; for a summary of the discussion
see Joseph (2000). Joseph also mentions a third problem, viz. the semantic and
syntactic shift of οὐ ‘not’ to όχι ‘no’. I will here go into this matter in some detail.
(i) The accent shift may be due to influence from ναίχι, the emphatic vari-
ant of ναί ‘Yes’. Interestingly, the grammarian Herodian (2th c. AD) observes
(De Pros. cath. 3.1, 506, 2) that οὐχί has an illogical oxytone accent, for just like
ναίχι it ought to be barytone (ἐχρῆν δὲ καὶ τὸ οὐχί βαρύνεσθαι, i.e. to be not-
oxytone), because it is a lengthened form of oxytone οὔ (a form, one might add,
which, like ναί, is predominantly found in answers). Herodian’s observation
returns in the much later (9th–11th c.?) Commentaria in Dionysii Thracis Artem
Grammaticam, Scholia Marciana 432.5 Τὸ οὐχί παράλογον κατὰ τὸν τόνον, ἔδει γὰρ
βαρύνεσθαι, ὡς τὸ ναίχι, likewise Scholia Londinensia 563.16. The accent shift may
reflect the functional shift from proposition particle to answering particle dis-
cussed in the main text, a phenomenon that is also found in mergers like οὔκουν,
an emphatic variant of οὐ, from οὐκ οὖν, where the accent has shifted from
the particle to the negative, distinguishing it from οὐκοῦν; ‘in οὔκουν the pre-

38 Cp. the parallel constructions in Modern Greek mentioned in n. 34.

does ancient greek have a word for ‘no’? 353

dominant element is οὐκ, in οὐκοῦν, οὖν’ (Denniston 1954: 430). Similar parallel
functional and accents shifts are shown by connective ἀλλά ‘but’, from ἄλλα ‘in
other respects’, and adverbial σφόδρα ‘exceedingly, very much’, from the adjec-
tive σφοδρός ‘vehement.’ Be that as it may, the paroxytone form οὔχι is extremely
rare. Checking the TLG for οὔχι brought to light just three instances; the gram-
marians’ advice was apparently not heeded universally. The first instance is to
be found in Joannes Chrysostomus (4th c. AD), De incomprehensibili dei natura
(= Contra Anomoeos, homiliae 1–5). 2.288: Ὁρᾷς ὅτι οὔχι πρὸς τὴν δημιουργίαν τῶν
ἐν τῇ γῇ μόνον, ἀλλὰ καὶ πρὸς … This seems to be the reading of the MSS, but
there is a variant οὐ, see the edition in Sources chrétiennes tome 28bis, Paris
1970. Elsewhere in Joannes the form οὐχί is found. I should add that the accent,
like indeed all accents, in all probability was added much later (9th–10th c.),
as part of the Byzantine μεταχαρακτηρισμός, the transliteration from uncials to
minuscules. There are only two other instances of οὔχι, both in the philosopher
Isaac Comnenus, (11th c. AD), at De providentia et fato 5.13 and 15. Again, else-
where in this author οὐχί is found. The rarity of οὔχι can perhaps be explained
as being due to the growing importance of the form ὄχι in early and medieval
Byzantine Greek, which is partly visible in our texts, see next point. Or are we
perhaps dealing with errors? In that case there is no intermediate form οὔχι
between οὐχί and ὄχι.
(ii) As for the single ο- in ὄχι/όχι, Joseph rightly points out (Joseph 2000: 208)
that ‘from Ancient Greek ου, normally a Modern Greek [u] develops’. He sug-
gests influence from Turkish yo ‘no’ and yok ‘there is not, no’. If so, this can only
have been as a reinforcement of a change that had already occurred long before
the Turkish domination of the Greek world, for ὄχι is found e.g. in the epic of
Digenis Akritis (originally from the 12th c. AD), the romance of Libistrus and
Rhodamne (13th–14th c.) and the Polemos tis Troados (14th c.). (See also below.)
Moreover, the change from ου- to ο is found in classical Greek. For examples see
Threatte (1980: 351–352). To be sure, Threatte observes (p. 352) that ‘[e]xamples
of Ο for ου are rare by 355–350, and only two … are later than 345 BC.’ But these
two are found, interestingly, in the negatives ὀκ (οὐκ) and ὀδέ (οὐδέ). Both forms
occur also in Greek papyri from the Roman period, see Gignac (1976: 212), ὀκ
in a papyrus from the 2nd/3d c., ὀδέ in papyri of AD 112 and the 2nd/3d c. The
change from ου to ο is also found in οὖν and several other words, see Gignac
ibidem.
The first occurrence of ὀχί in a literary text is found in the famous inscrip-
tion of Diogenes of Oenoanda (first half of 2nd c. AD), in fragment 13 II.10 of
the ed. by M.F. Smith (Smith 1993): τὴν οὖν ἀπόφασιν ὁρῶμεν αὐτοῦ ταπεινήν, ἀλλ᾽
ὀχὶ αὐτόν (‘So it is its image which we see low, not the sun itself’, transl. Smith).
Smith observes, however, that this may well be an error, since there are in this
354 chapter 17

particular part of the inscription more errors. Even so, it is a significant error,
for it points to uncertainty about the correctness of the spelling ΟΥΧΙ. (Which
is found elsewhere in the inscription, viz. in fr. 70.II.13.) Also, οχι is probably
present in a documentary papyrus of AD241, POxy. 19.2231, line 15 (ὀ⟨ὐ⟩χ̣ι ̣̀ τό̣ τ̣ε)̣ ,
where the editors have put οὐχί in the text, while adding in a note: ‘the scribe
would appear to have written οχι.’ All in all, the ο in ὀχί/ὄχι/όχι may be the result
of an undercurrent in post-classical Greek which resisted the pressure for ου
from the learned language, e.g. of the Second Sophistic.
(iii) The shift of οὐ ‘Not’ to όχι ‘No’. For this development, and its relationship
with the development οὐ > οὐδέν > δέν see the main body of the text. Here I add
a few particulars about όχι as an answering particle.
The first signs of οὐχί as an answering particle (‘No’), alongside its frequent
use as a constituent negative (as in the sentence from Joannes Chrysostomus
quoted at (i) and, in the form οὐχί, in countless other places), may be visible
already in post-classical Greek, e.g. in the Testamentum Abrahae (1st c. AD) 5,
l. 36, where we find the sequence οὐχί … οὐ, with emphatic οὐχί in front posi-
tion: εἶπεν πρὸς αὐτήν· Οὐχὶ, … οὐκ ἔστιν οὕτως ὡς σὺ λέγεις: (‘He said to her: “No,
it is not as you said”’). For όχι as an independent answering particle in Medieval
Greek see the entry in Kriaras’ lexicon (Kriaras 2010). Significantly, ὄχι and δέ(ν)
are also found combined, as in Damascenus Stoudites (16th c.), Or. 33.659 Σὲ
λέγει νὰ ὀμόσῃς ποτέ σου ψεύματα ἢ ἀλήθειαν; Ὄχι· δὲν λέγει τίποτε·. However,
if the date of the source is reliable the sequence Οὐχί … δέν may already be
present in the Vita Sanctae Marinae (3, 1.22), from the 6th c. AD: ἡ κόρη ἀπε-
κρίθη καὶ τὸν εἶπε· Οὐχὶ, αὐθέντη πατέρα μου, δὲν θέλω ἔλθη καθὼς λέγεις (‘The girl
answered and said to him: “No, master, my father, I do not want to go there in
the way you describe …”’). For δέν without οὐ(χί) see the main text § 6. For the
combination of answering particle and proposition negative see also the end
of section 7.

References

Asher, N. and Brian, R., ‘Negative Bias in Polar Questions’, in E. Maier, C. Bary and
J. Huitink (eds), Proceedings of ‘Sinn und Bedeutung’ 9 (Nijmegen 2005) 30–43.
Ball, M.J., The Celtic Languages (London/New York 1993).
Beekes, R.S.P., Comparative Indo-European Linguistics: An Introduction (Amsterdam/
Philadelphia 1995).
Beekes, R.S.P., Etymological Dictionary of Greek, with the assistance of L. van Beek (Lei-
den 2010).
Bernini. G., ‘Attempting the Reconstruction of Negation Patterns in PIE’, in A. Giaca-
does ancient greek have a word for ‘no’? 355

lone-Ramat, O. Carruba, G. Bernini (eds), Papers from the 7th International Confer-
ence on Historical Linguistics (Amsterdam 1987) 57–69.
Buridant, C., Grammaire nouvelle de l’ancien français (Paris 2000).
Chantraine, P., Dictionnaire étymologique de la langue grecque: histoire des mots (Paris
1968–1980).
Cowgill, W., ‘Greek ou and Armenian ǒc’, Language 36 (1960) 347–350.
Croft, W., ‘The Evolution of Negation’, Journal of Linguistics 27 (1991) 1–27
Denniston, J.D., The Greek Particles, rev. by K.J. Dover (Oxford 1954).
Dik, S.C., The Theory of Functional Grammar, part 2 (Berlin/New York 1997).
Giannakidou, A., ‘Weak and Strong Polarity: The Evidence from Greek’, in A. Alexiadou,
Geoffrey Horrocks and Melita Stavrou (eds), Studies in Greek Syntax (Dordrecht
1999) 113–135.
Gignac, F.T., A Grammar of the Greek Papyri of the Roman and Byzantine Periods, vol. II:
Morphology (Milano 1976).
Holton, D., Mackridge, P., Philippaki-Warburton, I., Greek: A Comprehensive Grammar
of the Modern Language (London 1997).
Horn, L.R., A Natural History of Negation (Chicago/London 1989).
Hudson, R.A., ‘The Meaning of Questions’, Language 51 (1975) 1–32.
Jannaris, A.N., An Historical Grammar Chiefly of the Attic Dialect (London 1897; repr.
Hildesheim 1968).
Jespersen, O., Negation in English and Other Languages (København 1917).
Joseph, B., ‘On the Development of Modern Greek όχι “No!”’, in: C. Schaner-Wolles,
J. Rennison, F. Neubarth (eds), Naturally! Linguistic Studies in Honour of Wolfgang
Ulrich Dressler presented on the occasion of his 60th birthday (Torino 2000) 207–214.
Kriaras, E., Λεξικό της μεσαιωνικής ελληνικής δημόδους γραμματείας (Thessaloniki 1969–).
Kühner, R., Stegmann, K., Ausführliche Grammatik der lateinischen Sprache, II. Teil: Sat-
zlehre. 2 Bände (Hannover 1912–1914; repr. 1955 and 1962, with corr.).
Lyons, J., Semantics (Cambridge 1977).
Moore, R.W., Comparative Greek and Latin Syntax (London 1934; repr. Bristol 1999).
Orlandini, A., Négation et argumentation en latin (Louvain-Paris 2001).
Quirk, R., Greenbaum, S., Leech, G., Svartvik, J., A Comprehensive Grammar of the
English Language (London/New York 1985).
Rijksbaron, A., ‘Negatives in Questions (and Answers): The Case of οὐκοῦν (…) οὐ. Or:
Does Ancient Greek Have a Word for “No”?’ in K. Loudová and M. Záková (eds), Early
European Languages in the Eyes of Modern Linguistics (Brno 2009) 261–271.
Romero, M. and Han, Ch.-H., ‘On Negative “Yes/No” Questions’, Linguistics and Philos-
ophy 27 (2004) 609–658.
Romero, M., ‘Two Approaches to Biased Yes/No Questions’, in J. Alderete, Ch.-H. Han
and A. Kochetov (eds), Proceedings of the 24th West Coast Conference on Formal Lin-
guistics (Somerville, MA 2005) 352–360.
356 chapter 17

Rowlett, P., Sentential Negation in French (New York/Oxford 1998).

Smith, M.F., The Epicurean Inscription/Diogenes of Oinoanda, ed. with intr., tr. and notes
by M.F. Smith (Napoli 1993).
Smyth, H.W., Greek Grammar, rev. by G.M. Messing (Cambridge MA 1956).
Sweet, H., New English Grammar: Logical and Historical, part I (Oxford 1892).
Thesleff, H., Yes and No in Plautus and Terence (Helsingfors-Leipzig 1960).
Threatte, L., The Grammar of Attic Inscriptions, vol. I: Phonology (Berlin 1980).
Wackernagel, J., Vorlesungun über Syntax, Zweite Reihe (Basel 1928).
Zeijlstra, H.H., Sentential Negation and Negative Concord (Utrecht 2004).
 chapter 18

The Treatment of the Greek Middle Voice by the

Ancient Grammarians

Abstract

This chapter presents a survey of the treatment of the Greek middle voice by both
Greek and Latin grammarians. It shows that they on the whole treated the middle voice
as a ‘rest’ category into which all forms that in one way or another do not conform
to the patterns of active forms could be put. The mechanical approach of the gram-
marians is contrasted with philosophical discussions of the related notions of ‘agency’
and ‘patiency’ in Aristotle, Simplicius and Plotinus, which make a number of valuable
observations.

1 Dionysius Thrax and His Scholiasts

The point of departure* for my observations on the treatment of the middle

by the ancient grammarians will be the famous, and deceivingly simple, sen-
tence in Dionysius Thrax:1 διαθέσεις εἰσὶ τρεῖς, ἐνέργεια, πάθος, μεσότης· ἐνέργεια
μὲν οἷον τύπτω, πάθος δὲ οἷον τύπτομαι, μεσότης δὲ ἡ ποτὲ μὲν ἐνέργειαν ποτὲ δὲ
πάθος παριστᾶσα, οἷον πέπηγα διέφθορα ἐποιησάμην ἐγραψάμην (1, 1, 48–49).2 | 428
Dionysius’ formulation of the ‘middle’ as the διάθεσις ‘which sometimes
signifies activity and sometimes affectedness’ is ambiguous and has been a
constant source of confusion, for this definition makes not clear whether the
‘middle’ diathesis consists of verbs whose morphology is not in accordance

* I am indebted to Prof. C.J. Ruijgh and Mr. H. Mulder for their criticism and suggestions, and
to the Netherlands Organization for the Advancement of Pure Research (Z.W.O.), for subsi-
dizing part of this research.
1 My taking this passage as a starting-point does not imply that I view the passage as the oldest
testimony of the ancient grammatical tradition on “voice”. The date as well as the authen-
ticity of the little treatise which has been transmitted under the name of Dionysius Thrax
are—again—the subject of much controversy. For a detailed discussion I refer to the articles
by di Benedetto (1958; 1959) and to the summary of his conclusions in Pinborg (1975: 104ff.).
2 References to the Greek grammarians are made according to the page and line in the edition
by Uhlig-Schneider-Hilgard, Grammatici Graeci; so 1, 1, 48–49 is the ‘partis primae volumen
primum’, pages 48–49.

© koninklijke brill nv, leiden, 2019 | doi:10.1163/9789004386129_020

 358 chapter 18

with their meaning—i.e. verbs that have active forms but passive meaning
and vice versa—or verbs that individually have forms that may have active
as well as passive meaning. In the first case the term ‘middle’ relates to the
anomalous behaviour of certain verbal endings, in the second, to that of cer-
tain verbs. To all appearance Dionysius takes μεσότης in the first sense: πέπηγα
and διέφθορα are perfects with active endings—nowadays sometimes called
secondary perfects—with passive meaning, from πήγνυμαι and διαφθείρομαι,
respectively; by the same token ἐποιησάμην and ἐγραψάμην must be considered
as having active meaning. Observe, in this connection, that the intransitive use
of πέπηγα and διέφθορα (and also of πήγνυμαι and διαφθείρομαι, for that matter)
is not mentioned. It is, in fact, one of the striking features of the Greek gram-
marians’ treatment of voice that the important group of intransitive ‘middle’
verbs that correspond to active so-called causative verbs (e.g. διαφθείρω : δια-
φθείρομαι, ἵστημι : ἵσταμαι) is not recognized as a separate group. The same view
concerning πέπηγα and ἐποιησάμην is found in Heliodorus’ scholion (1, 3, 401,
29ff.): πέπηγα is a φωνή ἐνεργητική yet expresses πάθος (‘ἴσον γὰρ ἐστι τῷ πέπη-
γμαι’), whereas ἐποιησάμην is a φωνή παθητική yet expresses ἐνέργεια (‘ἀντὶ τοῦ
429 ἐποίησα’).3 However, Heliodorus | also has some alternative approaches, one of
the many testimonies of the considerable embarrassment felt by the scholiasts
vis-à-vis this matter. At 1, 3, 401, 22ff. he informs us that we are dealing with μέση
διάθεσις ‘when the same form relates to (χωρῇ εἰς) action as well affectedness,
as is the case with βιάζομαι’, since we can say both βιάζομαί σε and βιάζομαι ὑπὸ
σοῦ. This, then, is a verb-oriented analysis: a form like βιάζομαι ‘combines’, as it
were, active and passive meaning. But then he presents an, again, completely
different third approach, according to which διάθεσις is μέση, when the same
ending either expresses πάθος or ἐνέργεια. This is the case with -μην; for verbs
ending in this form on the one hand express action, as ἐγραψάμην and ἐφάμην
and on the other affectedness, as ἐτριψάμην and ἠλειψάμην. The latter two are
explicitly said to have the same meaning as ‘real’ passive ἐτρίφθην and ἠλείφθην.
Here, then, we have a second ending-oriented analysis: forms in -μην either have
active meaning or they have passive meaning. -μην combines, again, active and
passive meaning, but in different verbs. The most striking feature of these analy-
ses is that the middle forms are not seen as having a meaning of their own, but
as mere doublets of ‘real’ active and passive forms. This feature is almost univer-
sally present, as we will see further on. Notice that the opinions of Heliodorus

3 The omission of διέφθορα may not be due to chance: this form is particularly ill suited to illus-
trate ‘middle’ in the ‘Dionysian’ sense, since in actual practice it also has active meaning, at
least in 5th century Greek. In Homeric Greek both πέπηγα and διέφθορα have, indeed, only
passive (or, rather, intransitive) meaning.
the treatment of the greek middle voice 359

in this passage are only remotely related to Dionysius’ definition. They are, per-
haps, more directly influenced by Apollonius (see below), especially by his lost
Ῥηματικόν.
So far then the following opinions on ‘middle’ were seen to be held:
1) the ‘middle’ diathesis comprises verbs that either have active endings but
passive meaning (πέπηγα) or passive endings but active meaning (ἐποιη-
σάμην); | 430
2) it comprises verbs with passive endings only, that may have active as well
as passive meaning (βιάζομαι);
3) it comprises verbs with passive endings only, that partly have active mean-
ing (ἐγραψάμην) and partly passive meaning (ἐτριψάμην).
Notice, finally, that in this scholion three diatheseis only are recognized and
that the middle verbs with active meaning are all transitive verbs. ‘Activity’
would seem to be equated with ‘transitivity’, cp. note 6. The other scholia are
partly similar to those of Heliodorus and partly divergent. One noteworthy
point of divergence is that these scholia often have more than three diathe-
seis, the others being οὐδέτερον and ἐμπεριεκτικόν. Μεσότης, too, is sometimes
viewed, again, in a different way. Thus, the scholion on p. 1, 3, 245, 32 defines
‘middle’ as the διάθεσις which signifies both the other διαθέσεις, which means,
as appears from the context, that it follows a βιάζομαι-approach, but it then
goes on rather surprisingly, by giving ἐγραψάμην as an example of such a verb.4
It is also, at least if the relevant text can be trusted, the only ancient source
for the indirect-reflexive meaning of the middle, which in modern treatments
plays such an important role. At 1, 3, 246, 5 we read: τὸ … ἐποιησάμην δηλοῖ, ὅτι
ἐμαυτῷ ἐποίησά τι which is given as an illustration of the familiar μέση δὲ ἡ πῇ
μὲν ἐνέργειαν πῇ δὲ πάθος δηλοῦσα. Since this explanation of ἐποιησάμην would
be quite unique as an example of ‘active middle’ and | since ἐποιήθη in the next 431
line is presented as an instance of ‘passive middle’, which is, again, unparalleled,
editors are possibly right in suspecting a textual corruption.5 This scholiast is,

4 This scholion rightly stresses that the syntax makes clear whether we are dealing with active
meaning or with passive meaning. The latter is brought about by the addition of ὑπὸ σοῦ. Since
this would result in the unacceptable ἐγραψάμην ὑπὸ σοῦ, Hilgard corrects into ⟨γράφομαι ὑπὸ
σοῦ⟩, but in view of the strange choice of γράφομαι in the first place, I wonder whether Hil-
gard is not ‘hypercorrecting’ this scholion. Alternatively, ἐγραψάμην may have been viewed as
a ‘real’ passive, comparable to ἐγράφην, just as ἐτριψάμην was considered equivalent to ἐτρί-
φθην.
5 Unless the scholiast has mixed up everything completely; cp. the preceding note. Observe that
Lucian, too, appears to be aware of the indirect reflexive value of the middle. In Soloecista
12, while discussing the difference between καταδουλοῦν and καταδουλοῦσθαι, he notes that
καταδουλοῦν is ἑτέρῳ and καταδουλοῦσθαι, ἑαυτῷ. (This interpretation hinges, however, on the
 360 chapter 18

finally, one of the many ancient grammarians who recognise a fourth διάθε-
σις viz. the οὐδετέρα (‘neutrum’), defined as ‘which does not signify action nor
affectedness’; the verbs given as examples are ζῶ πλουτῶ δύναμαι βούλομαι.
This category has a clear parallel in the Stoic οὐδέτερον given in Diog. Laërt.
7, 64, although the definition of οὐδέτερον given there is syntactic rather than
semantic: οὐδέτερα δ᾽ ἐστί (sc. κατηγορήματα ‘predicates’) τὰ μηδετέρως ἔχοντα,
viz. neither ὀρθά (‘transitive’) nor ὕπτια (‘reversed’, i. e. passive), examples
being the semantically active but syntactically intransitive φρονεῖ and περιπα-
τεῖ. Observe that the terms ὀρθός and ὕπτιος are present, too, in the scholia
on Dionysius, e.g. 1, 3, 247, 10, where we are told that ‘the’ philosophers—i.e.
apparently the Stoics—used the term ὀρθή for the ἐνεργητική and ὑπτία for the
παθητική.6
There are two further scholia in the Scholia Vaticana that deserve our atten-
432 tion. First, the scholion at 1, 3, 246, 7ff. | assures us that there are five diathe-
seis, if one is accurate (κατὰ τὴν ἀκρίβειαν). Besides active and passive it has:
οὐδέτερον—examples: ζῶ, πλουτῶ, and also πυρέσσω, not, however, δύναμαι
and βούλομαι—, μέση—examples: πέπηγα, ἐγραψάμην, i.e. the ‘middle’ is ap-
proached as it is in Dionysius—, and, finally, ἐμπεριεκτική, which contains verbs
that show both diatheseis, the examples being ‘βιάζομαι ὑπὸ σοῦ’ and ‘πορεύομαι
διὰ σέ’. The latter group apparently consists of verbs with passive endings that
may have either passive or active meaning, although the examples given exem-
plify only the passive use (πορεύομαι is, of course, a doubtful passive). Not much
ἀκρίβεια, here.
The second scholion (by Stephanus?) is more orthodox and recognizes only
three diatheseis, on the slightly less orthodox analogy, however, of the number
of the nominal genders.7 The analogy is carried on in that ‘active’ is equated
with ‘masculine’, ‘passive’ with ‘feminine’ and ‘middle’ with ἢ ἑκάτερον ἢ οὐδέ-
τερον. The latter is explained ingeniously. Just as a neuter noun like τέκος may
refer to a boy or to a girl, so βιάζομαι may have ‘active’ and ‘passive’ meaning,

presence of τό, added by Hemsterhuys. Macleod, in his Loeb-edition, accepts τό, while taking
καταδουλοῦσθαι as passive, which is impossible).
6 This suggests that the ‘active’ voice was considered to express not ‘activity’ as such, but tran-
sitive activity. There was, indeed, a strong tendency to view the active voice in this way; cp.
e.g. Charisius’ statement (1, p. 165, 3 K.): ‘ubi est qui facit, necesse est ut sit qui patitur’. This
view was strongly opposed by other philosophical schools than the Stoa; see below p. [441].
Note that Chrysippus wrote a treatise περὶ ὀρθῶν καὶ ὑπτίων. (Diog. Laert. 7, 192). Frede (1978)
has a full discussion of the question whether there is such a thing as ‘Stoic grammar’.
7 Less orthodox, that is, in Greek grammar. The term ‘genus verbi’ is, of course, quite common
in Latin grammar.
the treatment of the greek middle voice 361

and just as βέλος is not only morphologically, but also semantically neuter, so
πλουτῶ, it is implied, does neither express ἐνέργεια nor πάθος. This scholion,
then, does know of an ούδέτερον-category but ranges it under ‘middle’. All in all,
the additions provided by these scholia to the information of Heliodorus are
perplexingly variegated. Their main feature is the introduction of a οὐδετέρα
διάθεσις for verbs that have neither active nor passive meaning. This διάθεσις is
perhaps set up as a direct counterpart to ‘middle’ verbs like βιάζομαι, that have
both active and passive meaning, also called ἑκάτερον by one scholion. | 433
All things considered, the discussion of the ‘middle’ voice in Dionysius and
his followers can hardly be called satisfactory. There is a constant hesitation
concerning verbs that ‘behave improperly’, i.e. whose meaning is not what it
may expected to be on the basis of their morphology. The primacy of the ‘nor-
mal’ verbs that have either active morphology and active, or, rather, transitive,
meaning, or passive morphology and passive meaning, apparently was so dom-
inating that the verbs whose morphology and meaning are at variance were a
permanent source of uneasiness. Sometimes, to be sure, the analysis as such is
fairly sensible, e.g in the case of βιάζομαι. More often, however, a tendency pre-
vails to view the deviant verbs merely as doublets of ‘normal’ active and passive
forms (cp. ἐγραψάμην and ἐτριψάμην), with the—possible—exception of the
indirect-reflexive analysis of ἐποιησάμην. Particularly unfortunate, I think, was
the introduction of non-productive forms like διέφθορα and πέπηγα as typical
examples of ‘middle’. In some cases this led to a very narrow conception of ‘mid-
dle’, e.g. in Choeroboscus, see below. The ultimate problem with the category
of ‘middle’ would seem to be that it is too wide: διέφθορα, ἐγραψάμην, βιάζομαι
and ἐτριψάμην can only be brought together under one heading at the cost of a
considerable loss of clarity.

2 Apollonius Dyscolus

In a passage on certain anomalies in the verbal system, Apollonius also dis-

cusses the ‘middle’. At 2, 2, 296, 2 he has the following general remark τὰ …
καλούμενα μέσα σχήματα συνέμπτωσιν ἀνεδέξατο ἐνεργητικῆς καὶ παθητικῆς δια-
θέσεως. The middle forms, then, ‘received a falling together of the active and
the passive diathesis’. This is primarily to be taken in the sense of opinion (3)
on p. [430], i.e. passive endings | may have either active or passive meaning. 434
The examples given are ἐλουσάμην, ἐποιησάμην and ἐτριψάμην, of which ἐλουσά-
μην and ἐτριψάμην have passive, and ἐποιησάμην active meaning. At least, this
is implied; Apollonius does not say expressis verbis that ἐλουσάμην has passive
meaning. Again, then, it seems that the middle forms are not given a mean-
 362 chapter 18

ing of their own, ἐλουσάμην is simply a by-form of ἐλούσθην, etc.8 Some lines
further on Apollonius also illustrates ‘middle’ with some Homeric peculiari-
ties, viz. the perfect participles κεκοπώς and πεπληγώς and impf. ὁρᾶτο. The first
two are so-called secondary perfects (cp. πέπηγα and διέφθορα in Dionysius, p.
[428] above); they should, therefore, have passive meaning, yet are construed
with an object. The latter also holds for ὁρᾶτο. Here, then, we find a βιάζομαι-
like approach (cp. nr. (2) on p. [430]): the syntax decides whether such forms
have passive or active meaning.9 The treatment of κεκοπώς and πεπληγώς pre-
supposes, I think, that Apollonius, too, like Dionysius, considered forms like
πέπηγα, i.e. forms with active endings but passive meaning, as ‘middle’. So we
find with him the same classificatory principles as with Dionysius and his scho-
liasts.
In the remainder of the περὶ συντάξεως ‘middle’ is not discussed separately
435 again. There are only two more or less | marginal remarks, first at 2, 2, 395, 3,
where he raises the question whether all verbs have all three diatheseis, and
where there is a strong suggestion that Apollonius saw the position of μεσό-
της as similar to that of οὐδέτερον in the nominal system; and secondly at 2, 2,
398, 5ff., where βιάζομαι, μάχομαι, χρῶμαι are called middle, because they have
passive forms and active meaning. Apollonius does not mention that βιάζομαι
may also have passive meaning. In the same passage he also gives πέταμαι, ἄγα-
μαι, δύναμαι and ἔραμαι as ‘non-passive’ passives, without indicating, however,
whether they are to be considered ‘middle’. Probably they are, for they appear,
more or less as an afterthought, at the end of the passage on the ‘middles’ βιά-
ζομαι, μάχομαι and χρῶμαι. As for possible other diatheseis, Apollonius calls
γηρῶ, φθίνω, κοπιῶ etc. verbs expressing αὐτοπάθεια (‘self-affectedness’), the lat-
ter being a subcategory of παθητική. This is, therefore, no separate diathesis.10
Οὐδέτερον is not recognized by Apollonius; verbs like ζῶ, ὑπάρχω and also περι-
πατῶ, have with him a vaguely ‘active’ status. Nor does he know of ἐμπεριεκτικόν.

8 In a note, Uhlig translates ἐλουσάμην by ‘lavatus sum mea ipsius opera’, but this reads far
too much into the text. Nor do his references to Macrobius and Choeroboscus corroborate
his translation, for both explicitly say that middle forms like ἐλουσάμην have strictly pas-
sive meaning. Cp. Macrobius, on ἠλειψάμην, ἡσάμην, ἐλουσάμην: ‘nihil aliud significant nisi
πάθος: nam hoc est ἠλειψάμην quod ἠλείφθην’ (5, 627, 38 K.) and Choeroboscus (4, 2, 208,
6): τὸ … ἐλουσάμην πάθος παρίστησιν, ἀντί … τοῦ … ἐλούσθην.
9 For κεκοπώς and πεπληγώς cp. Chantraine (958: 397 and 426) for ὁρᾶτο Bechert (1964). In
reality, κεκοπώς and πεπληγώς never have passive meaning in Homer. Possibly one should
accept the variant readings κεκοπών and πεπληγών, being Aeolic perfect participles.
10 If the criterium for assigning a verb to ‘middle’ is the presence of active endings with pas-
sive meaning, then verbs like κοπιῶ, γηρῶ etc. could in principle have been called ‘middle’
as well, just as e.g. διέφθορα.
the treatment of the greek middle voice 363

All in all Apollonius’ discussion of ‘middle’ does not differ much from that
by Dionysius and his commentators. The most salient point with him, too, is
the non-assignment of a proper meaning to the ‘middle’ diathesis.

3 Choeroboscus

For the larger part Choeroboscus presents the by now familiar picture (4, 2, 9,
7 ff.). For verb forms like ἐγραψάμην and ἐλουσάμην, he supports an approach
which concentrates on the verbal endings: ἐγραψάμην has active meaning | 436
(ἀντὶ τοῦ ἔγραψα), ἐλουσάμην passive meaning (ἀντὶ τοῦ ἐλούσθην). A novelty
is that he does so for the secondary perfects as well; since his examples of
perfect middles are chosen according to the same principle τέτυπα is given
as an instance of a ‘middle’ perfect with active meaning (ἀντὶ τοῦ ἔτυψα),
τέτηκα as one of a ‘middle’ perfect with passive meaning (ἀντὶ τοῦ ἐτάκην).
τέτυπα is very strange indeed: if anything, it should have ‘passive’ meaning,
being a secondary perfect, like πέπηγα. But then this form is completely arti-
ficial; it is part of the verbal canons given by Theodosius, where τύπτω is the
paradigm verb, but it is not found in actual Greek; incidentally, neither is the
‘active’ perfect τέτυφα. Compare Veitch’s remark (1887: 654) ‘1. perf. τέτυφα, 2.
perf. τέτυπα … we have never seen except in grammars’.11 The ‘double mean-
ing’ approach of verbs like βιάζομαι is also found in Choeroboscus (cp. 4, 2,
99, 28ff.), as it is in Theodosius (4, 1, 49, 25), although there is one modi-
fication, since βιάζομαι in constructions like βιάζομαί σε is sometimes called
μέσον, not ἐνεργητικόν (e.g. 4, 2, 99, 32). Finally, Choeroboscus discusses the
question whether the ‘middle’ should be considered a diathesis at all, rather
than simply being classified as either active or passive, just as, in the nom-
inal system, there is a group of κοινά words, i.e. words that are either mas-
culine or feminine (e.g. ὁ and ἡ ἵππος). Choeroboscus denies the validity of
this analogy: the κοινά-nouns are not formally distinct from ‘real’ masculine | 437
and feminine nouns, whereas the ‘middle’ has some forms of its own (e.g.
τέτυπα, ἐτυψάμην, τύψομαι). Therefore it is correctly set up as a separate diathe-
sis.

11 According to Theodosius (4, 1, 49, 17ff., cp. also Choeroboscus 4, 2, 98, 30ff.) ‘the older
grammarians’ did not recognize a separate ‘middle’. On strictly morphological grounds
these—unidentified—grammarians reckoned the -α perfects, as well as the -ειν pluper-
fects, with passive meaning, among the active perfects and pluperfects, and the -μην
aorists, as well as the -σομαι futures, among the passive aorists and futures. On the same
morphological grounds present and imperfect middle were called ‘passive’ forms.
 364 chapter 18

As to other diatheseis, Choeroboscus tells us (4, 2, 101, 16 ff.) that verbs like ζῶ,
πλουτῶ and ὑπάρχω by some are called οὐδέτερα, because they denote neither
action nor affectedness, whereas there are also people who consider οὐδέτερον,
in turn, as being in reality μέσον. As to the latter issue, Choeroboscus points out
that this cannot be correct, since these verbs are not characterized morpholog-
ically as ‘middles’. Nor does he recognize the existence of a separate οὐδετέρα
diathesis, on the ground, again, that it has no ‘middle’ forms.12 Since these verbs
have active endings they should be classified as actives. This is, to all appear-
ances, the same analysis as that proposed by Apollonius.

4 The Latin Grammarians

4.1 Macrobius
Macrobius’ approach of the middle (5, 627, 5 ff., K.)13—called commune in
Latin, neutrum is never used—strongly resembles that of Heliodorus (cp. p.
[430]): middle verbs or forms are βιάζομαι (active and passive meaning), ἠλει-
438 ψάμην (passive meaning, = ἠλείφθην), ἐγραψάμην (active meaning, = | ἔγραψα),
πέπηγα (passive meaning, = πέπηγμαι). But he also gives πεπληγώς and κεκο-
πώς, which stem from Apollonius. Verbs like κήδομαι, μάχομαι, ἄγαμαι, i.e. both
verbs with a σα-aorist and verbs with a θη-aorist, are all viewed as similar to
ἐγραψάμην: they are called μέσα, but they have active meaning only, i.e. they are
construed with an object.14 This, too, may go back to Apollonius (cp. p. [434]).
Like some Greek scholiasts (cp. pp. [430–431]), he has a fourth diathesis, οὐδέτε-
ρον, also called ἀπολελυμένον.15 This category only consists of verbs with active
forms, but the verbs concerned may have ‘active’ (e.g. τρέχω, ἀριστῶ, περιπατῶ)
as well as ‘passive’ meaning (e.g. νοσῶ, ὀφθαλμιῶ). The crucial factor for being
considered ‘neutral’ or ‘absolute’ would seem to be the absence of an object.16

12 Another argument adduced by him against the setting up of an οὐδέτερον (4, 2, 102, 1ff.) is
of a purely logical or taxonomical nature. If there is such a thing as an οὐδέτερον, which
place should it be given? It cannot take the third place, for in that case the middle would
have the fourth position, which would mean that the οὐδέτερον would be part of the mid-
dle; and this is false. Nor can it have the fourth position, for then it would partake of three
other diatheseis, which is ἄτοπον, since οὐδέτερον (‘neither’) presupposes two other cate-
gories.
13 References are to volume and page of the Grammatici Latini, ed. H. Keil.
14 He tells us that the Latin counterparts of these verbs are called deponentia, a term which
has no Greek equivalent.
15 Curiously enough, this term is not found in the Greek grammarians.
16 Observe that the content of the category οὐδέτερον is not uniform. It varies from active
the treatment of the greek middle voice 365

Macrobius, then, would seem to summarize more or less the pre-Choero-

boscan ancient treatment of the middle: βιάζομαι, ἠλειψάμην, ἐγραψάμην, πέπη-
γα, belong to the ‘Dionysian’ tradition, κεκοπώς, πεπληγώς, as well as μάχο-
μαι, ἄγαμαι etc., on the other hand, to the ‘Apollonian’ tradition. The οὐδέτε-
ρον diathesis is another Dionysian element. The unfortunate τέτυπα, τέτυφα of
Choeroboscus are lacking. | 439

4.2 Charisius and Priscianus

Charisius and Priscianus both have an extensive discussion of the Latin voices;
occasionally Greek data are taken into account as well. (Cp. 1, 164ff. K and 2,
376ff. K, respectively). They both have the categories commune (= middle in the
sense of (3) on p. [430]), neutrum and deponens. Curiously enough, Charisius
mentions yet another voice-category, viz. that of the habitiva (= verbs express-
ing an action which ‘remains within itself’?), comprising verbs like cresco,
nascor, cado, labor (Char. 1, 164 Κ.). This category partly coincides with the cat-
egory of verbs expressing αὐτοπάθεια (‘self-affectedness’) of Apollonius (cp. p.
[434] above).

5 Conclusion: a Comparison with Philosophical Approaches

To conclude this overview of the treatment of the middle in the ancient gram-
marians it can be said, I think, that the verdict given on the discussion in
Dionysius and his scholiasts (cp. p. [433]) is valid for the other grammarians,
too: it is unsatisfactory, because too many different phenomena are brought
together under one single heading. Thereby, the middle is made into a waste-
paper basket for every verb form which in some way or other does not conform
to the pattern active forms—active meaning, passive forms—passive meaning.
More seriously, this led to the view that these incongruous forms were mere
doublets of the canonical active and passive forms and had, thus, no status
of their own. In late antiquity this approach culminated in the setting up of
completely fictive forms like τέτυπα. All this is the more remarkable because
firstly, the separate status of reflexive actions, if not the notion ‘reflexivity’

forms with non-active but also non-passive meaning (e.g. ζῶ, πλουτῶ, cp. p. [432]), to pas-
sive forms with non-active and non-passive meaning (δύναμαι, βούλομαι, cp. p. [432]) and
active forms with active meaning (ἀριστῶ, περιπατῶ, τρέχω) as well as passive meaning
(νοσῶ, ὀφθαλμιῶ). It will be recalled that the latter group was called αὐτοπάθεια by Apollo-
nius. The only common feature of these verbs is that they have no object. Notice, in this
connection, that the Stoic οὐδέτερον also contains object-less verbs: φρονεῖ and περιπατεῖ.
 366 chapter 18

as such, was not unknown to ancient grammarians. Apollonius, for instance,

explicitly mentions constructions like ἐμαυτὸν ἔπαισα (2, 1, 1, 70, calling the
440 pronoun | αὐτοπαθές, i.e. ‘self-affected’); secondly, a verbal category αὐτοπάθεια
(‘self-affectedness’) was recognized by Apollonius, at least, containing verbs
like γηρῶ, πάσχω, φθίνω (but, curiously enough, not διαφθείρομαι); and because,
finally, in a number of philosophical writings middle forms are analysed in a
way that would seem to provide a fruitful basis for further discussion. This holds
notably for the direct-reflexive use of the middle and for intransitive verbs
like διαφθείρομαι. In the Stoic passage already cited (Diog. Laert. 7, 64) we find,
alongside the ὀρθά, ὕπτια and οὐδέτερα, a fourth category, viz. that of the ἀντιπε-
πονθότα, the ‘anti-passives’,17 which is defined as follows: ἀντιπεπονθότα δέ ἐστιν,
ἐν τοῖς ὑπτίοις, ἃ ὕπτια ὄντα ἐνεργήματα [δέ] ἐστιν, οἷον κείρεται· ἐμπεριέχει γὰρ
ἑαυτὸν ὁ κειρόμενος.18 This, then, appears to be a clear instance of the reflexive
middle; notice especially the presence of ἑαυτόν.19
As to the intransitives, a number of valuable suggestions may be found in
Simplicius’ commentary on Aristotle’s Categories, the chapter on ποιεῖν καὶ
441 πάσχειν, in Aristotle | himself, notably in the Physics, and in Plotinus’Enneades.
The relevance of these authors for the present issue lies in their very thorough
discussion of the notions ‘agency’ and ‘patiency’ and of the notion ‘change’.20
Some noteworthy features of this often very subtle semantic discussion are:
– The question whether ποιεῖν is always directed towards some entity outside
the ποιῶν or may exist absolutely (the latter view was not endorsed by the
Stoics and the grammarians, cp. fn. 6); cp. e.g. Simpl. p. 296, 299, 300.21 On

17 This term is also used in modern linguistic studies, be it in a different way, namely in con-
nection with ergative languages. Compare Comrie (1981: 108ff.) and Dik (1981: 123ff.).
18 Steinthal’s discussion of κείρεται (1890: 1, 299–300) is rather confusing. He views κείρο-
μαι primarily as a ‘reflexive causativum’ and translates by ‘ich lasse mich scheren’, reading
ἐμπαρέχει instead of ἐμπεριέχει. On the other hand, he has the translation ‘ich scheere
mich’ as well, although he denies that κείρομαι is reflexive. Notice, incidentally that Long’s
app. crit. in the Oxford-edition wrongly attributes ἃ ὕπτια (the mss have ἀνύπτια) and the
seclusion of δέ to Hicks, the editor of the Loeb-edition. Steinthal already has these alter-
ations; it is not clear to whom they are due in the last resort.
19 A reflexive approach of κείρομαι/ἐκειράμην is also present, I think, in Phrynichus (cp.
Rutherford 1881: 368: καρῆναι καὶ ἐκάρην φασί, καὶ εἶναι τούτου πρὸς τὸ κείρασθαι διαφοράν, τὸ
μὲν γὰρ ἐπὶ προβάτων τιθέασι, κείρασθαι δὲ ἐπὶ ἀνθρώπων. Compare also Lucian, Soloecista,
6 where the use of καρῆναι instead of κείρασθαι is criticized.)
20 Aristotle’s opinions on this matter are treated extensively in the valuable monograph by
S. Waterlow (1982). Cp. also Steinthal (1890: 300 ff.).
21 References are to the pages of Kalbfleisch’s edition in the Commentaria in Aristotelem
graeca series. The term used for ‘absolute’ is ἀπόλυτος, (e.g. Simpl. p. 296, 29 and p. 313,
5 (ἀπόλυτος ἐνέργεια)), which was not, as will be recalled, found in the Greek grammari-
ans.
the treatment of the greek middle voice 367

p. 299 we find περιπατεῖν and τρέχειν as instances of absolute ποιεῖν (οὐ πρός
τι) and the general remark: οὐ πᾶσα δὲ ἐνέργεια πρὸς πάσχον. The latter type
of ‘acting’ is sometimes called πράττειν, defined as: ‘whose effects exist in
their occurring (ἐν τῷ γίγνεσθαι)’, examples being flute-playing and dancing
(Simpl. p. 318, 29), as opposed to ποιεῖν, ‘whose effects remain present after
the action’, e.g. carpenting and building (Simpl. p. 318, 27).
– The question whether there is such a thing as absolute affectedness, i.e.
πάθος or πεῖσις without any agent at all.22 Verbs discussed in this connec-
tion are θερμαίνεσθαι, ψύχεσθαι (cp. Simpl. p. 310), ἰοῦσθαι (‘rust’) and σήπε-
σθαι (Plot. Enn. 6, 1, 19 and 20; Simpl. p. 315. Also relevant is Arist. Phys.,
ch. III and VII). Many would deny that there is such a thing: either there
is an outside agent (or changer)—this | may be the influence of wind and 442
humidity, e.g. in the case of ‘rust’ and ‘rot’—, or there is something in the
natural constitution of a given entity which is responsible for the being
affected of that entity, as is the case with λευκαίνεσθαι ‘become white’, of a
swan. Plotinus would seem to favour an approach of πάσχειν which does
not hinge on the presence of an outside agent, whether animate or inani-
mate: μὴ γὰρ οὐδὲ ᾖ τῷ παρ᾽ ἑτέρου χαρακτηρίζειν τὸ πάσχειν· ἔστι γάρ τι καὶ
παρ᾽ ἑαυτοῦ καὶ οὐ πᾶν παρ᾽ ἑτέρου, παρ᾽ ἑαυτοῦ μὲν τὸ σήπεσθαι, παρ᾽ ἑτέρου
δὲ τὸ τύπτεσθαι. This view, however, is not held by Simplicius (cp. p. 316, 6 and
p. 325); he may be following Aristotle, who remarks: (Phys. 8, 2, 253a10 ff.) that
many seemingly ‘inherent movements’ (ἐνοῦσα κίνησις) are, in fact, deter-
mined by the environment (τὸ περιέχον). This holds also for a process like
‘waking up’ (ἐγείρεσθαι), because it is the digestion of, and the appetite for,
food which cause an animate being to wake up (cp. also Phys. 8, 6, 259b
10ff.).
None of these issues is discussed in the ancient grammarians, who had, I think,
a somewhat mechanical view of active, passive and, especially, middle, and
whose main concern, as regards the middle voice, was to dispose of a num-
ber of anomalies of the Greek verbal system, a not too strange phenomenon, of
course, since they were first of all interested in criticism and philological expla-
nation of the canonical classical texts. More astonishing is, perhaps, that mod-
ern grammars have not been influenced by the philosophical analyses, either,
although it would seem that a full treatment of active and passive could profit
considerably from the observations made by the authors mentioned above.23

22 Again, called ἀπόλυτος, e.g. Plot. Enn. 6, 1, 20, but also καθαρά (πεῖσις), e.g. Simpl. p. 310, 23.
Observe, in this connection, that the discussion makes use of the Stoic terms ὀρθός and
ὕπτιος, Simpl. p. 310, 14.
23 In general linguistic literature, though, a number of Aristotelian observations have been
368 chapter 18

References

Ancient Authors
Aristotle, Physics. Transl. by P.H. Wickstead and F.M. Cornford (London/Cambridge,
MA 1934).
Diogenes Laertius, Vitae philosophorum. Ed. H.S. Long (Oxford 1964).
Diogenes Laertius, Lives of eminent philosophers. With an English translation by R.D.
Hicks (London/Cambridge, MA 1925).
Grammatici graeci. Ed. G. Uhlig, R. Schneider, A. Hilgard. 4 vol. (Leipzig 1883–1901, repr.
Hildesheim 1965).
Grammatici latini. Ed. H. Keil et al. 8 vol. (Leipzig 1857–1870).
(Phrynichus), The new Phrynichus, being a revised text of the Ecloga of the grammarian
Phrynichus. Ed. by W.G. Rutherford (London 1881).
Plotinus, Opera. Edd. P. Henry et H.R. Schwyzer, 3 vol. (Oxford 1964–1982).
Simplicius, In Aristotelis Categorias commentarium. Ed. C. Kalbfleisch (Berlin 1907) (=
Comm. in Aristotelem graeca, vol. VIII).

Other
Bechert, J., Die Diathesen von ἰδεῖv and ὁρᾶv bei Homer, 2 Bände (München 1964).
Benedetto, V. di, ‘Dionisio Trace e la Techne a lui attributa’. ASNP 27 (1958) 169–210; 28
(1959) 87–118.
Chantraine, P., Grammaire homérique, Tome 1: Phonétique et morphologie (Paris 1958).
Comrie, B., Language Universals and Linguistic Typology (Oxford 1981).
Dik, S.C., Studies in Functional Grammar (London/New York 1981).
Dowty, D.R., Word Meaning and Montague Grammar (Dordrecht 1979).
Frede, M., ‘Principles of Stoic Grammar’, in J.M. Rist (ed.), The Stoics (Berkeley 1978)
27–75.
Garcia Gual, C., El sistema diatético en el verbo griego (Madrid 1970).
Pinborg, J., ‘Classical Antiquity: Greece’, in Th.A. Sebeok (ed.), Current Trends in Linguis-
tics, vol. 13 (The Hague 1975) 69–126.
Rijksbaron, A., The Syntax and Semantics of the Verb in Classical Greek (Amsterdam
1984).

at the base of classifications of verb meanings (cp. Dowty (1979: 51ff.), and the literature
quoted there); not, however, as far as I know the observations on ‘agency’ and ‘patiency’
in the Physics. Pinborg (1975: 107) draws attention to the potential value of the Neopla-
tonic commentators for semantic matters. A modern monograph on the Greek voices
is Garcia Gual (1970); they are also discussed in some detail in Rijksbaron (1984: 126–
157).
the treatment of the greek middle voice 369

Steinthal, H., Geschichte der Sprachwissenschaft bei den Griechen und Römern (Berlin
1890–1891, repr. Hildesheim 1971).
Veitch, W., Greek Verbs Irregular and Defective (Oxford 1887, repr. Hildesheim 1967).
Waterlow, S., Nature, Change and Agency in Aristotle’s Physics (Oxford 1982).
 chapter 19

A Question of Questions: peusis, erôtêsis and

[Longinus] Περὶ ὕψους 18.1

Abstract

Starting with the opening sentence of Ps.-Longinus’ Περὶ ὕψους 18, which speaks of
τὰς πεύσεις τε καὶ ἐρωτήσεις, this chapter sets out to determine the difference between
these two words for ‘question’. After a review of modern translations of the sentence
from ps.-Longinus and their shortcomings, it is shown through a careful reading of
passages from ancient grammarians that both are technical linguistic terms and that
ἐρώτησις is the equivalent of ‘yes/no-questions’, while πεῦσις corresponds to English
‘wh-questions’. Underlining Ps.-Longinus’ fondness for leçons par l’exemple, it is noted
that the two opening questions of Περὶ ὕψους 18 are themselves examples of these two
distinct types of question.

Chapter 18 of Pseudo-Longinus περὶ ὕψους, on rhetorical questions, opens with

the following two questions: Τί δ᾽ ἐκεῖνα φῶμεν, τὰς πεύσεις τε καὶ ἐρωτήσεις; ἆρα
οὐκ αὐταῖς ταῖς τῶν σχημάτων εἰδοποιίαις παρὰ πολὺ ἐμπρακτότερα καὶ σοβαρώτερα
συντείνει τα λεγόμενα;1 These questions are themselves generally taken as exam-
ples of rhetorical questions; for the author’s fondness of operating by la leçon
par l’exemple see e.g. Russell here.2 But what about the distinction between
πεύσεις and ἐρωτήσεις? Russell notes in his commentary ad loc.: ‘The technical
writers distinguish between πύσμα and ἐρώτημα. ἐρωτήματα can be answered
by yes or no, while πύσματα need longer explanations; this seems originally to
have been a distinction of Stoic logic, cf. Diog. Laert. 7.66. L treats the terms as
synonymous.’3 In his translation (see note 1) Russell has ‘inquiries and ques-
tions’.4 Other translators treat these words in the same vein: Rhys Roberts

1 ‘What are we to say of inquiries and questions? Should we not say that they increase the
realism and vigour of the writing by the actual form of the figure?’ (translation Russell 1965).
Unless indicated otherwise, the other translations are mine.
2 Russell (1964).
3 Russell does not tell us how he arrived at this view. ‘Synonymy’ is a concept that is perhaps
invoked too easily in literary studies. ‘It is by now almost a truism that absolute synonymy is
extremely rare—at least as a relation between lexemes …’ (Lyons 1995: 60–61).
4 Should these also be considered ‘synonymous’?

© koninklijke brill nv, leiden, 2019 | doi:10.1163/9789004386129_021

a question of questions 371

(21907): ‘questions and interrogations’, Prickard (31946): ‘Questions and Interro-

gations’,5 Grube (1957): ‘rhetorical questions and inquiries’, Fyfe-Russell (1995):
‘the figures of inquiry and interrogation’, Mazzucchi (1992): ‘le interrogazioni e
le domande’.6 As for the ‘technical writers’ simply mentioned by Russell, one of
these is quoted by Hoogland (1936) in his commentary, but for πύσμα he unfor-
tunately leaves out an important part (see below). Hoogland’s quotation runs
(= Alexander Rhetor, De figuris Spengel III 24ff.): Ἐρώτημά ἐστι, πρὸς ὃ ἀνάγκη
ἀποκρίνασθαι κατ᾽ ἀπόφασιν ἢ κατάφασιν οὕτω, ναὶ ἢ οὔ … πύσμα δέ ἐστι, πρὸς ὃ
διεξοδικῶς ἀπαντῆσαι δεῖ καὶ διὰ πλειόνων κτἑ. (For a translation see note 12.) This
text is apparently the source of Russell’s contention that πύσματα ‘need longer
explanations’, cp. διεξοδικῶς … καὶ διὰ πλειόνων.
There are, however, three dissenters: von Scheliha (1938), Lebègue (1939) and
Brandt (1966). Lebègue translates πεύσεις with ‘demandes’ (and ἐρωτήσεις with
‘interrogations’); but πεῦσις is not ‘demande’ (‘request’). The two German schol-
ars are outside the main stream. Von Scheliha’s translation runs: ‘bestimmte
und unbestimmte Fragen’, whereby she comes halfway the truth. Brandt, on
the other hand, translates πεύσεις τε και ἐρωτήσεις with ‘die Bestimmungs- und
Entscheidungsfragen’, apparently trying to avoid the impression that they are
synonymous, and with this, rather technical, translation he hits the nail on the
head. (Remarkably enough, he has no note with his translation.) For in the
German grammatical terminology, ‘Bestimmungsfragen’ and ‘Entscheidungs-
fragen’7 are identical with English wh-questions and yes-no questions, respec-
tively, and these are meant here. In fact, of Russell’s note only the first part
is correct. While ἐρωτήματα can indeed be answered by yes or no,8 πύσματα

5 Actually, ‘inquiries’ and ‘interrogations’ are rather strange, since they suggest that there is such
a thing as a ‘rhetorical inquiry’ or a ‘rhetorical interrogation’, which to my knowledge do not
exist.—Prickard adds in a note: ‘As all the examples are of Question and Answer it seems not
improbable that one of the two substantives has replaced the word ‘Answers’ in the original.’
The same idea occurs in other scholars.
6 Amazingly enough, Mazzucchi has not made use of Hoogland’s edition and commentary.
Hoogland (1936) has followed Prickard’s suggestion (see note 5), and translates ‘de volgende
figuren, die van vraag en antwoord.’ In his recent Dutch translation Op de Coul (2000) simply
ignores one of the two words, for he translates: ‘retorisch gebruikte vragen’.
7 ‘Bestimmungsfragen’ could be rendered as ‘specifying questions’. They are also, and in fact
more commonly, called ‘Ergänzungsfragen’ (lit. ‘filling-in questions’ or ‘supplementary ques-
tions’). ‘Entscheidungsfragen’ = ‘decision questions’.
8 There are, of course, other possibilities, e.g. adverbs like πάνυ, or (part of the) original ques-
tion may return in the answer, like γράφω, after the question γράφεις; a possibility mentioned
by A.D. De constr. 2.51 (= p. 161, line 12 Uhlig). See also De constr. 1.30 (= p. 28, line 1 Uhlig),
which deals with τὰ πευστικά, with Lallot’s note (1997): ‘A la suite des Stoïciens … A. distingue
soigneusement entre erṓtēsis, “question totale” … et peûsis, púsma “question partielle” …’
 372 chapter 19

do not just need longer explanations, they denote a wholly different type of
question, and correspond exactly to what in grammars of English are called
734 wh-questions, i.e. | questions introduced by who, where, why etc. D.L. 7.66, men-
tioned by Russell, gives only part of the truth. This passage, which in turn is a
quotation from Diocles Magnes (= Stoic. Veterum Fragm. II 186, Hülser fragm.
874)9 τὸ δὲ ‘ἆρά γ᾽ ἡμέρα ἐστίν;’ ἐρώτημα. πύσμα δέ ἐστι πρᾶγμα πρὸς ὃ συμβολι-
κῶς οὐκ ἔστιν ἀποκρίνεσθαι, ὡς ἐπὶ τοῦ ἐρωτήματος, Ναί, ἀλλὰ ⟨δεῖ⟩ (add. Arnim)
εἰπεῖν ‘οἰκεῖ ἐν τῷδε τῷ τόπῳ’.10 The presence of ἐν τῷδε τῷ τόπῳ makes us expect
that this sentence is an answer to a question with ποῦ, and this is borne out
by two other passages on πύσματα, viz. S.E. M. 8.70 (= SVF II 187, Hülser 876): …
πύσματα, ἅπερ λέγοντες πυνθανόμεθα, οἷον ‘ποῦ οἰκεῖ Δίων;’ and especially Ammo-
nius in Int. p. 200, 5ff. (Hülser 904): πυσματικὴν δὲ πρὸς ἣν τὸ μὲν ναί καὶ τὸ οὔ
χώραν οὐκ ἔχουσι, λέξεως δὲ ἄλλης, ἐνίοτε δὲ καὶ λόγου πλείονος δεῖ τῷ ἐρωτωμένῳ
πρὸς τὴν ἀπόκρισιν· διδαχθῆναι γάρ τι περί τινος βουλομένων ἐστὶν ἡ πυσματικὴ ἐρώ-
τησις, οἷον ‘πότε ἦλθες; τίς καλεῖ; ποῦ οἰκεῖς; πόθεν δῆλον ὅτι ἀθάνατος ἡ ψυχή; διὰ
τί ἡ Μαγνῆτις λίθος ἕλκει τὸν σίδηρον;’11 When we have a look now at the part on
πύσμα that was left out by Hoogland (see above), we note that Alexander, too,
uses a ‘wh-question’ to illustrate a πύσμα, since καὶ διὰ πλειόνων is followed by:
ὡς ἔχει τὸ τοιοῦτον, ‘ἐμὲ δέ, ὦ τριταγωνιστά, τὸ τίνος φρόνημα λαβόντα τούτοις συμ-

9 Hülser translates πύσμα with ‘Bestimmungsfrage’ and ἐρώτημα with ‘Entscheidungsfrage’.

10 ‘The expression “Is it day?” is an erôtêma. A pusma is an expression that cannot be
answered by a sign, as in the case of an erôtêma, Yes, but one should say “He lives in that
place.” ’ The meaning of συμβολικῶς is not entirely clear. I take it to mean ‘by means of a
sign representing a fuller answer’. Likewise Hülser: ‘mit einem Kürzel’. Thus, after a ques-
tion like γράφεις … the answer ναί represents (‘symbolizes’) the more complete γράφω; cf.
also note 8. Hicks (1958), however, renders πρὸς ὃ συμβολικῶς οὐκ ἔστιν ἀποκρίνεσθαι, ὡς
ἐπὶ τοῦ ἐρωτήματος, Ναί, as ‘… to which we cannot reply by signs, as you can nod Yes to
an interrogation …’ In view of the presence of ἀποκρίνεσθαι this seems unlikely, although
the sign can, indeed, be a nod, as is observed by Ammonius in Int., p. 199, 23; he speaks of
συμβολικῶς κατανεύσαντα ἢ ἀνανεύσαντα ‘approving or denying symbolically by a nod’. On
Ammonius see further below.
11 ‘(They say) that a pusmatic question is a question with respect to which there is no room
for yes and no; to give an answer the person questioned needs another word, and some-
times a longer expression. For the pusmatic question belongs to people who want to get
instruction about something, like “When did you arrive? Where do you live? On which
ground is it clear that the soul is immortal? Why does the magnete attract iron?”’ To which
Ammonius adds, quite appropriately: καὶ ὁ μὲν τὴν πυσματικὴν ἐρώτησιν ἐρωτώμενος κύριός
ἐστι τοῦ ἀποκρίνασθαι ὅπερ ἂν ἐθέλῃ τῶν προσαρμόττειν τῇ ἐρωτήσει δυναμένων ἐναριθμήτων
τυγχανόντων, ἢ ἀληθεύων δηλονότι ἢ καὶ ψευδόμενος … (‘Someone who is asked a pusmatic
question is free to answer whatever he likes from the countless things that can suit the
question, whether true or false …’ (as opposed to the erôtêsis, which can only be answered
affirmatively or negatively)).
a question of questions 373

βουλεύειν ἔδει;’12 Incidentally, the distinction between the two question types
is not a distinction ‘of Stoic logic’, but is in nuce already present in Aristotle,
although Aristotle does not use different names for them; they are just exam-
ples of ἐρωτήματα. See Top. 158a: Οὐ δοκεῖ δὲ πᾶν τὸ καθόλου διαλεκτικὴ πρότασις
εἶναι, οἷον ‘τί ἐστιν ἄνθρωπος;’ ἢ ‘ποσαχῶς λέγεται τἀγαθόν;’ ἔστι γὰρ πρότασις διαλε-
κτικὴ πρὸς ἣν ἔστιν ἀποκρίνασθαι ‘ναί᾽ ἢ’ ‘οὔ’• πρὸς δὲ τὰς εἰρημένας oὐκ ἔστιν.13 All
in all, then, it would seem that the distinction between the two question types
was firmly established in philosophy (in which it probably originates), as well
as among grammarians and rhetoricians.14
Returning to our Longinus text, we may note, finally, that the two open-
ing questions of ch. 18 are themselves not only rhetorical questions, but also
examples of the question types under discussion. The first question, Τί δ᾽ ἐκεῖνα
φῶμεν, τὰς πεύσεις τε καὶ ἐρωτήσεις; is a πεῦσις,15 while the second one, ἆρα oὐκ

12 The full text of the relevant passage from Alex. runs: Ἐρώτημά ἐστι, πρὸς ὃ ἀνάγκη ἀπο-
κρίνασθαι κατ᾽ ἀπόφασιν ἢ κατάφασιν οὕτω, ναὶ ἢ οὔ, οἷον ἐξῆλθες ἐπὶ τὴν μάχην ἢ οὔ; καὶ ὡς
Δημοσθένης, πότερον ταῦτα ποιῶν ἔλυε τὴν εἰρήνην ἢ οὔ; ἀνάγκη γὰρ καὶ πρὸς τοῦτο ἢ ναὶ ἢ οὒ
ἀποκρίνασθαι. πύσμα δέ ἐστι, πρός ὃ διεξοδικῶς ἀπαντῆσαι δεῖ καὶ διὰ πλειόνων, ὡς ἔχει τό τοιοῦ-
τον, ἐμὲ δέ, ὦ τριταγωνιστά, τὸ τίνος φρόνημα λαβόντα τούτοις συμβουλεύειν ἔδει; (‘An erôtêma
is ⟨an expression⟩ that must be answered negatively or affirmatively, as follows, yes or no.
Examples are: “Did you march out to the battle or not?” And, from Demosthenes, “Did he
try to undo the peace by these acts or not?” For it is necessary in the second case, too, to
answer yes or no. To a pusma, on the other hand, one should reply in a more detailed and
more elaborate way, as in the following case: “And as for me, you third part player, in which
mental state should I give them advice?” ’).
13 ‘It is generally agreed that not every universal can form a dialectical proposition, for exam-
ple “What is man?”, or “In what various senses can the good be used?” For a dialectical
proposition is one to which it is possible to answer “yes” or “no”, whereas to the above
questions this is impossible.’ (Forster 1960).
14 Some other sources for the distinction are Apsines Rh., Spengel-Hammer I 305, Theon
Progymn. Spengel II, 96–97, Anonym. De figuris Walz VIII, 703–704, the Scholia in Hes.
Th. 463 (= SVF II 190, Hülser 902), Alex. Aphr. in Top., p. 539, 18 (= SVF II 191, Hülser 903),
the Suda, ss.vv. ἀξίωμα, κατηγόρημα, πυνθάνεσθαι, πύσμα(τα), and the Etym. Magnum, s.v.
τίς. The distinction is not present, however, in Dionysius Thrax; or, rather, the only ques-
tion words mentioned by him are τίς ποῖος πόσος πηλίκος, and these are examples of the
‘erôtêmatikon, which is also called peustikon’ (Ars gramm. 12.80 = p. 39, 1–2 Uhlig). There
is no mention, then, of ἆρα and the yes-no type. Ἆρα is mentioned all right, but in the
section on the aporrhêmatikoi (sc. sundesmoi), i.e. the ‘dubitative conjunctions’. For the
difficulties involved in this classification see Lallot (1989: 244–247). Dionysius’ position is
criticized by his scholiasts, e.g. Stephanus (Sch. D.T. p. 239, 14ff. Hilgard).
15 Or a πύσμα. Πεῦσις/πύσμα and ἐρώτησις/ἐρώτημα seem to have been used indiscriminately.
Incidentally, it is worth noticing that the use of precisely the terms πεῦσις/πύσμα for spec-
ifying questions may be connected with the usage of the related verb, πυνθάνομαι. From
an investigation of Plato it became clear that if (δια)πυνθάνομαι is construed with a depen-
dent question (32 examples) this is in 93% of the cases a πῶς, τί, etc. question, as against
374 chapter 19

αὐταῖς ταῖς τῶν σχημάτων εἰδοποιίαις παρὰ πολὺ ἐμπρακτότερα καὶ σοβαρώτερα
συντείνει τὰ λεγόμενα; is an ἐρώτησις. A translation should reflect the fact that
these are technical terms. In part following the German terminology, I would
suggest ‘specifying question’ for πεῦσις (‘substitution question’ is another pos-
sibility), and ‘decision question’ or ‘yes-no question’ for ἐρώτησις.16

References

Editions etc. of [Longinus]

Brandt, R. Longinus, Vom Erhabenen: Griechisch und deutsch (Darmstadt 1966).
Fyfe, W.H., Russell, D.A. Longinus: On the Sublime (Cambridge, MA/London 1995).
Grube, G.M.A. Longinus: On Great Writing (On the Sublime) (Indianapolis 1957).
Hoogland, J.P. Longinus: Over het verhevene (Groningen 1936).
Lebègue, H., Du sublime (Paris 1939).
Mazzucchi, C.M., Dionisio Longino: Del sublime (Milano 1992).
Op de Coul, M., Longinus: Het sublieme (Amsterdam 2000).
Prickard, A.O., Longinus: On the Sublime. 3rd ed. (Oxford 1946).
Rhys Roberts, W., Longinus: On the Sublime. 2nd ed. (Cambridge 1907).
Russell, D.A., ‘Longinus’: On the Sublime (Oxford 1964).
Russell, D.A., Longinus: On Sublimity (Oxford 1965).
von Scheliha, R., Longinus: Die Schrift vom Erhabenen (Berlin 1938).

only 7 % of εἰ-questions, representing yes-no questions. For ἐρώτησις/ἐρώτημα the picture

is different; it seems that the related verb, ἐρωτάω/ἠρόμην, has no marked preference for
either question type. Be that as it may, the two types are clearly opposed at Pl. Grg. 462c10:
ἤδη πέπυσαι παρ᾽ ἐμοῦ ὅτι φημὶ αὐτὴν εἶναι, ὥστε τὸ μετὰ τοῦτο ἐρωτᾷς εἰ οὐ καλή μοι δοκεῖ
εἶναι.
16 There are also echoes of the distinction in several Latin texts, e.g. Aquila Romanus §§11–
12 (= Halm, Rhetores latini minores pp. 256ff.; already mentioned by Hoogland): ‘Ἐρώτημα,
interrogatum’, e.g. Fuistine illo in loco? ‘Πύσμα, quaesitum’, e.g. Qua igitur ratione bellum
geremus? Lausberg (1973: § 770) and Martin (1974: 285) mention the equivalence of quae-
situm and πύσμα, but were apparently not aware of the precise function of this type of
questions. Another term for πύσμα was percontatio or percunctatio, cp. Goetz and Gun-
dermann (1888: vol. II s.v. ‘percontatio’). For further references see the Thesaurus Linguae
Latinae s.v. ‘percontatio’ 2.b. Cf. also the juxtaposition of percontando atque interrogando
at Cic. Fin. 2.2. As to percontor, observe that, in Cicero, the verb percontor or percunctor,
if construed with a dependent question, is only found with a specifying question (five
instances). Perhaps this use mirrors that of πυνθάνομαι in Greek, see the preceding note.
There are also two cases of direct specifying questions with percontor, de Orat. 2.255: ‘per-
contatur ita: “quanti addictus?” ’ and Sen. 20: ‘ “Cedo qui vestram rem publicam tantam
amisistis tam cito?” sic enim percontantibus …’.
a question of questions 375

Other Works
Hicks, R.D. (ed.), Diogenes Laertius: Lives of Eminent Philosophers, 2 vols. (Cambridge,
MA/London 1958).
Goetz, G., Gundermann, G., Glossae latinograecae et graecolatinae (Leipzig 1888).
Hülser, K., Die Fragmente zur Dialektik der Stoiker, 4 Bände (Stuttgart 1987/1988).
Lallot, J., La grammaire de Denys le Thrace (Paris 1989).
Lallot, J. (ed.), Apollonius Dyscole: De la construction, 2 tomes (Paris 1997).
Lausberg, H., Handbuch der literarischen Rhetorik (München 1973).
Lyons, J., Linguistic Semantics (Cambridge 1995).
Martin, J., Antike Rhetorik (München 1974).
Tredennick, H., Forster, E.S. (eds), Aristotle: Organon, Posterior Analytics, Topica (Cam-
bridge, MA/London 1960).
 chapter 20

The Xenophon Factory: One Hundred and Fifty

Years of School Editions of Xenophon’s Anabasis

Abstract

This chapter offers an overview of school commentaries on Xenophon’s Anabasis

published in Germany, Great Britain, the United States and elsewhere, starting from
Krüger’s 1830 edition (the first to present notes in the vernacular and not in Latin).
It considers what can be learnt from them about the place of the Anabasis in school
education in these various countries, and offers a number of reflections on what
could/should be expected of school editions. After noting that there are very consid-
erable differences between the commentaries under review in terms of, among other
things, the density of coverage and the treatment of grammatical phenomena, it then
looks in detail at how they interpret two problematic items in the first sentence of
the Anabasis, viz. the genitives Δαρείου καὶ Παρυσάτιδος and present-tense γίγνονται.
It concludes that there is a bewildering number of views on offer, and that much of the
information and analysis provided is repetitive and unreliable.

1 Introduction

In 1862, shortly after Easter, thirteen-year-old Ulrich von Wilamowitz-

Möllendorf (1848–1931) was admitted as an ‘Untertertianer’ to the venerable
Landesschule Pforta in the province of Saxony.* At home he had received his
primary education, which in the final years also included Greek and Latin. He
now entered a school that had an outstanding reputation and a long tradition
in the field of classical education. Yet young Wilamowitz was less than enthu-
siastic about his Greek lessons:

das Griechische war nicht mehr als ein Anhang zum Latein. Formen-
lehre ohne jeden Hauch von sprachwissenschaftlich orientierter Gram-

* This article originates from a seminar at the University of Amsterdam during the autumn
term of 1999. I am indebted to Frederik Bakker, Annemarie Bronneman, Matthias Haentjens,
Guus van der Kraan, Yvette van der Raad, Josine Schrickx, Judith Stavast, and Aart van Wijk
for their contributions to that seminar.

© koninklijke brill nv, leiden, 2019 | doi:10.1163/9789004386129_022

the xenophon factory 377

matik, Pauken der nutzlosen Akzente, nachher jene ungriechische, auf

den unpassenden lateinischen Leisten geschlagene Syntax. … Auswahl
der Lektüre ungeschickt, Xenophon immer weiter,1 Lysias, Arrian. Eine
Rettung war erst, daß die Odyssee in der Untersekunda, die Ilias in der
Obersekunda durchgelesen wurden mußte.2

Wilamowitz’s verdict must have been influenced by his teacher, who fares badly
at his hands. In his memoirs Wilamowitz does not tell us what his school cur-
riculum looked like, but it will not have been much different from the cur-
riculum that is printed in the tercentenary anniversary edition of the Pforta
Schoolbook from 1843.3 In | that year 61 hours were devoted to Latin, distributed 236
over five classes, and 27 to Greek, again distributed over five classes. For a better
understanding of the system in which Wilamowitz received his education, and
of the role of the classics in general and Xenophon in particular in that system,
it is necessary to say a few words about its background.

2 Nineteenth-Century Education, in Germany and Elsewhere

From 1812, the organization of school teaching in Germany was determined

by a new curriculum, which in turn was the result of the Laws on Education
established by Wilhelm von Humboldt. In 1810, at the end of a brief career
as the Prussian Minister of Education, this multi-talented scholar4 and diplo-
mat had profoundly reformed the educational system. His reform was based
on the following premise: ‘Es gibt, philosophisch genommen, nur drei Sta-
dien des Unterrichts: Elementarunterricht, Schulunterricht, Universitätsunter-
richt.’5 The ‘Schulunterricht’ comprised advanced education, and was subdi-

1 An allusion to ἐντεῦθεν ἐξελαύνει? Incidentally, the Greek text had to be translated into
Latin!
2 ‘Greek was no more than an appendix to Latin. Accidence with no trace at all of a linguis-
tically oriented grammar, hammering home the useless accents, next that un-Greek syntax
which was modeled after the unfitting Latin one. … Reading selections unsuitable, Xenophon
ever going on, Lysias, Arrian. Our only salvation was that the Odyssey had to be read in the
“Untersekunda” and the Iliad in the “Obersekunda” ’ (Wilamowitz-Möllendorf 1928: 71).
3 Kirchner (1843).
4 Especially in the field of linguistics. The Danish linguist Otto Jespersen considered Humboldt
‘one of the profoundest thinkers in the domain of linguistics.’ He is of course also known as
the founder of the University of Berlin.
5 ‘There are, philosophically speaking, only three stages of education: elementary education,
school education, university education.’
 378 chapter 20

vided into the ‘Bürgerschule,’ a kind of professional school, and the ‘Gelehrten-
schule,’ whose objective it was to prepare for higher education. At the end of
that school

[ist] der Schüler reif, wenn er soviel bei anderen gelernt hat, daß er
nun für sich selbst zu lernen imstande ist. Sein Sprachunterricht z. B.
ist auf der Schule geschlossen, wenn er dahin gekommen ist, nun mit
eigener Anstrengung und mit dem Gebrauch der vorhandenen Hilfsmit-
tel jeden Schriftsteller, insoweit er wirklich verständlich ist, mit Sicherheit
zu verstehen, und sich in jede gegebene Sprache, nach seiner allgemeinen
237 Kenntnis vom Sprachbau überhaupt, leicht und schnell einzustudieren.6 |

Humboldt’s school system was introduced with remarkable speed and has been
immensely influential, also outside Germany, to a certain extent up to the
present day. One important consequence of this reform was that the ‘Gymna-
sium’ became a ‘pre-academic’ institution, and this created an enormous need
of ‘Hilfsmittel,’ whose availability Humboldt apparently took for granted. In the
case of Xenophon, who already before the reform had been a favorite elemen-
tary author, this led to a real explosion of editions with ‘Erklärungen,’ whether
or not ‘für den Schulgebrauch.’ Remarkably enough, this development was not
confined to Germany but took place, almost simultaneously, in other coun-
tries as well. A common feature of these editions is that the notes, with a few
exceptions, were no longer given in Latin, but in the vernacular. Thus, in the
German-speaking countries the language of the notes was, from Krüger’s 1830
Anabasis edition onwards, German.7 With this, and more particularly with the
second edition of his commentary, Krüger set the fashion for all later commen-
taries.

6 ‘The pupil is mature when he has learned so much with others that he is now capable of learn-
ing all by himself. Thus, his linguistic education at school is finished when he has reached the
point that he understands with certainty, using his own forces and the available aids, any
author, in as far as he can really be understood, and may get acquainted easily and quickly
with any given language, using his general knowledge of the structure of language at large’
(Humboldt 1809: 12 ff.).
7 I have found a few references to this edition in German, but it is not mentioned in any of the
Dutch University libraries. The language of the edition by Graff 1842 is probably Latin, but I
was not able to establish this as a fact. Greek (school-)grammars in German had been avail-
able for quite some time, e.g., that of Ph. Buttmann, of which the 10th edition appeared in
1837.
the xenophon factory 379

3 School Editions of the Anabasis

On the assumption that Wilamowitz, presumably at the end of his first year,
in the so-called ‘Obertertia,’ started reading the Anabasis (as is strongly sug-
gested by the words ‘immer weiter’ in the above quotation),8 his teacher could
choose from no fewer than five school editions: Kühner 1852, 4Krüger (i.e., the
fourth edition) 1854, 3Hertlein 1857, 2Vollbrecht 1861, 2Matthiä 1859 (see the
table below). Regrettably, Wilamowitz does not tell us which edition his school
used, nor what the ‘ungriechische’ syntax was which so provoked his wrath.
It goes without saying that it is no small task to establish in what way and to
what extent the elementary stages of the learning of Greek influence the later
views of a professional classicist, apart from | the stimulating effect that may or 238
may not have been brought about by teacher, reading, and coursebook. Such
influences cannot be ruled out, for some things learned in school may have
a very persistent life indeed. The ‘vividness’ allegedly expressed by the historic
present (see also below), or the idea that ‘the’ aorist expresses ‘pastness,’ are per-
haps cases in point. Be that as it may, it definitely made a difference whether
the one or the other edition was used: the number of the notes, their layout,
degree of detailedness, clarity, scholarly qualities9—on all these points there
are considerable differences between the school editions. Thus, the pupil who
used Rehdantz’s edition was bound to get a completely different idea of the
functions of the genitive and the present indicative from the pupil with Krüger
on his desk.
Below I shall treat in some detail the quality of the grammatical notes of
these and other editions. But first, I present in tabular form a survey of a large
number of school editions of the Anabasis. By way of introduction to this sur-
vey, perhaps a few words about the dominant position of Germany, notably
with respect to the US, are in order. An eloquent and unequivocal testimony
of the German influence on the American classical world is to be found in
B.L. Gildersleeve’s article ‘English and German Scholarship’; in this article, the
great American classicist (1831–1924) looks back at his formative years.10 ‘Such
philological schooling as I have had is wholly Teutonic,’ he says bluntly, because

8 It appears from the anniversary book (above, n. 3) that in 1843 the Memorabilia were on
the programme, but these may of course have been replaced by the Anabasis in the course
of the 20 years between 1843 and 1863.
9 In this connection it is worth noticing that up to the present day no scholarly commentary
on the Anabasis has been published, on which school editions might have drawn.
10 Gildersleeve (1930: 365).
 380 chapter 20

his ‘American teachers did not understand their business.’11 Actually, this ‘Teu-
tonic’ schooling was not self-evident, for in spite of the Revolution, whose
echoes still resounded loudly in the first half of the nineteenth century, the
American classical world was very much oriented towards Great Britain, and
initially this also applied to Gildersleeve. But he went to Germany (Bonn) for
an academic training in Greek and Latin, and it was only there that he got an
education that was worthy of that name. Reacting to an article by Gilbert Mur-
239 ray, which also | deals with the learned classical worlds in Germany and Great
Britain, Gildersleeve summarizes his views as follows:

it is not to be denied that for the American classical teacher who wishes
to fit himself for his work in life the only sensible course is to familiarize
himself with German methods, and in my day that could only be encom-
passed in Germany itself.

In the same article Gildersleeve pronounces in passing a harsh verdict on the

phenomenon of the school edition:

an Oxford Don once lamented to me the modern mania for writing books.
If he meant the run of school editions he was quite right. Most of them
are absolutely negligible for the advanced student, and no book ought to
be published that does not contain some individual contribution to what
is already known.12

How true his observation on the ‘run of school editions’ is, becomes apparent
from the following survey. The question whether these editions should be con-
sidered ‘negligible’ will be addressed after the survey.

11 In Gildersleeve’s youth, school editions of the Anabasis in English were, in the US, still
scarce. In the 1840s a teacher could choose between Cleveland, published in 1834, and
Owen, in 1843.
12 Incidentally, Gildersleeve’s ire was not so much directed at school editions, as to certain
(unnamed) British classicists, for he continues as follows: ‘But one waxes impatient at the
reputations that have been gained in England by infinitesimal productions. Where else on
God’s earth would a man gain immortality by an Introductory Lecture?’
 the xenophon factory 381

table 1 A survey of (school) editions of the Anabasis in Germany, the United States, Great
Britain, and some other countries, 1826–200013

Period of Germany United States Great Britain Other countries

publication

1820–183014 [1826, C.G. Krüger,

Halis Saxonum]
1830–1840 [1830, K.W. Krüger; 1834, C.D. Cleve- [1830, F.C. Belfour,
notes in German] land,15 Boston London]
[1834, 2Belfour]
[1835, A. Negris,
Edinburgh] | 240
1838, J.T.V. Hardy,16
London [1839,
D.B. Hickie, Books I–
II, London]
1840–1850 [1842, G. Graff, [1843, J.J. Owen,17 [1841, C.S. Stanford,
Leipzig] New York: Appleton Dublin]
1845, K.W. Krüger, & Co.] [1844, 2J.A. Phillips,
Berlin, bei K.W. Krü- Dublin; year of 1st ed.
ger (in German; unclear]
‘zweite Ausgabe’)

13 I may have missed a couple of editions, but on the whole I believe that the survey is fairly
complete, at least for the nineteenth century and for Germany, the US, and Great Britain.
I am indebted to Dr. T. Rood (Oxford) for pointing out to me the existence of many British
editions from the nineteenth century which I initially had overlooked, and to Mr. M. de
Bakker (Oxford) for providing me with copies of some of these editions. Of many editions,
not only the American and British ones, but, e.g., also those of Hertlein and Bersi, I have
been able to consult the copies in Widener Library (Harvard University). I am indebted
to the Widener staff for their assistance in collecting the material. Prof. H.T. Wallinga
(Utrecht) was so kind as to lend me his copy of Brownrigg’s edition.
Square brackets around a given name indicate that the edition concerned was not seen
by me (which unfortunately occurred quite often, especially in the case of many British
editions from the first half of the nineteenth century). Some editions cover only selections
from the Anabasis; this is sometimes but not always indicated in the survey.
14 Apart from Krüger’s edition there appeared, in the first decades of the nineteenth century,
still many other editions, all with notes in Latin; sometimes these were new, sometimes
reprints of eighteenth century editions. Some examples are: 1825, Bornemann, Lipsiae;
1827, Poppo, Lipsiae; 1821, Schneider, Oxonii; 1825, Hutchinson, Glasgow (originally pub-
lished 1735). Since these editions played no role in schools, I have not included them in
the survey.
15 ‘Professor of the Latin Language and Literature in the University of the City of New York.’
16 ‘Assistant Master in University College School.’
17 ‘D. D., Ll.D., Formerly Principal of the Cornelius Institute, and now Professor of the Latin
and Greek Language and Literature in the Free Academy in New York City.’
 382 chapter 20

Table 1 A survey of (school) editions of the Anabasis (cont.)

Period of Germany United States Great Britain Other countries

publication

[1849, 2Krüger] [1847, J.F. Mac-

[1849, F.K. Hertlein, michael,18 London:
Berlin: Weidmann] G. Bell & Sons]
[1847, J.T. White, Lon-
don]
[1848, Ch. Anthon,19
London]
[1848, J.S. Fergusson
(or Ferguson), Edin-
burgh]
[1849, 2Hickie, Lon-
241 don, ‘rev. ed.’] |
1850–1860 [1850, 3Krüger] 1857, J.R. Boise,21 [1850, 2Fergusson] France:
[1851, K. Matthiä, New York: Appleton [1852, 2Anthon, new [ca. 1860, F. Dübner,
Quedlinburg] & Co. ed. rev. by T. Doran] Paris]
1852, R. Kühner, [1853, Browne,22
Gothae, notes trans. of Hertlein
in Latin [1854, 1849 with add’l notes,
2Hertlein] Arnold’s School Clas-
1854, 4Krüger, Berlin: sics]
K.W. Krüger’s Ver- [1853, 2Macmichael]
lagsbuchhandlung; [1853, J.T.V. Hardy &
simultaneously pub- E. Adams, London
lished in London (‘portions only’)]
and New York 1857, [1854, 4Fergusson]
3Hertlein [1854–1855,
[1857, F. Voll- H. Young,23 London]
brecht,20 Leipzig: [1857, 3Macmichael,
Teubner] ‘new ed.’]
1859, 2Matthiä

18 ‘B. A., Trinity College, Cambridge; Head Master of the Grammar School, Ripon.’
19 ‘Ll.D., Professor of the Greek and Latin Languages in Columbia College, New York, and
Rector of the Grammar School.’ This edition was apparently first published in London,
and only afterwards in New York (see below under 1868).
20 ‘Rector zu Otterndorf.’
21 ‘Professor of Greek in the University of Michigan.’
22 ‘Prebendary of Chichester.’
23 ‘Second Master of the Grammar School, Guildford.’
 the xenophon factory 383

Table 1 A survey of (school) editions of the Anabasis (cont.)

Period of Germany United States Great Britain Other countries

publication

1860–1870 [1861, 2Vollbrecht] 1867, 2Owen, pref- 1860, Macmichael

[1863, C. Rehdantz, ace dated 1861 (‘new edition’)
Berlin: Weidmann] (‘Since that time [1860, 6Fergusson]
[1865, 3Vollbrecht] [viz., 1843] about [1866, 9Fergusson]
1865, L. Breitenbach, 35,000 copies have [1866, 2White]
Halle: Verlag der been printed and [1866, K.S. Isbister,
Buchhandlung des sold’) London (Books I–III)]
Waisenhauses 1868, 3Anthon [1868, Macmichael
[1869, 4Vollbrecht] (‘new edition’)]
[1868, J.S. Watson,
London]
1870–1880 [1871, 6Krüger] [1875, A. Crosby, [1872, Macmichael
[1873, 5Vollbrecht] New York] (‘new edition’)]
1877, 6Vollbrecht (‘up [1877, A. Pretor,24
to and incl. the 5th Cambridge: CUP.,
ed. 20000 copies repr. 1878]
sold’) [1879, R.W. Taylor,
London] | 242
1880–1890 1888 6Rehdantz, [1889, F.W. Kelsey26 1880, W.W. Good- Austria: [1884,
bearb. von O. Car- & A.C. Zenos,27 win28 & J.W. R. Hansen, Wien]
nuth25 Boston] White,29 London: Italy: [1886, A. Bersi,
[1889, E. Bachof, Macmillan Torino]
Paderborn] 1881,2Pretor (repr.
1889. 7Krüger, besorgt 1882, 1885, 1891, 1894,
von W. Pökel 1900)
1882, A.S. Walpole,30
London: Macmillan
(repr. 1885, 1888, 1891,
1894, 1896, 1898, 1900,
1902, 1903, 1906, 1909)
1882, H.R. Parker,31
Dublin (Book I)

24 ‘Fellow of St. Catharine’s College, Cambridge.’

25 ‘Professor Dr., Direktor des städtischen Gymnasiums in Danzig.’
26 ‘Ph.D.’
27 ‘M. A.’
28 ‘Ph.D., Eliot Professor of Greek Literature in Harvard College.’
29 ‘Ph.D., Assistant Professor of Greek in Harvard College.’
30 ‘M. A., Formerly Scholar of Worcester College, Oxford.’
31 ‘Ll.D. Headmaster Methodist College Belfast. Ex-scholar and Senior Classical Moderator,
Trinity College Dublin.’
 384 chapter 20

Table 1 A survey of (school) editions of the Anabasis (cont.)

Period of Germany United States Great Britain Other countries

publication

1890–1900 [1890, A. Weidner, 1892, Kelsey-Zenos [1892, A.H. Allcroft Austria: [1890,
Leipzig: Freytag] (‘third edition’ but & F.L.D. Richardson, A. Weidner, Wien
[1895, F.G. Sorof, actually = ed. of London (Book I)] (same as the Ger-
Berlin: Teubner] 1889) [1894, C.E. Brown- man ed., but with
[1896, 9Vollbrecht, 1895, 4Kelsey-Zenos rigg,33 London: another publisher)]
bes. von W. Voll- (‘fourth edition,’ Blackie & Sons] [1894, 2Weidner]
brecht32] but now revised; [1885–1892, J. Mar- France: [1896,
repr. 1948 as ‘sixth shall,34 Edinburgh] P. Couvreur,35
edition’) [1896, 2Allcroft & Paris]
Richardson] Holland: [1899,
[1897, Fergusson & J. Mehler & H.O. de
W. Coutts, Edinburgh Jong, Utrecht]
(‘new ed.’)] Italy: [1893, 2Bersi;
‘of the 1st ed. 4000
copies sold’]
1900–1910 [1905, 4Weidner] 1901, G.M. Ed- France: 1900,
1907, 10Vollbrecht wards,36 Cambridge: 2Couvreur
CUP (Book I) Italy: 1905 1905,
1902, 2Brownrigg 3Bersi (repr. in
243 [1908, Goodwin- 1912) |
White, unalter. repr.
of 1880?]
1910–1920 1912, 7Rehdantz- 1912?, M.W. Mather Austria: [1919,
Carnuth, bearb. v. & J.W. Hewitt, s.l. 5Weidner] France:
E. Richter [1913, 5F. Dürrbach,
Paris]
1920–1930 1929, 8Sorof, bes. von
G. Sorof37
1930–1940 Holland: [1933,
H. Rogge, Zwolle]

32 ‘Dr., Professor am Christianeum zu Altona.’

33 ‘M. A., Headmaster of Magdalen College School, Oxford.’
34 ‘Rector of the Royal High School.’
35 ‘Ancien Maître de Conférences à la Faculté des Lettres de l’Université de Lille.’
36 ‘M. A., Fellow and Tutor of Sidney Sussex College, Cambridge.’
37 ‘Geh. Stud. Rat Prof. Dr., Oberstudiendirektor i.R.’
 the xenophon factory 385

Table 1 A survey of (school) editions of the Anabasis (cont.)

Period of Germany United States Great Britain Other countries

publication

1940–1950 [1947, Ed. Borne- 1948, Kelsey & Holland: [1947,

mann, Frankfurt/M. Zenos; on the A.M. Bent,
(‘Auswahl’)] titlepage ‘sixth edi- Nijmegen]
[1949, A. Clausing tion,’ but actually
(‘Auswahl’)] unaltered repr. of
4th ed.
1950–1960 [1952, 2Clausing]
[1954, Bornemann
(‘neue Ausgabe’5)]
[1956, 6Bornemann]
[1956, 3Clausing]
1960–1970 [1960, 4Clausing] 1962, Mather & Holland: 1963, 7Bent
[1962, 9Bornemann] Hewitt (reprint)
[1963, 10Bornemann]
1970–1980 [1974, 14Bornemann] 1980?, J. Antrich38
[1979, 23E. Krämer, and St. Usher,39 Bris-
Münster; printing tol: BCP40
history could not be
traced]

What can we learn from this survey of what with only a little exaggeration 244
may be called ‘the Xenophon factory’? (I confine myself to the nineteenth and
early twentieth centuries.41) In Germany, the high point of the Anabasis school
editions falls between 1850 and 1870, in the US slightly later, between 1860
and 1880. In Great Britain, on the other hand, one can hardly speak of a high
point, for during the whole of the nineteenth century the school editions of
the Anabasis kept pouring out in all parts of the British Isles. This is connected
with another difference. In Germany there appear, from 1840 onwards, three
editions that for some or a long (and sometimes a very long) time will dominate
the market. In Germany, Krüger’s editions, of which the 1845 edition should be

38 ‘Head of Classics, Godalming Sixth Form College.’

39 ‘Senior Lecturer in Classics, Royal Holloway College, University of London.’
40 On the titlepage: ‘Xenophon. The Persian Expedition.’ Actually, however, it is a small
anthology.
41 Observe that after the Second World War in Germany two new Anabasis-commentaries
appeared, of which that by Ed. Bornemann became very popular. Unfortunately, from lack
of material I could not compare these with their predecessors.
 386 chapter 20

considered the first ‘real’ edition, gets heavy competition from Vollbrecht and
Rehdantz, for the first time in 1857 and 1863, respectively. Of Krüger, Vollbrecht,
and Rehdantz there appeared respectively in toto seven, ten, and twelve (usu-
ally revised) editions. Kühner’s 1852 edition was something of an anomaly, since
the notes were in Latin. This may explain why there appeared just this edition:
for pupils it was probably unfit. (His fellow editors made extensive use of it.) But
in Great Britain and Ireland the situation was different. Although there were a
number of editions which remained popular for several decades (Macmichael,
Fergusson, Pretor, Walpole), the market was certainly not dominated by these.
Every now and then yet another edition appeared. Between 1830 and 1900 the
Xenophon factory in the British Isles employed all in all some 22 authors, that
in Germany eleven.
Some other conclusions impose themselves. In Great Britain, the wide vari-
ety of editors was paralleled by an equally wide variety of publishers. In Ger-
many, on the other hand, the classical market, both the academic and the
schoolmarket, was largely in the hands of two big publishing companies,
namely Weidmann and Teubner. The history of the first went back to 1680,
while Teubner, established in 1811, started publishing classical texts from 1821.42
All big publishers profited enormously from Humboldt’s reform, and it was only
245 to be expected that they would include a favorite, and thereby | commercially
very attractive, school author like Xenophon in their list.43 Also, the success
of Krüger, a dangerous outsider who published his books with his own firm,
will not have gone unnoticed. That the competition was heavy, and commer-
cial success bitterly fought, can be inferred from the stray remarks that we find
in some Introductions. Thus, Vollbrecht notes in 1877 that of his preceding edi-
tions 20,000 copies have been sold, apparently a fact to boast about. From the
survey it also appears that it was very unwise to publish your edition with a
small, provincial, publishing house. See the editions by Matthiä (1851, reprinted
once), Breitenbach (1865, no reprint) and Bachof (1889, no reprint).

42 There is a direct link with the measures taken by Humboldt: see Garzya (1983: 19).
43 For such responses by publishers to changes in the curriculum there are also modern par-
allels. Thus, in Holland the reform of the classics in secondary schools which started in
the 1970s was followed by a whole new series of coursebooks and annotated editions; the
format of the latter was completely different from the earlier school editions, since the
emphasis was now on providing continuous help in studying the text, notably by giving
very copious lexical annotations. Likewise, as Roy Gibson pointed out to me, the ‘insis-
tence in the Cambridge “green and yellow series” on “commenting on texts as literature”
is in good part a response to the changing demands of the syllabus in secondary schools,
where less emphasis is now placed on knowledge of the language.’ See further Stephens
(2002: 84–85) and Henderson (2002: 213, 218–219).
the xenophon factory 387

Furthermore, it is worth noticing that Weidmann all of a sudden changed

commentators: in 1857 Hertlein had published his third edition of the Anaba-
sis, but only six years later he is replaced by Rehdantz. The reason is a matter
of guesswork. It cannot have been lack of success, for within 10 years there had
appeared three editions by Hertlein. Nor was he deceased, for he kept editing
other texts for Weidmann. A comparison with his successor, Rehdantz, suggests
that Hertlein’s notes may have been too summary, and that Weidmann would
lose the battle with the new edition by Vollbrecht, that had been published in
1857 by its arch-rival Teubner.
Another remarkable fact is that at the end of the century Teubner simulta-
neously published two Anabasis-editions: from 1895 onwards Vollbrecht with
his own publisher gets competition from Sorof, who is very successful, too,
and eventually ousts Vollbrecht: in 1907 the latter’s edition is revised for the
last time. The same happened in Great Britain and the US: in Great Britain,
Macmillan was the publisher of both Goodwin–White 1880 (reprinted once)
and Walpole (from 1882), and Cambridge University Press published both Pre-
tor | (at least until 1900) and Edwards (in 1901); for the US see Owen 1843 and 246
Boise 1857. Had the earlier editions become too learned, perhaps because of
changes in the curriculum in the ‘Gelehrtenschule’ and its Anglo-Saxon coun-
terparts? In the case of Teubner, the latter explanation is not unlikely, for Sorof’s
edition no longer covers the whole Anabasis, but only a selection. Also, while
the notes are very copious they are also very brief, and often they are just trans-
lations.

4 Cultural Differences

In the footnotes to the survey I have indicated how the editor presents himself.
It appears that there are considerable cultural differences between on the one
hand Germany, and on the other the US and Great Britain. In the US and Great
Britain the editor nearly always mentions his credentials;44 in Germany this is
very exceptional. In Germany, it was apparently self-evident that someone who
edited the Anabasis, or indeed any classical text, was qualified to do so, while in
the other countries an appeal to the auctoritas of the editor was needed. Only
in the course of the nineteenth century does the German practice change, see,
e.g., Carnuth 1888 and W. Vollbrecht 1896.

44 There is only one exception, Mather and Hewitt 1962, but this may be a feature of the
reprint. I have not been able to consult the original edition of ca. 1912.
 388 chapter 20

There is another clear and striking difference, now between Germany and
the British Isles. Whereas in the latter, especially in the 1880s and 1890s, many
popular editions are simply reprinted, often at a remarkable pace (cf. notably
Pretor 1881 and Walpole 1882; but there are exceptions, such as Macmichael
and Fergusson), in Germany this never occurs, strictly speaking. Always the fact
that some edition has been sold out leads to the next edition being ‘verbessert,’
possibly even ‘sehr verbessert.’45 Often the reasons for correcting the previous
edition are mentioned in the Preface: these may be the experiences gained
by the use of the book in the classroom, or the appearance of new publica-
tions on the Anabasis, or of other editions. Commercially, this publishing policy
was both risky (readers could not go on for ever purchasing new editions) and
247 attractive (one really | could not afford to ignore the latest insights). Was the
scholarly conscience of the German editors and their audience better devel-
oped? Or was this situation rather due to the competition in Germany being
heavier? If one editor decided to revise his edition, the others naturally could
not stay behind. Or perhaps the role of the public authorities was the decisive
factor. In Germany, the educational material was prescribed for all schools by
the provinces, and the authorities possibly saw to it that this material was not
outdated.

5 ‘Für den Schulgebrauch’—or is There More to It?

All editions mentioned in the survey are, practically speaking, school editions.
One might expect this to be mentioned, e.g., on the title page. Also, one may
expect a preface, which gives some information about the intended audience,
and accordingly about the structure of the edition and the nature of the notes.
I have scrutinized the editions on these points, and for a number of them I
present the results in the next table. In this table, I also indicate whether the
editors ever refer to each other, or to a grammar.

45 In this respect, too, the situation in the US was similar to that in Germany, although in the
US the success of most editions was limited; only Kelsey and Zenos’s edition was revised
and reprinted several times.
 the xenophon factory 389

table 2 Some characteristics of the editions46

Edition ‘Für den Schul- Preface with References to other editions/ grammars?
gebrauch’ vel objectives, etc.?
sim.
on title page?

Krüger 1845 no no other ed.: –

gramm.: Krüger
Kühner 1852 no yes other ed.: ‘superiores editores’ in general gramm.:
Kühner
4Krüger 1854 no no other ed.: ‘Ze.’
gramm.: Krüger
Vollbrecht 1857 yes yes other ed.: ‘alle Arbeiten früheren Herausgeber’
gramm.: ‘die verschiedensten’
2 Vollbrecht 1861 yes yes, justification as in first ed. | 248
of changes with
respect to 1st ed.;
sim. in next edd.
Anthon 1867 yes, extensive no the only ed. with a bibliography (up to 1845);
does not mention his American colleagues Cleve-
land and Owen
Goodwin-White no yes other ed.: general remark about predecessors47
1880 gramm.: ‘notes adapted to Goodwin’s Greek
Grammar’
Pretor 1881 no yes other ed.: Kühner, Schneider, Vollbrecht, Borne-
mann, Macmichael gramm.: Madvig
Walpole 1882 yes yes other ed.: notably those by Rehdantz, Krüger,
Vollbrecht; Goodwin & White (‘the best of the
English editions’) gramm.: Goodwin
Bersi 1886 no yes other ed.: Krüger, Breitenbach, Kühner, Matthiä,
Rehdantz, Vollbrecht gramm.: ?
6Rehdantz– no no, just a few no references? (printing history not mentioned
Carnuth 188848 remarks about in the ‘preface’)
the corrections
7Krüger-Pökel no no other ed.: a great many, also from the eighteenth
1889 century, without any bibliographical information
gramm.: Krüger | 249

46 Sometimes I refer to an edition that I have not seen; in these cases the relevant informa-
tion was supplied by the preface(s) of earlier editions.
47 ‘It is of course impossible in notes like these to give special credit for all interpretations
which are wholly or partly borrowed. No small erudition, indeed, would now be required
to trace every valuable remark on the Anabasis to its original source or sources. The Edi-
tors must therefore express, once for all, their obligation to the long and familiar line of
commentators on Xenophon, whose diligence has rendered further originality well-nigh
impossible.’ See further Kraus (2002: 11–13, 16–17).
48 The first edition, the work of Rehdantz alone, appeared in 1863; I have not seen it.
 390 chapter 20

Table 2 Some characteristics of the editions (cont.)

Edition ‘Für den Schul- Preface with References to other editions/ grammars?
gebrauch’ vel objectives, etc.?
sim.
on title page?

Kelsey-Zenos no yes other ed.: Rehdantz, Vollbrecht, Matthiä, Taylor

1889 (= 1892) gramm.: Goodwin, Hadley-Allen
2Bersi 1893 no yes, justification other ed.: as in 1886
of changes with gramm.: C. (= Curtius?)
respect to 1st ed.,
sim. in 1905 ed.
4Kelsey-Zenos no yes, very brief other ed.: as in 1889
1895 justification of gramm.: Goodwin, Hadley-Allen, also Goodwin,
changes with Moods & Tenses
respect to 1889
ed. (= 1892)
Couvreur 1896 no yes, extensive other ed.: Vollbrecht, Rehdantz-Carnuth,
Goodwin-White
gramm.: Croiset & Petitjean
Brownrigg 1902 no no other ed.: –
gramm.: –
Mather-Hewitt no yes other ed.: –
1912 (?) gramm.: Smyth, Hadley-Allan, Babbitt, Goodwin,
Goodell
Sorof 1929 yes yes, but very other ed.: –
brief gramm.: –
Bent 1947 no yes, but very other ed.: –
brief gramm.: van Oppenraaij-Vermeulen
Antrich-Usher in the introduc- yes other ed.: –
1980 (?) tion: for students gramm.: –

It is clear that some editions pretended to be more than just a school edi-
tion. This especially applies to those of Krüger. There is not a single edition by
Krüger where we read ‘für den Schulgebrauch’ on the title page, and he must
have assumed, in fact, that by the sheer quantity, the degree of detail, and the
250 quality of his notes his editions were in a class of their own.49 |

49 This is confirmed in an unexpected way by the Preface to Vollbrecht’s edition from 1877. In
this Preface, Vollbrecht thanks a number of colleagues and reviewers for their comments,
with one exception, an anonymous reviewer who ‘die Krüger’sche Ausgabe, deren Werth
… niemand bezweifelt, als Muster hervorhebt’ (‘who extolls Krüger’s edition, whose value
nobody doubts, as a model’). While Vollbrecht acknowledges, then, the value of Krüger’s
edition, he adds in a parenthesis that it is valuable ‘nur für Lehrer’ (‘only for teachers’). Krü-
the xenophon factory 391

From the table it appears that the editors, generally speaking, did not ignore
each other. Only Krüger, again, shows idiosyncratic behavior. To be sure, start-
ing with the 1854 edition he occasionally refers to others, but in an absolutely
impenetrable way; thus, some notes are followed by ‘Ze.’ To whom or what this
refers remains a mystery.50 The simplest, and rhetorically most effective, way
of letting it transpire that your notes were the apogee of a long tradition was,
of course, to mention your predecessors—but not by name. This is done by
Kühner, Vollbrecht (whose ‘alle Arbeiten’ is not correct: see below), and most
fully and frankly by Goodwin-White (quoted above, n. 47). As indicated in the
survey, just one single editor gives in his preface a regular bibliography, Anthon
from 1867 (1847).
On closer inspection, another feature strikes the eye: while the nineteenth-
century British and American (and also the French and the Italian) editors
frequently refer to their German counterparts, the reverse never occurs. The
explanation is simple: not only must the German editors have felt that they
would learn nothing from these editions (if they knew them at all), but they
could not read English in the first place. Wilamowitz is clear on this point:
‘Französich sollte gelernt werden. … Englisch lag noch ganz allgemein außer-
halb des Horizontes; es gab wohl unter den Gästen, die in Kobelnik erschienen,
niemanden, der es auch nur von ferne kennte.’ The gymnasium brought no
improvement: ‘Englisch kannte das Preußische Gymnasium nicht.’51 Wilam-
owitz regarded himself as a very privileged exception, because he had had the
opportunity to learn English as an extra-curricular activity. His remarks are con-
firmed by Gildersleeve, who observes, ‘Few of the German classical scholars of
my day even pretended to know English’;52 during his studies in Germany, on
the rather rare occasions when the use of an English text could not be avoided,
he acted as an interpreter. | 251
Finally, I should point out that in the course of the twentieth century the
nature of the editions changes. Normally, the editions I consulted no longer
mention any predecessors, or grammars, for that matter. The most telling exam-

ger’s prestige is also apparent from the fact that his edition of 1854 was simultaneously
published in Berlin and (in German!) in London and New York.
50 In the seventh edition (but perhaps also in earlier ones, not seen by me) it appears that
an edition by Zeune is meant, without further details. This is probably Johann Karl Zeune,
who in the second half of the eighteenth century published editions of several works by
Xenophon (with notes in Latin).
51 ‘French had to be learned. … English at that time [ca. 1860] lay quite generally outside our
horizon; among the guests that made their appearance at Kobelnik [his parental home]
there was presumably not a single one who knew it even remotely. … English was unknown
in the Prussian gymnasium’ (Wilamowitz 1928: 59, 78).
52 Gildersleeve (1930: 366).
 392 chapter 20

ples are the editions by Sorof (at least that of 1929) and by Brownrigg (1902) and
Antrich–Usher (no date, probably around 1980). A novice to the field cannot
but have received the impression that these editions were the result of genera-
tio spontanea.
After this introduction I now pass on to a discussion of the linguistic mer-
its of these editions, which will be illustrated from the very beginning of the
Anabasis.

6 Xenophon, Anabasis 1.1.1–4: What Are the Problems?

Things to be Required of a School Edition53

A school edition, and indeed any edition with notes, should meet at least two
requirements. First, in an ideal situation, the editor should know the aver-
age level of the intended audience, and adapt the quantity, extensiveness, and
degree of detail of his notes to that level. Although, in the case of Xenophon,
such a uniform starting point will have been rather exceptional, there was one
certainty: almost always the Anabasis was the first Greek text to be read in the
gymnasium and its look-alikes elsewhere, and the im- or explicit level of the
edition should reckon with this fact. This issue will be discussed presently. Sec-
ond, the notes should do two jobs at the same time: they should give immediate
help for a local problem, thereby enabling the pupil to go on; and they should
anticipate, if applicable, the presence of the same or a similar problem later
in the text, and add a remark to this effect, or—perhaps more conveniently—
refer back to the earlier explanation upon arrival at the second (third, etc.)
occurrence of the problem, or—still more conveniently—they should repeat
252 the same explanation. I have not examined the editions | systematically on this
score, but generally speaking I have the impression that they meet this require-
ment.
For the sake of completeness I mention yet another, and far more intricate,
issue, that sooner or later confronts every commentator: what to do with a phe-
nomenon for which no real explanation exists, either in grammars or in other
commentaries, whether on the Anabasis or on other texts? It goes without say-
ing that a pupil may feel slightly frustrated if time and again his attention is

53 I am aware that the school edition as described above no longer exists, at least not for sec-
ondary schools. From recent times I have found (for Xenophon) only one descendant of
the traditional school edition (Antrich—Usher), but this is meant for beginning students.
The demands made here upon a school edition are, therefore, made retrospectively. See
more generally on audience requirements Kraus (2002: 8–9); on the anticipation of prob-
lems see the Index of Gibson and Kraus (2002), s.v. surprise.
the xenophon factory 393

drawn to some peculiarity, only to be told that ‘there is no really satisfying

explanation for this phenomenon.’ Should it be ignored, then? An attentive
pupil might very well realize that there is something the matter, and ask a ques-
tion to that effect. Whereupon the teacher—unless he has found the answer
himself—is forced to reveal that there is as yet no answer to the question. Since
for a ‘cursory’ interpretation such phenomena often are not really a stumbling
block, this may lead to the wholly mistaken idea that they do not have to be
commented upon. On the very first page of the Anabasis there are at least three
such phenomena54 (line references are to the OCT edition):
(a) the occurrence of παῖδες δύο (1.1) alongside the dual τὼ παῖδε (1.3)
(b) Κῦρον without article (1.5) alongside ὁ Κῦρος (1.7)
(c) the oblique optative ἐπιβουλεύοι in 1.12 alongside tense and mood of direct
speech, as in ἔσται in 1.15.
All three phenomena are by no means rare; on the contrary, they are wide-
spread throughout the Anabasis (not to mention other texts). In all three cases
a generally accepted explanation is lacking, although for (b) and (c) some pro-
posals may be found in a few grammars and special studies.55 As for our school
editions, only παῖδες δύο is sometimes provided with a note, which generally
has no value at all, except for signaling the phenomenon.56 The presence or
absence of | the article with Κῦρος (and indeed with other proper names), and 253

54 Still seen from the perspective of a nineteenth-century editor. Ideally, a modern variant
of the school edition should also deal with word order phenomena like ἠσθένει Δαρεῖος
(§ 1), ἀναβαίνει … ὁ Κῦρος and ἐτελεύτησε Δαρεῖος (§ 3), on the one hand, and Τισσαφέρνης
διαβάλλει and ἡ δὲ μήτηρ … ἀποπέμπει (§ 4), on the other.
55 E.g., for (b) in Zucker (1899); in my inaugural lecture (‘On Definite Persons,’ Rijksbaron
2001) more may be found on the article with proper names. [See also this volume, chap-
ter 16.] For (c) see K-G 2, 555, Anm. 3 and Amigues (1977: 270).
56 Thus, Rehdantz-Carnuth: ‘παῖδες δύο: aber 4, 1, 19 der Dualis: δύο καλώ τε κἀγαθὼ ἄνδρε
τεθνᾶτον’; Goodwin-White: ‘τὼ παῖδε: dual; but above, the plural’; Mather-Hewitt: ‘παῖδες
δύο: a plural noun is often used with δύο.’ Couvreur (‘Δύο rend inutile l’emploi du duel du
substantif’) is simply wrong, since δύο is also found with the dual, cf. the instance quoted
by Rehdantz. Only Walpole has at first sight an interesting observation; on closer inspec-
tion, however, it does not hold. He notes: ‘The pl. is used w. δύο because no stress is laid on
the fact of there being only two; there had been thirteen.’ An investigation carried out by
Guus van der Kraan during the seminar mentioned in the first note in this chapter, how-
ever, led to the conclusion that, in the Anabasis, δύο in the nominative and the accusative
overwhelmingly prefers a plural noun (30 times plural, as against two cases of the dual,
4.1.19, 4.3.10.) So it is the dual rather than the plural that has to be explained. Remarkably
enough, this also applies to τὼ παῖδε ἀμφοτέρω: 1.1.1 is the only instance in the Anabasis of
ἀμφοτερ- appearing in the dual, alongside seven instances of the plural. Perhaps the com-
bination of the dual and ἀμφοτέρω emphasizes the idea that Artaxerxes and Cyrus had to
be present as a pair, so: ‘He wanted both his sons to be present together.’
 394 chapter 20

the possible differences between the use, in oratio obliqua, of the oblique opta-
tive and the tense and mood of direct speech, are nearly always passed over in
silence. (There is just one exception: Parker 1882.)
Now, to return to the question of the ‘average level of the intended audience,’
it appears that, in spite of the fact that all commentators wrote for beginners,
in actual practice their opinions differ widely as to what is a grammatical prob-
lem, and should therefore receive a note, and what is not a problem. There are
remarkably few things which worry all commentators. Below, I give first of all a
survey of phenomena from An. 1.1–4 that are de facto treated in six editions that
I consider representative of the whole group; then follows a more detailed dis-
cussion of the notes that all commentators have on the first two constituents
(Δαρείου καὶ Παρυσάτιδος and γίγνονται).

Frequency of the Notes in Four Popular Nineteenth-Century (Krüger,

Vollbrecht, Rehdantz, Walpole) and Two Twentieth-Century (Bent,
Antrich–Usher) Editions of Anabasis 1.1–4 (up to and Including ἀντ’
ἐκείνου)
Altogether, some 49 words and constructions, which I shall not enumerate here,
are provided with a note. In only nine cases do all six editions mentioned have
a note, viz. on:57 §1 γίγνονται, τελευτὴν τοῦ βίου; § 2 ἐποίησε, καὶ … δέ, ὡς φίλον,
τῶν Ἑλλήνων; §3 ὡς ἀποκτενῶν, ἐξαιτησαμένη; §4 ὅπως ἔσται. In seven cases five
editions have a note (on §1 Δαρείου; §2 παρὼν ἐτύγχανε, λαβών; § 3 διαβάλλει
254 ὡς, | ἐπιβουλεύοι, ἀποπέμπει; §4 [ὁ δ᾽] ὡς). In four cases four editions have a
note (on §2 μεταπέμπεται, μὲν … δέ; §3 κατέστη, αὐτόν [with ἐξαιτησάμενη]). In
the remaining 29 cases the number of editions that have something to say on a
given problem lies between one and three.
Leaving numbers aside, even if the commentators see the same problem,
the nature of their notes shows the most extraordinary variation. By way of
‘warming up’ for the detailed treatment of Δαρείου καὶ Παρυσάτιδος and γίγνον-
ται below, I shall now discuss the way in which the six selected commentaries
deal with three other phenomena in the passage in question.
(a) §2 καὶ … δέ. A note here is more or less de rigueur, since this combina-
tion of particles occurs about 265 times in Xenophon, and all six com-
mentaries do, indeed, have a note. Most of them confine themselves to
observing, correctly, that the word between καί and δέ is emphasized, and
do not commit themselves as to the syntactic function of the particles

57 Below, in the discussion of Δαρείου καὶ Παρυσάτιδος and γίγνονται, it will become clear that
when the other editions are taken into account, the picture changes again. Thus, Hertlein
has a note neither on Δαρείου καὶ Π., nor on γίγνονται.
the xenophon factory 395

(Krüger, Vollbrecht, Rehdantz, Bent). Only Walpole and Antrich–Usher

have a note on this function. The latter take it that δέ is the connec-
tive particle, while Walpole thinks it is καί (‘and moreover’), although he
leaves the possibility open that it is the other way round. To my mind only
Walpole’s analysis—which, incidentally, can also be found in Kühner-
Gerth—is correct.58
(b) §3 ὡς ἀποκτενῶν. Four commentaries note that there is a difference be-
tween the future participle with and without ὡς. Only Bent and Antrich–
Usher believe, quite wrongly, that the absence or presence of ὡς is irrel-
evant: ‘het part, fut., met of zonder ὡς, heeft finale betekenis’ (Bent); ‘ὡς
+ fut. part., expressing purpose’ (Antrich–Usher). This carelesness is puz-
zling: why did they not just take over Krüger’s note, who in this case has
by far the clearest wording (‘ἀποκτενῶν allein würde den Zweck als bloss
erzählt bezeichnen; mit ὡς wird er als Absicht, als Gedanke des Artaxerxes
ausgesprochen’)?59
(c) §3 ἐξαιτησαμένη. Krüger, Vollbrecht, Rehdantz, and Walpole rightly point
to the value of the middle, ‘Medium des Interesses,’ in | Krüger’s words, 255
which is appropriate here, since Parysatis is on Cyrus’ side all the time.
Again, Bent and Antrich–Usher take a different position, since they are
silent about the middle; again, this is puzzling. By ignoring the difference
between αἰτεῖν and αἰτεῖσθαι, they not only miss the opportunity to add a
subtle nuance to the interpretation of the text, but may also induce the
students to believe that active and middle are six of one, half a dozen of
the other.
Finally, I pass on to discussing the explanatory force of the notes on the first
two constituents of §1, Δαρείου καὶ Παρυσάτιδος and γίγνονται.

The Various Explanations in Chronological Order

(a) Δαρείου καὶ Παρυσάτιδος

Just a reference to a grammar in

– Cleveland 1834 (‘See F[isk] Rule XVII’)
– Owen 1867 (to the grammar of Sophocles)60

58 See Rijksbaron (1997) [= chapter 15 in this volume].

59 ‘ἀποκτενῶν alone would present the purpose as simply narrated; with ὡς the objective is
formulated as a thought of Artaxerxes.’
60 Most of his notes are phrased as a question, of the type ‘What is this mid. voice equivalent
to?,’ followed by references to one or more grammars.
 396 chapter 20

Merely a translation in
– Hardy 1838 (‘Of-Darius and of-Parysatis are-born children two’)61
– Macmichael 1860 (‘To Darius … are born two sons. Literally, “are born of ” ’)
– Anthon 1868 (‘Of Darius and Parysatis are born two sons’)
– Antrich–Usher ca. 1980 (‘Dareios and Parysatis had’)

We are dealing here with a genitive of origin, according to

– Boise 1857 (‘lit. there are born of Darius etc … gen. of origin or author’ + ref-
erence to Kühner’s grammar)
– Vollbrecht 1877, 1907 (‘Gen. des Urspr.’). No translation
– Parker 1882 (‘Goodwin Gr. 169 calls this a genitive of possession and trans-
lates ‘are born (belonging to) Darius.’ It is rather a genitive of source, origin,
256 which is a kind of gen. of ablation.’) |
– Rehdantz–Carnuth 1888 (‘Genitiv des Ursprungs’). No translation
– Kelsey–Zenos 1892, but disguised as a question (‘why gen.?’) + ref. to Good-
win’s Grammar (‘gen. of source’); in 1895 the question has disappeared; in
both cases no translation
– Couvreur 1896 (‘génit. de provenance’). No translation
– Edwards (1901) (‘“are the children of Darius and Parysatis,”—genitive of ori-
gin’)
– Bersi 1905 (‘Questo genitivo è detto d’ origine’). No translation
– Mather-Hewitt ca. 1912 (‘genitive of source’). No translation

… or rather with a genitivus possessivus, in the opinion of

– Krüger 1854 (‘eine Art des Besitzes: sie wurden (Söhne) des D.’)
– Krüger–Pökel 1889 (‘eine Art des Besitzes: sie wurden (Söhne) des D.’)
– Goodwin–White 1889 (‘predicative genitive of possession’). No translation
– Sorof 1929 (‘Gen. poss., deutsch: dem Dareios … werden (wurden) geboren,
eig. zwei dem Dar. und der Par. gehörende Söhne werden geboren’)
– Bent 1963 (‘gen. possessivus om de oorsprong aan te geven’).62 No translation

61 Hardy’s book contains a ‘literal and interlinear translation’ of chapters 1–6 of the first book,
and a vocabulary; there are no notes.
62 ‘Possessive gen. to indicate the origin.’ He refers to §116.1 of the grammar of Van Oppen-
raaij—Vermeulen. There one reads: ‘Possessivus: geeft de bezitter of oorsprong aan; dik-
wijls in verbinding met εἰμί en γίγνομαι’ (‘possessivus: indicates the possessor or origin;
often in connection with εἰμί and γίγνομαι’), followed by one (artificial) example with ἐστιν:
τὸ φεύγειν τῶν δειλῶν ἐστιν. Actually, however, An. 1.1.1 is mentioned in §117.1, as an instance
of the genit. originis. In reprint after reprint the incorrect reference has been repeated.
the xenophon factory 397

ad hoc explanations are given by

– Krüger 1845 (‘Der Ge. steht als ob die Eigennamen gleich folgten’).63 No
translation
– Breitenbach 1865 (‘Die Genitive … hängen von παῖδες δύο ab’). No translation

No note at all is given by

– Hertlein 1857 | 257
– Matthiä 1859 (just a reference to the grammatical ‘Anhang,’ which I have not
been able to consult)
– Kühner 1852 (but implicitly a genitive of origin)
– Pretor 1877
– Walpole 1882
– Brownrigg 1902

(b) γίγνονται

Just a translation is given by

– Breitenbach 1865 (‘es werden geboren’)

It is taken as an historic present

A. Without further comment by

– Brownrigg 1902 (‘the verb is historic present’)
– Couvreur 1896 (‘présent dit historique’ + reference to the grammar of Croiset
and Petitjean, which I have not been able to find)

B. With further details by

(i) ‘occurs more often in Greek than in …’
– Boise 1857 (‘which is more common in Greek than in Latin’) + a reference to
the [school?] grammar of Kühner
– Sorof 1929 (‘welches im Griech. häufiger als im Deutschen gebraucht wird’)
– Bent 1963 (‘Het praes. hist. komt in het Gr. veel meer voor dan in het Lat.,
waardoor de betekenis is afgesleten en het weinig of geen meer nadruk heeft
dan aor. of impf.’)

(ii) ‘lebhaft,’ ‘vivid’

– Krüger 1845 (‘hist. Pr.’ + a reference to his own grammar (p. 164 of the

63 ‘The ge. stands as if the proper names followed immediately.’

 398 chapter 20

1873 ed.), where one reads: ‘Mit Lebhaftigkeit macht der Geist Vergangenes
zu ideeller Anschauung der Gegenwart; und so steht das sogenannte his-
258 torische Präsens im Griechischen viel häufiger als im Deutschen’)64 |
– Krüger 1854 (‘Das historische Präsens gebraucht der lebhafte Grieche [in
Hauptsätzen] viel häufiger als wir und selbst als der Lateiner’)65
– Macmichael 1860 (‘To give liveliness to a narrative, the historian [in Greek
as in other languages] will often conceive and speak of past events as if he
had them before him in course of action’)
– Anthon 1868 (‘animation to a narrative … more vividly before the mind’)
– Krüger–Pökel 1889 (‘Das historische Präsens …’ etc., as in 1854)
– Goodwin–White 1889 (‘were born’ + a reference to Goodwin’s grammar,
where the hist. present is probably considered ‘vivid’, just as in the 2nd ed.
(see next item); the first ed. I could not consult)
– Kelsey–Zenos 1892, 1895 (‘historical present’ + reference to Goodwin’s gram-
mar, §1252, ‘used vividly for the aorist’)
– Antrich–Usher ca. 1980 (‘“Dareios and Parysatis had”; pres. tense for lively
narrative as the story begins’)

(iii) ‘dauernd gültig,’ ‘permanently valid’

– Kühner 1852 (‘Praesens saepius usurpatur, ubi actio verbi e praeterito tem-
pore pertingit in praesens (entsprossen sein und abstammen)’) He compares
the tragic use of τίκτειν, γεννᾶν, θνήσκειν (todt sein); also ἀδικεῖν, νικᾶν. ‘Inuria
γίγνονται h. 1. pro praesenti historico habetur, cuius usus ab hoc quidem loco
alienissimus est.’66
– Vollbrecht 1877 (‘das Präs. von einem Ereignisse, das für die Geschichte
dauernd gültig ist’)
– Parker 1882 (‘“are born.” The historical present is often used with verbs of
birth, the relationship being looked on as permanent.’)
– Edwards 1901 (‘For the idiomatic use of the present tense representing the
259 effects of a past event lasting into present time | cf. note on § 7,67 and Pro-
pertius v.i.77 me creat “he is my father.”’)
– Vollbrecht 1907 (‘das Präs. von einem Ereignisse, das dauernd gültig ist’)

64 ‘With vividness the mind turns the past into an imaginary view of the present, and there-
fore the so-called historic present is much more common in Greek than in German.’
65 ‘The lively Greek uses the historic present (in main clauses) far more often than we and
even than the Roman.’
66 ‘The present is rather often used when the verbal action extends from the past into the
present time’; ‘It is not correct to take γίγνονται here as a historic present, whose use is
totally different from that in our passage.’
67 Where we are directed to § 1: ‘[w]ith the idiomatic φεύγω “I am in exile” cf. note on §1.’
the xenophon factory 399

– Bersi 1905 (‘[the pr.] ha talvolta un valore tutto speciale, perché indica
un’azione che principiata in passato si estende coi sui effetti al presente’)68

(iv) cannot choose between (ii) and (iii)

– Pretor 1877 (‘This is commonly regarded as an instance of the historic pres-
ent, by which additional reality is given to the narrative of a past event.
Kühner however rejects this theory, and explains the tense by the fact that
the action of the verb is represented as continuing down to the present time’)

(v) genealogical present

– Rehdantz–Carnuth 1888 (‘“stammen”; genealogisches Präsens’)
– Walpole 1882 (‘were born. Rehdantz calls this a genealogical present’)

(vi) ‘annalistic’ present

– Mather–Hewitt ca. 1912 (‘historical present, here better called the annalistic
or notebook present, which is used in diaries or notebooks to record inci-
dents, especially births, deaths, and accessions’)

No note at all is given by:

– Cleveland 1834
– Hertlein 1857
– Matthiä 1859
– Owen 1867

Comments
One of the enduring charms of our profession is that there is, just as with the
humanities in general, much room for divergent opinions. | But I submit that 260
the reader will agree with me when I, after surveying this inventory, conclude
that in our case the diversity is positively bewildering and, in fact, rather alarm-
ing, especially in the case of γίγνονται. How can it be that there is so little
consensus of opinion about the interpretation of the first sentence of one of
the most widely read classical works? And how is it possible that many com-
mentators see no problems where their colleagues do see them? I will not go
into all the proposals, but try rather to create some order in this muddle by
making clear what is, in my opinion, the correct explanation for each of these

68 ‘sometimes has quite a special value, since it expresses an action that has started in the
past but extends into the present together with its effects.’ Note the similarity to Kühner’s
words.
 400 chapter 20

items, an explanation that—and this is perhaps an encouraging thought—is

included in the above notes.

a Δαρείου καὶ Παρυσάτιδος

No real explanation of the genitive is given by those commentators who com-
bine, either in a translation or in a note, a dative with a genitive interpretation;
these are Macmichael and Sorof, and, implicitly, also Antrich–Usher. These
commentators apparently started from the fact that usually the getting or hav-
ing of children is expressed by the dative; since the latter, however, unfortu-
nately is not there, they take refuge in the clincher ‘ “literally” the text runs …
but we should translate as if …,’ as is done by Macmichael, or translate with
a dative (Sorof), or eschew the problem by translating ‘Darius and Parysatis
had’ (Antrich–Usher)—which, incidentally, is the current translation of εἰμί +
dative, and clarifies nothing. On top of this, Sorof’s translation is simply impos-
sible: as if a story could begin (I translate in my turn): ‘Two sons belonging to
Darius and Parysatis were born’! This nonsense is also due to the fact that Sorof
completely ignores the word order of the Greek text, something that more often
than not results in unsatisfactory interpretations.
The possessive genitive of Krüger 1854 and 1889, and, presumably inspired
by him, of Goodwin–White, cannot be defended either. Krüger’s reasoning is so
bizarre and unclear that I shall discuss it in some detail in an Appendix, which
includes Krüger 1845. Bent is alone in his belief that the possessive genitive is
used to express the origin.
All this means that of the explanations proposed only the one that sees the
genitive as a genitive of origin (or source) is tenable. This view can be supported
261 by a number of parallels, that in part are | also adduced by Krüger, although he
believes that they support his own view (for these see the Appendix).

b γίγνονται
The discussion of just this present could easily fill a full-scale article. The com-
mentators mentioned at B(i) above have thrown in the towel straightaway: they
dish up some trivia that do not further our knowledge of the Greek historic
present one bit. As for Bent, moreover, who claims that the meaning of the
historic present had completely worn off, he fails to tell us what that original
meaning was.
The commentators at B(ii) confront us with a particularly tough fellow: the
‘vivid’ historic present. The assumptions behind Krüger’s note are especially
worth noticing. In the 1845 edition we are told that ‘with vividness the mind
turns the past into an imaginary view of the present, and therefore the so-called
historic present is much more common in Greek than in German.’ In the next
the xenophon factory 401

editions the implication of ‘therefore’ has been made explicit: now ‘the’ Greek
is simply ‘lively.’ Observe that now the Romans, too, get a scolding. Yes, perhaps
they are lively when compared with the Germans, but still not lively enough!
What all champions of the vivid present overlook—or, worse, ignore—is the
fact that the first four paragraphs of chapter 1 contain no fewer than seven other
historic presents besides γίγνονται, viz., μεταπέμπεται, ἀναβαίνει, διαβάλλει, πεί-
θεται, συλλαμβάνει, ἀποπέμπει, βουλεύεται. Naturally, if the commentators were
true to their views, these ought to be as many signs of vividness, but most of
them wisely keep silent on this score.69 So much vividness in such a short pas-
sage is perhaps a bit too much, even for the ‘lebhafte Grieche.’ Only Macmichael
and Antrich–Usher add that μεταπέμπεται is ‘used for vividness’ as well, but
they too fail to tell us how we should take ἀναβαίνει and the rest. It is precisely
because of the high concentration of historic presents here that a commenta-
tor should be very cautious and should not snatch too quickly at ‘vividness’ to
explain the use of γίγνονται, etc. | 262
The group at B(iii), headed by Kühner, takes a totally different view. However,
in spite of the fact that Kühner in the most vigorous terms denies that γίγνονται
is a historic present, Kühner-Gerth do treat it as such,70 and more specifically
as a ‘chronicle-like historic present.’ Kühner’s original view is untenable, since
the use postulated by him in this passage, which certainly does exist, belongs to
direct speech, and is especially frequent in tragedy. This usage, too, is discussed
extensively in Kühner-Gerth.71 As for Vollbrecht, observe that he does not speak
about the present being the reference point, but hesitates between ‘perma-
nently valid for the story’ (1877) and ‘permanently valid’ tout court. In spite of
this difference, both interpretations look like variants of that of Kühner. If he
means, in fact, the same as Kühner, the objections raised to the latter’s interpre-
tation of course apply to that of Vollbrecht as well. Of the other presents in our
passage only ἀναβαίνει gets a note, both in Kühner and Vollbrecht; they take it
as ‘vivid.’ On the other presents they remain silent. As for Edwards he, too, has a
note only on ἀναβαίνει (‘historic present’), adding that it must be distinguished
‘from the use of the present commented upon in § 1.’
B(v). On the ‘genealogical present’ I can be very brief. Unless this term
covers the same interpretation as that of Mather-Hewitt (see next category),

69 Incidentally, with the sole exception of Walpole, no commentator has a note on ἁθροίζον-
ται. Since this use of the tense, being a generic/habitual present, is totally different from
the other eight presents, a note definitely seems appropriate here.
70 (1898–1904) 1.134. This must be due to Gerth, for in the second edition of Kühner’s gram-
mar (1870) γίγνονται at An. 1.1.1 is not treated as such. In fact, it is not discussed at all.
71 (1898–1904) 1.145 ff.
 402 chapter 20

Rehdantz’s category seems to have been created for the occasion. Rehdantz has
not a single note on the other presents.
B(vi). From the above it will have become clear that the interpretation
as ‘annalistic’ present, which strictly speaking is also given by Kühner-Gerth
(‘chronicle-like historic present’),72 is to my mind basically correct, even though
the term ‘annalistic’ does not tell us much about the function of γίγνονται (see
below). Mather–Hewitt also have a note on μεταπέμπεται, which they take as a
‘genuine historical present,’ without indicating what such a ‘genuine’ historical
present is. On ἀναβαίνει and διαβάλλει they have something similar. Actually, all
eight historic presents here have the same function. Far from expressing ‘vivid-
ness,’ they present, in a small compass, those events of Cyrus’ life and career
263 that are of crucial importance for | the story proper. In other words, the passage
1.1–1.4 βουλεύεται is not part of the story but only the introduction to the story.
This use might perhaps be called ‘annalistic,’ provided that this term covers
the function just mentioned. The story proper begins only with the imperfects
ὑπῆρχε and ἀπεπέμπετο of §§4–5, of which the former gives the general back-
ground of Cyrus’ actions, and the latter the first real action on the part of Cyrus
following his βουλεύεσθαι. In this story, too, historic presents occur, first with § 9
δίδωσιν. Basically, this present has the same function as those of §§ 1–4: δίδωσιν
conveys the idea that the giving in question was an important event. The effect
achieved by this present, however, is very different, because it is followed by a
passage which tells us why this event was so important.73

7 Conclusion

I quoted above (p. [239]) a rather unflattering statement by Gildersleeve con-

cerning school editions. Although the possibility cannot be ruled out that the
above discussion presents too negative a picture of the editions consulted
by me, it is to be feared that Gildersleeve’s harsh judgment is all too justi-

72 Also by S-D 2, 272.

73 For a recent discussion of the historic present see Sicking and Stork (1997: 131–169); on the
historic present in Xenophon see especially 147–156. Summarizing their analyses, Sicking
and Stork observe that Xenophon uses the hist. pres, ‘for organizing his narrative,’ since
it distinguishes ‘matter that relates to what is the writer’s main concern’ from the other
elements of the story (156). This view may already be found in earlier studies, notably
Eriksson (1943), especially 11, on our passage: ‘Ein historisches Präsens [ist] in der Lage,
ohne vorbereitende Präterita abrupt die benötigte sachliche Unterlage einer Erzählung
mitzuteilen’ (‘A historic present may mention abruptly, without any preparatory past
tenses, the factual foundation of a narrative’).
the xenophon factory 403

fied. Actually, as the preceding discussion may have illustrated, the situation
is even worse: not only are the editions ‘almost negligible for the advanced
student,’ for the beginner, too, they virtually all show serious shortcomings,
either for Δαρείου καὶ Παρυσάτιδος or for γίγνονται, or for both constituents,
the most widespread one being the failure to explain γίγνονται (and the other
historic presents of 1.1–4) in a satisfactory way.74 As for the ‘original | contribu- 264
tion’ demanded by Gildersleeve: if only the commentators, especially the more
recent ones, of course, had done some research, or at least consulted the major
grammars75 and the specialized studies, instead of repeating an idée reçue for
the umpteenth time!

Appendix: Krüger on Δαρείου καὶ Παρυσάτιδος76

2Krüger 1845
Δ. καὶ Π.: “Man sagt παῖδές εἰσι, γίγνονταί τινι; der Ge. steht als ob die Eigennamen
gleich folgten: Δαρείου καὶ Π. γίγνονται Ἀρταξέρξης καὶ Κ. vgl. Gr. 47, 6, A. 5.” This,
the very first note, is far from clear. Krüger seems to be implying that in view of
the presence of παῖδες we should expect the dative; the genitive, however, has
apparently been preferred to the dative because of the proper names, although
these strictly speaking ought to follow directly after the genitive, i.e., without
παῖδες. Since we are referred to his grammar,77 we may expect to find more
information there on the relationship between genitive and proper names. In
the relevant section, the ‘Ge. der Angehörigkeit’ with a ‘Prädikatsverbum’ is
discussed; a subcategory of this genitive is the ‘Gen. der Geschlecht und Geburt-
sort bezeichnet.’ I give all his examples from prose, and maintain his way of
quoting the Greek text, which is, as will become clear presently, extremely mis-
leading: Πατρὸς τίνος ἐστὶ καὶ μητρὸς ὁ Ἔρως … [read: Πατρὸς τίνος ἐστὶ καὶ μητρός

74 Of the editions discussed, that by Mather—Hewitt has to my mind the best notes both on
Δαρείου καὶ Παρυσάτιδος and on γίγνονται; in fact, it is the only one with a reliable note on
both constituents. This edition, too, however, fails to discuss the historic presents of An.
1.1–4 together, and shows some idiosyncracies in other notes, e.g., that on τελευτὴν τοῦ βίου
in 1.1.1 (‘an end of his life’).
75 Clear clues for the correct interpretation of both the genitive and the hist. present can be
found in K-G.
76 I have added this rather detailed treatment of Krüger’s views to make it clear that the high
esteem accorded both to his Anabasis editions (cf. above, n. 49) and to his grammar, is not,
to put it mildly, in all respects justified. I quote the text of Krüger’s notes in their original,
often painfully succinct, form.
77 I made use of the fifth edition (1873).
 404 chapter 20

… (sc. ὁ Ἔρως, to be supplied from the context)]. Πλ. συ. 203, a. ἀγαθῶν [read:
τοιούτων] ἐστε προγόνων. Ξε. ἀν. 3, 2, 14. Παυσανίας γένους τοῦ βασιλείου ἦν [read:
ἄνδρα γένους τοῦ βασιλείου ὄντα], Θ. 1, 132, 1. οὐδὲν θαυμαστὸν τῶν ἀγαθῶν πατέρων
φαύλους υἱεῖς γίγνεσθαι. Πλ. Πρω. 328 [sic; = 328c]. Θουκυδίδης οἰκίας μεγάλης ἦν
[read: οἰκίας μεγάλης ἦν (sc. ὁ Θουκυδίδης)]. Πλ. Μένων 94 [sic; = 94d], Ξενόφων
πόλεως μεγάλης ἦν [read: σὺ πόλεως μεγίστης εἶ!]. Ξε. ἀν. 7, 3, 19. Τηρεὺς καὶ Τήρης
265 οὐ τῆς αὐτῆς Θρᾴκης ἐγένοντο. [Here, I refrain from giving the actual | Greek text,
which is completely different.] Θ. 2, 29, 2 [= 3]. Notice that not only is there in
the actual Greek of the passages quoted by Krüger not a single proper name,
but even in the form in which he quotes them there is just one genitive which
is followed by a proper name (Pl. Smp. 203a). The value of the reference ad An.
1.1 “vgl. Gr. 47, 6, A. 5” is, thus, nil. Surprisingly, Krüger does not mention two
almost exact parallels to our sentence, Hdt. 1.102.1 Δηιόκεω δὲ παῖς γίνεται Φραόρ-
της and 4.160.1 Τούτου δὲ τοῦ Βάττου παῖς γίνεται Ἀρκεσίλεως—perhaps because
here, just as at An. 1.1.1, παῖς is present?
There remains the intriguing question as to why Krüger does not consider
taking Δαρείου καὶ Π. as a genitive of origin/source. The reason must be that,
according to K. (in the same note), this genitive is only found in combination
with ἐξ or ἀπό. However, the strict distinction made by him between ἐξ and ἀπό
+ genitive (“Ἐξ oder ἀπό hinzugefügt bezeichnen Geburt oder Abstammung”)
and the simple genitive (which designates “Geschlecht und Geburtsort”) seems
artificial and far-fetched—what, for instance, may be the difference between
‘Abstammung’ and ‘Geschlecht’?—and leads in actual practice, as his analysis
of An. 1.1.1 shows, to unacceptable solutions.
In the next editions of Krüger things have not exactly improved.

4Krüger 1854
Δ. καὶ Π.: Here, the first part of the note is identical with that of 1845. Now,
however, after “Ἀρταξέρξης καὶ K.,” Krüger adds: “Dieser Ge. bezeichnet eine Art
des Besitzes: sie wurden (Söhne) des D. Ar. Ach. 47: Ἀμφίθεος Δήμητρος ἦν καὶ
Τριπτολέμου· τούτου δὲ Κελεὸς γίγνεται”. This is followed by the reference to his
grammar. This is not a real improvement. First, Krüger blurs the issue by speak-
ing about “a kind of possession.” Second, his analysis is faulty on two counts: (a)
to whom exactly does the ‘sie’ refer, in “sie wurden (Söhne) des D.”? Apart from
παῖδες δύο there simply is no other candidate for the subject function; (b) how
can the ‘sie,’ whoever they may be, become somebody’s sons, as if they change
fathers?78 Third, the line from Acharnenses is a rather dubious parallel, since

78 Except of course by adoption. Indeed, γίγνομαι + gen. can be used to express change of
the xenophon factory 405

the text is probably corrupt. And even if the text is correct, only the second
part | illustrates his point, for in the first part the proper name precedes rather 266
than follows the genitive(s).

7Krüger–Pökel 1889
Δ. καὶ Π.: the note is identical with that of 1854, but there are two additions.
“Ἀρταξέρξης καὶ K.” is followed by “vgl. Her. 3, 160,” and the reference to the
grammar is followed by a second reference, now to “Dial. 48,3,2”; this concerns
the second part, in which Krüger presents a grammar of the dialects, especially
the epic and Ionic dialects. The passage cited from Hdt. (3.160.2) has: Ζωπύρου
δὲ τούτου γίνεται Μεγάβυξος, and somewhat further, Μεγαβύξου δὲ τούτου γίνε-
ται Ζώπυρος. These sentences do, at last, illustrate his original contention, viz.,
that in constructions with a genitive, proper names follow the genitive. (Recall,
however, that according to Krüger in all these cases we are dealing with pos-
sessive genitives.) As for the reference to the dialect part of the grammar: there,
εἰμί + dative is discussed, with a brief additional remark about the genitive with
πατήρ, which is, therefore, irrelevant for our sentence at X. An. 1.1.1.
All in all, Krüger’s analysis of An. 1.1.1 is needlessly complicated (and initially
supported by evidence that has been tampered with).79 As the passages men-
tioned above show, there are two—or (if the line from Acharnenses is correct)
three—different ways to express descent using γίγνομαι/εἰμί and the simple
genitive:
– genitive of the parent—γίγνομαι—name of the child (Hdt. 3.160.2)
– name of the child—gen. of the parent—εἰμί—gen. of the other parent (Ar.
Ach. 47)
– genitive of the parent—παῖς—γίγνομαι—name of the child (Hdt. 1.102.1,
4.160.1)
In view of these constructions an explanation of Δαρείου καὶ Παρυσάτιδος, in
Δαρείου καὶ Παρυσάτιδος ~ γίγνονται παῖδες δύο, as a genitive of origin seems
unobjectionable, presenting us with a fourth variant:
– genitive of the parent—γίγνομαι—παῖς—name of the child. | 267
For a balanced discussion of the simple genitive and the genitive with ἀπό and
ἐξ see also K-G 1, 374, Anm. 3.

ownership, e.g., at Th. 5.5.1 ἐγένετο Μεσσήνη Λοκρῶν τινὰ χρόνον, ‘For some time Messene
came into the possession of the Locrians.’
79 Fortunately, in his revised edition of Krüger’s grammar, G.W. Cooper has completely aban-
doned Krüger’s views ([1998] 173). He takes Δαρείου καὶ Π. as an ablatival genitive.
406 chapter 20

Works Cited

Editions of the Anabasis referred to in the tables above are not included here.

Amigues, S., Les subordonées finales par ὅπως en attique classique (Paris 1977).
Cooper, G.W., Attic Greek Prose Syntax, after K.W. Krüger (Ann Arbor 1998).
Eriksson, K., Das Praesens historicum in der nachklassischen griechischen Historiogra-
phie (Lund 1943).
Garzya, A., Omaggio a B.G. Teubner (Napoli 1983).
Gibson, R.K., Kraus, C.S. (eds), The Classical Commentary: Histories, Practices, Theory
(Leiden 2002).
Gildersleeve, B.L., ‘English and German Scholarship,’ in C.W.E. Miller (ed.), Selections
from the Brief Mention of Basil Lanneau Gildersleeve (Baltimore/London 1930) 364–
376 [Orig. published 1917].
Henderson, J., ‘The Way We Were: R.G. Austin, In Caelianam’, in Gibson and Kraus 2002,
205–234.
von Humboldt, W., Der königsberger und der litauische Schulplan (Berlin 1809), also
published in W. von Humboldt, Bildung des Menschen in Schule und Üniversität, ed.
K. Püllen. (Heidelberg 1964).
Kirchner, C., Die Landesschule Pforta in ihrer geschichtlichen Entwicklung: seit dem
Anfange des XIX. Jahrhundert bis auf die Gegenwart; mit einem Grundriss von Pforta:
Einladungsschrift zur dritten Saecularfeier ihrer Stiftung den 21. Mai 1843 (Naumburg
1843).
Kraus, C.S., ‘Introduction: Reading Commentaries/Commentaries as Reading’, Gibson
and Kraus 2002, 1–28.
Krüger, K.W., Griechische Sprachlehre für Schulen, zweites Heft: Syntax, fünfte Auflage
(Leipzig 1873)
Rijksbaron, A., ‘Adverb or Connector? The Case of καὶ … δέ,’ in A. Rijksbaron (ed.), New
Approaches to Greek Particles (Amsterdam 1997) 187–208 [ch. 15 in this volume].
Rijksbaron., Over bepaalde personen. Amsterdam 2001.
Sicking, C.M.J., P. Stork., ‘The Grammar of the So-Called Historical Present,’ in E.J. Bak-
ker (ed.), Grammar as Interpretation: Greek Literature in its Linguistic Contexts (Lei-
den 1997) 131–169.
Stephens, S., ‘Commenting on Fragments’, in Gibson and Kraus 2002, 67–88.
von Wilamowitz-Möllendorf, U., Erinnerungen 1848–1914 (Leipzig 1928).
Zucker, A., Beobachtungen über den Gebrauch des Artikels bei Personennamen in Xeno-
phons Anabasis (Nürnberg 1899).
Index Rerum

absolute tense 63, 65–68 relative tense 16–18

see also tense semelfactive 22, 23n, 83, 91, 242, 263
accentuation telicity 127, 129, 130, 131
shifts 352–353 verba dicendi 24–29
value of mss evidence 108–109n argument (syntactical role) 229n, 281–
see also Index Graecitatis 282
accomplishment: see Aktionsart article
accusative anaphoric 324, 333
of duration of time 289–290 forming noun phrases 279–280
achievement: see Aktionsart marker of definiteness 283, 285n, 291
adverb position in noun phrases with οὗτος
affirmative 339, 343–345 223–236
evaluative 139 with proper names 319–333, 393–394
intensive 226 τό with temporal adverbs 279–292
locative 116, 118 aspect 5–9
manner 116 see also aorist stem, perfect stem, present
vs. accusative noun phrases 289– stem
290 asyndeton 104, 224–230, 235, 298n
vs. connector 293–317 augment
see also Index Graecitatis unaugmented past tenses 107–108, 162
agent: see semantic functions
Aktionsart 32, 82, 102–103, 113, 126, 131 change of speaker (in dialogue) 215, 218,
anaphoric use 224–236 325–327
strong 226–230, 230–232, 235
temporal modifiers 12, 27n dative
weak 226–227, 228n, 232–234, 235 ‘possessive’ with εἶναι/γενέσθαι 256n,
also see article 269–271, 275, 400, 403, 405
ancient grammarians 332, 357 proper names with and without article
answer 335–356, 370–371 322–323, 328n
‘no’ and ‘not’ in 344–347, 349–350 declarative sentence 102, 179
antecedent 234n diathesis (active, middle, passive) 43–44,
aorist stem 47–48, 357–358, 361–365, 395
aspectual value 40–41, 70, 83, 125–132, direct speech 50, 72, 112, 134, 167
242 Dramatis Personae 180n, 182n, 183
complexive 33, 74, 161, 192, 195
constative 49–50, 70 exclamation 102, 221
countable actions 49n, 283n expressions of sorrow 254–277
gnomic 151, 162, 188–189, 203, 204, 205, see also Index Graecitatis
206, 207
habitual/omnitemporal contexts 69, 151, focalisation 136, 189, 204
204 see also Substitutionary Perception
indicative 41, 60, 68–73, 78 Focus 226, 227, 309n, 326n, 339
ingressive 32–34, 66 Fortsetzung 226, 228, 230, 231
negation 22 Free Indirect Speech/Thought 135, 139–
non-past use 68–71, 146–147 140
participle 33–36 see also Substitutionary Perception
408 index rerum

Functional Grammar 1, 39, 53, 291n, 340n middle voice 357–369

see also argument, sattelite, semantic see also diathesis
functions direct-reflexive 366
future perfect stem indirect-reflexive 260n, 359, 361, 395
aspectual value 40 intransitive 358, 366
indicative 41, 66 morphologically active perfects 43
future stem passive meaning 358–365
aspectual value 40–41 reflexive causativum 366n
indicative 41, 65, 71–72, 187n, 201, 240– momentaneous verb see Aktionsart
242, 246n
in Linear B 150 narration/narrativity 133, 139n, 141–146
generic use: see habitual/omnitemporal alternation of tenses in 50, 99, 105, 118,
use 146–151
habitual/omnitemporal use 65, 201, negation 335–356
207 and aspect choice 10, 21–23, 25
participle with and without ὡς 395 historic present not combined with 103,
vs. future perfect stem 66n 107n
‘no’/‘not’: see answer; negation
genitive nominative
of comparison, with πρότερον 285 proper names with and without article
proper names with and without article 322–323, 328n, 330
322–323, 328n number
with expressions of sorrow 266, 277 plural vs. dual 393–394
Δαρείου καὶ Παρυσάτιδος (at X. An. 1.1.1) singular vs. plural first person 187, 194
395–396, 400, 403–405 singular vs. plural in expressions of sor-
row 255–256, 258, 259, 263, 268n,
historic present: see present stem 269, 270, 275, 277

imparfait/passé simple (French tenses) 8, optative 72, 124n, 159n, 240–242, 246n, 393–
73n, 138–139, 140–141, 145, 146n, 147n, 394
167–168 ordinal, ordinator 289–290, 291
imperfect: see present stem
imperative parenthesis 113–114, 297n
aspect 80–97 participle
implicature (Grice) 70, 146 aspect 21, 23, 29–30, 34–36, 50n, 69, 76,
infinitive 78, 114, 151, 154, 160n, 192n, 213, 274
articular 279 dominant 261
aspect 49–50n, 82n, 83–84, 96n, 125–132 future with ὡς 395
dynamic 83, 124–126 particle 3
injunctive 162–163, 191n answering: see answer
inquit formula 210–222 with historic present 104
internally affected: see semantic functions see also Index Graecitatis
passive voice
messenger speech 99–122 passim, 149, 170– historic present rare in 103, 106
184 in perfect stem 43–48, 50n, 52–53
basic scheme 171 see also diathesis, middle voice (passive
beginning with a simple sentence 176– meaning)
179 past progressive (English tense) 73n, 138,
beginning with ἐπεί 171–176, 179–184 141, 167
index rerum 409

patient: see semantic functions imperfect 15, 22, 25, 26, 33, 41, 50, 82n,
Perception Indicator 136, 158, 160, 163, 165 214n, 105, 191, 198, 204, 363, 402
see also Substitutionary Perception iterative 22
perfect stem 39–58 expressing simultaneity 30, 269n
active intransitive 43–44 of consecutive action 82n
active transitive 44–45, 46 unaugmented 101, 107–131
aspectual value 40–41, 64, 66, 67 absent from Linear B 149–150
development of in Homer and after 43– as past tense 146
48, 53–54 at start of narratio 147–149
emphasizing responsibility 47–48, 49, discourse function 60–79
51, 53–54 στεῖχον in Hes. Th. 10 186, 189, 191,
indicative 41, 64–65 204
not a double of aor. ind. 48–52 focalising use 189–191, 204–205
intensive 42n framework use 16–19, 71, 73–77, 146,
passive 43–48, 50n, 52–53 150, 192
pluperfect 14, 26, 41, 50n, 67, 71, 72, 76, in Substitutionary Perception 136,
105, 146, 150, 151, 154, 155, 190n, 363n 151–165
subject-oriented vs object-oriented 42– iterative use: see generic/habitual/omni-
52 temporal use
person negated 21–23, 103
first-person vs. third-person narration participle 33–36
156–157, 163, 189 ‘protocol’ use 87, 90, 91, 97
first person future 187 relative tense 15–16, 63
in historic present 102 timeless use: see generic/habitual/omni-
in inquit formulae 210–211 temporal use
pluperfect: see perfect stem verba dicendi 24–30, 193, 214–216, 219
present stem γίγνονται in X. An. 1.1.1 397–399, 400–
aspectual value 40–41, 71, 81–82, 125– 402
132, 150–151, 242 presupposition 243, 244, 249, 251, 252,
atelicity 129, 131, 150 333n
conative 82, 150 in Substitutionary Perception 135
durative 32, 82, 83, 90, 127, 136, 150, 155, proper names 139, 159n, 161, 228, 319, 323–
167, 242, 283 324n, 397, 403–405
indicative 41, 61–64 with or without article 282n, 283n, 319–
future use 63–64 333, 393
generic/habitual/omnitemporal use Präterit (German tense) 138, 141
41n, 61–65, 69, 82, 126, 127, 129, 150,
155, 158, 162, 163n, 188–191, 192n, question 97, 102, 216, 218, 221, 299, 335–
195, 203, 204, 205, 206, 207, 247– 356
248 introduced by τί οὖν οὐκ 63n
historic present 9n, 62, 64, 76, 77, 99– rhetorical 90, 91, 102n
122, 149, 151, 161n wh-question 371–373
not a marker of ‘vividness’ 101, yes-no question 371–374
379 also see answer
not normally with negation 103
marks decisiveness 99–101, 160 reflexivity 365
rare in subordinate clauses 103 relative clause 188
semantic features 101–107 digressive vs. restrictive 233–234
γίγνονται in X. An. 1.1.1 397–403 relative time/tense 7n, 11–12, 63, 67
410 index rerum

repetition Theme 298n

in answers 339, 343–344 Topic 223–236, 231, 234–235, 224, 298n,
Rosetta Stone 50–51, 54–56 317
New 228–230, 308
sattelite 291n Resumed 224, 228–236
first-, second- and third-order entities Topic shift 304, 307, 312, 315
264, 266, 268, 270n, 272n, 275
semantic functions 43, 44, 46–48, 52–53 verbal system 5, 39–41, 61, 361, 367
sentence type: see exclamation, question, verbs
declarative sentence of approaching 274
subjunctive 72, 124n, 126n, 187n, 194, 201n, of giving 258–259
202, 205 of reaching 256, 272–274
+ ἄν 238–252 of speaking 23–30, 112, 193, 214–216,
generic/habitual/omnitemporal use 219
246–250 of violence/seizing 256, 259n, 276
Substitutionary Perception 133–141, 151–167 of volition 124–125
voice: see diathesis, middle voice
tense 6–7
see also absolute tense, aorist stem, etc. Window Opener 133, 138, 151, 159, 161,
turn-taking: see change of speaker 166
Index Graecitatis

Ancient Greek

ἄγγελος 170, 180n, 181, 182n ἐπεί οὐ 22

ἄλγος/ἄλγεα 254–255, 258, 261, 262, 264, ἔπειτα/τὸ ἔπειτα 290
270, 272, 274, 276 ἐρώτα 87–88
ἀληθῆ, ἀληθέστατα (λέγεις) 213–214, 337 ἐρώτησις 370–374
ἀκροᾶσθαι/ἀκροάσασθαι 124–131 ἔστι ταῦτα / ἔστιν οὕτω 337, 342
ἀναγίγνωσκε/ἀνάγνωθι 84n εὕδειν 102–103
ἀνάγκη 338, 342 εὑρίσκειν 161n
ἀπαγορεύειν 81 ἔφη/ἔφην 210–222
ἀπειπεῖν 81 ἔφη ὑπολαβών 212–213
ἀποκρίνεσθαι/ἀποκρίνασθαι 86–87, 88–89 ἔχειν 102–103, 257–258
αὐτίκα/τὸ αὐτίκα 290
αὐτοπάθεια 362, 365–366 ἤ 285
ἄν 238–252 ἦ δ’ ὅς/ἦν δ’ ἐγώ 210–222
αὐτάρ 201–202 ἦλθον, εἶδον, ἐνίκησα 143
ἄχος/ἄχεα 259, 263, 270, 272, 274, 276 ἤρετο 221

γάρ 104–105, 113, 116 θεράπων 170, 180, 181

γε 339, 343 θρῴσκειν 119
γίγνεσθαι/γενέσθαι 269–271, 395–396, 400,
403–405 ἵστασθαι 154n
γίγνονται (in X. An. 1.1.1) 397–399, 400–402
καὶ 104, 293–317
Δαρείου καὶ Παρυσάτιδος (in X. An. 1.1.1) καὶ … αὖ 303–304
395–396, 400, 403–405 καὶ … γε 303–304
δέ 104, 215–216, 226, 228, 230–235, 293–317 καὶ … δέ 293–317, 394–395
διάθεσις 357–369 καὶ … δή 302–304
διδόναι 258–259 κακόν/κακά 258, 261, 268–269, 270
διηγεῖσθαι/διηγήσασθαι 147 καλεῖ/κάλει 109–112, 116
δοκεῖ μοι 338, 342 κεῖσθαι 102–103, 120–122
κῆδος/κήδεα 258, 259, 261, 263–265, 270, 272,
ἐάν 238–252 274, 277
εἰ 239–242, 246n, 252 κινδυνεύει 338, 342
see also ἐάν κυνεῖ/κύνει 101, 108, 117–118
εἰἑν 216–217 κυρεῖ/κύρει 108, 110–111, 113–116, 117–119
εἰκός (γε) 338, 342
εἶναι 102–103, 233, 269–271 λαβεῖν 271–272
εἰπέ 80–97 λέγε 80–97
εἰς + adv. 290
ἐλέγχειν/ἐλέγξαι 86–87, 89, 90–91, 93 μάλα (καὶ μάλα, μάλα γε, etc) 339, 343
ἑλεῖν 271–272 μέν νυν 76
ἐναγώνιος 101n μένειν 102–103
ἐνέργεια 357–361, 363, 366–367 μεσότης (μέση) 357–360
ἐπεάν 12, 238, 243–252 μεταξύ 82, 151
ἐπεί 13, 15, 23, 171–176, 179–184 μόρος 255
412 index graecitatis

ναί 339, 343 πένθος 258, 259, 261, 267, 270, 272, 274
νομίζειν 102–103 πεῦσις/πύσμα 370–374
νυ 192 pe–re (φέρει/φέρε) 150
νῦν 50n πῆμα/πήματα 261–263, 268, 274, 277
πόνος 258, 259, 261, 267, 270
ὁμολογῶ 338 πρότερον/τὸ πρότερον 279–292
ὁ 319–333 πρῶτον/τὸ πρῶτον 287
see also article πῶς δ’ οὐ / πῶς γὰρ οὐ 339, 343
ὀϊζύς 258, 259, 261, 266–267, 270, 272, 274
ὅπως 171n σχολή 126, 129–131
ὀρθῶς 337
ὅταν 12, 205, 238, 244–246, 252 τε (‘épique’) 188
ὅτε 11n, 20 –τερος 285n, 287
οὐ 336–345 τί οὖν οὐκ 63n
(οὔ/οὐδαμῶς/οὐ δῆτα/etc.) 343–347 τιθέναι 260–261
οὐδὲ … δέ 313–314 τοι 200
οὐδέτερον 359–362, 364, 365n τοίνυν 218
οὐκοῦν 336–347
οὗτος 223–236 φάθι 87n
φαίνεται 338, 342
πάθος 357–362, 367 φημί 338
παντάπασι (παντάπασί γε, etc.) 339, 343 φοιτᾷ vs. φοίτα 109–110
πάνυ γε 339, 343
παύεσθαι 81, 151 ὡς + future participle 395

Modern Greek

δέν/δε(ν) 348–349 όχι 348–349, 351, 352–354

κανείς/κανένας 348n τίποτα 348n

Index Locorum

Aeschines Anonymus
In Timarchum (1) De figuris (Walz)
43 148n VIII 703–704 373n
De falsa legatione (2)
62 129 Apollonius Dyscolus
In Ctesiphontem (3) De contructione (Uhlig)
59–60 129 2.2.237.9 366
Epistulae 2.2.296.2 361
7.4 127 2.2.395.3 362
2.2.398.5ff. 362
Aeschylus De constructione (Lallot)
Agamemnon 1.30 (=2.2.28.1) 371n
1055 130 2.51 (=2.2.161.12) 371n
Persae 3.21 (=2.2.286) 81n
313 108
376 108 Apollonius Rhodius
416 108 4.1381 191n
458 108
506 108 Apsines
Choephori Rhetorica (Spengel–Hammer)
738 10 I 305 373n
Prometheus Vinctus
6 54n Aquila Romanus (Halm)
11–12 374n
Alexander Aphrodisiensis
in Aristotelis Topicorum libros octo commen- Aristophanes
taria Acharnenses
539.18 373n 409 130
Ecclesiazusae
Alexander Rhetor 477 311
De figuris (Spengel) 888 308
III 24 ff. 371, 372–373 Equites
870 346
Alcaeus Lysistrata
fr. 2D 196n 6 311
Nubes
Ammonius 1462 296
In Aristotelis de Interpretatione commentarius Pax
(Hülser) 243–244 310
200.5 ff. 372 246 310
250 310
Andocides 523 311n
1.9 128, 129 718 311
1.14 339n Ranae
1.69 127 30 62
164 310
414 index locorum

Ranae (cont.) 57.4 300

271 323n 59.126 314
608 323n
Vespae Diocles Magnus
8 346 =Stoic. Veterum Fragm. II 186 (Hülser)
652 82n 874 372

Aristoteles Diogenes Laertius

Metaphysica 7.64 360, 366
1034a17 350n 7.66 370, 372
Physica
8.2.253a10 ff. 367 Dionysius Thrax
8.6.259b10 ff. 367 (Uhlig–Schneider–Hilgard)
Topica 1.1.48–49 357
158a 373 (Lallot)
12.80 (1.1.39.1–2) 373n
Charisius (Keil)
1.164 ff. 365 Elias
1.165.3 360n In Aristotelis Categorias commentaria
146 216n
Choeroboscus (Uhlig–Schneider–Hilgard)
4.2.98.30 ff. 363n Etymologicum Magnum
4.2.99.28 ff. 363 τίς 373n
4.2.101.16 ff. 364
4.2.208.6 362n Euripides
Fragmenta
Cicero fr. 495.36 108
De domo sua Alcestis
51 345n 158 ff. 172–173, 181
De finibus 176 117n
2.2 374n 183–184 117n
De oratore 184 103
2.255 374n Andromache
De senectute 1085 ff. 149, 173, 181
20 374n 1153 102n
1159 108n, 122n
Demosthenes Bacchae
8.23 127 50–51 238–240
9.55 127 333 308
9.70 300 409–411 196n
18.2 128, 129 677 ff. 177
18.148 99n, 101n 723–729 117–119
21.13–18 148n 760 105n
21.189 300 767 108
34.5 ff. 148n 940 244
42.1 302 1043 ff. 175, 181
43.2 129 1066 108
47.3 128, 129 1084 108
52.11 302 1134 108
index locorum 415

Cyclops 1400 ff. 178

199 200 Phoenissae
Electra 946 308
686–687 65 1090 ff. 174–175, 182
774 ff. 174, 181 1207 ff. 175n
777 119n 1219 ff. 178–179
1117 307n, 310 1359 ff. 174–175, 182–183
Hecuba 1410 105n
521 ff. 176–177 1428 ff. 174–175, 183
580 108 1458 105n
963 113–114 Supplices
1153 108 650 ff. 177
Helena
605 ff. 177–178 Eusthatius
1526 ff. 174, 180 Ad Homeri Iliadem (Van der Valk)
1584 112 3.623.23 196n
Heraclidae
800 ff. 173, 180 Herodotus
825 112 1.1.1 234
Hercules Furens 1.1.2 227n
923 ff. 173, 181–182 1.5.4–6.1 148n
Hippolytus 1.6.1 66, 227–228
760 117n 1.6.2 225
1173 ff. 176 1.7.3 33, 288
1212 103n 1.9.3 243
1236–1237 103 1.10.2 63, 100
Ion 1.13.1 245
1122 ff. 174, 183 1.13.2 225n, 229
1205 108 1.14.2 225–226, 230–231
Iphigenia Aulidensis 1.17.1 150
414 ff. 178 1.24.1 225n
1543 ff. 171n, 175, 181 1.31.5 225n
1579 117n 1.50.2 231n
Iphigenia Taurica 1.50.3 225n, 230n
261 ff. 174 1.60.4 225n, 228
330 108n 1.62.1 225n
334 108n 1.65.1 225n
625 245 1.66.4 225n, 233–234
1327 ff. 174, 181 1.70.2 225n
1395 105n, 107n, 117n 1.71.1 17, 18
Medea 1.71.3 239–240
1136 ff. 173, 181 1.72.1 282
1141 100–101, 108, 115–116, 1.74.2 69, 225n, 233n
117, 118 1.75.2 17
1207 108, 117n, 122n 1.75.3 17n
1243 296n 1.77.1 21n
Orestes 1.77.2 28
257 119 1.85.4 233n
866 ff. 149, 177 1.96.2 225n, 229
416 index locorum

Herodotus (cont.) 3.50.2 25n

1.114.1 68, 150 3.54.1 19
1.133.4 250 3.56.2 19
1.141.2 23 3.86.1 31n, 234
1.142.4 225n, 231–232 3.88.1 23
1.143.3 225n 3.113.1 241
1.144.3 225n 3.126.2 26n
1.146.3 225n, 231 3.135.1 28
1.149.2 225n 3.159.1 67, 281n
1.157.2 22 4.1–5.23.1 75
1.157.3 225n, 233n 4.5–82 76
1.160.4 225n, 233n 4.43.1 23
1.162.2 225n 4.68.4 308
1.163.1 225n 4.82 76, 77
1.165.2 21n 4.89.1 29
1.176 18 4.89.3 73
1.176.3 225n 4.115.1 13
1.183.3 225n 4.117 23
1.184 225n 4.119.1 27
1.187.3 225n, 233n 4.143.1 76, 77
1.188.2 225n 4.144 77
1.191.6 67 4.144.3 77
1.194.1–4 246–247 4.154.4 30
1.198 65 4.163.1 74
1.199.3–4 250 4.167.3 77
1.201 225n, 226, 233n 4.200 77
1.205.1–206.1 18 4.205 77
1.206.1 151 5.1 77
1.208.1 67 5.11 76
1.214.1 229 5.14 77
2.1.2 17 5.15.3 31
2.2.1 83 5.27.1 33
2.10.3 46–47 5.30.2 20
2.29.2 246 5.42.2 28
2.41.4 32 5.57.2 29
2.44.1 309 5.63.3 26
2.94.1–2 248 5.79.2 31
2.113.1 31n 5.81.2 27
2.113.2 235n 5.92β–η.4 72
2.121β–γ.1 16 5.92β.1 72
2.121ε.4 34 5.99.1–100.1 18
2.118.4 22 5.104.2 22
2.147.2 33, 83n 5.108.1 8n
2.162.3 28 6.13.1 26n
3.3.3 19, 244 6.19.3 34n
3.4.1 17 6.25.1 26n
3.36 23 6.27.2 76
3.39.1 19 6.63.2 281n
3.41.2 30 6.86.1 28
index locorum 417

6.89.1 26 7.168.1 25n

7.1.2 29 7.169 35
7.3.2 225n 7.171.2 34
7.4 33 7.180 225n, 233n
7.6.2 225n, 226, 232, 233n 7.181.1 14
7.6.4 29 7.184 16
7.8–10 71 7.186.1 225n
7.8α.2 68, 147 7.190 225n
7.8δ.1 34 7.198.1 225n
7.10.1 21n 7.211.2 23
7.10β.1–2 68, 147 7.211.3 23
7.10δ.2 64 7.212.2 22
7.10ε 68, 189 7.213.2 29
7.10η.1 23 7.213.3 225n
7.15.1 29 7.217.1 227
7.16.1 23 7.225.3 225n
7.27.1 225n 7.220.4 29
7.49.5 29 7.222 23
7.58.3 226–227n 7.223.4 35n
7.59.2 225n, 233 7.229.1 28
7.59.3 225n 7.234.1 28
7.70.2 225n 8.7.1 310n
7.72.1 225n 8.25.1 27
7.75 288 8.37.2 31n
7.80 225n 8.70.1 29
7.83.1 225n 8.76.1 26n
7.86.2 23 8.80.2 244
7.87 225n 8.90.1 73
7.89.1 23 8.94.2 31n
7.89.2 225n 9.18.2 27
7.91 225n, 233n 9.20 22
7.99.1 21n 9.69.2 31n
7.103.3 21n 9.80.2 23
7.109.2 225n, 233n 9.93.2 192n
7.111.2 225n, 233n 9.98.2 21
7.113.2 225n 9.94.1 235n
7.117.2 225n 9.122.2 64
7.123.2 226
7.136.1 29 Hesiodus
7.142.3 225n Opera et Dies
7.143.1 225n 43–46 202
7.144.2 225n 633–635 192n
7.145.1 29, 225n Theogonia
7.150.2 30n 1–115 185–208
7.152.1 30n 1–21 192
7.153.1 225n, 233n 1–10 162, 186, 191n
7.155.2 225n 1 187
7.164.1 225n 2–8 191
7.166 23 2 188
418 index locorum

Theogonia (cont.) 60 196

4 188, 204 62–62 188
5–9 188 62 195, 196n, 197
8 191 63–67 196
9–34 187 63 188, 198
9–10 189 66 195
9 189 67 195
10–23 189 68–79 196
10–11 199 68–70 163, 204
10 133, 189, 191, 204 68–71 187n, 199
11–21 199 68 196, 198, 199, 204
22–23 189 69 199
22 192–193, 198–199, 204 70 199
24 192–193 71–74 187n, 199
25 193 71 198n, 199, 204
30–34 194n 72 198
30–31 192n 75 199, 204
30 192, 200, 203 75–79 187n
31–32 193 79–92 200
31 192, 203 80–103 187n
32 203 80–92 199, 200
33–34 193, 203 81–82 200
33 187, 203, 204 81 201
34 195 82 200
35–36 187n 83–84 199, 201
35 194 83 200
36–37 195, 202 87 202
36 194, 195n, 196n 90 203
37–67 187n, 193 93 200, 203
37 195 94–96 200
39–40 199, 201 94 200
39 195 95 201
40 195 96–97 200
42 195 96 201
43–46 203 97 199, 201
43–45 204 98–103 201
44 193, 195, 202 99–101 204
45 195 99 202, 203
46 195 100–101 202
47 202 100 204
50–51 204 102–103 203
50 202 102 201n, 202
52 195 103 203
53–60 187n 104–115 187n
53 195, 196n 104–106 203
55 195 104–105 204
56–57 195 104 203
60–68 198 108 188
60–61 196, 203 113 188
index locorum 419

114–115 204 9.412–416 146–147

114 202, 203 9.412 68
191–192 198n 9.591–592 270
192–193 198n 10.89 258n
201–202 198n 10.131 36
10.145 45n, 271
Homerus 10.245 258n
Ilias 10.279 258n
1.163–164 200 10.438 44
1.188 269 11.106 192
1.219 215n 11.347 262, 268, 274
1.234–235 45 11.362–363 274
1.254 273 11.407 199
1.362 273 11.413 261n
1.386 193 11.657 44
1.417 267 12.272 311n
1.444–445 258 12.392 269
1.467 268 13.73 311n
2.42 267 13.86–87 269n
2.135 43 13.239 270n
2.195 261n 13.416 266n
2.272–274 45 13.417 269
2.375 255 13.464 273
3.97–98 272 13.581 271
3.288–291 201, 202 14.80 269
3.412 257 14.458 269
3.351 261n 14.459 268
3.354 261n 14.475 271
4.169 269 14.480–481 270
4.456 270n 15.90 43
5.297–310 16 15.202 193
5.394 271 15.235 268
5.470 36 15.245 273
5.667 257 15.720–721 260
5.759 269 16.11 21
6.110 193 16.22 271
6.116 193 16.55 269
6.261–262 43 16.101–111 16
6.282 268 16.262 261n
6.355 273n 16.508 269
6.413 259, 269 16.515–520 273
6.489 198n 16.548–549 271–272
7.36 268 16.552 271
7.113 311n 16.581 269
7.124 273 16.599–602 271
7.173 311n 16.651 267
7.375 311n 17.37 260, 267
9.229 262n 17.97 199
9.249 269 17.99 268n, 274
420 index locorum

Ilias (cont.) 1.103–112 161n

17.157–158 260n 1.342 273
17.351 311n 1.392 202
17.538–539 271n 2.41 272
17.687–688 261 2.45–46 193
18.8 264n 2.45 272n
18.52–53 264 2.163–164 274
18.53 265 2.163 268n
18.64 273 3.100 262
18.73 273 3.220 263
18.79 ff. 265 3.291–293 191n
18.88–89 270 4.35 267n
18.395 272 4.236–237 259
18.429–431 264, 265 4.391 311n
18.431 263 4.812 267n
19.40 193 5.5 265, 266
19.125 271 5.58 ff. 161n
19.178 311n 5.63–75 154n
19.307 273 5.206–208 263
19.367 273 5.289 273
20.28 311n 5.301–302 263
20.282 273 5.336 254
20.293 269 6.60 311n
21.12 ff. 261n 6.169 273
21.92–93 270 7.81–135 190n
21.137 268 7.81–137 151–153
21.249 268 7.104–132 189
21.522–525 260 7.132 190
21.562 199 7.195 262
22.43 273n 7.211 257
22.52–53 256 7.213 268
22.122 199 7.218–219 267
22.288 268 7.218 257n
22.385 199 7.219 257n
22.389–390 201, 202 7.241–242 258
22.421 268 7.269–270 266
22.422 255 8.81 268n
22.488 270 8.149 266
24.91 257, 263 8.263 154n
24.105 257 8.266 199n
24.419 44 8.267–266 199n
24.525–528 255, 259 8.367–377 202
24.547 261, 262n 8.367 199n
24.708 273 8.480–481 193
24.741 267 8.480 200
Odyssea 8.488 193
1.10 187 8.553 198n
1.32–34 255 9.12 265
1.76 193 9.15 265n
index locorum 421

9.37 265n 23.210 259

9.381 154n 23.211–212 266n
9.402 154n 23.224 273
9.423 274 23.306–308 260, 265n
9.535 262 24.226ff. 161n, 267
10.376 257n 24.231 267
11.115 262n 24.233 257n, 267
11.167 257 24.462 269
11.208 269
11.369 265 Hymni Homerici
11.376 265 Demeter (2)
11.382 265 1 187n
11.481–482 257 Apollo (3)
11.542 265 1–13 186, 191n
11.555 260n 5 191n
11.582–584 265n Hermes (4)
11.593–594 265n 1–7 195n
12.231 261 13–18 198n
13.301 258n 20–102 198n
13.302 311n 23 196n
14.32 263 Aphrodite (5)
14.47 265 260–261 186, 191n
14.185 265, 266 6.3–4 197
14.197 265, 266 6.6–13 197
14.270 270 6.6 197
14.310–312 262 6.14–16 197
15.223 198 9.8 187n
15.256–257 198–199 11.1 187n
15.313–317 202 13.1 187n
15.342 267 16.1 187n
15.344–345 257, 274 19.10–15 186, 191n
15.399–401 265 19.27–29 186, 191n
15.487 265 19.40–45 197
16.205 193 22.1 187n
17.286–287 259n 25.1 187n
17.287 258 26.1 187n
17.439 270 28.1 187n
17.446 261
17.555 266 Isaeus
17.472–473 259n De Philoctemone (6)
17.477 270 2 129
18.272 246n
18.324 257n Isocrates
18.376–377 202 4.112 130
19.167–168 259n 14.6 129
19.471 271 15.12 127
19.512 259 15.20 127
20.195–196 258 15.28 129
21.207 193 18.4 148n
422 index locorum

Isocrates (cont.) 23b9 130

19.4 148n 40b2 82n
21.1 148n Charmides
164b 345n
[Longinus] Cratylus
De sublimitate 401c 281n
18.1 370 Critias
25 101n 112a 281n
Crito
Longus 48b 305n
1.2.1–3 160–161 Demodocus
383b3 129
Lysias Epistulae
1.6 113–114, 148 7.349b2 213n
1.22 147n Euthydemus
7.4 148 273e 286
10.4 114 277a3 216n
12.55 127 277a4 216n
12.58 50n 277a6 216n
13.5 148 277b2 216n
13.79 127 283b 288
16.9 128 288c 286
18.4 50n 293c 343
19.5 297 289c9 213n
19.11 128 Eryxias
30.1 129 400e 308
30.24 48 Euthyphro
6c 290n
Macrobius 6d 286
Excerpta Grammatica (Keil) 7c 337
5.627.5 ff. 364 14a 284
5.627.38 362n 15c 342
Gorgias
Philostratus 447a 288
Imagines 448a5 124, 125
1.26.1 196n 448d7 125
451d5 93n
Pindarus 461b–481b 86
Isthmia 461b1 87
4.35–39 46 461b3–462b2 87
Nemea 461d1 87
3.26 199 462a2 87n
462a4 87
Plato 462b 340n
Alcibiades II 462b1 87, 88
147b8 297 462b3–467c1 87–89
Apologia 462b3 87, 94n
20e3–5 84 462b5 87
21a5 84 462b6–10 88n
index locorum 423

462c3 87 476a 281n

462c6–d4 87 476a2 91
462c10 374n 476a7 91
462d8 87, 88 476a8 89n
462d10–11 87n 476b1–2 89n
463a5 88 476b2 91
463b8 88 477b1 91
463c1 ff. 88 479c4 92
463c8 88 479d7 92
463e3 88 480b6 92n
463e5 88 480d6–7 92n
464b2 88 488c2 127
465d7 88 492d5 93n
465e3 88 493d1 126
466a1 88 494b7 93n
466a2 88 495a2 93n
466a7–8 88 498e10 92n
466c3 88 505d1 126
466e9 88 506c 66
467c1 88–89 506c4 93n
467c5–470c7 89 506d 303
467c5–468e5 89 511d5 125n
468c7–8 89 513a8 126
468d6 89 515a7 93n
468e6–469e8 89 515c 284
470a1 89 516d6 93n
470a4 93n Hipparchus
470c1 89 226b 349n
470c7–474c3 89–91 Hippias Minor
470c7–8 90 363c5 125n
470c8 90 367a7 125n
471d5 ff. 90 Ion
472c4–5 90 530c2 125n
472c5 91 530d6–7 131
472d3 90 530d9 125
472d4 90 531b7–9 128
473a3 ff. 90 531e 338
473a4 90 533c2–3 125, 129
473a10 90 536d6 131
473b7 90 536d8 131
473b9 90, 91 538a 342
473d3 91 538b8 125
473e6 91 540e1 125n
474b1–2 91 541a5 342n
474c4–479c6 91–92 Laches
474c4 91, 93n, 94 179b 61
475a 307–308 180d6–8 83
475d4–5 91 180e 320n
475d6 126 181b 307
424 index locorum

Laches (cont.) 65b8 331n

184a 63 65c–d 338
185d–e 339 67d3 331n
188a 285 68e1 331n
188c 304 75a 281n, 287
194e 339 84c3 332
199d 316 86d5 331n
Leges 90c–d 339
629e 281n 91b 289
653c–d 280 92a2 331n
657c 350n 92a4 331n
670d 281n 92a6 331n
833a–b 301 94c 339
962c 345 105e 343–344
Lysis 105e8 217n
203a1–b2 211–212, 219 110d 316
205d4 127 115b1 217n
207b 151 117a8 217n
207b1 82n 117d1 332
211d 63, 65 Phaedrus
212d 282 227b8 129
221d 287 232e 284–285
Menexenus 237c6 194n
235c9 96n 251b 281n
235e1 96 253e5–255a1 133, 164, 204–205
236a7 96n 256c 281n
236b 284 259a 242
236b4 96n Philebus
236b7 96 13b 299
236c1–10 95 13e1 93n
236d2–3 96 14b3 93n
242d 283n 14c6 93n
Meno 14e2 93n
72c7 125n 15a4 93n
73a 346 16c4 93n
84a6 125n 21a7 93n
85d6 125n 23a7 93n
87e 125n, 302, 303n, 349n 23d6 93n
89a 344 25c2–4 94
98d 344 25c4 93n
Parmenides 25e6 93n
126b 281 27e4 93n
Phaedo 28d1 93n
59c8 331 29a6–8 94
60c8 213n 29a8 93n
60d8–62a1 219–220 30a2 93n
61c8 221 31d3 93n
63a2 331–332 32e4–5 94
64c3 212 33a7 93n
index locorum 425

33a8–9 95 436c4 94n

33c 284 436e 308
36e1 93n 444a3 214
40c6 316 445c3 94n
41c10 93n 445d2 94n
43a9 93n 450b–487a 94n
43d 343 451b5 94n
45d5 93n 451b8 94n
46d7 93n 457c6 94n
48c5 93n 459c7 214
49c9 93n 463e7 214
50d6 93n 467b7 214
52d10 93n 471e3 94n
53d2 93n 472b2 82n, 94n
53d8 93n 473c10 94n
53e2 93n 475b10 214
54b2 93n 495b7 214
61d6 93n 500b8 129
64b9 93n 504c9 218
65e1 93n 518e 304
Politicus 542c 339
257c8–10 84n 564b4 218
Protagoras 566a 340n
310e 286 584b1 218
316e 281n 601c14 214
321b 281n 608d 84n
331b 314n Sophista
332c 338 218a 281n
336b7 212 247e 290n
353d 290 Symposium
360c6–7 338n 176e 281n
Respublica 185e4 82n
332a6 214, 216n 186e 316
334d2 214 189a1 332n
336a8 214n 189a7 332n
337c2 216n 193d 290
343e 309 194e4ff. 96n
350c10 ff. 217 210e 290n
372c4 214 Theaetetus
372c5 216n 144e 338
373d 288 171e 300
380c6 218 202b 296
392c9 215n Timaeus
395b7 214 20c 290n
402b 343
406c5 129 Plautus
406d4 129 Pseudolus
427e–445e 94n 318 345n
430c1 214
426 index locorum

Captiva 1001 105n

262 347n 1226–1227 103, 112
1236–1240 120–121
Plotinus Electra
Enneades 35 105n
6.1.19–20 367 715–716 108
725 102n
Plutarchus 732 102n
Caesar 738 102n
50.2–3 142–143 746 102n, 103
757–761 121–122
Priscianus (Keil) Oedipus Coloneus
2, 376 ff. 365 825 49n
854 49n
Quintilianus 1606–1607 108
Institutio Oratoria 1624 108
4.2.12 143n 1626 112, 116
1627 111
Rosetta Stone decree 1655 111
Preamble, ll. 9–36 50–51, 54–56 Oedipus Tyrannus
775–780 105
Scholia 812 103
in Hesiodi Theogonia 1245 111–112
463 373n 1246 109
in Dionysium Thracem (Hilgard) 1249 108
239.14 ff. 373n 1255 109–110, 112
1268 ff. 105n
Sextus Empiricus Philoctetes
Adversus mathematicos 371 108, 110–111, 113–116
8.70 372 923–924 48
936–940 49
Sophocles 1360–1363 306
Ajax 1362 293
308 107 Trachiniae
764 105n 413–414 346n
773 112n 701 121
816 130 765 112
1107 82n 767 110, 112
1185 290n 785–786 106
Antigone 796 110, 112
254 113–114 904 108
305–314 199n 915 108
391 199n
404–407 102–103 Suda
407 108n ἀξίωμα 373n
423 103 κατηγόρημα 373n
432–433 108n πυνθάνεσθαι 373n
432 295n πύσμα(τα) 373n
442 52n
index locorum 427

Suetonius 6.86.1 287

Caesar 6.91.3 63
37.2 143 7.18.2 283n
7.56.3 295, 313n
Terentius 7.70.7 189
Adulescens 8.1.4 82n
569 347 8.8.2 281n
8.67.3 295, 313n
Theodosius
Canones (Uhlig–Schneider–Hilgard) Vergilius
4.1.49.17 ff. 363n Aeneis
4.1.49.25 363 1.439–454 165–166
4.2.99.25 363
Xenophon
Theon Agesilaus
Progymnasmata (Spengel) 10.1 308–309
II 96–97 373n Anabasis
1.1.1–4 392–405
Theophrastus 1.1.1 285n, 324
Characteres 1.1.2–10 324
23.3 307n 1.1.2 328
23.4 307n 1.1.3 328n
23.6 307n 1.1.4 328n
23.7 307n 1.1.7 328
1.1.8 309, 328
Thucydides 1.1.9 328
1.24 70 1.1.10 328
1.25.3 305n 1.2.5 156
1.26.5 82n 1.2.10 329
1.36.1 281 1.2.14 329
1.53.2 193 1.3.1 329n
1.63.2 155 1.4.12 328n
1.91.5 308 1.5.1–4 157
1.132.4 205, 313 1.5.3 190
1.144.2 296 1.5.5 105
2.17.4 306n 1.5.6 155, 190
2.21.3 127 1.6.1–5 324–325
2.36.1 295, 313 1.6.2 328n
2.48.3 241 1.6.5 328n
2.64.6 290 1.6.7–9 324–326
2.84.1–3 158–159 1.6.10 325n, 328n
2.93.4 82n 1.7.9 325n, 328n
3.87.2 281n, 287 1.8.2 299
4.24.2 294n, 295, 313n 1.8.13 328n
5.24.2 48 1.8.20 313
5.26.1 47 1.9.28 156
5.115.4 281n 1.10.1 103n
6.17.3 127 1.10.8 103
6.71.2 294n, 295, 305, 313 2.3.1 156
428 index locorum

Anabasis (cont.) 3.3.19 306

2.6.3 103–104n 4.2.22 130
2.6.5 328n 4.3.12 130
2.6.6 156 4.5.8 305
3.1.5 323n 4.5.46 305n
3.1.9 323n 7.1.1 312
3.1.13 104 7.1.32 300n, 301n
3.1.34 327 7.2.10 310
4.1.15 103 7.5.50 130
4.1.17 130 8.2.25 308
4.7.13 105n 8.3.48 130
4.8.18 82n 8.4.17 301n
5.1.9 130 Hiero
5.3.9 301 10.5 130
5.6.15 302 Historia Graeca
6.4.1–6 190 2.1.21 155, 191
6.4.14 330n 3.2.31 103n
7.1.30 314 5.2.37 305n
7.4.6 106 6.5.16 150
7.7.17 246n Memorabilia
7.7.55 330n 1.6 130
7.8.21 330n 1.1.15 302
Cyropaedia 4.4.23 344
1.4.26 311–312 Oeconomicus
1.6.17 130 1.3 312
2.1.4 68 Symposium
2.1.16 130 2.9 298
2.4.17 312n