Form and Function in Greek Grammar: Linguistic Contributions to the Study of Greek Literature 9789004386129, 9004386122

Albert Rijksbaron is internationally known as one of the leading scholars of the Ancient Greek language, whose work has

313 86 2MB

English Pages xviii+428 [447] Year 2018

Report DMCA / Copyright

DOWNLOAD PDF FILE

Table of contents :
Form and Function in Greek Grammar: Linguistic Contributions to the Study of Greek Literature
‎Contents
‎Acknowledgements
‎Albert Rijksbaron—CV and Publications
‎Introduction
Note on Editorial Practice
1 A Review of: H. Hettrich, Kontext und Aspekt in der altgriechischen Prosa Herodots
1 Introduction
1.1 Aspect as a Universal Category
1.2 Aspect versus Tense
1.3 Subjectivity; Freedom of Choice; Sentence and Context
2 Hettrich’s Analysis
3 Conclusion
Bibliography
2 The Greek Perfect: Subject versus Object
1 Introduction
2 The Views of Wackernagel and Chantraine
3 The Perfect in Homer
4 The Perfect after Homer, and the Rise of the -κ- perfect
5 Perfect and Aorist
6 Conclusion
Rosetta Stone Preamble, Lines 9–36
Bibliography
3 The Discourse Function of the Imperfect
1 Introduction
2 The Temporal Properties of the Indicatives
2.1 Present Indicative
2.2 Perfect Indicative
2.3 Future Indicative
2.4 Future Perfect
2.5 Imperfect
2.6 Pluperfect
2.7 Aorist Indicative
2.8 Conclusion of Sections 2.1.–2.7.
3 The Discourse Function of the Imperfect in Herodotus
4 Some conclusions
Bibliography
4. Sur les emplois de λέγε et εἰπέ chez Platon
1 Introduction
2 λέγε et εἰπέ chez Platon
2.1 Le Gorgias
2.2 Le Philèbe
2.3 Le Ménexène
3 Conclusion
Références
5 On False Historic Presents in Sophocles (and Euripides)
1 The Function of the Historic Present
2 Some Syntactic and Semantic Features of the Historic Present
3 Some Conclusions
4 The Unaugmented Imperfect (and Aorist)
5 A List of Historic Presents in Sophocles That May Actually Be Imperfects; Text and Apparatus Lloyd-Jones & Wilson, with Additions
5.1 Discussion
5.2 Discussion, Continued
6 Historic Presents in Euripides That May Be Imperfects
6.1 Discussion
7 General Conclusions
Appendix: Historic and Actual Presents
References
6 Ἀκροᾶσθαι or ἀκροάσασθαι (Plato, Ion 530d9)?
Ἀκροᾶσθαι
Ἀκροάσασθαι
Σχολή + Present Infinitive
Σχολή + Aorist Infinitive
Conclusion
Bibliography
7 The Imperfect as the Tense of Substitutionary Perception
Narrativity and the Imperfect
Greek Narrative and the Imperfect
Substitutionary Perception in Greek
Epilogue: an Example from Latin
Conclusion
References
8 How Does a Messenger Begin His Speech? Some Observations on the Opening Lines of Euripidean Messenger Speeches
1 Introduction
2 Survey of Content
3 Messenger Speeches That Open with an ἐπεί-clause
3.1 List of Plays
3.2 What Do the ἐπεί-clauses Refer to? Is the Addressee Acquainted with the Event Referred to?
3.3 Conclusion
4 Messenger Speeches That Open with a Simple Sentence
4.1 List of Plays
4.2 Conclusion
5 How Does the Messenger Know Where to Begin His Speech?
6 Conclusion
References
9 Discourse Cohesion in the Proem of Hesiod’s Theogony
Conclusion
Appendix to στεῖχον, line 10: a Focalising Imperfect in the Myth of Plato’s Phaedrus
The Structure of the Proem of Hesiod’s Theogony
Bibliography
10 On the Syntax and Pragmatics of inquit Formulae in Plato’s Narrated Dialogues
1 Frequency and Distribution of the inquit Formulae ἔφη/ἔφην and ἦ δ᾽ ὅς/ἦν δ᾽ ἐγώ
2 ἔφη/ἦ δ᾽ ὅς and ἔφην/ἦν δ᾽ ἐγώ in Context
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’
3.2 ἀληθῆ, ἀληθέστατα (λέγεις), inquit
3.3 Εἰἑν, inquit
3.4 τοίνυν, inquit
3.5 Results of the Above Investigation
4 The Opening Section of Lysis
5 Application to an Extended Passage: Plato, Phaedo 60d8–62a1
Works cited
11 Sur quelques différences entre οὗτος ὁ + substantif, οὗτος δὲ ὁ + substantif, ὁ δὲ + substantif + οὗτος chez Hérodote
1 οὗτος ὁ + substantif
2 οὗτος δὲ ὁ + substantif
3 ὁ δὲ + substantif + οὗτος
4 Conclusion
Références
12 Sur les emplois de ἐάν et ἐπεάν (à propos d’Euripide, Bacchae 50–51)
1 Subordonée en ἐάν à valeur future: ‘éventualité’ ou ‘attente’?
2 Subordonnée en ἐπεάν à valeur future: la ‘certitude’
3 Ἐάν et ἐπεάν ‘génériques’: ‘réalisation occasionnelle’ et ‘réalisation nécessaire’
4 Conclusion
Références
13 The Syntax and Semantics of Expressions of Sorrow and Related Concepts in Homer
Introduction
1 Ἄχεα etc. as Object
1.1 Constructions with ἔχειν
1.2 Constructions with διδόναι and Other Verbs of Giving
1.3 Constructions with τιθέναι
1.4 Constructions with πῆμα/πήματα as Object
1.5 Some Particulars
2 The Same Words as Subject, but Mostly in the Singular
2.1 Of εἶναι and γενέσθαι, + Dative; Also of Verbless Sentences
2.2 Of Momentaneous Verb Forms of Seizing: λαβεῖν, ἑλεῖν
2.3 With Verbs of Reaching
2.4 With Verbs of Approaching
Select bibliography
14 The Meaning and Word Class of πρότερον and τὸ πρότερον
1 Introduction
2 Some (Morpho-)syntactic Observations
3 Τὸ πρότερον and πρότερον
4 Some Syntactic-Semantic Differences
5 Further Examples
5.1 Of τὸ πρότερον with a Form of the Aorist Stem; = ‘the previous time’
5.2 Of τὸ πρότερον with a Form of the Present Stem; = ‘during the preceding time’
5.3 Of πρότερον = ‘before, earlier’
6 The Word Class of πρότερον and τὸ πρότερον
6.1 From a Traditional Point of View
6.2 From a Functional Point of View
References
15 Adverb or Connector? The Case of καὶ … δέ
1 Introduction
2 The Use of καὶ … δέ Connecting Single Words
2.1 καὶ … δέ Connects Two Items |
2.2 καὶ … δέ Connects Two Items in a Series of Three or More Items
3 καὶ … δέ Connects Clauses or Sentences
3.1 Preceded by τε or καί
3.2 Other Cases
3.3 καὶ … δέ = ‘and also/too’?
3.4 οὐδὲ … δέ
3.5 μὲν … καὶ … δέ
4 Conclusion
Appendix: καὶ … δὲ ὡσαύτως, and Other Constructions Involving ὡσαύτως
References
16 Sur l’article avec nom propre
1 Article et nom de personne en grec ancien, en général…
2 … et dans l’Anabase de Xénophon en particulier
3 Κῦρος contre ὁ Κῦρος
4 Les alternances des prises de parole
5 La position spéciale de Cyrus
6 Les noms de personne dans la Cyropédie
7 Article et nom de personne chez Platon
8 Conclusion
Références
17 Does Ancient Greek Have a Word for ‘No’? The Evidence from οὐκοῦν (…) οὐ Questions
1 Preliminaries
2 Questions with Bare οὐκοῦν and the Answers to οὐκοῦν
3 The Semantics of οὐ in οὐκοῦν Questions
4 Questions with οὐκοῦν … οὐ and the Answers to οὐκοῦν … οὐ
5 Latin and Other Parallels
6 Modern Greek
7 Semantic and Pragmatic Features of ‘no’ and ‘not’ in Answers
8 Summary
Appendix: οὐχί and οὔχι/ὄχι/όχι (Continuation of Footnote 33)
References
18 The Treatment of the Greek Middle Voice by the Ancient Grammarians
1 Dionysius Thrax and His Scholiasts
2 Apollonius Dyscolus
3 Choeroboscus
4 The Latin Grammarians
4.1 Macrobius
4.2 Charisius and Priscianus
5 Conclusion: a Comparison with Philosophical Approaches
References
19 A Question of Questions: peusis, erôtêsis and [Longinus] Περὶ ὕψους 18.1
References
20 The Xenophon Factory: One Hundred and Fifty Years of School Editions of Xenophon’s Anabasis
1 Introduction
2 Nineteenth-Century Education, in Germany and Elsewhere
3 School Editions of the Anabasis
4 Cultural Differences
5 ‘Für den Schulgebrauch’—or is There More to It?
6 Xenophon, Anabasis 1.1.1–4: What Are the Problems?
7 Conclusion
Appendix: Krüger on Δαρείου καὶ Παρυσάτιδος
Works Cited
‎Index Rerum
‎Index Graecitatis
‎Index Locorum
Recommend Papers

Form and Function in Greek Grammar: Linguistic Contributions to the Study of Greek Literature
 9789004386129, 9004386122

  • 0 0 0
  • Like this paper and download? You can publish your own PDF file online for free in a few minutes! Sign Up
File loading please wait...
Citation preview

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 Introduction

xi

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

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

185

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

15

Adverb or Connector? The Case of καὶ … δέ

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

279

293

Acknowledgements The author and editors wish to acknowledge the kind permission of the relevant 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. 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: 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 Dialogues’, 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 Concepts in Homer’, adapted by A. Rijksbaron from ‘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: 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 internazionale 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-laNeuve: 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 universitaires, 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 Commentary: Histories, Practices, Theory (Leiden: Brill, 2002) 235–267

Albert Rijksbaron—CV and Publications 1

Curriculum Vitae

1968 1969

doctoraalexamen (MA), cum laude, University of Amsterdam 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 European 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 Mnemosyne (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), Miscellanea tragica in honorem J.C. Kamerbeek (Amsterdam) 293–309. [ch. 8]

1979 *Review of H. Hettrich, Kontext und Aspekt in der altgriechischen Prosa Herodots, Lingua 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 geschiedenis 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 linguistiques 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 Weisglas (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. Chantraine, 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 Odysseevertaling 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 functie 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: Cultuurhistorische 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 (Lunteren).

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

2002 *‘The Xenophon Factory: One Hundred and Fifty Years of School Editions of Xenophon’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 philosophies (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: Mélanges 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 Languages in the Eyes of Modern Linguistics (Brno) 261–271.

2010 ‘Voorwerpen als personages: De revolver en de pijp van Maigret’, in A. van Heerikhuizen, I.J.F. de Jong, M. van Montfrans (eds), Tweede levens: Over personen en personages 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 between 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 methodological 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 syntactic constructions co-occur in a particular literary text, apparently without any clear difference in meaning or function between them. A thorough examination of the occurrences of these linguistic alternatives in their (literary) context 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. Functional Grammar has been a constant and important inspiration to Rijksbaron’s work. He knew Simon Dik well personally and he belongs to the first generation of linguists (together with, for example, his close colleagues, the Latinists Machtelt Bolkestein and Harm Pinkster) who applied the theory and contributed to its further development. Rijksbaron’s work occupies a prominent place in the field of Greek linguistics. His Syntax and Semantics of the Verb in Classical Greek has become a standard reference work for Greek linguists; unusually for an academic monograph in this field, the book has gone through three editions and an American reissue 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 volume), 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 edition with commentary of Plato’s Ion. Rijksbaron is no doubt the most recognizable voice of ‘Dutch pragmatic criticism’ (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 traditions 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 publications 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 coursebooks, 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, syntax, 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 correct 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 commentary 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 dealing 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 unchanged. One paper has been translated from the original Dutch (ch. 2, ‘The Greek Perfect: Subject versus Object’), and one is a completely new adaptation 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 Concepts 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 maintain 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 accidentally skipped). Square brackets are also used to indicate newly added crossreferences 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 Philology 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 editions 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 (aoristic) 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 tenseform 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 anteriority. 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, Curtius 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

224

chapter 1

verbal categories. For a—non-exhaustive—survey one may consult the bibliography 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 phenomena 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 number of language-specific realisations.2 Nor are there | any a priori reasons for the often applied method of assigning, within one language, the same aspectual value to all forms formed upon a certain verbal stem, since in actual usage these forms may occur in quite different syntactic and semantic environments. 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 separated 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 Linguistics. 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 startingpoint 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 sentences 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 auszudrücken, 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 constituency’ 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 internal 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 simultaneity. Notice that he does not speak at all about the presence of the temporal conjunction when.

8

225

chapter 1

1.3 Subjectivity; Freedom of Choice; Sentence and Context It is often said that in a language having aspectual distinctions, the languageuser may make a free, subjective choice out of the aspectually differing verbforms. 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 come to choose form x rather than form y. This is | connected with another phenomenon. 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 particular 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 languageuser 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 message 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 temporal 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 Tempusstämme 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 variation, 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 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 ‘Teilkomplex 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 distinction 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).

226

10

227

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 simplified. 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 different functions; cf. also 1.1 and note 3. On pp. 12–13 H. mentions the problem of aspectual versus temporal categories; he notes that most discussions take the primary importance of aspect for granted, and he rightly stresses that the treatment of questions of tense/ 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 ‘subjectivity’ 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

11

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 suffixy: 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. 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

13

14

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. 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 reckoned 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 certainly 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). 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 examples 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 examples 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 consideration 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 semantics.

12

228

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 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 concentrate 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, irrespective 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

16

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.]. 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 excellence 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 relationships. 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/during 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 practice 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 sufficient 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 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. Godley. The Greek text is that of the Oxford-edition by C. Hude and J.E. Powell.

229

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 buying 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 relationship 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

230

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 contexts (called ‘neutral’ by H.) where there are no situational indications as to the 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 information is conveyed. One example: (7) ὃς ἐπειδὴ ἡ νηῦς ἡλίσκετο [impf.] ἐς τοῦτο ἀντεῖχε μαχόμενος ἐς ὃ κατεκρεοργήθη 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 temporal 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 former 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 subordinate 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 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.

231

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 …

232

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 example: he saw a 10-pound note on the floor. To put it otherwise, the imperfect furnishes a framework within which other actions may occur.20 As to the second point, although H. claims that he makes use of the context 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 conclusions as to the uses of the tenses in other constructions. H. himself applies the results arrived at for sentences consisting of temporal clause and main clause, to suprasentential units in four different passages, 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 clauseconstructions, 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 limited 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, 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 21

To be sure, within certain limits, cf. pp. 6–7. 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, however, 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 (–). παρασκευαζομένου 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 preparing 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

233

chapter 1

(11) Ἅρπαγος … ἐποιέετο [impf.] στρατηίην ἐπὶ Κᾶρας καὶ Καυνίους καὶ Λυκίoυς [long interruption] [174.1] οἱ μέν νυν Κᾶρες ἐδουλώθησαν ὑπὸ Ἁρπάγου (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 ‘expedition’ interpretation. Two more examples: (12) Κῦρος … ἐποιέετο [impf.] ἐκ τοῦ ἐμφανέος ἐπὶ τοὺς Μασσαγέτας στρατηίην (–). 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 formation 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, referring 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 sometimes refers to an action that stands on its own, having no temporal relationship 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 Λακεδαιμόνιοι 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) | 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 thereafter. (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 βασιληίην, 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 framework 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

234

20

235

chapter 1

examples where the theory put forward by him as to the relationship between 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 verlaufende Handlung dar, in die das ἦλθον hineinstösst. Die Hauptsatzhandlung ἐτύγχανε ἐών ist hier im Widerspruch zu dem in 2.7 Festgestellten das B-Glied [i.e. the ‘subordinate’ clause], dessen Tempusstamm vom Zeitverhältnis 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 Handlung’, 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 Naxians [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 constituents 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 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 BGliedern) 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 speaking, 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 24

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). 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 terminative 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. ἐβούλετο 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.: without 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.

236

22

237

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 explanations, 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 ‘nobody was putting out against them’ may have | been made at several, otherwise 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 signals 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 occasions. 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 negation. 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. (participles) 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 precedes), 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 ἐσκύλευον 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

27

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. 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’.

238

24

239

chapter 1

(a) the ‘Akt des Sagens’ (the ‘action of speaking’), including the enduring situation 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 distinctions, at least, have to be reckoned with: (1) When the present stem of a verbum dicendi that is directed towards eliciting 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 uttering 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 ‘simultaneously’, 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 eliciting 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 implication 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 indirect, i.e. not directed to the person asking; (c) sometimes no reaction occurs; in that case the demand, etc. is presented as such: possible reactions are considered irrelevant.28 28

Of course, when the subject of the subordinate clause and the main clause are co-

25

a review of: h. hettrich

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 difference 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 examples 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 | (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., ἀπεπέμπετο in 3.50.2 or φράζοντες in 7.168,1 as “Handlung ohne Fortwirken” (p. 107).

240

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 relationship 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 dealing 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 distinguishing 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 Thessalians 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 syntactic 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 | (36) Αἰγινῆται δὲ (–) Θηβαίων δεηθέντων [ptc. aor.] πόλεμον ἀκήρυκτον Ἀθηναίοισι 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 δεηθέντων [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 σπανιώτερον (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 concerns 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 containing 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.

241

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)

242

When Patarbemis … summoned Amasis, he, being then on horseback, lifted his leg and made an unseemly gesture, and bade the messenger take that token back to Apries | (41) ὡς δὲ … Λευτυχίδης … ἀπαίτεε [impf.] τὴν παραθήκην οἱ [δ᾽] Ἀθηναῖοι προφάσις 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, debarring 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. | (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 κελευόντων. (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’); ἐπιλεγόμενος 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 ἐπιλέγομαι (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.

243

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 irrelevant.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 verbal action. H. connects this phenomenon with these verbs being ‘non-durative’ (or, as I would prefer, terminative or punctual), which makes that, for example, 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 following sentences:

244

(49) ὡς δὲ ἐγίνετο [impf.] ἐν τῷ πελάγει, ἀποσιεύμενος τὴν ἐξόρκωσιν τοῦ Ἐτεάρχου 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, ankom32

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 attributive 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

34

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 + locative 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.). 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

245

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 terminative 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, simple 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 following 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 number 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 36

H. himself makes a remark to this effect on p. 69 in connection with νικῶ. 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 different 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 durativetransformative verbs (‘transformative’ equals more or less ‘terminative’) as ‘aoristo finitivo’ (others use terms like ‘confective’, ‘complexive’), with durativenon-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’, ἄγω ‘lead, bring’, ποιέω ‘make’. In the second case, the aorist signals the startingpoint 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 modifier that indicates the extension of the action or process, as in: (54) συνήνεικε αὐτὸν Δαρεῖον, βασιλεύσαντα [ptc. aor.] τὰ πάντα ἕξ τε καὶ τριήκοντα 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 temporal 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 Maeandrius who had been king of Samos The other examples are: 1.7.3 and 2.147.2. In this connection another objection 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

246

34

247

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 ingressive aorist only seems to occur in contexts where a statement precedes to the effect that some other king has died. | I conclude this review with some short remarks on the first section of chapter 8 (Partizip und Nebensatz-Konstruktionen ausserhalb der A-B-Beziehung).37 In this section, H. discusses constructions involving participles and subordinate 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 participle does not refer to an independent action standing in a temporal relationship 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 beiden 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

38

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. 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. mentions 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 temporal 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 successive 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 functioning 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 examples 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).

248

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 temporally. That we, in last analysis, interpret the participial clause as specifying the manner in which the σχεῖν is brought about, and, thereby, as occurring simultaneously 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 participle, 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

249

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 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 temporal 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 context 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 careless in choosing his material. I finally note that H.’s book compares favourably with many linguistic studies, in that he presents full lists of examples, which enables the reader to immediately 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: Satzlehre. 2 vols. (Hannover 1912, repr.). Maslov, Ju.D., ‘Вопросы глагольного вида в современом зарубешном языкознании/ Voprosy glagol’nogo vida v sovremennom zarubežnom jazykoznanii [Questions 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 neusprachlichen 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 Nederlandse 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 Reference 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 syntactic 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 awareness 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 responsibility.

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 scholars (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 – – – – –

404

chapter 2

the present stem the aorist stem the perfect stem the future stem the future perfect stem

(e.g. βουλεύε-σθαι) (e.g. βουλεύσα-σθαι) (e.g. βεβουλεῦ-σθαι) (e.g. βουλεύσε-σθαι) (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 continues 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 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 ‘completion’ 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 various values on the basis of the active indicative of παιδεύω:

2 Middle rather than active forms are given, since these show the various morphological features 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).

41

the greek perfect: subject versus object

action is completed not-completed situated in + resulting state present

πεπαίδευκα

past

ἐπεπαιδεύκη

future

πεπαιδευκὼς ἔσομαι

– resulting state παιδεύω4 ἐπαίδευσα

ἐπαίδευον 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 regarding 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, however, 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 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)’.

405

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 classical 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 indicative 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 intransitive 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 summarized by Schwyzer-Debrunner (2, 263) as follows: ‘… für die spätere Entwicklung 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 … Resultativperfekt 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 der Abgeschlossenheit zur Hauptsache wird, entsteht das erzählende Vergangenheitsperfekt (–). What is not immediately clear from this quotation (for it is nowhere mentioned 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 passive 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 Wackernagel. 8 Middle verbs often have a morphologically active perfect, e.g. γίγνομαι : γέγονα, σήπομαι : σέσηπα, etc.

406

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

407

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 Wackernagel, and especially Chantraine, pass by an essential semantic difference, with 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 10 11 12

Or rather, in more precise terms: the Agent is involved in the action of which the state expressed by the perfect is the result. This corresponds to Dik’s ‘Processed’ (Dik 1978: 37 ff.). To be sure, Wackernagel acknowledges the function of the passive perfect (1904: 7), but he does not pursue this point any further. 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 perfect 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 ‘objectoriented’ 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 4

chapter 2

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)

408

He observes (1927: 22): ‘Le parfait exprime bien encore un état; mais ce n’est 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 passive 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 ensuing 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 Pindar 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 However, 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 | 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 Achelous, 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 illustrative 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).

409

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 orators, 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 transitive 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

410

Perfect and Aorist

In this section another element of Chantraine’s view will be scrutinized, namely the supposed rapprochement between perfect and aorist (indicative) (cp. above p. [405]). | According to Chantraine the perfect gradually lost its stative 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 passages 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 Mountainside, 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 mentioned 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 16

The aorist is used, then, with so-called ‘countable’ actions; cp. Armstrong (1981). 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

411

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, 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 Hellenistic 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 transitive verbs, which all depend on ἐπειδή: εὐεργέτηκεν, ἀνατέθεικεν, ὑπομεμένηκεν, intransitive πεφιλανθρώπηκε, one puzzling aorist ἀφῆκεν,19 κεκούφικεν:

17

18

19

constative ἔδρασα, merely admits that she has done them. See also Rijksbaron 2006: 36, note 2]. 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 alternation 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 further 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. 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. Perhaps a hearing/writing error for ἀφεῖκεν? In general, however, the writing is very careful. 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, illustrating 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 dealing 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 discussed above (p. [410]). The transitive perfect active expresses a state that concerns 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 respective values of perfect and aorist. Observe, finally, that here, too, the perfect is not used as a narrative tense. |

20

21

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’ Parkinson (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 ΑΝΑΤΕΘΕΙΚΕΝ, 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.] 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. By using the perf. δεδώκασιν for the compensation provided by the gods the priests also stress the responsibility and merits of the gods.

412

52 6

chapter 2

Conclusion

I hope to have shown that the views of Wackernagel and Chantraine concerning the perfect transitive active cannot stand. The state expressed by this perfect primarily is the state of the Agent-subject, rather than that of the Patientobject. 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 differences 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

23

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 Comrie (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. 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 Chantraine 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 observations 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 ascertainable 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 elsewhere 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 transitive verbs is so rare in Homer, whereas intransitive and passive perfects are quite common? As an explanation I venture the following, somewhat speculative, hypothesis, which has a social, perhaps even, ethical foundation.25 In Homer, the intransitive perfect active expressing the state of the subject 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 responsible 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, this factor was the growing awareness of, and interest in, the role of the individual, 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 25

26

Cp. Dik (1978: ch. 5). 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.’ 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).

413

54

chapter 2

century that the importance of the collective begins to diminish. The beginnings 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 Φιλοπατόρων, 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 καθάπερ 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 πεφιλανθρώπηκε 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 πάντες 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 πλήθει, ἀφῆκεν, ............. καὶ τοὺς ἐν ταῖς φυλακαῖς (14) ἀπηγμένους καὶ τοὺς ἐν αἰτίαις ὄντας ἐκ πολλοῦ χρόνου ἀπέλυσε ............... τῶν ἐνκεκλ⟨η⟩μένων· προσέταξε .................. δὲ καὶ τὰς προσόδους τῶν ἱερῶν καὶ τὰς διδομένας εἰς αὐτὰ κατ᾽ ἐνιαυτὸν συντάξεις σιτικάς (15) τε καὶ ἀργυρικάς, 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

28

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)). 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 δὲ καὶ τοὺς καταπορευομένους ἔκ τε τῶν μαχίμων καὶ τῶν καὶ μέγας· πρ⟨ο⟩σέταξεν .... . . . ............. ἄλλων τῶν ἀλλότρια (20) φρονησάντων ἐν τοῖς κατὰ τὴν ταραχὴν καιροῖς κατελθόντας 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 ἀξιολόγοις περιέλαβεν, .................... τοῦ τε Νείλου τὴν ἀνάβασιν μεγάλην ποιησαμένου ἐν τῶι ὀγδόωι ἔτει καὶ εἰθισμένου κατακλύζειν τὰ (25) πεδία κατέσχεν ................ ἐκ πολλῶν τόπων ὀχυρώσας 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, καθάπερ [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 παραγενόμενος 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 ἀργυρίου 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 δειγματισμὸν 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 ζώιοις 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 πανηγύρεων 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 κατεσκεύασεν ......................... χορηγήσας εἰς αὐτὸ χρυσίο⟨υ⟩ τε κ[αὶ ἀργυρί]ου (34) καὶ λίθων πολυ-

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 Eucharistos, the Offspring of King Ptolemy and Queen Arsinoë, the Gods Philopatores, 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 benevolent 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 foreign 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 number; 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 portion [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 commanded 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 rebellious 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 Hermes 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, having 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 honour, 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 Tedeschi 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 temporal 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 discourse 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 ‘narrative 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 candidate, 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 discussed.

© 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 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

238

62

239

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 omnitemporal use the nominal constituents typically have generic reference. Also, habitual states of affairs may in principle come to an end, becoming past habitual states of affairs; | omnitemporal states are, rather, universally valid or ‘temporally 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 altogether’ (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 | (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 relative 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 crosslinguistic 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).

240

64

241

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 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 πιέζεται 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 perfect 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: | (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 WestIndies. 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).

242

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. Comrie 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

243

2.5 Imperfect One example will suffice, since the picture presented by the imperfect is very 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

11

12

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). 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). I use ‘aspectual’ in a rather informal way (and neglect voice differences): it roughly covers 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 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 contextual 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.

13

to a large extent cover phenomena of a temporal nature. Thus, ‘ongoingness’ may imply ‘simultaneity’, just as ‘closedness’ may imply ‘anteriority’. Compare Comrie (1976: 56), who calls a comparable English construction a ‘perfect-inthe-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.

244

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 primarily 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 downright 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/omnitemporal (18),14 a future ((19), (20)15 and (21)16), a past, both an | immediate (22) 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

15

16 17 18

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, however, 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. 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 Handlung 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, Barentsen 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. For my taking this aorist as having future reference see note 3, on a similar use of the present. These two differ in that the former typically occurs in direct speech and the latter in narrative discourse. See also Rijksbaron (1984: 28 ff.). 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’.

246

70

247

chapter 3

temporally fixed. This means, that, in principle, the aorist indicative is characterized 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 reinterpret 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 imperfect. 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 feature that they all simply occurred before the speech moment (or the ‘writing’ 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 … ξυνῴκισαν … 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

20

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 iterativum’, (c) when the counterfactual ἄν is present, and (d), when it occurs in unrealizable wishes. I cannot go into this matter any further, here. 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 barbarians, 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 narrative 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 imperfect 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: | (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 22 23

To put it somewhat differently: the imperfect is the sure sign that we are dealing with a narrative (cf. Longacre and Levinsohn 1978: 107). ‘Time anchor’ is used, be it in a slightly different way, by e.g. Bertinetto (1986). See also note 25. It should be noticed, incidentally, that the aorist is conspicuously absent from Plato’s discussion 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 distinguished from the present indic, and imperf. primarily by its being formed from a different stem. See Rix (1976: 239).

248

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 common, 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 provided by the two imperfects of ex. (29): (29) Κορινθίοισι … ἦν πόλιος κατάστασις τοιήδε· ἦν ὀλιγαρχίη … (Hdt. 5.92β.1) The Corinthian state was ordered in the following way. The Few ruled …24 |

249

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 Samothracian 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 | 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 imperfect can also be shown to be of major importance on what may be called the

25

26

27

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. 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. 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.]

250

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 Darius and his generals.

3

251

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 welldefined, 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 presently.30 The imperfect συνήγειρε | 28

29

30

Cf. Longacre and Levinsohn’s remark (1978: 107): ‘Significant chunks of the whole verb system 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). 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 follows, 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 συνήγειρε. A simultaneous imperfect συνήγειρε οn the micro-level occurs in: ὁ δὲ Ἀρκεσίλεως … συνήγειρε 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 κατεστρέφετο 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 ταύτης 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, ἐντελλόμενος … 5.15.3 οἱ δὲ Παίονες … παρεδίδοσαν σφέας αὐτοὺς τοῖσι Πέρσῃσι. οὕτω δὴ Παιόνων Σιριοπαίονές 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 διαπεραιωθεὶς ἀπίκετο ἐς τὰς Σάρδις. |

252

76

253

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 function: 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 completion 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 activities 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 κατεστρέψατο/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 outside 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 ἐγίνετο, 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 information 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 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 γρά32

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.

254

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/Philadelphia 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 Cascio & 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’emploie 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 raisonnement 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 continuative, 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 nettement 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). | In the Philebus, on the other hand, a dialogue that for the greater part is an argumentative 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 ‘immediative’ (‘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 conspicuously 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 peutêtre 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 participe, 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 explicitement 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.

152

82

153

chapter 4

(ii) Dans son emploi temporel l’adverbe μεταξύ ‘en plein milieu de’ est toujours accompagné d’un participe PR ou d’ une autre forme du thème de 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-accomplissement’ 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 commun 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 souvent 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 ‘Aussitôt 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 effectivement 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 décidée, 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 réalisation 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 diffère 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 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 (notamment 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 suivante. 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 référence 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.

154

84

155

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, 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 subjonctif 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 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

156

λέγε et εἰπέ chez Platon

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

Euthphr. Ap. Cri. Phd. Cra. Tht. Sph. Plt. Prm. Phlb. Smp. Phdr. Alc. 1 Alc. 2 Chrm. La. Ly. Euthd. Prt.

λέγε

εἰπέ

4 2 3 7 7 14 16 8 – 27 4 6 6 – 6 3 2 – 1

7 6 2 1 5 5 4 1 3 1 6 9 5 – 6 2 5 9 7

d’ impératifs AO et PR d’ une façon comparable. À noter qu’on ne trouve jamais, chez les orateurs, 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.

157

86

chapter 4

(cont.)

Grg. Men. Hp.Ma. Hp.Mi. Ion Mx. R. Ti. Criti. Lg. Totaux

158

λέγε

εἰπέ

9 4 6 1 – – 27 2 – 29 194

21 16 6 4 3 3 17 – – 3 167

Tandis qu’il y a un certain nombre de dialogues où les deux formes sont représentées plus ou moins en proportion égale, il y en a d’ autres où les différences 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 discussion formelle, où λέγε et εἰπέ alternent avec des verbes ‘de discussion’, notamment ἐλέγχειν/ἐλέγξαι et ἀποκρίνεσθαι/ἀποκρίνασθαι ; aussi tiendrai-je également 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 représentés 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 intervient, 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 apercevoir, 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’ accomplissement), 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 11

Dans εἰ… ἐπανορθώσασθαι αὐτὸν βούλει (462a2), l’ aoriste exprime que Socrate vise ici à la correction effective des points contestables. 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-

159

88

160

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 κολακείας 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 combine 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 impératifs 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 οὖν ἐγώ φημι | τὴν ῥητορικὴν εἶναι, ἀκήκοας, 465d7, voir aussi συχνὸν λόγον ἀποτέτακα 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 élaborer 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 questionneur; 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 passage 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 ‘complètement’ 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 processus 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 première 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 remarquer 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 13

‘Polus, defeated, falls back on an argumentum ad hominem’, Dodds dans son commentaire ad loc. 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).

161

90

162

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 itérative-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 largement à réfuter Socrate, ἐξελέγξαι. Noter l’apparition de l’ inf. AO: non seulement Polos va passer à la réfutation, mais cette réfutation sera faite une fois pour toutes. Néanmoins, après leur discussion, Socrate explique par une question 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 διοίσουσιν ἀλλήλων. L’AO exprime que Socrate veut en arriver à examiner à fond et 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 νομίζοντος 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 réfutation’, 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᾽ ἀδύνατον· 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 ἀδύνατον 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 sémelfactive. 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 questionnement, sinon son interrogatoire, en 474c4: Λέγε δή μοι… | 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 renforcée par la subordonnée ὥσπερ ἂν εἰ ἐξ ἀρχῆς σε ἠρώτων : ‘réponds-moi successivement 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 fondamental: 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 ἀνθρώπων 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 établis.

163

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.

164

2.2 Le Philèbe Je passe maintenant au Philèbe, où il y a une majorité écrasante de λέγε (voir 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 16

17

18

Ici et ailleurs (p. ex. Grg. 498e10) συλλογίζομαι = ‘énumérer ensemble’, ‘faire le bilan ensemble’, plutôt que simplement ‘reckon up’, ‘recapitulate’ (Dodds). 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 ?’ 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 surtout 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. 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)ἐλέγχειν, 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 ἀποκρίνεσθαι: 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 fournit 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 20

21

22

14b3, 14e2, 15a4, 23a7, 52d10. Peut-être Diès a-t-il le caractère ‘expositionnel’ en vue quand il parle d’une ‘discussion prolongée’. 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 préciser 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. 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. Socrate : 451d5, 470a4, 474c4, 492d5, 494b7, 495a2, 515a7, 516d6; Calliclès; 506c4.

165

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)

166

Socrate, en revanche, ne dit jamais λέγε μόνον, λέγε ou λέγε σαφέστερον. Quand 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

24

Il est intéressant de noter (voir note 20) que l’ emploi de λέγε dans la partie de la République 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. 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 totalement 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: | (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 Ἀσπασίας 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 rapporterPR 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. Méridier) 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’impératif (‘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’.

167

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, modifiée) |

168

Dans tout ce passage les formes de εἰπ-, pour autant qu’ elles soient employées absolument, ont le sens spécialisé et idiomatique de ‘prononcer une oraison (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 oraison (qui est celle d’Aspasie), il emploie naturellement λεγ- (ἔλεγε… οὑτωσί, 236d2–3).

26

27 28

Noter que le sens de ‘prononcer une oraison’ peut être rattaché à la valeur d’accomplissement. 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, préciser’. Pour λεγ- voir aussi 236b4, 236b7 et 236d2–3. 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

3

97

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 démarche 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 raisonnement, 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 réalisation 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.

169

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 Introduction 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-prohibitivus 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. (Amsterdam 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 argument, 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 determining 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 events. At varying intervals the flow of these past tenses may be interrupted by the occurrence of a present indicative form, traditionally called the historic present, sometimes also narrative present. An important function of these

1 Occasionally, the historic present occurs in what may be called mixed narrative-argumentative 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

128

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 …

129

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 indications of ‘decisiveness’ than the historic present. Below, I will try to neutralize 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 unexplained. 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 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 Hervorhebung 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 children. (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 straightaway 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 quotation 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. | 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).

130

102

chapter 5

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

131

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 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 historic 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); ἀνακωχεύει (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 πρέσβυς) 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 Sophocles 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 special 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

104 132

chapter 5

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 καί τε ἀλλά asyndeton γάρ

23 3 – – 9 –

24 19 1 2 7 5

84 18 1 2 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 Anabasis 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. Immediately after … In this example γάρ introduces an embedded narrative, explaining what the συμβάντα mentioned in the preceding sentence consisted of.10 This function of

10

ἐνταῦθα οὐκέτι πείθεται, ἀλλ᾽ ᾤχετο πλέων εἰς Ἑλλήσποντον. Observe that here, too, the negated historic present is balanced by a positive statement. 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 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 imperfect 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 …

11

12 13

he had himself, a dream that eventually led to his appointment as the commander-in-chief of the Greeks. 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’. 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. Cf. e.g. ἐπικατερρίπτουν at X. An. 4.7.13, where the embedded narrative is formally announced 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.

133

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 earth, or made him leap into the air |

134

In fact, the basic function of γάρ is to provide this type of background information.14 The two uses can in principle be distinguished by the following ruleof-thumb: regular γάρ presupposes an implicit question like ‘how was this possible?’ (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 information – (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 pro14

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 ὑποζυγίων 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 introduce background information. | 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 augment 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 acknowledged. 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 coronis 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

16

The special information status of negative statements may be illustrated from the following 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. Although verbs beginning with a long vowel or diphthong pose similar problems of accentuation. An example is E. IT 1395 ὠθεῖ Kirchhoff: ὤθει L, see also notes 18 and 27.

135

108

136

137

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: Lautensach (1899: 165–174; ‘Fehlen des syllabischen Augments’), Page’s note in his commentary from 1938 on Medea 1141 (strangely enough, Page does not mention Lautensach), and Bergson (1953: 121–128; ‘The omitted augment in the messengers’ 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 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 τροποῦτο, 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.’ Lautensach 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 following possible instances,17 where the imperfect has some, overwhelming or even exclusive manuscript support.18 For Sophocles these are: |

17

18

I confine myself to forms where a change of accent or verb ending changes the interpretation. 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. 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 mentioned 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)

5

109

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 scholarship. 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, EllendtGenthe, Kamerbeek, Moorhouse] : φοίτα p, coni. Blaydes [Roussel, Dain-Mazon]

(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)

138

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 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, DainMazon, 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, Hermann]

(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) |

5.1 Discussion All in all this gives us a maximum of six new instances of unaugmented imperfects 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 scruples. 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.

139

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.

140

(14), Trachiniae 767 Here, we are not dealing with a change of accent, but of verb ending. With the 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 preferred.

20 21

But unlike ἵετο, which rather has immediative meaning. 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 corrupt 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 regular 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]. | 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 following 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

23

‘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 languages. So we may safely assume that Greek, too, did not know this use.’ Before Hermann, Brunck had already observed: ‘Sane res ipsa manifesto tempus imperfectum 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. Apparently, Hermann shared Brunck’s view of the augment here.

141

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 happens 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 präsentischem 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 virtually 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 historic 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 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 messenger’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 byform 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 25 26

Ed. of 1812, originally published in 1801, and also in those of 1821 and 1825. The edition of Lascaris (1494) and the Aldina (1503). ‘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.

143

116

144

Brunck for rejecting a present tense form of the MSS, while elsewhere he introduced 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 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 conjecture 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

145

chapter 5

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 introducing 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 many sides’). |

on false historic presents in sophocles (and euripides)

6

117

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 blowing 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 speaking) |

146

And the women at the appointed time of day began to wave their thyrsoi 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 subsequent 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 argumentum 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 meaning: 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 29

In the Leipzig edition of 1794. ‘That the tenses show such variation in narratives is well-known to people who have read the tragedians just a bit more attentively.’

119

on false historic presents in sophocles (and euripides)

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 galopping 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

S. OT 780 (ex. (7)) OT 1245 (ex. (12)) OT 1255 (ex. (13)) Tr. 767 (ex. (14)) Tr. 796 (ex. (15))

30 31

32

Manuscripts

OCT

This article

καλεῖ κάλει φοιτᾷ majority of the MSS; φοίτα minority προσπτύσσετο κάλει

καλεῖ καλεῖ φοιτᾷ

καλεῖ κάλει φοίτα

προσπτύσσεται καλεῖ

προσπτύσσετο καλεῖ

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’). 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. Observe, however, that Dain-Mazon, in their Budé-edition, read an imperfect at OT 1245, 1255 and Tr. 767.

147

120

chapter 5

(cont.)

Ph. 371 (ex. (16)) OC 1626 (ex. (17)) E. Med. 1141 (ex. (22)) Med. 1207 (note 27, ex. (ii)) IT 1395 (note 27, ex. (iii)) Ba. 728 (ex. (23))

148

Manuscripts

OCT

This article

κύρει καλεῖ κύνει κύνει ὤθει κύρει

κυρεῖ καλεῖ κυνεῖ κυνεῖ ὠθεῖ κυρεῖ

κύρει καλεῖ κύνει κύνει ὤθει κύρει |

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 Sophocles 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 indispensable 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 translations 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 therefore 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 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-

149

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 générale, 2, (Paris 1974) 67–78. Bergson, L., ‘The Omitted Augment in the Messengers’ Speeches of Greek Tragedy’, Eranos 51 (1953) 121–128. Campbell, L., Paralipomena Sophoclea (London 1907). Eriksson, K.V., Das Präsens historicum in der nachklassischen griechischen Historiographie (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 Commentary: 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

34

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 βραχεῖ/τεύχει, where φέροντες of course echoes φέρουσιν at 759. 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 Linguistic 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 commentary 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 meaning 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 Olympiodorus (18.12 Norvin)’. Why λαβεῖν is more appropriate Dodds does not say. Burnet preferred λαμβάνειν. And this is understandable enough, for two reasons. 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 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 variation 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.

262

126 493d1 513a8 505d1

263

chapter 6

μεταθέσθαι F : μετατίθεσθαι B T W f ἀπεργάσασθαι F : ἀπεργάζεσθαι cett. 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 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 predecessors) 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 elsewhere, 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) meaning; in that case the infinitive denotes a general line of conduct. Typical examples 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, Xenophon, Andocides, Lysias, Isocrates, Demosthenes, Isaeus, Aeschines. Ἀκροάομαι does not occur in Aeschylus, Sophocles, Euripides, Herodotus and Xenophon. It does occur in Aristophanes, but not in the infinitive. 6 For activity and accomplishment, used below, see Rijksbaron (1989: 17ff.).

ακροασθαι or ακροασασθαι (plato, ion 530d9)?

127

ταῦτ᾽ ἐστὶν ἃ οὗτος λέγων τε καὶ ᾄδων ἀναγκάζει καὶ ἡμᾶς ἀκροᾶσθαι—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 κρείττους καλεῖς καὶ δεῖ ἀκροᾶσθαι τοῦ ἰσχυροτέρου τοὺς ἀσθενεστέρους— in general; note δεῖ and the generic articles τούς, τοῦ and τούς. | 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 ἀπολογήσομαι—unbounded (durative) ‘listening’ tout court. Lys. 13.79 ἀνάγκη δὲ ἦν στρατηγοῦ ἀνδρὸς ἀκροᾶσθαι, εἴπερ ἔμελλον σωθήσεσθαι—the necessity to listen (obey) existed in general; note indefinite, generic, στρατηγοῦ ἀνδρός. Isoc. 15.20 Ὧν χρὴ μεμνημένους μὴ προπετῶς πιστεύειν τοῖς τῶν κατηγόρων λόγοις, μηδὲ μετὰ θορύβου καὶ χαλεπότητος ἀκροᾶσθαι τῶν ἀπολογουμένων—in general, with implications for the actual situation; note the generic article in τῶν κατηγόρων and τῶν ἀπολογουμένων. Ly. 205d4

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 ἀκρόασιν, 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 frequently 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

264

128

265

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 passage. | Typical examples are: Lys. 16.9

δέομαι οὖν ὑμῶν μετ᾽ εὐνοίας ἀκροάσασθαί μου (i.e. my defence). ποιήσομαι 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 ἀντίδικον 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 ἀγωνιζομένων 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 πράγματος 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 ποιήσασθαι, 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 ἄριστον καὶ εὐορκότατον νομίζητε εἶναι. Also relevant is

266

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 ἀκροασαμένους ἡμῶν διὰ τέλους, ὅ τι ἂν ὑμῖν ἄριστον καὶ εὐορκότατον νομίζητε 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 (‘binding’) 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. Generally 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 frequently 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 ποιήσασθαι 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. | From other authors:8

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

267

130

chapter 6

οὔτοι θυριαίᾳ τῇδ᾽ ἐμοὶ σχολὴ πάρα | τρίβειν· … εἴ τῳ καὶ λογίζεσθαι σχολή καταβαίνειν δ᾽ οὐ σγολή. 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 A. A. 1055 S. Aj. 816 Ar. Ach. 409 X. An. 5.1.9

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 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). 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

268

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 ἀγαθὸν αὑτοῖς μηδέν … | Also relevant is: Isoc. 4.112 … ἐπαυσάμεθ᾽ ἀλλήλους ἐλεοῦντες. οὐδενὶ γὰρ τοσαύτην σχολὴν παρέλιπον 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 another’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 ἀκροάσασθαί 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 ἀκροάσασθαι 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.

269

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 Typology 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 narrative 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 perception). 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 Conference 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 scholars,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 seeing 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 country 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 statement 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 following 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 difference. 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. | Sentence (4) illustrates the phenomenon of (3) in ‘real’ English. Finally, sentence (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).

334

136

335

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 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), ‘narrated 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 Perception 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 secondary 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, circled, 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’). | (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 Mutter Hentjens Wirtschaft hinauf. In der Kastellgasse blieb Martin stehen: ‘Haben sie dir ein anständiges Zeugnis gegeben?’ (–), sie traten ein. Es

336

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 beiden Männer rieben sich die Hände, … (H. Broch, Esch oder die Anarchie. Frankfurt/M., Fischer Bücherei 1954, pp. 6–7 (origin. 1931–1932))

337

All three fragments illustrate SP to varying degrees, sometimes with and sometimes 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 presence 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 characters, 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 | dû s’endormir passes into Free Indirect Thought and next, with était, into Substitutionary Perception (note also the evaluative adjective obscure), but passes then again into exclusive narrator perception with le doigt de Maigret avait hésité. 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 Sentence (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 | 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

11 12

Both features are often ignored in discussions of narrators and characters. Thus, concerning 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 mention 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 minimalist definition of narrative might be: a perceived sequence of non-randomly connected events.’ (On ‘sequence’ see further below). Fludernik, too (1993: 449), considers the perception part especially important, arguing that narration is structured around ‘the script of experiencing or viewing, rather than telling events’ (emphasis original). In a narrative, the narrator is the default source of proper names. See also below, on sentence (f). Alternatively, it could be argued that the three characters have receded into the background to such an extent that these two sentences describe the general setting and atmosphere 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 énonciation (i.e. by the narrator—AR); ou bien: il se pencha et constata: c’était un crapaud, alors on a le style indirect libre.’

338

339

140

340

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 perceptual 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 strategy 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 character, 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 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 Substitutionary 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 14

Some illuminating observations about these matters may be found in Alavoine (1999). 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 perceptual implication of the imparfait (past progressive/German preterite), its consciousness factor, is primary and that … its function in the pattern of incidence 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 sentence (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 ‘single, 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 (external) 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 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 quotation 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

16

17

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 farreaching implications (Bakker 1997). See further below, fn. 65. 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. Stretching a bit the meaning of fio, for ‘gardens’ do not happen, of course.

341

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 relatively seldom found. Normally, the main story-line is continuously interrupted by various kinds of flashbacks and points of background information.’

342

Such pure narrative discourses are indeed exceptional, for it is the ‘interruptions’ 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 overshadows it’ (Forster 1974: 60).18 Upon this Fludernik comments (2009: 79): ‘There is 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 transforms 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 understandable. 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 19 20

21

‘… is a plot’ would nowadays rather be formulated as ‘has a plot’. Observe that Fludernik equates ‘narrative’ with ‘plot’. I would rather prefer the equation ‘story + plot = narrative.’ 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.’ 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 Caesar’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 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 Ponticus triumphus: (12) Pontico triumpho inter pompae fercula trium uerborum praetulit titulum veni vidi vici non acta belli significantem sicut ceteris, sed celeriter confecti notam. (Suet. Caes. 37.2)23 Apparently Suetonius, for one, was well aware that the events of this ‘titulus’ 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

23

24

‘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.’ ‘In his Pontic triumph he displayed among the show-pieces of the procession an inscription 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). Interestingly Quintilian, in the part of his Institutio Oratoria where he discusses the narratio 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.

343

144

344

chapter 7

for writing about Der Zauberberg Weinrich observes (1972: 24): ‘Thomas Mann hat mehrfach seine Überzeugung ausgedrückt, daß die Tempus-Form Imperfekt 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 advancement 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 denoting 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 forehead, 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:

345

(14) Elle (: Cunégonde) rencontra Candide en revenant au château, et rougit: Candide rougit aussi. Elle lui dit bonjour d’ une voix entrecoupée; et Candide lui parla sans savoir ce qu’il disait. Le lendemain, après le dîner, comme on sortait de table, Cunégonde et Candide se trouvèrent derrière un paravent; Cunégonde laissa tomber son mouchoir, Candide le ramassa; elle lui prit innocement la main, le jeune homme baissa innocement 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 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 edition of 1928, Paris)26 Notice, incidentally, the parody in the reference to Leibniz’ le meilleur des mondes 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, preparations, 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, ‘Narrative 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 | 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 27 28

29 30

For ‘das ganz ungewöhnliches Übergewicht des Passé simple’ in this satire see the pertinent remarks by Weinrich (1971: 96–97). That were published in his Théodicée (1710); Candide is from 1759. Interestingly, while a novel consisting of just simple pasts or passés simples is hardly conceivable, 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 Weinrich (1971: 97): ‘Flauberts Romane wie überhaupt die erzählende Literatur des Realismus und Naturalismus sind … durch ein auffälliges Überwiegen des Imparfait in der Erzählung gekennzeichnet.’ An interesting recent study of narrative, which strongly focuses on the importance of characters in analysing narratives, is Altman (2008). Although present tense narratives are also possible, not only as an eyewitness account

346

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

347

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 framework for what follows, and then proceed to its use in Substitutionary Perception. 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 interpretation. The point is that the aorist can be shown to behave in ways that 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

31

32

of, for instance, a football match, but also of past events, especially in first person narratives. 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. 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. ‘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 | 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, referring 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

33

34

35 36

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). ‘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). ‘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). ‘For this cause I have now summoned you together, that I may impart to you my purpose’ (transl. Godley). 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. At least, for Lysias; for διηγήσασθαι, which occurs outside the exordium, see 1.22. Out-

348

148

chapter 7

Some examples:

349

(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 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 Λακεδαιμονίων 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 summarizes the prehistory of the narrative sensu stricto that will follow. Notice, in

38

39

40

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 usually 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 catalogue 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. ‘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). ‘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 transitions 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 …’). ‘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. | For further examples see 3.5, 10.4 (with a somewhat different introductory sentence), 12.4, 16.4 (probably) and 32.4.

350

(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 (–) 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᾽· 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

863

1084

(III) That narratives and imperfects are intimately connected can also be shown from a quite different angle. If the presence of imperfects signals narrativity 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 documentary texts, non-narrative texts par excellence, such as contracts, inventories, decrees. Thus, in the Linear B tablets there are no certain instances of the

41 42 43

For a comprehensive study of the messenger speeches in Euripides see De Jong (1991). ‘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). ‘Explain …’—‘When we had come to …, we spent three shining circuits of the sun in looking around and … This caused suspicion …: the people … gathered in knots and circles’ (transl. Kovacs).

351

150

352

chapter 7

imperfect, nor of the pluperfect, for that matter. There is a maximum of 2 possible 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 imperfect are really imperfects. See for this whole subject the recent and very thorough 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. Ultimately, 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 narrative 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 ‘invade’: ἐπολιόρκεε τὴν Μίλητον τρόπῳ τοιῷδε. ὅκως μὲν εἴη ἐν τῇ γῇ καρπὸς ἁδρός, 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 45 46

‘there he was playing in the road with others of his age.’ ‘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) ‘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 Substitutionary Perception. |

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 Window Opener.49 (26)

85

90

47 48 49

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 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, ἀργύρεον δ᾽ ἐφ᾽ ὑπερθύριον, χρυσέη δὲ κορώνη. χρύσειοι δ᾽ ἑκάτερθε καὶ ἀργύρεοι κύνες ἦσαν, ‘stop striving after the things that you are striving after.’ ‘next Menexenus steps in from the court, in the middle of his game.’ For the narrative technique displayed in this passage see also De Jong (2001).

353

152

95

354

100

105

110

115

120

125

355

130

chapter 7

οὓς Ἥφαιστος ἔτευξεν ἰδυίῃσι πραπίδεσσι δῶμα φυλασσέμεναι μεγαλήτορος Ἀλκινόοιο, ἀθανάτους ὂντας καὶ ἀγήρως ἤματα πάντα. ἐν δὲ θρόνοι περὶ τοῖχον ἐρηρέδάτ᾽ ἔνθα καὶ ἔνθα, ἐς μυχὸν ἐξ οὐδοῖο διαμπερές, ἔνθ᾽ ἐνὶ πέπλοι λεπτοὶ ἐΰννητοι βεβλήατο, ἔργα γυναικῶν. | ἔνθα δὲ Φαιήκων ἡγήτορες ἑδριόωντο πίνοντες καὶ ἔδοντες· ἐπηετανὸν γὰρ ἔχεσκον. 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, 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 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, 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. 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, ἄλλας δὲ τραπέουσι· πάροιθε δέ τ᾽ ὄμφακές εἰσιν 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 σκίδναται, ἡ δ᾽ ἑτέρωθεν ὑπ᾽ αὐλῆς οὐδὸν ἵησι πρὸς δόμον ὑψηλόν, ὅθεν ὑδρεύοντο πολῖται.

the imperfect as the tense of substitutionary perception

135

153

τȏι’ ἄρ᾽ ἐν Ἀλκινόοιο θεῶν ἔσαν ἀγλαὰ δῶρα. ἔνθα στὰς θηεῖτο πολύτλας δῖος Ὀδυσσεύς. αὐτὰρ ἐπεὶ δὴ πάντα ἑῷ θηήσατο θυμῷ, καρπαλίμως ὑπὲρ οὐδὸν ἐβήσετο δώματος εἴσω. εὗρε δὲ Φαιήκων ἡγήτορας ἠδὲ μέδοντας σπένδοντας δεπάεσσιν … (Hom. Od. 7.81–137)50 |

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 Substitutionary Perception makes a quite spectacular entrance into European literature.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

51

‘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 threshold … 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 strongcompounded 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 fetching 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). A much shorter example, without omnitemporal presents, occurs already before this passage, viz. at 5.55–74, for which see De Jong (2001) on 63–75.

356

154

357

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 mention 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 pluperfects,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 belonging 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 courtyard (95–102; notice 95 ἐν δὲ θρόνοι, and cp. ὑπ᾽ αὐλῆς οὐδόν at 130), with the 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 διοτρεφέας 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, neither of which can be seen by Odysseus, is told by the omniscient narrator in omnitemporal present (and perfect) indicatives (103–131 ὑψηλόν).54 With ὑδρεύοντο 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 evaluative 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

53 54

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. 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. 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 pluperfects, 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). | The next examples, from prose, show the working of Substitutionary Perception in a more unobtrusive way.

358

(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 χοίνικας Ἀττικὰς ἐχώρει. κρέα οὖν ἐσθίοντες οἱ στρατιῶται διεγίγνοντο. (X. An. 1.5.6)57 |

55

56

57

‘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.) ClassenSteup ad loc. already observed: ‘Impf., indem die lokalen Umstände in die Vergangenheit der Erzählung hineingezogen werden.’ ‘… 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’). ‘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

359

156

360

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 combined 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 Dingen, welche in der Gegenwart noch fortbestehen, um den ehemaligen Zustand zu berücksichtigen.’ – Krüger 18897 refers the reader to his note at 1.4.9 ἐνόμιζον: ‘ἐνόμιζον für νομίζουσι. Der lebhafte Grieche drückt öfter als wir nur aus, wie etwas zur Zeit der erzählten Vorgänge gefunden wurde, auch wenn es sich dauernd so verhält.’ – 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 constatait’ 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 of a third-person, hetero-diegetic, narrative, with an anonymous, | covert, narrator. 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-

58

of two Attic choenices. The soldiers therefore managed to subsist by eating meat’ (transl. Brownson). 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 Fortbestehen der Handlung in der Gegenwart, sondern verstetzt sich in den Zeitpunkt der Vergangenheit 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 Perception. 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 στρουθοὶ 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 ἕστασαν· (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 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 ἀφικνοῦνται > ἐπὶ τὸν Μάσκαν ποταμόν, τὸ εὖρος πλεθριαῖον. ἐνταῦθα ἦν πόλις ἐρήμη … (X. An. 1.5.1–4)61

59 60

61

For the distinction between ‘covert’ and ‘overt’ narrators see e.g. Fludernik (2009: 21–22). 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 firstperson 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 narrator 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 character. 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 important characteristics of first-person narratives. ‘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

361

158

362

chapter 7

This fragment exhibits basically the same uses of imperfect and omni-temporal present as in (26), but on a much larger scale. Observe that in (30) some imperfects of Substitutionary Perception are accompanied by Perception Indicators: ὁμαλές and ὥσπερ in sentence (b), εὐώδη in (c), παραπλήσια and ἁπαλώτερα 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 did taste their meat? | So far in the majority of my examples the imperfect of SP occurred in a context 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 omnitemporal 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

62

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.’ Words like ὥσπερ and παραπλήσια, and morphological elements like the comparative suffix –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 προσπεσόντες 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

64

65

‘The Athenians, formed in line, sailed round and round them, and forced them to contract 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 signal. [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 expectation 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, swearing, 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). 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). 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 concludes: ‘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, furthermore, 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 distance of the Athenians (and vice versa, but this is irrelevant here, although with δόκησιν παρέχοντες αὐτίκα ἐμβαλεῖν we for a moment pass over to perception 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 Perception Indicators βοῇ, λοιδορίᾳ, τῶν παραγγελλομένων and τῶν κελευστῶν, and the imperfects, of course:67 Phormio’s perception hits the turmoil that is developing 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

365

(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 ἐπιπλανώμενος καὶ πόα μαλθακή, | καθ᾽ ἧς ἔκειτο τὸ παιδὶον. (c) ἐνταῦθα ἡ αἲξ θέουσα συνεχὲς ἀφανὴς ἐγίνετο πολλάκις καὶ τὸν ἔριφον ἀπολιποῦσα τῷ βρέφει παρέμενε. (d) φυλάττει > τὰς διαδρομὰς ὁ Λάμων οἰκτείρας ἀμελούμενον τὸν ἔριφον, (e) καὶ μεσημβρίας ἀκμαζούσης κατ᾽ ἴχνος ἐλθὼν ὁρᾷ τὴν μὲν αἶγα πεφυλαγμένως περιβεβηκυῖαν, μὴ ταῖς χηλαῖς βλάπτοι πατοῦσα, τὸ δὲ ὥσπερ ἐκ μητρῴας

66

67

68

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. 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.’ The present participles (ἤδη) οὖσαι and προσκειμένων belong also to Phormio’s Substitutionary 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. 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, subjunctives, 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 summarizes 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 παιδίον: 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, 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 πρόσεισιν 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 infantboy.’ 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 elements, 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

70

71

‘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 comings 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). 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. 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).

366

162

chapter 7

I conclude this part with two texts where commentators and translators have seriously misunderstood the text by not recognizing the Substitutionary Perception function of the imperfect. First, the proem of Hesiod’s Theogony.

367

(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 ἐννύχιαι στεῖχον περικαλλέα ὄσσαν ἱεῖσαι (Hes. Th. 1–10)72 | After omnitemporal ἔχουσιν in line 2, ὀρχεῦνται in 4, and gnomic (= omnitemporal) ἐνεποιήσαντο and ἐπερρώσαντο in lines 7 and 8, which are due to the omniscient 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 solutions. 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 publication (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 secondary endings. But this injunctive use does not exist. Cp. Beekes (1995: 245): 72

73

‘Let us begin to sing from the Heliconian Muses, who possess the great and holy mountain 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 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 thick invisibility, by night they walked, sending forth their very beautiful voice …’ (Translation Most, except for στεῖχον in l. 10). 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 character 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 narrative 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 75

76

Or perhaps rather the singer of lines 1–8, for strictly speaking there has so far been no narrative. 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). ‘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).

368

164

369

chapter 7

scribed by means of present and (gnomic) aorist indicatives.77 But in between 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, … ἀναγκάζει … 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, χρεμετίζων, 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

370

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 English translations: |

77

78

Although this is a description of non-past, omnitemporal and therefore iterative (mythical) 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 particularized (τότε), as in a ‘real’ narrative. ‘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 forward … 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 shamelessly’, 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, 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 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/seeing 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 Substitutionary Perception; moreover, Aeneas, too, is covered by a cloud. | (37) ‘o fortunati, quorum iam moenia surgunt!’ Aeneas ait et fastigia suspicit urbis. infert > se saeptus nebula (mirabile dictu)

371

166

chapter 7

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, reginam opperiens, … (Verg. Aen. 1.437–454)79 |

440

372

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 perception 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

80

‘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). 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): Substitutionary 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 perception (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 | discussed 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 narrative 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 perceptions, 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 character, 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 experiencing, 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).

373

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 racconto 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 (Berlin/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. Brisard (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 Reference 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 substantial portion of them start with an ἐπεί-clause. It is demonstrated that, being presupposition-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 information 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 onstage 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 ‘messenger speeches’.*,1 In these a character who is generally called an ἄγγελος ‘messenger’ 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., respectively.—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 translation ‘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 persons 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. |

2

Survey of Content

Of a total of 24 messenger speeches in the tragedies,3 15 open, either immediately 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 difference in content. My argument proceeds in the following manner: in section 3 I describe three characteristics of ἐπεί-clauses at the beginning of a messengerspeech, 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.), Andromache (1085ff.), Heracles (923ff.), Ion (1122ff.), Electra (774 ff.), Iphigenia in Tau3 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 messenger speeches in Phoenissae: see below, note 13. 4 Kamerbeek, in his commentary ad S. OT 1241, draws attention to similar openings in Sophocles 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 Herakles’ one-line report in 1140.

294

172

295

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 ἐπείclauses, 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 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 messenger 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, however, 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 cho7 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 continuative, 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 Chicagotranslation’. At Ion 1122 and Ba. 1043 where the Chicago-translation has no temporal conjunction, 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 … ἥκουσαν). 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 messenger is referring to than anyone else, | since it is Medea herself who sent the children away.10 Heraclidae 800ff. The servant, to Alcmene: ‘When we had drawn our own troops up …’ The ἐπείclause 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 journey 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 11

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. 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 precedes the main clause. This ‘dislocation’ of the ἐπεί-clause may be due to the contrastive μὲν … δέ structure of the passage as a whole.

296

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 prepare the banquet (651–668).

297

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 Electra 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 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 himself, 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 presence 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 shining 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]. | 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 reference 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 concerns 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 second question, there are two exceptions: in Phoenissae I and III, the addressee

13

14 15

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. This is properly speaking the sequel to lines 1356–1424. But again the subject changes. The messenger’s failure to make known in the preceding lines the main burden of his news, that Iphigenia is safe, is exceptional.

298

176

299

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 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. Hippolytus’ 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 Talthybius.17 | 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 messenger’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

18

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 …’ 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].

300

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;

301

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 resembles 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 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 trimeters, 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 20 21

So also Erdmann (1964: 173–174). But which does not occur on stage. 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 ἐπείclause but with a simple declarative sentence, can be shown to open with completely new information, which does not pick up information | that has 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 person 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 circumstances 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 23

Not the first speech of the IT, see p. [297]. Of course this can be seen as a continuity-device.

302

180

303

304

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 indication 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 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 Hyllus in the battle. A θεράπων is also present, witness e.g. vv. 720ff. If we are dealing with the same θεράπων on both occasions, the reference to 709 ff. is easy.25 |

24

25

A slight inconsistency remains: ἤλθομεν in 1530 indicates that the messenger accompanied 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). 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) proposed 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 messenger 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 principal 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 knowledge 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 ὁπαδὸς 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 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 27 28

Cf. also 1329 ἡμᾶς μέν, οὓς σὺ … συμπέμπεις. The event referred to in the ἐπεί-clause here lies outside the play, cf. p. [296]. 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.

305

182

306

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 implicitly. (1) The ἐπεί-clause refers to Menoeceus’ suicide, which we are to suppose 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 something 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 not the same as the messenger of 1067ff., | the latter being Eteocles’ armourbearer (cf. 673–674).29 Yet this is not indisputable: one indication to the contrary might be that the second messenger in v. 1461 speaks of Eteocles as δεσπότην 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 possibly 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 difficulty, 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 expression 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 Xuthus and Ion. But how does the servant know that the chorus knows? Apparently 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 requirement, that the messenger has to know where to begin, is satisfied. In three 30 31 32

Of course, this does not hold for the audience, since for them the story broke off at just this point. Note that until v. 1194 (ἐπίμπλαμεν) there is no explicit indication that the slave has witnessed the events he is describing. Fourteen, not fifteen, since the first speech of the IT has a position of its own.

307

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’, meaning 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 messenger 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 Reference 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 establishes 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 identified 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 principle, 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 suggested 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 inconsistencies 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 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.

242

Thalmann 1984: 227, fn. 9

I will now review the relevant tense forms and adverbs in the form of a running 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.

5

10

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 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 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 characteristic 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. | 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 creates 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 hortative 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, Dichterweihe; 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.

243

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.

244

2 ἔχουσιν, 4 ὀρχεῦνται ‘Omnitemporal’ (Faulkner 2005: 66), rather than ‘timeless’ (e.g. Thalmann following West, see above), ‘eternal’ (Stoddard 2004: 130) states of affairs. ‘Timeless’ 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 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 omniscient 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 possession 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 temporally 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 nettement 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 reinterpreted 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 φθονήσας φόβον ἐμβάλη ἢ βροντήν, δι’ ὦν ἐφθάρησαν ἀναξίως ἑωυτῶν. | 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 ‘gnomiccum-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 Hesiod (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 (…) στρωφῶ9 Smith (2002: 71) uses the term ‘inferred perception’ for this phenomenon, and calls the sentences 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.’

245

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

130

246

ἐν δὲ δύω κρῆναι ἡ μέν τ᾽ ἀνὰ κῆπον ἅπαντα σκίδναται, ἡ δ᾽ ἑτέρωθεν ὑπ᾽ αὐλῆς οὐδὸν ἵησι πρὸς δόμον ὑψηλόν, ὅθεν ὑδρεύοντο πολῖται. 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 evaluative 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 11

Cp. De Jong (2001: 176–177) on Od. 7.81–135. 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 pondering 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 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 activities 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. | 12

13 14

15

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. For the idea cp. A.R. 4.1381 Μουσάων ὅδε μῦθος, ἐγὼ δ᾽ ὑπακουὸς ἀείδω | Πιερίδων. West, in his commentary ad locum, considered στεῖχον a ‘typifying’, that is to say a timeless, 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 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 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 meaning 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, ingenious but not quite convincing, analysis of the beginning of h. Ap. see Bakker (2002). Cp. Clay (2003: 54): ‘In effect, the imperfect στεῖχον conveys precisely and vividly the

247

248

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] signale 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 happen 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

16

17

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 activity, 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 Εὐήνιος 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 ἀνύσσας. Pucci (2007: 43–44) confines himself to enumerating the various interpretations of lines 9–10. 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 probably 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). 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 ἔμμοροί εἰσι καὶ αἰδοῦς, οὕνεκ᾽ ἄρα σφέας | οἴμας Μοῦσ᾽ ἐδίδαξε, φίλησε δὲ φῦλον ἀοιδῶν, also 8.488.

249

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 follow (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 πρῶτος κελόμην θεὸν ἱλάσκεσθαι· | Ἀτρεΐωνα δ᾽ ἔπειτα χόλος λάβεν (…), χόλος λάβεν describes Agamemnon’s reaction to Achilles’ κελόμην. The aorist lacks this feature, 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).

250

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, however, that the narrative is not yet finished, for the poet resumes it at line 68, quo vide.

251

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 subjunctive 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 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

20 21

22

Whatever the exact meaning of ταῦτα περὶ δρῦν ἢ περὶ πέτρην, it can hardly stand for speaking 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). Pucci (2007: 76) mainly discusses its use in the Iliad. Most translates ‘come then’, but τύνη is not a hortative particle like ἄγε. 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. 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 interpretation 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 relative 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 digression 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 εἰσαναβαίνων.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 24 25

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.’ 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. 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 παλίσκιον, ἔνθα Κρονίων | νύμφῃ ἐϋπλοκάμῳ μισγέσκετο νυκτὸς ἀμολγῷ, | ὄφρα κατὰ γλυκὺς ὕπνος ἔχοι λευκώλενον Ἥρην.

252

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

253

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 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 presentarsi 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

27 28 29

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 immediately 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 καλλιστευομένα 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. 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? ‘They then went towards Olympus, exulting in their beautiful voice, and around them the black earth resounded as they sang’ (translation Most, adapted). 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 summit 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 interpretations 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)

45

αὐτάρ ἐπεὶ δὴ πάντα περὶ χροΐ κόσμον ἔθηκαν 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

45

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, δεῖξε δὲ κοῦρον ἑόν·31

In this case, too, the newborn god is not active himself, cp. lines 42 and 45: it is Hermes who brings Pan to the assembly of the gods.

30

31

‘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 …’ (translation Sargent). ‘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).

254

198

255

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. ἐδίδαξαν (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» (notamment chez soi), rarement aus sens de «venir»’, Lexikon des frühgr. Epos s.v. νίσομαι: ‘“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

33 34

35

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). According to Verdenius τότε refers to στεῖχον (line 10), but this view is less likely, since it ignores the presence of ἐδίδαξαν. 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’. ἐπῆλθε at line 256 merely repeats (σχεδόθεν (…)) ἤλυθεν (line 223), there is no temporal progression. 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 soliloquies 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᾽ (…) ἄειδον turns lines 71–74 into songs heard by Hesiod, while ultimately ταῦτα also comprises the song of the Muses at lines 11–21.36 | 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.

36

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. 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.

256

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 discussion 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.

257

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 audience (…).’ 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 element 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 phrasing 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 adjective: 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:

100

εἰ γάρ τις καὶ πένθος ἔχων νεοκηδέι θυμῷ ἄζηται κραδίην ἀκαχήμενος, αὐτὰρ ἀοιδὸς 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 conditional 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., however 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. Denniston (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 singular 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.

258

202

259

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 conditional 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, 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 ἔχουσιν, 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 ἀοιδός (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 | 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.’

260

204

chapter 9

Conclusion

261

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 interdependencies. Thus, the imperfect verb forms which express a sound (ἄειδον, line 75), or are accompanied by verb forms expressing a sound (στεῖχον περικαλλέα 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 recurrent 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 ἐόντων; | – 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 Γιγάντων | 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 ἀοιδήν· | 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) (…) πορεύεσθον (…), 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 συντεθέντος 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 translations: 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 + subjunctive, 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 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: ‘now/seeing that they were close, it lowers its head and (…) pulls shamelessly.’ |

The Structure of the Proem of Hesiod’s Theogony

5

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 καί τε λοεσσάμεναι τέρενα χρόα Περμησσοῖο

1 Metatextual formula 2–8 Hymn to the Muses part I attributive section: description of some characteristic activities of the Muses on Mount Helicon (omnitemporal presents: 2 ἔχουσιν, 4 ὀρχεῦν-

262

263

206

10

15

20

25

30

35

40

45

50

ἀκροτάτῳ Ἑλικῶνι χοροὺς ἐνεποιήσαντο, καλοὺς ἱμερόεντας, ἐπερρώσαντο δὲ ποσσίν. ἠ᾽ Ἵππου κρήνης ἠ᾽ Ὀλμειοῦ ζαθέοιο ἔνθεν ἀπορνύμεναι κεκαλυμμέναι ἠέρι πολλῷ ἐννύχιαι στεῖχον περικαλλέα ὄσσαν ἱεῖσαι, ὑμνεῦσαι Δία τ᾽ αἰγίοχον καὶ πότνιαν Ἥρην Ἀργείην, χρυσέοισι πεδίλοις ἐμβεβαυῖαν, κούρην τ᾽ αἰγιόχοιο Διὸς γλαυκῶπιν Ἀθήνην Φοῖβόν τ᾽ Ἀπόλλωνα καὶ Ἄρτεμιν ἰοχέαιραν 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 Γαῖάν τ᾽ Ὠκεανόν τε μέγαν καὶ Νύκτα μέλαιναν ἄλλων τ᾽ ἀθανάτων ἱερὸν γένος αἰὲν ἐόντων. αἵ νύ ποθ᾽ Ἡσίοδον καλὴν ἐδίδαξαν ἀοιδήν, ἄρνας ποιμαίνονθ᾽ Ἑλικῶνος ὕπο ζαθέοιο. τόνδε δέ με πρώτιστα θεαὶ πρὸς μῦθον ἔειπον, 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, 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, σφᾶς δ᾽ αὐτὰς πρῶτόν τε καὶ ὕστατον αἰὲν ἀείδειν. ἀλλὰ τίη μοι ταῦτα περὶ δρῦν ἢ περὶ πέτρην; τύνη, Μουσάων ἀρχώμεθα, ταὶ Διὶ πατρὶ ὑμνεῦσαι τέρπουσι μέγαν νόον ἐντὸς Ὀλύμπου, εἴρουσαι τά τ᾽ ἐόντα τά τ᾽ ἐσσόμενα πρό τ᾽ ἐόντα, φωνῇ ὁμηρεῦσαι, τῶν δ᾽ ἀκάματος ῥέει αὐδὴ ἐκ στομάτων ἡδεῖα· γελᾷ δέ τε δώματα πατρὸς Ζηνὸς ἐριγδούποιο θεᾶν ὀπὶ λειριοέσσῃ σκιδναμένῃ, ἠχεῖ δὲ κάρη νιφόεντος Ὀλύμπου δώματά τ᾽ ἀθανάτων· αἱ δ᾽ ἄμβροτον ὄσσαν ἰεῖσαι θεῶν γένος αἰδοῖον πρῶτον κλείουσιν ἀοιδῇ 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· αὖτις δ᾽ ἀνθρώπων τε γένος κρατερῶν τε Γιγάντων ὑμνεῦσαι τέρπουσι Διὸς νόον ἐντὸς Ὀλύμπου Μοῦσαι Ὀλυμπιάδες, κοῦραι Διὸς αἰγιόχοιο.

chapter 9 ται, gnomic aorists: 7 ἐνεποιήσαντο, 8 ἐπερρώσαντο) passing into a narrative 9–34 Personal narrative: Hesiod hears the Muses singing on the slopes of Mount Helicon (imperfect + verb of sound: 10 στεῖχον (…) ὑμνεῦσαι)

(…) they bestow on him the ability to sing (Dichterweihe) (aorist: 22 ἐδίδαξαν ἀοιδήν) …

33 (…) and urge him to use his newly acquired ability to sing of the gods and first of all of themselves (imperfect: 33 ἐκέλοντο) 35–36 Metatextual formulas 36 ταί—67 Hymn to the Muses part II attributive section: description of the hymnic activities of the Muses on Olympus, who sing of the gods and of mankind (omnitemporal presents: 37 τέρπουσι, 39 ῥέει, 40 γελᾷ, 42 ἠχεῖ, 44 κλείουσιν, 51 τέρπουσι, 66 μέλπονται, 67 κλείουσιν) [264]

discourse cohesion in the proem of hesiod’s theogony

55

60

65

70

75

80

85

90

95

τὰς ἐν Πιερίῃ Κρονίδῃ τέκε πατρὶ μιγεῖσα Μνημοσύνη, γουνοῖσιν Ἐλευθῆρος μεδέουσα, λησμοσύνην τε κακῶν ἄμπαυμά τε μερμηράων. ἐννέα γάρ οἱ νύκτας ἐμίσγετο μητίετα Ζεὺς νόσφιν ἀπ᾽ ἀθανάτων ἰερὸν λέχος εἰσαναβαίνων· ἀλλ᾽ ὅτε δή ῥ᾽ ἐνιαυτὸς ἔην, περὶ δ᾽ ἔτραπον ὧραι μηνῶν φθινόντων, περί δ᾽ ἤματα πόλλ᾽ ἐτελέσθη, 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 ἐν θαλίῃς· ἐρατὴν δὲ διὰ στόμα ὄσσαν ἰεῖσαι μέλπονται, πάντων τε νόμους καὶ ἤθεα κεδνὰ ἀθανάτων κλείουσιν, ἐπήρατον ὄσσαν ἱεῖσαι. αἳ τότ᾽ ἴσαν πρὸς Ὄλυμπον, ἀγαλλόμεναι ὀπὶ καλῇ, ἀμβροσίῃ μολπῇ· περὶ δ᾽ ἴαχε γαῖα μέλαινα ὑμνεύσαις, ἐρατὸς δὲ ποδῶν ὕπο δοῦπος ὀρώρει νισομένων πατέρ᾽ εἰς ὅν· ὁ δ᾽ οὐρανῷ ἐμβασιλεύει, αὐτὸς ἔχων βροντὴν ἠδ᾽ αἰθαλόεντα κεραυνόν, κάρτει νικήσας πατέρα Κρόνον· εὖ δὲ ἔκαστα ἀθανάτοις διέταξε νόμους καὶ ἐπέφραδε τιμάς. ταῦτ᾽ ἄρα Μοῦσαι ἄειδον Ὀλύμπια δώματ᾽ ἔχουσαι, ἐννέα θυγατέρες μεγάλου Διὸς ἐκγεγαυῖαι, Κλειώ τ᾽ Εὐτέρπη τε Θάλειά τε Μελπομένη τε Τερψιχόρη τ᾽ Ἐρατώ τε Πολύμνιά τ᾽ Οὐρανίη τε Καλλιόπη θ᾽· ἡ δὲ προφερεστάτη ἐστὶν ἁπασέων. ἡ γὰρ καὶ βασιλεῦσιν ἅμ᾽ αἰδοίοισιν ὀπηδεῖ. ὅντινα τιμήσουσι Διὸς κοῦραι μεγάλοιο γεινόμενόν τε ἴδωσι διοτρεφέων βασιλήων, τῷ μὲν ἐπὶ γλώσσῃ γλυκερὴν χείουσιν ἐέρσην, τοῦ δ᾽ ἔπε᾽ ἐκ στόματος ῥεῖ μείλιχα· οἱ δέ νυ λαοὶ πάντες ἐς αὐτὸν ὁρῶσι διακρίνοντα θέμιστας ἰθείῃσι δίκῃσιν· ὁ δ᾽ ἀσφαλέως ἀγορεύων αἶψά τι καὶ μέγα νεῖκος ἐπισταμένως κατέπαυσε· τούνεκα γὰρ βασιλῆες ἐχέφρονες, οὕνεκα λαοῖς βλαπτομένοις ἀγορῆφι μετάτροπα ἔργα τελεῦσι ῥηιδίως, μαλακοῖσι παραιφάμενοι ἐπέεσσιν· ἐρχόμενον δ᾽ ἀν᾽ ἀγῶνα θεὸν ὣς ἱλάσκονται αἰδοῖ μειλιχίῃ, μετὰ δὲ πρέπει ἀγρομένοισι. τοίη Μουσάων ἱερὴ δόσις ἀνθρώποισιν. ἐκ γάρ τοι Μουσέων καὶ ἑκηβόλου Ἀπόλλωνος 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

207

53–62 Mousogony: narrative digression within the attributive section (non-gnomic aorists: 53 τέκε, 60 ἔτεκε, iterative imperfect: 56 μίσγετο)

67 End of attributive section on the Muses 68–79 Narrative about Hesiod and the Muses, continued: τότε refers back to 22 ἐδίδαξαν, return of the Muses to Olympus (their abode, 63) as perceived by the poet (reappearance (cp. 10) of imperfects and verbs of sound: 68 ἴσαν (…) ἀγαλλόμεναι ὀπὶ καλῇ, 69–70 ἴαχε γαῖα (…) ὑμνεύσαις, 70 δοῦπος ὀρώρει, 75 ἄειδον) 71–74 Interrupted by brief attributive section on Zeus (omnitemporal present: ἐμβασιλεύει) + brief narrative digression (nongnomic aorists: διέταξε, ἐπέφραδε) [265] 79–80 Brief attributive section on Calliope, the most prominent of the Muses (omnitemporal presents: 79 ἐστίν, 80 ὀπηδεῖ), which triggers an attributive section-like description of the Muses’ activities among mortals through kings (81–92) and singers (94–103); like the Muses, the latter sing of mankind and the gods (verbless clauses: 88 ἐχέφρονες, 93 τοίη, 96 ὄλβιος; omnitemporal presents: 83 χείουσιν, 84 ῥεῖ, 85 ὁρῶσι, 89 τελεῦσι, 91 ἱλάσκονται, 92 πρέπει, 95 ἔασιν, 97 ῥέει, 102 ἐπιλήθεται, 103 μέμνηται; gnomic aorists: 87 κατέπαυσε, 103 παρέτραπε; generic future: 101 ὑμνήσει)

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· 105 κλείετε δ᾽ ἀθανάτων ἱερὸν γένος αἰὲν ἐόντων, οἳ Γῆς ἐξεγένοντο καὶ Οὐρανοῦ ἀστερόεντος, Νυκτός τε δνοφερῆς, οὕς θ᾽ ἁλμυρὸς ἔτρεφε Πόντος. εἴπατε δ᾽ ὡς τὰ πρῶτα θεοὶ καὶ γαῖα γένοντο καὶ ποταμοὶ καὶ πόντος ἀπείριτος οἴδματι θυίων 110 ἄστρά τε λαμπετόωντα καὶ οὐρανὸς εὐρὺς ὕπερθεν· [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.

104–115 Salutation and prayer: having executed the second part of the order of lines 33– 34, viz. to sing of the Muses, Hesiod now salutes the Muses and asks them to give him ἀοιδή so that he can execute the first part of their order, viz. to sing of the gods and their birth

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. Hasselgård, 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 (Baltimore/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, accompanied 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) Smp. 3d person narrative (Aristodemus) embedded in 1st person narrative (Apollodorus) Euthd. 1st person narrative (Socrates) Prt. 1st person narrative (Socrates) total

244 93 (φάναι: 57)

3 11

63 10

8 17

201 128 666

52 40 106

39 8 120

66 43 134

1 These consist of dialogues with an introductory ‘dramatic’ conversation, plus a narrated philosophical discussion, so-called ‘framed’ dialogues (Phaedo, Symposium, Euthydemus, Protagoras), and dialogues with an introduction that is part of the narrative: Parmenides, Charmides, 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 (Griswold 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

211

on the syntax and pragmatics of inquit formulae (cont.)

narrative throughout

ἔφη

ἔφην ἦ δ᾽ ὅς/ἥ ἦν δ᾽ ἐγώ

Prm. 3d person narrative (Pythodorus), that is embedded in another 3d person narrative (Antiphon), which is embedded in a 1st person narrative (Cephalus) Chrm. 1st person narrative (Socrates) Ly. 1st person narrative (Socrates) R. 1st person narrative (Socrates)

24 (φάναι: 43)







88 82 1064

19 2 76

21 13 188

60 54 601

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 | distribution of the inquit formulae. But it is of course the way they function in context, 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 formula. 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.

399

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 Hippothales … 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 enclosure 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.)

400

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 number 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 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

213

on the syntax and pragmatics of inquit formulae

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 combination 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 participle, 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 discussion. 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 looking for further tests to establish a difference between the two inquit formulae. |

401

ἀληθῆ, ἀληθέστατα (λέγεις), inquit

3.2

ἔφη

ἔφην

ἦ δ᾽ ὅς

ἦν δ᾽ ἐγώ

BARE ἀληθῆ (λέγεις)

Phd. Prm. Smp.

12 1 (φάναι 2) – (φάναι 1)

– – –

– 2 2 3

– – 1 (Socrates to Diotima) – – – –

3 26 2

Chrm. Ly. Euthd. Prt.

1 (Phaedo) – 3 (Socrates to Diotima) – – – 1 (Socrates to Protagoras)

– – – –

4 9 1 1

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.)

R.

ἔφη

ἔφην

ἦ δ᾽ ὅς

ἦν δ᾽ ἐγώ

BARE ἀληθῆ (λέγεις)

54



5 332a6, 395b7, 444a3, 463e7, 495b7 5

3 372c4, 430c1 467b7

4 334d2, 459c7 475b10, 601c14

4

50

TOTAL 74 5 (without φάναι)

402

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, 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 completed, 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, formally 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 continuative 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 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 meaning and denote a complete speech act;13 (3) (pragmatics) by the combination

10

11

12

13

the aorist view of ἔφη(ν). For the meaning development of φημί (orig. ‘proférer des paroles magiques ou sacrées’) see Fournier (1946: 8 ff.). 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.’ 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). 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 precedes (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. Ameis-Hentze-Cauer, in their commentary on Iliad 1.219, take ἦ in Homer, which occurs

403

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. Εἰἑν, inquit

3.3

Εἰἑν, ἔφη Εἰἑν, φάναι Εἰἑν, ἔφην Εἰἑν, ἦ δ᾽ ὅς Εἰἑν, ἦ δ᾽ ἐγώ BARE Εἰἑν Phd. Prm. Smp. Chrm. Ly. Euthd. Prt. R. TOTAL

14

15

16

3 – – – – – – I16 4

– 1 1 – – – – — 2

– – – – – – – – –

1 – – – – – – – 1

– – 2 2 6 5 1 10 26

2 1 – 1 – – 3 6 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 »’. 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. 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 something 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 questioning but provides a comment on the way of questioning. In other words: the questioning itself comes temporarily to a stand-still. 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 readily 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 discussion, for non-continuative ἦν δέ clears the way for introducing a new subject 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 διωμολογησάμεθα 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 διωμολογησάμεθα; note also that the rounding-off function of εἰἑν is here reinforced by τοῦτο … οὕτω κείσθω.

17 18

For the aspirated ε cp. Chantraine 1968–1980: s.v. 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.

405

218

chapter 10

3.4

τοίνυν, inquit

Phd. Prm. Smp. Chrm. Ly. Euthd. Prt. R. TOTAL

406

τοίνυν, ἔφη

τοίνυν, ἔφην

τοίνυν, ἦ δ’ ὅς

τοίνυν, ἦνδ’ ἐγώ

τοίνυν

9 – 2 4 – – 1 11 27

1 – – 2 – 2 2 3 504c9, 564b4, 584b1 10

5 – – 2 – 1 – – 8

– – – 7 1 2 1 53 64

5 5 2 4 1 4 44 65

Note again the preponderance of ἦν δ᾽ ἐγώ. As a matter of fact, τοίνυν, ἦν δ᾽ ἐγώ, 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

20

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). For the use of the concept ‘move’ in discourse analysis see Kroon (1995: 65–66).

on the syntax and pragmatics of inquit formulae

4

219

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, 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, spoken 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 physical 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᾽ ὁπωστιοῦν 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

407

220

chapter 10

ἅμα λέγων ταῦτα καθῆκε τα σκέλη ἐπὶ τὴν γῆν, καὶ καθεζόμενος οὕτως ἤδη τὰ λοιπὰ διελέγετο. c8 τοίνυν ἔφη καὶ δ

408

61d3 Ἤρετο οὖν αὐτον ὁ Κέβης· Πῶς τοῦτο λέγεις, ὦ Σωκρατες, τὸ μὴ θεμιτὸν εἶναι ἑαυτὸν βιάζεσθαι, ἐθέλειν δ᾽ ἂν τῷ ἀποθνῄσκοντι τὸν φιλόσοφον 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 ἀποκτεινύναι, 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 today; 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 something. (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 Simmias 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 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 τοίνυν (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).

409

222

410

chapter 10

proper, and this, I suggest, is why they do not have inquit formulae. Observe 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 Literature (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 arbitraire, mais qu’elle exprime des différences sémantico-pragmatiques spécifiques. Ces différences portent sur la ‘saillance’ du (référent du) nom aussi bien que des informations 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 importance 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 different 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

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 semantic-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 previous sentence; there is, thus, asyndeton. In such cases both the Topic and the contents 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 situation 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. (2)

120

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 asyndétiquement 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 + substantif), 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 arbitraire mais exprime des différences sémantico-pragmatiques bien spécifiques. Mis à part quelques observations générales sur la syntaxe de ces groupes anaphoriques, 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 (combinaisons 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 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 ἀναθήματα 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 γνώμην 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.

121

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), puisqu’ils écrivent, 2, 343 γ: ‘Sehr häufig aber (wird das Asyndeton angewendet) auch bei einem Demonstrative, das auf einen vorhergehenden Begriff zurückweist, wo keine Folgerung, sondern bloss eine Fortsetzung der Rede ausgedrückt 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 + substantif + 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 anaphorique 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 + substantif. |

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 habitant 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.

123

228

chapter 11

Le groupe οὗτος ὁ Κροῖσος vient immédiatement après l’ introduction de Crésus 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 Crésus 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 participant, 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 taille… Ils revêtirent cette femme… |

124

9 10

Pour les différentes façons dont un Topique Nouveau peut être introduit voir Dik (1989: 268ss.). 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 explicite). Il se sert de cette échelle pour définir la notion de ‘champ d’identification’ (‘identification 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 remarquer 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 Φραόρτεω. 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’ asyndète 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 ἀπόγονον 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 quatrième 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 introduite 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

125

chapter 11

tandis que l’absence de δέ signale à nouveau qu’ on n’ a pas affaire à une ‘Fortsetzung’, 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, respectivement), ainsi que toutes les informations données sur cette entité, sont détachées du contexte et obtiennent une certaine | indépendance. Par là, le narrateur 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)

12

13

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’. 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’. Pour être complet, j’ ajoute que l’ asyndète se trouve parfois dans des phrases qui commencent 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, ἕλκουσαν 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 offrandes abondantes; … Ce Gygès est le premier Barbare à notre connaissance qui ait consacré des offrandes à Delphes, après Midas fils de Gordias… 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 participant 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 employer une formule plus prudente, comme ‘pas plus saillante que les informations 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

15

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. 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 πρῶτος (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.

126

232

chapter 11

celles-là, au point de vue de la langue, diffèrent tout à fait des précédentes et s’accordent entre elles

127

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 narrateur 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. 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 ἐπεκαλέοντο 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 relativement peu d’intérêt. Quant à οὗτοι, il a ici, en raison de sa postposition, une fonction 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 informations ‘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 caractère 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 ἐκόμισαν, 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) | (20) ἑσσωθέντες δὲ τῇ συμβολῇ, ὅσοι αὐτῶν ἐζωγρήθησαν, πέδας τε ἔχοντες· τὰς ἐφέροντο 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 apportées, 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

17

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’arrièreplan 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. Dans le corpus examiné. Dans 1.187.3 ἦν ἀκίνητος (= οὐκ ἐκινέετο) n’exprime pas un état.

128

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 digressive, 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 fonction 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 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 toujours 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’, ‘précisément 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 + substantif + 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.

129

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 démontrer 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 propositions 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 indication 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 fundamental difference between ‘future’ ἐάν-clauses and ‘generic’ ἐάν-clauses. In clauses having 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

132

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 + subjunctive 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 possible 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 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’ arrêtent 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 traiter 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 subjonctif + 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 problème. 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 subjonctif + 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 Verwirklichung’. 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 réaliser 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 Verwirklichung’, 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 Ausdrücksweisen 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 ‘probabilité’ 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 manifestement 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 locuteur se trouve : c’ est cette situation qui l’ amène à s’ attendre à la réalisation de la condition.

133

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)

134

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 Schlussfolgerung’ 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 locuteur 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, Sandanis 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 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’ optatif, 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 conditionnelles 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 limitée dans le contexte. Voir par exemple: (3) ἐγὼ δὲ οἷόν τε ἐγίγνετο λέξω, καὶ ἀφ᾽ ὧν ἄν τις σκοπῶν, εἴ ποτε καὶ αὖθις ἐπιπέσοι, 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 observer 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 (violation 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 rhétoriquement 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 ἐλάσσονας, 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. | Un exemple net des différences entre le subjonctif et l’ optatif est :

135

242

chapter 12

(5) ΣΩ. Σχολὴ μὲν δή, ὡς ἔοικε· καὶ ἅμα μοι δοκοῦσιν ὡς ἐν τῷ πνίγει ὑπὲρ κεφαλῆς 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. Pragmatiquement 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 suffit 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 condition 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 temporelles 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. |

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 certain 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 comportement 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 question. 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 conditionnelles et les temporelles est souvent ignorée, voir la remarque suivante de

136

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 beaucoup 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)

137

Eh bien, mère, quand moi je serai grand, je mettrai l’ Égypte sens dessus 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 πείθωνται, 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, convaincus 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 considérer 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 Aristide, considère tout naturellement l’exécution de son ordre comme allant de soi. Une fois de plus ἤν serait étrange. L’incertitude exprimée par ἢν ‘disjonctif’ 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 probable, puisqu’il vient de demander à Aristide d’annoncer lui-même les bonnes nouvelles. Les conditionnelles ἢν μὲν πείθωνται… ἢν δὲ αὐτοῖσι μὴ πιστὰ γένηται 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 luimême; 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 conditionnelle, 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 3

chapter 12

Ἐάν et ἐπεάν ‘génériques’: ‘réalisation occasionnelle’ et ‘réalisation nécessaire’

Dans l’analyse que je viens de donner de l’emploi du subjonctif dans les conditionnelles 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 conditionnelles)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 cependant 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 ἀπίοιμεν 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 réalisation 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 pratique 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 conjonctions 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 cargaison, 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) | 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 spéculations 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 conditionnelles 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 nullement 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 subjonctif 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.

139

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)

140

Du coup, tout ce qui est dit après le prédicat générique ποιεῦσι, que l’ on pourrait 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 συλλέξωνται, mais également pour les participes, p. ex. le participe aoriste κόψαντες. 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 raisonnement 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 postériorité. Reste à savoir sous quel rapport les conditionnelles se distinguent des temporelles dans ce type de discours. Tout d’abord on constate que ces subordonnées 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 possibilité. 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 certaine, 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 fonctionne 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 nécessairement, 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 contingence 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 appeler 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 ‘quantification relative non proportionnelle’ est employée par Dik (1989: 153) dans son analyse des cas comme: | 6 Il est intéressant de noter que Powell, dans son Lexicon s.v. ἤν 2, parle de ‘repeated contingencies in the present’.

141

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 illustrées par deux autres phénomènes. Comme ἐάν réfère à une situation disjonctive, 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, ἀποσιωσαμένη 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. | 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 vaudrait 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 subjonctif 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 rupture 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. |

142

143

252 4

chapter 12

Conclusion

En guise de conclusion, je présente le schéma suivant :

type de subordonnée conditionnelles: trait commun : décrivent une situation disjonctive : la condition peut ou non se réaliser

temporelles : traits communs: – établissent un lien temporel; – présupposent en principe l’état de choses de la subordonnée

type de discours

εἰ + ind. fut.

εἰ + opt.

ἐάν + subj.

ἐπεάν, ὅταν + subj.

discours (in)direct : la subord. porte sur l’avenir

cond. neutre ; le locuteur ne s’ exprime pas sur la probabilité de réalisation de la condition

la réalisation de la condition est simplement possible

la réalisation de la condition relève d’une forte possibilité

réalisation de l’état de choses quasi-certaine (effet combiné de la valeur semiprésuppositionnelle de la conjonction et de la valeur de ‘forte possibilité’ du subjonctif. N.B. Puisqu’il s’agit de l’avenir, la conjonction ne peut avoir ici qu’une valeur semi-présuppositionnelle)

descriptions : l’état de choses de la subordonnée fait partie d’un état de choses hyperonyme

manque

manque

réalisation possible implique: l’ état de choses de la subordonnée se réalise parfois pendant la durée de l’ état de choses hyperonyme (‘quantification non exhaustive’)

l’état de choses se produit à toute occasion possible pendant la durée de l’état de choses hyperonyme (‘quantification 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 (Dordrecht 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 construction 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 comments. 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 possessions. 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 number 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 expressions 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 discussion 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 plural (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 standard 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 construction 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

1

Ἄχεα etc. as Object

1.1

Constructions with ἔχειν

257

(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 locative 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 speaking. 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 speaking 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 speaking) 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. Presumably, 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, spoken by Andromache, referring to herself). These possessive constructions are rather remarkable, because, as we will see later, sg. ἄχος usually is a ‘schmerzlicher 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

10

The plural κακά, too, occurs as object of διδόναι, however not with a god but with the stomach 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). A unique expression is found at Od. 19.167–168: ἦ μέν μ᾽ ἀχέεσσί γε δώσεις | πλείοσιν ἢ ἔχομαι (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 represented 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 1.3

chapter 13

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 sorrow 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

12

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’). 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 largescale 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 constructions 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 altogether 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 speaking, 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

14

15

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. 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.). 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: Odysseus) 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 speaking) 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

17 18

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. Cp. Od. 11.115: δήεις δ᾽ ἐν πήματα οἴκῳ (‘In your house you’ll find misery’). Mawet’s (1979: 100) remark: ‘Πῆμα figure encore comme complément de verbes transitifs 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 metrical 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

19

nature, examples being εὑρίσκειν and δήειν), shows that she does not realise that these uses strongly point to πῆμα having a concrete nature. 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 different: 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 expression plus abstraite, plus intellectuelle de la souffrance qu’ ἄλγος [read: ἄλγεα]’. Refining her observation, by making use of the semantic-ontological terms discussed 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 unobservable and cannot be said to occur or to be located either in space or time. Thirdorder 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 uncertainty 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 (second-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

22

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). 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 speaking about his vicissitudes at sea)24 and my dear heart rejoiced, ill-fated me, for I would still meet with much hardship.

23 24

Cp. e.g. Il. 13.416: οἱ ὤπασα πομπόν (‘I have sent him an escort’). 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 Eumaios)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 ‘mourning’.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 grudgingly 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), τειρόμενος (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

26

And also by that of a god, Od. 4.35. Cp. also Od. 4.812 καί με κέλεαι παύσασθαι ὀϊζύος ἠδ᾽ ὀδυνάων (‘and you bid me cease from my grief and the many sorrows’), spoken by Penelope to Athena-Iphthime. 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 brilliant 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. μυθήσασθαι 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-WestHainsworth 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 addressing 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 therefore 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 (43)

chapter 13

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 speaking; 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 untakeable, not to be endured (…) They went straight for the Danaans … As subject of λαβεῖν, ἑλεῖν occur: ἄλγος, ἄχος, πένθος. These constructions express 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 (51)

With Verbs of Reaching32 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 speaking; acute physical pain) She saved me when I suffered much at the time of my great fall 31 32

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. 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 entities, in the second case as (semi-) first-order entities. De Boel (1988: 89ff.) argues, convincingly, 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 ἀκαχήμενον. 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 cannot 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 feeling)) Oh, for shame. Great sorrow comes on the land of Achaia.

33

34

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. 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, elsewhere 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 ἱκάνει. 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 phenomenon. 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 36

I have put πῆμα between parentheses, because this is a dubious example. See note 6. 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 πεῖραρ 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

3 1.

2.

3.

4.

5. 6. 7. 8.

9.

275

Main Conclusions 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 excellence 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. 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. 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 concerned is typically caused by an event. Again, πῆμα is not used in this way. 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. Ἄλγος, κῆδος, ἄχος, πένθος, ὀϊζύς, πόνος may occur as the subject of a verb of reaching, mostly expressing a lasting feeling. Πῆμα and κακόν may be the subject of a verb of approaching. Κακόν/κακά are used as replacers for the more specific terms. Ἄλγος/ἄλγεα, ἄχος/ἄχεα, κῆδος, ὀϊζύς, πένθος, πήματα and πόνος are socalled second-order entities (abstract action nouns). Κήδεα are usually third-order entities (abstract ‘non-action’ noun); they are typically mental representations of ἄλγεα. Πῆμα comes close to being a first-order entity (concrete noun). Κακόν/κακά may be all three types of entities. 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 differences, that are not sufficiently acknowledged in the standard dictionaries 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 feeling, 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, possibly 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 perfect: ‘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, feeling’, 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): ‘vexing 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 differences 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 difference 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 πρότερον 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 presentstem 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 morphological categories (case, gender, number), or semi-developed noun phrases, | having 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), participles (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

442

280

443

chapter 14

properties.2 And thirdly there is the use where we find what looks like the definite 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 – 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 εἰρῆσθαι 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 surprisingly, of course, τὸ πρότερον must be considered adverbial if the argument | 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 ἐκείνου. (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 references; 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).

444

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

445

Τὸ πρότερον and πρότερον

Above (ex. (3)), I added ‘wrongly’ to Fowler’s translation ‘before’ of τὸ πρότερον, 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 distinction 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: ‘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

9 10

There are exceptions, though. Thus, Diès, in his Budé-edition, translates ‘lors de mon premier 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. Cp. also Bailly s.v. πρότερος ‘neutre adv. πρότερον … ou τὸ πρότερον … auparavant’. A typical example of this belief is Goodwin (1894: 206): ‘Proper names may take the article; 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 article, it seems a sound procedure to start from the assumption that its function 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 πρότερον 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 context. 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 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:

11 12 13 14

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. Cp. e.g. Mx. 242d, Th. 7.18.2. This use well illustrates the ‘countability’ of states of affairs in the aorist, cp. Armstrong (1981). Cp. the definition of ‘fois’ in Le Petit Robert: ‘moment du temps où un événement, conçu comme identique à d’ autres événements, se produit.’ 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.

446

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 ἔγνωσαν (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 operative. 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 theoretical issues involved here.18 | 16 17

18

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). 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. 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 important 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 referential 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-

448

286

chapter 14

I conclude this discussion with some other examples illustrating the differences.

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 …’,

19

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. ‘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 | Athenians … 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’.

449

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 ἐκαλέοντο … Στρυμόνιοι (Hdt. 7.75) These took the name of Bithynians after they crossed over to Asia; before that they were called … Strymonians |

450

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 Λυδοῦ 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 understood 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 ‘general 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 | Second, there is the morphological status of the ending -ον. Naturally, establishing 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 22

In their discussion of the ‘Ordnungszahlwörter’, Kühner-Blass mention πρότερος, but only in its use as ‘the first of two’ (1890: 1, 622). 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 connection with causal τί; ‘why?’, see Rijksbaron (1976: 10–18).

451

290

452

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 nominals, 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 governed by the preposition εἰς;25 for εἰς τὸ ἔπειτα see e.g. Smp. 193d, for ἑς τὸ αὐτίκα Th. 2.64.6. It is true | that the combination with εἰς is not found with τὸ πρότερον, 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 discussion 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 24 25

26

The discussion of the classification criteria in Pinkster (1972: 18–33) is still relevant for this matter. The meaning of these phrases is another matter, which will be ignored here. 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 languages, 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). Or is there a semantic reason, εἰς being oriented towards the future?—Theoretically, ὕστερον 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 προτεραία 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 ‘quantificational 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 following definition of ‘ordinators’: ‘By “ordinators” we understand such elements | as first, second, third, … last. These elements presuppose that the intended referent 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. Tedeschi 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 (Hannover 1890). Pinkster, H., On Latin Adverbs (Amsterdam 1972). 27

28

It is true that in Functional Grammar predication operators are confined to ‘the grammatical (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 elements, 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. Instead of ‘similar entities’ I would prefer ‘identical entities’. See also note 12.

453

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 Reference 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 combination καὶ … δέ. It defends the view of Denniston and Kühner-Gerth that the combination 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 considered 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 precedes.’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

188

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 saying 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 apodictic statement on p. 162: ‘Except in the apodotic use, δέ is always a connective.’ 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 unexplained. Denniston might also have turned to Kühner-Gerth for support. They mention 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 andererseits, 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, dagegen’, 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 KühnerGerth 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 |

(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 fitting … 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, sondern 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 numbers 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 discrepancies 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 formal 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 characteristic of most adverbs see Pinkster (1972: 75–76).

189

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 adjectives 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 adversative’ (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 principle 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) beautiful. 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 monograph, 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 unacceptable: *justified but fitting, *gerechtvaardigd maar gepast; interestingly, however, 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 combinations 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 adverbial 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 explanatory 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,

10

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. 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

191

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 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 semantically 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 combination 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 elements’ (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

12

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 information 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). Kroon calls this ‘mixture of (formal) conjunction and (semantic or pragmatic) disjunction’ ‘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 | 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 concedes 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].

192

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 usually 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

15

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. 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 halfmiler,—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 treatment 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. | (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, connected 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.

194

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 δεύτερον 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 Cephesiades 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)

195

Then let us see … what sort of things they are that profit us. Health, let us 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 little; 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 differ 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

17

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 Vorliegendes … vorgestellt wird hin’. At Men. 87e δή ‘no doubt contains sarcasm’, as Bluck ad loc. observes. Dodds explains γε as: ‘ “yes and …”, adding an unexpected item to the enumeration’, but this seems rather ad hoc.

304

196

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 semantically. 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 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 connector. 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 connectors 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 indicates that καί should be taken as the connector. Consider (22) and (23):

18 19

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. 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 μισόλογος, 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 | 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 example 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 understood, 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

21

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 supposed 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. 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.

197

306

chapter 15

δέ should be given the ‘on the other hand’ or ‘for that matter’ meaning discussed 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.

198

3.2 Other Cases The example from Sophocles (without preceding τε) that played such an important role in Denniston’s discussion of καὶ … δέ, should be analysed in the same 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 preceding 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 themselves 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 preceding clause, of someone else who considered the speaker (Lysimachus) wellwishing. 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 construction 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. | 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 ἀμφοτέροις 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 excellence 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.

199

308

chapter 15

So.: Thus when of two fair things one is fairer, the cause is that it surpasses 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 connector. 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 connector, 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 probably 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), corresponding 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:

200

(27) ταῦτα γὰρ οὐχ ὥσπερ εἰ …, καὶ εἰ νόσου δὲ πολεμίοις ἐμπεσούσης κρατήσειεν, 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

25

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 ἐμακάρισε (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. 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 coniunctum, 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 Attische 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 suggests 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

26

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. 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

201

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 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 adverbially 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

28

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. 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 wellwisher; 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 identical or similar, while the entities involved are different. Indeed, this explains why in example (29) ‘and you too’ is in principle fully acceptable: ὡς ἀπόλλυσαι 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

30

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. 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 ἐϋκνήμιδας 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.).

202

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

203

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) 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 preceding 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 overall ‘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 32

That the presence of αὐτός alone is not sufficient for this effect appears e.g. from X. Cyr. 2.4.17. 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· ἐλευθέρωσίν 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 mentions 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 matter of fact, this was the case; for …’ On this analysis, the ἐπυνθάνοντο-clause is answered by ἀλλ᾽ οὐδ᾽ ὣς etc.33 | 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’.

204

314

chapter 15

represents other people’s words, but voices his own comments. By the transition 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 combination 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 suggest 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

35

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. 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 KühnerGerth 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 πρέπον 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 principle 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. | I further argued that those cases where at first sight an ‘and also’ interpretation 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.

206

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 ὡσαύτως 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 ὡσαύτως (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 λειότητα … (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 pragmatically 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 omitted; 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 reasoning 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 | 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 opposites. 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).

207

318

chapter 15

Rijksbaron, A., ‘Sur quelques différences entre οὗτος ὁ + substantif, οὗτος δὲ ὁ + substantif 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 (Amsterdam 1980) 101–125. Wakker, G.C., ‘Emphasis and Affirmation: Some Aspects of μήν in Tragedy’, in A. Rijksbaron (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 investigation 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 accentuate 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 comment l’article fonctionne-t-il dans cet emploi, si fonction il y a ? Voilà la question 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 cohôte 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 grammaires 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 opinions 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

244

chapter 16

présent avec un nom propre c’est une caractéristique qui « belongs to popular 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 raisons 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.

245

Il fait suivre cette remarque par une série de remarques très diverses sur la présence ou l’absence de l’article dans toute sorte de genres. Maurice Bizos, en 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 μειράκια… 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 homonymes. – 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 conclusion 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 perspective 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 particulier 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 signification 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 remarquable 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 présence 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 prosateurs, 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 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 Anabasis, [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.

246

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 suivante:

nominatif

génitif

Κῦρος11

ὁ Κῦρος

Κύρου

τοῦ Κύρου

68

21

54

1

datif Κύρῳ 39

accusatif

τῷ Κύρῳ Κῦρον12 τὸν Κῦρον 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 11

12 13

J’ ai laissé de côté les résumés (non-authentiques) au début des livres 2, 3, 4, 5 et 7. Y compris 1.3.1, où Hude avec une partie des manuscrits lit Κῦρος; d’autres éditeurs préfèrent la variante ὁ Κῦρος. Le texte sans l’ article est probablement préférable. Voir aussi infra, n. 27. 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. 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 apparaît dans un discours direct, c’est-à-dire dans le texte d’ un personnage parlant. 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 mentionnées dans un discours direct sont | normalement présentes dans la situation 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 chapitres, 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 consécutivement 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

15

16

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»). Quant aux 79 cas du nom de « Cyrus » que l’ on trouve dans le récit après sa mort (voir cidessus), 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 activement engagé dans la préparation de son expédition. Ici donc Cyrus est exceptionellement réintroduit dans le récit comme personne vivante. L’ emploi fréquent du nom « Cyrus », dans un texte où il est de loin le Topique du discours

247

324 248

chapter 16

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 nécessaire. 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 regroupée 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 Orontas procède à son act perfide (ὁ δ᾽ Ὀρόντας… γράφει ἐπιστολὴν παρὰ βασιλέα ὅτι

17

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» («currentlyinactive », 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 referential expressions»), où l’ identification du référent ne pose aucun problème, je renvoie à Maes (1996 : 9–11). 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 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

1′ 2

2′ 3 3′ 4 4′ 5 5′

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; ἔφη ὁ Ὀρόντας. 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 ἔδωκάς μοι καὶ ἔλαβες παρ’ ἐμοῦ; καὶ ταῦθ’ ὡμολόγει ὁ Ὀρόντας. τί οὖν, ἔφη ὁ Κῦρος, ἀδικηθεὶς ὑπ’ ἐμοῦ νῦν τὸ τρίτον ἐπιβουλεύων μοι φανερὸς γέγονας ; εἰπόντος δὲ τοῦ Ὀρόντα ὅτι οὐδὲν ἀδικηθείς, ἠρώτησεν ὁ Κῦρος αὐτόν· Ὁμολογεῖς οὖν περὶ ἐμὲ ἄδικος γεγενῆσθαι ; Ἦ γὰρ ἀνάγκη, ἔφη ὁ Ὀρόντας. ἐκ τούτου πάλιν ἠρώτησεν ὁ Κῦρος· Ἔτι οὖν ἂν γένοιο τῷ ἐμῷ ἀδελφῷ πολέμιος, ἐμοὶ δὲ φίλος καὶ πιστός ; 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 ressortir 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

249

326

250

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 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 intéressant 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 combinaison 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

19 20

21

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. Il y a toutefois une variante ὁ Κῦρος. Pour τὸν Ὀρόνταν dans 1.6.10 voir infra, n. 25. Bien que cet ordre des mots semble avoir un caractère formulaire, je pense que pragmatiquement 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 discours (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 exceptionelle 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 μαχεῖσθαι, 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. 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 Ὀρόντας, 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 préférable, 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 : Ἱερώνυμος 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 Anabasis», 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 approfondi.

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 23

Avant Zucker, Schmidt (1890 : 54) avait déjà écrit : «Nomen proprium quod coniungitur cum ἔφη in nullo scriptore sine articulo inveni ». 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.

251

328

252

chapter 16

ennemis qu’amis de Cyrus, nous sont présentés. Aucun de ces personnages 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 ἀποστῆναι 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 Τισσαφέρνους 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

25

Il faut remarquer que lorsque ce même épisode est rapporté dans la nécrologie de Cléarque (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. À noter aussi qu’ ici la présence de ὁ Κῦρος a été préparée exceptionellement par la combinaison 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 partie 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, pourrait 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 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 organisateur, 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 βαρβάρων. ἐκέλευσε δὲ τοὺς Ἕλληνας… ταχθῆναι καὶ στῆναι, … ἐτάχθησαν οὖν ἐπὶ τεττάρων·… 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 exclusivement 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 préparatives 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, inspectait 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 également des autres noms propres dans l’Anabase, mais en | faisant cela je dépasserais 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

27

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. 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…).

253

254

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

255

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 Cyropédie 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 détachées, dans lesquelles Cyrus ou bien entre en débat avec d’ autres personnes, ou bien est présenté comme la personne centrale de situations nouvelles; le premier 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 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 prédomine, dans le récit continu de l’Anabase, c’est Κῦρος seul. 28 29

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. Et pour d’ autres écrivains, bien entendu. Pour quelques observations sur Platon voir §7.

sur l’ article avec nom propre

7

331

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 participe 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 dialogue, 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 διήγησις (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 Σωκράτης 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 uniform 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 alternances des prises de parole34. Quant à Κέβης sans article, il y en a huit exemples,

30 31 32

33 34

À vrai dire la situation dans le Banquet est plus complexe, parce que le narrateur, Apollodore, raconte à son tour ce qu’ il a entendu d’ Aristodème. À son tour Burnet a utilisé un article de J.I. Beare dans Hermathena 9 (1896), p. 197 sqq. En 86d5, dans une situation tout à fait semblable, T est de nouveau le seul MS qui a Σιμμίας, mais curieusement Burnet, tout comme le nouveau Oxford Classical Text, préfère ici la variante ὁ Σιμμίας, sans aucune explication dans son commentaire. La présence ici de ὁ s’ explique bien comme un effet de la formule ἔφη ὁ Κέβης, très fréquente dans le Phédon, voir ci-dessous. 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,

256

332

chapter 16

dont sept dans les mots d’un participant à la discussion (y compris 63a2, mentionné 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 question 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 spéciale 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, Criton 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

257

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 analysées ci-dessus. Chez Xénophon et Platon, la fonction de l’article avec un nom de personne 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

35

notamment ἦ δ᾽ ὅς. Les conditions d’ emploi de toutes ces formules sont tout à fait obscures. 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.] On a une extension de cet emploi « de relief» dans un cas comme Phdr. 269e: Κινδυνεύει… 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? » (pareillement en d’ autres langues, comme l’ allemand, l’ anglais, l’italien, le néerlandais…) ressemble 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 anaphorique, mais par la combinaison avec un nom propre, qui est, lui aussi, employé anaphoriquement, il y a un effet pragmatique spécial, qu’ on pourrait appeler «anaphore forte». Autre conclusion: la présence de l’ article avec un nom propre est une caractéristique de textes narratifs (dialogue rapporté, récit historique)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 question 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 anaphoric ». 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 présupposition 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: ἐπιχειρήσωμεν 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 + substantif 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 Xenophons 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 Tantalus: 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 1

chapter 17

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 Rembrandt?, to which I will return in more detail below. Greek is of course a goldmine of such phenomena.1 One needs only think of the combinations οὐδεὶς … οὐ nobody … not and οὐκ … οὐδείς not … nobody, and of constructions like ἀπαγορεύω 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 Denniston 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 referees 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 preliminary 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 affirmative 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

3

4 5

6

Liddell-Scott-Jones do not specify what they mean by ‘negative answer’. They must have considered 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. 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.). Unless indicated otherwise the examples are from the very rich corpus of interrogative discourse provided by Plato. 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. 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, ἠπιστοῦμεν; :: 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 ‘certainly/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 νεανίσκων; :: 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 Question (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 Expression, 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, Diognetus, 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 affirmative 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 semantics 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:

9

10

11

12

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). ‘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.’ 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. 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. 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 (positive) 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 Hudson (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 corresponding 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

14

15

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.’ 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 relationship: if the speaker questions the truth of ~p, he must consider p true. 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 πραττόμενα 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.

16

this seems to have been extremely rare, in Plato, at least. Normally, the interlocutor cooperates. (‘[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.’) 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 languages 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 Question (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 particle, but as not, i.e. a proposition negative, which in answers functions as an elliptic substitute for the entire underlying negative proposition of the question: (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 asseverative 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. Further 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., importantly, 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 20

For the Romance languages see further n. 20 and section 5. 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 temperately …? :: 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

22 23

24

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. In the analysis of οὐ advocated here, the No of Henderson’s Loeb translation should rather be I’m not. 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 proposition: 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. 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 particle.

5

Latin and Other Parallels

Ancient Greek is by no means alone among Indo-European languages in having 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? :: NegPtcle’) 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 proposition negative ne and the answering particle njet) and Albanian ( jo). Outside

25 26

27 28

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’. And haud, which has a far more restricted use than non. See Kühner-Stegmann 1, 813–816. Haud is not found in answers. 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. Orlandini (2001) does not discuss the problems involved in establishing the function of non. 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 differentiation 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 proposition 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 plau29

30

31

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. Jespersen’s Cycle is still fully alive in French, see Du tout for Pas du tout in Êtes-vous originaire 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 Jespersen’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. 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 negative, 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 Modern 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 particles 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:

32

33

34

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). 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 IndoEuropean see Croft (1991). 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. 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 preceding 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 correctness 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 particle 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

35 36

37

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.’ 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. 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’). 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 replacer 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 particular 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:

non

1) > nen > ne > ne … pas negative proposition particle (La vie de saint Alexis, end of 11th c.); modern spoken French: pas (through Jespersen’s Cycle); ne retained in expressions like je ne saurais le dire 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 probably 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 variant 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 notoxytone), 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 discussed 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 adjective 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 grammarians’ 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, elsewhere 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 suggests 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 inscription 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 position: 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 ἀπεκρίθη 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 (Leiden 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 Conference 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: Satzlehre. 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 Philosophy 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 Linguistics (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 grammarians 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, sentence 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 | 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 subsidizing 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 authenticity 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

428

358

429

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 certain 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 grammarians’ treatment of voice that the important group of intransitive ‘middle’ verbs that correspond to active so-called causative verbs (e.g. διαφθείρω : διαφθείρομαι, 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 πέπηγμαι’), whereas ἐποιησάμην is a φωνή παθητική yet expresses ἐνέργεια (‘ἀντὶ τοῦ 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 analyses 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 universally 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 illustrate ‘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, perhaps, 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 (ἐποιησάμην); | 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 meaning (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 diatheseis, 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 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 Hilgard 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 ἐτρίφθην. 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

430

431

360

432

chapter 18

finally, one of the many ancient grammarians who recognise a fourth διάθεσις 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 περιπατεῖ. 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 attention. First, the scholion at 1, 3, 246, 7ff. | assures us that there are five diatheseis, 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 approached 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 exemplify 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 οὐδέτερον. 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 transitive 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. | 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 ‘normal’ verbs that have either active morphology and active, or, rather, transitive, meaning, or passive morphology and passive meaning, apparently was so dominating 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 prevails 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 ‘middle’, 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

433

Apollonius Dyscolus

In a passage on certain anomalies in the verbal system, Apollonius also discusses the ‘middle’. At 2, 2, 296, 2 he has the following general remark τὰ … καλούμενα μέσα σχήματα συνέμπτωσιν ἀνεδέξατο ἐνεργητικῆς καὶ παθητικῆς διαθέσεως. 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. The examples given are ἐλουσάμην, ἐποιησάμην and ἐτριψάμην, of which ἐλουσάμην 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-

434

362

435

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 peculiarities, 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 βιάζομαιlike 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 presupposes, 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 scholiasts. In the remainder of the περὶ συντάξεως ‘middle’ is not discussed separately 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 μεσότης 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, ἄγαμαι, 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’ βιάζομαι, μάχομαι and χρῶμαι. As for possible other diatheseis, Apollonius calls γηρῶ, φθίνω, κοπιῶ etc. verbs expressing αὐτοπάθεια (‘self-affectedness’), the latter being a subcategory of παθητική. This is, therefore, no separate diathesis.10 Οὐδέτερον is not recognized by Apollonius; verbs like ζῶ, ὑπάρχω and also περιπατῶ, have with him a vaguely ‘active’ status. Nor does he know of ἐμπεριεκτικόν.

8

9

10

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 passive 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. 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. If the criterium for assigning a verb to ‘middle’ is the presence of active endings with passive 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 | (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 artificial; 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 meaning’ 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 modification, 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 nominal system, there is a group of κοινά words, i.e. words that are either masculine 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 | 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 diathesis.

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 pluperfects, 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.

436

437

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 morphologically 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 appearances, the same analysis as that proposed by Apollonius.

4

438

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), ἠλειψάμην (passive meaning, = ἠλείφθην), ἐγραψάμην (active meaning, = | ἔγραψα), πέπηγα (passive meaning, = πέπηγμαι). But he also gives πεπληγώς and κεκοπώς, 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, οὐδέτερον, 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

13 14 15 16

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 middle; 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 categories. References are to volume and page of the Grammatici Latini, ed. H. Keil. He tells us that the Latin counterparts of these verbs are called deponentia, a term which has no Greek equivalent. Curiously enough, this term is not found in the Greek grammarians. 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-Choeroboscan ancient treatment of the middle: βιάζομαι, ἠλειψάμην, ἐγραψάμην, πέπηγα, belong to the ‘Dionysian’ tradition, κεκοπώς, πεπληγώς, as well as μάχομαι, ἄγαμαι etc., on the other hand, to the ‘Apollonian’ tradition. The οὐδέτερον diathesis is another Dionysian element. The unfortunate τέτυπα, τέτυφα of Choeroboscus are lacking. | 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 expressing an action which ‘remains within itself’?), comprising verbs like cresco, nascor, cado, labor (Char. 1, 164 Κ.). This category partly coincides with the category 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 grammarians 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 wastepaper 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 passive 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 Apollonius. 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.

439

366

440

441

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 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 ἀντιπεπονθότα, 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 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 18

19

20 21

This term is also used in modern linguistic studies, be it in a different way, namely in connection with ergative languages. Compare Comrie (1981: 108ff.) and Dik (1981: 123ff.). Steinthal’s discussion of κείρεται (1890: 1, 299–300) is rather confusing. He views κείρομαι 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 alterations; it is not clear to whom they are due in the last resort. 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.) Aristotle’s opinions on this matter are treated extensively in the valuable monograph by S. Waterlow (1982). Cp. also Steinthal (1890: 300 ff.). 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 grammarians.

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 connection are θερμαίνεσθαι, ψύχεσθαι (cp. Simpl. p. 310), ἰοῦσθαι (‘rust’) and σήπεσθαι (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 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 inanimate: 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, determined 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 number 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 explanation of the canonical classical texts. More astonishing is, perhaps, that modern 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

23

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. In general linguistic literature, though, a number of Aristotelian observations have been

442

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 Linguistics, 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 Neoplatonic 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 examples 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 questions’.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 Interrogations’,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 unfortunately 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 scholars 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 ‘Entscheidungsfragen’7 are identical with English wh-questions and yes-no questions, respectively, 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 questions’). ‘Entscheidungsfragen’ = ‘decision questions’. 8 There are, of course, other possibilities, e.g. adverbs like πάνυ, or (part of the) original question 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

734

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 wh-questions, i.e. | questions introduced by who, where, why etc. D.L. 7.66, mentioned 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 συμβολικῶς 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 Ammonius 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 ἐρώτησις, 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 10

11

Hülser translates πύσμα with ‘Bestimmungsfrage’ and ἐρώτημα with ‘Entscheidungsfrage’. ‘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 question 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. ‘(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 sometimes 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 examples 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 διαλεκτικὴ 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 opening 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

13

14

15

The full text of the relevant passage from Alex. runs: Ἐρώτημά ἐστι, πρὸς ὃ ἀνάγκη ἀποκρίνασθαι 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 τοιοῦτον, 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?” ’). ‘It is generally agreed that not every universal can form a dialectical proposition, for example “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). 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 question 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). 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 specifying 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 dependent 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 possibility), 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).

16

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. 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 quaesitum 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 Gundermann (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: ‘percontatur 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 (Cambridge, 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 considerable 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 WilamowitzMöllendorf (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 enthusiastic about his Greek lessons: das Griechische war nicht mehr als ein Anhang zum Latein. Formenlehre 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 curriculum looked like, but it will not have been much different from the curriculum 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 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 diplomat had profoundly reformed the educational system. His reform was based on the following premise: ‘Es gibt, philosophisch genommen, nur drei Stadien des Unterrichts: Elementarunterricht, Schulunterricht, Universitätsunterricht.’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 linguistically 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.’

236

378

chapter 20

vided into the ‘Bürgerschule,’ a kind of professional school, and the ‘Gelehrtenschule,’ whose objective it was to prepare for higher education. At the end of that school

237

[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 Hilfsmittel jeden Schriftsteller, insoweit er wirklich verständlich ist, mit Sicherheit zu verstehen, und sich in jede gegebene Sprache, nach seiner allgemeinen 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 ‘Gymnasium’ 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 elementary 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 countries 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 commentaries.

6 ‘The pupil is mature when he has learned so much with others that he is now capable of learning 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 available for quite some time, e.g., that of Ph. Buttmann, of which the 10th edition appeared in 1837.

the xenophon factory

3

379

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 suggested 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 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 perhaps 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 survey, 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

9 10

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. 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. Gildersleeve (1930: 365).

238

380

239

chapter 20

his ‘American teachers did not understand their business.’11 Actually, this ‘Teutonic’ 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 Murray, 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 encompassed 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 considered ‘negligible’ will be addressed after the survey.

11

12

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. 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?’

381

the xenophon factory 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 publication

Germany

United States

Great Britain

1820–183014

[1826, C.G. Krüger, Halis Saxonum] [1830, K.W. Krüger; notes in German]

1834, C.D. Cleveland,15 Boston

[1830, F.C. Belfour, London] [1834, 2Belfour] [1835, A. Negris, Edinburgh] | 1838, J.T.V. Hardy,16 London [1839, D.B. Hickie, Books I– II, London] [1841, C.S. Stanford, Dublin] [1844, 2J.A. Phillips, Dublin; year of 1st ed. unclear]

1830–1840

1840–1850

13

14

15 16 17

[1842, G. Graff, Leipzig] 1845, K.W. Krüger, Berlin, bei K.W. Krüger (in German; ‘zweite Ausgabe’)

[1843, J.J. Owen,17 New York: Appleton & Co.]

Other countries

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. 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 published 1735). Since these editions played no role in schools, I have not included them in the survey. ‘Professor of the Latin Language and Literature in the University of the City of New York.’ ‘Assistant Master in University College School.’ ‘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.’

240

382

chapter 20

Table 1 Period of publication

A survey of (school) editions of the Anabasis (cont.)

Germany

United States

[1849, 2Krüger] [1849, F.K. Hertlein, Berlin: Weidmann]

241 1850–1860

18 19

20 21 22 23

[1850, 3Krüger] [1851, K. Matthiä, Quedlinburg] 1852, R. Kühner, Gothae, notes in Latin [1854, 2Hertlein] 1854, 4Krüger, Berlin: K.W. Krüger’s Verlagsbuchhandlung; simultaneously published in London and New York 1857, 3Hertlein [1857, F. Vollbrecht,20 Leipzig: Teubner] 1859, 2Matthiä

1857, J.R. Boise,21 New York: Appleton & Co.

Great Britain

[1847, J.F. Macmichael,18 London: G. Bell & Sons] [1847, J.T. White, London] [1848, Ch. Anthon,19 London] [1848, J.S. Fergusson (or Ferguson), Edinburgh] [1849, 2Hickie, London, ‘rev. ed.’] | [1850, 2Fergusson] [1852, 2Anthon, new ed. rev. by T. Doran] [1853, Browne,22 trans. of Hertlein 1849 with add’l notes, Arnold’s School Classics] [1853, 2Macmichael] [1853, J.T.V. Hardy & E. Adams, London (‘portions only’)] [1854, 4Fergusson] [1854–1855, H. Young,23 London] [1857, 3Macmichael, ‘new ed.’]

Other countries

France: [ca. 1860, F. Dübner, Paris]

‘B. A., Trinity College, Cambridge; Head Master of the Grammar School, Ripon.’ ‘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). ‘Rector zu Otterndorf.’ ‘Professor of Greek in the University of Michigan.’ ‘Prebendary of Chichester.’ ‘Second Master of the Grammar School, Guildford.’

383

the xenophon factory Table 1

A survey of (school) editions of the Anabasis (cont.)

Period of publication

Germany

United States

Great Britain

1860–1870

[1861, 2Vollbrecht] [1863, C. Rehdantz, Berlin: Weidmann] [1865, 3Vollbrecht] 1865, L. Breitenbach, Halle: Verlag der Buchhandlung des Waisenhauses [1869, 4Vollbrecht]

1867, 2Owen, preface dated 1861 (‘Since that time [viz., 1843] about 35,000 copies have been printed and sold’) 1868, 3Anthon

1870–1880

[1871, 6Krüger] [1873, 5Vollbrecht] 1877, 6Vollbrecht (‘up to and incl. the 5th ed. 20000 copies sold’)

[1875, A. Crosby, New York]

1880–1890

1888 6Rehdantz, bearb. von O. Carnuth25 [1889, E. Bachof, Paderborn] 1889. 7Krüger, besorgt von W. Pökel

[1889, F.W. Kelsey26 & A.C. Zenos,27 Boston]

1860, Macmichael (‘new edition’) [1860, 6Fergusson] [1866, 9Fergusson] [1866, 2White] [1866, K.S. Isbister, London (Books I–III)] [1868, Macmichael (‘new edition’)] [1868, J.S. Watson, London] [1872, Macmichael (‘new edition’)] [1877, A. Pretor,24 Cambridge: CUP., repr. 1878] [1879, R.W. Taylor, London] | 1880, W.W. Goodwin28 & J.W. White,29 London: Macmillan 1881,2Pretor (repr. 1882, 1885, 1891, 1894, 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 25 26 27 28 29 30 31

Other countries

242 Austria: [1884, R. Hansen, Wien] Italy: [1886, A. Bersi, Torino]

‘Fellow of St. Catharine’s College, Cambridge.’ ‘Professor Dr., Direktor des städtischen Gymnasiums in Danzig.’ ‘Ph.D.’ ‘M. A.’ ‘Ph.D., Eliot Professor of Greek Literature in Harvard College.’ ‘Ph.D., Assistant Professor of Greek in Harvard College.’ ‘M. A., Formerly Scholar of Worcester College, Oxford.’ ‘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 publication

Germany

United States

Great Britain

Other countries

1890–1900

[1890, A. Weidner, Leipzig: Freytag] [1895, F.G. Sorof, Berlin: Teubner] [1896, 9Vollbrecht, bes. von W. Vollbrecht32]

1892, Kelsey-Zenos (‘third edition’ but actually = ed. of 1889) 1895, 4Kelsey-Zenos (‘fourth edition,’ but now revised; repr. 1948 as ‘sixth edition’)

[1892, A.H. Allcroft & F.L.D. Richardson, London (Book I)] [1894, C.E. Brownrigg,33 London: Blackie & Sons] [1885–1892, J. Marshall,34 Edinburgh] [1896, 2Allcroft & Richardson] [1897, Fergusson & W. Coutts, Edinburgh (‘new ed.’)]

1900–1910

[1905, 4Weidner] 1907, 10Vollbrecht

Austria: [1890, A. Weidner, Wien (same as the German ed., but with another publisher)] [1894, 2Weidner] France: [1896, P. Couvreur,35 Paris] Holland: [1899, J. Mehler & H.O. de Jong, Utrecht] Italy: [1893, 2Bersi; ‘of the 1st ed. 4000 copies sold’] France: 1900, 2Couvreur Italy: 1905 1905, 3Bersi (repr. in 1912) |

1910–1920

1912, 7RehdantzCarnuth, bearb. v. E. Richter

1920–1930

1929, 8Sorof, bes. von G. Sorof37

1901, G.M. Edwards,36 Cambridge: CUP (Book I) 1902, 2Brownrigg [1908, GoodwinWhite, unalter. repr. of 1880?]

243

1930–1940

32 33 34 35 36 37

1912?, M.W. Mather & J.W. Hewitt, s.l.

Austria: [1919, 5Weidner] France: [1913, 5F. Dürrbach, Paris]

Holland: [1933, H. Rogge, Zwolle]

‘Dr., Professor am Christianeum zu Altona.’ ‘M. A., Headmaster of Magdalen College School, Oxford.’ ‘Rector of the Royal High School.’ ‘Ancien Maître de Conférences à la Faculté des Lettres de l’Université de Lille.’ ‘M. A., Fellow and Tutor of Sidney Sussex College, Cambridge.’ ‘Geh. Stud. Rat Prof. Dr., Oberstudiendirektor i.R.’

385

the xenophon factory Table 1

A survey of (school) editions of the Anabasis (cont.)

Period of publication

Germany

United States

1940–1950

[1947, Ed. Bornemann, Frankfurt/M. (‘Auswahl’)] [1949, A. Clausing (‘Auswahl’)]

1948, Kelsey & Zenos; on the titlepage ‘sixth edition,’ but actually unaltered repr. of 4th ed.

Holland: [1947, A.M. Bent, Nijmegen]

1950–1960

[1952, 2Clausing] [1954, Bornemann (‘neue Ausgabe’5)] [1956, 6Bornemann] [1956, 3Clausing] [1960, 4Clausing] [1962, 9Bornemann] [1963, 10Bornemann] [1974, 14Bornemann] [1979, 23E. Krämer, Münster; printing history could not be traced]

1962, Mather & Hewitt (reprint)

Holland: 1963, 7Bent

1960–1970

1970–1980

Great Britain

Other countries

1980?, J. Antrich38 and St. Usher,39 Bristol: BCP40

What can we learn from this survey of what with only a little exaggeration 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 39 40 41

‘Head of Classics, Godalming Sixth Form College.’ ‘Senior Lecturer in Classics, Royal Holloway College, University of London.’ On the titlepage: ‘Xenophon. The Persian Expedition.’ Actually, however, it is a small anthology. 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.

244

386

245

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 (usually 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 variety of editors was paralleled by an equally wide variety of publishers. In Germany, 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 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 commercial 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 editions 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 43

There is a direct link with the measures taken by Humboldt: see Garzya (1983: 19). For such responses by publishers to changes in the curriculum there are also modern parallels. 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 ‘insistence 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 Anabasis, 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 simultaneously 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 Pretor | (at least until 1900) and Edwards (in 1901); for the US see Owen 1843 and Boise 1857. Had the earlier editions become too learned, perhaps because of changes in the curriculum in the ‘Gelehrtenschule’ and its Anglo-Saxon counterparts? 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 translations.

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.

246

388

247

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 publications 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 attractive (one really | could not afford to ignore the latest insights). Was the scholarly conscience of the German editors and their audience better developed? 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.

389

the xenophon factory table 2

Some characteristics of the editions46

Edition

‘Für den Schulgebrauch’ vel sim. on title page?

Preface with objectives, etc.?

References to other editions/ grammars?

Krüger 1845

no

no

Kühner 1852

no

yes

4Krüger 1854

no

no

Vollbrecht 1857

yes

yes

2 Vollbrecht 1861

yes

other ed.: – gramm.: Krüger other ed.: ‘superiores editores’ in general gramm.: Kühner other ed.: ‘Ze.’ gramm.: Krüger other ed.: ‘alle Arbeiten früheren Herausgeber’ gramm.: ‘die verschiedensten’ as in first ed. | 248

Anthon 1867

yes, extensive

Goodwin-White 1880

no

Pretor 1881

no

Walpole 1882

yes

Bersi 1886

no

6Rehdantz– Carnuth 188848

no

7Krüger-Pökel 1889

no

yes, justification of changes with respect to 1st ed.; sim. in next edd. no the only ed. with a bibliography (up to 1845); does not mention his American colleagues Cleveland and Owen yes other ed.: general remark about predecessors47 gramm.: ‘notes adapted to Goodwin’s Greek Grammar’ yes other ed.: Kühner, Schneider, Vollbrecht, Bornemann, Macmichael gramm.: Madvig yes other ed.: notably those by Rehdantz, Krüger, Vollbrecht; Goodwin & White (‘the best of the English editions’) gramm.: Goodwin yes other ed.: Krüger, Breitenbach, Kühner, Matthiä, Rehdantz, Vollbrecht gramm.: ? no, just a few no references? (printing history not mentioned remarks about in the ‘preface’) the corrections no other ed.: a great many, also from the eighteenth century, without any bibliographical information gramm.: Krüger | 249

46 47

48

Sometimes I refer to an edition that I have not seen; in these cases the relevant information was supplied by the preface(s) of earlier editions. ‘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 Editors 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). 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 Schulgebrauch’ vel sim. on title page?

Preface with objectives, etc.?

References to other editions/ grammars?

Kelsey-Zenos 1889 (= 1892) 2Bersi 1893

no

yes

no

other ed.: Rehdantz, Vollbrecht, Matthiä, Taylor gramm.: Goodwin, Hadley-Allen other ed.: as in 1886 gramm.: C. (= Curtius?)

4Kelsey-Zenos 1895

no

Couvreur 1896

no

yes, justification of changes with respect to 1st ed., sim. in 1905 ed. yes, very brief justification of changes with respect to 1889 ed. (= 1892) yes, extensive

Brownrigg 1902

no

no

Mather-Hewitt 1912 (?)

no

yes

Sorof 1929

yes

Bent 1947

no

Antrich-Usher 1980 (?)

in the introduction: for students

yes, but very brief yes, but very brief yes

250

other ed.: as in 1889 gramm.: Goodwin, Hadley-Allen, also Goodwin, Moods & Tenses

other ed.: Vollbrecht, Rehdantz-Carnuth, Goodwin-White gramm.: Croiset & Petitjean other ed.: – gramm.: – other ed.: – gramm.: Smyth, Hadley-Allan, Babbitt, Goodwin, Goodell other ed.: – gramm.: – other ed.: – gramm.: van Oppenraaij-Vermeulen other ed.: – gramm.: –

It is clear that some editions pretended to be more than just a school edition. 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 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, starting 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 nineteenthcentury 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 außerhalb 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 Wilamowitz regarded himself as a very privileged exception, because he had had the opportunity to learn English as an extra-curricular activity. His remarks are confirmed 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. | 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-

50

51

52

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. 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). ‘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). Gildersleeve (1930: 366).

251

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 generatio spontanea. After this introduction I now pass on to a discussion of the linguistic merits of these editions, which will be illustrated from the very beginning of the Anabasis.

6

252

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 average 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. Second, 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 the same explanation. I have not examined the editions | systematically on this score, but generally speaking I have the impression that they meet this requirement. 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 phenomenon for which no real explanation exists, either in grammars or in other commentaries, whether on the Anabasis or on other texts? It goes without saying 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 secondary 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 problems 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 question 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 widespread throughout the Anabasis (not to mention other texts). In all three cases a generally accepted explanation is lacking, although for (b) and (c) some proposals 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

54

55

56

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. 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, chapter 16.] For (c) see K-G 2, 555, Anm. 3 and Amigues (1977: 270). 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 inspection, 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, however, 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 combination 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.’

253

394

chapter 20

the possible differences between the use, in oratio obliqua, of the oblique optative 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 problem, 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 discussion of the notes that all commentators have on the first two constituents (Δαρείου καὶ Παρυσάτιδος and γίγνονται).

254

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 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 γίγνονται 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 combination of particles occurs about 265 times in Xenophon, and all six commentaries 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 connective 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 KühnerGerth—is correct.58 (b) §3 ὡς ἀποκτενῶν. Four commentaries note that there is a difference between the future participle with and without ὡς. Only Bent and Antrich– Usher believe, quite wrongly, that the absence or presence of ὡς is irrelevant: ‘het part, fut., met of zonder ὡς, heeft finale betekenis’ (Bent); ‘ὡς + fut. part., expressing purpose’ (Antrich–Usher). This carelesness is puzzling: 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, 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 59 60

See Rijksbaron (1997) [= chapter 15 in this volume]. ‘ἀποκτενῶν alone would present the purpose as simply narrated; with ὡς the objective is formulated as a thought of Artaxerxes.’ 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.

255

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’)

256

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’ + reference to Kühner’s grammar) – Vollbrecht 1877, 1907 (‘Gen. des Urspr.’). No translation – Parker 1882 (‘Goodwin Gr. 169 calls this a genitive of possession and translates ‘are born (belonging to) Darius.’ It is rather a genitive of source, origin, 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 Goodwin’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 origin’) – 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 62

Hardy’s book contains a ‘literal and interlinear translation’ of chapters 1–6 of the first book, and a vocabulary; there are no notes. ‘Possessive gen. to indicate the origin.’ He refers to §116.1 of the grammar of Van Oppenraaij—Vermeulen. There one reads: ‘Possessivus: geeft de bezitter of oorsprong aan; dikwijls 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 | – 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.’

257

398

258

– –

– – –

– –

259

chapter 20

1873 ed.), where one reads: ‘Mit Lebhaftigkeit macht der Geist Vergangenes zu ideeller Anschauung der Gegenwart; und so steht das sogenannte historische 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 grammar, §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 tempore 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 effects of a past event lasting into present time | cf. note on § 7,67 and Propertius v.i.77 me creat “he is my father.”’) – Vollbrecht 1907 (‘das Präs. von einem Ereignisse, das dauernd gültig ist’) 64 65 66

67

‘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.’ ‘The lively Greek uses the historic present (in main clauses) far more often than we and even than the Roman.’ ‘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.’ 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 present, 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 incidents, 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 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 alarming, 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 commentators 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.

260

400

chapter 20

items, an explanation that—and this is perhaps an encouraging thought—is included in the above notes.

261

a Δαρείου καὶ Παρυσάτιδος No real explanation of the genitive is given by those commentators who combine, 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 having of children is expressed by the dative; since the latter, however, unfortunately 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 impossible: 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 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 commentators 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, πείθεται, 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 passage 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 commentator should be very cautious and should not snatch too quickly at ‘vividness’ to explain the use of γίγνονται, etc. | 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 ‘permanently 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 interpretation 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

70 71

Incidentally, with the sole exception of Walpole, no commentator has a note on ἁθροίζονται. 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. (1898–1904) 1.134. This must be due to Gerth, for in the second edition of Kühner’s grammar (1870) γίγνονται at An. 1.1.1 is not treated as such. In fact, it is not discussed at all. (1898–1904) 1.145 ff.

262

402

263

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 ‘vividness,’ they present, in a small compass, those events of Cyrus’ life and career 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 background 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 concerning 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 justi72 73

Also by S-D 2, 272. 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 | contribution’ 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 Geburtsort 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 misleading: Πατρὸς τίνος ἐστὶ καὶ μητρὸς ὁ Ἔρως … [read: Πατρὸς τίνος ἐστὶ καὶ μητρός

74

75 76

77

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’). Clear clues for the correct interpretation of both the genitive and the hist. present can be found in K-G. 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. I made use of the fifth edition (1873).

264

404

265

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 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 Φραόρτης 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 speaking 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 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 γίνεται 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 possessive 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 mentioned 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. | For a balanced discussion of the simple genitive and the genitive with ἀπό and ἐξ see also K-G 1, 374, Anm. 3.

79

ownership, e.g., at Th. 5.5.1 ἐγένετο Μεσσήνη Λοκρῶν τινὰ χρόνον, ‘For some time Messene came into the possession of the Locrians.’ Fortunately, in his revised edition of Krüger’s grammar, G.W. Cooper has completely abandoned Krüger’s views ([1998] 173). He takes Δαρείου καὶ Π. as an ablatival genitive.

266

267

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 Historiographie (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. Bakker (ed.), Grammar as Interpretation: Greek Literature in its Linguistic Contexts (Leiden 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 Xenophons Anabasis (Nürnberg 1899).

Index Rerum absolute tense 63, 65–68 see also tense accentuation shifts 352–353 value of mss evidence 108–109n see also Index Graecitatis accomplishment: see Aktionsart accusative of duration of time 289–290 achievement: see Aktionsart adverb affirmative 339, 343–345 evaluative 139 intensive 226 locative 116, 118 manner 116 vs. accusative noun phrases 289– 290 vs. connector 293–317 see also Index Graecitatis agent: see semantic functions Aktionsart 32, 82, 102–103, 113, 126, 131 anaphoric use 224–236 strong 226–230, 230–232, 235 temporal modifiers 12, 27n weak 226–227, 228n, 232–234, 235 also see article ancient grammarians 332, 357 answer 335–356, 370–371 ‘no’ and ‘not’ in 344–347, 349–350 antecedent 234n aorist stem aspectual value 40–41, 70, 83, 125–132, 242 complexive 33, 74, 161, 192, 195 constative 49–50, 70 countable actions 49n, 283n gnomic 151, 162, 188–189, 203, 204, 205, 206, 207 habitual/omnitemporal contexts 69, 151, 204 indicative 41, 60, 68–73, 78 ingressive 32–34, 66 negation 22 non-past use 68–71, 146–147 participle 33–36

relative tense 16–18 semelfactive 22, 23n, 83, 91, 242, 263 telicity 127, 129, 130, 131 verba dicendi 24–29 argument (syntactical role) 229n, 281– 282 article anaphoric 324, 333 forming noun phrases 279–280 marker of definiteness 283, 285n, 291 position in noun phrases with οὗτος 223–236 with proper names 319–333, 393–394 τό with temporal adverbs 279–292 aspect 5–9 see also aorist stem, perfect stem, present stem asyndeton 104, 224–230, 235, 298n augment unaugmented past tenses 107–108, 162 change of speaker (in dialogue) 215, 218, 325–327 dative ‘possessive’ with εἶναι/γενέσθαι 256n, 269–271, 275, 400, 403, 405 proper names with and without article 322–323, 328n declarative sentence 102, 179 diathesis (active, middle, passive) 43–44, 47–48, 357–358, 361–365, 395 direct speech 50, 72, 112, 134, 167 Dramatis Personae 180n, 182n, 183 exclamation 102, 221 expressions of sorrow 254–277 see also Index Graecitatis focalisation 136, 189, 204 see also Substitutionary Perception Focus 226, 227, 309n, 326n, 339 Fortsetzung 226, 228, 230, 231 Free Indirect Speech/Thought 135, 139– 140 see also Substitutionary Perception

408 Functional Grammar 1, 39, 53, 291n, 340n see also argument, sattelite, semantic functions future perfect stem aspectual value 40 indicative 41, 66 future stem aspectual value 40–41 indicative 41, 65, 71–72, 187n, 201, 240– 242, 246n in Linear B 150 generic use: see habitual/omnitemporal use habitual/omnitemporal use 65, 201, 207 participle with and without ὡς 395 vs. future perfect stem 66n genitive of comparison, with πρότερον 285 proper names with and without article 322–323, 328n with expressions of sorrow 266, 277 Δαρείου καὶ Παρυσάτιδος (at X. An. 1.1.1) 395–396, 400, 403–405 historic present: see present stem imparfait/passé simple (French tenses) 8, 73n, 138–139, 140–141, 145, 146n, 147n, 167–168 imperfect: see present stem imperative aspect 80–97 implicature (Grice) 70, 146 infinitive articular 279 aspect 49–50n, 82n, 83–84, 96n, 125–132 dynamic 83, 124–126 injunctive 162–163, 191n inquit formula 210–222 internally affected: see semantic functions messenger speech 99–122 passim, 149, 170– 184 basic scheme 171 beginning with a simple sentence 176– 179 beginning with ἐπεί 171–176, 179–184

index rerum middle voice 357–369 see also diathesis direct-reflexive 366 indirect-reflexive 260n, 359, 361, 395 intransitive 358, 366 morphologically active perfects 43 passive meaning 358–365 reflexive causativum 366n momentaneous verb see Aktionsart narration/narrativity 133, 139n, 141–146 alternation of tenses in 50, 99, 105, 118, 146–151 negation 335–356 and aspect choice 10, 21–23, 25 historic present not combined with 103, 107n ‘no’/‘not’: see answer; negation nominative proper names with and without article 322–323, 328n, 330 number plural vs. dual 393–394 singular vs. plural first person 187, 194 singular vs. plural in expressions of sorrow 255–256, 258, 259, 263, 268n, 269, 270, 275, 277 optative 72, 124n, 159n, 240–242, 246n, 393– 394 ordinal, ordinator 289–290, 291 parenthesis 113–114, 297n participle aspect 21, 23, 29–30, 34–36, 50n, 69, 76, 78, 114, 151, 154, 160n, 192n, 213, 274 dominant 261 future with ὡς 395 particle 3 answering: see answer with historic present 104 see also Index Graecitatis passive voice historic present rare in 103, 106 in perfect stem 43–48, 50n, 52–53 see also diathesis, middle voice (passive meaning) past progressive (English tense) 73n, 138, 141, 167

409

index rerum patient: see semantic functions Perception Indicator 136, 158, 160, 163, 165 see also Substitutionary Perception perfect stem 39–58 active intransitive 43–44 active transitive 44–45, 46 aspectual value 40–41, 64, 66, 67 development of in Homer and after 43– 48, 53–54 emphasizing responsibility 47–48, 49, 51, 53–54 indicative 41, 64–65 not a double of aor. ind. 48–52 intensive 42n passive 43–48, 50n, 52–53 pluperfect 14, 26, 41, 50n, 67, 71, 72, 76, 105, 146, 150, 151, 154, 155, 190n, 363n subject-oriented vs object-oriented 42– 52 person first-person vs. third-person narration 156–157, 163, 189 first person future 187 in historic present 102 in inquit formulae 210–211 pluperfect: see perfect stem present stem aspectual value 40–41, 71, 81–82, 125– 132, 150–151, 242 atelicity 129, 131, 150 conative 82, 150 durative 32, 82, 83, 90, 127, 136, 150, 155, 167, 242, 283 indicative 41, 61–64 future use 63–64 generic/habitual/omnitemporal use 41n, 61–65, 69, 82, 126, 127, 129, 150, 155, 158, 162, 163n, 188–191, 192n, 195, 203, 204, 205, 206, 207, 247– 248 historic present 9n, 62, 64, 76, 77, 99– 122, 149, 151, 161n not a marker of ‘vividness’ 101, 379 not normally with negation 103 marks decisiveness 99–101, 160 rare in subordinate clauses 103 semantic features 101–107 γίγνονται in X. An. 1.1.1 397–403

imperfect 15, 22, 25, 26, 33, 41, 50, 82n, 214n, 105, 191, 198, 204, 363, 402 iterative 22 expressing simultaneity 30, 269n of consecutive action 82n unaugmented 101, 107–131 absent from Linear B 149–150 as past tense 146 at start of narratio 147–149 discourse function 60–79 στεῖχον in Hes. Th. 10 186, 189, 191, 204 focalising use 189–191, 204–205 framework use 16–19, 71, 73–77, 146, 150, 192 in Substitutionary Perception 136, 151–165 iterative use: see generic/habitual/omnitemporal use negated 21–23, 103 participle 33–36 ‘protocol’ use 87, 90, 91, 97 relative tense 15–16, 63 timeless use: see generic/habitual/omnitemporal use verba dicendi 24–30, 193, 214–216, 219 γίγνονται in X. An. 1.1.1 397–399, 400– 402 presupposition 243, 244, 249, 251, 252, 333n in Substitutionary Perception 135 proper names 139, 159n, 161, 228, 319, 323– 324n, 397, 403–405 with or without article 282n, 283n, 319– 333, 393 Präterit (German tense) 138, 141 question 97, 102, 216, 218, 221, 299, 335– 356 introduced by τί οὖν οὐκ 63n rhetorical 90, 91, 102n wh-question 371–373 yes-no question 371–374 also see answer reflexivity 365 relative clause 188 digressive vs. restrictive 233–234 relative time/tense 7n, 11–12, 63, 67

410 repetition in answers 339, 343–344 Rosetta Stone 50–51, 54–56 sattelite 291n first-, second- and third-order entities 264, 266, 268, 270n, 272n, 275 semantic functions 43, 44, 46–48, 52–53 sentence type: see exclamation, question, declarative sentence subjunctive 72, 124n, 126n, 187n, 194, 201n, 202, 205 + ἄν 238–252 generic/habitual/omnitemporal use 246–250 Substitutionary Perception 133–141, 151–167 tense 6–7 see also absolute tense, aorist stem, etc. turn-taking: see change of speaker

index rerum Theme 298n Topic 223–236, 231, 234–235, 224, 298n, 317 New 228–230, 308 Resumed 224, 228–236 Topic shift 304, 307, 312, 315 verbal system 5, 39–41, 61, 361, 367 verbs of approaching 274 of giving 258–259 of reaching 256, 272–274 of speaking 23–30, 112, 193, 214–216, 219 of violence/seizing 256, 259n, 276 of volition 124–125 voice: see diathesis, middle voice Window Opener 133, 138, 151, 159, 161, 166

Index Graecitatis Ancient Greek ἄγγελος 170, 180n, 181, 182n ἄλγος/ἄλγεα 254–255, 258, 261, 262, 264, 270, 272, 274, 276 ἀληθῆ, ἀληθέστατα (λέγεις) 213–214, 337 ἀκροᾶσθαι/ἀκροάσασθαι 124–131 ἀναγίγνωσκε/ἀνάγνωθι 84n ἀνάγκη 338, 342 ἀπαγορεύειν 81 ἀπειπεῖν 81 ἀποκρίνεσθαι/ἀποκρίνασθαι 86–87, 88–89 αὐτίκα/τὸ αὐτίκα 290 αὐτοπάθεια 362, 365–366 ἄν 238–252 αὐτάρ 201–202 ἄχος/ἄχεα 259, 263, 270, 272, 274, 276

ἐπεί οὐ 22 ἔπειτα/τὸ ἔπειτα 290 ἐρώτα 87–88 ἐρώτησις 370–374 ἔστι ταῦτα / ἔστιν οὕτω 337, 342 εὕδειν 102–103 εὑρίσκειν 161n ἔφη/ἔφην 210–222 ἔφη ὑπολαβών 212–213 ἔχειν 102–103, 257–258

γάρ 104–105, 113, 116 γε 339, 343 γίγνεσθαι/γενέσθαι 269–271, 395–396, 400, 403–405 γίγνονται (in X. An. 1.1.1) 397–399, 400–402

θεράπων 170, 180, 181 θρῴσκειν 119

Δαρείου καὶ Παρυσάτιδος (in X. An. 1.1.1) 395–396, 400, 403–405 δέ 104, 215–216, 226, 228, 230–235, 293–317 διάθεσις 357–369 διδόναι 258–259 διηγεῖσθαι/διηγήσασθαι 147 δοκεῖ μοι 338, 342 ἐάν 238–252 εἰ 239–242, 246n, 252 see also ἐάν εἰἑν 216–217 εἰκός (γε) 338, 342 εἶναι 102–103, 233, 269–271 εἰπέ 80–97 εἰς + adv. 290 ἐλέγχειν/ἐλέγξαι 86–87, 89, 90–91, 93 ἑλεῖν 271–272 ἐναγώνιος 101n ἐνέργεια 357–361, 363, 366–367 ἐπεάν 12, 238, 243–252 ἐπεί 13, 15, 23, 171–176, 179–184

ἤ 285 ἦ δ’ ὅς/ἦν δ’ ἐγώ 210–222 ἦλθον, εἶδον, ἐνίκησα 143 ἤρετο 221

ἵστασθαι

154n

καὶ 104, 293–317 καὶ … αὖ 303–304 καὶ … γε 303–304 καὶ … δέ 293–317, 394–395 καὶ … δή 302–304 κακόν/κακά 258, 261, 268–269, 270 καλεῖ/κάλει 109–112, 116 κεῖσθαι 102–103, 120–122 κῆδος/κήδεα 258, 259, 261, 263–265, 270, 272, 274, 277 κινδυνεύει 338, 342 κυνεῖ/κύνει 101, 108, 117–118 κυρεῖ/κύρει 108, 110–111, 113–116, 117–119 λαβεῖν 271–272 λέγε 80–97 μάλα (καὶ μάλα, μάλα γε, etc) 339, 343 μέν νυν 76 μένειν 102–103 μεσότης (μέση) 357–360 μεταξύ 82, 151 μόρος 255

412

index graecitatis

ναί 339, 343 νομίζειν 102–103 νυ 192 νῦν 50n ὁμολογῶ 338 ὁ 319–333 see also article ὀϊζύς 258, 259, 261, 266–267, 270, 272, 274 ὅπως 171n ὀρθῶς 337 ὅταν 12, 205, 238, 244–246, 252 ὅτε 11n, 20 οὐ 336–345 (οὔ/οὐδαμῶς/οὐ δῆτα/etc.) 343–347 οὐδὲ … δέ 313–314 οὐδέτερον 359–362, 364, 365n οὐκοῦν 336–347 οὗτος 223–236 πάθος 357–362, 367 παντάπασι (παντάπασί γε, etc.) πάνυ γε 339, 343 παύεσθαι 81, 151

339, 343

πένθος 258, 259, 261, 267, 270, 272, 274 πεῦσις/πύσμα 370–374 pe–re (φέρει/φέρε) 150 πῆμα/πήματα 261–263, 268, 274, 277 πόνος 258, 259, 261, 267, 270 πρότερον/τὸ πρότερον 279–292 πρῶτον/τὸ πρῶτον 287 πῶς δ’ οὐ / πῶς γὰρ οὐ 339, 343 σχολή 126, 129–131 τε (‘épique’) 188 –τερος 285n, 287 τί οὖν οὐκ 63n τιθέναι 260–261 τοι 200 τοίνυν 218 φάθι 87n φαίνεται 338, 342 φημί 338 φοιτᾷ vs. φοίτα 109–110 ὡς + future participle 395

Modern Greek δέν/δε(ν) 348–349

όχι 348–349, 351, 352–354

κανείς/κανένας

τίποτα 348n

348n

Index Locorum Aeschines In Timarchum (1) 43 De falsa legatione (2) 62 In Ctesiphontem (3) 59–60 Epistulae 7.4 Aeschylus Agamemnon 1055 Persae 313 376 416 458 506 Choephori 738 Prometheus Vinctus 6

148n 129 129 127

130 108 108 108 108 108 10 54n

Alexander Aphrodisiensis in Aristotelis Topicorum libros octo commentaria 539.18 373n Alexander Rhetor De figuris (Spengel) III 24 ff. Alcaeus fr. 2D

371, 372–373

196n

Ammonius In Aristotelis de Interpretatione commentarius (Hülser) 200.5 ff. 372 Andocides 1.9 1.14 1.69

128, 129 339n 127

Anonymus De figuris (Walz) VIII 703–704

373n

Apollonius Dyscolus De contructione (Uhlig) 2.2.237.9 366 2.2.296.2 361 2.2.395.3 362 2.2.398.5ff. 362 De constructione (Lallot) 1.30 (=2.2.28.1) 371n 2.51 (=2.2.161.12) 371n 3.21 (=2.2.286) 81n Apollonius Rhodius 4.1381

191n

Apsines Rhetorica (Spengel–Hammer) I 305 373n Aquila Romanus (Halm) 11–12 374n Aristophanes Acharnenses 409 Ecclesiazusae 477 888 Equites 870 Lysistrata 6 Nubes 1462 Pax 243–244 246 250 523 718 Ranae 30 164

130 311 308 346 311 296 310 310 310 311n 311 62 310

414 Ranae (cont.) 271 608 Vespae 8 652 Aristoteles Metaphysica 1034a17 Physica 8.2.253a10 ff. 8.6.259b10 ff. Topica 158a Charisius (Keil) 1.164 ff. 1.165.3

index locorum

323n 323n

Demosthenes 8.23 9.55 9.70 18.2 18.148 21.13–18 21.189 34.5 ff. 42.1 43.2 47.3 52.11

300 314

346 82n

Diocles Magnus =Stoic. Veterum Fragm. II 186 (Hülser) 874 372

350n

Diogenes Laertius 7.64 7.66

367 367 373

365 360n

Choeroboscus (Uhlig–Schneider–Hilgard) 4.2.98.30 ff. 363n 4.2.99.28 ff. 363 4.2.101.16 ff. 364 4.2.208.6 362n Cicero De domo sua 51 De finibus 2.2 De oratore 2.255 De senectute 20

57.4 59.126

345n 374n 374n 374n

127 127 300 128, 129 99n, 101n 148n 300 148n 302 129 128, 129 302

360, 366 370, 372

Dionysius Thrax (Uhlig–Schneider–Hilgard) 1.1.48–49 357 (Lallot) 12.80 (1.1.39.1–2) 373n Elias In Aristotelis Categorias commentaria 146 216n Etymologicum Magnum τίς 373n Euripides Fragmenta fr. 495.36 Alcestis 158 ff. 176 183–184 184 Andromache 1085 ff. 1153 1159 Bacchae 50–51 333 409–411 677 ff. 723–729 760 767 940 1043 ff. 1066 1084 1134

108 172–173, 181 117n 117n 103 149, 173, 181 102n 108n, 122n 238–240 308 196n 177 117–119 105n 108 244 175, 181 108 108 108

415

index locorum Cyclops 199 Electra 686–687 774 ff. 777 1117 Hecuba 521 ff. 580 963 1153 Helena 605 ff. 1526 ff. 1584 Heraclidae 800 ff. 825 Hercules Furens 923 ff. Hippolytus 760 1173 ff. 1212 1236–1237 Ion 1122 ff. 1205 Iphigenia Aulidensis 414 ff. 1543 ff. 1579 Iphigenia Taurica 261 ff. 330 334 625 1327 ff. 1395 Medea 1136 ff. 1141 1207 1243 Orestes 257 866 ff.

176–177 108 113–114 108

1400 ff. Phoenissae 946 1090 ff. 1207 ff. 1219 ff. 1359 ff. 1410 1428 ff. 1458 Supplices 650 ff.

177–178 174, 180 112

Eusthatius Ad Homeri Iliadem (Van der Valk) 3.623.23 196n

173, 180 112

Herodotus 1.1.1 1.1.2 1.5.4–6.1 1.6.1 1.6.2 1.7.3 1.9.3 1.10.2 1.13.1 1.13.2 1.14.2 1.17.1 1.24.1 1.31.5 1.50.2 1.50.3 1.60.4 1.62.1 1.65.1 1.66.4 1.70.2 1.71.1 1.71.3 1.72.1 1.74.2 1.75.2 1.75.3 1.77.1 1.77.2 1.85.4 1.96.2

200 65 174, 181 119n 307n, 310

173, 181–182 117n 176 103n 103 174, 183 108 178 171n, 175, 181 117n 174 108n 108n 245 174, 181 105n, 107n, 117n 173, 181 100–101, 108, 115–116, 117, 118 108, 117n, 122n 296n 119 149, 177

178 308 174–175, 182 175n 178–179 174–175, 182–183 105n 174–175, 183 105n 177

234 227n 148n 66, 227–228 225 33, 288 243 63, 100 245 225n, 229 225–226, 230–231 150 225n 225n 231n 225n, 230n 225n, 228 225n 225n 225n, 233–234 225n 17, 18 239–240 282 69, 225n, 233n 17 17n 21n 28 233n 225n, 229

416 Herodotus (cont.) 1.114.1 1.133.4 1.141.2 1.142.4 1.143.3 1.144.3 1.146.3 1.149.2 1.157.2 1.157.3 1.160.4 1.162.2 1.163.1 1.165.2 1.176 1.176.3 1.183.3 1.184 1.187.3 1.188.2 1.191.6 1.194.1–4 1.198 1.199.3–4 1.201 1.205.1–206.1 1.206.1 1.208.1 1.214.1 2.1.2 2.2.1 2.10.3 2.29.2 2.41.4 2.44.1 2.94.1–2 2.113.1 2.113.2 2.121β–γ.1 2.121ε.4 2.118.4 2.147.2 2.162.3 3.3.3 3.4.1 3.36 3.39.1 3.41.2

index locorum

68, 150 250 23 225n, 231–232 225n 225n 225n, 231 225n 22 225n, 233n 225n, 233n 225n 225n 21n 18 225n 225n 225n 225n, 233n 225n 67 246–247 65 250 225n, 226, 233n 18 151 67 229 17 83 46–47 246 32 309 248 31n 235n 16 34 22 33, 83n 28 19, 244 17 23 19 30

3.50.2 3.54.1 3.56.2 3.86.1 3.88.1 3.113.1 3.126.2 3.135.1 3.159.1 4.1–5.23.1 4.5–82 4.43.1 4.68.4 4.82 4.89.1 4.89.3 4.115.1 4.117 4.119.1 4.143.1 4.144 4.144.3 4.154.4 4.163.1 4.167.3 4.200 4.205 5.1 5.11 5.14 5.15.3 5.27.1 5.30.2 5.42.2 5.57.2 5.63.3 5.79.2 5.81.2 5.92β–η.4 5.92β.1 5.99.1–100.1 5.104.2 5.108.1 6.13.1 6.19.3 6.25.1 6.27.2 6.63.2 6.86.1

25n 19 19 31n, 234 23 241 26n 28 67, 281n 75 76 23 308 76, 77 29 73 13 23 27 76, 77 77 77 30 74 77 77 77 77 76 77 31 33 20 28 29 26 31 27 72 72 18 22 8n 26n 34n 26n 76 281n 28

417

index locorum 6.89.1 7.1.2 7.3.2 7.4 7.6.2 7.6.4 7.8–10 7.8α.2 7.8δ.1 7.10.1 7.10β.1–2 7.10δ.2 7.10ε 7.10η.1 7.15.1 7.16.1 7.27.1 7.49.5 7.58.3 7.59.2 7.59.3 7.70.2 7.72.1 7.75 7.80 7.83.1 7.86.2 7.87 7.89.1 7.89.2 7.91 7.99.1 7.103.3 7.109.2 7.111.2 7.113.2 7.117.2 7.123.2 7.136.1 7.142.3 7.143.1 7.144.2 7.145.1 7.150.2 7.152.1 7.153.1 7.155.2 7.164.1 7.166

26 29 225n 33 225n, 226, 232, 233n 29 71 68, 147 34 21n 68, 147 64 68, 189 23 29 23 225n 29 226–227n 225n, 233 225n 225n 225n 288 225n 225n 23 225n 23 225n 225n, 233n 21n 21n 225n, 233n 225n, 233n 225n 225n 226 29 225n 225n 225n 29, 225n 30n 30n 225n, 233n 225n 225n 23

7.168.1 7.169 7.171.2 7.180 7.181.1 7.184 7.186.1 7.190 7.198.1 7.211.2 7.211.3 7.212.2 7.213.2 7.213.3 7.217.1 7.225.3 7.220.4 7.222 7.223.4 7.229.1 7.234.1 8.7.1 8.25.1 8.37.2 8.70.1 8.76.1 8.80.2 8.90.1 8.94.2 9.18.2 9.20 9.69.2 9.80.2 9.93.2 9.98.2 9.94.1 9.122.2 Hesiodus Opera et Dies 43–46 633–635 Theogonia 1–115 1–21 1–10 1 2–8 2

25n 35 34 225n, 233n 14 16 225n 225n 225n 23 23 22 29 225n 227 225n 29 23 35n 28 28 310n 27 31n 29 26n 244 73 31n 27 22 31n 23 192n 21 235n 64

202 192n 185–208 192 162, 186, 191n 187 191 188

418 Theogonia (cont.) 4 5–9 8 9–34 9–10 9 10–23 10–11 10 11–21 22–23 22 24 25 30–34 30–31 30 31–32 31 32 33–34 33 34 35–36 35 36–37 36 37–67 37 39–40 39 40 42 43–46 43–45 44 45 46 47 50–51 50 52 53–60 53 55 56–57 60–68 60–61

index locorum

188, 204 188 191 187 189 189 189 199 133, 189, 191, 204 199 189 192–193, 198–199, 204 192–193 193 194n 192n 192, 200, 203 193 192, 203 203 193, 203 187, 203, 204 195 187n 194 195, 202 194, 195n, 196n 187n, 193 195 199, 201 195 195 195 203 204 193, 195, 202 195 195 202 204 202 195 187n 195, 196n 195 195 198 196, 203

60 62–62 62 63–67 63 66 67 68–79 68–70 68–71 68 69 70 71–74 71 72 75 75–79 79–92 80–103 80–92 81–82 81 82 83–84 83 87 90 93 94–96 94 95 96–97 96 97 98–103 99–101 99 100–101 100 102–103 102 103 104–115 104–106 104–105 104 108 113

196 188 195, 196n, 197 196 188, 198 195 195 196 163, 204 187n, 199 196, 198, 199, 204 199 199 187n, 199 198n, 199, 204 198 199, 204 187n 200 187n 199, 200 200 201 200 199, 201 200 202 203 200, 203 200 200 201 200 201 199, 201 201 204 202, 203 202 204 203 201n, 202 203 187n 203 204 203 188 188

419

index locorum 114–115 114 191–192 192–193 201–202 Homerus Ilias 1.163–164 1.188 1.219 1.234–235 1.254 1.362 1.386 1.417 1.444–445 1.467 2.42 2.135 2.195 2.272–274 2.375 3.97–98 3.288–291 3.412 3.351 3.354 4.169 4.456 5.297–310 5.394 5.470 5.667 5.759 6.110 6.116 6.261–262 6.282 6.355 6.413 6.489 7.36 7.113 7.124 7.173 7.375 9.229 9.249

204 202, 203 198n 198n 198n

200 269 215n 45 273 273 193 267 258 268 267 43 261n 45 255 272 201, 202 257 261n 261n 269 270n 16 271 36 257 269 193 193 43 268 273n 259, 269 198n 268 311n 273 311n 311n 262n 269

9.412–416 9.412 9.591–592 10.89 10.131 10.145 10.245 10.279 10.438 11.106 11.347 11.362–363 11.407 11.413 11.657 12.272 12.392 13.73 13.86–87 13.239 13.416 13.417 13.464 13.581 14.80 14.458 14.459 14.475 14.480–481 15.90 15.202 15.235 15.245 15.720–721 16.11 16.22 16.55 16.101–111 16.262 16.508 16.515–520 16.548–549 16.552 16.581 16.599–602 16.651 17.37 17.97 17.99

146–147 68 270 258n 36 45n, 271 258n 258n 44 192 262, 268, 274 274 199 261n 44 311n 269 311n 269n 270n 266n 269 273 271 269 269 268 271 270 43 193 268 273 260 21 271 269 16 261n 269 273 271–272 271 269 271 267 260, 267 199 268n, 274

420 Ilias (cont.) 17.157–158 17.351 17.538–539 17.687–688 18.8 18.52–53 18.53 18.64 18.73 18.79 ff. 18.88–89 18.395 18.429–431 18.431 19.40 19.125 19.178 19.307 19.367 20.28 20.282 20.293 21.12 ff. 21.92–93 21.137 21.249 21.522–525 21.562 22.43 22.52–53 22.122 22.288 22.385 22.389–390 22.421 22.422 22.488 24.91 24.105 24.419 24.525–528 24.547 24.708 24.741 Odyssea 1.10 1.32–34 1.76

index locorum

260n 311n 271n 261 264n 264 265 273 273 265 270 272 264, 265 263 193 271 311n 273 273 311n 273 269 261n 270 268 268 260 199 273n 256 199 268 199 201, 202 268 255 270 257, 263 257 44 255, 259 261, 262n 273 267 187 255 193

1.103–112 1.342 1.392 2.41 2.45–46 2.45 2.163–164 2.163 3.100 3.220 3.291–293 4.35 4.236–237 4.391 4.812 5.5 5.58 ff. 5.63–75 5.206–208 5.289 5.301–302 5.336 6.60 6.169 7.81–135 7.81–137 7.104–132 7.132 7.195 7.211 7.213 7.218–219 7.218 7.219 7.241–242 7.269–270 8.81 8.149 8.263 8.266 8.267–266 8.367–377 8.367 8.480–481 8.480 8.488 8.553 9.12 9.15

161n 273 202 272 193 272n 274 268n 262 263 191n 267n 259 311n 267n 265, 266 161n 154n 263 273 263 254 311n 273 190n 151–153 189 190 262 257 268 267 257n 257n 258 266 268n 266 154n 199n 199n 202 199n 193 200 193 198n 265 265n

421

index locorum 9.37 9.381 9.402 9.423 9.535 10.376 11.115 11.167 11.208 11.369 11.376 11.382 11.481–482 11.542 11.555 11.582–584 11.593–594 12.231 13.301 13.302 14.32 14.47 14.185 14.197 14.270 14.310–312 15.223 15.256–257 15.313–317 15.342 15.344–345 15.399–401 15.487 16.205 17.286–287 17.287 17.439 17.446 17.555 17.472–473 17.477 18.272 18.324 18.376–377 19.167–168 19.471 19.512 20.195–196 21.207

265n 154n 154n 274 262 257n 262n 257 269 265 265 265 257 265 260n 265n 265n 261 258n 311n 263 265 265, 266 265, 266 270 262 198 198–199 202 267 257, 274 265 265 193 259n 258 270 261 266 259n 270 246n 257n 202 259n 271 259 258 193

23.210 23.211–212 23.224 23.306–308 24.226ff. 24.231 24.233 24.462 Hymni Homerici Demeter (2) 1 Apollo (3) 1–13 5 Hermes (4) 1–7 13–18 20–102 23 Aphrodite (5) 260–261 6.3–4 6.6–13 6.6 6.14–16 9.8 11.1 13.1 16.1 19.10–15 19.27–29 19.40–45 22.1 25.1 26.1 28.1

259 266n 273 260, 265n 161n, 267 267 257n, 267 269

187n 186, 191n 191n 195n 198n 198n 196n 186, 191n 197 197 197 197 187n 187n 187n 187n 186, 191n 186, 191n 197 187n 187n 187n 187n

Isaeus De Philoctemone (6) 2

129

Isocrates 4.112 14.6 15.12 15.20 15.28 18.4

130 129 127 127 129 148n

422 Isocrates (cont.) 19.4 21.1

index locorum

148n 148n

[Longinus] De sublimitate 18.1 25

370 101n

Longus 1.2.1–3

160–161

Lysias 1.6 1.22 7.4 10.4 12.55 12.58 13.5 13.79 16.9 18.4 19.5 19.11 30.1 30.24

113–114, 148 147n 148 114 127 50n 148 127 128 50n 297 128 129 48

Macrobius Excerpta Grammatica (Keil) 5.627.5 ff. 364 5.627.38 362n Philostratus Imagines 1.26.1 Pindarus Isthmia 4.35–39 Nemea 3.26 Plato Alcibiades II 147b8 Apologia 20e3–5 21a5

196n

46 199

297 84 84

23b9 40b2 Charmides 164b Cratylus 401c Critias 112a Crito 48b Demodocus 383b3 Epistulae 7.349b2 Euthydemus 273e 277a3 277a4 277a6 277b2 283b 288c 293c 289c9 Eryxias 400e Euthyphro 6c 6d 7c 14a 15c Gorgias 447a 448a5 448d7 451d5 461b–481b 461b1 461b3–462b2 461d1 462a2 462a4 462b 462b1 462b3–467c1 462b3 462b5 462b6–10

130 82n 345n 281n 281n 305n 129 213n 286 216n 216n 216n 216n 288 286 343 213n 308 290n 286 337 284 342 288 124, 125 125 93n 86 87 87 87 87n 87 340n 87, 88 87–89 87, 94n 87 88n

423

index locorum 462c3 462c6–d4 462c10 462d8 462d10–11 463a5 463b8 463c1 ff. 463c8 463e3 463e5 464b2 465d7 465e3 466a1 466a2 466a7–8 466c3 466e9 467c1 467c5–470c7 467c5–468e5 468c7–8 468d6 468e6–469e8 470a1 470a4 470c1 470c7–474c3 470c7–8 470c8 471d5 ff. 472c4–5 472c5 472d3 472d4 473a3 ff. 473a4 473a10 473b7 473b9 473d3 473e6 474b1–2 474c4–479c6 474c4 475a 475d4–5 475d6

87 87 374n 87, 88 87n 88 88 88 88 88 88 88 88 88 88 88 88 88 88 88–89 89 89 89 89 89 89 93n 89 89–91 90 90 90 90 91 90 90 90 90 90 90 90, 91 91 91 91 91–92 91, 93n, 94 307–308 91 126

476a 476a2 476a7 476a8 476b1–2 476b2 477b1 479c4 479d7 480b6 480d6–7 488c2 492d5 493d1 494b7 495a2 498e10 505d1 506c 506c4 506d 511d5 513a8 515a7 515c 516d6 Hipparchus 226b Hippias Minor 363c5 367a7 Ion 530c2 530d6–7 530d9 531b7–9 531e 533c2–3 536d6 536d8 538a 538b8 540e1 541a5 Laches 179b 180d6–8 180e 181b

281n 91 91 89n 89n 91 91 92 92 92n 92n 127 93n 126 93n 93n 92n 126 66 93n 303 125n 126 93n 284 93n 349n 125n 125n 125n 131 125 128 338 125, 129 131 131 342 125 125n 342n 61 83 320n 307

424 Laches (cont.) 184a 185d–e 188a 188c 194e 199d Leges 629e 653c–d 657c 670d 833a–b 962c Lysis 203a1–b2 205d4 207b 207b1 211d 212d 221d Menexenus 235c9 235e1 236a7 236b 236b4 236b7 236c1–10 236d2–3 242d Meno 72c7 73a 84a6 85d6 87e 89a 98d Parmenides 126b Phaedo 59c8 60c8 60d8–62a1 61c8 63a2 64c3

index locorum

63 339 285 304 339 316 281n 280 350n 281n 301 345 211–212, 219 127 151 82n 63, 65 282 287 96n 96 96n 284 96n 96 95 96 283n 125n 346 125n 125n 125n, 302, 303n, 349n 344 344 281 331 213n 219–220 221 331–332 212

65b8 65c–d 67d3 68e1 75a 84c3 86d5 90c–d 91b 92a2 92a4 92a6 94c 105e 105e8 110d 115b1 117a8 117d1 Phaedrus 227b8 232e 237c6 251b 253e5–255a1 256c 259a Philebus 13b 13e1 14b3 14c6 14e2 15a4 16c4 21a7 23a7 23d6 25c2–4 25c4 25e6 27e4 28d1 29a6–8 29a8 30a2 31d3 32e4–5 33a7

331n 338 331n 331n 281n, 287 332 331n 339 289 331n 331n 331n 339 343–344 217n 316 217n 217n 332 129 284–285 194n 281n 133, 164, 204–205 281n 242 299 93n 93n 93n 93n 93n 93n 93n 93n 93n 94 93n 93n 93n 93n 94 93n 93n 93n 94 93n

425

index locorum 33a8–9 33c 36e1 40c6 41c10 43a9 43d 45d5 46d7 48c5 49c9 50d6 52d10 53d2 53d8 53e2 54b2 61d6 64b9 65e1 Politicus 257c8–10 Protagoras 310e 316e 321b 331b 332c 336b7 353d 360c6–7 Respublica 332a6 334d2 336a8 337c2 343e 350c10 ff. 372c4 372c5 373d 380c6 392c9 395b7 402b 406c5 406d4 427e–445e 430c1

95 284 93n 316 93n 93n 343 93n 93n 93n 93n 93n 93n 93n 93n 93n 93n 93n 93n 93n 84n 286 281n 281n 314n 338 212 290 338n 214, 216n 214 214n 216n 309 217 214 216n 288 218 215n 214 343 129 129 94n 214

436c4 436e 444a3 445c3 445d2 450b–487a 451b5 451b8 457c6 459c7 463e7 467b7 471e3 472b2 473c10 475b10 495b7 500b8 504c9 518e 542c 564b4 566a 584b1 601c14 608d Sophista 218a 247e Symposium 176e 185e4 186e 189a1 189a7 193d 194e4ff. 210e Theaetetus 144e 171e 202b Timaeus 20c Plautus Pseudolus 318

94n 308 214 94n 94n 94n 94n 94n 94n 214 214 214 94n 82n, 94n 94n 214 214 129 218 304 339 218 340n 218 214 84n 281n 290n 281n 82n 316 332n 332n 290 96n 290n 338 300 296 290n

345n

426

index locorum

Captiva 262

347n

Plotinus Enneades 6.1.19–20

367

Plutarchus Caesar 50.2–3

142–143

Priscianus (Keil) 2, 376 ff.

365

Quintilianus Institutio Oratoria 4.2.12

143n

Rosetta Stone decree Preamble, ll. 9–36

50–51, 54–56

Scholia in Hesiodi Theogonia 463 373n in Dionysium Thracem (Hilgard) 239.14 ff. 373n Sextus Empiricus Adversus mathematicos 8.70 372 Sophocles Ajax 308 764 773 816 1107 1185 Antigone 254 305–314 391 404–407 407 423 432–433 432 442

107 105n 112n 130 82n 290n 113–114 199n 199n 102–103 108n 103 108n 295n 52n

1001 1226–1227 1236–1240 Electra 35 715–716 725 732 738 746 757–761 Oedipus Coloneus 825 854 1606–1607 1624 1626 1627 1655 Oedipus Tyrannus 775–780 812 1245 1246 1249 1255 1268 ff. Philoctetes 371 923–924 936–940 1360–1363 1362 Trachiniae 413–414 701 765 767 785–786 796 904 915 Suda ἀξίωμα κατηγόρημα πυνθάνεσθαι πύσμα(τα)

105n 103, 112 120–121 105n 108 102n 102n 102n 102n, 103 121–122 49n 49n 108 108 112, 116 111 111 105 103 111–112 109 108 109–110, 112 105n 108, 110–111, 113–116 48 49 306 293 346n 121 112 110, 112 106 110, 112 108 108

373n 373n 373n 373n

427

index locorum Suetonius Caesar 37.2

143

Terentius Adulescens 569

347

Theodosius Canones (Uhlig–Schneider–Hilgard) 4.1.49.17 ff. 363n 4.1.49.25 363 4.2.99.25 363 Theon Progymnasmata (Spengel) II 96–97 373n Theophrastus Characteres 23.3 23.4 23.6 23.7

307n 307n 307n 307n

Thucydides 1.24 1.25.3 1.26.5 1.36.1 1.53.2 1.63.2 1.91.5 1.132.4 1.144.2 2.17.4 2.21.3 2.36.1 2.48.3 2.64.6 2.84.1–3 2.93.4 3.87.2 4.24.2 5.24.2 5.26.1 5.115.4 6.17.3 6.71.2

70 305n 82n 281 193 155 308 205, 313 296 306n 127 295, 313 241 290 158–159 82n 281n, 287 294n, 295, 313n 48 47 281n 127 294n, 295, 305, 313

6.86.1 6.91.3 7.18.2 7.56.3 7.70.7 8.1.4 8.8.2 8.67.3

287 63 283n 295, 313n 189 82n 281n 295, 313n

Vergilius Aeneis 1.439–454

165–166

Xenophon Agesilaus 10.1 Anabasis 1.1.1–4 1.1.1 1.1.2–10 1.1.2 1.1.3 1.1.4 1.1.7 1.1.8 1.1.9 1.1.10 1.2.5 1.2.10 1.2.14 1.3.1 1.4.12 1.5.1–4 1.5.3 1.5.5 1.5.6 1.6.1–5 1.6.2 1.6.5 1.6.7–9 1.6.10 1.7.9 1.8.2 1.8.13 1.8.20 1.9.28 1.10.1 1.10.8 2.3.1

308–309 392–405 285n, 324 324 328 328n 328n 328 309, 328 328 328 156 329 329 329n 328n 157 190 105 155, 190 324–325 328n 328n 324–326 325n, 328n 325n, 328n 299 328n 313 156 103n 103 156

428 Anabasis (cont.) 2.6.3 2.6.5 2.6.6 3.1.5 3.1.9 3.1.13 3.1.34 4.1.15 4.1.17 4.7.13 4.8.18 5.1.9 5.3.9 5.6.15 6.4.1–6 6.4.14 7.1.30 7.4.6 7.7.17 7.7.55 7.8.21 Cyropaedia 1.4.26 1.6.17 2.1.4 2.1.16 2.4.17

index locorum

103–104n 328n 156 323n 323n 104 327 103 130 105n 82n 130 301 302 190 330n 314 106 246n 330n 330n 311–312 130 68 130 312n

3.3.19 4.2.22 4.3.12 4.5.8 4.5.46 7.1.1 7.1.32 7.2.10 7.5.50 8.2.25 8.3.48 8.4.17 Hiero 10.5 Historia Graeca 2.1.21 3.2.31 5.2.37 6.5.16 Memorabilia 1.6 1.1.15 4.4.23 Oeconomicus 1.3 Symposium 2.9

306 130 130 305 305n 312 300n, 301n 310 130 308 130 301n 130 155, 191 103n 305n 150 130 302 344 312 298