Textual Strategies in Ancient War Narrative: Thermopylae, Cannae and Beyond 9789004383340, 9004383344

Citation preview

Textual Strategies in Ancient War Narrative

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 29

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

Textual Strategies in Ancient War Narrative Thermopylae, Cannae and Beyond

Edited by

Lidewij van Gils Irene de Jong Caroline Kroon

LEIDEN | BOSTON

The Library of Congress Cataloging-in-Publication Data Names: Gils, Lidewij van, editor. | Jong, Irene J. F. de, editor. | Kroon, Caroline, editor. Title: Textual strategies in ancient war narrative : Thermopylae, Cannae and beyond / edited by Lidewij van Gils, Irene J.F. de Jong, Caroline Kroon. Description: Leiden ; Boston : Brill, 2018. | Series: Amsterdam studies in classical philology ; volume 29 | Includes bibliographical references and index. Identifiers: LCCN 2018043297 (print) | LCCN 2018044721 (ebook) | ISBN 9789004383340 (ebook) | ISBN 9789004383333 (hardback : alk. paper) Subjects: LCSH: War in literature. | Greece–History, Military. | Rome–History, Military. | Livy–Criticism and interpretation. | Herodotus–Criticism and interpretation. | Thermopylae, Battle of, Greece, 480 B.C. | Cannae, Battle of, Italy, 216 B.C. Classification: LCC DE61.W35 (ebook) | LCC DE61.W35 T49 2018 (print) | DDC 355.02/0938–dc23 LC record available at https://lccn.loc.gov/2018043297

Typeface for the Latin, Greek, and Cyrillic scripts: “Brill”. See and download: brill.com/brill‑typeface. ISSN 1380-6068 ISBN 978-90-04-38333-3 (hardback) ISBN 978-90-04-38334-0 (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.

Contents Preface vii Notes on Contributors 1

viii

Introduction 1 Lidewij van Gils, Irene de Jong and Caroline Kroon

Part 1 Thermopylae 2

Thermopylae: Herodotus versus the Legend Hans van Wees

19

3

A Narratological Comparison of Herodotus and Diodorus on Thermopylae 54 Mathieu de Bakker

4

Narrative and Identity in Thermopylae (Herodotus 7.201–7.239) Antonis Tsakmakis

5

Herodotus’ Handling of (Narratological) Time in the Thermopylae Passage 113 Irene de Jong

6

Herodotus and Thucydides: Distance and Immersion Rutger Allan

131

Part 2 Cannae 7

Livy on Cannae: a Literary Overview Stephen Oakley

157

8

Discourse-Linguistic Strategies in Livy’s Account of the Battle at Cannae 191 Lidewij van Gils and Caroline Kroon

91

vi

contents

9

Who Knows What Will Happen Next? Livy’s fraus Punica from a Literary Point of View 234 Dennis Pausch

10

Livy’s Use of Spatial References in the Cannae Episode: from Structure to Strategy 253 Lidewij van Gils

11

ET RATIO ET RES: Characterization of Roman Conduct through Speech Representation in the Battle of Cannae 273 Michel Buijs

12

Words When It’s Time for Action: Representations of Speech and Thought in the Battles of Cannae and Zama 293 Suzanne Adema

Part 3 Beyond Thermopylae and Cannae 13

Thermopylae and Cannae: How One Battle Narrative Enriches Another 319 Mathieu de Bakker and Michiel van der Keur

14

The Great and the Small: Thermopylae and Sphacteria Adriaan Rademaker

15

Force, Frequency and Focalisation: the Function of Similes in the Battle-Narrative of Vergil, Aeneid 10 359 Stephen Harrison

16

Parallel Plotlines: the Function of Similes in the Battle Narrative of Vergil, Aeneid 10 (2) 376 Michiel van der Keur Index

395

342

Preface This book is based on contributions to and discussions at a conference in October 2014 in Amsterdam about Textual Strategies in Greek and Latin War Narrative. This conference was the last in a series of workshops, in which linguists and narratologists working in the field of Classics ‘joined forces’ in order to cross-fertilize narratological and linguistic approaches to the study of Ancient Greek and Latin texts. This joint venture was initiated in 2005 within the National Research School for Classical Studies in the Netherlands, OIKOS, when senior and junior researchers from six Dutch Universities and a number of associate members from Belgium, Germany, the United Kingdom and the USA came together at a conference in honour of the retirement of Albert Rijksbaron, professor of Greek Linguistics at the university of Amsterdam. The results of that conference were published as The Language of Literature (Allan & Buijs, Leiden 2007). Six more workshops called Linguists and Narratologists Joining Forces were organised. In 2010 Caroline Kroon and Irene De Jong initiated a combined narratological and linguistic research program Ancient War Narrative, with financial aid from the Netherlands Organization for Scientific Research (NWO). The program included projects on ekphrasis in Ancient Greek Literature (Ancient Greek ekphrasis. Between description and narration. Koopman, Leiden 2018), reported speech and thought in Latin War narrative (Words of warriors. Speech and Thought in Latin War Narratives. Adema, Leiden 2017), and the role of the narrator in Latin War Narrative (Latin War narrative. Textual Strategies in Caesar, Sallust and Tacitus. Stienaers forthc.). The present volume on Textual Strategies in Ancient War Narrative: Thermopylae, Cannae and Beyond, forms the conclusion of that program, in which its results are tested on particular instances of Greek and Latin war narrative and confronted with historical and philological approaches, which are also included. We would like to thank the Netherlands Organisation for Scientific Research (NWO) and the Dutch National Graduate School in Classical Studies (OIKOS) for their generous financial support, and Michiel van der Keur, Mike Kruijer and Nicky Vos for assisting us during the editorial process. Lidewij van Gils, Irene de Jong, Caroline Kroon Amsterdam, January 2018

Notes on Contributors Suzanne Adema is Assistant Professor of Latin at the University of Amsterdam. Her publications are characterized by a combined narratological and linguistic approach to Latin epic and historiography. She coordinates a research project on Greek and Latin Learning and Instruction. Rutger Allan is a Lecturer in Ancient Greek at the Vrije Universiteit Amsterdam. He has published on a variety of topics in Ancient Greek linguistics relating to verbal semantics and discourse pragmatics. He has a special interest in cognitive linguistic and narratological approaches to Greek narrative texts. He is the author of The Middle Voice in Ancient Greek. A Study in Polysemy (2002) and co-editor of the volumes The Language of Literature (2007) and The Greek Future and its History (2017). Mathieu de Bakker is University Lecturer of ancient Greek at the University of Amsterdam. He has published on the Greek historians and orators and is co-author of The Cambridge Grammar of Classical Greek (2018). Michel Buijs (Ph.D. 2003, Leiden University) teaches Greek and Latin at Utrecht University, as well as through his own company, Classix. He is author of Clause Combining in Ancient Greek Narrative Discourse (2005) and co-editor of The Language of Literature: Linguistic Approaches to Classical Texts (2007) and The Historical Present in Thucydides: Semantics and Narrative Function (2011). His research interests include discourse linguistics and description of the Ancient Greek and Latin languages. He is currently writing, as co-author, a new Dutch course book for Ancient Greek language acquisition at universities. Lidewij van Gils obtained her PhD in 2009 with an analysis of the narrationes in Cicero’s speeches. Currently, her focus is on Latin historiography and the linguistic aspects of common ground. She is Assistant Professor of Latin at the Vrije Universiteit Amsterdam.

notes on contributors

ix

Stephen Harrison is Fellow and Tutor in Classics at Corpus Christi College, Oxford and Professor of Latin Literature in the University of Oxford. He is author and/or editor of many books on Latin literature and its reception, especially on Horace, Vergil and Apuleius, most recently of a commentary on Horace Odes 2 (CUP) and Victorian Horace: Classics and Class (Bloomsbury), both 2017. Irene de Jong is Professor of Ancient Greek at the University of Amsterdam. She specializes in the analysis and interpretation of the forms and functions of narrative (Homer, Herodotus, and Greek narrative at large), making use of the modern theory of narratology. Publications include A narratological commentary on the Odyssey (Cambridge 2001), Homer Iliad Book XXII (Cambridge 2012), and Narratology and Classics. A Practical Guide (Oxford 2014). Michiel van der Keur is Assistant Professor of Latin at the University of Groningen. He obtained his PhD in 2015 with a commentary on Silius Italicus’ Punica 13. His research interests include Flavian epic, intertextual techniques, genre crossings and narrative structure. Caroline Kroon Ph.D. 1995, is Professor of Latin at the University of Amsterdam. She is author of the monograph Discourse Particles in Latin (Amsterdam, 1995), and of many articles on discourse-linguistic topics. Her current research is characterized by an interdisciplinary approach crossing the border between linguistics and literary studies. Stephen Oakley Ph.D. 1984, is Kennedy Professor of Latin in the University of Cambridge and Fellow of Emmanuel College; previously he taught at the University of Reading. He is the author of A commentary on Livy, book VI–X (Oxford, 1997–2005). Dennis Pausch wrote his PhD (Biographie und Bildungskultur, Berlin 2004) and his second book (Livius und der Leser: Narrative Strukturen in Ab Urbe Condita, München 2011), which was awarded the Bruno Snell Prize of the Mommsen-Gesellschaft, at Gießen University and during his research stay in Edinburgh as Feodoy Lynen Fellow of the Alexander von Humboldt Foundation. After that, he taught Latin at Regensburg University, before he took over the chair of Latin Literature at Dresden in 2014.

x

notes on contributors

Adriaan Rademaker is Assistant professor of Ancient Greek at Leiden University. His research focuses on linguistic analyses of Greek literary texts (discourse analysis) and pays special attention to the rhetorical and persuasive strategies employed in Greek literary discourse. Antonis Tsakmakis is Associate Professor of Greek at the University of Cyprus. His research interests include Greek Historiography and Old Comedy. He is the author of Thukydides über die Vergangenheit (Tübingen 1995) and co-editor of Brill’s Companion to Thucydides (Leiden 2006), and Thucydides between History and Literature (Berlin—New York 2013). He has also authored a series of textbooks for the teaching of Greek in High School (Nicosia, 2011–2014). Hans van Wees is Grote Professor of Ancient History in the Department of History at University College London. His books include Greek Warfare: myths and realities (2004) and Ships and Silver, Taxes and Tribute: a fiscal history of archaic Athens (2013). He has (co-)edited eight volumes on a range of subjects, including The Cambridge History of Greek and Roman Warfare (2007) and (with Nick Fisher) ‘Aristocracy’ in the Ancient World: re-defining Greek and Roman elites (2015).

chapter 1

Introduction Lidewij van Gils, Irene de Jong and Caroline Kroon

1

War Narrative

War has been a prominent theme in art since the earliest cultures.1 Most recently, the War on Terror has led to a new body of literature, including the use of new genres such as ‘blogs’, but we also find poetry, drama, short stories, novels, journals, diaries, memoirs, letters, graphic novels and comics expressing war experiences. For ancient times, the genres most connected to the topic of war are epic and historiography, but in tragedy, lyric and the ancient novel, too, war is never far away or, when it is, peace is emphatically praised as if to confirm the exceptionality of its presence. Each war may have its specific features in terms of weapons used, types of conflict, and cultures and landscapes involved, but the most dramatic aspects of war are psychological and, therefore, universal: human frailty and heroism, suffering and sacrifice, loyalty and betrayal, love and hatred, reasons for wanting to live or die, beliefs in luck or fate, and, of course, the continuous presence of all-permeating fear. The universality of these themes explains why war literature of ancient times continues to appeal to modern audiences, even literature without a particular interest in the historical context of a conflict. Stories about the Trojan war, the revolt of Spartacus or the battle of Thermopylae still inspire filmmakers, and translations of and monographs about these war narratives attract many readers.2 They are still worth telling as great historical events or as personal or national tragedies. The long-lasting practice of writing war literature has produced canonical texts for particular wars and has led to literary topoi for various aspects of warfare.3 Modern examples of canonical war literature are, for instance, Hem-

1 See e.g. Calloway 2015 for an overview of scholarship on war literature. 2 Famous movies about ancient battles are, for instance, Spartacus (1960, directed by Stanley Kubrick), Troy (2004, directed by Wolfgang Petersen) and 300 (2006, directed by Zack Snyder). The battle of Cannae is known as the biggest defeat of the Roman army, but in spite of its cinematic potential, Hollywood has not yet proved susceptible to the attraction of its story. 3 See e.g. McLoughlin 2009; Sherry 2005; and (for ancient war narrative) Bakogianni & Hope 2015.

© koninklijke brill nv, leiden, 2019 | doi:10.1163/9789004383340_002

2

van gils, de jong and kroon

ingway’s A Farewell to arms and Remarque’s Im Westen nichts Neues for World War I and the Diary of Anne Frank for World War II. For ancient examples, the Trojan War as immortalized by Homer immediately springs to mind, for the Greco-Persian Wars in the fifth century BC including the legendary battle of Thermopylae Herodotus, and for the Punic Wars between Rome and Carthage, including the famous defeat at Cannae, Livy. These canonical texts and their depictions of war may be seen as part of our collective memory or, at least, of the literary tradition of war narrative which continues to influence our thinking about this both glorifying and horrifying aspect of human society. The paradox of war narrative as the expression of one of the most horrific experiences in human existence and one of the richest artistic practices has been frequently noted. In fact, one of the literary topoi of war literature is that the atrocities of war cannot and maybe should not be rendered in words.4 On the other hand, we may find a variety of reasons why war literature needs to be written and read. War literature may aim at informing the reader about great historical events, at sharing with the audience the personal histories involved, or at entertaining the reader with an exciting story. Some texts have been written to cope with personal experiences, others to prevent new wars and still others to explain how a war should be organized or endured. Of course, one aim does not exclude the other, and, as has long been recognized by both historians and literary scholars, it is essential to take into account why and for whom a particular war narrative is written. Another question with regard to ancient war narratives is to what extent they are truthful. Here, the notions of historicity and realism should be distinguished. Historicity in the modern sense of historical truthfulness is not necessarily the highest aim in ancient historiography. Rather, a coherent, likely, persuasive and even pleasant story was demanded. In order to achieve this goal certain literary elements, techniques or topoi were commonly employed (in the case of war narrative for instance a general’s exhortation speech or the portrayal of the enemy), and a lack of reliable sources was no reason to dispense with such elements.5 Realism, on the other hand, is a highly relevant concept, since it concerns the question of whether a war narrative depicts events in such a way as to allow readers to recognize things and hence consider a narrative

4 See McLoughlin 2009: 15 for a discussion of the inadequacy of depicting war because of such factors as complexity, first-hand experience and ethics. 5 See e.g. Fornara 1983 for an introduction to the genre of ancient historiography and White 1973, 1978, 1987 for the modern historian’s use of narrative devices.

introduction

3

persuasive.6 Ancient historiographical works do convey such a sense of realism, precisely because they include the last words of a general or the thoughts of an enemy (which to a modern audience are unhistorical elements).

2

Two Case Studies: Thermopylae and Cannae

This multidisciplinary volume discusses two canonical war narratives from antiquity: the battle of Thermopylae in Herodotus’ Histories (book 7, sections 138–239) and the battle of Cannae in Livius’ History of Rome (book 22, sections 34–61). The narratives about these battles have had an impact which almost surpasses the impact of the battles themselves.7 Who does not know of the three hundred Spartans led by Leonidas who were killed by a massive Persian force under Xerxes’ command at the pass of Thermopylae in 480 BC? Yet, the defeat at Thermopylae did not decide the confrontation between Greeks and Persians. Similarly, the famous Roman defeat in the war against the Punic general Hannibal at Cannae in 216 BC did not lead to Carthaginian rule over Rome. These two war narratives are especially interesting since they both concern defeats. The authors had to present painful yet well-known events from history and this called for particularly creative writing. While all narrators generally employ ‘textual strategies’, as we call them, in these instances Herodotus’ and Livy’s narrative strategy was particularly called for. They had to decide, for instance, whether to tell from a distance or to immerse their readers into events, what details to include about the landscape and material aspects of the battle, which events to select and how to order them in order to stress or cover up strategic behavior by characters, whether to portray their characters as individuals or as stereotypes and, finally, how to explain—and partly soften—the disastrous outcome of the battles. The aim of this volume is to bring together philological, historical, narratological and linguistic perspectives on these two battle narratives, in an attempt to uncover the various textual strategies employed by their authors. Individual chapters offer close readings and comparisons of the battle narratives of Thermopylae and Cannae. On the one hand, the chosen texts merit such an in-depth analysis because of their significance for their original recipients and 6 See for ancient views on realistic storytelling for instance Quint. Inst. 2.3.4, Rhet. ad Herr. 1.13, Cic. Inv. 1.27, as discussed in Woodman 1988. 7 For recent literature on Thermopylae see, for instance, Cartledge 2006; Foster & Lateiner 2012; Matthew & Trundle 2013; Priestley 2014. Recent publications on Livy’s narrative of Cannae include Daly 2002; Levene 2010; and Mineo 2015.

4

van gils, de jong and kroon

their impact on later historical and literary versions of these battles. We thus aim at offering classicists and other readers of these ancient texts new material and ideas which can enrich their own interpretation. On the other hand, the focus on two passages may bring to the fore the strength of the various perspectives and may test whether they can complement and reinforce each other. We hope the example we set here will lead to more combined efforts to investigate ancient texts. A first attempt to broaden the perspective from the genre of ancient historiography to other classical genres (e.g. epic), and from Thermopylae and Cannae to other battles, can be found in the last part of this volume. The strategies to tell about war in classical epic partly overlap with those of ancient historiography but, most interestingly, they also differ. The observations in this last subset of chapters both sharpens our view of what might be typical for the two selected battles and for ancient historiography and stimulates discussion on potentially universal aspects of representing war in textual forms of art.

3

Narratology, Discourse Linguistics and Classics

The study of classical texts has a history which is as old as the texts themselves. Each generation interprets texts anew, inspired by contemporaneous phenomena or new methodologies. The new methodologies which are central in this volume are narratology, which studies the structural features of narratives, and discourse linguistics, which looks at the function of language as communicative act. The introduction of these theories has led to new questions, and, we hope and believe, will lead to new observations and interpretations of the classical texts. In this introductory chapter we offer a basic overview of narratological and discourse linguistic research in the field of Classics and introduce some basic concepts and theses which underly the analyses in chapters to follow. In view of the multidisciplinary readership of this book, it seems useful to explain some of the more technical vocabulary. It is precisely the merit of both narratology and discourse linguistics to offer clearly defined concepts with which to analyse texts. These concepts make it possible to critically observe, compare and discuss the textual strategies underlying war (and other) narratives. 3.1 Narratology and Classics Narratology is very much an offspring of formalism and structuralism and has become one of the most influential literary-critical theories introduced into classics in the twentieth century. Central to the development of narratology

introduction

5

have been Stanzel 1979, Genette [1972] 1980; Bal [1988] 1997; Fludernik 1996; and Herman 2009. The narratological apparatus enables a systematic investigation of narrative texts and includes topics such as the types and functions of narrators and narratees, the temporal aspects of a story, setting, characterization, perspective (focalization), and the representation of speech. Narratology has been introduced in classics by Winkler 1985 in a study of the narrator in Apuleius and by De Jong 1987 in a study of narrators and focalizers in Homer. Since then, a great number of narratological studies have appeared on virtually all narrative genres in antiquity,8 a multi-volume narratological history of ancient Greek literature is being written (De Jong, Nünlist & Bowie 2004; De Jong & Nünlist 2007; De Jong 2012; Van Emde Boas & De Temmerman 2018) and comprehensive discussions of narratology and classics are given in Grethlein & Rengakos 2009 and De Jong 2014. 3.2 Discourse Linguistics and Classics In the field of linguistics, classical scholarship has been strongly influenced by the functional theories of language which were developed from 1970 onwards, in particular the theory of Functional Grammar (Dik 1978; Dik 1997) and its successor Functional Discourse Grammar (Hengeveld & Mackenzie 2008). These functional frameworks have as their main theoretical principle that language is an instrument for communicative interaction within a specific communicative context. As a consequence of this principle, classicists started to study all kinds of linguistic phenomena that had been more or less neglected in the sentencebased form of analyzing syntax current until then. Landmarks in this development are Rijksbaron 1984 on Ancient Greek, Pinkster 1984/1990 on Latin, and Kroon 1995 on Latin discourse particles. This last study, and a number of articles by the late Machtelt Bolkestein,9 introduced into classical scholarship the use of discourse linguistics, a field of research concerned with the organization of language ‘beyond the level of the sentence’. It is precisely this broad focus on text and context which brings discourse linguistics close to literary analysis in a shared effort to enhance the interpretation of classical texts.10

8 9 10

Overviews can be found in Schmitz [2002] 2007: 55–75; and De Jong 2014: 6–15. See, for instance, Bolkestein 1987, 2000, 2002. Illustrations of the importance of linguistic analysis for literary interpretation can for instance be found in volumes edited by Bakker 1997; Allan & Buijs 2007; and Bakker & Wakker 2009. See also Rijksbaron 1991; Risselada 1993; Buijs 2004; Adema 2008; Van Gils 2009; Rose 2013; Van der Keur 2015; Adema 2017; Van Emde Boas 2017; and a number of articles by Kroon (2002; 2004; 2007; 2012).

6

van gils, de jong and kroon

3.3 Joined Forces Over the course of the last three decades, and as a result of the development within Greek and Latin linguistics to study language from the perspective of discourse rather than individual sentences, Dutch classical linguists and literary scholars have started to explore their common ground, as well as the strength of the complementarity of their approaches. Classical texts are approached with a toolbox that contains both linguistic and narratological instruments, in the expectation that in this way fruitful contributions can be made to existing philological disputes or new aspects of texts can be detected. The present volume can be seen, among other things, as a test to explore how far the cooperation of narratological and linguistic classicists has progressed, and whether we are indeed able to detect new aspects in and provide new perspectives to, in this case, famous Greek and Roman war narratives.11 3.4 Historiography and Narratology The two central texts in this volume belong to the historiographical genre. A question discussed by both literary theorists and historians is whether historiography may be analyzed with the same critical instruments as those employed for the study of fictional narrative.12 After all, both the purported content (nonfiction) and the primary goal (didactic) of historical texts differ from fictional narrative. However, ancient historiography was traditionally seen, already in antiquity, as very much a literary genre and the opposition therefore is less strict than in modern historiography. This having been said, it needs to be acknowledged that the ‘story world’ of a historical narrative somehow relates to an actual world which may be known from other sources. The existence of such sources, whether of a written or material form, has led (some) narratologists to add another layer, called material, to the three well-known narratological layers fabula-story-text (see Table 1.1). The abstract layer of fabula represents the chronological series of events such as narratees can reconstruct when reading a narrative. In historiography, it is relevant to relate this series of events to the layer of the material: the in principle endless number of events which together constitute history and which

11

12

The development of combining literary and linguistic models is not exclusively found in the field of Classics, but part of a broader movement. See, for instance, Banfield 1982; Prince 1982; Ehrlich 1990; Fleischman 1990; Fludernik 1993, 1996, 2000; Tolliver 1990; Emmott 1997; Semino & Culpeper 2002; more recently e.g. Dancygier 2012; Harrison et al. 2014; Burke 2014; Toolan 2016. See, for instance, Barthes [1967] 1981; White 1973; Genette 1991; Cohn 1999; Walsh 2003; De Jong 2014: 167–195.

introduction table 1.1

7

Narratological layers

Material historical events such as are recoverable from sources Fabula chronological series of events Story presentation of events in certain order and from a specific point of view by a focalizer Text wording by a narrator

are (partly) recoverable from sources such as texts, oral history, and physical objects. The material may be referred to by the narrator, when he refers to his sources or includes autoptical testimonies. In the present volume, these concepts are highly relevant to the discussions of Van Wees, Oakley and De Bakker who all compare the Thermopylae and Cannae narratives with other versions of the events which Herodotus and Livy may have used as their sources. In the case of intertextual relations, too, the domain of literary sources may be seen as part of the material and two chapters operate mainly at this particular layer, viz. those by Rademaker and De Bakker & Van der Keur. At the narratological layer of the story, three more points are at issue: order, rhythm and focalization. Where order is concerned, historiography is assumed to display a functional use of analepsis (‘flashback’) and to employ mainly a summarizing rhythm as opposed to fictional narratives in which analepses are used for ‘aesthetic concerns or formal experimentation’ and the default rhythm is scenic.13 This difference is of course not absolute, in particular not where the highly literary form of ancient historiography is concerned. In this volume, Tsakmakis, De Jong, and Kroon & Van Gils analyze time and the use of tenses in war narratives. With regard to focalization, one could question the relevance of the concept of focalization by a character for historiography, since it is hardly imaginable that a historian could know what went on in a historical character’s mind. In ancient historiography, however, we find many examples of embedded focalization by a character and here, as in many other respects, the foundational role of the Homeric epics as the model par excellence for historians seems a decisive factor.14 Aspects of focalization used as part of particular textual strategies are dealt with in this volume by Allan, De Jong, Van Gils, Buijs, Adema, Harrison, and Van der Keur.

13 14

Cohn 1999: 116. Strasburger 1972.

8

van gils, de jong and kroon

At the narratological layer of the text, even in the case of historiography the narrator should not be equated with the author, because Herodotus and Livy, too, construct a narratorial persona. The narrator, his purposes and the narratee are treated in this volume in the chapters by Tsakmakis, Allan, Rademaker, Pausch, Harrison, and Van der Keur.

4

Historiography and Discourse Linguistics

It is especially at this narratological layer of the text (see above, table 1.1) that the narratological and discourse-linguistic approaches complement each other. In narratology, whenever the narratorial persona refers to himself or to the hic et nunc of himself and his narratee, these references are analyzed as signs of an overt (as contrasted to a covert) narrator.15 Discourse linguists, by contrast, use the terms narrator and narrative in a more restricted way, in accordance with their linguistically based view that texts which may be called narrative on a comprehensive level, actually consist of a mix of various kinds of narrative and non-narrative textual components.16 Thus, in cases where the narratologist would speak of an overt narrator, the discourse linguist would usually identify a non-narrative component in the text, characterized by a nonnarrative text type (or discourse mode, see below).17 Discourse linguists define the strictly narrative parts of the text in terms of the temporal succession of (usually past tense) events and situations, which together constitute a particular story-world. This story-world can be seen as an alternative mental world, parallel to (but essentially distinct from) the communicative situation of the speaker and his audience. This latter ‘world’ comes to the fore most explicitly in the non-narrative parts of the text with which

15 16 17

For the terms overt and covert narrator see De Jong 2014: 27, 30. These components may be as small as single words and as large as complete texts. The discourse linguist views the narrator as one out of many possible roles the speaker or writer of a discourse may play. Whenever a speaker is found referring to the communicative situation he shares with the addressee (an overt narrator in narratological terms), the discourse linguist no longer analyzes that reference as narrative, and the speaker is not referred to as narrator anymore. The choice between types of narrator (overt versus covert) or types of discourse mode (narrative or non-narrative) may seem like a mere terminological issue, but it is not easy to choose for one or the other set of terms, as the different terminologies reflect differences in research interests; narratologists have developed a refined model of types of narrators for a large variety of typically narrative genres, whereas discourse linguists attempt to apply their methodological and terminological apparatus also to texts that are (predominantly) non-narrative.

9

introduction

the narrative parts tend to alternate. In these non-narrative parts we find, for instance, comments, evaluations, feelings, general information, or even promises or directives which are grounded in the communicative situation of the speaker and his addressee. Linguistically, the non-narrative parts of the text behave differently from the text parts that are narrative in a strict sense. By paying attention to, for instance, tense, mood, pronouns, negation and particles, discourse linguists are able to uncover the more indirect and strategic ways in which historiographers and other ‘story-tellers’ let story-world and communicative situation intermingle and alternate in their texts, sometimes even within the scope of a single sentence. Several chapters (Allan, Adema, Van Gils & Kroon) make use of two discourse-linguistic theories which, in combination, have proven to be especially fruitful for uncovering and understanding the subtle structuring and layering of historiographical and other narrative texts: a theory involving the prototypical structure of natural narrative18 and a theory involving discourse modes.19 The theory involving the prototypical structure of natural narrative is summarized in Table 1.2. table 1.2

Prototypical structure of a natural story or episode

Abstract Orientation Complication Peak Resolution Coda Evaluation (any position)

18

19

summary of the content or point of the narrative previous history; introduction of time, place, circumstances and main participants a conflict arises, build-up of tension Climax: the conflict is maximally tangible and near a solution the conflict is resolved, substituted by another conflict or remains permanently unresolved summarizing bridge to time of narrating evaluation of the narrative or of elements of the narrative, often conveying, or pertaining to, the point of the narrative

This theory was first designed by Labov 1972, in his sociolinguistic research of natural, spontaneous narratives; see for the application to classical texts e.g. Adam 1998; Allan 2009; and Kroon 2015. For the linguistic concept of discourse mode, see especially Smith 2003; for its application to classical texts see e.g. Kroon 2007; Adema 2007, 2008; Allan 2007, 2009; Van Gils 2009.

10

van gils, de jong and kroon

Certain components typically pertain to the narratological layer of the story (see table 1.1 above), together building up the story-world: orientation, complication, peak and resolution. Together they form a sequence starting with a stable situation, which is unsettled by some form of incident or conflict, and consequently develops towards a climax (peak), followed by a new situation. In narratological terms, this is the plot structure at the level of the ‘story’. In the abstract and coda the transition is made from the communicative situation to the story-world and vice versa. Clearly, not all narratives comply with this prototype, and this holds a fortiori for a genre like historiography. But precisely because in cross-cultural and cross-linguistic research the prototypical structure of natural narrative has proven to be a rather stable factor, it is relevant to point out variations or exploitations of this schema in historiographical texts, and point to their possible rhetorical effects. The second linguistic theory which is especially prominent in current linguistic research of classical literature concerns text typological features of coherent discourse units, so-called discourse modes. This theory is based on the observation that texts are usually not monolithic, but commonly consist of alternations of different local text types. Ancient historiographers, for instance, do not only narrate, but may also describe, inform, or comment. An especially important distinction as to the genre of historiography is the distinction between a ‘narrating’ mode of presentation and a ‘discursive’ mode, initially drawn attention to by Weinrich 1964/1985 and Benveniste 1966, in the context of their studies on tense usage in narrative.20 Later scholars have used a variety of other terms to refer to the same or comparable distinctions (see Table 1.3).21 In a ‘narrating’ discourse mode, the narrator creates a—usually past—storyworld, which consists of successive events which are connected by a close temporal and/or causal relationship. This is the default mode in prototypical narratives: without it, there is no story. In a ‘discursive’ mode, the narrator may also refer to past events, but here the events are not necessarily sequentially related to one another within a particular, self-contained story world. Rather, the past events referred to are presented as currently relevant facts, which are each individually related to the actual communicative situation of the speaker/writer 20

21

The argument of Benveniste and Weinrich sets out from the linguistic observation that in order to give an adequate account of the use of tenses in e.g. French and German narrative texts, it is necessary to distinguish between two different systems of tense: ‘discursive’ tenses and ‘narrating’ tenses. More recent linguistic scholarship has proposed a variety of additional discourse modes, and has also based its observations on non-narrative and non-literary texts. Smith 2003, for instance, also distinguishes an informative mode and a descriptive mode, in addition to a narrative and a report mode.

11

introduction table 1.3

Two main discourse modes in narrative texts

Weinrich (1964) Benveniste (1966) Chafe (1994) Smith (2003) Fludernik (2012)

‘narrating’ (pertaining to story-world)

‘discursive’ (pertaining to communicative situation)

erzählte Welt histoire immediate mode narrative mode story-telling mode

besprochene Welt discours displaced mode report mode interlocutionary mode

and his audience, and may also be coordinated with present and future events. This type of telling is, for instance, characteristic of news reports: the journalist does not tell us a story, but reports a number of relevant events and situations which have obtained in the distant or recent past, may hold in the present, or will obtain in the future. The ‘discursive’ discourse mode is also prominently present in the historiographical genre, sometimes in a quite explicit form, but more often in a form that is at first sight hardly distinguishable from the ‘narrating’ discourse mode. Careful linguistic and narratological analysis may help to detect these and other subtleties of the composition of the text, which—as a number of chapters in this volume will show—may also be highly interesting for the interpretation of the text in terms of its underlying values, morality and ideology. For instance, the combination of historical present tense, short and syntactically noncomplex sentences, visual detail, and embedded focalization is typical for a narrative peak, conveyed by the ‘narrating’ mode. Likewise, the co-occurrence of, for instance, interactional particles, perfect tense, and evaluative expressions is characteristic of the discursive mode and is typically found in evaluations. In the chapters by Allan, Adema, and Kroon & Van Gils it is illustrated how an analysis along these mixed linguistic and narratological lines may clarify the rhetorical organization and presentation of the text, and as such may enhance its interpretation. This does not mean that we plead in this volume for the development of one overarching narratological-linguistic model of analysis. Such a model would probably be reductive rather than productive. We do think, however, that much is to be gained by the collaboration of narratologists and linguists.

12 5

van gils, de jong and kroon

Outline

The chapters in this volume are arranged into three subsets. The first subset focuses on the battle of Thermopylae, as presented (mainly) by Herodotus. Van Wees presents an overview of the main historical issues of this battle by comparing Herodotus’ account with the quite different tradition found in other writers. This historical approach is complemented by De Bakker with a narratological analysis of the accounts in Herodotus and Diodorus Siculus. The next chapters continue exploring Herodotus’ narrative techniques. Tsakmakis discusses the narrative devices used to create a glorifying image of the Spartans, while De Jong focuses on Herodotus’ handling of (narratological) time in the Thermopylae narrative. As a last chapter in this subset, Allan provides a linguistic analysis of the ‘distance’ or ‘immersion’ allowed to the narratee. In the second subset, the battle of Cannae as presented by Livy is central. The first chapter by Oakley illuminates the main characteristics of Livy’s account in a comparison with the version found in Polybius. Next, Van Gils & Kroon analyze, from a discourse-linguistic perspective, the narrative structure of the Cannae episode within the broader context of Livy’s book 22. This is followed by a chapter in which Pausch explores the literary motif of Punica fraus in Livy’s books 21 and 22, paying particular attention to its culmination in Cannae. Van Gils offers a narratological and linguistic analysis of space in the Cannae episode, with attention to the strategic uses of space and its role in characterization. Buijs discusses Livy’s use of represented and reported speech for characterization and the thematic opposition of ratio and fortuna. Speech is also the subject of the contribution by Adema, whose analyses show how the various types of speech contribute to Livy’s narrative aims in his account of the battles of Cannae and Zama. The third and last subset of chapters deals with ‘comparisons’ although of two different natures: on the one hand we find two chapters comparing two battle narratives resulting in observations about phenomena related to intertextuality, genre and authorial choices. On the other hand, the stylistic device of a comparison, better known as simile, is discussed as a genre specific feature of war narrative in epic poetry. First, De Bakker & Van der Keur bring together the two battles under discussion in this volume by showing how Livy’s narrative of the Roman defeat at Cannae can be seen in the light of the Greek tradition on Thermopylae. Second, Rademaker compares Herodotus’ account of Thermopylae with Thucydides’ narrative of the Spartan defeat at Sphacteria exploring the ways in which Thucydides’ narrator persuades the narratee to accept his belittling of the Athenian victory over the Spartans. Next, Harrison discusses the function of similes in Vergil’s Aeneid 10 and demon-

introduction

13

strates how the similes make an important contribution to its narrative texture. And finally, Van der Keur addresses the role of similes in establishing secondary plot lines, which guide the narratee’s interpretation of the main narrative.

Bibliography Adam, G., Tense and Aspect in Roman Historiographic Narrative: A Functional Approach to the Prose of the Memoria Rerum Gestarum, dissertation (Chicago 1998). Adema, S.M., ‘Discourse Modes and Bases in Vergil’s Aeneid’, in R.J. Allan & M. Buijs (eds.), The Language of Literature: Linguistic Approaches to Classical Texts (Leiden 2007) 42–64. Adema, S.M., Discourse Modes and Bases. A Study of the Use of Tenses in Vergil’s Aeneid, dissertation (Amsterdam 2008). Adema, S.M., Words of Warriors. Speech and Thought in Latin War Narratives (Leiden 2017). Allan, R.J., ‘Sense an sentence complexity. Sentence structure, sentence connection, and tense-aspect as indicators of narrative mode in Thucydides’ Histories’, in Allan and Buijs (eds.), The language of Literature. Linguistic approaches to classical texts (Leiden 2007), 93–121. Allan, R.J., ‘Towards a Typology of the Narrative Modes in Ancient Greek: Text types and narrative structure in Euripidean messenger speeches’, in S.J. Bakker & G.C. Wakker (eds.), Discourse Cohesion in Ancient Greek (Leiden 2009), 171–204. Allan, R.J. & Buijs, M. (eds.), The language of Literature. Linguistic Approaches to Classical Texts (Leiden 2007). Bakker, E.J. (ed.), Grammar as Interpretation. Greek Literature in its Linguistic Contexts (Leiden 1997). Bakker, S.J. & Wakker, G.C. (eds.), Discourse Cohesion in Greek (Leiden 2009). Bakogianni, A. & Hope, V.M. (eds.), War as Spectacle: Ancient and Modern Perspectives on the Display of Armed Conflict (London 2015). Bal, M., Narratology. Introduction to the Theory of Narrative ([1988] Toronto 1997). Banfield, A., Unspeakable Sentences: Narration and Representation in the Language of Fiction (Boston 1982). Barthes, R., ‘The Discourse of History’, Comparative Criticism 3 ([1967] 1981) 7–20. Benveniste, E., Problèmes de Linguistique Générale (Paris 1966). Bolkestein, A.M., ‘Discourse functions of predications: The background/foreground distinction and tense and voice in Latin main and subordinate clauses’, in J. Nuyts & G. de Schutter (eds.), Getting One’s Words into Line. On Word Order and Functional Grammar (Dordrecht 1987) 163–178.

14

van gils, de jong and kroon

Bolkestein, A.M., ‘Discourse Organization and Anaphora in Latin’, in S.C. Herring, P. van Reenen & L. Schøsler (eds.), Textual Parameters in Older Languages (Amsterdam 2000) 107–137. Bolkestein, A.M., ‘Linguistic reflection of discourse structure in Latin’, in A.M. Bolkestein, C.H.M. Kroon, H. Pinkster, H.W. Remmelink & R. Risselada (eds.), Theory and Description in Latin Linguistics. Selected Papers from the XIth International Colloquium on Latin Linguistics (Amsterdam 2002) 17–28. Buijs, M., Clause Combining in Ancient Greek Narrative Discourse. The Distribution of Subclauses and Participial Clauses in Xenophon’s Hellenica and Anabasis (Leiden 2004). Burke, M. (ed.), The Routledge Handbook of Stylistics (London 2014). Calloway, C., War in Literature and Drama (Oxford 2015). Cartledge, P., Thermopylae: the Battle that Changed the World (London 2006). Chafe, W., Discourse, Consciousness, and Time: The Flow and Displacement of Conscious Experience in Speaking and Writing (Chicago 1994). Cohn, D., ‘‘I Doze and I Wake’: The Paradox of Simultaneous Narration.’ Paper read at the Nice conference of the Society for the Study of Narrative Literature (Nice 1991). Cohn, D., The distinction of Fiction (Baltimore 1999). Daly, G., Cannae: the Experience of Battle in the Second Punic War (London 2002). Dancygier, B., The Language of Stories. A Cognitive Approach (Cambridge 2012). Dik, S., Functional Grammar (Amsterdam 1978). Dik, S., The Theory of Functional Grammar (Berlin 1997). Ehrlich, S., Point of View: A Linguistic Analysis of Literary Style (London 1990). Emde Boas, E. van & Temmerman, K. de (eds.), Characterization in Ancient Greek Literature (Leiden 2018). Emde Boas, E. van, Language and Character in Euripides’ Electra (Oxford 2017). Emmott, C., Narrative Comprehension: A Discourse Perspective (Oxford 1997). Fleischmann, S., Tense and Narrativity (Austin 1990). Fludernik, M., The Fictions of Language and the Languages of Fiction: The Linguistic Representation of Speech and Consciousness (London 1993). Fludernik, M., Towards a ‘Natural’ Narratology (London 1996). Fludernik, M., ‘Genres, Text Types, or Discourse Modes—Narrative Modalities and Generic Categorization’, Style 34.2 (2000) 274–292. Fludernik, M., ‘Narratology and Literary Linguistics’, in R. Binnick (ed.) The Oxford Handbook of Tense and Aspect, Oxford Handbooks online, DOI: 10.1093/oxfordhb/ 9780195381979.013.0002 (Oxford 2012) Fornara, C.W., Archaic Times to the End of the Peloponnesian War (Cambridge 1983). Foster, E. & Lateiner, D. (eds.), Thucydides and Herodotus (Oxford 2012). Genette, G., Fiction et diction (Paris 1991). Genette, G., Narrative Discourse: An Essay in Method ([1972] Ithaca N.Y. 1980).

introduction

15

Gils, L.W. van, Argument and Narrative. A Discourse Analysis of Ten Ciceronian Speeches, PhD thesis (Amsterdam 2009). Grethlein, J. & Rengakos, A. (eds.), Narratology and Interpretation. The Content of Narrative Form in Ancient Literature (Berlin 2009). Harrison, C., Nuttall, L., Stockwell, P. & Yuan, W., Cognitive Grammar in Literature (Amsterdam 2014). Hengeveld, K. & Mackenzie, J.L., Functional Discourse Grammar: A Typologically-based Theory of Language Structure (Oxford 2008). Herman, D., Basic Elements of Narrative (Chichester 2009). Jong, I.J.F. de, Narrators and Focalizers: the Presentation of the Story in the Iliad (London 1987). Jong, I.J.F. de, Narratology and Classics. A Practical Guide (Oxford 2014). Jong, I.J.F. de, Nünlist, R., & Bowie, A. (eds.), Narrators, Narratees, and Narratives in Ancient Greek Literature (Leiden 2004). Jong, I.J.F. de & Nünlist, R. (eds.), Time in Ancient Greek Literature (Leiden 2007). Jong, I.J.F. de (ed.), Space in Ancient Greek Literature (Leiden 2012). Keur, C.M. van der, A Commentary on Silius Italicus’ Punica 13. Intertextuality and Narrative Structure, PhD thesis (Amsterdam 2015). Koopman, N., Ancient Greek Ekphrasis: Between Description and Narration (Leiden 2018). Kroon, C.H.M., Discourse Particles in Latin. A study of nam, enim, autem, vero, and at (Amsterdam 1995). Kroon, C.H.M., ‘How to write a ghost story? A linguistic view on narrative modes in Pliny Ep.7.27’, in D. Shalev & L. Szawicky (eds.), Donum Grammaticum: Studies in Latin and Celtic Philology and Linguistics in Honour of Hannah Rosén (Louvain-la-Neuve 2002) 189–200. Kroon, C.H.M., ‘The effect of the echo. A text linguistic approach to Catullus Carmen 63’, Mnemosyne 57.5 (2004) 629–650. Kroon, C.H.M., ‘Discourse modes and the use of tenses in Ovid’s Metamorphoses’, in R.J. Allan & M. Buijs (eds.), The language of Literature. Linguistic Approaches to Classical Texts (2007) 65–92. Kroon, C.H.M., ‘Anchoring as a communicative device in Roman historiography: a discourse-linguistic perspective’, Proceedings International Conference Anchoring in Antiquity (Ravenstein 2015) [internet publication: http://www.ru.nl/oikos/ anchoring‑innovation/anchoring‑scholarship/anchoring‑antiquity‑international ‑conference/] Kroon, C.H.M., ‘Voce voco. Some text linguistic observations on Ovid Heroides 10’, Mnemosyne 65.2 (2012) 238–250. Labov, W., Language in the Inner City. Studies in the Black English Vernacular (Philadelphia 1972).

16

van gils, de jong and kroon

Levene, D.S., Livy on the Hannibalic War (Oxford 2010). Matthew, C. & Trundle, M. (eds.), Beyond the Gates of Fire (Barnsley 2013). McLoughlin, K. (ed.), ‘War and words’, in K. McLoughlin (ed.), The Cambridge Companion to War Writing (Cambridge 2009) 15–24. Mineo, B. (ed.), A Companion to Livy (Malden 2015). Pinkster, H., Latijnse Syntaxis en Semantiek (Amsterdam 1984). Pinkster, H., Latin Syntax and Semantics (London 1990). Priestley, J., Herodotus and Hellenistic Culture. Literary Studies in the Reception of the Histories (Oxford 2014). Prince, G., Narratology: the Form and Function of Narrative (Berlin 1982) Rijksbaron, A., Grammatical Observations on Euripides’ Bacchae (Amsterdam 1991). Rijksbaron, A., The Syntax and Semantics of the Verb in Classical Greek (Amsterdam 1984). Risselada, R., Imperatives and Other Directive Expressions in Latin (Leiden 1993). Rose, P.J., A Commentary on Augustine’s De cura pro mortuis gerenda. Rhetoric in Practice (Leiden 2013). Schmitz, T.A., Modern Literary Theory and Ancient Texts: An Introduction ([2002] Malden 2007). Semino, E. & Culpeper, J., Cognitive Stylistics. Language and Cognition in Text (Amsterdam 2002). Sherry, V. (ed.), The Cambridge Companion to the Literature of the First World War (Cambridge 2005). Smith, C.S., Modes of Discourse. The Logical Structure of Texts (Cambridge 2003). Stanzel, F.K., Theorie des Erzählens, ([1979] Göttingen 1982). Strasburger, H., Homer und die Geschichtsschreibung (Heidelberg 1972). Tolliver, J., ‘Discourse Analysis and the Interpretation of Literary Narrative’, Style 24.2 (1990) 266–281. Toolan, M., Making sense of Narrative Text: Situation, Repetition, and Picturing in the Reading of Short Stories (New York 2016). Walsh, R., ‘Fictionality and Mimesis: Between Narrativity and Fictional Worlds’, Narrative 11.1 (2003) 110–121. Weinrich, H., Tempus (Stuttgart 1964). White, H., Metahistory. The Historical Imagination in 19th-Century Europe (Baltimore 1973). White, H., The Content of the Form: Narrative Discourse and Historical Representation (Baltimore 1987). Winkler, J.J., Auctor and Actor. A Narratological Reading of Apuleius’ Golden Ass (Berkeley 1985). Woodman, A.J., Rhetoric in Classical Historiography (London 1988).

part 1 Thermopylae

∵

chapter 2

Thermopylae: Herodotus versus the Legend Hans van Wees

Not everyone has always been impressed with the Spartans’ performance in the Battle of Thermopylae. ‘The Mede came from the ends of the earth to the Peloponnese before you offered any resistance worth the name’, delegates from Corinth supposedly told Sparta at a conference in 432 BC (Thuc. 1.69.5), implying, of course, that what happened at Thermopylae did not amount to ‘resistance worth the name’.1 The Corinthians here treated the annihilation of Spartan troops in 480 BC not in the usual manner as a glorious moral victory but as evidence of poor leadership by Sparta. In stark contrast, the huge war effort at sea by the Athenians in the same year led Herodotus to proclaim them ‘the saviours of Greece’ (7.139), a boast they themselves were only too happy to repeat as evidence of outstanding leadership which entitled them to rule over other Greeks (for instance Thuc. 1.73–1.74). In the late 430s, when Herodotus was writing his Histories, stories about the Persian War were not simply entertainment or merely of academic interest, but carried a powerful political charge. By telling the story of a battle or campaign in a certain way, one was inevitably taking a political position: this is why Herodotus expected his opinion that Athens, not Sparta, had done most to defeat the Persians to be ‘resented by most people’ (7.139.1).2 Two quite different stories about Thermopylae were told in antiquity, and the political dimension is vital to understanding both. One version is the full and seemingly sober account by Herodotus. The other survives only in relatively short summaries and is in many respects so unlikely that it has been dubbed ‘the Legend’.3 It portrays Leonidas and his men as a suicide squad who end their lives, not defending the pass, but raiding the Persian camp in an attempt to assassinate Xerxes. Diodorus of Sicily gives a fairly detailed outline of this Legend, while briefer versions containing the same essential elements appear in Plutarch’s The Malice of Herodotus and Justin’s Epitome of Pom1 As pointed out by Grundy 1901: 273. 2 For the politics of Persian War historiography, see Osborne 2009: 231–234. Cf. Hammond 1996; he, however, greatly oversimplifies in arguing that Herodotus’ version of Thermopylae represented a pro-Athenian, anti-Spartan view. 3 Hignett 1963: 125, 371.

© koninklijke brill nv, leiden, 2019 | doi:10.1163/9789004383340_003

20

van wees

peius Trogus’ Philippic Histories. Diodorus’ main source for archaic and classical Greek history was Ephorus, and the other authors may have drawn directly or indirectly on the same source.4 A common assumption among scholars has been that Ephorus himself invented large parts, if not all, of the Legend,5 and that Herodotus’ account represents an older version of events, based at least in part on first-hand testimony, oral or written, and thus much more reliable. Twenty years ago, however, Michael Flower argued that the main elements of Ephorus’ version, including the Spartan suicide mission and night-raid on the Persian camp, derived from a much earlier source: a long narrative elegy by Simonides, composed soon after the battle. Flower’s main concern in his paper was to reassess Ephorus’ merits as a historian, and he therefore went on to argue that Simonides’ version of events might well have been largely accurate, and that Ephorus should not be condemned for preferring it to Herodotus’ account.6 In what follows, I shall build on Flower’s essential insight that the tradition found in Ephorus may have been formulated well before Herodotus’ day, but take it in a different direction. This chapter argues that, despite deriving from an early source—whether Simonides’ elegy or other poetry or oral tradition—Ephorus’ story was indeed a ‘Legend’, and was originally put about by Sparta to salvage political credit from a campaign that could easily have been regarded as a disaster due to poor leadership.7 Secondly, and most importantly, Herodotus’ account was not independent of this older Legend, but derived from it. All the notorious problems of his narrative can be explained as the result of Herodotus’ attempt to reconcile the Legend with other evidence at his disposal and to tone down the most blatant elements of Spartan selfjustification. The extent of historically reliable material in Herodotus thus turns out to be even smaller than is usually assumed, and we are left with only the barest outline of ‘what really happened’ at Thermopylae.

4 Flower 1998: 365–366, 370–371; Hammond 1996: 1–4; and see in this volume De Bakker. 5 Specifically the concluding night-raid on the Persian camp. Flower 1998: 366 n. 9 and Trundle 2013: 30 n. 27, quote modern authors who have declared this a ‘fantasy’ or ‘fiction’ by Ephorus; add e.g. Cartledge 2006: 146. Others just ignore it: e.g. Balcer 1995; Cawkwell 2005. However, Green 1996: 139, suggests the story may have a factual core (cf. Holland 2005: 282; Strauss 2004: 36: ‘so improbable it might even be true’). 6 Flower 1998: 372–379; his ideas are adopted by Trundle 2013: 32–34, and De Bakker, this volume. 7 In other words, the story arose as ‘post-battle spin’, as one of the referees for the Press suggested it should be called instead of ‘the Legend’.

thermopylae: herodotus versus the legend

1

21

Leonidas: Kamikaze or Advance Guard?

An immediately puzzling feature of the Thermopylae campaign is that so few troops were sent out to defend the pass. Modern scholars have sometimes argued that these small numbers would in fact have been adequate, but neither Herodotus nor the sources representing the Legend take this view. According to Herodotus, the numbers were small because the forces were merely an advance guard, while according to the Legend the number of Spartans was small because they were a suicide squad. King Leonidas, according to Diodorus’ version of the Legend, ‘announced that only 1,000 were to follow him on the expedition’. When the Ephors told him that this was too small a number, he revealed to them in secret that his mission was, not to stop the Persian advance, but ‘to die for the freedom of all’. ‘And so, if 1,000 take the field, Sparta will be more famous when they have fallen, but if the Lacedaemonians march out with a general levy, Lacedaemon will be utterly destroyed’ (Diod. Sic. 11.4.2–11.4.4). Plutarch agreed that Leonidas’ men were on a suicide mission, but openly so rather than in secret: ‘before they set out, they competed in their own funeral games and their own fathers and mothers watched’. Leonidas told his wife and the other widows-to-be that they should ‘marry good men and give birth to good men’ (32 = Mor. 866b; cf. 225a).8 Neither of these two authors explained why Leonidas and his men were so certain that they would die, but Justin’s Epitome made this clear. The Spartans had had advance warning of the Persian invasion, through a covert message from their deposed king Demaratus, living in exile at the Persian court (2.10.13– 2.10.17). They had therefore consulted the Delphic oracle, which foretold that ‘either a king of the Spartans or the city must fall’ (2.11.8). Leonidas duly decided to sacrifice himself and told his men ‘when setting out for war’ that they should ‘make up their minds to die’ (2.11.9). Perhaps Diodorus’ and Plutarch’s omission of the oracle implies that Ephorus did not cite it, but even so his and their story is premised on the oracle’s warning, which provides the only reason why Leonidas’ expedition was supposedly a suicide mission from the start. Without the oracle, Leonidas’ conviction, before he even leaves home, that death is inevitable would imply a firm belief that the Greeks could not possibly defeat the Persian army, and that the best the Spartans could hope for was a noble

8 Matthew 2013b: 73, suggests that the passages in Diodorus and Plutarch where Leonidas allusively says that 1,000 men are enough ‘for the task at hand’ refer to the holding the pass—but the context, cited above, makes it perfectly clear that the ‘task at hand’ is to die.

22

van wees

death. This is hardly the impression that the Spartans sought to make, or indeed the impression that the narrative of the Persian War as a whole sought to create. With the oracle, by contrast, the story conveys a very different, much more positive and powerful message: Leonidas and his men made a conscious decision to sacrifice their lives for the sake of an ultimate Greek victory. The Legend behind all three accounts must therefore have included the oracle.9 Herodotus told a very different story. When the Greek force arrived at Thermopylae, they sent messengers to the Locrians and Phocians to announce ‘that they themselves had come as an advance guard (prodromoi) ahead of the others, and that the rest of the allies were expected any day’ (7.203.1). They insisted that resistance was not futile: Xerxes was mortal and vulnerable to misfortune, like any man (7.203.2). This had the desired effect of bringing the Locrians, who had already surrendered to Xerxes (7.132; Diod. Sic. 11.4.6), and the Phocians over to the Greek side. Herodotus goes on to elaborate the point in his own voice: The Spartans sent the men with Leonidas first so that the other allies would join the campaign when they saw them, and not go over to the Persians if they heard of delay. Later—because the Karneia impeded them—they intended to come to their aid with a general levy (pandēmei) as fast as possible after celebrating the festival and leaving a guard in Sparta. The rest of the allies planned to do the same thing themselves, for at the same time the Olympic Games coincided with these events, so they sent advance guards, not expecting the war at Thermopylae to be decided so soon. Hdt. 7.206

When the Persians arrived, the Peloponnesians wanted to withdraw to the Isthmus, but in order to placate the Phocians and Locrians, Leonidas ‘voted to stay there and to send messengers to the cities with an order to come to their aid, since they were too few to hold off the Persian army’ (7.207). We are later reminded that the battle of Thermopylae coincided with the Olympic Games (8.26) and informed that immediately afterwards a Peloponnesian general levy of ‘many tens of thousands’ did muster at the Isthmus: ‘the Olympia and Karneia had already passed’ (8.71–8.72; cf. 8.40; 9.10).

9 Clarke 2002: 69–72, notes parallels with legends of kings who sacrifice themselves for the safety of their countries, such as Erechtheus and Kodros in Athens.

thermopylae: herodotus versus the legend

23

Herodotus could hardly have stated more emphatically that sending only a small number of troops to Thermopylae was not intended to minimise losses from a suicide mission but was as much as the Spartans and Peloponnesians could do while they were hampered by religious taboos. In line with this view, he suggested that events unfolded quickly, while these taboos were in force. The leading men of the Greeks met at the Isthmus of Corinth (7.173) and decided to make a stand at Thermopylae and Artemisium (7.175). ‘When they heard that the Persian was in Pieria’, in the mountains immediately north of Thessaly, ‘they broke up from the Isthmus and marched out’ (7.177), taking their positions ‘as quickly as possible’ (7.178.1). This chronologically vague narrative suggests that the Greeks were caught out by the speed of the Persians, who reached the northern border before Greek plans had been implemented, so that troops were raised in a hurry at an inconvenient time. Accordingly, Herodotus did not believe that the Spartans stayed at Thermopylae because they always intended to die, but said twice that Leonidas regarded it as ‘not seemly’ (ouk euprepeōs) for himself ‘and the Spartiates present’ (7.220.1) and ‘not noble’ (ou kalōs, 7.220.2) to leave their post once they had taken it, and twice that Leonidas sought ‘great fame’ for himself (mega kleos, 7.220.2) and ‘fame for the Spartans alone’ (7.220.4). Given the military values associated with Sparta, these were perfectly adequate reasons for the advance guard not to retreat when reinforcements failed to arrive, without implying any suicidal motive. So far, then, Herodotus’ version of events seems plausible and completely independent of the Legend. Yet at this point it emerges that Herodotus was aware of the oracle mentioned by Justin which formed an integral part of the Legend. Leonidas had a third reason for staying to fight: ‘so that Sparta’s prosperity would not be wiped out’ (7.220.2), as foretold by an oracle that Herodotus quoted in full (7.220.3–7.220.4). The Legend is thus acknowledged but only belatedly, as one of several factors to which Leonidas was ‘giving consideration’ (epilegomenon), and without acknowledgement of the implication that Leonidas and his men always intended to die. Instead, the story is subordinated to Herodotus’ own view that the troops at Thermopylae were an advance guard. The secret warning from Demaratus that caused the Spartans to consult the oracle in the first place also features in Herodotus’ account, but is even more radically postponed. Not until he has brought the story of the battle to a close does Herodotus suddenly announce: ‘I am going back to a point in the story that I omitted earlier’ (7.239.1). Details follow of how Demaratus smuggled out his message and how Leonidas’ wife Gorgo discovered it (7.239.2–7.239.4). It is unlikely that Herodotus had forgotten the story and then inserted it as soon as he remembered, without ever revising his text. Rather, he included it because it

24

van wees

was a story too good and well-known to omit altogether, but he deliberately told it out of sequence because he did not want to give as much weight to the oracle, and its implication that the Spartans always intended to die at Thermopylae, as it would have carried if he had started his narrative with this story.10 Herodotus knew at least this part of the Legend but chose to tell a different story.11 Here, then, we have two competing answers to the question why the Spartan contingent at Thermopylae was so small. This was a matter of acute political importance. The 1,000 Lacedaemonian hoplites present according to the Legend amounted to only 10% of the number who fought at Plataea a year later (Hdt. 9.10–9.11, 9.28), and compared very unfavourably with the 100 % mobilisation at the same time by the Athenians, who sent all 200 of their warships, manned by most of the adult male population, to face the Persians at Artemisium and Salamis.12 Such limited mobilisation by Sparta might suggest to a critic that the Spartans had not been as committed to the defence of Greece as one would expect of the recognized leader of the coalition against Persia. Sparta needed to justify its actions at Thermopylae in order to vindicate its leadership. The Legend solved this problem by claiming that the Spartans always knew that they would have to fight to the death, so that their commitment of 1,000 men was not a sign of inadequate leadership but on the contrary an extraordinarily great, generous, act of self-sacrifice—a tithe of Spartan manpower, as it were. Diodorus spelled out the point: ‘it is incumbent on the leaders of Hellas to die willingly in competition for first rank’ (11.9.1). The story of the oracle gave this sacrifice still greater meaning: rather than a merely symbolic gesture, noble but futile, it had been the only act that could ensure the ultimate victory of the Greeks. And thanks to Demaratus’ message, Leonidas had decided to make his sacrifice before the other Greeks even knew that the Persians were going to attack. In short, Spartan leadership had been exemplary.13 Impressive as it was, the Legend had serious weaknesses as an explanation of the small numbers defending Thermopylae. Even if one were to accept that 10

11

12 13

For a different take on the role of the oracle within the story, see in this volume De Jong. Others have found the delay of the Demaratus-story so odd that they have dismissed 7.239 as a later addition to Herodotus’ original text: see e.g. Macan 1908 ad loc. The detail that the Spartiates selected for the mission were all men ‘who had living sons’ (7.205.2) is conceivably another reflection of the suicide-mission story, insofar as it suggests an expectation that they need living heirs because they will not return. Athenian ships at Artemisium: Hdt. 8.1, 14; still all deployed at Salamis: 8.44, 8.46 (in each case including 20 ships manned by Athenian settlers at Chalcis). Moreover, as De Bakker notes in this volume, Diodorus explicitly rates Thermopylae as a more ‘decisive’ battle in the Persian War than those which followed (11.11.5).

thermopylae: herodotus versus the legend

25

the Spartans really did receive an oracle demanding the death of one of their kings, the obvious way to fulfil destiny would have been for Leonidas to sacrifice his own life by charging into the enemy ranks in an act of Roman-style devotio, and thereby clear the way for a victory by the rest of the Spartan and allied army over the Persians. Why order another 1,000 men to commit suicide alongside him? Why not send additional troops to fight with the gods’ favour once the king had fallen? Most importantly, the oracle fails to explain why Sparta’s Peloponnesian allies, too, sent only small contingents, ranging from 80 to 500 hoplites, typically only between one-third and one-fifth of the number they deployed a year later at Plataea.14 These troops were clearly not suicide squads and did not stay to fight to the death (7.207, 7.219–7.220). Clearly, the Legend on this point represented an ingenious retrospective justification of events, not a genuine explanation of what happened.15 Herodotus’ claim that the troops at Thermopylae were merely an advance guard is at first sight more plausible insofar as it accounts for the small size of all units, not just Sparta’s. He attributes this explanation to the Spartans themselves, who sent messengers to the cities of Locris and Phocis—and one imagines to Athens and the cities of Boeotia, too—in order to reassure their allies that reinforcements were on the way. Such promises would have been conveyed to public meetings across Central Greece, and Herodotus may well have spoken to men who were present on these occasions—the youngest of them were probably only about 15 years older than he was. Similarly, there would have been plenty of eyewitnesses to Leonidas’ subsequent sending of messengers to ask for immediate reinforcements, both in the army and in the cities which received his request. So whatever the Spartans said after the battle, it seems likely that beforehand they excused the small numbers which turned up at the pass by claiming that many more would follow ‘any day’. Yet this story was not without its own weaknesses, since it did not in itself explain why the promised reinforcements never materialised or indeed why the Peloponnesians did not send larger contingents in the first place. Here Herodotus relied on an explanation which he did not attribute to the Spar-

14

15

Numbers at Thermopylae: Hdt. 7.202; cf. Paus. 10.20.1–10.20.2; at Plataea: Hdt. 9.28. Tegea increased its troops from 500 to 1,500, Arcadian Orchomenos from 120 to 600, Phleious from 200 to 1,000, Mycenae from 80 to 400 (jointly with Tiryns); Corinth from 400 to 5,000. Pace e.g. Cartledge 2006: 129–131, who suggests the Japanese kamikaze as a parallel. See also Matthew 2013b: 83–85, on the oracle as ‘a later piece of propaganda (possibly Spartan in origin)’, though it is not clear to me what he means by ‘later’ or by describing it repeatedly as an ‘insertion into Herodotus’ text’.

26

van wees

tans but stated in his own voice, and implied in the rest of his narrative: the Karneia and Olympia festivals ‘were in the way’ for the entire duration of the Thermopylae campaign, hampering both initial mobilisation and later reinforcement. Herodotus himself, however, elsewhere provided a great deal of detailed information that was incompatible with this notion. If that information was correct then Herodotus’ account of mobilisation for Thermopylae was almost as much of a retrospective justification of the actions of the Spartans and Peloponnesians as the Legend was, and only slightly less false.16 The problem with Herodotus’ story is essentially that the religious inhibitions which he cited were more limited and short-lived than he implied, while the campaign lasted longer than he suggested. Moreover, his subsequent account of the mobilisation of the Greek fleet conflicts with what he tells us about the simultaneous mobilisation of the Greek army. Herodotus indicated that the Olympic Games were still in progress on the day after the end of fighting at Thermopylae (8.26), and since this was at the time a four-day festival, this would mean it started no earlier than the three days of combat in the pass and could not have affected any earlier mobilisation. The Olympic truce which preceded the festival prohibited Elis from making war and others from attacking Elis and would explain why the Eleans were the only Peloponnesians who were not represented at either Thermopylae or Artemisium but did take part in the general levy afterwards (8.72). This is presumably the basis for Herodotus’ idea that the Olympia ‘were in the way’. However, the truce did not ban warfare in general, and accordingly did not prevent the Athenians from deploying their entire fleet or the Locrians from mobilising their general levy (7.203.1), so the Olympia cannot explain why the Arcadians showed up only in small numbers. Since the non-Dorian Arcadians would not have been affected by the Dorian Karneia festival either, Herodotus in effect has no valid explanation at all for their limited mobilisation.17 The Karneia likewise offer only a very partial explanation for the small numbers of Dorians at Thermopylae. Herodotus elsewhere reported that Spartan troops arrived too late to assist the Athenians in the Battle of Marathon because they had had to wait for the full moon before they could march out (6.106.3– 6.107.1). It is inconceivable that the rule applied to every month, which would have made it next to impossible for the Spartans to wage war at all, as Plutarch 16 17

Cf. Grundy 1901: 273–276. It has nevertheless been widely accepted: Green 1996: 111–112, for example, calls it ‘no more than the plain truth’. See Lämmer 1982–1983 for the Olympic truce. Note also the Arcadian ‘deserters’ offering their military service to Xerxes while the Olympia are in progress: Hdt. 8.26.

thermopylae: herodotus versus the legend

27

vehemently pointed out (Mor. 861e). The rule must have applied only in the month Karneios, when Spartans and other Dorians did genuinely try to avoid marching out on campaigns (Thuc. 5.54, 5.75–5.76), and when the Karneia festival was celebrated for 9 days, ending at the full moon in the middle of the month.18 So for a period of 15 days, from the start of the month to the end of the festival, it would have been taboo for Sparta, Corinth, Phleious and Mycenae to send out troops. The Olympic Games also ended on the full moon, so the fighting at Thermopylae would have coincided with the Karneia just as it coincided with the Games. Since Leonidas made his request for reinforcements 7 days before he and his men were massacred,19 this request must have come squarely within the sacred period of Karneios and probably during the Karneia festival itself. So it seems entirely plausible that Sparta and a few allies delayed sending reinforcements for religious reasons. By the same token, however, these religious reasons could not have applied to the initial mobilisation. Despite Herodotus’ suggestion that everything happened very fast, the chronological details he provides later imply that the first mobilisation took place more than two weeks before Leonidas sent for help. The Greeks marched out ‘when the Persian was in Pieria’ (7.177), and Xerxes with one-third of his army spent ‘many days’ in Pieria to ‘clear the Macedonian mountain so that the army could pass through it’ (7.131); the rest of the Persian army waited at Therma until the road was clear, then took 14 days to reach Thermopylae, where they spent 7 days.20 The final advance of the Persian army from Therma thus started about 21 days before the end of the Karneia, and therefore 6 days before the start of the month Karneios. The mobilisation of the Greeks started earlier still, by however ‘many days’ the Persians spent clearing the roads into Thessaly, which must surely mean no less than 10 days before the start of the sacred month. No religious taboo could have been in force, therefore, and this is confirmed by yet more detailed information later provided by Herodotus, which shows that the launch of Dorian naval forces, at the same time as land forces were sent 18 19 20

For detailed discussion of chronological issues, see Sacks 1976; Matthew 2013b: 65–69. He did so when the Persian army arrived at Thermopylae, which was 4 days before the fighting started (Hdt. 7.210), and the fighting lasted 3 days (see below). As implied by Herodotus’ account of the movements of the Persian fleet, which left Therma 11 days after Xerxes and the army did (7.183.2) and took 5 days to get from Therma to Aphetae (7.183.3, 7.188.2, 7.191.2, 7.193), by which time the army had already arrived in Malis ‘on the third day’, i.e. two days earlier (7.196). The precise interpretation of the temporal indications has been much debated, but the differences amount to no more than a few days and do not affect our present argument: see e.g. Macan 1908: 272–276; Hignett 1963: 379–385; Sacks 1976; Burn 1984: 395–396; Lazenby 1993: 118–123.

28

van wees

to Thermopylae, was unaffected by religious restrictions. Sparta and Corinth sent considerable numbers of warships to Artemisium, as did other Dorian cities which also observed the Karneia. Crucially, these contingents were of the same size or not much smaller than those sent to Salamis after the end of the festival. Sparta sent 10 triremes, with 2,000 crew members, to Artemisium and 16 to Salamis; Corinth 40 triremes on both occasions; the Dorian cities of Megara, Aegina, Sicyon, Epidaurus and Troizen between them sent 63 ships to Artemisium and 80 to Salamis (Hdt. 8.1, 8.43). The slightly larger numbers were surely due simply to the fact that Salamis was very much closer to home than Artemisium for all these cities. The alleged restrictions imposed by the Karneia thus had no discernible impact on the size of these fleets.21 Where Herodotus found all this detailed material on the composition of the Greek fleets and the chronology of Persian campaign we do not know. It is clear that he did not simply invent these details, given that they contradict the story that he wanted to tell. He must have had oral or written sources on these matters which he regarded as sufficiently authoritative to cite despite the problems they posed for his preferred narrative. If these sources were correct, then the sacred periods account for the absence of allies from Elis and for the failure of Spartans and other Dorians to send reinforcements shortly before the battle— but they do not explain why the Dorians mobilised such small units in the first place or why they sent no further troops over the next 10 days or more, or why the Arcadian cities between them sent only 2,120 hoplites to Thermopylae (7.202) although they later ‘all’ joined a general levy of ‘many tens of thousands’ Peloponnesians at the Isthmus (8.71–8.72). Herodotus’ version of events thus clearly cannot be true. I would suggest that he took three well-attested historical facts—the Spartan proclamation that the forces arriving at Thermopylae were only the forerunners of a much larger force; the Olympic Games preventing Elis from mobilising; and the Karneia preventing the Dorians from immediately answering Leonidas’ call for help— and stretched these into a general explanation for the small forces of hoplites defending the pass. Perhaps he was unaware of the chronological difficulties with this interpretation, or perhaps he did realise but chose to gloss over the problem. Either way, he evidently meant to give an explanation more plausible than the Legend but still credit the Spartans and their allies with doing their very best to defend Thermopylae, although unluckily hampered by legitimate, respectable, religious scruples.22

21 22

As pointed out by e.g. Hammond 1996: 12; cf. Matthew 2013b: 65–69. Cf. Matthew 2013b: 65–69, who suggests that the religious taboos are ‘later insertions in

thermopylae: herodotus versus the legend

29

In sum, both the Legend and Herodotus regarded the number of Peloponnesian and especially Spartan troops sent to Thermopylae as inadequate, and each in their own way tried to explain the shortfall. Herodotus’ explanation had slightly more of a basis in history than the Legend, but fundamentally both accounts put justification above historical accuracy.

2

One Thousand or Three Hundred?

Just how small was the Spartan contingent? Diodorus spoke of ‘1,000 Lacedaemonians’, as we have seen, not of the famous 300 Spartans. A little later, he clarified that ‘1,000 Lacedaemonians and with them 300 Spartiates’ were present (11.4.5), which in context clearly means ‘including 300 Spartan citizens’. By implication, the other 700 were perioikoi, also liable to serve in the Lacedaemonian army.23 The same number appeared already in Isocrates’ Archidamos (6.99–6.100), in a context where the author exaggerated the impossible odds Spartans were prepared to face: at Dipaea, a single rank fought tens of thousands of Arcadians; in the Battle of Champions, 300 Spartans defeated ‘all Argives’ (whereas in the usual version they merely defeated 300 picked men); at Thermopylae, 1,000 stood against an army of 700,000. It would have suited Isocrates’ rhetoric to focus on the 300, and the fact that he did not do so suggests that the story of the 1,000 Lacedaemonians was well-established by 380BC, when he first alluded to it in Panegyricus (4.90). The Legend as told by Ephorus a generation later claimed that all 1,000, including the 700 perioikoi, had been on a suicide mission and died alongside the Spartans.24 Insofar as the Legend served to vindicate Spartan leadership, it gave all Lacedaemonians a share of the credit. The number of Lacedaemonians cited is likely to be accurate. If the Spartans had had the license to invent a number, they would surely have claimed that the force they sent was much larger, and there would have been no need to create the Legend at all. It would in any case have been surprising if only Spartiates

23

24

the text’, though he also claims that they ‘are likely part of the legend which grew almost instantaneously’ (and does not explain how this tallies with the legend of the suicide mission). On perioikoi, see e.g. Van Wees 2004: 83–85; Mertens 2002. Ctesias, Persica fr. 13.28 refers to a force of 300 Spartiates and 1,000 perioikoi—but at Plataea. Helots were also present at Thermopylae (Hdt. 7.229.1; 8.25.1), in addition to the 700 perioikoi (not instead of them, as assumed by Matthew 2013a: 23; 2013b: 67 n. 38). Leonidas’ words to the Ephors in Diodorus, as quoted above, assume that all 1,000 must die (contra e.g. Flower 1998: 368: ‘they did not stay to perish with the 300 Spartans’).

30

van wees

had been sent into action at Thermopylae, since mobilisation of perioikoi was normal practice in Sparta.25 Moreover, an inscription set up at Thermopylae, surely soon afterwards, implied the presence of 1,000 Lacedaemonians: ‘Here once fought against three million/ four thousand men from the Peloponnese’ (7.228.1). The figure of 4,000 Peloponnesians is given also by the Legend, which broke it down into 1,000 Lacedaemonians and 3,000 others (Diod. Sic. 11.4.5– 11.4.6; Justin 2.11). The implication is that the number of 1,000 Lacedaemonians was cited from an early stage of the tradition onwards. Herodotus nevertheless took a different line. ‘Three hundred hoplites of the Spartiatai’ headed his list of defenders of Thermopylae (7.202). Leonidas ‘selected the established 300 men who had living sons’ (7.205.2). Herodotus boasted that he had ‘learnt the names of all 300’ (7.224.1), including the one or two ‘of the 300’ who survived (7.229.1, 7.232; 9.71), and often referred to the men as specifically ‘Spartiates’, not just ‘Lacedaemonians’.26 He could not have been more insistent that the Spartan force included 300 full citizens but no perioikoi. The same situation is presupposed by the words he attributed to Demaratus, who told Xerxes after the battle that the ‘city of Sparta’ had ‘about 8,000 men who are all the equals of the men who fought here; the rest of the Lacedaemonians are not their equals but still good men’ (7.234.2). By implication, only Spartiates fought at Thermopylae, and indeed perioikoi were categorically not brave enough to fight to the death in the Spartan manner. Again Herodotus took this view although he was aware of the different story told by the Legend. He had Demaratus warn Xerxes before the battle that being outnumbered would not deter the Spartans: ‘if they happen to march out with 1,000 men, these will fight you’ (7.102.3; 7.103.1). This might be a randomly chosen number, but is likely to represent a snippet of the Legend. Herodotus also knew the inscription about 4,000 Peloponnesians facing the Persian hordes, and his own list of Peloponnesian contingents adds up to 2,800 men (7.202; cf. Paus. 10.20.2), close enough to the 3,000 non-Spartan Peloponnesians of the other version to suggest that he followed the Legend on this point. The consequence of insisting that only 300 Spartans and no other Lacedaemonians fought at Thermopylae, however, was that his own total fell far short 25

26

Hignett 1963: 116–117; Lazenby 1993: 134, object that perioikoi ought to have provided the same number of men as the Spartans, as at Plataea; however, we cannot assume that the same proportions would apply to the highly selective levy at Thermopylae and the much larger-scale levy at Plataea (probably a general levy: see van Wees 2017). ‘Spartiates’: Hdt. 7.211.3; 7.220.1, 3; 7.224.1; 7.228.2; ‘Lacedaemonians’: 7.211.3; 7.218.2; 7.222.1; 7.226.1; 7.227. The latter term of course includes Spartiates: Dienekes is called a ‘Spartiate’ (7.226.1), then a ‘Lacedaemonian’ (7.226.2); the famous epitaph ‘for the Spartiates’ (7.228.2) is summarized as ‘for the Lacedaemonians’ (7.228.3).

thermopylae: herodotus versus the legend

31

of the 4,000 of the inscription. Herodotus appears to have been aware of the discrepancy and to have tried to resolve it by reinterpreting the inscription. He claims that 4,000 Greeks died at Thermopylae, and that these included not only 300 Spartans but also 700 Thespians and many helots (8.25.1), as well as various allies who fell in the fighting before the final day (7.228.1). Herodotus apparently took the ‘4,000’ of the inscription and interpreted it as the number of Greek casualties rather than the number of Peloponnesian forces. This interpretation is not tenable: the text clearly speaks of Peloponnesians who fought, not Greeks who died.27 Rejection of the Legend thus created difficulties which Herodotus was unable to solve. A clue as to why he nonetheless dismissed tradition on this point lies in his claim that he knew the names of ‘all 300’ (7.224.1). The obvious place to learn the names of the 300 was a monument at Sparta where these names were inscribed, and Herodotus visited Sparta at least once (3.55.2). The monument was constructed 40 years after Thermopylae, in 440 BC, when Leonidas was reburied at Sparta and annual games were instituted in his honour (Paus. 3.14.1). The new cult of Leonidas and the 300 was established on behalf of King Pausanias, a child of 5 or 6 years old at the time, whose father and grandfather had been badly compromised by convictions for treason. The glorification of Thermopylae clearly served to prop up a teetering dynasty by reminding everyone of its past glories—Leonidas being the child-king’s nearest relative with an untainted military reputation.28 The inscription of the names of 300, not 1,000, men implies that by 440 BC the Spartans, or at any rate the supporters of young Pausanias,29 preferred a version of the story which asserted Spartiate superiority by ignoring or denying the presence of perioikoi. A reason for this preference will have been that the Legend acquired an extra political dimension after the Messenian Revolt of the 460s, during which the rebels included a few perioikoi (Thuc. 1.101). This crisis will have driven the Spar-

27 28

29

See Lazenby 1993: 148 (ibid. 135 his attempts to reconcile Herodotus and the inscription create more problems than they solve). Further discussion of the inscription below. King Pausanias was still a minor in 427BC (Thuc. 3.26) and thus an infant when he succeeded his father Pleistoanax, who was exiled in or shortly after 446 BC (on the grounds that he had been bribed to abort an invasion of Attica: Thuc. 1.114; 2.21; 5.16–5.17). King Pausanias’ grandfather, the Regent Pausanias, had notoriously died in disgrace and lost all credit for his victory at Plataea. His great-grandfather Cleombrotus had done no more than dismiss the allied army from the Isthmus in 480 on account of an ominous solar eclipse (Hdt. 9.10), leaving Cleombrotus’ brother Leonidas as the nearest reputable ancestor. Sources associated with this particular political faction in Sparta might explain Herodotus’ notably sympathetic portrayal of the discredited Regent Pausanias, and his openly hostile portrayal of Cleomenes, Leonidas’ and Cleombrotus’ half-brother and rival.

32

van wees

tiates to make stronger claims for their own superiority and their right to rule. If it is true that another unit of 300 Spartans fought to the death during this revolt (Hdt. 9.64), this parallel massacre will have further encouraged an exclusive focus on the fate of the 300 Spartiates alone. Herodotus evidently chose to rely on the detailed and seemingly authoritative information provided by the new monument of Leonidas and the stories related in speech or song at the annual games. In principle, this was a reasonable choice for a historian to make, but in practice the Legend was on this particular point less politically charged and more reliable.

3

The First Two Days of Battle: Close Combat or Archery?

The initial fighting at Thermopylae as related by Diodorus followed a clear pattern of escalation and increasing domination by the Spartans. The first wave of attack came from the Medes, along with ‘the brothers and sons of those who had fallen at Marathon, in the belief that these would be most committed to taking revenge on the Greeks’ (11.6.3–11.6.4). When these retreated with heavy casualties, the next two contingents in line, Kissians and Sacae ‘selected for their excellence’, charged but were also defeated (11.7.2–11.7.3). Then Xerxes sent in the Immortals, his very best troops, again to no avail. By the end of the day, only ‘a few men had fallen among the Greeks’ (7.4). On the second day, Xerxes raised the stakes still further and sent in the bravest and most daring men ‘of all nations’ (11.8.1), who ultimately tried to flee but were blocked by the rest of the army which forced them to renew the fight (11.8.3). Whereas on the first day the fighting was done by ‘the Greeks’ at large, on the second day the Spartans did all the work: ‘the men around Leonidas’ were so determined that they ‘did not make way for the usual troops to take their share of fighting in turn’ but fought all day (11.8.2–11.8.3).30 In short, a day of increasingly intense attacks held off by the Greeks collectively was followed by a day on which the Spartans single-handedly took on the best of the rest of the world, and won. The relatives of the dead of Marathon also featured in Justin, whose account was otherwise vague and referred only to ‘useless multitudes’ defeated by the Greeks, over the course of three, rather than two, days (2.11.2–2.11.4). Plutarch included no details of the first two days of fighting at all. A neat pattern of escalation and clear Spartan dominance also shaped the idiosyncratic version of 30

Although ‘the men around Leonidas’ elsewhere includes all Greek troops, the refusal of these men to let others take their turn only makes sense if Diodorus is here referring to the 1,000 Lacedaemonians fighting all day without being relieved.

thermopylae: herodotus versus the legend

33

events presented by Ctesias in the early fourth century BC (but known only from a late antique summary).31 The point of this aspect of the Legend is obvious: in combat, the Spartans acquitted themselves of their leadership obligations in an exemplary manner by shouldering most of the burden of combat with astonishing success. Herodotus, by contrast, was determined to play down any notion of escalation and to keep the Spartans in their place alongside other Greeks. He introduced the Spartans before the fighting even began, but went out of his way to stress that they were not alone: a scout who saw the Spartans calmly exercising or doing up their hair, ‘could not see the entire army’, only the part of it stationed in front of the wall, ‘and it happened that at the time the Lacedaemonians were stationed outside’ (7.208.2). On the first day of fighting, Medes, Kissians and ‘others’ were easily defeated by ‘the Greeks’ (7.210), then the Immortals confidently took over but did no better against ‘the Greeks’, among whom the Spartans fought especially well (7.211). Herodotus concludes that the Persians were unable to succeed whether they attacked ‘by units or en masse’ (7.211.3). There is no escalation here, and all the Greeks contribute. For the second day, Herodotus reports only that the Persians did no better than before (7.212.1) and that the Spartans were not the only Greeks to fight: ‘the Greeks were marshalled by unit (taxis) and by nation (ethnos) and each played its part in the fight’ (en merei hekastoi emachonto, 7.212.2). It is possible, and even likely, that Herodotus drew on first-hand evidence. In other contexts, he was able to cite the eyewitness testimony of two Greeks on the Persian side, and these men, or others like them, or their contemporaries on the Greek side, could have told Herodotus a good deal about the two days of battle when the whole Greek army was present.32 At the same time, Herodotus’ peculiar emphases suggest that he was not just neutrally recording what he had 31

32

Ctesias fr. 13.27: ‘Xerxes sent Artapanos with 10,000 men to attack Leonidas the general of the Lacedaemonians at Thermopylae, and the Persian horde was cut down while 2 or 3 Lacedaemonians were killed. Then he ordered an attack with 20,000 and these too were defeated. Then they were whipped into battle, and although they were whipped they were still defeated. On the next day, he ordered them to fight with 50,000 men, and when that came to nothing, he put an end to the fight.’ Ctesias (ibid.) also had the story of Demaratus’ secret message, but from the summary it is not clear if he otherwise followed the Legend. Dikaios son of Theokydes, an Athenian exile, ‘highly regarded among the Medes’, on good terms with Demaratus (8.65), and Thersander, a leading man of Orchomenos (9.16). Thersander told his story orally to Herodotus in person, and the same is likely to be true of Dikaios, although we cannot exclude a written source for the latter’s story. Both are cited for predictions of Greek victory, at Salamis and Plataea, and it is surely for the contentious nature of their stories that they are cited by name, and are said to have had many witnesses, while informants for ‘factual’ elements of the story are not named.

34

van wees

heard but tacitly modifying an excessively schematic and Sparta-centred early version of the Legend. In the Histories, the Spartans still received more credit than anyone else, but they were brought down to a more human level and made to share the glory with their allies. This version will have been more palatable to Herodotus’ audiences in the many cities that found themselves in rivalry, or indeed at war, with Sparta and were keen not to have the Spartans monopolise the story of Thermopylae.33 The precise nature of the fighting is another matter on which Herodotus differs from Diodorus,34 but the two accounts do have in common an exclusive focus on hand-to-hand combat which is historically implausible. Diodorus had wave upon wave of troops charging at the Greeks, only to be driven back in close combat (11.6.4, 7.2, 7.4), ‘hurling themselves upon the Greeks with great density and violence’ while ‘the men around Leonidas fenced themselves in and made their close formation (sustasis) much like a wall’ (11.8.2). The Greeks won partly because of their prowess (aretai) and partly because they were protected by large shields while their opponents with ‘small shields and peltai’ were more vulnerable (11.7.2–11.7.3). Herodotus likewise imagined that Persian tactics were to advance and try to take the Greek position by storm (7.210–7.212). The Medes and Kissians were defeated simply because they were cowards—‘many people, but few men’ (7.210.2)—but the Persian Immortals lost because they had ‘shorter spears than the Greeks’ (7.211.2). The implication is presumably that they were not cowards but merely disadvantaged; in his account of the Battle of Plataea Herodotus explicitly says that the Immortals were a match for the Spartans ‘in courage and strength’ (9.62). Another factor in both battles is that the Spartans—not the Greeks in general—displayed greater skill: they were ‘experts fighting against non-experts’ (en ouk epistamenoisi machesthai exepistamenoi, 7.211.3; cf. 9.62). It is highly surprising that in both versions the Persians tried for two days to take the pass by storm and engaged in continuous fighting at close quarters and in close order. Herodotus had already told us that Persians, Medes and Kissians were equipped not only with short spears and daggers, but also with ‘large bows and reed arrows’, and were protected by iron scale armour and large but light shields called gerra (7.61–7.62). He later tells us how they used this armament at Plataea: they set up a wall of shields, and from behind this 33 34

As far as the narrative of the first two days of fighting is concerned, therefore, it is Herodotus rather than Diodorus who is more ‘panhellenic’, pace De Bakker, this volume. See De Bakker’s interesting analysis in this volume of Diodorus’ military vocabulary, suggestive of the fourth century or later. Even if Ephorus and Diodorus described the fighting in the language of their time, they may have retained the substance of an older tradition.

thermopylae: herodotus versus the legend

35

screen rained down arrows on the opposition; when the enemy charged, they put down their bows and fought hand-to-hand from behind their shields (9.61– 9.63). The bow was thus their primary offensive weapon, with spears used in stationary defence. If this is how the Persians fought in the open plain, they had all the more reason to adopt the same tactics in an assault on an easily defensible pass—especially one with a fortification wall built across it (see below). The Greeks at Thermopylae ought to have suffered a two-day barrage of missiles rather than a series of hand-to-hand battles resulting from attempts to storm their position.35 On the final day of fighting at Thermopylae, it was indeed Persian archery that won the day. First, Herodotus reported, the Phocians guarding the mountain path were driven from their position when ‘they were hit by many dense volleys of arrows’ shot by the Immortals (7.218.3). Next, the last Greek survivors were defeated because ‘the barbarians overwhelmed them with missiles’ (katechōsan hoi barbaroi ballontes, 7.225.3). Finally, in flashback, we hear that a Trachinian had warned that the Persians would blot out the sun with their arrows, to which the Spartiate Dienekes had coolly replied: ‘Good news. If the Medes blot out the sun, we will be fighting in the shade’ (7.226). The Legend had a very different account of the Spartans’ last stand, as we shall see, but also ended with them being overwhelmed by missiles, without any hand-to-hand combat. The Greeks, in turn, will surely have relied at least in part on missiles to defend themselves. Herodotus noted that the Greek forces spent their time at Thermopylae before the Persians arrived rebuilding the ruined ‘Phocian Wall’ across the pass (7.176.3–7.176.5, 7.225.2–7.223) and that at any given time only 35

For Persian tactics at Plataea, see Konijnendijk 2012 (contra Matthew 2013a: 18–19, 25– 26, a caricature of how Persians fought). Lazenby 1993: 138, rightly notes that the Persians ‘should have been able to hold off and wear the Greeks down with arrow fire’; others consider the possibility of missile combat but dismiss it: e.g. Cartledge 2006: 142. One of the referees for the Press raised a series of objections to my argument at this point: (a) we must trust Herodotus’ account of close combat because he drew on eyewitness accounts; (b) ‘Persian tactics were variable’, and Persians had discovered at Marathon that archery did not work against Greeks; (c) the Persians had to attack at close range on the second day because the first day had shown that the Greeks ‘had to be drawn out of their position’. My answers are: (a) Herodotus had both the Legend (perhaps in the form of a detailed poetic narrative) and eyewitness evidence at his disposal, and may have drawn more on the former than the latter; (b) any lessons supposedly learned at Marathon had evidently been forgotten by day 3 (see below) and at Plataea in 479, when Persian archery was once again prominent; (c) there is no change in Persian tactics between the first and second day in Herodotus or Diodorus, and it is this immediate and exclusive reliance on storm tactics that makes both accounts implausible.

36

van wees

part of the Greek forces was stationed outside this wall (7.208.2; 7.223.2). The majority stationed inside the wall would hardly have sat still waiting for their turn but manned the battlements and supported their comrades with missile fire. The Spartiates’ helot attendants (7.229.1), many of whom also fell in action (8.25.1), will have been light-armed, as they were at Plataea (9.10.1, 9.28.2, 9.29.1), and the same is true for any unmentioned attendants which other Greek hoplites brought along. The Locrian general levy may even have consisted of archers and slingers, judging by their role in the Iliad (13.712–13.722). And yet, the stories of the first two days of combat mentioned no Persian or Greek firing a single arrow. These narratives must at the very least be highly selective. Such selectivity served the interests of Greek hoplites in general and Spartan hoplites in particular. Exposure to prolonged missile fire was among the hoplite’s greatest frustrations, as illustrated by the famous complaint of a survivor of the Spartan defeat at Sphacteria in 425 BC that ‘spindles’, i.e. arrows, do not distinguish between brave men and cowards (Thuc. 4.40). For the Spartans, to acknowledge that light-armed were more significant in the defence of Thermopylae than hoplites would have meant giving most of the credit to their helots, which was unthinkable. In order to rank as a heroic Spartan and Greek action, therefore, the fighting at Thermopylae had to be represented as a pitched battle in which Greek hoplites proved their superiority in close combat.36 Herodotus inadvertently gave away the game and undermined the overall picture of pitched battle when he offered one illustration of superior Spartan skill: They would turn their backs and repeatedly pretend to run away en masse, and the barbarians who saw them fleeing would advance shouting and clattering, but when they were overtaken they would turn about to face the barbarians and as they turned back they slaughtered a countless mass of Persians. 7.211.3

The feints here executed by the Spartans are elsewhere regarded as typical lightinfantry tactics, and require much open space to be effective. Peltasts or psiloi run at a hoplite phalanx, fire their missiles and run back again, in the hope of provoking the hoplites to break their formation so that they can be picked off individually. Herodotus seems to posit a radical role reversal: Spartans, the

36

See Hunt 1998 for the exclusion of helots and slaves from military historiography.

thermopylae: herodotus versus the legend

37

ultimate hoplites, operate like light infantry against a phalanx of barbarians. One might detect literary games being played here, but I would argue that for once we get a glimpse of how the fight at Thermopylae was really fought. The Spartans’ mobile tactics would have been impossible against troops charging against them while they had their backs almost literally against the wall, but would have been feasible and effective if the Persians formed a shield wall at firing distance and kept the defenders of the pass under constant missile bombardment. ln that case, the Greeks could do little except sally out in hit-and-run attacks and provoke the Persians into breaking rank and coming closer to the Greek fortification so that the rallying hoplites, and the uncredited Greeks on the wall, could do some damage.37 In short, Herodotus’ illustration of Spartan skill fits much better in a context of missile warfare than in a Persian storming of the pass, which suggests that this detail came from an eyewitness account. But although Herodotus had this information, and although everything he knew about Persian equipment and Persian tactics as well as various anecdotes and simple common sense suggested that the action ought to have been dominated by missile combat, he clung to the idea that the Persians had for two days made no use of their superior firepower but relied on storm tactics which resulted in a lengthy handto-hand battle for which they were not well equipped. So powerful was hoplite ideology that Herodotus accepted the Legend’s deeply implausible claim that Greeks and Persians each gave up their respective military advantages and marched out to fight in open battle.

4

The Last Stand: in the Persian Camp or Behind the Phocian Wall?

The Legend ends with an undoubtedly fictional assault on the Persian camp.38 When they found out during the night about the encircling movement of the enemy, they got up and made for the camp and the tent of the king in order to kill him and to be killed around him. They advanced right up to the tent, killing whoever got in their way and routing the rest. When

37 38

Lazenby 1993: 138, similarly explains the passage as reflecting anti-missile tactics. ‘Fictional’ because factors of time and space make it impossible, as discussed below, not because ‘hoplites were not commandos, and such an attack would be quite unparalleled’ (Lazenby 1993: 142): Herodotus in fact relates a very close parallel, which may even have been a model for the Spartan story: a night-raid by 600 Phocian hoplites on a Thessalian camp ‘not many years’ before 480 BC, in which 4,000 Thessalians were killed (8.27).

38

van wees

they did not find Xerxes, they searched for him in the great, wide open, camp, and as they wandered around they were slain with difficulty by the barbarians who surrounded them on all sides. Plut. Mor. 866a

Justin said that after failing to catch Xerxes the Spartans turned to massacring Persians well into the next day, until they fell ‘not conquered but exhausted by conquering’, inflicting on Xerxes ‘two wounds’, i.e. this night raid as well as their resistance at the pass (2.11.11–2.11.19). Diodorus told the tale in most detail (11.9.2–11.10.4), including Leonidas’ famous exhortation to his men to eat and fortify themselves for a long fight, ‘for tonight we dine in Hades’ (11.9.4), as well as the detail that after daybreak the Persians became aware of the small numbers of Greeks but still did not dare face them in hand-to-hand combat. Instead, ‘surrounding them on the flanks and in the rear, shooting arrows and throwing javelins from every direction, they killed them all’ (11.10.4). If the Spartans had managed to find Xerxes, he noted, ‘the entire war might have come to a quick end’ (11.10.3). Before launching his final attack, Leonidas ordered the other Greeks to make their escape (Diodorus 11.9.1; Justin 2.11.5–2.11.7). Justin claimed that only 600 men, referred to as ‘Spartans’ and ‘Lacedaemonians’, stayed and attacked the Persian camp (2.11.15). Diodorus said that Leonidas ‘kept beside him only the Thespians’ who together with the Lacedaemonians amounted to ‘no more than 500 men’ and launched the attack against one million Persians (11.9.2; 11.11.2). Plutarch insisted that Thebans also joined the attack (Mor. 864ef). The Greek numbers are surprisingly small, and presumably imply that the rest of the Lacedaemonians and Thespians had fallen in combat on previous days.39 This night raid by the last remaining few hundred men was, it seems to me, an integral part of the Legend from the outset rather than grafted onto the story at a later stage. The Legend posited that a Spartan king had to die to save Sparta and Greece, but the oracle did not say that he and his men had to die merely standing their ground in the pass, and indeed gave no rationale for the death of 1,000 Lacedaemonians. Self-sacrifice in a purely defensive action might be a noble symbolic gesture but an offensive action aimed at killing Xerxes was

39

Justin’s number of 600 may reflect some confusion with, or borrowing from, the story of the Phocian night-raid mentioned in the previous note. For the role of the Thebans in this final battle, see below. Flower 1998: 372, 373 assumes that Diodorus’ ‘500’ includes 300 Spartans and 200 Thespians, but nothing suggests that the 700 perioikoi had left (above, n. 14), and Diodorus never gives the number of Thespians (700 in Herodotus). Paus. 10.20.2 uniquely says that the Mycenaeans also stayed to the end.

thermopylae: herodotus versus the legend

39

potentially of more practical use and indeed could have ended the war at once. A kamikaze attack on Xerxes’ tent was an outcome which followed logically from the premise of a Spartan suicide mission involving not just the devotio of Leonidas but the annihilation of his entire unit.40 The purpose of this element of the Legend was, as ever, to assert that the Spartans had done everything one could ask of a leader. They had not just, by order of the oracle, sacrificed their king to secure the ultimate victory for Greece, but given their lives in a tactical offensive which might have ended the war then and there. What is more, Leonidas had not been abandoned by his allies but ordered them to save themselves for future battles. Not only did the Legend thereby absolve the other Greeks of cowardice, but it claimed that the Spartans had remained in control of its Peloponnesian League at all times. The presence of the Thespians and Thebans until the very end, however, does not serve the Legend’s purpose, but rather constitutes an awkward exception to its message of uniquely Spartan leadership and sacrifice. We may infer that the Legend had to accommodate the presence of these Boeotian units since it was a historical fact too well known to be ignored. Herodotus’ version of events was fundamentally different insofar as it featured no night raid but ended with a defensive last stand behind the fortification wall. This account seems at first sight independent of the Legend. Yet a closer look reveals some odd features, which suggest that Herodotus did, after all, know this part of the Legend, too, and in his narrative attempted to reconcile it with an alternative version of events. He began by suggesting that the allies left because they simply did not want to defend Thermopylae. The ‘other Peloponnesians’ had wanted to leave as soon as the Persian army appeared at the pass, but had been made to stay by Leonidas (7.207). Now, about to be surrounded, they were ‘split’ on what to do. ‘Some departed and scattered as each returned to their city, while others stayed with Leonidas’ (7.219.2). This version of the story offered a radical challenge to the Legend, but Herodotus pulled back from it. He went on to note that alterna40

Both referees for the press questioned the link between the suicide mission and the night attack, arguing that the latter ‘might … be a later elaboration’, i.e. presumably an invention by Ephorus after all (contra Flower 1998). However, the suicide mission needs the night raid because this version of story (unlike Herodotus’ version) offers no other rationale for the deaths of all Lacedaemonians. Moreover, the night raid predates Herodotus, whose account reveals hints of it: see below. Finally, it is not at all clear why Ephorus, or any author later than Herodotus, would have invented this unbelievable episode, and modern scholars have offered no explanation beyond a generic and quite unjustified dismissal of fourth-century historians as supposedly capable of writing any old rubbish (see n. 4, above).

40

van wees

tively ‘it is said’ that Leonidas ordered the Greeks to leave; he listed arguments in support (7.220–7.221) and concluded that this story was more likely than that ‘those who departed left in such disorderly fashion because of a difference of opinion’ (7.220.4). While raising a more critical version of events, Herodotus thus opted again to retain, and actively defend, the Legend that placed the Spartans and their allies in a more flattering light. A major oddity in Herodotus’ version is the sequence of events during the night before the final battle. In Diodorus, the Greeks were warned by a deserter who arrived during the night; this prompted Leonidas to embark immediately on his midnight raid (11.8.5). In Herodotus, deserters likewise arrived ‘when it was still night’, and at the same time the diviner Megistias saw signs that ‘dawn would bring death’ (7.219.1).41 Bizarrely, Herodotus’ Greeks did not respond at all to the bad omens or the bad news. Only when, ‘in third place, the day scouts came running down from the heights, when it was already broad daylight’ (7.219.1), did they meet in council to decide whether to run or fight. For good measure, Herodotus reports that the Persians, too, waited until well into the new day: Xerxes made a libation to the rising sun but did not send his troops into action ‘until the time when the agora is at its fullest’, i.e. mid- or late morning, in order to give the troops which had gone around the Greek position time to come down from the mountain (7.223.1). For the Greeks to receive two early warnings but ignore them and go back to sleep until they got a third warning when the encirclement was nearly complete, would be deeply negligent behaviour, but there is not a hint in Herodotus that we are meant to see it as such. Instead, what we have here is surely an awkward conflation of two different stories. One is the Legend that the Greeks were warned at night and immediately counter-attacked. The other is that the Greeks remained unaware of Persian movements until the day scouts reported well after dawn, at which point they got ready to fight or to leave before the Persians attacked later in the morning. Herodotus thus tried to incorporate— to incongruous effect—as much as possible both the events and timings suggested by the imaginative Legend and those suggested by a more prosaic source. The same thing happens in Herodotus’ strange narrative of the fighting which ensued. The Greeks around Leonidas, aware that they were making an expedition that would end in death, went out much further than before into the

41

Our summaries of the Legend do not mention Megistias, but the fact that he had his own epitaph at Thermopylae (7.228.3: see below) suggests that he had a role in it.

thermopylae: herodotus versus the legend

41

wider parts of the pass. The bulwark of the wall protected them, and during the previous days they had sallied out to fight in the narrows. But at this time, when they engaged outside the narrows, a great number barbarians fell … [The Greeks] displayed all the strength they had against the barbarian, acting heedlessly and self-destructively. 7.223.2, 7.223.442 Leonidas himself ‘and many notable Spartiates’ fell in this charge (7.224.1), and ‘there was a great othismos between the Persians and Lacedaemonians over the body of Leonidas until the Greeks by their excellence dragged the body away and routed their opponents four times’ (7.225.1). However, when Persian troops arrived to their rear, ‘then the conflict changed in nature: they retreated to the narrow part of the road, crossed the wall, and took position on the hill … This hill stands in the approach where now the stone lion stands over Leonidas’. Here they were surrounded and overwhelmed ‘by missiles’ (7.225.2–7.225.3). This very last stage of the fighting on the hill would make sense if the last few Greeks had been staying behind the wall and were caught in the rear. But the preceding sequence does not make sense, and also makes nonsense of what happens at the end. Why the initial move forward into more open terrain? By advancing, the Spartans gave up both of their crucial advantages: the narrowness of the pass—only 15 metres wide, so that fewer than 20 hoplites side by side could completely block the passage—and the vicinity of the defensive wall. A forward move achieved nothing except to make it easier for the enemy to outflank and surround them. Unlike a raid on the Persian camp, it would have been a suicidal action for no conceivable gain.43 If we accept for the sake of argument that this is how they chose to die, why the subsequent retreat behind the wall? One might argue that they lost their suicidal drive once Leonidas had fallen, but then they could have pulled back towards the wall and resumed their previous, safer, tactics, rather than cross the wall and abandon it altogether at the very moment that the area behind the wall was no longer relatively safe, because it was filling up with Persians attacking from the rear. Again the Greeks seem determined to make things easy for the enemy, for no discernible gain.

42

43

This translation of the final clause, parachreōmenoi te kai ateontes, is suggested by Clarke 2002: 66–70, who plausibly argues that it hints at criticism of suicidal tactics. If so, this is another indication of Herodotus’ rejection of the Legend’s kamikaze mission. Lazenby 1993: 146, suggests that by going on the attack the Spartans hoped to intimidate the Persians into keeping their distance and thereby give themselves a chance to escape. No one has come up with a better explanation, and its sheer implausibility shows that the supposed Spartan attack cannot be accounted for in realistic tactical terms.

42

van wees

This bizarre scenario was, I suggest, a construction by Herodotus based on the Legend but adjusted in the light of other evidence for topography and chronology. The latter told him that no fighting had taken place until it was day, so he could not accept the story of the night-raid on the camp. The topography of the battlefield showed that such an attack would in any case have been physically impossible. Xerxes’ camp was at Trachis, the only part of the region which had a sizeable plain (Hdt. 7.199, 7.201), or according to Diodorus on the River Spercheios (11.5.4–11.5.5) north of Trachis. Either way, the camp was miles away from the Greek position at the so-called ‘Middle Gate’ of Thermopylae, and it could be reached only through the so-called ‘West Gate’, an even narrower pass, with a single-cart track, 15 stades (2.7km) west of Thermopylae (7.200). Clearly, the Persians guarded the West Gate heavily, and a few hundred Greeks could not have made their way through, let alone cross another couple of miles of the plain of Trachis, without being overwhelmed long before they came anywhere the Persian camp. Herodotus’ exceptionally detailed topographical description of the site (7.199–7.201) serves as a tacit refutation of the Legend on this point. Moreover, the Greek memorial at Thermopylae, a statue of a lion, stood on the hill behind the Phocian wall, which implied that the final battle had taken place here, not somewhere in Trachis. Herodotus’ topographical information is perfectly accurate and presumably derives from his own visit to the site. The story that the final fight did not begin until the morning of the third day was certainly not invented by Herodotus himself, since he was keen to hang on to the midnight omens and deserters of the Legend. His sources could be Peloponnesian soldiers who saw the ‘day scouts running down from the heights’ and abandoned the pass, and Greeks in Xerxes’ entourage who may have been invited to attend the king’s ‘libation to the rising sun’ before he launched his final assault (7.223.1). From the evidence at his disposal, then, Herodotus inferred that the Spartans had had no time or space to move far forward from their defensive position at the wall, and that they had ultimately been massacred on the hill. But rather than conclude that the Spartans had died while defending the wall and that the Legend was utterly false, he again tried to salvage something from it. So he posited that Leonidas had indeed taken the offensive against the Persians, as the Legend claimed, and died in an epic charge, but only during the day, and not beyond the West Gate. As a corollary, he had to posit that the final survivors had withdrawn behind the wall. Thus he was able to reconcile the Legend with most of the evidence that he encountered, but ironically this rationalisation entailed some quite irrational Spartan tactics. Finally, Herodotus accepts that the Thespians stayed with the Spartans to the end (7.222, 7.227) but claims that 400 Thebans stayed only because Leonidas

thermopylae: herodotus versus the legend

43

forced them and that they surrendered as soon as the king was dead, only to be branded as slaves by the Persians (7.205, 7.222, 7.233). Plutarch rehearses in detail everything that is implausible about this episode (Mor. 864c–865f; 866d– 867b) and claims that it was inspired by personal hostility to Thebes: ‘before Herodotus not a single person knew this story’ (867a). A more likely motive for Herodotus’ undeniably anti-Theban story is suggested by the name of the Theban commander: Leontiades son of Eurymachos. The son of this man, another Eurymachus, was according to Herodotus in charge of 400 Thebans who in 431 BC tried to occupy Plataea (7.233.2). This was the event that triggered the Peloponnesian War and, judging by Thucydides’ account (2.2.3), Herodotus’ version of what happened was not exactly accurate, but the close parallel forged by Herodotus is a clue to the point of the story. It was surely an Athenian tale put about in time of war in order to deprive the Thebans of any role in the defence of Greece against Persia, by claiming that they had contributed nothing but a token, coerced, act of resistance. The Athenians made similar accusations against their other main rival, Corinth, whose admiral Adeimantos allegedly had had to be bribed by Themistocles to fight at Artemisium (Hdt. 8.5) and had fled before the battle of Salamis (Hdt. 8.94).44 Despite his endorsement of Athens as ‘saviour of Greece’, Herodotus felt the need to record that no other Greeks accepted the latter claim (8.94.4), but he did uncritically accept the Athenian story about the Thebans as if it were historical fact. This final departure from the Legend thus aligned Herodotus with contemporary Athenian propaganda, but at the same time helped preserve the exclusivity of the glory claimed by Sparta.

5

The Legend and the Histories

The story that Leonidas and his men were on a suicide mission and gave their lives in a kamikaze attack on the Persian camp was developed in answer to those who would challenge Sparta’s position of leadership. Sparta’s limited deployment of troops could be blamed for the loss of Thermopylae—not to mention the ensuing devastation of Phocis (8.31–8.33), Thespiae, Plataea (8.50) and Attica (8.51–8.56) as well as the surrender of Locris and Boeotia—and needed to be justified. This Legend was well-known to Herodotus, who adopted 44

The stories about Adeimantos were surely influenced also by the actions of his son Aristeus in the 430s: Aristeus was a leading figure in the run-up to the Peloponnesian War until his capture and execution by Athens without trial or burial in 430 BC (Thuc. 1.60–1.63, 1.65; 2.67), an event of which Herodotus was aware (7.137.3).

44

van wees

a good deal of it, tacitly modified the rest in the light of other evidence, and in so doing created a more credible and balanced story, but one which nevertheless in large parts remained a fiction legitimating Sparta’s actions in 480BC. When did the Legend emerge? The political incentive to ‘spin’ the story in Sparta’s favour must have been acute within a matter of weeks, when the loss of Thermopylae contrasted badly with Athenian-led victory in the Battle of Salamis. Spartan leadership in the wider Greek world was structurally difficult to maintain in the face of newly overwhelming Athenian naval superiority (Hdt. 8.3–8.5, 8.58–8.64; cf. 7.161) and was soon openly challenged by Athens’ creation of its own alliance. Within the Peloponnese, many Arcadians were resisting Spartan authority already in 479BC, judging by how few joined Sparta at Plataea,45 and a few years later they tried to overthrow Sparta’s hegemony by military means: first the Tegeans declared war, with support from Argos, then ‘all the Arcadians except the Mantineans’ fought Sparta in the Battle of Dipaea (9.35). The pressure on Sparta to defend its leadership was thus immediate and continuous.46 The Spartans had a good deal of leeway to invent. The initial message from Demaratus, the oracular consultation and the rest of the decision-making process would have been confined to a small Spartan elite. In the absence of survivors, the only Greek witnesses to what happened in the last stand were on the Persian side and would hardly have enjoyed much authority in Greece—until much later, when the likes of Herodotus began to examine the past critically. Peloponnesian allies might have objected that they played a greater role in the initial fighting, and more importantly that no night-raid could have taken place because no one left Thermopylae until it was broad daylight. But they had no incentive to do so: the Legend made them, too, look good insofar as it gave them a share in a glorious (if imaginary) hoplite battle against the Persians, and gave them an excellent excuse for their failure to die alongside the Spartans. By co-opting its allies into the Legend, Sparta made sure that no one in the Peloponnese had any cause to challenge it, and indeed that the story countered divisions within the Peloponnesian League by showing all members working in harmony.

45

46

Hdt. 9.28.3–9.28.4: Tegean and Orchomenian contingents only; 9.77: Mantineans (and Eleans) arrived too late (and punished their commanders). ‘The rest of Arcadia’ had at Thermopylae supplied almost half of the Arcadian forces (7.202), but here remains unaccounted for. Contra Hammond 1996: 9, 18, arguing that Sparta had no need to defend its record before the First Peloponnesian War began in 460 BC, when the Legend (minus night-raid, 5–8) was first formulated, to be recorded soon afterwards by a prose author.

thermopylae: herodotus versus the legend

45

Hints that the Legend was promoted at the earliest opportunity appear in the memorials inscribed at Thermopylae, as cited by Herodotus. ‘Stranger, report to the Lacedaemonians that here we lie, obeying their orders’ (7.228.2). Most translations have some variation on ‘Go tell the Spartans’, but the inscription says ‘Lacedaemonians’, including the perioikoi. ‘Obeying their orders’ might allude to the principle that a Spartan must never abandon his post, but in the context of the Legend might suggest that the dead executed an order to undertake a suicide mission. An inscription added privately by the poet Simonides for his friend Megistias pointedly mentioned Spartan leadership and hinted that the Spartans ordered their allies to leave. ‘This is the memorial of famous Megistias, who once was killed by the Medes when they crossed the River Spercheios, a seer who knew full well at this time that the demons of death were drawing near but could not bear to abandon the leaders of Sparta’ (7.228.3). Most remarkable is the inscription for the allied dead, already cited: ‘Here once fought against three million 4,000 men from the Peloponnese’. This is not actually an epitaph in form, like the other two, and it did not actually commemorate the dead: far fewer than 4,000 were buried here,47 and they included Locrians, Phocians and Boeotians as well as ‘men from the Peloponnese’. Instead, it was a memorial of battle which glorified all Peloponnesians who turned up, and only the Peloponnesians. This extraordinary pseudo-epitaph suggests a determined effort to claim that not only the Spartans but also their Peloponnesian allies had done their duty in defence of Greece. Moreover, as noted above, Diodorus—and Ephorus before him, presumably—quoted a verse ‘encomium’ for the men who fell at Thermopylae, attributed to Simonides. This poem was evidently sung in ‘a sacred enclosure’ where ‘an altar is their tomb’, i.e. as part of a form of hero cult for some or all of the dead (11.11.6; Simonides fr. 531 Page). This song of praise may have included an account of the deeds of these heroes, with many a nod towards Homeric battle narrative, as in Simonides’ elegy on the Battle of Plataea (frgs. el. 10–17 West).48 Simonides’ poem may thus have constituted an early and authoritative articulation of the Legend. If Herodotus’ picture of the struggle over the body of Leonidas sounds like something out of the Iliad, it may be because he drew these details from a narrative by Simonides that adopted ‘epic’ features.49 47

48 49

Only the relatively small proportion of 3,000 Peloponnesians (excluding the separately buried Lacedaemonians) who fell in combat on the first two days. For Herodotus’ mistaken inference that 4,000 Greeks (including Lacedaemonians) died, see above. Flower 1998: esp. 373–375, citing Homeric parallels. If Simonides told a version of the Legend, his fight over Leonidas’ body was set in the

46

van wees

In sum, the Legend was in all likelihood told by Sparta and its supporters from the immediate aftermath of the Persian War onwards. In response to a grave internal challenge, the revolt of Messenian helots and some perioikoi in 464BC, the official version of the Legend celebrated at Sparta excluded all Lacedaemonians except the 300 Spartiates. These stories, and other traditions about events of the Persian War, were developed and promoted because they served a political agenda, already in late 480 BC and still just as much when Herodotus collected his information up to 50 years later. The blatantly propagandistic elements of the Legend may be one reason why Herodotus chose not simply to repeat it in his Histories. In rewriting the story, however, Herodotus did not make it anti-Spartan or pro-Athenian. He incorporated as many elements of the Legend as possible, presented a different but still highly positive image of Sparta where he did depart from the Legend, and never explicitly rejected any part of it—as he did with Athens’ slurs against Corinth— but only disagreed tacitly. An account intended to be critical of Sparta would not have been so diplomatic and could have told a much less favourable story.50 Herodotus’ modifications of the Legend were surely motivated primarily by historical rather than political objections to the original tale. As we have seen, his account included a good deal of detail, especially on chronology but also on other aspects of the campaign, that had no place in the Legend and was not of his own invention because it contradicted the story he chose to tell. If he nevertheless regarded such material as sufficiently reliable to require corrections to the Spartan story, it must surely have come from the sort of first-hand testimony, whether oral or written, that he occasionally cited explicitly. Herodotus’ method was to reconcile such seemingly authoritative first-hand evidence with an ‘official’ Spartan story that seemed equally authoritative. His narrative is a compromise between the two.51

50

51

Persian camp, and Herodotus changed the location. Another poetic touch may be the detail that Xerxes jumped up from his throne three times as he watched the battle (7.212.1). Later prose versions of the Legend may of course have changed and added details, such as the vocabulary of combat (noted above, n. 33) or the name of the deserter who warned Leonidas, Tyrrhastiadas ‘of Cyme’, Ephorus’ home town, yet contra De Bakker in this volume I would argue that Diodorus/Ephorus did not structurally ‘innovate’ but reverted to the older Legend. Contra e.g. Hammond 1996: esp. 18; Grundy 1901: 314 (‘Herodotus gives what was, no doubt, the authorised version’). The ‘authenticating’ strategies identified by De Bakker in this volume as much more prominent in Herodotus than in Diodorus confirm that Herodotus is engaging in independent critical analysis while Diodorus rehearses a traditional story. The ‘dualism’ in Herodotus’ portrait of Leonidas detected by Baragwanath (2008: 64–79) closely reflects this historical method: the ‘heroic strand’ consists of elements retained from ‘the Legend’, the ‘human strand’ represents Herodotus’ own ideas.

thermopylae: herodotus versus the legend

6

47

‘What Really Happened’ at Thermopylae?

If we are to get closer to the reality of events at Thermopylae, we should take from the Legend only those elements that cannot be explained as political propaganda and take from Herodotus only those elements that cannot be explained as borrowings from the Legend or as his own speculative, Spartafriendly, solutions to the problems it posed. What remains may well derive from first-hand testimony which, although not necessarily true or unbiased, offers a story not filtered through a pro-Spartan political prism. Above all, the chronological detail that Herodotus provided often contradicted both the Legend and his own version of events, but tells a believable story in its own right, much less flattering to the Spartans and Peloponnesians. First of all, the report that the Greeks did not mobilise until ‘the Persian was in Pieria’52 is plausible in itself and does not serve to glorify any of the allies. Despite the claim that Sparta, thanks to Demaratus, knew from the outset what Xerxes’ plans were—and the claim that Xerxes’ primary objective was revenge on Athens for the Persian defeat at Marathon—it must have been uncertain in summer 480 BC how far and in what direction the Persian army and fleet would go as they took nearly two months to travel very slowly across Thrace and Macedonia, territories under Persian control for more than 20 years already.53 Not until the Persians started clearing roads across the Pierian mountains was it beyond doubt that they would be invading Greece. Bearing in mind also that it would have been a great financial and logistical strain for the Greeks to keep an army and fleet deployed but inactive for a couple of months, everything supports the idea that mobilisation took place only at the last possible moment.54 Secondly, if the chronological details found in Herodotus are correct, Spartans and other Dorians were not inhibited by the Karneia when they first mobilised, and had at least another 10 days to send reinforcements before the start of the month Karneios made this difficult. The Eleans were constrained by the Olympic truce, but the Arcadians were not, and could have sent any number of men at any time. The implication is surely that the Spartans and many other 52 53

54

Setting aside the allegedly very much earlier but short-lived campaign to Tempe (Hdt. 7.174), which may or may not be historical. The slow pace of travel is implied by Herodotus’ comment that the journey from Abydos to Attica had taken 4 months, including one whole month for the crossing of the Hellespont (8.51); deducting 8 days for the march from Thermopylae to Attica (8.23–8.25, 8.66), 7 days at Thermopylae and 14 days for the march from Therma to Thermopylae (above), leaves 2 months in Thrace and Macedonia, including ‘many days’ in Pieria. The Legend apparently claimed that mobilisation took place much earlier, when Xerxes was still at Doriscus: Diod. Sic. 11.4.1, 11.5.1; cf. Hdt. 7.59–7.106.

48

van wees

Peloponnesians sent so few troops simply because they made a strategic decision not to commit more manpower to the defence of Central Greece. Sparta sent the smallest number compatible with honouring its obligations as leader of the alliance, while the Arcadians sent as few as compatible with minimally meeting their obligations to Sparta.55 The same chronological details which disprove Herodotus’ explanation of the limits of Greek mobilisation do suggest that the Spartans, at least, may genuinely have intended to send more troops about a month later, after the Karneia, but were caught out by an acceleration of the Persian advance. The plan to rebuild the Phocian Wall, mentioned by Herodotus but not featured in the Legend, fits with the idea that the Spartans were anticipating a lengthy occupation of Thermopylae. So does the logistical detail—rare in Herodotus— that they made arrangements for provisioning, from the nearby village of Alpenoi (7.176.5). However, having taken nearly 3 months to get from the Hellespont to Macedonia, the Persian army now crossed Macedonia and Thessaly in only two weeks, a quick pace of about 10 miles a day.56 This forced Leonidas to send for reinforcements sooner, during the taboo period of the Karneia. The Dorian cities did not respond, and neither did the Arcadians, although they were at the time the only Peloponnesians not under any religious restrictions. It does not necessarily follow that those Peloponnesians who did mobilise proposed to abandon Thermopylae and Artemisium as soon as the Persian army and fleet arrived, and had to be forced to stay by their Spartan com55

56

Contra the ingenious but untenable thesis advanced by Matthew (2013b), who argues that the advance guard was designed to be large enough to hold up the Persian army for two weeks, long enough for logistical constraints to cripple the Persians through hunger, thirst and illness, at which point Greek reinforcements were scheduled to arrive via a different route (through Phocis) in order to finish off the enfeebled survivors. Polyaenus’ anecdote (Strat. 1.32.3; Excerpta 12.4) about Leonidas devastating the land of an enemy city clearly does not refer to the Thermopylae campaign and does not show that Leonidas turned Trachis it into a ‘denuded wasteland’ before the arrival of the Persians (pp. 76–77). The Persian army may well have experienced logistical difficulties, but with the River Spercheios nearby and the vast plains of Thessaly under their control, it is hard imagine that these would have been severe (pp. 77–81). Above all, if the Persian army did find itself in dire straits, there was nothing to stop them moving on from Thermopylae via the alternative route that they eventually did take (see below) rather than wait for Greek reinforcements to occupy that road (pp. 75–76, 81–83). Hignett 1963: 113, estimates the distance covered in 14 days at 140 miles. Herodotus provided no details of the stages of this march, in contrast to his minute description of the march up to Therma (7.108–7.126), perhaps to reinforce the impression of speed (rather than because he ‘temporarily lost interest’: Hignett 1963: 107), even if he slightly undermined that impression by noting that Xerxes took the time to hold horse races in Thessaly (7.196).

thermopylae: herodotus versus the legend

49

manders (7.207; 8.4–8.5). At least some elements of the story about the Greek navy are highly coloured by pro-Athenian bias and demonstrably untrue,57 which casts doubt on the parallel story about Thermopylae, too. The Spartan and Peloponnesian preference for defending the Isthmus once Thermopylae and Artemision had been lost is another matter, because by that time Phocis, Euboea and Attica had been evacuated while Boeotia and presumably Locris had surrendered, so there was little left to defend outside the Isthmus, but it may nevertheless have encouraged Athenians and others to allege that this was their real intention all along. Herodotus’ chronology further shows that the four-day delay between the arrival of the Persian army and the start of fighting was caused not by Xerxes’ arrogant assumption that the Greeks would just run away and subsequent fury that they dared resist (Hdt. 7.210.1), but by the storms which delayed the arrival of his fleet (7.188–7.193). Once the ships did arrive, Xerxes launched a simultaneous attack by land and sea. As we have argued, one probable piece of eyewitness testimony, as well as other accounts of Persian armament and tactics, indicate that the Persian army did not try for two days to take the pass by storm only to be repeatedly defeated in pitched battles: this was a fictional product of hoplite ideology. The attack predominantly took the form of missile bombardment, while the Greek defence relied on the fortification of the Phocian Wall and occasional hit-and-run sallies against the line of Persian archers. In the light of these tactics, the number of 1,000 Persian dead seen by the naval personnel who came over on the day after battle is not as ‘ridiculous’ as Herodotus thought (8.25.2), and it is rather the claim that another 19,000 had been buried secretly and covered over with leaves (8.24) which is unlikely. During the night of the second day, Xerxes sent his elite force of ‘Immortals’ over the mountains to attack the Greek position from the rear. The Legend blamed this on a gratuitous act of treachery by a Trachinian, and on the cowardice of the Phocians who were supposed to guard this path but abandoned it without a fight (Hdt. 7.213–7.218; Diod. Sic. 11.8.4–11.8.5). It seems unlikely that it had not occurred to the Persians themselves to make the local population show them alternative routes over the mountains, especially since the Persian

57

Clearly false is the story is that Themistocles had to bribe the Spartan commander of the fleet, Eurybiades, and the Corinthian commander Adeimantos in order persuade them to keep the ships at Artemisium (8.5). Eurybiades and Adeimantos did not need bribing because they had compelling reasons of their own to stay here: a withdrawal would have meant allowing the Persian fleet to sail along the Locrian coast right past Thermopylae and attack their fellow-Spartans and -Corinthians from behind—a point ignored by Athenian propaganda and missed by Herodotus.

50

van wees

army eventually marched south, according to Herodotus, by a third mountainous route, via Doris into Phocis rather than via Thermopylae into Locris (8.31), a route known to modern scholarship as the ‘Great Isthmus Corridor’, and much used in later campaigns.58 The latter episode reveals that the Legend omitted an entire dimension of the campaign: if 1,000 Phocians defended the path which led around Thermopylae and was a minor route into Phocis (7.217), many more must surely have been posted to defend the major route via Doris, without any Spartan or Peloponnesian help. We hear nothing about this, except that the Phocians abandoned their territory when the Persians advanced; ‘some Phocians fleeing into the mountains’ were caught and killed, ‘some women’ were gang-raped (Hdt. 8.33). It is possible, therefore, that the Persian army had spent its six days in Trachis not only attacking Thermopylae but also exploring other routes and clearing the way from Doris into Phocis. On the seventh day, they were able to surround the Greeks in the pass, and to storm the fortification from both sides, co-ordinating the timing of the charge, as Herodotus’ sources explicitly said, while launching a final naval assault at Artemisium as well. Herodotus’ sources also told him that most of the Greek troops chose to abandon the pass, and even if he preferred the Legend which put a more positive spin on it, this does seem a plausible reaction, especially from the Arcadians who had been reluctant to defend Thermopylae in the first place. A spontaneous retreat also helps explain why the Thespians and Thebans stayed to fight: unlike the Arcadians, these Boeotians had nowhere else left to go if the Persians could not be stopped at Thermopylae. The surviving Thespians had to evacuate their city and flee to the Peloponnese (8.50). The 400 Thebans, despite Herodotus’ hostile story, were surely volunteers ‘from the other party’, as Diodorus’ put it (11.4.7), i.e. political opponents of the pro-Persian regime in Thebes, who could not go home if a Persian victory kept their enemies in power.59 The Greek forces thus seem to have split according to how important the defence of the pass was to each contingent—with the exception of the Spartans, who presumably stayed because as leaders they felt obliged to do so.60 This, then, was a genuine act of sacrifice on which they soon built their much-inflated Legend. 58 59

60

Well-discussed by Londey 2013; for details of the various routes, see Sánchez-Moreno 2013. Cf. Thuc. 3.62.3–3.62.5: Thebans blame the oligarchic regime for surrender to Persia. If 1,000 Malians were present at Thermopylae (so Diod. Sic. 11.4.7, but not mentioned by Herodotus), they will also have been an exiled faction, since Malis had already been occupied. As several scholars have suggested: e.g. Grundy 1901: 309–310; Green 1996: 139–140. Burn 1984: 418, 420–421, suggested that Leonidas did send away the other troops and stayed as a rear guard to help them escape (and one of the referees for the Press regarded it as

thermopylae: herodotus versus the legend

51

No witnesses on the Greek side survived to report what happened to those who stayed to fight in the final battle. Greeks in Xerxes’ entourage will have been able to observe only that intense fighting took place around the wall, and the fighting on the hill behind the wall was witnessed only by the Persian Immortals. Herodotus’ narrative must therefore be entirely fictional. His description of struggle over Leonidas’ body was probably borrowed from Simonides’ poem, while his evocation of Spartans and Thespians somehow fighting with ‘hands and teeth’ against missiles must be the work of his own imagination. The irony that the Persians, now that they finally did resort to storm tactics, are represented at last as using archery against the Greeks is another product of hoplite ideology, which refused to accept that Persian heavy infantry, however numerous, could defeat its Greek counterpart. From a Persian point of view, the action at Thermopylae will have seemed an unqualified success, a well-coordinated operation that had run like clockwork and managed to occupy an easily defensible position in only three days, thanks to an inadequate commitment of Greek troops—not just at Thermopylae but at other routes south—and divisions among the defenders. From a Greek point of view, it must initially have looked much the same and Spartan leadership will have been much criticised. The sacrifice of Leonidas and 1,000 men, however, provided the basis for a reinterpretation of events so successful that even Herodotus, for all his careful thought and research, was unable to abandon it. From the fourth century BC onwards, the Legend rather than Herodotus’ account established itself as the accepted version of events, as a story that had set out to redeem Sparta became a story told for the glorification of Greece at large.

Bibliography Balcer, J.M., The Persian Conquest of the Greeks, 545–450 BC (Konstanz 1995). Baragwanath, E., Motivation and Narrative in Herodotus (Oxford 2008). Burn, A.R., Persia and the Greeks. The Defence of the West, c. 546–478BC ([1962] London 1984). Cartledge, P., Thermopylae. The Battle that Changed the World (London 2006).

obvious that ‘those men died in a rearguard action’), but although this could in principle have explained the deaths of the Spartans, neither Herodotus nor the Legend actually presents the action in this way: in both versions, the other Greeks have ample time to leave, and the Spartans stand their ground although there are no longer any retreating Greeks to cover.

52

van wees

Cawkwell, G., The Greek Wars. The Failure of Persia (Oxford 2005). Clarke, M., ‘Spartan atē at Thermopylai: Semantics and Ideology at Herodotus, Histories 7.234’, in A. Powell & S. Hodkinson (eds.), Sparta: Beyond the Mirage (Swansea 2002) 63–84. Flower, M.A., ‘Simonides, Ephorus and Herodotus on the Battle of Thermopylae’, CQ 48.2 (1998) 365–379. Green, P., The Greco-Persian Wars ([1970] Berkeley 1996). Grundy, G.B., The Great Persian War and Its Preliminaries. A Study of the Evidence, Literary and Topographical (London 1901). Hammond, N.G.L., ‘Sparta at Thermopylae’, Historia 45.1 (1996) 1–20. Hignett, C., Xerxes’ Invasion of Greece (Oxford 1963). Holland, T., Persian Fire. The first world empire and the battle for the West (London 2005). Hunt, P., Slaves, Warfare, and Ideology in the Greek Historians (Cambridge 1998). Konijnendijk, R., ‘“Neither the Less Valorous or the Weaker”: Persian Military Might and the Battle of Plataia’, Historia 61.1 (2012) 1–17. Lämmer, M., ‘Der sogenannte Olympische Friede in der griechischen Antike’, Stadion 8–9 (1982–1983) 47–70. Lazenby, J.F., The Defence of Greece, 490–479 BC (Warminster 1993). Londey, P., ‘Other Battles of Thermopylae’, in C. Matthew & M. Trundle (eds.), Beyond the Gates of Fire (Barnsley 2013) 138–149. Macan, R.W., Herodotus. The Seventh, Eighth and Ninth Books. Volume II: Appendices, Indices, Plans (London 1908). Matthew, C., ‘Towards the Hot Gates: the Events Leading up to the Battle of Thermopylae’, in C. Matthew & M. Trundle (eds.), Beyond the Gates of Fire (Barnsley 2013a) 1–26. Matthew, C., ‘Was the Defence of Thermopylae in 480 BC a Suicide Mission?’, in C. Matthew & M. Trundle (eds.), Beyond the Gates of Fire (Barnsley 2013b) 60–99. Mertens, N., ‘ouk homoioi, agathoi de: the perioikoi in the classical Lakedaimonian polis’, in A. Powell & S. Hodkinson (eds.), Sparta: Beyond the Mirage (Swansea 2002) 285– 303. Osborne, R., Greece in the Making, 1200–479 BC. ([1996] London 2009). Sacks, K.S., ‘Herodotus and the Dating of the Battle of Thermopylae’, CQ 26.2 (1976) 232– 248. Sánchez-Moreno, E., ‘Communication Routes in and around Epicnemidian Locris’ and ‘Mountain passes in Epicnemidian Locris’, in J. Pascual & M.-F. Papakonstantinou (eds.), Topography and History of Ancient Epicnemidian Locris (Leiden 2013), 279– 335, 337–359. Strauss, B., Salamis: the Greatest Naval Battle of the Ancient World, 480BC (London 2004). Trundle, M., ‘Thermopylae’, in C. Matthew & M. Trundle (eds.), Beyond the Gates of Fire (Barnsley 2013) 27–38.

thermopylae: herodotus versus the legend

53

Wees, H. van, Greek Warfare: Myths and Realities (London 2004). Wees, H. van, ‘Luxury, Austerity and Equality in Sparta’, in A. Powell (ed.), A Companion to Sparta (Malden 2017) 202–235.

chapter 3

A Narratological Comparison of Herodotus and Diodorus on Thermopylae Mathieu de Bakker

1

Introduction: Two Different Traditions That Have Much in Common*

In this contribution I compare Herodotus’ account of Thermopylae (5th c. BCE) with that of Diodorus Siculus (1st c. BCE, Diodorus from now on) by way of a narratological analysis. My aim is to bring to light the main differences in their narrative strategies and thereby draw attention to each historian’s agenda in presenting his material. Diodorus’ version in the Historical Library (11.4.1–11.12.1) is based on an altogether more encomiastic version of the Thermopylae events, dubbed ‘the Legend’ by Van Wees in this volume. It agrees with the longer report of Herodotus in various respects, such as the duration of the battle (three days), the betrayal of the Greeks by a fellow-Greek, and the subsequent withdrawal of all other allies except Spartans and Thespians. We also witness similarities in terms of detail, such as the number of Greek fighters (appr. 7,000, Hdt. 7.202– 7.203.1; Diod. Sic. 11.4.5–11.4.7), the involvement of the Cissians on the first day of the battle (Hdt. 7.210.1; Diod. Sic. 11.7.2), and the preliminary advice of the Spartan exile Demaratus to Xerxes (Hdt. 7.209; Diod. Sic. 11.6.1–11.6.2). The two accounts, however, also contain salient differences, most notably in the presentation of Leonidas and his Spartans. In Herodotus’ version, the Spartan king decides to sacrifice himself and his Spartiates on hearing that their position has been betrayed (Hdt. 7.219.2–7.222), whereas Diodorus describes the Spartan expedition to Thermopylae as a suicide mission right from the beginning (Diod. Sic. 11.4.3–11.4.4). In the account of the actual battle, Herodotus presents the Spartans as defensive heroes, who hold their ground against a majority of Persians. They make their last stand in front of the pass before retreating to the * I thank Lidewij van Gils and Irene de Jong for their useful comments on an earlier version of this paper, and Hannah Kousbroek for correcting my English. All Herodotus references are taken from Hude’s OCT 19273, but the statistics presented in the tables of part II are based on Legrand’s Budé text (1932–1954), which is available via the TLG (stephanus.tlg.uci.edu). For Diodorus, the Teubner edition by Fischer and Vogel was used (1888–19063).

© koninklijke brill nv, leiden, 2019 | doi:10.1163/9789004383340_004

a narratological comparison of herodotus and diodorus

55

low mound (κολωνός) within it (Hdt. 7.223.2–7.225). Diodorus’ version relates how the Spartans venture on a kamikaze-like night raid into the enemy camp and almost succeed in killing Xerxes before being surrounded and shot with arrows (Diod. Sic. 11.9.3–11.10.4).1 The similarities and differences between the two accounts have led to discussions about the course of events at Thermopylae.2 In tandem with this, the interrelationship of the two accounts has been debated, in which Diodorus’ use of Herodotus has even been denied.3 This, however, cannot be maintained, as Diodorus mentions the Histories specifically at the end of his account of the Persian Wars: Ὁ μὲν οὖν Μηδικὸς ὀνομασθεὶς πόλεμος γενόμενος διετὴς τοῦτο ἔσχε τὸ πέρας. τῶν δὲ συγγραφέων Ἡρόδοτος ἀρξάμενος πρὸ τῶν Τρωικῶν χρόνων γέγραφε κοινὰς σχεδόν τι τὰς τῆς οἰκουμένης πράξεις ἐν βίβλοις ἐννέα, καταστρέφει δὲ τὴν σύνταξιν εἰς τὴν περὶ Μυκάλην μάχην τοῖς Ἕλλησι πρὸς τοὺς Πέρσας καὶ Σηστοῦ πολιορκίαν. This, then, was the end of the so-called Median War, which had lasted two years. Among the authors [describing it] is Herodotos, who, beginning at a point before the Trojan War, composed in nine books an account of just about all public events that took place in the inhabited world, bringing his account to a close with the battle between the Greeks and the Persians at Mykale, and the siege of Sestos. 11.37.6, translation Peter Green

1 For an elaborate overview of the similarities and differences between the two accounts, see Albertz 2006: 81–92. 2 See Matthew 2013a: 23–24, with references to further scholarship on the historical aspects of the battle and, in this volume, Van Wees’ reconstruction of the events based upon a careful analysis and comparison of the existing Thermopylae traditions. In this debate one should always be extremely careful to distinguish between historiographical tradition and historical truth. As for the latter, Flower points out that ‘given that those Greeks who could have given the most authoritative account of the final struggle were all killed and that fifth- and fourth-century writers, such as Herodotus and Ephorus, viewed the events of the Persian Wars through the lens of Homeric epic, any attempt to reconstruct ‘what actually happened’ is inherently problematic’ (Flower 1998: 375, compare Albertz 2006: 33–39). To this one can add the general human tendency to exaggerate numbers and skills of one’s opponents in warfare. That Herodotus himself was aware of this can be deduced from his description of Xerxes’ manipulation of the number of Persian casualties at Thermopylae (Hdt. 8.24– 8.25). 3 Murray 1972: 210. Contra Flower 1998: 365–366. Observe that Diodorus mentions Herodotus

56

de bakker

This passage can be seen as a tribute to Herodotus, whose work Diodorus will no longer need now that he has finished his account of the Persian Wars and moves on to more recent times. Diodorus considers Herodotus his predecessor in writing universal history (note κοινὰς σχεδόν τι τὰς τῆς οἰκουμένης πράξεις, ‘almost all public events in the inhabited world’), and evidently consulted his work when he collected the material for his account of the Persian Wars. Within Diodorus’ Thermopylae narrative, traces of the Herodotean account (or, less likely, of a common source) can be found, in spite of the significant differences in content. At discourse level, Diodorus copies Herodotus’ περιχαρής ‘overjoyed’ to describe Xerxes’ reaction to Ephialtes’ offer to lead his men around the pass (Hdt. 7.215, Diod. Sic. 11.8.5). In the few cases where Herodotus uses this adjective, those who experience joy normally meet with disaster, and Diodorus appears to have copied it with the same connotation.4 Furthermore, Diodorus describes Xerxes’ aporia in similar terms as Herodotus (Diod. Sic. 11.8.4 ἀπορουμένου δὲ τοῦ βασιλέως ≈ Hdt. 7.213.1 ἀπορέοντος δὲ βασιλέος). We also find a typically Herodotean phrase in his version of Xerxes’ message to the Greeks: προσέταξε δ’ αὐτοῖς παραγγέλλειν, ὅτι βασιλεὺς Ξέρξης κελεύει τὰ μὲν ὅπλα πάντας ἀποθέσθαι … Diod. Sic. 11.5.4

And he ordered them to pass on the message that king Xerxes summoned all of them to lay down their arms … Xerxes orders his envoys to introduce his message with a reference to himself as the sender to emphasise his eminence. Such introductory self-references are familiar to Herodotus, who uses them in messages sent by Croesus (Hdt. 1.53.2; 1.69.1–1.69.2), Darius (Hdt. 5.24.1), Xerxes (Hdt. 7.150.2) and Mardonius (Hdt. 8.140α1).5 At a thematic and compositional level, there are overlaps between the two accounts, too. Both contain Laconic sayings (Hdt. 7.226.2, Diod. Sic. 11.4.3; 11.5.5; 11.6.2) and quotations of (Simonidean) poetry (Hdt. 7.228.1–7.228.3; Diod. Sic.

as a source for his Egyptian and early near Eastern material in books 1 and 2 (1.37.4; 1.37.11; 1.38.8; 1.69.7; 2.15.1; 2.32.2). Cf. Burton 1972: ad loc., Priestley 2014: 53. For a discussion of the differences between the two traditions, see Marincola 2007c: 115–117. 4 On the ominous character of περιχαρής in Herodotus, see Lateiner 1977 and Flory 1978. 5 Cf. De Bakker 2007: 52–53.

a narratological comparison of herodotus and diodorus

57

11.11.6). Both heroize the Spartans and Leonidas.6 Herodotus presents Leonidas’ genealogy all the way back to Heracles (Hdt. 7.204), and implicitly compares him to Hector in his choice to die gloriously once he is aware of the betrayal of his position (cf. Hom. Il. 22.297–22.305). When he falls, a fight unfolds over his body, which evokes similar scenes from epic poetry (Hdt. 7.225.1, compare the fights over the bodies of Sarpedon and Patroclus, resp. Hom. Il. 16.563–16.683 and 17.1–18.242). Diodorus’ version presents Leonidas as a hero like Achilles, who chooses Sparta’s glory above the lives of himself and his Spartiates (Diod. Sic. 11.4.4). Both versions award the Spartans their aristeiai. In Herodotus it is a defensive stand (compare the Greater Aias in the Iliad), whereas Diodorus makes them attack the Persian camp by night, which can be compared to Diomedes and Odysseus’s raid of the Trojan Camp in Iliad 10.7 Despite the evident similarities between the two versions, establishing their exact relationship remains difficult, as we do not know to what extent Diodorus followed other, no longer extant, sources when he wrote his account. We learn from his text that he knew of more than one Thermopylae tradition, as he refers to his sources in the plural (‘the writings of the historians and many of the poets’, Diod. Sic. 11.11.6). It is assumed that he took the account of the fourth century BCE historian Ephorus as his point of departure, but the view that he copied his sources verbatim without adding anything by himself is nowadays no longer maintained.8 Michael Flower, in his useful analysis of the sources of the Thermopylae tradition (1998), convincingly discredits some of the disparaging statements that have been made about Diodorus’ efforts. He suggests that encomiastic

6 For an overview of Homeric elements in Herodotus’ version of the battle of Thermopylae see Baragwanath 2008: 65 and n. 31 and Gainsford 2013. 7 Flower 1998: 374; Marincola 2007c: 117. 8 For Ephorus as Diodorus’ main source, see Hornblower 1994: 36–38; Flower 1998: 370; Marincola 2007b, 2007c: 110; Brown 2013: 110–113, Trundle 2013: 30 and Van Wees in this volume. The view that Diodorus slavishly copied his sources (e.g. Scherr 1933) and thus gives direct access to the lost works of historians like Ephorus and Timaeus has nowadays become more nuanced, e.g. by Priestley 2014: 14, 162–168, Gauger & Gauger 2015 and Muntz 2017, with pp. 14– 26 for an overview of the debate. For an attempt to prove Diodorus’ creative independence from his sources, see Sacks 1990, who argues that the Library reflects the intellectual attitudes of the late Hellenistic period (but see Fornara 1992 and Marincola 1997: 243, n. 134 for some critical observations), and Muntz 2017, who claims that particular thematic choices within its first three books reveal a late republican Rome context. Diodorus is nowadays considered an intermediary between Hellenistic historiography and the Second Sophistic (Wiater 2006 and Schmitz 2011: 244–245), and recent years have shown more scholarly output on Diodorus than before (a.o. Flower 1998; Clarke 1999; Green 2006, 2007; Marincola 2007b; Sulimani 2011).

58

de bakker

poetry could be the source of Leonidas’ suicide mission and the night raid on Xerxes’ camp—the most salient differences between the two versions. This poetry, written shortly after the Persian Wars, may have been comparable to the Simonides fragment quoted at the end of Diodorus’ account (Diod. Sic. 11.11.6). Ephorus, Flower argues, took over these elements with an eye to presenting the events in a new light, and through him they found their place in Diodorus’ Library as well as other authors.9 Flower’s theory about the influence of encomiastic poetry on Ephorus’ account can be tested when we compare different versions of the other battles of the Persian Wars. In Herodotus’ version of the battle of Salamis, the Greek galleys initially pull back in fear when the Persians come within view (Hdt. 8.84.1). In Aeschylus’ older version, however, the Greeks proudly confront the Persians in the narrows of Salamis (Aesch. Pers. 391–405).10 It is this older, encomiastic version that Diodorus prefers to Herodotus’ account (Diod. Sic. 11.18.2–11.19.3), and in the drafting of his Thermopylae narrative he may well have followed the same path.11 In this contribution, however, I will approach the question of the relationship between the two accounts, not by looking at the sources and their historical background (for this approach, see Van Wees in this volume), but instead by undertaking a comparative narratological analysis with an eye to laying bare the different views on the past that each historian seeks to bring across (§2). Following this I will investigate the type of discourse that Diodorus uses in those passages that seem to be innovations, based on a comparison with Herodotus’ older account (§3). With the help of these analyses, I will propose a hypothesis about the origins and backgrounds of these versions and their relationship (§4).

9 10

11

Flower 1998: 366–372. The other authors are Plutarch (Malice 866a) and Pompeius Trogus, for whom see Justin’s Epitome (2.11.12–2.11.18). Cf. Pelling 1997: 7–8, who warns against drawing conclusions from Aeschylus’ version as to the ‘historical truth’. In Plutarch’s version (Them. 14) Themistocles orders the Greeks to wait and not attack the Persians because a sea breeze rises that disrupts the Persian attack. Thus he follows Herodotus’ account in making the Greeks wait, but offers strategic reasons rather than fear as a motivation. Thus neither Diodorus’ account nor Ephorus’ lost version of the Persian Wars should be seen as a ‘deviation’ from a canonical version by Herodotus, as Marincola points out (2007c: 105–106).

a narratological comparison of herodotus and diodorus

2

59

Comparative Analysis

Below I present the results of a comparative narratological analysis between Herodotus’ and Diodorus’ versions of the events at Thermopylae in 480 BCE. The advantage of this approach in the case of historiographical texts is that it allows us to temporarily sidestep the need to look for sources that are, for the most part, ‘known unknowns’. Instead, the focus lies on the choices that the historiographers made themselves in presenting the material. Which material did they select? How much attention did they pay to it within their narrative? And how did they present it? These questions lie at the heart of any narratological research into ancient historiographical texts and may yield particularly interesting results when authors can be compared with one another.12 If they also treat the same topic, such comparisons become even more illuminating, as they reveal differences in choice and presentation that can be illustrated with statistic evidence. In the analysis below, I have singled out the narratological parameters that most clearly reflect the different outlooks of Herodotus and Diodorus upon the Thermopylae story. I will first look at the length of both accounts (2.1) and at the attention that each narrator pays to its various episodes (2.2). Next, I will look at the different ways in which the narrators manifest themselves and argue that each intends to present a version that is essentially polemic (2.3). I will end with some observations on the handling of time (2.4) and speeches and characterisation (2.5). 2.1 Length To begin with, a few words need to be said about the different lengths of the two versions, both in relation to one another, and to each historian’s account of the other great battles of Xerxes’ war. Both Thermopylae narratives are structured in a way that is typical of ancient Greek battle narratives. They include preliminary consultations, a catalogue of troops, an introduction of the leader(s) and a description of the battle itself. My analysis of the account begins with the conference of the Greek allies at the Isthmus (Hdt. 7.175.1; Diod. Sic. 11.4.1) and includes the aftermath of the battle, which Herodotus seals off with a series of anecdotes (Hdt. 7.226–7.239), and Diodorus with an encomiastic evaluation (Diod. Sic. 11.11.1–11.12.1). In Herodotus’ case, I exclude chapters 7.178 to 7.183 and 7.188 to 7.195 as they cover events related to Delphi and the Persian 12

For this comparative approach see the four volumes of Studies in Ancient Greek Narrative (De Jong, Nünlist & Bowie 2004; De Jong & Nünlist 2007; De Jong 2012; Van Emde Boas & De Temmerman 2017), each of which encourages narratological comparison between the authors that are selected for analysis.

60

de bakker

fleet that are not part of the Thermopylae story, but prepare the narratee for the ensuing episode about the battle of Artemisium. Thus counted, Diodorus’ version measures slightly less than two fifths of the length of Herodotus’ version.13 The brevity of Diodorus’ account compared to that of Herodotus has evident generic reasons. Diodorus writes universal history, and aspires to cover the entire known past from its origins to his own day. In doing so he aims to surpass earlier historians like Polybius and Posidonius, whose works are universal in scope, but lack overall inclusiveness.14 As Clarke points out, we learn from Diodorus’ work that he ‘aims at symmetry’ (στοχάζεσθαι τῆς συμμετρίας) in presenting past events. He applies a principle of due proportion, which enables him to keep control of a narrative that has a global scope.15 This argument of due proportioning is complicated, however, by the fact that Diodorus’ account of Thermopylae is still two to three times longer than his reports of the other battles of Xerxes’ war.16 His own evaluative comments give a clue as to why this is the case: δικαίως δ’ ἄν τις τούτους καὶ τῆς κοινῆς τῶν Ἑλλήνων ἐλευθερίας αἰτίους ἡγήσαιτο ἢ τοὺς ὕστερον ἐν ταῖς πρὸς Ξέρξην μάχαις νικήσαντας· τούτων γὰρ τῶν πράξεων μνημονεύοντες οἱ μὲν βάρβαροι κατεπλάγησαν, οἱ δὲ Ἕλληνες παρωξύνθησαν πρὸς τὴν ὁμοίαν ἀνδραγαθίαν. Moreover, anyone who argued that these men were also more responsible for achieving the freedom of the Greeks than the victors in subsequent battles against Xerxes would be in the right of it; for when the barbarians recalled their deeds, they were terror-struck, whereas the Greeks were encouraged to attempt similar acts of bravery. Diod. Sic. 11.11.5, translation Green

13 14

15

16

Herodotus 7.175–7.239 (with 7.178–7.183 and 7.188–7.195 excluded): 5637 words. Diodorus 11.4.1–11.12.1: 2190 words, so 38.9 % of the length of Herodotus’ account. Marincola 1997: 241–244; 1999: 312. For an overview of Diodorus’ most important predecessors, see Green 2006: 13–25. Diodorus’ contemporaries Strabo and Pompeius Trogus also aimed at overall inclusiveness (Clarke 1999: 255). Clarke 1999: 265, cf. Rubincam 1987 and Most 2011: 171. For the formula, see Diod. Sic. 1.9.1; 1.9.4; 1.29.6; 1.41.10; 4.5.4; 4.68.6; 6.1.3. Diodorus follows a principle set out by Polybius (29.12). Measured by page numbers in the Teubner edition of Vogel-Fischer: Thermopylae (Diod. Sic. 11.4.1–11.12.1): 12.5 pages; Artemisium (Diod. Sic. 11.12.2–11.13.5): 3.5 pages; Salamis (Diod. Sic. 11.15.2–11.19.6): 6 pages; Plataeae (Diod. Sic. 11.29.1–11.33.4): 6.5 pages; Mycale (Diod. Sic. 11.34.1–11.36.7): 4 pages.

a narratological comparison of herodotus and diodorus

61

In the tradition that Diodorus uses for his universal history, Thermopylae is believed to be—or has been framed as—the battle that changed the outlook of the clashing parties decisively, as the Greeks witnessed that the Persians could be defeated, and the Persians were terrified by the fighting power of the Greeks.17 This tradition awarded pride of place among the battles of the Persian Wars to Thermopylae, which accounts for the disproportionate attention Diodorus paid to it. This is different in Herodotus, who subscribes to the view that the Athenians were responsible for the survival of Greek freedom, as they refused, against all odds, to surrender to Xerxes or to migrate elsewhere (Hdt. 7.139.1–7.139.6). For him, the decisive battles are those of Salamis and Plataeae, whereas Thermopylae mainly deserves attention for its exemplary Spartan heroism.18 2.2 Proportion of Episodes The next question for comparison concerns the specific attention that is paid to the episodes of the Thermopylae story. As I indicated in the introduction, both versions are more or less similar in terms of structural content. This similarity allows us to divide the story into roughly similar episodes and compare their lengths to one another. In table (3.1) I give the references of each episode in both works. In table (3.2) an overview is given of the number of words that each historian spends on each episode, in both absolute and relative numbers. The figures in table (3.2) reveal a profoundly different emphasis in the works of the two historians. Part of this can be ascribed to Diodorus’ activities as a compiler. Herodotus’ account of Xerxes’ advance through Thessaly, Achaia and Malia (Hdt. 7.196–7.201, part of episode nr. 4), for instance, is absent in Diodorus, who summarises the king’s march from the Hellespont to the Malian Gulf in a few sentences. Herodotus narrates this episode at a more leisurely pace, which adds suspense and allows him to insert thematically relevant information. Donning the robe of a geographer, he makes his narratees follow the route of Xerxes and his army, pointing out villages, rivers, sanctuaries, and note17

18

Observe however 11.16.1–11.16.3 shortly afterwards, where the Greeks are described as ‘petrified’ when they face the Persian fleet at Salamis (Green 2006: 63, note 46, with reference to a private comment by Marincola). This appears to be one of the many inconsistencies in Diodorus’ Library, resulting either from carelessness or a lacking final round of revision (Green 2006: 29–30; 2007: 367). This is reflected in the lenghts of their narratives compared to the other battles. Measured by numbers of pages in the OCT edition of Hude: Salamis (Hdt. 8.40–8.99): 27 pages; Plataeae (Hdt. 9.19–9.89): 37 pages. Compare Thermopylae (Hdt. 7.175–7.177; 7.184–7.187; 7.196–7.239): 24.5 pages; Artemisium (Hdt. 8.1–8.26): 11.5 pages; Mycale (Hdt. 9.90–9.107): 10 pages.

62 table 3.1

de bakker Division of episodes of the Thermopylae narrative per author

Nr. of epis.

Description of episode

Herodotus

Diodorus

1 2 3

Conference at the Isthmus & introduction Introduction Greek leaders and Leonidas Catalogue of Greek troops

11.4.1 11.4.2–11.4.4 11.4.5–11.4.7

4

Xerxes’ advance, catalogue of his troops

5 6 7 8 9 10 11

Preparations and Demaratus’ advice First day of the battle Second day of the battle The betrayal of Ephialtes Third day of the battle: awareness and assembly Third day of the battle: the fighting Aftermath of the battle

7.175–7.177 7.204–7.205.1 7.202–7.203; 7.205.2–7.206 7.184–7.187; 7.196–7.201 7.207–7.209 7.210–7.212.1 7.212.1–7.212.2 7.213–7.217.1 7.217–7.222 7.223–7.225 7.226–7.239

table 3.2

Nr. of epis.

1 2 3 4 5 6 7 8 9 10 11 total

11.5.1–11.5.3 11.5.4–11.6.2 11.6.3–11.7.4 11.8.1–11.8.3 11.8.4–11.8.5 11.9.1–11.9.2 11.9.3–11.10.4 11.11.1–11.12.1

Number of words used per episode in Herodotus’ and Diodorus’ Thermopylae narratives

Description of episode

Conference at the Isthmus & introduction Introduction Greek leaders and Leonidas Catalogue of Greek troops Xerxes’ advance, catalogue of his troops Preparations and Demaratus’ advice First day of the battle Second day of the battle The betrayal of Ephialtes Third day of the battle: awareness and assembly Third day of the battle: the fighting Aftermath of the battle

Herodotus

Diodorus

Nr

%

Nr

%

418 125 331 1220 394 242 65 363 581 409 1489 5637

7.4 2.2 5.9 21.6 7.0 4.3 1.2 6.4 10.3 7.3 26.4 100

71 145 129 154 215 308 148 111 136 349 424 2190

3.2 6.6 5.9 7.0 9.8 14.1 6.8 5.1 6.2 15.9 19.4 100

a narratological comparison of herodotus and diodorus

63

worthy items in the landscape.19 Through this manner of narrating, he makes the landscape appear as it gradually unfolds before Xerxes’ eyes.20 As Thessaly, Achaia and Malia have given earth and water to the king (Hdt. 7.132.1), we share his perspective as he pays his first visit to these recent assets to his empire, and understand his curiosity and respect of local traditions, manifest in his avoidance of the property of Athamas and the honours on behalf of Zeus Laphystius (Hdt. 7.197.4). Meanwhile, the water given to the king is literally drunk by his army, which empties several rivers on its march towards the pass or uses their riverbeds to find its direction (Hdt. 7.196; 7.199–7.200). Furthermore, Herodotus subtly lowers a mythological backdrop behind his description with references to Athamas (Hdt. 7.197.1–7.197.3) and the burning Heracles, who was helped by the river Dyras (Hdt. 7.198.2). Both legendary heroes exemplify the typically Herodotean theme of ‘look to the end’, as they suffer a gruesome fate at the end of their careers. A thematic connection seems implied with Xerxes, who is king of the largest empire that ever existed, but whose aspirations will soon be dented by the Greeks, and with Leonidas and his Spartans, who will end their lives in heroic fashion in the battle about to begin. Not all differences in proportion between the episodes can, however, be explained by Diodorus’ compiling efforts. Diodorus pays considerably more attention to the actual fighting episodes (nrs. 6, 7 and 10), which, if added up, make up more than one third of his narrative (36,8 %).21 Herodotus pays less attention to the fighting (12,8%) and gives precedence to the diplomacy that surrounds the battles22 (observe that Diodorus spends fewer words on episodes 1, 8, 9). It is in the fighting episodes, then, that we encounter the most obvious innovation that Diodorus (or his sources) appears to have made to the existing Thermopylae tradition. Below (§§3–4) I will return to this subject and attempt an explanation by analysing Diodorus’ type of discourse in these passages.

19

20

21 22

Villages: Halus (7.197.1), Anticyra (7.198.2), Trachis (7.199), Anthele (7.200.2); rivers: Onochonus (7.196), Epidanus (7.196), Spercheius (7.198.2), Dyras (7.198.2), Melas (7.198.2), Asopus (7.199), Phoenix (7.200.1); sanctuaries: temple of Zeus Laphystius (7.197.1), temple of Demeter Amphictyonis (7.200.2), temple of Amphictyon (7.200.2); items in the landscape: plain around Malian Gulf (7.198.1; 7.199), Eniana valley (7.198.2), Thermopylae (7.200.1). For this typical Herodotean ‘hodological’ way of describing space, see Purves 2010: 118–158. Below I discuss the effects of this manner of narrating on Herodotus’ self-presentation. See Albertz 2006: 40: ‘Die topographische Beschreibung ist … mit den Bewegungen der historischen Akteure verbunden, und die Landschaft zwischen Gebirge und Meer wird dadurch den Lesern sukzessive erschlossen.’ Cf. Albertz 2006: 86. Cf. Immerwahr 1966: 69, 285–305, who points out that for Herodotus the sections that precede and follow the action are usually more important than the battle itself.

64

de bakker

2.3

The Voice of the Narrator: Polemics and Historiographical Authentication Turning to the voice of the narrator, I will now compare the ways in which Herodotus and Diodorus authenticate their texts so as to persuade the narratee of their factual correctness. It is well known that ancient historiographers not only narrate, but also respond, often critically, to the work of colleagues, both predecessors and contemporaries, whose views they seek to contest.23 In consequence, the historiographical narrator often manifests himself more openly in his text. In furnishing comments on the events and the traditions that describe them, he points at an ambition not only to preserve great events from forgetfulness, but also to persuade his narratee of the historical correctness of his presentation. The Herodotean narrator, who, in de Jong’s words, ‘poses as an epideictic speaker’24 poignantly surfaces in the Thermopylae narrative, to begin with in the evaluation of the various versions that have come to the historian’s ear.25 Thus he rejects the alternative versions that lay the blame for the betrayal of the Greeks not on Ephialtes, but on Onetes and Corydallus (οὐδαμῶς ἔμοιγε πιστός ‘in no way credible to me’, Hdt. 7.214.1). In a similar, overt way, he defends the version in which Leonidas himself decided to send away the other Greeks so as to frame the heroic death of himself and his fellow Spartiates (ταύτῃ … τὴν γνώμην πλεῖστός εἰμι, ‘I am thoroughly convinced about this’ Hdt. 7.220.2). As evidence (μαρτύριον, Hdt. 7.221), he refers to the seer Megistias, who disobeyed Leonidas’ order to leave, but whose son survived and—we may assume— reported the event. Other instances are the traditions about the lone survivor Aristodemus, where Herodotus mentions a different scenario had both he and his comrade Eurytus followed the same path (Hdt. 7.229.2), and the case of Xerxes’ decapitation of Leonidas’ corpse, an act so contrary to Persian customs that Herodotus considers it an indication (τεκμήριον, cf. Hdt. 7.238.2) that the king ‘felt the gravest anger against Leonidas of all men when he was alive’ (πάντων δὴ μάλιστα ἀνδρῶν ἐθυμώθη ζῶντι Λεωνίδῃ, Hdt. 7.238.2).

23 24

25

See Marincola 1997: 217–257, with ample references to older literature. De Jong 2004: 107 and compare De Jong 1999: 227–229. Other publications that discuss this aspect of the Herodotean narrator are Darbo-Peschanski 1987; Dewald 1987; Marincola 1987 and Branscome 2013. Herodotus’ critical engagement here may be related to an intensive contemporary debate about the course of affairs during the Persian Wars. Herodotus seems to approach the contested issues with an empirical methodology that is derived from the Hippocratic school of medicine, for which see Müller 1981; Thomas 2000 and Hollmann 2011.

a narratological comparison of herodotus and diodorus

65

Next to these overt evaluations, the historian tries to authenticate his narrative by creating an impression of accuracy, for instance in the calculations of Xerxes’ troops (ὡς ἐγὼ συμβαλλόμενος εὑρίσκω ‘as I conclude from my calculations’, Hdt. 7.184.1, cf. Hdt. 7.187.2 εὑρίσκω … συμβαλλόμενος ‘I conclude by calculating’) and in his confident assurance that he knows the names of all three hundred Spartiates (7.224.1, ἐπυθόμην τὰ οὐνόματα, ἐπυθόμην δὲ καὶ ἁπάντων τῶν τριηκοσίων ‘I learnt their names, I learnt them even of all the three hundred’). This authentication strategy also appears in the many details that Herodotus adds to his narrative, such as the width of the passes in Trachis (ἡμίπλεθρον ‘half a plethrum’, Hdt. 7.176.2) and Thermopylae (ἁμαξιτός ‘a car’s length’, Hdt. 7.176.2) and the distance between the mouth of the river Spercheius and the river Dyras (twenty stades, Hdt. 7.198.2).26 He also gives a specific number of days (four, Hdt. 7.210.1) that Xerxes allows to pass for the Greeks to retreat. Furthermore, he refers to some highborn Achaemenids among the casualties, and displays his expert knowledge of the dynasty by indicating their relationships with Xerxes (Hdt. 7.224.2).27 Another aspect is the variant versions included in the Thermopylae narrative, which conform to his self-imposed duty to report what is told, even if he does not attach belief to it (cf. Hdt. 7.152.3). Following this principle, he refers to the alternative traditions that seek to blame Onetes and Corydallus for the betrayal of the Greeks (Hdt. 7.214.1), and that make Leonidas, and not the non-Spartan allies themselves, responsible for their withdrawal from the battlefield (Hdt. 7.219.2– 7.220). In the post-battle anecdotes, Herodotus refers to alternative versions of Aristodemus’ survival (Hdt. 7.229–7.230), and of the motivations for Demaratus to warn the Spartans about Xerxes’ plans to invade Greece (Hdt. 7.239.2).28 Akin to this use of variant versions are φασι/λέγεται-statements, which Herodotus uses either to give an impression of completeness in his report (for instance by referring to the tradition about Pantites, another Spartiate survivor, Hdt. 7.232), or to frame noteworthy events, such as the memorable words spoken by Dienekes:

26

27

28

See furthermore Hdt. 7.198.2 (distance of twenty stades between river Dyras and river Melas); 7.199 (distance of five stades between river Melas and Trachis); 7.199 (width of Trachinian plain, 22,000 plethrums); 7.200.1 (fifteen stades from Phoenix river to Thermopylae). Observe that although Herodotus claims to know the names of all the Spartiates who died at Thermopylae (Hdt. 7.224.1) the actual names that he records here are those of high-born Persians. For the device of variant versions in Herodotus, see Groten 1963.

66

de bakker

τὸν τόδε φασὶ εἰπεῖν τὸ ἔπος … Hdt. 7.226.1

Of him, they claim that he made the following statement … ταῦτα μὲν καὶ ἄλλα τοιουτότροπα ἔπεά φασι Διηνέκεα τὸν Λακεδαιμόνιον λιπέσθαι μνημόσυνα. Hdt. 7.226.2

This and other sayings of such kind they claim that Dienekes the Spartan left as memorials. The use of these φασι/λέγεται-statements need not necessarily indicate that Herodotus finds these traditions less trustworthy or wants to distance himself from their content. Rather, given that Dienekes’ sayings are so remarkable for their typically Laconic fearlessness, Herodotus explicitly earmarks them as belonging to the traditions that came to his ear, and in mentioning ‘one of the Trachinians’ (τευ τῶν Τρηχινίων, Hdt. 7.226.1) as Dienekes’ partner in conversation, implicitly accounts for their survival. Given that almost all Spartans died in the pass, their record could only survive via a non-Spartan intermediary. Thus in mentioning the Trachinian, Herodotus anticipates possible questions about his historical method.29 Herodotus’ concern with the authentication of his narrative becomes particularly clear when we compare the number of proper names, toponyms and ethnonyms, with those that are recorded in Diodorus’ version. In the table below, both the number of different names in each category is given (each item is recorded once only, even when it appears more than once) as well as the total number of names (all names are counted as they appear in the narrative). Furthermore, the number of names that are unique within the entire work is given. These statistics show that Herodotus’ version contains a significantly larger number of different names compared to the account in Diodorus (while the latter is approximately two fifths of the length of Herodotus’ version, it contains only one fifth as many different names). This holds especially for proper names, of which Herodotus mentions forty seven, including thirteen that are not found elsewhere in the Histories, for instance Dithyrambus son of Harmatides (Hdt. 7.227), the greatest fighter of the Thespians. Diodorus refers to 29

Similar λέγεται-statements accompany the remarkable anecdotes of Xerxes jumping from his throne during the battle (7.212.1) and Demaratus’ secret letter to the Spartans, which was discovered by Gorgo, Leonidas wife (7.239.4). Cf. De Bakker 2007: 173–177.

a narratological comparison of herodotus and diodorus table 3.3

67

Number of proper names, toponyms and ethnonyms in Herodotus’ and Diodorus’ Thermopylae narratives. (See Appendix 1 for a full overview.)

Category

Hdt.

Diod. Sic.

Difference in percentage

total number of words

5637

2190

38.9%

number of different proper names number of different toponyms number of different ethnonyms total number of different names

47a 52 47 146

6 12 15 33

12.8% 23.1% 31.9% 22.6%

total number of proper names total number of toponyms total number of ethnonyms total number of names

151a 145 191 487

37 28 79 144

24.5% 19.3% 41.4% 30.0%

proper names unique in work toponyms unique in work ethnonyms unique in work total names unique in work

13 7 3 23

2 2 1 5

15.4% 28.6% 33.3% 21.7%

a This includes names of gods and heroes and excludes patronymics if the father does not play a role in the Thermopylae narrative. Of the lengthy twenty-itemed genealogy of Leonidas (7.204), only the names of Leonidas’ father Anaxandrides (also mentioned in Hdt. 7.205.1) and of Heracles are included.

six proper names in his narrative, of whom only one, Tyrrhastiadas from Cyme (Diod. Sic. 11.8.5) is otherwise unknown and unique in his work.30 We spot a similar tendency in the other categories. Whereas the two accounts differ relatively slightly in the total number of names used (with Diodorus mentioning 30% of Herodotus’ names, while his account is 38.9 % the length of that of Herodotus), Herodotus’ text displays a much wider variation of names and a greater attention to detail and accuracy in presenting names. Thus, Herodotus specifies the city of Orchomenos as ‘Arcadian’ (Hdt. 7.202) to distinguish it from Boeotian Orchomenos, and the Locrians as ‘Opuntians’ (Hdt. 7.203.1), to distinguish them from the Epizephyrian Locrians. Furthermore, Herodotus includes 30

The other unique name in Diodorus’ version is the poet Simonides, but he was well known. Tyrrhastiadas’ origin from Cyme points to the Cymaean Ephorus as Diodorus’ source. See Flower 1998: 368 and n. 22 for references to earlier scholars who observed this.

68

de bakker

local toponyms, such as Chutri (the hot springs in the pass, Hdt. 7.176.3), and Pylae (for Thermopylae, Hdt. 7.201), as well as local traditions such as the myth of Athamas told to Xerxes ‘by the locals’ (Hdt. 7.197.1–7.197.3). Such specific or local information is absent in Diodorus’ account, who also leaves out patronymics (not included in table 3.3), and genealogical data of the kind that Herodotus inserts in the cases of Leonidas (Hdt. 7.204) and the fallen Achaemenids (Hdt. 7.224.2). The only category in which Diodorus surpasses Herodotus is that of ethnonyms (41.4%). The exception can be explained as resulting from the large number of references to ‘Greeks’ (30) and ‘Persians’ (17) in Diodorus’ account, who are often placed in opposition to one another within the same sentence.31 This approach points at a crucial difference in outlook between the two texts: whereas Herodotus emphasises the diversity of the Greeks, and mentions the various participating Peloponnesian and Boeotian troops individually, Diodorus underlines their unity as ‘Greeks’ in their war against ‘the Persians’. His version of the battle of Thermopylae is written from a panhellenic perspective, whereas Herodotus’ version is written against the background of Greek interstate war, and accordingly highlights their quarrels and internecine rivalry.32 It is exactly in this respect that Diodorus’ version shows traces of a polemic against Herodotus’ account (and possibly other versions). Diodorus presents the Thebans who fight alongside the Greek allies at Thermopylae as representatives of the anti-Persian party who supported the Greek alliance (Diod. Sic. 11.4.7). In Herodotus’ version, Leonidas suspects the Thebans of medism and treats them as hostages (Hdt. 7.205; 7.222). Their behaviour in the battle confirms their disingenuous attitude (Hdt. 7.233). Herodotus also presents the threat of medism among the non-Peloponnesian Greeks as a motivation for the Spartans to send out their troops to Thermopylae (7.206.1). In Diodorus’ version, the decision is taken for the purpose of protection of the anti-Persian alliance (11.4.1). Furthermore, Diodorus omits references to debate and controversy within the Greek alliance. He does not refer, for instance, to the promises made to the Locrians (see Hdt. 7.203.1), which turn out to be empty when the 31 32

Excluded from these numbers are references to the Persians as οἱ βάρβαροι (13 times). If included, the total number of ethnonyms in Diodorus would even be larger. As observed by e.g. Immerwahr 1966: 273–277, 286–287 and Albertz 2006: 90–91. Van Wees in this volume takes a different view, arguing that Herodotus in paying attention to the allies who fought at Thermopylae tacitly responded to an older version that sought to ascribe the heroic achievements at Thermopylae exclusively to Sparta. Unlike Diodorus, however, Herodotus does not give credit to the Greeks as a whole, but thematises their rivalry and lack of mutual trust in his Thermopylae narrative (as he does elsewhere in his account of the Persian Wars).

a narratological comparison of herodotus and diodorus

69

allies become fearful and urge Leonidas to withdraw, much to the Locrians’ anger (Hdt. 7.207). In Diodorus’ version the Locrians desert the Persians out of free will and act in panhellenic interests (Diod. Sic. 11.4.6). Diodorus also does not refer to the fear of the Greeks at Thermopylae and the conference in which they discuss the option of retreating (see Hdt. 7.207). Absent, too, are the references to Greek internecine fighting that Herodotus includes in his narrative, such as the legendary war between Thessalians and Phocians in which Thermopylae had been a battle ground too (Hdt. 7.176.4; 7.215), and the assault of Plataeae by the Thebans that marked the beginning of the Peloponnesian War (Hdt. 7.233.2, cf. Thuc. 2.2–2.6).33 Furthermore, Diodorus does not pay attention to the Spartiate survivors. In his version, the Spartans encourage Leonidas ‘with one voice’ (μιᾷ φωνῇ, Diod. Sic. 11.9.3) and choose to fight ‘with united determination’ (μιᾷ γνώμῃ, Diod. Sic. 11.11.1). On the whole, Diodorus’ version is more encomiastic, and irons out the less praiseworthy aspects of Greek behaviour at Thermopylae that Herodotus mentions. The difference in outlook between the two accounts even surfaces in the poetry that both authors quote in praise of the fallen. Whereas Herodotus’ first epigram praises all allied Peloponnesians who fought in the battle (Hdt. 7.228.1), the second the fallen Spartans only (Hdt. 7.228.2) and the third the individual Megistias (Hdt. 7.228.3), Diodorus quotes a Simonides poem that praises the heroism of all Greeks together, united under their leader Leonidas (Diod. Sic. 11.11.6). Like Herodotus, Diodorus uses evaluative comments in his narrative (or derives them from his sources), which aim at bolstering his authority and lending credibility to the version that he presents. Although the large number of metahistoriographical comments that are found in Herodotus’ version are absent, he uses φασι/λέγεται-statements in referring to the traditions about the emptying of rivers by Xerxes’ army (Diod. Sic. 11.5.3) and the reply that Demaratus gives to Xerxes (Diod. Sic. 11.6.2), and gives alternative motivations for Xerxes to send the Medes into battle first: εἴτε δι’ ἀνδρείαν προκρίνας αὐτοὺς εἴτε καὶ βουλόμενος ἅπαντας ἀπολέσαι· Diod. Sic. 11.6.3

either preferring them for their courage or because he wanted to kill them all.

33

Observe the reference here to the Thebans as ‘400’ (Hdt. 7.233.2)—an ironic echo of the Spartan ‘300’? The reference to Cythera in Demaratus’ speech (Hdt. 7.235.2) may also fit within this theme (see De Bakker and Van der Keur in this volume).

70

de bakker

The presentation of alternative motivations is a common strategy in Herodotus, who often uses the second alternative to hint at darker, less noble aspects of his characters.34 Diodorus (or his sources) appear to have used the exact same strategy here. 2.4 The Handling of Time A few words should be dedicated to the handling of time in both accounts.35 Whereas each version presents the events of the battle in a relatively straightforward chronology, Herodotus’ more extensive report contains on average far more anachronies (i.e., more references backwards—analepseis—and forwards—prolepseis—in time). This holds for both internal and external anachronies (the former referring to earlier and later events within the Thermopylae story, and the latter to events that took place before it or happened later). These figures again highlight the complexity of Herodotus’ narrative compared to Diodorus.36 In contrast to Diodorus, Herodotus adds analepseis about the legendary past to his narrative, as well as at least one prolepsis about a contemporary event, the Theban assault on Plataeae in 431 BCE.37 Both types of anachrony are absent in Diodorus’ account, whose shorter narrative focuses more on the actual fighting (see above). For Herodotus, anachronies allow him to use the aforementioned strategy of lowering backdrop of Greek internecine discord behind the narrative of the Persian Wars. References to warfare among the Greek states in both legendary and contemporary times are found in his Thermopylae narrative, whereas Diodorus (or his sources) chose to omit this aspect in his account of the Persian Wars.

34 35

36

37

For examples of this narrative strategy in the case of Leonidas’ character, see Baragwanath 2008: 64–78. For an in-depth analysis of Herodotus’ handling of time in the Thermopylae passage, see De Jong in this volume. For an overview of Herodotus’ handling of time in general, see Rood 2007. In relative terms, the difference (with Diodorus’ number of anachronies amounting to 23.1 % of the number found in Herodotus) is larger than the relative difference in length (38.9 %). See Appendix 2 for an overview. See Hdt. 7.233.2 (Eurymachus killed during the Theban assault on Plataeae, cf. Thuc. 2.2– 2.6). External analepseis referring to the legendary past: Hdt. 7.176.4–7.176.5 (wall of the Phocians); Hdt. 7.197.1–7.197.3 (Athamas legend); Hdt. 7.198.2 (Heracles relieved by the river Dyras); Hdt. 7.215 (the discovery of the Anopaea path by the Malians during the legendary war between Thessaly and Phocis).

a narratological comparison of herodotus and diodorus table 3.4

Author

Anachronies in Herodotus’ and Diodorus’ accounts of Thermopylae

Analepseis

Prolepseis

Internal External Internal External Herodotus Diodorus

71

11 1

18 6

2 0

8 2

Total 39 9 (23.1 %)

2.5 Speeches and Characterisation The last narratological categories to discuss are those of speeches and characterisation. The ascription of speeches to historical characters enlivens the narrative, as the voice of the narrator temporarily fades into the background and other opinions are brought to the fore. Speeches also give relief to the narrative, since the information brought forward by the narrator is discussed or looked upon from a different, usually more limited, perspective. In this respect, speeches play an important role in explaining how events came about, especially in the early historians Herodotus and Thucydides, who usually refrain from explicit observations on historical causation, and prefer to show the way in which history is ‘played out’ by staging their characters in conversations and debates.38 In doing so, they empower their readers, who are invited to reflect upon the complexity of historical events.39 This attitude to the insertion of speeches changes in the Hellenistic Era. Polybius criticizes the invented speeches of earlier historians, singling out Timaeus for venomous criticism (Polyb. 12.25–12.28).40 He advocates the use of speeches only when their content can be confirmed in line with the truth (Polyb. 29.12; 36.1). Diodorus reflects on the incorporation of speeches in similar terms in the preface to his twentieth book, though like Polybius, he admits that he should not entirely ban them from his work, considering their use appropriate when they accompany great deeds, or help in explaining paradoxical events (Diod. Sic. 20.1–20.2).41

38

39 40 41

For Herodotus, see e.g. Waters 1966; Pelling 2006; De Bakker 2007 and Scardino 2007. For Thucydides, see e.g. Pelling 2000. For a recent volume on speeches in ancient historiography see Pausch 2010. See Baragwanath 2008: 22–26. Green 2006: 18–19. See also Marincola 2007a: 129, who points out that Diodorus, in treating the same period as his predecessors, might have avoided their speeches.

72

de bakker

table 3.5

Overview of speeches in Herodotus’ and Diodorus’ accounts of Thermopylae. (See Appendix 3 for an overview. The citations of poetry in Hdt. 7.228 and Diod. Sic. 11.11.6 are not included.)

Type

Herodotus

or. recta or. obliqua total

Word count

Number

Word count

Number

Word count

9 9 18

686 327 1013

2 10 12

51 256 307

22% 111% 66.7%

7.4% 78.3% 30.3%

Amount of oratio recta and oratio obliqua compared to overall narrative

Total nr. of words

Herodotus Diodorus

Relative difference

Number

table 3.6

Author

Diodorus

Total nr. of words in oratio recta

Total nr. of words in oratio obliqua

Total nr. of words in speech

Abs.

Relat.

Abs.

Relat.

Abs.

Relat.

Abs.

Relat.

5637 2190

100% 100%

686 51

12.2% 2.3%

327 256

5.8% 11.7%

1013 307

18.0% 14.0%

This change in attitude to the incorporation of speeches is visible in the number and word count of the speeches in oratio recta that both historians award to their characters in their Thermopylae narratives. Whereas Herodotus presents nine speeches in oratio recta, which together make up 12.2 % of his narrative,42 Diodorus presents only the conversation between Xerxes and Demaratus (Diod. Sic. 11.6.1–11.6.2) in oratio recta (2.3% of his narrative). In all other instances, as tables 3.5 and 3.6 indicate, Diodorus prefers the use of oratio obliqua to present the communication between his characters. The oratio obliqua nonetheless allows him to add an individual flavour, for instance in the case of Leonidas’ Laconic statement to the Ephors (Diod. Sic. 11.4.3) and the message that Xerxes sends to the Greeks (Diod. Sic. 11.5.4, see above). Apart from differences in number and method of presentation, the speeches in both accounts are also used for different purposes of characterisation. In 42

The figure is roughly similar to the overall percentage of speeches in oratio recta in de Histories, which is 14.5 % (De Bakker 2007: 28).

a narratological comparison of herodotus and diodorus

73

selecting the conversation between Xerxes and Demaratus as the only passage to present in oratio recta, Diodorus encourages reflection upon the character of Xerxes, who ridicules Demaratus’ warning that the Greeks will fight resiliently on behalf of their freedom (Diod. Sic. 11.6.2). Diodorus appears to have derived Xerxes’ response from a conversation between the king and Demaratus that Herodotus presents when the king has crossed the Hellespont (Hdt. 7.101–105). Here, Xerxes laughs about Demaratus’ warnings that the Spartans will anyway engage the Persian army, regardless of their numerical inferiority. Diodorus (or one of his sources) has transferred the gist of this famous conversation to the narrative of Thermopylae and changed Xerxes’ apparently friendly laughter in response to Demaratus’ advice (Hdt. 7.103.1: γελάσας; 7.105: ἐς γέλωτά ἔτρεψε, and observe the statement in 7.105 that Xerxes was not angered at all by Demaratus’ replies) into ridicule of Demaratus’ person (καταγελάσας αὐτοῦ, Diod. Sic. 11.6.2). This befits Diodorus’ characterisation of Xerxes in general. We find the same traits as in Herodotus, but Diodorus has magnified them. Thus, in Diodorus’ version, Xerxes threatens his special forces with death if they fail to force a breakthrough (Diod. Sic. 11.8.1), whereas in Herodotus he only uses the whip to force his troops to fight (Hdt. 7.223.3). Such a magnifying strategy is also visible in the characterisation of Leonidas. In Diodorus’ version, Leonidas is presented as a hero tout court and receives full praise for his achievements. This is different in Herodotus, who, as Baragwanath has argued, presents us with an ambiguous Leonidas, whose heroic characteristics are juxtaposed with more mundane traits, for instance in his attitude to the Thebans, whom he suspects of having pro-Persian sympathies (Hdt. 7.205.3).43 Furthermore, in Herodotus’ version, the last, heroic fight is not set in motion by Leonidas himself, but the result of Greek deliberations in which the anger of the Phocians and Locrians about the imminent departure of the allied Greeks persuades Leonidas to stay (Hdt. 7.207). Diodorus appears to characterise Xerxes and Leonidas more in black-and-white terms, with Leonidas encomiastically described as the great hero, and Xerxes as the villain. Herodotus, on the other hand, allows for more shades of subtlety in the characterisation of his protagonists, and singles out more participants in the battle for praise (Dienekes, Megistias, the Thespians). His more complex strategy of characterisation also surfaces in the case of Xerxes, whose vulnerable position as king is highlighted when his relative Achaemenes steps in to urge him not to adopt Demaratus’ advice to send his fleet to Cythera (Hdt. 7.236.1– 7.236.3). Throughout the narrative of Xerxes’ invasion, Herodotus shows the king surrounded by advisors with different interests, who often give contrary advice and thereby undermine his royal authority. Here, Achaemenes advo43

Baragwanath 2008: 64–78.

74

de bakker

cates against a strategy that could have worked out well. It involves the splitting up of the land army and the fleet, as the Greeks were forced to do in Salamis, where it did not hamper their success. Throughout his narrative, Herodotus highlights the odd role of Xerxes’ advisors to explain why his expedition foundered. 2.6 Summary The comparison above, which focused on five narratological parameters, shows Herodotus’ account to be richer and more complex in terms of not only length, but also digressions, anachronies, speeches and characterisation. These differences can be explained by Diodorus’ compiling efforts and his adaptation of the material to the tastes of his time, which we see for instance in his reluctance to include speeches in oratio recta. Furthermore, Diodorus’ version pays more attention to the fighting episodes, and has a consistent panhellenic outlook, as opposed to Herodotus, who lowers a backdrop of Greek internecine discord behind his Thermopylae narrative. It is in these latter aspects that Diodorus’ text seems to innovate the tradition and critically respond to Herodotus’ version. To further explore their relationship, a closer look at these aspects is needed.

3

Diodorus’ Discourse in the Fighting Episodes

To begin with, the question arises whether it is possible to learn more about the background of Diodorus’ fighting episodes, which is the first salient difference between the two existing versions. To answer this question, it is instructive to look at the actual discourse, and try to identify the origins of the expressions that are used. By way of example I look at a passage that describes the disadvantages of Persian weaponry in a confined space: ἀσπίσι γὰρ καὶ πέλταις μικραῖς οἱ βάρβαροι χρώμενοι κατὰ μὲν τὰς εὐρυχωρίας ἐπλεονέκτουν, εὐκίνητοι γινόμενοι, κατὰ δὲ τὰς στενοχωρίας τοὺς μὲν πολεμίους οὐκ εὐχερῶς ἐτίτρωσκον, συμπεφραγμένους καὶ μεγάλαις ἀσπίσι σκεπαζομένους ὅλον τὸ σῶμα, αὐτοὶ δὲ διὰ τὰς κουφότητας τῶν σκεπαστηρίων ὅπλων ἐλαττούμενοι πυκνοῖς τραύμασι περιέπιπτον. Diod. Sic. 11.7.3

For the barbarians employed small shields and targets, which gave them an advantage in open terrain, as they allowed them to move easily. On a narrow front, however, they found it difficult to wound their enemies who

a narratological comparison of herodotus and diodorus

75

stood close-packed side by side, their huge shields protecting their entire bodies, while they themselves, at a disadvantage because of the lightness of their protective armour, suffered countless wounds. translation Green, adapted

Some expressions in this sentence are found in Herodotus’ Histories, but the meaning in which Diodorus uses them appears to be post-Herodotean. Thus the word πέλτη originally indicates a light, Thracian shield and is found in this specific meaning in Herodotus (Hdt. 7.75.1). In this passage, however, it refers to a light shield in general (carried by a πελταστής), without a specific Thracian connotation, a shift in meaning that is only found after Herodotus.44 Similary, Herodotus uses the word εὐρυχωρίη ‘open space’ twice (Hdt. 4.71.4; 8.60β), but Diodorus (or his sources) may derive its specific use in opposition to στενοχωρία (also in Diod. Sic. 20.29.10) from Thucydides (Thuc. 7.49.2). The same holds for πλεονεκτέω, which Herodotus uses once to indicate Themistocles’ greed (8.112.1). In fourth century BCE military prose, the verb is found without this negative connotation to indicate the advantage that one gains in battle.45 συμφράσσω ‘pack closely together’ is found once in Herodotus (Hdt. 4.73.2), but Diodorus uses it in a specifically military meaning that is familiar to Xenophon and Polybius.46 Other expressions are post-Herodotean tout court. εὐκίνητος (‘agile, mobile’) originates from medical and philosophical discourse, and may have found its way into historiography via Polybius.47 εὐχερῶς ‘indifferent to harshness’, and later ‘easy’ has a similar tradition, though it is attested in Sophoclean tragedy.48 Via medical, philosophical and oratorical literature, it appears to have found its way into Polybius, who bequeathed it to the historiographical tradition.49 44

45

46 47

48 49

It occurs in the inscriptions of the Parthenon treasures (IG I3 349–351 and 353–357), the earliest of which (IG I3 349) dates from 428/427BCE. In literature, the non-Thracian meaning is found for the first time in Euripides (Bacch. 783, and compare, possibly, Fr. 530 Nauck), and in Xenophon’s Anabasis (1.10.12; 2.1.6; 5.2.29; 6.1.9; 6.1.10; 7.4.7). For the earliest instances of this use, LSJ refers to Thuc. 4.61.5, 4.62.3 and 4.86.6. In all three cases, however, a negative connotation cannot be excluded. For an unambiguous example, see Xen. Cyr. 7.1.33, where the Egyptians have the advantage over the Persians because of better weaponry. Compare Xen. Cyr. 1.6.35; Hell. 7.1.33; Pl. Lach. 183a1. Xen. Hell. 1.1.7; Polyb. 2.69.9; 10.14.12. Cf. Palm 1955: 105. For instances in medical and philosophical literature, see Hippoc. Aph. 3.17; Pl. Ti. 56a (twice). Polybius uses the word in a military sense: 1.33.7; 1.40.7; 1.60.8; 3.43.2, just like Diodorus: 5.34.5; 5.34.6; 15.44.2; 20.95.1, with the exception of 2.6.6, where it describes a garment. For other instances of εὐχερής in poetry, see Soph. Phil. 519; 875. For instances in medical, philosophical and oratorical literature, see, resp. Hippoc. Ep. 18;

76

de bakker

A similar path is followed by the abstract noun κουφότης ‘lightness’,50 and by σκεπάζω ‘protect’, a verb found mainly in the philosophers, but also used by Xenophon and Polybius in a military context.51 The adjective σκεπαστήριος ‘defensive’ seems to be a contemporary coinage, also used by Dionysius of Halicarnassus and Philo Judaeus.52 The same holds for the expression πυκνοῖς τραύμασι, which is not attested earlier than Diodorus, although the adjective πυκνός is as old as Homer.53 This brief look at the discourse shows that Diodorus’ military terminology finds its origins in post-Herodotean times, most notably the fourth century BCE, where it was either first used in the works of Thucydides and Xenophon, or— via the philosophers and medical writers—inherited by Polybius. This finding, then, is in harmony with Palm’s general assessment of Diodorus’ style as typically ‘hellenistic’: Alles das oben Gesagte scheint somit auf das 4. Jahrh. als die wahrscheinliche Geburtszeit des hellenistischen Prosastils hinzuweisen. Palm 1955: 20654 The observations about the background of the military discourse tie in with the second main difference between Herodotus’ and Diodorus’ versions, namely the latter’s distinctly panhellenic outlook (as discussed in § 2.3). This, too, appears to be a typically fourth century BCE focus,55 a period in which intellectuals like Isocrates gave voice to a sentiment that sought to stress the importance of Greek unity in the face of non-Greek invaders, using the Persian Wars as an example. Marincola in this respect observes that fourth-century BCE historians (who served Diodorus as sources) reverse the Herodotean theme of the fragility of the Greek alliance

50

51 52 53 54

55

Pl. Phd. 117c4; Aeschines Tim. 179; the word occurs thirty times in Polybius (e.g. 1.32.4; 10.47.4), and—apart from here—twelve times in Diodorus (1.36.3; 1.84.4; 8.6.2; 11.43.3; 11.73.2; 13.31.3; 16.20.4; 19.93.6; 19.95.7; 31.4.1; 31.27.5; 31.38.1). For κουφότης in poetry, see Eur. Fr. 119 Nauck; Eupolis Fr. 96 Austin. Most instances are found in the philosophical works of Plato (e.g. Lg. 897a7; Ti. 65e6) and Aristotle (e.g. Cael. 309a28; Metaph. 1020b10). For the philosophers, see e.g. Arist. GA 785a27; PA 658b6; Thphr. CP 4.12.8. For Xenophon, see Eq. 12.5; 8 (twice); Mem. 3.10.9. For Polybius, see 1.22.10; 8.4.4; 10.13.2; 21.28.4; 30.25.9. E.g. Dion. Hal. Ant. Rom. 2.38.3; 5.24.1; 14.9.2; Philo Opif. 84; Her. 203. For a parallel in Diodorus’ Library, see 17.34.6 τραυματιζόμενοι πυκνῶς. To Palm’s already extensive list of exclusively Hellenistic terminology, I add the use of the opposition between προαίρεσις ‘choice beforehand’ and ἀποτέλεσμα ‘result afterwards’ (Diod. Sic. 11.11.2), which Diodorus appears to derive from Polybius (Polyb. 2.39.11). Pace Van Wees in this volume (see above n. 32).

a narratological comparison of herodotus and diodorus

77

with a decidedly panhellenic perspective that plays out in two ways: first, individual states claim to have fought the Persians not for themselves but “on behalf of Hellas”; and second, the narratives that we find emphasize Greek unity and Greek aggressiveness against the Persians. Marincola 2007c: 11556 Having established, then, that the two main differences between Herodotus’ and Diodorus’ versions of Thermopylae may well originate from the fourth century BCE, I will now attempt to place the two traditions within their wider contexts.

4

The Origins of the Two Thermopylae Traditions and Their Mutual Interrelationships: a Hypothesis

As a conclusion to this analysis, I propose a hypothesis about the origins of and relationship between Herodotus’ and Diodorus’ versions of Thermopylae based on the outcome of the narratological comparison and the analysis of Diodorus’ lexicon in the fighting passages. Some aspects of this hypothesis are impossible to prove, but in order to present a coherent picture of the Thermopylae tradition, I will include them nonetheless, whilst indicating their speculative nature.57 As we have seen in the narratological comparison, Herodotus’ version shows traces of polemic and in general a high degree of narrative strategies that seek to authenticate the events as told in the narrative (§ 2.3). We may assume, then that Herodotus responded to traditions that sought to glorify the deeds of the Greeks.58 Such traditions might have come into existence already shortly after the battle, when poets like Simonides were commissioned to write songs of praise and epitaphs were inscribed at the site in remembrance of the fallen. This yielded encomiastic versions of the battle of Thermopylae that glorified

56

57

58

See also Flower 1998: 368, and above (§ 1). Compare in this respect also Isocrates’Panegyricus 85, where the rivalry between Sparta and Athens is mentioned as a source of mutual inspiration during the Persian Wars, in which they pushed one another to great deeds. Some aspects of my hypothesis can be reinforced by the model of the relationship between Herodotus’ and Diodorus’ versions of Thermopylae developed by Van Wees in this volume. Compare Herodotus’ observation about the life of Cyrus. From the various versions that existed about the life of this Persian King, Herodotus chose the one that did not ‘glorify’ (σεμνοῦν) his achievements (Hdt. 1.95.1).

78

de bakker

Greek achievements in terms familiar from epic and lyric poetry, which could account for a tradition in which the Spartans died whilst making a night raid into the enemy camp.59 In the second half of the fifth century BCE, and against the background of increasing tensions between the Greek states, Herodotus questioned these traditions and drafted a more controversial account of the Thermopylae events. He sought to authenticate his version by including a detailed description of the battlefield, a catalogue of troops, and a slow-paced presentation of the events. He also chose to comment on the events in his own voice, and insert variant versions to further endorse his historiographical authority and thereby lend credibility to his account. Typical features of his polemical version are the emphasis on Greek discord, which surfaces a number of times throughout the narrative, and the questioning of motivation, as exemplified in his portrait of Leonidas, who appears to frame his heroic death. The contemporary circumstances of Greek interstate warfare may have prompted Herodotus to elaborate these themes—as he does elsewhere in his work—and shed a sobering light upon the heroic achievements of the past. He might even, though implicitly, have targeted the encomiastic tradition of the night raid by mentioning the lion statue in honour of Leonidas at the site in the pass where the last Spartans fell (Hdt. 7.225.3), a typical instance of Herodotean autopsy, which awards evidential value to monuments that can be seen in his own times.60 A typical Herodotean theme serves as a pointer towards a potential explanation of Herodotus’ choices. Twice in the Thermopylae narrative, Herodotus mentions fathers who die in battle, but leave sons behind who survive. This is the case with the Spartans whom Leonidas selects (Hdt. 7.205.2), and with Megistias, who sends his son away from the battlefield once he knows that the Spartans are doomed (Hdt. 7.221). When the battle is over, the theme recurs, but this time in reverse order: it is the father Leontiades, the branded leader of the medising Thebans, who is allowed to live on, whereas his son dies at the beginning of the Peloponnesian War, when the Thebans send a force of ‘400’ to take the city of Plataeae by surprise (Thuc. 2.2–2.6). Here, Herodotus turns the motif around, stressing the unnatural situation of war, in which ‘fathers bury their sons’ (cf. Hdt. 1.87.4), and connects it to the Greek infighting of his own time. In doing so, he hints at a reason for toning down the encomiastic 59 60

See Flower 1998: 366–372, Van Wees in this volume and the introduction to this chapter (§ 1). Van Wees in this volume makes the same argument. For parallels of this use of monuments in the Histories see e.g. the votive gift of Arion the singer at cape Taenarum (Hdt. 1.24.8) and the statues of Cleobis and Biton in Delphi (Hdt. 1.31.5).

a narratological comparison of herodotus and diodorus

79

aspects of the canonical Persian War narratives. In his view, the Greeks did not deserve much praise, as they fought one another rather than a foreign enemy. Thus his version of the battle of Thermopylae can be considered an antidote against Greek feelings of superiority in relation to their recent past. Moving from Herodotus’ to Diodorus’ version, we have observed that the latter pays more attention to the military aspects of the battle, and uses a terminology of which the origins date back to the fourth century BCE. This ties in with changes in the historiographical genre that occurred between the fifth and fourth century BCE. Compared to Herodotus, it was Thucydides who cut down on ethnographical and geographical digressions, which account for much of the material in Herodotus’ Thermopylae version, such as the layout of the battlefield (Hdt. 7.176) and the geography of Achaia and Malia as focalised by Xerxes (Hdt. 7.196–7.201). Instead, Thucydides included extensive descriptions of the battles themselves, which he claimed to have witnessed or heard about from eyewitnesses (Thuc. 1.22.2).61 Successors like Ephorus appear to have followed in his footsteps and to have retrospectively applied such descriptions also in the case of narratives of non-contemporary battles like Thermopylae. Instead of admitting, in Herodotean fashion, that an accurate representation of all incidents on a battlefield is impossible (for instance Hdt. 8.87.1), they chose to authenticate their narratives with more elaborate battle scenes and use contemporary military terminology to support their authority. Thus it is no coincidence that the narrative of the night raid in Diodorus contains a motif known from Thucydides, who in his account of the night-battle of Epipolae at Syracuse mentions the chaos that ensues from the failing recognition of passwords (Thuc. 7.44.4–5). Diodorus’ version describes a similar scene, in similar terms, in Xerxes’ camp (Diod. Sic. 11.10.2).62 Furthermore, we have seen that the emphasis in Diodorus’ account on Greek concord during the battle of Thermopylae may reflect fourth century BCE panhellenic ideals. This can be interpreted as a polemical response against Herodotus’ version, which thematises Greek discord and disloyalty.63 Another way in which Diodorus’ version reacts against Herodotus is in magnifying the crucial importance of the battle of Thermopylae for the survival of Greece, presenting it as an example of heroism to the other Greeks which encouraged them to continue fighting. Ultimately, this view may have gone back to the Spartans, 61 62 63

Momigliano 20143: 34. Flower 1998: 366; Marincola 2007c: 116–117. Albertz 2006: 67–72 points at the panhellenic tendencies in the Thermopylae tradition that we encounter in the Attic Orators (Lysias Epitaphios 20–47, and Isocrates Panegyricus 51–99).

80

de bakker

as they sought to counter the impression that the Greeks owed their freedom and liberation from Persian overlordship to the Athenians.64 Regardless of the question of whether Ephorus was schooled by Isocrates,65 he may well have taken a more pro-Spartan, panhellenic, and military account of Thermopylae as point of departure in his universal history.66 Three centuries later, Diodorus decided to follow Ephorus’ version and made Thermopylae the turning point of the Persian Wars. The structure of his version and his choice of words reveal engagement with Herodotus’ account. As a compiler, however, he reduced the length of the account according to the principle of ‘symmetry’, and, following Polybian practice, reduced the use of speeches in oratio recta. Much in harmony with the tastes of his contemporary Roman intellectual audience, he praised the heroic self-sacrifice of Leonidas and his Spartans against the barbarians from the east, and presented them as moral exempla of disciplined soldiers who will never leave their ranks, regardless of the course of the battle.67 Livy would write about Roman defeats in a similar manner, glorifying those who refused to give up, and vilifying those who lost heart.68

64

65 66 67

68

If we believe Thucydides, the Athenians were all too keen to underline this before the Archidamian war (e.g. Thuc. 1.73–1.74). Herodotus himself subscribes to this view, though admits its unpopularity (Hdt. 7.139). Furthermore, Van Wees in this volume plausibly suggests that the Spartans themselves felt the need to justify their limited commitment to the Greek defence in the decades after the Persian Wars, and added heroic elements to the Thermopylae tradition so as to make their rather unimpressive contribution look like a generous act of self-sacrifice on behalf of Greece. Flower 1994: 42–62 argues that Theopompus and Ephorus were not students of Isocrates, an idea, he argues, that was based on Hellenistic biographical invention. Marincola 2007b: 172–174 observes that Ephorus’ strong interest in moralism can be seen as typical of the fourth century BCE. Marincola 2007b: 179 demonstrates that creating (moral) paradigms of the characters who had made history was an essential feature of the genre of universal history, which makes it seamlessly tie in with the contemporary taste in Roman historiography (e.g. Sallust and Livy). The Roman background to Diodorus’ work also surfaces, according to Evans (2012), in an attitude to tyrants more positive compared to earlier historians (cf. Muntz 2017: 191–214). This could result from Diodorus’ contemporary experiences in first century BCE Rome, where Caesar and Octavian brought peace after decades of civil war. See De Bakker and Van der Keur in this volume.

a narratological comparison of herodotus and diodorus

81

Appendix 1: Proper Names, Toponyms and Ethnonyms in Herodotus’ and Diodorus’ Versions of the Battle of Thermopylae69 I Proper Names Herodotus: (47, mentioned 151 times in Thermopylae narrative in total) Achaemenes (236.1; 237.1), Alexander of Macedon (175.1), Alpheus son of Orsiphantus (227), Amphictyon (200.2), Anaxandrides (204; 205.1), Aristodemus (229.1 (2×); 229.2 (2×); 230; 231 (2×)), Artanes (224.2 (2×)), Artemis (176.1), Athamas son of Aeolus (197.1; 197.3; 197.4), Athenades (213.2; 213.3), Cercopi (216), Chilon (235.2), Cleombrotus (205.1), Cleomenes (205.1 (3×); 239.4), Corydallus (214.1; 214.2), Cytissorus son of Phrixus (197.3 (2×)), Darius (224.2 (4×)), Demaratus son of Ariston (209.1; 234.1 (2×); 234.3; 237.1; 237.3; 239.2 (2×)), Demeter (200.2), Demophilus son of Diadromes (222), Dienekes (226.1; 226.2), Dithyrambus son of Harmatides (227), Dorieus (205.1 (2×)), Ephialtes son of Eurydemus (213.1; 213.3 (2×); 214.2 (2×); 214.3; 215; 218.2; 218.3; 223.1; 225.1), Eurymachus son of Leontiades (233.2), Eurytus (229.1 (2×)), Gorgo (239.4), Habrocomes (224.2), Heracles (176.3; 198.2; 204; 220.4), Hydarnes (211.1; 215 (2×); 218.2; 218.3), Hyperanthes (224.2), Hystaspes son of Arsames (224.2), Ino (197.1), Leonidas (204; 205.1; 205.3; 206.1; 207; 208.1; 217.2; 219.2; 220.1; 220.2; 220.4; 221; 222 (3×); 223.2; 224.1; 225.1; 225.2; 228.1; 229.1; 233.1; 238.1; 238.2; 239.4), Leontiades son of Eurymachus (205.2; 233.1; 233.2), Maron son of Orsiphantus (227), Megistias (219.1; 221; 228.3; 228.4), Melampus (221), Onetes son of Phanagores (214.1; 214.2; 214.3), Pantites (232), Phratagoune (224.2), Phrixus (197.1), Pythia (220.3), Sandoces (196), Simonides son of Leoprepes (228.4), Xerxes (186.2; 187.1; 187.2; 196; 197.1; 197.4; 201; 208.1; 208.3; 209.1; 209.2; 209.5 (2×); 215; 223.1; 223.2; 225.1; 233.2; 234.1; 234.3; 236.1 (2×); 237.1; 238.1; 238.2; 239.2), Zeus (197.1; 220.4). Mentioned just once in Histories: Alpheus son of Orsiphantus; Amphictyon; Cercopi; Demophilus son of Diadromes; Dithyrambus son of Harmatides; Eurymachus son of Leontiades; Habrocomes; Hyperanthes; Ino; Maron; Pantites; Phratagoune; Phrixus. Diodorus: (6, mentioned 37 times in Thermopylae narrative in total)

69

All references in Herodotus to book 7, in Diodorus to book 11. Patronymics built on names that do not play a role elsewhere in the narrative are not included. Neither is the genealogy of Leonidas (Hdt. 7.204), Anaxandrides excepted.

82

de bakker

Demaratus (6.1; 6.2), Eurybiades (4.2), Leonidas (4.2; 4.6 (2×); 4.7; 5.5; 6.4; 7.1; 7.2; 8.2; 8.4; 8.5; 9.1; 9.2; 9.3; 9.4; 10.1; 10.2; 10.4; 11.6), Simonides (11.6), Tyrrhastiadas (8.5), Xerxes (5.1; 5.3 (2×); 5.4 (2×); 6.2; 7.4; 8.1; 8.4; 10.3; 10.4; 11.5; 12.1). Mentioned just once in Library: Simonides, Tyrrhastiadas II Toponyms Herodotus: (52, mentioned 145 times in Thermopylae narrative in total) Achaia (196 (2×); 197.1; 198.1; 198.2), Aea (Colchis) (197.3), Aeolia (176.4), Alpeni (176.2; 176.5; 216; 229.1), Anopaea path (216 (2×)), Anthele (176.2; 200.2), Anticyra (198.2; 213.2), Aphetae (196), Arcadia (202), Artemisium (sanctuary) (175.2; 176.1 (2×); 177), Asia (184.1; 184.4; 184.5; 185.3), Asopus (river) (199; 200.1 (2×); 200.2; 216; 217.1), Chutri wells (176.3), Cythera (235.2), Delphi (239.1), Dyras (river) (198.2), Epidanos (river) (196), Euboea (176.1), Europe (185.1), Greece (175.2; 176.2; 176.5; 177; 203.2; 209.2; 235.3 (2×); 239.1; 239.2), Halos (197.1), Histiaea (175.2), Isthmus (175.1; 177; 207; 235.4 (2×)), Korinth (202), Lacedaemon (220.3; 220.4; 231; 234.2; 239.4), Laconia (235.1; 235.3), Magnesia (176.1), Malia (198.1 (2×); 201), Melampygus rock (216), Melas (river) (198.2; 199), Oete (176.3), Onochonus (river) (196), Orchomenos (Arcadia) (202), Peloponnese (202; 207; 228.1; 235.4; 236.2), Phoenix (river) (176.2; 200.1 (3×); 200.2), Phlious (202), Pieria (177), Plataeae (231), Pylaie (213.2), Seats of Cercopi (216), Sepias (186.2), Sicily (205.1), Skiathos (176.1), Sousa (239.2), Sparta (204; 206.1; 209.4; 220.2; 220.4; 228.3; 229.1; 229.2; 230; 232; 234.2), Spercheius (river) (198.2; 228.3), Thermopylae (175.1; 175.2 (2×); 176.2; 176.3; 177; 184.1; 186.2; 200.1; 200.2; 201 (2×); 205.2; 206.2; 207; 213.1; 219.1; 233.1; 234.1), Thessaly (175.1; 196 (3×); 198.1; 208.1; 213.2; 232), Thracian Sea (176.1), Thrace (185.1 (2×); 185.2), Trachinian rocks (198.1), Trachis (176.2; 199 (3×); 201; 203.2). Mentioned just once in Histories: Chutri wells; Dyras (river); Epidanos (river); Melampygus rock; Oete; Seats of Cercopi; Trachinian rocks. Diodorus: (12, mentioned 28 times in Thermopylae narrative in total) Acanthus (5.1), Artemisium (4.1), Euboea (4.1), Europe (5.2), Greece (4.1; 9.1; 9.2; 11.1; 11.6), Lacedaemon (4.4), Malian Gulf (5.2), Marathon (6.4), Sparta (4.4; 11.6), Spercheius (river) (5.4), Thebes (4.7), Thermopylae (4.1; 4.2; 4.5; 4.6 (2×); 4.7 (2×); 5.4; 6.3; 6.4; 10.4; 11.6). Mentioned just once in Library: Malian Gulf, Spercheius (river)

a narratological comparison of herodotus and diodorus

83

III Ethnonyms Herodotus: (47, mentioned 191 times in Thermopylae narrative in total) Acarnanian (221), Achaeans (185.2; 197.1; 197.2; 197.3), Aeginetans (203.1), (A)enianans (185.2; 198.2), Amphictyons (200.2; 213.2; 228.4), Anticyran (214.1), Arabians (184.4), Arcadians (202), Athenians (203.1), Boeotians (202), Bottiaeans (185.2), Brygans (185.2), Carystian (214.1), Chalcidican (185.2), Cissians (210.1), Dolopans (185.2), Eordans (185.2), Greeks (175.1; 175.2; 176.5; 177; 185.1; 196 (2×); 201 (2×); 202; 203.1; 205.3 (2×); 207; 209.4; 210.2; 211.2 (2×); 212.2; 213.1; 214.2; 219.1; 219.2; 223.2; 225.1; 225.2; 233.1 (2×); 234.1; 236.1; 239.4), Indian (187.1); Lacedaemonian(s) (204; 208.1; 208.2; 209.2; 211.3; 213.2; 213.3; 218.2; 222; 225.1; 226.1; 226.2; 227; 228.2; 228.3; 234.1; 234.2 (2×); 235.3; 236.3; 238.1; 239.1; 239.2 (2×); 4), Libyans (184.4), Macedonians (185.2), Magnetans (185.2), Malian(s) (196; 213.1; 214.3; 215 (2×); 216), Mantineans (202), Medes (184.2; 207; 210.1; 210.2; 211.1; 211.2; 226.1; 226.2 (2×); 228.3; 239.2), Myceneans (202), Oeteans (217.1), (Opuntian) Locrians (203.1; 207; 216), Paeonians (185.2), Peloponnesians (207; 235.4), Perrhaebans (185.2), Perseidans (= Persians) (220.4), Persian(s) (177; 184.2; 202; 207; 211.1; 211.3 (2×); 212.2; 214.1; 217.1; 218.1; 218.2; 218.3; 219.1; 224.1; 224.2; 225.1; 229.1; 233.1; 236.3; 238.2), Phocians (176.4 (2×); 203.1; 207; 212.2; 215 (2×); 217.2 (2×); 218.1 (2×); 218.2; 218.3 (2×)), Pierans (185.2), Plataeans (233.2 (2×)), Pylagori (213.2; 214.2), Sacae (184.2), Spartiate(s) (202; 206.1; 211.3; 220.1; 220.3; 220.4; 224.1; 226.1; 228.2; 229.2; 231; 235.2), Tegeans (202), Thebans (202; 205.2; 222 (2×); 225.2; 233.1; 233.2), Thespians (202; 222 (2×); 226.1; 227), Thesprotians (176.4), Thessalian(s) (176.4 (3×); 196; 215; 233.2), Thracians (185.2), Trachinian(s) (175.2; 199; 201; 213.2; 214.2; 217.1; 226.1; 226.2). Mentioned just once in Histories: Brygans, Eordans, Oeteans Diodorus: (15, mentioned 79 times in Thermopylae narrative in total) Cissians (7.2), Cymaean (8.5), Greeks (4.1 (2×); 4.5; 4.6; 4.7; 5.4; 5.5; 6.1 (2×); 6.2 (2×); 6.3; 6.4 (2×); 7.1; 7.2; 7.4 (2×); 8.2; 9.1 (3×); 9.3; 10.1; 10.3 (2×); 10.4; 11.1; 11.5 (2×)), Lacedaemonian(s) (4.2; 4.4; 4.5; 6.2; 9.1 (2×); 9.2), Laconian (6.1), Locrians (4.6; 4.7), Malians (4.7), Medes (6.3 (2×); 6.4 (2×); 7.2 (2×)), Persians (4.1; 4.6; 4.7 (2×); 5.4 (2×); 6.2 (2×); 7.4; 8.4; 8.5 (2×); 9.3 (2×); 9.4; 10.4; 11.1), Phocians (4.7), Sacae (7.2), Spartiate(s) (4.2; 4.5; 6.1; 9.1), Thebans (4.7), Thespians (9.2), Trachinian (8.4; 8.5 (2×); 9.3; 10.1). Mentioned just once in Library: Cissians

84

de bakker

Appendix 2: Overview of Anachronies in Herodotus’ and Diodorus’ Versions of the Battle of Thermopylae Herodotus Internal Analepseis (11) 7.212.2: Phocians had been sent up to guard the Anopaea path. 7.219.1: defectors from the Persians had announced the betrayal by Ephialtes. 7.223.1: Ephialtes had advised Xerxes to postpone his attack until Hydarnes’ troops had circumvented the pass. 7.223.2: reference to fighting during previous two days. 7.226.1–7.226.2: the memorable statements that Dienekes uttered to a Trachinian before the battle. 7.229.1: the discord between Eurytus and Aristodemus about how to act when the Spartan position appeared to be lost. 7.230: alternative version of 7.229.1, with Aristodemus as a messenger. 7.233.1–7.233.2: surrender of the Thebans to the Persians, and their subsequent branding. 7.234.1: Xerxes referring to Demaratus’ previous advice (7.209). 7.237.2: Xerxes referring to Demaratus’ previous advice (7.209). 7.238.2: Xerxes’ anger with Leonidas when he was still alive. External Analepseis (18) 7.176.3–7.176.5: wall of the Phocians. 7.184.3: the recruitment of 3.000 pentekonters (referring back to 7.97). 7.197.1–7.197.3: Athamas legend. 7.198.2: Heracles relieved by the river Dyras. 7.203.1–7.203.2: message to the Locrians and Phocians with a request to support the Greek alliance. 7.205.1: how Leonidas had become king. 7.205.3: Leonidas’ request to the Thebans to send support. 7.206: Allied planning of war efforts at the time of the dispatchment to Thermopylae. 7.208.1: What Xerxes was told about the Spartans at Thermopylae when he was still in Thessaly. 7.209.2 Demaratus refers to his previous advice to Xerxes (7.101–7.104). 7.215: the discovery of the Anopaea path by the Malians during the war between Thessaly and Phocis. 7.220.3–7.220.4: the oracle that announces the destruction of Sparta or the death of its king.

a narratological comparison of herodotus and diodorus

85

7.224.2: when Artanes married off his daughter Phratagoune to Darius, and thereby left his estate to Darius. 7.233.1: the Thebans had given earth and water to Xerxes. 7.235.2: Demaratus refers to a statement made by Chilon about Kythera (this could contain an implicit prolepsis). 7.236.2: Achaemenes refers to the wrecking of 400 Persian ships at Magnesia. 7.239.1: the consultation of the oracle of Delphi (see 7.220.4). 7.239.1–7.239.4: how the Spartans were the first of the Greeks to learn about Xerxes’ invasion via a hidden message sent by Demaratus, which was discovered by Gorgo, Leonidas’ wife. Internal Prolepseis (2) 7.175.2: the trapping of the Greeks at Thermopylae by the Persians via the Anopaea path. 7.206.2: it was not expected that the battle of Thermopylae would come so soon. External Prolepseis (8) 7.213.2: Ephialtes’ flight and bounty placed upon his head. 7.213.2–7.213.3: the killing of Ephialtes by Athenades. 7.225.2: the contemporary kolonos at Thermopylae with the monument for Leonidas. 7.228.1–7.228.4: the erection of the stelae and epigrams on behalf of the fallen Greeks. 7.231: the disgrace that befell Aristodemus at home. 7.231: Aristodemus fighting hard at Plataeae to compensate for Thermopylae. 7.232: the return home and suicide of Pantites, who was sent as a messenger to Thessaly. 7.233.2: Eurymachus killed during Theban assault upon Plataeae (cf. Thuc. 2.2– 2.6). Diodorus Internal Analepsis (1) 11.11.1–11.11.4 evaluation of the feelings of the Spartans before engaging the Persians, as well as the perplexity of the Persians about the way in which the battle unfolded. External Analepseis (6) 11.4.2–11.4.4: Leonidas’ meeting with the Ephors. 11.4.6: Locrians had given earth and water to Persians, but changed their minds when Leonidas arrived.

86

de bakker

11.5.1–11.5.2: Xerxes’ march from Abydus to Malian Gulf. 11.6.1: Demaratus fled away from Sparta. 11.6.3: Medes were toppled by the Persians. 11.6.4: reference to the battle of Marathon. External Prolepseis (2) 11.11.5: reference to further Persian wars. 11.11.6: reference to the poets who praised the deeds of the Greeks at Thermopylae and of Leonidas in particular.

Appendix 3: Speeches in Herodotus’ and Diodorus’ Versions of the Battle of Thermopylae (incl. Number of Words per Speech That the Narrator Ascribes to the Character)70 Herodotus Oratio Recta (9) 7.209.2–7.209.4 (Demaratus to Xerxes, 104 words); 7.209.5 (Demaratus to Xerxes, 16 words); 7.220.4 (Oracle of Delphi to Spartans, 50 words); 7.234.1 (Xerxes to Demaratus, 33 words); 7.234.2 (Demaratus to Xerxes, 44 words); 7.234.3 (Xerxes to Demaratus, 21 words); 7.235.1–7.235.4 (Demaratus to Xerxes, 159 words); 7.236.1–7.236.3 (Achaemenes to Xerxes, 150 words); 7.237.1–7.237.3 (Xerxes to Achaemenes, 109 words). Oratio obliqua (9) 7.197.1–7.197.3 (Achaeans to Xerxes, 147 words); 7.203.1–7.203.2 (Greek alliance to Locrians and Phocians, 82 words); 7.209.5 (Xerxes to Demaratus, 8 words); 7.220.3 (Oracle of Delphi to Spartans, 12 words); 7.221 (Megistias to the other Greeks, 4 words); 7.226.1 (Trachinian to Dienekes, 18 words); 7.226.2 (Dienekes to Trachinian, 24 words); 7.233.1 (Thebans to Persians, 26 words); 7.239.4 (Gorgo to Spartans, 6 words). Diodorus Oratio recta (2) 11.6.1 (Xerxes to Demaratus, 14 words); 11.6.2 (Demaratus to Xerxes, 37 words).

70

Not included are the records of speech acts, e.g. constructions with κελεύω + inf.

a narratological comparison of herodotus and diodorus

87

Oratio obliqua (10) 11.4.3 (Ephors to Leonidas, 6 words); 11.4.3 (Leonidas to Ephors, 20 words); 11.4.4 (Ephors to Leonidas, 7 words); 11.4.4 (Leonidas to Ephors, 46 words); 11.5.4 (Xerxes to the Greeks, 35 words); 11.5.5 (Greeks to Xerxes, 39 words); 11.8.1 (Xerxes to his special forces, 14 words); 11.8.4 (Trachinian to Xerxes, 30 words); 11.9.1 (Greeks to fellow Greeks, 16 words); 11.9.1 (Leonidas to the Greeks, 43 words).

Bibliography Albertz, A., Exemplarisches Heldentum. Die Rezeptionsgeschichte der Schlacht an den Thermopylen von der Antike bis zur Gegenwart (Munich 2006). Asheri, D., ‘General Introduction’, in O. Murray & A. Moreno (eds.), A Commentary on Herodotus Books I–IV (Oxford 2007) 1–56. Bakker, M.P. de, Speech and Authority in Herodotus’ Histories, PhD thesis (Amsterdam 2007). Baragwanath, E., Motivation and Narrative in Herodotus (Oxford 2008). Branscome, D., Textual Rivals: Self-Presentation in Herodotus’ Histories (Ann Arbor 2013). Brown, A.R., ‘Remembering Thermopylae and the Persian Wars in Antiquity’., in C.A. Matthew & M. Trundle (eds.), Beyond the Gates of Fire. New Perspectives on the Battle of Thermopylae (Barnsley 2013) 100–116. Burton, A., Diodorus Siculus. Book I. A Commentary (Leiden 1972). Clarke, K., ‘Universal Perspectives in Historiography’, in C.S. Kraus (ed.), The Limits of Historiography. Genre and Narrative in Ancient Historical Texts (Leiden 1999) 249– 279. Darbo-Peschanski, C., Le discours du particulier: essai sur l’enquête hérodotéenne (Paris 1987). Dewald, C., ‘Narrative Surface and Authorial Voice in Herodotus’Histories’, in D. Boedeker & J. Peradotto (eds.), Herodotus and the Invention of History (Buffalo 1987) 147– 170. Emde Boas, E. van & Temmerman, K. de (eds.), Characterization in Ancient Greek Literature. Studies in Ancient Greek Narrative. Volume IV (Leiden 2017). Evans, R., ‘Nostalgia in Diodorus Siculus (16.82.5 & 16.83.1): Wishing for the Tyrant of Integrity?’, in P. Bosman (ed.), Corruption and Integrity in Ancient Greece and Rome. (Pretoria 2012) 45–62. Flory, S., ‘Laughter, Tears and Wisdom in Herodotus’, AJPh 99.2 (1978) 145–153. Flower, M., Theopompus of Chios (Oxford 1994). Flower, M., ‘Simonides, Ephorus, and Herodotus on the Battle of Thermopylae’, CQ 48.2 (1998) 365–379.

88

de bakker

Fornara, C.W., ‘(review of) Diodorus Siculus and the First Century. Kenneth S. Sacks.’ CPh 87.4 (1992) 383–388. Gainsford, P., ‘Herodotus’ Homer: Troy, Thermopylae, and the Dorians’, in C.A. Matthew & M. Trundle (eds.), Beyond the Gates of Fire. New Perspectives on the Battle of Thermopylae (Barnsley 2013) 117–137. Gauger, B. & Gauger, J.-D, Die Fragmente der Historiker: Ephoros von Kyme und Timaios von Tauromenion (Stuttgart 2015). Green, P., Diodorus Siculus, Books 11–12.37.1. Greek History 480–431B.C.—the Alternative Version (Austin 2006). Green, P., ‘Diodorus Siculus on the Third Sacred War’, in J. Marincola (ed.), A Companion to Greek and Roman Historiography (Malden 2007) 363–370. Groten, F.J., ‘Herodotus’ Use of Variant Versions’, Phoenix 17.2 (1963) 79–87. Haillet, J., Diodore de Sicile: Bibliothèque Historique. Livre XI (Paris 2001, Budé edn.). Hollmann, A., The Master of Signs: Signs and the Interpretation of Signs in Herodotus’ Histories (Washington 2011). Hornblower, S., ‘Introduction’, in S. Hornblower (ed.), Greek Historiography (Oxford 1994) 1–72. Immerwahr, H.R., Form and Thought in Herodotus (Cleveland 1966). Jong, I.J.F. de, ‘Aspects narratologiques des Histoires d’Hérodote’, Lalies 19 (1999) 217– 275. Jong, I.J.F. de, ‘Herodotus’, in I.J.F. de Jong, R. Nünlist & A. Bowie (eds.), Narrators, Narratees, and Narratives in Ancient Greek Literature. Studies in Ancient Greek Narrative. Volume I (Leiden 2004) 101–114. Jong, I.J.F. de, (ed.), Space in Ancient Greek Literature. Studies in Ancient Greek Narrative. Volume III (Leiden 2012). Jong, I.J.F. de, Nünlist, R., & Bowie, A. (eds.), Narrators, Narratees, and Narratives in Ancient Greek Literature. Studies in Ancient Greek Narrative. Volume I (Leiden 2004). Jong, I.J.F. de & Nünlist, R. (eds.), Time in Ancient Greek Literature. Studies in Ancient Greek Narrative. Volume II (Leiden 2007). Lateiner, D., ‘No Laughing Matter: a Literary Tactic in Herodotus’, TAPhA 107 (1977) 173– 182. Marincola, J., ‘Herodotean Narrative and the Narrator’s Presence’, in D. Boedeker & J. Peradotto (eds.), Herodotus and the Invention of History (Buffalo 1987) 121–137. Marincola, J., Authority and Tradition in Ancient Historiography (Cambridge 1997). Marincola, J., ‘Genre, Convention, and Innovation in Greco-Roman Historiography’, in C.S. Kraus (ed.), The Limits of Historiography. Genre and Narrative in Ancient Historical Texts (Leiden 1999) 281–324. Marincola, J., ‘Speeches in Classical Historiography’, in J. Marincola (ed.), A Companion to Greek and Roman Historiography, vol. I (Malden 2007a) 118–132.

a narratological comparison of herodotus and diodorus

89

Marincola, J., ‘Universal History from Ephorus to Diodorus’, in J. Marincola (ed.), A Companion to Greek and Roman Historiography, vol. I (Malden 2007b), 171–179. Marincola, J., ‘The Persian Wars in Fourth-Century Oratory and Historiography’, in E. Brigdes, E. Hall & P.J. Rhodes (eds.) Cultural Responses to the Persian Wars. Antiquity to the Third Millenium (Oxford 2007c) 105–125. Matthew, C., ‘Towards the Hot Gates: the events leading to the Battle of Thermopylae’, in C. Matthew & M. Trundle (eds.), Beyond the Gates of Fire (Barnsley 2013a) 1–26. Matthew, C., ‘Was the Greek defence of Thermopylae in 480 BC a suicide mission?’, in C. Matthew & M. Trundle (eds.), Beyond the Gates of Fire (Barnsley 2013b) 60–99. Momigliano, A., ‘The Place of Herodotus in the History of Historiography’, in R.V. Munson (ed.), Oxford Readings in Classical Studies. Herodotus: Volume I. Herodotus and the Narrative of the Past ([1958] Oxford 2014) 31–45. Most, G., ‘Principate and System’, in T.A. Schmitz & N. Wiater (eds.), The Struggle for Identity: Greeks and their Past in the First Century BCE (Stuttgart 2011) 163– 179. Müller, D., ‘Herodot—Vater des Empirismus?’, in G. Kurz, D. Müller & W. Nicolai (eds.), Gnomosyne: menschliches Denken und Handeln in der frühgriechischen Literatur (Munich 1981) 299–318. Muntz, C.E., Diodorus Siculus and the World of the Late Roman Republic (Oxford 2017). Murray, O., ‘Herodotus and Hellenistic Culture’, CQ 22.2 (1972) 200–213. Palm, J., Über Sprache und Stil des Diodoros von Sizilien. Ein Beitrag zur Beleuchtung der hellenistischen Prosa, dissertation (Lund 1955). Pausch, D. (ed.), Stimmen der Geschichte: Funktionen von Reden in der antiken Historiographie (Berlin 2010). Pelling, C.B.R., ‘Aeschylus’ Persae and History’, in C.B.R. Pelling (ed.), Greek Tragedy and the Historian (Oxford 1997) 1–19. Pelling, C.B.R., Literary Texts and the Greek Historian (London 2000). Pelling, C.B.R., ‘Speech and Narrative in the Histories’, in C. Dewald & J. Marincola (eds.), The Cambridge Companion to Herodotus (Cambridge 2006) 103–121. Priestley, J., Herodotus & Hellenistic Culture. Literary Studies in the Reception of the Histories (Oxford 2014). Purves, A.C., Space and Time in Ancient Greek Narrative (Cambridge 2010). Rood, T., ‘Herodotus’, in I.J.F. de Jong & R. Nünlist (eds.), Time in Ancient Greek Literature. Studies in Ancient Greek Narrative, Volume II (Leiden 2007) 115–130. Rubincam, C.I.R., ‘The Organization and Composition of Diodorus’ Bibliotheke’, EMC 31.3 (1987) 313–328. Sacks, K.S., Diodorus Siculus and the First Century (Princeton 1990). Scardino, C., Gestaltung und Funktion der Reden bei Herodot und Thukydides (Berlin 2007).

90

de bakker

Scherr, A., Diodors XI. Buch. Kompositions- und Quellenstudien, dissertation (Tübingen 1933). Schmitz, T.A., ‘The Image of Athens in Diodorus Siculus’, in T.A. Schmitz & N. Wiater (eds.), The Struggle for Identity: Greeks and their Past in the First Century BCE (Stuttgart 2011) 235–251. Sulimani, I., Diodorus’ Mythistory and the Pagan Mission (Leiden 2011). Thomas, R., Herodotus in Context: Ethnography, Science and the Art of Persuasion (Cambridge 2000). Trundle, M., ‘Thermopylae’, in C.A. Matthew & M. Trundle (eds.), Beyond the Gates of Fire. New Perspectives on the Battle of Thermopylae (Barnsley 2013) 27–38. Waters, K.H. ‘The Purpose of Dramatisation in Herodotos’, Historia 15.2 (1966) 157–171. Wiater, N., ‘Geschichtsschreibung und Kompilation. Diodors historiographische Arbeitsmethode und seine Vorstellungen von zeitgemäßer Geschichtsschreibung’, RhM 149.3/149.4 (2006) 248–271.

chapter 4

Narrative and Identity in Thermopylae (Herodotus 7.201–7.239) Antonis Tsakmakis

1

Introduction

The battle of Thermopylae can be considered a rather inconsequential military event. For the Persians it was a successful operation: the King’s land forces could surpass the last physical obstacle on their way from Northern Greece to Attica and the Peloponnese, after having overpowered a small number of Greek fighters who defended the mountain pass till the end. Despite its minor scale and the defeat of the Greeks Herodotus devotes an extensive section to the incident which draws a favourable picture of the Greeks (7.201–7.239).1 His large-scale narrative competes with heroic epic inasmuch as it creates heroes, fosters moral and political values, and creates an almost mythical version of recent history. The explicit or implicit characterization of the protagonists epitomizes the Greek soul as opposed to the barbarian mentality, while, above all, it demonstrates Spartan excellence: for the first time in Herodotus’ work, it amply unfolds what would become the stereotypical Spartan identity. Thus, Herodotus’ narrative is of paramount importance for the formation of the emblematic image of “the Spartan”, which has proven to be very influential right up until this day. This chapter explores the narrative and linguistic devices which are instrumental in constructing a heroic image of the Spartans, and especially of their king, Leonidas. We show that the interpretatio spartana of the events plays a critical role in the positive evaluation of death in battle, an idea which is meticulously prepared in the preceding narrative. Death is presented as the

1 On Thermopylae as an example of a glorious defeat, see Dillery 1996; Baragwanath 2008: 74– 75; and De Jong in this volume. For Herodotus’ disagreement with other sources and the problems concerning the historicity of some details or their evaluation see in this volume Van Wees and De Bakker. For the identification of the sources of Spartan traditions about the royal families of the city in Herodotus, see Hooker 1989: 134–135; Forsdyke 2002: 531–533; Lombardo 2005 with bibliography; see also Vaniccelli 2013: 16 (on the efforts of the Greek cities to promote their own version of history in the 5th century).

© koninklijke brill nv, leiden, 2019 | doi:10.1163/9789004383340_005

92

tsakmakis

dominant theme of the Thermopylae passage; its importance is further foregrounded through narrative anticipations and embedded focalization: characters are repeatedly presented as being confronted with their imminent death. These narrative characteristics are also supported by linguistic choices such as the use of tenses, especially the pluperfect, which is found remarkably often in the passage.

2

The Idea of Sparta

In the many pages from the beginning of Herodotus’ work to the middle of book 7, Sparta is repeatedly acknowledged as a leading force in the Peloponnese, as well as one of the two principal cities of mainland Greece.2 However, its military superiority and the heroic performance of its men (standard elements of Spartan tradition and propaganda) are not directly demonstrated. Instead, the readers are extensively informed about Sparta’s constitution and customs; there is abundant material about the royal houses: genealogy, diplomacy, successful but also disappointing enterprises, gossip, intrigue. Much emphasis is given to Cleomenes’ career and to his rivalry with Demaratus, who eventually lost his throne and fled to Persia.3 Demaratus now accompanies Xerxes in his campaign against Greece, hoping for his restoration. It is this former king of Sparta who, in his capacity as Xerxes’ consultant, becomes the first broadcaster of Spartan excellence in his exchange with the Persian monarch at the beginning of the campaign (7.101–7.105), a passage which introduces a number of motives which play a key role in the Thermopylae narrative. Herodotus reports a lengthy conversation between Xerxes and Demaratus, which takes place after the crossing of the Hellespont. The conversation links the description of Xerxes’ army (7.61–7.99) with the narrative of his march through Northern Greece (7.108–7.132), which practically meant the submission of all populations. Xerxes invites Demaratus to give his opinion about whether the remaining Greeks will dare to face troops of such phenomenal size. In the course of their exchange, he also remarks that his own soldiers will fight bravely because they are compelled to do so—in his view, a further advantage for the Persian side. Demaratus asserts that all Greeks excel in virtue, yet singles out the Spar2 This idea is most powerfully suggested by being attributed to a third, neutral party. In book 1, Croesus tries to find out which of the two cities is stronger (1.56.2); similarly in book 6, Aristagoras visits both cities looking for support for his revolt (5.38.2, 5.55, 5.97); see also 7.133.1 on Darius’ heralds to Athens and Sparta who were equally mistreated by both cities. 3 See Munson 1993.

narrative and identity in thermopylae (herodotus 7.201–7.239)

93

tans, whose character he draws with much detail. He lets the king know that they are extremely skilled warriors, and that they are not deterred by the number of their enemies; they can encounter and defeat multiple forces. And, above all, they do not feel free to desert and, in contrast to the Persians, their freedom is not constrained by a master, but by their sense of duty, which depends on the observance of their nomos (law, custom). The nomos prescribes that they should never abandon their position in the battlefield; thus Spartan warriors are determined to win or to die. Demaratus’ words foreshadow Thermopylae. First, there is an explicit link of this conversation with Thermopylae, as it is alluded to in a new exchange between the two men, which takes place before the fighting starts. Second, there is a thematic correspondence between Demaratus’ claims and the narrative: the battle account confirms both the fact that, despite their small number, the Greeks are victorious until they are surrounded as a result of Ephialtes’ treason and that the Spartans do not desert but die to the last man.

3

The Triumph of Death—Highlighted, Anticipated, Focalized, Dimmed

Prior to any other information about the outcome of the battle and its impact on the Greek side, the significance of Thermopylae for the invaders is suggested by the narrator’s hint (7.184.1) that Thermopylae and Artemision (where the Persian navy suffered damage after a storm)4 marked a turning point in Xerxes’ march through Greece, as both locations became the theatre of the first Persian casualties in the war. The parallelism of Thermopylae with Artemision (‘until the whole host reached this place and Thermopylae it suffered no hurt’)5 undermines the reading of the forthcoming narrative as an unambiguous success for the Persian side, while it also foreshadows, albeit faintly, the barbarians’ final defeat. This feeling is enhanced by the avoidance of terms denoting victory and defeat in the whole section. What is Thermopylae, after all, if not a Greek defeat and a Persian victory? If for the Persians Thermopylae causes the first casualties and entails a first, alarming sign of their vulnerability, for the Greeks it becomes the Iliadic setting of a deliberate encounter with heroic death. This is artfully announced in the pre-battle section (7.201–7.209) of the narrative. Xerxes sends a scout to 4 On the incident (and on Xerxes’ strategy), see Wallinga 2005: 81–86. 5 Translations are from Godley’s Loeb edition, occasionally modified. References to the text without book number are from book 7.

94

tsakmakis

spy upon the Greek camp (7.208). Upon his return the scout reports that the Spartans are preoccupied with physical exercise and combing their hair. This behaviour is incomprehensible to the king: ‘when Xerxes heard that, he could not understand the truth, namely, that the Lacedaemonians were preparing to slay to the best of their power or be slain’ (7.209.1: ἀκούων δὲ Ξέρξης οὐκ εἶχε συμβαλέσθαι τὸ ἐόν, ὅτι παρασκευάζοιντο ὡς ἀπολεόμενοί τε καὶ ἀπολέοντες κατὰ δύναμιν). In this one statement various focalizations intersect: a) Xerxes’ failure to correctly assess the Greeks’ behaviour (οὐκ εἶχε συμβαλέσθαι), b) the Greeks’ physical and mental preparation (παρασκευάζοιντο) ‘to kill and be killed’, and c) the omniscient narrator’s labelling of his own interpretation of the Spartans’ behaviour as ‘the truth’ (τὸ ἐὸν). This artful use of focalization proves an effective tool for the characterization of the protagonists but also for an authoritative interpretation of the narrative as a whole—an interpretation which is further enhanced by the echo of a Homeric expression: the polyptoton ἀπολεόμενοί τε καὶ ἀπολέοντες reminds us of a similar Iliadic figure.6 The killing of Persians by Greeks (anticipated by ἀπολέοντες in 7.209.1) dominates the first section of the battle narrative (7.210–7.212). After having waited for four days, Xerxes attacks, but all his efforts to repulse the Greeks from the pass prove disastrous. In this first section, the narrative progress is straightforward, without any (narratological) anachronies. The arrival of Ephialtes (7.213.1) is the turning point which changes both the course of the fighting and the style of the narrative. An excursus on the traitor’s later fortune (7.213.2– 7.213.3) separates the first from the second, lengthier section of the battle narrative (7.215–7.233). For the most part, the latter consists of digressions from the main narrative: these include descriptions, proleptic and analeptic references, narratorial comments and argumentative sections, discussions of alternative versions and citations of poetic lines (dated both prior—an oracle—and posterior to the main storyline—three epigrams on the deceased).7 This complex,

6 Cf. Il. 4.450–1 = 8.64–65 ἔνθα δ’ ἅμ’ οἰμωγή τε καὶ εὐχωλὴ πέλεν ἀνδρῶν | ὀλλύντων τε καὶ ὀλλυμένων, ῥέε δ’ αἵματι γαῖα, ‘then there were mingled the groaning and the crowing of men killed and killing, and the ground ran with blood’; 11.80–83 ὁ δὲ νόσφι λιασθεὶς | τῶν ἄλλων ἀπάνευθε καθέζετο κύδεϊ γαίων, | εἰσορόων Τρώων τε πόλιν καὶ νῆας Ἀχαιῶν | χαλκοῦ τε στεροπὴν, ὀλλύντας τ’ ὀλλυμένους τε, ‘he drew away from the others and sat down apart, glorying in his splendour, and looking out over the Trojans’ city and the ships of the Greeks, at the flash of bronze and the men killing and being killed.’ 7 The function of these texts (also in Herodotus) becomes evident from the following remark by Mikalson 2003: 66–67: ‘For those investigating Greek religious views, these epigrams, epitaphs, and other contemporary epitaphs are remarkable for their lack of any mention of an afterlife, of the gods, or of anything we might term “religious.” They are almost exclusively “this-worldly,” and, like the dedications, primarily memorialize the virtues and accomplish-

narrative and identity in thermopylae (herodotus 7.201–7.239)

95

artful composition does not necessarily add to the historicity of Herodotus’ version of the Thermopylae episode; but the coherence of the applied techniques does contribute forcefully to the characterization of the protagonists, as well as strongly suggesting a particular interpretation of the events. Consequently, the narrative style of the passage plays an important part in the creation of meaning and urges the readers to adopt the narrator’s ideological stance. The excursus on Ephialtes contains the first explicit assertion of the final outcome. The traitor’s introductory characterization asserts that he was the person who ‘slew’ (διέφθειρε) the Greeks at Thermopylae. This expression anticipates the outcome of the battle and establishes the ‘death of the Greeks’ as the core idea of the narrative. At the same time, it diminishes the Persians’ active role in the slaying (see below). Further anticipations of dying, always by means of embedded focalization, accumulate as the narrative proceeds: – The Phoceans who guard the mountain passageway prepare themselves for their end as they notice that the enemy is approaching (7.218.3: παρεσκευάδατο ὡς ἀπολεόμενοι). However, the Persians lose interest in them, as soon as they realize that they are not Spartans. This ironic relief produces suspense: the Persians are principally interested in the Spartans, and this only means that the latter’s death is imminent. Eventually, the inexorable end is announced to the Spartans and the other Greeks in a threefold way (which has a striking impact on the reader, who is invited to sympathize with the text-internal recipients of these forecasts): – Their upcoming death is explicitly foretold by the seer Megistias, who ‘advised them of the death that awaited them in the morning’ (ἔφρασε τὸν μέλλοντα ἔσεσθαι ἅμα ἠοῖ θάνατον); his prediction is then corroborated by deserters from the Persian lines who arrive during the night and announce that the Persians are coming over the mountain; finally, the news is confirmed by Greek watchers at dawn (all 7.219.1).8 Although this overwhelming evidence about the imminent peril confronts the Greeks with an acute dilemma—to stay (and die) or to leave (and live)—

ments of these warriors.’ For a discussion of the anachronies in the Thermopylae episode, see also the chapter by de Jong in this volume. 8 The last two anticipations do not mention death explicitly, they are, however, unambiguously introduced as additional items (ἐπὶ δὲ καὶ αὐτόμολοι ἦσαν … τρίτοι δὲ …) to a list of hints which was headed by Megistias’ prophecy, whose only subject was ‘death’. Hence the new signals are not dissociated from the idea of death but, rather, they explain the way Megistias’ prophecy will be fulfilled.

96

tsakmakis

Leonidas and 300 Spartans, together with the Thespians (and the Thebans who were forced to stay) decide to confront the enemy by defending the pass. Arrangements for the fighting are once more referred to as preparations for death: – The seer Megistias refuses to abandon Leonidas, but sends his son away, so that he would not ‘perish with’ (συναπόληται) his father and the Spartans (7.221). – the Thespians remain and ‘die with’ (συναπέθανον) Leonidas and his men (7.222). – The final, heroic evacuation of their initial station is referred to (7.223.2) as an ‘exit towards death’ (ὡς τὴν ἐπὶ θανάτῳ ἔξοδον). In the above passages, death is singled out as the essence of the story. The narrative itself of the final battle justifies this expectation in which abstract labels for the narrated action such as ‘fight’, ‘battle’, ‘war’, ‘victory’, ‘defeat’ are consistently avoided.9 On the other hand, in the battle narrative of Thermopylae, death is more anticipated than it is reported. As is clear from the passages just cited, those anticipations are often focalized by the characters themselves. More specifically, the perception of the events as ‘death’ inherently reflects the perspective of those who were expecting their own death in battle. But while anticipations of death are mostly expectations attributed to the Greeks who have opted for fighting and dying, in the narrative peak of the passage—more specifically, the final scene when the last Greek fighters perish (7.225.2–7.225.3)—there is no explicit reference to ‘dying’ or ‘death’ or any synonymous terms, which contrasts with the explicit death notices in the first part of the account of the last combat: 7.224–7.225.1. Instead, the death of the last Spartan warriors is merely suggested by the fact that they are ‘buried under’ the missiles (and stones?) hurled at them. Such reluctance to name death at this peak in the story is not only to be explained by reference to Herodotus’ suggestive writing style, which owes much to the poetic tradition: a hitherto salient theme—and hence sufficiently present in the mind of the readers—is suppressed, enabling them to activate their own imagination. The decisive turn of the battle is, rather, presented in a way which, again, reflects the perceptions of the last Greek warriors: they first realize that the squadron guided by Ephialtes has already broken its way through the mountain 9 It is only in the concluding phrase 7.234.1 that the theme of the narrative is summarized as ‘fight’ or ‘struggle’ (ἠγωνίσαντο). The narrator now seems to opt for a more general and neutral expression, as he is referring to the Thermopylae section as a whole, from the perspective of the overall composition of his work.

narrative and identity in thermopylae (herodotus 7.201–7.239)

97

path, a fact which, according to the narrator, immediately (ἤδη) changed the course of the battle (7.225.2). This is the first occurrence of any word referring to ‘fighting’: now fighting becomes more important for the Greeks than dying. Death has already been irreversibly decided, yet the defenders do not remain passive, awaiting their end in a fatalistic way. The narrative tellingly shows that their only concern now is fighting—there is neither the need nor the time for decisions, expectations, or any mental preparation. Although the narrator no longer presents events via the embedded focalization of the Greek warriors, he does take up his standpoint near them and thus presents events from their spatial position until the end (7.225.3): ‘in that place they defended themselves with their swords, as many as yet had such, and with fists and teeth; till the foreigners overwhelmed them with missile weapons, some attacking them in front and throwing down the wall of defence, and others standing around them in a ring.’ (ἀλεξομένους μαχαίρῃσι, τοῖσι αὐτῶν ἐτύγχανον ἔτι περιεοῦσαι, καὶ χερσὶ καὶ στόμασι κατέχωσαν οἱ βάρβαροι βάλλοντες, οἳ μὲν ἐξ ἐναντίης ἐπισπόμενοι καὶ τὸ ἔρυμα τοῦ τείχεος συγχώσαντες, οἳ δὲ περιελθόντες πάντοθεν περισταδόν). The first participial clause (ἀλεξομένους … στόμασι) refers to the action of the Greeks which lasted till the end, but had, of course, started at the very beginning of the clash. Κατέχωσαν draws our attention to the essential action which eventually wore down their resistance; it is a progressive action, which is centered around the victims, more precisely around their corporeal state. It is only then that we learn that they had been completely surrounded and attacked from all sides (βάλλοντες … περισταδόν). Under such extreme conditions it is psychologically very likely that the desperately fighting men could only be fully aware of and accept this fact, when they had already been confronted with the physical consequences of it upon their own bodies. What is more, the fact that this action is reported last suggests that it is likely to have lasted longer: in the end, there are no more Greeks to be ‘seen’ in the scene; only the barbarians, who have filled the site of battle. Still, the men do not ‘die’, as one cannot observe and narrate one’s own death. The perfective verb form κατέχωσαν does not contradict this interpretation since the action denoted by the verb is not absolutely co-extensive with ‘dying’; the reader is still insecure as to whether this verb is all that the narrator has to say on the subject. An idea of suspension and delay (which leads to the consciousness of dying) can follow the action of being buried under the stuff thrown at the fighting men. The final impression is an acoustic one: the alliteration of the initial π in the last three words περιελθόντες πάντοθεν περισταδόν has the narrative end with an acoustic image, as the last impressions of the still living warriors were likely to be acoustic rather than visual. The theme of death is artfully developed in the Thermopylae passage. The deliberate fight to the death of the Greeks contrasts with the unheroic death

98

tsakmakis

of others. The Persians are forced to fight unwillingly and, consequently, they die unwillingly (7.223.3).10 As for Ephialtes, the anticipation of his deserved elimination suggests that, according to both the human and divine rule of just retribution, death is retaliated with death.11 At the same time, the detailed analepsis not only underlines the determining role of the betrayal for the outcome of the episode, but also suggests the idea of reversal, which proves to be a significant motive of the whole section. On the linguistic level, the barbarians are denied the active role of ‘killing’ the Greeks, especially the Spartans, thus allowing the deliberate character of the latter’s death to appear more clearly. The Greeks ‘fall’ (7.211.3; 7.224.1), but the Persians never ‘kill’ or ‘slay’ them—on the contrary Greeks ‘kill’ Persians (κατέβαλλον 7.211.3; διεργάζοντο 7.224.1). There are, however, significant exceptions when Greeks are ‘slain’: the Thebans who had medized but were later killed by Xerxes (7.233.2), and the traitor Ephialtes (7.213.3). We may conclude that in the battle narrative of Thermopylae the salience of death is suggested both by explicit references and by the narrative style itself. Typical elements of a battle narrative such as discussion of strategy, pre-battle movements, conduct of war, verbal exhortation, tactics, goals, the result (in terms of victory and defeat) are directly related or subordinated to the central theme of death, which (after its artful introduction in the pre-battle section) unfolds in four stages. First, death is used as an equivalent label for the result of the conflict, as the narrator states from his perspective that Ephialtes became guilty of the death of the Greeks. Second, in a series of anticipations focalized (mainly) by characters, the Greeks are shown to opt deliberately for a heroic death in battle. Third, the battle narrative itself is full of references to the casualties on both sides, although the vocabulary used spares the Greeks from ‘being killed’ by the barbarians. Finally, in the last scene which constitutes the narrative peak of the battle (not least because of its exhaustiveness and richness in detail), events are narrated in an order which seems to reproduce the experience of the fighting Spartans, thus prompting the reader to sympathize with the dying heroes.

10 11

For this kind of compulsory fighting, see Aristotle’s discussion in Eth. Nic.1116a36–b2, who presents it as an example of the lack of bravery. Ephialtes is prosecuted by the Greeks but his eventually killed for another reason; but his murderer still received the award. The story suggests that he deserved to die for more than one reason.

narrative and identity in thermopylae (herodotus 7.201–7.239)

4

99

Leonidas: Plans, Intentions, Expectations—and Surprise

Although the expectation to die is repeatedly attributed to the Greeks, collectively or individually, it is worth remarking that it is never explicitly attributed to the central figure of the narrative, Leonidas, who is ultimately the mastermind of this heroic self-sacrifice.12 It is therefore worthwhile examining Leonidas’ portrayal in a more systematic way.13 The Spartan king is introduced in 7.204 as the leader who was mostly respected—or admired14—(θωμαζόμενος) among the Greek generals: θωμαζόμενος picks up a keyword of Herodotus’ proem (ἔργα μεγάλα τε καὶ θωμαστά).15 Since this admiration is not explained, θωμαζόμενος can be regarded as a kind of narrative marker addressed to the readers, a sign that a key subject is touched upon and an important narrative is to be expected. The readers are entitled to assume that the narrative will include sufficient evidence for the character of the hero. However, readers will have to wait for quite some time to comprehend fully why Leonidas was admired. The reason of the Greeks’ admiration for the Spartan king will never be overtly and directly expressed but readers will be provided with indications for admiring Leonidas. Suspense but also surprise plays a role in this process of gradual elucidation. As will be shown, even when readers have every reason to believe that they already know what is necessary, new evidence will be made available to them, which will supplement the description of the king’s character, and will add to the interpretation of the whole section. The solemn presentation of Leonidas’ genealogy up to Heracles only partly satisfies the readers’ need for more information about Leonidas’ high esteem among the Greeks. Despite the rhetorical potential of Leonidas’ genealogy,16 it was well-known that all Spartan kings claimed to be a descendant of Heracles but, admittedly, not all of them were equally popular. Yet, the readers’ interest in this question soon yields to curiosity about a further assertion, namely that Leonidas became king ‘in an unexpected way’. This claim is explained in the ensuing excursus (7.205.1): ‘For since he had two elder brothers, Cleomenes and Dorieus, he had renounced all thought of the kingship. But when Clemenes died without male issue, and Dorieus was dead too

12 13 14 15 16

Ferrario 2014: 57–61, 82–84. On the implicit and explicit characterization of Leonidas, see Zali 2015: 73–75. See Powell 1938 ad loc. sub II. For a detailed discussion of the use and meaning of θωμάζειν in Herodotus, see Hunzinger 1995; cf. also Asheri 2007: 15 n. 39. Herodotus seems to have been interested in Leonidas’ genealogy, not in reproducing a Spartan king-list; cf. Möller 2001: 252–253.

100

tsakmakis

(having met his end in Sicily), so it came about that the succession fell to Leonidas, because he was older that Anaxandrides’ youngest son Cleombrotus, and moreover had Cleomenes’ daughter to wife’. In this context (which is familiar to Herodotus’ readers from earlier narratives on Cleomenes and Dorieus) the labelling ἐξ ἀπροσδοκήτου is not entirely justified. Leonidas’ becoming a king could not have been positively foreseen from as early as his birth, but it occurred neither suddenly nor against expectation. Thus, the motif of the unexpected seems to function as another narrative marker, a sign that this motif will also play a key role in the narrative, especially in connection to Leonidas. Consequently, the readers are prepared to respond to the narrative in an appropriate way. The next piece of information on Leonidas is that he was married to the daughter of Cleomenes.17 Cleomenes’ daughter is not unknown to the readers. Herodotus had reported in 5.48 (and again in 5.51) that Cleomenes only had one child, a daughter named Gorgo, who, at the age of eight or nine, was present at her father’s meeting with Aristagoras, warning him against the troubles that were likely to stem from the Milesian’s proposals (‘the stranger will ruin you, o father, unless you go now’). Readers may reasonably assume that Leonidas’ wife is the girl who had once safeguarded Sparta. Here, however, in contrast to the flood of names of the king’s male ancestors, the name of Cleomenes’ daughter, who, in addition, is alluded to in a way that suggests that her marriage played a role in Leonidas’ becoming a king, is omitted; the task of identification is left to the readers. As becomes evident later, neither the reference to Leonidas’ wife nor the absence of more details is fortuitous. Herodotus withholds the information which is to be disclosed in an analepsis in the concluding chapter of the Thermopylae section. This handling of information about the woman’s identity is a further device that draws attention to the importance of Leonidas’ introduction and connects the beginning to the end. The woman’s name will be revealed—or, for attentive readers, confirmed— in the last chapter of the book (7.239)18 in the context of a further achievement. We now learn that before the Persian invasion a message was sent to the Spartans by their former king Demaratus, revealing to them Xerxes’ plans.19 For reasons of safety, the message was carved on a wooden tablet under the wax. It was Gorgo who came up with the idea to scrub off the wax and uncover the written text. This message provoked the immediate interrogation of the 17 18 19

On royal marriages and endogamy in Sparta, see Pomeroy 2002: 73–74. The authenticity of the chapter is defended by Corcella 1985, who, however, explains some minor stylistic peculiarities as traces of a late addition to the text by Herodotus himself. On Demaratus’ role in the work, see Boedeker 1987.

narrative and identity in thermopylae (herodotus 7.201–7.239)

101

Delphic oracle by the Spartans about the matter. The reply to their query was the oracle that had been cited in 7.220.4, namely that either Sparta would be destroyed or its king had to die, which was the reason why Leonidas had refused to leave Thermopylae. Thus, the new story, which is classified by the narrator as ‘marvelous’ (τρόπῳ θωμασίῳ; the adjective reminds us of Leonidas’ introduction as θωμαζόμενος, a further link joining the end of the narrative with its beginning), can be associated with an important part of the narrative and provides information which is fundamental for its interpretation. This technique has a foregrounding effect. By withholding the information about the prehistory of the oracle, Herodotus highlights the true background of Leonidas’ decision to die. Eventually, it is the readers who are now exposed to the experience of the unexpected. The meaning of the story is not exhausted in the moral grandeur of the Greek fighters. While readers were left to believe that they had privileged knowledge about the outcome of the battle thanks to repeated anticipations, they now realize that their knowledge was insufficient for a proper understanding of the passage; they still have more to learn about the true reason of the battle. In addition, only now are they fully prepared to evaluate the principal characters. Furthermore, they can realize in retrospect that even if all Greeks were reconciled with the idea of dying in battle, this does not correspond with what was going on in Leonidas’ mind. For him, death was not only acceptable as a duty but it was even more: his self-sacrifice had been planned and desired. Once this is fully acknowledged by the readers, they are even more convinced that Thermopylae was not a defeat but, rather, the fulfilment of a wish, the accomplishment of a conscious, and a preconceived plan. The reversal of the course of the fighting no longer appears as adversity. For Leonidas it was the welcome opportunity for saving Sparta.20 By providing essential background information on the past after the episode has been completely narrated, the narrator urges his readers to review the narrative and re-assess the meaning of the events, under consideration of the additional details now available to them, in full awareness of their wide-ranging implications. While the focalization via the Greeks up to the end of the narrative had caused the illusory impression of an adequate interpretation of the

20

This also secures his glory (kleos), as Herodotus explicitly points out. On the importance of this motif in Herodotus, see Nagy 1987; on kleos in Thermopylae, see Baragwanath 2008: 69–74. For the Spartans death is not the end. It brings kleos, while the deceased Spartans all have children who guarantee continuity. Conversely, the Persians suffer a reversal of their boastful arrogance: the selected fighters whom Xerxes called ‘the Immortals’ are defeated by the Greeks and perish.

102

tsakmakis

events, the new data would prompt the readers to review the story from a different perspective. As the new information which determines the new reading provides knowledge of what was going on in Leonidas’ mind, the king becomes the true, implicit focalizer of the narrative. The events now have to be filtered through the consciousness of Leonidas; the readers re-construct his thoughts, feelings and aims at each stage. As being the one who knew more, Leonidas appears as the true ‘author’ of the narrative in this second reading.21 Now, admiration for Leonidas is not as the result of him having been a descendant of Heracles nor of reasons that were known only to the Greeks who were gathered in Thermopylae. Admiration crops up spontaneously, once the readers realize that he deliberately sought to oppose the enemy and fall in battle in order to save Sparta.

5

Surrounded by Pluperfects

As we have noted, the interpretation of the Thermopylae passage as a Spartan aristeia which reveals and extols the heroic characteristics of Sparta’s men (with special focus on the role of their king Leonidas) is supported by the repeated suggestion that the core subject of the passage is death, and by the consistent use of embedded focalization, usually in combination with narrative anticipation (mostly of death). I would now like to point out that salient motifs of the section which are highlighted through these narrative devices (retrieval of the oracle which recommended the death of a Spartan king; planning of the self-sacrifice; mental preparation to be killed) are in some cases associated with the use of a tense which is rather infrequent in discourse, namely the pluperfect. A closer examination reveals that the use of this tense in the whole section is remarkable for three reasons: a) it is used more frequently than in any other part of Herodotus’ work; b) it is repeatedly used to indicate mental states; and c) it is stylistically highlighted through repetition, as in two instances where pluperfect forms are repeated within a few lines. The average frequency of pluperfect forms in 7.201–7.239 is 4,2 instances for every 1000 words (17 instances in 4061 words), whereas the average for the whole work is about 2 forms in every 1000 words. Hence, the pluperfect occurs more than twice as often than in the other parts of Herodotus’ work. No individual book displays a comparable frequency.22 This suggests that the use of

21 22

Cf. Pelling 2006: 94. The next more frequent usage is found in books 7, 8 and 9. In the remaining part of book 7

narrative and identity in thermopylae (herodotus 7.201–7.239)

103

the pluperfect in the narrative on Thermopylae is considerably more important than in any other part of Herodotus’ work. The pluperfect refers to a past action or state (P) and relates it to another event in the past (O) which is posterior to P but anterior to the speech situation. Consequently, a pluperfect form establishes a relationship between two past actions or states. The use of the pluperfect suggests that P functions as background information for O; P became relevant to O at some moment in the past. For the purpose of our analysis, it is not of primary importance whether this information concerns an action which took place before O or a state which lasted (at least) up to the moment of O (but whose exact beginning or cause further in the past does not have to be precisely specified). In the narrative of Thermopylae, a number of pluperfect forms convey information which is focalized by the protagonists. A past event (P) is recollected, directly or indirectly, in order to elucidate someone’s decision or reaction in the story (O). In numerous instances, such occurrences of embedded focalization concern the commanders who play an important role in the narrative, Xerxes and Leonidas. Thus, in 7.208.1 Xerxes’ sending of a scout to spy upon the Greek camp is motivated through a backward reference: ‘for while he [Xerxes] was yet in Thessaly, he had heard (ἀκηκόεε) that some small army was here gathered, and that its leaders were Lacedaemonians, Leonidas a descendant of Heracles among them’.23 What Xerxes had heard about the Spartans, had motivated the sending of the scout. The behaviour of the Spartans will cause the Persian scout to wonder (ἐθώμαζε); thereafter, Xerxes will be the recipient of the latter’s experiences, which are introduced through a new pluperfect (ὀπώπεε, 7.208.3). For the readers, the information conveyed by the pluperfects is not as troubling as it is for Xerxes. Herodotus’ readers are already aware of the Spartans’

23

the average is 3 pluperfect forms in every 1000 words, and in Books 8 and 9 about 2,8. However, the importance of the pluperfect in the Thermopylae passage is greater than these figures suggest, as the high frequency of pluperfects in book 7 is partly due to the regular occurrence of a special category of pluperfects (of which there is only one instance in the Thermopylae narrative). In the remaining part of book 7, a total of 17 pluperfect forms belong to standard expressions used to denote the state of troops which ‘have accomplished the necessary (military) preparations’, ‘have been arrayed’, etc. (ἐτετάχατο, ἐσκευάδατο, ἐστάλατο, κατειλίχατο, ἐσεσάχατο, κεκοσμημένοι ἦσαν, ἐνεπεπορπέατο, ἐνεδεδύκεσαν, etc.). ἀκηκόεε δὲ ἔτι ἐὼν ἐν Θεσσαλίῃ ὡς ἁλισμένη εἴη ταύτῃ στρατιὴ ὀλίγη, καὶ τοὺς ἡγεμόνας ὡς εἴησαν Λακεδαιμόνιοί τε καὶ Λεωνίδης ἐὼν γένος Ἡρακλείδης. Further instances 7.223.1 (ἐπέσταλτο); in secondary clauses: … (ἐγεγόνει); 7.208 (ἔλεγε πρὸς Ξέρξην τά περ ὀπώπεε πάντα); 7.229.1 (ὡς μεμετιμένοι … ἦσαν … καὶ κατεκέατο).

104

tsakmakis

virtue and have also already been informed about Leonidas’ genealogy. In addition, the readers know that Leonidas was admired by the Greeks. This admiration is fully justified by the Spartans’ determination to fight despite their disadvantage in terms of size, even if the readers are provided with more reasons later on in the episode. Having already discussed the function of such motifs in the preceding section, we can contrast our ‘reading’ of the text with Xerxes’ attempt to cope with the same perceptions. The use of pluperfects suggests a temporal interval which was available to Xerxes (and his scout) in order to process the content of his perceptions (what he had heard and what his scout had seen and now reports) and understand the character of their opponents. The king and his spy are not reacting spontaneously to an event; it is suggested that they had enough time in order to cope more successfully with the challenge they faced. And this detail makes Xerxes’ failure to understand Greek mentality even more deplorable. Besides, the king had already been warned by Demaratus. Demaratus had called his attempt to illuminate Xerxes about the Greeks an ἀγὼν μέγιστος, ‘a major struggle’. This figurative expression indicates the difficulties of such an attempt and suggests the distance which separates the worlds of Persians and Greeks. The two nations are separated by a deep cultural gap, and Herodotus’ work highlights this fact by illustrating how the Persians try in vain to understand the Greeks. In his effort to understand Spartan behaviour, Xerxes recurs to the past and retrieves any relevant information. His failure to cope successfully with this aspect of reality proves futile. The narrative illustrates eloquently his limits of comprehension. Yet, in doing so, it also monitors the differences and helps define Greek (and Spartan) identity through its opposition with ‘the other’, Persian mentality. Furthermore, Xerxes’ inability to make the right inferences from experience invites Herodotus’ readers to develop the opposite stance, namely to assess correctly the meaning of the past, i.e. to draw the appropriate conclusions from historical information. Xerxes’ intellectual shortcomings contrast with Leonidas’ cognitive activity as it is presented in the section. The information conveyed by pluperfects is of paramount importance in this respect. First, an essential piece of information, namely the oracle, which is crucial for the interpretation of the whole narrative and the evaluation of its protagonists, is included in an analepsis, in which the pluperfect is used in order to provide background information for Leonidas’ decision to fight in Thermopylae (7.220.3): ‘for when the Spartans enquired of the oracle concerning the war at its very first beginning, the Pythian priestess had prophesied to them that either Lacedaimon should be destroyed by the foreigners, or that its king should perish’ (ἐκέχρηστο γὰρ ὑπὸ τῆς Πυθίης τοῖσι Σπαρτιήτῃσι χρεωμένοισι περὶ τοῦ πολέμου τούτου αὐτίκα

narrative and identity in thermopylae (herodotus 7.201–7.239)

105

κατ’ ἀρχὰς ἐγειρομένου, ἢ Λακεδαίμονα ἀνάστατον γενέσθαι ὑπὸ τῶν βαρβάρων ἢ τὴν βασιλέα σφέων ἀπολέσθαι).24 Here for the first time the readers are offered a concrete explanation for Leonidas’ decision to stay in Thermopylae and to die with his fellow Spartans. The analeptic reference to the past is attached to Herodotus’ narratorial comment that Leonidas intended to die in order to acquire glory and save Sparta’s prosperity. Thus, the passage justifies both Leonidas’ introductory characterization (his high esteem among the Greeks) and Demaratus’ claims about the Spartan’s nomos. The importance of the information conveyed by the verb in the pluperfect will become even more evident in the last chapter of the book, where more details about the past story is made available to the readers. Second, prior to Leonidas’ crucial decision to fight in Thermopylae, the Greek commander was credited with a further important decision which is again explained through a pluperfect form: Leonidas insisted that the Thebans participated in the campaign, because they ‘had been forcefully accused of medizing’ (ὅτι σφέων μεγάλως κατηγόρητο μηδίζειν). Although the Thebans did not participate wholeheartedly, Leonidas’ determined reaction to the allegations was psychologically and practically advantageous, a reaction that illustrates his integrity and tough leadership. In contrast to Xerxes, he is able to correctly assess information and take action accordingly. Pluperfect forms also appear in Leonidas’ first introduction. In a secondary clause, it is reported that Leonidas became a king, because he ‘had been born’ before Cleombrotus (πρότερος ἐγεγόνεε, 7.205.1). Most importantly, Leonidas’ non-expectation of becoming a king is expressed in the introduction with a pluperfect (7.204): ‘for since he had two elder brothers, Cleomenes and Dorieus, he had renounced all thought of the kingship’ (διξῶν γάρ οἱ ἐόντων πρεσβυτέρων ἀδελφεῶν, Κλεομένεός τε καὶ Δωριέος, ἀπελήλατο τῆς φροντίδος περὶ τῆς βασιληίης). This expression introduces a spatial metaphor for the mind, more exactly a martial metaphor for a cognitive process (the same verb will be used with its literal meaning three times in the battle narrative: 7.208.3; 7.211.3; 7.212.2). Thus, the essence of Leonidas’ intellectual activity is foreshadowed with this first of four pluperfects which appear in passages about Leonidas (three of which directly refer to his thoughts). Leonidas’ non-anticipation of a later event here contrasts with (and, therefore, once more foregrounds) his later mental control over the events. The present instance is not to be evaluated as an intellectual failure. It rather confirms the selfless personality of the future king, who will deliberately sacrifice his life (and his kingship) for the sake of the city.

24

The oracle is also analysed in the chapter by de Jong in this volume.

106

tsakmakis

In any case, the focus on Leonidas’ mind foreshadows the dilemmas and the difficult decisions the king will face. Before focusing on some unusual or marked uses of the pluperfect, we will briefly discuss some more standard occurrences, which reinforce the impression that the properties of this tense are amply exploited and combined with some dominant narrative characteristics of the section on Thermopylae. Against all expectations, in the brief digressive narrative on the path which Ephialtes showed to Xerxes, (7.215) the pluperfect is not used in the opening phrase to provide background information for the event which is referred to in the digression (as in the English translation), rather it is used in the concluding sentence: ‘Now this path has been discovered by the Malians of the country, who guided the Thessalians thereby into Phocis, at the time when the Phocians sheltered themselves from attack by fencing the pass with a wall; thus early had the Malians shown that it could avail nothing’ [literally: ‘it had been shown by the Malians’]25 (τὴν δὲ ἀτραπὸν ταύτην ἐξεῦρον μὲν οἱ ἐπιχώριοι Μηλιέες, ἐξευρόντες δὲ Θεσσαλοῖσι κατηγήσαντο ἐπὶ Φωκέας, τότε ὅτε οἱ Φωκέες φράξαντες τείχεϊ τὴν ἐσβολὴν ἦσαν ἐν σκέπῃ τοῦ πολέμου· ἔκ τε τόσου δὴ κατεδέδεκτο ἐοῦσα οὐδὲν χρηστὴ Μηλιεῦσι). The effect of the pluperfect is to make the whole passage appear as background information for the events of the main story line. The tense suggests someone who is located on the temporal level of the main narrative and perceives the ‘uselessness’ of the path as a fact which is still relevant to his time (ἐκ suggests an initial time boundary, but τόσσου adds a durative connotation). For the readers, a question which emerges automatically is whether this situation has remained unchanged or whether a reversal is to be expected. Of course, the readers are equipped with the necessary knowledge to answer this question immediately. In 7.225.1 the use of the pluperfect is accompanied by a signal of a reversal right away: ‘nor was [literally: had been] there an end to this melee till the men with Ephialtes came up’ (τοῦτο δὲ συνεστήκεε μέχρι οὗ οἱ σὺν Ἐπιάλτῃ παρεγένοντο). Here, focalization through an observer who is located in O and perceives the event (arrival of the enemy) which terminated P is suggested by the temporal modifier ‘till’. In both cases, however, the pluperfect indirectly draws the readers’ attention to the narrative peak of the whole section, the final slaughter of the Greeks. It may have become obvious that in the Thermopylae section the pluperfect is regularly used in passages which are internally focalized, explicitly or (as in 25

Stein 1874–1883: ad loc. (followed by Godley in his Loeb translation) takes the pass to be subject of the imperfect. This seems absurd, as it would imply criticism of the Greeks’ strategy and Leonidas’ decision to defend it. Herodotus did not share the views of modern historians.

narrative and identity in thermopylae (herodotus 7.201–7.239)

107

the last two cases discussed) implicitly. Some further remarkable instances will reinforce this impression. As we noted earlier, the pluperfect is regularly used in descriptions of troops to indicate their equipment, physical position, external appearance (see footnote 22). Such a use is also found in the Thermopylae section in the following passage (7.212.2): οἱ δὲ Ἕλληνες κατὰ τάξις τε καὶ κατὰ ἔθνεα κεκοσμημένοι ἦσαν; a further instance occurs in a secondary clause (7.238.2): ‘so they who were [literally: had … been] thus charged did as I have said’ (οἳ μὲν δὴ ταῦτα ἐποίευν, τοῖσι ἐπετέτακτο ποιέειν). Far more frequent in this section, however, are pluperfect forms which are used in the context of pre-battle arrangements of regiments to denote mental states. In two instances, a verb which is commonly used to express military preparations26 is used in the context of mental preparation. Both cases concern the expectation of death. In 7.218.3, the Phoceans who guard the path are supposed to have prepared themselves to die (a salient motif of the narrative), as they notice that the enemy is approaching: ‘and the Phoceans, assailed by showers of arrows, and supposing that it was they whom the Persians had meant from the first to attack, fled away up to the top of the mountain and prepared there to perish’27 (οἱ δὲ Φωκέες ὡς ἐβάλλοντο τοῖσι τοξεύμασι πολλοῖσί τε καὶ πυκνοῖσι, οἴχοντο φεύγοντες ἐπὶ τοῦ ὄρεος τὸν κόρυμβον, ἐπιστάμενοι ὡς ἐπὶ σφέας ὁρμήθησαν ἀρχήν, καὶ παρεσκευάδατο ὡς ἀπολεόμενοι). Against our justified expectations, however, the men will survive, as the Persians show no interest in the Phoceans. This relief, however, is only a temporary one, as the next use of the same pluperfect verb form will not be accompanied by a similarly felicitous end. When the Greeks have to take a final decision, those who stay will be given the same pluperfect: ‘and presently they parted asunder, these taking their departure and dispersing each to their own cities, and those resolving to remain where they had been with Leonidas’ (οἳ μὲν ἀπαλλάσσοντο καὶ διασκεδασθέντες κατὰ πόλις ἕκαστοι ἐτράποντο, οἳ δὲ αὐτῶν ἅμα Λεωνίδῃ μένειν αὐτοῦ παρεσκευάδατο, 7.219.2). It is more than clear that staying means dying. Thus, once more the end of the state expressed by the pluperfect, the event of the main narrative which justifies the choice of tense, will be death. The use of παρεσκευάδατο in two corresponding scenes invites the readers to compare not only the Phoceans’ and the Spartans’ mental preparation to die, but also to relate each preparation to the outcome.28 This will also be the case 26 27

28

Παρεσκευάδατο: 5.65.1; 9.97; 9.100.1; ἐσκευάδατο: 7.62.2; 7.67.1; 7.86.1. The ambiguity between perfective and imperfective is reflected in various translations: e.g. ‘here they were ready to meet death’ (Macaulay) or ‘were preparing themselves with the intention that they would be destroyed’ (Shlomo). What is also interesting is the contrast of this use of the pluperfect with an earlier refer-

108

tsakmakis

in a further pair of corresponding pluperfects to be discussed. In this case, a verb which denotes mental activity and planning will be used in a literal sense, but even this use is not free from ambiguity and pragmatic obscurities. The two verb forms under discussion are used to frame in ring composition the report about the plans of the Greeks who were not present in Thermopylae (7.206.2– 7.207.1): ‘the rest of the allies had been planning [my translation; Godley has had planned] to do the same likewise; for an Olympic festival fell due at the same time as those doings; wherefore they sent their advance guard, not supposing that the war at Thermopylae would so speedily come to an issue. Such had been their intent’ (ὣς δὲ καὶ οἱ λοιποὶ τῶν συμμάχων ἐνένωντο καὶ αὐτοὶ ἕτερα τοιαῦτα ποιήσειν· ἦν γὰρ κατὰ τὠυτὸ Ὀλυμπιὰς τούτοισι τοῖσι πρήγμασι συμπεσοῦσα· οὔκων δοκέοντες κατὰ τάχος οὕτω διακριθήσεσθαι τὸν ἐν Θερμοπύλῃσι πόλεμον ἔπεμπον τοὺς προδρόμους. οὗτοι μὲν δὴ οὕτω διενένωντο ποιήσειν.) It is not clear whether the perfect of the verb (δια)νοέομαι is supposed to denote a state (‘have in mind’, ‘be inclined’, in a similar way as perfects like οἶδα, δέδοικα etc.) or an event (‘have conceived a plan’, ‘have decided’). In any case, however, the choice of pluperfect instead of imperfect (which would also be possible) dissociates the action P denoted by the pluperfect from the main narrative line. Thought and intention are implicitly contrasted in the acuteness of the situation faced by Leonidas and his men, who, therefore, seem to be isolated.29 The remoteness suggested by the choice of the pluperfect (instead of the imperfect), also raises the question (and increases the doubt) whether the intentions expressed by the pluperfect were eventually fulfilled—and if not, what might have caused the reversal of those plans. Was their abandonment deliberate or not? These are questions which will remain pending for a long while and which will prepare the readers for acknowledging the necessity of revisiting the narrative.

29

ence to the Phocaeans (and to their different mission): 7.212.2 οἱ δὲ Ἕλληνες κατὰ τάξις τε καὶ κατὰ ἔθνεα κεκοσμημένοι ἦσαν, καὶ ἐν μέρεϊ ἕκαστοι ἐμάχοντο, πλὴν Φωκέων· οὗτοι δὲ ἐς τὸ ὄρος ἐτάχθησαν φυλάξοντες τὴν ἀτραπόν. ὡς δὲ οὐδὲν εὕρισκον ἀλλοιότερον οἱ Πέρσαι ἢ τῇ προτεραίῃ ἐνώρων, ἀπήλαυνον, ‘but the Greeks stood arrayed by battalions and nations, and each of these fought in its turn, save the Phocians, who were posted on the mountain to guard the path. So when the Persians found the Greeks in no wat different from what the day before had shown them, they drew off from the fight.’ (7.212.2). While we would have expected the usual τετάχαται, the aorist ἐτάχθησαν implies that the narrator is interested in the fact that the men were sent to the path, but avoids moving the camera to the other locality where they were placed. The implication of a second temporal moment would inevitably distract the reader’s attention, an effect which was not intended there. Apart from this narrative, ἐνένωτο occurs in Herodotus only in 1.77, where it is used for a cancelled intention. In the same context ὀρμέατο (1.83) is also used.

narrative and identity in thermopylae (herodotus 7.201–7.239)

109

Finally, we will discuss in some detail two occurrences of the same verb form in two subsequent episodes. Both refer to the state of public contempt imposed upon two Spartans who failed to die in Thermopylae: ‘When Aristodemus returned to Lacedaemon, he was disgraced and dishonoured; this had been the manner of his dishonour, that no Spartan would give him fire, nor speak with him; and they called him for disgrace, Aristodemus the coward. But he repaired all that was laid to his charge in the fight at Plataeae’ (ἀπονοστήσας δὲ ἐς Λακεδαίμονα ὁ Ἀριστόδημος εἶχε ὄνειδός τε καὶ ἀτιμίην· πάσχων δὲ τοιάδε ἠτίμωτο· οὔτε οἱ πῦρ οὐδεὶς ἔναυε Σπαρτιητέων οὔτε διελέγετο. ὄνειδος δὲ εἶχε ὁ τρέσας Ἀριστόδημος καλεόμενος. ἀλλ’ ὃ μὲν ἐν τῇ ἐν Πλαταιῇσι μάχῃ ἀνέλαβε πᾶσαν τὴν ἐπενειχθεῖσαν αἰτίην, 7.231–7.232). The verb form ἠτίμωτο can express both the action (passive) and the state (medium) of ‘being deprived’ of public rights.30 Here the stative meaning is clearly prevalent, as the verb fills in the details of an event mentioned with an imperfect (εἶχε ὄνειδός τε καὶ ἀτιμίην) and is followed by three more imperfects (formally responding to the attached participle πάσχων τοιάδε). The reason why the pluperfect ἠτίμωτο is preferred here instead of an imperfect is presumably that the state which is denoted by the verb is at some stage terminated, and this expectation has to be suggested by the verb form, which, again, separates the past state P from the main narrative line O. In the immediately following episode (7.232) a further instance of the same pluperfect is found: λέγεται δὲ καὶ ἄλλον ἀποπεμφθέντα ἄγγελον ἐς Θεσσαλίην τῶν τριηκοσίων τούτων περιγενέσθαι, τῷ οὔνομα εἶναι Παντίτην· νοστήσαντα δὲ τοῦτον ἐς Σπάρτην, ὡς ἠτίμωτο, ἀπάγξασθαι (‘it is said too that another of the three hundred, whose name was Pantites, was saved alive, carrying a message into Thessaly; he also returned to Sparta, but being there dishonoured hanged himself’). The repetition of the verb forms links the two episodes and invites the readers to compare them. Thus, the readers cannot fail to realize that survivors had to suffer dishonour in Sparta, but also that this state produced further results, namely death (the salient motif of the narrative), though in a different way. This confirms the power and inescapability of the Spartan nomos.31 On the other hand the death of the two men is essentially different. Although the first survivor was not a deserter but had been dismissed because of his illness, his wish to die led him to fight bravely and fall at the next opportunity offered to him. His case becomes a foil for the evaluation of the second man. This is 30 31

On the contempt against survivors in Sparta, see Powell 1988: 237–238. For the significance of the nomos, see Branscome 2013: 69–71. Baragwanath 2008: 76 advocates a different understanding of the episodes, namely that they expose the nomos as ‘exaggerated’.

110

tsakmakis

how the nomos ultimately works: it expects from everybody the highest degree of personal responsibility. Reality itself shows that its standards are realistic. According to Herodotus, this is Sparta—just as modern versions of the legend: dying as responsible humans, not killing as superheroes.

6

Conclusion

In 7.139 Herodotus had expressed his opinion that even if all Spartans had died defending their city and Greece, their sacrifice would have been useless if the Athenians had abandoned Athens and had not fought in Salamis. The salvation of Greece, according to the historian, is primarily due to Themistocles’ military plan. On the whole, Herodotus’ narrative of the Persian Wars highlights two outstanding exploits of the Athenians: Marathon, from which Sparta was absent, and Salamis, where the battle was won thanks to Themistocles. Sparta’s main contribution was Pausanias’ leadership in Plataeae, where both Spartan and Athenian troops excelled. The lost battle of Thermopylae restores the balance between the two major cities of Greece by displaying a Spartan aristeia. Herodotus denies the Persians the satisfaction of victory but gives the Greeks the satisfaction of glory. The Persians do not fight bravely and cannot overpower the Greeks until they are offered help by Ephialtes. They are even incapable of understanding the behaviour and the mentality of their opponents. Greek values remain incomprehensible to them. Thus, through the collision with ‘the other’, the Thermopylae passage becomes the place in Herodotus’ work where the Greek character, and above all, the Spartan identity are defined. The account of the battle is a complex, an artful narrative which glorifies Sparta and Leonidas. A military defeat of the Greeks is treated as an opportunity to demonstrate their moral superiority. Narrative devices contribute essentially to this effect. The central theme of the narrative is repeatedly presented via the embedded focalization of characters: it is death—not fighting, victory etc., as one would reasonably expect. At the same time, anachronies contribute to the framing of the events in a way which relates them to the final victory of the Greeks. The last piece of information provided to the readers refers to events prior to any related event of the narrative, and links Leonidas’ death to the prosperous outcome of Sparta’s defence against the Persians. Both focalization effects and temporal manipulation are enhanced through the use of the pluperfect, which occurs more frequently than in any other passage of Herodotus. On the whole, the coherence of thematic, narrative and linguistic elements is an indication of the highly elaborate character of the narrative, which, even

narrative and identity in thermopylae (herodotus 7.201–7.239)

111

if it perhaps reflects traditions about Sparta of Herodotus’ own time, distinguishes itself as a piece of almost poetic—epic—historiography.

Bibliography Asheri, D., ‘General Introduction’, in O. Murray & A. Moreno (eds.), A Commentary on Herodotus Books I–IV (Oxford 2007) 1–56. Baragwanath, E., Motivation and Narrative in Herodotus (Oxford 2008). Boedeker, D., ‘The Two Faces of Demaratus’, in D. Boedeker & J. Peradotto (eds.), Herodotus and the Invention of History, Arethusa 20.1/20.2 (1987) 185–201. Branscome, D., Textual Rivals. Self-Presentation in Herodotus’ Histories (Ann Arbor 2013). Chantraine, P., Histoire du parfait grec (Paris 1926). Corcella, A., ‘Erodoto VII, 239: Una “interpolazione d’autore”?’, ASNP 15.2 (1985) 313–491. Dillery, J., ‘Reconfiguring the Past: Thyrea, Thermopylae and Narrative Patterns in Herodotus’, AJPh 117.2 (1996) 217–254. Ferrario, S. Brown, Historical Agency and the ‘Great Man’ in Classical Greece (Cambridge 2014). Forsdyke, S., ‘Greek History, ca. 525–480 BC’, in E.J. Bakker, I.J.F. de Jong & H. van Wees, (eds.), Brill’s Companion to Herodotus (Leiden 2002) 521–549. Godley, A.D., Herodotus, with an English Translation by A.D.G., III, Books V–VII (London 1938). Hooker, J.T., ‘Spartan Propaganda’, in A. Powell (ed.), Classical Sparta: Techniques behind her Success (Oklahoma 1989) 122–141. Hunzinger, C., ‘La notion de thôma chez Hérodote’, Ktèma 20 (1995) 47–70. Lombardo, M., ‘Erodoto sulle Termopili: Leonida, Demarato e l’ideologia Spartiata’, in Giangiuglio, M. (ed.), Erodoto e il “modello” erodoteo (Trento 2005) 173–192. Mikalson, J.D., Herodotus and Religion in the Persian Wars (Chapel Hill 2003). Möller, A., ‘The Beginning of Chronography: Hellanikus’ Hiereiai’, in N. Luraghi (ed.), The Historian’s Craft in the Age of Herodotus (Oxford 2001) 241–262. Munson, R.V., ‘Three Aspects of Spartan Kingship in Herodotus’, in R.M. Rosen & J. Farrell (eds.), Nomodeiktes. Greek Studies in Honour of Martin Ostwald (Ann Arbor 1993) 39–54. Nagy, G., ‘Herodotus the Logios’, in D. Boedeker & J. Peradotto (eds.), Herodotus and the Invention of History, Arethusa 20.1/20.2 (1987) 175–184. Paradiso, A., ‘Gorgô et les manipulations de la fonction’, in S. Boehringer & V. Sebillote Cuchet (eds.), Des femmes en action: l’individu et la fonction en Grèce antique (Paris 2013) 39–51. Pelling, C.B.R., ‘Homer and Herodotus’, in M.J. Clarke, B.G.F. Currie & R.O.A.M. Lyne,

112

tsakmakis

(eds.), Epic Interactions: Perspectives on Homer, Virgil, and the Epic Tradition Presented to Jasper Griffin by Former Pupils (Oxford 2006) 75–104. Pomeroy, S., Spartan Women (Oxford 2002). Powell, A., Athens and Sparta. Constructing Greek Political and Social History from 478BC ([1988] London 2008). Powell, J.E., A Lexicon to Herodotus (Cambridge 1938). Smith, C.S., Modes of Discourse. The Logical Structure of Texts (Cambridge 2003). Stein, H., Herodotos, erklärt von H.S., 5 vols. (Berlin3–51874–1883). Vaniccelli, P., Resistenza e intesa (Bari 2013). Wallinga, H.T., Xerxes’ Greek Adventure. The Naval Perspective (Leiden 2005). Zali, V., The Shape of Herodotean Rhetoric: A Study of the Speeches in Herodotus’ Histories with Special attention to Books 5–9 (Leiden 2015).

chapter 5

Herodotus’ Handling of (Narratological) Time in the Thermopylae Passage Irene de Jong

1

Introduction: (Narratological) Time and the Structure of the Histories*

A hot issue in Herodotean scholarship is the structure, or rather the apparent lack of structure, of his Histories. Scholars have been troubled by the ‘möglichst bunte Folge Assoziationen unsachlicher Art’1 and the ‘Zerrissenheit und Vielspaltigkeit dieses Geschichtswerkes, seiner aus ungezählten Wirklichkeitspartikeln mosaikartig zusammengestuckelten Erscheinungsform’,2 which means that at first sight the narratees feel ‘buried under an avalanche of facts and at the same time utterly lost in a landscape bewilderingly criss-crossed and looped by stories without discernible paths or sense of structured connection’.3 Various explanations have been proposed to account for this situation: the storyteller Herodotus’ penchant for entertaining his narratees which would make him insert anecdotes even if they are unrelated to his main story;4 the genesis of his work as a series of independent logoi;5 or his archaic mentality which simply is not interested in or capable of a coherent story.6 Fortunately, there have also been more appreciative responses to Herodotus’ way of structuring his tale. Scholars have argued that the Histories achieves unity through analogy, which means that episodes are—paradigmatically— connected via their use of the same recurrent story-patterns or themes, such as

* I would like to thank Mathieu de Bakker and Caroline Kroon for their comments, Nina King for checking my English. 1 Howald 1923: 128. 2 Focke 1927: 47. 3 Gould 1989: 42. Cf. Thomson 1935: 228, who compares the reader to a voyager on a ship on the river of Herodotus’ Histories; and Lateiner 1989: 13. 4 See e.g. Fränkel [1924] 1960: 87; Von Fritz 1967: 450; and Griffiths 2006: 132. Dionysius of Halicarnassus in his Pomp. 3 also speaks of Herodotus introducing stories ‘to add charm to the narrative’, but on the whole considers Herodotus’ Histories a unity. 5 See e.g. Jacoby 1913. 6 See e.g. Fränkel [1924] 1960; and Focke 1927.

© koninklijke brill nv, leiden, 2019 | doi:10.1163/9789004383340_006

114

de jong

the crossing of a geographical boundary by an oriental conqueror, an inquiring king, the rise and fall of a ruler or the tragic warner.7 Others have pointed out how personal relationships, such as kinship, guestfriendship and revenge, may link episodes in syntagmatic or linear fashion.8 Finally, Herodotus’ structure has been looked at in terms of his handling of time. Dionysius of Halicarnassus in his Letter to Pompey 3 already notes that Herodotus relates ‘some events as a sequel, taking up others (προσαναλαμβάνων) as missing links in the story’ but never breaks ‘the continuity of the narrative’. Dionysius here more or less anticipates the modern narratological notion of the analepsis, when a narrator recounts an event not at its proper chronological place but at a later moment in the form of a flashback. The Histories abound in such analepses and to a lesser degree prolepses9 and this led me, in a study from 2001, to call its structure anachronical: a largely chronological main story is interrupted by anachronies, prolepses which create (a forward looking) tension and especially analepses which provide (backward looking) background information; between them these anachronies create cohesion.10 To be sure, the three structural devices just listed, analogy, personal relationship, and time, often operate at the same time. In this chapter I will analyse Herodotus’ handling of time in the episode of Thermopylae (7.175–7.239). Obviously I cannot discuss all prolepses and analepses, and I thus have selected five examples that will make clear in particular, I hope, the benefits of a narratological approach (with a strong linguistic basis) for the interpretation of the Histories.11

7

8 9

10

11

E.g. Immerwahr 1966 and Munson 2001: 45–73, esp. 46 (‘comparison and analogy are fundamental strategies by which the text of the Histories organizes its material. Because the logos contains so many story elements that escape the network of causal connections of the plot, classification and the comparative approach that classification entails provide a powerful glue’). E.g. De Romilly 1971–1972; Gould 1989: 42–62; and Lateiner 1989: 126–144. See Pohlenz 1937: 69–73, 83–88; Gould 1989: 64–65 (‘the narrative mode of explanation’); Lateiner 1989: 114–125; and Grethlein 2010: 196–204. For analepses, see Huber 1965: 96–99; Lang 1984: 6–7; for prolepses Van Groningen 1953: 39–42. On analeptic explanation as a hallmark of all Greek historiography, see Hau 2014: 254–257. De Jong 2001. The terms anachrony, prolepsis and analepsis derive from Genette [1972] 1980: 35–36. Together they belong to the aspect of narratological time that is called order: the order in which events are presented in a story often differs from the (reconstructed) chronological order of the fabula; see De Jong 2014: 78–87. This chapter presents a preliminary version of sections from a Narratological commentary on Herodotus Histories, which I am working on at present.

herodotus’ handling of (narratological) time

2

115

The Prolepsis about the Anopaea Path

Thermopylae has barely been mentioned for the first time, as the place where the Greeks decide to make a stand against the Persians, when the narrator says: τὴν δὲ ἀτραπόν, δι’ ἣν ἥλωσαν οἱ ἁλόντες Ἑλλήνων ἐν Θερμοπύλῃσι, οὐδὲ ᾔδεσαν ἐοῦσαν πρότερον ἤ περ ἀπικόμενοι ἐς Θερμοπύλας ἐπύθοντο Τρηχινίων. 7.175.2

And as regards the path by which those of the Greeks who were trapped and slain in Thermopylae were trapped and slain,12 they did not know that it existed before, arriving in Thermopylae, they heard about it from the Trachinians.13 When one consults the commentaries on this passage, Stein writes ‘ἥλωσαν οἱ ἁλόντες, zur Sache c. 213ff.’; Macan grumbles that ‘the statement here is one of those very hard to stomach’ since Greeks would know that there always is a way round a mountain, and suggests that it is an ‘apologetic note’ by which Herodotus excuses the defeated Greeks; and How-Wells have no comment at all. Looking at the sentence with a narratological eye, we may observe that we are dealing with two prolepses. In the first place, the narrator, right at the start of the Thermopylae episode, reveals its outcome: the Greeks will be defeated (the prolepsis, as often, is in the past tense: ἥλωσαν οἱ ἁλόντες Ἑλλήνων ἐν Θερμοπύλῃσι). The narrator will recount this defeat in his chapters 7.213–7.233, but his narratees already seem to know how the Greeks are going to be defeated, since he refers to the Anopaea path, even though it is mentioned here for the first time, as ‘the (well-known or notorious) path by which the Greeks were trapped and slain’. The function of this prolepsis is to cast a tragic light on all subsequent actions of the Greeks, who are doomed from the start but do not themselves know this (I will come back to this point in section 5). It also mitigates the

12

13

The verb ἁλίσκομαι in Herodotus, when used of persons, means ‘being caught’. Since the Greeks are not just trapped but killed, and in view of the parallel in 7.213.1 (Ephialtes, showing the Anopaea path to Xerxes, διέφθειρε τοὺς ταύτῃ ὑπομείναντας Ἑλλήνων, ‘destroyed the Greeks who stayed there’), I take it here also to have its Homeric meaning of ‘being slain’ (cf. e.g. Il. 17.506). There is one parallel for this combined meaning: in 1.191.6 the narrator refers to the Babylonians who are ‘caught and killed’ (ἑαλωκότων … ἑαλωκότας) by the Persians, who invade their city via a channel of the Euphrates. The text is that of Hude (OCT), translations are my own.

116

de jong

(potential) dishonour of defeat by indicating that the Greeks are worsted not because of some lack of courage but because of a treacherous path. A second prolepsis, linguistically marked by the use of a temporal expression (οὐδὲ … πρότερον ἤ περ), announces that the Greeks will find out about the path when they arrive in Thermopylae. This is a completing prolepsis14 in that this moment is not recorded in the main story. In chapter 7.212.2 we simply hear that the Phocians ‘were stationed on the mountain in order to guard the path’, the moment of the Trachinians telling the Greeks about its existence having been passed over by the narrator. The function of this second prolepsis is to excuse the Greeks for choosing a strategic position which will turn out not to be a watertight one.15

3

The Analepses about the Phocian Wall

The narrator goes on to describe the Thermopylae pass and mentions a wall next to its natural scenery: ἐδέδμητο δὲ τεῖχος κατὰ ταύτας τὰς ἐσβολάς, […] (4) ἔδειμαν δὲ Φωκέες τὸ τεῖχος δείσαντες, ἐπεὶ Θεσσαλοὶ ἦλθον ἐκ Θεσπρωτῶν οἰκήσοντες γῆν τὴν Αἰολίδα, τήν περ νῦν ἐκτέαται. ἃτε δὴ πειρωμένων τῶν Θεσσαλῶν καταστρέφεσθαί σφεας, τοῦτο προεφυλάξαντο οἱ Φωκέες. […] (5) τὸ μέν νυν τεῖχος τὸ ἀρχαῖον ἐκ παλαιοῦ τε ἐδέδμητο καὶ τὸ πλέον αὐτοῦ ἤδη ὑπὸ χρόνου ἔκειτο. τοῖσι δὲ αὖτις ὀρθώσασι ἔδοξε ταύτῃ ἀπαμύνειν ἀπὸ τῆς Ἑλλάδος τὸν βάρβαρον. 7.176.3–7.176.5

a wall had been built across that entry, […] (4) The Phocians had built the wall for fear of the Thessalians, when they came from Thesprotia to dwell in the Aeolian land that they now possess. Because the Thessalians were trying to subdue them, the Phocians used that wall for their protection. […] (5) The ancient wall had been built long ago and its larger part was already lying in ruins because of the passing of time. The Greeks decided to rebuild it again and at that place keep off the barbarian from Greece.

14

15

A completing prolepsis (or analepsis) means that the prolepsis recounts an event not recorded in the main story (and hence completes that main story), as opposed to a repeating prolepsis (or analepsis) which recounts an event also recorded in the main story (and hence repeats that main story); see De Jong 2014: 81–82. I thus agree with Macan 1908: ad loc. that this second prolepsis has an apologetic undertone.

herodotus’ handling of (narratological) time

117

Commentators note that the ruins of this wall are still to be seen and further content themselves with a ‘cf.’ (How-Wells: ‘For the wall cf. ch. 7.208, 7.223, 7.225’)16 or an ultra brief comment (Stein: ‘ἔκειτο: sie hatte sich als unzureichend erwiesen (c. 215)’). I suggest that much more is to be made of Herodotus’ mention of the wall at this moment in his story. The rebuilt wall will play a role in the ensuing battle (chapters 7.223 and 7.225). But why would the narrator take the trouble to recount its history in an analepsis, linguistically marked by pluperfects at the beginning and end (ἐδέδμητο)17 and past-in-the-past aorists (ἔδειμαν, προεφυλάξαντο)?18 Perhaps alert narratees would already now sense that there is an analogy between what happened in the past and what is happening in the present of the main story: the Phocians built a wall to keep off invading Thessalians, just as the Greeks now rebuild that wall to keep off invading Persians. The relevance of what happened in the past becomes clear in a second analepsis, inserted at the moment when Ephialtes betrays the Anopaea path to Xerxes: τὴν δὲ ἀτραπὸν ταύτην ἐξεῦρον μὲν οἱ ἐπιχώριοι Μηλιέες, ἐξευρόντες δὲ Θεσσαλοῖσι κατηγήσαντο ἐπὶ Φωκέας, τότε ὅτε οἱ Φωκέες φράξαντες τείχεϊ τὴν ἐσβολὴν ἦσαν ἐν σκέπῃ τοῦ πολέμου· ἔκ τε τοσοῦδε κατεδέδεκτο ἐοῦσα οὐδὲν χρηστὴ Μηλιεῦσι. 7.215

That path had been discovered by local Malians, and having discovered it they had guided the Thessalians against the Phocians, at that time when the Phocians having fenced off the pass with a wall guarded themselves against war. So long ago it [the path] had been shown by the Malians to be pernicious. This analepsis, linguistically marked by past-in-the-past aorists (ἐξεῦρον, κατηγήσαντο), a temporal adjunct (τότε), and a pluperfect (κατεδέδεκτο), primarily concerns the Anopaea path but it also hints at the role of the wall in the ensuing

16 17

18

For an excellent discussion of the notoriously vague ‘cf.’ of commentaries, which in fact can refer to many different things, see Gibson 2002. Cf. Macan 1908: ad loc.: ‘one of the most genuine pluperfects, temporally, in Hdt., for it is related not to the date of writing but to the date given in the narrative’. For the pluperfects in the Thermopylae episode, see also the chapter of Tsakmakis in this volume. For the past-in-the past interpretation of the aorist, see Rijksbaron 2002: 20.

118

de jong

events: just as Malians in the past showed the Anopaea path to the Thessalians and thus, it is to be assumed, invalidated the defensive potential of the wall,19 so now the Malian Ephialtes reveals the path to the Persians and, the narratees understand, will cause the same effect.20 Indeed, in the final battle the Greeks, at the moment they see the Persians coming down from the path, will withdraw behind the wall but the Persians attacking them in front ‘throw down the defensive wall’ or surround them (7.225) and kill them all. The two analepses about the wall are linked to the main story by analogy, which gives them a foreshadowing force. We see two of the structural devices mentioned in my introduction working in conjunction.

4

The Analepsis about Leonidas’ Kingship

Having set the stage where the battle will take place the Herodotean narrator goes on to list the Greek contingents that await Xerxes, and their generals. The other generals are dealt with in summary fashion (‘all these contingents had their generals, each one his own’), but the commander-in-chief Leonidas is introduced at length: […], ὁ δὲ θωμαζόμενος μάλιστα καὶ παντὸς τοῦ στρατεύματος ἡγεόμενος Λακεδαιμόνιος ἦν Λεωνίδης ὁ Ἀναξανδρίδεω τοῦ Λέοντος τοῦ Εὐρυκρατίδεω […] τοῦ Ἡρακλέος, κτησάμενος τὴν βασιληίην ἐν Σπάρτῃ ἐξ ἀπροσδοκήτου. (7.205.1) διξῶν γάρ οἱ ἐόντων πρεσβυτέρων ἀδελφεῶν, Κλεομένεός τε καὶ Δωριέος, ἀπελήλατο τῆς φροντίδος περὶ τῆς βασιληίης. […] οὕτω δὴ ἐς Λεωνίδην ἀνέβαινε ἡ βασιληίη, […]. 7.204–7.205.1

[…], he that was most admired and was the leader of the whole army was the Spartan Leonidas, the son of Anaxandrides, the son of Leon, the son of Eurycratides, […], the son of Heracles, who had obtained the kingship in Sparta unexpectedly. (7.205.1) For since he had two older brothers, Cleomenes and Dorieus, he had renounced all thought to be-

19

20

The Herodotean narrator refers again to the Thessalian invasion of Phocis in another analepsis in 8.27–8.28 (cf. ἐσβαλόντες … ἐς τοὺς Φωκέας), and there it is clear that it was successful. I follow Macan 1908: ad loc. and How & Wells [1912] 1928: ad loc. in taking οὐδὲν χρηστή to mean ‘in no way useful’= ‘pernicious’ and as referring to the path. Stein [1889] 1908: ad loc. takes it as ‘not useful’ and as referring to the pass.

herodotus’ handling of (narratological) time

119

come king. [But both brothers died young.] And so the kingship devolved onto Leonidas, […]. All commentators note that Herodotus’ insertion at this place of Leonidas’ long and distinguished genealogy serves as a mark of honour. But why does the narrator include an analepsis about how Leonidas ‘unexpectedly’ had become king of Sparta, linguistically marked by the push particle γάρ21 and a pluperfect (ἀπελήλατο)?22 As so often, the relevance of an analepsis remains implicit and has to be supplied by the narratees themselves. In this case that relevance becomes clear only later, but then is, I think, unmistakable for narratees who are steeped in Herodotus’ way of thinking, especially his conviction that human fortune is never stable (cf. for instance 1.5, 1.32). For in chapter 7.220 it will turn out that there was an oracle which prescribed that for Xerxes to be (eventually) stopped, either Sparta or its king, significantly referred to in genealogical terms which recall the present passage (ἀφ’ Ἡρακλέους … γενέθλης … βασιλῆ), must perish. King Leonidas has no option but to die for his country. Thus, if his succession to the throne at first may have seem an unexpected boon for him, it will later be seen to bring with it grave consequences.23 Perhaps it is for this reason that we find the imperfect ἐς Λεωνίδην ἀνέβαινε ἡ βασιληίη rather than an aorist, such as we find in a closely similar context: ἀποθανόντος δὲ Δαρείου ἡ βασιληίη ἀνεχώρησε ἐς τὸν παῖδα τὸν ἐκείνου Ξέρξην, ‘when Darius died the kingship went over to his son Xerxes’ (7.4). The imperfect suggests that we have not heard the last of Leonidas’ kingship.24

21 22

23

24

See Slings 1997. A push particle is a particle that marks a speaker entering an embedded sequence, as against a pop particle that marks his return to the embedding sequence. This analepsis complements an earlier one at 5.39–5.48, where the narrator had told how the Spartan king Anaxandrides had four sons with two different wives and how the first two (Cleomenes and Dorieus) died. Only now does he tell that it was thus the third son, Leonidas, who unexpectedly became king. A different evaluation of the ‘unexpectedly’ is given by Baragwanath 2008: 65: ‘The story of his succession, displaced from a natural position at 5.41 to 7.204, so as to introduce the Thermopylae narrative, is […] idealizing: removing Leonidas from the ranks of the many in the Histories we have witnessed wrangling egotistically for power, it presents him as one to whom kingship came unexpectedly (ἐξ ἀπροσδοκήτου, 7.204). This in turn heralds the narrative strand portraying Leonidas as servant of his country, which later culminates with the account of his conscious decision—quite Hector- or Achilles-like—to sacrifice himself, in his case for Sparta’s sake.’ See also the chapter of Tsakmakis in this volume. For the discourse function of the Greek imperfect, which raises expectations that more is to follow, see Rijksbaron 1988. In his response to the oral delivery of this chapter, Rijksbaron pointed out that at 7.4 we find an imperfect as variant reading (ἀνεχώρεε), in my view an inferior reading, since Xerxes’ kingship will not recur as a topic of discussion.

120 5

de jong

The Prolepsis about Ephialtes

The Persians attack the Greek contingents twice and try to force them away from the Thermopylae pass but are not successful. Now Ephialtes enters the scene and reveals the Anopaea path to Xerxes, thus sealing the fate of the Greeks. At this climactic moment the narrator interrupts his story: ὕστερον δὲ δείσας Λακεδαιμονίους ἔφυγε ἐς Θεσσαλίην, καί οἱ φυγόντι […] ἀργύριον ἐπεκηρύχθη. χρόνῳ δὲ ὕστερον, κατῆλθε γὰρ ἐς Ἀντικύρην, ἀπέθανε ὑπὸ Ἀθηνάδεω ἀνδρὸς Τρηχινίου· […] Ἐπιάλτης μὲν οὕτω ὕστερον τούτων ἀπέθανε. 7.213.2

Later he [Ephialtes] fled to Thessaly, fearing the Spartans, and while in exile a price was put on his head. Much later, having returned to Anticyra, he died by the hand of a Trachinian called Athenades. […] In such a manner Ephialtes later died. In this explicitly marked prolepsis (thrice featuring the temporal adjunct ὕστερον) the narrator recounts, at the very moment of his successful intervention with Xerxes, Ephialtes’ future fate: he first flees to Thessaly but upon his return home is killed. There are more prolepses in the Histories that look ahead to the punishment of characters at the moment they perform their crime or to their death at a moment of triumph.25 In the present case, there is additional satisfaction in the way Ephialtes will die in that his murderer is attracted by a ransom put on his head, just as Ephialtes himself betrayed the Greeks because he expected, and undoubtedly got, a reward from Xerxes (7.213.1). The narrator might have recounted the later fate of Ephialtes at the end of the Thermopylae passage, when he reports what happened to two Spartans who survived the battle (7.229–7.232). But for someone who is keen on showing the principle of the reversal of fortune at work in history it is of course much more effective to anticipate Ephialtes’ end now, at the height of his ‘success’.26

25 26

See e.g. 6.66; 6.71–2; 7.7; and 9.37.4–38.1, and Rood 2007: 127. The same device (but with a tragic rather than moralistic undertone) is found in Homer, e.g. Il. 16.799–16.800: at the height of Hector’s success (the moment he slays Patroclus) the narrator anticipates his imminent death.

herodotus’ handling of (narratological) time

6

121

The Analepsis/Prolepsis of the Oracle

The Greeks detect that they are surrounded by the Persians and most of the contingents now decide to leave. Leonidas chooses to stay on the basis of an oracle: (1) λέγεται δὲ ⟨καὶ⟩ ὡς αὐτός σφεας ἀπέπεμψε Λεωνίδης, μὴ ἀπόλωνται κηδόμενος· αὐτῷ δὲ καὶ Σπαρτιητέων τοῖσι παρεοῦσι οὐκ ἔχειν εὐπρεπέως ἐκλιπεῖν τὴν τάξιν ἐς τὴν ἦλθον φυλάξοντες ἀρχήν. (2) ταύτῃ καὶ μᾶλλον τὴν γνώμην πλεῖστός εἰμι· Λεωνίδην, ἐπείτε ᾔσθετο τοὺς συμμάχους ἐόντας ἀπροθύμους καὶ οὐκ ἐθέλοντας συνδιακινδυνεύειν, κελεῦσαί σφεας ἀπαλλάσσεσθαι, αὐτῷ δὲ ἀπιέναι οὐ καλῶς ἔχειν· μένοντι δὲ αὐτοῦ κλέος μέγα ἐλείπετο, καὶ ἡ Σπάρτης εὐδαιμονίη οὐκ ἐξηλείφετο. (3) ἐκέχρηστο γὰρ ὑπὸ τῆς Πυθίης τοῖσι Σπαρτιήτῃσι χρεωμένοισι περὶ τοῦ πολέμου τούτου αὐτίκα κατ’ ἀρχὰς ἐγειρομένου, ἢ Λακεδαίμονα ἀνάστατον γενέσθαι ὑπὸ τῶν βαρβάρων, ἢ τὸν βασιλέα σφέων ἀπολέσθαι. (4) [oracle in original hexameters]. ταῦτά τε δὴ ἐπιλεγόμενον Λεωνίδην καὶ βουλόμενον κλέος καταθέσθαι μούνων Σπαρτιητέων, ἀποπέμψαι τοὺς συμμάχους μᾶλλον ἢ γνώμῃ διενειχθέντας οὕτω ἀκόσμως οἴχεσθαι τοὺς οἰχομένους. 7.220.1–7.220.4

(1) It is ⟨also⟩ told that Leonidas himself sent them away, because he cared for them not to be killed. But for himself and the Spartans present (he said) it was not seeming to leave the post which they had come to guard in the first place. (2) I am very much of that same opinion. When Leonidas perceived the allies to be faint of heart and not willing to run risks with him, he ordered them to go but (said that) for him to go was not honourable. When he [Leonidas] would stay at his post, great glory would be left behind and the good fortune of Sparta would not be blotted out. (3) For it had been prophesied by the Pythia to the Spartans, when they consulted her about that war right after it had started, that either Sparta would be destroyed by the barbarians or their king would die. [oracle in original hexameters] (4) (It is said that) taking into consideration that oracle and wanting to lay up a store of glory only for the Spartans Leonidas sent away the allies rather than let them go away in disarray because they differed in opinion. How-Wells write in their commentary: It is clear that H. in these chapters aims at excusing the allies for deserting Leonidas by explaining that his death was fated by heaven and foretold by

122

de jong

the oracle […] it was a convenient excuse for all concerned, for the Athenians who had urged pushing forward the line of defence to Artemisium and Thermopylae, for the Spartans who had sent but inadequate support to their heroic king, and for the Peloponnesian allies who had failed him in the hour of trial. But the oracle is plainly a vaticinium post eventum. The idea that the oracle is a vaticinium post eventum is embraced by most scholars and this may well be true.27 But since Herodotus has chosen to present the oracle in his text, we must ask ourselves what its function is within his narrative of Thermopylae. Let us first take a closer look at how the oracle is presented. In the preceding chapter Herodotus had briefly told how the Greeks, having found out that they are surrounded by the Persians, hold a council in which most allies decide to leave Thermopylae (7.219.2). He now presents an alternative version of that council,28 which he knows from reported narrators (λέγεται), in which it is Leonidas himself who ordered most allies to go away. He emphatically endorses the alternative version (ταύτῃ καὶ μᾶλλον τὴν γνώμην πλεῖστός εἰμι) and then recounts the deliberations of the council once again, now in the new version: when Leonidas sensed that his allies were afraid he ordered them to go but (said) that for him to leave was not honourable. It is not clear whether the AcI-construction Λεωνίδην … κελεῦσαί is the content of the alternative version (that is, depends on λέγεται) or of the narrator’s opinion (that is, depends on τὴν γνώμην πλεῖστός εἰμι), the ambiguity actually expressing how much the two coincide and how much Herodotus endorses that version. In the sentence that follows, ‘When he would stay at his post, great glory would be left behind (ἐλείπετο) and the good fortune of Sparta would not be blotted out (ἐξηλείφετο)’, the narrator enters Leonidas’ mind and tells us what he thought (but did not say: see below). A linguistic argument in favour of this analysis in terms of embedded focalization is the use of the imperfects ἐλείπετο and ἐξηλείφετο: ‘So überlegte nach H.’s Meinung der König; daher das Imperfekt’ (Stein, followed by Macan).29

27 28 29

Cf. also Macan 1908: ad loc. and Legrand 1963: 189–190, who contends that the oracle serves to exculpate Leonidas for his decision to sacrifice 300 men. As he does so often, see Groten 1963; Lateiner 1989: 76–90; and Marincola 1997: 280– 286. We are dealing with implicit embedded focalization when there is no verb of seeing, thinking, feeling, or speaking but the focalization of a character can still be argued on the basis of the presence of evaluative words, particles, or the use of tenses and moods,

herodotus’ handling of (narratological) time

123

The idea that staying at his post would bring Leonidas glory is not strange, but that it also means that ‘Sparta’s good fortune would not be blotted out’ is puzzling. The narrator thus inserts, for the benefit of his narratees, a (completing) analepsis about an oracle that explains Leonidas’ reasoning. The analepsis is linguistically marked by the push particle γάρ and the use of a pluperfect (ἐκέχρηστο). It recounts how the Spartans had received an oracle ‘at the start of the war’, when Xerxes decided to invade Greece,30 which indicated that either Sparta was to be destroyed or its king had to die since Xerxes would not be stopped before destroying one of the two. At the end of the analepsis the narrator returns to his main story via Leonidas’ embedded focalization (ταῦτά ἐπιλεγόμενον καὶ βουλόμενον …). In this way he smoothly integrates the analepsis into the story: it begins as a piece of information inserted for the narratees but ends as the content of Leonidas’ thoughts.31 This is a technique often to be observed in the Histories: background information is first presented as a kind of analeptic footnote by the narrator but at the end is revealed to be (also) the focalization of one of the characters. In this way potentially digressive elements are firmly integrated into the narrative after all.32 Leonidas’ embedded focalization after Herodotus’ presentation of the oracle (7.220.4) largely repeats that before it (7.220.2): he wants ‘to lay up a store of glory only for the Spartans’ (κλέος καταθέσθαι μούνων Σπαρτιητέων ≈ κλέος μέγα ἐλείπετο). In both cases the embedded focalization seems to represent Leonidas’ hidden thoughts: he speaks about his concern for his allies’ life and the Spartan rule never to retreat but he thinks of the oracle and the kleos it could bring himself and the Spartans.33

30 31

32 33

see De Jong [1987] 2004: 102–123. Rijksbaron 2012 amply argues for the use of Greek imperfect as a marker of embedded focalization or, as he calls it, ‘substitutionary perception’. In 7.239.1 the narrator again refers to the oracle and indicates that the Spartans solicited it at the moment they heard that Xerxes had decided to march against Greece. Macan 1908: ad loc. gives a genetical analysis: ‘ταῦτα … ἐπιλεγόμενον resumes the construction interrupted by the insertion of the oracle. Perhaps the versified oracle was not in the first draft of Hdt.’s work.’ Cf. e.g. 3.34–3.35; 5.89. For the use of embedded focalization to present a character’s hidden thoughts in epic, see De Jong 1994. The device is also regularly found in the Histories, e.g. 1.10.2. A slightly different take is that of Baragwanath 2008: 69–70: ‘Thus Leonidas is presented as motivated in sending the Greek away in part by his care for their own safety and for the survival of Spartan eudaimonia (in accordance with his awareness of an oracle dating from the very beginning of the war), but particularly—this is the most insistent theme—by his concern for keeping up appearances.’

124

de jong

Having looked at the oracle’s presentation, let us now proceed to discuss the question of why the narrator mentions it here. For the ancient historian Matthew, the oracle’s position is one more reason to consider it a vaticinium post eventum: If the purpose of Herodotus’ narrative was to demonstrate the bravery of Leonidas in marching off to a battle from which he knew he was destined never to return, it is curious that the ‘oracle’ is not detailed before the king’s actual departure from Sparta with the advance force—the time at which he would have heard it and the most logical, and chronological place for the passage to sit within the text if it ever actually happened […] (my italics, IdJ)34 His reasoning is that the most logical place for the oracle to be mentioned would have been at the moment when the Spartans heard it, and hence its insertion here, so much later in Herodotus’ story and text, exposes it as a later invention. But what is the most logical place for an oracle in the Histories? The Herodotean narrator actually has two places to present oracles: at the moment when they are pronounced, that is, as an event of the main story, or at the moment when they are (about to be) fulfilled, as an analepsis. The first method is employed for instance in the case of Croesus and the oracle of the mule: when he is pondering whether to attack the Persians he consults the oracle of Delphi and hears that his reign will last until a mule becomes king of the Persians (1.55). The fulfilment of the oracle follows when Croesus is defeated by Cyrus, the son of a Persian woman and a Median man and hence a hybrid person like a mule (1.84–1.86). We find the second method in the case of the oracle about the deaf-mute son of Croesus who one day will speak, which is presented in an analepsis at the very moment when the boy speaks and the oracle is fulfilled (1.85). In the case of Leonidas’ oracle, too, Herodotus opts for the second method, and thereby turns it into the culmination of a very careful build up. While, as we saw earlier, the narratees are aware of the Spartans’ doom right from the start of the Thermopylae episode, Leonidas only gradually realises the predicament he is in. In 7.205.2, he selects three hundred men who already have sons (and the continuity of whose families is hence secured), which indicates that he realises that the mission is going to be a very dangerous one. It is dangerous because the

34

Matthew 2013: 84.

herodotus’ handling of (narratological) time

125

Greeks have far less men than the Persians in general and now especially, since a large part of the Spartans and of the Greek allies are still occupied with the Carnean and Olympian festivals. In 7.207, when the Persians have drawn near the pass, the Greeks panic and it is only because the Phocians and Locrians are incensed by the idea of retreat that Leonidas votes to stay, but he sends messengers to ask for more troops, realising that they were ‘too few to beat off the Persians’. In 7.209, the Spartans combing their hair is explained by both narrator and the exiled Spartan king Demaratus as indicating that they are ready ‘to put their lives at risk’ (κινδυνεύειν τῇ ψυχῇ) and ‘to slay or be slain’ (ὡς ἀπολεόμενοί τε καὶ ἀπολέοντες κατὰ δύναμιν). In these three passages the Spartans seriously reckon with the possibility that they will die, as is the case in any battle, but they are not certain to die, let alone planning to die.35 The turning point comes when they are informed about the Persian circuit via the Anopaea path. Now Leonidas realises that those who stay will certainly die (cf. ἀπέπεμψε Λεωνίδης, μὴ ἀπόλωνται κηδόμενος and Λεωνίδης ἀποπέμπων, ἵνα μὴ συναπόληταί σφι) and recalls the oracle that indicates that by dying he could save his city. By presenting the oracle here at this dramatic moment the narrator gives it maximal weight and effect. So much for the effective presentation of the oracle in a carefully positioned analepsis. But its narrative potential is not yet exhausted. For the oracle also functions as a prolepsis, obviously of Leonidas’ death and—perhaps less immediately obviously—of later Greek success (How-Wells speak of an ‘omen of future victory’). The oracle, by announcing that Xerxes will be stopped when either Sparta or its king is destroyed, suggests that Leonidas’ death will eventually lead to Xerxes’ defeat. The narratees can be expected to think of Salamis, where Xerxes himself will be defeated, and especially Plataea, where his general Mardonius will be defeated. In other words, the oracle hints at a connection between the defeat of the Greeks at Thermopylae and their later victories at Salamis and Plataea. This connection, here only implied, becomes explicit36 when the Persians have been defeated at Salamis and the Spartans receive another oracle: they are to demand ‘retribution for the murder of Leonidas’ (δίκας τοῦ Λεωνίδεω φόνου: 8.114.1). They send messengers to Xerxes and demand ‘retribution for the murder, because he had killed their king while he was defending Greece’ (φόνου δίκας, ὅτι σφέων τὸν βασιλέα ἀπέκτεινας ῥύομενον τὴν Ἑλλάδα: 8.114.1), but 35 36

For the very different story of Thermopylae as a suicide mission right from the start, see the chapters of Van Wees and De Bakker in this volume. What follows is largely based on Dillery 1996: 242–245 and Asheri 1998, who do not, however, include the oracle at Thermopylae in their discussion.

126

de jong

Xerxes starts laughing and, pointing at his deputy Mardonius who happens to be standing next to him, says: ‘τοιγὰρ σφι Μαρδόνιος ὅδε δίκας δώσει τοιαύτας οἵας ἐκείνοισι πρέπει.’ 8.114.2

‘Ok, that is why37 Mardonius here will give them such retribution as befits them.’ His formulation is ambiguous, thereby becoming oracular itself:38 Xerxes means that Mardonius will ‘repay’ the Spartans by inflicting another defeat like Thermopylae on them, but the narratees are also able to understand his words as announcing that Mardonius himself will ‘make amends’ to the Spartans by being killed and his troops being defeated at Plataea. This of course is what happens, and after the battle of Plataea the narrator confirms that this Greek victory is the fulfilment of the oracle: ἐνθαῦτα ἥ τε δίκη τοῦ φόνου τοῦ Λεωνίδεω κατὰ τὸ χρηστήριον τοῖσι Σπαρτήτῃσι ἐκ Μαρδονίου ἐπετελέετο […] 9.64.1

Then the retribution for the slaying of Leonidas was being paid in full to the Spartans by Mardonius according to the oracle […] Both the echo of the leitmotiv δίκη τοῦ φόνου (≈ δίκας … φόνου: 8.114.1 ≈ φόνου δίκας: 8.114.1 ≈ δίκας: 8.114.2) and the analeptic reference to the oracle (κατὰ τὸ χρηστήριον) underscore that Plataea is the Greek revenge for Thermopylae. Shortly after, the Spartan general Pausanias also connects the two battles:

37

38

The particle τοιγάρ is only here found in prose (perhaps in 3.3.3, where some MSS. have it instead of τοιγάρ τοι; I owe this parallel to Mathieu de Bakker). It marks a strong logical connection, usually in reaction to a request; cf. Denniston 1954: 565–566. I take its force to be that because Leonidas had died ‘saving Greece’, i.e. (from the Persian point of view) had been resisting him, Mardonius will punish the Spartans. Cf. 8.100.2 where Mardonius says of the Spartans at Thermopylae: οἵ τε ἡμῖν ἠντιώθησαν, ἔδοσαν δίκας, ‘those who resisted us, paid a penalty’. See Bowie 2007: 208–208, who calls his words ‘a κληδών, a chance utterance that turns out to be prophetic in a way not intended by the speaker’, and points at the verb δέκεσθαι, ‘a technical verb for “recognising” an oracular remark’.

herodotus’ handling of (narratological) time

127

‘Λεωνίδῃ δέ, […], φημὶ μεγάλως τετιμωρῆσθαι, ψυχῇσί τε τῇσι τῶνδε ἀναριθμήτοισι τετίμηται αὐτός τε καὶ οἱ ἄλλοι οἱ ἐν Θερμοπύλῇσι τελευτήσαντες.’ 9.79

‘As for Leonidas, […], I hold that he is fully avenged, and he is both himself honoured and all those that died in Thermopylae by the countless souls of these men here.’ The victory at Plataea is thus carefully framed as the Spartans’ revenge for their defeat at Thermopylae. The oracle of Thermopylae (7.220) is the first stage of this framing, suggesting as it does that Leonidas’ death and the defeat of the Spartans form a necessary prelude to victory: Xerxes can only be stopped when Leonidas dies. The reasoning here seems to be that the victory at Thermopylae will entice the Persians to continue their invasion of Greece, only to be (twice) defeated and completely annihilated during their long retreat home after those defeats. Herodotus’ narrative strategy brilliantly complements that of the Greeks on the battlefield!

7

Conclusion

I have argued that the Herodotean narrator carefully inserts analepses and prolepses at places where they are most effective. They are always relevant to the main story, the connection being of a causal, thematic or analogous nature. The relevance can be made explicitly clear by the narrator or it can remain implicit and then has to be detected by the narratees themselves. And the relevance can be understood right away or only at a later moment. The prolepsis about the Spartans’ defeat at Thermopylae at the very start of the episode casts a tragic light on all their exertions; the analepsis about the Phocian wall is connected to the main story via analogy and thereby acquires a foreshadowing function; the analepsis about how Leonidas acquired his kingship ‘unexpectedly’ triggers associations with the Herodotean theme of the reversal of fortune and hints that becoming king of the Spartans will have unexpected consequences, as it does when it turns out that it involves dying for his country; the prolepsis about Ephialtes revolves around the central Herodotean theme of retribution and gloatingly anticipates his demise at the very moment of his ‘success’; the oracle is effectively presented as an analepsis at the moment when it is (about to be) fulfilled by Leonidas at Thermopylae and at the same time functions as a prolepsis of Xerxes’ invasion of Greece coming to an end at

128

de jong

Salamis and Plataeae, thus forming part of Herodotus’ careful framing of Greek defeat as a prelude to Greek victory. Herodotus’ use of analepses and prolepses is only one aspect of his handling of time in the Thermopylae passage. There is also his effective use of narrative pace, for instance the dramatic scenic rhythm in the final stage of the battle, the Spartans fighting with swords, fists and teeth, and the strategic ellipsis (or nonnarration) of their actual slaughter by the Persians at its end (7.225.3). There is the powerful reference to the narrator’s own time, when he notes that the Spartans for their final stand withdraw to the hill ‘where now (νῦν) stands the stone lion in honour of Leonidas’ (7.225.2), pointing out to his narratees a lieu de mémoire that they may know from first-hand experience.39 And there is his consistent association of the setting of the battle, the landscape of Trachis, with the deeds of Heracles in the past (7.176.3; 7.198.2; 7.216), in order to underscore the heroism of Heracles’ descendant Leonidas (7.204). All in all, it is clear that far from being a rambling, archaic or unstoppable storyteller Herodotus knows perfectly well what he is doing. If his narrative does not always offer the kind of exact or reliable information that modern historians would like, it deserves to be taken seriously in and for itself and to be analysed by any modern theory that can help lay bare its subtleties.

Bibliography Asheri, D., ‘Plataea vendetta del Termopile: alla origine di un motivo teologico erodotea’, in M. Sordi (ed.), Responsibilità, perdono e vendetta nel mondo antico (Milan 1998) 65–86. Baragwanath, E., Motivation and Narrative in Herodotus (Oxford 2008). Bowie, A.M., Herodotus Histories Book vii (Cambridge 2007). Denniston, J.D., The Greek Particles ([1934] Oxford 1954). Dillery, J., ‘Reconfiguring the Past: Thyrea, Thermopylae and Narrative Patterns in Herodotus’, AJPh 117.2 (1996) 217–254. Focke, F., Herodot als Historiker (Stuttgart 1927). Fränkel, H., ‘Eine Stileigenheit der frühgriechischen Literatur’, in Wege und Formen frühgriechischen Denken ([1924] München 1960) 40–96. Fritz, K. von, Die griechische Geschichtsschreibung (Berlin 1967). Genette, G., Narrative Discourse: An Essay in Method ([1972] Ithaca, NY 1980). Gibson, R.K., ‘“Cf. e.g.”: a Typology of Parallels and the Role of Commentaries on Latin 39

And which backs up his implicit rejection of the alternative ‘Legend’ of the Spartans’ suicide mission; the point is made by both Van Wees and De Bakker in this volume.

herodotus’ handling of (narratological) time

129

Poetry’, in R.K. Gibson & C.S. Kraus (eds.), The Classical Commentary. Histories, Practices, Theory (Leiden 2002) 331–357. Gould, J., Herodotus (London 1989). Grethlein, J., The Greeks and Their Past: Poetry, Oratory, and History in the Fifth Century BCE (Cambridge 2010). Griffiths, A.H., ‘Stories and Story-Telling in the Histories’, in C. Dewald & J. Marincola (eds.), The Cambridge Companion to Herodotus (Cambridge 2006) 130–144. Groningen, B.A. van, In the Grip of the Past: Essay on an Aspect of Greek Thought (Leiden 1953). Groten, F.J., ‘Herodotus’ Use of Variant Versions’, Phoenix 17 (1963) 79–87. Hau, L.I., ‘Stock Situations, Topoi and the Greekness of Greek Historiography’, in D. Cairns & R. Scodel (eds.), Defining Greek Narrative (Edinburgh 2014) 241–259. How, W.W., & Wells, J., A Commentary on Herodotus, volume II (Books V–IX) ([1912] Oxford 1928). Howald, E., ‘Ionische Geschichtsschreibung’, Hermes 58.2 (1923) 113–146. Huber, L., Religiöse und politische Beweggründe des Handelns in der Geschichtsschreibung des Herodots (Tübingen 1965). Immerwahr, H.R., Form and Thought in Herodotus (Cleveland 1966). Jacoby, F., ‘Herodot’, in Pauly-Wissowa-Kroll (eds.), Realencyklopädie der klassischen Altertumswissenschaft, Suppl. II (Stuttgart 1913) 205–320. Jong, I.J.F. de, Narrators and Focalizers. The Presentation of the Story in the Iliad ([1987] London 2004). Jong, I.J.F. de, ‘Between Word and Deed: Hidden Thoughts in the Odyssey’, in I.J.F. de Jong & J.P. Sullivan (eds.), Modern Critical Theory & Classical Literature (Leiden 1994) 27–50. Jong, I.J.F. de, ‘The Anachronical Structure of Herodotus’Histories’, in S.J. Harrison (ed.), Texts, Ideas and the Classics (Oxford 2001) 93–116. Jong, I.J.F. de, Narratology and Classics. A Practical Guide (Oxford 2014). Lang, M.L., Herodotean Narrative and Discourse (Cambridge 1984). Lateiner, D., The Historical Method of Herodotus (Toronto 1989). Legrand, Ph.-E., Hérodote Histoires, Livre VII (Paris 1963). Macan, R.W., Herodotus, the Seventh, Eighth, and Ninth Books with Introduction and Commentary (London 1908). Marincola, J., Authority and Tradition in Ancient Historiography (Cambridge 1997). Matthew, C.A., ‘Towards the Hot Gates: the Events Leading up to the Battle of Thermopylae’, in C.A. Matthew & M. Trundle (eds.), Beyond the Gates of Fire: New Perspectives on the Battle of Thermopylae (Barnsley 2013) 1–26. Munson, R.V., Telling Wonders: Ethnographic and Political Discourse in the Work of Herodotus (Ann Arbor 2001).

130

de jong

Pohlenz, M., Herodot, der erste Geschichtsschreiber des Abendlandes (Leipzig 1937). Rijksbaron, A., ‘The Discourse Function of the Imperfect’, in A. Rijksbaron, H.A. Mulder & G.C. Wakker (eds.), In the Footsteps of Raphael Kühner (Amsterdam 1988) 237–254. Rijksbaron, A., The Syntax and Semantics of the Verb in Classical Greek. An Introduction (Amsterdam 2002). Rijksbaron, A., ‘The Imperfect as the Tense of Substitutionary Perception’, in P. da Cunha Correa e.a. (eds.), Hyperboreans. Essays in Greek and Latin Poetry, Philosophy, Rhetoric and Linguistics (Sao Paulo 2012) 331–377. Romilly, J. de, ‘La vengeance comme explication historique dans l’oeuvre d’Hérodote’, REG 84 (1971) 314–337. Rood, A.T., ‘Herodotus’, in I.J.F. de Jong & R. Nünlist (eds.), Time in Ancient Greek Literature (Leiden 2007) 115–130. Slings, S.R., ‘Adversative Relators between PUSH and POP’, in A. Rijskbaron (ed.), New Approaches to Greek Particles. Proceedings of the Colloquium held in Amsterdam, January 4–6, 1996, to honour C.J. Ruijgh on the Occasion of his Retirement (Amsterdam 1997) 101–129. Stein, H., Herodotos, vierter Band (Buch VII) ([1889] Berlin 1908). Thomson, J.A.K., The Art of the Logos (London 1935).

chapter 6

Herodotus and Thucydides: Distance and Immersion Rutger Allan

In studies on Greek historiography, Herodotus and Thucydides are often pitted against one another with regard to their styles, narrative techniques, scientific methodology, thematic interests and general worldview.1 The idea that the two historians fundamentally differ in their overall perspective on history has recently been given a new impulse by Grethlein (2013). While Herodotus, according to Grethlein, tends to portray history from a distanced, retrospective point of view, safely guiding the reader through the narrative with only a few major surprises, Thucydides engages the reader to assume a more empathetic and participatory stance, inviting us to relive the events from the perspective of the historical agents—as if the outcome of the events were still open. Herodotus’ history, according to Grethlein, is teleological and based on the advantages of hindsight, while Thucydides tries to make us relive the experiences of the historical actors. As Grethlein puts it: (1) Of course, Thucydides writes in hindsight and readers with some knowledge of the Peloponnesian War will remember the outcome of the sea battles. Nonetheless, the narrative compels the reader to witness the events as if they were just unfolding. History is always written retrospectively, but Thucydides enlists numerous narrative techniques to restore presentness in the past.2 In characterizing Thucydides in this way, Grethlein stands in a long tradition which at least started with Plutarch’s famous appreciation of Thucydides’ style.

1 Recent scholarship has become more sensitive to the important similarities between the works of the two major Greek historians. The recent volume Thucydides and Herodotus (Foster & Lateiner 2012) is an example of this converging tendency of scholarly approaches towards the two historians. 2 Grethlein 2013: 40.

© koninklijke brill nv, leiden, 2019 | doi:10.1163/9789004383340_007

132

allan

(2) ὁ γοῦν Θουκυδίδης ἀεὶ τῷ λόγῳ πρὸς ταύτην ἁμιλλᾶται τὴν ἐνάργειαν, οἷον θεατὴν ποιῆσαι τὸν ἀκροατὴν καὶ τὰ γινόμενα περὶ τοὺς ὁρῶντας ἐκπληκτικὰ καὶ ταρακτικὰ πάθη τοῖς ἀναγινώσκουσιν ἐνεργάσασθαι λιχνευόμενος. Plut., On the Fame of the Athenians, 347A

Thucydides, at least, always strives for this vividness, desiring to make the reader a spectator, and to evoke the events as equally amazing and upsetting to the readers as to the eye-witnesses. Thucydides’ capacity to make the reader feel as if he or she is present at the historical events is also stressed by Connor, who has been very influential in the development of Thucydidean scholarship of the last three decades. In his seminal article of 1985 on Thucydides’ narrative technique he stresses the ‘experiential’ and ‘participatory’ aspect of the historian’s writing: (3) ‘We do not usually think of Thucydides as a writer who keeps drawing his readers into the narrative of events until they feel they are themselves present, actually experiencing them. […] We do not let ourselves be caught up in the vicarious experience of the actions described but we should.’3 Connor’s formulation (‘we do not usually think of Thucydides as a writer who etc.’) makes it clear that his observation was at that time, some three decades ago, not as obvious to Thucydidean scholars as it appears to be today.4 That Herodotus and Thucydides as narrators approach their historical subjects from a fundamentally different perspective appears to be a generally accepted idea. But which properties of their narrative styles explain this difference? In current literary and art studies, the capacity of works of art to create the illusion of being present in the world that is represented by the work of art is often referred to by the term immersion. The term immersion has been introduced into literary studies by MarieLaure Ryan,5 who is working in the field of cognitive narratology, but the concept soon also attracted lively interest outside literary studies, for example in the study of the visual arts, film and computer games.6 Immersion is the mental state of being absorbed in a virtual world to the extent that one experiences it— 3 4 5 6

Connor 1985: 9. See also Bakker 2005: 159–167; Morrison 2006. Ryan 1991. E.g. Ryan 2001, Wolf et al. 2013.

herodotus and thucydides: distance and immersion

133

up to a point—as if it were the actual world, or, in Ryan’s words: ‘(…) immersion is the experience through which a fictional world acquires the presence of an autonomous, language-independent reality populated with live human beings’ (Ryan 2001: 14). It should be noted that the experience of immersion is never complete: an immersed reader or spectator will always retain a certain degree of self-awareness and awareness of the real world surrounding him or her, as well as an awareness that the story world is not the actual world.7 An important reason for the increasing interest in immersion across various academic disciplines is the fact that immersion is not bound to one particular medium but it may involve all sensory channels such as sight, hearing, and touch. What is more, the experience of immersion is not restricted to the activity of our senses only: it also crucially involves the mental simulation of motor action, emotion and cognition; it may even involve body motion, when the immersed subject is physically interacting with the objects in the virtual world. (One may think, for example, of a moving flight simulator used for the training of pilots.) The intensely immersive effect literature may have on readers can obviously not be reduced to one specific property of the text. A wide range of number of textual features are responsible for the immersive quality of a literary text. I consider the following set of linguistic and narratological features to be crucial to the reader’s experience of immersion in the story world.8 I categorize the immersive features under five headings: verisimilitude, perspective, transparency, interest and emotional involvement and the Principle of Minimal Departure. (I)

Verisimilitude: The text evokes a life-like (vivid) mental representation of persons, objects, actions and their setting.9 A life-like representation (a.) focuses on concrete, physical objects or agents and their motion; (b.) provides graphic sensory details; (c.) provides detailed spatial information;

7 See for this point also Allan, forthc. b. 8 This inventory of textual features is an expansion of the list of textual features given by Ryan 2001: 130–137. It is apparent that the concept of immersion and the specific textual features associated with immersion show similarities to ancient rhetorical notions such as ἐνάργεια (see Walker 1993; Plett 2012), ἐναγώνιος (Ooms & De Jonge 2013) and ἔκφρασις (Webb 2009). For the relationship between immersion and the ancient terminology, see Allan, De Jong & De Jonge (2014, 2017). In Allan (forthc. a), I discuss a number of passages from Homer (Il. 6.466– 470 and 23.362–372) and Thucydides (7.71.1–3), mentioned by the ancient critics as examples of ἐνάργεια, in terms of their immersive characteristics. 9 Texts that evoke a specific location at a specific time and provide rich descriptions of sensory

134

allan

(d.) progresses at a relatively slow pace (scene narration: narration time approximates narrated time); (e.) advances in chronological order (no flashbacks/flashforwards) (f.) activates (cultural) knowledge schemas (cognitive frames and scripts) that enable the addressee to ‘complete the picture’ mentally.10 (II) Perspective: The text chooses its deictic centre within the described scene. Preferably, it takes the perspective of (is focalized by) a character with whom the addressee may identify and feel empathy.11 Specific linguistic indications of this perspective shift are: (a.) proximal (‘here’ and ‘now’) deixis (e.g. the use of the historical present); (b.) imperfective aspect (to indicate an ‘internal viewpoint’); (c.) character-oriented vocabulary (III) ‘Transparency’ of the text:12 The text directs the addressee’s attention to the storyworld, that is, it defocuses from the text itself as a medium and from the narrator as a mediating voice: in other words, the artificiality of the text is concealed, the narrator remains invisible. More specifically, we will typically find (a.) no metanarrative elements (e.g. narrator comments) (b.) direct speech;13

10

11 12 13

details and (vigorous) action/movement are likely to engender embodied simulation effects (Sanford & Emmot 2012: 155–156). For the role of cognitive schemata (frames, scripts, scenarios) in the comprehension of literary texts, see Stockwell 2002: 75–89 and Sanford & Emmott 2012: 12–44. Schemata are cognitive representations that arise from peoples’ recurrent embodied experiences (cf. Gibbs 2005: 138). For example, mentioning a single word like cat is already able to activate a host of sensorimotor as well as emotional experience from memory, including a cat’s shape, its colour(s), the sound it makes when it miaows, purrs or hisses, the touch of its fur when one strokes it, how it feels to be scratched, feelings of love (provided one is a cat-lover) or hate, the usual point of view and distance from which one perceives and interacts with cats, etcetera. The activation of these experientally rich, multimodal, embodied cognitive schemas enables the reader to ‘connect the dots’ and to flesh out the mental simulation of the narrated situation, giving the reader the feeling that a description is complete even though it is actually relatively ‘gappy’ and only provides a number of strategically selected sensory details. See also Ryan’s emotional immersion (2001: 148–157). For the requirement of transparency of the medium for immersion, Ryan 2001: 4, 56–58. Psychological evidence for the effect of vividness/immediacy/involvement associated with direct speech is presented by Sanford & Emmott 2012: 181–190.

herodotus and thucydides: distance and immersion

135

(c.) no elements drawing attention to the conventionality of the textual (literary) genre. (IV) Interest and emotional involvement: The theme of the text is of strong interest to the addressee. The text contains elements steering the addressee’s emotional response (e.g. subjective-evaluative and emotive vocabulary). The text (segment) is crucial to the main story line (e.g. a Peak), creates suspense, and is serious rather than comic. (V) Principle of Minimal Departure:14 The storyworld should not (or only minimally) depart from the ‘real world’ as we know it: the story world should be internally consistent and subject to the same rules as ‘real life’. Departures are tolerated if they can be explained, for example, by generic conventions. The experience of immersion can be represented graphically as in Fig. 6.1. The reader feels as if he or she is mentally transported into the story world (arrow) which is within the focus of the reader’s attention (thick circle), while the real world (including the reader’s self-awareness) and world of narration are psychologically backgrounded (dashed circles).15 The experience of being immersed in the story is never complete and it should, rather, be seen as a gradient phenomenon. A reader can be immersed to a higher or lower degree, depending on such factors as the reader’s general or temporary psychological predisposition towards being absorbed in a storyworld and the presence of immersion-inducing features in the text (passage). The higher the number of immersive features (and the lower the number of immersion-disrupting features) in a particular text passage, the stronger the feeling of immersion is. The degree of immersion is, however, also dependent on the diversity of immersive features. It is, for example, not enough for a text passage to contain an abundance of sensory details to be immersive—a reader should also care about the people and objects that are introduced into the storyworld.16 In order to be 14 15 16

Ryan 1991: 51. The figure is a somewhat elaborated version of a figure presented by Dannenberg 2008: 24. A description too much fraught with irrelevant details is likely to be detrimental to immersion since it does not accord with the way in which we normally look at an object or observe a situation. When we observe an object or spatial setting in a natural, everyday situation we will tend to focus our attention to the perceptually more salient elements of the observed object, to elements that are relevant to our usual, everyday way of interaction with the observed object, or to elements that are otherwise (e.g. emotionally) relevant to

136

allan

figure 6.1 The reader’s experience of immersion into the storyworld

immersive, therefore, a text should also contain elements that trigger an emotional involvement of the reader with the events in the storyworld. Key factors in this emotional dimension of immersion are such narrative techniques as shifts in perspective towards a viewpoint within the story world (e.g. character focalization) encouraging empathy with a story character, and the creation of a feeling of suspense (‘What will happen next?’). The intensity of the immersive experience, in sum, is dependent on a tight and well-balanced interplay of immersive features.

the observer. When textual descriptions provide more details than necessary for a coherent mental simulation (note the crucial role of cognitive schemas here) the reader will experience an increased processing effort and the description will feel as an artificial literary construct which draws attention to the text as an artistic medium rather than facilitating the reader’s mental representation of the described object. In Allan (forthc. c.) I argue that Homer’s description of Calypso’s cave is immersive due to its wealth of graphic perceptual details in combination with the fact that it is presented as a simulated observation: the narrator guides the narratee’s gaze in a carefully organized way through the abundant details of the scene, reflecting how we would normally observe a (static) scene in everyday life. This natural, embodied, way of presentation effects that the description, despite its abundance in details, is easy to process.

herodotus and thucydides: distance and immersion table 6.1

137

Some linguistic and narratological features

Thermopylae Sphacteria (Hdt. 7.200–7.233) (Thuc. 4.30.4–4.38) Words Actual Presenta Historical Present Present Perfect Imperfect Aorist Pluperfect Particles ἀλλά γάρ δή δῆθεν μέν μέντοι οὖν/ὦν Direct discourse Indirect discourse Flashforward Flashback Narratorial comment Narratorial summary

3324 27 5 2 120 40 7

1358 0 3 0 43 30 2

9 34 23 1 50 2 3 2 60 4 14 25 10

3 8 1 0 20 0 2 1 17 0 1 1 1

a The numbers of the tenses and aspects in the table only include their occurrences in main clauses.

I return to our two historians. It seems clear that while Thucydides often consciously seeks to immerse the reader in the past events, Herodotus only rarely seems to do so. How differently the two historians approach their subject as a narrator can be observed by comparing two crucial passages from both works: the battle of Thermopylae (Hdt. 7.207–7.233) and the battle of Sphacteria (Thuc. 4.31–4.38). One of the reasons for comparing these two episodes is that they are fairly typical examples of how the two historians portray battles in their work. Another reason is that there are many similarities between the two accounts: in both cases an army is surrounded and locked up within a narrow space and eventually defeated. It has been argued by Edith Foster (2012)

138

allan

that Thucydides deliberately modeled his account after that of Herodotus’ account of Thermopylae. Thucydides himself also notes the similarities of the two events in 4.36.3. This is in fact the only moment in which Thucydides pops up in the narrative. Thucydides observes that, whereas the older generation of Spartans fought to the death, the younger generation sees no shame in surrendering to the arch-enemy. As a point of departure for my comparison of the two battle scenes, it may be insightful to examine some basic statistics concerning the two text passages. There are a number of features in which the two passages differ in a marked way: in the account of Thermopylae we find actual presents (i.e. the present that refers to the hic et nunc of the narrator) and present perfects, the particle δή occurs more frequently, we find more flashbacks and flashforwards, and there are more narratorial comments and more narratorial summaries. These features all point in the same direction: in the Thermopylae passage, the narrator is more visibly present than in the Pylos episode: he refers to states of affairs in his own time by using the present and the present perfect; he interacts with the narratee by using the particle δή; he interrupts the narrative in order to refer to events that have occurred anterior to Thermopylae and will occur after Thermopylae; he gives short summaries to structure the narrative and he comments on the events. Examples of flashbacks, flashforwards, summaries and narratorial comments are the following: (4) Flashbacks and flashforwards17 a. Αὐτοὶ γάρ σφεας οἱ Ἕλληνες ἐπεκαλέσαντο Hdt. 7.203.1

For the Hellenes had themselves summoned them. b. ἀκηκόεε δὲ ἔτι ἐὼν ἐν Θεσσαλίῃ ὡς ἁλισμένη εἴη ταύτῃ στρατιὴ ὀλίγη Hdt. 7.208.1

While he [Xerxes] was still in Thessaly, he had heard that a small army had gathered there.

17

For a discussion of these temporal anachronies in the Thermopylae episode, see the chapter of De Jong in this volume.

herodotus and thucydides: distance and immersion

139

c. […] καὶ διέφθειρε τοὺς ταύτῃ ὑπομείναντας Ἑλλήνων. Hdt. 7.213.1

[Ephialtes told Xerxes of the path through the mountains to Thermopylae] and thereby he caused the destruction of the Hellenes waiting there. Θεσπιέες δὲ ἑκόντες μάλιστα, οἳ οὐκ ἔφασαν ἀπολιπόντες Λεωνίδην καὶ τοὺς μετ’ αὐτοῦ ἀπαλλάξεσθαι, ἀλλὰ καταμείναντες συναπέθανον· Hdt. 7.222

The Thespians remained quite willingly. They refused to go away leaving Leonidas and his men, but they remained and died with him. (5) Summaries a. Οὗτοι μὲν δὴ οὕτω διενένωντο ποιήσειν. Hdt. 7.207

They intended to act in that way. b. Τότε μὲν οὕτω ἠγωνίσαντο. Hdt. 7.212.1

They had fought then in that way. c. χρόνον μὲν οὖν τινὰ ὀλίγον οὕτω πρὸς ἀλλήλους ἠκροβολίσαντο· Thuc. 4.34.1

Now they had skirmished in that way for a little while. (6) Narratorial comments a. Λακεδαιμόνιοι δὲ ἐμάχοντο ἀξίως λόγου, […] Hdt. 7.211.3

The Lacedaemonians fought memorably, […] b. ὁ δὲ κολωνός ἐστι ἐν τῇ ἐσόδῳ, ὅκου νῦν ὁ λίθινος λέων ἕστηκε ἐπὶ Λεωνίδῃ. Hdt. 7.225.2

140

allan

The hill is in the entrance where now the stone lion stands in honour of Leonidas. c. καὶ οἱ Λακεδαιμόνιοι βαλλόμενοί τε ἀμφοτέρωθεν ἤδη καὶ γιγνόμενοι ἐν τῷ αὐτῷ ξυμπτώματι, ὡς μικρὸν μεγάλῳ εἰκάσαι, τῷ ἐν Θερμοπύλαις, ἐκεῖνοί τε γὰρ τῇ ἀτραπῷ περιελθόντων τῶν Περσῶν διεφθάρησαν, οὗτοί τε ἀμφίβολοι ἤδη ὄντες οὐκέτι ἀντεῖχον, ἀλλὰ πολλοῖς τε ὀλίγοι μαχόμενοι καὶ ἀσθενείᾳ σωμάτων διὰ τὴν σιτοδείαν ὑπεχώρουν, καὶ οἱ Ἀθηναῖοι ἐκράτουν ἤδη τῶν ἐφόδων. Thuc. 4.36.3

The Lacedaemonians thus placed between two fires, and in the same dilemma, to compare small things with great, as at Thermopylae, where the defenders were cut off through the Persians getting round by the path, being now attacked in front and behind, began to give way, and overcome by the odds against them and exhausted from want of food, retreated. Transl. Crawley

The observation that Herodotus is prominently present in his narration, while Thucydides remains reticent, is in itself not a new one.18 However, what is relevant for my argument is that the narrator who makes his presence felt disturbs the reader’s immersive experience. The narrator’s intrusions detract from the feeling of being plunged into the events as they make the narratee fully aware that the storyworld is no more than an artificial construct. As a result, a narratee will be inclined to assume a detached rather than an engaged attitude towards the story world. In addition to the narrator’s covertness, there are other factors that are beneficial to immersion. I will discuss two passages from the Sphacteria-episode that illustrate Thucydides’ superb ability to give his battle descriptions an immersive quality.19 In the first scene, the Athenians rush to attack the Spartans.

18 19

For the contrast between Herodotus as an overt and Thucydides as a covert narrator, see e.g. De Jong, Nünlist & Bowie 2004. Elsewhere, I discuss Thucydides 7.71.1–3 (Allan, forthc. a.), 2.84.1–3, 5.116.3–4, 7.75.6–7, 7.87.5–6 (Allan, forthc. b.) in terms of their immersive qualities.

herodotus and thucydides: distance and immersion

141

(7) γενομένης δὲ τῆς βοῆς ἅμα τῇ ἐπιδρομῇ ἔκπληξίς τε ἐνέπεσεν ἀνθρώποις ἀήθεσι τοιαύτης μάχης καὶ ὁ κονιορτὸς τῆς ὕλης νεωστὶ κεκαυμένης ἐχώρει πολὺς ἄνω, ἄπορόν τε ἦν ἰδεῖν τὸ πρὸ αὑτοῦ ὑπὸ τῶν τοξευμάτων καὶ λίθων ἀπὸ πολλῶν ἀνθρώπων μετὰ τοῦ κονιορτοῦ ἅμα φερομένων. τό τε ἔργον ἐνταῦθα χαλεπὸν τοῖς Λακεδαιμονίοις καθίστατο· οὔτε γὰρ οἱ πῖλοι ἔστεγον τὰ τοξεύματα, δοράτιά τε ἐναπεκέκλαστο βαλλομένων, εἶχόν τε οὐδὲν σφίσιν αὐτοῖς χρήσασθαι ἀποκεκλῃμένοι μὲν τῇ ὄψει τοῦ προορᾶν, ὑπὸ δὲ τῆς μείζονος βοῆς τῶν πολεμίων τὰ ἐν αὐτοῖς παραγγελλόμενα οὐκ ἐσακούοντες, κινδύνου τε πανταχόθεν περιεστῶτος καὶ οὐκ ἔχοντες ἐλπίδα καθ’ ὅτι χρὴ ἀμυνομένους σωθῆναι. 4.34.2–4.34.3

The shouting accompanying their onset confounded the Spartans, unaccustomed to this mode of fighting; dust was rising from the newly burnt wood, and it was impossible to see in front of one with the arrows and stones flying through clouds of dust from the hands of numerous assailants. The Spartans had now to sustain a difficult conflict; their caps would not keep out the arrows, and points of javelins had broken off as they were struck. They themselves were unable to retaliate, being prevented from using their eyes to see what was before them, and unable to hear the words of command for the hubbub raised by the enemy; danger encompassed them on every side, and there was no hope of any means of defence or safety. transl. Crawley, adapted

There are a number of immersive features in this passage. First, the passage provides an abundance of graphic details: the rising cloud of ashes, the arrows flying through the air, the points of javelins broken off when they hit the harnesses, the shouting of the enemy. The passage also contains a considerable number of prepositions and adverbial phrases (marked blue in the text) that specify the spatial dimensions of the events. What is also typical of immersive text is the slow progression of time (scene narration). In this passage, there is no perceivable progression of time. Apart from the aorist ἐνέπεσεν in the first clause, the tense used in the main clauses is the imperfect: ἐχώρει, ἦν, καθίστατο, ἔστεγον, εἶχον (bold in the text). There is one pluperfect ἐναπεκέκλαστο. The use of this pluperfect, however, is not fundamentally different from the surrounding imperfects since it expresses a resulting state ‘had broken off’, instead of an event ‘broke off’. These imperfects and pluperfect express situations that are occurring simultaneously to one another and do not reach an endpoint. They serve to portray the scene from an internal viewpoint, suggesting that one is looking at the scene while it is unfold-

142

allan

ing before one’s eyes.20 The absence of progression is also reflected by the use of τε as a connective particle, which marks the close thematic continuity between the sentences. The use of the imperfect as a means to present the scene from an internal viewpoint brings us to the issue of focalization. Through whose eyes are we observing the scene? The word that prepares the shift in perspective is ἔκπληξις in the first sentence. We are told that the Spartans are in panic because they are unaccustomed to the disorderly way of fighting. The mentioning of the Spartans’ fear prompts us to interpret the following description as focalized by the Spartans. The transition to the Spartans’ viewpoint is sudden but seamless. The coordinated clause καὶ ὁ κονιορτὸς … ἐχώρει πολὺς ἄνω at once plunges us, through the use of the imperfect tense, into an internal viewpoint. The imperfects continue through the following main clauses. From then on, we perceive the scene through the Spartans’ desperate gaze, which moves through the scene, focusing on the rising dust cloud, on the arrows and stones, on their caps, on the broken javelin points and on the noise made by the Athenians. Another conspicuous feature of the passage is the high number of negative expressions: ἄπορον τε ἦν ἰδεῖν, οὔτε γὰρ οἱ πῖλοι, εἶχόν τε οὐδέν, οὐκ ἐσακούοντες, οὐκ ἔχοντες ἐλπίδα. The negations stress the desperate situation the Spartans find themselves in. Negations in narrative are typically used to deny an (implicit) expectation that might exist on the part of the narratee or a character. The Spartans expected to be able to see in front of them and they thought their caps would protect them against the arrows—but their expectation is constantly thwarted. That we are inside the psyche of the Spartans is also suggested by the use of subjective-evaluative terms such as ἄπορον, χαλεπόν, πολεμίων and κινδύνου (marked green in text 7). Only from the Spartans’ perspective can the Athenians be seen as enemies and the situation as dangerous. It is worth noting here that Thucydides virtually always uses the terms πολέμιος and κίνδυνος in character language, and not in narrator text.21 The same goes for modal expressions such as the adjective ἄπορος ‘impossible’ and the verb ἔχω + infinitive ‘be able to’.22 The fact that there is no verb of thinking, saying or perception that signal the shift to the Spartans’ perspective makes it a case of implicit embedded focalisa20 21 22

For this mode of narration in Thucydides, see Bakker 1997 (who calls it the ‘mimetic mode’), Allan 2009 and Allan 2013 (‘descriptive mode’). See Allan 2013: 381. In (cognitive) stylistic literature, modal expressions are often mentioned as indicators of character-bound viewpoints: see, e.g., Simpson 1993; Fowler 1996: 167; and Short 1996: 287.

herodotus and thucydides: distance and immersion

143

tion. More precisely, we are dealing with a case of free indirect perception.23 The narrator gives us unrestricted access to the Spartans’ perceptions: we perceive the scene through the eyes and ears of the Spartans. This shift to the Spartans’ perspective is ‘free’ as it is not indicated explicitly by a verb of perception and it is ‘indirect’ since the tense is shifted towards the past in harmony with the surrounding narrative. In the literature on free indirect perception, the predominance of the imperfect (or similar tenses such as the past progressive) as a signal of an internal viewpoint on the events is often mentioned.24 Free indirect perception can be described in the terminology of mental space theory as a blend of two viewpoints.25 The viewpoints of the narrator and that of the character are integrated into one blended viewpoint: on the one hand, the past tense and the use of 3rd person reference (instead of the 1st person) are signs that the narrator still functions as a deictic center. The language used (syntax, vocabulary and style) can be ascribed to the voice of the narrator Thucydides: it is not in any way meant as an attempt to accurately verbalize the actual thoughts of the Spartans, for example, by the use of expressive markers, colloquialisms, syntactic irregularities (suggestive of oral language) or the Doric dialect.26 On the other hand, the sustained use of the imperfect and the subjective-evaluative terms are indications of the character’s perspective. 23

24

25 26

Free indirect perception is also called represented perception, substitutionary perception, narrated perception, erlebte Wahrnehmung, style indirect libre de perception; see also Chatman 1978: 204; Brinton 1980; Fludernik 1993: 299–303; Rijksbaron 2012; Allan 2013: 379–380. In modern literary language, spatial and temporal adverbs tend to align with the character’s viewpoint rather than with the narrator’s (see e.g. Nikiforidou 2010). Thus we find here and now in combination with past tenses, e.g. He was driving through Las Vegas. He was to marry here today. In our passage, the deictic adverb ἐνταῦθα is ambiguous: it might be distal ‘then, in that situation’ (narrator viewpoint) but also proximal ‘here, in this situation’ (character viewpoint). Although ἐνταῦθα more frequently expresses distal deixis, there are also cases in which the adverb refers to a location close to the deictic center, e.g. Ar. Av. 1184–1185: κἄστ’ οὐ μακρὰν ἄπωθεν, ἀλλ’ ἐνταῦθά που/ ἤδη ’στίν (‘And he is not far off, but already somewhere nearby’). Crawley’s translation of ἐνταῦθα in passage 7 with ‘now’ suggests that he reads it as a proximal deictic. For the wide-spread phenomenon of blended viewpoints in discourse, see e.g. Dancygier 2005; 2009: 183–196; and Dancygier & Sweetser 2012. Another indication of the narrator’s viewpoint is the use of the full noun phrase Λακεδαιμονίοις instead of a simple pronoun αὐτοῖς as we find elsewhere in the passage. The smooth and rough breaths of the 3rd person pronouns in this passage should not be taken at face-value. Manuscripts are notoriously unreliable in this respect. The rough breathing of πρὸ αὑτοῦ, for example, has been introduced by Stephanus; the MSS. have αὐτοῦ. If the reflexive form πρὸ αὑτοῦ and σφίσιν αὐτοῖς are sound they would be an additional indication of the subjective viewpoint of the Spartans. According to Kühner-Gerth (1898: 562), reflexive pronouns can be used in infinitive constructions ‘wenn dieselben aus der Seele des Subjektes im Hauptsatze, also als Gedanken desselben, ausgesagt werden’. In the case

144

allan

In the prelude to this passage in the Sphacteria-episode, there is another striking example of perspectivized narration. Thucydides presents us with a blend of multiple viewpoints which is even more complex than the one discussed above. (8) Δημοσθένους δὲ τάξαντος διέστησαν κατὰ διακοσίους τε καὶ πλείους, ἔστι δ’ ᾗ ἐλάσσους, τῶν χωρίων τὰ μετεωρότατα λαβόντες, ὅπως ὅτι πλείστη ἀπορία ᾖ τοῖς πολεμίοις πανταχόθεν κεκυκλομένοις καὶ μὴ ἔχωσι πρὸς ὅτι ἀντιτάξωνται, ἀλλ’ ἀμφίβολοι γίγνωνται τῷ πλήθει, εἰ μὲν τοῖς πρόσθεν ἐπίοιεν, ὑπὸ τῶν κατόπιν βαλλόμενοι, εἰ δὲ τοῖς πλαγίοις, ὑπὸ τῶν ἑκατέρωθεν παρατεταγμένων. κατὰ νώτου τε αἰεὶ ἔμελλον αὐτοῖς, ᾗ χωρήσειαν, οἱ πολέμιοι ἔσεσθαι ψιλοὶ καὶ οἱ ἀπορώτατοι, τοξεύμασι καὶ ἀκοντίοις καὶ λίθοις καὶ σφενδόναις ἐκ πολλοῦ ἔχοντες ἀλκήν, οἷς μηδὲ ἐπελθεῖν οἷόν τε ἦν· φεύγοντές τε γὰρ ἐκράτουν καὶ ἀναχωροῦσιν ἐπέκειντο. Τοιαύτῃ μὲν γνώμῃ ὁ Δημοσθένης τό τε πρῶτον τὴν ἀπόβασιν ἐπενόει καὶ ἐν τῷ ἔργῳ ἔταξεν· Thuc. 4.32.3–4.32.4

The tactics of Demosthenes had divided them into companies of two hundred, more or less, and made them occupy the highest points in order to paralyze the enemies by surrounding him on every side and thus leaving him without any tangible adversary, exposed to the cross-fire of their host; plied by those in his rear if he attacked in front, and by those on one flank he moved against those on the other. Wherever they went the enemies would be behind them, and these light-armed assailants, the ones hardest to deal with; arrows, darts, stones, and slings making them formidable at a distance, and there being no means of getting at them at close quarters, as they would have an advantage by fleeing, and the moment their pursuer turned they were upon him. Such was the idea that inspired Demosthenes in his conception of the descent, and presided over its execution. Transl. Crawley, adapted

of narratorial perspective, according to Kühner-Gerth (1898: 563), the regular 3rd person pronoun (i.e. the oblique cases of αὐτός) is used. The fact that in ἄπορόν τε ἦν ἰδεῖν τὸ πρὸ αὑτοῦ, the antecedent of αὑτοῦ is not the grammatical subject of the matrix clause but is to be understood as an implicit dative to ἄπορον ἦν is not unusual (see Kühner-Gerth 1898: 562–563, Anm. 2). For the reflexive pronoun expressing the subjectivity of the character in free indirect style in modern languages, see Brinton 1995.

herodotus and thucydides: distance and immersion

145

figure 6.2 Temporal structure of Th. 4.32.3–4

Demosthenes commands the troops to occupy the highest point on Sphacteria (aorist διέστησαν). A final ὅπως-clause (in italics) states Demosthenes’ aim with this surrounding manoeuvre—to lure the Spartans into difficulty by surrounding them, to prevent the Spartans from being able to confront them in battle and to expose the Spartans to their missiles. Final clauses are indicators of implicit embedded focalization.27 The auxiliary μέλλω has two semantic aspects. It has a temporal dimension in that it signals that the situation referred to by the infinitive is posterior to the time referred to by μέλλω. But it also has an epistemic modal dimension: it indicates that, at the time referred to by μέλλω, it is likely that the situation referred to by the infinitive will be realized.28 The imperfect aspect ἔμελλον signals that the presence of likelihood should be seen as a temporal framework (background) within which the main event in the previous sentence (διέστησαν) occurred. In this particular case, the imperfect aspect ἔμελλον suggests that it is Demosthenes who makes the epistemic assessment that the Spartans would always have the Athenians in behind them, wherever they would go (κατὰ νώτου τε αἰεὶ ἔμελλον αὐτοῖς, ᾗ χωρήσειαν, οἱ πολέ-

27

28

See De Jong 1987: 111, 118. Note that the use of the subjunctive rather than the (oblique) optative signals that the deictic center (ground) has shifted from the time of the narrator to the time of Demosthenes, which now functions as the temporal vantage point (surrogate ground) from which the content of the final clause is viewed. The ground is a term used in Cognitive Grammar to refer to the ensemble of ‘the speaker, the hearer, the speech in which they participate, and their immediate circumstances (e.g. the time and place of speaking)’ (Langacker 2008: 78). See Basset 1979; Ruijgh 1985; Wakker 2006. In the terminology of Mental Space Theory, μέλλω is a space-builder: it projects a time frame which is posterior to the Viewpoint Space (the time of Demosthenes’ tactical manoeuvres) and whose events or properties are presented as a prediction from that Viewpoint Space (Fauconnier 1997: 77–78). For μέλλω as an epistemic modal auxiliary, see also Allan 2017.

146

allan

μιοι ἔσεσθαι) and that this assessment is the motivation for positioning his men on the highest point. In other words, ἔμελλον signals that the content of the complement clause is focalized by Demosthenes. As there is no embedding verb of saying or thinking, we are dealing with a case of free indirect thought, a blend between the narrator’s voice and Demosthenes’ focalization.29 That Demosthenes is indeed the focalizer is also confirmed by the concluding sentence which explicitly states that we have been inside Demosthenes’ mind: Τοιαύτῃ μὲν γνώμῃ ὁ Δημοσθένης τό τε πρῶτον τὴν ἀπόβασιν ἐπενόει καὶ ἐν τῷ ἔργῳ ἔταξεν. Free indirect discourse (i.e. speech and thought) has been the object of a huge amount of scholarship.30 It is sometimes characterized as a ‘dual-voice’ mode of narration, or, more recently as a blend between the narrator’s and character’s viewpoints.31 The past tense and the use of 3rd person pronouns (instead of 1st person) are signs that the narrator still partly functions as a deictic center. On the other hand, there are other indications in the text that point to the character’s viewpoint. In narratological theory, it is commonly held that free indirect discourse is a ‘fuzzy’ concept in the sense that there is no discrete set of formal linguistics features with which one can distinguish free indirect discourse from other modes of speech and thought representation. Text segments very often show ambiguity as to their mode of representation: in many cases, a reader’s decision to interpret a stretch of text as free indirect discourse may depend on subjective intuitions based on contextual cues. Be that as it may, there are several linguistic features that are widely recognized as strong indicators of free indirect discourse, such as proximal adverbial (‘here’ and ‘now’) deixis (taking the character as deictic center), expressive markers (goodness! wow!), exclamatory clauses, questions, oral style, (epistemic) modals (perhaps, probably, must, can, seem, ought to), subjective-evaluative expressions (good, beautiful) and character-oriented designations (daddy, friend).

29 30

31

That the imperfect of ἔμελλον + infinitive can be analyzed as a marker of free indirect discourse (in this and other passages) has already been noted by Basset (1979: 154 and passim). The most important (cognitively-oriented) narratological work on free indirect discourse is Fludernik 1993, who also gives a useful overview of the earlier literature. More recent cognitively-oriented approaches to speech and thought representation (including free indirect discourse) are Palmer 2004 (who criticizes the traditional speech category approach towards a broader approach that captures how readers make inferences about fictional minds using textual clues) and Vandelanotte 2009 (who also distinguishes a new category called Distancing Speech and Thought Representation (DIST)). A formal semantic approach to free indirect discourse is Eckardt 2014. For Latin, see Adema 2017: 22–33. For an account of free indirect discourse as a blend of viewpoints, see Dancygier 2005; 2012; Nikiforidou 2010 and some of the contributions in Dancygier & Sweetser 2012.

herodotus and thucydides: distance and immersion

147

A common conception of free indirect discourse is to view it as a region on a gradual scale between pure narrative and direct discourse.32 Fludernik, for example, proposes a scale in which the following fuzzily bounded gradations can be distinguished:33 (9) Scale of speech and thought representation (A) Pure narrative (B) Narrated perception (C) Psycho-narration (D) Free indirect discourse (E) Indirect discourse (F) Direct discourse Another influential account of speech and thought representation is that of Short, who proposes a similar, but double, scale: one of speech and one of thought representation: 34

figure 6.3 Scale of Speech and Thought Representation

The reason why Short makes a distinction between speech and thought representation is that the effect associated with the two parallel scales is different. According to Short, this difference in effect can be ascribed to a difference in what is the norm in narrative fiction. For speech representation, direct speech is the norm as it is directly accessible to the listener and claims to be a faithful representation of the speaker’s words. For thought representation, however, indirect thought is the norm since other people’s thoughts are not accessible and therefore not reported verbatim (and generally not verbally formulated 32 33 34

E.g. Chatman 1978; Fludernik 1993; Short 1996; Leech & Short 2007. Fludernik 1993: Ch. 5. Short 1996: Ch. 10; Leech & Short 2007: Ch. 10.

148

allan

by the person thinking in the first place). This difference in norm gives rise to a difference in effect between free indirect speech and free indirect thought. Free indirect speech is a deviation of the norm in the direction of a representation in which the narrator is more in control. The narrator’s voice is interposed between the character’s words and the narratee. Free indirect speech, therefore, typically has a distancing (occasionally even ironic) effect. Free indirect thought, however, has the opposite effect: it is a deviation from the norm towards less narratorial control and towards a more immediate (‘vivid’) representation of a character’s thoughts. It thus enhances the illusion that we are actually inside the character’s mind following the character’s flow of thoughts and feelings as it occurs and we are encouraged to sympathize with his or her viewpoint.35 Since free indirect thought involves a transfer to a point of view internal to the storyworld, often accompanied by an empathetic identification with a character, it can be employed as an effective immersive narrative technique. To return to our passage in Thucydides, it is clear from a number of textual indicators (such as ἔμελλον and τοιαύτῃ μὲν γνώμῃ […]) that the text should be interpreted as representing Demosthenes’ thoughts. The fact that there is no explicit cognitive verb embedding the thought content makes it clear that we are dealing here with an—in Greek literature—very rare case of free indirect thought.36 However, the passage turns out to be even more intriguing if we consider its wording. The vocabulary used suggests that the content of the clause is also focalized by the Spartans. The terms οἱ πολέμιοι ‘enemies’ and οἱ ἀπορώτατοι ‘hard to deal with’ referring to the Athenians betray the subjective point of view of the Spartans. Only from the point of view of the Spartans can the 35

36

See Fludernik 1993: 5–6; Short 1996: 314–316; Leech & Short 2007: 274–279. An experimental study by Sanders & Redeker (1993) suggests that readers experience free indirect thought as more lively and suspenseful. The occurrence of free indirect discourse in Greek literature is a debated issue. For example, Beck 2012 detects various types of free indirect discourse in Homer, but this may be due to her relatively broad definition of the concept. De Jong 1987 does not recognize a category of free indirect speech in Homer; likewise Bakker 2009: 121: ‘No systematic grammar of free indirect discourse is detectable in Homer’. In earlier publications, Bakker did seem to recognize the occurrence of free indirect discourse in Thucydides: in Bakker 1997b: 40– 41 and Bakker 2005: 163, Thuc. 7.70.7 is cited as a prime example of free indirect discourse. However, in Bakker 2009: 121–122, he labels the passage rightly as a form of indirect speech: the two infinitival clauses (connected by τε … καί) in τοῖς μὲν Ἀθηναίοις βιάζεσθαί τε τὸν ἔκπλουν ἐπιβοῶντες καὶ περὶ τῆς ἐς τὴν πατρίδα σωτηρίας νῦν, εἴ ποτε καὶ αὖθις, προθύμως ἀντιλαβέσθαι are dependent on the speech verb ἐπιβοῶντες (‘[T]he Athenians they urged to force the passage out, and now if ever to show their mettle and lay hold of a safe return to their country.’ Trans. Crawley).

herodotus and thucydides: distance and immersion

149

Athenians be seen as enemies or as hard to deal with.37 Demosthenes, in other words, is imagining what the Spartans would be thinking when surrounded by the Athenians. This perspective shift is all the more striking considering the fact that only in the preceding final ὅπως-clause the term πολεμίοις had been used to refer to the Spartans from Demosthenes’ perspective. This means that we are in fact dealing with a blend of three viewpoints (triple focalization): the narrator’s, Demosthenes’ and the Spartans’. The next sentence is marked by γάρ as an explanation of the preceding sentence. The imperfect tenses ἐκράτουν and ἐπέκειντο signal that the situations (still only existent in Demosthenes’ imagination) should be interpreted as simultaneous with (and as an explanation of) the preceding relative clause (οἷς μηδὲ ἐπελθεῖν οἷόν τε ἦν). The Athenians could not be approached because they would either have an advantage over the Spartans by fleeing or the Athenians would be able to attack the Spartans as soon as they would retreat. There is no reason to assume that the triple blend of viewpoint is not continued here: the added explanation can be interpreted as expressing the view of Demosthenes, of the Spartans and of the narrator.38 The multiple levels of conception (focalization), with a number of embedded conceptualizers (focalizers) can be graphically represented as in Fig. 6.4. The narrator (N) has a conception of (left dashed arrow) a field of view (F1)39 in which Demosthenes (D) has a conception of (second dashed arrow) a field of view (F2) in which the Spartans (S1) have a conception of (third dashed arrow) a field of view (F3) in which the Spartans themselves (S2) are in battle with the Athenians. The two passages discussed above show Thucydides capacity to immerse the reader in the narrative by making himself invisible, by slipping seamlessly from one perspective into another and by creating complex blended perspectives. The difference with Herodotus’ narrative style can be appreciated fully when

37

38 39

The modal expression οἷόν τε ἦν also suggests that we are dealing here with subjective character language. It occurs 20 times in Thucydides, but never in neutral narrative focalized by the narrator: it occurs 4 times in direct speech (1.80.3, 1.82.6, 2.44.3, 7.14.2), 13 times in indirect speech (1.91.7, 4.22.3, 5.10.3, 6.78.2, 6.101.2, 7.1.1, 7.20.2, 7.42.3, 7.47.3, 7.60.2, 7.68.6, 8.60.3, 8.92.3), 2 times in character focalization (4.53.3, 7.60.4) and once in an authorial comment (2.48.3). In Thucydides, γάρ often introduces a motivation focalized by a character ‘for (he thought that) …’ (Hornblower 1994: 134. Cf. also De Jong 1987: 112). The term field of view is taken from Cognitive Grammar. It refers to the maximal scope of awareness (“mental reach”) which a conceptualizer may apprehend at a given moment (Langacker 2008). Compare also the notion of field of vision which features in Jahn’s approach to focalization (Jahn 1996).

150

allan

figure 6.4 Multiple levels of focalization in Th. 4.32.3–4

we compare the Thucydidean passages with the passage constituting the climax of the Battle of Thermopylae. (10) Δόρατα μέν νυν τοῖσι πλέοσι αὐτῶν τηνικαῦτα ἤδη ἐτύγχανε κατεηγότα, οἱ δὲ τοῖσι ξίφεσι διεργάζοντο τοὺς Πέρσας. Καὶ Λεωνίδης τε ἐν τούτῳ τῷ πόνῳ πίπτει ἀνὴρ γενόμενος ἄριστος, καὶ ἕτεροι μετ’ αὐτοῦ ὀνομαστοὶ Σπαρτιητέων, τῶν ἐγὼ ὡς ἀνδρῶν ἀξίων γενομένων ἐπυθόμην τὰ οὐνόματα, ἐπυθόμην δὲ καὶ ἁπάντων τῶν τριηκοσίων. Καὶ δὴ ⟨καὶ⟩ Περσέων πίπτουσι ἐνθαῦτα ἄλλοι τε πολλοὶ καὶ ὀνομαστοί, ἐν δὲ δὴ καὶ Δαρείου δύο παῖδες, Ἀβροκόμης τε καὶ Ὑπεράνθης, ἐκ τῆς Ἀρτάνεω θυγατρὸς Φραταγούνης γεγονότες Δαρείῳ· ὁ δὲ Ἀρτάνης Δαρείου μὲν τοῦ βασιλέος ἦν ἀδελφεός, Ὑστάσπεος δὲ τοῦ Ἀρσάμεος παῖς· ὃς καὶ ἐκδιδοὺς τὴν θυγατέρα Δαρείῳ τὸν οἶκον πάντα τὸν ἑωυτοῦ ἐπέδωκε, ὡς μούνου οἱ ἐούσης ταύτης τέκνου. Ξέρξεώ τε δὴ δύο ἀδελφεοὶ ἐνθαῦτα πίπτουσι μαχόμενοι, ⟨καὶ⟩ ὑπὲρ τοῦ νεκροῦ τοῦ Λεωνίδεω Περσέων τε καὶ Λακεδαιμονίων ὠθισμὸς ἐγίνετοπολλός, ἐς ὃ τοῦτόν τε ἀρετῇ οἱ Ἕλληνες ὑπεξείρυσαν καὶ ἐτρέψαντο τοὺς ἐναντίους τετράκις. Hdt. 7.224–7.225.1

By this time most of them had had their spears broken and were killing the Persians with swords. Leonidas, proving himself extremely valiant, fell in that struggle and with him other famous Spartans, whose names I have learned by inquiry since they were worthy men. Indeed, I have learned by inquiry the names of all three hundred. Many famous Persians also fell there, including two sons of Darius, Abrocomes and Hyperanthes, born to Darius by Phratagune daughter of Artanes. Artanes was the brother of king Darius and son of Hystaspes son of Arsames. When

herodotus and thucydides: distance and immersion

151

he gave his daughter in marriage to Darius, he gave his whole house as dowry, since she was his only child. Two brothers of Xerxes accordingly fought and fell there. There was a great struggle between the Persians and Lacedaemonians over Leonidas’ body, until the Hellenes by their courageous prowess dragged it away and routed their enemies four times. Herodotus tells us that Leonidas and many Persians die, among them two sons of Darius and brothers of Xerxes. Three times a historical present (in italics) is used to mark these deaths as the climax of the battle.40 However, Herodotus’ aim here is clearly not to let us re-experience the events as if we are present. He does not depict the scene in any (spatial or visual) detail, the speed of the narration is high (summary narration rather than scene narration) and we are not informed of the protagonists’ thoughts.41 Herodotus himself is also prominently present as a retrospective, omniscient narrator. He tells us that he learned all the names of the 300 by inquiry. The two sons of Darius, Abrocomes and Hyperanthes receive a biographical digression in the style of Homer. In their descriptions of the Battles of Thermopylae and Sphacteria, Herodotus and Thucydides employ very different, in many ways contrasting, narrative styles. While Herodotus encourages the reader to keep a retrospective distance towards the past, Thucydides seduces us, at crucial moments, to immerse ourselves into the world of the past. Thucydides’ capacity to immerse the reader is based on a tight interplay of textual factors, such as descriptive detail, spatial language and the invisibility of the narrator. But perhaps the most important factor are sophisticated techniques used by Thucydides to manipulate the perspective from which we observe the events, inviting us to experience the events from the point of view of the historical actors.

Bibliography Adema, S.M., Speech and Thought in Latin War Narratives. Words of Warriors (Leiden 2017). Allan, R.J., Jong, I.J.F. de & Jonge, C.C. de, ‘Homerus’ narratieve stijl: enargeia en immersion’ [Homer’s narrative style: enargeia and immersion], Lampas 47 (2014) 202–223.

40 41

For the historical present’s function (among other functions) as a marker of narrative Peaks in episodes, see Allan 2007; 2009; 2011; 2013. See also Foster 2012: 201.

152

allan

Allan, R.J., Jong, I.J.F. de & Jonge, C.C. de, ‘From Enargaia to Immersion: the Ancient Roots of a Modern Concept’, Style 51.2 (2017) 34–51. Allan, R.J., ‘Sense and sentence complexity: sentence structure, sentence connection, and tense-aspect as indicators of narrative mode in Thucydides’ Histories’, in R.J. Allan & M. Buijs (eds.), The Language of Literature: linguistic approaches to classical texts (Leiden 2007) 93–121. Allan, R.J., ‘Towards a Typology of the Narrative Modes in Ancient Greek: Text Types and Narrative Structure in Euripidean Messenger speeches’, in S.J. Bakker & G.C. Wakker (eds.), Discourse Cohesion in Greek (Leiden 2009). Allan, R.J., ‘The Historical Present in Thucydides: Capturing the Case of αἱρεῖ and λαμβάνει’, in J. Lallot, A. Rijksbaron, B. Jacquinod & M. Buijs (eds.), The Historical Present in Thucydides. Semantic and Narrative Function (Leiden 2011) 37–63. Allan, R.J., ‘History as Presence. Time, Tense and Narrative Modes in Thucydides’, in A. Tsakmakis & M. Tamiolaki (eds.), Thucydides Between History and Literature (Berlin 2013) 371–379. Allan, R.J., ‘The History of the Future: Grammaticalization and Subjectification in Ancient Greek Future Expressions’, in F. Lambert, R.J. Allan & T. Markopoulos (eds.), The Greek Future and its History (Louvain-La-Neuve 2017) 43–72. Allan, R.J., ‘Persuasion by Immersion’, in T. Liao & A. Vatri (eds.), The Language of Persuasion (Leiden, forthc. a). Allan, R.J., ‘Narrative Immersion: Some linguistic and narratological aspects’, in J. Grethlein & L. Huitink (eds.), Narrative and Experience (Berlin, forthc. b). Allan, R.J., ‘Construal and Immersion: A cognitive Linguistic Approach to Homeric Immersivity’, in P. Meineck (ed.), The Routledge Handbook of Classics and Cognitive Theory (London, forthc. c). Bakker, E.J., ‘Verbal Aspect and Mimetic Description in Thucydides’, in E.J. Bakker (ed.), Grammar as Interpretation. Greek Literature in its Linguistic Contexts, Mnemosyne Supplement 171 (Leiden 1997) 7–54. Bakker, E.J., Pointing at the Past. From Formula to Performance in Homeric Poetics (Washington, D.C. 2005). Bakker, E.J., ‘Homer, Odysseus, and the Narratology of Performance’, in J. Grethlein & A. Rengakos (eds.), Narratology and Interpretation (Berlin 2009). Basset, L., Les emplois périphrastiques du verbe grec μέλλειν (Lyon 1979). Beck, D., Speech Presentation in Homeric Epic (Austin 2012). Brinton, L., ‘“Represented Perception”. A Study in Narrative Style’, Poetics 9.4 (1980) 363– 381. Brinton, L., ‘Non-anaphoric Reflexives in Free Indirect Style: Expressing the Subjectivity of the Non-Speaker’, in D. Stein & S. Wright (eds.), Subjectivity and Subjectivisation. Linguistic Perspectives (Cambridge 1995) 173–194. Chatman, S., Story and Discourse. Narrative Structure in Fiction and Film (Ithaca 1978).

herodotus and thucydides: distance and immersion

153

Connor, W.R., ‘Narrative Discourse in Thucydides’, in M.H. Jameson (ed.), The Greek Historians, Literature and History. Papers presented to A.E. Raubitschek (Stanford 1985) 1–17. Dancygier, B., & Sweetser, E. (eds.), Viewpoint in Language. A Multimodal Perspective (Cambridge 2012). Dancygier, B., ‘Blending and Narrative Viewpoint: Jonathan Raban’s Travels Through Mental Spaces’, Language and Literature 14.2 (2005) 99–127. Dancygier, B., The Language of Stories. A Cognitive Approach (Cambridge 2012). Dannenberg, H.P., Coincidence and Counterfactuality (Lincoln 2008). Eckardt, R., The Semantic of Free Indirect Discourse (Leiden 2004) Fauconnier, G., Mappings in Thought and Language (Cambridge 1997). Fludernik, M., The Fictions of Language and he Language of Fiction. The Linguistic Representation of Speech and Consciousness (London 1993). Foster, E., ‘Thermopylae and Pylos, with Reference to the Homeric Background’, in E. Foster & D. Lateiner (eds.), Thucydides and Herodotus (Oxford 2012) 185–214. Fowler, R., Linguistic Criticism ([1986] Oxford 1996). Gibbs, Jr., R.W. Embodiment and Cognitive Science (Cambridge 2005). Grethlein, J., Experience and Teleology in Ancient Historiography (Cambridge 2013). Grethlein, J., ‘The presence of the past in Thucydides’, in A. Tsakmakis & M. Tamiolaki (eds.), Thucydides Between History and Literature (Berlin 2013) 371–379. Hornblower, S., ‘Narratology and Narrative Techniques in Thucydides’, in S. Hornblower (ed.), Greek Historiography (Oxford 1994) 131–166. Jahn, M., ‘Windows of Focalization. Deconstructing and Reconstructing a Narratological Concept’, Style 30.2 (1996) 241–267. Jong, I.J.F. de, Narrators and Focalizers: the Presentation of the Story in the Iliad (Amsterdam 1987). Jong., I.J.F. de, Nünlist, R. & Bowie, A. (eds.), Narrators, Narratees, and Narratives in Ancient Greek Literature (Leiden 2004). Kühner, R. & Gerth, B., Ausführliche Grammatik der griechischen Sprache, II.1 (Hannover 1898). Langacker, R.W., Cognitive Grammar. A Basic Introduction (Oxford 2008). Leech, G. & Short, M., Style in Fiction. A Linguistic Introduction to English Fictional Prose. Second Edition ([1981] London 2007). Morrison, J.V., Reading Thucydides (Columbus, OH 2006). Nikiforidou, K., ‘Viewpoint and Construction Grammar’, Language and Literature 19.3 (2010) 265–284. Ooms, S. & Jonge, C.C. de, ‘The Semantics of ΕΝΑΓΩΝΙΟΣ in Greek Literary Criticism’, CPh 108 (2013) 95–110. Palmer, A., Fictional Minds (Lincoln 2004). Plett, H., Enargeia in Classical Antiquity and the Early Modern Age (Leiden 2012).

154

allan

Ruijgh, C.J., ‘Review Les emplois périphrastiques du verbe grec μέλλειν, by L. Basset (1979)’, Lingua 65.4 (1985) 323–333. Rijksbaron, A., ‘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 2012) 331–375. Ryan, M.-L., Possible Worlds, Artificial Intelligence, and Narrative Theory (Bloomington 1991). Ryan, M.-L., Narrative as Virtual Reality. Immersion and Interactivity in Literature and Electronic Media (Baltimore 2001). Ryan, M.-L., Narrative as Virtual Reality 2. Immersion and Interactivity in Literature and Electronic Media (Baltimore 2015). Sanders, J. & Redeker, G., ‘Linguistic Perspective in Short News Stories’, Poetics 22.1/22.2 (1993) 69–87. Sanders, J. & Redeker, G., ‘Perspective and Representation of Speech and Thought in Narrative Discourse’, in G. Fauconnier & E. Sweetser (eds.), Spaces, Worlds and Grammar (Chicago 1996) 290–317. Sanford, A.J., & Emmott, C., Mind, Brain and Narrative (Cambridge 2012). Short, M., Exploring the Language of Poems, Plays and Prose (London 1996). Simpson, P., Language, Ideology and Point of View (London 1993). Vandelanotte, L., Speech and Thought Representation in English. A Cognitive-functional Approach (Berlin 2009). Walker, A.D., ‘Enargeia and the Spectator in Greek Historiography’, TAPhA 123 (1993) 353–377. Wakker, G.C., ‘Future Auxiliaries or Not’, in E. Crespo, J. de la Villa & A.R. Revuelta (eds.), Word Classes and Related Topics in Ancient Greek. Proceedings of the Conference on ‘Greek Syntax and Word Classes’ held in Madrid on 18–21, June 2003, (Louvain-la-Neuve 2006), 237–525. Webb, R., Ekphrasis, Imagination and Persuasion in Ancient Rhetorical Theory and Practice (Farnham 2009). Wolf, W., Bernhart, W. & Mahler, A. (eds.), Immersion and Distance. Aesthetic Illusion in Literature and Other Media (Amsterdam 2013). Zwaan, R.A., ‘The Immersed Experiencer: Toward an Embodied Theory of Language Comprehension’, in B.H. Ross (ed.), The Psychology of Learning and Motivation 44 (New York 2004), 35–62.

part 2 Cannae

∵

chapter 7

Livy on Cannae: a Literary Overview Stephen Oakley

1

Introduction*

Livy’s account of Hannibal’s famous victory at Cannae, its preliminaries and its aftermath occupies the final third of his twenty-second book. It is a powerful and compelling narrative that brings one of his most powerful and compelling books to a close, a narrative so shaped that its climax comes not with the horrific carnage on the battlefield itself but in the responses of the Carthaginians and Romans to Hannibal’s victory.1 To provide background for the more detailed textual analyses that follow, this essay will examine some of Livy’s main themes.2 In a manner time-honoured in Livian studies, I shall use a comparison with Polybius (whose narrative is extant for the preliminaries to the battle, the battle itself, and for a small part of its aftermath) to clarify some of the characteristics of Livy’s narrative. Their narratives are summarised in the table in the appendix. No one has imagined that Polybius was Livy’s only source for Cannae; rather, if (as I believe) he did consult Polybius directly, material derived from Polybius is blended with material derived both from other writers, amongst whom Coelius Antipater is likely to have been the most notable, and from his own imagination. Although the comparison of the two writers remains illuminating even if Livy did not use Polybius directly, it becomes particularly pertinent if he did use him, since any divergence from Polybius then becomes evidence for Livy’s planning of his narrative.3 We shall see that for the prelim-

* This essay originated as a ‘key-note’ lecture at the conference on Thermopylae and Cannae; I thank the audience for discussion of it and the editors, Dr J. Briscoe, and Professor D.S. Levene for further corrections and improvements. The bibliography on Cannae is large and continues to grow (see e.g. Clark 2014: 62–81): since my concern here is with Livy as a writer, I discuss historical problems only when they have a bearing on literary or historiographical questions. I do not specify the author in references to Livy, the book number in references to Polybius book 3. 1 ‘Twice as many chapters are devoted to depiction of Roman reaction after Cannae as to the battle itself’ (Walsh 1961: 171). 2 I owe a particular debt to the fine analysis of Bruckmann 1936: 70–103. 3 An enormous amount of scholarly energy has been expended on attempts to identify Livy’s

© koninklijke brill nv, leiden, 2019 | doi:10.1163/9789004383340_008

158

oakley

inaries to the battle Livy’s account is so different from Polybius’ that he can hardly have made much, if any use of him, that for its opening phases it is so close that he must have been his prime source, and that for its final stages there are important differences. What Polybius’ account of the Cannae campaign offered above all was informed comment on strategy, tactics, and generalship, and Livy availed himself of it where it mattered most, for the opening of the battle. On Roman politics and the behaviour of Rome’s leading citizens Polybius had less to offer, and therefore it is not surprising that to find material for his account of the quarrel between Varro and Paullus, its disastrous outcome, and the tragic death of Paullus, he looked elsewhere.

2

Structure

The schematic table shows just how much more material there is in Livy than Polybius. Livy’s text itself offers one very clear point of division, with a paragraph in which Cannae is compared to the defeat by the Gauls at the Allia:

sources. As the extensive doxography in Luterbacher 1875: 1–3 shows, most of the positions that have been supported in subsequent scholarship had already emerged by 1875. In books 21–22 Livy’s narrative is often very similar to that of Polybius, but some details that cannot come from Polybius in passages otherwise very close to Polybius have made many scholars doubt that Livy worked with Polybius’ text in front of him. For these details see e.g. Luterbacher 1875: 32–35 or Kahrstedt 1913: 145–146. Therefore it has been argued that either Coelius Antipater (thus e.g. Kahrstedt 1913: 152–153, 212; Klotz 1940–1941: 101–108, 111–119, 144–150, 291– 293; Walsh 1961: 124–125; and Jumeau 1964: esp. 327–333; Livy’s numerous references to Coelius prove that he consulted him often) or another writer (thus e.g. Soltau 1894a: 17 and passim; 1894b: 82–83, who argued for Claudius Quadrigarius; and De Sanctis 1917: 180–181, 190, who did not name the writer) created a narrative very similar to Polybius’, either because he himself used Polybius (thus e.g. Soltau, De Sanctis) or because he used the same sources as Polybius (thus e.g. Kahrstedt and Jumeau); that portions of Livy’s narrative that are close to Polybius come from either Coelius or this other writer; and that portions that have no Polybian counterpart come from Coelius (if he were not the ‘polybianizing’ author) or later annalists, especially Valerius Antias, or both. Klotz did concede that Livy may occasionally have checked Polybius on certain points. However, that Livy did use Polybius directly as a source throughout much of books 21–22 had been argued briefly by Lachmann 1822–1828: 2.46 (and elsewhere), was established by e.g. Peter 1863 (see esp. pp. 46–48 on Cannae) and Luterbacher 1875, and has been reaffirmed recently by Levene 2010: 126–163; and it is much harder than some scholars allow to distinguish what in Livy comes from Coelius from what comes from e.g. Antias (thus, rightly, Luterbacher 1875: 50, anticipating more recent comment). That Appian, Hannibalica 17.73–28.120, gives a good indication of what stood in late annalistic narratives is quite likely but not certain, since Appian’s narrative may reflect both his own innovations and those made by other writers later than Livy.

livy on cannae: a literary overview

159

haec est pugna ⟨Cannensis⟩, Alliensi cladi nobilitate par, ceterum ut illis quae post pugnam accidere leuior, quia ab hoste est cessatum, sic strage exercitus grauior foediorque. fuga namque ad Alliam sicut urbem prodidit, ita exercitum seruauit: ad Cannas fugientem consulem uix quinquaginta secuti sunt, alterius morientis prope totus exercitus fuit. 50.1–50.3

This is the battle of Cannae, equal in repute to the disaster at the Allia: although it was less momentous in respect of what happened after the battle because of the inactivity of the enemy, it was more grievous and frightful in respect of the massacre of the army. For although the flight at the Allia betrayed the city, it saved the army; at Cannae scarcely fifty men followed the fleeting consul, but virtually the whole army was with the other consul who died. Introduced with a summative phrase that recalls an obituary notice,4 this small section divides the account of the battle from the account of its aftermath (50.4–61.15). A similar summative phrase is found in Polybius’ narrative (117.1). The absence of Livy’s Latin sources means that we cannot say for certain whether he took the idea for this summative remark from Polybius, but, if he did, he made some innovations: in Polybius it is in a slightly different position, and the account of the aftermath of Cannae which it divides from the account of the battle itself seems to have been of much less thematic weight than in Livy. Interpreters of Livy have often found it convenient to make another division: chapters 38.1–43.11 (or perhaps 45.4) deal with the preliminaries to the battle; chapters 45.5–50.3 with the battle itself. This helps analysis and corresponds also to the point at which Livy seems to make greater use of Polybius, but the narrative itself offers no clear point of division.5

3

The Preliminaries to the Battle

Polybius makes much of these preliminaries, but Livy far more, and here above all we find material that has no counterpart in Polybius, much of which is likely to come from Coelius and annalistic writers.

4 See my note on 6.20.14 (Oakley 1997–2005: 1.567). 5 For slightly differing divisions see e.g. Bruckmann 1936: 71 and Burck 1950: 92.

160

oakley

3.1 Some Factual Matters There are seven important factual matters, all of relevance for the literary interpretation of Livy, that Polybius and Livy handle differently,6 and only for the first do most modern historians find Livy’s narrative more reliable.7 (a) In Polybius, Paullus is the main focus of attention and is viewed as the leading Roman commander; in Livy, Varro is elected initially and presides over the election of Paullus. Polybius here reveals either his relative lack of interest in Roman politics, or a bias deriving from his knowledge of Paullus’ son (L. Aemilius Paullus, the victor at Pydna) and grandson (Scipio Aemilianus), or both, whilst Livy offers much information for the invention of which there was no motive.8 Nevertheless, what Livy says about Varro includes also a very large amount of historiographical and literary colouring which is found also in Appian (Hann. 17.73– 18.82). (b) In Polybius, after the election of the consuls for 216, because the new consuls are recruiting supplementary forces, Cn. Servilius Geminus, consul in 217 and now proconsul, continues to lead the Roman army facing Hannibal; in Livy, as in Appian (Hann. 17.74–17.76), there is no reference to Servilius’ command after the election of the consuls. More probably, the proconsuls did stay in command whilst the consuls recruited, and the recruiting took a considerable time. (c) Both authors tell us that the Romans recruited a very large army for 216 (Polyb. 107.9–107.15, Liv. 36.1–36.5, where Livy reports that his sources varied on the size of the reinforcements). In Polybius, the new recruits are trained by Servilius at the beginning of the year (106.4) and later regarded by Paullus as trained (108.6–109.4); in Livy (41.5), Hannibal refers to the rawness of the troops, saying that two thirds of them are recruits. It seems more likely that the Roman forces were trained before the battle.9 (d) In Polybius Hannibal moves from Gereonium to Cannae quite soon after the arrival of the consuls (107.1–107.2); since Cannae had been a Roman supply depot, the Romans have difficulty in feeding themselves (107.3– 107.6, perhaps 111.4). In Livy Hannibal moves rather later, and it is Hanni-

6 When the accounts are so different one may be tempted to question the value of a comparison, but the greater historical plausibility of Polybius’ account makes it a very useful base against which to characterize the less plausible innovations in Livy. 7 Compare the remarks of Breska 1889: 6. 8 For the difficulties of Polybius’ narrative, see Münzer 1934: 683. 9 Emphasis on the rawness of the Roman recruits helps to provide some exculpation for their defeat.

livy on cannae: a literary overview

161

bal who is short of food because of Fabian tactics (40.8–40.9, 43.2–43.5); see also App. Hann. 17.77–18.78. (e) Polybius’ account of what happened in Apulia may be divided up into the events of eight days, as the table shows. Livy records first a skirmish near Gereonium in which the Romans have the upper hand on a day in which Paullus was in command and reluctant to let the battle escalate; then Hannibal’s abortive ambush on the next day;10 then, after an indeterminate number of days, Hannibal’s move to Cannae, in which he pretends to be using a stratagem similar to that which had earlier failed; then Hannibal’s provocation of the Romans; and finally the battle itself. Polybius places a skirmish with a similar outcome at 110.4–110.7, but on a day in which Varro commanded, just before the battle of Cannae. Doubtless Livy or his sources placed it earlier because they had brought the consuls to the theatre of the war too soon; their placing Paullus in command allowed them to make Varro criticize him with typical lack of foresight.11 And it is very likely that the account of Hannibal’s failed stratagem was invented, in part to illustrate Hannibal’s fraus and Varro’s naivety; that the notion that he tried the reverse trick a little later arose from a doublet;12 and that Varro did not behave in the crass and stupid way in which Livy suggests. ( f ) In Livy, as we shall see, Fabius Maximus’ wish for ‘Fabian’ tactics, Varro’s criticism of these tactics, and Paullus’ attempt to adopt them loom large.13 In Polybius, the senate wishes to effect a decisive resolution of the war as soon as possible, and there is no division of opinion between those sensible few who wish to adopt ‘Fabian’ tactics and the foolhardy who wish to come to battle swiftly.14 The question of delay arises only just before the battle, when Paullus is not convinced that the Romans are about to fight on suitable terrain (3.110. 2–3.110.4, 3.110.7, 3.112.2). Polybius clearly sided with Paullus: he draws attention to Varro’s inexperience (3.110.3), and will later praise Paullus’ heroic death (3.116.9) and criticize Varro’s flight and

10 11 12 13

14

In App., Hann. 18.79–18.82, these two events are fused. See Bruckmann 1936: 76. Thus e.g. Soltau 1894a: 18; 1894b: 82. Perhaps not all traces of the version in which it is the wish of the Romans to effect a decisive battle have disappeared from Livy: Fabius predicts that Paullus will be isolated and that his troops will side with Varro (39.7); when Hannibal plots his ambush, the whole army sides with Varro in wishing to fall into the trap (42.7); and at 43.8, a significant passage, all except Servilius support Varro against Paullus in his wish to follow Hannibal to Cannae. As Burck 1951: 91, 95 notes, the essential aggressiveness of Roman plans is clear only in Polybius.

162

oakley

general conduct (3.116.13).15 Unless one counts the delay on the eve of the battle desired by Paullus as an instance of ‘Fabian’ tactics, these tactics are used only to train troops at the beginning of the year, and even then they are not associated with Fabius Maximus himself. Indeed, after he had laid down his dictatorship in 217, Fabius is not mentioned in what survives of Polybius’ narrative for 216. Given that Polybius had no axe to grind against Fabius, it is difficult to resolve the discrepancy except by concluding that the senate had no collective wish to adopt ‘Fabian’ tactics in 216.16 (g) In Polybius blame and abuse of Varro is largely limited to what has been described in the previous paragraph, but Livy’s whole narrative is shaped so as to throw as much blame as possible on Varro. How much of this fiction was created by Livy, and how much by earlier sources, there is no knowing: most probably a substantial part of it is owed to the Roman annalistic tradition, but in such a key episode Livy is likely to have made changes, perhaps substantial, to the material that he inherited. 3.2 Suspense and Delay Livy’s inclusion of much more material than Polybius introduces suspense and delay: every educated Roman would have known that the consular year of Varro and Paullus brought with it Rome’s most famous defeat, but only eleven chapters after Varro’s election does the main battle actually begin. During these chapters Livy builds up an atmosphere of gloom and foreboding: the new oath taken by the troops prepares us for a dangerous campaign out of the ordinary (38.2–38.5);17 Fabius Maximus warns Paullus against Varro and predicts a disaster greater than Trasimene (39.1–39.22); Paullus presages his own death (40.3); fortuna gives Rome victory in the skirmish, thereby encouraging Varro’s rashness (41.1); there is a pestis looming over the Romans (42.1); the Romans head to Cannae urgente fato ‘with fate pressing’ (43.9). Into all this the tale of Hannibal’s abortive ambush (41.1–42.12) fits neatly: the narrative appears to be hurtling us towards Roman disaster before being turned around by a classic Livian περιπέτεια introduced by forte (42.10),18 a reversal that prolongs the suspense. 15

16 17 18

The reference to the dispute allows Polybius a characteristic moment of didacticism (110.3): ἦν ἀμφισβήτησις καὶ δυσχρηστία περὶ τοὺς ἡγεμόνας, ὃ πάντων ἐστὶ σφαλερώτατον ‘there arose a dispute and unpleasantness between the leaders, which of all things is the most dangerous’. The essential correctness of Polybius’ view of these matters was reasserted by Erdkamp 1992. See Bruckmann 1936: 72; Klotz 1949: 62–63. For full analysis of Livy’s structuring of this episode, see Bruckmann 1936: 76–77; also Burck 1950: 95. For forte introducing a περιπέτεια, see Oakley 1997–2005: 1.416.

livy on cannae: a literary overview

163

3.3 Varro, Fabius, Paullus, and Hannibal Livy’s narrative of Cannae invites contrasts and comparisons, most involving Varro, who is at the centre of a nexus of themes. When Varro is introduced at 25.18–26.4, as the only senator prepared to support the unprecedented law that Minucius, Master of the Horse in 217, should be made equal in command to the dictator Fabius, the narrative oozes implicit disapproval and is replete with the normal prejudices of the Roman propertied and educated classes. Inter alia, we are told that Varro joins in the attacks on Fabius Maximus, whom the earlier narrative has shown in a good light.19 Many of these themes recur at 34.2, when Varro is introduced as standing for the consulship for 216. There follows at once the speech in which Varro’s relative Q. Baebius Herennius makes various sweeping accusations against the senate (34.3–34.11): that the consuls of 217 had deliberately remained with the army so that they could not hold the elections; that they had deliberately prevented the dictator L. Veturius Philo from holding the elections so that there would be interreges and the elections would be in the power of the patres; that Hannibal had been brought into Italy by the nobility; that Fabius had deliberately subtracted two legions from those with which he and Minucius were fighting in order that he could be seen to be Minucius’ rescuer; and, finally, that the consuls had dragged out the war. The presentation of these accusations in Herennius’ voice invites some response from readers. Since the narrative has shown us in general nothing but disaster afflicting those who charge into battle and try to defeat Hannibal and in particular nothing to support this interpretation of Fabius’ recent dictatorship—in it Minucius, Fabius’ Master of the Horse, is elevated to hold an imperium equal to that of the dictator, engages Hannibal, and is saved only by the intervention of Fabius himself (23.9–30.10)—, it is hard not to feel that the accusations are absurd, even if one admires the rhetorical skill with which Herennius manages to put a perverse color on the interpretation of events. Notable in this respect is his account of the appointment and resignation of Veturius Philus as dictator comitiorum habendorum causa (‘dictator appointed to hold the elections’), since in the narrative (33.9–34.1) this had seemed like

19

Since some of Varro’s earlier magistracies will have been mentioned in book 20, it is not absurd to argue that the introduction of the character sketch is surprising and owed to a narrative such as Coelius’ which was concerned only with the Hannibalic War and in which, therefore, this is Varro’s first appearance; thus e.g. Münzer 1934: 681. However, this may underestimate Livy’s artistry: if Varro’s aedileships and praetorships had drawn no comment from him, this is a perfectly sensible place for an introductory charactersketch.

164

oakley

the routine reporting of an appointment and resignation such as we find often in Livy’s first decade. The speech and the whole episode reminds us of episodes in the Struggle of the Orders as portrayed in earlier books, in which the patricians were often accused, sometimes with justice, of deliberately manufacturing wars.20 Varro is contrasted strongly with Quintus Fabius Maximus, who dominates the middle section of book 22, with the narrative voice reporting Hannibal’s appreciation of his tactics (12.5, 31.10). It reports also opposition to them from many Romans, who wished to see a decisive resolution of the war. Since Fabius was not present at Cannae, the dominance that he exerted over the narrative in the middle section of the book could not be easily repeated at its end, but he appears at 34.2, where he is said to have kept Varro in check; at 38.7, where he is criticized by Varro; at 38.13–39.22, where he gives a long speech of advice to Paullus; at 41.9, where the narrative voice imputes to Hannibal the idea of repeating the stratagem that he had used against Fabius in the previous year; at 44.5, where his tactics are again criticized by Varro; at 49.10, where the dying Paullus tells Lentulus to inform Fabius that he lived and died mindful of his advice; and at 55.4–56.1, where he is singled out as offering the advice that calms the senate and people. He gives the fullest exposition of his tactics in his speech to Paullus, notably at 39.13–39.15.21 The narrative voice implicitly supports the rightness of this Fabian strategy after it has been adopted by Paullus: it tells us at 43.1–43.5 that lack of supplies makes Hannibal face the prospect of mutiny among his troops and that he himself even contemplated flight to Spain, and at 43.6–43.9 that he is forced to move from Gereonium to Cannae.22 As we have seen, all this is very different from what Polybius tells us.

20

21

22

Cf. also 31.6.4 (with Briscoe 1973: 70–71). Yet perhaps there is one respect in which the speech should not seem entirely perverse: the senate had sent threatening embassies to Hannibal and Carthage at the time of the capture of Saguntum at the beginning of the war (21.6.8, 21.9.3–21.11.2, 21.18.1–21.19.5), the second of which had included Fabius Maximus as an ambassador and had led to a Roman declaration of war; one could argue that the effect of these embassies had been to invite Hannibal’s attack on Italy. On the echo in this speech of Ennius’ famous lines about Fabius, see recently Hardie 2012: 259–262. Skutsch 1985: ad loc. (pp. 530–531) reviews suggested contexts for Ennius’ lines and finds Badian’s suggestion of a speech of Aemilius Paullus in his Ligurian campaign of 181 BC the most attractive guess. If this is correct, Livy would make Fabius in a speech to Paullus echo words spoken in Ennius by Paullus’ son about Fabius himself. For Hannibal’s move as a vindication of Paullus’ strategy, see Bruckmann 1936: 78, and Burck 1950: 94; Burck observes that Livy’s imparting this information before the battle allows his subsequent narrative to be read in a tragic light.

livy on cannae: a literary overview

165

Varro is also set against Paullus, who had been persuaded to stand as his partner in the consulship by the nobility, who were looking for a man who was Varro’s equal (35.2–35.3). The result is described at 35.4: Is proximo comitiali die, concedentibus omnibus qui cum Varrone certauerant, par magis in aduersandum quam collega datur consuli On the next day on which an assembly could be held all the men who had stood against Varro gave way and Paullus was given to the consul more to make a pair for sparring than as a colleague. As so often in Livy, par contains a hint of gladiatorial imagery,23 and this sentence establishes a conflict between the two men that runs right up to Paullus’ death at Cannae. First, the two espouse different approaches to the war in the contiones that they make before departure (38.6–38.13): Varro promises to finish the war, whilst Paullus says that it is unwise to make plans before one has visited the theatre of war. Then Fabius offers Paullus private advice (39.1–39.22), in which he starts by pointing out that Paullus will have as much difficulty with Varro as Hannibal (39.4–39.5). When the two consuls leave, they are escorted by opposing constituencies: Paullus by the leading senators, Varro by the plebs. When they arrive at Gereonium, Roman success in an initial skirmish provokes a different response: Paullus calls a halt for fear of ambushes; Varro is furious (41.1–41.3). Hannibal, we are told (41.5) is well aware of the discord between the Roman commanders. When he plans an ambush, Paullus urges caution (42.4), Varro, reflecting the mood of the troops, gives the signal for advancing, and is stopped only by unfavourable auspices that are followed promptly by the discovery of the ambush (42.8–42.12). When Hannibal moves to Cannae, almost everyone agrees with Varro in wishing to follow him; Paullus and the proconsul Servilius disagree in vain (43.8–43.9). When Hannibal entices the Romans with skirmishes there is once again discord in the Roman camp: Varro wants to fight, Paullus does not (44.4–44.7). And on the day of the battle, when Varro is in command, Paullus has to fight against his will (45.5). In his account of the battle itself, Livy records how Paullus fought heroically (49.1–49.12) but mentions Varro only when he flees (49.14). Finally, Varro is set against Hannibal. Livy’s readers have seen Sempronius at the Trebia, Flaminius at Trasimene, and Minucius at Gereonium behave as though they are putty in Hannibal’s hands.24 By criticizing Fabius, Varro aligns 23 24

See e.g. Oakley 1997–2005: 2.140. For Livy’s characterization of these ‘rash’ commanders, see e.g. Will 1983; Mineo 1997 ≈

166

oakley

himself, perhaps unwittingly, with Flaminius. Certainly Fabius himself sees it this way. We have seen that in his speech to Paullus (39.4–39.5) he explicitly compares Varro to Flaminius, making Varro come out worse. Hannibal then treats Varro much as had he treated Sempronius, Flaminius, and Minucius. When the consuls arrive with their extra forces, Hannibal, who has almost run out of supplies, rejoices, despite the increased size of the forces ranged against him. We are not told why but may presume that he knows that Varro wants to fight a decisive battle. When Varro is successful in the first skirmish (41.1–41.3), Hannibal is hardly bothered: he knows that this will increase the overconfidence of Varro and the new troops, and he knows, too, all about the discord between the commanders and the lack of experience of the Roman troops (41.4). Accordingly he plans an ambush, setting a trap into which Varro nearly walks (41.6–42.12). When this fails, he is forced to move to Apulia to seek supplies, but, despite Paullus’ protests, Varro makes the Roman army follow him (43.1–43.11). When the two armies have arrived at Cannae, Hannibal wants to fight on terrain that suits his forces; Varro duly obliges by responding to his provocation (45.1–45.8). The contrast between the good and bad Roman commanders in books 21 and 22 is told in language in which reason (ratio) and prudence (prudentia) are opposed to rashness (temeritas) and reliance on fortune ( fortuna). The theme reaches its climax in the narrative of Cannae: temeritas and its cognates are found at 38.12, 39.20, 41.1, 41.4, 43.1, 44.5, 44.7; consultus and inconsultus at 38.11, 43.1, 44.7; cauere, cautus and their cognates at 38.11, 39.20, 42.4, and 47.8; prouidus and improuidus and their cognates at 39.20, 39.22, 42.4, 44.7; ratio at 39.9, 10, 21; ferox at 38.6 and 41.4.

4

The Battle Itself

4.1 The Structures of the Narratives The table in which the narratives of Livy and Polybius are compared shows how similar is Livy’s narrative to that of Polybius from 45.5 to 47.10, where he describes the disposition of the two armies, the fighting of the cavalry at the beginning of the battle, the first clash of the infantry, and the famous pincer-movement with which Hannibal surrounded the Roman legions. Here he must have made direct use of Polybius as his prime source:25 this is one of

25

2006: 256–269; Bernard 2000: 62, 135–141 (for Varro), 209–210, 252, 269, 288–289; Levene 2010: 170, with further bibliography. Thus, rightly, e.g. Luterbacher 1875: 56; Hesselbarth 1889: 337–338.

livy on cannae: a literary overview

167

those places for which any one who wishes to argue that Livy took his Polybian material from Coelius or another Roman source would have to argue for copying virtually without change by either this Roman source from Polybius or Livy from the Roman source.26 All Livy’s differences from Polybius are compatible with direct use of Polybius, even though they may have been suggested by his reading of other sources.27 Polybius describes first the Roman troop dispositions (noting at 113.3 the extra thickness of the line of battle), then those of the Carthaginians (with a comment on the moon-like shape of Hannibal’s line of battle), then how Hannibal’s African and Iberian and Gallic troops were armed, then the positions taken up by the Carthaginian and Roman commanders. Livy keeps all this, except that he ignores the comment on the extra thickness of the Roman line, tells us where the Roman commanders stood immediately after his reporting of the arrangement of the Roman line (in some respects a more logical place for this information), and cuts the reference to the moon-like shape.28 He makes Maharbal and not Hanno the commander on the Carthaginian right, where the Numidian cavalry were stationed, presumably so as to prepare for Maharbal’s role with the cavalry later in the story;29 whether he found this in a source or made the change on his own initiative, there is no knowing. He has also cut Polybius’ reference to Marcus Atilius’ command at Cannae (114.6), as he cuts also the subsequent reference to his death (116.11): this corresponds to his report at 40.6 that Atilius returned home and to subsequent references to magistracies that he performed in the war (23.21.6, 24.11.6, 24.43.4).30 Polybius is wrong: since he has no references to the presence at Cannae of Marcus Minucius, Master of the Horse in 217, to whose death Livy refers (49.16), he probably confused the two Marci.31 At 46.9 Livy has made the minor addition, almost certainly from another source, that the wind was blowing in the face of the Romans,32 and at 47.10 he states that 26

27 28 29 30 31 32

Nevertheless, this view has been popular: see e.g. Klotz 1940–1941: 145; and Burck 1950: 91 (both think that Livy here drew on Coelius). For those like Kahrstedt 1913: 212 who believe that Livy used a source who drew on Polybius’ sources the difficulty is even greater: the close similarity to Polybius has to be explained despite three writers who are likely to have made changes: Polybius, Livy’s source, and Livy himself. Contra Cornelius 1932: 42–47, who argued that Polybian material here came to Livy through another source. He will refer to it at 47.8, and perhaps thought one reference was enough; see also Burck 1950: 96. Thus, rightly, Cornelius 1932: 46; Bruckmann 1936: 82–83. See e.g. Breska 1889: 6. Thus Walbank 1957: 442, 446. The full development of the motif at Sil. 9.486–9.523 and App. Hann. 22.99–22.100 is likely to reflect what stood in some of Livy’s sources and shows his restraint.

168

oakley

the Romans, now tired, were in effect entering a second battle against fresh troops—more on both these remarks later. After 47.10, although Livy broadly keeps to Polybius’ order of events, his detail is much less close to the Greek writer’s, and he must have drawn more heavily on his Roman sources.33 Appian, Han. 21.92–26.114, who recounts that Hannibal won the battle because of four stratagems (making use of the wind, making cavalry pretend to surrender themselves, making other troops pretend to flee, and hiding troops in a defile so that they could ambush Romans pursuing those pretending to flee) probably gives some idea of what could be found in these sources. Livy has no equivalent to Polyb. 116.1–116.4, where Paullus’ move from the right wing to the centre is described. Both writers then have an account of the cavalry battle between the Numidians and Rome’s allies (48.1– 48.6; Polyb. 116.5–116.8), but whereas Polybius draws attention to the skill of the Punic forces and especially of Hasdrubal, their commander, Livy reports the episode as an instance of Punic trickery, and without a comparison with Polybius it is hard to work out why it was important for the outcome of the battle.34 Both writers then describe the death of Paullus (49.1–49.13; Polyb. 116.9), but Livy at much greater length, emphasizing its tragic aspect and linking it to the earlier theme of the dispute over tactics. For anyone interested in historical reconstruction, his failure to describe Paullus’ move into the centre makes it unclear how, after being on the right when last mentioned (45.8), he could now be opposed to Hannibal (49.2), who commanded the Carthaginian middle (46.7); but history’s loss is literature’s gain, since Paullus’ final moments make for a graphic end to the battle narrative. Both writers then describe the end of the battle (49.13; Polyb. 116.10–16.11), but whereas Polybius has given a coherent account of the defeat of the Roman legions, Livy’s concentration on the cavalry makes it hard to understand quite what had happened.35 4.2 Explanations of Hannibal’s Victory In his final comment on the battle Polybius, who had noted earlier (110.2) that Paullus was unhappy with the terrain in which Varro proposed to fight because it favoured Hannibal’s cavalry, explains Hannibal’s success as due to the supe-

33

34 35

See e.g. Lachmann 1822–1828: 2.46; Luterbacher 1875: 56; and Hesselbarth 1889: 346–347. Kahrstedt 1913: 212; De Sanctis 1917: 190; and Klotz 1940–1941: 145 note the different texture of the narrative but have to explain the change differently. Thus, rightly, e.g. Peter 1860: 47. Well pointed out by e.g. Luterbacher 1875: 57. Luterbacher notes that Livy’s non-Polybian sources may have had Paullus commanding the Roman middle, as he does at App. Han. 19.86.

livy on cannae: a literary overview

169

riority of his cavalry (117.4). His narrative points also to three other factors that may be thought to have a bearing on Hannibal’s success: the damage caused by the dispute between those in command (110.2–110.3), the skill with which Hannibal used his moon-like line of battle (115.7–115.12), and Hasdrubal’s skill with the cavalry (116.7). Livy is rather different. He reduces reflection on Hannibal’s superiority in cavalry to (44.4): Hannibal spem nanctus locis natis ad equestrem pugnam, qua parte uirium inuictus erat, facturos copiam pugnandi consules, dirigit aciem lacessitque Numidarum procursatione hostes ‘Hannibal, hoping that the consuls would provide the opportunity to fight on terrain made for fighting with cavalry (in which part of his army he was invincible), arranged his line of battle and enticed the enemy with a charge of his Numidians’. Perhaps also relevant in this context is 22.49.1–22.49.5, where he describes how the Roman cavalry that accompanied Paullus dismounted when Paullus had become too weak to control his own horse. With his narrator’s voice Livy offers no judgement, but Hannibal (to whom it had been reported that Paullus had ordered his cavalry to dismount) makes a scathing comment (22.49.3): ‘Quam mallem, uinctos mihi traderet!’ ‘I should have preferred it if he had handed them over in chains!’.36 Livy removes (as we have seen) one of the references to Hannibal’s moonlike line of battle, and his less technical description of how the encirclement worked offers his readers less of an invitation to reflect on Hannibal’s skill;37 and we have just seen that he changes Polybius’ reference to Hasdrubal’s skill with the cavalry into an instance of Punica fraus ‘Carthaginian treachery’ (48.1). Instead, the thematic expansion of the prelude to the battle highlights above all the disunity in the state caused by the improvidence and folly of Varro, who follows a long line of other improvident commanders and who will be singled out in the final sentence of the book as the main cause of Rome’s disaster, and the disunity in the state caused by his political stance.38 Also explicitly responsible is the Punica fraus just mentioned and the urging of Fate (43.9). In this context two of Livy’s additions to Polybius are relevant: the wind blowing in the faces of the Romans (46.9), and the fact that the Roman troops, tired after defeating the Iberians and Celts, have to face fresh troops. Livy, always reluctant to say that competently commanded Roman infantry were defeated in a fair fight,39 seems to imply that the Romans lost and fled only because of their 36 37 38 39

I.e. Paullus might as well have bound his cavalry and handed them over. Contrast Pol. 115.7–115.12 with Liv. 47.6–47.10. Thus, rightly. Will 1983: 175–176; and Mineo 1997: 126 ≈ 2006: 257–258. Well brought out by Bruckmann 1936: passim; see p. 85 on Cannae. But Polybius never implies anything other than that the Romans fought bravely (see esp. 116.10).

170

oakley

being encircled in more than one place in the rear (47.9–10, 48.5–6), but his refusal to go into full details of how Hannibal’s moon-like formation actually worked, leaves it rather harder to work out exactly what happened on the field of battle—perhaps he felt that too much detail on Hannibal’s skill would work to the discredit of Rome’s troops.

5

The Aftermath40

5.1 General The third section of Livy’s narrative (50.4–61.15), which deals with the aftermath of the battle, does contain some of the same themes as Polybius’ parallel narrative, such as the revolt of Rome’s allies and the calmness of the senate, but it is much longer, and is packed with extra material. Livy probably drew his material for this section from Roman sources, refashioning it by means of his own creative imagination.41 A unifying theme of Livy’s account of the battle and its aftermath are his references to the two Roman camps (40.5–40.6, 44.2–44.3, 45.2, 45.6, 49.13, 50.4– 50.12, 52.1–52.3, 52.4–52.6, 59.1–60.27). These are mentioned also by Polybius (110.8–110.11, 112.3, 113.1–113.2), and doubtless the Romans really did have two camps. However, Livy uses this circumstance to great artistic effect. What had seemed to be a minor detail, a motif quietly in the background, turns out later in the narrative to be of the utmost importance: the behaviour of the men in these camps, in particular the escape of some, and the surrender of others, will lead to a debate that will become the climax of the book. By devoting so much space to recounting what happened in the immediate aftermath of the battle—the escape of some Romans, the surrender of others, the debate as to whether the captives should be ransomed—Livy ensures that the description of the battle of Cannae itself does not become the climax of the narrative.42 It soon becomes clear that it is the response to its outcome that matters more. Gone after the battle is the debate about Fabian tactics, gone 40 41

42

Different aspects of the text are studied in the analysis of Jaeger 1997: 99–105. But our knowledge of how Livy’s sources handled these events is as usual very restricted. That at least some of them were to be found in Coelius is shown by FRHist 15 F22 = Gell. 10.24.6, a version of Maharbal’s offer to lead his cavalry on Rome (compared with Livy’s version below). Acilius’ version of the story of the embassy of the captives is recounted at Cic. Off. 3.113–3.115 = FRHist 7 F2. Polybius (6.58.1–6.58.13) also told the story and may have influenced Livy’s account. Burck 1950: 92 sees this well, making the point that Polybius was interested more in military events, Livy in Roman character.

livy on cannae: a literary overview

171

the contrasting of the wise and improvident Roman. In their place the narrative raises other questions: how will the Romans respond to their defeat? how will Hannibal respond to his victory? Behind these questions lies an even more important theme. Livy has already reported Hiero of Syracuse as saying that (37.3) probe sciat magnitudinem populi Romani admirabiliorem prope aduersis rebus quam secundis esse ‘he knew well that the greatness of the Roman people evoked more admiration in adverse circumstances than in success’, but will Cannae confirm that this is true of the Roman character?43 And this question is made sharper by both explicit and implicit comparisons with the Carthaginians. The nature of the themes that will come to the fore in the final section of the narrative is seen most clearly at 54.7–54.11. There Livy says that he will not attempt to capture in words the panic and tumult at Rome when it was believed that there were no survivors from Cannae whatsoever. Part of the explanation for this may lie in his having already described such a scene after Trasimene (7.6–7.14), but his comment at 54.10–54.11 takes us to a deeper reason: nulla profecto alia gens tanta mole cladis non obruta esset. (11) compares cladem ad Aegates insulas Carthaginiensium proelio nauali acceptam, qua fracti Sicilia ac Sardinia cessere, uectigales44 ac stipendiarios fieri se passi sunt, aut pugnam aduersam in Africa, cui postea hic ipse Hannibal succubuit; nulla ex parte comparandae sunt nisi quod minore animo latae sunt Assuredly no other nation would not have been overwhelmed by such a weight of disaster. (11) Compare either the Carthaginians’ disaster in the naval battle at the Aegates islands (after this they departed from Sicily and Sardinia with their power broken and allowed themselves to become taxpayers and tributaries) or the defeat in Africa to which Hannibal himself later succumbed—there is no respect in which these ought to be compared, except that they were endured with lesser courage. In this final section the theme of unflinching Roman courage (animus) will be prominent, and a description of tumult and lamentation would have detracted from this. Much about the Roman national character, however, has emerged already from the account of the battle itself. Although some Romans will flee, many will 43 44

Compare the remarks of Burck 1950: 98–99. P (= Paris Lat. 5730), the only ms. with authority, reads in uectigales; I have adopted a deletion suggested by Briscoe 2016 ad loc.

172

oakley

fight bravely and be killed (47.4–47.10); the conduct of the consul Paullus, who heroically sacrifices himself, epitomizes this bravery (49.1–49.13). The ferocity of the Roman fighting spirit is shown by the gruesome description (51.5–51.9) of the Roman dead and half-dead, especially that of the Roman who died biting a Numidian who was on top of him. Implicit comment on the unbroken Roman spirit is to be found in the way the senate in general, and Fabius Maximus in particular, make calm preparations when it is believed that Hannibal is about to march on the city (54.7–55.8). Similarly, when Varro’s letter finally arrives and renewed lamentation breaks out, the senate puts a thirty-day limit on mourning and continues to make dispositions for the war as a whole (56.1–57.12). 5.2 The Two Roman Camps45 Dominant in this final section are the contrasting fates of those in the two Roman camps: these will illustrate both Rome’s attitude to the war and a flaw in Hannibal’s generalship. The account of how Sempronius Tuditanus led some from the smaller camp to the larger (50.4–50.12) has already provided an instance of Roman heroism and courage in the immediate aftermath of the battle; others had escaped from the larger camp; and the young Scipio’s stopping the flight overseas of Metellus and others is another example of Roman courage (53.3–53.13). Yet others had surrendered (52.1–52.4). Most of what the narrative voice tells us about the surrender of the two camps may be found in the following passages: septem milia hominum in minora castra, decem in maiora, duo ferme in uicum ipsum Cannas perfugerunt, qui extemplo a Carthalone atque equitibus nullo munimento tegente uicum circumuenti sunt 49.13

Into the smaller camp fled seven thousand men, into the larger ten thousand, and into the village of Cannae almost two thousand, who were immediately surrounded by Carthalo and the cavalry, since no fortification protected the village

45

The analyses of this episode given here and in Beltramini 2016 were conceived independently but are very similar to each other. I have added some references to Beltramini’s article. See also Jaeger 1997: 104–105 and Chaplin 2000: 57–62, 71 (part of a wider examination of Cannae as an exemplum at pp. 53–72).

livy on cannae: a literary overview

173

binis in castris cum multitudo semiermis sine ducibus esset, nuntium qui in maioribus erant mittunt, dum proelio deinde ex laetitia epulis fatigatos quies nocturna hostes premeret, ut ad se transirent: uno agmine Canusium abituros esse 50.4

When a half-armed mass had formed in the two camps without leaders, those who were in the larger camp sent a message: while night-time and rest held back the enemy who were tired by the battle and then by their celebratory dinner, they should cross to them; they would then leave in one column for Canusium spoliis ad multum diei lectis, Hannibal ad minora ducit castra oppugnanda et omnium primum brachio obiecto flumine eos excludit; (2) ceterum, omnibus ab labore uigiliis uolneribus etiam fessis, maturior ipsius spe deditio est facta. pacti ut arma atque equos traderent, in capita Romana trecenis nummis quadrigatis, in socios ducenis, in seruos centenis, (3) et ut eo pretio persoluto cum singulis abirent uestimentis, in castra hostes acceperunt traditique in custodiam omnes sunt, seorsum ciues sociique. (4) dum ibi tempus teritur, interea cum ex maioribus castris, quibus satis uirium aut animi fuit, ad quattuor milia hominum et ducenti equites, alii agmine alii palati passim per agros, quod haud minus tutum erat, Canusium perfugissent, castra ipsa ab sauciis timidisque eadem condicione qua altera tradita hosti 52.1–52.4

After spoil had been collected for much of the day, Hannibal led his forces forward to attack the smaller camp and first of all by blocking the way with a rampart he shut them off from the river. (2) Even so, because they were all tired with hard work, sentry-duty at night, and wounds, the surrender was made quicker than even he had expected. After they had agreed that they would hand over their arms and horses, that each Roman was to be valued at three hundred quadrigati, each ally at two hundred, and each slave at a hundred, (3) and that if this sum was paid they could depart with their basic underclothing, the enemy took them into their camp and they were all handed over to guards, with the Romans and allies kept separate. (4) In the meantime, while time was being wasted there, those who had sufficient energy or courage—namely four thousand infantry and two hundred cavalry—had fled from the larger camp, some in a column, some spread out through the countryside (which latter method was no

174

oakley

less safe), and had reached Canusium. The camp itself was handed over by the wounded and fearful on the same terms as the other. From this we learn that Hannibal shuts the smaller camp off from the river and hence its supply of water, and that because the Romans in it are exhausted with exertion, sleeplessness, and wounds, they surrender more quickly than he had expected. As for the larger camp, those who had enough courage (animi) or strength escape, those who are wounded or fearful (timidi) surrender themselves. In the context of escape from the camps, references to animus have been introduced already at 50.6 and 12. The spokesman for the captives who had been sent by Hannibal to Rome to argue for their own ransoming recounts these events rather differently: non enim in acie per timorem arma tradidimus sed cum prope ad noctem superstantes cumulis caesorum corporum proelium extraxissemus, in castra recepimus nos; (4) diei reliquum ac noctem insequentem, fessi labore ac uolneribus, uallum sumus tutati; (5) postero die, cum circumsessi ab exercitu uictore aqua arceremur nec ulla iam per confertos hostes erumpendi spes esset nec esse nefas duceremus quinquaginta milibus hominum ex acie nostra trucidatis aliquem ex Cannensi pugna Romanum militem restare, (6) tunc demum pacti sumus pretium quo redempti dimitteremur, arma in quibus nihil iam auxilii erat hosti tradidimus. 59.3–59.6

For we did not hand over our weapons in fear on the battlefield but, after we had dragged out the battle almost to nightfall standing on heaps of slaughtered bodies, we retreated to our camp. (4) For the rest of the day and for the following night, tired with hard work and wounds, we protected our rampart. (5) On the next day, when we had been deprived of water by a victorious army and there was no hope of breaking out through the packed ranks of the enemy and when we did not think it an outrage that, after 50,000 men from our line of battle had been butchered, some Roman troops should survive from the Battle of Cannae, only then did we agree the price for which we might be sent away if ransomed and did we hand over our arms, which were now of no help to us, to the enemy. That they did not hand themselves over on the field of battle is vouched for by the narrative (although their spokesman omits to say that many others had fought to the death, whereas they had fled to their camps), and we may imagine some truth in the notion that they had protected their uallum through-

livy on cannae: a literary overview

175

out the remainder of the day and night; but anyone who remembers 50.4 will feel uncomfortable: there the message from the larger camp is that the enemy was exhausted or feasting. Although the narrative voice nowhere offers similar information, although the desirability of moving from the smaller to the larger camp is contested by some of those in the smaller camp, who hold that (50.5) plena hostium omnia in medio essent ‘all the middle ground is full of enemy troops’, and although Tuditanus himself refers to enemy who (50.8) inordinati atque incompositi obstrepunt portis ‘out of their ranks and without any order make a noise at our gates’, those who choose to escape with Tuditanus do so with relative ease (50.10–50.11), and few readers of chapter 50 are likely to feel that the enemy made any very serious attempt to capture the smaller camp in the night. What is said about being shut off from water and there being no hope of escape is likewise true, but conspicuous by its absence is any reference to the escape of Tuditanus, who had specifically warned that they should go before the way was blocked by the movement of enemy forces at daylight (50.8–50.9); and the alert reader will remember that they had surrendered more swiftly than Hannibal had anticipated. There is a varying response in the Roman senate to the captives’ plea. Many are sympathetic, some suggesting that the captives should be redeemed at public expense, others at private expense, perhaps with the loan of money from the treasury. Then Manlius Torquatus is introduced by Livy’s narrative voice with a comment on how his character might be viewed:46 […] tum T. Manlius Torquatus, priscae ac nimis durae, ut plerisque uidebatur, seueritatis, sententiam ita locutus fertur 60.5

Then Titus Manlius Torquatus, a man (as most people thought) of oldfashioned and too unyielding strictness, is said to have spoken his opinion in this way.47

46 47

On Manlius’s role in Livy’s account of the Second Punic War, generally as a representative of strict old-fashioned views, see the brief but perceptive remarks of Pelling 1989: 203–205. uidebatur is the reading of e.g. mss RC; P1 has uideatur, Pc uideuatur. Were uideatur correct, then a response is invited from readers: many would think that Torquatus’ behaviour was too severe. This seems unnatural. The reading of Pc probably has authority only as a conjecture—by uideuatur the corrector seems to have meant uidebatur as offered by P’s mediaeval descendants—but offers much more natural sense: Manlius’ views seemed too severe to many of those in the senate at the time, something that is hardly unsurprising when many had just argued for the opposite point of view.

176

oakley

He ruthlessly exposes the untruths offered by the captives, referring to Sempronius close to the beginning of his speech (60.8); particularly scathing is his passage on the surrender of the camp (60.2–60.5), and his view that those who stayed in the smaller camp were cowards finds support from the earlier pronouncements of the narrative voice, which affirms that some of those who did not wish to escape to the larger camp approved the plan but lacked the courage to carry it out (50.5). But Manlius, too, is guilty of exaggeration. In the main narrative of chapter 50, when the message from the larger camp arrives, there is deliberation about whether to abandon the smaller camp, Tuditanus makes his rousing speech, and then he charges off (§10): haec ubi dicta dedit, stringit gladium cuneoque facto per medios uadit hostes ‘after he had said this, he drew his sword and arranging his men in a wedge-like formation, he rushed through the middle of the enemy’; in Manlius’ version the cowards fail to follow Tuditanus despite a night of urging (60.10): nocte prope tota P. Sempronius Tuditanus non destitit monere, adhortari eos, dum paucitas hostium circa castra, dum quies ac silentium esset, dum nox inceptum tegere posset, se ducem sequerentur: ante lucem peruenire in tuta loca, in sociorum urbes posse ‘for almost the whole night Publius Sempronius Tuditanus did not stop warning and exhorting them to follow him as their leader, while there were only a few enemy around the camp, while it was the time of rest and there was no noise, while night could cover their plan. He said that before dawn they could reach safe havens, the towns of their allies.’ Similarly, Manlius exaggerates by saying that Sempronius had to cut his way through his own side before the enemy (60.17) and by referring to the haste with which the smaller camp was surrendered: in the narrative, cited above, Hannibal spends much of the day collecting spoil and then turns to the smaller camp, which surrenders more swiftly than he had expected; in Manlius’ version the surrender takes place not long after dawn (60.24): orto sole ab hostibus ad uallum accessum; ante secundam horam, nullam fortunam certaminis experti, tradiderunt arma ac se ipsos ‘when the sun had come up, the enemy marched up to the rampart; before the second hour they handed over themselves and their weapons, having made no test of their luck in battle’.48

48

To bolster their arguments both the spokesman and Manlius cite exempla. Beltramini 2016: 176–182 shows that the spokesman’s alleged precedents for the ransoming of captives (after Rome had been sacked by the Gauls and during the Pyrrhic War; see 59.8) are weak, Manlius’ precedents of daring in desperate situations (Decius Mus in 343 and Calpurnius Flamma in 258; see 60.11) rather more cogent. He notes (following e.g. Chapman 2000) also how the use and misuse of such exempla in the history itself illustrates what Livy enjoins his readers to do in his Praefatio and that Manlius’ references to Sempronius show Sempronius in the process of becoming an exemplum.

livy on cannae: a literary overview

177

The exaggerations and half-truths on each side are typical of such speeches, both in Livy and the many other works of ancient literature in which speeches are included.49 These speeches allow the author’s skill at inventing arguments to be admired and present different voices giving differing interpretations of events for readers to ponder, accept, or reject. The narrative gives Manlius the final word, records that his sententia was adopted,50 and goes on to show that, after the senate had adopted it, Rome eventually will be victorious. Implicit is the conclusion that Rome did not ransom these captives, highly convenient though it would have been to do so, because, in Manlius’ words, these men lacked the animus (60.13) to save themselves. Yet Livy offers no narratorial judgement on the outcome of the debate, merely noting in 61.1 that financial as well as moral considerations affected the outcome; and his description of the pain felt by the relatives of the captives at 60.1–2 and especially 61.3 leaves no doubt about the magnitude of the decision.51 5.3 Hannibal’s Failure to Attack Rome In the brief dialogue that Livy gives to Aemilius Paullus and Cn. Lentulus before Paullus’ death, Paullus tells Lentulus (49.10) ‘Abi, nuntia publice patribus urbem Romanam muniant ac priusquam uictor hostis adueniat praesidiis firment’ ‘Go! Report in public to the senators that they should fortify the city of Rome and should secure it with garrisons before the victorious enemy arrives’, thereby introducing a new theme that will be important in the final chapters of book 22: the Roman defeat has been so comprehensive that the city of Rome was in grave danger of an assault from Hannibal. That in turn is linked closely to aspects of the theme ‘Can Rome continue the struggle? Or is the war now over?’ The first theme of the danger of an assault on Rome appears next in Livy’s comparison of Cannae to the Allia (50.1–2): Haec est pugna ⟨Cannensis⟩, Alliensi clade nobilitate par, (2) ceterum ut illis quae post pugnam accidere leuior, quia ab hoste est cessatum, sic strage exercitus grauior foediorque ‘This is the battle of Cannae, equal in repute to the disaster at the Allia: although it was less momentous in respect of what happened after the battle because of the inactivity of the enemy, it it was more grievous and frightful in respect of the massacre of the 49

50

51

Beltramini 2016: 182–191 analyses also 26.2.7–26.3.12 and 38.58.3–38.59.11, where a similar phenomenon is found. Livy likes to show what happens when history is used in rhetoric. See also Adema (this volume) for an analysis of the speeches in the Cannae-episode. In the words of Pelling 1989: 204, ‘Torquatus’ defence of tradition could not then be dismissed: it touched on the core of Roman pride, dignity, and resilience, the qualities so vital to Rome’s survival’. The silence of Livy’s narrative voice on matters of morality is more common than is sometimes realised.

178

oakley

army’, in which note especially cessatum. Both the theme of the danger of an assault on the city and the theme of whether Rome can continue the struggle then reappear in one of the most famous scenes in Latin literature, when Maharbal requests a rapid strike on Rome itself: Hannibali uictori cum ceteri circumfusi gratularentur suaderentque ut tanto perfunctus bello diei quod reliquum esset noctisque insequentis quietem et ipse sibi sumeret et fessis daret militibus, (2) Maharbal praefectus equitum, minime cessandum ratus, ‘Immo ut quid hac pugna sit actum scias, die quinto’ inquit ‘uictor in Capitolio epulaberis. sequere; cum equite, ut prius uenisse quam uenturum sciant, praecedam.’ (3) Hannibali nimis laeta res est uisa maiorque quam ut eam statim capere animo posset. itaque uoluntatem se laudare Maharbalis ait; ad consilium pensandum temporis opus esse. (4) tum Maharbal: ‘non omnia nimirum eidem di dedere. uincere scis, Hannibal; uictoria uti nescis.’ mora eius diei satis creditur saluti fuisse urbi atque imperio. 51.1–51.4

When others surrounded the victorious Hannibal and congratulated him and urged him after finishing such a war himself to take some rest for the remainder of the day and the following night and to grant the same to his tired troops, Maharbal, the commander of the cavalry, thinking that this was no time for dallying spoke, “On the contrary! So that you may know what has been accomplished in this fight, in four days time you will banquet on the Capitol. You follow; I shall lead the way with the cavalry so that they realise that I have arrived before they know that I am coming.” (3) For Hannibal this prospect appeared too good to be true and greater than he could grasp in his mind. Therefore he praised Maharbal’s inclination, but said that he needed time for weighing his plans. (4) Whereupon Maharbal said, “The gods have not granted everything to the same person. Hannibal, you know how to be victorious; you do not know how to use a victory.” It is generally believed that the delay on that day saved the city and its empire. cessandum picks up cessatum in the previous chapter, and the second word uictori sets up the main theme of the rest of the paragraph: what should happen after such a victory? We are fortunate that of Coelius Antipater’s version of this scene, which Livy almost certainly used, one sentence survives:52 52

For comparison of Livy and Coelius, see e.g. Badian 1966: 17; T.J. Cornell 2013: 3.127; J. Briscoe, ibid., 3.251, and Jaeger 2015: 69.

livy on cannae: a literary overview

179

Si uis mihi equitatum dare et ipse cum cetero exercitu me sequi, diequinti Romae in Capitolium curabo tibi cena sit cocta FRHist 15 F22 = Gell. 10.24.6

If you want to give me the cavalry and yourself to follow me with the rest of the army, I shall make sure that at Rome in four days time a dinner is cooked for you for eating on the Capitol. A comparison with Maharbal’s first words in Livy is instructive: while Coelius’ Maharbal requests, Livy’s uses a more urgent imperative; Livy’s ut quid hac pugna sit actum scias points to the real significance of Cannae; Coelius’ curabo tibi cena sit cocta suggests the mundanities of preparing a meal, whereas Livy’s epulaberis suggests a celebratory banquet; and uictor, which has no equivalent in Coelius, continues the theme of uictori, is pointed in a context in which final victory is won only when the Capitol is taken, and looks forward to the famous Vincere scis, Hannibal, uictoria uti nescis.53 Maharbal’s point of view receives some endorsement from Livy’s narrative voice in the final sentence quoted,54 and further endorsement comes from the Romans. When news of the defeat arrives the praetors summon the senate and did not doubt that (55.2) deletis exercitibus hostem ad oppugnandam Romam, quod unum opus belli restaret, uenturum ‘After destroying our armies the enemy would come to attack Rome, the one task in the war that remained’, where uenturum looks back to the same word at 51.2. After Hannibal has captured the smaller camp, his bargaining with those captured is brutally summarized thus by the narrative voice (52.1): Dum ibi tempus teritur (‘While time was wasted there’), and we are told that while the bargaining continued more men were able to escape to Canusium from the larger camp. The notion of time-wasting 53

54

Something of Cato’s version of these events in his Origines may be recovered from FRHist. 5 F78 (= Gell. 10.24.7) and 79 (= Gell. 2.19.9). From F79 it emerges that on the day after the battle Hannibal decided to take the advice of his cavalry commander that he should march on Rome but is told that he was too late: the Romans were now alerted. Livy’s version, in which Maharbal replies at once, is more dramatic. Unfortunately the small compass of the fragment of Coelius means that we cannot be sure whether this innovation is due to Livy or Coelius; similarly not enough of Cato arrives to determine whether Livy, though using Coelius as his primary source, also alludes to him (but Jaeger 2015: 69 could be right that Livy’s imperative sequere was suggested by Cato’s mitte). But note the qualification that comes with satis. Modern historians tend to agree that Hannibal, who had difficulty besieging even smaller cities, would have achieved nothing with an attack on Rome, protected as it was by its massive Servian walls. This point of view is never encouraged by the text. For a recent full discussion of Hannibal’s strategy in the early years of the Second Punic War see Fronda 2010: 34–52; more briefly, id. 2011: 249–254.

180

oakley

recurs in letters of Varro that are sent to the senate (56.2–56.3): while Varro collected survivors, Poenum sedere ad Cannas, in captiuorum pretiis praedaque alia nec uictoris animo nec magni ducis more nundinantem ‘The Carthaginian was sitting around at Cannae, haggling over the price of captives and over other booty and was displaying neither the resolution of one victorious nor the manner of a great leader’. And it recurs when Livy’s narrative voice turns to introduce the tale of the ambassadors who were sent to Rome to negotiate their own ransom (58.1): Namque Hannibal secundum tam prosperam ad Cannas pugnam uictoris magis quam bellum gerentis intentus curis (‘Hannibal, after so successful a battle at Cannae, being preoccupied with the concerns of a victor rather than one who was still waging war […]’). Therefore one lesson raised by Livy’s narrative of Cannae is what it takes to win a great war, tanto perfungi bello (51.1). Varro had earlier said (41.3) that the war could have been won (debellari) had he not been shackled; the text reveals that this is a disastrous delusion and that he cannot win even a battle. Hannibal at least knows how to win a battle but not a war. But a nation with a character like Rome’s will eventually be victorious. Another lesson may be that there is a time for haste and a time for caution: before the battle Varro made the wrong choice, after the battle Hannibal. However, given the results of Varro’s earlier over-zealous action, there is some irony in his criticism of Hannibal’s tardiness. 5.4 Roman and Carthaginians The book ends by contrasting Roman and Carthaginian behaviour. Carthaginians were often portrayed by Romans as treacherous and perfidious, and Hannibal was the exemplar of Carthaginian values par excellence. His treachery has been a recurrent theme of the narrative since it was introduced in Livy’s opening character sketch of him: Has tantas uiri uirtutes ingentia uitia aequabant, inhumana crudelitas, perfidia plus quam Punica, nihil ueri, nihil sancti, nullus deum metus, nullum ius iurandum, nulla religio 21.4.9

Colossal vices equalled the man’s very great virtues: an inhumane cruelty, treachery that was more than Carthaginian, the holding of nothing truthful, nothing sacred, no fear of the gods, no respect for an oath, no religious scruple. In the Cannae campaign we meet it first in the ambush by which he attempts to catch the Romans (41.6–43.1), at the end of which we are told by the narrator

livy on cannae: a literary overview

181

that Hannibal … nequiquam detecta fraude in castra rediit ‘Hannibal returned into his camp with his deceitful plan made vain by detection’. The ablative absolute is nicely ambiguous, since it may be focalized either as a narratorial comment or through Hannibal as reflecting his own view of his recent activity. At 43.6 he plays another trick, setting up what appeared to be an identical ambush but in fact moving his army to Apulia. At 58.5 he offers to ransom captured Roman cavalry for a higher sum than had been agreed when they had surrendered themselves. In the battle itself the Numidian cavalry offer an egregious example of what the narrative voice (48.1) calls Punica fraus when they take up arms after surrendering themselves and attack the Romans in the rear. By contrast there is no treachery on the side of the Roman state. In Livy’s main narrative Roman captives pledge that, if they go as envoys to Rome, they will return. Then one of their number returns almost immediately after departure, pretending to have forgotten something, before setting out again; after the embassy, he goes home rather than back to Hannibal, claiming to have kept his vow. He is termed (58.8) minime Romani ingenii homo ‘a man hardly with a Roman nature’ and sent back bound to Hannibal. An alternative version reported at 61.5–61.10 tells the tale differently. Ten ambassadors claim that they are entitled to stay in Rome, having fulfilled their obligation to Hannibal. They are shamed by a censorial nota and either commit suicide or languish removing themselves from the public gaze. This second version shows the Roman response to be equally disapproving but perhaps more respectful of the letter of the law. In their differing approaches to fraus, we have only an implicit comparison between the two nations. Far more explicit is 54.10–54.11, quoted already, in which the Roman and Carthaginian attitudes to defeat are contrasted. And it is this theme with which Livy ends the book. The defection of many of Rome’s allies has just been announced, adding woe upon woe, when Varro returns: quo in tempore ipso adeo magno animo ciuitas fuit ut consuli ex tanta clade, cuius ipse causa maxima fuisset, redeunti et obuiam itum frequenter ab omnibus ordinibus sit et gratiae actae quod de re publica non desperasset; (15) qui si Carthaginiensium ductor fuisset, nihil recusandum supplicii foret. 61.14–61.15

At this very time the state was so great hearted [but suggesting also ‘was endowed with such courage’] that all ranks went out to meet the consul returning from a disaster which was of such magnitude and for which he had been chiefly responsible, and thanks were offered that he had not

182

oakley

despaired of the state. (15) But if he had been a Carthaginian commander, no punishment could have been denied him. Silius (10.605–10.639) and Plutarch (Fab. 18.4–18.5) make Fabius Maximus welcome Varro, and doubtless this version was found in Livy’s annalistic sources; perhaps Livy eschews it because he wishes to focus attention on the collective Roman animus55 and the unity of the state at this critical juncture, a unity that was an important precondition of ultimate victory in the war.56 The book, however, is not the end of the war, and twice the narrative looks forward to the eventual Roman victory. The statue of Victoria offered by Hiero and dedicated in the temple of Jupiter Optimus Maximus (37.5, 37.12) is an omen of Rome’s eventual victory. And when Publius Cornelius Scipio, later victor over Hannibal at Zama, is said to have stopped Metellus from fleeing overseas and abandoning Rome, the narrative voice reminds us that he is the fatalis dux huiusce belli (53.6). Therefore, even in the hour of Rome’s greatest defeat, Livy reminds us of her eventual victory; and though the narrative offers no prompting, it would be a very ignorant student of Cannae who does not reflect on the fact that Publius Cornelius Scipio Aemilianus, the grandson of the Paullus who died at Cannae, will sack Carthage in 146BC.

Appendix

Polybius

Livy

3.106.1–3.106.2. Election of Varro and Paullus to the consulship. Of the two, Paullus is singled out for appointing the consuls of the previous year proconsuls.

22.33.9–22.34.2. When the consuls say that they cannot return to carry out the elections for 217, the senate ask for a dictator to be appointed for the task. When he is said to have been appointed with a flaw in the auspices, an interregnum ensures. 34.2–34.7. The elections are hotly contested. Many plebeians support the candidature of a new man, C. Terentius Varro, and the tribune of the plebs, Q. Baebius Herennius, makes a strong speech on his behalf, in which he criticizes the nobles for starting and prolonging the war. After Varro is elected, the nobility urge M. Aemilius Paullus, an opponent of

55 56

See Hoffmann 1942: 45. The scene contrasts strikingly with the political divisions that attended the departure of the consuls at 40.4. See further Mineo 1997: 126–127 ≈ 2006: 269.

183

livy on cannae: a literary overview (cont.) Polybius

106.3–106.5. Paullus oversees the recruiting of new troops. Cn. Servilius Geminus, one of the proconsuls, is urged not to engage in full-scale conflict but to train the troops by skirmishing and guerrilla warfare.

106.6–106.8. Troop dispositions are made for the war in Cisalpine Gaul, Sicily, and Spain.

106.9–106.11. Servilius carries out his instructions, and the Romans campaign successfully.

Livy the plebs, in part because of their criticism of his earlier consulship, to stand. He is elected. Praetors too are elected. 36.1–36.5. Large increase in number of Roman forces

36.6.7–36.6.8. Procuration of prodigies. 36.9. Legates from Paestum are thanked for their offer to help out Roman finances with gold, but the offer is refused. 37.1–37.13. An embassy comes from Hiero of Sicily. The statue of Victory that he sends is placed on the Capitol as an omen. Hiero’s envoy requests Roman attacks on Africa from Sicily. [Livy defers reference to the Roman forces in Cisalpine Gaul until book 23] [At 32.1–32.3 Livy has referred to the guerilla warfare carried out by Atilius Regulus and Servilius Geminus, but as consuls rather than proconsuls.] 38.1–38.5. A new form of military oath is sworn by Roman troops and administered by themselves, that they will not flee. 38.6–38.15. Varro and Paullus hold contiones before their departure. Varro criticizes Fabian tactics and promises to fight and win at once. Paullus says that it is foolish to decide on tactics before arrival at the scene of war. 39.1–39.22. Speech of Fabius Maximus, warning Paullus of the dangers posed by Varro and encouraging him to persist in Fabian tactics. 40.1–40.3. Paullus’ diffident response: in effect saying that he would do his best but was uncertain that he would be successful. 40.4–40.6. Varro and Paullus join the Roman army opposed to Hannibal. It is divided into two camps, with Servilius Geminus put in charge of the smaller. 40.7–40.9. Hannibal is running out of supplies and so pleased by the arrival of the consuls. 41.1–42.12. The Romans win a minor success in a skirmish, but, to Varro’s dismay, Paullus stops them from turning this into a full-scale battle. Hannibal knows about the discord in the Roman camp, and plans an ambush, by withdrawing his forces from

184

oakley

(cont.) Polybius

107.1–107.5. Hannibal moves from Gereonium, his base since 217, into Apulia and captures the village of Cannae, where the Romans had piled up a depot of grain. This put the Romans at a disadvantage: they had lost supplies, and from Cannae Hannibal commanded the surrounding territory. 107.6–108.1. Finding that the changed circumstances did not allow him to carry on with guerrilla tactics because of the difficulty of supplying his troops, Servilius writes to Rome asking for advice. The Romans decide to fight a decisive battle against Hannibal. Servilius is asked to hold himself in check and to await the arrival of the consuls. Of these two, the Romans pinned most hope in Aemilius Paullus because of the prowess of his life. They decide not to fight with just one consul and his standard complement of two legions and allies but to send both consuls, each with four legions and a correspondingly increased complement of allies. On his departure Paullus is urged to fight the decisive battle nobly. 108.2–109.13. Day 1. Paullus (implicitly) in command. On joining the Roman forces in Apulia, Paullus addresses them in a speech, in which he lays the blame for previous defeats on the use of inexperienced troops, on their arrival at the Trebia just before the battle, and their not even being able to see the enemy at Trasimene. Now experienced and trained, knowing their enemy well, commanded by two consuls, and knowing how their families would be affected by the result, they should be confident of victory. 110.1. Day 2. Varro (implicitly) in command. The Romans march closer to the enemy. 110.1–110.3. Day 3. Paullus (implicitly) in command. As they advanced closer to the enemy, Paullus did not want to fight in terrain that was flat and devoid of undergrowth because it would favour the Carthaginians, whose strength lay in their cavalry. Varro took a different view, because of

Livy his camp and making it look as though he had fled. Varro and the Roman troops wish to plunder the camp and pursue. Paullus is able to stop them only because the auspices are unfavourable. 43.1–43.11. Now very short of supplies, and with his troops threatening mutiny, Hannibal retreats. The Romans initially suspect an ambush but soon realize that the retreat is real and follow Hannibal to Cannae in Apulia.

44.1–44.7 Hannibal wants to fight in terrain suiting his cavalry and provokes the Romans: Varro wants to fight, Paullus to resist. [45.1–45.4 happens on the same day]

185

livy on cannae: a literary overview (cont.) Polybius his inexperience—and there was an argument between the commanders, (§ 3) ὃ πάντων ἐστὶ σφαλερώτατον. 110.4–110.7. Day 4. Varro explicitly in command. Varro leads out the Roman forces. A skirmish ensued in which the Romans did better. 110.8–110.11. Day 5. Paullus (implicitly) in command. Judging it safe neither to fight nor to retreat, Paullus made two camps, the larger, with two thirds of the Roman forces, next to the Aufidus, the smaller on the other side of the river ten stades away. His plan was that the men in the smaller camp would protect those who had gone out of the other camp to forage and would trouble the Carthaginians. 111.1–111.11. Hannibal decides to fight and makes a speech to his men, reminding them that victory would make them lords of Italy. 112.1. Day 6. Varro (implicitly) in command. A day of inactivity. 112.1–112.5. Day 7. Paullus (implicitly) in command. On the next day, Hannibal leads out his forces, but, seeing that lack of supplies would soon force Hannibal to move camp, Paullus does not respond. Hannibal withdraws most of his troops but sends Numidian cavalry to harass those foraging from the smaller Roman camp. Varro and the mass of the Roman forces were greatly disturbed and did not like the delay. 112.6–112.9. In the meantime, news that the decisive battle was soon to be fought had made the whole of Rome apprehensive, with many prodigies being reported. 113.1–113.6. Day 8. The day of the battle. Varro in command. Varro leads out his forces to fight. Those from the larger camp cross the river and are joined by those from the smaller camp. On the right wing he places the Roman cavalry, the Roman infantry next to them, in a thicker formation than usual, and the allied cavalry on the left. Light-armed troops were placed in the front. The force comprised 60,000 foot and 6,000 cavalry. 113.6–113.9. Hannibal sends his slingers and lightarmed troops across the river, and then the rest of his army. On his left wing he placed his Spanish and

Livy

45.1–45.4. Later in the day Hannibal draws up his line of battle and sends Numidians to attack Roman water-collectors with his Numidians. The Romans want to fight, but Paullus, in command on that day, resists.

45.5–45.8. On the next day, Varro, now in command again, draws up the Roman line of battle. Those from the larger camp cross the river and are joined by those from the smaller camp. On the right wing he places the Roman cavalry, the Roman infantry next to them, then the allied infantry, and the allied cavalry on the far left. Paullus commanded the Roman right, Varro the left, the proconsul Servilius the centre. 46.1–46.3. Hannibal crosses the river and draws up his line of battle. On his left wing he placed his Spanish and Gallic cavalry, opposite the Roman

186

oakley

(cont.) Polybius

Livy

Gallic cavalry, opposite the Roman cavalry, next to them half of his heavily-armed African troops, next to them his Spanish and Gallic infantry, next to them the remainder of his African infantry, and then on his right wing his Numidian cavalry. He then made the detachments of the Iberians and Gauls in the middle advance slightly, so that his formation was moon-shaped, with the African infantry in reserve. 114.1 The Africans were armed with Roman spoils. 114.2–114.4. The Gauls and Spaniards had similar shields but different swords: the Gauls long and the Spanish short. They had a frightening appearance, the Spaniards because of their purple tunics, the Gauls because of their naked bodies. 114.5. Hannibal’s troops comprised 40,000 infantry and about 10,000 cavalry. 114.6. Paullus commanded the Roman right, Varro the left, the two proconsuls Atilius and Servilius the centre. 114.6–114.7. Hasdrubal commanded the Carthaginian left, Hanno the right, and Hannibal with his brother Mago the centre. 114.8. The Roman army faced south, the Carthaginian north.

cavalry, on his right wing Numidian cavalry. Gallic and Spanish infantry were in the centre, flanked by on each side by African troops.

115.1. The light-armed troops begin the battle. 115.1–115.4. The cavalry on the Carthaginian left/Roman right follow. The constricted space because of the river means that they end up fighting like infantry, ferociously. The Roman cavalry is driven back. 115.4–115.12. The infantry battle begins, and the Romans drive back the Gauls and Spaniards in the centre of the Punic line. (Polybius refers again to its being moon-shaped). As the Gauls and Spaniards retreat, the Romans follow and the Africans fight them on their flanks and soon surround them. The Romans now have to fight a second battle but because they are surrounded they cannot fight in their normal formation. 116.1–116.4. Paullus, who had been stationed on the right, sees that the middle of the battlefield will

46.4. The Africans were armed with Roman spoils. 46.5–46.6. The Gauls and Spaniards had similar shields but different swords: the Gauls long and the Spanish short. They had a frightening appearance, the Spaniards because of their purple tunics, the Gauls because of their naked bodies.

47.7–47.8. Hasdrubal commanded the Carthaginian left, Maharbal the right, and Hannibal with his brother Mago the centre. 46.8. The Roman army faced south, the Carthaginian north. 46.9. The wind called the Volturnus blew in the faces of the Romans. 47.1–47.3. The light-armed troops begin the battle. 47.2–47.3. The cavalry on the Carthaginian left/Roman right follow. The constricted space because of the river means that they end up fighting like infantry, ferociously. The Roman cavalry is driven back. 47.4–47.10. The infantry battle begins, and the Romans drive back the Gauls and Spaniards in the centre of the Punic line. (Livy mentions that it was shaped like a moon.) As the Gauls and Spaniards retreat, the Romans follow and the Africans fight them on their flanks and soon surround them. The Romans now have to fight a second battle at a considerable disadvantage, because they are surrounded and because the enemy are fresh.

187

livy on cannae: a literary overview (cont.) Polybius prove decisive and moves there to encourage his troops. Hannibal, who had been in the centre all the time, does likewise. 116.5. The Numidian cavalry on the Carthaginian right engage the allied cavalry on the Roman left. They cause little damage but keep the Roman cavalry out of the battle. 116.6–116.8. By now the cavalry on the Carthaginian left, commanded by Hasdrubal, which had been pursuing the defeated Roman cavalry, returns victorious and joins forces with the cavalry on the Carthaginian right. The allied cavalry flee. Hasdrubal then lets the Numidians pursue the fleeing cavalry of Rome’s allies, a task for which they were well equipped both in number and style of fighting, whilst he leads his own cavalry to help the African infantry. He then falls on the rear of the Roman infantry, causing dismay. 116.9. Paullus dies bravely in the thick of the fighting.

116.10–116.11. The Romans continue to resist until they were completely encircled. At which point they were killed, the consuls of the previous year being among the casualties. 116.12–116.13. In the meantime, the Numidian cavalry chased and killed the fleeing Roman cavalry. 117.12. A few Romans flee to Venusia, including Varro, whose flight and whole holding of office were shameful.

117.1–117.3. A summative comment leads into reflection on the scale of the battle and a description of the casualties on the Roman side. 117.4–117.5. Polybius reflects on the advantages of Hannibal’s being superior in cavalry. 117.6. The few Carthaginian casualties are described.

Livy

48.1–48.6. The allied cavalry on the Roman left is by now engaged, and the Carthaginians engage in Punica fraus. Around 500 Numidians hiding swords under their normal attire pretend to desert to the Romans and are asked to wait behind the Roman line with their shields and javelins on the ground. When everyone was occupied with the battle raging around about, they seize scuta and attack the Roman line in the rear. In some places they cause panic; elsewhere they are resisted stoutly. Hasdrubal sends them to chase those fleeing and joins the Gauls and Spaniards to the Carthaginians, who are tired more by slaughter than by fighting.

49.1–49.12. In the other part, Paullus is wounded at the beginning of the battle. When he can no longer ride a horse, those around him dismount. The battle is hopeless and eventually the Romans are driven back. Cn. Lentulus offers Paullus the chance to flee, but he prefers to die rather than face trial by the people again; he tells Lentulus to tell Fabius that he carried out his orders. 49.13. This leads to flight everywhere.

49.1–49.14. 7,000 Romans flee into the smaller camp, 10,000 into the larger, some to the village of Cannae, Varro and some cavalry to Venusia. 49.15–49.18. The massive Roman casualties are listed. 50.1–50.3. A summative comment introduces a comparison of Cannae to the battle of the Allia.

188

oakley

(cont.) Polybius 117.5–117.11. Paullus had left 10,000 troops in the larger camp to attack Hannibal’s camp during the battle. This they did, but when Hannibal was victorious elsewhere he drove them back to their own camp, from which they surrendered. 117.12. The Numidian cavalry kill many Roman fugitives. 118.1–118.9 Polybius describes (i) the defection of Rome’s allies in southern Italy, (ii) the Romans’ expectation that Hannibal would march on their city, (iii) the defeat of a Roman army in Gaul, and (iv) the thoroughness of the senate in making dispositions. He then comments that despite her loss of military reputation Rome would yet recover her power in Italy, largely thanks to the excellence of her constitution. This leads into a discussion of first what was happening in Greece and the eastern Mediterranean at this time and then of the Roman and other constitutions.

Livy

50.4–50.12. A message comes from the larger camp urging those in smaller camp to join them and then to escape to Canusium. Many resist, but P. Sempronius Tuditanus makes a rousing speech and leads 600 men to escape. 51.1–51.4. Maharbal vainly urges Hannibal to march on Rome. 51.5–51.9. The dead and almost dead Romans are gruesome sight; notable is the ferocity of one who bit his opponent as he died. 52.1–52.6. Swift surrender of smaller Roman camp. 4,000 men escape from large camp to Canusium; surrender of large camp. Hannibal buries his dead. 52.7–53.13. Busa looks after the Romans at Canusium; when L. Caecilius Metellus encourages despair, Scipio makes the Romans swear to fight on. 54.1–54.6. Varro comes from Venusia to join forces with those at Canusium. 54.7–54.11. Livy does not describe the effect of the disaster at Rome; he compares the effect of the disaster on Rome with that of disasters on Carthage. 55.1–56.1. The senate meets to calm the city; Fabius offers advice. 56.2–57.12. Troop dispositions and procuration of religious matters at Rome. 58.1–58.9. Hannibal agrees to ransom his Roman prisoners; an embassy is sent to the senate; the ambassadors pledge to return; one of them, minime Romani ingenii homo, returns at once and then sets out again. 59.1–59.19. Speech of legates asking for freedom 60.1–60.4. The senate wavers. 60.5–60.27. Speech of Manlius asking for no compromise. 61.1–61.4. The ambassadors return. The one who had imagined himself free is handed over to Hannibal.

189

livy on cannae: a literary overview (cont.) Polybius

Livy 61.5–61.10. A variant version of the embassy and of the story of the legates who refused to hand themselves over to Hannibal. 61.10–61.15. Rome’s allies revolt but Roman spirit remains undaunted.

Bibliography Badian, E., ‘The early historians’, in T.A. Dorey (ed.), Latin historians (London 1966) 1– 38. Beltramini, L., ‘Narrazione ed exemplum in Livio’, Eikasmos 28 (2016) 171–194. Bernard, J.-E., Le portrait chez Tite-Live: essai sur une écriture de l’histoire romaine (Brussels 2000). Breska, A. von, Quellen Untersuchungen im 21. bis 23. Buche des Livius (Programm der Luisenstädtischen Oberrealschule zu Berlin) (Berlin 1889). Briscoe, J., A commentary on Livy books xxxi–xxxiii (Oxford 1973). Bruckmann, H., Die römischen Niederlagen im Geschichtswerk des T. Livius, dissertation (Münster 1936). Burck, E., Einführung in die dritte Dekade des Livius (Heidelberg 1950). Chaplin, J.D., Livy’s exemplary history (Oxford 2000). Clark, J.H., Triumph in defeat. Military loss and the Roman Republic (New York 2014). Cornelius, F., Cannae: das militarische und das literarische Problem (Leipzig 1932). Cornell, T.J. (ed.), The fragments of the Roman historians (Oxford 2013). De Sanctis, G., Storia dei Romani III. II (Turin 1917). Erdkamp, P., ‘Polybius, Livy and the Fabian strategy’, AncSoc 23 (1992) 127–147. Fronda, M.P., Between Rome and Carthage: southern Italy during the Second Punic War (Cambridge 2010). Fronda, M.P., ‘Hannibal: tactics, strategy and geostrategy’, in D. Hoyos (ed.), A companion to the Punic Wars (New Malden 2011) 242–259. Hardie, P.R., Rumour and renown: representations of fama in western literature (Cambridge 2012). Hesselbarth, H., Historisch-kritische Untersuchungen zur dritten Dekade des Livius (Halle 1889). Hoffmann, W., Livius und der zweite punische Krieg (Berlin 1942). Jaeger, M., Livy’s written Rome (Ann Arbor 1997).

190

oakley

Jaeger, M., ‘Urban landscape, monuments, and the building of memory in Livy’ in B. Mineo (ed.), A companion to Livy (New Malden 2015) 65–77. Jumeau, R., ‘Un aspect significatif de l’exposé livien dans les livres XXI et XXII’, in M. Renard & R. Schilling (eds.), Hommages à Jean Bayet, Collection Latomus lxx (Brussels 1964) 309–333. Kahrstedt, U., Geschichte der Karthager III (Berlin 1913). Klotz, A., Livius und seiner Vorgänger (Leipzig 1940–1941). Klotz, A., ‘Dichtung und Wahrheit in der livianischen Erzählung der Schlacht bei Cannae’, Gymnasium 56 (1944) 58–70. Lachmann, F., De fontibus historiarum T. Livii (Göttingen 1822–1828). Levene, D.S., Livy on the Hannibalic War (Oxford 2010). Luterbacher, F., De fontibus librorum xxi et xxii Titi Livi, dissertation (Strasburg 1875). Mineo, B., ‘L’interprétation livienne de l’histoire: le récit des défaites romaines de la Trébie à Cannes’, REL 75 (1997) 113–128. Mineo, B., Tite-Live et l’histoire de Rome (Paris 2006). Münzer, F., ‘Fabius 116’, in [eds.] Paulys Real-Encyclopädie der classischen Altertumswissenschaft VI.2 (Stuttgart 1912) 1814–1830. Münzer, F., ‘Terentius 83’, in [eds.] Paulys Real-Encyclopädie der classischen Altertumswissenschaft (Stuttgart 1934), VA.1: 680–690. Oakley, S.P., A commentary on Livy, books vi–x (Oxford 1997–2005). Pelling, C.B.R., ‘Plutarch: Roman heroes and Greek culture’, in M. Griffin & J. Barnes (eds.), Philosophia togata. Essays on philosophy and Roman society (Oxford 1989) 199–232. Peter, C., Livius und Polybius. Ueber die Quellen des xxi und xxii. Buchs des Livius (Halle 1860). Skutsch, O., The annals of Quintus Ennius (Oxford 1985). Soltau, W., Die Quellen des Livius im 21. und 22. Buch (Programm Zabern 1894a). Soltau, W., Livius’ Quellen in der III. Dekade (Berlin 1894b). Walbank, F.W., A historical commentary on Polybius: volume 1, commentary on books I–VI (Oxford 1957). Walsh, P.G., Livy: his historical aims and methods (Cambridge 1961). Will, W., ‘Imperatores victi. Zum Bild besiegter römischen Consuln bei Livius’, Historia 32.2 (1983) 173–182.

chapter 8

Discourse-Linguistic Strategies in Livy’s Account of the Battle at Cannae Lidewij van Gils and Caroline Kroon

1

Introduction*

A familiar observation about ancient historiographical texts is that they generally display ‘emplotment’: the assembly of a series of events into a narrative with a plot.1 The skillful ways in which the Roman historian Livy fashions his source material into a narrative have been much discussed by modern scholarship, and from various angles and perspectives.2 Most recently, in a stimulating monograph on Livy’s books on the Hannibalic war (AUC books 21–30), David Levene has shown how Livy, by means of his personal selection and organization of the source material, is able to communicate particular interpretations of the war, demonstrating once more that coherent narratives may be rhetorically highly compelling.3 In the present volume, Oakley also draws attention to the rhetorical force of Livy’s historiographical narrative, for instance when he observes that ‘Livy’s whole narrative in the Cannae episode of book 22 is shaped so as to throw as much blame as possible on Varro’.4 Another example of rhetorically driven * This chapter was written as part of the NWO-project Ancient War Narrative. A Combined Discourse-Linguistic and Narratological Approach (NWO-project 360-30-190). We wish to thank Suzanne Adema, Irene de Jong and Rodie Risselada for critically reviewing the draft version of the chapter, and all members of the Amsterdam ‘Cannae Lab’ for the stimulating discussions we had while reading the text of Livy Book 22. 1 The term emplotment was coined by Hayden White in his 1973 work Metahistory to describe the way in which modern historians necessarily fashion their source material into narrative. See Cuddon 2013, s.v. ‘emplotment’. 2 See e.g. Burck [1934] 1964 and [1950] 1962; Bruckman 1936; Walsh [1961] 1989; Luce 1977; Fuhrmann 1983; Feldherr 1998; Chaplin 2000; Krafft 2007; Pausch 2011. 3 Levene 2010. For the same view on the rhetorical force of narrative from a linguistic-anthropological perspective, see e.g. Ochs 2005: 278. 4 Oakley, this volume, p. 162 For the rhetoric of classical historiography see e.g. Woodman 1988; Marincola 1997; Laird 2009. Cf. also Conte [1987] 1994: 372, who places Livy’s history in a diachronic development in which ancient historiography, ‘instead of being ‘investigation’ of the truth, could become a rhetorical activity’.

© koninklijke brill nv, leiden, 2019 | doi:10.1163/9789004383340_009

192

van gils and kroon

organization of the narrative is discussed in the contribution by Pausch. In the context of a discussion on Livy’s techniques for enhancing the engagement of the audience, Pausch points out the way in which the historian strategically places the list of prodigia of the year 218 BC at the end of book 21, and thus in close proximity to the list of prodigia of the year 217 at the beginning of book 22, ‘presumably in order to double their frightening impact’.5 In this chapter we would like to take Levene’s general observations and conclusions as a starting point for a detailed discussion of the narrative organization in Livy’s book 22, paying special attention to the episode of the battle at Cannae (22.34.1–22.61). In this discussion we will focus on the various strategies that the historian employs in order to create a coherent and rhetorically compelling narrative, which not only guarantees the audience’s sustained attention, but also makes the audience receptive to the particular ideological and moral stances from which the narrative is presented. The views we present here are based on the results of an in-depth discourselinguistic analysis of the complete text of book 22, and strongly corroborate observations made by Levene 2010 and Oakley (this volume) on prominent themes in Livius’ account of this traumatic episode in Rome’s war against Hannibal.6 Both Levene and Oakley draw attention to the importance of the theme of discordia ordinum (the struggle of the orders and the disunity of the state), which in books 21 and 22 is strongly associated with the theme of contrast between the good and bad Roman commanders.7 This theme is conveyed by a particular lexicon in which reason (ratio) and prudence (prudentia) are opposed to rashness (temeritas) and reliance on fortune ( fortuna).8 Another theme that appears to permeate book 22, also mentioned by Levene and Oakley, is the theme of deception ( fraus, especially punica fraus) as opposed to faith ( fides).9 In our chapter we will show, among other things, how the particular organization of the narrative of book 22 supports these themes. As such it may contribute to the view that Livy is preoccupied, in this part of the decade,

5 Pausch, this volume 241. For a comparable observation see Levene 2010: 37. 6 For an introduction to the relevant concepts used in the analysis, we refer to the Introduction to this volume. 7 The rash commanders in books 21–22, who lead their armies to defeat against Hannibal, are Sempronius (Allia), Flaminius (Trasimene), Minucius (hill near Gereonium), and Varro (Cannae). The good commanders are, especially, Fabius Maximus and Aemilius Paullus. 8 See Oakley in this volume 166 for an overview of the Latin words in book 22 that are connected to this theme: temeritas and cognates; consultus—inconsultus; cauere, cautus and their cognates; prouidus, improuidus and their cognates; ratio; ferox. See Buijs in this volume for Livy’s treatment of this theme by means of speech representation. 9 This theme is e.g. present at 41.6–43.1; 43.6; 48.1; 58.8. See Oakley in this volume, p. 166.

discourse-linguistic strategies in livy’s account of the battle

193

with explaining away the Roman defeat at Cannae by highlighting Roman discordia and Punic fraus as excuses.10 In the first part of the chapter we will introduce and explain the narrative strategies and devices that appear to be dominant in Livy’s book 22 (§ 2), and give an illustration by means of an analysis of the organizationally quite straightforward mini-episode of Abelux, the clever Spaniard from Saguntum who managed to hand over to the Romans the Spanish hostages that were held captive by the Carthaginians in Saguntum (22.22.6–22.22.21) (§ 3). This analysis occasions a digression on the strategic use of the historic present in § 4. In the second part of our contribution we will provide, along the same lines, an elaborate analysis of the much more complex episode of the battle at Cannae (§5). This will lead, in §6, to the conclusion that the narrative of Livy’s book 22 is a carefully wrought unity, in which a heterogeneous set of textual strategies and linguistic devices is put to use in such a way as to keep the audience optimally interested and amenable to the historian’s particular presentation, interpretation and evaluation of the events.

2

Livy’s Interweaving of ‘Annalistic Material’ and the Main Story Line

In his aforementioned study on the Hannibalic war, Levene makes a convincing case for considering the third decade of Livy’s Ab Urbe Condita a single story.11 The very beginning of book 21, which seems to function as a preface to books 21–30 as a whole, already points strongly in this direction:12 (1) In parte operis mei licet mihi praefari, quod in principio summae totius professi plerique sunt rerum scriptores, bellum maxime omnium memorabile, quae unquam gesta sint, me scripturum, quod Hannibale duce Carthaginienses cum populo Romano gessere. […] et adeo uaria fortuna belli ancepsque Mars fuit, ut propius periculum fuerint qui uicerunt. LIV. 21.1.1–21.1.2

10 11

12

See Levene 2010: 262–263, referring to Bruckmann 1936. Also Pausch (this volume, 243). Levene 2010: 15–17 extensively defends his position building on earlier studies, notably, but not solely by Burck [1950] 1962; Walsh 1973; and Luce 1977 on the episodic structure of Livy’s historiography. Here as in the following the Latin text is that of Dorey’s Teubneriana (1971), the translation that of the Loeb edition (Foster 1929), sometimes slightly adapted.

194

van gils and kroon

In this preface to a part of my history I may properly assert what many an historian has declared at the outset of his entire work, to wit, that the war which I am going to describe was the most memorable of all wars ever waged—the war, that is, which, under the leadership of Hannibal, the Carthaginians waged with the Roman People. […] And so variable were the fortunes of the war and so uncertain was its outcome that those who ultimately conquered had been nearer ruin. By means of these specific opening words the historian already creates an important expectation on the part of his audience which will remain active and dominant throughout the third decade: the audience may expect an extended continuing story line on a single war, figuring Hannibal in the role of main antagonist and pivot around which the story will evolve. Levene observes, and rightly we think, a general narrative ‘arc’ in the 3rd decade.13 This arc starts with the rise of Hannibal and his invasion of Italy in books 21 and 22, where Hannibal’s victories become increasingly devastating. In the next three books the Romans begin to recover, but are still faced with a number of significant defeats. At the end of book 25, exactly in the middle of the Hannibal books, a balance in defeats and victories is achieved,14 after which, from book 26 onwards, the Romans definitely seem to have recovered, and will gain a series of victories in Italy, Spain and Africa. Book 30, in reversing the events of book 21, could be seen as a natural closure of the entire decade. This appears to be quite a straightforward narrative organization, providing a global ‘arc of narrative tension’ on the highest level of narrative analysis. However, keeping track of this main story line feels, in Levene’s words, ‘bewilderingly difficult’, as ‘we move from scene to scene, from theatre of war to theatre of war, sometimes picking up threads from a book or more earlier’.15 When we move our focus from the decade as a whole to the internal structure and coherence of the single book, more particularly book 22,16 the narrative organization appears not to be less complex and bewildering. Although the impression of a continuing main story line dealing with the war against Hannibal in Italy is maintained, it is difficult to speak of one single story which is presented as one single narrative arc of tension. What we see is rather a diversity of related or (seemingly) unrelated minor story lines, which do or

13 14 15 16

Levene 2010: 15–17. Recapture of Syracuse, but defeat and death of the Scipio brothers in Spain. Levene 2010: 63. Levene leaves the internal structure of book 22 largely out of account.

discourse-linguistic strategies in livy’s account of the battle

195

do not have themselves the form of fully elaborated narrative arcs of tension, and which alternate with, or are intertwined with, elements that are generally described as ‘annalistic material’. This ‘annalistic material’ refers to formal elements of the Roman political year, such as the election of new magistrates and their taking up of office, the reception of embassies, reports of deaths of important persons, and prodigies. These elements tend to get a fixed position in the organization of the single book (e.g. the election of new consuls at the end of a book, and their taking up office at the beginning of a new book), as was probably customary for a prior generation of historians.17 The narrativity of annalistic events is usually low.18 The following—abridged—passage from the beginning of book 22 serves as an illustration of how Livy adroitly manages to interweave ‘annalistic material’ of the years 217 and 216 BC and elements of the main story line about the war between Hannibal and the Romans. Our main focus in this chapter will be on the second half of book 22, but the illustrations in § 2-§ 4 below are taken from the first half, mostly in chronological order, in order to prepare both methodologically and by way of content our analysis of the battle of Cannae (22.34– 22.61). Iam uer adpetebat; itaque Hannibal ex hibernis mouit […] 22.1.4 per idem tempus Cn. Seruilius consul Romae idibus Martiis magistratum iniit. […] 22.1.19 postremo Decembri iam mense ad aedem Saturni Romae immolatum est […] 22.2.1 dum consul placandis Romae dis habendoque dilectu dat operam, Hannibal profectus ex hibernis, quia iam Flaminium consulem Arretium peruenisse fama erat, cum aliud longius, ceterum commodius ostenderetur iter, propiorem uiam per paludes petit

(2) 22.1.1

LIV. 22.1.1–22.2.1

22.1.1

17 18

Spring was now drawing on, and accordingly Hannibal moved out of his winter encampment […]

Levene 2010: 34–45 convincingly illustrates how in the Hannibal books the annalistic chronology is often subordinated to a thematically governed book structure. Narrativity is a concept used to distinguish narrative text types from, for instance, argumentative or reportive ones. See Herman 2007.

196

van gils and kroon

22.1.4

About the same time, on the Ides of March, Gnaeus Servilius began his consulship in Rome […] 22.1.19 Finally—the month was now December—victims were slain at the temple of Saturn in Rome […] 22.2.1 While the consul was occupied in Rome appeasing the gods and levying troops, Hannibal, who had left his winter quarters, heard that Flaminius, the other consul, had already arrived at Arretium; and so, though another route, longer to be sure but less difficult, was pointed out to him, he took the shorter route through the marshes The first sentence of book 22 contains clear linguistic indications that the main story line is being resumed here, the combination of iam and the imperfect tense (appetebat) being a common linguistic strategy for signalling a ‘fade in’, and for building up narrative tension towards the occurrence of a new action or episode in the main story line.19 The first action of this new episode follows in the next sentence: as expected now that spring has arrived, Hannibal moves out of his winter encampment (Hannibal ex hibernis mouit) from northern Italy towards the central regions where the Roman army was encamped (not cited). Before we are told, however, what happened next, the main story line gives way to a passage which conveys prototypical ‘annalistic material’: Consul Gnaeus Servilius taking up office in Rome on the Ides of March, and the occurrence and expiation of a large number of prodigies in the ensuing months, ending, in the month of December, with a public offering and feast at the temple of Saturn (22.1.4–22.1.19). Next, in 22.2.1, we are taken smoothly back in time and switch from the events of the political and religious year in Rome to the main story line of Hannibal’s offensive in Italy, the transition being facilitated by a dum-clause containing a few more items from the ‘annalistic’ list of events,20 as well as by the recapitulative phrase profectus ex hibernis, repeating

19

20

The cinematographic term ‘fade in’ is used for the phenomenon of turning the camera to an event or situation that has already started while the camera was still at a different location. Chausserie-Laprée 1969: 497–517 speaks of ‘iam d’ouverture et de préparation’. See for this narrative technique also Kroon & Risselada 2002, 2004. The transitional, coherence-creating function of the clause is also enhanced by the fact that the dum-clause has a personal subject (consul), whereas the activities referred to in the immediately preceding text lack a specific, individualized subject. The dum-clause

discourse-linguistic strategies in livy’s account of the battle

197

ex hibernis mouit (22.1.1) and bringing us back to the theatre of war in mid-Italy. As we will explain later, the present tense form petit in 22.2.1. (the first present tense after a long series of perfects) has, as is often found in Livy, a prospective, open-ended force and provides the signal that here we are dealing with a consequential action that will propel the first of a number of smaller, local narrative arcs of tension in book 22: the dramatic episode of Hannibal’s march through the marshes that will be recounted in the remainder of paragraph 2. In terms of discourse mode, this first chapter of book 22 might be considered a sort of hybrid.21 On the one hand the passage displays the characteristics of a continuing narrative in that it contains, also in the ‘annalistic digression’, various characteristic features of causal-chronological storytelling, for instance the use of temporal coherence markers like per idem tempus, inde, haec ubi facta and postremo. Moreover, there appear to be a few attempts to tell the events and facts from a story-internal perspective, as is for instance the case in the sentence (not quoted in 2 above) augebant metum prodigia ex pluribus simul locis nuntiata (‘men’s fears were augmented by the prodigies reported simultaneously from many places’, 22.1.8).22 At the same time, however, the passage altogether lacks an arc of tension, and the temporal markers used are deliberately vague and merely serve to integrate the annalistic material loosely into the frame of the main story line, hence giving the passage the appearance of a coherent and continuous narrative.23 As readers we hardly have the impression that we are really drawn into a story-world, in the sense of being mentally transferred to an alternative world parallel to the hic and nunc of the historian and his audience. The suggestion of temporal continuity, moreover, seems to be broken in the sentence postremo Decembri iam mense ad aedem Saturni Romae immolatum est (‘finally—the month was now December—victims were slain at the temple of Saturn in Rome’, 22.1.19), where with Decembri iam mense an exact date is provided which refers to a considerably later moment in the year than the reference time of the main story line, which is early spring.24

21 22 23

24

clearly has forward-linking capacities in that it introduces here one of the protagonists of the main story line that is about to restart. For the concept of discourse mode, see the Introduction to this volume, p. 10. The imperfect tense and the mentioning of an emotive response are typical features of a story-internal perspective. See also Allan, this volume pp. 141–142. Chausserie-Laprée 1969: 69 ff. speaks of clichés de liaison. See also McDonald 1957: 155–159. Levene 2010 §1.2 provides an extensive discussion of how Livy organizes his narrative of the Hannabalic war in terms of chronology. See also Levene 2010: 49–50, who gives a slightly different interpretation of the phrase Decembri iam mense.

198

van gils and kroon

All in all we might conclude that in the first chapter of book 22 we are dealing with a hybrid text type which contains features of both the narrating and the discursive discourse mode.25 This mixture of modes, conveniently reflecting and combining here the two main ‘layers’ of the text (annalistic parts and main story line), might be regarded as a textual strategy for rhetorical purposes, meant to guarantee the coherence of the text and, hence, to retain the audience’s attention also in between the individual narrative sub-arcs. In our discussion of the episode of the battle of Cannae we will see more examples of this strategy. But first we will now turn our attention to the strategic organization of the narrative sub-arcs.

3

Narrative Arcs of Tension and the Story of Abelux as an Illustration26

The use of narrative subarcs is a particularly characteristic feature of Livy’s book 22, as we will show. We find these subarcs in various shapes and sizes, ranging from a single complex sentence to a full-blown, self-contained ministory.27 Examples of such mini-stories in the first half of book 22 are the episode of the battle at lake Trasimene (22.4.2–22.7.5), the episode of the battle in Spain at the mouth of the river Ebro (22.19.1–22.21.8), and the episode of the clever Spanish hostage Abelux (22.22.6–22.22.21).

25

26 27

For the distinction between narrating mode and discursive mode of presentation, see the Introduction to this volume. In the discursive mode the story-world is temporarily suspended, and the communicative situation of the historian and his audience comes to the fore. Undiluted discursive mode, in which the historian is overtly present and ‘in discussion’ with his audience, is relatively rare in book 22. Livy hardly gives explicit comments or judgements on the recounted facts and events, and becomes visible as a historian only at major moments in the narrative organization, or in passages in which he discusses his sources. For examples of the former, see 22.42.10 di prope ipsi eo die magis distulere quam prohibuere imminentem pestem Romanis and 22.43.10 ad nobilitandas clade Romana Cannas urgente fato profecti sunt; for an example of the latter, 22.36.1 Exercitus quoque multiplicati sunt; quantae autem copiae peditum equitumque additae sint, adeo et numero et genere copiarum uariant auctores ut uix quicquam satis certum adfirmare ausus sim. A remarkable instance of metanarrative comment is 22.54.8 Nunquam salua urbe tantum pauoris tumultusque intra moenia Romana fuit. Itaque succumbam oneri neque adgrediar narrare quae edissertando minora uero faciam. See Adema 2008: 175–178 for a discussion of this episode in terms of discourse modes and use of tenses. A good example of one complex sentence in the first half of the book is 22.2.10–22.2.11,

discourse-linguistic strategies in livy’s account of the battle

199

Inserted substories like these comply quite neatly with theoretical models of storytelling, such as the influential model originally proposed by Labov 1972. These models describe the prototypical structure of natural stories in terms of a number of successive standard elements which together form some sort of ‘arc of tension’, displaying a rising action, climax, falling action, and resolution. These arcs may or may not be framed by a so-called abstract and coda, in which the transition is made from the communicative situation (the here and now of speaker/writer and hearer/reader) to the story-world and vice versa. In table 8.1 we repeat the distinctions introduced in the Introduction to the present volume, which are largely based on Labov’s system. When we apply this theory to the story of Abelux, we get quite a clear view of the historian’s skills as a narrator, and of the ways in which he connects selfcontained mini-stories with the overarching main story line. The first sentence of the episode is clearly meant as an announcement of a story that the historian is about to tell, and for which he tries to capture the audience’s attention by highlighting its relevance: (3) [Abstract] eo uinculo Hispaniam uir unus sollerti magis quam fideli consilio exsoluit. LIV. 22.22.6

From this constraint Spain was released by the machinations—more clever than honest—of one man. It is to be noted that in this first stage of the process of tension building (the so-called Abstract) the discourse mode involved is usually not narrating in a strict sense, but discursive: the speaker merely draws attention to the topic of a story that he is about to tell, and is not yet involved in creating that world. In the next stage (Orientation), the reader is actually drawn into this new, secondary reality or world, in which he will become more and more involved and curious about the story’s outcome, and in which, as a narratee, he may himself actively build up expectations, form opinions and draw conclusions, under the guidance of an overtly or more covertly present narrator. In the story of Abelux, this second stage starts immediately after the introductory sentence quoted under (3) above:

where the entire story of how Hannibal lost the sight of one of his eyes is told in the form of one single periodic sentence.

200

van gils and kroon

table 8.1

Prototypical structure of a natural story or episode

Abstract Orientation Complication Peak Resolution Coda Evaluation (any position)

summary of the content or point of the narrative previous history; introduction of time, place, circumstances and main participants a conflict arises, build-up of tension climax: the conflict is maximally tangible and near a solution the conflict is resolved, substituted by another conflict or remains permanently unresolved summarizing bridge to time of narrating evaluation of the narrative or of elements of the narrative, often conveying, or pertaining to, the point of the narrative

(4) [Orientation] Abelux erat Sagunti nobilis Hispanus, fidus ante Poenis, tum, qualia plerumque sunt barbarorum ingenia, cum fortuna mutauerat fidem. ceterum […] id agebat, ut quam maxumum emolumentum nouis sociis esset. LIV. 22.22.6–22.22.7

Abelux was his name, and he was a noble Spaniard of Saguntum. Loyal hitherto to the Phoenicians, he had now—as barbarians are for the most part prone to do—altered his allegiance with the alteration in their fortunes. But […] he proposed to benefit his new allies to the utmost extent of his ability. As is common in Orientations, we find imperfect and pluperfect tense forms (erat, mutauerat), the semantics of which are especially apt to describe the setting (time, place, protagonists, circumstances) within which one or more complications are expected to arise. As appears from the inserted comment qualia plerumque sunt barbarorum ingenia, the narrator is still relatively ‘overt’, steering the narratee’s interpretation of the story into a certain direction. The recurrence in the Orientation of the concept fidus/ fides—introduced in the Abstract by means of the words fideli consilio,—corroborates this observation (see the words fidus ante Poenis; fortuna mutauerat fidem). After this Orientation we get a clear build-up of tension in the form of a first Complication, Abelux’s decision to deliver up the hostages to the Romans:

discourse-linguistic strategies in livy’s account of the battle

201

(5) [Complication] circumspectis igitur omnibus, quae fortuna potestatis eius poterat facere, obsidibus potissimum tradendis animum adiecit, eam unam rem maxime ratus conciliaturam Romanis principum Hispaniae amicitiam. LIV. 22.22.8

And considering everything that fortune could put into his power, he inclined for choice to deliver up the hostages, believing that this was the one thing that would most effectively secure for the Romans the friendship of the Spanish leaders. After this start of the narrative proper, the arc of tension continues in an upward direction with a second Complicating Action, this time conveyed by a verb form in the historic present tense, adgreditur: (6) [Complication] sed cum iniussu Bostaris praefecti satis sciret nihil obsidum custodes facturos esse, Bostarem ipsum arte adgreditur. LIV. 22.22.9

But since he knew that the men guarding the hostages would do nothing without the orders of Bostar, the governor, he artfully approached Bostar himself. The act of approaching Bostar is not presented as the kind of event that is sufficiently newsworthy in itself to be independently stored as such in the narratee’s long-term memory (as the perfect tense form adgressus est would have been), but merely opens a narrative window towards another event, scene or episode that is still to come, in this case the ensuing scene of Abelux’s meeting with Bostar, which will turn out to be the Peak of the story. In this sense the historic present tense can be seen as a forward pointing coherence device: the historic present forms a signal for the narratee that for a yet undetermined period of time the act or event referred to has to be kept mentally active and is not to be stored as a particularly relevant event in itself on the narratee’s ‘mental hard disk’.28 As already suggested above, and as we will explain in § 4 with more examples, Livy often seems to use the inherent open-endedness of the present tense (e.g. as opposed to the ‘factive’ perfect tense) as a textual strategy to raise the narrative tension and to direct the narratee’s attention for-

28

For a cognitive explanation of this use of the present, see Van Gils & Kroon (forthc.).

202

van gils and kroon

wards, therewith stimulating the narratee’s active anticipation of something important or remarkable that is still to come. This is also clearly the case with adgreditur in (6). The next element in Labov’s model of narrative organization, the Peak, may count as the ultimate stage of tension building. The narratee is prototypically incited in this stage to immerse himself completely into the story-world, forgetting as much as possible about his own reality and that of the narrator, and identifying himself with one or more of the characters.29 The rightmost column of the table in Appendix 1 on pp. 229–230 provides the linguistic and narratological features that prototypically occur in this type of immersive story-telling. In the story of Abelux, the Peak is clearly recognizable by the occurrence of both indirect and direct speech: the use of directly reported speech can be seen as a very powerful immersion device. Abelux is talking here to the prefect Bostar, who keeps, on account of Hannibal, the Spanish hostages imprisoned: (7) [Peak] ‘Obsides’ inquit ‘in ciuitates remitte. id et priuatim parentibus, quorum maximum nomen in ciuitatibus est suis, et publice populis gratum erit. uolt sibi quisque credi, et habita fides ipsam plerumque obligat fidem. ministerium restituendorum domos obsidum mihimet deposco ipse, ut opera quoque impensa consilium adiuuem meum et rei suapte natura gratae quantam insuper gratiam possim adiciam.’ LIV. 22.22.12–22.22.14

‘Send back the hostages to their homes,’ said Abelux. ‘That will at once be grateful personally to their parents, who are the people of most consequence in their own states, and to their tribes in general. Everyone wishes to be trusted: confide in people, and almost always you confirm their confidence in you. The task of restoring the hostages to their homes I request for myself, that I may work, as well as counsel, for the furtherance of my plan, and to an act that is gracious in itself lend such added grace as I am able.’ 29

The three stages of tension building and gradual involvement of the audience described here also play a role in a recent experimental research by Bjørner, Magnusson & Nielsen 2016 on cinematographic storytelling, which nicely complements and corroborates our own research. The study shows how people watching a 3D animation film get only progressively involved in a story, by subsequently crossing a number of cognitive boundaries, and also how their degree of immersion lowers when all kinds of obstacles to narrative immersion from outside the story-world are brought in. The researchers propose a circular model, called ‘The Wheel of Immersion’ to describe the dynamic process of different levels of involvement viewers can be in while watching a film.

discourse-linguistic strategies in livy’s account of the battle

203

It is clear that immersive Peaks of the type illustrated in (7) may play an important role in keeping the narratee’s attention and in rousing his special interest. Significantly, however, the Peak of the Abeluxstory does not coincide with what might be considered the climax of the story in terms of relevance for the course of events on a higher level of the narrative (its relevance for the broader ‘narrative arc’, the war against Hannibal). From that point of view, the moment of Abelux’s handing over of the hostages to the Romans, later on in the story, might have been a better candidate for treatment as an immersive climax.30 However, judging from the actual placement of the Peak, this mini-story apparently serves quite specific communicative goals in addition to recounting the most relevant facts of the Hannibalic war. The emphasis on fides, both in the introductory stages of the story—see the italicized words in the passages quoted under 3 and 4—and especially in its immersive peak—see the italicized words in 7—makes clear, we think, that local arcs of narrative tension such as the mini-story of Abelux are created because of their thematic relevance. One of the important themes in book 22 is the virtue of fides, and especially the infidelity and treachery on the Punic side, the Punica fraus.31 The strategic positioning of the Peak of the story, and the concomitant narrative strategy of immersion, clearly serve to highlight this theme in the story of Abelux. After this Peak the rest of the story (22.22.15–22.22.18) is told in a significantly lower key and in quite big steps, marked by present tense forms of the type illustrated by adgreditur in passage (6) above. Abelux secretly informs Scipio of his plans (expromit), returns to Saguntum (redit), and leads the hostages and their guardians into the trap (in praeparatas sua fraude insidias ducit), without Bostar being aware of Abelux’s false intentions. These present tense verb forms lead to the outcome (Resolution) of the story in 22.22.18: the actual handing over of the hostages, first to the Romans and then to their families and friends. The Resolution is told in passive perfect forms which clearly round off the narrative proper (perducti; acta), after which Livy evaluates this outcome by stressing that the gratitude of the Spaniards to the Romans was much bigger than the gratitude to the Carthaginians could ever have been: (8) in castra Romana perducti; cetera omnia de reddendis obsidibus, sicut cum Bostare constitutum erat, acta per eum ⟨eo⟩dem ordine, quo si Carthaginiensium nomine sic ageretur. maior aliquanto Romanorum gratia fuit in re pari, quam quanta futura Carthaginiensium fuerat. LIV. 22.22.18–22.22.19 30 31

For the idea that certain events are more ‘tellable’ than others, see Baroni 2013. For punica fraus see also the contribution by Pausch, this volume, p. 235.

204

van gils and kroon

They were then conducted to the Roman camp. The remainder of the plan for the restoration of the hostages to their friends was carried out, through the agency of Abelux, exactly as he and Bostar had agreed, and everything was done as it would have been if he had been acting in the name of the Carthaginians. The gratitude which the Romans won under such circumstances was much greater than the Carthaginians would have enjoyed. With the end of the story of Abelux the episode in Spain comes to an end, after which the historian returns to his main story line by means of the following transitional sentence, in which in a subtle manner also the book’s central themes of good leadership and of temeritas versus prudentia and ratio is touched upon: (9) Haec in Hispania [quoque] secunda aestate Punici belli gesta, cum in Italia paulum intervalli cladibus Romanis sollers cunctatio Fabii fecisset; LIV. 22.23.1

Such was the course of events in Spain in the second summer of the Punic war. In Italy meanwhile the defeated Romans had been afforded a little breathing space by Fabius’s wise policy of holding back.

4

Digression: Immersion and the Use of the Historic Present in Livy

One of the interesting things we observed when analysing the text of Livy 22 by means of our linguistic-narratological instrument, is that there are not many moments that on account of a specific clustering of features might count as particularly immersive, that is, as causing a high degree of immersion of the narratee in the story-world, for instance by drawing the narratee into subjective or emotional involvement with a character’s perspective or psychology. In the mini-story of Abelux we have analysed the passage of Abelux approaching Bostar as immersive narrative, especially due to the occurrence—always significant—of direct speech (see 7 above).32 High immersivity in narrative is commonly held to co-occur with a disappearance of overt signs of a mediating narrator (such as coherence particles 32

Other instances of direct speech in book 22 can be found in 39 (Fabius’ speech to Paullus), 49 (death of Paullus) and in various scenes after the battle at the end of book 22 where the fate of the surviving Roman soldiers is discussed.

discourse-linguistic strategies in livy’s account of the battle

205

and a complex sentence structure), and, especially, with the use of the historic present tense, which seems to turn the narrative into a vivid eye-witness account in which the narratee is mentally engaged.33 This so-called eye-witness use of the historic present can be illustrated with an excerpt taken from the section in book 22 that precedes the story of Abelux, and which also deals with the war in Spain. The historian describes here, in quite a detailed way, the chaos among the Carthaginian sailors when they are suddenly confronted with a rapidly approaching Roman fleet. In narratological terms, such a passage in which time of narration and story time more or less seem to coincide is called a ‘scene’:34 (10) uixdum omnes conscenderant, cum alii resolutis oris in ancoras euehuntur, alii, ne quid teneat, ancoralia incidunt, raptimque omnia ⟨ac⟩ praepropere agendo militum apparatu nautica ministeria impediuntur, trepidatione nautarum capere et aptare arma miles prohibetur. LIV. 22.19.10

Hardly were they all on board, when some cast off the hawsers and swung out on to their anchors, and others—that nothing might detain them— cut the anchor cables, and, in the hurry and excessive haste with which everything was done, the soldiers’ gear interfered with the sailors in the performance of their tasks, and the confusion of the sailors kept the soldiers from taking and fitting on their armour. As said, this type of highly immersive passages, in which the use of the present tense can be seen as an indication that the addressee is invited to become fully engaged in the story and to immerse himself in the perspective of one or more characters on the spot, is extremely rare in book 22.35 We do find historic presents in Livy regularly, but almost all of them are of the type illustrated in (2) and (6) above (petit; adgreditur), and function as a kind of instruction for the narratee to open a window to, for instance, a new narrative arc, and to actively engage in forming expectations about what will follow in the text. 33 34

35

E.g. Pinkster 1990: 225; Rossi 2004: 131; Adema 2008; Allan, De Jong & De Jonge 2014; Toolan 2016; Grethlein & Huitink 2017. See also Allan, this volume. Walsh 1963: 185–198 speaks of ‘episodes’ which may contain ‘emotional reactions by a vivid, imaginative and often imaginary reconstruction of crowd scenes’ (185). The graphic effect, Walsh continues, is achieved through ‘asyndeton, short clauses, accumulation of words and expressions, historic presents and historic infinitives’. We have found only four or five other examples, and all of them less convincing than example (10).

206

van gils and kroon

Another example is coniungit in (11), also taken from the Spanish episode in book 22. In the text preceding this example Livy has briefly told how Publius Scipio arrived in the harbour of Terragona with a large fleet. (11) ea classis ingens […] portum Tarraconis ex alto tenuit. ibi milite exposito profectus Scipio fratri se coniungit, ac deinde communi animo consilioque gerebant bellum. LIV. 22.22.2–22.22.3

This fleet […] dropped anchor in the harbour of Tarraco. There Scipio disembarked his troops and set out to join his brother; and from that time forward they carried on the war with perfect harmony of temper and of purpose. The semantic value of simultaneity of the present tense clearly does not have anything to do here with creating the impression of an eye-witness account (or with the reader identifying with a character in an immersive narrative mode of presentation).36 The present tense rather seems to underline, we think, that the narrator takes along the narratee in constructing the narrative, taking a step that is not necessarily new or surprising in itself, but which is needed here on which to build a new narrative arc. The arc anticipated in (11) will prove to be a major one: a small part of it is given in the next chapter (first action of the two brothers), but most of it will extend even beyond the boundaries of book 22. The view that the historic present tense somehow has a prospective force and stimulates an active anticipation on the part of the narratee of what is to follow, seems to be corroborated here by the presence of the imperfect tense form gerebant in the following sentence, which has the same semantic open-endedness as the present tense. Both verb forms have a tension raising effect here, which would have been absent had the perfect tense been used. The perfect tense form coniunxit would have made the event of the reunion of the Scipio brothers into a mere subsequent item in a chronological series of historical facts, without any structuring or tension-building effects. Example (11) forms, moreover, a good illustration of the coherence and tension-creating force of certain thematic strands, even across individual books. It is no coincidence, we think, that in addition to the use of the prospective historic present coniungit and the tension-building imperfect gerebant, 36

See Pinkster 2015: 395 for the semantic value of simultaneity of the present tense. In Van Gils & Kroon (forthc.) a more elaborate argumentation is provided for this interpretation of the historic present.

discourse-linguistic strategies in livy’s account of the battle

207

we also find, in the expression communi animo consilioque, a reference to the important concordia/discordia theme that dominates the first book of Livy’s third decade: although the narrative is now heading straight for the defeat at Cannae, the narratee seems to be invited here to already look ahead of that, towards the eventual recovery of the Romans and the salutary virtue of concordia. Various textual strategies are clearly working together here in leading the audience through the multi-layered structure of the third decade. Finally, a particularly clear example we want to give here of the tensionbuilding use of the historic present tense in the first half of book 22 is (12). After the episode of the defeat at Lake Trasimene, the historian inserts a long passage on the reactions in Rome and the decisions made there. With the sentence quoted in (12) he finally returns to the main story line of the campaign against Hannibal, and it is here that after a long time we encounter again a historic present, ducit: (12) Dictator exercitu consulis accepto a Fuluio Flacco legato per agrum Sabinum Tibur, quo diem ad conueniendum edixerat nouis militibus, uenit. inde Praeneste ac transuersis limitibus in uiam Latinam est egressus, unde itineribus summa cum cura exploratis ad hostem ducit, nullo loco, nisi quantum necessitas cogeret, fortunae se commissurus. LIV. 22.12.1–22.12.2

The dictator, after taking over the consul’s army from Fulvius Flaccus, his lieutenant, marched through the Sabine country to Tibur, where he had given the new levies notice to assemble on a certain day. From Tibur he marched to Praeneste, and striking across the country came out into the Latin Way, and then, reconnoitring the roads with the utmost circumspection, advanced in the direction of the enemy, though resolved nowhere to commit himself to fortune, except in so far as necessity might compel him. Like in (11) above, the present tense opens an important new window, raising expectations for things to come. The anticipatory nature of the passage is corroborated and made explicit by the future participle commissurus, and gains even more in weight by the reference to one of the other important themes of the book: prudentia, as opposed to temeritas (see the italicized words itineribus summa cum cura exploratis). It is obviously that it is here that the rising arc of tension towards Cannae starts. The use of a perfect tense duxit would have presented this action as merely another narrative step after egressus est, instead of marking its relevance for future events in the story.

208 5

van gils and kroon

The Battle of Cannae

After this overview of Livy’s narrative method in the first half of book 22 of his History, we may now proceed to our analysis of the account of the battle of Cannae. How does Livy employ the aforementioned and other narrative strategies in his presentation of the defeat at Cannae? 5.1 The Beginning of the Battle Defeats have been presented before in Livy’s historiography, and the topic in itself is well suited for a presentation in one continuous narrative arc which starts with the first strategic plan and ends with its announcement in Rome. In spite of this narrative potential, we do not see such a single narrative arc in Livy’s treatment of the battle of Cannae, but rather a series of connected episodes which may or may not have their own narrative arc of tension. We will argue that in the Cannae episode, as in the first half of book 22, Livy’s narrative strategies have two intended effects on the reader: on the one hand they highlight important themes, such as discordia, fides, fortuna and temeritas, and on the other hand they provide strong cohesive ties between otherwise loosely connected episodes. The first thing we need to decide upon for our analysis is where exactly the Cannae episode starts. Since there is not one narrative arc in which the battle of Cannae is presented, it is possible to choose various starting points.37 The battle proper starts at 22.47.1 with the typical battle cry (clamore sublato), and finishes in 22.49.18 with the equally typical account of the number of people injured, killed, put to flight and made captive. When taking a slightly broader view, however, we could pinpoint the start of the Cannae-story at the arrival of the consuls in Cannae (22.44.1); or, also including the preparation of the enemy, we could start with Hannibal’s decision to move southwards when he faced famine and rebellion in his previous camp at Gereonium (22.43.1); on account of the inevitability of the encounter of the two armies in 216, the story could also be taken as starting when the consuls Varro and Paullus leave Rome for their armies (22.40.4), or even before their departure with the speeches about how to wage a war against Hannibal (22.38.1). And we may want to start even earlier, and to include in the Cannae-story the four chapters preceding 22.38, where we are successively informed about the elections of the consuls Varro and Paullus (22.34, 22.35), about the levy of the army (22.36), about certain prodigies and embassies (22.36), and about the embassy of Hiero from Sicily (22.37).

37

Oakley (this volume: 182) lets the story start in 22.33.9, as we do as well, see below.

discourse-linguistic strategies in livy’s account of the battle

209

In the following subsections we will analyze the battle narrative in a strict sense (22.47–22.49) as a complex, narrative unit, but we will also look at the larger story (starting at the end of 22.33 and ending at 22.61), and show how Livy uses various narrative strategies to present the elections, prodigies, and embassies in 217 together with the war preparations and the battle of Cannae in 216 BC as correlated events in Roman history. We therefore take the consular elections for the year 216 as the starting point of the story about the battle of Cannae. The idea that the outcome of elections could be the starting point of a military defeat is easily explained within the logic of ancient historiography, in which the outcome of great events, like battles or conquests, could and should be related to the conduct of individuals, usually the generals of an army.38 This particular view on causes and effects in complex societies may be problematic from our modern point of view on the goals of historiography, but if one takes into account one of the most important objectives in ancient historiography, namely to teach later generations about morally good and bad leadership, the focus on the political and military leaders is quite understandable. It is therefore natural that the story about Rome’s most traumatic defeat should begin with the election of the discordant consuls who would command the Roman army against Hannibal. The theme discordia is present from the first introduction of the elections till the final stages of the battle.39 If we look at the larger narrative structure of book 22, the elections are presented as a very problematic political episode in between the two main defeats of book 22—Lake Trasimene and Cannae—and, in spite of their chronological place in the year 217BC, appear to be more strongly connected to Cannae than to the past and contemporary events of the year 217. Livy does not interrupt the tragic episode of the Lake Trasimene or the following episodes about the prudent dictator Fabius and his rash master of the horse Minucius with a section on the elections in Rome.40 Instead, the typically annalistic material of electoral outcomes is presented only after the reader has heard about the armies 38 39

40

See Cic. De orat. II.63. E.g. comitia habita magno certamine patrum ac plebis (22.34.1) and at the end of the battle, Paullus explicitly says he would rather die than become an accusator collegae (22.49.11). Two examples of concordia just before the elections (summa inter se concordia in 22.32.1 and the delayed construction of the temple of Concordia in 22.33.7) may be interpreted as anticipating by contrast the upcoming discordia with its disastrous consequences for the war. Also in section 2 above we have seen that some typically annalistic events throughout 217 were mentioned clustered instead of scattered throughout the main story line about Hannibal’s advancement into Italy.

210

van gils and kroon

having retired in their winter camps (22.32.4). Probably the elections had taken place already by that time.41 A reader will not easily recognize this deviation from the chronological order, as Livy loosely relates his information about the Roman winter camps to information about envoys arriving in Rome.42 After some elaboration about this embassy, an equally vague connection brings the reader to yet another short annalistic report, this time about remarkable punishments, and ending with the city praetor arranging the building of a new temple.43 The next subtle connection is that this same praetor had written to the consuls with a request to hold elections for the new consuls.44 Without noticing, we have made a chronological step backwards in this sentence.45 All in all, in terms of narrative strategies, Livy apparently wants to present the consular elections as the start of the Cannae-story rather than connect them to the events of the preceding year 217. From a narrative strategic point of view, the difficult process of organizing consul elections for the year 216 anticipates the problematic year itself, especially with regard to Roman leadership.46 5.2 Types of Narrative Arcs in the Cannae Episode Each event in Livy’s History, whether it is the election of new consuls or the heat of the battle, may in essence be either summarily reported in the discursive mode or presented in more detail in the narrative mode. Why certain events are presented as full-blown narrative episodes is an intriguing, but difficult question to answer.47 Based on the analysis presented in this chapter, we see that in book 22 Livy chooses the more elaborate, narrative option for events which best illustrate his main themes, like fraus Punica or Roman discordia, and the discursive presentation for the other events which need to be 41 42 43 44 45 46

47

A chronological reconstruction of the consular elections for the year 216 can be found in Sumner 1975. Cum ad Gereonium iam hieme impediente constitisset bellum, Neapolitani legati Roman uenere. (22.32.4) Per eosdem dies speculator Carthaginiensis (22.33.1). Ab eodem praetor … litterae ad consules missae (22.33.9). See Levene 2010: 36–63 for a discussion of various examples of subtle non-chronological presentation of events and vague temporal expressions to link non-related events. This observation fits in with Levene’s general conclusion on Livy’s third decade (2010: 52): ‘while Livy is aware of his calendrical anachronisms, and generally ensures that they do not overtly damage the main chronological sequence of his narrative, he is simultaneously prepared to allow them to maintain a narrative effect according to their place in his ‘textual’ year rather than according to when they actually occurred.’ Reasons for presenting events in a narrative episode may be their immersive potential or simply the level of detail in available sources, but neither seem to play a role in the case of Livy’s book 22.

discourse-linguistic strategies in livy’s account of the battle

211

covered by the annalistic historiographer, especially the events of the political years 217 and 216BC in Rome. Narrative episodes in the Cannae-story, and elsewhere, are distinguished by the presence of a narrative arc, which prototypically contains the fixed set of components summed up in table 8.1 above (section 3), namely Abstract, Orientation, Complication, Peak, Resolution and Coda.48 When we look at Livy’s use of the narrating mode in the second half of book 22, we find only a few episodes told with a complete narrative arc and most of them told after the defeat.49 Other episodes contain narrative components, but these are of a much more complex type than the prototypical Labovian pattern and seem to be intentionally structured in such a way as to continually increase the narrative tension. The patterns we find in the second half of book 22, and which may also be combined, are of the following kind: (i) An episode containing all elements of a narrative arc (ii) The periodic sentence (which contains a narrative arc within the scope of one syntactic unit) (iii) A short narrative episode (which contains a complete narrative arc in a few sentences) (iv) An extended ascending narrative arc (which contains an abstract, orientation and complication) (v) A change of perspective halfway through a story (which has the effect of starting a new narrative arc) In the last two types, we find ‘Labovian’ narrative arcs which are incomplete either because they lack the final components of Resolution and Coda (type iv), or because they lack the initiating components of Abstract and Orientation (type v). We will see below how the historian in book 22 combines especially the patterns (iv) and (v) combining an ascending narrative arc with a change of perspective, to the effect that halfway through a complication the story is interrupted by a new viewpoint on the situation. In the following sections, we treat the various episodes in the broader Cannae-story as composed of narrative, discursive and hybrid elements, paying specific attention to narrative patterns and strategies.50 The results of this analysis reveal narrative and historiographical choices which, we believe, are 48 49

50

See also the Introduction to this volume. The Abelux-episode discussed in section 3 of this chapter illustrates such a typical narrative arc. Before the end of the battle at Cannae, complete narrative arcs are found for the Punic ruse at the end of 48 and, immediately afterwards, the dying Paullus in 49. See Appendix 2 for examples of complete narrative arcs after 49. See Appendix 2 for a schematic overview of the narrative structure of book 22, based on our analysis.

212

van gils and kroon

directly relatable to the historian’s particular ideological and moral stances and to his efforts to keep the reader involved. 5.3 Elections, Armies, Prodigies, Embassies (22.34–22.37) In 22.34 the plain discursive presentation of events, which continues the text type of 22.33, soon makes place for an ascending narrative arc:51 (13) Consulibus prorogatum in annum imperium. interreges proditi a patribus C. Claudius Appi filius Cento, inde P. Cornelius Asina. in eius interregno comitia habita magno certamine patrum ac plebis. LIV. 22.34.1

The authority of the consuls was extended for a year. To be interrex the Fathers named Gaius Claudius Cento, the son of Appius, and after him Publius Cornelius Asina. The latter conducted an election, which was marked by a bitter struggle between patricians and plebeians. All three sentences contain passive perfect tense predicates—a typical feature of the discursive mode, see Appendix 1—and lack any form of narrative tension. However, the last sentence clearly hints at a story by announcing a major conflict (magno certamine) between the senators and the plebs. The experienced reader of these sentences will supposedly recognize the dominant theme of Roman discordia here, which heightens his expectation of an upcoming story. A discursive sentence announcing a thematically relevant event may function as an Abstract of a narrative. Such abstracts do not necessarily summarize the main event of the story, but may point to a significant aspect of it, which in Livy’s book 22 is usually one of the central themes of the decade, such as discordia, temeritas and fraus, or their opposites.52 The abstract in 22.34.1 is indeed followed by the next component in the Labovian narrative model, an Orientation. The plebeian protagonist of this narrative episode is introduced in the remaining part of 22.34 with his name, background and character in imperfect tenses and with his thoughts in indirect speech. (14) C. Terentio Varroni, quem sui generis hominem, plebi insectatione principum popularibusque artibus conciliatum, ab Q. Fabi opibus et dictato51

52

The theme of consular elections starts at 22.33.9 with a few discursive sentences about the difficult epistolary communication between the senate in Rome and the consuls who stay outside Rome with their armies. See also examples 3 above and 20 below, where the abstracts point at a story about temeritas.

discourse-linguistic strategies in livy’s account of the battle

213

rio imperio concusso aliena inuidia splendentem, uolgus [et] extrahere ad consulatum nitebatur, patres summa ope obstabant, ne se insectando sibi aequari adsuescerent homines. LIV. 22.34.2

Gaius Terentius Varro had endeared himself to the plebeians—the class to which he himself belonged—by invective against the leading men and the usual tricks of the demagogue. The blow he had struck at the influence and dictatorial authority of Fabius brought him the glory which is won by defaming others, and the rabble was now striving to raise him to the consulship, while the patricians opposed the attempt with all their might, lest men should acquire the custom of assailing them as a means of rising to their level. At 22.35, the Complication starts when the character Varro, with his provocative behavior and his atypical background for a consul, reaches this political rank as a result of his popularity at the elections:53 (15) Cum his orationibus accensa plebs esset, […] C. Terentius consul unus creatur, ut in manu eius essent comitia rogando collegae. LIV. 22.35.1–22.35.2

When the plebs had been inflamed by these harangues, […] Gaius Terentius was the only consul elected, and the assembly called to choose a colleague for him was therefore under his control. The event of the election of the popular Terentius Varro who opposed the leading elite, leaving all his aristocratic competitors far behind, is the first in five capita to be presented with a historic present tense (creatur). This historic present is clearly of the type discussed above with regard to petit in example (2), adgreditur in (6), coniungit in (11) and ducit in (12): the choice for a present tense instead of a perfect tense makes clear that the election of Varro is not a mere subsequent fact in an annalistic account of historical events, but rather a consequential event in the main story line.54 This means that creatur starts a

53

54

Historical veracity of Livy’s presentation is not our object of study: whenever we discuss historical events as treated by Livy, these should be taken as being part of the story world and not as historical reality. Compare the present tense form creatur here with creati in 22.35.5 which is clearly

214

van gils and kroon

narrative arc, in this case one that will lead to the disastrous battle of Cannae. The preceding discourse has prepared the reader for this narrative development and it is immediately clear that it complicates the course of events. However, the expected next step, the satisfaction of the plebs about this election, which would count as a resolution to this narrative arc, does not (yet) follow. Instead, the ascending narrative arc is interrupted, due to a radical switch to another point of view, namely the perspective of the nobilitas: (16) tum experta nobilitas parum fuisse uirium in competitoribus eius, L. Aemilium Paulum, qui cum M. Liuio consul fuerat, ex damnatione collegae, ex qua prope ambustus euaserat, infestum plebei, diu ac multum recusantem ad petitionem compellit. LIV. 22.35.3

The nobles, finding that Varro’s competitors had not been able to command the necessary strength, thereupon obliged Lucius Aemilius Paullus to stand, though he held out long and earnestly against their importunity. He had been consul together with Marcus Livius, and the condemnation of his colleague—from which he had not himself escaped unscathed— had embittered him against the plebs. Within one single sentence we find out about the deliberations of the nobiles and their decision to put forward Aemilius Paullus as the only candidate for the position of second consul. This periodic narrative sentence contains an Orientation with regard to the thoughts of the nobilitas (experta … eius) and an Orientation with regard to the newly introduced character of Lucius Aemilius Paullus (L. Aemilium Paulum … infestum plebei), and ends with a Complicating (and consequential) new action within the situation sketched: nobilitas Paulum ad petitionem compellit. Periodic sentences like these build up narrative tension through two or more embedded clauses and they often contain character motivations for the particular course of action that is conveyed by the finite verb form of the sentence. The change in perspective from the plebs to the nobilitas is both a break and a continuation of the narrative arc about the election of Varro. In fact, the two perspectives of Varro and his supporters on the one hand and the nobilitas and their leaders on the other will continue to alternate as simultaneous

included in an ‘annalistic’ context and lacks the tension-raising capacities that creatur has.

discourse-linguistic strategies in livy’s account of the battle

215

narrative arcs with conflicting objectives.55 Later on, we will find a third perspective, namely that of Hannibal.56 In the story about Cannae, the continuous alternation of these three perspectives stimulates the curiosity of the reader about the outcome of multiple ascending narrative arcs, in which the component called Complicating Action is never followed by a Peak or Resolution, but by a switch to one of the other two perspectives—as is also the case in (16) above. Often the sentence before such a change in perspective contains a historic present tense, which stresses the open-endedness of the current narrative arc and gives the effect of continuing tension.57 These present tense forms, such as creatur in (15) and compellit in (16), usually refer to logical and well-motivated actions of a character or group, which will appear to have major consequences for the following course of events. It is remarkable how Livy uses, or rather abuses, the forward looking capacity of the historic present to keep the reader alert just before radically moving perspective, leaving the ascending narrative arc without a proper peak, let alone a resolution. The effect is clearly that of a suspenseful story of multiperspectivalism in which the parallel universes of Hannibal, the plebeian consul with soldiers and plebs, and the senatorial consul with the nobilitas all contribute to illustrate the major themes of discordia, temeritas and fraus, or their opposites.58 The story about the turbulent consular elections of 216, which started in 22.34.1 with a discursive sentence containing an Abstract (example 13), continued with an Orientation and Complication and was interrupted by a change in perspective, ends (22.35) with information about the elected praetors and consuls. The sentences are again typically discursive here, with the passive voice, no individual agents and a summarizing rhythm.59 In terms of narrative coherence, these sentences present the Resolution and Coda of the story of the elections.

55 56 57

58

59

See also Varrone indignante (22.41.3), et consul alter (22.42.4), and Quod quamquam Varro (22.42.9). For Hannibal’s view as a sudden break in perspective, see 22.40.7, 22.41.4, 22.43.1, 22.44.4, 22.45.1, 22.46.1; or back from Hannibal to the Romans, see 22.42.1, 22.44.1. In addition to creating unresolved narrative tension, the combined perspectives also provide the narratee with more information than the characters (dramatic irony). See Pausch (this volume, 241) for the suspense created by presenting two opposing views. Examples of the historic present just before a change of perspective and hence apparently at the end of an episode are terga uertunt (22.47.4), integram pugnam ineunt (22.47.10), ad persequendos mittit Hispanos et Gallos pedites adiungit (22.48.6). For an overview of the most characteristic features of the discursive mode in Latin historiographical prose, see the leftmost column of Appendix 1 to this chapter.

216

van gils and kroon

Next, at 22.36, we are informed about the levy of the Roman armies and about alarming prodigies. Apart from an implicit and weak temporal relation with the preceding story about the elections and an explicit, but vague temporal relation with the following story about the Sicilian ambassadors (per eosdem dies, 22.37.1), there is no relation between the military preparations and prodigies in 36 and the surrounding discourse.60 The first sentence illustrates well the typical features of the discursive mode of presentation: (17) Exercitus quoque multiplicati sunt; quantae autem copiae peditum equitumque additae sint, adeo et numero et genere copiarum uariant auctores, ut uix quicquam satis certum adfirmare ausus sim. LIV. 22.36.1

The armies also were augmented. But how large were the additions of infantry and cavalry I should hardly venture to declare with any certainty—so greatly do historians differ in regard to the numbers and kinds of troops. The narrative pace is summarizing and lacking in details, and we find the passive voice (multiplicati sunt) without any reference to individual agents. The present tense (uariant) is a ‘real’ present tense in that it refers to the hic-et-nunc of the author and the audience: the historian explicitly discusses the source of his information (auctores) in this sentence and even refers to himself (ausus sim). There is no individual agent with personal features and no arc of tension. The prodigies in the second part of the chapter are told in the same, discursive way (passive voice, no individual agents, no narrative tension, general statements). The effect of this mode of presentation is clearly more matter-of-fact and less emotionally involving for the reader. In the discursive passages the historiographer presents himself as an author who can be trusted to include all relevant annalistic facts after thorough personal consideration of the sources. 5.4 Words of Soldiers and Generals (22.38–22.40.4) The passage in book 22 that follows is characterized by the conventional insertion of (reported) speech. The words of combatants and especially of generals are a typical part of war narrative.61 They may be quoted in direct speech, as happens with Fabius’ premonitory speech (22.39), or presented in indirect speech as we see with Paullus’ answer (22.40). An even more reduced pre60 61

See also section 2 above. See Adema 2017, and in this volume 293.

discourse-linguistic strategies in livy’s account of the battle

217

sentation of speech is the summarizing treatment of the new oath of the soldiers (22.38). Direct speech especially is a forceful instrument for involving the reader and, as is also demonstrated by Buijs (this volume), for bringing to the fore important themes uttered by characters who are mostly affected by the course of history themselves. In terms of narrative tension, the speech of Fabius at 22.39 may, as a whole, be seen as a rising narrative arc which brings other characters, and especially the addressee Paullus, in a situation which is typical of a ‘Complication’.62 Paullus’ answer resolves the tension by acknowledging the points Fabius has made. It is to be noted that the speech of Fabius highlights several of the themes we have established above as dominant, namely discordia between generals, fear of fraus punica and an admonition towards prudentia. As a whole this set of speeches, coming right after the story of the elections, is yet another foreshadowing of the looming disaster. There is apparent thematic coherence in the presentation of events anticipating the battle of Cannae. However, the events themselves (elections, speeches and the following arrival at the encampment in 22.40) are only loosely connected and do not constitute one narrative arc. 5.5 Arrival and First Hostilities in Gereonium (22.40.5–22.42) One of the aspects in which literary storytelling typically differs from everyday oral storytelling, is the presentation of multiple perspectives. We have seen already that in his presentation of the consular elections, the narrator easily switches between the short-sighted view of the masses and the cautious point of view of the nobilitas (example 16). In the episode in which the Roman army arrives at Gereonium, a third point of view is added: the strategic thoughts of Hannibal.63 An interesting question is how exactly such a changing of perspective might interfere with a narrative arc. We refrain here from a theoretical digression on the various ways in which a change of perspective may interrupt, expand or complicate a narrative arc. Instead, we merely indicate the effect in terms of reader involvement: an arc is, as it were, interrupted by a new point of view and the reader is implicitly invited to balance the two points of view and form a critical attitude towards at least one of them. 62

63

A premonitory speech is expected to heighten the tension because the speaker highlights an uncomfortable or dangerous truth for the addressee who is expected to react in either a contrastive or proactive way (peak) or simply by accepting the inevitability of the situation (resolution). See Van Gils in this volume: 265 for the spatial indications of the various points of view.

218

van gils and kroon

A first example of a shift in perspective to Hannibal is found when Varro and Paullus arrive at the camp near Hannibal and take control of their legions: (18) ut in castra uenerunt, […] Geminum Seruilium in minoribus castris legioni Romanae […] praeficiunt. Hannibal […] tamen aduentu consulum mire gaudere. non solum enim nihil […] superabat […]. LIV. 22.40.5–22.40.8

When they got to the camp […] Geminus Servilius was put in command of the smaller camp […]. Hannibal […] was nevertheless greatly rejoiced at the coming of the consuls. For not only were the spoils exhausted […]. The consuls are, exceptionally enough, presented as unanimously taking a decision: they put consul Geminus Servilius in command of the smaller camp (praeficiunt, a historical present marking, as we saw before, a consequential action), but before we hear about the consequences, our viewpoint is directed toward Hannibal, who is oddly happy to see the Roman army augmented. Hannibal’s joy is explained in a mixed narrative-discursive sentence.64 In the following episode, which forms the overture to the actual battle at Cannae, we also find an example of mixed, narrating-discursive telling: (19) Ceterum temeritati consulis ac ⟨prae⟩propero ingenio materiam etiam fortuna dedit, quod in prohibendis praedatoribus tumultuario proelio ac procursu magis militum quam ex praeparato aut [in]iussu imperatorum orto haudquaquam par Poenis dimicatio fuit. ad mille et septingenti caesi, non plus centum Romanorum sociorumque occisis. ceterum uictoribus effuse sequentibus metu insidiarum obstitit Paulus consul, cuius eo die—nam alternis imperitabant—imperium erat, Varrone indignante ac uociferante emissum hostem e manibus debellarique, ni cessatum foret, potuisse. LIV. 22.41.1–22.41.3

But even Fortune furnished material to the recklessness and over-hasty temper of the consul. The repulse of a foraging party had led to a general melee, which came about from the soldiers rushing forward to attack the enemy, rather than from any plan or orders on the part of the generals; 64

Narrative elements are the imperfect tense and spatial indications from Hannibal’s relative viewpoint, and discursive elements enim, references to later sources, counterfactual mood.

discourse-linguistic strategies in livy’s account of the battle

219

and in this the Phoenicians by no means held their own. About seventeen hundred of them were slain and not more than a hundred of Romans and allies. But the consul Paullus, who was in command that day—for they commanded on alternate days—was fearful of an ambuscade and checked the victors in their headlong pursuit, despite the angry remonstrances of Varro, who cried out that they had let the enemy slip through their hands and that they might have brought the war to a conclusion if they had not relaxed their efforts. In the above passage the historian recounts the first skirmishes between the Romans and the Carthaginians after the arrival of the new consuls Paullus and Varro in the Roman camp in central Italy. In the preceding paragraph we have seen Hannibal greatly rejoicing at the arrival of the consuls and at the advantageous prospect of a combat, pressed hard as he is by a lack of supplies and the expected desertion of his Spanish troops. The first sentence of the passage in (19) at first sight seems to prepare the reader for the start of a new narrative subarc of tension, and, hence, for a narrating mode of presentation: the clause ceterum temeritati consulis ac praepropero ingenio materiam etiam fortuna dedit (‘But even Fortune furnished material to the recklessness and over-hasty temper of the consul’) has clear characteristics of an Abstract, mentioning the significant themes of temeritas and fortuna. What follows, however, is not a gradual building up of a narrative arc of tension (via an Orientation and Complication rising towards a Peak), by means of which the narratee may become more and more involved in a story-world that is being evoked. Rather, the discursive mode of presentation seems to continue, with the historian supplying a number of relevant facts without elaborating upon them in an ‘episodic’, story-telling way: there has been a fight in which the Carthaginians were hardly a match (haudquaquam par Poenis dimicatio fuit), and the casualties at the Carthaginian side outnumbered by far the Roman casualties (ad mille et septingenti caesi, non plus centum Romanorum sociorumque occisis). Typically discursive features in the passage are the use of the ‘factive’ perfect tense, the lack of a clear temporal sequentiality of the reported events, the mentioning of approximate, large numbers, the use of negation (haudquaquam) and comparison (magis … quam), and the low degree of ‘agentivity’.65 The paragraph seems to end, however, in a mixed narrating65

This low degree of agentivity appears from the use of the abstract verbal noun dimicatio, the passive verb form caesi, and the lack of specific, individualized protagonists. For an overview of features of the discursive mode in Latin historiographical prose, see Appendix 1 to this chapter.

220

van gils and kroon

discursive way, with the return of the protagonists Paullus and Varro, and the mentioning of their (re)actions—in Varro’s case even being rendered in the form of an indirect speech. This touch of narrativity in an otherwise discursive context might be explained in terms of a coherence-creating narrative strategy, similar to the hybrid presentation in example 2 in section 2 above. What we see in example 19 is that Livy, by adding a few well-chosen narrative ingredients at the beginning and at the end, brings to the fore the entire cluster of themes that dominate book 22, especially discordia consulum, ratio versus fortuna, and temeritas versus ratio/prudentia. Instead of blowing up the fortuitous and relatively unimportant Roman victory recounted here into a full mini-story with its own narrative arc of tension (which at this point in book 22—shortly before Cannae— would probably have raised the tension too high), the historian chooses to recount the battle largely in the authoritative, matter-of-fact style of the discursive mode. Its particular framing, however, makes unmistakably clear how the passage thematically coheres with book 22 as a whole, and keeps the audience’s attention alive and focused. At the end of this mixed narrative-discursive account of the small victory for the Romans, we again get the viewpoint of Hannibal (22.41.4), who evaluates his loss as an opportunity to provoke the temeritas of the ‘rougher consul and the recruits’. Hannibal’s view smoothly connects the Resolution of this lost skirmish to the Orientation of a new narrative arc, about an attempt to trick the Romans into believing that Hannibal has left with his army behind the hills. The historic presents relinquit, condit, traducit together ‘open up’ the complication. The narrative arc of this story about Punic fraus contains a change in perspective to the Roman soldiers once Hannibal’s army is hidden during the night (22.42.1 Ubi inluxit) and a Peak when the soldiers start to shout (22.42.3 clamor) and consul Paullus pleads for prudentia (22.42.4 Paullus etiam atque etiam dicere prouidendum praecauendumque esse).66 The historic infinitive underlines the inconclusiveness of the situation. But with the following sentence (postremo), the immersive effect ends through some clear text structuring elements from the narrator’s perspective and with the historic present mittit (22.42.4) the narrator marks a new direction in the course of events: Paullus sends praefect Statilius to check out the situation and Statilius’ report (renuntiat) confirms Paullus’ fear of an ambush.67 The unexpected effect of Statilius’ 66

67

The effect of reader immersion which is typical of a Peak is achieved through auditive and visual details of the situation and the inclusion of character perspectives on the situation. The immersive effect of the preceding sentence ends through some clear indications that

discourse-linguistic strategies in livy’s account of the battle

221

words on the soldiers and their leader Varro is summarized by the narrator: they are determined to go to Hannibal’s camp and, in spite of Statilius’ warning, fall in Hannibal’s trap. The two completely contrasting reactions to Hannibal by Varro and his soldiers on the one hand and Paullus and Statilius on the other are another illustration of the Roman discordia which will also determine the outcome of the battle of Cannae. At this point in the course of events, however, the Romans are still saved by the gods. The surprising denouement to this story of fraud is told in the form of a mini-episode with one sentence for every Labovian narrative component: (20) [ABSTRACT] di prope ipsi eo die magis distulere quam prohibuere inminentem pestem Romanis: [ORIENTATION] nam forte ita euenit ut, cum referri signa in castra iubenti consuli milites non parerent, serui duo, Formiani unus, alter Sidicini equitis, qui Seruilio atque Atilio consulibus inter pabulatores excepti a Numidis fuerant, profugerent eo die ad dominos; [COMPLICATION] deductique ad consules nuntiant omnem exercitum Hannibalis trans proximos montes sedere in insidiis. [RESOLUTION AND CODA] horum opportunus aduentus consules imperii potentes fecit, cum ambitio alterius suam primum apud eos praua indulgentia maiestatem soluisset. LIV. 22.42.10–22.42.12

On that day, it might almost be said, the very gods put off, but did not prevent, the calamity that impended over the Romans: for it chanced that when the consul ordered the standards back into the camp and the soldiers were refusing to obey him, two slaves appeared on the scene, one belonging to a Formian, the other to a Sidicinian knight. They had been captured by the Numidians, along with other foragers, in the consulship of Servilius and Atilius, and on that day had escaped back to their masters. Being conducted to the consuls, they stated that Hannibal’s entire army was lying in ambush just over the nearest hills. Their opportune arrival restored the authority of the consuls, when one of them, by running after popularity, and by unprincipled indulgence, had impaired their prestige—beginning with his own—amongst the soldiers.

the narrator’s structuring and evaluating perspective is back: postremo, seditio, and the full denomination of Marium Statilium praefectum.

222

van gils and kroon

The Abstract of this story announces how the gods themselves postponed the imminent disaster of the Roman defeat. The story reinforces the two important themes of discordia and fraus: only divine intervention could temporarily help the two discordant consuls against Hannibal and the prudent consul was, of course, right in suspecting Hannibal’s insidiae.68 The use of the historic present (nuntiant) in the Complication of this short narrative episode marks the future relevance of this action.69 The consequences are summarized in the next sentence which functions as Resolution and Coda to this story.70 A ministory like this seems to be the perfect form to tell a thematically relevant story which does not pertain to the main story line. 5.6 Movement to Cannae and Preparations for Battle (22.43–22.45) After his failed attempt to ambush the Romans, Hannibal decides to move to Cannae and the consuls Paullus and Varro follow shortly afterwards. This episode is told quite elaborately with a narrative arc starting with an Orientation and moving forward with a Complicating action (statuit). This would be a typical context for a historical present. The form statuit is ambiguous, but based on our analysis of similar contexts we propose to interpret it as a present tense, and not a perfect tense: the consequences of Hannibal’s decision to move southwards will include the fatal battle at Cannae, a consequence explicitly anticipated here by the narrator.71 In this passage, as elsewhere in historiographical narrative, perfect tense forms are the unmarked way to tell a chronologically structured story. Similar to the situation in Gereonium (praeficiunt, 22.40.7), a present tense action by the two consuls (communiunt, 22.44.1) opens up a new narrative arc, the rare plural form of which promises well for future events. This hope vanishes quickly when the perspective switches to Hannibal again who knows perfectly well how to push the Roman leaders in the direction of disagreement (dirigit aciem lacessitque … hostes, 22.44.4). Hannibal’s consequential action (note the historic presents) has the desired consequences. And while the Roman consuls contest each other (discordia consulum, 22.44.5), Hannibal provokes them further by sending his Numidians in a surprise attack (mittit, 22.45.2). The expected reaction by Varro, as soon as he is in command, does not keep the reader waiting: the armies are prepared for the final battle with the consuls Varro 68 69 70 71

See Pausch (this volume, 237–240) for the role of insidiae in books 21 and 22. See § 4 above. The aftermath of the peak is told (Resolution), but we also find a comment about the general significance of this story (Coda). ad nobilitandas clade Romana Cannas urgente fato profecti sunt (22.43.10).

discourse-linguistic strategies in livy’s account of the battle

223

and Paullus on respectively the left and right wing and the ex-consul Geminus Seruilius in the middle.72 5.7 Description of the Armies and the Battle (22.46–22.49) The episode of the actual battle starts with a description of the enemy with imperfect tenses and visual details (22.46), which, in terms of Labov, can be seen as an Orientation to the battle itself. Battle narratives are a traditional ingredient of ancient historiography and tend to contain typical elements which are recognized, and maybe even expected as such by the audience.73 The description of the armies is one of those elements, and others are the battle cry (clamore sublato, 22.47.1) which signals the beginning of the fighting and the description of simultaneous fighting scenes told in a series of connected episodes from left to right wing or vice versa.74 Also the fate of the survivors and a view of the battlefield when the battle is over can be considered typical elements, or historiographical topoi. In the case of the battle of Cannae, we find three battle episodes: right wing, left wing and back to the right wing. The right wing of the Romans (of which Paullus was commander, but this is not mentioned) is described first in discursive mode (22.47.1). The combat leads to the remarkable flight of the Roman cavalry (Romani equites terga uertunt, 22.47.3). This unfavourable moment in the right wing is presented in a historic present tense, as if, we hypothesize, this was not the end, but rather the beginning of a new complication. A perfect tense would have suggested that the flight was a conclusive step in the narrative, whereas the historic tense suggests that the flight is not the end of the story. The fight continues indeed, but the Romans are unable to overcome the tactics of the Punic army. The battle description of the right wing finishes with an open end: the Romans enter a new fight (integram pugnam ineunt, 22.47.10), but the reader’s expectation is subtly directed towards a negative end through the meaningful classification of the fight as unfair (iniquam): the tired Romans are confronted with fresh and strong troops. The continuation of the story is an example of how a rising narrative arc with an open end may be followed by a change in perspective and a new episode. The left wing description incorporates the story of a Punic ruse (22.48). We find a

72 73 74

Consules cornua tenuerunt, Terentius laeuum, Aemilius dextrum: Gemino Seruilio media pugna tuenda data (22.45.8). See, for instance, Walsh 1963: 197–204 for stock techniques in Livy’s battle descriptions and Hau 2014 for battle topoi in Greek historiography. Livy closely follows the information found in Polybius in the battle scene. See Oakley (this volume, 182–189).

224

van gils and kroon

hint about a story in the Abstract-like first sentence (Punica fraude, 22.48.1), an Orientation in which the main characters are introduced (Quingenti ferme Numidae etc., 22.48.2), and a Complication in the historic present (repente … desiliunt … considere iubentur, 22.48.3). The reader knows it is a story about fraud, so he expects a surprise attack any moment, and this expectation is fulfilled (adoriuntur, 22.48.4) in a Peak. In the third and last battle episode (22.49), we return to the right wing (parte altera pugnae), where Paullus sits heavily wounded. This section contains much direct speech and the rhythm of the narrative accordingly slows down to scenic level. The narrative of the dying Paullus contains a rising arc with Lentulus’ approach to Paullus and his offer to help him escape. The Peak of this scene and, we believe, of the Cannae story at large, is formed by Paullus’ elaborate answer to Lentulus in which he refuses Lentulus’ offer, but shows typical Roman values, like uirtus, dignitas and constantia even in the moment of death. After the Peak, Paullus’ death is described matter-of-factly as a Resolution to the story, in the discursive mode. In a next scene, the Romans flee in all directions ( fugiunt, 22.49.13). The historic present fugiunt opens up a window about the outcome of this flight, which is, indeed, immediately elaborated upon in more detail: in the following sentences the facts about the whereabouts of the fleeing Romans and allies are given. Note that in this case, the ‘window’ that was opened is not filled in with a narrative, but with a list of events in discursive mode. At the end of the battle description (22.46–22.49) we find a Coda with the words Haec est pugna Cannensis (22.50.1) and a comparison to another major Roman defeat (at the river Allia around 390BC). (21) Haec est pugna ⟨Cannensis⟩, Alliensi cladi nobilitate par, ceterum ut illis, quae post pugnam accidere, leuior, quia ab hoste est cessatum, sic strage exercitus grauior foediorque. LIV. 22.50.1–22.50.2

Such was the battle of Cannae, a calamity as memorable as that suffered at the Allia, and though less grave in its results—because the enemy failed to follow up his victory—yet for the slaughter of the army even more grievous and disgraceful.

discourse-linguistic strategies in livy’s account of the battle

225

5.8 Aftermath of the Battle (22.50–22.54) The aftermath of the battle is given more space than the battle itself.75 This is remarkable in itself, but the details of this narrative episode are even more noteworthy: the lack of formal leaders and the chaotic situation of the defeated Romans is told with special interest in the deliberations and discussions of the various groups who need to decide on their own what to do. In 22.50, we hear about contrasting proposals in indirect and direct speech and certain Romans, like Sempronius Tuditanus, are a clear exemplum of courageous words and deeds. Hannibal also needs to decide what to do right after his victory and again the various options are presented in direct and indirect speech by Maharbal and Hannibal respectively (22.51). Hannibal’s choice is explicitly evaluated by the narrator as an exemplum of a wrong decision in terms of warfare leadership. Hannibal decides not to proceed immediately to Rome, but to let his soldiers plunder the victims and take the surviving Romans as hostages. The next scene, the Carthaginians revisiting the site of the carnage, starts off with a historic present (insistunt, 22.51.5). The view of the battlefield is terrible even for the enemy ( foedam etiam hostibus spectandam stragem, 22.51.5) and, as if it were through their eyes, Livy gives us a horrific description of this view. The graphic quality of this description makes it one of the most immersive passages of book 22. Note that not the historic present tense, but rather participle and imperfect and pluperfect tenses are used to give the graphic quality to this scene.76 (22) quosdam et iacentis uiuos succisis feminibus poplitibusque inuenerunt, nudantis ceruicem et reliquum sanguinem iubentes haurire; inuenti quidam sunt mersis in effossam terram capitibus, quos sibi ipsos fecisse foueas obruentisque ora superiecta humo interclusisse spiritum apparebat. LIV. 22.51.7–8

Some were discovered lying there alive, with thighs and tendons slashed, baring their necks and throats and bidding their conquerors drain the remnant of their blood. Others were found with their heads buried in holes dug in the ground. They had apparently made these pits for themselves, and heaping the dirt over their faces shut off their breath. 75 76

See also Oakley (this volume, 170) for a discussion of this observation. Walsh 1963: 181 has pointed out the graphic presentation of this ‘rare scene of horror’ in Livy. Another immersive passage is 22.2.5–22.2.10, again with imperfect tenses rather than the historic present.

226

van gils and kroon

The next decision regards the capture of the smaller camp and trade of hostages. With ducit et excludit (22.52.1) the narrator opens up a new discursive window: a great number of quantitative data are provided and the fate of groups and not that of individuals is told. Predicates are in perfect tense, often passives, or are actual present tense forms. An individual note is reserved for a generous lady called Busa who provided shelter, food and clothes to the Romans who were valorous enough to escape to Canusium. Canusium provides the context for a narrative episode about Roman leadership (22.53). First, we hear about a number of military tribunes and other young noble Romans who deliberate about the best course of action; the young Scipio is given the role of good exemplum by convincing the others through words, oaths and threats to defend the Roman Republic. Next, the reader is informed about the situation in Venusia, where consul Varro finds himself still alive. In a hybrid mode of presentation facts are summarized and a lack of narrative tension is combined with typically narrative elements like imperfect tenses and individual agents. As soon as Scipio and the other leader Appius hear about Varro’s presence in Venusia, they send him a message to ask him what he wants them to do. The historic present (mittunt, 22.54.5) opens a narrative window which is quickly ‘filled in’ with Varro’s decision to move to Canusium. The narrator ends his treatment of the surviving Roman troops with a section on the situation in Rome where people are still unaware of the fate of the survivors. With negations and strong personal comments, the narrator is extraordinarily visible at this point in his work.77 Especially noteworthy is his personal evaluation that no other people would have been able to rise again after such a major defeat, comparing this defeat even to the future downfall of Hannibal in Africa. With this comment at the end of 22.54, the narrator brings his readers back to the primary narrative arc of the third decade which was explicitly started in the first paragraph of book 21 (example 1 above). Where the start was a typical Abstract, this is a typical (intermediate) Coda in which evaluations are common and the mode of presentation is clearly discursive. 5.9 Reactions in Rome (22.55–22.61) The impact of the battle of Cannae on the Roman politics and population is described in the last sections (22.55–22.61) of book 22. We learn about the desperation of their personal and political losses, but also their harshness towards 77

Ne has quidem reliquias (22.54.7); Numquam salua urbe (22.54.8); succumbam oneri neque adgrediar narrare (22.54.8); nulla profecto alia gens (22.54.10); compares cladem ad Aegates insulas … aut pugnam aduersam in Africa (22.54.11); nulla ex parte comparandae sunt (22.54.11).

discourse-linguistic strategies in livy’s account of the battle

227

the Roman hostages who are kept prisoner by Hannibal mainly through indirect and direct speeches. The narrator connects the various speeches with discursive sentences in which the difficult situation in Rome is emphatically brought to the fore. Direct speeches in the form of rhetorically well-formed orations, such as delivered by Fabius (22.39), the embassy of prisoners (22.59), and the response to their request by a severe Roman senator (22.60), are a format well suited to argue, by the utterances of involved characters, the key themes. Moreover, orations slow down the narrative rhythm and heighten narrative tension because of their persuasive goal. In the last section of book 22 (22.61) the prisoners’ dilemma is resolved: the harsh verdict is pronounced to the envoys (triste responsum, 22.61.3) that no money will be spent to get the captives back to their families. In two sentences a story is told about an unnamed envoy who deceitfully returned to his home ( fallaci reditu, 22.61.4) but was sent back to Hannibal by a unanimous decision of the Romans: there is no place for fraud in Roman society, even if the alternative is surrendering to the enemy.78 Finally, the historian includes a completely different version of the story about the prisoner’s embassy in indirect speech (est et alia fama, 22.61.5) and after this alternative version he comments on the strangeness of the existence of different versions. The last passage of book 22 contains a second Coda to the story of Cannae.79 It is a sign of Livy’s narrative art that he manages to make us think of this battle as one coherent story (ea clades, 22.61.10) and evaluate the impact or magnitude of it in a typical Coda, even though there has not been one single narrative arc to describe the defeat of the Romans at Cannae. (23) Quanto autem maior ea clades superioribus cladibus fuerit, uel [de] ea res indicio ⟨est, quod fides socio⟩rum, quae ad eam diem firma steterat, tum labare coepit, nulla profecto alia de re, quam quod ⟨de⟩sperauerant de imperio. LIV. 22.61.10

For the rest, how greatly this disaster exceeded those that had gone before is plain from this: the loyalty of the allies, which had held firm until the day of Cannae, now began to waver, assuredly for no other reason than because they had lost all hope of the empire.

78 79

See also Oakley, 180–182 for this passage. We already saw a Coda in Liv. 22.50.

228 6

van gils and kroon

Conclusion

In this chapter, we have investigated the textual strategies and associated linguistic devices in book 22 of Livy’s Ab urbe condita. The results of this analysis strongly suggest that two motivations are behind most presentational choices of the historian: to keep the audience alert and interested, and to highlight specific moral themes. In the first part of this chapter, we have introduced and illustrated with examples from the first half of book 22 a number of narratological and linguistic concepts which can help us to discover Livy’s use of textual strategies in the story about Cannae. In the second part of this chapter, we have analysed the story about Cannae in the second half of book 22, making use of the concepts introduced in part 1. The larger Cannae-story has been divided into seven episodes and for each episode the presence of particular textual strategies has been analysed. The most important textual strategies we have discerned are the announcement of a thematically relevant event (Abstract), the insertion of narrative subarcs (subsequent increase and decrease of narrative tension, along the components Orientation-Complication-Peak-Resolution in terms of Labov), the change of perspective halfway through a narrative arc, the rare use of immersive narrative and the frequent use of a hybrid narrative-discursive mode of presentation. In order to detect the various textual strategies, we have made use of a narratological-linguistic instrument which is summarized in Appendix 1. The linguistic devices we have particularly focused on are tenses, coherence markers, and point of view. Also relevant for our analysis, though less systematically discussed, have been active-passive variation, collective vs. individual agents, narrative rhythm, the presence of emotions or visual details, and lexical references to key themes like fraus, fides, prudentia and temeritas. In the analysis of book 22, the frequent use of the historic present to open up a narrative window (or Complication, in terms of Labov) was particularly revealing. Events presented in the historic present tense are typically wellmotivated actions by one of the main characters, which are expected to have relevant consequences. We have coined the term ‘consequential action’ to describe this type of events. These consequences are often given in the ensuing text. The reader’s expectation linked to a consequential action is, however, not always fulfilled: the narrator may insert a sudden switch in perspective. The tension about relevant consequences remains, but is temporarily put on hold because a new narrative arc is introduced. Another recurring characteristic of Livy’s text is his use of typically narrative features (e.g. temporal connectives, the imperfect tense, references to emo-

discourse-linguistic strategies in livy’s account of the battle

229

tions) in passages which lack a narrative arc. In such passages we find the historian reporting a number of events in a summarizing and evaluative manner and without creating expectations about a particular outcome, as a straightforward narrative would have done. In order to keep the audience attentive, the report is, so to speak, ‘narrativized’. A last observation regards the thematic relevance of (in)direct speech. The insertion of a conversation between two characters, as happens in the scene of Abelux and Bostar (22.22) and of Lentulus and the dying Paullus (22.49), seems to typically mark the Peak in a narrative arc. The words of these characters contain valuable moral lessons for the reader, and seem to point at the very reason why the historian has inserted the story. The Peaks of both stories do not coincide with historically pivotal events, but with exemplary characters showing their moral standard to a fellow character, and indirectly to the reader. All in all, we would conclude that Livy’s decisions as to narrative structuring (where to put peaks, what events to be magnified, and what other events to be played down, mentioned in passing or not referred to at all) may be partly driven by aesthetic considerations, but that, at least in book 22, many of his choices seem to be strategies for taking the audience along a well-considered construction of moral and ideological themes.80

Appendix 1: Prototypical Features of Modes of Presentation in Latin Historiography

point of view progression coherence marking events rhythm quantities nouns semantic agent

80

Discursive mode

Neutral narrating mode

(low/no immersion)

(moderate immersion)

Immersive narrating mode (total immersion)

speaker (historian) argumentative (ceterum, itaque, igitur, ergo) listed quick big numbers possible abstract, collective collective or well-known

narrator narrative (tum, eo die) summarized quick or scenic big numbers possible collective, concrete character or collective

character impressionistic (hic, repente) detailed scenic or slower small numbers concrete character

For a comparable conclusion with respect to Livy’s third decade more generally, see Levene 2010: 31, who refers to Lipovsky 1981.

230

van gils and kroon

(cont.) Discursive mode

Neutral narrating mode

(low/no immersion)

(moderate immersion)

syntactic subject referential chain transitivity tense

evaluative or abstract noun full denotation low transitivity actual present; perfect

character or collective series, light references high transitivity perfect & imperfect; historic present

voice embedded speech

passive no

active indirect discourse

Immersive narrating mode (total immersion) character series, light references high transitivity historic present; imperfect; historic infinitive active (in)direct discourse

Appendix 2: Narrative Elements in the Presentation of the Battle of Cannae (22.34–22.61)81

Presence of narrative arc 34.1 35.3 35.5 36.6 37.1

Varro becomes popular and is elected consul Paullus is put forward as consular colleague The result of the elections and composition of the armies Prodigies frighten the Roman citizens Arrival at the senate of a Sicilian embassy

38.1 38.6

The soldiers take a new oath Commander’s speeches by Varro and Paullus

38.12 40.1 40.5 40.7

Fabius takes the floor and gives a speech Paullus reacts to Fabius The consuls arrive at the Roman camp Hannibal sees and rejoices at the arrival of the consuls and new soldiers The situation of Hannibal is tough Swift Roman victory in chaotic melee Paullus’ holds back the Roman soldiers Hannibal’s view on the Roman victory is positive Hannibal prepares an ambush

40.8 41.1 41.3 41.4 41.6

81

Arc type iv Arc type ii No arc: discursive No arc: hybrid No arc: hybrid, incl. (in)direct speech No arc: hybrid No arc: discursive, incl. ind. speech Arc type iv: dir. speech Arc type v: ind. speech Arc type iv No arc: hybrid No arc: hybrid No arc: discursive No arc: hybrid Arc type v, incl. ind. speech Arc type iv

See pp. 210 for the five types of narrative patterns we distinguish and the difference between a discursive and narrative-discursive (hybrid) presentation of events.

discourse-linguistic strategies in livy’s account of the battle

231

(cont.) Presence of narrative arc 42.1 42.4 42.7 42.10– 42.12 43.1 43.7 43.10 44.1 44.4 44.5 45.1 45.5 46.1 47.1 48.1 49.1 50.1 50.4 51.1 51.5 52.1 53.1 54.1 54.5 54.7 55.1 58.1 59.1 60.1 61.1 61.4 61.5 61.10

Romans discover the empty Carthaginian camp Paullus tries to check the soldiers and sends Statilius Statilius’ report inflames rather than restrains the soldiers and worsens the conflict between Varro and Paullus Escaped slaves reveal ambush Hannibal returns, but decides to leave for southern Italy Statilius confirms Hannibal’s departure; Romans follow Hannibal puts his camp in a favourable place The consuls put up their camp Hannibal provokes the enemy The consuls disagree about the best course of action

Arc type v Arc type v, incl. ind. speech Arc type iv Arc type iii

Arc type iv No arc: discursive Arc type v Arc type v Arc type v No arc: discursive, incl. indir. speech Hannibal lets Numidians attack some Romans Arc type v Varro decides to prepare for battle Arc type v Description of Hannibal’s army Arc type iv Start of the battle and right wing fighting No arc: hybrid Left wing: Punic ruse Arc type i Right wing again: dying Paullus Arc type i, incl. dir. speech Haec est pugna Cannensis Coda to main story Deliberations in the Roman camps and outbreak to Canusium Arc type i, incl. (in)d. speech Deliberations in Hannibal’s camp Arc type i, incl. dir. speech The Carthaginians view the battle field and collect spolia Arc type iv Hannibal easily conquers the Roman camps No arc: discursive Deliberations in Canusium Arc type iv, incl. (in)dir. speech In Venusia Varro and many soldiers are welcomed generously No arc: hybrid The soldiers of Canusium and Venusia are united in Canusium Arc type iii The situation in Rome No arc: discursive Deliberations in the senate No arc: discursive (incl ind. speech) Hannibal’s treatment of the prisoners No arc: hybrid Embassy of prisoners plead their case in Rome Direct speech Reaction of the masses and response by Manlius Torquatus Hybrid + direct speech The reaction to the harsh verdict No arc: hybrid The runaway prisoner is brought back to Hannibal Arc type ii Alternative facts (alia fama) Indirect speech Coda to major defeat No arc: discursive

232

van gils and kroon

Bibliography Adema, S.M. Speech and Thought in Latin War Narratives. Words of Warriors. Amsterdam Studies in Classical Philology, 24. (Leiden 2017). Adema, S.M., Discourse Modes and Bases. A Study of the Use of Tenses in Vergil’s Aeneid, dissertation (Amsterdam 2008). Allan, R.J., Jong, I.J.F. de & Jonge, C.C. de, ‘Homerus’ narratieve stijl: enargeia en immersion’, Lampas 47 (2014) 202–223. Baroni, R., ‘Tellability’, in Hühn, Peter et al. (eds.), The living handbook of narratology (Hamburg 2013), accessed 11 March 2018, http://wikis.sub.uni‑hamburg.de/lhn/ index.php/Tellability. Bjørner, T., Magnusson, A. & Nielsen, R.P., ‘How to Describe and Measure Obstacles of Narrative Immersion in a Film? The Wheel of Immersion as a Framework’, Nordicom Review 37.1 (2016) 101–117. Bruckmann, H., Die römischen Niederlagen im Geschichtswerk des T. Livius, dissertation (Münster 1936). Burck, E., Die Erzählkunst des T. Livius ([1934] Berlin 1964). Burck, E., Einführung in die dritte Dekade des Livius ([1950] Heidelberg 1962). Chaplin, J.D., Livy’s Exemplary History (Oxford 2000). Chausserie-Laprée, J.P., L’expression narrative chez les historiens latins. Histoire d’un style (Paris 1969). Conte, G.B., Latin Literature: A History ([1987] Baltimore 1994). Cuddon, J.A., ‘emplotment’, in [ed(s)], A Dictionary of Literary Terms and Literary Theory, Blackwell Reference online (Oxford 2013), accessed [26-1-2018], [www .blackwellreference.com]. Dorey, T.A., Titi Livi: ab urbe condita XXI–XXII (Leipzig 1971). Feldherr, A., Spectacle and Society in Livy’s History (Berkeley 1998). Foster, B.O., Livy V: Books XXI–XXII (London 1929). Fuhrmann, M., ‘Narrative Techniken im Dienste der Geschichtsschreibung (Livius, Buch 21–22). Eine Skizze’, in E. Lefèvre & E. Olshausen (eds.), Livius. Werk und Rezeption. Festschrift für Erich Burck (München 1983) 19–29. Gils, L.W. van & Kroon, C.H.M., (forthc.) Involving the audience. A cognitive-linguistic approach to the Historic Present in Livy and Tacitus. Grethlein, J. & Huitink, L., ‘Homer’s Vividness: An Enactive Approach’, JHS 137 (2017) 67–91. Hau, L.I., ‘Stock Situations, Topoi and the Greekness of Greek Historiography’, in D. Cairns & R. Scodel (eds.), Defining Greek Narrative (Edinburgh 2014) 241–259. Herman, D. (ed.), The Cambridge Companion to Narrative (Cambridge 2007). Krafft, P., ‘Hannibals Perücken. Motivik und Erzählstruktur von Livius 22,1’, RhM 150 (2007) 67–85.

discourse-linguistic strategies in livy’s account of the battle

233

Kroon, C.H.M. & Risselada, R., ‘Iamque dies infanda aderat. Time management, historiografie en de geschiedenis van een narratieve techniek’, Lampas 37.3 (2004) 191–202. Kroon, C.H.M. & Risselada, R., ‘Phasality, polarity, focality; a feature analysis of the Latin particle iam’, Belgian Journal of Linguistics 16 (2002) 63–78. Labov, W., Language in the Inner City (Philadelphia 1972). Laird, A., ‘The Rhetoric of Roman Historiography’, in A. Feldherr (ed.), The Cambridge Companion to the Roman Historians (Cambridge 2009) 197–213. Levene, D.S., Livy on the Hannibalic War (Oxford 2010). Lipovsky, J., A historiographical study of Livy, books VI–X (New York 1981). Luce, J.T., Livy. The Composition of his History (Princeton 1977). Marincola, J., Authority and Tradition in Ancient Historiography (Cambridge 1997). McDonald, A.H. The Style of Livy. The Journal of Roman Studies 47.1–2 (1957) 155–172. Ochs, E., ‘Narrative Lessons’, in A. Duranti (ed.), A Companion to Linguistic Anthropology (Oxford 2005) 269–289. Pausch, D., Livius und der Leser. Narrative Strukturen in ab urbe condita (München 2011). Pinkster, H., Latin Syntax and Semantics (London 1990). Pinkster, H., The Oxford Latin Syntax. Volume 1, The Simple Clause (Oxford 2015). Rossi, A., ‘Parallel lives: Hannibal and Scipio in Livy’s third decade’, TAPhA 134.2 (2004) 359–381. Sumner, G.V. Elections at Rome in 217B.C. Phoenix 29 (1975) 250–259 Toolan, M., Making sense of narrative text (London 2016). Walsh, P.G., Livy. His Historical Aims and Methods ([1961] Cambridge 1989). White, H., Metahistory: The Historical Imagination in Nineteenth-century Europe (Baltimore 1973). Woodman, A.J., Rhetoric in Classical Historiography: Four Studies (London 1988).

chapter 9

Who Knows What Will Happen Next? Livy’s fraus Punica from a Literary Point of View Dennis Pausch

1

Introduction or: Dear Reader, Beware of Unreliable Enemies!*

Livy is usually studied—and read, I suppose, at least by classicists—in a diachronic, one could even say a ‘diatextual’ way, that is with the focus on how his version of a particular event interrelates with the versions of his predecessors, whether their texts have actually been preserved or have merely been reconstructed (the second perhaps being the more frequent case). The obvious reason for this is, of course, that we hope to improve our knowledge of ‘what really happened’ by comparing the various versions we have. I am far from belittling this historical approach which brought forth important results and remains highly useful. My own take on Livy, however, is a more literary one and tries to read his text in a ‘synchronic’ and, as it were, continuous way, focussing on how our reading of a particular event is affected by our knowledge of similar episodes related earlier in Ab Urbe Condita (or even later, if we presume a repeated reading). To be sure, approaches like these had been undertaken before, not least in the classical works of the so-called ‘literary school’ written by Erich Burck, Patrick G. Walsh and James T. Luce around the middle of the 20th century.12 Yet, many of them dealt primarily with Livy’s assumed pro-Roman bias or other problems related to history, whereas a fuller analysis of Livy’s narrative technique by genuine literary categories has only just begun to be developed during the last decades.2 Apart from the first pentad which usually receives the lion’s

* This paper profited very much from the remarks made by my respondent Christoph Pieper (Leiden) and by the other participants in the discussion. I would like to thank them as well as the organisers most sincerely for an inspiring conference and a most welcoming atmosphere. Last, but certainly not least, I feel deeply obliged to Nicolas Wiater (St Andrews) for doing his very best to make my English more readable. 1 See Burck [1934] 1964; Walsh 1961; and Luce 1977. 2 For studies dealing with Ab Urbe Condita as a whole see Jaeger 1997; Feldherr 1998; Chaplin 2000; and Pausch 2011.

© koninklijke brill nv, leiden, 2019 | doi:10.1163/9789004383340_010

livy’s fraus punica from a literary point of view

235

share of attention and research,3 it is above all the so-called third decade that suggests itself for narrative analysis, since Rome’s war against Hannibal constitutes a coherent and well-structured unit in which the way the story is told is additionally made explicit by the narrator in several instances.4 My aim in this chapter is to extend the focus on Cannae in this volume by reading Livy’s account alongside two of the earlier Roman defeats, at the Trebia (218 BC) and at Lake Trasimene (217BC), in order to explore how these descriptions influence the reaction of a reader to ‘our’ battle. I am particularly interested in the question of how information about the further course of the events is given to the historical figures, on the one hand, and to the reader, on the other: are they presented as part of a stratagem the success of which remains unclear at the time or are they narrated with hindsight, as already well-known historical facts? This will, I hope, lead to a better understanding of Livy’s narrative technique in battle descriptions in general, but also, perhaps, cast some new light on the much-debated fraus Punica as well as the question of how hostile Livy’s portrait of Hannibal really is in the third decade. One methodological problem needs to be discussed in advance, at least briefly: how much previous knowledge of the historical developments of the second war against Carthage should we assume for a reader at the end of the first century BC?5 Even if we narrow this down to a reader from Rome or romanized Italy with a common upper-class education, thus supposing that Livy imagined that he was in a dialogue with a reader very similar to himself, it is difficult to guess how much detailed knowledge of, for example, the battle of Cannae we can presuppose. That Livy takes the disastrous outcome for granted can be inferred from a number of previews given both implicitly and explicitly.6 The same, however, does not seem to apply to the full particulars of neither the course of events of Hannibal’s campaign in Italy nor of his major encounters with his Roman opponents. Whether or not Livy’s assumptions about previous historical knowledge proved to be true for his actual readers, we will never know. But since psycholinguistic studies have shown that temporary uncertainty about the outcome of a story is induced even in readers who in principle 3 See e.g. Miles 1995; Vasaly 2015; and the relevant contributions to Mineo 2015: 243–326. 4 See esp. Burck [1950] 1962; Fuhrmann 1983; and Levene 2010. 5 The time of composition of the third decade is hard to determine exactly, but if Liv. 28.12.12 presupposes the end of the Cantabrian wars in 19BC (and not only the end of the first campaign of 27–25BC: see Händl-Sagawe 1995: 9), that gives us at least a terminus post quem for the later books which is more or less congruent with Livy’s average working speed as calculated by Mensching 1986: 574. 6 Cf. esp. Liv. 43.9: ad nobilitandas clade Romana Cannas urgente fato (for more on this see below).

236

pausch

already know its end,7 this subtlety probably does not matter too much. In any case, Livy does not seem to aim at a kind of suspense that is based on lack of knowledge of the final outcome, but on keeping the readers wondering how this final outcome is actually realized (“Was-Spannung” as opposed to “WieSpannung”, as German terminology has it).8 With this in mind, we will now follow Hannibal’s southward march. We will have a look at how his two other major victories over Roman forces are presented to Livy’s Roman readers and then turn to Cannae. Seen in this light, it becomes even more apparent that this battle is not only the most important military event in these books, but also the focal point of a wide-ranging literary pattern that includes a surprising variation at its end and thus should be recognized and appreciated by an attentive reader. But before embarking on our journey, however, a quick glance at how the Carthaginian general is introduced by the narrator at the end of the proem to the third decade is in order. This portrait of Hannibal is reminiscent of Sallust’s characterisation of Catiline at a similar point of the narrative in his monograph.9 Just like the Roman conspirator, as the reader has been informed in advance, Hannibal combines highly admirable (especially military) virtues with their very opposite: has tantas uiri uirtutes ingentia uitia aequabant: inhumana crudelitas, perfidia plus quam Punica, nihil ueri, nihil sancti, nullus deum metus, nullum ius iurandum, nulla religio. Liv. 21.4.9

These admirable qualities of the man were equalled by his monstrous vices: his cruelty was inhuman, his perfidy worse than Punic; he had no regard for truth, and none for sanctity, no fear of the gods, no reverence for an oath, no religious scruple!10 As has been noted before, this pointedly sharp portrait with its rather exaggerated rhetoric does not correspond entirely to the much more nuanced way

7 8 9 10

Cf. Gerrig 1989; for an adaptation to the Odyssey see Schmitz 1994. On this differentiation reaching back to Berthold Brecht see e.g. Anz 2003: 465. Cf. Liv. 21.4.3–21.4.10 and Sall. Cat. 5.1–5.8 see e.g. Walsh 1961: 104; and Levene 2010: 99–104, who also points to the portrait of Jugurtha in Sallust’s other monograph. Here as in the following the Latin text is basically that of Dorey’s Teubneriana (1971), the translation that of Foster’s Loeb (1929), although slight modifications have been necessary.

livy’s fraus punica from a literary point of view

237

in which Hannibal’s achievements are presented in the following books.11 This makes it all the more apparent that this passage is designed to influence the reaction of the reader to the subsequent narration.12 One of its effects—the one I am interested in here—is that the reader is warned about Hannibal’s stratagems13 right from the start, whereas the historical characters will have to learn their lesson the hard way. Thus the reader knows not only which aspects of the following account to pay special attention to, but will also be able to see—or at least guess—more than the historical figures are able to. This is already a good strategy to create and maintain suspense throughout the story, but Livy does even more to enhance the difference between historical actors and his readers, as his account of the battle at the Trebia demonstrates.

2

The Ambush at the Trebia: Being Blind to Stratagems

After crossing the Alps and prevailing against the Romans in a first encounter at the Ticinus river, Hannibal had already taken position on the southern bank of the Po at the end of 218 BC. In the meantime, the second consul, Sempronius Longus, arrived at the scene and joined his forces with P. Cornelius Scipio, who had been wounded at the Ticinus river and had only just been rescued by his son with the same name (an opportunity readily taken by the narrator to insert a short prolepsis of the end of the war at Zama).14 Encouraged by a successful equestrian skirmish,15 Sempronius is eager to fight as soon as possible, while

11

12 13

14

15

See e.g. Walsh 1961: 104; Christ 1968: 470; Levene 2010: 228–236; and Hoyos 2015: esp. 370– 373. On the origins of the negative picture of Hannibal in the Roman ‘propaganda’ during the war see Stepper 2006: esp. 404; and Chassignet 2009. On this effect with regard to the portrait as a whole see Cipriani 1987. On the moral evaluation of stratagems by Roman readers see Levene 2010: 228: “The ‘Hannibalic model’, in this context, is not a clear-cut or uncontroversially bad one. The Romans, like the Greeks before them, had an ambivalent reaction to the use of strategic devices in battle rather than a straight-forward head-on conflict. In many contexts one finds the two placed in sharp antithesis, and very often with the implication that the use of stratagems is deceptive and morally inferior to fighting directly without them. But there is also a regular sense that deliberation and wisdom in fighting is superior to merely rushing in head-long, and terms like consilium, which may be used for deceptive stratagems, tend even in those contexts to have positive rather than negative overtones.” Cf. Liv. 21.46.8: hic erit iuuenis, penes quem perfecti huiusque belli laus est, Africanus ob egregiam uictoriam de Hannibale Poenisque appellatus; see e.g. Levene 2010: 14; on Scipio in the third decade in general Rossi 2004. Cf. Liv. 21.52.9–21.52.11.

238

pausch

his colleague is more cautious and strongly disagrees.16 We thus have the same constellation here as later before the battle at Cannae (even though the dispute between the consuls at the Trebia is actually more likely to have been modelled on the one at Cannae as the more significant event in historical terms).17 Furthermore, in both episodes the narrator exploits this constellation in the same way to create suspense by presenting two contrary options and prompting the reader to hope—in vain—that the Romans will choose the right one. The reader’s sensation of an imminent danger rising against the Romans becomes even more intense, with the narrator’s detailed account of the actions of the Carthaginians which is strongly focalized through Hannibal. Hannibal learns about the dispute between his opponents and immediately realizes the advantage of this situation:18 cum ob haec taliaque speraret propinquum certamen et facere, si cessaretur, cuperet speculatoresque Galli, ad ea exploranda, quae uellet, tutiores quia in utrisque castris militabant, paratos pugnae esse Romanos rettulissent, locum insidiis circumspectare Poenus coepit. Liv. 21.53.11

For these and similar reasons he hoped that a battle would soon be fought, and was eager, should there be any hesitation, to force it on. And so, when his Gallic scouts—who were safer for gathering the information that he wanted because there were men of that nation in both camps— had reported that the Romans were prepared to fight, the Phoenician began to look about for a place in which to lay an ambush. Ambush, insidiae, is one of the keywords used by Livy throughout the following two books.19 But as this is the first instance that the Romans are confronted with Hannibal’s artifice, the narrator provides some more detail about how such stratagems are prepared:

16 17

18 19

Cf. Liv. 21.53.1–21.53.7. On the motif of the ‘contrasting consuls’ in the books 21 and 22 see e.g. Fuhrmann 1983: 27–28, esp. 27: “Auf römischer Seite operiert fast immer ein Paar von Befehlshabern, die ebenso regelmäßig miteinander im Streit liegen, da sie sich nach Charakter und Haltung auf das Bestimmteste voneinander unterscheiden.”; Will 1983; and Porod 1989: 210–218. Cf. Liv. 21.53.7–21.54.3. On historical authenticity and possible sources see Händl-Sagawe 1995: 331–333. On the range of meanings for insidiae see Wheeler 1988: 84–87, esp. 84: “In truth, insidiae is much more a generic term for stratagem than often thought.”

livy’s fraus punica from a literary point of view

239

erat in medio riuus praealtis utrimque clausus ripis et circa obsitus palustribus herbis et, quibus inculta ferme uestiuntur, uirgultis uepribusque. quem ubi equites quoque tegendo satis latebrosum locum circumuectus ipse oculis perlustrauit, ‘hic erit locus’ Magoni fratri ait ‘quem teneas. delige centenos uiros ex omni pedite atque equite cum quibus ad me uigilia prima uenias; nunc corpora curare tempus est.’ ita praetorium missum. mox cum delectis Mago aderat. ‘robora uirorum cerno’ inquit Hannibal ‘sed uti numero etiam, non animis modo ualeatis, singulis uobis nouenos ex turmis manipulisque uestri similes eligite. Mago locum monstrabit, quem insideatis; hostem caecum ad has belli artes habetis.’ Liv. 21.54.1–21.54.3

Between the two camps was a water-course, shut in by very high banks on either side and overgrown all round with marsh-grass and the underbrush and brambles with which uncultivated land is usually clothed. When Hannibal, riding over the ground himself, saw that this place afforded sufficient cover even for cavalry, he said to his brother Mago, ‘This will be the place for you to hold. Choose out a hundred men from all the infantry and a hundred from the cavalry, and come with them to my quarters at the first watch. It is time now to sup and rest.’ With that he broke up the council. In a little while Mago presented himself with his picked men. ‘I see the stoutest of my men,’ said Hannibal, ‘but that your numbers too may be strong to match your bravery, choose, each of you, from the squadrons and the maniples, nine others like yourselves. Mago will point out to you the spot where you are to lie in ambush; you have an enemy who is blind to these stratagems.’20 By way of the metaphor of seeing and not seeing—hostem caecum ad has belli artes habetis—, Livy has Hannibal point out the decisive aspect of the situation: the Carthaginians, and the reader with them, are now aware of the ambush, while the Romans are not, and they will not learn about it until it is too late.21 A 20 21

On historical authenticity and possible sources see Händl-Sagawe 1995: 331–333. For another reading of this passage with regard to suspense see Fuhrmann 1982: 25: “Die Darstellung der Schlacht an der Trebia enthält ein eindrucksvolles Beispiel perspektivischer Berichterstattung. Livius teilt zwar mit, daß Hannibal den Römern eine Falle stellt (21,54,1–5), er verschweigt indes, wie diese funktionieren soll. So ahnt der Leser das drohende Unheil, ohne zu wissen, wie es sich vollziehen wird (er ist also in Spannung versetzt); Livius stellt sodann die Schlacht selber aus der beschränkten Perspektive des unterlegenen Opfers dar, und das Funktionieren der Falle wird in dem Augenblick deutlich, in dem sie zuschlägt (21,55,9).”

240

pausch

similar effect is achieved in Polybius’ version of the same incident.22 In the third book of the Histories the corresponding observation is not put into the mouth of one of the characters but made by the narrator himself. As Livy’s technique is surely more refined and effective in literary terms, this passage not only shows that Livy knew Polybius’ account of the Hannibalic war,23 but also that he tried to surpass his predecessor not least by creating a more lively and engaging narrative.24 Thus Livy gives in his description a kind of textbook example not only of how an ambush is prepared, but also of how it ought to be narrated in order to have maximum effect on the reader. This is all the more remarkable, since the ambush turns out not to be the crucial factor in the Roman defeat on that cold winter day. In fact, it plays a rather minor role given the large number of mistakes attributed to Sempronius by the Roman tradition,25 and is dealt with in no more than twenty words in the subsequent account of the battle itself.26

3

Too Much Fog at Lake Trasimene: Lessons Still Not Learned

The preceding observations have shown that it is highly probable that the emphasis put on Hannibal’s cleverness in the above episode is designed to alert the reader to dangers yet to come in the following parts of the narrative. Such premonition will be fulfilled early enough indeed. But in contrast to Roman readers in Augustan times, their fellow countrymen in the late third century BC continue to be blind to these stratagems, even after Hannibal’s use of several disguises during the winter27 and his surprise attack at Placentia in

22 23

24

25

26 27

Cf. Pol. 3.71.2–4. According to the older communis opinio strongly influenced by the paradigm of sourcecriticism, Livy had not used Polybius directly for books 21 and 22: see e.g. Tränkle 1977: 193–241, and Briscoe 1993: 40. In contrast to the traditional view, David Levene was able to show that the differences between their accounts are deliberate variations and corrections on the part of Livy, for whom the Histories acted not only as historical source, but as literary model as well: see Levene 2010: 135–155, using the crossing of the Alps as his example. Cf. Liv. 21.54.6–21.56.9; see also Polyb. 3.72–3.74. On the exculpatory tendency in the Roman tradition about the Trebia see e.g. Bruckmann 1936: 59–65; and Burck [1950] 1962: 75–77; but now also Levene 2010: 263–267, who convincingly argues that none of the exculpatory elements is depicted as a decisive factor by Livy. Cf. Liv. 21.55.9: et Mago Numidaeque, simul latebras eorum improuida praeterlata acies est, exorti ab tergo ingentem tumultum ac terrorem fecere. Cf. Liv. 22.1.1–22.11.4; see Burck [1950] 1962: 79; and Krafft 2007: 67–85.

livy’s fraus punica from a literary point of view

241

early spring.28 A difference in knowledge between characters and readers, as it is highlighted by the Livian narrator here, is one of the most efficient techniques to enhance the engagement of any audience with a text, play or movie. Another prominent strategy aiming at the emotional involvement of the reader is employed here as well, as the imminent danger is invoked by the large number of bad omens in the current29 as well as the previous year.30 Contrary to Livy’s usual practice, both lists of prodigia are reported at the end of the year 218 BC and thus in close proximity to each other, presumably in order to double their frightening impact.31 They concern, as is usual, Rome in general, but also the newly elected consul Gaius Flaminius specifically, who will be in charge of the legions in Italy. His inaugural sacrifice in particular fails entirely and ends up with the sacrificial animal tearing itself away from the servants and spattering its blood upon the bystanders: id a plerisque in omen magni terroris acceptum (‘most people took it as a sign of great horror’).32 All this leads to a forceful foreshadowing of the next disaster just about to happen and prompts the reader to be alert to Hannibal’s next move. This contrasts effectively with Flaminius’ ill-suited behavior during his pursuit of the Carthaginian army through Umbria, as the reader realizes soon afterwards: first, from Hannibal’s point of view, who had recently lost half his eyesight33 but apparently sees more than his Roman opponents even as a semicaecus: et iam peruenerant [sc. the Carthaginians] ad loca nata insidiis, ubi maxime montes Cortonenses in Trasumennus subit. uia tantum interest perangusta, uelut ad ⟨id⟩ ipsum de industria relicto spatio; deinde paulo latior patescit campus; inde colles adsurgunt. ibi castra in aperto locat, ubi ipse cum Afris modo Hispanisque consideret; Baliares ceteramque leuem armaturam post montes circumducit; equites ad ipsas fauces saltus tumulis apte tegentibus locat, ut, ubi intrassent Romani, obiecto equitatu clausa omnia lacu ac montibus essent. Liv. 22.4.2–22.4.4

28 29 30 31 32 33

Cf. Liv. 21.59. Cf. Liv. 22.1.8–22.1.20. Cf. Liv. 21.62, esp. 1: Romae aut circa urbem multa ea hieme prodigia facta aut, quod euenire solet motis semel in religionem animis, multa nuntiata et temere credita sunt. See Fuhrmann 1983: 25–28; Levene 1993: 38–43; Engels 2007: 189; and Levene 2010: 37–38. Cf. Liv. 21.63.13–21.63.14. On how Flaminius was made a ‘scapegoat’ by Roman tradition see e.g. Meißner 2000. Cf. Liv. 22.2.10–22.2.11.

242

pausch

And now they had reached a spot designed by nature for an ambuscade, where Trasumennus approaches closest to the mountains of Cortona. Between them is nothing but a very narrow track, as though room had been left expressly for this purpose; the ground then widens into a little plain; beyond this the hills rise steeply. At this point he laid out a camp in the open, for himself and his African and Spanish troops only; the Baliares and the rest of his light armed forces he led round behind the mountains; the cavalry he stationed near the entrance to the defile, where some hillocks formed a convenient screen for them, so that when the Romans should have entered the pass, they might block the road, and trap the entire army between the lake and the mountains. After these preparations, the focalization changes once more, pointedly, to the frustratingly incautious Roman consul: Flaminius cum pridie solis occasu ad lacum peruenisset, inexplorato postero die uixdum satis certa luce angustiis superatis, postquam in patentiorem campum pandi agmen coepit, id tantum hostium quod ex adverso erat, conspexit: ab tergo ac super caput ⟨haud⟩ deceptae insidiae. Poenus ubi, id quod petierat, clausum lacu ac montibus et circumfusum suis copiis habuit hostem, signum omnibus dat simul inuadendi. qui ubi, qua cuique proximum fuit, decucurrerunt, eo magis Romanis subita atque improuisa res fuit, quod orta ex lacu nebula campo quam montibus densior sederat agminaque hostium ex pluribus collibus ipsa inter se satis conspecta eoque magis pariter decucurrerant. Romanus clamore prius undique orto quam satis cerneret se circumuentum esse sensit, et ante in frontem lateraque pugnari coeptum est, quam satis instrueretur acies aut expediri arma stringique gladii possent. Liv. 22.4.5–22.4.7

Flaminius had reached the lake at sunset; the next morning, without reconnoitring, and scarcely waiting for broad daylight, he passed through the defile. As the column began to spread out on the more open ground, they caught sight of those enemies only who were right in front of them; the ambush in their rear and that above them they failed to perceive. The Phoenician had now gained his object, the Romans were hemmed in between the mountains and the lake and their escape cut off by his own troops, when he made the signal for all his forces to attack at once. As they charged down, each at the nearest point, their onset was all the more sudden and unforeseen inasmuch as the mist from the lake lay less thickly on

livy’s fraus punica from a literary point of view

243

the heights than on the plain, and the attacking columns had been clearly visible to one another from the various hills and had therefore delivered their charge at more nearly the same instant. From the shouting that arose on every side the Romans learned, before they could clearly see, that they were surrounded; and they were already engaged on their front and flank before they could properly form up or get out their arms and draw their swords. Whereas Hannibal’s ambush played only a minor part in the description and justification of the Roman defeat at the Trebia, Livy’s account of the battle of Lake Trasimene, on the contrary, places special emphasis on the idea of an enormous trap succeeding perfectly. This impression is highlighted, among other things, by the prominence of the theme of seeing and not seeing in the narration.34 The Romans’ lack of sight caused by the darkness or the mist rising from the water is mentioned several times and is effectively juxtaposed with the spectacular description of the return of the sunlight after the end of the fighting.35 The notorious fog, however, has a primarily literary function and adds atmosphere to the picture. This time, the Romans are not defeated by nature but by their own lack of foresight as embodied, as it were, in the traditional scapegoat figure Flaminius.36 Yet even after the appointment of Fabius Maximus as dictator, they continue to be inferior to Hannibal. Despite the fact that the famous ‘Cunctator’ does indeed exercise much more care, not only Fabius Maximus himself (in the Callicula mountains)37 but also his magister equitum Minucius (in Apulia)38 are outwitted again by the Carthaginian general.

34

35

36

37 38

The contrast between the historical figures who are not able to see, but only hear what is happening around them (cf. esp. Liv. 22.5.3: et erat in tanta caligine maior usus aurium quam oculorum), and the narrator who is able to give a clear account of the battle even from a bird’s-eye view may contain a ‘metaliterary’ comment on the limits of autopsy held in high regard by Polybius and other ancient writers. Cf. Liv. 22.6.9: inclinata denique re, cum incalescente sole dispulsa nebula aperuisset diem, tum liquida iam luce montes campique perditas res stratamque ostendere foede Romanam aciem. On the exculpatory tendency see e.g. Bruckmann 1936: 65–70; Burck [1950] 1962: 80–82; and Levene 2010: 267–270, who argues that Livy puts special emphasis on the death of the consul as the main reason for the defeat. Cf. Liv. 22.16.4–22.17.7. Cf. Liv. 22.28.1–22.29.6.

244

pausch

4.1 The Empty Camp on the Way to Cannae: a Pattern Broken? Nevertheless, as time passes, the Romans finally seem to learn their lesson—or at least, those whom Livy regards as the better part, namely, Fabius Maximus and his senatorial friends. Suspense, however, is maintained once more (apart from more omina reported at the end of the year)39 by the juxtaposition of conflicting opinions held by the consuls of 216 BC: this time it is Terentius Varro who has taken over the part of both Sempronius and Flaminius and wants to bring about a speedy decision through an open battle, whereas his opponent Aemilius Paullus is the true heir of the Cunctator’s cautiousness.40 To make the constellation even more evident, Livy employs the literary technique of character speech and puts a long address to Aemilius Paullus into the mouth of Fabius himself.41 These 500 plus words of oratio recta contain not only a preview of the next major defeat from the perspective of the historical figures,42 but also end with a powerful admonition not to continue to be blind to Hannibal’s artfulness: omnia non properanti clara certaque erunt; festinatio improuida est et caeca (“All things will be clear and definite to one who does not hurry. Haste is improvident and blind.”).43 Although this warning is addressed mainly to the reader (Fabius’ only listener is Aemilius Paullus who required no further warning), it does have some beneficial effect at the level of the narrative in that the previous pattern of Roman behaviour is interrupted in the end. After losing yet another equestrian skirmish, Hannibal seizes the opportunity to set another trap for the newly encouraged Romans. Once more we do not only learn about this danger immediately through internal focalization of the Carthaginian general: itaque locum et tempus insidiis aptum se habere ratus (“Thus he thought place and time suitable for a stratagem”);44 but we also follow his preparations in detail (the camp 39 40 41

42

43 44

Cf. Liv. 22.36.6–22.36.9. Cf. Liv. 22.38.9–22.38.12; see e.g. Burck [1950] 1962: 93: “Der Gegensatz der beiden Konsuln wirkt als beträchtliches Spannungsmoment für den weiteren Aufbau des Berichtes.” Cf. Liv. 22.39.1–22.39.22; see Burck [1950] 1962: 94: “Damit kommt dieser Rede eine erhöhte kompositionelle Bedeutung zu, löst sie doch im Leser eine Spannung aus, die noch über die bange Erwartung der bevorstehenden Entscheidungsschlacht auf die folgenden Kriegsjahre, ja fast bis aufs Kriegsende hinweist.”; and Treptow 1964: 139–147. Cf. Liv. 22.39.8: atqui si hic [sc. Terentius Varro], quod facturum se denuntiat, extemplo pugnauerit, aut ego rem militarem, belli hoc genus, hostem hunc ignoro, aut nobilior alius Transumenno locus cladibus erit. (“And yet, if he [Terentius Varro] fights at once, as he declares that he intends to do, either I know nothing of military science, of the nature of this war, and of our enemy, or another place will be more notorious than Trasumennus for our overthrow.”) Cf. Liv. 22.39.22. Cf. Liv. 22.41.1–22.41.6, esp. 6.

livy’s fraus punica from a literary point of view

245

is abandoned, but enough booty is put on display to lure the enemy inside, fires are left burning and the soldiers are hidden in the surrounding hills).45 As could be expected, this leads to another dispute between the consuls in which Varro seems to eventually prevail.46 At the very last moment, however, Paullus succeeds in holding the Roman soldiers back by pointing to unfavourable auspicia and the testimony of two slaves recently escaped from the Carthaginians.47 This surprising deviation from the usual pattern is further emphasized by a change of perspective to Hannibal who sees that his ambush failed48 and is thus forced to give up his position and retreat towards Cannae. If that did prompt any reader to hope that things would turn out favourably for the Romans after all, he would have been disillusioned rather soon by the preview implied in the narrator’s comment a few lines later on the Romans’ decision to embark on an immediate pursuit of Hannibal which is made against the advice of Aemilius Paullus, of course. This, Livy remarks, happens ad nobilitandas clade Romana Cannas urgente fato (to make Cannae famous through a Roman defeat by the will of destiny).49 At this point, I think, it is evident that although there is almost no suspense directed towards the outcome of the battle itself, there is significant tension regarding the way this outcome will come into being and the role Hannibal’s stratagems will play in this process. 4.2 In the Middle of all Malice: on Being Circumvented Once More Now that we have reached the plains of Cannae, what should we expect to happen? And how will we be instructed by the narrator on this occasion? Will there be another speech by Hannibal to one of his generals? Or will Paullus once more see through the schemes of his opponent, and will we be allowed to survey the battle ground through his eyes? As is well known, Livy offers us none of these options. After another debate among the consuls (leading to a final ‘what-if situation’ prompting the reader to imagine what would have happened if Paullus had got the upper hand),50 the narrator leads us medias in res by way of a detailed account of the formations of the Roman and Carthaginian forces, including a description of the equipment of the Carthaginian soldiers 45 46 47 48 49

50

Cf. Liv. 22.41.7–22.41.9. Cf. Liv. 22.41.1–22.41.7. Cf. Liv. 22.41.8–22.41.12. Cf. Liv. 22.43.1: postquam postquam motos magis inconsulte Romanos quam ad ultimum temere euectos uidit, nequiquam detecta fraude in castra rediit. Cf. Liv. 22.43.9: cum utriusque consulis eadem quae ante semper fuisset sententia, ceterum Varroni fere omnes, Paulo nemo praeter Seruilium, prioris anni consulem, adsentiretur, [ex] maioris partis sententia ad nobilitandas clade Romana Cannas urgente fato profecti sunt. Cf. Liv. 22.44.1–22.45.4.

246

pausch

and a report on the weather of the day.51 The sun is shining in Southern Italy on this 2nd of August 216 BC and no mist is blocking the view as it had at Lake Trasimene. Nevertheless, the sight of the Romans is impaired by the dust driven by the strong wind from the dry plain into their faces.52 No explicit mention is made of Hannibal’s intentions. But no reader who made his way from the beginning of book 21 to the lowlands of Apulia could possibly expect the ‘master of malice’ to abstain from his tricks at Cannae of all places. Once more, this disquieting feeling on the part of the reader is significantly enhanced by the contrast with the Romans’ complete lack of awareness of possible stratagems on the part of the enemy. On the contrary, they start the fight with great confidence—clamore sublato (47.1)—and are further motivated by the advance of the infantry in the centre, despite the defeat of their cavalry on the right wing:53 tandem Romani, diu ac saepe conisi, aequa fronte acieque densa impulere hostium cuneum nimis tenuem eoque parum ualidum, a cetera prominentem acie. inpulsis deinde ac trepide referentibus pedem institere, ac tenore uno per praeceps pauore fugientium agmen in mediam primum aciem inlati, postremo nullo resistente ad subsidia Afrorum peruenerunt, qui utrimque reductis alis constiterant media, qua Galli Hispanique steterant, aliquantum prominente acie. qui cuneus ut pulsus aequauit frontem primum, deinde cedendo etiam [in] sinum in medio dedit, Afri circa iam cornua fecerant inruentibusque incaute in medium Romanis circumdedere alas; mox cornua extendendo clausere et ab tergo hostes. hinc Romani, defuncti nequiquam [de] proelio uno, omissis Gallis Hispanisque, quorum terga ceciderant, [et] aduersus Afros integram pugnam ineunt, non tantum [in] eo iniquam, quod inclusi aduersus circumfusos, sed etiam quod fessi cum recentibus ac uegetis pugnabant. Liv. 22.47.5–22.47.10

But at last the Romans, by prolonged and frequent efforts, pushing forward with an even front and a dense line, drove in the wedge-like formation which projected from the enemy’s line, for it was too thin to be strong; and then, as the Gauls and Spaniards gave way and fell back in confusion, pressed forward and without once stopping forced their way 51 52 53

See Liv. 22.45.5–22.46.9. See Liv. 22.46.9: uentus—Volturnum regionis incolae uocant—aduersus Romanis coortus multo puluere in ipsa ora uoluendo prospectum ademit. See Liv. 22.47.1–22.47.3.

livy’s fraus punica from a literary point of view

247

through the crowd of fleeing, panic-stricken foes, till they reached first the centre and ultimately—for they met with no resistance—the African supports. These had been used to form the two wings, which had been drawn back, while the centre, where the Gauls and Spaniards had been stationed, projected somewhat. When this wedge was first driven back so far as to straighten the front, and then, continuing to yield, even left a hollow in the centre, the Africans had already begun a flanking movement on either side, and as the Romans rushed incautiously in between, they enveloped them, and presently, extending their wings, crescent-wise, even closed in on their rear. From this moment the Romans, who had gained one battle to no purpose, gave over the pursuit and slaughter of the Gauls and Spaniards and began a new fight with the Africans. In this they were at a twofold disadvantage: they were shut in while their enemies ranged on every side of them; they were tired, and faced troops that were fresh and strong. The only explicit comment given by the narrator so far consists of one adverb: “the Romans rushed incautiously (incaute) in between.” A single word might indeed be enough for a reader already experienced in fraus Punica. But the narrator’s taciturnity is surprising nevertheless. At this stage, Polybius who also gives no precise explication of Hannibal’s tactic in advance54 states at least that things had now developed as the Carthaginian had wanted (ἐξ οὗ συνέβη κατὰ τὴν Ἀννίβου πρόνοιαν).55 Livy’s reader, on the other hand, gets no information either about the plans of the generals or about the concrete proceedings on the battle field. This has often been explained by Livy’s lack of military experience and is thus considered as an unintentional deficiency of his account.56 One could also argue, however, that the lack of information to be observed here is a deliberate narrative strategy by which Livy wants to affect the reaction of his Roman readers: They can recreate more easily the desperate situation of

54 55

56

But his description of Hannibal’s order of battle includes several implicit hints about what the Carthaginian general might have had in mind: cf. Polyb. 3.113.8–3.113.9. Cf. Polyb. 3.115.11–3.115.12: ἐξ οὗ συνέβη κατὰ τὴν Ἀννίβου πρόνοιαν μέσους ἀποληφθῆναι τοὺς Ῥωμαίους ὑπὸ τῶν Λιβύων κατὰ τὴν ἐπὶ τοὺς Κελτοὺς παράπτωσιν. (As a result, it turned out in accordance with Hannibal’s foresight that the Romans, after their pursuit of the Celts, had been trapped between the Libyans). Cf. e.g. Burck [1950] 1962: 95–96: “Die Größe der strategischen Planung Hannibals und seiner taktischen Manöver, die den Namen Cannae unsterblich und zum Vorbild aller Umfassungsschlachten gemacht haben, kann der Leser aus dem Livianischen Bericht nur schwer ermessen; hierfür ist die Beschreibung der Schlacht durch Polybios unerläßlich.”

248

pausch

their historical counterparts who are circumvented and on their own as they, too, are left alone by the narrator without any explanation for what is happening. The assumption that Livy aims at a ‘mirroring effect’ by intentionally not mentioning Hannibal’s idea behind the decisive manoeuvre at Cannae can be reinforced, if we now take a look at the stratagem that immediately follows (and is missing from Polybius interestingly enough).57 It is presented, in a clear contrast, in the most explicit way by the narrator: iam et sinistro cornu Romanis, ubi sociorum equites aduersus Numidas steterant, consertum proelium erat, segne primo et a Punica coeptum fraude. quingenti ferme Numidae praeter ⟨sol⟩ita arma telaque gladios occultos sub loricis habentes, specie transfugarum cum ab suis parmas post terga habentes adequitassent, repente ex equis desiliunt parmisque et iaculis ante pedes hostium proiectis in mediam aciem accepti ductique ad ultimos considere ab tergo iubentur. Liv. 22.48.1–22.48.3

By this time the Roman left, where the cavalry of the allies had taken position facing the Numidians, was also engaged, though the fighting was at first but sluggish. It began with a Punic ruse. About five hundred Numidians, who, in addition to their customary arms and missiles, carried swords concealed under their corselets, pretended to desert. Riding over from their own side, with their bucklers at their backs, they suddenly dismounted and threw down bucklers and javelins at the feet of their enemies. Being received into the midst of their ranks they were conducted to the rear and ordered to fall in behind. This time, we do not need to read further to find out what is about to happen. The fraus Punica is named at the very beginning of the passage and the Numidians are described as merely species transfugarum; we also learn in advance about the additional weapons hidden in their armour and thus achieve a much fuller view of the situation than the historical figures themselves who were part of it. After the Carthaginian’s insidiae, the Romans are lost, and the narration now focuses on the dramatic death of Aemilius Paullus58 and some other 57 58

Cf. Liv. 22. 48.1–48.4. For a discussion of possible sources (perhaps Valerius Antias) see Klotz 1949: 64–65; Burck [1950] 1962: 97; and Gärtner 1975: 13–15. Cf. Liv. 22.49.6–22.49.13; see e.g. Burck [1950] 1962: 98: “Den größten Raum nimmt bezeichnenderweise der letzte Abschnitt ein, in dem das Militärische zurücktritt …, das hohe

livy’s fraus punica from a literary point of view

249

poignant episodes suitable for involving Livy’s Roman readers emotionally and perhaps for raising their spirits again—at least ever so slightly.59

5

Outlook: Explaining Defeats or Engaging the Reader?

Helping his Roman readers cope with the Roman defeat is indeed one of Livy’s primary concerns. The battle descriptions in Ab Urbe Condita are therefore often analysed from such a perspective. Viewed in this light, Hannibal’s skilful stratagems and his fraus Punica can be seen as key factors in an attempt to exculpate the Romans of the third century BC: they would, of course, have been victorious if the Carthaginians had granted them a fair fight on equal ground. Hannibal’s stratagems can, therefore, be counted among what Heinz Bruckmann, in his seminal study on Roman defeats in Livy, called ‘Entlastungsfaktoren’ (‘exculpatory factors’).60 Yet, as, for example, David Levene has shown in his 2010 monograph on the Hannibalic war, there is more to Livy’s battle descriptions than alleviating his readers’ dismay about the disasters of the past by blaming the enemy for cheating: All of this shows that interpreting Livy’s battles primarily in terms of his provision of patriotic excuses for defeat, as Bruckmann does, is not satisfactory. Livy’s position is more complex—more opaque, indeed—than that. The excuses that he provides are largely derived from Polybius, and far from accepting them as the primary explanation for Roman losses, he rewrites the battles in order to minimize their actual effect on the sequence of events. The reader is invited to recognize Polybius’ account underlying Livy’s, but also to reject it.61

59

60

61

Ethos und das nationale Verantwortungsgefühl des sterbenden Consuls aber mit ergreifender Anschaulichkeit hervortreten.” Cf. Liv. 22.49.14–22.50.12. On the exculpatory tendency see e.g. Bruckmann 1936: 70–103; Klotz 1949; Burck [1950] 1962: 90–102; and Gärtner 1975: 13–18. On the role played by Cannae in the following books of the third decade see Chaplin 2000: 54–72. See Bruckmann 1936: esp. 124–125: “Die Folge ist, daß er … überhaupt in allen Berichten über römische Niederlagen niemals eine bessere kämpferische Leistung der römischen Gegner anerkennt, geschweige denn irgendwie betont herausarbeitet. Eine entscheidende Rolle spielt dabei in der Gruppe dieser Niederlagen eine Reihe von verschiedenartigen Tatsachen (z.B. temeritas ducis; discordia ducum; fatum; insidiae), die wir unter dem Namen der ‘Entlastungsmomente’ zusammenzufassen versuchten.”; for a summary that is more easily accessible than his original Münster dissertation from 1936 see Bruckmann 1967. Levene 2010: 274.

250

pausch

In addition to his observations, even more attention should be paid to the literary side of the battle descriptions in the third decade of Ab Urbe Condita. This is a wide subject and in this paper only a small segment could be singled out. Nevertheless, the narrative route leading to Cannae might offer a good example of the more nuanced view on Livy’s work resulting from a literary analysis. At this point, it might be useful to recall the main points of the argument presented here: a recognizable pattern is introduced by the narrator of how Hannibal manages to outwit his historical opponents repeatedly who, in addition, understand what is happening to them much later than the contemporary readers at each time. After it is established, the pattern is varied, since in the case of Cannae the same expectations that have been raised initially at the Trebia and then have been fulfilled in an exemplary way at Lake Trasimene are finally belied—or so it seems—as Hannibals’ ruse of the empty camp is seen through by the Romans and thus fails. Yet this success in learning will not prove sufficient for military victory. On the contrary, the Carthaginian general will continue to play tricks on his enemies. Among these, the feigned desertion of the Numidians is explicitly mentioned by the narrator. This ambush is successful, but turns out to be of secondary importance for the final outcome of the battle. Its prominence can be explained, however, from a narratological perspective, as the empty camp and the feigned desertion are both described explicitly as artifices and thus provide a kind of framework for the main stratagem, the encircling of the centre of the Roman infantry which, unlike the previous two, remains implicit and is explained neither in advance nor even in retrospect. And yet it is this ruse that the narrator as well as historical tradition unanimously identify as the decisive factor for the defeat. Its prominence, however, is further enhanced by the means of this framing technique as well as by the fact that it becomes the focal point of a wide-ranging literary pattern beginning at the Trebia at least. In this way of telling the story, the motif of fraus Punica surely plays an important part and undeniably contributes to a negative depiction of Hannibal. Yet it seems to me, that Livy is not so much interested in an historical evaluation of the character of the Carthaginian general or in the retelling of dated ethnic stereotypes, but primarily uses this motif for the creation and for the maintenance of suspense as well as for the involvement of the reader in the narration—two goals that are important for his way of presenting history in general.62 We might even go so far as to say that the narrator is himself leading his readers into a trap at Cannae by not giving them his usual and, hence,

62

See e.g. Pausch 2011: esp. 251–254.

livy’s fraus punica from a literary point of view

251

expected, explanations before the fact. In so doing, he puts the readers into a position resembling the one of the Romans at the time of the events instead of granting them their usual privileged point of view. They can thus re-experience the deception of the historical figures and get involved in the story and their own history in a very emphatic way. This specific narrative technique may help to explain this particular defeat to a contemporary audience and make it even harder for them to criticize the men in charge at the time. But this, of course, will depend on each reader’s readiness to let himself be drawn into the story.

Bibliography Anz, T., ‘Spannung’, Reallexikon der deutschen Literaturwissenschaft 3 (Berlin 2003) 464–467. Burck, E., Die Erzählkunst des T. Livius ([1934] Berlin 1964). Burck, E., Einführung in die dritte Dekade des Livius ([1950] Heidelberg 1962). Briscoe, J., ‘Livy and Polybios’, in W. Schuller (ed.), Livius. Aspekte seines Werkes (Konstanz 1993) 39–52. Bruckmann, H., Die römischen Niederlagen im Geschichtswerk des T. Livius, dissertation (Münster 1936). Bruckmann, H., ‘Die römischen Niederlagen im Geschichtswerk des Titus Livius’, in E. Burck (ed.), Wege zu Livius (Darmstadt 1967) 298–309. Chaplin, J.D., Livy’s Exemplary History (Oxford 2000). Chassignet, M., ‘L’image des Barcides chez les historiographes latins de la Republique: naissance d’une tradition’, in J. Pigón (ed.), The Children of Herodotus: Greek and Roman Historiography and Related Genres (Newcastle 2009) 206–218. Christ, K., ‘Zur Beurteilung Hannibals’, Historia 17.4 (1968) 461–495. Cipriani, G., ‘La chiamata dell’eroe (Commento a Livio 21,1–5)’, Aufidus 2 (1987) 3–28. Dorey, T.A., Titi Livi: ab urbe condita XXI–XXII (Leipzig 1971). Engels, D., Das römische Vorzeichenwesen (753–27 v. Chr.). Quellen, Terminologie, Kommentar, historische Entwicklung (Stuttgart 2007). Feldherr, A., Spectacle and Society in Livy’s History (Berkeley 1998). Foster, B.O., Livy V: Books XXI–XXII (London 1929). Fuhrmann, M., ‘Narrative Techniken im Dienste der Geschichtsschreibung (Livius, Buch 21–22). Eine Skizze’, in E. Lefèvre & E. Olshausen (eds.), Livius. Werk und Rezeption, Festschrift für Erich Burck (München 1983) 19–29. Gärtner, H.A., Beobachtungen zu Bauelementen in der antiken Historiographie besonders bei Livius und Caesar (Wiesbaden 1975). Gerrig, R.J., ‘Suspense in the Absence of Uncertainity’, Journal of Memory and Language 28.6 (1989), 633–648.

252

pausch

Händl-Sagawe, U., Der Beginn des 2. Punischen Krieges. Ein historisch-kritischer Kommentar zu Livius Buch 21 (München 1995). Hoyos, D., ‘Rome and Carthage in Livy’, in B. Mineo (ed.), A Companion to Livy (Chichester 2015) 369–381. Jaeger, M., Livy’s written Rome ([1997] Ann Arbor, 2009). Klotz, A., ‘Dichtung und Wahrheit in der livianischen Erzählung von der Schlacht bei Cannae’, Gymnasium 56 (1949), 58–70. Krafft, P., ‘Hannibals Perücken. Motivik und Erzählstruktur von Livius 22,1’, RhM 150 (2007), 67–85. Levene, D.S., Religion in Livy (Leiden 1993). Levene, D.S., Livy on the Hannibalic War (Oxford 2010). Luce, J.T., Livy. The Composition of his History (Princeton 1977). Mensching, E., ‘Zur Entstehung und Beurteilung von Ab Urbe Condita’, Latomus 45 (1986), 572–589. Meißner, B., ‘C. Flaminius, oder: Wie ein Außenseiter zum Sündenbock wurde’, in K.-J. Hölkeskamp & E. Stein-Hölkeskamp (eds.), Von Romulus zu Augustus. Große Gestalten der römischen Republik (München 2000) 92–105. Miles, G.B. Livy: Reconstructing Early Rome (Ithaca 1995). Mineo, B. (ed.), A Companion to Livy (Chichester 2015). Pausch, D., Livius und der Leser. Narrative Strukturen in ab urbe condita (München 2011). Porod, R., ‘Die livianischen Bewertungskriterien in den Hannibalbüchern’, GB 16 (1989), 203–227. Rossi, A., ‘Parallel lives: Hannibal and Scipio in Livy’s third decade’, TAPhA 134 (2004), 359–381. Schmitz, T.A., ‘Ist die Odyssee ‘spannend’? Anmerkungen zur Erzähltechnik des homerischen Epos’, Philologus 138.1 (1994), 3–23. Stepper, R., ‘Politische Parolen und Propaganda im Hannibalkrieg’, Klio 88.2 (2006), 397–407. Tränkle, H., Livius und Polybios (Basel 1977). Treptow, R., Die Kunst der Reden in der 1. und 3. Dekade des livianischen Geschichtswerkes, dissertation (Kiel 1964). Vasaly, A., Livy’s Political Philosophy Power and Personality in Early Rome (Cambridge 2015). Walsh, P.G., Livy. His Historical Aims and Methods ([1961] Cambridge 1989). Wheeler, E.L., Stratagem and the Vocabulary of Military Trickery (Leiden 1988). Will, W., ‘Imperatores victi. Zum Bild besiegter Consuln bei Livius’, Historia 32.2 (1983), 173–183.

chapter 10

Livy’s Use of Spatial References in the Cannae Episode: from Structure to Strategy Lidewij van Gils

1

Introduction*

Spatial awareness is crucial in warfare, as Quintus Fabius Maximus points out to consul Paullus in the famous speech, recorded in Liv. 22.39. He says Paullus should be more afraid of his consular colleague Varro than of the enemy Hannibal, because Varro dangerously lacks interest in spatial and other strategic information: (1) hic [sc. Varro] […] priusquam castra uideat aut hostem, insanit. Liv. 22.39.6

This man […] is raving before he has seen the camp or the enemy.1 Varro himself is told by the narrator to have repeatedly said in popular assemblies in Rome: (2) bellum […] se quo die hostem uidisset perfecturum. Liv. 22.39.7

that he would bring the war to conclusion on the very day he got sight of the enemy. Paullus responds to Varro’s harangues in a short speech just before their departure to the battle field: (3) mirari se quindi qui dux priusquam aut suum aut hostium exercitum locorum situm naturam regionis nosset, iam nunc togatus in urbe sciret quae sibi agenda armato forent […] Liv. 22.39.8 * Special thanks go to Irene de Jong for her feedback on a first draft of this paper. 1 Translation of all examples is by A.M. Baker 1834. © koninklijke brill nv, leiden, 2019 | doi:10.1163/9789004383340_011

254

van gils

that he wondered how any general, before he knew anything of his own army, or that of the enemy, the situation of the places, or the nature of the country, even now while in the city, and with the gown on, could tell what he must do when in arms. It is made clear from the outset of the Cannae episode that space needs to be taken seriously, which is exactly what we will do in this chapter.2 A combination of narratological and linguistic categories will serve to analyse the role of space in Livy’s presentation of the battle of Cannae. First, I distinguish the spatial references within the narrative world from those of the discourse situation, outside the narrative (section 2). Next, I discuss the textstructuring functions of space in the narrative in the Cannae-episode (section 3). I dedicate the remainder of the chapter to strategic uses of space in the Cannae narrative (section 4) and relate spatial references to specific character viewpoints (section 5) and to cognitive processes of orientation (section 6). From a focus on character viewpoints and orientation, at the end of this chapter I switch to the nature of the spatial object itself (section 7) and end with a conclusion (section 8).

2

The Discourse World of Livy’s Intended Audience versus the Story World of Cannae

In a narrative text at least two worlds need to be distinguished, each with its own spatio-temporal centre: the discourse world and the narrative world.3 In Livy’s text, the discourse world of Livy’s addressees is that of a Roman reading audience.4 The society of the first century AD provides shared knowledge and expectations with regard to historical facts, moral judgements, literary conventions and other types of common ground. In this discourse world, Livy may ‘anchor’, as it were, the Cannae episode in the common ground of his audience, and refer to it as a well-known Roman disaster.5

2 The literary choices of Livy to portray Varro as a strategic failure, in contrast to Paullus, is not incompatible with Polybius’ version of this episode, but Polybius loathes Varro’s inexperience and flight and does not concentrate on his (lack of) strategic insight. See in this volume Oakley (this volume, 157–161) for a comparison of the two authors. 3 See Introduction. 4 See Pausch 2011: 13, 17–46 for an analysis of Livy’s contemporary reading audience; and Levene 2010: 65–74 for a discussion of the common ground of his reader, assumed by Livy. 5 Anchoring Innovation is the award winning research agenda of OIKOS, the National Research

livy’s use of spatial references in the cannae episode

255

(4) Haec est pugna Cannensis, Alliensi cladi nobilitate par. Liv. 22.50.1

Such is the battle of Cannae, equal in celebrity to the defeat at the Allia. The author relies on his audience’s knowledge about the places Cannae and Allia and their fame due to Roman defeats. This type of sentence is not part of the story world, but part of the discourse world of the author and his primary addressees.6 In this paper, I will not deal with spatial references to this discourse world, but instead focus on the narrative world that is created by the narrator.

3

Text-Structuring Spatial References

The omniscient narrator of the battle of Cannae structures the story in key episodes centred on a few main characters. One of the structuring devices used is formed by spatial references.7 A transition to a new episode may be indicated by a change of location, usually from some distant viewpoint, as in the following example: (5) Iam et sinistro cornu Romanis, ubi sociorum equites aduersus Numidas steterant, consertum proelium erat. Liv. 22.48.1

Now also in the left wing of the Romans, in which the allied cavalry were opposed to the Numidians, the battle was joined. This sentence comes after a long episode (section 47) about the Roman right wing. The spatial reference sinistro cornu takes us, the narratees, to another side of the battle field and with the pluperfect tense it takes us a little bit back in time since both wings will have started to fight at the same time. This type of transition to a new episode is associated to a bird’s-eye view on the events and

School in Classical Studies, the Netherlands. For the concept of “anchoring”, see the website http://www.ru.nl/oikos/anchoring‑innovation/ under further reading. 6 The implied knowledge of his audience about the Roman defeats is not necessarily accurate. By using this sentence, Livy creates an ‘ideal addressee’ who shares knowledge with the speaker about Cannae and Allia. 7 See also Walsh 1961: 181; and Levene 2010: 75–78.

256

van gils

it is a common device in ancient narrative, including historiography. Another example of this text-structuring device can be found in the beginning of section 49. (6) Parte altera pugnae Paullus, quamquam primo statim proelio funda grauiter ictus fuerat, tamen et occurrit saepe cum confertis Hannibali […] Liv. 22.49.1

In the other part of the field Paullus, although he had received a severe wound from a sling at the very outset of the battle, nevertheless repeatedly opposed himself to Hannibal, with his men in close formation, […] We also find another spatial type of transition: characters arriving at a new location, as in example 7. (7) Ut in castra uenerunt, permixto nouo exercitu ac uetere, castris bifariam factis […] Liv. 22.40.5

When they had arrived at the camp, the old and new troops being united, they formed two distinct camps […] The sentences before this example refer to the departure of the two consuls from Rome, ending the former episode. The change of location from Rome to the castra signals a new deictic centre and a next step in the narrative.8 The temporal progression naturally and implicitly coincides with the time that was needed for the journey. Similar examples may be found in the beginning of sections 43 and 44: (8) Hannibal […] nequiquam detecta fraude in castra rediit. Liv. 22.43.1

Hannibal […] returned to the camp, his stratagem having been detected and rendered idle.

8 See Richardson 1990: 110–111 for this type of transitions to a new scene in Homer and in novels.

livy’s use of spatial references in the cannae episode

257

(9) Consules satis exploratis itineribus sequentis Poenum, ut uentum ad Cannas est et in conspectu Poenum habebant, bina castra communiunt, […] Liv. 22.44.1

The consuls, after making a sufficient reconnaissance of the roads, followed the Phoenicians and when they came to Cannae, having the enemy in view, they fortified two camps […] In all text-structuring uses of spatial references, we find that the new location in se has been anticipated by the preceding context: after the right wing comes a left wing and after departure from Rome, comes arrival at the castra. Apart from such fairly common and unobtrusive spatial references with a structuring function, we also find in Livy’s Cannae episode instances with a strategic function, as we will see in the remainder of this chapter.

4

Strategic Spatial References

As Paullus and Quintus Fabius Maximus had already recognized, spatial awareness is indispensable for a general. Hannibal turns out to be superior in precisely this aspect, in contrast to the Roman general Varro. Hannibal’s final victory depends on his strategic interpretation of places and moving entities as tools or obstacles: he recognizes them as appealing or appalling, not only from his personal point of view, but also with a keen eye for the viewpoint of his enemy. When Hannibal arrives at Cannae, he positions his camp in such a way that the dusty wind will not hinder his men, not only during their stay, but also during the battle: (10) Prope eum uicum Hannibal castra posuerat auersa a Volturno uento, qui campis torridis siccitate nubes pulueris uehit. Liv. 22.43.10

Hannibal had pitched his camp near that village, with his back to the wind Volturnus, which, in those plains, which are parched with drought, carries with it clouds of dust. Let us take a closer look at the spatial reference castra posuerat auersa a Volturno uento. The pluperfect (posuerat) presents this event as prior to the setting of the action, creating with this analepsis a narrative world distinct from

258

van gils

the setting of the action.9 The analepsis refers to Hannibal’s strategic positioning of his army, which is commented upon by the narrator with an informative sentence about the Volturnus followed by a proleptic remark on the expediency of this position in the future battle. Quite explicitly, the narrator connects the outcome of the battle in Cannae to Hannibal’s strategic choice for his encampment by stating that Hannibal’s disposition would be ‘particularly salutary when the troops formed up for battle, facing in the opposite direction, with the wind blowing only on their backs, and ready to fight with enemies half-blinded by the dust driven into their faces.’10 The presence of more than one world in narrative is not limited to analepses. Such alternative worlds, also known as frames, may consist also of embedded focalization, window views, dreams, etcetera.11 A few lines later, we find another example of a frame. As soon as the Romans have pitched their camps, Hannibal draws up his troops in battle order (setting of the action) in the hope (the start of a frame) that the Romans could be provoked to fight. (11) Hannibal spem nanctus locis natis ad equestrem pugnam, qua parte uirium inuictus erat, facturos copiam pugnandi consules, dirigit aciem lacessitque Numidarum procursatione hostes. Liv. 22.44.4

Hannibal, entertaining a hope that the consuls would not decline a battle in this tract, which was naturally adapted to a cavalry engagement, in which portion of his forces he was invincible, formed his line, and provoked the enemy by a skirmishing attack with his Numidians. In his mind Hannibal envisions the Romans provoked to fight. This is an alternative story world to the one presented by the narrator. In this example, we get to see Hannibal’s view of the space that surrounds him. Hannibal sees a ‘tract which was naturally adapted to a cavalry engagement’. This strategic view on space turns out to be typical for the character Hannibal. His embedded focalization, the start of which was signalled by spem nanctus, not only provides a

9 10

11

See De Jong 2015:78–87 for an introduction of the concept of analepsis in classical literature. Id cum ipsis castris percommodum fuit, tum salutare praecipue futurum erat, cum aciem dirigerent, ipsi auersi, terga tantum adflante uento, in occaecatum puluere offuso hostem pugnaturi. (22.43) See De Jong 2014:107 for a distinction between Setting and Frames; and Ryan 1991 for an elaborate study of Possible Worlds in Narratives.

livy’s use of spatial references in the cannae episode

259

motif for his actions (dirigit lacessitque), but it portrays Hannibal as a scheming character and his view guides the reader’s view of the events. The word hostes referring to the Romans is not problematic once the reader accepts seeing things from Hannibal’s standpoint. But exactly how close does the reader come to Hannibal’s view of his surroundings and to what extent does his view include visible details? In the next section, the standpoints of narrator and characters in the Cannae-episode are subjected to closer inspection.

5

Narratorial Standpoints and Discourse Modes

Returning to the sentence of example 10, in which Cannae is identified as the new location of Hannibal’s camp, I would like to analyse the spatial distance between the narrated and the narratee. The summarizing rhythm (posuerat castra), the reference to large unities (castra) and absence of visible details reflect a large distance between the standpoint and what is viewed. In terms of the narratological tripartite division in panoramic, scenic and close-up standpoints, the first one is clearly the most suited here.12 The panoramic viewpoint appears to be standard in this episode. The battle of Cannae does not present many scenes and even fewer close-ups. Exceptions are the speeches in the beginning of the story, and the direct discourse of Lentulus and Paullus at the end of the battle. Most of the narrative consists of a panoramic view of the movements of the armies and the rhythm is, consistently, summarizing.13 A panoramic standpoint is conceivable for a character when this character is in a high position (typically the walls of a city, a palace roof or a hill) and, in fact, Hannibal may have taken such a position.14 However, such a distanced stand12

13

14

See De Jong 2014: 60–65 for an introduction of the narratorial panoramic standpoint, the actorial panoramic standpoint, the narratorial and actorial scenic standpoints and the close-up. The terms perspective and point-of-view are not used, in favour of focalisation and standpoint. A panoramic standpoint is typically used for the presentation of movements in a more summarizing fashion, whereas the details of a close-up may need more narrating time than narrated time, in other words a slow-down. In a scene the narrated time coincides with the narrating time as is the case with direct discourse, for instance, and the distance for a scene is more or less what is needed to overhear a conversation without interfering. A scene seems to be prototypically a chain of events witnessed by the narrator. The teichoskopia in Iliad 3.161–3.242 is a famous example of actorial panoramic standpoint. In Riggsby 2009:155–156 this type of panoramic overview of a battlefield is classified as ‘tactical space’ in opposition to ‘geographic space’ in which a large territory is introduced (Gaul, for instance) and ‘strategic space’ which consists of the interconnected routes (network) of relevant sites.

260

van gils

point especially when combined with a summarizing rhythm is usually the narrator’s. Even more, such a narrating style comes close to reporting events rather than narrating them. In other words, the shift in standpoint from the narrator to Hannibal without losing a distanced, reporting narrative style is remarkable and worth investigating. In order to get a firmer grip on Livy’s way of presenting the story and on alternative presentations in other war narratives, we may use the discourseanalytical model as developed in Amsterdam.15 Within this model, we distinguish Descriptive, Narrative and Discursive Discourse Modes based on the internal structure of a longer stretch of text. Within Description, the various clauses are related to each other through spatial relations, whereas a Narrative passage is characterized by temporal relations between the clauses. When a narrator presents observations or events as relevant in isolation or as part of an argumentation this is called the Discursive mode. It is possible to identify prototypical combinations with distance and rhythm, but by no means does this scheme imply that other combinations are impossible. (12) Table 10.1: Prototypical (not exclusive) combinations of distance, rhythm and discourse mode Distance

Rhythm

Discourse mode

panoramic summary discursive scene scene narrative close-up slow down description

The table above expresses a prototypical co-occurrence of the panoramic view, a summarizing rhythm and the discursive mode. But is this the case in my example 10? Does this sentence present Hannibal’s positioning of his castra as an isolated, meaningful event? Or are temporal relations between the clauses in this passage dominant? Or are spatial relations dominant? In order to answer this question, we must look at the larger passage. (13) Cum utriusque consulis eadem quae ante semper fuisset sententia, ceterum Varroni fere omnes, Paulo nemo praetor Seruilium, prioris anni consulem, 15

This approach to ancient texts has been developed mainly by a number of Latin and Greek linguists working in close collaboration at the Vrije Universiteit Amsterdam: Kroon 2007; Adema 2008; Allan 2009; Van Gils 2009.

livy’s use of spatial references in the cannae episode

261

adsentiretur, ex maioris partis sententia ad nobilitandas clade Romana Cannas urgente fato profecti sunt. Prope eum uicum Hannibal castra posuerat auersa a Volturno uento, qui campis torridis siccitate nubis pulueris uehit. Id cum ipsis castris percommodum fuit, tum salutare praecipue futurum erat cum aciem dirigerent, ipsi auersi terga tantum adflante uento in occaecatum puluere offuso hostem pugnaturi. Liv. 22.43.8–22.43.11

The consuls persisted in the same opinions they ever entertained, but nearly all acquiesced with Varro, and none with Paullus except Servilius, the consul of the former year. In compliance with the opinion of the majority, they set out, under the impulse of destiny, to render Cannae celebrated by a Roman disaster. Hannibal had pitched his camp near that village, with his back to the wind Volturnus, which, in those plains which are parched with drought, carries with it clouds of dust. This circumstance was not only very advantageous to the camp, but would be a great protection to them when they formed their line, as they, with the wind blowing only on their backs, would combat with an enemy blinded with the thickly blown dust. Temporal relations are typical of a Narrative Discourse Mode and if we look at the sentence preceding Hannibal castra posuerat, there is a meaningful temporal connection between the departing Roman armies (profecti sunt) to Cannae and Hannibal’s previously setting up of his castra at that place. A counter argument to temporal sequence, however, is the change in location and subject (prope eum uicum Hannibal) which indicates at least a small rupture in the narrative. In defence of a Descriptive relation, one could point at the location Cannae, mentioned in the preceding sentence, which is filled in with visible details in the Hannibal castra posuerat-sentence. This shows how difficult it may be to choose the most relevant coherence relation when a set of sentences share spatial, temporal and argumentative frames. My cautious suggestion would be that the sentence under scrutiny (castra posuerat) is neither predominantly narrative nor descriptive, but part of a discursive passage. Its coherence is, in my opinion, mainly based on the implicit argumentation that Cannae became synonymous with disaster because of Hannibal’s strategic positioning of his camp. Moreover, the sentence after Prope eum uicum […] nubis pulueris uehit evaluates precisely this point (percommodum fuit) from the narrator’s omniscient point of view. An evaluation may concern any type of discourse mode, but with the discursive mode it is certainly a natural combination.

262

van gils

The strategic use of space is, as we have already anticipated with example 11, typical for Hannibal in the Cannae episode. His view is, like that of the narrator, mostly panoramic rather than scenic. But how do we distinguish between the narrator’s panoramic view and that of Hannibal? Are there any linguistic clues which point to one standpoint rather than another? In example 11 the use of the word hostes (enemies) for the Romans may be considered a good indication that we share Hannibal’s standpoint. And how about the use of a spatial indication like castra auersa a uento (the camp with his back to the wind) in example 10: does the interpretation of this reference need a particular standpoint? The next section provides answers to these and similar questions from a cognitive linguistic framework.

6

Frames in Cognitive Linguistics

Spatial references have been studied by cognitive linguists like Levinson in terms of their universal semantic features.16 Levinson introduced absolute, intrinsic and relative frames. The word ‘frame’ appears for the second time in this chapter but, unfortunately, it has a totally different meaning within cognitive linguistics. In section 4, the narratological use of the term frame indicated a location other than the setting of the main story. Narratological frames are, as we have seen, embedded thoughts or an analepsis. Within cognitive linguistics, frame is used as a manner to conceptualize spatial references. With the present chapter, I do not propose a new model but, rather, I integrate influential narratological and linguistic approaches to space in a narrative. Therefore, rather than solving it, I simply point at this incongruence in terminology. (14) Table 10.2: Cognitive Linguistic Frames Frame

Needed

Absolute the army east of the river – Intrinsic the army in front of the enemy orientation Relative the army at the left side of the river viewpoint

16

Levinson 2003.

livy’s use of spatial references in the cannae episode

263

An absolute frame does not need any particular viewpoint in order to be understood. The intrinsic frame needs a reference point (the enemy in the example) with a natural orientation, for instance an upside and a downside, or a front and a back, but again no particular viewpoint is needed. In the case of a relative frame, we need to know the viewpoint in order to interpret the reference. The end of section 46 contains examples of all types of frames within a few sentences: (15) Duces cornibus praeerant, sinistro Hasdrubal, dextro Maharbal; mediam aciem Hannibal ipse cum fratre Magone tenuit. Sol, seu de industria ita locatis seu quod forte ita stetere, peropportune utrique parti obliquus erat, Romanis in meridiem, Poenis in septemtrionem uersis; Liv. 22.46.7–22.46.8

The generals commanding on the wings were Hasdrubal on the left, Maharbal on the right; Hannibal himself, with his brother Mago, had the centre. The sun—whether they had so placed themselves on purpose or stood as they did by accident—was, very conveniently for both sides, on their flanks, the Romans looking south, the Phoenicians north. At the time of Livy, praesum meant ‘to command’, but etymologically speaking, it represents an intrinsic frame.17 The verb praesum originally meant ‘to be in front of’, an example of an intrinsic frame which needs the back / front orientation of the army in order to be understood. With the spatial references sinistro (on the left) and dextro (on the right) a relative frame is needed; there are two possible standpoints: either we adopt Hannibal’s standpoint, or we view his army from the point of view of the Romans, which logically inverts the interpretation of right and left. There are no contextual clues which point to a particular standpoint, but there seems to be a common ground in the historiographical genre to describe the position of the wings from the point of view of the described army.18 This principle can be illustrated by a sentence taken from the following section in Livy, where we hear that ‘the Gallic and Spanish

17 18

Maltby 1991: 493; and De Vaan 2008: 485 about the etymology of prae and prae-compounds. With an intrinsic view standpoint, there is no alternative interpretation with regard to the spatial references, while with a relative standpoint there may be a preferred ‘quasiintrinsic’ standpoint, but other standpoints are possible. A relative standpoint may also be explicitly mentioned.

264

van gils

horse which formed the left wing engaged with the Roman right in a combat’.19 Left is interpreted from the Punic standpoint, and right from the Roman standpoint. The change in standpoint occurs without any marking because of the mentioned principle of taking the relevant army as the standard standpoint. Hannibal’s position is presented as keeping the mediam aciem (the centre). An intrinsic frame is needed to interpret this spatial reference: an army has a central part regardless of a specific standpoint. After presenting the position of the Punic commanders, the narrator provides us with the orientation of the Punic and the Roman armies with an absolute frame with the words in meridiem (south) and in septemtrionem (north). In table 10.3, I summarize this analysis. (16) Table 10.3: Analysis of spatial references Text Duces cornibus praeerant sinistro Hasdrubal dextro Maharbal; mediam aciem Hannibal ipse tenuit Romanis in meridiem Poenis in septentrionem uersis.

‘intrinsic’ relative relative intrinsic absolute absolute

Absolute spatial reference frames are not very frequent in the Cannae-passage. Usually the narratee is invited to interpret spatial references either from an intrinsic perspective or a relative perspective. The intrinsic examples usually have the Roman soldiers as an orientation point, but in example 10 the Punic castra provide the intrinsic frame to understand auersa a uento.20 An intrinsic orientation point is not related to a standpoint, as is shown also in example 17: (17) […] Afri circa iam cornua fecerant irruentibusque incaute in medium Romanis circumdedere alas, mox cornua extendendo clausere et ab tergo hostes. Liv. 22.47.8

19 20

deinde equitum Gallorum Hispanorumque laevum cornu cum dextro Romano concurrit. Liv. 22.47. intra pedites (22.45); ab tergo clausere (22.47); aversam Romanam aciem (22.48).

livy’s use of spatial references in the cannae episode

265

[…] the Africans had already begun a flanking movement on either side, and as the Romans rushed incautiously in between, they enveloped them, and presently, extending their wings, crescent-wise, even closed in on their rear. This spatial reference ab tergo uses the Romans as an intrinsic orientation point. However, it is clearly not their point of view, as the use of hostes for the Romans indicates. A relative frame is always linked to a specific standpoint; left and right need such a standpoint, and so does trans (on the other side, to the other side). The river Aufidus and the mountains near Cannae are obstacles which occur frequently with trans, either as preposition (trans flumen) or as prefix (traduxerat). The following table shows the changing standpoints needed for a correct interpretation of trans. (18) Table 10.4: Relative frames, some examples Section Example

Viewpoint

42 43 44 45 45 46 46

Hannibal Romani Roma Hannibal Varro Hannibal Hannibal

trans proximos montes omnibus ultra castra transque montes exploratis castris, quae posita trans Aufidum erant Hannibal Numidas trans flumen mittit Varro instructus copias flumen traduxit transgressus flumen traduxerat

The frequency of Hannibal’s standpoint is noteworthy, since the Roman narrator and his narratees are supposed to relate more to the Romans. Let’s take a closer look at one of the sections where a change in standpoint occurs. In section 43, the situation in the Punic camp (simply referred to as castris) pushes Hannibal to make a decision (statuit). As the (implicit) subject of a focalizing verb, Hannibal is the logical standpoint for spatial references within the focalized content:21 The relative references inde, hoste and transfugia, in fact, all refer to Hannibal’s standpoint. 21

Focalization has been introduced in Classical scholarship with De Jong 1987. The concept of embedded thought and embedded speech is captured in all its varieties with the term focalization. See De Jong 2014: 47–72 for an introduction to the concept.

266

van gils

(19) Cum haec consilia atque hic habitus animorum esset in castris, mouere inde statuit in calidiora atque eo maturiora messibus Apuliae loca, simul ut, quo longius ab hoste recessisset, eo transfugia impeditiora leuibus ingeniis essent. Profectus est nocte ignibus similiter factis tabernaculisque paucis in speciem relictis, ut insidiarum par priori metus contineret Romanos. Sed per eundem Lucanum Statilium omnibus ultra castra transque montes exploratis, cum relatum esset uisum procul hostium agmen, tamen de insequendo eo consilia agitari coepta. Liv. 22.43.5–22.43.7

Such being the plans in agitation, and such the state of feeling in the camp, he resolved to depart thence into the regions of Apulia, which were warmer, and therefore earlier in the harvest. Thinking also, that the farther he retired from the enemy, the more difficult would desertion be to the wavering. He set out by night, having, as before, kindled fires, and leaving a few tents to produce an appearance; that a fear of an ambuscade, similar to the former, might keep the Romans in their place. But when intelligence was brought by the same Lucanian Statilius, who had reconnoitred every place on the other side of the mountains and beyond the camp, that the enemy was seen marching at a distance, then plans began to be deliberated on about pursuing him. Hannibal’s strongest weapon seems to be his capacity to assume the perspective of his enemies. In fact the Roman perspective of being held back through fear is envisaged and nourished by Hannibal before it becomes reality. Hannibal has left (profectus est) leaving behind some tents (paucis tabernaculis relictis) with the intention to delay the enemy (ut … contineret) by counting on their fear (par priori metus). After Hannibal’s anticipation of the Roman reaction to his departure, the perspective swiftly turns to the Romans. The narratee now looks at Hannibal’s departure from the Roman point of view, as is clear from the prepositions ultra and trans, the adverb procul and the verb insequor. And from this Roman standpoint, we find out that Hannibal’s anticipation of the Roman reaction had been correct. In concomitance with the relative spatial frames the word ‘hostis’ in this example contains the same implicit standpoint. First, the enemy is the Roman army (ab hoste recessisset) and shortly afterwards the hostes refer to Hannibal’s men (hostium agmen). With example 19, we have seen how a linguistic analysis of spatial references helps to identify implicit perspectives in a narrative. The main perspectives are the Romans and in particular their consuls Paullus and Varro and contrasted to

livy’s use of spatial references in the cannae episode

267

them we find ourselves repeatedly sharing a viewpoint with Hannibal. Hannibal’s view of space is radically different from the Roman view, especially from the view of Varro and his soldiers. I would like to take our spatial analysis one step further in order to grasp this strategic difference and focus on how spatial elements are perceived in strategic terms, i.e. as an opportunity, an instrument, an obstacle or a threat. In the next section, we will study some entities and locations and how their spatial features are conceived of by the main participants of the story.

7

Strategic Interpretations of Space

A division of spatial references in entities and locations is artificial, especially in cases like castra, which may be both. For the present text, I propose the following continuum with, on the one end, unmovable locations (Roma, Cannae, the river, the mountain) and on the other end moving individuals. In between, I put moveable places (castra) which are sometimes conceived of as places and sometimes as moveable objects and moving groups (wings of an army) which are sometimes moved by their commander but, especially, on the battle field, they seem to move autonomously, like individuals. The table under 20 summarizes the four options: (20) Table 10.5: Locations and entities Location Moving Autonomous Indivisible Roma castra ala Hannibal

+ + + –

– + + +

– – + +

– – – +

Roma, the river Aufidus and other territorial spaces typically function as location. They may be viewed as an obstacle (the river during the cavalry fight) by the narrator, but also as a tool (the mountain to hide behind) by Hannibal. At the beginning of Livy’s story of Cannae (22.39), Q. Fabius Maximus points at the hostile space (in terms of inhabitants) for Hannibal and the favourable space of the Romans. Fabius is clearly capable of a strategic view on his surroundings both from his own, and from his adversary’s viewpoint. Hannibal, from his point of view, does not count on the support of inhabitants or other

268

van gils

such privileges. Instead, he makes use of natural locations, almost as if nature is there as a strategic weapon.22 We see how the same territory is used with different strategies by the Roman commanders and Hannibal respectively. Within this shared territory, camps are moved around by their commanders, serving various strategic goals. The Roman generals argue about moving their castra, while Hannibal builds, shows and hides his castra as malleable tools in his authoritarian hands. However, castra are not only viewed as the tools of generals, but also as locations for individuals. It may not only be a safe place where one could wait or return to (especially for the Romans) or a potential threat (the castra of the enemy in the cautious eyes of Paullus), but it could also be a prison from which to escape (for Hannibal himself and the Spanish soldiers when food starts to become scarce). And when Hannibal hides behind the mountains and leaves his castra with fires burning and booty visible, the place is viewed as a treasure to plunder by the Roman soldiers, but as an ambuscade to approach and enter with care by Paullus and Statilius. As in the case with territorial, unmovable places, also in the case of moveable places like castra, we find a contrast between Hannibal’s creative views of his surroundings, and the short-viewed perspective of the Roman soldiers and their general Varro. Of yet still different type are groups of soldiers, like wings, within an army. They are viewed as strategic groups which may be used as a tool by their general (who places, leads or sends them to a specific place) and they may move without an apparent controller during combat. In the last case they move, flee, attack, advance or retreat like an individual. In example 21 a group of soldiers acts on their own accord rather than from any plan or orders on the part of the generals, and the narrator clearly marks the unusualness of the situation with the comparison magis … quam. (21) Ceterum temeritati consulis ac praepropero ingenio materiam etiam fortuna dedit, quod in prohibendis praedatoribus tumultuario proelio a procursu magis militum quam ex praeparato aut iussu imperatorum orto haudquaquam par Poenis dimicatio fuit. Liv. 22.41.1

22

In 22.41, Hannibal hides his army behind the mountains and in 22.43, he decides to bring his men to warmer places.

livy’s use of spatial references in the cannae episode

269

But even Fortune furnished material to the recklessness and over-hasty temper of the consul. The repulse of a foraging party had led to a general mellay, which came about from the soldiers rushing forward to attack the enemy, rather than from any plan or orders on the part of the generals; and in this the Phoenicians by no means held their own. The fight described in example 21 ends with many more Phoenician dead than Romans, but when the victorious Romans want to go in pursuit of their enemy, we see that Paullus, just in time, holds them back (uictoribus obstitit), against their own and Varro’s will.23 Consul Paullus is portrayed more often as opposing the wish of his eager soldiers than as leading them into battle, in contrast to Varro who does lead the Roman army where the soldiers want to go. Hannibal easily commands various parts of his army by keeping them in line, sending them across the river or back again. Moreover, in the heat of the battle, Hannibal’s troops apparently effortlessly attack and envelop their adversaries or deceive them by faking surrender. Even in the case of large groups of hundreds of soldiers, their movements are described as if they concerned those of an individual.24 Actual individuals like Hannibal, Varro, Statilius and Paullus also stay, attack or flee, but unlike parts of the army, Hannibal and Paullus are both also portrayed as sitting, a posture suitable for contemplation. All protagonists view and contemplate space in their own subjective way: Q. Fabius Maximus viewed space in terms of friendly or hostile populations, Hannibal looked for support in the natural surroundings, and Varro and Paullus seem mostly occupied with moving (Varro) or resisting movement (Paullus). The focalising potential of individual characters in the story brings us back to Q. Fabius Maximus who warned Paullus against Hannibal’s strategic view and Varro’s lack of it, as discussed in the beginning of this chapter.

8

Conclusion

In this chapter, the use of space in the Cannae passage has been the subject of a narratological and linguistic analysis in order to increase our awareness of the role of the strategic uses of space by characters and by the narrator. After

23 24

Ceterum victoribus effuse sequentibus metu insidiarum obstitit Paullus consul […] Varrone indignante ac vociferante emissum hostem e manibus debellarique potuisse. (22.41) See, for instance, Afri circumdedere (22.47) or Quingenti Numidae desiliunt (22.48).

270

van gils

a short treatment of spatial references to the discourse situation and spatial references in a mainly text-structuring function, we had a look at strategic uses of space. With this analysis, I have drawn attention to the impact of spatial references for the alignment of the narratee with a particular perspective and to the implicit characterization of the main characters by showing their different attitudes in viewing and manipulating their environment. The Cannae episode in Livy offers many examples of strategically used spatial references. In fact, the narrative tension in the passage about the battle of Cannae in Livy is partly due to the different interpretations by the main characters of their spatial environment. We have Hannibal’s strategic view of the landscape as being apt for ambushes, flights, sieges or attacks, but also the Roman soldiers’ greedy and short-sighted reliance on their first impression of the presence or absence of the hostile army. The two Roman commanders, Varro and Paullus, repeatedly disagree about the advantages and risks of the various encampments and possible battlefields. It seems reasonable to conclude that in Livy’s account of the defeat of Cannae, spatial information in character’s perspective does not primarily describe the actual locations but, rather, portrays the main characters with their contrastive ambitions and differing strategic visions.25 I started my chapter with the speech of Q. Fabius Maximus and I would like to end with two more quotes from this general who advised Paullus before the battle of Cannae. (22) Haec una salutis est uia, L. Paule, quam difficilem infestamque ciues tibi magis quam hostes facient. Liv. 22.39.17

This is the only path of safety, Lucius Paullus, which your countrymen will render more difficult and dangerous to you than their enemies will. (23) Duobus ducibus unus resistas oportet. Resistes autem, aduersus famam rumoresque hominum si satis firmus steteris, si te neque college uana Gloria neque tua falsa infamia mouerit. Liv. 22.39.18

25

Characterizing the main characters is one of the six functions of space which have been distinguished in narratological studies, others being thematic, mirror, symbolic, psychologizing and personification. See, for instance, De Jong 2012.

livy’s use of spatial references in the cannae episode

271

You alone must resist two generals; and you will resist them sufficiently if you stand firm against the report and the rumours of men; if neither the empty glory of your colleague, and the unfounded calumnies against yourself, shall move you. Fabius tells Paullus there is only one way to safety and that is to be unmovable. These examples show yet another aspect of spatial analysis: metaphorical uses of space in argumentative discourse. However, this topic falls outside the scope of this chapter, which has been limited to investigating the narratological functions of spatial references.

Bibliography Adema, S.M., Discourse Modes and Bases, dissertation (Amsterdam 2008). Allan, R.J., ‘Towards a Typology of the Narrative Modes in Ancient Greek: Text types and narrative structure in Euripidean messenger speeches’, in S.J. Bakker & G.C. Wakker (eds.), Discourse Cohesion in Greek (Leiden 2009) [171–204]. Allan, R.J. & M. Buijs (eds.), The language of Literature. Linguistic approaches to classical texts (Leiden 2007). Feldherr, A. (ed.), The Cambridge Companion to the Roman Historians (Cambridge 2009). Gils, L.W. van, Argument and Narrative, dissertation (Amsterdam 2009). Jong, I.J.F. de, & Nünlist, R., ‘From Bird’s Eye View to Close-up. The Standpoint of the Narrator in the Homeric Epics’, in A. Bierl, A. Schmidt & A. Willi (eds.), Antike Literatur in neuer Deutung (Leipzig 2004) 62–83. Jong, I.J.F. de, Space in ancient Greek literature: studies in ancient Greek narrative. (Leiden 2012). Jong, I.J.F. de, Narratology & Classics. A Practical Guide (Oxford 2014). Kroon, C.H.M., ‘Discourse modes and the use of tenses in Ovid’s Metamorphoses’, in R.J. Allan & M. Buijs (eds.), The Language of Literature. Linguistic approaches to Classical Texts (Leiden 2007) 65–92. Levene, D.S., Livy on the Hannibalic War (Oxford 2010). Levinson, S.C., Space in Language and Cognition (Cambridge 2003). Maltby, R., A Lexicon of Ancient Latin Etymologies (Leeds 1991). Pausch, D., Livius und der Leser. Narrative Strukturen in ab urbe condita (München 2011). Richardson, S., The Homeric narrator (Nashville, Tenn. 1990). Riggsby, A., ‘Space’, in Feldherr (ed.), The Cambridge Companion to the Roman Historians: (Cambridge 2009) 152–165.

272

van gils

Roth, J., ‘Siege Narrative in Livy: Representation and Reality’, in S. Dillon & K. Welch (eds.), Representations of War in Ancient Rome (Cambridge 2006) 49–68. Ryan, M.L., Possible Worlds, Artificial Intelligence, and Narrative Theory (Bloomington 1991). Vaan, M. de, Etymological Dictionary of Latin and the other Italic Languages (Leiden 2008). Walsh, P.G., Livy: His Historical Aims and Methods (Cambridge 1961).

chapter 11

ET RATIO ET RES: Characterization of Roman Conduct through Speech Representation in the Battle of Cannae Michel Buijs

1

Introduction

In the account of the battle of Cannae, the Livian narrator overtly refers to the negative ending of the story he is narrating by means of a narratorial prolepsis:1 (1) ⟨ex⟩ maioris partis sententia ad nobilitandas clade Romana Cannas urgente fato profecti sunt. Liv. 22.43.9

The will of the majority prevailed, and they set forward, under the urge of destiny, to make Cannae famous for the calamity which there befell the Romans.2 In the coda to this narrative episode, the disaster is evaluated: (2) Haec est pugna ⟨Cannensis⟩, Alliensi cladi nobilitate par, ceterum ut illis quae post pugnam accidere leuior, quia ab hoste est cessatum, sic strage exercitus grauior foediorque. Liv. 22.50.1–22.50.2

Such was the battle of Cannae, a calamity as memorable as that suffered at the Allia, and though less grave in its results—because the enemy failed to follow up his victory—yet for the slaughter of the army even more grievous and disgraceful. 1 Compare Luce 1971: 295, on episodes of early Roman history in the AUC: ‘Indeed, it is imperative that the reader know in advance the outcome of such episodes; otherwise he cannot properly feel or appreciate the force of the narrative’. 2 The Latin text is that of the OCT; translations of Livy are based on B.O. Foster, Livy books xxi– xxii, with an English translation, London 1929, with slight adaptations.

© koninklijke brill nv, leiden, 2019 | doi:10.1163/9789004383340_012

274

buijs

After nobilitate par (cf. ad nobilitandas Cannas in [1]) and the opposition levior-gravior, the last word of the sentence, foedior, stands out as an additional evaluative term; it reminds the narratee of the praefatio to the Ab Urbe Condita:3 (3) hoc illud est praecipue in cognitione rerum salubre ac frugiferum, omnis te exempli documenta in inlustri posita monumento intueri; inde tibi tuaeque rei publicae quod imitere capias, inde foedum inceptu, foedum exitu, quod uites. Liv. praef. 10

What chiefly makes the study of history wholesome and profitable is this, that you behold the lessons of every kind of experience set forth as on a conspicuous monument; from these you may choose for yourself and for your own state what to imitate, from these mark for avoidance what is shameful in the conception and shameful in the result. Therewith, the Livian narrator has successfully incorporated this for a Roman audience negative episode of Roman history into his works, viz. as a negative example that is not to be followed.4 In the praefatio, the narratee is explicitly invited to ‘look at’ (te intueri) the lessons history has to offer as if looking at a monument,5 ‘which teaches (documenta) by providing paradigms of behaviour’.6 In my contribution to the present volume I hope to show that the Livian narrator uses reported speech as an effective means to show by what moral or strategic failure the Roman defeat came about.

2

Speech Representation, Character Portrayal and the Cunctandi Ratio

Throughout the Cannae episode, Roman conduct is characterized by words related to ratio or the absence of it: temeritas and cognates, consultus/inconsultus, cauere/cautus and cognates, providus/improvidus and cognates, ratio and

3 On Livy’s preface, see Moles 2009, with literature in note 2; on §10 of the preface, see Moles 2009: 70–75. Moles 2009: 65 speaks of ‘the crucial historical and historiographical claims of sections 9–10’ of the preface. 4 For other strategies in dealing with negative exempla in Livy, see Stem 2007. 5 On the visual language of the preface, see Feldherr 1998: 1–19. 6 Moles 2009: 73.

et ratio et res

275

ferox.7 The following table provides for each of these terms an overview of its occurrences (first column) and an indication of the target of the evaluation (second column): table 11.1

Characterization of Roman conduct

temeritas and cognates 22.38.12 22.39.20 22.41.1 22.41.4 22.43.1 22.44.5

in general in general (with hint at Varro) of the consul (sc. Varro) of the ferocior consul (sc. Varro) of the Romans of Sempronius and Flaminius (Paullus blames Varro)

reported speech by Paullus represented speech by Fabius narrator reported thought of Hannibal reported sight of Hannibal reported speech of Paullus

consultus/inconsultus 22.38.11 22.43.1 22.44.7

consultus in general inconsultus, of the Romans inconsultus, of Roman conduct (pugna)

reported speech by Paullus reported sight of Hannibal reported speech by Paullus

cauere/cautus and cognates 22.38.11 22.39.20 22.42.4 22.47.8

cautus, in general cautus, of Paullus cavere, of Roman conduct (advise) incautus, of Roman conduct

reported speech by Paullus represented speech by Fabius reported speech by Paullus narrator

consideratus/providus/improvidus and cognates 22.39.20 22.39.22 22.42.4 22.44.7

consideratus, of Paullus improvidus, of festinatio providere, of Roman conduct (advise) improvidus, of Roman conduct (pugna)

represented speech by Fabius represented speech by Fabius reported speech by Paullus reported speech by Paullus

7 Here I would like to thank professor Stephen Oakley for his keynote lecture given at the conference ‘Textual Strategies in Greek and Latin War Narrative’, held October 23–25, 2014 in Amsterdam, and for his handout entitled ‘Livy on Cannae: a literary overview’, from both of which I have benefited much.

276

buijs

Table 11.1

Characterization of Roman conduct (cont.)

ratio 22.39.9 22.39.10 22.39.21

of Roman conduct, against Hannibal of Roman conduct, in general of Roman conduct, relating to Paullus

represented speech by Fabius represented speech by Fabius represented speech by Fabius

of contiones by Varro of the consul (sc. Varro)

narrator reported thought of Hannibal

ferox 22.38.6 22.41.4

As the third column of table 11.1 shows, the Livian narrator characterizes Roman conduct overwhelmingly through represented speech or reported speech/ thought/sight. A crucial role is played by the long, represented speech delivered by Fabius Maximus to the consul Paullus before the two consuls head for Cannae (22.39). A question to be asked in cases like this, is: why do we get such a long speech at all? The answer to this question should be: to prime the primary narratees of the Ab Urbe Condita in a quasi-objective way with themes and motives that will play an important role in the sequel of the narrative.8 Especially since Fabius’ speech is most probably made up by Livy,9 we are allowed to regard the speech as an important structural element in the story that is about to begin.10 Space forbids to represent the speech in full length here; the following overview presents its major features. Fabius starts by stating that if the two consuls had been comparable, to the extent that they had both been good consuls (22.39.2 duo boni consules) or two bad ones (22.39.2 mali), his words would have been superfluous. The words ‘good’ and ‘bad’ are introduced, and as it is, seeing what Varro is like and what Paullus is like (22.39.3 nunc et collegam tuum et te talem uirum intuenti), Fabius can deliver a speech to Paullus only, thereby qualifying him as ‘good’ and the

8 9 10

Canter 1913: 32 draws attention to the fact that the speech of Q. Fabius Maximus urges vigilance and the cunctandi ratio as the proper campaign against Hannibal. On the ‘Ungeschichtlichkeit der Fabiusrede’, see Bruckmann 1936: 74. See Burck 1950: 93–94, who speaks of ‘eine erhöhte kompositionelle Bedeutung’ of Fabius’ speech, which also provides us with ‘eine Bewertung der Persönlichkeiten und Ziele beider Consuln […], die Livius über den Moment hinaus festgehalten wissen will.’

et ratio et res

277

other, Varro, as ‘bad’.11 Fabius continues by explicitly calling Paullus a good man and a good citizen (22.39.3 et uirum bonum et ciuem); these qualities in Paullus will however be to little purpose, if ‘the state is lame on the other side and evil counsels enjoy the same rights and the same authority as good’ (22.39.3 si altera parte claudente re publica malis consiliis idem ac bonis iuris et potestatis erit). Here, with malis consiliis … ac bonis, the terms ‘good’ and ‘bad’ are narrowed down from being applied to a person in general to his conduct or tactics (consilia). The theme of decision-making has been introduced. Fabius compares Varro with Hannibal (22.39.4–22.39.512 and 22.39.18) and further states that Varro is, and has always been, out of his wits (22.39.6). According to Fabius, the one and only way of conducting a war with Hannibal is the way in which he himself conducted it (22.39.9 una ratio belli gerendi aduersus Hannibalem est qua ego gessi). The word ratio recurs in 22.39.10, when Fabius continues: ‘Not only the event—that schoolmaster of fools—teaches us this, but the same reasoning (ratio) which held good then will hold unchanged, so long as circumstances remain the same’ (Nec euentus modo hoc docet— stultorum iste magister est—sed eadem ratio quae fuit futuraque donec res eaedem manebunt, immutabilis est). In 22.39.11–22.39.15, Fabius points at the opportunities offered by the fact that the Romans will fight in their own country, while Hannibal is on foreign soil and ‘subsists on the plunder of each day’ (22.39.13 in diem rapto uiuit), and already more of his soldiers have perished by starvation than by the sword, and the few that are left have no longer any food (22.39.14). Therefore, he asks a rhetorical question (22.39.15),13 summing up his entire point, viz. that it is better to refrain from immediate action and wait (sedendo): ‘Can you doubt then that if we sit still we must gain the victory over one who is growing weaker every day and is destitute of provisions, of replacements, and of money?’ (Dubitas ergo quin sedendo superaturi simus eum qui senescat in dies, non commeatus, non supplementum, non pecuniam habeat?). At the end of his speech, Fabius sums up his argument: one should not hurry, there should be no haste (22.39.22 Omnia non properanti clara certaque erunt; festinatio improuida est et caeca; ‘All things will be clear and definite to one who

11 12 13

Bruckmann 1936: 73 points out that the reader may be influenced by Fabius’, in itself, subjective opinion on Varro, especially since Varro is not given the floor. Cf. Levene 2010: 189, n. 57: ‘The suggestion that it is Varro who is the hostis, while Hannibal is merely the adversarius, is especially striking’. Canter 1917: 139 states about the rhetorical question in Livy’s direct speeches that ‘[i]t occurs most frequently when the speaker is impulsive and passion raises high’ and that ‘[f]or the opposite reasons it is used […] not at all by Q. Fabius urging the cunctandi ratio’. Apparently Canter does not take this question as a rhetorical one.

278

buijs

does not hurry. Haste is improvident and blind’). His advice is not that Paullus should refrain from taking action, but simply that in acting, reason instead of fortune should lead him—Nec ego ut nihil agatur ⟨hortor⟩ sed ut agentem te ratio ducat, non fortuna (22.39.21). Fabius’ speech is easily summarized: a good consul would act in accordance with the public interest and his own conscientiousness; good leadership is characterized by good counsel; good counsel is brought about by ratio; in the situation at hand ratio leads one to the conviction that it is better to wait than to act in a rash fashion. Paullus is allotted one last oratio before he sets out; it is given in indirect speech (22.40.1–22.40.3). What should strike us is the fact that Paullus considers what Fabius said was true (uera), albeit difficult to put into practice (magis … quam facilia factu—as the res to be narrated will unequivocally show), given the fact that his colleague is a turbulent and headstrong man. Just before the fighting will start, the two consuls are, once more, characterized through reported speech: Paullus on a par with Fabius as the prudent, reasoning leader, and Varro as unadvised and audacious. The speech is of grim nature (haud sane laeta—Paullus merely hopes for the best), and foreshadows the negative tale that is about to begin. By way of contrast, let us look at the way in which Polybius, one of Livy’s predecessors in writing about Cannae, characterizes the two consuls: (4) ἐν ᾧ καιρῷ καὶ Λεύκιος Αἰμίλιος περιπεσὼν βιαίοις πληγαῖς ἐν χειρῶν νόμῳ μετήλλαξε τὸν βίον, ἀνὴρ πάντα τὰ δίκαια τῇ πατρίδι κατὰ τὸν λοιπὸν βίον καὶ κατὰ τὸν ἔσχατον καιρόν, εἰ καί τις ἕτερος, ποιήσας. Polyb. 3.116.914 It was at this point that Lucius Aemilius fell, in the thick of the fight, covered with wounds: a man who did his duty to his country at that last hour of his life, as he had throughout its previous years, if any man ever did. (5) ὀλίγοι δέ τινες εἰς Οὐενουσίαν διέφυγον, ἐν οἷς ἦν καὶ Γάιος Τερέντιος ὁ τῶν Ῥωμαίων στρατηγός, ἀνὴρ αἰσχρὰν μὲν τὴν ψυχὴν ἀλυσιτελῆ δὲ τὴν ἀρχὴν τὴν αὑτοῦ τῇ πατρίδι πεποιημένος. Polyb. 3.116.13

14

The text of Polybius is that of T. Büttner-Wobst, Polybii historiae, Leipzig: Teubner; translations of Polybius by E.S. Shuckburgh, Polybius, Histories, London/New York, 1889.

et ratio et res

279

Some few escaped into Venusia, among whom was Gaius Terentius, the Consul, who thus sought a flight, as disgraceful to himself, as his conduct in office had been disastrous to his country. In the midst of his narrative, the Polybian narrator characterizes either consul in an apposition opening with the word ἀνήρ and closing with a participle of the verb ποιέω: ‘a man who did/had made …’, thereby telling his addressee what kind of man either consul had been at the moment they are brought off the stage. On the basis of the preceding overview of the Livian way of characterizing both protagonists we should conclude 1. that inserting Fabius’ speech provides the Livian narrator with the opportunity to frame Paullus as a good man and a good leader possessing all the qualities belonging to good leadership, and Varro as his opposite; 2. that this procedure aims at putting all the blame for the disaster to come on Varro;15 3. that therewith the Livian narrator achieves his persuasive goal of teaching his narratee that from the study of history one can learn to avoid what is shameful in conception and shameful in result, as set out in praefatio 10.16 2.1 Speech Representation II: the Death of Aemilius Paullus At the end of the account of the battle, the disaster is complete. Fabius’ precepts are once more referred back to, by the dying Lucius Aemilius Paullus. The death of Aemilius Paullus constitutes a remarkable scene. Paullus, wounded as he is, is not able to control his horse, and his horsemen dismount. They stand their ground as good as they can, but are either cut down or put to flight. With the dispersion of those who survived, the episode seems to have come to an end, but a small scene that in the grand scheme of things seems rather irrelevant 15

16

Cf. Walsh 1961: 231: ‘The eloquent and prophetic oration attributed to Fabius and the fatalistic words of Aemilus were clearly composed long after the battle in denigration of Varro; Polybius has no mention of these forebodings, and Livy is acquainted with his version’; see also Walsh 1961: 72; and Bruckmann 1936: 79–80. Varro’s background is sketched in 22.25.18–22.26.4, his popularity with the plebeians in 22.34.2; only at the end of book 22, long after the account of the battle has ended, the Livian narrator explicitly holds Varro responsible for the disaster at Cannae (22.61.14: ex tanta clade, cuius ipse causa maxima fuisset: ‘from that defeat for which he himself (sc. Varro) had been chiefly responsible’). Cf. Luce 1971: 275: ‘Livy’s desire to teach a moral lesson is clearly responsible for the selection, ordering and emphasis of material. The narrative corresponds to the programme announced in the Preface.’ On the role of rhetoric in Roman historiography in general, see Laird 2009, especially 204–208 for a discussion of the exchange of orations between Fabius Maximus Cunctator and Scipio Africanus in Livy 28.40.1–28.45.9; on structure in Livy’s speeches, see Luce 1993.

280

buijs

follows: we are presented with a dialogue between Gnaeus Lentulus, a tribune of the soldiers, and Lucius Aemilius Paullus, who is sitting on a stone, covered with blood: (6) Inde dissipati omnes sunt, equosque ad fugam qui poterant repetebant. Cn. Lentulus tribunus militum cum praeteruehens equo sedentem in saxo cruore oppletum consulem uidisset, ‘L. Aemili’ inquit, ‘quem unum insontem culpae cladis hodiernae dei respicere debent, cape hunc equum, dum et tibi uirium aliquid superest ⟨et⟩ comes ego te tollere possum ac protegere. Ne funestam hanc pugnam morte consulis feceris; etiam sine hoc lacrimarum satis luctusque est.’ Ad ea consul: ‘tu quidem, Cn. Corneli, macte uirtute esto; sed caue, frustra miserando exiguum tempus e manibus hostium euadendi absumas. Abi, nuntia publice patribus urbem Romanam muniant ac priusquam uictor hostis adueniat praesidiis firment; priuatim Q. Fabio L. Aemilium praeceptorum eius memorem et uixisse [et] adhuc et mori. Me in hac strage militum meorum patere exspirare, ne aut reus iterum e consulatu sim ⟨aut⟩ accusator collegae exsistam ut alieno crimine innocentiam meam protegam.’ Haec eos agentes prius turba fugientium ciuium, deinde hostes oppressere; consulem ignorantes quis esset obruere telis, Lentulum in tumultu abripuit equus. Liv. 22.49.5–22.49.12

The survivors were now all dispersed, and those who could attempted to regain their horses and escape. Gnaeus Lentulus, a tribune of the soldiers, as he rode by on his horse, caught sight of the consul sitting on a stone and covered with blood. ‘Lucius Aemilius,’ he cried, ‘on whom the gods ought to look down in mercy, as the only man without guilt in this day’s disaster, take this horse, while you have still a little strength remaining and I can attend you and raise you up and guard you. Make not this battle calamitous by a consul’s death; even without that there are tears and grief enough.’ To this the consul answered, ‘All honour, Cornelius, to your manhood! But waste not in unavailing pity the little time you have to escape the enemy. Go, and tell the senators in public session to fortify the City of Rome and garrison it strongly before the victorious enemy draws near: in private say to Quintus Fabius that Lucius Aemilius has lived till this hour and now dies remembering his precepts. As for me, let me breathe my last in the midst of my slaughtered soldiers, lest either for a second time I be brought to trial after being consul, or else stand forth the accuser of my colleague, blaming another in defence of my own innocence.’ While they were speaking, there came up with them first a crowd of fleeing Romans,

et ratio et res

281

and then the enemy, who overwhelmed the consul, without knowing who he was, beneath a rain of missiles. Lentulus, thanks to his horse, escaped in the confusion. The pace of the narrative suddenly slows down,17 and the Livian narrator spends much attention on the death of the consul Aemilius Paullus. In contrast, just a few final words are spent on Varro, who (22.49.14 seu forte seu consilio; ‘whether by accident or by design’) had not joined any throng of fugitives, but fled to Venusia with some fifty horsemen. The death of Aemilius Paullus receives considerably more attention. Again we might ask ourselves the question: Why are we presented with a dialogue at all? The point of the pictorial scene18 is that by opting for a dialogue, Aemilius is enabled to utter the words: ‘Go, and tell the senators in public session to fortify the City of Rome and garrison it strongly before the victorious enemy draws near: in private say to Quintus Fabius that Lucius Aemilius has lived till this hour and now dies remembering his precepts’. Aemilius Paullus gives two instructions here: one to be announced publice and one privatim, namely to Quintus Fabius Maximus. Aemilius Paullus lived and now dies remembering Fabius’ precepts (praeceptorum eius memorem). The Livian narrator apparently opts for scenic narration to be able to quote the final words of Aemilius Paullus in direct speech, and therewith establishes a ring composition: the two speeches frame the narrative account of the battle of Cannae proper. The narratee is reminded of Fabius’ advice ‘ut agentem te ratio ducat, non fortuna’ (22.39.21; cf. praeceptorum eius memorem) and the characterization of Roman conduct throughout as presented in Table 11.1, at the moment he has been told what disastrous events happened and has been shown how they came about.

3

The Res: the Story Proper

The events of the battle itself (22.40.4sqq.) clearly illustrate the importance of the difference in character between the two consuls, and at the same time makes clear how Fabius was right in claiming that good leadership is based on ratio.19 I shall primarily focus on the distribution of verbs and nouns of cognition among the rivalling parties. 17 18 19

The fact that Gnaeus Lentulus is asyndetically introduced into the narrative is concomitant with the sudden change from straightforward narrative to scenic rhythm. Bruckmann 1936: 86 speaks of ‘[e]ine bildhafte Szene’. Cf. Burck 1950: 94: ‘So nahe lag also unmittelbar vor Cannae dank der richtigen Taktik des

282

buijs

When Varro and Paullus get to the camp, the consuls of the year before are quickly dealt with: they sent Marcus Atilius back to Rome and put Geminus Servilius in command of the smaller camp. What interests us here is the reaction of Hannibal: (7) Hannibal quamquam parte dimidia auctas hostium copias cernebat, tamen aduentu consulum mire gaudere. Non solum enim nihil ex raptis in diem commeatibus superabat sed ne unde raperet quidem quicquam reliqui erat, omni undique frumento, postquam ager parum tutus erat, in urbes munitas conuecto, ut uix decem dierum, quod compertum postea est, frumentum superesset Hispanorumque ob inopiam transitio parata fuerit, si maturitas temporum exspectata foret. Liv. 22.40.7–22.40.9

Hannibal, though he perceived that the forces of his enemies were augmented by a half, was nevertheless greatly rejoiced at the coming of the consuls. For not only were the spoils exhausted on which his men had subsisted from day to day, but there was not even any district left for them to spoil; for when it appeared that the farms were no longer safe, the corn had everywhere been carried into the walled towns, and in consequence there was barely grain enough left—as was afterwards discovered—to last ten days, and the Spaniards, for want of food, had made ready to desert, if the Romans had only waited till the time was ripe. Embedding Hannibal’s focalization in his narrator-text, the Livian narrator informs the narratee of what Hannibal saw (cernebat) and thus that he realised that his adversaries had gained strength and that therewith his own situation had deteriorated, as well as Hannibal’s reaction to the new situation: he rejoiced (gaudere). The fact that Hannibal greatly rejoiced can, under the circumstances, be called strange. Therefore, in a long sentence marked by enim, the narrator explains Hannibal’s mental disposition, and endorses it by referring to external, albeit anonymous, authority (quod compertum postea est). The last words of the citation, si maturitas temporum exspectata foret, do not mean ‘till the time was ripe for the Romans to fight’, but ‘till the time had come to Fabius, die Paulus fortzuführen gewillt war, die Chance eines großen Erfolgs über Hannibal, ja einer Kriegswendung, wenn nicht Varro diese Siegesmöglichkeit zunichte gemacht hätte.’ Levene 2010: 197 notes on Fabian tactics: ‘In the context of Cannae Fabius is clearly correct, but as the debate in book 28 will show, his approach is misguidedly narrow in the context of the war as a whole.’

et ratio et res

283

full maturity’, i.e. if the Romans had waited till the others deserted, and had not attacked or fought before.20 It should be noted that in his single speech before he left, Paullus had already announced that he would follow a cautious course (22.38.11 se, quae consilia magis res dent hominibus quam homines rebus, ea ante tempus immatura non praecepturum; ‘that he would not anticipate, before they ripened, those plans with which circumstances provided men but which men could not well impose on circumstances’). The verbal echo of ante tempus immatura in si maturitas temporum exspectata foret is conspicuous, and it should be noted, too, that Paullus spoke of consilia that had not fully matured, so that, by this very verbal echo, the Livian narrator reminds his narratee that Paullus’ line of conduct would have been the right one. Moreover, the situation is more or less as Fabius Maximus had predicted in his speech (see especially 22.39.13–22.39.15), and even worse for Hannibal—as was afterwards discovered. By now, the narratee knows Paullus and Fabius were right: the Romans should have waited a little longer. The upshot is that both the narratee and Hannibal have full understanding of the situation and its implications, whereas the Roman troups dramatically do not. Next, the Livian narrator brings Fortune ( fortuna) into play, and overtly refers to the rashness and over-hasty temper of the consul Varro—this being one of the few instances (see Table 11.1) where the Livian narrator characterizes Varro by referring to his temeritas:21 (8) Ceterum temeritati consulis ac praepropero ingenio materiam etiam fortuna dedit, quod in prohibendis praedatoribus tumultuario proelio ac procursu magis militum quam ex praeparato aut iussu imperatorum orto haudquaquam par Poenis dimicatio fuit. Ad mille et septingenti caesi, non plus centum Romanorum sociorumque occisis. Liv. 22.41.1–22.41.2

But even Fortune furnished material to the recklessness and over-hasty temper of the consul. The repulse of a foraging party had led to a general

20

21

Weissenborn/Müller ad maturitas temporum: ‘die Reife (volle Entwickelung) der Zeitmomente = ‘der rechte Zeitpunkt’; vgl. Zon. 9.1: εἴ γε καὶ τὸ βραχύτατον ἐπεσχήκεσαν (die Konsuln), ἀπόνως ἐκράτησαν ἄν’ (if only they had waited for even the shortest period of time, they would have won easily). For the narrator-oriented narrative mode adopted here, note materiam etiam fortuna dedit, functioning as an ‘Abstract’ (Labov 1972), and the high pace of the narration. See also Van Gils & Kroon (this volume) for a discussion of Labovian narrative phases in the Cannae episode.

284

buijs

mellay, which came about from the soldiers rushing forward to attack the enemy, rather than from any plan or orders on the part of the generals; and in this the Phoenicians by no means held their own. About seventeen hundred of them were slain and not more than a hundred of Romans and allies. The Romans achieve a small successful result. The fact that the success of the Romans is brought about by fortuna (instead of by ratio/consilium), is underlined by the fact that the general mellay came about from the soldiers rushing forward to attack the enemy, rather than from any plan or orders on the part of the generals. On the role of fortuna in general, Levene notes:22 ‘Fortuna rarely appears in battles as a meaningful causal agent,’ and provides the following footnote: One exception is 22.41.1–2, where a minor Roman victory is directly referred to as something that happened by fortuna. In this case the chief point appears to be that is was indeed effectively a random result: the soldiers are not acting under direction of their generals, and the battle is described as ‘chaotic’ (22.41.1 tumultuario). But this is clearly atypical of the battles of the Hannibalic War (my italics). The fact that this is an atypical battle scene indicates that something special is going on. The narrator focuses on the opposition ratio vs. fortuna, and reminds the narratee of Fabius’s words of advice spoken to Paullus: ut agentem te ratio ducat, non fortuna (22.39.21 ‘that reason and not fortune should be your guide’). Then the Roman consuls start quarrelling again; Paullus at least displays some element of ratio, as from fear of an ambuscade he checks the victors in their headlong pursuit, despite the vociferous remonstrances of Varro, who cries out that they had let the enemy slip through their hands and that they might have brought the war to a conclusion if they had not relaxed their efforts. Interesting again is Hannibal’s reaction to his defeat: (9) Hannibal id damnum haud aegerrime pati; quin potius credere23 uelut inescatam temeritatem ferocioris consulis ac nouorum maxime militum esse. Et omnia ei hostium haud secus quam sua nota erant: dissimiles discordesque imperitare, duas prope partes tironum militum in exercitu esse.

22 23

Levene 2010: 285. OCT: credere PCRMDA: gaudere Pluygers, audacter.

et ratio et res

285

Itaque locum et tempus insidiis aptum se habere ratus, nocte proxima nihil praeter arma ferente secum milite castra plena omnis fortunae publicae priuataeque relinquit, transque proximos montes laeua pedites instructos condit, dextra equites, impedimenta per conuallem mediam traducit, ut diripiendis uelut desertis fuga dominorum castris occupatum impeditumque hostem opprimeret. Crebri relicti in castris ignes, ut fides fieret dum ipse longius spatium fuga praeciperet falsa imagine castrorum, sicut Fabium priore anno frustratus esset, tenere in locis consules uoluisse. Liv. 22.41.4–22.41.9

Hannibal was not greatly disconcerted by this reverse; indeed he believed that the hook should have been baited, as it were, for the rashness of the more impetuous consul, and especially for that of the new soldiers. All the circumstances of his enemies were as familiar to him as his own: that their generals were unlike each other and were at loggerheads, and that nearly two-thirds of their army consisted of recruits. Believing, therefore, that place and time were favourable for a ruse, he left his camp full of every sort of public and of private riches, and putting himself at the head of his troops, who carried nothing but their weapons, marched over the nearest ridge, drew up the infantry in ambush on the left, and the cavalry on the right, and made the baggage-train pass through the valley between, intending to fall upon the enemy whilst they were preoccupied and encumbered with the pillage of the camp, which would seem to them to have been deserted by its owners. He left a large number of fires burning, as though he had sought by means of this illusory appearance of an encampment to hold the consuls to their positions—as he had cheated Fabius the year before—till he could gain as long a start as possible in his retreat. Hannibal’s reaction to the situation is presented by the narrator by means of a second and third historical infinitive (after 22.40.7 gaudere, here pati and credere). It is telling that the historical infinitives attributed to Hannibal thus far all three present his mental disposition, and that the two other historical infinitives we find in the account of the battle, attributed to the Romans, are verbs of discontent and turmoil: (10) Paullus etiam atque etiam dicere prouidendum praecauendumque esse. Liv. 22.42.4

Paullus kept insisting on the need for watchfulness and circumspection.

286

buijs

(11) Inde rursus sollicitari seditione militari ac discordia consulum Romana castra. Liv. 22.44.5

This caused the camp of the Romans to be once more in anxiety because of strife amongst the soldiers and dissension between the consuls. To all appearances, the Livian narrator uses the historical infinitive to characterize the Romans as quarrellers and Hannibal as the rational combatant,24 using his ratio in the way the Romans—according to Fabius, and Aemilius, note prouidendum praecauendumque—should have done.25 The contrast could not have been greater. In [9], Hannibal seems to be in complete control of the situation. The Livian narrator informs his narratee that all the circumstances of Hannibal’s enemies were as familiar to him as his own.26 He is calculating, using his ratio, when believing (ratus)27 that place and time are favourable for a ruse. The actions intended to deceive the Romans, i.e. his stratagem, are presented by three historical presents:28 he leaves (relinquit) his camp full of every sort of public and of private riches, and putting himself at the head of his troops, who carried nothing but their weapons, marched over the nearest ridge, draws up (condit) infantry in ambush on the left, and the cavalry on the right, and makes the baggage-train pass through (traducit) the valley between. In order to paint Hannibal as the rational combatant, the omniscient narrator even informs the narratee of Hannibal’s considerations and intentions in two utclauses. In the remainder of the account of the battle, Hannibal remains in control of the situation, even when the situation turns out not to be favourable for him, and verbs and nouns of cognition are overwhelmingly attributed to him: he sees things, conceives hope, considers, and acts accordingly—the actions

24 25 26

27 28

Cf. Bruckmann 1936: 76. On Hannibal’s rationality, see Levene 2010: 152–154. Hannibal’s use of ratio is described by Polybius at 3.47.6–3.48.12. For the use of the historical infinitive (or infinitive of narration) in characterization and in marking dominant themes, see Viljamaa 1983: 71–76. Cf. Bruckmann 1936: 76: ‘In dieser Kenntnis Hannibals dürfen wir also ein drittes Moment neben dem Streit der Führer und dem Wirken der fortuna sehen, das die Lage der Römer noch verschlechtert.’ Note that the noun ratio is derived from the perfect participle ratus. See Van Gils & Kroon (this volume) for an analysis of the historical presents in this passage.

et ratio et res

287

often being highlighted by historical presents. Some illustrative examples (in the following examples, historical presents marking Hannibal’s stratagem are green): (12) Hannibal postquam motos magis inconsulte Romanos quam ad ultimum temere euectos uidit, nequiquam detecta fraude in castra rediit. Ibi plures dies propter inopiam frumenti manere nequit, nouaque consilia in dies non apud milites solum mixtos ex conluuione omnium gentium sed etiam apud ducem ipsum oriebantur. Nam cum initio fremitus, deinde aperta uociferatio fuisset exposcentium stipendium debitum querentiumque annonam primo, postremo famem, et mercennarios milites, maxime Hispani generis, de transitione cepisse consilium fama esset, ipse etiam interdum Hannibal de fuga in Galliam dicitur agitasse ita ut relicto peditatu omni cum equitibus se proriperet. Cum haec consilia atque hic habitus animorum esset in castris, mouere inde statuit [historical present?] in calidiora atque eo maturiora messibus Apuliae loca, simul ut, quo longius ab hoste recessisset, eo transfugia impeditiora leuibus ingeniis essent. Profectus est nocte ignibus similiter factis tabernaculisque paucis in speciem relictis, ut insidiarum par priori metus contineret Romanos. Liv. 22.43.1–22.43.6

Hannibal, perceiving that the Romans, although they had acted illadvisedly, had not proceeded to the extremity of rashness, returned to the camp, his stratagem having been detected and rendered idle. There, however, the scarcity of corn forbade his remaining many days, and new plans were daily forming, not only amongst the soldiers, the mingled offscourings of every race on earth, but even in the mind of the general himself. For when the men, with murmurs at first and afterwards with loud clamours, demanded their arrears of pay, and complained at first of the scarcity of corn, and finally of being starved; and when the report went round that the mercenaries—particularly those of Spanish blood—had resolved on going over to the enemy, they say that even Hannibal himself had thoughts of abandoning all his infantry and saving himself and his cavalry by escaping into Gaul. Such being the projects that were entertained in camp and such the temper of his soldiers, he decided to move from his present quarters to Apulia, where the climate was warmer and in consequence of this the harvest earlier; at the same time it would be the more difficult, the greater their distance from the enemy, for those of his followers who were fickle to desert. He set out in the night, after making up some fires, as before, and leaving a few tents standing where they

288

buijs

would be seen, so that the Romans might be withheld from following him through fear of an ambush, as before. The fact that the Romans had not pushed their rashness to the extreme, yet had acted inconsiderately, is focalized by Hannibal (uidit), whose stratagem has failed and now returns to his camp. In the remainder of the passage, the Livian narrator informs his narratee of the stressful situation in Hannibal’s camp and the ensuing decision of the latter to depart for Apulia—and the considerations on the basis of which this decision was taken.29 Again, the narratee, unlike the Roman troops, possesses the same amount of information as Hannibal does. (13) Hannibal spem nanctus locis natis ad equestrem pugnam, qua parte uirium inuictus erat, facturos copiam pugnandi consules, dirigit aciem lacessitque Numidarum procursatione hostes. Liv. 22.44.4

Hannibal had conceived a hope that the consuls would give him an opportunity of fighting in a place that was formed by nature for a cavalry action, in which arm he was invincible. He therefore drew out his men in battle array and ordered the Numidians to make a sally and provoke the enemy. Once more, the Livian narrator informs his narratee about Hannibal’s mental disposition (spem nanctus) and stratagem (historical presents dirigit and lacessit). (14) Dum altercationibus magis quam consiliis tempus teritur, Hannibal ex acie, quam ad multum diei tenuerat instructam, cum in castra ceteras reciperet copias, Numidas ad inuadendos ex minoribus castris Romanorum aquatores trans flumen mittit. Liv. 22.45.1–22.45.2

While they [sc. the Romans] wasted time, rather quarrelling than consulting, Hannibal withdrew the rest of his troops, whom he had kept in line till far on in the day, into his camp, and sent the Numidians across the river to attack the men from the smaller Roman camp who were fetching water.

29

Cf. Bruckmann 1936: 78: ‘[D]er Grund, den Livius für Hannibals Abmarsch angibt, ist ein Beweis der Richtigkeit der Kriegstaktik des Paulus.’

et ratio et res

289

Hannibal’s stratagem is disclosed (historical present mittit), while in the dum-clause, the Romans are, in contrast, presented as quarrelling, rather than consulting, and wasting time. Here, themes and motives as introduced in represented/reported speech merge: prudent conduct based on ratio (caute ac consulte, 22.38.11; see Table 11.1) is impossible in view of the differing character of the two consuls, and so are decisions based on a considering together (consilia), and quarrelling ensues. The Romans are, in fact, constantly presented by the Livian narrator as quarrelsome; words belonging to the word field ‘to consult’ abound.30 Let us look back at 22.43: (15) Sed per eundem Lucanum Statilium omnibus ultra castra transque montes exploratis, cum relatum esset uisum procul hostium agmen, tum de insequendo eo consilia agitari coepta. Cum utriusque consulis eadem quae ante semper fuisset sententia, ceterum Varroni fere omnes, Paulo nemo praeter Seruilium, prioris anni consulem, adsentiretur, ⟨ex⟩ maioris partis sententia ad nobilitandas clade Romana Cannas urgente fato profecti sunt. Prope eum vicum Hannibal castra posuerat aversa a Volturno vento, qui campis torridis siccitate nubes pulveris vehit. Id cum ipsis castris percommodum fuit, tum salutare praecipue futurum erat cum aciem dirigerent, ipsi aversi terga tantum adflante vento in occaecatum pulvere offuso hostem pugnaturi. Liv. 22.43.7–22.43.10

But when the same Lucanian, Statilius, had made a thorough reconnaissance beyond the camp and on the other side of the mountains, and had reported seeing the enemy on the march a long way off, then the question of pursuing him began to be debated. The consuls were each of the same mind as they had always been; but Varro had the support of almost everybody, Paullus of none except Servilius, the consul of the year before. The will of the majority prevailed, and they set forward, under the urge of destiny, to make Cannae famous for the calamity which there befell the Romans. This was the village near which Hannibal had pitched his camp, with his back to the Volturnus, a wind that brings clouds of dust over the drought-parched plains. Such a disposition was very convenient for the camp itself and bound to be particularly salutary when the troops formed up for battle, facing in the opposite direction, with the wind blowing only

30

Note that consul, consilium, and inconsultus are all derivatives of consulo, ‘to consult’.

290

buijs

on their backs, and ready to fight with enemies half-blinded by the dust driven into their faces. The lack of agreement in the Roman camp is, by means of the narratorial prolepsis quoted as example [1] (ad nobilitandas clade Romana Cannas), directly linked to the disaster that is about to befall them. Moreover, the narrator contrasts the debating Romans to the cunning Hannibal, by informing his narratee in the very next sentence about Hannibal’s choice of a favourable position for his camp and for the battle to come. A final example showing the disaccord among the Romans is the following, in which the lack of agreement is pushed to the limit: there is even lack of communication among the Roman consuls (nihil consulto collega), and Aemilius Paullus has to assist in a plan (consilium) of his colleague that he does not even approve of: (16) Itaque postero die Varro, cui sors eius diei imperii erat, nihil consulto collega signum proposuit instructasque copias flumen traduxit, sequente Paulo quia magis non probare quam non adiuuare consilium poterat. Liv. 22.45.5

The next morning, therefore, Varro, whom the lot had made commander for that day, hung out the signal, without saying a word of the matter to his colleague, and, making his troops fall in, led them over the river. Paullus followed him, for he could more easily disapprove the plan than deprive it of his help.

4

Conclusion

The Livian narrator uses represented/reported speech to introduce elements that turn out to be of seminal importance for the way he wants to present his version of the battle of Cannae. In speech, the characters of the main Roman protagonists are portrayed, and attention is drawn to the theme of deliberate decision making on the basis of ratio instead of on the basis of fortune: good leadership is characterized by good counsel and good counsel is brought about by ratio. Close-reading has revealed that, surprisingly, Hannibal takes counsel time and again, even or especially in situations that are not favourable to him, while the Romans, because of Varro’s temper, act in a rash fashion where waiting would have been better. Aemilius Paullus would have kept Fabius’ precepts, as the dialogue at the end of the battle episode shows, but the res prevented him

et ratio et res

291

to do so. Therewith, Hannibal’s victory is easily explained: of the three main participants, Paullus is first framed as good and then shown to be prudent but not capable, Varro is first framed as bad and then shown to be not prudent and therefore not capable, while Hannibal is characterized through his considerations and actions as prudent and capable. That Aemilius is presented as the good leader and Varro as the bad one is not a revolutionary invention in character portrayal that can be ascribed to the Livian narrator—compare the presentation of both consuls by the Polybian narrator. What is special is that the Livian narrator shows, rather than tells, how the Roman defeat came about by making use of character portrayal through speech, underscored by verbs and nouns of cognition in his narrative, and by the distribution of historical infinitives. This showing instead of telling is, I contend, in line with one of the aims in writing history the Livian narrator had in mind, inviting his narratee in the praefatio to the work as a whole to look at the lessons history has to offer as if looking at a monument, and to choose for himself and for his own state what to imitate, and to mark for avoidance what is shameful in the conception and shameful in the result. The Roman conduct at Cannae was certainly foedum inceptu and foedum exitu, even more so for the slaughter of the army. We may conclude that the Livian narrator focuses on aspects of strife among the Romans, on the one hand, and on deliberation and tactics on the part of Hannibal, on the other, and that by keeping his narratee on a par with Hannibal as far as insight in the situation at hand is concerned, he shows how Hannibal behaves as the Romans should have behaved,31 if only they had followed Fabius’ advice.

Bibliography Bruckmann, H., Die römischen Niederlagen im Geschichtswerk des T. Livius (BochumLangendreer 1936). Burck, E., Einführung in de dritte Dekade des Livius (Heidelberg 1950). Canter, H.V., ‘Livy the Orator’, CJ 9.1 (1913) 24–34.

31

The conclusion that roles are reversed may seem daring, but cf. Luce 1971: 269 on Book 5: ‘Livy takes great pains to make the contrast on either side of each peripeteia as sharp and as striking as possible. One method is by exaggeration, even to the point of paradox. Contrast is carried so far that on occasion Livy comes near to depicting a reversal of identity. Before the first peripeteia the Romans are more like the Gauls that themselves […]. Similarly, Roman characteristics belong (temporarily) to the Gauls […]. But at the final denouement the reversal is complete.’

292

buijs

Canter, H.V., ‘Rhetorical Elements in Livy’s Direct Speeches: Part I’, AJP 38 (1917) 125–151. Feldherr, A., Spectacle and Society in Livy’s History (Berkeley 1998). Labov, W., Language in the Inner City. Studies in the Black English Vernacular (Philadelphia 1972). Laird, A., ‘The Rhetoric of Roman Historiography’, in A. Feldherr (ed.), The Cambridge Companion to the Roman Historians (Cambridge 2009) 197–213. Levene, D.S., Livy on the Hannibalic War (Oxford 2010). Luce, T.J., ‘Design and Structure in Livy: 5.32–55’, TAPhA 102 (1971) 265–302. Luce, T.J., ‘Structure in Livy’s Speeches’, in W. Schuller (ed.), Livius: Aspekte seines Werkes (Konstanz 1993), 71–85. Moles, J.L., ‘Livy’s Preface’, in Jane D. Chaplin & Christina S. Kraus (eds.), Livy. Oxford Readings in Classical Studies (Oxford 2009) 49–87. Stem, R., ‘The Exemplary Lessons of Livy’s Romulus’, TAPhA 137.2 (2007) 435–471. Viljamaa, T., Infinitive of Narration in Livy. A Study in Narrative Technique (Turku 1983). Walsh, P.G., Livy: His Historical Aims and Methods (Cambridge 1961). Weissenborn, W. & Müller, H.J., Titi Livi Ab urbe condita libri. 4 Bd., [2 Heft]: Buch XXII ([1891] Berlin).

chapter 12

Words When It’s Time for Action: Representations of Speech and Thought in the Battles of Cannae and Zama Suzanne Adema

1

Introduction*

In battle, warriors should act, not speak or think. This, in any case, is the impression we get from Latin narratives about battle, in which speech and thought tend to be less frequent than in other types of episodes.1 The preference of narrators for actions over words in the thick of battle is a reason to give all the more significance to the speeches and thoughts that they do present in their battle episodes. Therefore, the topic of this article is formed by the speeches and thoughts in Livy’s episode about the battle of Cannae, in contrast to those in the episode about the battle of Zama.2 In general, speeches and thoughts in Latin battle episodes are confined to specific types, at specific moments. This article focuses on a set of eight conventional representations of speeches and thoughts in battle narratives.3 In the phase leading up to a battle, battle narratives contain intelligence, messages and other forms of communication between the two enemies (1), as well as strategic considerations of the generals (2). At the end of this phase, generals tend to present an exhortation speech to their troops (3), followed by a moment in which the troops reflect on the upcoming battle (4). Orders by generals to their troops may be found both in the preceding phase and dur-

* This article was written as part of the NWO-project Ancient War Narrative. A Combined Discourse-Linguistic and Narratological Approach (NWO-project 360-30-190), led by Prof. Dr. Caroline Kroon and Prof. Dr. Irene de Jong (VU University Amsterdam and University of Amsterdam). The article forms a brief illustration of the approach I follow in my study Speech and Thought in Latin War Narratives. Words of Warriors. (Adema 2017). 1 Adema 2017. 2 The corpus of this article is 22.40.5–22.50.1 and 30.29.1–30.35.11. 3 This set is based on studies of both epic and historiography, to wit Fingerle 1937; Highet 1972; Hansen 1993; Dominik 1995; Lendon 1999; D’Huys 1990; Ash 2002; Longley 2012; the index list on military topics in Oakley 2005 and the results of my research project Speech and Thought in Latin War Narratives.

© koninklijke brill nv, leiden, 2019 | doi:10.1163/9789004383340_013

294

adema

ing battle itself (5). The start of a battle is typically marked by cries of battle, which, among other aggressive utterances, may also be found during battle (6). At the height of battle, a narrator may focus on one or two warriors and present their thoughts or speeches (7). The phase after battle usually contains thoughts or speeches from generals or other fighters in which the battle is evaluated (8). Although these types of speeches and thoughts are typical for battle episodes, even these conventional types of speech and thought leave room for narrators to make choices with respect to occurrence, presentational form and narrative techniques.4 This means that speeches and thoughts in narrative are not only a means of communication between the characters in the story world, but also a means of communication between the narrator and his narratees.5 The form, function and even the mere occurrence of a speech or thought should be incorporated in our analysis and interpretation of this speech or thought itself, its role in the immediate context and its function in conveying the more general ideas of the narrator. In this article I present a number of linguistic and narratological tools we may use to analyse and interpret narratorial choices concerning speech and thought representation. I illustrate these tools by means of speeches and thoughts in Livy’s battles of Cannae and Zama. The comparison between Cannae and Zama shows that the Livian narrator varies in the occurrence, forms and functions of the aforementioned eight types of speeches and thoughts. The variation may, firstly, be explained from the differences between these battles. A Roman defeat calls for a different presentation than a Roman victory. However, I will argue that we may learn more than that from the comparison and that speech and thought representation is a means for the narrator to present his own analysis of the course of a battle in a quite implicit way.6 At Cannae, the internal discord between Paullus and Varro is thematised in the representations of their speeches and thoughts, while Hannibal is portrayed as an excellent and almost omniscient general. Speech and thought representation is thus one of the ways in which the narrator suggests that the loss at Cannae was due to the specific characters leading both armies. At Zama, the representations of speeches and thoughts portray Hannibal as an excel-

4 In Adema 2016, I provide arguments for this idea by analyzing exhortation speeches from Caesar’s Bellum Gallicum. 5 On the idea that historiographical narrators have their own rhetorical aims, see e.g. White 1987; Woodman 1988; Laird 2009; Pausch 2011. 6 The same view is taken by Buijs (see p. 274) who focuses on the presentation, often in (in)direct speech, of some major themes.

words when it’s time for action

295

lent general who met his match in providence in Scipio, thus contributing to explaining why Hannibal was finally defeated.7

2

Narratorial Choices

The eight types of speeches and thoughts in this article are classified on the basis of the goals of the characters who communicate within the story world.8 This level of classification is used as a starting point to analyse the choices the narrator makes on another level, the level of communication between himself and his narratees. On the latter level, we may distinguish between choices concerning the forms of representations of speech and thought (linguistics) and choices concerning the narrative techniques employed in connection with the speeches and thoughts (narratology). An analysis of both types of narratorial choices should be part of our interpretation of a particular speech or thought. 2.1 Narratorial Choices: Linguistic Forms This article focuses on three linguistic forms of speech and thought representation.9 These are direct, indirect and mentioned representations of speeches and thoughts.10 Indirect speech and thought can also be looked at in terms of the narratological concept of embedded focalization.11 Approaching embedded focalization from a linguistic angle, I analyse those forms of embedded focalization that may be syntactically defined as complements of verbs of saying or thinking.12

7 8

9

10

11 12

Rossi 2004 traces the correspondences between these two generals, arguing that their stories are fashioned in the Zama-episode as “parallel lives”. The specific level of the communication between characters is the subject of many studies concerning speech and thought representation in Latin narrative. These studies often discuss the types of speeches within one narrative work, for instance, analysing them by using categories such as prayers, commands and soliloquy. E.g. Highet 1972; Dominik 1994; cf. Fingerle 1937. I leave free indirect discourse out of consideration in this article. On free indirect discourse in Latin, see Bayet 1931; Mellet 2000; Rosén 2013, 2015; and Adema 2017. On this topic in general, see Fludernik 1993. In mentioned discourse (in some studies called reported discourse) the narrator refers to the content of a speech or thought by means of, for instance, a noun (e.g. de caede) or an anaphorical expression (e.g. ea res nuntiatur). For embedded focalization, see De Jong [1987] 2004: 101–148; Bal 2009: 145–164. On linguistic forms of representations of speech and thought in Latin see e.g. Schlicher 1905; Dangel 1995; Bolkestein 1996; Sznajder 2005; Utard 2004; Van Gils 2009; Rosén

296

adema

In battle narrative, we may expect elaborate forms of speech and thought (direct and indirect discourse) to present the speeches and thoughts during the phases before and after the fighting itself. As the focus during the battle itself lies on the actions, we may expect brief forms of speech and thought representation (mentioned discourse) during that phase. Furthermore, we have to bear in mind that the narrator also has the option not to include a speech or thought, a choice that is of significance at moments in which we would expect a representation of a speech or thought. In this article, I have tried to identify these moments by analysing the occurrence of conventional types of speeches and thoughts in battle narratives and by comparing and contrasting Livy’s presentation of Cannae and Zama. 2.2 Narratorial Choices: Narrative Techniques Representations of speech and thought may be used in combination with several narrative concepts and techniques.13 The pace of narratives may increase or decrease by means of a variation between different forms of speech and thought representations. Speech and thought representation may also play a role in managing the order of the narrative (in actorial analepsis and actorial prolepsis), when characters speak or think about past or future events. Some speeches and thoughts contain prospective elements.14 Messages are a type of speeches that may be used to indicate the transition of one location within the story world to another. When a messenger travels from one location to another, his message functions in relation to the narratological concept of space. Representations of speech and thought may, furthermore, have a characterizing function, as part of getting to know the general nature of characters, but may also function to explain specific actions by characters.15 A final relevant narrative technique pertaining to representations of speech and thought is described by the narratological concepts ‘what it’s like’ and ‘experientiality’. These concepts concern the way in which ‘stories highlight the impact of events on the mind or minds experiencing those events within a story world’, as Herman puts it.16

13 14

15 16

2013. On linguistic forms of embedded focalization in ancient Greek, see e.g. De Jong [1987] 2004: 101–148; De Bakker 2007; Beck 2012. The narratological concepts and their definitions in this article are mostly based on the theoretical frameworks of Bal [1985] 2009; and De Jong 2004, 2007, 2012. This is, for instance, a rather frequent technique in speeches and thoughts (of Caesar) in Caesar’s Bellum Gallicum, cf. Adema 2016. Hunter 1973 provides a study of Thucydides’ use of this particular technique. On the relation between representation of speech and thought and characterization in Livy, see Levene 2010: 174–176. Herman 2009: 37.

words when it’s time for action

297

Often, two or more narrative techniques occur in combination in a speech or thought representation. A speech may, for instance, characterize a character and his or her relations with other characters, while also slowing down the pace of the narrative and presenting a prolepsis.17

3

Speech and Thought Representations in the Battle Narratives of Cannae and Zama

The previous section showed that the Livian narrator, as any narrator, had several choices in his presentation and use of speeches and thoughts typical for battle scenes. The sections below analyse and interpret the choices he has made in the Cannae and Zama-episodes. 3.1 Intelligence and Messages War cannot be fought without intelligence.18 In the description of battle scenes messages and information keep generals up to date of the movements of the enemy.19 A narrator may use these messages to create smooth changes of location in his narrative. A report of events that have just happened in one location is in these cases brought by messengers or spies to another location. Messages often repeat events that have just been narrated, and narrators, therefore, are able to present their content in highly summarized form, viz. a mentioned speech (e.g. ea res nuntiata est). Thus, messages form a less intrusive, narrativized way to change the location than a narratorial description (however short) of a new location, and seem to function as a text structuring device.20 Livy’s narrator indeed uses messages and messengers in this spatial function in the Zama-episode. The narrative takes place in several locations, viz. Hadrumetum, the Carthaginian camp near Zama and the Roman camp near Zama.21 The change of space from Hadrumetum to the Carthaginian camp near

17 18

19 20 21

On the other hand, speech and thought representation are not the only textual means that narrators can use for these narrative techniques. Austin & Rankov 1998 present an overview of how Roman intelligence functioned in the Roman army. Cf. also Daly 2002: 113–114, who discusses the use of intelligence as one of the main responsibilities of a general. In this chapter, I focus on the functions of intelligence in narrative techniques, as part of the textual strategies of the narrator of AUC. This observation is based on my research on Caesar (Bellum Gallicum), Sallust (Bellum Iugurthinum), Vergil (Aeneis). Van Gils (this volume, 255) discusses several uses of spatial references, among which textstructuring spatial references. In addition, the location of Carthage is evoked by means of the thoughts of Hannibal.

298

adema

Zama is established by narrating the journey of Hannibal (contendit). The ensuing changes of space from Hannibal’s camp to Scipio’s camp and back again are integrated into the narrative by means of the praemissi speculatores who are brought before Scipio and, then, sent back to Hannibal: Iam Hadrumetum peruenerat Hannibal; unde, ad reficiendum ex iactatione maritima militem paucis diebus sumptis, excitus pauidis nuntiis omnia circa Carthaginem obtineri armis adferentium, magnis itineribus Zamam contendit. Zama quinque dierum iter ab Carthagine abest. Inde praemissi speculatores cum excepti ab custodibus Romanis deducti ad Scipionem essent, traditos eos tribuno militum, iussosque omisso metu uisere omnia, per castra qua uellent circumduci iussit; percunctatusque satin per commodum omnia explorassent, datis qui prosequerentur, retro ad Hannibalem dimisit. Hannibal nihil quidem eorum quae nuntiabantur—nam et Masinissam cum sex milibus peditum, quattuor equitum uenisse eo ipso forte die adferebant—, laeto animo audiuit, […]. Liv. 30.29

By this time Hannibal had reached Hadrumetum. From there, after he had spent a few days that his soldiers might recuperate from sea-sickness, he was called away by alarming news brought by men who reported that all the country round Carthage was occupied by armed forces, and he hastened to Zama by forced marches. Zama is distant five day’s marches from Carthage. Scouts who had been sent in advance from that position were captured and brought before Scipio by their Roman guards. Thereupon he turned them over to a tribune of the soldiers, and bidding them go and see everything without fear, he ordered them to be led about the camp wherever they wished to go; and after questioning them as to whether they had examined everything quite at their leisure, he sent them back to Hannibal, furnishing men to escort them. Hannibal did not indeed hear with joy any of the reports of his scouts, for they reported that Masinissa had also arrived that very day, as it happened, with six thousand infantry and four thousand cavalry.22 The passage illustrates that the exchange of intelligence is used in the Zamaepisode to switch between Hannibal and Scipio, showing that both generals have more or less equal knowledge of each other.

22

Translations (slightly adapted) are taken from Foster 1929 (Loeb).

words when it’s time for action

299

In the Cannae-episode, we find a different use of messages and intelligence. Relatively many instances of speech and thought representations contain intelligence or misinformation about either the Romans or the Carthaginians, but these messages do not primarily have a spatial function. Furthermore, we should distinguish between information received by the Romans and information received or, rather, known by Hannibal. When news travels from the Carthaginian camp to the Roman camp, emphasis is put on wrong interpretations (for instance, 42.2) or wrong uses of the intelligence by the Romans. The most explicit example is the moment that Statilius has reported in quite some detail about the situation in the Carthaginian camp, concluding that an ambush awaits the Romans. The very report that should withhold the Roman troops from action only incites them: Quae ad deterrendos a cupiditate animos nuntiata erant, ea accenderunt, […]. Liv. 22.42.5–22.42.7

The report, which had been made with the purpose of checking the soldiers’ greed, only inflamed it, […]. The narrator travels along with Marius Statilius to the Carthaginian camp and back again, but the main narratological function of this message seems to be a characterizing one, conveying that the Roman troops were prepared to walk into an ambush, knowingly. The troops were too eager to fight. The presentation of Hannibal’s information about the Romans, too, has a characterizing function, presenting Hannibal as a well-informed general. A switch from the Roman camp to Hannibal is usually not marked by means of an explicit message. Rather, it is marked by means of Hannibal’s name being put at the first position of the sentence (e.g. at 41.4). The information that Hannibal has about the Romans is mostly presented in the form of a mentioned or indirect thought, as is illustrated in the next passage: Et omnia ei hostium haud secus quam sua nota erant: dissimiles discordesque imperitare, duas prope partes tironum militum in exercitu esse. Liv. 22.41.5

All the circumstances of his enemies were as familiar to him as his own: that their generals were unlike each other and were at loggerheads, and that nearly two-thirds of their army consisted of recruits.

300

adema

Indirect discourse (dissimiles … esse) represents Hannibal’s knowledge of his enemies in this passage. This form of presentation of Hannibal’s information about the Romans suggests he does not need scouts since he simply knows everything. The presentation of intelligence at Cannae thus contributes to the characterization of Hannibal as being as familiar with the Roman army as he is with his own army. 3.2 Strategic Considerations In war and battle narrative, generals prepare for battle, and their preparations and strategic considerations may be presented in (in)direct speech or (in)direct thought. These strategic considerations occur in the phase preceding battle, which may be days before battle, in reaction to received intelligence, or immediately before battle. Strategic considerations of skilled generals may be expected to contain anticipations of the course of events in the upcoming battle. Thus, these considerations foreshadow the battle itself. At the same time, they characterize generals as either good or bad, depending on the correctness of their anticipations. In Livy’s Zama episode, strategic considerations of both Hannibal and Scipio are presented immediately before battle, for instance, when both are drawing up their army (30.33.1–30.33.7). The narrator begins with Scipio, who anticipates the use of elephants by Hannibal and sets up his ranks accordingly. This is made explicit in an indirect speech depending on dato praecepto. dato praecepto ut ad impetum elephantorum aut post directos refugerent ordines aut in dextram laeuamque discursu applicantes se antesignanis uiam qua inruerent in ancipitia tela beluis darent. Hannibal ad terrorem primos elephantos […] instruxit, […]. Liv. 30.33.3–30.33.4

under orders that, upon the charge of the elephants, they should either flee behind the ranks in the line, or else dashing to right and left and closing up to the maniples in the van, should give the beasts an opening through which they might rush among missiles hurled from both sides. Hannibal in order to create a panic drew up his elephants in front, […]. Scipio specifically instructs his troops how to react on an attack of elephants. In the immediately following clause Hannibal indeed sets up the elephants at the head of his troops, an act confirming Scipio’s strategical insight. In setting up the ranks and anticipating the strategies of the enemy, Hannibal

words when it’s time for action

301

and Scipio seem to be in a pre-battle contest of anticipation, of which Scipio is the clear winner.23 The focus of the narrator on the elephants does not only function to characterize Scipio as a provident general, it also foreshadows the upcoming battle in which the elephants play an important role. As soon as the Roman sign for battle sounds, the elephants are frightened and turn on their own ranks. The elephants that do attack the Roman troops are wounded by the spears of the uelites (skirmishers), just as Scipio had planned (AUC 30.33.13–30.33.16). In Livy’s Cannae-episode, the only strategic considerations at the Roman side are made long before battle. When Paullus and Varro were still in Rome, Paullus was presented with the best strategy by Fabius in a directly presented speech full of prospective elements (especially 39.9–39.17). Paullus’ reaction is presented in an indirect speech of considerable length (22.40.2).24 Paullus’ speech anticipates the problems that Varro will cause. This means that his worries do not concern the upcoming battle and that, in contrast to Scipio at Zama, Paullus is not concerned with the best strategies to win. Although Paullus shows that he has the skill to anticipate upcoming events, he does not use them in a way befitting a great general. He merely worries about the behaviour of Varro and does not even ponder on how to overcome the lack of skills of his colleague. His speech characterizes him as rather defeatist, a characterization to which the following rhetorical question about his fellow consul contributes: quid consuli aduersus collegam seditiosum ac temerarium uirium atque auctoritatis fore? Liv. 22.40.2

what power or influence then would a consul have over a turbulent and headstrong colleague? In the phase leading up to the battle itself, the prospective elements in Paullus’ speech indeed come true, as Varro and Paullus are too occupied with each other and their troops to ponder on the best approach and strategies. The lack of

23 24

Edlund 1967: 168. Klotz 1949: 63 concludes from the indirect form of the speech that Paullus is not foregrounded (eine knappe indirekte Rede, in der die Person nicht so sehr in den Vordergrund tritt). It is true that this speech is not presented as elaborately as Fabius’ direct speech. However, the indirect speech does contain several elements in which we may assume the focalization of Paullus (e.g. the rhetorical question).

302

adema

strategic considerations in these generals contributes to their characterization and brings the dangers of internal discord to the fore. Hannibal, therefore, is the only general who thinks things through at Cannae.25 We have seen that he has all the intelligence he needs, and also has the anticipatory skills required in a good general. This is illustrated when Varro and Paullus arrive at the Roman army camp. Immediately, Hannibal interprets their arrival as a favourable event: Hannibal quamquam parte dimidia auctas hostium copias cernebat, tamen aduentu consulum mire gaudere. Liv. 22.40.7

Hannibal, though he perceived that the forces of his enemies were augmented by a half, was nevertheless greatly rejoiced at the coming of the consuls. Hannibal’s joy foreshadows the positive outcome of the upcoming battle of Cannae, and characterizes him as a provident general. In none of his strategic considerations in this episode, however, Hannibal looks beyond this upcoming battle.26 Hannibal does not look at the bigger picture, and this portrayal might be interpreted as an explanation for his later defeat, a way in which the narrator makes it plausible that this excellent general would eventually lose the war. 3.3 Exhortations Perhaps the best known type of speech in battle narrative is the exhortation by the general. I will not go into the historicity of exhortation speeches, but consider them as a recurring element in the presentation of battles in historiography and epic.27 It is of course up to the narrator, how or even if he presents exhortations.28 The form may be a direct or lengthy indirect speech, or a mere mentioned speech. From a narratological perspective, it is interesting

25 26

27

28

The only exception is his contemptuous direct speech (22.49.3), which is explicitly presented by the narrator as ‘hearsay’ (dixisse ferunt). This is in contrast with Polybius (3.111.9), in which Hannibal exhorts his troops by saying that this victory will make them masters of all Italy. Hannibal does not look ahead in such a fashion in the Livian episode. Albertus 1908 presents examples of pre-battle exhortations in Greek and Latin historiography. The question of the historicity of exhortation speeches is addressed by e.g. Hansen 1993; Pritchett 1994; Daly 2002; Zoida 2007; Yellin 2008; Anson 2010. Zoido 2007 makes this observation about exhortations in Thucydides.

words when it’s time for action

303

to investigate whether exhortations foreshadow the course and importance of the upcoming battle. In Livy’s presentation of the battle of Zama, both Scipio and Hannibal are presented as repeatedly exhorting their troops. The first exhortation in this episode indeed contains a clear reference to the importance and decisiveness of the upcoming battle: In castra ut est uentum, pronuntiant ambo arma expedirent milites animosque ad supremum certamen, non in unum diem sed in perpetuum, si felicitas adesset, uictores. Liv. 30.32.1

Arrived at their camps, they both ordered their soldiers to have arms and their spirits in readiness for the final conflict to make them victors, if success attended them, not for one day but forever. The expression supremum certamen makes clear that this is the decisive battle. The exhortation is remarkable in that the indirect discourse simultaneously presents the exhortation of Scipio and that of Hannibal (pronuntiant ambo). Thus, the narrator seems to use the topos of the exhortation speech to characterize the equality in the skills of Scipio and Hannibal. In addition to this exhortation, more indirectly presented exhortations are presented in the moments before battle by both generals (30.32.6–30.32.11). When Hannibal sets up his troops, originating from many different places, he addresses the separate parts of his army with specific exhortations (30.33.8– 30.33.12). By these separate exhortations, the narrator seems to emphasize the heterogeneity of Hannibal’s troops.29 In his larger discussion on the historicity of exhortation speeches, Daly focuses on the language problems this would have caused. I would like to suggest that the heterogeneity of Hannibal’s troops, emphasized by these particular exhortations, might be presented by the narrator of AUC as one of the reasons for Hannibal’s loss (see also section 2.6).30 At Cannae, no exhortations are presented. Between 45.5 and 46.9, Varro, Paullus and Hannibal are busy setting up ranks. In Polybius’ account of Cannae, however, both the Carthaginian troops and the Roman troops are exhorted.31 29 30 31

Daly 2002: 141. Levene 2010: 243 goes even further, when suggesting that Livy here might be referring to the heterogeneity of the Roman army in his own days. See e.g. Daly 2002: 143 for a discussion. For a comparison of the texts of Polybius and Livy on Cannae, see Oakley (this volume, 182–189); and Buijs (this volume, 278–279), on Zama,

304

adema

The Roman exhortation in Polybius is done by Paullus (3.108–3.109), although not on the day of the battle itself, but several days earlier, before he finds out the terrain is not fit for battle. The absence of exhortations at Cannae in Livy, thus, seems a significant narratorial choice, one that could be interpreted as part of the portrayal of Paullus. In Livy, Paullus is portrayed as a general who is opposed to fighting, and thus it is unlikely that he presents a battle exhortation. 3.4 Commands Commands of generals to their troops are a natural type of speech in battle narrative, both in the phase preceding battle and during the fighting itself. Presented as (very) brief indirect speeches, commands are a way to enhance the authority of generals and to contribute to a fast pace of events.32 The effect on the pace of the narrative may be illustrated by means of a command given by Scipio at Zama: Quod Scipio ubi uidit, receptui propere canere hastatis iussit et sauciis in postremam aciem subductis principes triariosque in cornua inducit, […]. Liv. 30.34.11

When Scipio saw this he ordered the recall to be sounded at once for the hastati, and after withdrawing the wounded to the rear line, he led the principes and triarii to the wings, […]. Scipio orders the recall to be sounded and the execution of this order is left implicit, suggesting that, indeed, the recall was sounded. This command has an effect on the pace of the narrative. The clause receptui propere canere hastatis iussit may seem to narrate one event, a command, but, in fact, it informs the narratee of two events on the time-line, to wit Scipio’s command and the ensuing recall. Thus, the pace of the narrative is rather fast or even elliptic here, almost as if to emphasise Scipio’s decisiveness and swiftness in addition to contributing to his authority. Orders do not occur in the battle of Cannae itself. They do occur in the phase preceding battle, but the Roman troops refuse to obey them (22.42.11). Moreover, they demand to be given only the commands they want to execute:

32

see e.g. Edlund 1967. A recent narratological discussion of Polybius’ presentation of the battle at Zama can be found in Grethlein 2013. See also Dominik 1994: 191 (on commands and the advancement of narrative time in Statius’ Thebais).

words when it’s time for action

305

Clamor inde ortus ut signa proferri iuberent ducerentque ad persequendos hostes ac protinus castra diripienda, […]. Liv. 22.42.3

Next they began to clamour for the order to advance and to pursue the enemy and plunder the camp without delay, […]. The commands on the Roman side characterize both the consuls and the troops, contributing to the idea that the Roman troops were too eager to fight and not controlled by either consul. In contrast, Hannibal does not even need to give orders at Cannae: Hannibal … derigit aciem lacessitque Numidarum procursatione hostes. Liv. 22.44.4

Hannibal … deployed his battle line, and provoked his enemy with sudden charges from his Numidian troops. The phrase lacessitque Numidarum procursatione hostes, of course, implies a command, but it is not presented in the form of an indirect speech. This is illustrative for the whole episode, in which Hannibal hardly utters a word. This presentation of Hannibal directing his troops contributes to a characterization of Hannibal as a general who effortlessly—almost without saying anything— makes his troops do as he wishes. 3.5 Reflections on the Upcoming Battle by the Troops Before the fighting begins there tend to be moments in war narratives in which both sides recognize the decisiveness of the moment (cf. for instance, Caes. Gal. 7.85.2). These reflections by troops are not only highly suitable to foreshadow the outcome of the battle, they may also function to make the narratees ‘feel’ what it is like to be about to engage in battle. Livy’s narration of the battle of Zama has such a moment of anxious reflection of the soldiers: Anceps igitur spes et metus miscebant animos; contemplantibusque modo suam modo hostium aciem, cum oculis magis quam ratione pensarent uires, simul laeta, simul tristia obuersabantur. Liv. 30.32.4

Consequently a wavering between hope and fear confused their spirits; and as they surveyed now their own battle-line, now that of the enemy,

306

adema

while weighing their strength more by the eye than by calculation, the bright side and at the same time the dark was before their minds. The comparable moment at Cannae lacks an evocation of the emotions of the troops, but, instead, focuses on the physical distress of the Roman troops: Ventus—Volturnum regionis incolae uocant—aduersus Romanis coortus multo puluere in ipsa ora uoluendo prospectum ademit. Liv. 22.47.1

A wind—which those who live in those parts call Volturnus—beginning to blow against the Romans carried clouds of dust right into their faces and prevented them from seeing anything. The element of ‘what it’s like’ here is not expressed by means of indirect thought, but takes the form of a description of the physical circumstances of the fight.33 The Roman army does not reflect on the upcoming battle. This is in contrast with the situation in Zama, a contrast that is underscored in details about the vision of the troops in both battles. The troops in Zama observe, survey and compare the enemy and themselves (contemplantibus, pensarent, obseruabantur). In Cannae, the troops do not see anything (prospectum ademit) and therefore cannot focus, physically nor mentally. 3.6 Battle Cries and Expressions of Aggression and Contempt During battle, brief forms of speech may be presented to convey the aggression of the fighters, or even their contempt as they slay their enemies. We may expect these expressions to contribute to the idea of what war is like. A typical expression of aggression in battle narrative is the shouting (clamor) of the troops immediately after the signum belli, marking the beginning of battle. Both Livy’s episode of Cannae (for instance, 22.47.1) and Zama contain

33

See Daly 2002: 156–206 for a reconstruction of the experience of Cannae. We might have expected more emphasis on the aspect of ‘what it’s like’ in the representations of speech and thought in the presentation of this battle, given the observation by White 1987: 204 that ‘[a]bove all, there is that preoccupation with psychological considerations— the joy, spirit, and determination of the victors, the depression, fear, and madness of the defeated—which holds so high a place in Livy’s conception of the historian’s craft.’ However, the narrator seems to emphasize the depression, fear and madness in the aftermath of the battle (both on the battlefield and in Rome) rather than during the battle (cf. Oakley, this volume, 171).

words when it’s time for action

307

such shouts. In the latter episode the battle cry expresses more than the aggression of the troops: congruens clamor ab Romanis eoque maior et terribilior, dissonae illis, ut gentium multarum discrepantibus linguis, uoces. Liv. 30.34.1

a harmony in the shouting of the Romans, which consequently was greater in volume and more terrifying; on the other side discordant voices, as was natural from many nations with a confusion of tongue. In this excerpt, the adjectives congruens and dissonae enhance the contrast between the unison of the Roman troops on the one hand and the heterogeneity of Hannibal’s troops on the other.34 The expression dissonae uoces repeats the motive of heterogeneity found in Hannibal’s earlier exhortations. Hannibal spoke to each of the different nations in his army in their own language (30.33.8–30.33.12, see above), and here all these nations utter cries of war in their own tongue. The Cannae-episode contains a less typical expression of aggression. The narrator refers to a contemptuous remark by Hannibal, when he is informed that the Roman cavalry has dismounted from their horses: tum denuntianti cuidam iussisse consulem ad pedes descendere equites dixisse Hannibalem ferunt ‘Quam mallem, uinctos mihi traderet’. Liv. 22.49.3

At this Hannibal, being told by someone that the consul had ordered his troopers to dismount, is said to have exclaimed: “How much better if he had handed them over to me in fetters!” The remark not only conveys contempt, but also portrays Hannibal, again, as the superior general. Earlier, he immediately interpreted the arrival of Varro and Paullus as a favourable event (22.40.7), and throughout the episode Hanni-

34

This difference between the Romans and Carthaginians is also part of Polybius’ presentation of the battle of Zama (15.12.8–15.12.9). D’Huys 1990 discusses this battle cry in Polybius’ presentation of Zama as a (Homeric) topos. D’Huys argues that Polybius uses this and other topoi not only as a literary means, but also as a means to emphasise the importance of certain moments (a use he refers to as ein historische Zweck and historische ‘Nützlichkeit’).

308

adema

bal is up-to-date about the actions of the Romans (especially at 22.41.5). Hannibal’s quick-witted reaction to the news about the consul shows that he, in the aftermath of this battle, is able to identify immediately the implications of the actions of his opponents.35 3.7 Focus on Individual Warriors In battle narratives narrators may focus on one or two warriors, singling them out by taking the narratorial standpoint of the close-up and decreasing narrative pace.36 These moments thus become climactic moments in the presentation of the battle. In the Zama-episode, the narrator does not single out one particular moment by means of this specific type of presentation. There is, however, a clear example of a climactic moment in the battle phase of the Cannae-episode: the dialogue between Gnaeus Lentulus and the dying Paullus.37 An important narratological function of this dialogue seems to be the confirmation of earlier prospective elements.38 Before battle, when Paullus and Varro were still in Rome, Paullus had been warned for his fellow consul by Fabius (22.39.9– 22.39.17). When he dies, Paullus explicitly refers back to this moment: Abi, nuntia publice patribus, urbem Romanam muniant ac priusquam uictor hostis aduenit praesidiis firment; priuatim Q. Fabio, L. Aemilium praeceptorum eius memorem et uixisse adhuc et mori. Liv. 22.49.10–22.49.11

Go, and tell the senators in public session to fortify the City of Rome and garrison it strongly before the victorious enemy draws near: in private say to Quintus Fabius that Lucius Aemilius has lived till this hour and now dies remembering his precepts.

35 36 37

38

The primary narrator embeds this remark in a construction depending on ferunt, thus making use of a reported narrator, De Jong 2004: 108; cf. De Bakker 2007: 160–177. Longley 2012. Although this is indeed a climactic moment in the battle phase of the Cannae-episode, it does not seem to be the climax of the Cannae-episode as a whole. Oakley (this volume, 171) argues that Livy ensures that the battle does not become the climax of the episode, by devoting much space to the immediate aftermath of the battle (50.1–61.15). For a different view, see Van Gils & Kroon (this volume, 224). Furthermore, Paullus’ speech foreshadows the Roman fear, since he sends Lentulus away to report the defeat to the senators (21.49.10), see also Oakley (this volume, 164).

words when it’s time for action

309

Paullus’ dying words refer once more to the discordia between himself and Varro. Thus, they seem to underscore one of the factors to which the narrator contributes the defeat at Cannae. Furthermore, his dying words introduce a new theme into the narrative.39 Paullus sends Lentulus away to report the defeat to the senators, an action that makes the narratees anticipate the perils for Rome and the reactions of its inhabitants. The speech, itself a climactic moment in battle, thus combines the preliminaries and aftermath of this battle. 3.8 Debrief In its final phase, a battle narrative typically contains speeches in a debriefing situation or thoughts in which generals reflect on the past battle. These speeches and reflections may be expected to reconsider earlier strategic considerations, thus referring to earlier prospective elements. By acknowledging that the course of battle was as the general expected, these speeches and thoughts also contribute to portraying the anticipatory and strategic skills of the generals. This use of post-battle reflection is found after the battle of Zama, when Scipio and his military experts look back admiringly at the way in which Hannibal had drawn up his line: et confessione etiam Scipionis omniumque peritorum militiae illam laudem adeptus singulari arte aciem eo die instruxisse: elephantos in prima fronte quorum fortuitus impetus atque intolerabilis uis signa sequi et seruare ordines, in quo plurimum spei ponerent, Romanos prohiberent; […] Liv. 30.35.5

And even by Scipio’s admission and that of all the military experts he had achieved this distinction, that he had drawn up his line that day with extraordinary skill: the elephants in the very front, that their haphazard charge and irresistible strength might prevent the Romans from following their standards and keeping their ranks, upon which tactics they based most of their hopes; […]. This passage is the beginning of their debriefing speech and shows how positively the experts evaluate Hannibal’s use of elephants. The speech clearly has a function in the characterization of Hannibal, portraying him as an excel-

39

As Oakley (this volume, 177) observes too.

310

adema

lent general who showed his strategical insight at one more occasion.40 Scipio, however, had anticipated his strategies. Thus these reflections form an indirect praise of Scipio, too, and underscore his skills. The defeat of Hannibal had been prepared already by the narrator at Cannae, in Hannibal’s strategic considerations and during the debriefing situation in the Carthaginian camp. The strategic considerations of Hannibal preceding the battle of Cannae were all very much focused on the upcoming battle and not on the war as a whole (see section 2.3). In the debriefing situation, Maharbal urges Hannibal to act, but Hannibal is reluctant: itaque uoluntatem se laudare Maharbalis ait; ad consilium pensandum temporis opus esse. tum Maharbal: ‘Non omnia nimirum eidem di dedere: uincere scis, Hannibal, uictoria uti nescis’. mora eius diei satis creditur saluti fuisse urbi atque imperio. Liv. 22.51.3–22.51.4

And so, while praising Maharbal’s goodwill, he declared that he must have time to deliberate regarding his advice. Then said Maharbal, “In very truth the gods bestow not on the same man all their gifts; you know how to gain a victory, Hannibal: you know not how to use one.” That day’s delay is generally believed to have saved the City and the empire. Hannibal’s indirect speech shows that Hannibal can only stick to his main strategy of thinking things through. Maharbal’s two speeches, of which one is quoted, are presented directly, and thus highlighted by the narrator, who makes clear in the ensuing sentence that Maharbal’s analysis has become generally accepted. In his presentation of the battle of Zama as a whole, the narrator uses several representations of speech and thought to underscore that Hannibal was an excellent general and could have conquered Rome. The debrief of the battle at Cannae makes explicit that while his strategic considerations won Hannibal Cannae, they cost him the war. The moments right after Cannae should have been, for Hannibal, a time for action, not for words.

40

The primary narrator refers to Zama as Hannibal’s ultimum uirtutis opus (30.35.10).

words when it’s time for action

4

311

Conclusion

Narrators have choices in their presentation of the speeches and thoughts of characters. This implies that they may use the speeches and thoughts of characters to convey their own visions, explanations and ideologies. In this article, I hope to have provided and illustrated a number of linguistic and narratological tools we may use to analyse and interpret these choices, aimed at interpreting speeches and thoughts on the level of the narrator and his narratees. Using the episodes of Cannae and Zama as my corpus, I focused on the differences and similarities in the occurrence of eight conventional types of speeches and thoughts in battle episodes, viz. messages, strategic considerations, commands, reflections of warriors, expressions of aggression, speeches in ‘close-ups’ during the battle and debriefs. Messages in battle narrative may be expected to link two threads of the story, switching between two army camps. In the Zama-episode, brief presentations of messages are indeed used to narrativize changes in space, but in the Cannae-episode messages and the availability of intelligence function to portray differences between Hannibal on the one hand and Varro and Paullus on the other, creating a contest in knowledge easily won by Hannibal. Strategic considerations are a type of speech and thought in battle narrative that is specifically suitable to foreshadow the outcome of the battle, while also characterizing the generals and their anticipatory skills. Hannibal is portrayed as a provident general at Cannae, but one who focuses on Cannae alone. No strategic considerations of Varro are presented and Paullus is portrayed as a defeatist general who only anticipates the behaviour of his colleague and not that of his enemy. In Zama, the functions of characterization and prospective elements are combined in the narrator’s treatment of Hannibal’s use of elephants and, especially, Scipio’s anticipation of this use. Commands enhance the authority of generals, a function they indeed seem to have in the portrayal of Scipio. This conventional type of speech in war narrative is, however, used quite differently in the Cannae-episode, in which Roman troops cease to obey commands and even instruct their general on what commands they should issue. At Cannae, commands are used to portray troops rather than generals, thus contributing to the analysis of the narrator that a lack of discipline and leadership was an important factor in the defeat at Cannae. Reflections of warriors on the upcoming battle and battle cries occur at a moment at which a narrator might want to convey the experience of battle to his narratees and highlight the impact of the events on the minds of the fighters. Comparison of the Zama-episode and the Cannae-episode shows that this

312

adema

moment indeed is used by the narrator to underscore what it’s like to inhabit the story world. In Cannae, however, the narrator does not use speeches and thoughts, but describes the physical circumstances for the troops. In Zama, the narrator does use representations of thought. These thoughts also illustrate that representations of speeches and thought usually have more than one narratological function, as they also often foreshadow the outcome of battle, underscoring the significance of a battle in the course of the war as a whole. A narrator may focus on individual warriors at the height of battle. In this type of environment, representations of speeches and thoughts may be used to refer back to prospective elements and underscore their fulfilment. This is particularly illustrated in Paullus’ dying words on the battlefield of Cannae. Debriefs, too, may contain such references to earlier prospective elements, as do the thoughts in which Scipio reflects on the course of the battle of Zama. The debriefing moment after the Cannae-episode in the Carthaginian camp does not refer back to earlier prospective elements, but rather presents an explicit prolepsis to the eventual defeat of the Carthaginians by the Romans. In this respect, speech and thought representation seems to have a role in the overall structuring of the narrative. In short, the narrator seems to use all these conventional types of speeches and thoughts to underscore his own analysis of a defeat and a victory respectively, thus communicating with his narratees. By means of these speeches and thoughts, he conveys that the characters of the generals and Roman troops were an important factor both in the inevitable defeat at Cannae and in the admirable victory at Zama. In Cannae, the speeches and thoughts help to convey the idea that the generals, even Hannibal, were too occupied with each other and the upcoming battle to ponder on the war as a whole. In Zama, the speeches and thoughts accentuate that excellent generals and armies met, realised what was at stake and fought. There, Scipio could demonstrate his superiority due to the brilliance of his enemy.

Bibliography Adema, S.M., ‘Encouraging Troops, Persuading Narratees. Pre-battle Exhortations in Caesar’s Bellum Gallicum as a Narrative Device’, in S. Farrington & V. Liotsakis (eds.), The Art of History (Boston 2016) 219–240. Adema, S.M., Speech and Thought in Latin War Narratives. Words of Warriors. (Amsterdam 2017). Albertus, J., Die Parakletikoi in Der Griechischen Und Römischen Literatur (Strassburg 1908).

words when it’s time for action

313

Anson, E., ‘The General’s Pre-Battle Exhortation in Graeco-Roman Warfare’, G&R 57.2 (2010) 304–318. Bal, M., Narratology. Introduction to the Theory of Narrative ([1985] Toronto 2009). Banfield, A., Unspeakable Sentences. Narration and Representation in the Language of Fiction (Boston 1982). Bayet, J., ‘Le style indirect libre en latin’, RPh 57/58 (1931, 1932) 327–342 & 5–23. Beck, D., Speech Presentation in Homeric Epic (Austin 2012). Berry, D.H. & Erskine, A., ‘Form and Function’, in D.H. Berry & A. Erskine (eds.), Form and Function in Roman Oratory (Cambridge 2010) 1–19. Bolkestein, A.M., ‘Reported Speech in Latin’, in T. Janssen & W.v.d. Wurff (eds.), Reported Speech. Forms and Functions of the Verb (Amsterdam 1996) 121–140. D’Huys, V., ‘Χρήσιμον καὶ τερπνόν in Polybios’ Schlachtschilderungen. Einige literarische Topoi in seiner Darstellung der Schlacht bei Zama (XV 9–16)’, in H. Verdin, G. Schepens & E. de Keyser (eds.), Purposes of History (Louvain 1990) 267–288. Daly, G., Cannae: The Experience of Battle in the Second Punic War (London 2002). De Bakker, M.P., Speech and Authority in Herodotus’ Histories (Amsterdam, 2007). De Jong, I., Narrators and Focalizers. The Presentation of the Story in the Iliad ([1987] London 2004). De Jong, I. & Nünlist, R. (eds.), Time in Ancient Greek Literature (Leiden 2007). De Jong, I., Nünlist, R. & Bowie, A. (eds.), Narrators, Narratees, and Narratives in Ancient Greek Literature (Leiden 2004). De Jong, I.J.F., Narratology and Classics. A Practical Guide (Oxford 2014). Dominik, W.J., Speech and Rhetoric in Statius’ Thebaid (Hildesheim 1994). Edlund, I. ‘Before Zama: A Comparison between Polybios’ and Livy’s Descriptions of the Meeting between Hannibal and Scipio’, Eranos 65 (1967) 148–168. Fingerle, A., Typik der Homerischen Reden (Munich 1939). Fludernik, M., The Fictions of Language and the Languages of Fiction: The Linguistic Representation of Speech and Consciousness (London 1993). Gils, L.W. van, Argument and Narrative. A Discourse Analysis of Ten Ciceronian Speeches (Amsterdam 2009). Grethlein, J., ‘The Unthucydidean Voice of Sallust’, TAPhA 136 (2006) 299–327. Hansen, M.H., ‘The Battle Exhortation in Ancient Historiography. Fact or Fiction?’, Historia 42.2 (1993) 161–180. Herman, D., Basic Elements of Narrative (Oxford 2009). Highet, G., The Speeches in Vergil’s Aeneid (Princeton 1972). Hunter, V.J., Thucydides: The Artful Reporter (Toronto 1973). Keitel, E., ‘Otho’s Exhortations in Tacitus’ ‘Histories’’, G&R 34.1 (1987) 73–82. Klotz. A., ‘Dichtung und Wahrheit in der livianischen Erzählung von der Schlacht bei Cannae’, Gymnasium 56 (1949) 58–170. Laird, A., ‘The Rhetoric of Roman Historiography’, in A. Feldherr (ed.), The Cambridge Companion to the Roman Historians (Cambridge 2009) 197–213.

314

adema

Lendon, J.E., ‘The Rhetoric of Combat. Greek Military Theory and Roman Culture in Julius Caesar’s Battle Descriptions’, ClAnt 18.2 (1999) 273–329. Levene, D.S., Livy on the Hannibalic War (Oxford 2010). Longley, G.M., ‘Battle narratives’, Encyclopedia of Ancient History (Online resource 2012). Luce, T.J., ‘Structure in Livy’s speeches’, in W. Schuller (ed.), Livius: Aspekte seines Werkes (Konstanz, 1993) 71–87. Marincola, J., ‘Speeches in Classical Historiography’, in J. Marincola (ed.), A Companion to Greek and Roman Historiography (Oxford 2007) 118–132. Mellet, S., ‘Imparfait et discours rapporté’, Oratio soluta, Oratio Numerosa. Études Luxembourgeoises d’histoire et de littérature Romaines 1 (1998) 117–125. Miller, N.P., ‘Dramatic Speech in the Roman Historians’, G&R 22.1 (1975) 45–57. Oakley, S.P., A Commentary on Livy Books VI–X. Volume III: Book IX (Oxford 2005). Pausch, D., ‘Der Feldherr Als Redner Und Der Appell an Den Leser: Wiederholung Und Antizipation in Den Reden Bei Livius’, in D. Pausch (ed.), Stimmen Der Geschichte. Funktionen Von Reden in Der Antiken Historiographie (New York 2010) 183–209. Pausch, D., Livius Und Der Leser: Narrative Strukturen in Ab Urbe Condita, (Munich 2011). Pelling, C., ‘Seeing through Caesar’s Eyes: Focalisation and Interpretation’, in J. Grethlein & A. Rengakos (eds.), Narratology and Interpretation (Berlin 2009) 507–526. Pritchett, W.K., Essays in Greek History (Amsterdam 1994). Rosén, H., ‘About Non-Direct Discourse: Another Look at Its Parameters in Latin’, Journal of Latin linguistics 12.2 (2013) 231–268. Rosén, H., ‘Le continuum des discours rapportés en latin du discours indirect vers le discours direct: schémas grammaticaux, critères lexicaux, et stratégies rhétoriques’, De Lingua Latina 10 (2015) 1–33. Rossi, A., ‘Parallel Lives: Hannibal and Scipio in Livy’s Third Decade’, TAPhA 134.2 (2004) 359–381. Schlicher, J., ‘The Development of Caesar’s Narrative Style’, CPh 31 (1936) 212–224. Schlicher, J., ‘The Moods of Indirect Quotation’, AJPh 26.1 (1905) 60–88. Sznajder, L., ‘Stratégies de Prises en charge énonciatives dans le discours indirect’, in G. Calboli (ed.), Latina Lingua! Proceedings of the Twelfth International Colloquium on Latin Linguistics (Bologna, 9–14 June 2003) (Rome 2005) 749–761. Treptow, R., Die Kunst der Reden in der ersten und dritten Dekade des Livianischen Geschichtswerkes (Kiel 1964). Tsitsiou-Chelidoni, C., ‘History Beyond Literature: Interpreting the ‘Internally Focalized’ Narrative in Livy’s Ab Urbe Condita’, in J. Grethlein & A. Rengakos (eds.), Narratology and Interpretation (Berlin 2009) 527–554. Ulmann, R., Étude sur le style des discours de Tite Live (Oslo 1929). Utard, R., Le discours indirect chez les historiens Latins (Louvain 2004). Walsh, P.G., Livy: His Historical Aims and Methods (Cambridge 1961).

words when it’s time for action

315

White, H., The Content of the Form: Narrative Discourse and Historical Representation (Baltimore 1987). Woodman, A.J., Rhetoric in Classical Historiography (London 1988). Yellin, K., Battle exhortation: The Rhetoric of Combat Leadership (Columbia 2008). Zoido, J.C.I., ‘The Battle Exhortation in Ancient Rhetoric’, Rhetorica 25.2 (2007) 141–158.

part 3 Beyond Thermopylae and Cannae

∵

chapter 13

Thermopylae and Cannae: How One Battle Narrative Enriches Another Mathieu de Bakker and Michiel van der Keur

1

Introduction*

When Tlepolemus, son of Heracles, and Sarpedon, son of Zeus, meet on the Trojan battlefield, the former refers to his father’s capture of Troy with only a handful of ships and comrades. Not intimidated, Sarpedon acknowledges Heracles’ just punishment of Priam’s father Laomedon, but announces that in contrast to the past, today’s victory will be Trojan. Indeed, when Zeus’ son and grandson throw their spears, Tlepolemus is mortally wounded (Hom. Il. 5.627– 5.662). The dialogue between the heroes illustrates that already in the oldest example of European literature, great battles do not stand on their own. They are conceptualized in terms of earlier famous battles, both by those who report them, and by those who are involved, although their outcome may be different: Tlepolemus is not able to repeat his father’s feat. The Greek historian Herodotus followed Homer in lowering a backdrop of precedents behind his narrative of the great battles of the Persian Wars. Some of these precedents took place on the very spot where the later battles were fought. In his description of the battle of Thermopylae, for instance, he refers to the earlier war between Phocians and Thessalians that led to the building of a wall across the entrance of the pass (Hdt. 7.176.4–7.176.5; 7.215). The Greeks decide to rebuild this wall to defend themselves against the Persians.1 Alternatively, Herodotus refers to earlier battles in the same campaign, such as in the case of Salamis, where he favourably compares the bravery of the Persians with their less glorious performance at Artemisium (Hdt. 8.86). Herodotus makes his characters, too, introduce past battles in their speeches. At Plataeae, for instance, the Athenians claim the honorific position on the wing with a reference to their success against the Persians at Marathon (Hdt. 9.27.5–9.27.6).2 * We wish to thank the editors of this volume for their suggestions and directions, which much helped us in rewriting this paper and sharpening its arguments. We thank Hannah Kousbroek for correcting our English. 1 See in this volume the chapters by De Jong and De Bakker. 2 Herodotus also makes the Athenians and their opponents in the debate, the Arcadians, refer

© koninklijke brill nv, leiden, 2019 | doi:10.1163/9789004383340_014

320

de bakker and van der keur

Herodotus’ successor Thucydides ascribes similar claims to the Athenians, who mention Marathon and Salamis to exemplify their unconditional and disinterested willingness to sacrifice themselves on behalf of the Greek community (Thuc. 1.73.4–1.73.5).3 In his own voice, Thucydides compares the crucial battle at Sphacteria to the events at Thermopylae (Thuc. 4.36.3) and likewise evokes the battle of Salamis when he describes the final Athenian stand in the great harbour of Syracuse (Thuc. 7.69.3–7.71).4 The Roman historian Livy wrote his narrative of the Hannibalic War, the greatest foreign threat to Rome in recent history, for an audience that had a good knowledge of earlier historiography, as most members of the Roman elite were schooled in a tradition of historical exempla that could be used as moral guidance and invoked in their speeches.5 It is not surprising, then, that in the preface of book 21, we encounter an allusion to Thucydides’ Histories in Livy’s claim that his war is ‘by far the most memorable of all those ever conducted’ (bellum maxime omnium memorabile quae umquam gesta sint, Liv. 21.1.1; compare Thuc. 1.1.1: ἀξιολογώτατον τῶν προγεγενημένων ‘most memorable compared to all earlier ones’) as no states had ever fought one another that were ‘healthier in terms of resources’ (ualidiores opibus, Liv. 21.1.2; compare Thuc. 1.1.1: ἀκμάζοντές … ᾖσαν … ἀμφότεροι παρασκευῇ τῇ πάσῃ ‘they went (to war) at their peak in terms of all their resources’).6 In presenting the topos of the unprecedented magnitude of his subject, Livy places himself in the historiographical tradition of his Greek predecessors. In this chapter we take a closer look at the way in which he uses their works in shaping the narrative of the battle of Cannae. We hope to show that apart from scenes that are typical of any ancient war narrative, such as arduous river crossings and councils of war,7 the Cannae narrative evokes the Greek tra-

3 4 5

6 7

to older battles that were fought in a legendary past. The Athenians reject those in favour of their recent performance at Marathon. For a more in-depth discussion of this passage, see Flower & Marincola 2002: ad 9.27.1–9.27.6. See furthermore Grethlein 2006 for a similar use of the legendary past in the embassy scene at Gelon’s court (Hdt. 7.153–7.163). In general Marathon becomes a lieu de mémoire throughout Greek literature, see Jung 2006. See in this volume Rademaker. These were primarily taken from their own past (see e.g. Van der Blom 2010: 12–17), but the Roman practitioners of historiography commonly rendered their national past with an eye on Greek history, too; see e.g. Wiseman 1979: 23–24; Momigliano 1990: 107. For more general literature on the Roman reception of the Greek historians, see Marincola 1997; Canfora 2006; Dillery 2009; O’Gorman 2009; and Samotta 2012. For the topos, see Marincola 1997: 34–43, and for Livy’s use of Thucydides, Rodgers 1986, with 336 n. 6 referring to 21.1. For a recent study of such structural conventions and topoi in battle descriptions in ancient historiography, see Lendon 2017a and 2017b.

thermopylae and cannae

321

ditions on the battle of Thermopylae and, in some cases, specifically recalls Herodotus’ account.8 It will be shown that the majority of these points of contact can be found in the aftermath of Cannae, in the evaluation of the battle and its significance, which is the actual focal point of Livy’s Cannae narrative.9 First, we will identify three specific allusions and discuss their effects within the immediate context (§2).10 Next, we will discuss broader, thematic similarities and show to what extent the lowering of the Thermopylae tradition as a backdrop enriches our evaluation of the Cannae events as told by Livy (§3). In general we will argue that—by making use of themes and narrative strategies derived in particular from Herodotus’ account—he tried to make his audience sensitive to issues of contemporary relevance and pointed at ways in which the course of the crucial events in Hannibal’s war against Rome should be read.11

2

Specific Allusions to Thermopylae

In this paragraph we discuss three specific allusions that Livy appears to make to the Thermopylae traditions as told by Herodotus and Diodorus Siculus.12 With so much of early Roman historiography lost to us (including the works of, e.g., Silenus and Coelius Antipater on the Second Punic War), it cannot be said with certainty that Livy would be the first author to evoke Thermopylae in a narrative of the battle of Cannae.13 In Polybius’ earlier version of the Cannae episode, however, allusions to Thermopylae are absent, which suggests an innovation on the part of Livy, whose stylistic and narrative choices in elaborating this comparison are anyway his own.

8 9 10

11 12 13

For case-studies of allusions to Herodotus in other passages in Livy’s Ab Urbe Condita, see Schubert 1991 and Meulder 2005. See in this volume Oakley. In this chapter we reserve the term allusion for unambiguous references by Livy to his predecessors (including, but not limited to, verbal echoes). For a discussion of the various forms allusions can take, see Hinds 1998: 17–51. For the historiographer’s rhetorical use of color to suit his material to his own program and interpretation of the past, see Wiseman 1979 and Woodman 1988. For a comparison between the accounts of Herodotus and Diodorus on Thermopylae, see in this volume De Bakker. For a brief overview of Livy’s sources, see in this volume Oakley n. 3.

322

de bakker and van der keur

2.1 Aemilius Paulus’ Last Stand and Death Speech From Homer onwards, battle narratives tend to spell out the deaths of great warriors, usually with details about their last stand, and, in some cases, final speeches that highlight their awareness of the imminent end of their lives. Classic examples are Patroclus and Hector in Homer’s Iliad, Leonidas in Herodotus’ Histories, and Nicias in Thucydides’ Histories.14 Livy’s Cannae narrative is no exception, as it devotes a paragraph to Aemilius Paulus’ last stand and final speech. This passage however contains elements that are familiar from the Thermopylae tradition, and—in Paulus’ final speech—a significant allusion. The scene begins with the moment that Paulus makes his cavalry dismount to continue fighting on foot, which the narrator marks as a point of no return: equitum pedestre proelium, quale iam haud dubia hostium uictoria, fuit, cum uicti mori in uestigio mallent quam fugere, uictores morantibus uictoriam irati trucidarent quos pellere non poterant. Liv. 22.49.4

A battle on foot of the cavalry ensued, of such a kind that it was now no longer in doubt that the enemy would prevail, with the vanquished preferring death on the spot to running away, and the victors, angry at those who delayed their victory, hewing down those whom they could not turn to flight. As in the case of the Spartans and Thespians at Thermopylae, the Romans around Paulus know that they are doomed, but choose to die gloriously instead of surrendering. Livy reports a tradition that ascribes to Hannibal a memorable statement about the consul’s decision to dismount: tum denuntianti cuidam iussisse consulem ad pedes descendere equites dixisse Hannibalem ferunt: ‘quam mallem, uinctos mihi traderet’. Liv. 22.49.3

It is reported that Hannibal at that point said to someone who announced that the consul had ordered his cavalry to dismount: ‘How much more would I have liked him to hand them over to me in chains!’

14

Patroclus: Hom. Il. 16.844–16.854. Hector: Hom. Il. 22.297–22.305; 22.338–22.343; 22.356– 22.360. Leonidas: Hdt. 7.220.2 (compare Diod. Sic. 11.4.4, 11.9.1 and 11.9.4). Nicias: Thuc. 7.85.1.

thermopylae and cannae

323

By embedding Hannibal’s ironic15 exclamation in anonymous indirect speech ( ferunt), Livy employs a narrative strategy familiar from Herodotus, who sometimes treats such short, memorable statements in a similar manner, and thereby qualifies them as belonging to a collectively held tradition, granting them independent authority. In Herodotus’ Thermopylae narrative, Dieneces’ memorable statements (Hdt. 7.226) are embedded in a similar way, and Diodorus (or his sources) copies the strategy by embedding Demaratus’ advice to Xerxes (Diod. Sic. 11.6.2) in an anonymous indirect speech.16 Hannibal’s statement anticipates his strategy of raising money and undermining Roman morale by selling his prisoners, a theme that Livy elaborates on in his narrative of the aftermath of the battle (see below). Though deliberate references to Thermopylae in the introduction to Paulus’ last stand and in Hannibal’s exclamation cannot be established with certainty, an evident allusion can be identified in the ensuing dialogue of the mortally wounded Paulus with the tribune Lentulus, who offers him a horse and his personal protection:17 tu quidem, Cn. Corneli, macte uirtute esto […]. abi, nuntia publice patribus urbem Romanam muniant ac priusquam uictor hostis adueniat praesidiis firment; priuatim Q. Fabio L. Aemilium praeceptorum eius memorem et uixisse [et] adhuc et mori. me in hac strage militum meorum patere exspirare, […]. Liv. 22.49.9–22.49.11

‘Be blessed, Gnaeus Cornelius, for your virtue […]. Go tell the senators publicly to fortify the city of Rome and to strengthen it with garrisons before the victorious enemy arrives; and tell Quintus Fabius privately that Lucius Aemilius has hitherto lived and now dies faithful to his precepts. Allow me to expire here in this massacre of my soldiers, […].’ Livy makes Paulus’ words echo the epigram that was set up at Thermopylae in honour of the fallen Spartans: 15

16 17

Cf. Foster in the 1929 Loeb translation, p. 359, note 4: ‘an ironical intimation that, since the consul’s order amounted to depriving his troopers of any hope of escape, he might as well have surrendered them at once and saved Hannibal all further trouble.’ See also Walsh 1961: 104. See in this volume Adema for a discussion of Hannibal’s statement from a text internal perspective. See in this volume De Bakker. See in this volume Adema for an interpretation of this dialogue within the context of Livius’ Cannae episode itself.

324

de bakker and van der keur

Ὦ ξεῖν’ ἀγγέλλειν Λακεδαιμονίοις ὅτι τῇδε κείμεθα τοῖς κείνων ῥήμασι πειθόμενοι. Hdt. 7.228.2

Stranger, go tell the Spartans that here we lie honouring their precepts. Like the epigram that honours Spartan resilience and self-sacrifice in battle, Paulus asks his addressee to deliver (abi nuntia ≈ ἀγγέλλειν) a message of uncorrupted principles on his part (praeceptorum eius memorem ≈ τοῖς κείνων ῥήμασι πειθόμενοι). Like the Spartans, he insists that he has kept to these principles up to his death (mori ≈ κείμεθα). Stylistically, too, there are points of contact, as Livy makes Paulus raise the solemnity of his last words by greeting Lentulus with the old-fashioned formula tu quidem … macte uirtute esto18 and by referring to Rome with the periphrastic urbem Romanam (‘the city of Rome’). In contrast to the epigram that honours the Spartiates, however, the message that Paulus sends to Rome is divided into a public and private part (publice … priuatim). This difference between the two messages also accounts for the variation in form, with Paulus’ goodbye to Fabius presented as a private message of an individual, whereas the epigram at Thermopylae concerns a public message on behalf of a collective. The public part of Paulus’ message is one of foresight, as he warns the Romans to provide garrisons in the expectation that Hannibal will now set his eyes on the city. The private part of his message is concerned with his obedience to Fabius’ precepts, and refers to the advice that the dictator offered Paulus when he set out on the arduous task of fighting the Carthaginians and keeping his fellow consul Varro in check at the same time (Liv. 22.39). The private part of Paulus’ death speech, then, creates ring composition with Fabius’ earlier advice, and further sharpens the contrast between the two consuls, which had already been exposed in the intervening narrative itself.19

18

19

Vallet 1966: ad loc. ‘formule appartenant au langage religieux, employé à date ancienne dans une prière accompagnant une offrande ou un sacrifice’. If we follow this observation, the self-chosen heroic death of Paulus could be conceived of in terms of a sacrifice needed to ensure victory. Compare, in this respect, Leonidas’ role at Thermopylae. According to Herodotus, it was predicted that Sparta would pay with the life of a king for its survival against the Persians, and Leonidas took this into account when he decided to fight to the death (Hdt. 7.220.3–7.220.4). According to Walsh 1961: 231, Livy took the speeches in this episode from his predecessor Valerius Antias, preferring ‘a love of oratory’ to more sober accounts without speeches, such as Polybius’ one. Although Livy may have followed Antias’ narrative manners, we do not see why he could not himself be responsible for the composition and insertion

thermopylae and cannae

325

Paulus dies, ‘buried under an avalanche of missiles’ (consulem … obruere telis, Liv. 22.49.12). This is different in Polybius’ version of the battle, who reports Paulus’ death ‘in close combat’ (ἐν χειρῶν νόμῳ, Polyb. 3.116.9). It appears as if Livy makes Paulus die in the same way as the Spartans died at Thermopylae, buried under an avalanche of arrows, in both Herodotus’ and Diodorus’ version (cf. Hdt. 7.225.3—observe that Livy echoes Herodotus’ κατέχωσαν … βάλλοντες; cf. Diod. Sic. 11.10.4). Together with the echo of the epigram on behalf of the Spartans, this correspondence reinforces our claim that Livy deliberately ‘Thermopylizes’ the Cannae story, and monumentalizes the death of Aemilius Paulus by raising him to the same standard as the Spartans, who make a heroic last stand before they fall prey to Xerxes’ troops. Polybius does not report any words spoken from Paulus to Lentulus in his account of the consul’s death, but he awards him a short character evaluation in which he praises his righteousness and unbroken principles: Λεύκιος Αἰμίλιος περιπεσὼν […] μετήλλαξε τὸν βίον, ἀνὴρ πάντα τὰ δίκαια τῇ πατρίδι κατὰ τὸν λοιπὸν βίον καὶ κατὰ τὸν ἔσχατον καιρόν, εἰ καί τις ἕτερος, ποιήσας. Polyb. 3.116.9

Lucius Aemilius fell […] and passed away, a man who performed all sorts of righteous deeds, more than anyone else, on his fatherland’s behalf throughout his life and during his ultimate stand. Polybius and Livy appear to report a tradition that seeks to exonerate Paulus from blame for the course of events at Cannae. Evidently, this tradition presents Paulus as a man of old-fashioned patrician mores, who prefers a glorious death on the battlefield to ignominious surrender and captivity. In his evaluation, Polybius refers to the consul’s righteousness throughout his life and during his final stand. His use of a formula familiar from Greek honorary inscriptions of the Roman Era (τὰ δίκαια ποιέω) gives the evaluation the flavour of an epitaph, a small verbal ‘monument’ erected on Aemilius Paulus’ behalf in the narrative.20

20

of the speeches in his narrative. Adema’s contribution in this volume points in the same direction. For the emphasis on the opposition between Varro and Paulus, see section 3.2 below. Cf. Schulte 1909: 218. For the formula on inscriptions, see for instance IG II/III2 1034, 10; 1036, 13; 1343, 20–21; IG V, 1, 1145, 17–18 (all from first c. BCE).

326

de bakker and van der keur

In contrast to Polybius, Livy does not evaluate the consul’s character in his own words, but makes Paulus voice these Roman principles by himself in the private part of his message to Fabius Maximus. Like Polybius, he couches his message in the solemn language of epigraphy, but unlike him, he does not use a common epigraphical formula but, as we saw, alludes to the inscriptional heritage of the battle of Thermopylae. In doing so, he awards Paulus the role of protagonist in a story of glorious defeat that enables the Romans to regain courage, reorganize their defence, and continue the fighting against Hannibal, much in the same way as the dogged Spartan resistance against the Persians at Thermopylae serves as an inspiration for the Greek allies, and ultimately brings victory (Diod. Sic. 11.11). 2.2 Manlius Torquatus’ Address to the Senate A second allusion to Thermopylae is found in the speech that Livy ascribes to T. Manlius Torquatus (Liv. 22.60.6–22.60.27). This speech seals the debate in the senate about the appropriate response to a request from a group of captive Romans to be ransomed.21 As Livy describes in 22.50 and 22.52, the behaviour of these Romans prior to their capture had been far from glorious. After the battle, they seek refuge in their camp. Although they are encouraged to reach friendly territory under the shelter of the night, most of them refuse to do so. Only a small band of 600 escapes, led by the tribune P. Sempronius Tuditanus, who holds up the noble death of Paulus as an example (Liv. 22.50.7). The next morning, those who stayed in the camp surrender to Hannibal without a fight (Liv. 22.52.1–22.52.2), who later allows ten of them to go to Rome and beg the Senate to ransom all prisoners. Initially, the Senate seems willing to raise the money to secure their release, but then Torquatus objects to his colleagues’ views and scorns the captives’ un-Roman behaviour. In his speech, he mentions examples of earlier Roman generals who refused to surrender in spite of being outnumbered: si, ut auorum memoria P. Decius tribunus militum in Samnio, si, ut nobis adulescentibus priore Punico bello Calpurnius Flamma trecentis uoluntariis, cum ad tumulum eos capiendum situm inter medios duceret hostis, dixit: ‘Moriamur, milites, et morte nostra eripiamus ex obsidione circumuentas legiones’, si hoc P. Sempronius diceret, nec uiros equidem nec Romanos uos ducerem, si nemo tantae uirtutis extitisset comes. Liv. 22.60.11–22.60.12

21

See also Oakley’s discussion of this scene in this volume.

thermopylae and cannae

327

‘If Publius Sempronius had spoken words like (as our grandfathers remembered) those of Publius Decius, the military tribune, in Samnium, or such words as Calpurnius Flamma spoke in the first Punic War (when we were boys) to his three hundred volunteers, when he led them out straight through the enemies in order to capture a hill: “Let us die, soldiers, and by our death free the surrounded legions from being besieged”, if Publius Sempronius had spoken thus, I would consider you neither men nor Romans if no one had offered himself as comrade of such bravery.’ Both examples recall in their own manner the Spartan spirit at Thermopylae, especially in the account of Diodorus, who presents Leonidas’ expedition to confront the Persians as a suicide mission with an eye on the common salvation of the Greeks (Diod. Sic. 11.4.4–11.4.5; 11.11.1; 11.11.5).22 The allusion is signalled most clearly by the reference to Flamma’s ‘three hundred’, who are ordered to occupy a mount (tumulus—compare the hill, κολωνός, where the Spartans make their ultimate stand in the pass of Thermopylae, Hdt. 7.225.2–7.225.3) in the midst of the enemy so as to distract the attention of the Carthaginians and allow the Roman armies a means to escape.23 The example of Decius refers to a story told by Livy in an earlier book (Liv. 7.34–7.36). Serving as a tribune during the first Samnite War, P. Decius [Mus] saved the consul’s army from entrapment by the enemy by occupying a hill with part of the army. Unlike Flamma’s three hundred, Decius and his men survived, escaping through the enemy camp during the night, much like Sempronius, and causing havoc and fear. Not less significant is the association with self-sacrificial deuotio that Decius’ name evokes. Later, as consul during the Latin War, he received a vision that the army whose commander would devote himself and the army of the enemy to the gods of the underworld, would obtain victory; by choosing this heroic death, Decius allowed the other Romans to escape and retain their liberty.24 His co-consul in that war was Titus Manlius Imperiosus Torquatus, ancestor of the Torquatus in 22 23

24

See in this volume Van Wees and De Bakker. See Livy Periochae book 17 and compare Frontinus, Str. 1.5.15. An explicit comparison between Flamma and Leonidas is made by Cato the Elder in Orig. 4.7a (Chassignet = 83 [Peter], apud Gell. 3.7.19); Krebs 2006 argues that Cato used Herodotus as his model. Compared to Cato, Livy has adjusted the number of Romans from four hundred to three hundred to match the Spartans even more closely. See also Ruffing 2013: 202–203, who refers to Lucius Ampelius’ observation (in his Liber Memorialis 20.5) that both numbers and heroism of Flamma’s and Leonidas’ troops were the same. Liv. 8.6, 8.9. For the same act by his son Publius Decius Mus I in the Third Samnite War (298–290 BCE), see Liv. 10.28.12–10.28.18.

328

de bakker and van der keur

Liv. 22.60.25 Mus’ choice recalls the oracle given to the Spartans before Xerxes’ invasion, namely that they would either lose their king or their freedom to the Persians. Herodotus quotes this oracle in a crucial passage of his report, when the Greeks at Thermopylae are faced with a choice between escaping in safety and making a final, heroic stand. According to Herodotus, the oracle is the key motivation for Leonidas to stay (Hdt. 7.220.3–7.220.4): it is in the knowledge that the Spartans will retain their liberty that Leonidas himself opts for glory.26 By evoking Leonidas’ choice and the self-sacrifice of the three hundred Spartans in Torquatus’ commentary on the aftermath of Cannae, Livy draws attention to other passages in this part of his narrative that have elements in common with the Thermopylae tradition. Some resemblance can be assumed, for instance, between Sempronius’ (as well as Decius’) heroic escape through enemy territory under the shelter of night, and the night raid upon Xerxes’ camp by Leonidas’ men in the alternative Thermopylae tradition as told by Diodorus (Diod. Sic. 11.9.3–11.10.4).27 More importantly, however, a thematic similarity can be observed between Livy’s account of the Cannae survivors and that of Herodotus, who deals with the fate of Aristodemus, one of the Spartiate survivors of Thermopylae, and that of the Theban hostages in the aftermath of the battle (Hdt. 7.229–7.233). Both Herodotus and Livy raise the moral question of how to treat the survivors of a defeat who have not behaved with courage, and both use the strategy of presenting variant versions to home in on the shame that falls upon those who have opted for inglorious survival (Hdt. 7.229–7.230; Liv. 22.61). Furthermore, both historians emphasize that the survivors could have chosen a more heroic alternative. In the case of Aristodemus, Herodotus refers to his partner Eurytus, who—despite being blinded by disease (Hdt. 7.229.1)—orders his slave to lead him to the battle. In Livy’s case, the nightly escape under Tuditanus reveals that those who surrender could have chosen a different route (Liv. 22.50.6–22.50.12). In dealing with the post-battle fate of the survivors, both historians refer to their status as pariahs. Herodotus mentions that Aristodemus was deprived of his citizen rights (ἀτιμίη) upon coming home, and was called a ‘trembler’

25

26 27

He and Decius decided that no Roman soldier was to leave his post under penalty of death, which forced him to execute his own son (Liv. 8.7). His descendant in 22.60 shows a similar ‘Manlian’ discipline. Livy appears to have chosen a Torquatus to voice the rigid, Spartan line of ‘no surrender’. For the practice in Roman historiography to associate an entire gens with certain traits and character types, see Walsh 1961: 88–91 and Wiseman 1979: 25. For Leonidas’ framing of his own heroic death, see Baragwanath 2008: 64–78 and the contribution of De Jong in this volume. For the fictional nature of this episode, see in this volume Van Wees and De Bakker.

thermopylae and cannae

329

(τρέσας, Hdt. 7.231), with no one willing to kindle his fire or greet him. Livy refers to similar incidents of social excommunication in his account of the fate of the captive Romans who came to negotiate about the ransom (Liv. 22.61). In the first version of this account, the entire group is sent back to Hannibal, with one of them in chains, as he has attempted to deceive the Carthaginians, who had sent them forth under an oath that they would return (Liv. 22.61.1– 22.61.4). In the second version, all ten deceive the Carthaginians, but the Senate allows them to stay. They are so much laden with shame, however, that some avoid the Forum during their remaining lives, whereas others commit suicide (Liv. 22.61.9). Suicide is also the path followed by Pantites, another Spartan whom Herodotus reports as having survived Thermopylae (Hdt. 7.232). The ultimate form of shame is left for the Theban regiment, whom Leonidas suspected of pro-Persian sympathies and therefore kept with him throughout the battle (Hdt. 7.205.3; 7.222). They indeed defect to the Persians and are afterwards branded with ‘royal marks’ (στίγματα βασιλήια, Hdt. 7.233.2) as lasting evidence of their submission to Xerxes.28 As in the case of the death of Aemilius Paulus, it is instructive to compare Livy’s version with that of Polybius, who singles out the senate’s decision on Hannibal’s prisoners as exemplary of Roman discipline and steadfastness, even in the severest adversity.29 For this purpose he has placed it not in his narrative of the aftermath of Cannae, but at the end of his digression on the Roman constitution (Polyb. 6.58). The last part of this digression is devoted to a comparison with the constitutions of other states like Sparta and Carthage, and with this anecdote Polybius wishes to explain what caused the Romans to prevail against Hannibal. Without referring to individual speakers, Polybius summarizes the arguments of the prisoners and the response of the senate in indirect speech: ἐνομοθέτησαν ἢ νικᾶν μαχομένους ἢ θνήσκειν, ὡς ἄλλης οὐδεμιᾶς ἐλπίδος ὑπαρχούσης εἰς σωτηρίαν αὐτοῖς ἡττωμένοις. Polyb. 6.58.11

They set the standard that they either had to be victorious in battle or had to be killed, as there was no means of survival to be expected when they had been defeated.

28 29

For the conflicting traditions about the role played by the Thebans during the Persian invasion, see also in this volume De Bakker. For a systematic comparison between both versions, see Oakley in this volume.

330

de bakker and van der keur

Polybius presents the story as an example of unbroken Roman morale when faced with hardship, and both he and Livy concur in considering Roman steadfastness the ultimate cause for victory over the Carthaginians. But whereas Polybius uses the story to illustrate the unparalleled strength of the Roman constitution, Livy presents the debate at a moment crucial for Rome’s survival, and awards a pivotal role to Torquatus in persuading the senate. He adds elements of the Thermopylae tradition to illustrate that like the Spartans after Thermopylae, the Romans were defeated on the battlefield, but undefeated in their minds and ready to face further challenges.30 2.3 Biting the Enemy: When Heroism Turns into Horror A third point of contact between Cannae and the Thermopylae tradition concerns a detail of the way in which the battles themselves are fought. Livy reports the following incident in the aftermath of the battle of Cannae, when the victorious Carthaginians inspect the battlefield: praecipue conuertit omnes subtractus Numida mortuo superincubanti Romano uiuus naso auribusque laceratis, cum manibus ad capiendum telum inutilibus in rabiem ira uersa laniando dentibus hostem exspirasset. Liv. 22.51.9

What disturbed all of them mostly was a Numidian who was drawn away alive from under a dead Roman who lay on top of him. The Numidian’s nose and ears were torn apart, for the Roman had, when his hands were no longer able to pick up a weapon, with anger turned to frenzy, expired whilst rending his opponent with his teeth.31 In the case of the battle of Thermopylae, Herodotus also refers to a fight with teeth by the last Spartans who defend their position on the hill: ἐν τούτῳ σφέας τῷ χώρῳ ἀλεξομένους μαχαίρῃσι, τοῖσι αὐτῶν ἐτύγχανον ἔτι περιεοῦσαι, καὶ χερσὶ καὶ στόμασι κατέχωσαν οἱ βάρβαροι βάλλοντες, […]. Hdt. 7.225.3

30

31

Polybius also refers to one delegate who committed perjury and asked for asylum in Rome. The senate, however, objects to his fraudulent behaviour and sends him back to Hannibal in chains (Polyb. 6.58.12). This concurs with Livy’s first version (Liv. 22.61.1–22.61.4). For Silius Italicus’ dramatic version of this scene, see Pun. 6.41–6.53.

thermopylae and cannae

331

There they defended themselves with their knives, at least those that still happened to have them, and with their bare hands and teeth, whilst the Persians buried them under an avalanche of arrows. Both incidents of biting—almost unique in the historiographical tradition32— appear to underline the unbroken fighting spirit of the Romans and the Spartans in the face of a certain death. In contrast to serene words of uncorrupted principles, they embody instead the ultimate, sordid consequence of such principles when making a heroic last stand against the enemy. With all means of defence exhausted, Spartans and Romans resort to the use of their teeth as a means to inflict wounds. Although the ‘biting-in-battle’ motif in Cannae may have been independently attested, Livy’s addition of such a salient detail again suggests engagement with the Thermopylae tradition as it is found in Herodotus. A possible reason for the inclusion of the motif can be found in the treatise On the Sublime, which praises Herodotus’ description of the Spartans’ last stand on the hill as an instance of ‘extravagant phrasing’ (ὑπερβολή, 38.3–38.4) used to underline the greatness of extraordinary events. Its graphic nature is explained as resulting from the wish to heighten the pathos of the event in accordance with its greatness and importance (ἐπειδὰν [αἱ ὑπερβολαὶ] ὑπὸ ἐκπαθείας μεγέθει τινὶ συνεκφωνῶνται περιστάσεως, On the Sublime, 38.3), with the author looking for a powerful image to emphasize the intensity of the battle. If Livy was indeed inspired by Herodotus when he included this reference to the biting Roman, we are dealing not only with an allusion, but also with the copying of a literary technique.

3

Larger Thematic Connections between Cannae and Thermopylae

Now that a number of correspondences between Livy’s account of Cannae and the Greek Thermopylae tradition have been established, it is time to look at the role of the Cannae episode within Ab Urbe Condita as a whole. We hope to show that Livy’s presentation of Cannae as a ‘Thermopylae’ helps him in ascribing to the Roman defeat the same pivotal and programmatic function as Thermopylae had in the Greek tradition on the Persian Wars (notably Herodotus): at the 32

The only parallel we have found is in Cassius Dio’s narrative of the Battle of the Vosges (38.49.3) between Caesar and Ariovistus’ Germans in 58 BCE. Here, it is said that the Germans, when deprived of their weapons, used their ‘hands and teeth’ (ταῖς τε χερσὶ καὶ τοῖς στόμασιν).

332

de bakker and van der keur

moment of greatest defeat, future victory is already anticipated, but at the same time, the unified defence against a common foreign enemy also carries the spark of later internal strife. 3.1 Defeat as Anticipation of Future Victory In section 2.2, we focused on the parallels between the Spartan and Roman reactions to their respective defeats; but the aftermath narratives also display other correspondences. Immediately after the battles, the victorious foreign leaders (Xerxes and Hannibal, respectively) are advised by subordinates on how to exploit the victory with an aggressive strategy. In Herodotus, the Spartan exile Demaratus recommends sending part of the Persian fleet to the island of Cythera, from where they would be able to attack Spartan territory directly;33 the Spartans would thus be kept from helping the rest of Greece against the enemy. Xerxes is dissuaded by his brother Achaemenes, admiral of the Persian navy, who admonishes him to keep the army and fleet together so that they can support each other. Livy presents a similar pair of recommendations, one cautious and one more aggressive. Hannibal is advised by many to grant himself and his men some rest after the battle of Cannae, but the cavalry commander Maharbal urges him to press on to Rome itself, for ‘in five days you will dine as victor on the Capitoline’ (Liv. 22.51.2 die quinto … uictor in Capitolio epulaberis).34 Hannibal is not convinced, however, pleading the need for time to consider. Both authors present the aggressive (but ignored) advice as having been correct, in hindsight.35 Maharbal’s famous retort that Hannibal knows how to obtain a victory but not how to use it (Liv. 22.51.4 uincere scis, Hannibal; uictoria uti nescis) is immediately confirmed by the narrator’s statement that this one day of delay is thought to have been enough to save Rome (Liv. 22.51.4 mora eius diei satis creditur saluti fuisse urbi atque imperio).36 Later, Livy even reports Hannibal’s own regrets when he was forced to return to Africa to confront Scipio: 33 34 35

36

Hdt. 7.235.2–7.235.3. For the background of this bon mot in the historiographical tradition, see Hoyos 2000. At 22.58.1, Livy again portrays Hannibal as a general who prematurely considers himself victorious (uictoris magis quam bellum gerentis intentus curis). Most modern scholars agree, however, that Hannibal’s decision not to attack Rome was prompted by strategic reasons; see e.g. Lazenby 1978: 85–86 and 1996 (the march would have taken weeks; Hannibal would face a protracted siege; taking Rome had probably never formed part of his plans); Shean 1996 (the major problems of supplying the army). For a discussion of the narrator’s endorsement of Maharbal’s view, see in this volume Oakley and Adema.

thermopylae and cannae

333

respexisse saepe Italiae litora, et deos hominesque accusantem in se quoque ac suum ipsius caput exsecratum, quod non cruentum ab Cannensi uictoria militem Romam duxisset. Liv. 30.20.7

They say he often looked back at the shores of Italy, and accusing gods and men he poured curses also on himself and his own head, that he had not led his men still blood-spattered after the victory at Cannae straight to Rome. Herodotus has incorporated the vindication of Demaratus’ strategy within the advice itself. The mention of Cythera recalls—for the contemporary audience—the effective strategy used by the Athenians in 424BCE, when Nicias took the island and used it as a base from which to raid the Laconian coast, delivering a further blow to Spartan morale (Thuc. 4.55); the Spartans were unable to match the naval superiority of the Athenians.37 Demaratus also lists the disadvantages that the Persian army would experience if his counsel were not followed—all Peloponnesian states will put up a unified defence at the Isthmus, promising a much harder battle than Thermopylae. Such a battle never took place, but to the audience, Xerxes’ disregard for Demaratus’ advice is nevertheless a form of dramatic irony; moreover, the narrator already observed (Hdt. 7.139) that the Greek defence at the Isthmus would have failed if the Athenians had left the alliance and the Persian navy were free to conquer the Peloponnese. The decision to keep the army and fleet together (as advocated by Achaemenes) proved disastrous when the Persian navy, with a view to working in concert with the land forces marching to the Isthmus, engaged the Greeks at Salamis. Demaratus’ words, like Maharbal’s, are the more remarkable for being inserted at the height of success for the invading army. On the one hand, these ignored recommendations serve as compelling counterfactuals, inviting the audience to consider how history would have developed if the more aggressive strategy had been used.38 On the other, they highlight the fact that the ‘winning’

37

38

For the dating of Herodotus’ passage in the light of Nicias’ conquest in 424 BCE, see Fornara 1971: 33–34; and again 1981: 151 in response to the objections in Cobet 1977: 6–7. Demaratus invokes the authority of Chilon, one of the Seven Sages, who had warned his fellow Spartans against the island; if the connection with the events of 424 BCE is accepted, it appears that Herodotus has credited these words as a uaticinium ex eventu to the sage to underscore the correctness of Demaratus’ advice. Cf. Ruffing 2013: 206. The motif that the defenders only narrowly escape defeat is often used in a peripeteia

334

de bakker and van der keur

strategy (as it is presented) was not followed. At the moment of triumph, the victorious leader fails to secure complete victory; or, in the view of the defending Greeks and Romans, the moment of defeat already carries a promise of future victory.39 Accordingly, when the tables are turned, both authors look back on the earlier defeat. At Hdt. 9.64 and 9.79, Mardonius’ death and the Persian defeat at Plataeae are presented as compensation for Leonidas’ demise at Thermopylae.40 Livy, similarly, looks back on Cannae when the tables have turned in the Hannibalic War. When Hannibal’s brother Hasdrubal has perished with his entire army at the battle of the Metaurus, the historiographer concludes: nunquam eo bello una acie tantum hostium interfectum est, redditaque aequa Cannensi clades uel ducis uel exercitus interitu uidebatur. Liv. 27.49.5

Never before in that war were in a single battle so many of the enemy killed, and their defeat was regarded as equal repayment for that at Cannae, both in the demise of the commander and of his army. The reversal of fortune is thematized again prior to the battle of Zama, when Hannibal says to Scipio quod ego fui ad Trasumennum, ad Cannas, id tu hodie es (Liv. 30.30.12), ‘what I was at Trasimene and at Cannae, that you are today’. Scipio already has an important role to play in the aftermath narrative of Cannae. In the camp at Canusium, where the Roman survivors had escaped to after the defeat, some nobles led by M. Caecilius Metellus consider aban-

39

40

(a reversal of fortune), a common narrative topos in historiography, for which see Burck 1934: 210–217. Surprise is a fundamental element in a peripeteia; in Livius’ account of the aftermath of Cannae, Maharbal’s eager hopes for a swift victory and his incredulity at Hannibal’s refusal to march on Rome reflect the Roman view that this was one of the turning points in the war. For another instance of peripeteia in the Cannae narrative, see in this volume Oakley. See also in this volume De Jong, where she discusses Ephialtes’ future downfall as being announced at the height of his ‘success’. For the teleological approach to historiography adopted by both Herodotus and Livy, see Grethlein 2013. See for a full discussion De Jong in this volume. She also discusses Hdt. 8.114, where Xerxes promises the Spartans that ‘Mardonius here will give them such retribution as befits them’; the king envisages Mardonius’ victory but unwittingly prophesies his demise. Livy uses the same textual strategy of inserting a boastful but ambiguous comment which turns out to be an ironic prophecy of defeat, when he makes his Varro proclaim that he will put an end to the war on the first day he catches sight of the enemy (Liv. 22.38.7 se quo die hostem uidisset [bellum] perfecturum).

thermopylae and cannae

335

doning Italy (Liv. 22.53.5). They are stopped by Scipio, who forces them to swear not to desert the Roman republic.41 The youthful Scipio, here for the first time in his future role as supreme commander (Liv. 22.53.2), through force of character effectively restores the harmony that was sorely lacking under Varro and Paulus (see below).42 His decisive action and his virtuous and patriotic conduct foreshadow his future triumph over Carthage; Livy stresses the point by calling Scipio the fatalis dux huiusce belli (Liv. 22.53.6) at the decisive moment, when others want to deliberate, but the future victor decides to act instead.43 3.2 Defeat as Anticipation of Future Internal Strife In both authors, defeat thus contains the seeds of future victory. However, it also contains the seeds of future strife. In their descriptions of the heroic last stand against the foreign invader, Herodotus and Livy include elements that suggest the later internal wars of the Greeks and Romans, respectively. We have seen above that Demaratus’ reference to Cythera may evoke the actual strategy used by the Athenians in 424BCE during the Archidamian War; similarly, Herodotus seems to have included a veiled hint about Athenian naval expeditions around the Peloponnese during the same war in Achaemenes’ counterspeech, who argues that the Persian fleet should not be used to ‘sail around the Peloponnese’ (περιπλέειν Πελοπόννησον, Hdt. 7.236.2).44 In his account of the war against the Persians, Herodotus shows that, while the need for a unified defence is clear to the Greeks, the future rivalry between their more prominent cities has already been set in motion.45 In the battle of Thermopylae, the theme of internal strife centres primarily around the Thebans, who fight at the side of the Spartans against their will and defect to the enemy afterwards (Hdt. 7.205.3; 7.233). As a final note, Herodotus adds an interesting piece of information about the Theban leader Leontiades:

41 42

43 44 45

For the historicity of the oath scene, see Zimmermann 1997. His conduct has a counterpart in Herodotus’ Themistocles, when the Athenians argue that they should leave Attica altogether and settle elsewhere (Hdt. 7.143.3, ἐκλιπόντας χώρην τὴν Ἀττικὴν ἄλλην τινὰ οἰκίζειν, which mirrors the second half of Metellus’ misguided plan, deserta Italia ad regum aliquem transfugiant, Liv. 22.53.5). There, Themistocles convinces the assembly to confront the Persians at sea instead. See also Oakley in this volume. A notion already mentioned (but rejected) by Macan 1908: lxvii. Cf. Harrison 2002: 567 ‘The theme of the tension between the two leading cities of Greece, Athens and Sparta, is one that runs throughout the Histories.’ See also in this volume De Bakker.

336

de bakker and van der keur

τοῦ τὸν παῖδα Εὐρύμαχον χρόνῳ μετέπειτα ἐφόνευσαν Πλαταιέες στρατηγήσαντα ἀνδρῶν Θηβαίων τετρακοσίων καὶ σχόντα τὸ ἄστυ τὸ Πλαταιέων. Hdt. 7.233.2

His son Eurymachus was in later time killed by the Plataeans when he led four hundred Theban men and occupied the town of Plataeae. This sentence refers to Eurymachus’ attempt to seize Plataeae by surprise during peacetime, an action that opened the hostilities in the Peloponnesian War in 431 BCE (Thuc. 2.2–2.6). One dishonourable Theban action—their defection at Thermopylae—is here thus connected to another one—their attack upon Plataeae. The connection is made through a family tie; Eurymachus was the son of the Theban leader at Thermopylae. The theme of internal strife is also present in Livy. The Cannae episode opens with the tumultuous elections in 216 BCE, which were known for the struggle between the senate and the people (Liv. 22.34.1 comitia habita magno certamine patrum ac plebis), a struggle reminiscent of the later opposition of optimates and populares. Both parties present the other as opponents, or even as enemies of Rome. In a lengthy reported speech, the tribune Q. Baebius Herennius suggests that the war is being drawn out on purpose by the patricians, and that the plebeian nobles have made common cause with them (Liv. 22.34.3– 22.34.11); his allegation that the nobles, seeking war, had brought Hannibal to Italy (Liv. 22.34.4 ab hominibus nobilibus per multos annos bellum quaerentibus Hannibalem in Italiam adductum) casts the Roman elite as the real enemy.46 Against Varro, the people’s favourite, the nobiles put forward Aemilius Paulus as sole competitor—as a skilled opponent rather than as a colleague (Liv. 22.35.4 par magis in aduersando quam collega datur consuli).47 The opposition between the two consuls is far less pronounced in Polybius.48 In 3.108–3.109, Paulus is given an exhortatory speech in which he stresses the differences with the previous battles and declares himself optimistic about the Roman chances in the upcoming battle. In Livy, this optimistic confidence is 46 47

48

The accusation is repeated by Varro at 22.38.6 bellum arcessitum in Italiam ab nobilibus. For Herennius’ speech, see Oakley in this volume. Their opposition is already clear from the way in which each is introduced: Varro had endeared himself to the people (Liv. 22.34.2 plebi … conciliatum), Paulus was hostile towards the people (Liv. 22.35.3 infestum plebei). Setting out from Rome, Paulus is escorted by the foremost senators, Varro the plebeian consul by the plebs (Liv. 22.40.4). For a detailed discussion of the antagonism between Varro and Paulus in Livy, see in this volume Oakley and Buijs. See in this volume Oakley.

thermopylae and cannae

337

transposed to the troops themselves, who are heartened by Fabius’ recent successes (Liv. 22.36.5). By now, numerous signs have been given that Varro will not follow Fabius’ strategy, however, and the misplaced confidence of the troops strongly contrasts with the gloomy demeanour of Livy’s Paulus. Among the differences with the previous battles emphasized by Polybius’ optimistic Paulus is the fact that earlier the Roman generals marched to a decisive battle without even having set eye upon the enemy (μήδ’ ἑωρακότας τοὺς ἀνταγωνιστὰς παρατάττεσθαι καὶ συγκαταβαίνειν εἰς τοὺς ὁλοσχερεῖς κινδύνους, Polyb. 3.108.7); this is precisely the objection which Livy’s pessimistic Paulus voices in response to Varro’s claim that he would defeat Hannibal as soon as he caught sight of him: mirari se quidni qui dux priusquam aut suum aut hostium exercitum locorum situm naturam regionis nosset, iam nunc togatus in urbe sciret quae sibi agenda armato forent, [et] diem quoque praedicere posset qua cum hoste signis conlatis esset dimicaturus. Liv. 22.38.9–22.38.10

He marvelled—and why not?—that a general, before he knew either his own army or that of the enemy or the lie of the land or the character of the country, should now already know, while still in the city and in civilian clothes, what he should do when in the field, and could even predict the day on which he would do battle with the enemy. Taking several elements from Polybius’ characterization of Paulus, Livy has thus created an entirely different version to fuel his narrative of two consuls at odds and a nation in conflict with itself. The struggle between Varro (representing the people) and Paulus (representing the senate) remains an important motif throughout Livy’s Cannae episode. The discord between the consuls is emphasized just after an initial Roman victory, when Paulus, to Varro’s indignation, refuses to pursue the enemy, and Hannibal is said to be aware that the consuls are dissimilis discordesque (Liv. 22.41.5); the word discordia, evocative of civil strife, is used again just prior to the battle of Cannae to describe the disturbance in the Roman camp (Liv. 22.44.5 inde rursus sollicitari seditione militari ac discordia consulum Romana castra). But the opposition of the two consuls is most clearly thematized in Fabius’ programmatic speech to Paulus (Liv. 22.39), which colours the episode as a whole. Fabius points out that Varro will prove just as much an opponent as Hannibal—indeed a more dangerous one, since Paulus will have to contend with Hannibal only in battle, but with Varro at all times (Liv. 22.39.4–22.39.5). The sentences build towards a climax in which the suggestion of civil war is strongest, as Fabius observes that against Hannibal,

338

de bakker and van der keur

Paulus will have his troops to fight for him, but that Varro will attack him with his own troops (Liv. 22.39.5 Varro dux tuis militibus te sit oppugnaturus).49 At Cannae, Rome fights Rome.50 And not for the last time.

4

Conclusion

When Churchill addressed the House of Commons on 16th June 1940, in one of his most famous speeches, his memorable words ‘Let us therefore brace ourselves to our duties, and so bear ourselves, that if the British Empire and its Commonwealth last for a thousand years, men will still say, ‘This was their finest hour’’ may have evoked for his audience the immortal speech of Shakespeare’s Henry V before the Battle of Agincourt,51 implicitly reminding them of another time in which victory was obtained in the face of tremendous odds. The techniques used by politicians to shape our perception of the present can also be employed by historiographers in their rendering of the past. Using a specific turn of phrase, inserting a few small but significant details, or adopting a different focus in recounting history compared to an illustrious predecessor can be all that the historiographer needs to present the past in a new way, to add a new layer of understanding to a well-known story. In such a subtle way, Livy prompts his audience to read his narrative of the Roman defeat at Cannae in the light of the Greek tradition on Thermopylae. At a few points, more or less explicit allusions can be found; as we have seen, the last words of Aemilius Paulus echo the famous epigram for the Spartiates, and the two Roman examples of self-sacrifice held up by Manlius Torquatus together recall the heroism of Leonidas and his three hundred. These points of contact call attention to the larger thematic correspondences between the Livian Cannae episode and the various accounts of Thermopylae (chief among which is the Herodotean one). Livy shows that Rome is not outdone by Greece (both in a moral and a literary sense), and that Roman virtue 49

50

51

Later in the speech, Fabius repeatedly aligns Varro with Hannibal, and reiterates that Paulus will have to fight both at once; cf. Liv. 22.39.18 idem enim tui quod hostium milites volent; idem Varro consul Romanus quod Hannibal Poenus imperator cupiet. duobus ducibus unus resistas oportet. A striking detail in Livy’s description of the African troops serves as a graphic reminder of this suggestion of Romans fighting Romans: decked out with the spoils from the earlier victories at the Trebia and Lake Trasimene, they looked like a battle line of Romans (Liv. 22.46.4 Afros Romanam crederes aciem). For the specific reference (Henry V, Act 4, Scene iii, 57–60) and the use of the play by Churchill in his war rhetoric, see Deats 2004: 97–99.

thermopylae and cannae

339

at the moment of defeat can compare with the spirit of the Spartans—even when pushed to extremes, such as in the rejection of the survivors of the battle or the sordid need to fight with the teeth when all other means of defence are exhausted. For Cannae to be presented as a Thermopylae, strategic and tactical manoeuvres do not need to be identical, nor does Livy stage Roman clones of Herodotus’ Spartans. We have argued instead that Cannae mirrors Thermopylae in its literary role. This also extends beyond the episode itself. Just as in Herodotus the aftermath of Thermopylae anticipates the eventual Greek victory, Livy’s narrative of defeat at Cannae carries the seeds of the future Roman triumph. The other undercurrent is less optimistic, and suggests that in Rome’s struggle against the Carthaginian invaders, its future internal strife is already foreshadowed, not unlike the way in which Herodotus subtly suggests a comparison between the uneasy Greek alliance during the Persian Wars and the later struggle for hegemony between Athens and Sparta. The emphasis on the valour of the Spartans and of consul Paulus in defeat, and the grim determination shown both in Greece and Rome in answer to the enemy’s success, mark Thermopylae and Cannae as moral highpoints before the degeneration into civil strife. Whereas Polybius interrupts his narrative after Cannae with a long discussion of the virtues of Roman society, Livy instead brings the class struggles between senate and people to the fore. Like Herodotus, the Roman historiographer lived in an age in which recent internecine warfare had cast its ominous shadow deep into the glorious past.

Bibliography Baragwanath, E., Motivation and Narrative in Herodotus (Oxford 2008). Blom, H. van der, Cicero’s Role Models: The Political Strategy of a Newcomer (Oxford 2010). Burck, E., Die Erzählungskunst des T. Livius (Berlin 1934). Canfora, L., ‘Thucydides in Rome and Late Antiquity’, in A. Rengakos & A. Tsakmakis (eds.), Brill’s Companion to Thucydides (Leiden 2006) 721–753. Cobet, J., ‘Wann wurde Herodots Darstellung der Perserkriege publiziert?’, Hermes 105.1 (1977) 2–27. Deats, S.M., ‘Henry V at War: Christian King or Model Machiavel’, in S.M. Deats, L.T. Lenker & M.G. Perry (eds.), War and Words: Horror and Heroism in the Literature of Warfare (Lanham 2004) 83–101. Dillery, J., ‘Roman Historians and the Greeks: Audiences and Models’, in A. Feldherr (ed.), The Cambridge Companion to the Roman Historians (Cambridge 2009) 77–107.

340

de bakker and van der keur

Flower, M.A. & Marincola, J. (eds.), Herodotus: Histories Book IX (Cambridge 2002). Fornara, C.W., ‘Evidence for the Date of Herodotus’ Publication’, JHS 91 (1971) 25– 34. Fornara, C.W., ‘Herodotus’ Knowledge of the Archidamian War’, Hermes 109.2 (1981) 149–156. Grethlein, J., ‘The Manifold Uses of the Epic Past: The Embassy Scene in Herodotus 7.153–63’, AJPh 127.4 (2006) 485–509. Grethlein, J., Experience and Teleology in Ancient Historiography. ‘Futures Past’ from Herodotus to Augustine (Cambridge 2013). Harrison, T., ‘The Persian Invasions’, in E.J. Bakker, I.J.F. de Jong & H. van Wees (eds.), Brill’s Companion to Herodotus (Leiden 2002) 551–578. Hinds, S., Allusion and Intertext: Dynamics of Appropriation in Roman Poetry (Cambridge 1998). Hoyos, D., ‘Maharbal’s Bon Mot: Authenticity and Survival’, CQ 50.2 (2000) 610–614. Jung, M., Marathon und Plataiai: Zwei Perserschlachten als “lieux de mémoire” im antiken Griechenland (Göttingen 2006). Krebs, C. ‘Leonides Laco quidem simile apud Thermopylas fecit. Cato and Herodotus’, BICS 49 (2006) 93–103. Lazenby, J.F., Hannibal’s war. A Military History of the Second Punic War (Warminster 1978). Lazenby, J.F., ‘Was Maharbal right?’, in T. Cornell, B. Rankov & P. Sabin (eds.), The Second Punic War. A Reappraisal (London 1996) 39–48. Lendon, J.E., ‘Battle Description in the Ancient Historians, Part I: Structure, Array, and Fighting’, G&R 64.1 (2017a) 39–64. Lendon, J.E., ‘Battle Description in the Ancient Historians, Part II: Speeches, Results, and Sea Battles’, G&R 64.2 (2017b) 145–167. Macan, R.W., Herodotus: the Capitals: Seventh, Eighth and Ninth Books, 3 vols. (London 1908). Marincola, J., Authority and Tradition in Ancient Historiography (Cambridge 1997). Meulder, M., ‘Hérodote et la prise de la ville latine de Gabies’, LEC 73 (2005) 109–156. Momigliano, A., The Classical Foundations of Modern Historiography (Berkeley 1990). O’Gorman, E., ‘Intertextuality and Historiography’, in A. Feldherr (ed.), The Cambridge Companion to the Roman Historians (Cambridge 2009) 231–242. Rodgers, B.S., ‘Great Expeditions: Livy on Thucydides’, TAPhA 116 (1986) 335–352. Rossi, A., Contexts of War. Manipulation of Genre in Virgilian Battle Narrative (Ann Arbor 2004). Ruffing, K., ‘300’, in B. Dunsch & K. Ruffing (eds.), Herodots Quellen—Die Quellen Herodots (Wiesbaden 2013) 201–221. Samotta, I., ‘Herodotus and Thucydides in Roman Republican Historiography’, in E. Foster & D. Lateiner (eds.), Thucydides and Herodotus (Oxford 2012) 345–378.

thermopylae and cannae

341

Schubert, W., ‘Herodot, Livius und die Gestalt des Collatinus in der LucretiaGeschichte’, RhM 134 (1991) 80–96. Schulte, A., De ratione quae intercedit inter Polybium et tabulas publicas, dissertation (Halle 1909). Shean, J.F., ‘Hannibal’s Mules: The Logistical Limitations of Hannibal’s Army and the Battle of Cannae, 216B.C.’, Historia 45.2 (1996) 159–187. Vallet, G. (ed.), T. Livi Ab urbe condita: Liber XXII (Paris 1966). Walsh, P.G., Livy. His Historical Aims and Methods (Cambridge 1961). Wiseman, T.P., Clio’s Cosmetics. Three Studies in Greco-Roman Literature (Leicester 1979). Woodman, A.J., Rhetoric in Classical Historiography (London 1988). Zimmermann, K., ‘Scipios Eid nach Cannae—eine propagandistische ‘Retourkutsche’?’, Chiron 27 (1997) 471–482.

chapter 14

The Great and the Small: Thermopylae and Sphacteria Adriaan Rademaker

Thucydides’ narrative of the battle on Sphacteria (Thuc. 4.26–4.41) consistently frames the important Athenian victory over the Spartan hoplites as an unexpected success that was greatly aided by the fact that the Athenian general Demosthenes took full advantage of the exceptional circumstances on the island.* Indeed, the narrator claims that the Athenian victory took the whole of the Greek world by surprise (Thuc. 4.40.1), and he further comments that the Spartan defeat was ‘something small’ compared to the legendary heroic defeat at Thermopylae (4.36.3). All this further detracts from the strategic and military achievement of—in particular—the Athenian politician and general Cleon. This paper explores the narrative techniques by means of which the narrator persuades his narratees to accept his interpretative frame; in particular, he consistently presents events from the point of view of their actual participants, in order to make his narratees see and feel what these actual participants saw and felt. In the early summer of 425 BC, an Athenian fleet led by Demosthenes is forced by a storm to land at Pylos in the South of the Peloponnese, and fortifies the peninsula.1 The Spartans react by sending a fleet and withdrawing their land army from Attica, in order to expel the Athenians from Pylos. The Spartans prepare to besiege Demosthenes and his men, but their preparations are interrupted by the arrival of Athenian ships from Zakynthos. The Athenians manage to isolate 420 hoplites (120 Spartiates among them) on the island of Sphacteria. The Spartans panic and start peace negotiations, but the Athenians reject these on the instigation of Cleon.2 * The author wishes to thank the organizers of the conference Textual Strategies in Greek and Latin War Narrative, Irene de Jong, Caroline Kroon, Suzanne Adema, Lidewij van Gils and Michel Buijs, as well as all participants for their contributions to the discussion. Needless to say, all errors and shortcomings remain mine. 1 On Thucydides’ technique of creating a narrative of casual acts having immense consequences, see Stahl 2006: 321–327. 2 For a detailed reconstruction of events during the opening of the campaign, see Strassler

© koninklijke brill nv, leiden, 2019 | doi:10.1163/9789004383340_015

the great and the small: thermopylae and sphacteria

343

Towards the end of the summer, Cleon is elected general; an Athenian army led by Cleon and Demosthenes beats the Spartans on Sphacteria; most of the Spartans are brought to Athens as hostages, and will be kept by the Athenians as prisoners of war until the peace of Nicias in 421, when they are finally released thanks to Nicias’ interventions.3 The Athenian victory at Sphacteria is among the greatest military successes of the Archidamian war, and indeed can be regarded as ‘one of the most decisive and valuable victories in Athenian history’.4 This implies that it is also one of the most painful defeats suffered by Sparta in the course of the Peloponnesian War.5 Surprisingly, Thucydides compares the Spartan defeat at Sphacteria to the heroic defeat of the Spartans at Thermopylae as described by Herodotus: καὶ οἱ Λακεδαιμόνιοι βαλλόμενοί τε ἀμφοτέρωθεν ἤδη καὶ γιγνόμενοι ἐν τῷ αὐτῷ ξυμπτώματι, ὡς μικρὸν μεγάλῳ εἰκάσαι, τῷ ἐν Θερμοπύλαις, ἐκεῖνοί τε γὰρ τῇ ἀτραπῷ περιελθόντων τῶν Περσῶν διεφθάρησαν, οὗτοί τε ἀμφίβολοι ἤδη ὄντες οὐκέτι ἀντεῖχον, ἀλλὰ πολλοῖς τε ὀλίγοι μαχόμενοι καὶ ἀσθενείᾳ σωμάτων διὰ τὴν σιτοδείαν ὑπεχώρουν. Thuc. 4.36.3

The Lacedaemonians now found themselves attacked from both sides. They found themselves in the same dilemma as at Thermopylae, to compare something small to something big. For the men at Thermopylae had perished when the Persians had found their way around them by the well-known path; on the present occasion, the Spartans were by now surrounded from both sides and did not hold up any longer. They were few in number and fighting many; weakened by the lack of food, they were forced to retreat.6

3 4 5 6

1990. Heitsch 1996: 24 has some good remarks about the opportunities that Athens missed by rejecting the peace offer. For a discussion of Cleon’s role in the capture of Sphacteria, see Woodhead 1960: 310–315 and Spence 1995: 420–425. Westlake 1968: 111. Thucydides explicitly compares the Spartan defeat at Pylos to that of the Athenians at Sicily (Th. 7.71.7). Cf. on this point Rengakos 2006: 299. All quotations from Thucydides follow the Oxford Classical Texts edition by H.S. Jones and J.E. Powell, (Oxford: Clarendon Press, 1942, revised edition, reprinted 1980). All translations are my own.

344

rademaker

Apparently, the comparison rests on the similarity between the conditions in which the Spartan hoplites found themselves: both the Three Hundred at Thermopylae and the 420 hoplites on Sphacteria found themselves surrounded by enemies and attacked from two sides; on Sphacteria, it is a small body of archers and light-armed soldiers that works its way around the Spartan stronghold and attacks from the rear. And in both cases, there is a ‘traitor’ who plays a decisive role in the defeat: at Thermopylae, it is Ephialtes who makes Xerxes aware of the path through the mountain (Hdt. 7.213.1); on Sphacteria, it is an unnamed Messenian general (Thuc. 4.36.1) who volunteers to attack the Spartans from the rear. But the comparison of Sphacteria with the legendary and heroic defeat at Thermopylae raises further questions. At Thermopylae, the Spartans suffered a heroic defeat against a numerically immensely superior Persian army, and the battle is perhaps the single most heroic moment in Spartan military history. By contrast, Sphacteria is also an important defeat, but seems a great deal less heroic at first sight, in that the Spartans surrendered themselves to Athens instead of fighting themselves to the death.7 In fact, the narrator himself claims that he is comparing ‘something small to something big’ (Thuc. 4.36.3 μικρὸν μεγάλῳ εἰκάσαι). This raises the question how the comparison with Thermopylae reflects on the dimensions and importance of the battle on Sphacteria, and how this comparison influences the narrator’s evaluation of the battle in his narrative. The phrase μικρὸν μεγάλῳ εἰκάσαι in fact provides an important clue to the persuasive strategy of the narrator throughout the Sphacteria episode: indeed, the narrator consistently stresses the significance of the chance circumstances on the island and the fact that Demosthenes took full advantage of these circumstances. By contrast, he almost completely downplays Cleon’s contribution to this victory. In his evaluation of the outcome of the battle, he famously calls Cleon’s promise to bring the Spartan hoplites to Athens within 20 days ‘insane’, μανιώδης (Thuc. 4.39.3). And he supports this surprisingly outspoken claim by showing that the Athenian expedition was an ill-planned undertaking with a fully unexpected outcome from every possible point of view. In his narrative of the Athenian assembly preceding Cleon’s expedition, he frames the whole enterprise as the result of an astonishingly irrational decision of the Athenian

7 On the unexpected surrender of the Spartans, cf. Von Fritz 1967: 696 and Rood 1998: 37: ‘The comparison suggests not the heroism of the defeated, but a revealing contrast: the Three Hundred at Thermopylae resisted to the death; the men on Sphakteria surrendered. And this contrast explains why the Greeks found the surrender surprising.’

the great and the small: thermopylae and sphacteria

345

assembly,8 driven by fear and anger: the Athenians fear a prolonged siege that will become hard to maintain during the winter (Thuc. 4.27.1), Cleon suggests that he could fix things on Sphacteria from fear of being exposed as a liar (Thuc. 4.27.3), and Nicias is irritated by Cleon’s taunts against his manliness (Thuc. 4.27.5). In his narrative of the battle itself, he presents the advantageous conditions in which the Athenians fought as greatly aided by the chance occurrence of a forest fire (Thuc. 4.29.2–4.29.3). Besides, he ascribes the success of the Athenians’ strategy to Demosthenes (Thuc. 4.29.2, 4.32.3, 4.32.4) and the Messenian general (4.36.1) rather than to Cleon. Finally, when evaluating the battle, he calls the Greeks at large to witness and signals that the victory came as an immense surprise to the whole of the Greek world. In what follows, I will discuss how the narrator manages to frame the Athenian victory as an unexpected success, and how he makes his narratees view the events through the eyes of its main participants9 as well as those of other Greeks from all over the Greek world, suggesting that the victory was a complete surprise from every possible point of view. Twice, the narrator comes in with evaluative remarks of his own. This occurs in 4.28.5, where he remarks that sensible Athenians welcomed the prospect of getting rid of Cleon, and in 4.39.3, where he states that Cleon’s promise was fulfilled, even though it was mad (μανιώδης).10 But more often, the narrator makes sure that he does not present his narratees with his own vision of events, but with that of its actual eyewitnesses. I will focus specifically on how the narrator presents events through the eyes of a number of secondary focalizers: first the Athenian citizens and politicians in the assembly, then the Athenian and Spartan soldiers on Sphacteria, and finally the Greeks at large. In narratological terms, the narrator makes extensive use of so-called character-bound focalization (or embedded focalization) in order to make his narratees share the point of view of those who actually observed the narrated events, and to induce the narratee to accept the judgement of these observers.11 By presenting events through the eyes of

8

9

10

11

On framing, seminal texts are Bateson 1972 and Goffman 1974. Johnson-Cartee 2005: 16 has good remarks on the function of framing in journalist narrative, remarks that by and large also seem to apply to historiography. One may also compare the contribution by Rutger Allan in the present volume. Allan discusses character-bound focalization as a device that helps to bring about ‘immersion’ of the narratee in the story world. Such remarks typically occur in so-called ‘Evaluation’ sections in narrative, for which see the Introduction to this volume. On the narrative structure of narrative episodes in Thucydides, see Allan 2007: 110–118, esp. 115. On embedded focalization, see De Jong 2014: 50–56 and Bal [1985] 2009: 149–153. On the use of embedded focalization in ancient historiography, cf. the general introduction, sec-

346

rademaker

these witnesses and participants, and by making his narratees share their surprise, the narrator suggests that the Athenian victory at Sphacteria was in fact completely unexpected and greatly aided by chance. In this context, the comparison of Sphacteria to Sparta’s most heroic defeat at Thermopylae serves to suggest that Sphacteria was really comparatively insignificant. This further detracts from the importance of Cleon’s victory.

1

The Athenian Assembly: Fear and Pride

Let us first consider the Athenian assembly (Thuc. 4.27–4.28). The participants in this assembly are consistently portrayed as driven by emotion and self-interest. First of all, the narrator presents the Athenian demos as driven by fear12 of the Spartans, and by second thoughts about their rejection of the peace negotiations: ἐν δὲ ταῖς Ἀθήναις πυνθανόμενοι περὶ τῆς στρατιᾶς ὅτι ταλαιπωρεῖται καὶ σῖτος τοῖς ἐν τῇ νήσῳ ὅτι ἐσπλεῖ, ἠπόρουν καὶ ἐδεδοίκεσαν μὴ σφῶν χειμὼν τὴν φυλακὴν ἐπιλάβοι […] πάντων τε ἐφοβοῦντο μάλιστα τοὺς Λακεδαιμονίους, ὅτι ἔχοντάς τι ἰσχυρὸν αὐτοὺς ἐνόμιζον οὐκέτι σφίσιν ἐπικηρυκεύεσθαι· καὶ μετεμέλοντο τὰς σπονδὰς οὐ δεξάμενοι. Thuc. 4.27.1–4.27.2

In Athens, people were informed about the army that it was in trouble, and that corn was sailed in to those on the island; accordingly, they did not know what to do and were afraid that the winter might overtake their blockade. […] Most of all, they feared the Spartans; because they thought that these had something of a firm position and would not send envoys to them anymore; and they regretted that they had failed to accept the treaty. In response to reports about the difficult situation of the army at Pylos, the Athenian citizens are at a loss (ἠπόρουν) and are afraid (ἐδεδοίκεσαν … ἐφοβοῦντο): they show an impressive example of the mindlessness of a body of democratic citizens, a ‘classic example of the ἀφροσύνη of democracy’.13

12 13

tion 1.2.2. Bakker 1997: 48–51 has good remarks on the use of character-bound focalization in Thucydides. For fear as a dominant driving force in Thucydides, see Desmond 2006: esp. 361–364. Cf. Woodhead 1960: 314.

the great and the small: thermopylae and sphacteria

347

Here, the narrator presents the situation at Pylos through the eyes of the Athenian commons, and pretends to have full insight into their fears and motives. He similarly pretends to know about the political motives of their leader, Cleon.14 And Cleon fares no better. In the present assembly, he is presented as driven by utterly selfish motives. At first he denies the sincerity of the messengers who reported that the Athenians on Pylos needed help, evidently from fear of being held responsible for the army’s plight because he had earlier rejected the peace negotiations. Then, when chosen as an examiner to inspect the situation at Pylos, he chooses the fuite en avant and suggests that the Athenians should send a military expedition instead: Κλέων δὲ γνοὺς αὐτῶν τὴν ἐς αὑτὸν ὑποψίαν περὶ τῆς κωλύμης τῆς ξυμβάσεως οὐ τἀληθῆ ἔφη λέγειν τοὺς ἐξαγγέλλοντας. […] [4] καὶ γνοὺς ὅτι ἀναγκασθήσεται ἢ ταὐτὰ λέγειν οἷς διέβαλλεν ἢ τἀναντία εἰπὼν ψευδὴς φανήσεσθαι, παρῄνει τοῖς Ἀθηναίοις, ὁρῶν αὐτοὺς καὶ ὡρμημένους τι τὸ πλέον τῇ γνώμῃ στρατεύειν, ὡς χρὴ […] πλεῖν ἐπὶ τοὺς ἄνδρας. Thuc. 4.27.3–4.27.4

Cleon noticed that the Athenians bore ill will against himself for having opposed the treaty. He claimed that the informants did not speak the truth. […] And when he realized that he would be forced to say the same things as the informants that he slandered, or to say the opposite and be exposed as a liar, he advised the Athenians, when he saw that they were somewhat more eager to send an expedition, that they should sail out against those men. The consistently negative portrayal of Cleon has led to a debate in the scholarly literature between those who claim that Thucydides willfully misrepresents Cleon’s merits, such as Woodhead (1960) and Westlake (1968), and those who, like Erbse (1989), are more inclined to accept this portrayal of Cleon as an idle boaster.15 Whatever the historical circumstances may have been, it seems clear 14 15

On Thucydides’ manner of inferring his characters’ motives from their actual behaviour, see Schneider 1974; and cf. Rengakos 2006: 282–283. Westlake 1968: 72–73 draws attention to the extent to which Thucydides claims to be able to read the mind of Cleon: ‘The attribution of undisclosed and almost wholly discreditable motives and feelings to Cleon is very remarkable. It is almost more remarkable that Thucydides tacitly claims to see into the mind of Cleon and to know precisely why he acted as he did at each stage of the debate. […] There can hardly be any doubt that he has inferred the quite considerable catalogue of motives and feelings included in the passage from what he knew, or claimed to know about the character of Cleon.’ Tsakmakis 2006:

348

rademaker

that Thucydides’ narrator consistently presents Cleon as driven by the desire for saving face.16 He does not at all seem to consider the situation of the Athenian army at Sphacteria; he simply does not want to be blamed for the whole affair. He has a keen grasp of the political implications of the situation (27.3: γνούς, 27.4: γνούς, ὁρῶν), but he fails to consider other interests than his own. Even Nicias does not come off much better in the present assembly. Cleon taunts him for his lack of manliness; in response, he suggests that he is ready to surrender his post to Cleon and seems to underestimate the Athenian’s willingness to appoint Cleon for the task:17 ὁ δὲ Νικίας τῶν τε Ἀθηναίων τι ὑποθορυβησάντων ἐς τὸν Κλέωνα, ὅτι οὐ καὶ νῦν πλεῖ, εἰ ῥᾴδιόν γε αὐτῷ φαίνεται, καὶ ἅμα ὁρῶν αὐτὸν ἐπιτιμῶντα, ἐκέλευεν ἥντινα βούλεται δύναμιν λαβόντα τὸ ἐπὶ σφᾶς εἶναι ἐπιχειρεῖν. Thuc. 4.28.1

Nicias noticed that the Athenians were muttering at Cleon that he should sail all the same [i.e. in spite of not being a general], if he thought it was going to be easy. At the same time he saw that Cleon was insulting him. So he encouraged Cleon to take whatever force he needed and give it a try, as far as the generals were concerned. Finally, the Athenian elite, the σώφρονες,18 are said to assent to the expedition in the hope, not so much of obtaining a military success, but of getting rid of Cleon:19

16 17 18

19

166 draws attention to the people’s laughter at Cleon and argues that ‘the people’s derision of Cleon seems justified here.’ Cf. also Hornblower 1987: 78. By contrast, Erbse 1989: 155– 156 claims that Thucydides is essentially reliable in his judgment on Cleon: ‘Gerade die in den Kapiteln 4,27–29 vorliegende Darstellung gibt in ihrer Eigenart […] das Verhalten Kleons, seine Prahlerei, seine Kühnheit und sein Glück in sachgerechter Weise wieder.’ Cf. Tsakmakis 2006: 170: ‘In the debate on Pylos, [Cleon’s] image as a public figure is ruined, since it becomes obvious that he desperately tried to conceal his inadequacies.’ On Nicias’ inability to persuade mass audiences, see Ober 2006: 152. While it may be true that τοῖς σώφροσι means ‘the sensible men’ (not ‘the conservative party’: Gomme 1956: 470), the phrase τοῖς σώφροσι τῶν ἀνθρώπων still suggests that Thucydides has a more or less well-defined subset of the Athenians in mind, and this subset evidently consists of citizens who disapprove of Cleon’s populist politics. See Rademaker 2005: 216–217. Tsakmakis 2006: 170 stresses that the troops that were sent with Cleon included allies only; no Athenian citizens were sent along (Th. 4.28.4).

the great and the small: thermopylae and sphacteria

349

τοῖς δὲ Ἀθηναίοις ἐνέπεσε μέν τι καὶ γέλωτος τῇ κουφολογίᾳ αὐτοῦ, ἀσμένοις δ’ ὅμως ἐγίγνετο τοῖς σώφροσι τῶν ἀνθρώπων, λογιζομένοις δυοῖν ἀγαθοῖν τοῦ ἑτέρου τεύξεσθαι, ἢ Κλέωνος ἀπαλλαγήσεσθαι, ὃ μᾶλλον ἤλπιζον, ἢ σφαλεῖσι γνώμης Λακεδαιμονίους σφίσι χειρώσεσθαι. Thuc. 4.28.5

The Athenians were somewhat overcome by laughter at his boasts. But still, these were welcome to the more sensible persons, because they reckoned that they would get either of two goods. They would either get rid of Cleon, what they hoped for most, or if their expectations were going to be thwarted, the Lacedaemonians would be subjected to themselves. Thus, the debate in the Athenian assembly is presented as an astonishingly irrational piece of decision-making.20 All actors in the narrative allow themselves to be driven by emotions and selfish motives. The Athenians fear the Spartans, Cleon tries to save face, Nicias is irritated by Cleon’s insult, and the Athenian elite fosters the hope of getting rid of Cleon.21 All this serves to suggest that the decision to send Cleon with a fleet to Sphacteria was not a well-planned decision, and that its eventual success was mainly due to sheer good luck.

2

The Situation on Sphacteria: after the Forest Fire

Good luck also plays a considerable role on the island itself.22 According to the narrator, one of Demosthenes’ soldiers had accidentally set fire to the wood on the island, and this helped greatly to make an invasion less dangerous: τῶν δὲ στρατιωτῶν ἀναγκασθέντων διὰ τὴν στενοχωρίαν τῆς νήσου τοῖς ἐσχάτοις προσίσχοντας ἀριστοποιεῖσθαι διὰ προφυλακῆς καὶ ἐμπρήσαντός τινος

20 21

22

On the Pylos debate as an example of the weaknesses of democratic decision-making, see Raaflaub 2006: 201–202. Woodhead 1960: 313–314, while surprisingly positive about Cleon, strongly draws out the reprehensible traits in the conduct of Nicias and the σώφρονες: ‘Yet the narrative, when one thinks it over, shows that the comments of the σώφρονες are unimaginably foolish, and that Nicias’ conduct is highly reprehensible. The latter is prepared to hand over an extremely important expedition in its entirety to the command of a man whom he believes to be incompetent and whom he expects and no doubt hopes will fail. The σώφρονες apparently viewed with pleasure the prospect of a disaster in which, though Cleon fell, many another brave Athenian might die beside him.’ On the role of τύχη in the Sphacteria episode, see Hunter 1973: 70–74.

350

rademaker

κατὰ μικρὸν τῆς ὕλης ἄκοντος καὶ ἀπὸ τούτου πνεύματος ἐπιγενομένου τὸ πολὺ αὐτῆς ἔλαθε κατακαυθέν. Thuc. 4.30.2

Demosthenes’ men had been forced by the lack of space to land on the extreme point of the island and take their meal there under the protection of guards. When someone by accident set fire to a small part of the wood, and a wind came up from that direction, the larger part of the wood was accidentally consumed by fire. Given that Demosthenes himself had suffered a defeat at the hands of the Aetolians as the result of a forest fire (3.98.2), there has been some speculation whether or not the present fire can really have been accidental.23 In any case, it seems evident that it is advantageous for the narrator to present the fire as accidental, since the unusual circumstances that the fire creates help to explain the unexpected Athenian success.24 The consequences of the fire also make themselves felt on the battlefield: γνόντες αὐτοὺς οἱ ψιλοὶ βραδυτέρους ἤδη ὄντας τῷ ἀμύνασθαι, καὶ αὐτοὶ τῇ τε ὄψει τοῦ θαρσεῖν τὸ πλεῖστον εἰληφότες πολλαπλάσιοι φαινόμενοι καὶ ξυνειθισμένοι μᾶλλον μηκέτι δεινοὺς αὐτοὺς ὁμοίως σφίσι φαίνεσθαι, ὅτι οὐκ εὐθὺς ἄξια τῆς προσδοκίας ἐπεπόνθεσαν, ὥσπερ ὅτε πρῶτον ἀπέβαινον τῇ γνώμῃ δεδουλωμένοι ὡς ἐπὶ Λακεδαιμονίους, καταφρονήσαντες καὶ ἐμβοήσαντες ἁθρόοι ὥρμησαν ἐπ’ αὐτοὺς καὶ ἔβαλλον λίθοις τε καὶ τοξεύμασι καὶ ἀκοντίοις, ὡς ἕκαστός τι πρόχειρον εἶχεν. [2] γενομένης δὲ τῆς βοῆς ἅμα τῇ ἐπιδρομῇ ἔκπληξίς τε ἐνέπεσεν ἀνθρώποις ἀήθεσι τοιαύτης μάχης καὶ ὁ κονιορτὸς τῆς ὕλης νεωστὶ κεκαυμένης ἐχώρει πολὺς ἄνω, ἄπορόν τε ἦν ἰδεῖν τὸ πρὸ αὑτοῦ ὑπὸ τῶν τοξευμάτων καὶ λίθων ἀπὸ πολλῶν ἀνθρώπων μετὰ τοῦ κονιορτοῦ ἅμα φερομένων. Thuc. 4.34.1–4.34.2

The light troops became aware that the Spartans were now more hesitant in their defense, and they themselves had taken far more courage because of what they saw, because it turned out that they were by far in the majority, and they grew more accustomed to the circumstances. The 23

24

See Wilson 1979: 103: ‘More surprising is the element of chance: why did not Demosthenes start a fire deliberately? Thucydides’ informant here is plainly either Demosthenes himself or someone close to him: there is virtually no chance of the facts being mistaken.’ Wylie 1993: 22–24 argues that the campaign was in fact ‘brilliantly conceived and carefully planned’ (ibid. 23).

the great and the small: thermopylae and sphacteria

351

Spartans did not seem as terrifying to them as before, because they had not immediately suffered as much as they had expected, as when they first went ashore in dejected spirits, because they were attacking Spartans. They now looked down upon them; loudly screaming, they collectively ran towards the Spartans and pelted them with stones, arrows, and spears, whatever anybody had at hand. When the shouting took place at the same time as the attack, panic befell the Spartan men, who were unaccustomed to this type of fighting. A large cloud of dust went up from the wood that had been recently burnt, and it was difficult to see what each man had in front of him because of the arrows and stones that were thrown by many people and flew through the air together with the ashes. The passage offers a lively insight into the minds of the soldiers on both sides. The narrator pretends to have full access to the thoughts and feelings of the soldiers in both armies;25 and this claim seems credible to the extent that Thucydides himself was of course an experienced general. The Athenians are at first ‘enslaved in their minds’ (Thuc. 4.34.1 τῇ γνώμῃ δεδουλωμένοι) at the prospect of having to fight with the Spartans. But when they see (ibid. γνόντες) that these fight less vigorously than usual, they become far more courageous (ibid. τοῦ θαρσεῖν τὸ πλεῖστον εἰληφότες). Conversely, the Spartans are struck with panic (34.2 ἔκπληξις … ἐνέπεσεν) because they are not accustomed to the exceptional conditions on the battlefield, conditions that are utterly unfit for hoplite battle. Thus, the forest fire and its consequences are said to have an enormous impact on the outcome of the battle. The Athenians see more than they would normally have seen, and take courage when they realize that the Spartans are not doing as well as they normally should; conversely, the Spartans see less than they normally would because of the dust, and are prevented from the kind of fighting at which they normally excel.26 This implicitly helps to explain why the Spartans lost a battle that they should have won under normal conditions, or so the narrator suggests.

25 26

On the ‘omniscience’ of the primary narrator and his access to the inner thoughts of his characters, see De Jong 2014: 56. Ober 2006: 149 stresses the contrast between the Athenians’ ability to adapt to the circumstances and the Spartans’ futile attempt to ‘fight a standard hoplite battle under impossible conditions.’

352 3

rademaker

Battlefield Tactics

However, even the exceptional conditions on the battlefield do not fully explain the Athenian victory. Battlefield tactics play a decisive role as well. And when it comes to tactics, the narrator quite consistently suppresses any contribution that Cleon may have made, giving all credit to Demosthenes and the Messenian general instead. It is Demosthenes who makes the first tactical moves that give the Athenians an advantage over their Spartan adversaries: Δημοσθένους δὲ τάξαντος διέστησαν κατὰ διακοσίους τε καὶ πλείους, ἔστι δ’ ᾗ ἐλάσσους, τῶν χωρίων τὰ μετεωρότατα λαβόντες, ὅπως ὅτι πλείστη ἀπορία ᾖ τοῖς πολεμίοις πανταχόθεν κεκυκλωμένοις. Thuc. 4.32.3

Demosthenes ordered them to divide themselves in groups of two hundred men or more, and in some spots less. They occupied the highest spots on the island, in order to create as much difficulty for the enemy as possible, as these would find themselves surrounded from everywhere. Even Westlake (1968), who is quite critical of Thucydides’ treatment of Cleon, seems to accept the narrator’s implicit suggestion that Cleon’s contribution to the military operation was minimal, and that the credits for the military success should go to Demosthenes rather than to him.27 However, it is left to an anonymous Messenian commander to make the decisive proposal to attack the Spartan army from the rear:

27

Cf. Westlake 1968: 70: ‘That Cleon rather than someone else was appointed to take command at Pylos does not appear to have decisively influenced the course of the military operations here, which were planned and conducted largely by Demosthenes.’ But he also notes that Thucydides praises Demosthenes only implicitly. ibid. 111: ‘Nowhere does Thucydides assess, or even refer to, the contribution of Demosthenes to one of the most decisive and valuable victories in Athenian history.’ Similarly, Marshall 1984: 20–21 concludes: ‘Demosthenes gets less credit than seems his due for the whole Pylos operation.’ See also Woodcock 1928: esp. 101–102. Another of Cleon’s enemies supports the idea that Demosthenes deserves the credit for the military success. In Aristophanes’ Knights 54– 57, Demosthenes is made to explain how ‘recently, when I had kneaded a Spartan cake at Pylos, the rogue somehow came from my back, snatched away the cake I had kneaded and served it up himself.’ (καὶ πρώην γ’ ἐμοῦ | μᾶζαν μεμαχότος ἐν Πύλῳ Λακωνικήν, | πανουργότατά πως περιδραμὼν ὑφαρπάσας | αὐτὸς παρέθηκε τὴν ὑπ’ ἐμοῦ μεμαγμένην).

the great and the small: thermopylae and sphacteria

353

ἐπειδὴ δὲ ἀπέραντον ἦν, προσελθὼν ὁ τῶν Μεσσηνίων στρατηγὸς Κλέωνι καὶ Δημοσθένει ἄλλως ἔφη πονεῖν σφᾶς: εἰ δὲ βούλονται ἑαυτῷ δοῦναι τῶν τοξοτῶν μέρος τι καὶ τῶν ψιλῶν περιιέναι κατὰ νώτου αὐτοῖς ὁδῷ ᾗ ἂν αὐτὸς εὕρῃ, δοκεῖν βιάσεσθαι τὴν ἔφοδον. Thuc. 4.36.1

As the battle drew on endlessly, the Messenian general approached Cleon and Demosthenes and told them that they struggled in vain. If they were prepared to give him part of the archers and the light troops to go around to the back of the enemy, by a road that he would be able to find himself, he thought he would be able to force an approach. The Spartans now find themselves in a situation similar to that at Thermopylae. Not only do they have to fight under bad circumstances; they also have to face the enemy on two sides, because of the ‘betrayal’ of the Messenian general who acts as a latter day Ephialtes. Thus, the Athenian victory at Sphacteria is presented as a military success that came about against all odds. The campaign was the result of utterly irrational decision making in the Athenian assembly, but it turned out to be successful thanks to the unusual circumstances on the island and, more decisively, thanks to the help of a smart commander among the allies.

4

The Reception of the News of the Spartan Defeat

How then, was the news of this victory received? In wrapping up his narrative, the narrator comes in with his famous authorial comment, and makes the claim that Cleon’s promise was fulfilled in spite of its sheer madness: οἱ μὲν δὴ Ἀθηναῖοι καὶ οἱ Πελοποννήσιοι ἀνεχώρησαν τῷ στρατῷ ἐκ τῆς Πύλου ἑκάτεροι ἐπ’ οἴκου, καὶ τοῦ Κλέωνος καίπερ μανιώδης οὖσα ἡ ὑπόσχεσις ἀπέβη: ἐντὸς γὰρ εἴκοσιν ἡμερῶν ἤγαγε τοὺς ἄνδρας, ὥσπερ ὑπέστη. Thuc. 4.39.3

The Athenians and the Peloponnesians withdrew their armies from Pylos and both returned home, and Cleon’s promise, insane as it may have been, turned out to be fulfilled: for he brought the men to Athens within 20 days, as he had promised. In exposing the ‘sheer madness’ of Cleon’s promise, the narrator drops his usual mask of impartiality and comes in with an evaluative remark of his own in

354

rademaker

order to minimize Cleon’s military merits.28 As the use of the particle καίπερ shows, the narrator claims that in spite of Cleon’s eventual victory, the fact remains that in the earlier assembly, victory was a crazy thing to promise.29 Thus he implies that Cleon’s success was not based on solid military expertise, but that he made a boastful promise that happened to be fulfilled by chance. And it seems that it is on account of this unexpected success that Cleon gets carried away: he counts on further similar successes and continues to oppose all peace initiatives.30 To back up his claim that Cleon’s promise was insane, the narrator calls the whole of Greece to witness, claiming that all the Greeks viewed the outcome of this battle as the biggest surprise in the entire Peloponnesian war:31 παρὰ γνώμην τε δὴ μάλιστα τῶν κατὰ τὸν πόλεμον τοῦτο τοῖς Ἕλλησιν ἐγένετο· τοὺς γὰρ Λακεδαιμονίους οὔτε λιμῷ οὔτ’ ἀνάγκῃ οὐδεμιᾷ ἠξίουν τὰ ὅπλα παραδοῦναι, ἀλλὰ ἔχοντας καὶ μαχομένους ὡς ἐδύναντο ἀποθνῄσκειν. Thuc. 4.40.1

Among the things that happened in the war, this was naturally the greatest surprise to the Greeks. For they had thought that the Spartans would not let themselves be forced by hunger or anything else to hand over their weapons, but that they would hold on to them and fight as best they could until they died.

28

29 30

31

While I think there is no denying Thucydides’ bias against Cleon (which may or may not be due to Cleon’s putative role in Thucydides’ banishment, cf. Rhodes 2006: 523), it is impossible to judge to what extent Thucydides’ judgement was justified. Many commentators have been severely critical of Thucydides. See for instance Woodhead 1960: 314: ‘Thucydides cannot avoid reporting that the ὑπόσχεσις ἀπέβη, but he can and does turn it sour by adding καίπερ μανιώδης.’ Cf. Westlake 1968: 75: ‘Throughout the latter part of his narrative of the Pylos episode all other considerations are subordinated to his desire to expose the unworthiness of Cleon.’ Spence 1995: 424 argues that there were sound military reasons to call Cleon’s promise insane, pointing to the extremely short deadline of twenty days, and the reputation of the military power of Spartan hoplite armies. Cf. Gomme 1956: 478: ‘It is Kleon’s promise that is called mad […] and this was mad.’ For καίπερ in participle phrases with concessive value, see Denniston [1934] 1954: 464–467 and Bakker 1988: 120–124 and 149 n. 67. Cf. Tsakmakis 2006: 177: ‘The fortuitous success in Pylos makes Cleon believe that the same strategy will guarantee further military triumphs, and therefore he opposes every peace initiative; in so doing he behaves like a gambler who is carried away by his luck and keeps on playing until he has lost everything.’ For the Sphacteria episode as a narrative of paradox and reversal with an unexpected outcome, see Connor 1984: 108–118.

the great and the small: thermopylae and sphacteria

355

The narrator claims that all the Greeks regarded the Athenian victory with astonishment, and he assures his narratees that it is quite understandable that they should do so (δή in 4.40.1).32 The Greeks had thought that the Spartans would keep up the high heroic standards set by their forefathers and fight themselves to death.33 However, these Spartans did not live up to the standard of their ancestors at Thermopylae; instead, they surrendered. The un-heroic, but arguably sensible, Realpolitik of the Spartan hoplites sets the seal on an Athenian victory that was, according to the narrator, unexpected from any point of view.

5

Conclusion: Comparing Great and Small

In the narrative of the Athenian campaign at Sphacteria, the narrator consistently downplays the significance of Cleon’s contribution to the Athenian victory. He presents the Athenians’ greatest military success as a complete surprise from every conceivable point of view. He does so by suggesting that the whole enterprise was the result of irresponsibly irrational decision-making in the assembly and by pointing out that the circumstances on the island after the forest fire were greatly in Athens’ favour. By attributing the strategic decisions behind the victory to Demosthenes and the Messenian general, he further downplays the role of the Athenian commander Cleon in particular. To back up his interpretation of the episode, the narrator makes us share the point of view of just about the whole of the Greek world in order to make sure that we can share his, and its, surprise. In the narrative of the assembly at Athens, we are made to share the fear of the Athenian δῆμος (Thuc. 4.27.1– 4.27.2); we see how Cleon recognizes the Athenians’ anger at himself and we share his fear of being exposed and his desire to save face (Thuc. 4.27.3–4.27.4), we share Nicias’ irritation at Cleon’s insult (4.28.1) and we share the hopes of Athens’ ‘decent citizens’ to get rid of Cleon (4.28.5). Read as a whole, the passage seems to suggest that none of the participants in the Athenian assembly made a balanced assessment of the pros and cons of the campaign, and that 32

33

Cf. Sicking 1993: 52: ‘it is possible to describe δή as a primarily ‘evidential’ particle which presents a statement as immediately evident to the senses or the understanding or as common knowledge. It thus implies that speaker and hearer are in the same position with respect to this statement.’ Here, the narrator implies that his narratees should be able to understand why all the Greeks were astonished at the Spartan defeat. The narrator here presents us with the embedded focalization of ‘the Greeks’ in general, as one can see from his use of the phrase παρὰ γνώμην and especially the verb of thinking ἠξίουν (4.40.1). Cf. De Jong 2014: 50–51.

356

rademaker

no one really believed in a successful outcome of the expedition. Instead, they were struck with fear, trying to save face, hurt in their pride, or even secretly hoping to get rid of Cleon. In the battle scene itself, we are similarly made to share the point of view of the soldiers in the two armies. We experience how the light troops of the Athenians unexpectedly gain confidence: at first, they are dejected by the prospect of having to fight the superior Spartans, but they take courage when these Spartans turn out to fight less vigorously than usual (4.34.1). We also share the panic of the Spartans, who were unaccustomed to the circumstances on the island (4.34.2). Like their fellow Greeks in general, the troops on Sphacteria seem to have anticipated a very different outcome of the battle, at least until the circumstances brought about a significant turn in Athens’ favor, and until the commander of the Messenians came up with the tactics that decisively put Sparta at a disadvantage. After the battle, the narrator assures us that Cleon’s promise to bring the Spartans to Athens within three weeks was ‘insane’ however successful it may have turned out to be (4.39.3). We are also made to share the point of view of the Greeks at large, and these regard the Athenian victory with complete and utter amazement (4.40.1). In this context, the comparison of Sphacteria with Thermopylae does more than just signal the similarity between two Spartan armies that were attacked from two sides. It suggests that the Spartan hoplites at Sphacteria did not live up to the high standards of courage traditionally associated with the Spartans, and exemplified most typically by the heroic defeat of the 300 under Leonidas.34 The narrative suggests that there were several factors that contributed to the Spartans’ poor performance. The fact that food had been in short supply, but most of all the conditions in which the fighting took place, and the ‘betrayal’ by the Messenian general, all these things helped to create exceptional conditions, in which the Spartans did far worse that they would normally have done. And this in turn serves to suggest that Cleon’s resounding victory was a stroke of good luck that was completely unexpected, or even totally undeserved.

34

Rood 1998: 37: ‘The flashback to Thermopylai […] relates to the disastrous position in which the Spartans found themselves: as in 480, so now, the only hope of keeping a position was destroyed when the enemy discovered a path leading to the rear. The comparison suggests not the heroism of the defeated, but a revealing contrast: the Three Hundred at Thermopylai resisted to the death; the men on Sphakteria surrendered. And this contrast explains why the Greeks found the surrender surprising.’

the great and the small: thermopylae and sphacteria

357

Bibliography Allan, R.J., ‘Sense and Sentence Complexity: Sentence Structure, Sentence Connection, and Tense-Aspect as Indicators of Narrative Mode in Thucydides’ Histories’, in R.J. Allan, & M. Buijs (eds.), The Language of Literature. Linguistic Approaches to Classical Texts (Leiden 2007) 93–121. Bakker, E.J., Linguistics and Formulas in Homer: Scalarity and the Description of the Particle per (Amsterdam 1988). 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. Bal, M., Narratology. An Introduction to the Theory of Narrative ([1985] Toronto 2009). Bateson, G., ‘A Theory of Game and Fantasy’, in G. Bateson (ed.), Steps to an Ecology of Mind (New York 1972) 177–193. Connor, W.R., Thucydides (Princeton N.J. 1984). Denniston, J.D., The Greek Particles ([1934] Oxford 1954). Desmond, W., ‘Lessons of Fear: A Reading of Thucydides’, CPh 101 (2006) 359–379. Erbse, H., Thukydides-Interpretationen (Berlin 1989) 131–177. Fritz, K. von, Die griechische Geschichtsschreibung I (Berlin 1967) 694–701. Goffman, E., Frame Analysis: An Essay on the Organization of Experience (London 1974). Gomme, A.W., A Historical Commentary on Thucydides (Oxford 1956). Heitsch, E., Geschichte und Situationen bei Thukydides (Stuttgart 1996). Hornblower, S., Thucydides (Baltimore 1987). Hunter, V.J.H., Thucydides the Artful Reporter (Toronto 1973). Johnson-Cartee, K., News Narrative and News Framing: Constructing political reality. (Lanham MD 2005). Jong, I.J.F. de, Narrators and Focalizers. The Presentation of the Story in the Iliad ([1987] London 2004). Jong, I.J.F. de, Narratology and Classics: A Practical Guide (Oxford 2014). Lallot, J., Rijksbaron, A., Jacquinod, & B., Buijs, M. (eds.), The Historical Present in Thucydides. Semantics and Narrative Function—Le présent historique chez Thucydide. Sémantique et fonction narrative (Leiden 2011). Marshall, M.H.B., ‘Cleon and Pericles: Sphacteria’, G&R 31 (1984) 19–36. Ober, J., ‘Thucydides and the Invention of Political Science’, in A. Rengakos & A. Tsakmakis (eds.), Brill’s Companion to Thucydides (Leiden 2006) 131–160. Raaflaub, K.A., ‘Thucydides on Democracy and Oligarchy’, in A. Rengakos & A. Tsakmakis (eds.), Brill’s Companion to Thucydides (Leiden 2006) 189–224. Rademaker, A., Sophrosyne and the Rhetoric of Self-Restraint: Polysemy & Persuasive Use of An Ancient Greek Value Term (Leiden 2005). Rengakos, A. & Tsakmakis, A. (eds.), Brill’s Companion to Thucydides (Leiden 2006).

358

rademaker

Rengakos, A., ‘Thucydides’ Narrative: The Epic and Herodotean Heritage’, in A. Rengakos & A. Tsakmakis (eds.), Brill’s Companion to Thucydides (Leiden 2006) 279–300. Rhodes, P.J., ‘Thucydides and Athenian History’, in A. Rengakos & A. Tsakmakis (eds.), Brill’s Companion to Thucydides (Leiden 2006) 523–546. Rood, T., Thucydides: Narrative and Explanation (Oxford 1998) 24–57. Schneider, C., Information und Absicht bei Thukydides. Untersuchung zur Motivation des Handelns (Göttingen 1974). Sicking, C.M.J. van, ‘Devices For Text Articulation in Lysias I and XII’, in C.M.J. Sicking, J.M. van Ophuijsen, Two Studies in Attic Particle Usage (Leiden 1993), 3–66. Spence, I.G., ‘Thucydides, Woodhead, and Kleon’, Mnemosyne 48 (1995) 411–437. Stahl, H.P., ‘Narrative Unity and Consistency of Thought: Composition of Event Sequences in Thucydides’, in A. Rengakos & A. Tsakmakis (eds.), Brill’s Companion to Thucydides (Leiden 2006) 301–334. Strasburger, H., ‘Thukydides und die politische Selbstdarstellung der Athener’, Hermes 86 (1968) 17–40. Strassler, R.B., ‘The Opening of the Pylos Campaign’, JHS 110 (1990) 110–125. Tsakmakis, A., ‘Leaders, Crowds and the Power of the Image: Political Communication in Thucydides’, in A. Rengakos & A. Tsakmakis (eds.), Brill’s Companion to Thucydides (Leiden 2006) 161–188. Wakker, G.C., ‘Modal Particles and Different Points of View in Herodotus and Thucydides’, in E.J. Bakker (ed.), Grammar as Interpretation. Greek Literature in its Linguistic Contexts (Leiden 1997) 215–250. Westlake, H.D., Individuals in Thucydides (Cambridge 1968). Wilson, J., Pylos 425 BC: A Historical and Topographical Study of Thucydides’ Account of the Campaign (Warminster 1979). Woodcock, E.C., ‘Demosthenes, Son of Alcisthenes’, HSPh 39 (1928) 93–108. Woodhead, A.G.W., ‘Thucydides’ Portrait of Cleon’, Mnemosyne 13 (1960) 289–317. Wylie, G., ‘Demosthenes the General: Protagonist in a Greek Tragedy?’, G&R 40 (1993) 20–30.

chapter 15

Force, Frequency and Focalisation: the Function of Similes in the Battle-Narrative of Vergil, Aeneid 10 Stephen Harrison

1

Introduction

In this paper I return to the analysis of the battle-narrative of Aeneid 10 a quarter-century after completing a commentary on that poetic book.1 Over that period, as the introduction to this volume shows, narratological ideas, then just emerging into the mainstream of the study of Latin poetry,2 have rightly become an established part of the literary landscape, alongside the interaction of linguistic and narratological approaches celebrated in this volume, and I am pleased to have the opportunity to go back to Vergil’s text with a somewhat more nuanced and theoretical perspective. My topic is the function of Homeric-style extended similes in the battle-narrative. I will examine this from three perspectives: (i) force, the effect of similes in characterisation, often through the evocation of Homeric models, (ii) frequency, their distribution and role in narrative variation and pace, and (iii) focalisation, the issue of whose views are presented of the images evoked in similes. Of these (i) owes something to the ideas on generic intertextuality and poetic memory propounded by Gian Biagio Conte,3 while (ii) and (iii) cover major concerns of mainstream narratological theory;4 (iii) also brings in the notion of reader-response.5 My emphasis in what follows will be on close textual analysis against this theoretical background; I will be particularly interested in the final section in the use of narrative theory to open up multiple possibilities of literary interpretation.

1 Harrison 1991 (completed early 1990). 2 In Harrison 1991 I did not allow enough weight to the work on focalisation in the Aeneid by Bonfanti 1985, and had not read Fowler 1990, which came out after I submitted my commentary to the publisher. Fowler 2000 reprints Fowler 1990 and several other classic narratological pieces. 3 See e.g. Conte 1986 and 2007. 4 On frequency, order and duration see e.g. Genette 1980: 33–160; Genette 1988: 21–40, on focalisation Genette 1980: 189–211; Bal 1985: 100–115; Genette 1988: 72–79. 5 For a convenient exposition see Tompkins 1980: ix–xxvi.

© koninklijke brill nv, leiden, 2019 | doi:10.1163/9789004383340_016

360 2

harrison

Overall Narrative Structure of Aeneid 10

Here I set out my segmentation of the action of Aeneid 10 into scenes as a reminder for the reader of the details of its plot.6 1–117 118–145 146–162 163–214 215–259 260–307 308–361

362–438

439–509

510–605

606–688

689–768 769–832

833–908

Divine council on Olympus Battle in Latium: the Trojans without Aeneas are hard-pressed Aeneas is seen returning to Latium at night with Etruscan allies Catalogue of Etruscans supporting Aeneas Aeneas meets his ships, now metamorphosed into nymphs, and prays to Cybele Aeneas lands amid Latin resistance—Tarchon is shipwrecked. The battle on the shore: the brief aristeia [personal slaying-episode] of Aeneas is countered by the Latin aristeiai of Clausus, Halaesus and Messapus, leading to an evenly poised situation in the battle. Pallas drives his Arcadians forward and has a major aristeia. Halaesus has a shorter aristeia in response but is then killed by Pallas. Fate forbids the meeting of Pallas and Lausus. Confrontation of Turnus and Pallas. The latter prays in vain to Hercules, who is reminded by Jupiter that Turnus too must die. Turnus slays Pallas and despoils his body. The poet comments on Turnus’ action and laments the death of Pallas. Aeneas, enraged by Pallas’ death, launches a furious aristeia, taking prisoners for future sacrifice at Pallas’ funeral and killing several suppliants. The Trojan camp is relieved. On Olympus, Jupiter allows Juno to remove Turnus temporarily from the battle; Juno lures Turnus on to a ship back to Ardea using a phantom image of Aeneas. Back on the battlefield, Mezentius distinguishes himself amongst the Latins with a successful aristeia. Aeneas meets Mezentius and wounds him; Mezentius’ son Lausus intervenes to save his father, and is killed by Aeneas. Aeneas expresses sympathy for Lausus’ pietas. Mezentius returns to battle in an attempt to avenge his son, and is killed by Aeneas.

6 For further material and references see Harrison 1991: xxvi–xxviii.

force, frequency and focalisation

3

361

Strategies of Variation

As I have underlined elsewhere,7 a prominent issue for the battle-narrative of the Aeneid is the need to maintain poetic and readerly interest. Despite general Roman enthusiasm for war in both life and literature, the Iliad’s long and often repetitious battle-scenes require variation for Augustan readers in an era of post-Callimachean aesthetics of subtle literary texture and polish. The most substantial move in the Aeneid is to restrict considerably the actual volume of fighting described: though the ‘proem in the middle’8 at the start of Aeneid 7 suggests that the second half of the poem will have war as its theme, the action is in fact carefully managed so that the major battle-narrative does not start until Aeneid 9, and only three of the six books of the poem’s second half are taken up with fighting for more than half of their length (9,10 and 12). It will be immediately obvious from the analysis of the book’s structure given in the last section that a key strategy of variation in Aeneid 10 is to ensure a regular change of scene. There are 13 scenes in 908 lines, the longest of which is 117 lines: nothing is too extended and the pace moves on swiftly. The book is thus not unlike a multi-poem poetic book in its segmentation into manageable episodes; we can compare the Eclogues, ten poems in 819 lines. Alongside this, the text shows a careful alternation of static and dynamic scenes: the slowpaced divine council is followed by a pair of rapid battle-scenes, succeeded by the static catalogue. The use of speeches is also an important tactic of variation within scenes once the battle-action has fully commenced: speech takes up ~ 20% of 10.405–10.908.9 The use of similes functions as a similar strategy of narrative variation within the main battle (similes take up ~ 10 % of 10.405–10.908). The density of similes in Book 10 overall is roughly comparable to the parallel phase of the battle-narrative of the Iliad which narrates the death of Patroclus; the all-action Iliad 16 (867 lines) has 20 similes (one for every ~44 lines), one of the densest frequencies of any Iliadic book,10 while Aeneid 10.118–10.908 (790 lines), the battle-narrative after the divine council, has 15 similes (one for every ~53 lines).

7 8 9 10

Harrison 1991: xxxi–xxxiii, with further references; see also Harrison 1988. For this key concept see Conte 2007: 219–231. For speeches in the Aeneid more generally see still the analysis by Highet 1972. Stanley 1993: 260.

362 4

harrison

Catalogue of Similes

It is useful at this point to set out the similes of Aeneid 10 and their chief Homeric models.11 10.97–10.99 10.134–10.137

10.264–10.266 10.270–10.275 10.356–10.361 10.405–10.411 10.454–10.456 10.565–10.570 10.603–10.604

10.641–10.642

10.693–10.696 10.707–10.718 10.723–10.729

11

The gods react to Jupiter’s speech like winds caught in a forest [cf. Iliad 14.398–14.400] The young Ascanius amid battle is like a precious jewel in its setting [cf. Iliad 4.141–4.145, Odyssey 6.232–6.235] The Trojans cheer Aeneas’ return like the cries of cranes [cf. Iliad 3.3–3.6] Aeneas’ head glows with flames like a comet [cf. Iliad 5.4–5.7, 18.205–18.206, 22.36–22.51] The two sides are locked in even combat like winds in a storm [cf. Iliad 9.4–9.8, 16.765–16.772] The Arcadians’ courage kindles like a fire in a cornfield [cf. Iliad 11.153–11.157, 15.605–15.606, 20.490–20.492] The charging Turnus is like a lion attacking a bull [cf. Iliad 5.161–5.165, 16.487–16.491, 17.542] Aeneas looks like the theomachic giant Aegaeon [no Homeric model—see below] Pair of mini-similes (Aeneas is like a stream in flood or a dark tornado) [cf. Iliad 5.87–5.88, 11.297] Pair of mini-similes (Aeneas’ phantom like a ghost or a dream) [no Homeric model] Mezentius is like a cliff resisting wind and waves [cf. Iliad 15.618–15.622] Mezentius is like a boar resisting hunters [cf. Iliad 11.414–11.420, 13.471–13.476, 17.61–17.69] Mezentius is like a lion hunting for prey [cf. Iliad 3.23–3.28, 12.299–12.308, 17.61–17.69; Odyssey 6.130– 6.136, 22.402–22.406]

For more discussion of detailed parallels see the relevant notes in Harrison 1991, and for an important overall treatment of Homeric elements in the narrative of Aeneid 10 see Barchiesi 1984 (English version: Barchiesi 2014).

force, frequency and focalisation

363

10.762–10.768 Mezentius is like Orion wading through the sea/walking on mountains [no Homeric model—see below] 10.803–10.810 Aeneas under pressure is like a farmer or traveller caught in a storm [cf. Iliad 15.170–15.173, 16.263–16.265]

5

Distribution of Similes between Scenes

Given that I have suggested above that similes play an important role in varying the pace of the narrative, it is worthwhile to consider their distribution over the individual scenes of the book, with some brief comment. 1–117

118–145 146–162 163–214 260–307 308–361 362–438 439–509 510–605 606–688 689–768 769–832 833–908

Only one short simile: in this scene the opportunity for similes is limited, given that the divine council necessarily mostly consists of speeches. Only one simile, but an important description of a major character. No similes (but a very brief scene) No similes (this catalogue provides a different kind of variation) Two similes close together, marking a key narrative moment One simile at the end marking key moment in the battle One long simile encapsulating the key theme of the scene One simile encapsulating the key theme of the scene Two similes (one long, one short double) expressing key theme of scene Only one short double simile (but variation is maintained through a number of speeches: 607–620, 622–627, 628–632) Four similes for Mezentius, who needs rapid characterization on his first major appearance (see further below). One long simile expressing key theme of scene (but a pair of speeches: 773–776, 811–813, and an authorial apostrophe: 791–793) No similes at all (but seven speeches: 811–812, 825–830, 845–856, 861–866, 875–876, 897–898, 900–906)

It is clear from this list that similes are generally evenly distributed within the book, with relatively little bunching (with the signal exception of the four similes for Mezentius at 689–768, see below). This allows each of them to have a clear function, often emphasising a key theme of the scene in which they

364

harrison

occur. I will now examine several key examples in detail, stressing their narrative functions, and considering how the application of Homeric and other intertextualities and (especially) of focalisation can open up interesting literary interpretations.

6

Some Individual Similes and Their Narrative Functions

At 10.134–10.138 the young Ascanius in the midst of the battle is compared to a precious jewel in its setting: qualis gemma micat fuluum quae diuidit aurum, aut collo decus aut capiti, uel quale per artem inclusum buxo aut Oricia terebintho lucet ebur; fusos ceruix cui lactea crinis accipit et molli subnectens circulus auro. Just as a jewel shines, which splits a frame of tawny gold, An ornament for neck or head, or just as by the artist’s skill Ivory gleams, inlaid in box or Orician terebinth: His milky neck takes the weight of his flowing hair, Along with the circlet which weaves it with soft gold. As already suggested, this simile presents a key character in the Aeneid and makes an important point; it dominates its short scene, thus acquiring particular emphasis. The young Ascanius, future founder of the gens Iulia, here in (temporary and supervised) command of the Trojans in his father Aeneas’ absence, is presented as a glamorous and beautiful boy, appropriate for the grandson of Venus, whose care for him has just been stressed (cf. 10.132 Veneris iustissima cura, ‘Venus’ most proper care’); the element of divine protection by a benevolent goddess is also implied by this passage’s imitation of the simile at Odyssey 6.232–6.235, where Athene improves Odysseus’ appearance as he is about to meet the princess Nausicaa like a craftsman overlaying silver with gold, a simile also drawn on at the parallel point in the Aeneid where Venus enhances Aeneas for his meeting with Dido at 1.589–1.593. This stress on vulnerability and divine protection is in interesting tension with the context here, where the Trojans are under considerable military pressure, something brought out by careful consideration of the elements of comparison here. The jewel-like handsome face of the boy is framed by his luxuriant hair, the primary comparandum, but the precious and vulnerable boy is also imagined as enclosed (inclusum) by

force, frequency and focalisation

365

his men: inter medios (10.132) here suggests not so much heroic participation as a non-combatant protected by others. This is consistent with Ascanius’ only earlier intervention in the battle in the previous book, where his slaying with the bow of the boastful Numanus Remulus was followed by Apollo’s congratulations, but also by the god’s firm instruction to keep out of the war (cf. 9.656 cetera parce, puer, bello, ‘for the rest, refrain from war, my boy’). This element of non-combatant status is underlined by the other Homeric model for this simile (Iliad 4.141–4.145), where the wounded Menelaus’ thigh which puts him out of his duel with Paris is compared to a woman staining ivory to make elegant horse-ornaments. All this emphasises a key point for the plot: Ascanius, though under divine protection as the carrier of the crucial Julian bloodline, is out of place in the heat of battle, and the Trojans sorely need Aeneas to return to take over leadership again in this time of pressure, an event which will occur in the next scene. Consideration of the focalisation of this simile leads to interesting possibilities of interpretation.12 The image might seem primarily to supply the view of Ascanius taken by his protective divine grandmother Venus, whose close interest in the battle and literal spectatorship of it is confirmed by 10.760 (hinc Venus, hinc contra spectat Saturnia Iuno, ‘on this side watches Venus, on that Juno in opposition’), and whose concern for Ascanius has just been stressed in the narrative at 10.132 (see above). On the other hand, the ecce, ‘behold’, of 10.133, immediately preceding the simile, sharply directs the attention of the reader to Ascanius’ beauty.13 This is not the only place in this book where the reader’s (putatively male) eye is drawn by the poet-narrator to the picture of an attractive young boy: at 10.325 the Italian Cydon is nearly slain while following his new partner Clytius, characterised as being at the age of greatest homoerotic attraction with new down on his cheeks (10.324).14 The quasi-feminine attractions of the young Ascanius for the Roman male are clear in the simile; the jewel to which he is compared is for a pendant or a diadem, both of which are firmly female accoutrements, of types linked with Greek rather than Roman culture.15 The narrative may offer us two different perspectives on the beautiful youth here, familial affection and erotic interest. This would be a further

12 13 14 15

For a pioneering use of this in Homeric analysis see De Jong 1985. For some good brief remarks on the simile from this perspective see Lovatt 2013: 281. For details see Harrison 1991: 158. For images of such artefacts see Higgins 1980, frontispiece, pl. 46 and 49A, and for the diadem in particular as Greek not Roman see Higgins 1980: 177 (for Roman necklaces with pendants see also Higgins 1980: 179–181).

366

harrison

example of Don Fowler’s ‘deviant focalisation’, providing an unexpected and ideologically alternative point of view.16 For the double perspective using the image of an attractive artefact we can compare Aeneid 1.589–1.593, already noted above, where Venus improves the appearance of Aeneas to meet Dido: namque ipsa decoram caesariem nato genetrix lumenque iuuentae purpureum et laetos oculis adflarat honores: quale manus addunt ebori decus, aut ubi flauo argentum Pariusue lapis circumdatur auro. For herself The mother had breathed on to her son a graceful head of hair And the rosy light of youth, and luxuriant beauty into his eyes: Just like the ornamentation skilled hands add to ivory, or when Silver or Parian marble is surrounded by tawny gold. Again the focalisation might seem in the first place to be that of the proud mother admiring her handiwork, given the evident analogy between Venus and the artisans of the simile, but there is a clear further view in the context, that of Dido herself, alluded to in the next line, as she sees Aeneas for the first time; her erotic interest is clearly marked a little later in the narrative (1.613 obstipuit primo aspectu Sidonia Dido, ‘Sidonian Dido was astonished at her first sight of him’). Again erotic and familial perspectives are combined. At 10.270–1.275 Aeneas’ head glows with flames like a comet as he enters the battle: ardet apex capiti tristisque a uertice flamma funditur et uastos umbo uomit aureus ignis: non secus ac liquida si quando nocte cometae sanguinei lugubre rubent, aut Sirius ardor ille sitim morbosque ferens mortalibus aegris nascitur et laeuo contristat lumine caelum.

16

See Fowler 1990.

force, frequency and focalisation

367

The peak on his head burned, and a grim flame poured down From its top, and the golden shield-boss disgorged terrible fire: Just as when bloody comets on a clear night Shine a mournful red, or the burning heat of Sirius, The bringer of thirst and disease to miserable mortals, Rises and saddens the sky with its gloomy light.17 Two Homeric models for the simile (Iliad 5.4–5.7, 18.205–18.206) suggest that Aeneas in battle is parallel to the devastating Diomedes and Achilles, whose heads are both enhanced with divine fire by Athene (Diomedes is compared in Book 5 to a scorching summer star during harvest, Achilles in Iliad 18 to beacon-fires in a war situation); the parallel with Achilles goes further, since both he and Aeneas in this book are entering the battle for the first time in their epics. Both these Homeric analogies imply that Aeneas’ fiery appearance is also a product of divine intervention, thus adding an element not made explicit in Vergil’s narrative. Once again interesting issues of focalisation arise: as often in battle-narrative, we need to think of the possible perspectives of the two opposing sides in the conflict. For the implied readers of the poem so far, who (we may assume) are in principle on Aeneas’ side,18 his flaming head is surely a positive sign of his divine backing: this is confirmed by earlier parallels such as the flaming head of Ascanius at Aeneid 2.682–2.684, and the flaming head of Augustus at Aeneid 8.680–8.681, both of which are clearly positive omens (on Augustus see further below). All three scenes significantly involve ancestors or members of the gens Iulia, and together can be seen as constituting an allusion to the sidus Iulium, the comet of 44 BC which was seen as the emblem of Julius Caesar’s apotheosis;19 as has been pointed out, the specific detail of the comet in the simile, though Homeric in origin (see below) seems to confirm this link.20 Aeneas is destined for divinity like his descendant Julius Caesar (cf. Aeneid 1.259–1.260).

17

18 19 20

For the reading tristis (conjectured by Gabriele Faerno) rather than the traditional cristis see Harrison 1991: 146 (where I failed to note that it is already found as a variant in the manuscripts of Ti. Claudius Donatus’ Interpretationes Vergilianae on this passage); it has been adopted by Conte 2009. For the term ‘implied reader’ (the reader imagined by the historical author) see Booth 1983: 71–76. See Williams 2003 for a full treatment of the Vergilian allusions to the sidus Iulium (I draw on her treatment of the passages treated here at pp. 8–9). Williams 2003: 9.

368

harrison

But we also need to consider the viewpoint of Aeneas’ enemies. Here a third Homeric model for the simile is relevant. At Iliad 22.26–22.31 Priam sees Achilles in deadly pursuit of Hector, and Achilles is compared by the poet to the baleful summer Dog-Star which brings fever to mortals. In context this clearly embodies Priam’s fears for his son as he faces his supreme adversary,21 fears which will be all too soon confirmed. In Aeneid 10 the simile has the same predictive capacity: Aeneas’ reappearance in the battle is clearly bad news for Turnus, who now like Hector has to face an enemy of superior power who is to bring about his death,22 and Turnus’ men no doubt appreciate this, as they do later in the poem when the two are scheduled to meet in the formal duel of Book 12 (12.216–12.218); to them Aeneas is a devastating and pernicious natural force like a comet or Sirius. This simile can therefore be seen as expressing two views: the recognition in the Roman reader that Aeneas has divine backing and that his victory will lead to the establishment of the gens Iulia and the glory of Julius Caesar, and the justifiable fears of Turnus and his men as their future conqueror Aeneas enters the battle. Here again we can find competing perspectives on the action expressed in the simile, this time the view of characters in the poem (intradiegetic, inside the narrative) and that of its implied readers (extradiegetic, outside the narrative).23 A similar ambiguity can be found at Aeneid 8.678–8.681: hinc Augustus agens Italos in proelia Caesar cum patribus populoque, penatibus et magnis dis, stans celsa in puppi, geminas cui tempora flammas laeta uomunt patriumque aperitur uertice sidus. On this side was Augustus Caesar, urging Italians to battle, With the senate, people, gods of home and the great powers, Standing on the lofty ship-stern: his triumphant temples Disgorged flames, and his father’s star is revealed on his head. Here the future Augustus is described at the battle of Actium, on the prophetic shield of Aeneas, not in a simile but in a proleptic ekphrasis, a feature which has a similar status in terms of notional distance from the main narrative.24 Once 21 22 23 24

See De Jong 2012: 65. See further Williams 1983: 65. Such internal/external contrast can also be deployed in the narrative analysis of ekphrasis: see e.g. Harrison 2001. For this feature see Harrison 2001.

force, frequency and focalisation

369

again there are interesting issues of focalisation: the explicit evocation of the sidus Iulium (patrium … sidus) and the positive tone of laeta suggests the perspective of the implied Augustan reader as well as the timeless perspective of the shield’s divine creator Vulcan, while it is clear that Aeneas himself cannot decipher this element of future history (cf. his reaction to the shield’s decoration at 8.730 rerumque ignarus imagine gaudet, ‘and ignorant of its subjectmatter he rejoices in the image’).25 Again we need to look to the other side in the battle: the flames which appear so propitious from a Julian/Augustan perspective are not so positive for Augustus’ Antonian enemies at Actium. While I would not agree with those who see this scene as undermining the heroic status of Augustus,26 the alternative focalisation of the opposition should not be ignored, especially for the original implied readership, many of whom would have been on the losing side in the civil war of Actium. At 10.454–10.456 Turnus charging at Pallas is compared to a lion attacking a bull: utque leo, specula cum uidit ab alta stare procul campis meditantem in proelia taurum, aduolat, haud alia est Turni uenientis imago. And as a lion, when it sees from a high vantage-point A bull standing at a distance on the plains planning battle Flies towards it, no other was the image of Turnus as he charged. The Homerically-learned first-time reader, who does not know that Turnus will shortly kill Pallas, can infer this outcome from the same simile’s repeated use in the Iliad, where the warrior compared to the lion regularly overcomes the one compared to the bull (cf. Iliad 5.161–5.165, 16.487–16.489, 17.542); the repeat reader, of course, along with the poet-narrator, clearly knows this impending result. Here, then, the simile has a clear proleptic function, and there is an element of dramatic irony in that case, since the reader knows more than the characters. The reader (whether first- or second-time) thus provides one possible focaliser of the simile, outside the narrative (extradiegetic), but there is another more obvious one inside the narrative (intradiegetic): the doomed Pallas, who is seeing Turnus move rapidly towards him with hostile intent. This is 25 26

For the different perspectives of Aeneas and Vulcan on the Shield see De Jong 2015: 903– 907. E.g. Boyle 1986: 93.

370

harrison

perhaps indicated by the word imago, ‘image’, which suggests a viewer within the story. Pallas, then, in the role of the bull, may already see that he is outclassed as a warrior by Turnus, in the role of the lion. The word imago seems to be used self-consciously here, since it is the standard Latin word for ‘simile’: here as in some other passages of Latin epic we can see a simile referring to its own status as an image within the narrative (meta-imagery).27 At 10.565–10.570 Aeneas is compared to the theomachic giant Aegaeon: Aegaeon qualis, centum cui bracchia dicunt centenasque manus, quinquaginta oribus ignem pectoribusque arsisse, Iouis cum fulmina contra tot paribus streperet clipeis, tot stringeret ensis: sic toto Aeneas desaeuit in aequore uictor ut semel intepuit mucro. Just like Aegaeon, who (they say) had a hundred arms And a hundred hands, and fire flaming from his fifty mouths And chests, when, against the thunderbolts of Jupiter, He was making such a din with so many identical shields, drawing so many swords: Just so Aeneas followed his rage to the end over the whole plain in victory, Once his sword-point had warmed up. Here there is no Homeric source, but the hundred-hander Aegaeon still seems to be derived from the Greek epic tradition in the form of the Titanomachy of Eumelus, in which he was depicted fighting against the gods.28 The ‘Alexandrian footnote’ dicunt29 suggests that the poet has made a careful choice of Eumelus’ version of Aegaeon here, since in Homer and Hesiod Aegaeon is presented by contrast as a helper of the gods (Iliad 1.401–1.406, Theogony 711–720). This simile is placed in the episode where Aeneas takes immediate indiscriminate revenge for the death of Pallas, and it has sometimes been argued that the poet is presenting Aeneas in an unambiguously unfavourable light, criticising him by a authorially-sanctioned comparison with a monstrous theomachic

27 28 29

For this technique see Harrison 2003. See West 2003: 224–225. See Hinds 1998: 1–16; and for Vergilian examples see Horsfall 1990.

force, frequency and focalisation

371

giant.30 However, as Gordon Williams has noted, ‘these details have no objective analogy in Aeneas, only in the terrifying image of him that the doomed Latins see’.31 In other words, we see here primarily the irrational focalisation of Aeneas’ enemies: this is supported by the context, where in the immediate sequel to the simile (10.570–10.574) the chariot-horses of the Italian warrior Niphaeus see the fearsome Aeneas and panic, throwing their master. Lastly, as already noted, the rapid series of similes used of Mezentius forms an interesting exception to the usual even distribution of similes over the battlenarrative of book 10; four of its fifteen similes concern him and occur in the short space of 10.693–10.770 at more than double the average frequency (see above). There is a clear narrative reason for this: the reader of the Aeneid so far has only seen Mezentius in person in brief glimpses in the Catalogue of Latins (7.647–7.648, 7.653–7.654) and in the fighting of Book 9 (9.521–9.522, 9.586– 9.589), though an extensive negative account of him has also been provided by Evander (8.470–8.504). This important character, who dominates the final section of Aeneid 10 as Aeneas’ chief enemy after the divinely-engineered disappearance of Turnus from the battlefield at 10.665–10.688, needs to be properly and rapidly characterised before his death at the end of the book; these similes, like Mezentius’ two pairs of speeches (cf. 10.739–10.741, 10.743–10.744, pithy and forceful, 10.845–10.856, 10.861–10.866, tragic and pathetic),32 are an important element in this process. The first three similes use standard Homeric material (see above), comparing Mezentius to a cliff resisting wind and waves (10.693–10.696),33 a boar resisting hunters (10.707–10.718), and a lion hunting for prey (10.723–10.729); the character’s physical prowess as a warrior is thus swiftly established, making him a suitable match for Aeneas on the battlefield. The fourth simile, by contrast, has no Homeric model, and is worth particular scrutiny (10.762–10.770): quam magnus Orion, cum pedes incedit medii per maxima Nerei stagna uiam scindens, umero supereminet undas, aut summis referens annosam montibus ornum 30 31 32 33

Cf. e.g. Bishop 1988: 321–322. Williams 1983: 180. See further Harrison 1991 on all these passages. A clear bellicose inversion of the same image used of the unsuccessfully pacific Latinus at 7.586–7.590.

372

harrison

ingrediturque solo et caput inter nubila condit, talis se uastis infert Mezentius armis. huic contra Aeneas speculatus in agmine longo obuius ire parat. As big as Orion, When he goes on foot, scything his way through the deepest waters Of Nereus, towering over the waves with his shoulder, Or when he carries an aged ash-tree from mountain-tops, And walks on the earth and hides his head amid the clouds: Such was Mezentius as he advanced with vast weapons. On the other side Aeneas, spying him in the long line some way off, Prepared to go to meet him. Here Mezentius is compared to the giant Orion wading through the sea or striding down mountains. Orion is already portrayed as a giant hunter mightily armed in the Underworld of the Odyssey (11.572–11.575), and his capacity to walk over the sea, given to him by his father Poseidon, may go back to Hesiod; it is significantly changed here into walking through the sea (see further below).34 Once again there are interesting issues of focalisation: the primary focaliser would seem to be Aeneas, whose spotting of Mezentius in the battle is specifically mentioned at 10.769 (speculatus); Mezentius seems giant-like to him. But why Orion? Orion, I suggest, comes to Aeneas’ mind as a traditional offender against the gods, who came to grief at the hands of Artemis;35 this fits the monstrous account already given of Mezentius to Aeneas by Evander in Book 8 (8.481–8.495), in which divine punishment is invoked for his sacrilegious crimes (8.484). There is a clear match with the comparison of Aeneas with Aegaeon just discussed: the two warriors are seen as impious giants by their enemies, and are thus neatly matched as opponents in the climactic duel of Vergil’s book. A further dimension to the simile also becomes available here. It has been well noted how the presentation of Orion here echoes the presentation of the Cyclops Polyphemus earlier in the Aeneid: at Aeneid 3.655–3.661 Aeneas narrates how the Trojans escape an encounter with the now blinded giant, who supports himself on a tree-trunk (3.659) and wades through the depths of the sea (3.664–3.665).36 It seems clear that in the simile of Aeneid 10 Vergil has 34 35 36

Cf. Eratosthenes Cat. 32 with Pàmias i Massana 2013: 286–287. See Apollodorus 1.4.3. See Glenn 1971; Polyphemus’ wading through the sea does not occur in Homer, but may

force, frequency and focalisation

373

changed the traditional capacity of Orion to pass over the sea’s surface into the ability of the Cyclops to wade through the deep in order to emphasise the link between Orion and Polyphemus;37 the additional element that Orion was famously blinded (Eratosthenes Cat. 32, Apollodorus 1.4.3) establishes another connection between the two figures. What is the effect of this carefully-manufactured link between Mezentius, Orion and Polyphemus on the reader’s view of Vergil’s simile? I suggest that it allows the reader to form a more sympathetic view of Mezentius than which is natural for his enemy Aeneas in the middle of the battle as he moves in for the kill, by evoking in the Orion-simile traces of the pathetic figure of the disabled Cyclops of Aeneid 3. This much is available to the poem’s first-time reader; but even more is available to the second-time or repeat reader, for whom the link between the Cyclops and Mezentius looks forward proleptically to Mezentius’ own imminent end, in which Mezentius himself becomes like the Vergilian Cyclops a pathetic and disabled figure in a masterpiece of transformational characterisation. It has been well argued that in his relatively brief appearance in Aeneid 10, Mezentius turns from monster to tragic and sympathetic bereaved father through his reaction to the death of his son Lausus,38 and the key role of the Cyclops in this transformation is confirmed by a further link of Polyphemus and Mezentius a little later in the book: at 10.861–10.866 Mezentius, about to return to the battle to meet certain death, speaks pathetically to his horse in a scene plainly based on Polyphemus’ Homeric address to his pet ram (Odyssey 9.446–9.460).39 The Orion-simile, I suggest, already contains the seeds of this transformation for the repeat reader through its links with Polyphemus. Here again the consideration of multiple focalisation, both that of Aeneas within the narrative and that of implied readers outside the narrative, allows the creation of a fuller and more rounded literary character.

7

Conclusion

In this paper I hope to have demonstrated how the similes of Aeneid 10 make an important contribution to its narrative texture. They are a key tool in the revival of Homeric-style battle-narrative for a Roman Augustan readership

37 38 39

have done so in another version of the story, to which Lucretius 1.200–1.201 seems to allude. I missed this important point in Harrison 1991. See especially Gotoff 1984. See the full account in Glenn 1971, summarised at Harrison 1991: 274–275.

374

harrison

which requires much more variation and polish in avoiding the real dangers of repetition and monotony in such material; they can also open up alternative perspectives on the action through considerations of multiple focalisations. Importantly, these focalisations are not always restricted to the views of characters in the narrative, though these are important; the perspectives and responses of various categories of implied readers are also highly significant. In calculating those perspectives, we need to take account of what those readers are likely to know about the poem’s plot and earlier literary models, what reader-response theory has called their repertory. Such an approach can add to and enrich our literary interpretation even of much-studied texts.

Bibliography Bal, M., Narratology: Introduction to the Theory of Narrative (Toronto 1985). Barchiesi, A., La traccia del modello: effetti omerici nella narrazione virgiliana (Pisa 1984). Barchiesi, A., Homeric effects in Vergil’s narrative (Princeton 2014). Bishop, J.H., The Cost of Power: Studies in the Aeneid of Virgil (Armidale 1988). Booth, W.C., The Rhetoric of Fiction ([1961] Chicago 1983). Boyle, A.J., The Chaonian Dove: Studies in the Eclogues, Georgics, and Aeneid of Virgil (Leiden 1986). Conte, G.B., The Rhetoric of Imitation (Ithaca 1986) Conte, G.B., The Poetry of Pathos (Oxford 2007). Conte, G.B., P. Vergilius Maro: Aeneis (Berlin 2009). Fowler, D.P., ‘Deviant Focalisation in Virgil’s Aeneid’, Proceedings of the Cambridge Philological Society 36 (1990) 42–63 [reprinted in Fowler 2000: 40–63]. Fowler, D.P., Roman Constructions (Oxford 2000). Genette, G., Narrative Discourse (Ithaca 1980). Genette, G., Narrative Discourse Revisited (Ithaca 1988). Glenn, J., ‘Mezentius and Polyphemus’, AJPh 92.2 (1971) 129–155. Gotoff, H.C., ‘The transformation of Mezentius’, TAPhA 114 (1984) 191–218. Harrison, S.J., Vergil: Aeneid 10 (Oxford 1991). Harrison, S.J., ‘Picturing the future: the proleptic ekphrasis from Homer to Vergil’, in S.J. Harrison (ed.), Texts, Ideas and the Classics (Oxford 2001) 70–90. Harrison, S.J. ‘Meta-Imagery: some self-reflexive similes in Latin epic’, in A.F. Basson & W.J. Dominik (eds.), Literature, Art, History: Studies on Classical Antiquity and Tradition In Honour of W.J. Henderson (Frankfurt 2003) 9–16. Higgins, R., Greek and Roman Jewellery ([1961] London 1980). Hinds, S.E., Allusion and Intertext: Dynamics of Appropriation in Roman Poetry (Cambridge 1998).

force, frequency and focalisation

375

Highet, G. The Speeches in Virgil’s Aeneid (Princeton 1972). Horsfall, N.M., ‘Virgil and the illusory footnote’, in F. Cairns (ed.), Papers of the Leeds International Latin Seminar: Sixth Volume (Leeds 1990) 49–63. Jong, I.J.F. de, ‘Fokalisation und die Homerischen Gleichnisse’, Mnemosyne 38.3/38.4 (1985) 257–280. Jong, I.J.F. de, Homer Iliad Book XXII (Cambridge 2012). Jong, I.J.F. de, ‘Pluperfects and the Artist in Ekphrases: From the Shield of Achilles to the Shield of Aeneas (and Beyond)’, Mnemosyne 68.6 (2015) 889–916. Lovatt, H., The Epic Gaze: Vision, Gender and Narrative in Ancient Epic (Cambridge 2013). Pàmias i Massana, J., Eratosthène de Cyrène: Catastérismes (Paris 2013). Stanley, K., The Shield of Homer. Narrative structure in the Iliad (Princeton 1993). Tompkins, J. (ed.), Reader-Response Criticism: From Formalism to Post-Structuralism (Baltimore 1980). West, M.L., The Epic Cycle (Oxford 2003). Williams, G.W., Technique and Ideas in Virgil’s Aeneid (New Haven 1983). Williams, M.F., ‘The Sidus Iulium, the divinity of men, and the Golden Age in Virgil’s Aeneid’, LICS 2.1 (2003) 1–29.

chapter 16

Parallel Plotlines: the Function of Similes in the Battle Narrative of Vergil, Aeneid 10 (2) Michiel van der Keur

1

Introduction*

In the preceding chapter, Harrison has focused on the function of the similes in Vergil’s Aeneid 10 in characterization, in varying the pace and nature of the narrative, and in presenting the different perspectives of the opposing sides in the epic through focalization. This chapter complements his discussion by addressing two other roles of the similes on a macrostructural level and their contribution to larger ‘textual strategies’, which allow the audience to recognize the significance of their context in the book and epic as a whole. Vergilian similes support the narrative by establishing a secondary plotline of their own, a sustained set of motifs, complementing the main narrative either through the imagery itself or through their intertextual models. The latter will be discussed in the first section, which will focus on the interaction of focalization and intertextuality in the similes, and the way in which they (mis)guide the interpretation and expectation of the reader. The analysis particularly studies the relation between the intertextual reading by the audience of the Aeneid and the story-internal development of Turnus’ character-perspective. His reading of events conflicts with the audience’s expectations, raised by Homeric models; earlier scholarship has drawn attention to the self-deception of the Italians and their leader Turnus in their perspective on the war.1 Here, we will especially consider the role of the similes in developing this incongruity in each of the three scenes in which Turnus figures in Aeneid 10, and how the intertextual quality of these similes points up his misguided behaviour. The second section elaborates on the imagery itself. Several similes in Aeneid 10 share the same conceptual domain, presenting the battle as a storm or a clash of winds. I will investigate how the similes’ imagery functions on a book

* I am much indebted to the editors and to Stephen Harrison for their suggestions and comments on an earlier draft of this chapter. 1 See particularly the seminal article by Anderson 1957.

© koninklijke brill nv, leiden, 2019 | doi:10.1163/9789004383340_017

parallel plotlines

377

level and how it contributes to establishing a sub-narrative that supports and mirrors the plotline.2

2

(Not) through Turnus’ Eyes

Aeneas’ arrival on the battlefield, accompanied by his army of allies, is one of the highlights of Aeneid 10. When Aeneas’ ship glides towards the shore and his arrival is witnessed by both the Trojans and the Rutulians, the reader’s expectation is raised that a decisive battle will soon follow. The phrase used for Aeneas, stans celsa in puppi (10.261), recalls Vergil’s earlier portrayal of Augustus attaining victory at Actium (8.680),3 suggesting that it will not be long for Aeneas’ own victory to come. The poet then describes the reactions of the two sides in the war, each with its own simile. The relieved Trojans raise a joyous cheer, for which the poet compares them to cranes fleeing a storm with loud cries (10.262–10.266). The primary intertext is Hom. Il. 3.3–3.7, also a simile used of the Trojans as battle commences.4 Since the subsequent passage in Iliad 3 does not actually bring a regular battle, but rather the duel between Paris and Menelaus, Vergil’s echo of the simile suggests that we should expect a similar showdown between the two leaders of the war in Latium very soon.5 Directly after the cranes-simile illustrating the Trojans’ cry, the narrator turns to the other side in the war. The anticipation of a duel is heightened even more in these lines, where, in a second simile, Aeneas’ head glowing with flames is compared to a comet or the Dog-star Sirius (10.272–10.275). As Harrison has also noted, the imagery alludes to the beginning of Iliad 22 with its Sirius-simile for Achilles in the build-up to the duel with Hector.6 For the present purpose, I wish to draw attention to the context of Vergil’s simile: 267

at Rutulo regi ducibusque ea mira uideri Ausoniis, donec uersas ad litora puppis respiciunt totumque adlabi classibus aequor. […]

2 Apart from a few short excursions, the question of how these similes relate to storm imagery in the rest of the Aeneid will not be discussed. For a study of the topic, see e.g. Hornsby 1970: 19–44. 3 See for a discussion Harrison 1991: 142; and Binder 1971: 225–226. 4 Harrison 1991: 143. 5 The details of Vergil’s simile deviate from Homer’s; this will be addressed further below. 6 See in this volume Harrison, 368.

378

van der keur

(Aeneas’ head glows with flames, like a comet or Sirius.) […] 276 haud tamen audaci Turno fiducia cessit litora praecipere et uenientis pellere terra. Verg. Aen. 10.267–10.277

But all this seems strange to the Rutulian king and to the Italians, until they look round to see ship-sterns turned towards the shore, and the whole ocean aglide with fleets. […] Yet bold Turnus has not lost his confidence to take the shore first and to drive the invaders from the land.7 In Homer, the Sirius-simile defines Priam’s viewing of the feared enemy who will soon kill his son. As Harrison has suggested, focalization is also important in Vergil’s scene. Both uideri and respiciunt indicate that the Rutulians are focalizers; the simile that follows arguably mirrors their viewpoint. Turnus’ response is presented in contrast (276 haud tamen) with the emotion implied by the simile—despite Aeneas’ ominous appearance, he remains confident. A comparison with the context of Homer’s simile is instructive. There, the simile is framed by references to Priam, and thus reflects his anxiety; he is filled with foreboding and fear by the sight of Achilles.8 However, whose point of view is exactly reflected by Vergil’s simile? The contrast between the simile and Turnus’ response in 276–277, a contrast that is not present in Homer, leaves room for interpretation. Is it Turnus’ own state of mind, as he notes the threat posed by Aeneas and yet (tamen) steels himself? Or is it the reaction of the other Rutulians, who dread the enemy’s arrival (and, in view of the Homeric intertext, may also fear for their champion, as Priam feared for Hector), whereas Turnus himself is oblivious of the danger? Perhaps the intertextual nature of the simile even primarily implies what Turnus’ reaction should have been, in contrast with his actual reaction. In any case, the simile is not only important for the depiction of Aeneas himself, but also for the characterization of the Rutulian

7 For all passages from Aeneid 10, I have printed Harrison’s translation (1991) with some adaptations; for quotations from other parts of Vergil’s oeuvre, the Loeb translation by Fairclough (rev. Goold) has been used. 8 See De Jong ad Hom. Il. 22.25–22.32: ‘the simile also conveys something of Priam’s feelings of fear; for him the star/Achilles is a ‘bad omen’, which/who will bring fever/misery to wretched mortals.’

parallel plotlines

379

hero, who as yet does not fully recognize that Aeneas is arriving as a second Achilles, and that he himself, by implication, will suffer Hector’s fate. Turnus’ ignorance of the role he is (re)playing is evident elsewhere as well; it will be seen that the similes continue to guide the reader’s interpretation and contrast with Turnus’ own perception. Confident that the Trojans will again be defeated in this new ‘Trojan war’, Turnus deceives himself into thinking that he is another Achilles.9 In his second appearance in Aeneid 10, his duel with Pallas (10.439–10.509), the Rutulian comports himself as such, or more specifically as the Achilles of Iliad 22, as his words at 10.442–10.443 suggest. Turnus tells his allies to step back, claiming Pallas for himself; he even wishes that Pallas’ father could be present to witness his son’s death.10 The combined phrases evoke the setting of Iliad 22: there, Achilles also signals his companions to stand back, lest another would steal his glory (22.205–22.207), and kills Hector while Priam is watching from the walls—Priam, who had been presented as the main spectator in the beginning of the book, as we have seen. Yet, as is well known, the rest of the duel between Pallas and Turnus evokes different Homeric intertexts, the main ones being Patroclus’ slaying of Sarpedon and Patroclus’ own death at the hands of Hector. The blend of intertexts in Vergil’s scene reflects the ambiguity as to which literary role each of the combatants Pallas and Turnus is replaying exactly. Part of this ambiguity rests upon the simile found at the beginning of combat, when Turnus charges at Pallas. The simile has also been discussed by Harrison in the preceding chapter; I cite it again for convenience: utque leo, specula cum uidit ab alta stare procul campis meditantem in proelia taurum, aduolat, haud alia est Turni uenientis imago. Verg. Aen. 10.454–10.456

And as a lion, when it sees from a high vantage-point a bull standing at a distance on the plains planning battle, flies towards it, no different was the image of Turnus as he charged.

9 10

Cf. e.g. Aen. 9.742 hic etiam inuentum Priamo narrabis Achillem, ‘you will tell Priam that here too an Achilles has been found’. See also Anderson 1957: 24. solus ego in Pallanta feror, soli mihi Pallas / debetur; cuperem ipse parens spectator adesset, ‘I alone go against Pallas, to me alone Pallas is due. I wish his father were here to see the sight’. As Harrison (1991: 186) observes, the second part of the sentence strongly recalls Pyrrhus’ slaying of Polites before the eyes of Priam (cf. 2.538–539); this points to Turnus’ similarity to that other arrogant successor of Achilles.

380

van der keur

As Harrison observes, the Homeric antecedents for this simile inform the reader that the lion—that is, Turnus—will be victorious. The most elaborate of these antecedents, and presumably the most relevant one here, is 16.487– 16.489, where Sarpedon dying at the hands of Patroclus is compared to a bull killed by a lion. Harrison plausibly suggests that there is an internal focalizer of the simile, in view of the use of imago (Turnus is being viewed). But the interpretation is complicated by the fact that the lion itself is also ‘viewing’ (an element that Vergil has added to the Homeric model): the lion sees the bull readying for battle, and Turnus is seen as such a lion. Who, exactly, is viewing whom? We should make two observations here. Firstly, Vergil presents each of the combatants in turn as viewer of the other. The viewing lion of the simile seems to counterbalance Pallas viewing Turnus in 10.445–10.447; the poet presents Turnus’ perception of Pallas while avoiding strict parallelism. Secondly, it is important to distinguish the act of viewing from focalization. The use of imago indicates that it is again Turnus who is being viewed as viewer (that is, the simile’s imagery is not Turnus’ focalization);11 the word uenientis (‘approaching’) also implies that the viewpoint of the narrator coincides with that of Pallas. We see Turnus as he approaches like a lion with a hungry stare in his eyes. Does the simile (and its loaded Homeric intertext) imply that Pallas realizes he is doomed? Or is the simile instead focalized by the fearful allies of the young warrior? Between Pallas’ acceptance of battling Turnus and the simile, we find the reaction of the Arcadians, whose blood runs cold in their veins (10.452); it is their viewpoint that has been offered last. Thus, this situation may be similar to the previous simile, as the reader joins the onlookers in anticipating the outcome, in contrast with the bravura of the doomed hero.12 Turnus’ ‘self-presentation’ as an Achilles is thus belied by the simile, which evokes the duel between Sarpedon and Patroclus; this Homeric scene is confirmed as an intertext in the subsequent lines, where Hercules is consoled in his grief for young Pallas by Jupiter, who recalls his own inability to save Sarpedon (10.467–10.472).13 Yet in establishing this intertext, the simile itself is unstable. 11

12

13

Cf. Brackert 1962: 166, who notes that the events are represented ‘als ein Ereignis, das andere sehen’. He maintains that in the simile, all the various perspectives in the scene return; while it is true that the narrator thus expresses the fact that Turnus is viewing as well, the simile does not represent his focalization. The fact that Pallas presents his spear cast as an attempt against all odds (10.458–10.459 si qua fors adiuuet ausum uiribus imparibus, ‘in the hope that chance would aid one who dared, though inferior in strength’) does, however, imply that he is not blind to the probable outcome; see Harrison 1991: 189. One wonders if Turnus is to be regarded as a Patroclus. In one sense, at least, he is: he too, has assumed the role of Achilles to his own doom.

parallel plotlines

381

The phrase haud alia est Turni uenientis imago is used to connect the simile to the story world. Harrison plausibly suggests that imago, as the Latin word for ‘simile’, is used in a metapoetical sense here—the simile is referred to as a simile.14 We might perhaps tease out a bit more. This simile, as the narrator says, is haud alia, ‘not at all different’. When read as a metapoetical signpost, this insistence on correspondence draws attention to the degree in which Vergil’s text actually corresponds with the Homeric model of Il. 16.487–16.489—and in fact, there are some differences. Apart from the lion’s ‘viewing’, the Roman poet has added another element, namely that it looks from a height, specula ab alta. It has been suggested by Brackert that this detail recalls a second simile in Iliad 16, of a lion slaying a boar high in the mountains.15 The context for this simile is Hector killing Patroclus. What, then, is the simile’s claim to being haud alia? The reader is invited to see that it is an amalgam of two important similes in Iliad 16, just as the entire scene is a blend of the duels between Sarpedon and Patroclus, and Patroclus and Hector. In effect, Pallas takes on the traits of both the doomed Sarpedon and the doomed Patroclus: he is mourned by the gods, and mourned by the hero who will avenge his death at the end of the epic.16 The simile anticipates Turnus’ victory here, but also his future defeat.17 The third and last appearance of Turnus in Aeneid 10 is at 10.633–10.688, where Juno creates a phantom image of Aeneas to lure the Rutulian leader away, in order to prevent his being killed at the hands of the (real) Trojan hero. The scene has been compared to the Iliadic one in which Apollo rescued Hector from Achilles (Hom. Il. 20.443ff.), postponing the actual duel between the two heroes, and the one earlier in the epic (5.449ff.) in which the same god created a phantom image of Aeneas to take the place of the wounded Trojan, whom Apollo transported to safety;18 we might also compare the end of Iliad 21, where again Apollo, disguised as Agenor, leads Achilles away from the routed Trojans, allowing them to reach the city in safety. It needs no explanation that Vergil has inverted the situation: it is now the deceived hero who is the one being saved. This inversion calls attention to the misguided behaviour of Turnus on the bat14 15 16 17

18

See in this volume Harrison, 370. specula ab alta ~ Il. 16.824 ὄρεος κορυφῇσι. See Brackert 1962: 167 with n. 13. Brackert 1962: 167. For a similar technique of anticipating future defeat at the moment of triumph in Herodotus and Livy, see in this volume De Jong, 126; De Bakker & Van der Keur, 331–332. We may also compare Verg. Aen. 9.756–9.761, where Turnus manages to enter the Trojan camp. This is one of his greatest moments in the epic, yet his bloodlust prevents him from opening the gates to his comrades and making his victory complete. See Harrison 1991: 221–222 with bibliography.

382

van der keur

tlefield; the first and last of the three Homeric intertexts identified above (from Iliad 20 and 21) once again present Turnus in his self-chosen role of Achilles (and a deceived one, at that). As in the two earlier scenes from Aeneid 10 we analysed above, the contrast between reality and Turnus’ perception of it is underscored by a simile, which compares the illusory Aeneas to a ghost or a dream image: gressumque effingit euntis, morte obita qualis fama est uolitare figuras aut quae sopitos deludunt somnia sensus. Verg. Aen. 10.640–10.642

[…] and she fashioned for it the step of a walking man, just like the shades which are said to flit about after death, or the dreams that delude the slumbering senses. The Alexandrian footnote fama est (10.641) alerts us to yet another Homeric intertext for this phantom.19 Turnus’ pursuit of what he imagines is Aeneas, plays against a well-known simile in Iliad 22, where Hector being chased by Achilles is compared to a man pursued by another man in a dream—the one cannot escape, the other cannot catch him (Hom. Il. 22.199–22.201). The second half of the Vergilian simile accentuates the connection between Homeric simile and the reality in the Aeneid: the image of Aeneas is akin to such a dream image. By inserting the phantom, that in Homer only figures in a simile, into his story world, the Roman poet emphasizes how much Turnus differs from his role model Achilles. The latter at first cannot catch the all-too-real Hector, but will eventually kill him. Turnus, on the other hand, chases an imaginary foe; his re-enactment of Achilles is just as illusory. For second-time readers, the dream simile in 10.641–10.642 adumbrates another reworking of the same Homeric simile.20 In Aeneid 12, desperation takes hold of Turnus when he has stopped running from Aeneas but then fails to strike his enemy; just so, says the poet, in our dreams we might seem to try to move in vain (12.908–12.912). Vergil has distributed the two men featuring in the dream of the Homeric simile over the scenes in books 10 and 12 to great effect: Turnus chasing Aeneas is a man who pursues a dream image, never able to catch his quarry; Turnus chased by Aeneas is like a dreamer who cannot

19 20

Harrison 1991: 227–228 connects fama est with Lucretian echoes in 641. Gransden 1984: 150–151.

parallel plotlines

383

escape. This technique, in which (elements of) Homeric similes are distributed over two different scenes, can also be discerned in another way. In the chase scene in Iliad 22, the dream simile is preceded by another one: Achilles chasing Hector is like a hound hunting a fawn (Hom. Il. 22.189–22.192). It is significant that this other simile returns in the parallel scene in Aeneid 12, Aeneas’ pursuit of Turnus (12.749–12.755)—the ‘real’ imitation of Achilles’ chase of Hector. There, the poet has expanded the Homeric original with the notion that the deer cannot escape, caught between the river and the frightening plumes used by the hunter.21 The two Homeric similes in Achilles’ chase of Hector—of the hunting dog and the man running in his dreams—have thus been distributed among the two contenders for being the epic successor to Achilles, and each simile has been adapted to reflect the hero’s success: Turnus pursues only fantasy, whereas Aeneas’ success is real. In each of the three scenes featuring Turnus in Aeneid 10, the similes help in drawing out a contrast between Turnus’ perception of his own literary role and his actual role. At the same time, these similes function in guiding the reader’s expectation. Harrison rightly draws attention to the need to maintain the reader’s interest throughout the long battle-narrative. One of the ways of doing this, is hinting at a swift conclusion. When Aeneas approached the shore earlier in Aeneid 10, the similes describing the armies’ reactions raised the expectation for a decisive confrontation between him and Turnus, for a counterpart to the Homeric duel between Achilles and Hector. The necessary ingredients are soon supplied: with Pallas, a Patroclus dies on Latian soil, and Aeneas accordingly transforms into a raging Achilles. The poem quickly— too quickly for Juno!—builds towards a Homeric clash between the main heroes. And we do indeed get a version of Iliad 22 in 10.633ff.—but it turns out to be Turnus’ misguided version. The ‘seed’ of the Sirius-simile, anticipating the climactic moment of a Vergilian Achilles, has borne unexpected fruit. The chase scene both ends Turnus’ narrative as self-appointed successor to Achilles and postpones the end in which Aeneas completes his role as Achilles by duelling Turnus and avenging Pallas. Although Turnus’ death is still fixed as the epic’s (Homeric) climax, we have suddenly left the clear and straight Iliadic path towards this climax with his temporary removal.22 The indefinite delay of the inevitable end creates suspense and a void for others (Mezentius, Camilla) to fill. The reader, consequently, is shaken out of the complacent 21 22

Cf. Williams 1973: 490. Naturally, the temporary removal of the main antagonist is itself a Homeric motif (cf. Il. 20.443 ff.), but the Vergilian narrative no longer follows the same order of events as the Iliadic precedent.

384

van der keur

interpretation that the earlier scenes had invited—of reading Vergil’s narrative solely along the familiar Iliadic pattern.

3

The Structure of Storms

In this second section we will look at the function of similes for Aeneid 10 as a whole. The book’s similes are important in highlighting key themes and varying the pace of the narrative, as Harrison has shown in the previous chapter. However, they also serve as a unifying factor in a book characterized by variety—in fact, they reflect the movement of the narrative and, through their imagery, they create a sub-narrative that helps in identifying the structure of the book.23 This unity is established through the common theme shared by many of the similes in Aeneid 10, referring to powerful natural forces, particularly storms. For this imagery, an important intertext is Vergil’s own Georgics and more specifically its storm description in 1.316 ff. and the subsequent passage (1.351ff.) on signs by which a coming storm might be known. The relevance of these two passages to the similes in Aeneid 10 has been well demonstrated by Briggs in his book on Vergil’s use of material from the Georgics in the similes of the Aeneid.24 The following argument will build upon his observations. The first part of Aeneid 10 builds up anticipation for the imminent battle. Towards the end of the divine council, when both Venus and Juno have said their piece, the book’s first simile sets the tone. The gods’ murmuring of agreement for either point of view is compared to a breeze rustling through the forest and unseen noises warning sailors of the winds to come (Aen. 10.97–10.99), recalling the first of the storm signs in the Georgics, described at 1.356–1.359.25 The simile spills into the narrative, for Jupiter ends the discussion with his voice that silences heaven; the winds subside and the sea is calmed.26 On the divine level, the storm dies down before it has begun. But as Jupiter’s speech makes 23 24 25

26

For Homer, the thematic use of similes and their interaction with the narrative is discussed in Moulton 1977. Briggs 1980: 81–91, especially 88–91. Aen. 10.97–10.99 ceu flamina prima / cum deprensa fremunt siluis et caeca uolutant / murmura uenturos nautis prodentia uentos, ‘just as when the first breezes roar when trapped in the woods, and send unseen rumblings rolling out, betraying to sailors the winds to come’ and G. 1.356–1.359 uentis surgentibus … [incipit] nemorum increbrescere murmur, ‘when the winds are rising, […] the woodland murmur begins to wax loud’. Aen. 10.103 tum Zephyri posuere, premit placida aequora pontus, ‘then the Zephyrs have come to rest, and the sea smoothes its waters to calm’.

parallel plotlines

385

clear, he now allows the war that he had opposed at the opening of the council; the storm on the human level is anything but averted. Aeneid 10 is not the first book to echo the storm signs of Georgics 1.351 ff. Before we study the further development of the ‘storm’ motif in the rest of book 10, it is useful to look briefly at its use in book 7, where the very beginning of the conflict had also been portrayed in such terms. As the Latin farmers are grimly drawn up against the Trojans to avenge the stag killed by Ascanius, the escalating situation is compared to a sea at the beginning of a storm—at the first wind, the surface starts to whiten with foam, and soon the billowing waves reach the sky. The simile recalls the other half of the same Georgic intertext discussed above, the first sign heralding a storm.27 Soon after, we find another simile in the catalogue of the Latin troops, where Messapus’ chanting soldiers are compared to swans: ceu quondam niuei liquida inter nubila cycni cum sese e pastu referunt et longa canoros dant per colla modos, sonat amnis et Asia longe pulsa palus. nec quisquam aeratas acies ex agmine tanto misceri putet, aeriam sed gurgite ab alto 705 urgeri uolucrum raucarum ad litora nubem. 700

Verg. Aen. 7.699–7.705

[…] as snowy swans among the moist clouds, when they return from feeding, and from their long throats utter their tuneful strains; the river and the reverberating Asian meadow echo far and wide. Nor would one think that mail-clad ranks were massed from that great column, but that high in the air a cloud of hoarse-voiced birds was pressing shoreward from the deep gulf. Commentators have stressed the models in heroic epic for this double simile;28 they have also argued that these two similes in close succession in fact

27

28

Aen. 7.528 fluctus uti primo coepit cum albescere uento (‘as when a billow begins to whiten under the wind’s first breath’) ~ G. 1.356–1.357 uentis surgentibus aut freta ponti incipiunt agitata tumescere (‘when the winds are rising, either the sea’s straits begin to heave and swell’). For the reading uento (preferable to the variant ponto), see Horsfall 2000: 348. For the place of the simile in the water imagery in Aeneid 7, see Harrison 1985: 99–102. 7.699–7.702: Hom. Il. 2.459–2.463, A.R. 4.1300–4.1302; 7.703–7.705: Hom. Il. 17.755–17.756, A.R. 4.238 ff.

386

van der keur

represent alternative drafts by the poet.29 The setting of the first simile, Asia … palus, recalls Il. 2.459–2.463 (just prior to the catalogue of the Greek ships), where the Greek army is compared to swans and other birds alighting on the meadows around the river Cayster. Homer’s passage stood model for the simile in Aeneid 7; but it is relevant that Vergil had earlier used the same image in Georgics 1. Seeing the birds quae Asia circum dulcibus in stagnis rimantur prata Caystri (‘such as, in Cayster’s sweet pools, rummage round about the Asian meadows’, G. 1.383–1.384) play in the water is a sure sign of a coming storm. It is worth noting that the double simile in Aeneid 7 contains echoes of other storm signs, too.30 The swans return from feeding, like the host of rooks in G. 1.381–1.382 e pastu decedens agmine magno / coruorum increpuit densis exercitus alis, ‘an army of rooks, quitting their pasture in long array, clang with serried wings’; not only does the phrase e pastu return, but agmine magno may also resonate in ex agmine tanto (notably outside the simile, of the Italians).31 Messapus’ men are also like a cloud of birds returning from the high sea to the shores (ad litora), which picks up G. 1.361–1.362 cum medio celeres reuolant ex aequore mergi / clamoremque ferunt ad litora, ‘when the fleet gulls fly back from mid-ocean, wafting their screams shoreward’; we find the same phrase ad litora, medio … ex aequore is picked up by gurgite ab alto and the birds’ cries suggested by clamorem resonate in raucarum. It will be obvious that, for the Georgic intertext, it does not matter whether these two similes in Aeneid 7 are alternative drafts or not; regardless which of the two similes might have made the final cut, the poet would have included an echo of Georgics 1. Alternatively, if the two were meant to go together, the intertext is even clearer. The Georgic storm signs serve as an intermediary between the (simile in the) Homeric troop catalogue and the Vergilian one. War is like a storm, and it is coming. It is not surprising that this imagery is employed again precisely in Aeneid 10. While the war was already raging in book 9, the arrival of Aeneas and his Arcadian and Etruscan allies (with their own catalogue of troops) promises a renewed clash on a grander scale. The first simile used at Aeneas’ return is that of the Trojans raising a cry like cranes fleeing before the winds (10.262–10.266); above, we have briefly looked at the Iliadic intertext (Il. 3.3–3.7). Whereas Homer focuses on the cranes’ migration to the south, where they will wage war upon the Pygmies, Vergil has privileged the first part: the birds’ flight before the

29 30 31

Fordyce 1977: 187; and Horsfall 2000: 457. The echoes of Georgics 1 are noted, but not commented upon, by Horsfall. Several editors read examine here; for ex agmine, see Horsfall 2000: 461.

parallel plotlines

387

storm.32 Here, I would like to focus not on what is left out, but rather on the now far more prominent connection between the cranes with their triumphant cry and the storm from which they have escaped. The passage from the Georgics once again serves as an intermediary text between the Iliad and the Aeneid: G. 1.374–1.375 [imbrum] surgentem uallibus imis aeriae fugere grues, ‘as [rain] gathers, the cranes flee before it high in the sky from the valleys’ depths’. The Georgic cranes are aeriae, echoing Homer’s ἠέριαι (Il. 3.7).33 In Aeneid 10, the first thing that is said of the cranes is that they dant signa, a phrase that seems deliberately ambiguous. Scholars have drawn attention to the martial connotations of the phrase, ‘give the signal for battle’, which creates another connection between the cranes and the Trojans and anticipates the battle that is to come.34 But in the light of the intertexts under discussion, arguably, the meaning ‘give (weather) signs’ is present as well.35 Two similes in Aeneid 10 have heralded the storm that is to come—one on the divine level towards the end of the Olympian debate, the other on the human level at the arrival of Aeneas and his army.36 Once the armies of the Trojans and Italians are locked in combat, the imagery reaches its logical climax, as the two armies are compared to battling winds: magno discordes aethere uenti proelia ceu tollunt animis et uiribus aequis Verg. Aen. 10.356–10.357

[…] as when warring winds in the great heaven raise battle with even spirit and strength

32 33

34 35 36

For a recent discussion of Vergil’s transformation of Homer’s crane simile, see Manoralaki 2012. Vergil adopted the sound, but not the sense, of ἠέριαι (‘early in the morning’). For this difference in meaning between the Homeric and the Vergilian epithets, see Conington ad G. 1.375. Harrison 1991: 144; Manolaraki 2012: 294. The storm signs are introduced as certis … signis in G. 1.351. In the lines immediately following the crane simile, the flash of Aeneas’ head and the accompanying comet simile (discussed in the first part of this chapter) may be interpreted as a third reference to the Georgic storm signs. In G. 1.365–1.367, on shooting stars (admittedly not the same as comets), we find comparable phrasing: the flames emanating from the head / comet (G. 1.367 flammarum longos a tergo … tractus ~ Aen. 10.270 tristisque a uertice flamma) are widely visible (G. 1.271 uastos … ignis ~ Aen. 10.367 flammarum longos … tractus); the white colour of the Georgic stars (1.367 albescere) is replaced with the more ominous red (Aen. 10.273 rubent).

388

van der keur

With this simile, Vergil picks up the opening line of his storm description in G. 1.318ff.: omnia uentorum concurrere proelia uidi. The echo is the culmination of the various storm signs. Both armies are storms; their anticipated clash is now realized. The discord of the winds also responds to the theme of discordia at the opening of the book (Aen. 10.9 quae contra uetitum discordia) with which the divine council also ended (Aen. 10.106 nec uestra capit discordia finem).37 The simile thus forms a fitting conclusion for the first part of Aeneid 10. The winds have finally clashed, for the time being resulting in a stale-mate. Scholars have noted that the battle narrative in the book is characterized by balance and symmetry.38 In the beginning of the battle, the forces are evenly matched, as Aeneas’ aristeia is countered by the valour of Clausus, Halaesus and Messapus (10.310–10.356); it is this balance that is emphasized by the above-mentioned simile of clashing winds. The equilibrium is continued when Pallas and Lausus lead their forces against each other (10.431 ducibusque et uiribus aequis, which has the same line ending as 10.357), and the narrator raises the possibility of an equally matched fight between these two youths; but Turnus disrupts the equilibrium when he replaces Lausus with himself and proceeds to kill Pallas in an ill-matched duel (10.459 uiribus imparibus). In the remainder of the book, there is still balance, but on a different level. Rather than being a contest of equally poised forces, the tide of battle heavily fluctuates, favouring now one side, now the other. Rutulian victories are counterbalanced by successes for the Trojans; the duel between Pallas and Turnus contrasts with the one between Lausus and Aeneas; Mezentius’ aristeia serves as a response to Aeneas’ heroics. Only at the end is Trojan martial superiority shown when Aeneas kills Mezentius in the book’s climax. These alternating successes are reflected in the storm imagery, and particularly in the representation of Aeneas and Mezentius. No longer are both sides portrayed as storms at the same time (as in the simile of 10.356–10.357); rather, sometimes it is one side that is represented as a storm and the other as bearing its brunt, sometimes it is vice versa. The last part of this chapter will discuss this new development in the storm motif in the second half of Aeneid 10. Following Pallas’ death, Aeneas rages across the battlefield in red-hazed fury. Towards the end of this scene, he is compared to a torrent or whirlwind (10.603–10.604 torrentis aquae uel turbinis atri / more furens). Aeneas himself has become the Georgic storm (G. 1.320 turbine nigro). Immediately after this 37 38

The civil war aspects inherent to discordia are, of course, also a Georgic theme; I will leave that out of the discussion here, however. For the structure and symmetrical balance of Aeneid 10, see for instance Harrison 1991: xxvi–xxvii and Barchiesi 1984: 55–73.

parallel plotlines

389

simile, we learn how his rage has helped his allies: the besieged Trojans finally sally forth from their camp (10.604–10.605). This passage interacts in an interesting way with the earlier crane simile at Aeneas’ arrival. The Trojans, saved by the advent of Aeneas, cheered like jubilant cranes that have escaped a storm; at the same time, however, the crane simile served to anticipate the coming storm, the storm that now allows their escape. Less than a hundred lines later, when Juno has removed Turnus from the field, Mezentius takes the Rutulian leader’s place—and his aristeia contains no less than four similes, due to the need for rapid characterization.39 These similes interact with each other, but also with the other similes in Aeneid 10. The first one is found at the very beginning of the scene, when Mezentius, beset by his Etruscan enemies, is compared to a cliff weathering the fury of the winds and the sea (10.693–10.696). This image of passive endurance is developed into defensive aggression in the second simile, which pictures him as a furious boar keeping the hunters at bay with his menacing ferocity (10.707– 10.718). When Mezentius has gone on the offensive, a third simile likens him to a lion assaulting a deer—another wild animal, but now with the roles of prey and predator reversed (10.723–10.728). With Mezentius’ aristeia, the balance of power is restored as the battle returns to a stalemate, and both parties now suffer in equal measure.40 The hero is presented as a force rivalling Aeneas. The fourth and last simile, which concludes the aristeia, suggests as much through its imagery. Mezentius is compared to Orion, the mythical giant, who wades through the deepest seas, or carries a full-grown ash tree from the mountains with his head amidst the clouds (10.763–10.767). This well-chosen simile interacts with many of the previous ones to suggest Mezentius’ development into being Aeneas’ main adversary. The simile concludes Mezentius’ aristeia, just as the description of Aeneas’ battle rage ended with one in 10.603–10.604.41 Harrison has already observed how the simile plays against Aeneas’ portrayal as Aegaeon in 10.565–10.568; their enemies view both heroes as intimidating

39 40

41

See in this volume Harrison, 362. 10.755–10.757 iam grauis aequabat luctus et mutua Mauors / funera; caedebant pariter pariterque ruebant / uictores uictique, neque his fuga nota neque illis, ‘And now fearsome Mars was evening out the grief and deaths on each side; they slew and fell in equal measure, victors and victims in turn, and flight was unknown to either one side or the other.’ Both similes are similarly introduced, with a reference to the hero moving across the battlefield; cf. 10.602–10.603 talia per campos edebat funera ductor / Dardanius (‘Such were the deaths the Trojan leader dealt out over the plains’) and 10.762–10.763 at uero ingentem quatiens Mezentius hastam / turbidus ingreditur campo (‘But now Mezentius, brandishing his mighty spear, strode stormily across the plain’).

390

van der keur

giants.42 The comparison with Orion also shows Mezentius’ progress since the beginning of the scene. Firstly, he has gone from being a hunted boar to impersonating the great hunter of myth. Secondly, while the first simile showed him as a cliff resisting the sea and storms, Orion ventures into the ocean and rears his head in the sky—conquering the elements rather than enduring them. However, the name Orion also calls forth associations other than the mythical giant. Guided by the imagery of the other similes in Aeneid 10, and by the four other occurrences of the name Orion in the epic, we may also think of the constellation Orion, which is, notably, the harbinger of the stormy season.43 When we allow for this association, Mezentius has gone from being a cliff resisting the storm to being the storm itself. Moreover, the constellation Orion would interact with the simile in the beginning of Aeneid 10, in which the appearance of Aeneas is compared to Sirius—Orion’s dog.44 The simile thus casts Mezentius as a worthy opponent fit to counterbalance Aeneas in every respect. We have seen that Aeneas’ victories, with which he allowed his side to gain the upper hand and freed his besieged compatriots, were concluded with the image of the hero as a whirlwind. Similarly, when Mezentius’ heroism puts the Latins back in a more favourable position, this is reflected in the similes’ imagery, which sees him evolve from enduring the storm to being a storm himself. The last simile of Aeneid 10 fits perfectly in this trend of images that mirror the flow of battle, as we witness Aeneas in the same position Mezentius found himself in at the beginning of his aristeia, withstanding a hail of enemy missiles:45 ac uelut effusa si quando grandine nimbi praecipitant, omnis campis diffugit arator 805 omnis et agricola, et tuta latet arce uiator aut amnis ripis aut alti fornice saxi, dum pluit in terris, ut possint sole reducto

42 43

44 45

See in this volume Harrison, 370. Cf. perhaps 10.767 caput inter nubila condit. The other four instances are 1.535, 3.517, 4.52 and 7.719, which all refer to the constellation; the first and third of these refer specifically to the storm which Aeneas had to endure in Aeneid 1. In the line introducing the simile, Mezentius’ epithet turbidus (translated by Harrison as ‘stormily’) also seems to guide the reader towards making the connection between Orion and storms (see OLD turbidus 1a and 1b for the use of the adjective with storms and constellations associated with stormy weather). In the model for Vergil’s Sirius simile, the star is also identified as such; cf. Hom. Il. 22.29 κύν’ Ὠρίωνος. Aen. 10.691–10.692 ~ 10.801–10.802.

parallel plotlines

810

391

exercere diem: sic obrutus undique telis Aeneas nubem belli, dum detonet omnis, sustinet […]. Verg. Aen. 10.803–10.810

And as when the storm-clouds plunge down in a shower of hail, every ploughman has fled from the fields and every farmer, and the traveller lies sheltered in a safe refuge, under the banks of a river or in the vault of a lofty rock, while it rains over the earth, so that they can occupy the day when the sun returns, just so Aeneas, inundated with missiles from all sides, weathered the storm-cloud of war until it should cease to thunder […]. Like Mezentius before him, Aeneas himself now has to endure a storm. However, there are important differences with the earlier simile. The men figuring in the simile are not like a cliff buffeted by the elements, but have taken shelter from the foul weather. They know that, eventually, the storm will pass and they will be able to return to their activities, just as Aeneas waits until the tide of battle will swing in his favour again. The farmer fleeing from his fields is a fitting Georgic image to conclude the storm imagery in this book and to bring out the contrast with Mezentius. Unlike a passive rock, the farmer and the traveller are able to anticipate a future change of weather; they know how to deal with storms and bide their time. Aeneas has not lost his storm qualities, either (10.802 furit Aeneas tectusque tenet se).46 He will defeat both Lausus and Mezentius shortly after the simile; but a more definite response to the imagery here will follow in book 12. There, Aeneas has transformed into the storm again. In the simile adorning his return to the battlefield (12.451–12.455), it is now the Latins who are the farmers; but they will not simply weather the storm. Instead, they fearfully recognize the signs and foresee the destruction that this storm will cause, just as the Latins realize Turnus’ impending defeat.47 The final storm image of the epic describes Aeneas’ spear as it travels through the air to bring doom to Turnus: 12.923 uolat atri turbinis instar, a close echo of the simile in 10.603–10.604 turbinis atri more furens. Immediately after the latter simile, Juno had intervened to rescue Turnus from the battlefield; with the use of the same image, the duel in Aeneid 12, which results in his death, is presented as 46

47

Briggs 1980: 90 notes ‘like the ploughman in the Georgics (1.259–1.267) who hones his blade indoors, Aeneas wisely prepares for battle by playing on Lausus’ vulnerability to furor [i.e. by taunting him], rather than being victimized by his own.’ See for a fuller discussion Briggs 1980: 91.

392

van der keur

the direct continuation of Aeneas’ battle rage in book 10. At the time, Turnus thought that he had the victor’s role to play in his own version of Iliad 22; but Juno clearly read the similes’ imagery far better than he did.

4

Conclusion

This chapter has discussed how the similes in Aeneid 10 contribute to two distinct textual strategies. The similes function not merely locally, but also serve a role in the larger narrative of the book. One such role is to set up an intertextual framework within which the rest of the narrative may be interpreted. As an example, we have studied the similes in each of the three scenes in which Turnus figures. In the first and last of these, the similes evoke Iliad 22. At one level, this intertext raises the expectation of a swift conclusion; but it also draws attention to Turnus’ misinterpretation of his own role in the epic. Focalization is important here: the informed reader perceives the conflict between Turnus’ bearing and epic reality. A second function of the similes that may be observed in Aeneid 10 is that they reflect the flow of the narrative. Storm signs abound in the similes in the first part of the book, as it builds towards the great clash between the two armies, presented as warring winds. The balance of forces is broken when Turnus kills Pallas; the remaining storm similes describe the alternating successes of both sides, with the defenders weathering the tempest unleashed by the aggressors. The imagery helps in setting up Mezentius as a worthy opponent of Aeneas; the last simile draws out Aeneas’ superiority, however, as a warrior who is prepared to bide his time. Therewith, it looks ahead at his future victories, both in this book and the conclusion of the epic.

Bibliography Anderson, W.S., ‘Vergil’s Second Iliad’, Transactions and Proceedings of the American Philological Association 88 (1957) 17–30. Barchiesi, A., La traccia del modello: effetti omerici nella narrazione virgiliana (Pisa 1984). Binder, G., Aeneas und Augustus. Interpretationen zum 8. Buch der Aeneis (Meisenheim 1971). Brackert, H., ‘Zu einigen Gleichnissen in Vergils Aeneis’, Euphorion 56 (1962) 165–173. Briggs, W.W., Narrative and simile from the Georgics in the Aeneid (Leiden 1980). Conington, J., The Works of Virgil, vol. 1: Eclogues and Georgics (London 1898). Fordyce, C.J., P. Vergili Maronis Aeneidos Libri VII–VIII (Oxford 1977).

parallel plotlines

393

Gransden, K.W., Virgil’s Iliad. An essay on epic narrative (Cambridge 1984). Harrison, S.J., ‘Vergilian similes: some connections’, Papers of the Liverpool Latin Seminar V (Liverpool 1985) 99–107. Harrison, S.J., Vergil Aeneid 10 (Oxford 1991). Hornsby, R.A., Patterns of action in the ‘Aeneid’: an interpretation of Vergil’s epic similes (Iowa City 1970). Horsfall, N.M., Virgil, Aeneid 7: a commentary (Leiden 2007). Manolaraki, E., ‘‘Aeriae grues’: crane migrations from Virgil to Statius’, CJ 107.3 (2012) 290–311. Moulton, C., Similes in the Homeric Poems (Göttingen 1977). Williams, R.D., The Aeneid of Virgil: Books 7–12 (London 1973).

Index Aeschylus Persians 391–405

58

Appian Hannibalic War 17.73–18.82 17.77–18.78 21.92–26.114

160 161 168

Coelius Antipater FRHist15F22=GELL.10.24.6 179 Diodorus of Sicily 11.4 11.5 11.6–11.8 11.6 11.7 11.8 11.9–11.10 11.9 11.10 11.11 11.18 11.20 11.37

19, 29, 30, 50, 54, 56, 57, 68, 69, 327 42, 56, 60–61, 69 34 32, 54, 56, 57, 69–70, 73, 323 32, 54, 74–76 32, 40, 49, 56, 67 38, 54–55, 328 24, 38, 69 79, 325 38, 45, 69, 326–327 58 71 55

7.102 7.103 7.131 7.139 7.150 7.161 7.173–7.178 7.175 7.176 7.177 7.184 7.188–7.193 7.196–7.201 7.196 7.197 7.198 7.199–7.201 7.199 7.200 7.201–7.209 7.201 7.202–7.203 7.202 7.203–7.207 7.203 7.204–7.205 7.204 7.205

Dionysius of Halicarnassus Letter to Pompey 3 114

7.206–7.207 7.206 7.207

Herodotus 1.53 1.55 1.69 1.85 3.55 5.24 5.48 6.106–6.107 7.61–7.62 7.101–7.105

7.208 7.208 7.209 7.210–7.212 7.210 7.211 7.212 7.213–7.218 7.213

56 124 56 124 31 56 100 62 34 92–93

7.214

30 30 27 19, 61, 333 56 44 23 115–116 35, 65, 68–69, 116–118, 128, 319 27 65, 93 49 61–63 63 63, 68 63, 65, 128 42, 63 42 42 93–94 42, 68 54 28, 30, 67 22 67, 68–69, 138 118–119 57, 68, 99, 105 30, 43, 68, 78–79, 99– 100, 105, 124, 329, 335 108 68 25, 39, 48–49, 69, 73, 125, 239 103, 138 33, 36, 103 54, 125 34, 94 33, 49, 54, 65 33, 36, 98, 239 33, 107, 239 49 56, 95, 98, 120, 139, 344 64, 65

396 Herodotus (cont.) 7.215 7.216 7.217 7.218 7.219–7.222 7.219–7.220 7.219 7.220–7.221 7.220 7.221 7.222 7.223 7.224–7.225 7.224 7.225

7.226 7.227 7.228 7.229–7.230 7.229–7.233 7.229 7.231–7.232 7.231 7.232 7.233 7.234 7.236 7.238 7.239 8.1 8.3–8.5 8.4–8.5 8.5 8.24 8.25 8.26 8.31–8.33 8.31 8.33 8.40 8.43 8.50 8.51–8.56

index

56, 69, 106, 117–118, 319 128 50 35, 95, 107 54 25, 65 39, 40, 95, 107 40 23, 64, 104–105, 121– 127 64, 78–79, 96 42, 43, 68, 96, 139, 329 36, 40–1, 54–55, 96, 98 150–151 30, 31, 41, 65, 68, 98 35, 41, 57, 78, 96–97, 106, 128, 239–240, 325, 327, 330 35, 56–66, 323 42, 66 30, 31, 45, 56, 69, 324 65, 328 328 30, 36, 64, 328 109–110 329 30, 65, 107 43, 68, 69, 98, 329, 335–336 30 73–74, 335 64 23–24, 65, 100–101 28 44 48–49 43 49 31, 36, 49 22, 26 43 50 50 22 28 43, 50 43

8.58–8.64 8.71–8.72 8.84 8.86 8.94 8.114 8.140 9.10 9.27 9.28 9.29 9.35 9.61–9.63 9.62 9.64 9.71 9.79

44 22, 28 58 319 43 125–126 56 22, 36 319 36 36 44 35 34 32, 126, 334 70 127, 334

Homer Iliad 2.459–463 3.3–7 4.141–145 5.627–662 16.487–489 16.563–685 16.1–762 22.26–31 22.189–192 22.199–201 22.205–207 22.297–305 Odyssey 6.232–235

386 377 364 319 381 57 57 368 383 382 379 57

Isocrates Archidamas 6.99–100 Panegyricus 4.90

29

364

29

Justin 2.11 2.15

30, 32, 38 38

Livy Praef. 10 7.34–36 21.1–2

274 327 195–197

397

index 21.1 21.4 21.53 21.54 22.4–7 22.4 22.7 22.12 22.19–22.21 22.19 22.22 22.23–22.30 22.23 22.25–22.26 22.31 22.32 22.33–22.34 22.33 22.34 22.35 22.36 22.37 22.38–22.39 22.38–22.40 22.38–43 22.38

22.39

22.40–22.42 22.40

22.41–22.42 22.41–22.43 22.41

22.42 22.43

193, 195–197, 320 180, 236 238 239 198 241–242 171 164, 207 198 205 193, 198–204, 206, 229 163 204 163 164 210 163 209, 212 163, 164, 208, 212–213, 215, 336 165, 208, 213–215, 336 160, 208, 216, 337 171, 182, 208, 216 164 216–217 159 162, 164–166, 207, 217, 275–276, 283, 289, 337 162, 164–166, 216– 217, 227, 253, 267, 270, 275–278, 281, 283–284, 308, 324, 337–338 217 161–162, 167, 207, 217– 218, 222, 256, 278, 281–282, 285, 301–302, 307 162, 166 180 160, 164–166, 180, 218, 220, 241–242, 268–269, 275–276, 283–286, 299, 308, 337 162, 165–166, 220–221, 275, 285, 299, 305 161–162, 164–166, 169,

22.44

22.45–22.47 22.45–22.50 22.45 22.46–22.49 22.46 22.47–22.49 22.47

22.48 22.49

22.50–22.54 22.50–22.61 22.50 22.51 22.52 22.53 22.54–22.55 22.54 22.55–22.56 22.55 22.56–22.57 22.56 22.58 22.59–22.60 22.59 22.60 22.61 23.21 24.11 24.43 27.49 30.20 30.29

181, 207, 256–257, 261, 265–266, 273, 275, 287–289–290 164–166, 169, 170, 207, 222, 257–258, 275, 286, 288, 305, 337 166, 168 159 165, 168, 170, 288, 290 223–224 167–169, 223–224, 263 209 166–167, 170, 172, 207, 223, 246, 264, 275, 306 168–170, 181, 223–224, 248, 255 164–165, 167–170, 172, 177, 224, 229, 256, 280, 280–281, 307–308, 322–323, 325 225 170 159, 170, 172–177, 224– 225, 255, 273, 326, 328 172, 178–179, 180, 225, 310, 330, 332 170, 172–173, 179, 226, 326 182, 226, 335 172 171, 181, 226 164 179 172 180 180, 181 170 174, 227 175–177, 227, 326, 328 177, 181, 209, 227, 328– 329 167 167 167 334 333 298

398

index

Livy (cont.) 30.30 30.32 30.33 30.34 30.35

334 303, 305 300 304, 307 309

Pausanias 3.14.1 10.20.2

31 30

Plutarch Fab. 18 Mor. 347a 864c–865f 864ef 866a 866b 866d–867b Polybius 3.106 3.107 3.108–3.109 3.108 3.110 3.111 3.112 3.113 3.114 3.115 3.116 3.117 6.58 Silius 10.605–10.639

182 132 43 38 37–38 20 43

160 160 160, 336 337 161, 168–169 160 161 167 167 169 161–162, 167–169, 278, 325 169 329

182

Simonides Fr. el. 10–17 (West) 45 Fr. 531 (Page) 45 Thucydides 1.1 1.73 1.101

320 320 31

2.2–2.4 2.2 4.26–4.41 4.27 4.28 4.29 4.30 4.32 4.34 4.36 4.39 4.40 4.55 5.54 5.75–5.76 7.69 Vergil Aeneid 1.589–593 2.682–684 3.655–661 7.699–705 8.678–681 8.680 10.9 10.97–99 10.106 10.134–138 10.132 10.261–266 10.262–266 10.270–275 10.272–277 10.310–356 10.324 10.356–357 10.431 10.439–509 10.454–456 10.565–568 10.565–570 10.603–604 10.604–605 10.633–688 10.693–696 10.693–770 10.707–718

336 43 342 345–347, 355 345, 348, 355 345 349–350 143–150, 345, 352 141–143, 239, 350, 356 138, 140, 320, 342–345, 353 344–345, 353, 356 36, 342, 354–356 333 27 27 320

364, 366 367 372 385 368 377 388 384 388 364 365 377 386 366 377–378 388 365 387 388 379–381 369 389 370 388 389 381–383 389 371–372 389

399

index 10.723–728 10.760 10.763–767 10.803–810 10.861–866 12.451–455 12.749–755

389 365 389 391–391 372 391 383

Georgics 1.318 1.320 1.356–359 1.361–362 1.374–375 1.381–384

388 388 384 386 387 386