A Late Antique Poetics?: The Jeweled Style Revisited 9781350346406, 9781350346437, 9781350346413

Table of contents :
Cover page
Halftitle page
Series
Title page
Copyright page
CONTENTS
ILLUSTRATIONS
CONTRIBUTORS
SERIES EDITOR PREFACE
PREFACE AND ACKNOWLEDGEMENTS
NOTE ON TEXTS AND TRANSLATIONS
ABBREVIATIONS
INTRODUCTION
The formal features of the jeweled style
The jeweled style and late antique aesthetics
PART I THE FORMAL FEATURES OF THE JEWELED STYLE
CHAPTER 1 THE DECADENT PREHISTORY OF THE JEWELED STYLE
Introduction: A chain of receptions
The language of jewels
The orientalism of the jeweled style
The aesthetics of decadence
From decadence to diversity
Conclusion: Towards a brighter future
CHAPTER 2 THE GREEK JEWELED STYLE
A Greek jeweled style in the third century?
The Greek jeweled style in the fourth and fifth centuries
Jeweled Christian prose and un-jeweled biblical poetry
CHAPTER 3 GILDING THE LILY: THE JEWELED STYLE IN PROSE PANEGYRIC
Non-metrical word patterning
Catalogues
Quotations and sententiae
Ecphrasis
Conclusion
CHAPTER 4 LEARNING THE JEWELED STYLE
Introduction
Schools in late antiquity
Classroom poetics in late antiquity
Polychromatic style
Variatio of vocabulary and theme
Description and structure
Conclusion
CHAPTER 5 QUANTITATIVE APPROACHES TO LATE ANTIQUE POETICS: ENUMERATION AND CONGERIES
Introduction
Methods and context
Methodological limitations and restrictions
Interpretation of findings
Conclusion
CHAPTER 6 THE JEWELED STYLE AND SILVER LATIN SCHOLARSHIP
Reclaiming the wood from the trees: Developments in scholarship on Silver Latin tree catalogues
The danger of discouraging contextual significance
The danger of overlooking the dynamics
Conclusion
CHAPTER 7 THE JEWELED STYLE IN EARLY MEDIEVAL LATIN POETRY
Aldhelm
Hisperica famina
Hucbald, Egloga de calvis
Conclusion
CHAPTER 8 DIGRESSION, VARIETY AND UNITY IN LATE LATIN POETRY
Digressions in ancient literary criticism and rhetoric
Variety and unity
Digressive rivers
Conclusion
PART II THE JEWELED STYLE AND LATE ANTIQUE AESTHETICS
CHAPTER 9 METAPHOR SQUARED
Arator, Epistula ad Florianum
Ennodius, Carmen 1.9
Merobaudes, Carmen 4
Merobaudes, Panegyricus poeticus
Metaphor squared as a literary technique
Conclusion
CHAPTER 10 AN ‘UNJEWELED’ CHRISTIAN STYLE? A LOOK AT AUGUSTINE’S CONFESSIONS
The jeweled style as a response to late antique ‘logoclasm’
Christian objections to the jeweled style
In quest of a genuinely Christian poetry
Augustine’s Confessions: A new poetic paradigm?
CHAPTER 11 THE CENTO AND SCRIPTURE: AN EARLY CHRISTIAN DEBATE OVER THE POETICS OF EXEGESIS
Competition and caution
Decoding Jerome: Ambrose as scriptural centonist
The poet-preacher Ambrose
De obitu Valentiniani as cento
Conclusion
CHAPTER 12 JEWELED SEA STORM DESCRIPTIONS IN ZENO OF VERONA (AND JUVENCUS)
The genera of Latin homiletics and the jeweled style
Sea storms in Zeno of Verona
Sea storms in biblical poetry: the case of Juvencus
Conclusion: Towards an assessment of jeweled Latin homiletics?
CHAPTER 13 ALLUSIVE CLUSTERS AND BIBLICAL CONFIGURATIONS IN DRACONTIUS’ DELAUDIBUS DEI: A CHRISTIAN JEWELED STYLE?
Stylistic patterns: Same techniques, new development
Clusters of allusions
A jeweled structure
The influence of patristic and biblical texts
Conclusion: A Christian jeweled style?
CHAPTER 14 VERGIL’S CHILDREN: PATTERNS IN CHRISTIAN CENTOS AND RESPONSES TO VERGIL’S FOURTH ECLOGUE
Cento de Ecclesia
Cento de Verbi Incarnatione
Conclusion
CHAPTER 15 ARCHITECTURAL ECPHRASIS IN VENANTIUS FORTUNATUS: BEYOND THE JEWELED STYLE
The villa and the church
Restorations
Light in late antique ecphrases of churches
Beyond the jeweled style
Conclusion
CHAPTER 16 THE JEWELED STYLE IN LATE ANTIQUE LATIN EPIGRAM
Introduction
Tessellated epigrams: Ennodius’ jeweled bishops and Christian ecphrastic epigram
Tessellation or exhaustion? Ausonian diadems, Claudian crystals and secular epigrammatic series
Conclusion
CHAPTER 17 THE JEWELED STYLE AND NEOPLATONISM
Introduction
The ‘unity’ of Macrobius’ Saturnalia
The number seven in Neoplatonism
An epistolary microcosmos: The letter collection of Q. Aurelius Symmachus
Conclusion
EPILOGUE: THE JEWELED STYLE IN CONTEXT
REFERENCES
INDEX RERUM
INDEX NOMINUM
INDEX LOCORUM

Citation preview

A LATE ANTIQUE POETICS?

i

sera tela: Studies in Late Antique Literature and Its Reception Series Editor: Marco Formisano, Ghent University, Belgium This new series challenges existing paradigms of subordinating late-antique texts to classical literature: instead, it places late-antique literature at the centre of inquiry, and its title plays on the ambiguity of tela, evoking both the drawing of weapons and the building of a web. Late antique literature becomes a vantage point for looking both forward to its reception and transformation in subsequent ages, and backward, by engaging in the new and exciting experiment of viewing classical literature through the lens of late antiquity. Each volume is animated by interpretive enthusiasm and experimentalism aiming at highlighting what is new, unique, and unprecedented, rather than focussing on relationships with the classical past. Series Board: Virginia Burrus, Syracuse University, USA Simon Goldhill, University of Cambridge, UK Jesús Hernández Lobato, University of Salamanca, Spain Ralph J. Hexter, University of California, Davis, USA Ingela Nilsson, Uppsala University, Sweden Paolo Felice Sacchi, Ghent University, Belgium Cristiana Sogno, Fordham University, USA James Uden, Boston University, USA Also available in the series: Epitomic Writing in Late Antiquity and Beyond, edited by Paolo Felice Sacchi and Marco Formisano

ii

A LATE ANTIQUE POETICS? THE JEWELED STYLE REVISITED

Edited by Joshua Hartman and Helen Kaufmann

iii

BLOOMSBURY ACADEMIC Bloomsbury Publishing Plc 50 Bedford Square, London, WC1B 3DP, UK 1385 Broadway, New York, NY 10018, USA 29 Earlsfort Terrace, Dublin 2, Ireland BLOOMSBURY, BLOOMSBURY ACADEMIC and the Diana logo are trademarks of Bloomsbury Publishing Plc First published in Great Britain 2023 Copyright © Joshua Hartman, Helen Kaufmann and Contributors, 2023 Joshua Hartman and Helen Kaufmann have asserted their right under the Copyright, Designs and Patents Act, 1988, to be identified as Editors of this work. For legal purposes the Acknowledgements on p. xiii constitute an extension of this copyright page. Cover image: kampee patisena/Getty Series design: Graham Robert Ward All rights reserved. No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage or retrieval system, without prior permission in writing from the publishers. Bloomsbury Publishing Plc does not have any control over, or responsibility for, any third-party websites referred to or in this book. All internet addresses given in this book were correct at the time of going to press. The author and publisher regret any inconvenience caused if addresses have changed or sites have ceased to exist, but can accept no responsibility for any such changes. A catalogue record for this book is available from the British Library. A catalog record for this book is available from the Library of Congress. ISBN:

HB: 978-1-3503-4640-6 ePDF: 978-1-3503-4641-3 eBook: 978-1-3503-4642-0

Series: sera tela: Studies in Late Antique Literature and Its Reception Typeset by RefineCatch Limited, Bungay, Suffolk To find out more about our authors and books visit www.bloomsbury.com and sign up for our newsletters.

iv

CONTENTS

List of Illustrations Notes on Contributors Series Editor Preface Preface and Acknowledgements Note on Texts and Translations Abbreviations Introduction Joshua Hartman and Helen Kaufmann

vii viii xi xiii xiv xv 1

Part I The Formal Features of the Jeweled Style 1

The Decadent Prehistory of the Jeweled Style Ian Fielding

11

2

The Greek Jeweled Style Fotini Hadjittofi

25

3

Gilding the Lily: The Jeweled Style in Prose Panegyric Catherine Ware

45

4

Learning the Jeweled Style Frances Foster

61

5

Quantitative Approaches to Late Antique Poetics: Enumeration and Congeries Joshua Hartman and Jacob Levernier

75

6

The Jeweled Style and Silver Latin Scholarship Ruth Parkes

91

7

The Jeweled Style in Early Medieval Latin Poetry Cillian O’Hogan

103

8

Digression, Variety and Unity in (Late) Latin Poetry Helen Kaufmann

115

Part II The Jeweled Style and Late Antique Aesthetics 9

Metaphor Squared Christoph Schubert

131

10 An ‘Unjeweled’ Christian Style? A Look at Augustine’s Confessions Jesús Hernández Lobato

143

11 The Cento and Scripture: An Early Christian Debate over the Poetics of Exegesis David Ungvary

159

12 Jeweled Sea Storm Descriptions in Zeno of Verona (and Juvencus) Francesco Lubian

173

13 Allusive Clusters and Biblical Configurations in Dracontius’ De laudibus dei: A Christian Jeweled Style? Elena Castelnuovo

187 v

Contents

14 Vergil’s Children: Patterns in Christian Centos and Responses to Vergil’s Fourth Eclogue Scott McGill

201

15 Architectural Ecphrasis in Venantius Fortunatus: Beyond the Jeweled Style Carole E. Newlands

215

16 The Jeweled Style in Late Antique Latin Epigram Bret Mulligan

231

17 The Jeweled Style and Neoplatonism Andreas Abele

245

Epilogue: The Jeweled Style in Context Michael Roberts

259

References Index Rerum Index Nominum Index Locorum

269 301 305 307

vi

ILLUSTRATIONS

Figures 1.1

Gustave Moreau, Salomé Dancing before Herod (1876), Hammer Museum. Wikimedia Commons 1.2 Gustave Moreau, The Apparition (1876), Harvard Art Museums. Wikimedia Commons 5.1 Chronological distribution of highly enumerative lines 5.2 Chronological distribution of wholly enumerative lines 7.1 Opening of the Egloga de calvis, Trier, Bibliothek des Bischöflichen Priesterseminars, MS 100, f. 44v., tenth century 15.1 Dagger handle and clasps from the Treasure of Pouan, Musée Saint-Loup, Troyes. Getty Images

14 15 84 85 111 229

Tables 5.1

Authors, corpora and enumeration scores

79

vii

CONTRIBUTORS

Andreas Abele is a tenured lecturer in Latin Philology in the Department of Classics at the University of Tübingen, Germany. In his PhD thesis, ‘Jeremias Drexel SJ: Iulianus Apostata Tragoedia’ (2018), he edited, translated and commented on a neo-Latin Jesuit school play on Julian the Apostate. It was followed by an edition of Matthäus Rader’s Drama de Divo Cassiano (2021). Recent publications include narratological studies on late Latin hagiography. Currently he is preparing a book on the letter collection of Q. Aurelius Symmachus. Elena Castelnuovo completed her PhD in late Latin poetry at the University of Milan, Italy, in 2018 with a dissertation on the reception of Horatian Sapphic poems in Prudentius, Ausonius and in early medieval hymns. Her research interests focus on late Latin poetry and its interactions with history, liturgy and art. She has published on the reception of Horace in late Latin literature, Claudian and Prudentius and is currently extending her area of expertise to include the poets of the fifth and sixth centuries, in particular Dracontius and Venantius Fortunatus. Ian Fielding is Associate Professor of Classical Studies at the University of Michigan in Ann Arbor, USA. He is the author of Transformations of Ovid in Late Antiquity (2017), and of several other articles about Latin poetry and its reception. His current book project, provisionally entitled Roman Satire and the Fall of Rome, highlights the influence of the satirist Juvenal in Edward Gibbon’s Decline and Fall of the Roman Empire (1776–88). Frances Foster teaches at the Faculty of Education, University of Cambridge, UK. Her work straddles the reception of antiquity at two distinct moments: late antiquity and the present day. Previous publications on Servius have appeared in The Classical Quarterly, Language and History, and forthcoming are chapters in Ars et Commentarius (ed. Garcea and Vallat) and Sicut Commentatores Loquuntur (ed. Tischer, Kuhn-Treichel and Poletti). Fotini Hadjittofi teaches in the Classics Department at the University of Lisbon, Portugal, and specializes in the Greek literature of late antiquity. She has published on the epic poems of Nonnus of Panopolis and Quintus of Smyrna, and on the rhetors Himerius and Choricius. Her wider research interests include Hellenistic poetry, Ovid and gender studies in the ancient world. She is currently working on Achilles in late antiquity. Joshua Hartman is Visiting Assistant Professor of Classics at Bowdoin College, USA. His research focuses on the relationship between literature and memory, especially during late antiquity. He has published articles on Greek and Roman literature, Roman cultural memory and classical reception in Latin America. viii

Contributors

Jesús Hernández Lobato is Associate Professor of Latin Language and Literature at the University of Salamanca, Spain. An expert on Sidonius Apollinaris, he has published extensively on late antique aesthetics and its reception. Among the books he has published is Vel Apolline muto: estética y poética de la Antigüedad tardía (2012); he is also the co-editor with Jaś Elsner of The Poetics of Late Latin Literature (2017) and with Óscar Prieto Domínguez of Literature Squared: Self-Reflexivity in Late Antique Literature (2020). Helen Kaufmann is an independent Latinist based in Oxford, UK, with a job in data management. Her research focuses on the Latin poetry of late antiquity. She has published a commentary on Dracontius’ Medea (2006) as well as articles exploring various aspects of late Latin poetics and is currently working on a monograph on space, place and identity in late Latin poetry. Jacob Levernier is a data scientist whose dissertation research focused on extracting values from natural language texts. As CLIR Bollinger Fellow in Library Innovation at the University of Pennsylvania, USA, their output included research and invited lectures on automated network analyses of medieval maps. Their current research contributions primarily centre on computer vision in healthcare. Francesco Lubian is Associate Professor of Latin Language and Literature at the University of Padua, Italy. His research interests lie mainly in late antique Latin literature: he has published on biblical paraphrase (Juvencus, the so-called poem of the Heptateuchos), Ambrose, Prudentius and the tradition of Latin epigram from Ausonius to the Carolingian age. He is preparing a commentary on Jerome’s Epistle 107 and an edition of the book of Numbers of the poem of the Heptateuchos. Scott McGill is Deedee McMurtry Professor in the Humanities at Rice University, USA. He is the author of four books, most recently a commentary on Aeneid 11 (2020). He has also co-edited three volumes, with a fourth, Forgery Beyond Deceit: Fabrication, Value, and the Desire for Ancient Rome, forthcoming. Current projects include a translation of the Aeneid and a commentary on the Vergilian centos. Bret Mulligan is an Associate Professor in the Department of Classics at Haverford College, USA. His research interests include Latin epigram, late antiquity, medical humanities, language pedagogy and digital humanities. His most recent books are The Poetry of Ennodius (2022) and The Crisis of Catiline: 63 bce (2023). Carole E. Newlands is Professor of Classics and Distinguished University Professor at the University of Colorado Boulder, USA. She has published widely in the field of Latin poetry, with a particular focus on Ovid and Statius. Her current interests are in late antique poetics and reception studies. Cillian O’Hogan teaches medieval Latin at the University of Toronto, Canada. His research interests are in late antique Latin poetry and its reception in the Middle Ages, medieval manuscript studies and Latin philology.

ix

Contributors

Ruth Parkes is Senior Lecturer in Classics at the University of Wales, Trinity Saint David, UK. Her research focuses upon the epic tradition and its reception in Latin literature, including neo-Latin texts. She has published extensively on the works of Statius, including a 2012 commentary on Book 4 of the Thebaid. More recently, she has added the works of Claudian to her research interests. Michael Roberts is the Robert Rich Professor of Latin Emeritus at Wesleyan University, USA. He is the author of The Jeweled Style (1989) and has published monographs on biblical epic (1985), the Peristephanon of Prudentius (1993) and the poetry of Venantius Fortunatus (2009), as well as a translation of Fortunatus’ poetry (2017). His translation of Avitus of Vienne’s poetry appeared in 2022. He is also the author of numerous articles on late Latin poetry. Christoph Schubert is Professor of Classics (Latin Language and Literature) at the Friedrich Alexander University at Erlangen, Germany. He has published monographs on Nero’s portrayal in Latin poetry (1998) and Minucius Felix’ Octavius (2014), respectively, as well as a range of articles on early Christian literature. His current projects include a commentary on Commodian’s Carmen apologeticum and an edition of Merobaudes. David Ungvary is Assistant Professor of Classics at Bard College, USA. His research revolves around questions pertaining to later Latin poetry, Christianization, conversion, and the interplay between literary and religious practices in late antiquity. His previous and upcoming publications include studies of late Roman patronage, ascetic writing in fifth-century Gaul and Visigothic penance. Catherine Ware is a lecturer in Classics at University College Cork, Ireland. Her research interests include late antique Latin epic, prose and verse panegyric. She is a member of the Panegyrici Latini Project and her most recent publication is A Literary Commentary on Panegyrici Latini VI (7): An Oration Delivered before the Emperor Constantine in Trier, c. ad 310 (2021).

x

SERIES EDITOR PREFACE

In recent years, the Greek and Latin literature of late antiquity has become the object of significant rediscovery, in line with heightened interest in the history and arts of the Mediterranean at a crucial period in the history of the Roman Empire. And yet, as the very term suggests, late antique texts are still often considered an appendix to classical literature and interpreted in relation to that literature, which provides the aesthetic standard by default. This new series aims at challenging such paradigms by placing late antique literature at the centre of inquiry, and its title joins the Latin sera (‘late’) with the ambiguous tela, evoking both the drawing of weapons and the building of a web. Late antique literature becomes a vantage point for looking both forward to its reception and transformation in subsequent ages, and backward, by engaging in the new and exciting experiment of viewing classical literature through the lens of late antiquity. sera tela offers a laboratory for new ideas animated by interpretive enthusiasm and experimentalism aiming at highlighting what is new, unique, and unprecedented, rather than focussing on relationships with the classical past, and embracing new sets of theoretical approaches and hermeneutic tools deriving from cultural studies, psychoanalysis, cultural materialism, and ecocriticism, as well as post-human, queer, and media studies, among others. This series is a home for literary studies but is also a transdisciplinary adventure that focuses on the novelty and radical otherness of late antique literature and culture, denaturalizing ‘lateness’ and highlighting instead a productive sense of anachronicity and belatedness that characterizes literary works of this period. Finally, sera tela makes of late antiquity another antiquity in which contemporary readers might see reflected their own cultural anxieties around otherness. What makes late antique poetics peculiar and different from that of the classical period? Is there even a specifically late antique literary aesthetic? More than thirty years after its publication, Michael Roberts’s book The Jeweled Style still represents an authoritative and original intervention; any discussion of late antique literature must engage with it. Appearing when late antique texts were scarcely read, especially in Anglo-American academia, The Jeweled Style was truly pioneering. And Roberts’s book is much more than a contribution to late Latin literary aesthetics: it is an elegant study that in certain respects is kindred to the types of texts it discusses, a kind of jewel in itself. Now, in this new volume edited by Joshua Hartman and Helen Kaufmann, we are treated to a rich variety of essays on what The Jeweled Style achieved, what it did not do, and what it still can do.

xi

Series Editor Preface

This book is a wonderful addition to the series sera tela, since the harmonious polyphony of its chapters introduces or reintroduces readers to of one of the most enduring critical foundations of late antique literary scholarship while at the same time opening up new perspectives. Marco Formisano Ghent University January 2023

xii

PREFACE AND ACKNOWLEDGEMENTS

This project began at a chance meeting between Scott McGill and Joshua Hartman at the Annual Meeting of the Society for Classical Studies (SCS) in 2017. The two scholars started thinking about organizing an SCS panel and, as their conversation developed, decided that it would be worthwhile to create new discussion of The Jeweled Style, not only because the panel would convene in 2019, marking the thirty-year anniversary of Michael Roberts’s influential monograph, but also because it had become the vade mecum for studying late antique poetry within those thirty years. Following the panel (‘30 Years of the Jeweled Style’), which was received with great interest at the Annual Meeting of the SCS in 2019, Helen Kaufmann joined the pair and later Scott McGill withdrew due to other commitments. Other contributors were found via invitation and an open call for contributions. To increase the coherence of the volume and the depth of its investigations, Joshua Hartman and Helen Kaufmann organized a conference (‘The Jeweled Style Revisited’) in Erlangen in June 2021. Though the conference had to be held online, it was a welcome opportunity to explore common themes, exchange ideas and test the jeweled style in preparation for this publication. We would like to thank the Deutsche Forschungsgemeinschaft (DFG) for the (virtual) support of the conference ‘The Jeweled Style Revisited’ (Erlangen, June 2021); Helen Kaufmann would also like to acknowledge the support of the Alexander von Humboldt Foundation in the form of a research fellowship at the University of Erlangen (2020–1), during which this volume took shape, and Joshua Hartman would like to mention his gratitude to Scott McGill for co-organizing the initial panel and to the Society for Classical Studies for hosting it. We are both very grateful to Marco Formisano for welcoming this volume into the sera tela series and for the opportunity to contribute to its mission. Furthermore, we are indebted to the staff at Bloomsbury, in particular Alice Wright, Lily MacMahon and Zoe Osman, for taking the volume so professionally and swiftly through the press. Finally, we would like to thank all contributors for entrusting us with their work and for sharing their ideas. In particular, we thank Michael Roberts for so cheerfully and generously accompanying and contributing to our joint enquiry. The adoption of the American spelling of ‘jeweled style’ throughout this volume may be read as a testimony to the influence of his scholarly work.

xiii

NOTE ON TEXTS AND TRANSLATIONS

Unless otherwise indicated, the text editions of the ancient texts quoted are those listed as preferences in the ThLL (for the Latin texts) and the LSJ (for the Greek texts). Secondly, all translations in this volume are the contributors’ own unless otherwise stated. Each contributor should therefore be considered the author of the translations in their chapter except where they explicitly quote from a printed translation. This includes translations from languages other than Latin and Greek.

xiv

ABBREVIATIONS

Greek authors and their works are abbreviated according to the conventions of the Oxford Classical Dictionary (OCD, 4th edn), Latin authors and their works according to those of the Thesaurus Linguae Latinae (ThLL) except in the case of Claudian’s and Ausonius’ works, where abbreviated work titles have been added to the numbers in Hall’s (1985) edition of Claudian and Green’s (1991) edition of Ausonius respectively. Slight variations to the standard abbreviations of the ThLL have also been made for the Panegyrici Latini (PL instead of Paneg.), Alcimus Avitus’ works (spir. hist. instead of carm. 1–5 and cons. instead of carm. 6) and Merobaudes’ poetic panegyric (pan. poet. instead of poet.). Finally, the anonymous Old Testament epic, conventionally attributed to a (probably fictitious) Cyprian from Gaul (Cyprianus Gallus), goes under the abbreviation Hept. (for ‘Heptateuch poet’) in this volume. Other abbreviations to note: ANRW Temporini, H. and W. Haase, eds (1972–), Aufstieg und Niedergang der römischen Welt, Berlin and Boston, MA: De Gruyter LSJ

Liddell, H. G., R. Scott and H. S. Jones (1996), A Greek-English Lexicon, 9th edn, Oxford: Oxford University Press

OLD

Glare, P. G. W. (2012), Oxford Latin Dictionary, 2nd edn, Oxford: Oxford University Press

PG

J.-P. Migne, ed. (1857–1866), Patrologiae cursus completus: series Graeca, Paris: Migne, Garnier and P. Geuthner

RAC

Klauser, T., E. Dassmann, G. Schöllgen and C. Hornung, eds (1950–), Reallexikon für Antike und Christentum, Stuttgart: Anton Hiersemann Verlag

ThLL

Thesaurus linguae Latinae, editus iussu et auctoritate consilii ab academiciis societatibusque diversarum nationum electi (1900–), Leipzig and Stuttgart: Teubner, Munich: Saur, and Berlin: De Gruyter

xv

xvi

INTRODUCTION Joshua Hartman and Helen Kaufmann

In 1989, the publication of Michael Roberts’s The Jeweled Style reshaped scholarly approaches to late antique Latin poetry. It resisted earlier claims that this poetry was derivative and mannerist by attempting to recover what late ancient sensibilities and tastes about poetry may have been. Roberts’s book asked scholars to take late antique poetry seriously and read it sympathetically, and his analyses have persuaded many to read late antique texts with a different set of aesthetic expectations than those they would bring to texts written several hundred years earlier. Through this argument, Roberts endorsed the idea that late antiquity was a distinct period in literary history and therefore needed to be understood and examined on its own terms. As Roberts mentions in his contribution to this volume, he conducted the research for The Jeweled Style in the early to mid 1980s, which makes for an interesting juxtaposition with the 1978 publication of Gordon Williams’s Change and Decline: Roman Literature in the Early Empire. Williams (1978: 2) declared that scholars who abandoned periodization and objective criteria through which works could be judged (and ‘decline’ perceived) abdicated their responsibility. Furthermore, like many studies of the 1970s and 1980s in the field of Latin literature, Williams’s book is largely concerned with the early imperial period: it presupposed distinct periods in Latin literature and saw the first century ce as a period of decline following the Augustan age. Williams’s impulse towards periodization is common to other critics of Latin literature in the 1980s. In Letteratura latina, for example, Gian Biagio Conte (1987) valorizes ‘return’ to an Augustan classicism. In addition, like Roberts, Karl Galinsky (1989) relies on comparative evidence from art history, which he uses to suggest that Ovid’s narrative techniques were in keeping with the Augustan age, although he also praises early imperial poets for their originality. Notably, however, he does not mention any poet later than Silius Italicus. In this scholarly climate, Roberts’s Jeweled Style rescued the Latin poetry of late antiquity by describing its typically late antique aesthetic features and showing how they can be evaluated in positive terms while maintaining the framework of periodization. Scholarly opinion has evolved considerably since Change and Decline, a monograph that resulted from the 1973 Sather Lectures. Philip Hardie’s recent book, Classicism and Christianity in Late Antique Poetry (2019), demonstrates this paradigm shift and strikes further contrast with Williams as it is based on Hardie’s 2016 Sather Lectures. Contrary to Williams’s insistence on periodization, Hardie issues a call to resist the temptation to structure Latin literary history using distinct periods and – with respect to late antique poetry – argues (2019: 223) that ‘the concern to define a period might itself be seen as a sign of the relative newness of the current attention to late antique literature, poetry in particular’. Instead, Hardie (2019: 3) proposes to read late Latin poetry like the poetry of the early empire. It is noteworthy, however, that his book focuses largely on the poets of 1

A Late Antique Poetics?

the late fourth and early fifth centuries: Paulinus of Nola, Prudentius and Claudian are the most frequently discussed authors, whereas Ausonius and later poets like Avitus of Vienne or Dracontius are mentioned more rarely. Meanwhile, Sidonius, Corippus and Venantius Fortunatus are largely confined to footnotes.1 Hardie’s choice of poets recalls J. B. Hall (1991: 361) in his (now infamous) review of The Jeweled Style, in which he praised Claudian and Prudentius while condemning Ausonius and Sidonius as secondrate authors. This leaves one to speculate that some poets may benefit more from reading strategies that are not specifically late antique (cf. Pelttari 2020: 137). The question of how to approach late Latin poetry is precisely why a return to The Jeweled Style is warranted and desirable thirty years after its publication. We must examine how this approach evolved and where it could still be evolving, testing its potential applicability to genres outside late Latin poetry, examining its links to the poetry of earlier and later periods and re-examining the jeweled style as a concept. The idea appealed to scholars from within and beyond the field of late Latin poetry, and we are delighted to present the outcomes of our joint investigation in this volume. Two fundamentally different, but not mutually exclusive understandings of the jeweled style emerge from our re-examination of the concept: a formalist approach and a late antique contextualization. The formalist approach defines the jeweled style through stylistic attributes such as highly visual descriptions of details within a regular framework, episodic narrative, exuberant variation, enumeration and juxtaposition (including antithesis and paradox). This approach considers the jeweled style to be separable from Latin poetry of the later Roman Empire and is close to chapter 2 of the Jeweled Style, in which Roberts discusses the influences on and development of a poetics that would ultimately be embraced by many late antique poets. Roberts’s third chapter, on the other hand, is at the centre of the contextualization perspective, which links the stylistic features of the jeweled style with other characteristics of late antique aesthetics such as formal experimentation, fragmentation and recomposition. We have not synthesized these two approaches into a unified definition of the jeweled style as this would likely have shut down new avenues of inquiry. Instead, we propose to apply the idea of re-wilding to our joint enquiry of Roberts’s metaphor. Commenting on the re-wilding of natural ecosystems, George Monbiot (2013: 20) writes that ‘rewilding . . . is resisting the urge to control nature and allowing it to find its own way’ and ‘rewilding has no end points, no view about what a “right” ecosystem or a “right” assemblage of species looks like’. Applying the concept to the jeweled style, we have stood back to observe what directions the jeweled style has taken in the hands of scholars within and outside this volume while recognizing that scholarly debate and research are impossible without some agreement about the object of enquiry. The volume also tests the limits of the jeweled style through investigations of literary history. Its contributions develop core concepts, engage in close readings and explore comparisons while laying the groundwork for future enquiries. Ecological rewilding attempts to sustain environments well into the future, but also entails a reimagining of what constitutes a successful future. We are confident that the jeweled style will continue to inform future readings and are curious to see what position in the food web it will assume, on what it may feed and which hostile 2

Introduction

competitors might attempt to feed on it. To continue the metaphor, we realize that such pressure from competing ‘natural enemies’ or predators is essential in ecology and we welcome further debate.

The formal features of the jeweled style Roberts’s original description of the jeweled style already implied its independent existence as a style, separate from the literature of a particular period. He (1989: 39–42) highlights its foundation in rhetoric and literary criticism referring, for example, to Quintilian’s definition and understanding of evidentia. In addition, he (1989: 61f.) points to the poets of the first century ce , in particular, Ovid, Lucan and Statius, as predecessors. Criticism of Roberts’s concept also turns on this point, arguing that the features of the jeweled style can be found elsewhere, and that, consequently, the jeweled style is not a particularly late antique phenomenon (e.g. Hall 1991: 360f. and Hardie 2019: 223–49). By contrast, the contributors in this volume who have chosen a formalist approach to the jeweled style improve our understanding of the relationship between the style of a text, its reception and its appreciation. In addition, they examine the jeweled style beyond late Latin poetry and explain how reading other texts as jeweled can inform an approach to the Latin poetry of late antiquity. The term ‘jeweled style’ predates Roberts’s book: Oscar Wilde’s novel The Picture of Dorian Gray (1890–1) deployed the phrase in a reference to Joris-Karl Huysmans’s À rebours (1884). As he examines the reception of late antique Latin in nineteenth-century English and French texts, Ian Fielding shows how perceptions of Roman decadence were shaped by contemporary orientalism. He examines how such associations still remain in the modern academy, and his investigation prompts reflection on new readings that could be generated if we approached the jeweled style as a product of the culturally hybrid world of late antiquity. Fotini Hadjittofi finds that, in Greek literature, the jeweled style can be found in prose (Achilles Tatius) long before late antiquity and only considerably later in poetry (Nonnus). She understands the jeweled style in Greek poetry as the opposite of a Homericizing style, analogous to the ‘sweet’ style of didactic, also understood in antiquity as a deviation from the Homeric model. Hadjittofi’s chapter encourages us to think about the rhetorical function of the jeweled style regardless of genre and prompts a crucial question: what – if anything – in late Latin poetry corresponds to the Homericizing style of Greek poetry? Catherine Ware’s enquiry into prose panegyric further investigates the connection between the jeweled style and rhetoric. Ware argues that many features of the jeweled style characterize epideictic oratory, a genre that had become central to late antique education and was essential to political careers. Prose and poetic style were becoming more similar, and thus the ornaments and tropes of epideixis embellished both poetry and prose. Ware draws attention to the rhetorical nature of late Latin poetry and identifies panegyric and epideictic oratory as influences on late antique aesthetics. 3

A Late Antique Poetics?

Frances Foster delves further into the educational background in rhetoric and literature that poets (and readers) of late Latin poetry would have encountered in the classrooms of grammatici and rhetores. By interpreting evidence from Servius and Ausonius, she demonstrates that while these elite teachers drew every student’s attention to features of the jeweled style, only the most advanced students were likely encouraged to write in it. Her study demonstrates how Vergil’s poetry was read as an exponent of the jeweled style and how it was used to instruct the most elite students to compose tasteful poetry in late antiquity. One of the most common areas of overlap between late antique rhetorical and poetic modes is enumeration, the subject of Joshua Hartman’s and Jacob Levernier’s enquiry. They conduct a quantitative analysis of its occurrences in Latin poetry, arguing that heavily enumerative passages are found more frequently in late antique poetry and that lines consisting almost entirely of enumeration are far more frequent in late antique authors. Their study offers quantitative corroboration of some of Roberts’s earlier hypotheses and suggests that extreme enumeration is rightly considered a distinct feature of the late antique poetic style. Enumeration and different reading strategies are core concerns of Ruth Parkes’s comparative study of tree catalogues in Flavian and late Latin poetry. She demonstrates that scholarly reactions to such catalogues have ranged from outright condemnation to appreciation and contextualization within the works in which they appear. Parkes’s study highlights the similarities between the reception of Flavian and late antique poetry alongside the risks of focusing on formal poetic features. Cillian O’Hogan offers chronological contrast to Parkes as he examines the jeweled style in early medieval poetry, in particular Aldhelm, Hucbald, and the anonymous Hisperica famina. His paper connects the jeweled style to intertextuality and playful poetry and establishes a strong link between early medieval and late antique poetry. Because some of the poems discussed consist almost entirely of ornament, placing them in dialogue with the jeweled style helps to understand the influence of late antique aesthetics as well as the development of medieval literary sensibilities. Concluding the first part of the volume, Helen Kaufmann examines literary unity, in particular digressions, and argues that variety helps us to read them as constituent parts of unified poems and poetic collections. Her study is based on river poems from Augustan and late antique poetry and suggests that late antique poets do not conceive of unity differently to their predecessors.

The jeweled style and late antique aesthetics The arguments of The Jeweled Style have found wide acceptance, and studies of late Latin literature – whether focused on aesthetics generally or on individual authors and works – regularly refer to Roberts’s book.2 The contributors in this volume who consider the jeweled style an expression of late antique aesthetics uncover links between its features and other elements that typify late antique poetic style. 4

Introduction

Christoph Schubert begins by examining novel combinations of traditional metaphors and metonymies in late Latin poetry, a phenomenon he refers to as ‘squaring metaphors.’ By simultaneously activating different meanings, these new expressions participate in the aesthetics of re-composition and accumulation through inherent semantic overload and creative re-use of traditional language. Schubert argues that this language communicates aesthetic ideals rather than content. According to Jesús Hernández Lobato, Augustine’s Confessions also uses language to transcend mere communication with readers. Hernández Lobato suggests that Augustine’s work represents both an apophatic text and a prose poem, in which the church father pioneers a new mode of poetic expression heavily influenced by the Psalms. Augustine’s opposition to the jeweled style manifests itself in plain language and a preference for repetition and parallelism over variatio and antithesis. Another four contributors address the Christian use and view of the jeweled style highlighting features that overlap with Christian artistic sensibilities such as accumulation, overload and the juxtaposition of decontextualized fragments. First, David Ungvary discusses Jerome’s objection to Ambrose’s cento-like sermons as part of a Christian opposition to a poetic style perceived as artificial and poorly suited to homiletics. While Ambrose embraced the radical decontextualization of scripture in his sermons, Jerome rejected this strategy claiming that it produced dubious exegesis. Ungvary’s study reveals a late antique debate about homiletics in dialogue with wider Christian views about how to read and compose texts of any kind. Homiletics also inform Francesco Lubian’s analysis of storm descriptions in the sermons of Zeno of Verona and Juvencus. He notes the limited scope of storm depictions in Juvencus’ New Testament epic and contrasts this with the flamboyant ecphrases in Zeno’s sermons. Like Ungvary, Lubian finds many features of the jeweled style in homiletics, and his analysis of Zeno’s style allows for a comparison with other late antique prose genres such as panegyric. The poetics of Christianity are also central to Elena Castelnuovo’s discussion of accumulation and allusive clustering in Dracontius’ De laudibus dei. She highlights juxtaposition of decontextualized fragments as an important structural feature of poetry, arguing that the apparent fragmentation caused by this juxtaposition find parallels in biblical texts such as the Psalms, liturgical hymns and sermons. Scott McGill also discusses clustering patterns, which he investigates in Christian centos. He suggests that a group of allusions to Vergil’s Fourth Eclogue that originally appear in Proba’s cento informs later centonists who combine the Fourth Eclogue and Proba’s cento itself in a nexus of Christian allusions. Overloaded with meaning, these centos demonstrate how Vergil’s lines transform into works that typify late antique aesthetics, newly marked by fragmentation and re-composition, allusive patterning and layered meanings. In keeping with the McGill’s theme of transforming the work of predecessors, Carole E. Newlands demonstrates how Venantius Fortunatus moved beyond Statius’ ecphrastic building descriptions in Book 1 of his Poems. Eschewing architectural description, 5

A Late Antique Poetics?

Fortunatus focuses instead on the light of churches. Newlands suggests that this emphasis on light may reflect the importance of gold in Merovingian art, its reduced colour schemes and the Christian meaning of light. Overall, her study complements chronologically and thematically the explorations of the Christian contributions to the jeweled style dating to the fourth and fifth centuries in this volume. Wider discussions of late antique aesthetics inform the final two contributions. In an analysis of Latin epigram, Bret Mulligan discusses two aspects of late antique Latin epigram that relate particularly closely to the jeweled style and late antique aesthetics: the connection between image and texts, which recalls the close connection between visual art and the jeweled style, and the tendency towards serial organization of epigram collections, exemplifying the principle of unity in variety. Meanwhile, Andreas Abele challenges the predominant understanding of late antique unity as fragmented by juxtaposing it against its Neoplatonic conceptualization as perfection and beauty. Applying arithmology to the letter collection of Symmachus, he shows how Neoplatonism provides the regular framework within which jewels can be placed. Abele’s argument roots aspects of the jeweled style in late antique philosophical trends and encourages a wider consideration of the philosophical underpinnings of literary style In the epilogue, Michael Roberts reviews both the volume’s findings and the scholarly climate of English-speaking academia around the time that he began writing his 1989 monograph. He re-examines the formal properties of the jeweled style through a theory of the descriptive (Philippe Hamon) and interprets the omnipresence of classical poetic language in terms of the Saussurian notion of langue. Finally, he elaborates the connection between the jeweled style and late antique ceremony, bridging the thirty-year gap between the origins of The Jeweled Style and its re-assessment. This survey of contributions reveals considerable overlap between the two understandings of the jeweled style as a set of formal features and as an expression of late antique aesthetics. Furthermore, numerous themes unite contributions within and across sections. The first is the reception of poetry that does not conform to strictly classical norms, for instance Ovid, Flavian epic, epigram, Nonnus or the exoticism of nineteenth century novelists (Parkes, Mulligan, Hadjittofi, Fielding). Similarly, the triangular relationship between Christianity, meaning and form brings together the contributions from Lubian, McGill, Ungvary, Castelnuovo, Newlands and Hernández Lobato. Meanwhile, enumeration connects other chapters (Parkes, Hartman and Levernier, Castelnuovo, Roberts) while variety and unity establish dialogue between yet another group of authors (O’Hogan, Mulligan, Kaufmann, Abele). Finally, the limits of communication (Schubert, Hernández Lobato), aural arrangements and performance (Ware, Lubian, Roberts, Schubert) and rhetorical training (Foster, Lubian) remain central to the analyses of diverse contributions. Overall, the following conclusions can be drawn from the studies in this volume: 1. The jeweled style has been used to describe a style typical of late Latin poetry and a series of formal features that may appear in texts of any period or language.

6

Introduction

2. The jeweled style combines excess and restraint in a unique way, be it excessive details within a regular frame or manifold variations of the same idea, but its excess has been more salient to critics than the frames that restrain it. 3. The jeweled style explores boundaries, for example those of communication, language, literary expression, sensory perceptions and reflexivity. 4. Within the Latin poetry of late antiquity, the balance tips slightly in favour of restraint for ‘early late antiquity’ (fourth and early fifth century ce ) and in favour of excess for ‘late late antiquity’ (late fifth to early seventh century ce ). 5. In the history of scholarship, the jeweled style has often been perceived as ‘other’. This perception implies deviation from a norm, often the classical norm. In periods which valued classicism, the reception of non-classical texts, including jeweled poetry, has been similarly negative. Conversely, the rehabilitation of earlier postclassical texts towards the end of the twentieth century has paved the way for the appreciation of late antique texts. 6. Even though the jeweled style is identifiable in other literature, the cultural, religious and aesthetic sensibilities of late antique audiences provided the formative elements that made the jeweled style a lasting and dynamic mode of expression capable of accommodating new experiences while preserving dialogue with earlier literature. While this investigation of the jeweled style has been a productive reassessment of its continued vitality as an idea, we anticipate that future examinations will modify these insights again and open up new avenues. It seems a particularly apt moment to establish conversations with other disciplines. For example, collaboration with scholars of late antique Near Eastern literatures could lead to exploration of the similarities and differences between the jeweled style and ornamentalism in Syriac, Hebrew and Aramaic poetry, dialogue with experts of comparative literature might investigate the idea of literary jewels across time and space and cooperation with archaeologists and art historians may link jeweled poetry with the material finds of late antique silver hoards, gold-threaded textiles and gleaming mosaics.3 Whatever new directions may arise from interdisciplinary approaches, they will proceed from the strong foundations that already exist for the jeweled style as a concept within late antique literary studies.

Notes 1. Paulinus occupies nearly three columns in the Index locorum (2019: 287f.), Prudentius more than two (2019: 289f.), while citations of Claudian occupy nearly two columns (2019: 282f.). Meanwhile, Ausonius accounts for less than one, Avitus about a quarter, Dracontius even less (2019: 281f.), and Sidonius and Corippus are mentioned only once each in a footnote according to the Index (2019: 290 and 283, respectively). Finally, Venantius Fortunatus appears more frequently in the footnotes, his only appearance in the body of the text perhaps tellingly occurs in the context of late antique paradox (2019: 167).

7

A Late Antique Poetics? 2. In discussions of late antique literary aesthetics, e.g. Formisano (2007: 279f.), Hernández Lobato (2012: 381–3), McGill and Pucci (2016b: 14), Elsner and Hernández Lobato (2017b: 17), as the standard reference e.g. Santelia (2014: 333), Trout (2015: 21), Onorato (2020: 60). 3. The first point was already noted by O’Hogan (2019: 312) and we regret that we did not succeed in starting this investigation despite various attempts. On the last point see the conference proceedings by Guggisberg and Kaufmann-Heinimann (in preparation).

8

PART I THE FORMAL FEATURES OF THE JEWELED STYLE

9

10

CHAPTER 1 THE DECADENT PREHISTORY OF THE JEWELED STYLE * Ian Fielding

Introduction: A chain of receptions In a footnote in a recent essay, Michael Roberts (2014: 128 n. 39) reveals that, when he published his generative study The Jeweled Style: Poetry and Poetics in Late Antiquity in 1989, he did not know that the phrase ‘jeweled style’ had been coined almost a century before by no less a stylist than Oscar Wilde. In chapter 10 of The Picture of Dorian Gray, Lord Henry Wotton sends Dorian an unidentified ‘yellow book’, written in a ‘curious jewelled style . . . that characterises the work of some of the finest artists of the French school of Symbolistes’ (Wilde [1890–1] 2005: 274).1 Wilde admitted during his first trial in 1895 that this ‘yellow book’ was modelled in part on Joris-Karl Huysmans’s À rebours (Against Nature, 1884), a masterpiece of the fin de siècle (Hyde 1962: 114; see also Wilde [1890–1] 2005: 362–4). Random though it may seem, there is more to this parallel between Roberts and Huysmans than meets the eye: Huysmans’s protagonist, the reclusive aesthete Jean des Esseintes, is characterized as a connoisseur of late Latin literature.2 Chapter 3 of À rebours is a lengthy catalogue of the Latin works on the shelves of des Esseintes’s library in his villa at Fontenay-aux-Roses, a suburb of Paris, where he gives pride of place to authors of the so-called decadence over classics like Vergil and Horace. The poets he singles out for praise include Ausonius, Rutilius Namatianus, Claudian, Prudentius and Venantius Fortunatus – all major representatives of Roberts’s late antique jeweled style. If Roberts was not aware that Wilde had already used the ‘jewelled style’ to allude to Huysmans, can these nineteenth-century authors possibly have had any influence on Roberts’s own readings of late Latin poetry? This question, I would suggest, brings into focus a serious crux for what can be termed the ‘hermeneutics of reception’. Roberts (1989: 5f.) was among the first to approach the history of Latin literature via the principles of Hans Robert Jauss’s Rezeptionsästhetik – above all, via the principle that literary works are best appreciated in relation to the aesthetic preconceptions of contemporary readers, their ‘horizon of expectations’ (Jauss 1982: 22–4). As Jauss explained when setting out his theses, however, the horizons of the past can only be surveyed from within the horizons of the present, and the perspective of the literary historian is always framed by the ways in which a text has been received in the time since it appeared (1982: 29f., citing Gadamer 1960 [2013]). For all that the field of classical reception studies has grown in the last thirty years, there has been little reconsideration of the degree to which expert readers can exert control over this ‘fusion of horizons’, as Jauss suggested. One of the few classicists to 11

A Late Antique Poetics?

critique this claim was Charles Martindale, who in the opening pages of Redeeming the Text (1993: 7) advanced the hypothesis that ‘our current interpretations of ancient texts, whether or not we are aware of it, are, in complex ways, constructed by the chain of receptions through which their continued readability has been effected’ (emphasis added). The ultimate implication of Jauss’s theory – in the terms of his favourite metaphor – is that one’s view of the past is necessarily determined by what lies beyond one’s own horizons. This essay aims to shed light on some links in the ‘chain of receptions’ that might connect Roberts’s late antique jeweled style with the decadent jeweled style of Huysmans. My enquiry draws inspiration from a recent turn in scholarship on late antiquity, led by Marco Formisano and Therese Fuhrer (2014), towards a more constructive engagement with the decadent associations that have loomed over the study of the period for a long time. This approach, informed by reception theory, recognizes that the nineteenthcentury fascination with Roman decline maintains a powerful residual influence over perceptions of late antiquity, even for people who have never heard of Huysmans (Formisano 2014: 9–13; see also David 2001). In a recent discussion of the reading of Latin literature in chapter 3 of À rebours, Scott McGill (2018: 87) has observed that ‘the nineteenth century, including the fin de siècle, was crucial to the development of the modern conception of a decadent late Roman Empire’ – adding, ‘the apologetic scholarship that . . . [Michael] Roberts represent[s] is a legacy of that way of thinking, because it is a response to it’. In what follows, I will show in more detail how Huysmans was reacting to certain cultural, intellectual and political trends that were formative for the modern academic discipline of classical studies and therefore continue to define the ‘horizon of expectations’ within which the Latin texts analysed in Roberts’s Jeweled Style are received. In order to achieve a more comprehensive understanding of these expectations, I argue, it will be important to expand our focus on jewels beyond metaphors of aesthetics and look carefully at what is reflected in their other facets.

The language of jewels Jewels had been common parlance for verbal ornament long before Wilde ever picked up a copy of À rebours, of course. Roberts (1989: 52–4) himself traced the idea of the jeweled style back to the Latin authors of the late first century ce . For example, Marcus Aper, the apologist for modern rhetoric in Tacitus’ Dialogus de oratoribus, argues that a public speaker should not decorate his house (as it were) with any old furniture, but rather, with eye-catching ‘gold and jewels’ (dial. 22.4: et aurum et gemmae, cf. Mart. 5.11.3f.). Writers associated with the decadence of the nineteenth century described their own artistry in much the same way. ‘I love this word decadence,’ remarked the poet Paul Verlaine, ‘all shimmering in purple and gold . . . it throws off bursts of fire and the sparkle of precious stones’ (Raynaud 1918: 64).3 The metaphor is also found in chapter 3 of À rebours, with reference to Lucan, whose epic verses, ‘plated with enamel’ and ‘lined with jewels’ (vers plaqués d’émaux, pavés de joaillerie), are the only Latin poetry from before the third century ce for which the narrator, des Esseintes, expresses any admiration.4 Huysmans 12

Decadent Prehistory of the Jeweled Style

thus presents Lucan as a Roman counterpart to Théophile Gautier, an influential figure of the generation before the decadent movement, whose collection of poems, Enamels and Cameos (Émaux et camées), first published in 1852, was another of the ‘poisonous’ books consumed by Dorian Gray in Wilde’s novel ([1890–1] 2005: 304f.).5 Lucan was identified by Roberts (1989: 61) as one of the most important forerunners of the jeweled style of late antiquity, but Gautier was more than just a model for later nineteenth-century writers like Huysmans: his preface to the 1868 edition of Charles Baudelaire’s Les fleurs du mal was, in David Weir’s (1995: 88f.) words, ‘a veritable ars poetica of decadence’ (see also Dowling 1986: 148–50). ‘For the poet,’ Gautier ([1868] 1986: 145f.) explained, commenting on Baudelaire’s penchant for strings of polysyllables, ‘words have, in themselves and apart from the meaning they express, their own beauty and value, like precious stones that are still uncut and not mounted in bracelets, in necklaces, or in rings’.6 This minute attention to the sensory qualities of verbal forms – viewed as if through a jeweler’s loupe – pushes language to the densest opacity of expression, a stylistic extreme which Gautier ([1868] 1986: 125) likens to the Latin of the late Roman Empire.7 Baudelaire ([1857] 2005: 1767–9) himself had invoked the poetics of ‘la dernière décadence Latine’ in a note on his experimental love lyric Franciscae meae laudes, which he composed in rhythmic triple rhymes reminiscent of medieval hymns (see also Evangelista 2019: 324f.). Gautier ([1868] 1986: 125) also applauded his preference for some of the same prose authors celebrated in chapter 3 of À rebours: Petronius and Apuleius, about whom more will be said below (p. 19), as well as Tertullian, ‘whose style has the black sheen of ebony’.8 As I will show, these ideals of extravagance and exoticism are elaborated on a grander scale in Huysmans’s treatment of late Latin literature.

The orientalism of the jeweled style I will return shortly to the Latin volumes in des Esseintes’s library, but in order to appreciate the significance of jewels in his decadent aesthetics, it is worth considering what else is displayed on his orange walls, lined in crushed morocco leather. Hung between his ebony bookcases are two paintings by the contemporary artist Gustave Moreau: Salomé Dancing before Herod and The Apparition (both 1876, see Figures 1.1 and 1.2). These artworks reflect what Charles Bernheimer (2002: 104–38, esp. 110–17) called the ‘Salomania’ of the fin de siècle – its captivation with Salomé, the bloodthirsty seductress of the Gospels. Wilde wrote a one-act tragedy, Salomé, in French – but not until he visited Paris in the last few months of 1891, too late to receive notice in À rebours.9 In chapter 14, however, Huysmans ([1884] 1959: 196–7) recounts how des Esseintes used to spend entire evenings contemplating the figure of Moreau’s bejewelled femme fatale, while reciting to himself the melancholy verses that Stéphane Mallarmé had composed for her in Hérodiade (1864–7).10 Of Huysmans’s most lavish descriptions of these two pictures, found in chapter 5, the following passage on The Dance of Salomé (Figure 1.1) offers an illustration: 13

A Late Antique Poetics?

Figure 1.1 Gustave Moreau, Salomé Dancing before Herod (1876), Hammer Museum. Image from Wikimedia Commons.

14

Decadent Prehistory of the Jeweled Style

Figure 1.2 Gustave Moreau, The Apparition (1876), Harvard Art Museums. Image from Wikimedia Commons.

With a withdrawn, solemn, almost august expression on her face, [Salomé] begins the lascivious dance which is to rouse the aged Herod’s dormant senses; her breasts rise and fall, the nipples hardening at the touch of her whirling necklaces; the strings of diamonds glitter against her moist flesh; her bracelets, her belts, her rings all spit out fiery sparks; and across her triumphal robe, sewn with pearls, patterned with silver, spangled with gold, the jewelled cuirass, of which every chain is a precious stone, seems to be ablaze with little snakes of fire, swarming over the mat 15

A Late Antique Poetics?

flesh, over the tea-rose skin, like gorgeous insects with dazzling shards, mottled with carmine, spotted with pale yellow, speckled with steel blue, striped with peacock green. (Huysmans [1884] 1959: 64) The scintillating vividness of this paragraph – just a small section of the description, which extends over several pages – is even more extraordinary in light of the fact that Moreau’s paintings had actually been purchased by private collectors soon after the exhibition at the Exposition Universelle in Paris in 1878; Huysmans only had access to black-and-white reproductions when he was writing (Bernheimer 2002: 210f. n. 16). It is clear, nonetheless, that the jewels in the jeweled style of À rebours are not merely metaphorical. The emphasis on the brilliant variation of colours radiating from the gemstones in Salomé’s ‘jewelled cuirass’ might bring to mind another biblical breastplate, which may be familiar to students of late Latin poetry as the paradigmatic example of Roberts’s (1989: 9–13) jeweled style: that is, the breastplate of Aaron, as described in the Heptateuchos attributed to Cyprianus Gallus (Hept. [Cypr. Gall.] exod. 1098–103; see Hartman and Levernier in this volume for the text and translation of the passage). With its meticulous juxtaposition of obscure vocabulary in regularly ordered clauses, mirroring in words the craftsmanship of the jeweller, Huysmans’s ecphrastic technique is practically late antique.11 The ‘excess of exactitude’ in his descriptions was already characterized as a ‘medieval’ quality in a 1908 essay by Arthur Symons ([1908] 2014: 78), who was the very first critic to write about the continental decadent movement for an Anglophone readership.12 In the passage quoted above, this luxurious excess can be taken as suggestive in itself of the dangerously effeminate sensuality that Salomé, clad in her jewel-studded armour, embodies (Meltzer 1987: 20f.).13 In the background of Moreau’s picture (Figure 1.1), by contrast, the figure of Herod, statuesque and almost entirely pallid apart from the jewels glinting on his own golden robe, seems to represent the threat of emasculation lurking beneath the sheen of precious stones. Roberts (1989: 12) comments in passing on the ‘oriental glamour’ that the names of rare stones lend to the jeweled style of late antique poetry, but the orientalism of Huysmans’s jeweled style is harder to overlook. As Edward Said (1978: 180) noted, the orient occupied a place of special fascination in the imaginations of French decadent writers, closely related to their peculiar tendencies to misogyny, deviancy and mysticism. Huysmans recognized in the accoutrements of Moreau’s Salomé a reference to the heroine of one of des Esseintes’s favourite novels, Gustave Flaubert’s Salammbô (1862), set in Carthage shortly after the First Punic War; Moreau’s paintings inspired Flaubert in turn to write Hérodias (1877), the last short story he published before his death.14 The model for Flaubert’s ancient exotic dancers was the Egyptian almeh Kuchuk Hanem, whom he had paid for sex while visiting the Upper Nile in March 1850.15 Flaubert was not the only French author of this period to exploit the opportunities presented to tourists by the expansion of European power across the Mediterranean: Gautier too wrote about his travels in Egypt and Algeria (Behdad 2019). Although France did not enlarge its overseas empire to the same extent as Britain in the nineteenth 16

Decadent Prehistory of the Jeweled Style

century, French interests were still advanced by private enterprises like the building of the Suez Canal, completed in 1869, which cut down the vast distances involved in the transportation of the precious stones, silks and other materials with which des Esseintes furnishes his villa in À rebours. What is distinctive about Huysmans’s orientalism, however, is that it does not simply portray the east as a place of escape from European inhibitions; rather, it savours the growing oriental influence in the culture of the European metropolis.16 Huysmans’s fixation with jewels can thus be understood as one aspect of what we might call his decadent eclecticism – his manner of appropriating aesthetic elements from various cultural traditions, especially those that he would have described as oriental. In his analysis of the Salomé paintings, for example, what he admires is not only the ambience of eastern splendour, but the blending of iconography from different civilizations: the architecture incorporates Byzantine and Islamic styles and the lotus in Salomé’s hand (see Figure 1.1) recalls the ancient religions of Egypt and India (Huysmans [1884] 1959: 63 and 66f.). He ([1884] 1959: 69) praises Moreau for ‘joining and fusing in one those legends which had originated in the Middle East only to be metamorphosed by the beliefs of other peoples . . . to justify his architectonic mixture, his sumptuous and unexpected combinations of dress materials’. Obviously, late antiquity is one historical period when these legends from the Middle East were transformed by the beliefs of people in the Roman Empire. As Stefano Evangelista (2018: 17) has pointed out, late Roman culture intrigued decadent writers like Huysmans and Wilde precisely because it was ‘belated and derivative, syncretic, cosmopolitan, heterogeneous, and layered with different histories’.17 For des Esseintes, certainly, part of the allure of late Latin literature is the distinctive foreign flavour that the language acquired through the influence of Christianity as it spread west across the Mediterranean, bringing ‘new words, unfamiliar constructions, unknown verbs, adjectives of super-subtle meaning, and . . . abstract nouns’ (Huysmans [1884] 1959: 46; see also McGill 2018: 96–8). These linguistic features are likened, not to jewels, but to another luxury import from the east – spices, with which the ‘ancient carcass’ of the Latin tongue was preserved.18

The aesthetics of decadence Most thinkers of the nineteenth century regarded Rome as the ultimate cautionary tale for modern European empires: a society so enfeebled by its adoption of oriental mores that it eventually gave way to the dual forces of ‘barbarism and religion’ – to borrow Edward Gibbon’s ([1788] 1994: 1068) well-known phrase.19 Huysmans, on the other hand, positively embraced this narrative of corruption and catastrophe. In addition to frequent references to the trappings and rituals of the early Catholic Church, À rebours contains numerous visions of invading tribes rampaging across Roman frontiers, laying waste to the boundaries separating civilization from barbarism – visions that, in the imagination of des Esseintes, exercise all the magnetism of fine jewellery. Thus, the exotic plants that he gathers in the entrance hall of his villa appear to form ‘a mass of green 17

A Late Antique Poetics?

weapons, over which floated, like barbarian battle-flags, flowers of crude and dazzling colours’ (Huysmans [1884] 1959: 101). In one of the novel’s most bizarre episodes, des Esseintes purchases a large tortoise, which he intends to let roam in his library, so that its dark brown carapace can bring out the orange and indigo hues of his decorative scheme. Unsatisfied with the effect, he decides to have the carapace embellished with gold; it comes back from the gilder ‘looking like a Visigothic shield, tegulated with shining scales by a barbaric artist’ (Huysmans [1884] 1959: 54). Des Esseintes continues to decorate the tortoise with an assortment of precious stones, until finally he finds it lying lifeless on the carpet, unable to bear the ‘dazzling luxury imposed on it’ – a strikingly macabre emblem of decadence (Huysmans [1884] 1959: 62). These images of aestheticized barbarian violence appear often in the descriptions of late Roman authors in chapter 3 – as, for instance, in this sketch of Claudian: A poet hammering out a brilliant, sonorous hexameter on his anvil, beating out each epithet with a single blow amid showers of sparks . . . With the Western Empire crumbling to its ruin all about him, amid the horror of the repeated massacres occurring on every side, and under the threat of invasion by the barbarians now pressing in their hordes against the creaking gates of the Empire, he calls Antiquity back to life, sings of the Rape of Proserpine, daubs his canvas with glowing colours, and goes by with all his lights blazing through the darkness closing in upon the world. (Huysmans [1884] 1959: 47) As with Lucan and Gautier in the passage I discussed before, Claudian here sounds like Paul Verlaine, mentioned briefly above, whose famous sonnet Langueur, published a year before À rebours in 1883, evokes ‘the empire at the end of decadence’, with the poet watching ‘blond barbarians pass by’ as he ‘composes idle acrostics in a style of gold’.20 Verlaine’s verses can be read as alluding to the Franco-Prussian War of 1870–1, which had ended quickly with victorious German soldiers marching through the streets of Paris, having inflicted a humiliating final blow to French pretensions of imperial power. This defeat drove home the uneasy feeling, which had been growing since the overthrow of Napoleon in the early nineteenth century, that France was experiencing its own decline and fall (Vance 1997: 263–6). It was not only French intellectuals like Verlaine who raised the spectre of the late Roman Empire: in the German states, this triumph was viewed as another moral victory for northern strength and self-discipline over a degenerate Latin race.21 As the example of Rome became more problematic across Europe, then, Huysmans and his fellow decadent writers tried to reclaim the negative stereotypes associated with their own national disgrace. From decadence to diversity This was also the geopolitical context in which classical studies was emerging in its current form as an academic discipline; it was by no means impervious to the nationalist 18

Decadent Prehistory of the Jeweled Style

politics that shaped the study of modern vernacular literatures in the nineteenth century.22 Ancient Greek and Latin literary history – much like French, German or English literary history (Perkins 1992: 1–3) – was treated as an enquiry into the character of a given ethnic group or nation. As a result, the authors of late antiquity – viewed as neither Roman nor Italian, Greek, French, Spanish or North African – sunk ever deeper into widening institutional cracks.23 To the extent that these writers held any interest for philologists in this period, it was as sources for the early history of Germanic peoples or for the development of Christianity in the era of the Church fathers.24 For des Esseintes’s remarkably promiscuous reading in later Latin, Huysmans’s only known guides were handbooks by patristic scholars, as well as Adolf Ebert’s Allgemeine Geschichte der Literatur des Mittelalters im Abendlande bis zum Beginne des XI. Jahrhunderts (1880) – the first volume of which had been translated into French in 1883, just in time for a few final checks before À rebours went to press (Ebert 1883).25 Ebert’s Geschichte (which also happens to appear in the bibliography of Roberts’s Jeweled Style) was unusual for its time in presenting medieval literature, both Latin and vernacular, as a unified cultural entity (Ziolkowski 1996: 515f.). Still, it would take an even more catastrophic eruption of European imperialist rivalries in the midtwentieth century for the most prized works in des Esseintes’s library to begin to be given serious attention from critics like Henri-Irénée Marrou, Ernst Robert Curtius and later Jacques Fontaine and Reinhart Herzog, who Roberts (in this volume) singles out as his scholarly models.26 From the first sentence of chapter 3, Huysmans ([1884] 1959: 40) frames his alternative canon – or, as Ralph Hexter (2010: 33) puts it, his ‘anti-canon’ – of Latin texts as a reproach to the conventional thinking of ‘minds drilled into conformity by repetitious university lectures’. Of course, the term ‘decadence’ had itself been usurped from scholars such as Désiré Nisard, whose influential 1834 study of the ‘decadent’ Latin poets urged contemporary French writers to model themselves on the geniuses of republican Rome, which was more like Athens (an independent city state), rather than the idle minds of imperial Rome, which was more like Alexandria (a multicultural, not to say oriental, capital).27 This is no doubt the sort of magisterial bluster at which Huysmans would have sneered. In À rebours, des Esseintes’s two favourite Latin stylists both speak to the intermixture of people in the Roman Empire: in the Satyricon, Petronius ‘borrows expressions from all the languages imported into Rome . . . foreigners in their barbaric lingo, shot with words from African, Syrian, and Greek’ while the prose of Apuleius rushes along ‘in a dense flood fed by tributary waters from every province . . . combining them all in a bizarre, exotic, almost incredible torrent of words’ (Huysmans [1884] 1959: 43f.).28 These two authors are not generally considered exemplars of the late antique jeweled style, but it is clear that what Huysmans valued in Latin literature was the sort of cultural exchange that brought precious stones into the Roman world in the first place. As Matthew Potolsky (2012: 86–91) has argued, then, Huysmans’s exaltation of texts that did not fit with the nineteenth century’s increasingly rigid national literary canons reads as a rejection of the ‘patriotic or political twaddle’ from which des Esseintes withdraws to his villa to escape (Huysmans [1884] 1959: 22). 19

A Late Antique Poetics?

Conclusion: Towards a brighter future It may be too easy to characterize Huysmans as little more than a provocateur, who was only drawn to the marginal authors of late antiquity by a compulsion to overturn contemporary canons of taste.29 My intention is not to hold him up as an ideal for Latinists to follow: the material I have presented makes clear that, for one thing, the jewels in À rebours are indicative of certain chauvinistic attitudes that are best left in the nineteenth century. I would argue, all the same, that there is valuable insight to be gained from looking at late Latin literature through the prism of Huysmans’s jeweled style. These decadent reflections reveal the broader conceptual framework of the late antique jeweled style – which, as I have shown, is not strictly aesthetic, but an intricate ideological construction that, in spite of much effort during the last thirty years, has yet to be dismantled.30 It is important to recognize that for many philologists of the nineteenth and twentieth centuries, the dazzling contrasts of rare stones so admired by des Esseintes signified not just an objectionable excess of verbal flamboyance, but also oriental luxury and, by association, gender non-conformity and ethnic miscegenation.31 Certainly, in the USA, where Roberts’s Jeweled Style was eventually conceived and published, not all the architects of American classical studies shared Huysmans’s taste for linguistic and cultural hybridity. In 1868, just a few months before the founding of the American Philological Association, the man who would later become its ninth president wrote: ‘Frenchmen are not good judges of race, and show an incredible philosophy in the matter of amalgamation’ (Gildersleeve [1868] 1890: 487).32 One can only infer what Basil Gildersleeve would have made of Huysmans’s essay on Latin literature in À rebours. Over more than three decades, Michael Roberts has helped an entire generation of scholars, who were trained to idealize classical Greece and Rome in a state of unadorned marmoreal whiteness, to acquire a richer appreciation for the polychrome poetics of late antiquity. In order to continue his project of creating more favourable conditions for the reception of late Latin texts in the decades to come, I think it will be necessary to reckon more directly with the institutional structures that have kept works of the jeweled style in the shadows for so long. These structures – the university departments with their degree programs, requirements and reading lists, around which the discipline of classics took shape in the latter half of the nineteenth century – have proved more enduring than the nationalist principles on which they were founded. As a critique of those principles and the modes of reading they inspired, Huysmans’s promotion of late antique authors who cannot be neatly circumscribed within the borders of modern European nation states was probably more radical in the 1880s than it is today. In other areas of the academy, literary studies have long since moved away from an emphasis on national traditions towards a more transnational approach, exploring the paths of migration by which people, like precious stones, have been extracted from their native lands and set alongside others in diverse new contexts.33 In this way, I suggest, Huysmans’s À rebours offers not just a glimpse of the decadent prehistory of the jeweled style, but also perhaps a compelling vision of its future. 20

Decadent Prehistory of the Jeweled Style

Notes *

My thanks go first and foremost to Michael Roberts, whose guidance and encouragement were invaluable at the beginning of my career. I am also grateful to the editors of this volume for their hard work, support and patience throughout the project. Finally, I would like to acknowledge the help of Bellina Gaskey in editing this essay for publication. Any remaining errors are mine alone.

1.

‘The style in which it was written was that curious jewelled style, vivid and obscure at once, full of argot and of archaisms, of technical expressions and of elaborate paraphrases, that characterises the work of some of the finest artists of the French school of Symbolistes.’ The earlier 1890 text, which appeared in Lippincott’s Monthly Magazine, read ‘the French school of Décadents’ (Wilde [1890–1] 2005: 103); the latter term was perhaps too redolent of moral depravity, in view of the scandal provoked by the novel’s initial publication.

2.

The reception of late antique texts in Huysmans’s novel has attracted attention from Latinists in recent years; see McGill (2018: 85–104) and also Shanzer (2009b: 927f.) and Uden (2018: 628–30).

3.

‘J’aime . . . le mot de décadence tout miroitant de pourpre et d’ors.’

4.

The French text is from Huysmans ([1884] 1983); the English translation is adapted from Huysmans ([1884] 1959). In spite of this ‘metallic brilliance’ of Lucan’s poetry, however, des Esseintes condemns ‘the poverty of its intellectual content’.

5.

As with Lucan (n. 4 above), des Esseintes is not overly enthusiastic about Gautier’s poetry in À rebours – although, Huysmans ([1884] 1959: 189) writes, he still appreciated it ‘in the same way that he appreciated rare jewels or precious substances’.

6.

‘Pour le poète, les mots ont, en eux-mêmes et en dehors du sens qu’ils expriment, une beauté et une valeur propres comme des pierres précieuses qui ne sont pas encore taillées et montées en bracelets, en colliers ou en bagues.’

7.

‘On peut rappeler . . . la langue marbrée déjà des verdeurs de la décomposition et faisandée du bas-empire romain’ (one may recall . . . the language of the late Roman Empire, already marbled with the greenness of decay and pungent). Compare Huysmans ([1884] 1959: 46 and 48f.) for similar metaphors of decomposition. On decadent perceptions of late Latin as the archetypal dead language see Dowling (1986: 151–3).

8.

On Tertullian compare Huysmans ([1884] 1959: 45f.) with McGill (2018: 94–7).

9.

The French text and the 1894 English translation by Lord Alfred Douglas are printed in Wilde ([1883–94] 2013). On Wilde’s play and earlier French treatments of Salomé see Townsend (2008).

10. Rather than Salomé, the name provided for the daughter of Herod and Herodias at Joseph. AJ 18.136, Mallarmé ([1862–71] 1959: 157) preferred on aesthetic grounds to call her Hérodiade – ‘that word’, he wrote in a letter of 1865, ‘sombre and red, like an open pomegranate’ (‘ce mot sombre et rouge, comme une grenade ouverte’). 11. On Huysmans’s pictorialism as ‘an uncanny, even an unconscious, double of a late antique jeweled style’ see Uden (2018: 630). 12. Symons ([1908] 2014: 76) also remarks that Huysmans possessed ‘so astonishing a mastery of description that he could describe the inside of a cow hanging in a butcher’s shop as beautifully as if it were a casket of jewels’. 13. More broadly, on the ways in which the fascination with Salomé in the fin de siècle reflects nineteenth-century anxieties about racial mixing see Burgers (2014).

21

A Late Antique Poetics? 14. On Salomé’s diadem, ‘like Salammbo’s, in the shape of a Phoenician tower’, see Huysmans ([1884] 1959: 66), apparently alluding to the description of Salammbô’s headdress at Flaubert ([1862] 1971: 50). On des Esseintes’s reading of Flaubert’s novel, see Huysmans ([1884] 1959: 182). On the mutual influence between Flaubert and Moreau see Meltzer (1987: 17–19). 15. Said (1978: 187) describes Kuchuk as the ‘prototype’ of Salammbô and Salomé. 16. Compare Orrells (2012: 217–23) on the ways in which Dorian Gray becomes orientalized in late Victorian London. 17. See also Evangelista (2019: 325). 18. Huysmans ([1884] 1959: 50): ‘the sixth century was represented by Fortunatus, Bishop of Poitiers, whose hymns and Vexilla Regis, carved out of the ancient carcass of the Latin language and spiced with the aromatics of the Church, haunted his thoughts on certain days’. On metaphors of linguistic decay see n. 7 above. 19. The early volumes of Gibbon’s history were first translated into French by N. M. Leclerc de Sept-Chênes – reportedly with the assistance of his tutee, the future King Louis XVI – before the latter volumes had even appeared in English; Decline and Fall remained influential in France, in the 1812 and 1828 editions of François Guizot, throughout the nineteenth century. 20. Verlaine’s poem was originally published in the journal of Le Chat noir on 26 May 1883. The first stanza (lines 1–4) reads: ‘Je suis l’Empire à la fin de la décadence, | qui regarde passer les grands Barbares blancs | en composant des acrostiches indolents | d’un style d’or où la langueur du soleil danse.’ On the parallel with Claudian see McGill (2018: 101f.). 21. See Marchand (2003: 159f.): ‘After 1870, “Rome”, more than ever, came to signify antinationalist tyranny, elitism and ultramontanism, and its symbolic defeat grew increasingly important to the establishment of German cultural autonomy.’ 22. According to the much-cited observation of Hobsbawm (1962: 135), ‘[t]he progress of schools and universities measures that of nationalism’. On the increasingly nationalistic turn that philology took after 1871 see Warren (2013: 120–4). 23. Celenza (2001) has claimed convincingly that the Latin authors of both late antiquity and the Italian Renaissance became marginalized in the nineteenth century as a result of the increasingly nationalistic orientation of the philological disciplines. This was also a period of growing separation between the study of Graeco-Roman and so-called oriental cultures see Marchand (2009: 72–4). 24. One could point to such projects in canon formation as the Auctores Antiquissimi series of the Monumenta Germaniae Historica (1826–66), Migne’s Patrologia Latina (1844–55) or the Corpus Scriptorum Ecclesiasticorum Latinorum (1866–). 25. Huysmans’s debt to Ebert was highlighted after his death by his collaborator, Remy de Gourmont (1909: 7f.). As Céard (1978) has shown, however, Huysmans only used Ebert when the manuscript of À rebours was almost finished – most of his information about late and medieval Latin authors was drawn from earlier histories. 26. In general, on the development of literary approaches to late antique texts in the nineteenth and twentieth centuries see Vessey (2008: 49–55). 27. For the parallels to Athens and Alexandria see Nisard (1834: 58–61) with Vance (1997: 251f.). Ten years later, Nisard (1844) assembled a modern French canon, on which see Guiney (2004: 104–6). 28. Compare Eduard Norden’s (1898: 2.600f.) infamous criticism of Apuleius’ prose style, which‚ he complained, overflows ‘like a wild torrent’ (‘der. . . wie ein wilder Strom sich selbst überstürzende . . . Stil’); see also Harrison (2013: 22f.). 22

Decadent Prehistory of the Jeweled Style 29. On Huysmans’s ‘inversion’ of the judgements of contemporary works of reference see e.g. Vance (1997: 263). 30. As Martindale (2005: 137f.) has stressed, ‘issues about canonicity . . . are not aesthetic . . . rather they have to do with the politics of reading, with institutions and education’. 31. To which connotations one could also add the aroma of Roman Catholicism, which was regarded in Protestant universities on both sides of the Atlantic as a peculiar vice of the so-called Latin races see Farrell (2001: 100–4). 32. He continues: ‘Those nations that keep their blood pure, that refuse to blend with an inferior stock, are those that are destined to the mastery.’ See the discussion of Habinek (1998: 15–32, p. 20 for the quotation of this passage). As Farrell (2001: 105) has observed, ‘[t]he traditional structures of Latin literary history . . . in their designation of postclassical latinity as debased, corrupt, substandard, weak-minded, and generally unworthy of notice . . . work to precisely the same ends as the sexually, racially, and religiously chauvinistic attitudes of nineteenth-century ethnography’. 33. Historians, too, have recently begun to call for global approaches to the study of late antiquity; see Humphries (2017).

23

24

CHAPTER 2 THE GREEK JEWELED STYLE * Fotini Hadjittofi

Michael Roberts’s ground-breaking monograph argued that a set of stylistic features became prominent in Latin poetry during the fourth and fifth centuries ce : brilliant verbal and/or visual patterning, episodic fragmentation, ecphrastic description, enumeration and paradoxical juxtaposition – together these constitute the jeweled style, which scholars of late Greek poetry have subsequently detected in poets spanning the whole period from Quintus of Smyrna (third century) to Nonnus of Panopolis (fifth century) and the latter’s sixth-century imitators.1 This contribution aims primarily to undertake an ‘archaeology’ of this style in late Greek poetry: it will first try to detect its presence in poetry, starting from the third century; it will argue that Greek prose is both an earlier and a more frequent vehicle for the jeweled style than Greek poetry; and finally, it will provide some preliminary explanations as to why much of Greek poetry throughout late antiquity continued to be written in a markedly archaic, Homericizing style, which can be conceived as the opposite of the jeweled style.

A Greek jeweled style in the third century? When exactly does a jeweled style emerge in Greek poetry? Can it really be detected as early as the third century? Ps.-Oppian’s Cynegetica, dedicated in the proem to emperor Caracalla, can be securely dated to the early third century. A few lines after the proem’s dedication, we find a very metapoetic, and very Callimachean, dialogue between the poet/narrator and the goddess who will inspire, and be the natural patron of, his poem: Artemis. The goddess accosts the poet and prescribes his subject matter using a style that is, at times, strikingly rhetorical. These are her parting words (1.35–40): μέλπε μόθους θηρῶν τε καὶ ἀνδρῶν ἀγρευτήρων· μέλπε γένη σκυλάκων τε καὶ ἵππων αἰόλα φῦλα, βουλὰς ὠκυνόους, στιβίης ἐϋκερδέος ἔργα· ἔχθεά μοι θήρεια λέγειν, φιλότητας ἀείδειν καὶ θαλάμους ἐν ὄρεσσιν ἀδακρύτοιο Κυθείρης καὶ τοκετοὺς ἐνὶ θηρσὶν ἀμαιεύτοιο λοχείης. Sing the battles of wild beasts and hunting men; sing of the breeds of hounds and the varied tribes of horses; the quick-witted counsels, the deeds of skilful tracking; tell me the hates of wild beasts; sing their friendships 25

A Late Antique Poetics?

and the bridal chambers of tearless Aphrodite upon the mountains, and the births without midwife among the beasts. (Translation adapted from Mair 1928: 7) Throughout his poem, Ps.-Oppian is very fond of anaphora: here we see it in two distichs (lines 35f. and 39f.). The latter distich repeats the exact same syntactical pattern in its two verses: the initial καί is followed by a plural accusative in -ους; then we have ἐν with a dative plural in -σιν; then a combination of adjective and noun in genitive singular, with the adjective ending in -οιο and the noun in -ης. With the last two words of the goddess, ἀμαιεύτοιο λοχείης (the unmidwifed birth), Ps.-Oppian coins a new adjective and creates a seeming contradiction that would later appeal to Nonnus of Panopolis, who uses the adjective to describe, among other paradoxical generations, the birth of Dionysus, unmidwifed and half-formed, from Semele’s burning body (Dion. 1.5).2 Before we rush to assume that the pseudo-Oppianic Artemis proclaims a new, jeweled style as well as a new poem, we should remember that, for Greek poets, style is wedded to genre, and each genre is defined by its first inventor, the πρῶτος εὑρετής. For Ps.Oppian’s didactic poem, this would be Hesiod. Indeed, in her very first words to the poet/ narrator, Artemis tells him, ‘Get up! Let us tread upon the rough path, / which no mortal has ever trodden with his song!’ (1.20–1: ἔγρεο, καὶ τρηχεῖαν ἐπιστείβωμεν ἀταρπόν, / τὴν μερόπων οὔπω τις ἑῇς ἐπάτησεν ἀοιδαῖς). Apart from containing obvious Callimachean echoes (see Costanza 1991), these lines also enclose an allusion to Hesiod in the ‘rough path’ that the poet is asked to tread which echoes the ‘rough path’ of virtue that Perses, Hesiod’s brother, is encouraged to follow in the Works and Days (290f.: οἶμος . . . τρηχύς). As Michael Paschalis (2000: 219) notes, the hunting ground, which is naturally ‘rough’, becomes metapoetic ground, as Ps.-Oppian traces the genealogy of his song through Callimachus back to Hesiod. By doing so, Ps.-Oppian acknowledges that Callimachus’ untrodden paths were already and always Hesiodean. The Cynegetica’s Hesiodean affiliation has consequences for its style. Hesiod did not compose the same kind of magniloquent epic as Homer. His style was recognized in antiquity as ‘smooth’ and sweet; Dionysius of Halicarnassus even calls it ‘girly’ (παρθενωπά).3 Hesiod’s interest in euphonics and rhythm is well illustrated, for example, in the rich anaphora, assonance and paronomasia that characterize the following four verses from the proem of the Works and Days: ῥέα μὲν γὰρ βριάει, ῥέα δὲ βριάοντα χαλέπτει, ῥεῖα δ’ ἀρίζηλον μινύθει καὶ ἄδηλον ἀέξει, ῥεῖα δέ τ’ ἰθύνει σκολιὸν καὶ ἀγήνορα κάρφει Ζεὺς ὑψιβρεμέτης, ὃς ὑπέρτατα δώματα ναίει. (Op. 5–8) For easily he strengthens, and easily he crushes the strong, easily he diminishes the conspicuous and increases the inconspicuous,

26

Greek Jeweled Style

easily he straightens the crooked and withers the proud – high-thundering Zeus, who dwells in the loftiest abode. (Translation adapted from Most 2018: 87) These are the same sound effects that Ps.-Oppian strives to achieve. Hesiod’s poetry has never been called jeweled, and there is no reason why such a label should be made to apply to Ps.-Oppian, who, apart from the repetitions and alliterations, does not exhibit other markers of the jeweled style such as the emphasis on visuality or paradox.4 The Greek poets of the third century display an ideology of continuity which needs to be taken into account here. Ps.-Oppian tells us that he is, ironically, treading an entirely new path while following into the footsteps of Callimachus, who is in turn revealed to be following into the footsteps of Hesiod. Some metrical and stylistic aspects do evolve over time (for example, Ps.-Oppian’s sentence structure is more nominal than Hesiod’s), but there is no clear rupture. Because the tradition in which Greek poets insert themselves already includes a poetic revolution (that of the Hellenistic period), even as they proclaim new departures, they can neither re-revolutionize the tradition without taking into account the Hellenistic transformations nor seamlessly continue the archaic tradition as if that revolution had not happened.5 Another third-century poet provides the most striking example of perpetuating not only the style but also the very persona of his archaic model. Quintus Smyrnaeus’ Posthomerica is an epic poem which bridges the gap between the Iliad and the Odyssey by narrating the events that took place in Troy after the death of Hector and up until the sacking of the city and the Achaeans’ departure. Lacking a proem, this epic presents itself as ‘still’ the Iliad, picking up the story from Hector’s funeral, exactly where the Iliad ended (see Maciver 2012: 29).6 In spite of attempts to see in Quintus (especially in some of his speeches) the influence of the rhetorical education that is the hallmark of the Second Sophistic,7 the Posthomerica is neither rhetorical in style nor in any other way a precursor of later aesthetics. The narrator of the Posthomerica not only speaks like Homer, reproducing Homeric language and style, but also inserts a pseudoautobiographical interlude whose details could easily be ascribed to ‘the real’ Homer; that is, he presents himself as receiving poetic inspiration while shepherding his flock on a hill in Smyrna, one of Homer’s traditional birthplaces. As scholars have often noted, these lines recall both Hesiod’s own autobiographical investiture in the Theogony (22–8) and Callimachus’ recollection of the Hesiodean scene in the Aetia (fr. 2.1f.).8 Even in this supremely Homeric epic, Quintus cannot just write ‘as if ’ he is Homer: he has to contend with Callimachean criticism of the ‘continuous song’ – the thematically and stylistically traditional epic that he is writing. That Quintus’ Homer (or Homeric persona) is also a ‘Hesiod’ (at least within the in-proem) could tell us something more about the Posthomerica’s style: not that it strives after euphonics and rhythm like Ps.-Oppian (who sits at the opposite end of the stylistic spectrum), but that it is heavily invested in gnomae and a general concern for propriety and moderation (stylistic and otherwise).9 In this case, the presence of Hesiod keys us 27

A Late Antique Poetics?

into an intellectualized didactic mode (with a focus on ethics), which is pervasive in the epic. Probably the last poem that can be dated to the third century (or the very early fourth century), Triphiodorus’ epyllion on the capture of Troy, represents a final ‘opportunity’ for tracing a jeweled style in Greek poetry before its emergence in Latin poetry.10 Like Quintus, Triphiodorus derives most of his vocabulary from Homer and deals with an eminently (post-)Homeric topic. His poem, however, is written on a smaller scale, is more Alexandrian in metre (see Whitby 1994: 119), and is more fond of lexical experiments including non-Homeric uses of Homeric words and apparent neologisms.11 The following four lines come from Triphiodorus’ only ecphrasis, the construction of the Trojan Horse by Epeius: ὀφθαλμοὺς δ’ ἐνέθηκε λιθώπεας ἐν δυσὶ κύκλοις γλαυκῆς βηρύλλοιο καὶ αἱμαλέης ἀμεθύσσου· τῶν δ’ ἐπιμισγομένων διδύμης ἀμαρύγματι χροιῆς γλαυκῶν φοινίσσοντο λίθων ἑλίκεσσιν ὀπωπαί. (69–72) The eyes he set, gem-eyes, in two circles of blue-green beryl and blood-red amethyst; as the colors mingled in a sparkle of double hue, the eyes were red enveloped in the blue-green gems. In just these four verses two words are never attested before Triphiodorus (λιθώπεας in l. 69 and αἱμαλέης in l. 70). The first one is a striking combination of λίθος (gem) and ὤψ (eye) that is doubly appropriate in its context, as the Horse’s eyes are made of gems and the gems themselves sparkle like eyes. Triphiodorus’ concern here is with showcasing language at the expense of visual clarity: we have to reach the last verse, in this convoluted sentence, to be able to understand what configuration of stones is implied: a circle of red amethysts surrounded by green beryl. Can we, then, speak of a jeweled style in Triphiodorus? We can probably speak of a specific jeweled passage, but Triphiodorus’ epyllion, as a whole, displays neither the fragmentation nor (outside the ecphrasis) any emphatic attention to visuality and detail. In many respects, Triphiodorus’ poetics and stylistic register create, as Calum Maciver (2020: 183) has recently put it, a paradoxical ‘collusion’ between the hyper-Homeric and the Alexandrian. Nor is Triphiodorus’ style as strikingly rhetorical as Ps.-Oppian’s. See, for example, how the Cynegetica describes the same combination of colours, green and red, in the eyes of the leopard: γλαυκιόωσι κόραι βλεφάροις ὕπο μαρμαίρουσαι, γλαυκιόωσιν ὁμοῦ τε καὶ ἔνδοθι φοινίσσονται, αἰθομέναις ἴκελαι, πυριλαμπέες. (3.70–2)

28

Greek Jeweled Style

The pupils glow in blue-green beneath the eyelids, blue-green at once and red within, flaming, bright like fire. (Translation adapted from Mair 1928: 119) Like Triphiodorus, Ps.-Oppian uses precious vocabulary (πυριλαμπέες is a rare word, first attested in Aratus, Phaen. 1040), but his verbal patterning is more rhythmical as the first two verses here begin with the same word and end in the same syllable. Ps.-Oppian and Triphiodorus adopt different elements of what we might call the jeweled style, but neither poet espouses that style entirely. And yet a jeweled style was arguably available in Greek literature – indeed in literature earlier than both Ps.-Oppian and Triphiodorus. To find it, however, we should look not to poetry but to prose. The second-century novel by Achilles Tatius displays most if not all of the characteristics of the jeweled style: fragmentation of the narrative by frequent digressions, paradoxical juxtapositions, short, often antithetical clauses and dazzling visuality expressed in brilliant verbal patterning.12 As a brief example that resonates with the description of the horse’s eyes in Triphiodorus, we can take the bridal necklace which appears in the second Book of the novel and which consists of a set of stones, three of which are placed on top of each other to form a pendant shaped like an eye: ἤριζον δὲ πρὸς ἀλλήλους οἱ λίθοι. ὑάκινθος μὲν ῥόδον ἦν ἐν λίθῳ, ἀμέθυσος δὲ ἐπορφύρετο τοῦ χρυσοῦ πλησίον. ἐν μέσῳ δὲ τρεῖς ἦσαν λίθοι, τὴν χροιὰν ἐπάλληλοι· συγκείμενοι δὲ ἦσαν οἱ τρεῖς· μέλαινα μὲν ἡ κρηπὶς τοῦ λίθου, τὸ δὲ μέσον σῶμα λευκὸν τῷ μέλανι συνυφαίνετο, ἑξῆς δὲ τῷ λευκῷ τὸ λοιπὸν ἐπυρρία κορυφούμενον· ὁ λίθος δὲ τῷ χρυσῷ στεφανούμενος ὀφθαλμὸν ἐμιμεῖτο χρυσοῦν. (2.11.2–4) The gems were at rivalry with one another. The jacinth was a rose in stone. The amethyst shone purple next to gold. In between were three gems of graded colour; the three of them were set together. The base of the gem was black; the one in the middle component was white woven with black; and next to the white the final crowning [stone] glowed like fire. The gem was wreathed with gold and looked like a golden eye. Achilles Tatius creates a dazzling verbal picture, crafted out of very short clauses, to describe the glittering ‘necklace of varied stones’ (περιδέραιον μὲν λίθων ποικίλων, 2.11.2). Achilles’ style of poikilia (not by any means unique to his ecphrases) is patently more jeweled than Triphiodorus’: as well as being rich in alliteration, it is also interested in lexical innovation (the verb ἐπυρρία, although not looking remarkable, is first attested here), and it is equally, if not more, fond of paradox. In the first line here, for example, the sentence ὑάκινθος μὲν ῥόδον ἦν ἐν λίθῳ reads at first sight, ‘the hyacinth was a rose in stone’.13 Achilles’ ‘rose’ refers not to the shape but the colour of this flower, so the author could have simply used a different term for ‘red’, had he not meant to catch his readers off 29

A Late Antique Poetics?

guard and make them wonder at the paradox of ‘the hyacinth was a rose’. Achilles’ notion that the gems rivalled each other (ἤριζον) while being set together (συγκείμενοι) is also typical of the concordia discors that Roberts (1989: 144–7) has set at the heart of late Latin poetics. If a jeweled style can be detected in Greek literature before the fourth century, its more natural expression is not in poetry but prose. As the following section goes on to trace this style in fourth- and fifth-century Greek poetry, the same tendency will be confirmed: prose will be shown to be just as or even more jeweled than poetry and importantly, it will have clearer and earlier precedents for its jeweled-ness.

The Greek jeweled style in the fourth and fifth centuries The most important Greek poet of the fourth century is without any doubt Gregory of Nazianzus. No justice can be done here to his enormous and multi-generic output spanning both poetry and prose. In his poetry, Gregory is able to adopt different styles (and metres) according to the genre in which he writes: he can be an epigrammatist, a poet of invective compositions in iambics and a didactic poet composing in hexameters.14 Like the third-century poets examined above, Gregory contends with both the archaic and the Hellenistic tradition, but, as he composes in a variety of modes and genres, he is able to pick up and adapt different poetic personas according to what he aims to achieve in each poem. In an elegiac poem that attacks his enemies, for example, he ‘becomes’ Callimachus: βαρὺς γὰρ αὐτοῖς, καὶ θράσους ἤμην γέμων. τέμνων ὁδοὺς ἀτρίπτους, ἐθῶν πατρῴων καὶ νόμων διαφθορεὺς, εἴπερ νόμος τὰ φαῦλα, πλούτου, τύφου τε, θρύψεως, φιλαρχίας, τῶν νῦν ἐπικρατούντων. (PG 37.1410.4–9 = carm. 2.1.68.15–20) I was too heavy for them and full of audacity. Carving untrodden paths, corrupting our ancestral customs and laws, (if what is bad can be law) wealth, vanity, debauchery, lust for power, everything that is now in charge. Gregory’s ‘untrodden paths’ (ὁδοὺς ἀτρίπτους) are obviously those of Callimachus,15 whose rebellious persona Gregory adopts here in order to figure himself as a revolutionary ecclesiastic trampling on ancestral customs. There are some markers of the jeweled style in this passage: the syntax relies on nouns and participles (five verses have no main verb) 30

Greek Jeweled Style

and there is a striking enumeration in line 19. However, we notice the jeweled style most prominently within Gregory’s poetry in a group of didactic poems entitled by their editors (Moreschini and Sykes 1997) as poemata arcana. These deal with theological topics and are written in hexameter. The following verses are from a poem entitled ‘On the Testaments and the Coming of Christ’: ἀλλὰ κενώσας ὃν κλέος ἀθανάτοιο Θεοῦ Πατρὸς Υἱὸς ἀμήτωρ αὐτὸς καὶ δίχα πατρὸς ἐμοὶ ξένος υἱὸς ἐφάνθη· οὐ ξένος, ἐξ ἐμέθεν γὰρ ὅδ’ ἄμβροτος ἦλθε βροτωθεὶς παρθενικῆς διὰ μητρὸς, ὅλον μ’ ὅλος ὄφρα σαώσῃ. καὶ γὰρ ὅλος πέπτωκεν Ἀδὰμ διὰ γεῦσιν ἀλιτρήν.

(40)

(PG 37.459.13–460.3 = arc. 8.39–44) but having emptied himself of his glory, the son without a mother of the ever-living God the Father, he revealed himself to me, a strange kind of son without a father; no, not strange, for on my account he who is immortal became a mortal through his virgin mother, so that He, whole, may save me wholly. For Adam had also fallen whole due to his sinful tasting [of the fruit]. Gregory revels in the many paradoxes of Christ’s incarnation, especially his having no mother in heaven and no father on earth. The richly adorned style, full of alliterations and polyptota (40f.: Πατρὸς Υἱὸς ἀμήτωρ . . . δίχα πατρός; 41f.: ξένος . . . οὐ ξένος, ἐξ ἐμέθεν; 42: ἄμβροτος . . . βροτωθείς; 43f.: ὅλον . . . ὅλος . . . ὅλος), emphasizes the juxtaposition of contradictory elements in Christ’s nature. How can we explain Gregory’s recourse to a markedly more jeweled style in these poems? Only part of the answer can come diachronically from within the poetic tradition: that is, because these poems are didactic, they can more easily be affiliated with the Hesiodean, ‘sweet’ style of epic poetry and thus embody the continuity of developments already under way in Ps.-Oppian. Another part of the answer, perhaps contextually and synchronically more relevant, is that these poems communicate directly with Gregory’s own prose, specifically with a group of theological orations in which we can also find this same jeweled style.16 A passage from Gregory’s Third Theological Oration compares particularly well with the lines cited above: ἐγεννήθη μέν, ἀλλὰ καὶ ἐγεγέννητο· ἐκ γυναικὸς μέν, ἀλλὰ καὶ παρθένου. τοῦτο ἀνθρώπινον, ἐκεῖνο θεῖον. ἀπάτωρ ἐντεῦθεν, ἀλλὰ καὶ ἀμήτωρ ἐκεῖθεν. ὅλον τοῦτο θεότητος. ἐκυοφορήθη μέν, ἀλλ’ ἐγνώσθη προφήτῃ καὶ αὐτῷ κυοφορουμένῳ, καὶ προσκιρτῶντι τοῦ λόγου, δι’ ὃν ἐγένετο. ἐσπαργανώθη μέν, ἀλλ’ ἀποσπαργανοῦται τὰ τῆς ταφῆς ἀνιστάμενος. (Or. 29.19) He was born – but he had already been born; from a woman – but a virgin. This [is] human, that divine. Without a father here, but with no mother there. All of this

31

A Late Antique Poetics?

[is] of the Godhead. He was carried in the womb – but he was recognized by the prophet [John the Baptist], himself still in the womb, leaping before the Word, for whose sake he came into being. He was wrapped in swaddling clothes [Lk. 2.41] – but he took off the swaddling clothes of the grave when he rose again. Similar to the language and style that Achilles Tatius uses in his novel, this is a very poetic prose, built on the basis of short, antithetical clauses which frequently omit any verb and are rich in repetition and alliteration.17 Gregory not only varies the lexicon he inherits from the Bible,18 but also uses the same striking vocabulary that features in his poetry to articulate the same paradoxes about Christ’s nature (e.g. ἀπάτωρ . . . ἀμήτωρ). Gregory of Nazianzus is by no means unique among Christian prose writers in employing such a jeweled style. Morwenna Ludlow (2020) has recently shown that such a style is also ubiquitous in many orations by Gregory of Nyssa and Basil of Caesarea, and in what follows (in the section on jeweled Christian prose below) it will even be found in Christian commentaries. Where Gregory is innovative is the degree to which he adapts this same style for his poetry. As suggested above, Greek poetry up until the early fourth century may display some elements of the jeweled style, but no single poem can be termed jeweled as straightforwardly as some of Gregory’s poems. Trying to establish whether Gregory composed his didactic poetry having his own prose in mind or even in a book next to him is likely a futile task,19 but it is important for what follows to bear in mind that when a full-fledged jeweled style emerges in Greek poetry, it does so in some of the poems of a prolific author who also addressed the same topics in jeweled prose. It is only in the fifth century, with the gargantuan Dionysiaca by Nonnus of Panopolis, that the jeweled style in Greek poetry finally comes into its own, and it is this specifically Nonnian jeweled style that we find later in the sixth century, in the brilliant ecphrases of Christodorus,20 John of Gaza21 and Paul the Silentiary,22 which, for lack of space, cannot be discussed here. Ecphrases and an ecphrastic mode of composition are prevalent in the Dionysiaca itself, whose extremely variegated style, episodic composition and attention to detail make it the most impressive (as well as the most extensive) embodiment of the jeweled style in Greek poetry. Rather than look at one of Nonnus’ ecphrases of works of art (the shield of Dionysus or the necklace of Harmonia), where questions of allegory and symbolism tend to interfere with stylistic considerations,23 a selection of verses from two descriptions of cities will be taken as an example here: the two coastal cities, Tyre in Book 40 and Beirut in Book 41, are presented in extravagant language that dwells on their paradoxical confounding of land and water.24 Tyre is introduced as a wonder (θάμβος, 40.316) and compared to a swimming girl who is, strikingly, immobile while embraced by Poseidon: νηχομένῃ δ’ ἀτίνακτος ὁμοίιος ἔπλετο κούρῃ, καὶ κεφαλὴν καὶ στέρνα καὶ αὐχένα δῶκε θαλάσσῃ, χεῖρας ἐφαπλώσασα μέση διδυμάονι πόντῳ, γείτονι λευκαίνουσα θαλασσαίῳ δέμας ἀφρῷ, καὶ πόδας ἀμφοτέρους ἐπερείσατο μητέρι γαίῃ. 32

(320)

Greek Jeweled Style

καὶ πόλιν Ἐννοσίγαιος ἔχων ἀστεμφέι δεσμῷ νυμφίος ὑδατόεις περινήχεται, οἷα συνάπτων πήχεϊ παφλάζοντι περίπλοκον αὐχένα νύμφης. καὶ Τύρον εἰσέτι Βάκχος ἐθάμβεε, τῇ ἔνι μούνῃ βουκόλος ἀγχικέλευθος ὁμίλεε γείτονι ναύτῃ . . . καὶ ἔβρεμεν εἰν ἑνὶ χώρῳ φλοῖσβος ἁλός, μύκημα βοῶν, ψιθύρισμα πετήλων, πεῖσμα, φυτόν, πλόος, ἄλσος, ὕδωρ, νέες, ὁλκάς, ἐχέτλη, μῆλα, δόναξ, δρεπάνη, σκαφίδες, λίνα, λαίφεα, θώρηξ. καὶ τάδε παπταίνων πολυθαμβέα ῥήξατο φωνήν· ‘νῆσον ἐν ἠπείρῳ πόθεν ἔδρακον;’

(325)

(335)

(Dion. 40.319–38) Immobile, she is like a swimming girl, and gives to the sea head and breast and neck, in the middle stretching her arms over the two waters, and whitening her body with foam from the sea beside her, while she rests both feet on mother earth. And Earthshaker holding the city in a firm bond floats all about, a watery bridegroom, as if clasping the neck of his bride in a splashing arm, embracing her. Still more Bacchus admired the city of Tyre; where alone the herdsman’s way was near the fisherman, . . . and in one place was the loud noise of the sea, the lowing of cattle, the whispering of leaves, rigging, trees, navigation, forest, water, ships, lugger, plowtail, sheep, reeds, sickle, boats, lines, sails, corselet. As he surveyed all this, he thus expressed his wonder: ‘How’s this – how do I see an island on the mainland?’

(320)

(325)

(335)

(Translation adapted from Rouse 1940: 177–9) The description is strikingly visual and erotic; it insists on an abundance of details that can overwhelm the reader (or listener) and culminates in three verses consisting solely of asyndeta (vv. 334–6), which create a cumulative effect.25 The description of Beirut in the following Book opens with another set of asyndeta,26 followed by a series of juxtaposed agricultural activities (which are not printed below), which then lead into the metaphor of the swimming girl embraced by Poseidon: ἔστι πόλις Βερόη, βιότου τρόπις, ὅρμος Ἐρώτων, ποντοπαγής, εὔνησος, ἐύχλοος, οὗ ῥάχις ἰσθμοῦ στεινὴ μῆκος ἔχοντος, ὅπῃ διδύμης μέσος ἅλμης κύμασιν ἀμφοτέροισιν ἱμάσσεται ὄρθιος αὐχήν· ...

(15)

33

A Late Antique Poetics?

πήχεϊ μυδαλέῳ περιβάλλεται ὑγρὸς ἀκοίτης, πέμπων ὑδατόεντα φιλήματα χείλεσι νύμφης· καὶ βυθίης ἀπὸ χειρὸς ὁμευνέτις ἠθάδι κόλπῳ ἕδνα Ποσειδάωνος ἁλίτροφα πώεα λίμνης δέχνυται, ἰχθυόεντα πολύχροα δεῖπνα τραπέζης, εἰναλίῃ Νηρῆος ἐπισκαίροντα τραπέζῃ

(30)

(Dion. 41.14–35) There is a city Beroë, the keel of human life, harbour of the Loves, firm-based upon the sea, with fine islands and finely green, with a ridge of isthmus narrow and long, and between the two seas the rising neck is beaten by the waves of both. ... and her watery husband embraces [her neck] with wet arm, (30) putting moist kisses on the bride’s lips; she, his bedfellow, accepts in her well-accustomed bosom from his watery hand Poseidon’s bride-gifts, the sea-bred flocks of the waters, the fishes of many colours for her banqueting-table, which dance on the table of Nereus in the brine. (Translation adapted from Rouse 1940: 197–9) The last three verses here are entirely taken up by an extravagant pleonasm about the fish that Poseidon gives to Beirut. This excessive string of nouns, adjectives and participles (which is very characteristic of Nonnus’ style) seems to have no point other than to draw attention to the language itself, to the poet’s mastering of his chosen style and to his own poetics. The ‘many-coloured’ (πολύχροα)27 fish resonate with the varied, multihued style – so fond of juxtaposing synonyms – that characterizes the poem as a whole.28 These three verses have no particles or conjunctions, thus forming a unit within themselves. The last verse is a four-word hexameter in which the first and last words are joined syntactically, further isolating the verse from what precedes and follows it. This type of verse, the versus tetracolus, can be a useful index for the jeweled style: since such verses usually include very long, often striking words and, comparatively speaking, tend to be epexegetical rather than narrative, it can be assumed that the higher their frequency, the more ostentatious the language of a text and the more fragmented its narration.29 Vergil’s Aeneid only has four such verses. Greek poetry apparently begins from a more jeweled ‘baseline’, as Homer has one versus tetracolus in every sixty-five. With Hesiod, the percentage increases to one in every forty-four verses; with Apollonius of Rhodes, it rises again to one in twenty-nine; with Ps.-Oppian, it reaches one in twenty-one; with Triphiodorus, one in seventeen; and, finally, in the Dionysiaca, one in every fifteen verses is a tetracolus. There is, then, a gradual development within the poetic tradition of specific techniques which allow Nonnus to articulate his poem in this highly ecphrastic, discontinuous mode. There is also a clearly increasing tendency to view the world as paradox and to

34

Greek Jeweled Style

express this view through antithesis and with striking lexical innovation. Beirut, for example, is described in a one-word paradox as ποντοπαγής (Dion. 41.16: fixed firmly in the sea). The adjective was apparently coined by Nonnus and is an absolute hapax; yet both its form and the type of thinking it expresses have precedents in Gregory of Nazianzus, a poet that Nonnus read and imitated (see D’Ippolito 1994 and Simelidis 2016: 298–307). Gregory, in the same poem mentioned above, describes Christ as an αὐτοπαγὴς βροτός (arc. 8.69: a self-formed mortal) – that is, one who mysteriously constituted himself in the womb of his mother. The Dionysiaca is a poem that ‘feeds’ on almost all previous poetic genres, incorporating hymns, epigrams, bucolic, didactic and of course mythological epic from Homer to Triphiodorus.30 Its style also participates in, and brings to its apex, an organic development (within the poetic tradition) towards an ever-more jeweled style – a development that received extra impetus from contact with prose, especially in the poetry of Gregory of Nazianzus, Nonnus’ important stylistic predecessor. Even in such an omnivorous mega-poem as the Dionysiaca, however, the direct influence of jeweled prose cannot be disregarded. Just as the Christian paradoxes of Gregory’s theology are equally conveyed within his poetry and his prose (and with recourse to the same jeweled style), the urban and natural paradoxes of late antique city encomia were communicated in both modes of expression: the genre of patria encompassed both poetry and prose (see Dagron 1984: 9–12 and Focanti 2016: 485f.). Scholars have often argued that Nonnus’ descriptions of Tyre and Beirut were directly inspired from Achilles Tatius, who had proclaimed Tyre a ‘city in the sea and an island on land’ (2.14.4: πόλις ἐν θαλάσσῃ καὶ νῆσος ἐν γῇ),31 and who inserted this paradoxical series of antitheses in his encomium of the Nile: Νεῖλος ὁ πολὺς πάντα αὐτοῖς γίνεται, καὶ ποταμὸς καὶ γῆ καὶ θάλασσα καὶ λίμνη. καὶ ἔστι τὸ θέαμα καινόν, ναῦς ὁμοῦ καὶ δίκελλα, κώπη καὶ ἄροτρον, πηδάλιον καὶ δρέπανον, ναυτῶν ὁμοῦ καὶ γεωργῶν καταγωγή, ἰχθύων ὁμοῦ καὶ βοῶν. ὃ πέπλευκας, φυτεύεις, καὶ ὃ φυτεύεις, τοῦτο πέλαγος γεωργούμενον. (4.12.1–5) The great Nile is everything to them – river and land and sea and lake. And it is a new sight: close together the boat and the hoe, the oar and the plough, the rudder and the sickle – the common lodging of sailors and farmers, of fishes and oxen. You sow where you have sailed, and where you sow it is a cultivated sea. Nonnus may indeed have read Achilles Tatius or even drawn inspiration from these particular passages,32 but the intermingling of earth and water, the figuring of coastlines as paradoxa and the praise of coastal cities as swimming nymphs would have likely been common if not ubiquitous in city encomia.33 In the fourth century, the rhetor Himerius, an author also responsible for a protreptic discourse on the value of poikilia (Or. 68), addresses a prose encomium to Constantinople:

35

A Late Antique Poetics?

σὲ μὲν καὶ Ποσειδῶν ὁ βασιλεὺς ὁ θαλάσσιος γλαυκοῖς περιβάλλει τοῖς κύμασιν, οἷά τινα νύμφην Ναΐδα, καὶ πανταχόθεν περιπτύσσει καὶ γέγηθε· σὲ δὲ καὶ Νηρηΐδων ἁλιπορφύρων χοροὶ ἄκροις ἐπισκιρτῶντες τοῖς κύμασι κύκλῳ περὶ πᾶσαν χορεύουσι. (Or. 62.2) Poseidon, king of the sea, surrounds you with his bluish waves, as if you were a Naiad nymph; he embraces you on every side and rejoices in you. Choruses of seapurple Nereids, leaping on the surface of the waves, dance around the whole of you in a circle. (Translation: Penella 2007: 46f.) Having proclaimed himself a new Pindar for a new Athens (Constantinople) at the beginning of the speech, Himerius consistently appropriates poetic vocabulary (here, ἁλιπορφύρων) and evokes dazzlingly jeweled images. In this oration, he will go on to compare the city’s walls to a golden diadem that the emperors have placed on her head. In a different speech, again for Constantinople, the city itself is compared to a golden necklace, in which the gems that are its various architectural wonders – the senate, baths and theatres – vie for the spectators’ attention (Or. 41.7).34 Himerius’ Constantinople, like the two Nonnian cities, is also a city paradoxically founded on water. Himerius describes how the city expanded towards the coast and ‘has turned what by its nature is rolling and constantly on the move [the sea] into something immobile’ (Or. 41.6: σαλεύουσαν φύσιν καὶ ἄστατον πεπηγέναι πεποίηκεν). Achilles Tatius, Himerius and Nonnus use a similarly jeweled style to praise similarly paradoxical landscapes. That the two prose authors are earlier (and Achilles Tatius significantly so) is a likely indication that it was in prose, not in poetry, that this style and the type of thinking it usually expresses were developed and sustained. To a significant extent the Greek jeweled style goes back to Gorgias, whose extremely figured speech was used as a vehicle precisely for paradoxical arguments such as, in his Encomium of Helen, ‘Even if she followed Paris willingly, Helen should be considered entirely innocent.’35 The Gorgianic style, later termed ‘Asianic’ by the Romans, would never go entirely out of fashion from the moment it was introduced to Athens in the fifth century.36 Literary fragments, as well as encomia surviving in inscriptions, from the Hellenistic period onwards attest to its continued appeal.37 It later became the style of choice for many orators of the Second Sophistic and other prose authors of the same period, such as Achilles Tatius.38 The development of a corresponding style in poetry must have been propelled by the growing parallel employment of both prose and poetry for the same functions.39 Gregory reflects on the same theological paradoxa in prose and poetry. Ecphrases and encomia, including patria and panegyrics, were equally written in prose and verse. It is significant that Nonnus’ Dionysiaca, for all its poetic ‘inter-genericity’ (cf. Harrison 2007), presents itself as a panegyric for Dionysus – an epic panegyric which, Nonnus alleges, Homer himself should have composed instead of singing the praises of Achilles.40 Nonnus pursues a poetics that is marked by discontinuity and rejection of (or 36

Greek Jeweled Style

fierce antagonism with) the poetic tradition – a stance that is markedly different from that of earlier Imperial poets.41 Nonnus’ antagonism with the poetic tradition can be construed as an adaptation for poetry of an anti-poetic discourse found in earlier prose panegyrics, which sought to denigrate Homer as the model panegyrist in order to elevate the contemporary speaker and his laudandus.42 Nonnus-the-panegyrist can attack Homer-the-panegyrist because his predecessors working in prose had used this exact same technique. The conceptualization of epic as panegyric, that is, as a declamatory genre which could be expressed in appropriately jeweled prose, allowed epic itself to assume an entirely and unashamedly jeweled guise.

Jeweled Christian prose and un-jeweled biblical poetry The jeweled style in Greek poetry remains mostly wedded to poetic genres that have corresponding prose iterations (the panegyric of Nonnus and the ecphrases of his imitators). The stylistic registers used by Greek biblical poetry seem to confirm this view. Poets writing on biblical topics choose to eschew the jeweled style in the main body of their compositions: the poems of Eudocia and Ps.-Apollinaris are, in different ways, decidedly Homeric, and Nonnus’ Paraphrase is significantly less jeweled than the same poet’s Dionysiaca (see below). Yet both Eudocia’s Homerocentones and the Metaphrasis Psalmorum by Ps.-Apollinaris are prefaced with jeweled hexameters, whose style markedly differs from the Homericizing compositions that follow (see Agosti 2001: 87–92 and 2019: 137). What can account for this change in stylistic register? The very genre of the preface provides at least part of the explanation. It has long been observed that verse prefaces grew out of the rhetorical prolaliae which preceded declamatory performances at least since the second century ce : before taking up his fictional persona, the declaimer would address the audience in his own voice and try to win their favour or address his speech’s composition and structure.43 When this practice was transferred to poetry, it would, eventually, give rise to more informal, iambic prefaces to hexameter poems.44 Iambic is, of course, the most conversational and un-poetic of all Greek metres. In the case of these earlier Christian compositions, the prefaces are not in iambics. They are, nevertheless, more prose-like precisely in that they are jeweled: the space of the declamatory preface afforded poets ‘permission’ to use a style that is more rhetorical and more closely affiliated with artistic prose than the very Homeric modes they would adopt in the rest of their poems. The reasons behind the avoidance of the jeweled style in Greek biblical poetry are complicated and cannot be addressed here in full, but must include both Christian suspicion against ostentation (material and linguistic)45 and also the scandal provoked by the composition of poetry by two arch-heretics: Arius and the historical Apollinaris, the first of whom wrote his notorious Thalia in an extremely rhetorical style.46 Yet, none of these factors stopped Gregory of Nazianzus from writing jeweled didactic emulating his own, jeweled theological orations. Genre is again significant: in paraphrasing or retelling the biblical stories in centos, poets must have consulted commentaries and 37

A Late Antique Poetics?

other Christian prose, but their project was plainly poetic. They aimed, that is, simply to render the Bible in verse and not primarily to expound or expand upon it in the manner of orations and commentaries. There can be no prose equivalent to these purely poetic enterprises. My final illustration of the extent to which the Greek jeweled style is predominantly tied to prose is Nonnus’ Paraphrase of the Gospel According to John. Penned by the same poet as the extravagant Dionysiaca, this is a surprisingly un-jeweled poem. That is not to say that the jeweled style is entirely absent from the poem: lacking a preface by the poet himself,47 the epic begins with a ‘hymn to the Logos’ which significantly expands upon John’s corresponding proclamation and which is reminiscent of Gregory’s didactic verses, especially as it also terms the Logos ‘a son without a mother’ (1.2: υἱὸς ἀμήτωρ). Such jeweled passages, however, are few and far between in the poem.48 The Paraphrase follows its Vorlage very closely, often limiting itself to lexical variation and the addition of adjectives to the biblical text. These adjectives can perform important theological work,49 but they are not enough for us to qualify the poem’s stylistic register as jeweled. Versus tetracoli appear less frequently in the Paraphrase than in the Dionysiaca (see Faulkner 2020: 29f.). The pace of the biblical narrative is respected and not interrupted by long digressions and ecphrases. There is but one full-fledged ecphrasis in the whole poem (the lanterns, in 18.18–24), and even that only takes up seven verses.50 This lack of attention to visuality is indeed striking, especially considering the poet’s own practices in the Dionysiaca. The healing of the man born blind is an apt example. Recounted in John 9.6, this miracle provided fertile ground for homilists and commentators to expound on the importance of vision and on the beauty of creation as proof of the wisdom and providence of the Creator. In the fourth century, John Chrysostom dedicates two homilies to the miracle, one of which waxes lyrical on the eye as the superior human organ (PG 59.307.46–308.23). Also from the fourth century is a homily by a less well-known bishop, Asterius of Amaseia, who takes the opportunity to insert an extended ecphrasis of the eye in his interpretation of the miracle. This ecphrasis unfolds over three paragraphs (7.3–5), dwelling on the eye’s glistening, mirror-like surface and its intermingling of colours (7.4.2: ποικίλοις τισὶ κύκλοις, a strange unity of varied circles). The ecphrasis concludes with a quasi-ethopoeia in which the author exclaims, as if in the persona of the newly seeing blind man: διὰ τούτου ἔχω γνῶσιν ἡλίου καὶ οὐρανοῦ κόσμον ἔμαθον καὶ κάλλος ἀστέρων ἱστόρησα καὶ γῆς ὑπόστασιν καὶ φύσιν θαλάσσης, σπερμάτων διαφορὰν καὶ φυτῶν ποικιλίαν καὶ χρωμάτων ποικίλην βαφήν, σκότους κατήφειαν καὶ φωτὸς λαμπήδονα καὶ πάντα ἁπλῶς ὅσα ὁ Θεὸς κτίσας ἐπῄνεσεν. (7.5.1) Because of it [the eye] I know the sun and I have learned the ornaments of sky and the beauty of the stars and the substance of the earth and the nature of the sea, the diverse kinds of seeds and the variety of plants, and the multiple hues of colours, the dejection of darkness and the brightness of light and simply everything that God made and praised. 38

Greek Jeweled Style

Asterius writes a sermon meant for oral delivery in front of an audience of certainly mixed capacities. His lively style, with the accumulation of short clauses and relatively simple vocabulary, is clearly intended to retain his listeners’ attention. Yet, we find a thematically similar passage in the commentary on John’s Gospel written by Cyril of Alexandria in the fifth century. Where Asterius sings the praises of the eye in the abstract, Cyril remarkably puts himself in the place of the healed blind man and proffers a fullfledged, jeweled ethopoeia. This is only its final part: ἰδοὺ λαμπρὸν ἡλίου περιαστράπτει με φῶς· ἰδοὺ ξένων θεαμάτων τὸν ἐμὸν ὀφθαλμὸν περιστοιχίζεται κάλλος. ἄρτι μόλις ἔγνων τὴν Ἱερουσαλήμ, καὶ βλέπω μὲν θεῖον ἐν αὐτῇ περιαστράπτοντα ναόν, ὁρῶ δὲ ἐν μέσῳ τὸ σεπτὸν ὄντως θυσιαστήριον· κἂν ἐπέκεινα πύλης εὑρεθῶ, ὅλην ἐν κύκλῳ περιαθρήσω τὴν Ἰουδαίαν, καὶ τί μὲν ὄρος ἐστί, τί δὲ φυτὸν ἐπιγνώσομαι. μεταμειβομένου δὲ τοῦ καιροῦ πρὸς ἑσπέραν, οὐκ ἔτι τὸν ἐμὸν ὀφθαλμὸν τὸ τῶν ἄνω θαυμάτων διαλήσεται κάλλος, οὐχ ὁ διαφανὴς τῶν ἀστέρων χορός, ἀλλ’ οὐδὲ τῆς σελήνης τὸ πάγχρυσον σέλας. (6.1 on Jn 9.15 = Pusey 1972: 2.165,25–166,8) Lo! the bright light of the sun is shining around me: lo! the beauty of strange sights surrounds my eye. A short time ago I scarcely knew what Jerusalem was like; now I see glittering in her the temple of God, and I behold in its midst the truly venerable altar. And if I stand outside the gate, I can look around on the country of Judea, and shall recognize one thing as a hill and another as a tree. And when the time changes to evening, my eye will no longer fail to notice the beauty of the wondrous objects on high, the brilliant company of the stars and the golden light of the moon. (Translation adapted from Randell 1885: 2.26f.) Nonnus had read Cyril’s commentary before composing his Paraphrase (see Livrea 1989: 25 and Spanoudakis 2014: 18f.). Yet, his rendering of this miracle inserts no ecphrasis or praise of the eye, no ethopoeia of the healed blind man and no amplification of the Gospel to highlight the power of vision to convey knowledge about the creation and the Creator. This is neither to say that Nonnus’ rendition is theologically disinterested, as Michele Cutino (2009: 237f.) claimed, nor that it lacks hermeneutic potential and remains at the level of a school exercise, as Martin Hose (2004: 29) thought in relation to Greek biblical poetry in general. The Paraphrase, relying on the exegesis of Cyril and others,51 turns this miracle into a second Genesis by recalling God’s creation of man: Jesus is said to mould new eyes for the blind man with the ‘familiar mud, out of manbegetting dust’ (9.33f.: ὀφθαλμοὺς τελέων νεοτευχέας ἠθάδι πηλῷ / ἐκ χοὸς ἀνδρογόνοιο,).52 The difference is that here the hermeneutic weight is carried by the adjectives (νεοτευχέας, ἠθάδι and ἀνδρογόνοιο are all significant) and not by a profusion of jeweled language and imagery as in Cyril. The above sample from the supposedly ‘dry’ genre of the commentary is clearly more jeweled than a poem written by the major representative of the jeweled style in Greek poetry. Nonnus was patently not aiming to 39

A Late Antique Poetics?

write a jeweled biblical poem – it is likely that such an enterprise would have been a generic impossibility in Greek poetry.

Conclusion This contribution has argued that a jeweled style emerged in Greek poetry at about the same time as in Latin poetry (during the fourth century), but it did so on a more limited scale, in only some of the poems of Gregory of Nazianzus and with the impetus of the same author’s jeweled prose. It was Nonnus’ Dionysiaca that brought a poetic jeweled style into vogue in the fifth century and influenced a series of imitators in the sixth century. This new style, however, did not sweep over the entire landscape of Greek poetry: it was neither embraced by all poets nor even adopted to the same extent for all poems by the same poet, as Nonnus’ different registers for the Dionysiaca and the Paraphrase suggest. Throughout late antiquity, poets could and did often choose to compose Homericizing poetry – a kind of poetry that has no clear equivalent in the Latin tradition, as it involves adopting a peculiar language and not just a style.53 The Orphic Argonautica is one such example of Homericizing poetry from the secular tradition. An archaic, Homericizing mode was from the beginning (in the Bodmer poems, from the early/mid-fourth century), and remained throughout late antiquity, the main style of choice for poets working on biblical topics.54 Greek poetry responded to the jeweled style in a genre-specific way, allowing this style to flourish in those genres that had equivalent prose iterations, from which the poems could derive, direct or indirect, motivation for and ‘legitimization’ of their jeweled-ness. An epic poem conceived as a panegyric, an ecphrasis in verse praising a public monument and a poetic preface corresponding to an orator’s prolalia all have obvious rhetorical functions which could be equally served by artistic, declamatory prose.55 Greek prose had a jeweled version long before Greek poetry acquired one, and it would continue to be the embodiment of the jeweled style par excellence.

Notes *

I thank the editors of this volume for their meticulous work on my chapter. I also gratefully acknowledge the financial support of FCT (Fundação para a Ciência e a Tecnologia), Portugal, through the project PTDC/LLT LES/30930/2017 (national funds).

1.

For the jeweled style in Quintus’ Posthomerica see the reading of Book 13 by Avlamis (2019: 184–6). It is commonplace to qualify Nonnus’ style as jeweled; see e.g. Agosti (2006) and De Stefani and Magnelli (2011: 557). See also below.

2.

See further Dion. 41.133 (birth of Eros); 48.841 (birth of Dionysus); Par. 3.36 (being reborn in baptism). For the theme of paradoxical generation in the Dion. see Hadjittofi (2016: 149–51).

3.

See Dion. Hal., De comp. 23.16–17 (εὔφωνά τε εἶναι βούλεται πάντα τὰ ὀνόματα καὶ λεῖα καὶ μαλακὰ καὶ παρθενωπά) and the analysis in R. Hunter (2009: 255 and 2014: ch. 6).

4.

For Ps.-Oppian as influenced by rhetoric see Whitby (1994: 111) and Silva Sánchez (1999).

40

Greek Jeweled Style 5.

Cf. specifically on Callimachus’ influence De Stefani and Magnelli (2011).

6.

For Quintus’ impersonation of Homer see now Greensmith (2020).

7.

For this argument see especially Bär (2010), with the antilogia of Maciver (2012b).

8.

See Greensmith (2020: 158–88) with further bibliography. For Homer as a shepherd before his poetic investiture in the Troad, in the vicinity of Achilles’ tomb, see Hermias’ commentary on Plato’s Phaedrus (243a = 1.80.14–26) – a neglected account, which I will analyse elsewhere.

9.

For Quintus’ gnomic style see Maciver (2012: 87–123).

10. Triphiodorus is dated thus on the basis of a papyrus fragment (P.Oxy. 2946) which cannot be later than the very early fourth century; see Miguélez Cavero (2013: 4–6). The relative chronology of Quintus and Triphiodorus is contested although most scholars now agree that Quintus is the earlier poet; see, recently, Carvounis (2019: xx–xxvi) and Greensmith (2020: 24–34). 11. See Miguélez Cavero (2013: 44–6) with further bibliography. 12. Whitmarsh (2020: 40) has called Achilles’ style in parts ‘prose-poetic’. 13. ὑάκινθος means both ‘jacinth’ and ‘hyacinth’, but the second meaning would have been the most natural and frequent one; cf. Whitmarsh (2020: 208). 14. See Cameron (2004: 333) and, especially on Gregory’s awareness of the classical genres and their norms, Kuhn-Treichel (2020). 15. See Aet. fr. 1.27f. (κελεύθους / ἀτρίπτους) and Poulos (2019: 54f.). On Gregory’s Callimacheanism cf. Simelidis (2009: 30–46). 16. Hose (2004: 24f.) argues that much of Gregory’s poetry is a ‘transposition’ of his own prose into verse: the Carmina moralia, for example, ‘transpose’ his paraenetic literature, and the poems about himself (Carmina de se ipso) ‘transpose’ his prose Apologia. A more extensive analysis than can be carried out here would be needed to confirm whether Gregory’s poetry consistently turns more jeweled in those cases where it ‘transposes’ jeweled prose (if the poetry indeed relies on the prose, which is not universally accepted; see n. 19 below). 17. On the musicality of Gregory’s prose, its ‘enumerative method’ and its influence on the style of Marino see Ward (1994). 18. E.g. the ‘leaping’ of John the Baptist is described in Luke 1.41 with the verb σκιρτάω, which Gregory turns into προσκιρτάω. Attestations of the latter verb in other authors are all later. Gregory uses it also in his Funeral Oration for Basil (Or. 43.75) and in two further speeches (Or. 38 and 39; PG 36.329.46 and 36.352.34). 19. For an attempt to read the poetry as depending on the orations see Keydell (1951), who goes so far as to treat the arcana as one continuous poem. Cf. also Hose’s (2004: 24f.) views, as reported above n. 16, and contra Moreschini and Sykes (1997: 58f.), who also note that as a didactic poet, Gregory is at once traditional and innovative. For Gregory’s apophatic poetics in the arcana cf. Meinel (2009). 20. On Christodorus’ Nonnian style see Bär (2012: 469–71); on his relationship with Homer, whom he calls his own father in Anth. Pal. 2.320–3, see among others Kaldellis (2007) and Höschele (2021). 21. On the date of John’s poem see Lauritzen (2015: xii-xviii), on its style Gigli Piccardi (2011). 22. On Paul’s Nonnian metre and style see Fayant (2003). 23. On the shield of Dionysus see, e.g. Spanoudakis (2014b) and Gagné (2019: 196–203). On the necklace of Harmonia see Miguélez Cavero (2017). 24. For the same paradox in late Latin poetry and in connection to the jeweled style see Hardie (2019: 173–85). Contrast Ps.-Oppian’s description of a diverted river in Cyn. 2.116–58,

41

A Late Antique Poetics? where such a paradoxical landscape could have been described but is not as Ps.-Oppian is more interested in providing mythological digressions; for stylistic analysis of part of this passage see Whitby (1994: 112–14). 25. On the ‘cumulative aesthetic’ in late antique art see Elsner (2004: 304–9), for that in literature see Roberts (1989: 66–121) and Miller (1998: 124–30). 26. A more impressive series of asyndeta in praise of Beirut, spanning eight verses, can be found in Dion. 41.143–50. 27. The same adjective is used for the rainbow in Dion. 2.203; for the ‘varied necklace’ crafted by Hephaestus in 5.580 (ποικίλον ὅρμον); and for Dionysus’ shield, again crafted by Hephaestus, in 25.387 (ἀσπίδα δαιδάλλουσα πολύχροον). 28. For Nonnus’ poetics of poikilia see e.g. Fauth (1981) and Miguélez Cavero (2008: 139–45 and 162–8). 29. See Bassett (1919), from whom the following statistics are also derived. 30. For brief overview see Lasek (2016). 31. See the commentaries by Simon (1999: 283f.), Accorinti (2004: 112–14 and 174–5), Chuvin and Fayant (2006: 156), all with further bibliography. Already in Chariton, Tyre is ‘a city founded in the sea’ (7.2.8.1: ἡ μὲν γὰρ πόλις ἐν θαλάσσῃ κατῴκισται). 32. For similar hymns and encomia of the Nile, which are not, however, as stylistically close to Nonnus as Achilles Tatius see e.g. an epigram by Philip of Thessalonica (Anth. Pal. 9.299) and a fragmentary hexameter hymn to the Nile (fourth century), edited by Cribiore (1995), who also notes that some vocabulary used in the hymn will later become standard in Nonnus. The fragmentary hexameter cosmogony found on P.Strasb. 481 is very probably part of a patrion of Hermoupolis Magna and likewise presents similarities with Nonnus’ vocabulary and with his account of the foundation of Beirut; see the recent edition of the text, with rich commentary, by Perale (2020: 33–83). 33. Also in the encomia for those who sponsored such building projects: an anonymous epigram praises a certain Venetius who surpassed Theseus and Pelops by turning Smyrna’s deep sea into dry land (Anth. Pal. 9.670.1: βυθὸν ἠπείρωσε). 34. For further commentary on this speech see Hadjittofi (2014). The motif of the dispersed gaze as a reaction to an architectural or urban wonder appears in two second-century authors: Aelius Aristides, whose praise of Smyrna includes a comparison of the city to a multicoloured necklace that draws the eye of the viewer into different directions (Or. 17.10–11) and Achilles Tatius 5.1.1–5, where Clitophon pronounces his eyes ‘vanquished’ at the sight of Alexandria. After Himerius, it also appears in Procop. Aed. 1.1.47–9; Paul. Silent. 296–9; Choric. Laud. Marc. 2.26. 35. Gorgias’ lost speech On Nature or On Non-Existence also put forward a very paradoxical thesis: it first denied the existence of reality itself; it then denied the possibility of knowing something about reality (if it existed) and communicating that knowledge (if it could ever be acquired). It is no wonder that such an apophatic argument and the style in which it must have been couched appealed to later theologians. 36. On the adoption of this essentially racist view of ‘Asian’ decadence in Eduard Norden’s influential study on Kunstprosa see Ludlow (2020: 13–18). 37. See e.g. Papanikolaou (2009) on a first- or second-century bce inscriptional encomium of Isis and (2012) on an inscription from Lycia, dating to the second century ce and featuring many hapax legomena and other rare terms, including the word θεοτόκος for a miracle in which the earth ‘gave birth’ to two stone images resembling Apollo and Artemis. 42

Greek Jeweled Style 38. For an example from the oratory of the early second century ce see the two declamations of Polemo, both written in the Gorgianic style and indulging in paradoxes; one speech, for instance, proclaims its subject ‘a dead man superior to death’ (2.2.31: νεκρὸν θανάτου κρείττονα) and an ‘ensouled corpse’ (2.12.5: σῶμα ἔμψυχον). 39. Roberts (1989: 49f.) briefly points out that in late antiquity the distinction between poetry and prose progressively broke down. 40. See the ‘proem in the middle’ of Book 25, with Henry (2020: 448–51). For the Dionysiaca as panegyric see Miguélez Cavero (2010) and for the late antique focalization of earlier epic poetry through panegyric Pollmann (1999: 63). 41. On Nonnus’ combative relationship with Homer see especially Shorrock (2001). 42. I analyse this technique in Hadjittofi (2021), also pointing out its genealogy in a historiographical trope first developed in Thucydides’ ‘Archaeology’. 43. See Zuenelli (2019: 46) with further bibliography. 44. In Greek poetry such iambic prefaces are only attested in the sixth-century poems of John of Gaza and Paul the Silentiary. 45. Roberts (1989: 142) argues that some Latin Christian poetry avoided the excesses of the jeweled style to refute ‘any accusation of self-serving stylistic virtuosity’. 46. Faulkner (2020: 1–31) attributes the extant Metaphrasis Psalmorum to the historical Apollinaris, but the question is far from settled. It is impossible to know from the brief mentions of Socrates of Constantinople (Hist. Eccl. 3.15f.) and Sozomen (Hist. Eccl. 5.18.3f.) which style the hexameter paraphrases would have adopted, but Sozomen does also mention lyric poems in the manner of Pindar, which are likely to have been jeweled compositions. 47. I explore the implications of this absence in Hadjittofi (2020). 48. For some such passages see Par. 3.16–47 (the paradox of the second birth in baptism), 5.79–82 (on the resurrection) and 19.139–45 (the Beloved Disciple is appointed as the son of the Virgin). 49. For an example of how the employment of single adjectives affects Nonnus’ Christology see Hadjittofi (2018). 50. On this ecphrasis, its links with the visual world of late antiquity and its symbolic interpretation see Agosti (2014: 159f.) and on its exceptional status in the poem Verhelst (2019: 12–14). 51. For full details and references see Hadjittofi (2020: 83 n. 70). 52. Cf. the emphatic hemistich ἀνέρος ἔπλασεν ὄμμα repeated in two consecutive verses at 9.30f. 53. Few scholars speak of ‘Vergilizing’ poetry, and one who does use this word refers almost exclusively to Proba’s cento (see Curran 2012). The imitation of Vergil by later poets never extends to, for example, continuing his epic or impersonating him, as Quintus impersonates Homer. 54. For the style of the Christian poems found on the Bodmer codex (Codex Visionum) see Agosti (2002). 55. Peirano Garrison (2019) rightly speaks of a dynamic and competitive interaction between Latin poetry and prose – a model that could well apply to the Greek tradition, too.

43

44

CHAPTER 3 GILDING THE LILY: THE JEWELED STYLE IN PROSE PANEGYRIC Catherine Ware

Roberts’s title assigns the jeweled style to late antique poetry, but a summary of the features he defines as jeweled applies equally well to the prose panegyrics of the period. At the time when the jeweled style was evolving in poetry, orators were striving for the heights of rhetorical invention, vying to outdo themselves and each other in their praise of the emperor. Late antique panegyric is a highly visual form, being in effect the text which accompanies an imperial ceremony. It is naturally episodic and ecphrastic. The narrative is subordinated to praise and halted so that the audience, invited by the orator, may gaze at the emperor. That the jeweled style features in panegyrics would seem self-evident. Isolating jeweled elements in prose encomia, however, is not straightforward. As Roberts (1989: 43 and 63f.) suggests, many elements of the jeweled style derive from epideixis and the showpiece of epideictic oratory is the panegyric. This is further complicated by the gradual shift in panegyric towards poetry and poetic ornamentation. Nevertheless, in the late antique collection of prose panegyrics, the Panegyrici Latini (see Appendix), one can observe elements of the jeweled style evolve and can trace how features which were already present in epideixis, and specifically in encomium, crossed over into poetry, both at the elemental level of the word and in the larger unit of the episode. Setting aside the concept of the jeweled style, there are sound reasons for considering panegyric when studying the development of late antique poetry. In comparison with forensic or deliberative oratory, epideictic had originally no practical purpose and was dismissed as having little more than entertainment value (Cic. de orat. 2.341). However, the Romans quickly found an increasing need for diplomatic and flattering speeches, and for training on how to acquire such rhetorical skills. Although little encomium has survived from before the late third century, it is evident from the letters of Pliny and Fronto and the handbooks of rhetoric that it was becoming an increasingly important part of court ceremonial and called for a particular style: luxuriant, flowery, decorative. It did not escape the writers on rhetoric that the style required for panegyric had much in common with poetry. All genres of oratory were becoming more ornate.1 Quintilian had used the metaphor flosculus to characterize the style of his own time which he saw as a threat to the proper modes of oratory: ne recentis huius lasciviae flosculis capti voluptate prava deleniantur (inst. 2.5.22: they must not be seduced by the ornaments of present-day excess, and grow soft with its corrupting pleasures). In rather similar language, he also warned against the threat of too much description and the danger of the ecphrasis.2 Like the enticing, decorative flosculi, elaborate description is a fault characteristic of young orators, being ‘digressive and excessive in far-fetched 45

A Late Antique Poetics?

descriptions, to which many are drawn, desiring to imitate poetic licence’ (inst. 2.4.3: [ratio narrandi . . .] sinuosa et arcessitis descriptionibus, in quas plerique imitatione poeticae licentiae ducuntur, lasciva). Quintilian rejected the licence of the poets, but Fronto, tutoring Marcus Aurelius on the writing of epideictic, said that verse gave the best assistance to composing speeches (p. 49,22f. [=ad M. Caes. 3.17.3]) and his contemporary Hermogenes declared that ‘all poetry is panegyric’ (Hermog. 389: πανηγυρικὸν γὰρ πρᾶγμα δήπουθέν ἐστι ποίησις ἅπασα). The orators of the Panegyrici Latini were vehement in their rejection of poetic licence (Ware 2017b), but they also recognized Fronto (PL 8[5]14.2) as the glory of Roman oratory, equal to Cicero and, presumably took his rhetorical advice to heart. Macrobius, writing in the early fifth century, would reach for the familiar metaphor of flos, in linking Pliny’s style (on which, cf. Gibson and Rees 2014: 149) with that of Symmachus: pingue et floridum in quo Plinius Secundus quondam et nunc . . . noster Symmachus luxuriatur (Sat. 5.1.7: the rich and florid style in which once Pliny the Younger and now . . . our Symmachus luxuriates). The hundred-year span of the Panegyrici Latini suggests that during the fourth century, prose panegyrics moved closer to poetry (Rees 2017). The final oration, that of Pacatus Drepanius (389 ce ), closes by including poetry within panegyric: accipiet a me argumentum poetica (PL 2(12)47.6: from me poetry will take its subject matter). According to Rees (2017: 343), this statement ‘narrows the gap between [Pacatus’] epideictic prose and poetry’ (cf. Ware 2017b: 367–9). It is not known if the growing poetic impulse in the Panegyrici Latini corpus was due to the taste of the editor or whether it was a natural development in keeping with more widespread changes in prose. In any case, during the fourth century, praise was being written in verse, if not in formal court panegyrics,3 and before the last of the Panegyrici Latini was written, Claudian had delivered his first verse panegyric, which would be the favoured form thereafter. While his poetry is now classified by many scholars as ‘panegyric epic’,4 Claudian exploited the naturally encomiastic element of Roman epic, specifically Vergilian epic, to write his honorands into the poetic imperium sine fine of the Roman Empire (Ware 2012: 27–31). His panegyric on Honorius’ fourth consulship, in fact, proves the truth of Pacatus’ prediction.5 Pacatus’ praise of Spain as the birthplace of Theodosius: cedat his terris terra Cretensis parvi Iovis gloriata cunabulis et geminis Delos reptata numinibus et alumno Hercule nobiles Thebae. Fidem constare nescimus auditis; deum dedit Hispania quem videmus (PL 2[12]4.5) Let the land of Crete give place to this land, Crete which boasts of being the cradle of the infant Jupiter, and Delos also, where twin divinities crawled, and Thebes, famous for its foster-child Hercules. We do not know the truth of what we have heard but Spain has given us a god whom we can see inspired Claudian: 46

Jeweled Style in Prose Panegyric

Herculis et Bromii sustentat gloria Thebas, haesit Apollineo Delos Latonia partu Cretaque se iactat tenero reptata Tonanti; sed melior Delo, Dictaeis clarior oris quae dedit hoc numen regio (IV cons. Hon. [8.]132–6) The glory of Hercules and Bacchus uplifted Thebes, Latonian Delos stayed fixed after the birth of Apollo and Crete boasts that the young Thunderer crawled there. But better than Delos, more renowned than Cretan shores is this land which gave this god. The quantity of mythological names alone makes the ornamental advantage of poetry obvious. Pacatus’ reference to the devoutly Christian Theodosius as deum dedit could be excused as a translation of his name (Saylor Rodgers 1986: 93f.) but in poetic praise, Claudian had the freedom to refer to his numen and link him with Hercules, Bacchus, Apollo and Jupiter, as well, perhaps, as suggesting a parallel with the greatest emperors (felicior Augusto, melior Traiano) in melior Delo, Dictaeis clarior (cf. Ware 2017b: 367f.). The jeweled style would seem a perfect match for panegyric, a genre designed to give pleasure and delight to the listener (Cic. part. 72). Fronto had advised Marcus Aurelius that there was no place for the plain style in epideixis. All must be utterly luxuriant: omnia ἐν τῷ ἐπιδεικτινῷ ἁδρῶς dicenda, ubique ornandum, ubique phaleris utendum; pauca τῷ μέσῳ χαρακτῆρι (p. 49,16f. [=ad M. Caes. 3.17.2]) In the epideictic speech everything must be said in luxuriant style, everywhere there must be ornament, everywhere trappings must be used. The medium style admits but sparingly of these. (Translation by T. R. Haines 1919: 1.107) Fronto’s language is itself luxuriant. Each of the balancing clauses, starting with omnia and pauca, has its Greek phrase, set like a precious stone in the plainer Latin background.6 Within is anaphora (ubique . . . ubique) and homoioteleuton (ornandum, utendum). This is language which anticipates the jeweled style in imperial panegyric. Phalerae, too, is the perfect metaphor in this regard, being both a personal adornment and an ornament of speech: ‘medallion’ is commonly used by scholars to describe the jeweled style (Fontaine 1981: 6, Elsner and Hernández Lobato 2017: 11). Another common metaphor for the jeweled style is met in the panegyric of 321, when the orator wishes for the ability to adorn his oration with all the ‘flowers of speech’, flosculos . . . dicendi (PL 4[10]30.3). The claim to lack talent is a commonplace of encomium, designed to draw attention to the orator’s skill and the use of flos, flosculus and cognates such as floridus or florens to describe the epideictic style goes back to Cicero (e.g. 47

A Late Antique Poetics?

orat. 19.65; 27.96). For Roberts (1989: 48–53), however, such words would be a key indicator of the jeweled style in late antique poetry, indicating the variatio of colour and carefully positioned words. This panegyric of 321 has been singled out by commentators as ‘florid’ (Nixon and Saylor Rodgers 1994: 336) and superabundant (Galletier 1952: 162), criticisms which suggest the presence of (and distaste for) the jeweled style. As the orator proceeds to his next theme, the adventus of Constantine into Rome, he comments that in describing the joy of the senate and Roman people, ‘no speech will be pleasing unless it is luxuriant’ (PL 4[10]30.4: nisi et ipsa lascivit, oratio), lascivit recalling Quintilian’s criticism of poetic and indulgent speech (inst. 2.5.22: ne recentis huius lasciviae flosculis capti voluptate prava deleniantur, cf. inst. 2.4.3). Far from striving for tasteful restraint, the orator is opting for the jeweled style with its flosculi and luxuriance.

Non-metrical word patterning To some extent, therefore, while ‘jeweled’ in the prose panegyrics has much in common with ‘poetic’, there is much that is ‘jeweled’ but not poetic. The most literal interpretation of the jeweled style and one of its most distinctive features is the careful patterning of words of brilliance. Roberts introduces his thesis by analysing the description of the breastplate of Aaron from the early fifth century Heptateuchos attributed to Cyprianus Gallus: Sardia prima loco, topazo adiuncta smaragdus, sapphirus hanc sequitur, cum qua carbunculus ardet, iaspisque viret fulvoque intermicat auro (Hept.[Cypr. Gall.] exod. 1098–1100; see p. 75 for the translation) In his lines, Cyprianus maps out the design of the breastplate, the name of each jewel placed carefully to recreate the literary artefact, with words of colour and brilliance added (ardet, viret, intermicat) and careful instructions as to how the whole is positioned (prima loco, adiuncta, sequitur, cum qua, intermicat) (Roberts 1989: 10f.). This is an ecphrasis in which structure and meaning unite. The reader is as much invited to see and marvel at the breastplate as a whole as to appreciate its components and wonder at the skill of the poet. The work of Optatian, writing in honour of Constantine, took this patterning as far as it could go, shaping the words and the poems themselves so that the letters could become signs and so part of a new meaning (Doria 1979: 65).7 The rigid metrical framework of poetry enables this elaborate patterning but the panegyrists worked within the restrictions of their prosaic genre. Consider this description of spring, taken from the panegyric of 297–8, which celebrates the dies imperii of Constantius and Galerius: o felix beatumque ver novo partu, iam non amoenitate florum nec viriditate segetum nec gemmis vitium nec ipsis tantum favoniis et luce reserata laetum atque venerabile, quantum ortu Caesarum maximorum! (PL 8[5]3.1) 48

Jeweled Style in Prose Panegyric

O Spring, fertile and blessed with new growth, now joyful and worthy of veneration, not so much because of the loveliness of your flowers or the greenness of your crops or the buds on the vines or the Zephyr breezes themselves and the light that is unleashed so much as for the creation of our magnificent Caesars! (Translation by Nixon and Saylor Rodgers 1994: 112) This is a mini-priamel, a catalogue leading to the culmination of a final point, a technique which creates its own framework. After the apostrophe of spring, the priamel begins: iam non amoenitate florum and is built up in three parallel comparisons, the structure created through the anaphora of nec: – nec viridate segetum – nec gemmis vitium – nec ipsis tantum favoniis These three clauses form a tricolon, the last being divided into two parts: – nec ipsis tantum favoniis – et luce reserata The nouns balance light and colour: pleasantness greenness buds (jewels) breezes

flowers crops vines light

The final balancing pair of laetum atque venerabile matches the opening felix beatumque. Tantum, set within the last element of the tricolon, creates a crossing link to the culmination of the priamel and the inauguration of the emperors: quantum ortu Caesarum maximorum and, with the closing ortu, completes the frame which began novo partu. Lacking the structure of poetic metre and designed to strike the ear rather than the eye, this scene cannot have the same visual impact as the poetic breastplate of Aaron but the same impulse is present. The words flores and gemmae are part of the description but also describe the technique of its creation, an early model for Cyprianus’ style.8 This description of spring develops from the orator’s praise of the dies imperii of Constantius and Galerius: 49

A Late Antique Poetics?

det igitur mihi, Caesar invicte, hodiernae gratulationis exordium divinus ille vestrae maiestatis ortus ipso quo inluxit auspicio veris inlustrior, cui dies serenus atque, ut celebrantes sensimus, ultra rationem temporis sol aestivus incaluit, augustiore fulgens luminis claritate quam cum originem mundi nascentis animavit; siquidem tunc inter illa rerum tenera primordia moderatus dicitur ne noceret ardentior, nunc certasse creditur ne maiestate vestra videretur obscurior. (PL 8[5]2.2) Then may that divine birth of Your Majesties, invincible Caesar, give me a beginning for today’s rejoicing, a birth brighter than the very beginning of spring which gave it light, for which the day was fair, and, as we who celebrated it felt, a summerlike sun warmed it beyond the expectation of the season, shining with a more majestic clarity of light than when it gave life at the creation of the world. As then midst those tender beginnings of life men say it was temperate so as not to do damage by burning too brightly, so now we believe it to have struggled not to seem less brilliant than Your Majesty. (Translation by Nixon and Saylor Rodgers 1994: 111) There is nothing here comparable to the poetic patterning of the spring episode, but the passage illustrates a feature which relates to the jeweled style and is very typical of the panegyrics: the word cluster. In this case, the words describe brightness: inluxit, inlustrior, dies serenus, sol aestivus incaluit, fulgens luminis claritate, ardentior. In the speech of 310, for example, clusters of words emphasize the divinity of the emperor, the imperium of Rome, and the fate of the barbarians who challenge the might of empire (Ware 2021: 30f.).

Catalogues In its more organized and patterned version, the word cluster becomes a word list. Following Quintilian, Roberts (1989: 40f.) distinguishes two different methods for making a description vivid: by a single image or ‘from a great many details’, ex plurimis (Quint. inst. 8.3.66), a mosaic-like approach which Roberts defines as leptologia.9 These details can be described in full sentences, but a characteristic of late antique literature is the list without adjectives or conjunctions. The poetry of Sidonius or Dracontius provides many examples.10 Pitying Sidonius’ audience, Piet Gerbrandy (2013: 75) suggests that the catalogues be read as ‘the desperate dying screams of a disintegrating social community’, although he acknowledges that contemporary tastes might be responsible (cf. also Cameron 1970: 285f). But the catalogue suggests rather that the writer is following Quintilian’s instructions to the letter. Sidonius summarizes the labours of Hercules to glorify the emperor: taurus, cerva, Gigas, hospes, luctator, Amazon, Cres, canis, Hesperides sint monimenta viri (carm. 13.11f.) 50

Jeweled Style in Prose Panegyric

though bull, deer, giant, host, the wrestler, Amazon, the Cretan beast, dog, and Hesperides are the hero’s monuments ending with the comparison ‘Alcides did these things once, now you are a second Hercules (carm. 13.15). The labours of Hercules are shorthand for heroism summarized by Hercules himself as he faces death in Ovid’s Metamorphoses (9.182–98), taking sixteen lines to list fourteen labours. Sidonius’ immediate model is an encomiastic passage from Claudian, which covers fourteen labours in twelve lines (rapt. Pros. 2 praef. 33–44), but to make his encomiastic point, Sidonius abbreviates even further. Hercules’ deeds are so well known that a single noun is enough for identification. The emperor’s heroism cannot be challenged when there is so much evidence that the orator doesn’t pause to add conjunctions. Similarly in the Panegyrici Latini, catalogues praise the emperor’s subjugation of the Franks: interfecit depulit cepit abduxit (PL 7[6]4.2: he killed, he drove out, he captured, he deported), and Moors: expugnasti, recepisti, transtulisti (PL 7[6]8.6: you assailed, you recaptured, you resettled). A string of verbs without linking conjunctions does not necessarily suggest late antique style: veni vidi vici is famously associated with Julius Caesar (Suet. Iul. 37.2: I came, saw, conquered).11 Asyndetic strings of nouns, however, are less classical. Maximian controls the barbarians with vastatione, proeliis, caedibus, ferro ignique (PL 10[2]7.6: by devastation, battle, slaughter, sword and fire) and Gaul, after decades of war, is devastated: inculta squalentia muta tenebrosa (PL 5[8]7.2: uncultivated, waste, silent, gloomy). Very often, the lists are of imperial virtues. The people praise Trajan for pietatem abstentiam mansuetudinem (PL 1[1]2.6: righteousness, restraint, gentleness) and he is said to be iustissimus humanissimus patientissimus (59.3: most just, most humane, most enduring). The pattern, usually of three or four virtues, continues throughout the corpus. Constantine and Constantius display continentia, fortitudo, iustitia, prudentia (PL 7[6]3.4: moderation, strength, justice, foresight) and Constantius is praised for valuing literature (litterae): continentiae modestiae vigilantiae patientiae magistras (PL 9[4]8.2: as the teacher of moderation, modesty, watchfulness, endurance).

Quotations and sententiae A feature of the panegyrics which is not found in poetry but which corresponds to the fragmentary nature of the jeweled style is the conspicuous insertion: a phrase or line which is to be viewed as separate. On the grand scale, this could be a letter from the emperor. In PL 9(4) of 297–8 ce , the letter is included for practical reasons, but the orator makes it clear that it is an ornament to his speech, ‘the words’ of the emperor being ‘divine’ and ‘the letter’ itself ‘sacred’ (PL 9[4]13.1: verba illa divina caelestesque litteras; 13.2: epistula sacra).12 On a smaller and perhaps more jeweled scale, separate from integrated allusion, is the introduction of quotations. Very few authors are considered worthy of this honour, but 51

A Late Antique Poetics?

Vergil is the poet of Rome, poeta Romanus, and a model for imperial encomium (Ware 2017). In describing how the barbarians tremble at the thought of the emperor, the orator calls on Vergil by name, and the authority of a direct quotation: illud quod de vestro cecinit poeta Romanus Iove, ‘Iovis omnia esse ’ (PL 11[3]14.2: that verse which the Roman poet sang of your Jupiter, ‘all things are filled with Jove’). The phrase Iovis omnia esse plena is the second half of the Vergilian line ab Iove principium Musae: Iovis omnia plena (ecl. 3.60) and is an ancient formula of dedication, particularly encomiastic dedication (Ware 2017b: 6). The encomiastic words of the poet of Rome are now reattributed to Maximian and attention is drawn to their positioning in the text, as though the phrase were a literary jewel.13 Aphorisms, sententiae, may also be inserted. These do not originate with the Gallic orations: Powell (2011: 386) describes Pliny’s panegyric as having a ‘lapidary style’ of epigram and sententia. Pliny certainly had a talent for this style and includes such wellturned phrases as: – neque enim umquam deceptus est princeps, nisi qui prius ipse decepit (PL 1[1]66.5: for never was a prince deceived unless he was the first deceiver) – non est princeps super leges sed leges super principem (PL 1[1]65.1: not ‘the prince is above the law’ but ‘the law is above the prince’). The other panegyrists are less adroit, but Constantine is praised as pater optimus, sed melior imperator (PL 4[10]4.1: the best of fathers, but a better emperor) and Pacatus describes life beneath the tyrant and the danger of expressing grief: nulla maior est poena quam esse miserum nec videri. (PL 2[12]25.5: There is no greater punishment that to be wretched and to appear not so). Of themselves, sententiae are not necessarily a feature of the jeweled style, but are discussed in rhetorical handbooks as an oratorical flourish.14 Even in Quintilian, however, they are described in familiar jeweled language: they are bright thoughts (lumina) which occur ‘particularly at the end of sentences’ (praecipueque in clausulis posita sententias) and which in his own time, are employed ‘without restraint’ (inst. 8.5.2: modo carent). To Quintilian, the examples from the Panegyrici Latini are what he defines as clausulae or conclusions: sed nunc aliud volunt, ut omnis locus, omnis sensus in fine sermonis feriat aurem. turpe autem ac prope nefas ducunt respirare ullo loco qui adclamationem non petierit. inde minuti corruptique sensiculi et extra rem petiti: neque enim possunt tam multae bonae sententiae esse quam necesse est multae sint clausulae. (inst. 8.5.13f.) Nowadays, however, something more is wanted, namely that every passage, every sentence, should strike the ear with its final phrase. Speakers think it a disgrace, almost a crime, to pause for breath at any point which does not call for applause. Hence a lot of little sentences, fragmented, affected, and irrelevant; for there cannot be as many good sententiae as there must be closures! (Translation by Russell 2001: 415) 52

Jeweled Style in Prose Panegyric

With minuti, corruptique sensiculi et extra rem petiti, we can see a move towards the jeweled style. The incorporation of visual detail characterizes both the jeweled style and panegyric, but one type of conspicuous insertion achieves this without words of brilliance and colour. The short slogan-like phrase, suggestive of the legends on coins and inscriptions, prompts recollection of the accompanying image.15 Such an image, associating the emperor with symbols of victory or a personified deity, can be evoked by a very simple caption: reciperata libertas (1[1]78.3), recepta libertas (6[7]6.1),16 aeterna victoria (6[7]11.5), continua felicitas (7[6]10.1). The series of such slogans in the panegyric of 310: qui eodem exercitu virtute capto, clementiae conservato, dum aedificandis classibus Britanniae reciperatio comparatur (PL 6[7]5.3: When that same army had been captured by his bravery and protected by his clemency, and while the fleets were being rebuilt in preparation for the recovery of Britain) can be matched by coins honouring VIRTUTI EXERCITUS or CLEMENTIA;17 the Arras medallion of 297, which shows Constantius approaching London and a kneeling personification of Britain matches the phrase Britanniae reciperatio.18

Ecphrasis Features such as the word cluster, catalogue or insertion which correspond to the fragmentary nature of the jeweled style ornament all sections of a panegyric, occurring equally in passages of narrative or more static description. However, the word positioning which most closely corresponds to Roberts’s definition of the jeweled style, with patterning of jeweled words, tends to occur within episodes. Roberts (1989: 38f.) argues that ecphrasis provided a point of departure for studying the poetics of late antiquity: it is also a feature of panegyric since the orator is constantly describing the scene before him.19 The classical definition of an ecphrasis is ‘a speech that brings the subject matter vividly before the eyes’ (Webb 2009: 1; cf. Quint. inst. 6.2.29). Does the description of spring (PL 8[5]3.1), quoted and discussed above, which embellished the emperors’ dies imperii, qualify? The description is certainly vivid, but the orator seems rather to be evoking the idea of spring rather than one particular day. His goal is not to make us see the flowers and crops but to apply their freshness and beauty to the majesty of the two new Augusti. It is an ecphrasis designed to make the invisible visible to the viewer. Compare this passage from 289 which describes the elevation of Maximian to Augustus: trabeae vestrae triumphales et fasces consulares et sellae curules et haec obsequiorum stipatio et fulgor, et illa lux divinum verticem claro orbe complectens, vestrorum sunt ornamenta meritorum, pulcherrima quidem et augustissima (PL (10[2]3.2)

53

A Late Antique Poetics?

Your triumphal robes and consular fasces, your curule thrones, this glittering crowd of courtiers, that light which surrounds your divine head with a shining orb, these are the trappings of your merits, very fine indeed, and most majestic. (Translation by Nixon and Saylor Rodgers 1994: 58) The imperial splendour of emperor and court is emphasized by word patterning. Parataxis focuses our attention on the details of the trappings, the robe, fasces and chair: – trabeae vestrae triumphales – et fasces consulares – et sellae curules which are followed by the two components of praise, stipatio et fulgor. Fulgor, the physical brightness of the imperial regalia, is linked to the metaphorical lux which embraces the emperor’s shining head. Visible and invisible marks of honour are joined as ornamenta meritorum, and in the final pairing of pulcherrima and augustissima. The ecphrasis is only the first part of the sentence. The orator continues: sed longe illa maiora sunt quae tu impartito tibi imperio vice gratiae rettulisti: admittere in animum tantae rei publicae curam et totius orbis fata suscipere et oblitum quodammodo sui gentibus vivere et in tam arduo humanarum rerum stare fastigio, ex quo veluti terras omnes et maria despicias vicissimque oculis ac mente conlustres ubi sit certa serenitas, ubi dubia tempestas, qui iustitiam vestram iudices aemulentur, qui virtutis vestrae gloriam duces servent, accipere innumerabiles undique nuntios, totidem mandata dimittere, de tot urbibus et nationibus et provinciis cogitare, noctes omnes diesque perpeti sollicitudine pro omnium salute transigere. (PL 10[2]3.3f.) But far greater are those services which you have rendered in place of thanks when the imperium was bestowed upon you: to admit into your heart the care of such a great state, and to take upon your shoulders the destiny of the whole world; to forget yourself, to speak, and live for the people, to stand on such a lofty summit of human affairs as to gaze down, as it were on every land and sea, and to survey in turn with eyes and mind where calm weather is assured, where storms threaten, to observe which governors emulate your justice, which commanders maintain the glory of your courage, to receive countless messengers from every quarter, to send out just as many dispatches, to worry about so many cities and nations and provinces, to spend all one’s nights and days in perpetual concern for the safety of all. (Translation by Nixon and Saylor Rodgers 1994: 58f.) Each of the two halves of this long sentence takes a different approach to creating a composite portrait of the emperor. The first half, describing the magnificence of Maximian, in its vocabulary and word positioning corresponds to the rules of the jeweled 54

Jeweled Style in Prose Panegyric

style, and deals with the splendour of his appearance. It is a static portrayal, with nouns and adjectives far outnumbering verbs. The second half is even more carefully structured, consisting of two mirroring sections: I sed longe illa maiora sunt quae tu impartito tibi imperio vice gratiae rettulisti: admittere in animum tantae rei publicae curam – et totius orbis fata suscipere – et oblitum quodammodo sui gentibus vivere – et in tam arduo humanarum rerum stare fastigio, ex quo veluti terras omnes et maria despicias vicissimque oculis ac mente conlustres – ubi sit certa serenitas, – ubi dubia tempestas, II – qui iustitiam vestram iudices aemulentur, – qui virtutis vestrae gloriam duces servent, accipere innumerabiles undique nuntios, totidem mandata dimittere, de tot urbibus et nationibus et provinciis cogitare, noctes omnes diesque perpeti sollicitudine pro omnium salute transigere. Careful word positioning makes this part jeweled in a different way. It is episodic, consisting of short clauses which come in groups of two or three, clearly matching in structure and content. If word patterning in the jeweled style can be described as mosaiclike, then each of these short, carefully positioned clauses is a tessera.20 Parallels are created through the introductory clause of the two sections, each beginning with an infinitive and linked by assonance: admittere . . . accipere. The anaphora in the first part, et . . . et . . . et and ubi . . . ubi is matched by qui . . . qui and totidem . . . tot in the second with further parallels in the homoioteleuton of infinitives. Within this meticulous structure are pairs of words, matched by similarity – terras et maria . . . oculis ac mente . . . Iudices . . . duces, noctes diesque – or dissimilarity – certa serenitas against dubia tempestas, the pairing iustitam vestram and virtutis vestrae gloriam adds variation through chiasmus and the addition of vestram/vestrae. The triplet of urbibus et nationibus et provinciis extends the pattern of doubling words and is also part of a word cluster emphasizing the extent of Maximian’s care for the world (tantae rei publicae . . . totius orbis . . . humanarum rerum . . . omnes et maria . . . innumerabiles undique nuntios, totidem . . . de tot urbibus et nationibus et provinciis . . . pro omnium). Unlike the static first section, the second part replaces nouns with verbs: the passive emperor decked in gold now becomes an active ruler with the series of infinitives (admittere, suscipere, vivere, stare . . . accipere, dimittere, cogitare, transigere) and an exhaustive catalogue 55

A Late Antique Poetics?

of his actions. The formulaic quality of the paired words – earth/sea, calm/storm, night/day – means that the audience can complete the phrases even as the orator is speaking so that they are step by step with him, even ahead of him, in seeing the picture created. What they see is more than reality. The splendour of the imperial regalia is impressive, but it needs the orator’s word to explain the ecphrasis and tell the audience what else they are seeing: that the fulgor of the imperial regalia is also the halo of power which surrounds the emperor’s godlike head.21 The emperor in majesty is further glorified by allusive language: in the Aeneid, Jupiter is the deity who looks down on the world from above: Iuppiter aethere summo despiciens mare velivolum terrasque iacentis litoraque et latos populos (Verg. Aen. 1.223–5) Jupiter, looking down from the top of the sky at the sail-bearing sea and the lands lying below, the shores and the people far and wide. In short, the picture of Maximian Augustus is created from far more than the splendid regalia he wears. All the techniques of rhetoric, structure, patterning and allusion are brought together to create something considerably more than an ecphrasis. Modifying the definition of ecphrasis as a means of bringing the subject matter vividly before the eyes, the orator employs the technique as a means of bringing the subject matter vividly before the eyes of the mind. In descriptions of the emperor, exegesis even takes the place of ecphrasis. In the speech of 307, the orator describes a painting in the imperial palace of Aquileia which shows the betrothal of the young Constantine to Fausta, Maximian’s daughter. What concerns the orator is the painter’s difficulty in seeing past the beauty of the emperor in order to portray the essence of Constantine and his love for Fausta (PL 7[6]6.4). This, too, is the role of the orator, not simply to describe, but to explain what the viewer is seeing. In the speech of 310, therefore, the orator points to brightness as the essence of the emperor’s physical appearance: imperator adulescens . . . in quo hic fulgor oculorum, haec veneranda pariter et grata maiestas praestringit simul et invitat adspectus (PL 6[7]17.1: a young emperor . . . in whom this brightness of gaze, this majesty, both august and appealing, at once dazzles and draws our eyes). The audience cannot look, but must look at the same time. As far as Constantine’s appearance goes, he is, in the same panegyric, the image of his father: tanta ex illo in te formae similitudo transivit, ut signante natura vultibus tuis impressa videatur (PL 6[7]4.3: such likeness of form passed from him to you that it seems imprinted on your face, as though moulded by nature), but also the image of Apollo (or Augustus; cf. Ware 2021: 327f.): vidisti teque in illius specie recognovisti, cui totius mundi regna deberi vatum carmina divina cecinerunt (PL 6[7]21.5: you saw and recognized yourself in the likeness of him, the one to whom the kingdoms of the whole world are owed in the divine songs of the poets). 56

Jeweled Style in Prose Panegyric

Although technically outside Roberts’s defining characteristics of the jeweled style, the acknowledged difficulties in expressing the ineffable is a concern of late antique poetics (Elsner and Hernández Lobato 2017: 6), and the attempts here to create a physical and metaphorical portrait of the emperor illustrate the cross-over between art and literature. Just as the emperor appears in art, formal and idealized, so he is depicted in the speeches. On coins and medallions, Constantine may be adorned with the jewels of majesty, but his eyes are his most striking feature, gazing forwards or upwards, suggesting insight or communication with the deity (Smith 1997: 198). Likewise, in other episodes which have great potential for ecphrasis – the imperial adventus or triumph – the orator usually has an agenda other than simple description. The speech of 321, one of the most visual and elaborate panegyrics of the collection, culminates in Constantine’s triumphal adventus at Rome after the victory of the Milvian Bridge. Rather than describing the magnificence of the scene, the orator recasts it as an allegory in which: Crime was mastered, Treachery conquered, Daring without its self-confidence and Insolence enchained, Fettered Fury and bloody Cruelty gnashed their teeth without the power to frighten; Pride and Arrogance were vanquished, Luxury was kept restrained and Lust bound with iron bonds. (PL 4[10]31.3; translation by Nixon and Saylor Rodgers 1994: 376f.) To aid the visualization of these personifications, the audience can turn to the Vergilian scene on which this is modelled: the chaining of Furor in Jupiter’s prophecy of Aeneid 1.294f.

Conclusion In conclusion, it is clear that the jeweled style exists in the prose panegyrics, separately from any poetic impulse. Rees’s (2017: 315) study of the poetics of the Panegyrici Latini lists the qualities he saw as defining poetic colour: citation of poets, other intertextual engagement with poetry, lexis with poetic associations, neologism, simile, ecphrasis, prosopopoiea, and apostrophe. This list overlaps with qualities associated in this paper with the jeweled style, but it is by no means identical and, as argued, many of the jeweled features are not found in poetry or are found with variation. The panegyrics then may be described as jeweled in their own right. How then do the three strands come together: poetry, panegyric and the jeweled style? All three flourish in the fourth century. Poetry is becoming more jeweled, panegyrics are becoming more poetic. That does not mean that poetic elements had gradually influenced the panegyrics so that panegyric finally became subsumed into epideictic poetry. Rather, the movement was the other way round. The passages discussed here in detail are from the earliest of the collection, written in the late third century, and the features which are particularly jeweled are also those which attracted criticism or praise from writers of rhetoric a century or two earlier. 57

A Late Antique Poetics?

In these third and fourth century orations is clearly displayed the orators’ delight in word patterning, structured in a rhetorical rather than metrical framework, in a cluster of synonyms or an asyndetic catalogue. The ‘lapidary style’ of Pliny’s panegyric covers more than a pause for an epigram or sententia: it includes the insertion of flagged quotation or a phrase suggestive of the legends on coins. These features on the whole have evolved from rhetoric and from the growing taste for elaboration described by Quintilian and Fronto. The influence of rhetoric on the jeweled style is not a new suggestion: Curtius (1990: 147f.) considers the ‘reciprocal relation between rhetoric and poetry’ in this period, Roberts (2009: 103) refers to the tropes of prose panegyrics which would become elaborate passages of description in Venantius Fortunatus’ poetry and Elsner and Hernández Lobato (2017: 6) introduce their study of late Latin poetics by noting that ‘all forms of writing are highly indebted to rhetorical training’. However, this paper argues that scholarship on late antique poetry should give greater weight to epideictic oratory. From the late first century, it was increasingly likely that an educated or ambitious Roman would need to deliver a panegyric at some stage of his career, whether to an emperor, or a governor or foreign power. The proliferation of handbooks shows that the professors of rhetoric took care to educate young Romans in this skill. Orators strove to outdo each other in technical excellence and an imperial panegyric would have received the greatest amount of adornment. As one orator put it, a speech before the emperor is something to be worked at for a long time and is not a place for extemporization (PL 6[7]1.1). In response to their desire to create something worthy of the emperor, the orators of the Panegyrici Latini crafted long and elaborate sentences, filled them with all the patterns of rhetorical tropes and with words of brilliance and colour. They link scene to scene, balancing praise with description (cf. MacCormack 1976: 46–54 and 1981: 6–12), pausing to create highly visual episodes. Therefore, when poetry began to be rewritten in the fourth century, it was written by people trained from infancy in this approach. Hermogenes’ sweeping ‘all poetry is panegyric’ needs, at the very least, qualification, but when it comes to description for the sake of description and the delight in recreating the visual in words, in meaning and in positioning, then it is difficult to deny that poetic passages of this nature, examples of the jeweled style, are very close to panegyric and may in fact be deeply indebted to it.

APPENDIX Panegyrici Latini The Panegyrici Latini corpus consists of eleven imperial orations from 289 to 389, in honour of emperors from Maximian to Theodosius, fronted by Pliny’s panegyric to Trajan (100 ce ). The Gallic orations are (more or less) in reverse chronological order. The final version of the collection, listed below, is thought to have been edited by Pacatus Drepanius, author of the speech to Theodosius. 58

Jeweled Style in Prose Panegyric

1(1)

100 ce

Pliny to Trajan (Rome)

2(12)

389 ce

Pacatus Drepanius to Theodosius (Rome)

3(11)

362 ce

Claudius Mamertinus to Julian (Constantinople)

4(10)

321 ce

Nazarius to Constantine (Rome)

5(8)

311 ce

Anon. to Constantine (Trier)

6(7)

310 ce

Anon. to Constantine (Trier)

7(6)

307 ce

Anon. to Maximian and Constantine (Trier)

8(5)

297–8 ce

Anon. to Constantius (probably Trier)

9(4)

297–9 ce

Eumenius to unknown official

10(2)

289 ce

Anon. to Maximian (Trier)

11(3)

291 ce

Anon. to Maximian (Trier)

12(9)

313 ce

Anon. to Constantine (Trier)

Notes 1.

The language of the jeweled style is anticipated in Marcus Aper’s assertion that a judge will not listen to an orator who does not entice him ‘with the brilliance of his epigrams or the shine and adornment of his descriptions’ (Tac. dial. 20.2: colore sententiarum aut nitore et cultu descriptionum), and compares the style to a building bright with marble and gold (Tac. dial. 20.2.6); on the qualities of an objection to postclassical style, see Dominik (1997).

2.

See the discussion of Webb (2009: 85), who concludes that ecphrasis in oratory should be discreet and not easily identifiable.

3.

Rees (2017: 344) notes Optatian, the Laudes Domini and encomiastic passages of Juvencus’ Libri Evangeliorum.

4.

On the sub-genre of ‘panegyric epic’ and Claudian’s influence on later writers see Schindler (2009) and Gillett (2012).

5.

Brolli (2013: 98) suggests that Sidonius was similarly inspired.

6.

The insertion of Greek words into Latin text is described by classical writers in jeweled style language and illustrates the gradual trend towards such mannerisms. Lucilius referring pejoratively to another’s oratorical style, had described the practice as mosaic-like, the inserted words being tesserulae (Lucil. 84f.) and Horace would criticize Lucilius for the same fault (sat. 1.10.20–3). For Fronto, however, it is an ornament.

7.

Cf. also Herzog and Schmidt (1993: 251), who see Ausonius’ diversity of forms and Optatian’s metrical experiments as a development of rhetorical exercises, and Squire (2017: 51), who argues that Optatian’s ‘materialistic aesthetics also reflects a peculiarly new, late antique interest in the interrelated mechanics of visual and verbal representational modes’.

8.

Roberts (2009: 143f.) compares this passage to Venantius Fortunatus’ jeweled praise of spring in carm. 3.9.9–32, showing how the familiar elements of the spring theme are already present in this abbreviated version. The elements, however, are also part of a display of rhetorical sophistication:

59

A Late Antique Poetics? Cicero had singled out such language as metaphors which even rustics used (orat. 81). 9.

A pull towards leptologia may also be seen in features such as anaphora and some forms of paraleipsis where the formula ‘I will pass over’ may actually be a means of inclusion rather than exclusion, e.g. PL 11(3)5.3f.

10. E.g. Sidon. carm. 5.207–10, 13.11f., 15.141–3; Drac. Romul. 2.24, 47, 5.34f. 11. Cf. Tacitus’ auferre trucidare rapere (Agr. 30.5: to plunder, to slaughter, to seize) and Cicero’s vastaret diriperat auferret (Phil. 10.12: he laid waste, he despoiled, he plundered). 12. For partial citations of imperial letters, cf. Auson. grat. [21.11.] 53, who introduces the short passage he quotes with munera tua verborum honore prosequeris; cf. also Lib. Orationes 14.30. 13. This is a rhetorical technique which goes back to Cicero: Quintilian (inst. 1.8.11) commends the practice which is designed to show off the orator’s learning but also to give pleasure as the voluptates of poetry are a pleasant relief from the asperitas of forensic rhetoric. 14. For an illustration of the cross-over between prose and verse panegyric see the sententiae of Claudian in a passage which imitates Pliny in Ware (2014: 323f.). 15. MacCormack (1976: 46 and 59f.) discusses how the words of the orators could be translated into images. For a full discussion of the link between panegyric and art, see MacCormack (1981). 16. Cf. OB LIBERTATEM RECEPTAM, OB REDDIT LIBERT in Hedlund (2008: 139). 17. For CLEMENTIA, see RIC v.ii 252; conservator was an epithet of Tetrarchs (see Rees 1993: 198). 18. MacCormack (1976: 47) suggests that phrases like Gallias tu, Caesar, veniendo fecisti (PL 8[5]6.1) could serve as a caption for such scenes. 19. The most common verb in the corpus after esse is videre; see L’Huillier (1992: 153f.). 20. The metaphor of mosaic to describe the careful patterning of words goes back to Lucilius and, variously attracting positive or negative criticism, is discussed by Horace, Cicero and Quintilian; see MacPhail (2003). 21. Cf. Symmachus’ simultaneous description and exegesis in his panegyric to Gratian: vidimus in sellis curulibus novum lumen, clementiae auspices incruentas secures, virtutis omina in fascibus laureatis (or. 3.2: in the curule chair we see new light, the bloodless axes are the auspices of mercy, the laurel-bedecked fasces are the omens of virtue).

60

CHAPTER 4 LEARNING THE JEWELED STYLE Frances Foster

Introduction The jeweled style, with its focus on visual and verbal patterning, and intense attention to detail, is a complex literary feature. An appreciation of that complexity relies on the reader’s familiarity with the literary tradition and their sensitivity to language, as well as a finely tuned awareness of a range of interconnected rhetorical effects. Composing literary texts, whether in prose or verse, that employ features of jeweled style, requires considerable further skills, including a competent and confident knowledge of the language and a breadth of appropriate vocabulary. This knowledge of literary tradition, sensitivity to language and the ability to manipulate language for varied and extravagant effects can only result from an extended literary education and rigorous training. Roberts (1989: 38) suggests that the late antique taste for a ‘densely textured play of repetition and variation’ developed from ‘scholastic traditions’, and particularly in the enarratio of the poets’ practices within schools, since schools formed a central part in constructing late antique literary taste and style. He (1989: 58) goes on to argue that the popularity of the characteristic fragmentary approach to poetic texts and analytic modes of composition arose out of the practices in schools, and that the late antique grammaticus ‘relished the opportunity to explain “hard words” ’. Much work has been undertaken on ancient education since The Jeweled Style was published in 1989, and a broader range of school texts from late antiquity are now available and well edited. This makes it possible to place educational writings in a stronger context and to evaluate the ways teachers may have taught literature and literary techniques in greater depth. It allows an evaluation of educational writing to go beyond the implicit assumption (arising from Henri Marrou’s survey of ancient education in 1956) that grammatici simply talked their way through Vergil’s texts in chronological order and minute detail. In this chapter, the key question I address is what educational writings can tell us about how students learnt both to appreciate and to emulate literary writing, and potentially to become practitioners of the jeweled style themselves. I focus on some of the key features which Roberts has argued define the jeweled style. In particular, I explore what we can discover about how students developed a sensitivity to visual patterning, how they built up a wide and varied vocabulary and how they gained an understanding of description and structure. I examine how teachers draw attention to literary features in order to enable their students to develop their own literary style. In so doing, I show how the jeweled style permeates late antique social and cultural capital, making it a crucial part of the aesthetics of late antique literature. 61

A Late Antique Poetics?

Two fourth-century authors provide the evidence for this study. Both were teachers: Servius worked as a grammaticus and literary commentator in Rome. Ausonius worked as a grammaticus and rhetor in Gaul as well as for the imperial household and held public office as a result of his imperial connections. Servius is identified as the author of an extensive and detailed commentary on Vergil’s works, which probably arose from his work in the classroom. In addition to his work in the classroom, Ausonius wrote poetry, and Roberts (1989: 17) labels Ausonius as ‘one of the most assured exponents’ of the jeweled style. While Ausonius has not left any teaching materials, a set of poems dedicated to teachers (mostly from Gaul) survives among his collected works. Servius and Ausonius represent elite teachers, and they were probably able to charge high fees. References in contemporary literature such as his appearance as a character in Macrobius’ Saturnalia indicate that Servius seems to have been very well respected for his expertise, and we can assume that a number of his students later became active members of the literary and cultural elite, including authors in their own right. Ausonius was invited to tutor the future emperor Gratian, but he also acknowledges other former students who went on to become wealthy and powerful. Although Servius is not known as a literary author, he nonetheless styles himself among the leisured literati in paratextual letters introducing short grammatical treatises. He claims to have access to a villa in Campania and leisure time to spend there writing about poetry (Serv. gramm. IV 468,3). Thus, both teachers present themselves as highly educated members of a literary elite – suggesting that the profession of teaching was not incompatible with a literary career. Both Servius and Ausonius taught students to understand and use a literary register of Latin through reading poetry – predominately, but not exclusively, that of Vergil. Servius’ commentary can tell us about many of the difficulties that late antique students faced when reading verse, since the literary register and rhetorical effects of verse were considerably removed from everyday Latin. It arose from his students’ needs and thus tells us more about how students interacted with jeweled features. By contrast, Ausonius’ poetry commemorating the teachers of his native region in Gaul reveals their own potential engagement with jeweled style. The commentary of Servius has reached us in two distinct forms: the text thought to be Servius’ own and an expanded form. Charles Murgia (2018: xi) describes how an unknown compiler ‘artfully fused’ Servius’ commentary with another commentary, adding, altering and deleting elements from the two sources. My focus in this chapter is on Servius in particular as a distinct late antique voice, and I therefore concentrate only on the text so far identified as Servius’ own, according to the editions by Thilo and Hagen (1881), Rand et al. (1946), Stocker and Travis (1965), Jeunet-Mancy (2012), Murgia and Kaster (2018) and Guillaumin (2019).

Schools in late antiquity Roman education lacked the structures and progression we take for granted in educational systems. William Harris (1989: 234) has observed that a three-stage division 62

Learning the Jeweled Style

between magistri, grammatici and rhetores is ‘far too rigid’. Carlotta Dionisotti (1982: 120) has questioned the extent to which Roman teachers ran distinct, separate establishments in accordance with these titles. The evidence suggests considerable fluidity and crossover between the three terms as students could choose freely between them and teachers could operate under multiple titles. For example, Libanius chose to pursue a career in rhetoric, but when his rhetor died, he elected to study with a grammaticus instead (Gibson 2008: xvii). Ausonius worked as both a grammaticus and a rhetor, and taught the future emperor Gratian in both capacities (praef. 1.26f.). This lack of clear distinction suggests that there may have been overlap in the curricular content of grammatici with that of rhetores. Since education was entirely private, and state interference was minimal, education was not regulated. As such, there may have been no expectation for a clearly divided and distinguished curriculum. I use the terms grammatici and rhetores as best fits the context in which they appear, but I do not mean to imply progression or specific distinction between them. Roman grammatici provided a literary education which was largely the preserve of wealthier citizens since it was expensive and time-consuming to obtain. Rhetores are usually attributed with providing training in declamation, but Bernstein (2013: 141) asserts that some level of declamatory training was probably a basic cultural competence for much of the audience of Flavian epic. Thus, it probably formed part of the curriculum at the literary level of grammatici as well. Raffaella Cribiore (2001: 187) has observed that those who experienced a literary education ‘attained the mental fitness and sense of identity’ which enabled them to be ‘recognised as a person of culture’. Since this type of education formed a common experience among those who underwent it, the skills that students learnt from it became a status marker by which members of the elite could recognize each other. The conservative nature of the curriculum in these schools enabled elites from different geographical parts of the Roman world to share a common cultural capital. Teresa Morgan (1998: 190) has noted that those who went on to practise the skills they had gained from the schools of the grammatici were ‘amongst the most prominent and influential members of their society’. The skills students learned at school helped them to exercise power and influence in adulthood both within the elite and beyond it. The costs of gaining an education from a grammaticus or rhetor – particularly an elite one in a major urban centre – would have been considerable. Students may have needed to move to another town specifically for the purpose of studying, if there was no grammaticus in their home town, or if the local grammaticus was not exceptionally well known and their parents were ambitious. This would have involved board and lodgings, over and above the teacher’s fee and any costs for materials. The poet Horace recalls how his father took him to Rome in the late Republic to study with a famous grammaticus, Orbilius. Edward Watts (2006: 5) points out that this type of literary education was a ‘sound investment’, since it not only distinguished the elite of Roman society, but it also bound them more closely together. This meant that gaining this type of literary education was a desirable commodity among elite and wealthy families across the late Roman world. To learn to read literary texts requires an understanding of classical Latin, a concept based on social class and education (Vincent 2016: 5). By late antiquity, the gap 63

A Late Antique Poetics?

between good spoken Latin and literary verse in a ‘classical’ register was considerable, and a literary education enabled students to learn to read and understand literary Latin, and, at the most advanced levels, to reproduce it in their own compositions. Some of our best evidence for how children in Roman schools might have learnt comes from the school scenes in a set of bilingual (Latin and Greek) language learning materials known as the Colloquia, edited by Eleanor Dickey (2012–15). The Colloquia were used and adapted for many centuries across both the Greek speaking East and the Latin speaking West, although the texts as we now have them are probably late antique in form. The Colloquia depict classrooms containing students at all levels, from the most basic to advanced, and tend to portray good, model students. In the Colloquium Stephani the child narrator describes undertaking a wide range of tasks, such as performing a recitation and writing out ‘everyday idioms’ (καθημερινά / cottidiana), thus covering different linguistic registers. The narrator then describes the process of learning poetry: ἐπερώτησα, καὶ διορθωθεὶς ἀνέγνωκα ἀνάγνωσιν τὴν ἐμήν, ἣν ἐμοὶ ἐξέθετο ἐπιμελῶς, ἕως νοήσαιμι καὶ πρόσωπα καὶ διάνοιαν ῥημάτων τοῦ ποιητοῦ. interrogavi, et emendatus legi lectionem meam, quam mihi exposuit diligenter, donec intelligerem et personas et sensum verborum au≤c>toris. I asked questions, and after I was corrected, I read my reading, which [the teacher] carefully explained to me, until I understood both the characters and the meaning of the poet’s words. (Hermen. Steph. coll. 17a–c) The child asks questions about the text prompting the teacher to explain the ‘meaning of the poet’s words’, because they find the literary language of poetry challenging. Reading Vergil’s poetry in school in late antiquity would have been as difficult for Roman students as reading Shakespeare is for British students today. The vocabulary, syntax, word order and grammar of literary language are far removed from the everyday language students use, even though the language of instruction is ostensibly the same as the language of the text. In this passage, the student needs the teacher’s help to understand the text so they can make sense of the action and perform a meaningful reading of the passage or write a response to it. Students did not just read literature in a passive manner – they were expected to do things with it and to respond to it. A great deal of schoolwork was oral rather than written (partly because of the technological difficulties of writing in the ancient school context) and so, students frequently performed readings of literary texts. They had to show not just correct rhythms and appropriate pauses in their reading, but also an understanding of character, declamation and performance. In the Colloquium Celtis, the narrator explains that some students read aloud at sight, while others prepare their performances:

64

Learning the Jeweled Style

λαμβάνω τόπον, καὶ ἄλλοι μετ’ ἐμοῦ σημερινόν, οἱ λοιποὶ ἀποφροντιμένον ἀποδιδοῦσι. accipio locum, et alii mecum extemporalem, ceteri accuratum reddunt. I receive a passage, and others along with me render it extemporaneously, the rest render it after careful preparation. (Hermen. Celtis coll. 33a–b) The narrator refers to two distinct skills, the ability to read at sight, demonstrating a level of fluency in reading, and preparing a performance. In addition to reading, students also learnt to compose their own responses to literary texts, both orally and in writing. The narrator describes how different students undertake different tasks: ἄλλος γράφει, ἠθοποιεῖ. alter scribit, alter meditatur. one writes, another practises / makes a speech in character. (Hermen. Celtis coll. 39c) Here, the Greek word ‘ἠθοποιεῖ’ points to ethopoeiae, exercises in which students composed and performed a prose speech in the persona of one of the characters from the text. In order to compose pieces in classical Latin – even in prose – students would have needed a solid understanding of the way the language works as well as a command of the stylistic features expected in this register. Such activities in composition provide the ideal opportunity for students to learn how to manage features of the jeweled style and put them into practice effectively. The Colloquium Celtis contains a significant and unusual amount of variation in vocabulary, providing several alternative expressions from which students might select the most appropriate. For example, the text suggests that students present their work to the teacher using six different expressions: δίκνυσον, δῖξον, πρόθες, σύναγε, †ἔνπαξον†, ἄξον, ἄγε, monstra, ostende, propone, admove, impinge, adduc, show! display! put forward! bring here! pin up! bring! (Hermen. Celtis coll. 26a–b). This level of repetition interrupts the narrative, making it harder to follow. Therefore, it must have held a pedagogical importance beyond the narrative. It provides the young learner with a set of synonyms on which they can draw to enhance variation in their own compositions, and it also models how a list (enumeratio) might be structured at a basic level. Dickey (2015: vol. 2, 157 and 159) has demonstrated that the vocabulary 65

A Late Antique Poetics?

lists are late antique and that the final form of the Colloquium Celtis seems to have emerged in late antique Gaul. This places the manual within Ausonius’ world, although it is impossible to identify the town in Gaul or the decade of late antiquity in which its final form emerged. It does, however, indicate that teachers may well have encouraged students to broaden their vocabularies and manage multiple synonyms, both skills that would help them to appreciate and reproduce features of the jeweled style. Scott McGill (2005: xviii) has suggested that students may also have been educated in verse composition, and that such instruction may have involved recasting Vergil’s poetry because of his centrality on the curriculum. Verse composition would have required students to develop significant confidence and skill in manipulating classical Latin while also keeping within the metrical boundaries of verse and employing the patterning and variation expected in verse. Such skills would have been supported by the extremely detailed knowledge of central literary texts which the curriculum instilled into successful students. For example, Augustine recalls how he discovered that his contemporary Simplicius seemed able to recite the Aeneid backwards, line by line (Aug. anim. 4.7.9). Even though Simplicius’ memory is meant to be remarkable, this episode indicates an attention to detail brought about by close study of literary texts, in which those details were as memorable as the whole. A confident command of a range of literary texts would have made composing responses to these texts easier, whether those responses were composed in prose or verse, and it would have increased the intertextual references a student was likely to employ. Therefore, composition exercises could have enabled students to demonstrate their skills at ‘brilliance and variation in detail’ (Roberts 1989: 12) and to perfect these techniques.

Classroom poetics in late antiquity Ausonius commemorates the grammatici and rhetores of Gaul through their skills in literary composition. For example, he tells us about the rhetor Luciolus, who was his condiscipulum atque magistrum / collegamque dehinc (prof. [11.]3.1f.: fellow student and teacher, and subsequently colleague), suggesting that their careers overlapped, but that Luciolus was slightly the senior of the two. Ausonius declares that Luciolus was faciundum doctumque (prof. [11.]3.3: eloquent and learned) because he excelled in literary composition: seu lege metrorum / condita seu prosis solveret orsa modis (prof. [11.]3.3f.: whether creating utterances bound by the rules of metre or freed by the rhythms of prose). What Ausonius admires about Luciolus is his skill in composing both literary prose and verse, through which he could display his learned eloquence. Ausonius, of course, demonstrates his own mastery of verse through his poetry about his fellow teachers. Servius provides relatively little overview of his approach to poetics in his commentary, perhaps because of the structure and purpose of the commentary – it focuses on the aspects of Vergil’s text that students found challenging or about which they needed specific advice. However, there are a few glimpses of his general approach at various 66

Learning the Jeweled Style

points, particularly when he evaluates how Vergil constructed his text. When Turnus chases the phantom of Aeneas that Juno created onto a boat in Aeneid 10, Vergil gives a brief background about the boat, and Servius comments: forte ratis descriptio per parecbasin facta, non enim a superioribus pendet, sed ante dictis adiungitur, ut ‘insula Sicanium iuxta latus’ [Aen. 8.416]. has autem descriptiones esse aptas et raras convenit, sicut etiam Horatius docet in arte poetica, dicens ‘purpureus late qui splendeat unus et alter adsuitur pannus’ [ars 15f.]: sunt enim ornatui. By chance the ship The description is made through digression, for it’s not dependent on earlier things, but it’s joined by words before, as ‘an island, next to the Sicinian shore’. But it’s agreed that these descriptions are fitting and rare, just as Horace teaches in his Poetic Art, saying ‘one or two purple pieces of cloth, stitched to glitter far and wide’, for they are for the purpose of embellishment. (Serv. Aen. 10.653) Servius notes that this is description through digression, and gives another example of Vergil’s use of this device elsewhere in the Aeneid, thus providing a helpful model by which to measure what a digression of this sort might look like. He praises such digressions on condition that they are appropriate and infrequent, validating his advice with a reference to Horace, and describing this feature as for the purpose of ‘embellishment’. Here, Servius highlights how Vergil has employed a poetic feature and shows students how it has been constructed for the greatest effect. However, he also points out that embellishments need to be used at the right moments and sparingly in order to achieve the full effect of glittering far and wide. The construction he provides allows students to imitate Vergil’s model when composing their own work, but makes them aware that they should not overuse this kind of description. Servius describes these poetic effects in the visual terms of late antiquity, particularly with the word ornatus, which indicates a decoration or adornment. He uses the same word to gloss the adjective Vergil uses to describe the jeweled coronet that the Trojans bring to Dido as a gift in Aeneid 1: bacatum ornatum margaritis. dicimus autem et haec margarita et hoc margaritum et haec margaris, quod Graecum est, quo modo Nais. sane multi separant gemmam a margarita, ut Cicero ‘nullam gemmam aut margaritam’ [Verr. II 4.1], et gemmas volunt dici diversi coloris, margaritas vero albas. Adorned with pearls Decorated with pearls. But we say ‘pearl’ in both the 1st declension feminine and 3rd declension neuter, because it is Greek, like Naiad. To be sure, many distinguish jewels from pearls, as Cicero ‘no gem or pearl’, and they mean that the term ‘gems’ is used [for jewels] of diverse colours but ‘pearls’ for white [ones]. (Serv. Aen. 1.655) 67

A Late Antique Poetics?

Servius begins this comment by explaining that the unusual adjective bacatum means ornatum margaritis (decorated with pearls). Servius concentrates on the visual aspect of the coronet, while at the same time ensuring that students understand the meaning of bacatum, which is only rarely attested elsewhere, mostly in glossaries (ThLL 2.1659,21– 31). Servius then clarifies how margarita declines since it is a Greek loan word which was probably not in everyday use, but formed part of the classical Latin register that students learnt. His subsequent explanation about the distinction between gemmae and margarita focuses on their respective colours. In his definition of the concept of the jeweled style, Roberts (1989: 52) cites Servius’ note to support the idea that jewels ‘are characterised by their varied colours and brilliance’. Servius wants his students to use language precisely to describe the visual appearance of various precious stones. He provides a comparative example from Cicero to show how the two words might be used together well. This suggests that Servius expected students to use classical words such as margarita in their own compositions and that he was training them to be aware of the visual effects they could create.

Polychromatic style Roberts (1989: 56) notes that ‘visual description lent itself best to the demonstration of that polychromatic exhaustivity’ which the jeweled style prized most highly. Ausonius displays his own skill in visual description through his praise of the declamatory skills shown by Minervius. Ausonius declares that many of Minervius’ former students went on to careers in the forum, and twice as many became senators. He praises Minervius: dicendi torrens tibi copia, quae tamen aurum, non etiam luteam volveret inluviem (prof. [11.]1.17f.: your ability in speaking was a rushing stream, but that skill was gold, and it did not even disturb the muddy sediment). Ausonius focuses on Minervius’ skills as a speaker (he does not record any literary texts by him), but his description visualizes the fluency and ‘golden’ qualities of his speech, in contrast to the ‘muddy sediment’ of other less successful orators, drawing attention to his own literary skills. Servius shows considerable sensitivity to Vergil’s use of colours, and he pays careful attention to the ways in which colours may work together to create polychromatic visual effects. At the start of Aeneid 7, Vergil describes the dawn in a traditional epic style, iamque rubescebat radiis mare et aethere ab alto / Aurora in roseis fulgebat lutea bigis (Aen. 7.25f.: and now the sea became red with rays and from the high upper air Dawn on a rosy chariot shone golden yellow). Servius explains: in roseis fulgebat lutea bigis ‘lutea’ est crocei coloris, ut ‘croceo mutabit vellera luto’: unde multi iungunt ‘inroseis’, id est non rubicundis, ne sit contrarium. tamen secundum Homerum dictum est, qui interdum ῥοδοδάκτυλον, id est rosei coloris digitos habentem, interdum κροκόπεπλον dicit Auroram, id est crocei coloris veste circumdatam. 68

Learning the Jeweled Style

Aurora on a rosy chariot shone golden yellow Golden yellow is the colour of saffron, as ‘the fleece will change to saffron yellow’ [ecl. 4.44]. Hence, many join together ‘in’ and ‘rosy’, that is, ‘non-rosy’ lest it marks a contrast to the yellow. But it’s said according to Homer, who sometimes says ‘ῥοδοδάκτυλον’, that is, having rosy coloured fingers. Sometimes he says Dawn is ‘κροκόπεπλον’, that is, surrounded by saffron coloured clothes. (Serv. Aen. 7.26) Servius considers both the colours Vergil describes, first defining the shade of yellow indicated by luteus as croceus (saffron yellow) supported by an example from the Eclogues, and then correcting the misconception of ‘many’ who form a single word from in roseis and read it as a negative, ‘not-rosy’. Finally, he provides two comparisons to Homer, describing Dawn in both red and yellow shades. Students at Servius’ school did not seem to have learnt classical (literary) Greek, but they were expected to understand some literary terminology (Foster 2016) and perhaps a few famous words and phrases from Homer. Servius demonstrates how Vergil combines the range of colours seen in the dawn sky into his single description. He uses the quotation from the Eclogues to demonstrate to students how colour words can be used to complement one another, since the adjectives luteus and croceus combine to indicate a particular shade of colour as the ram’s fleece changes with the seasons. He does not identify the ‘many’ who misread the words in roseis, and they may have been other commentators. But it is also possible that it was a common mistake students made when reading the verse, since there were no spaces between the Latin words in their text, and without training they may have felt the juxtaposition of red and yellow contradictory. Servius breaks down how both colours refer to famous Homeric descriptions, each relying on a flower (rose and crocus) for the colour, and a human element (finger and robe) for the aspect that is coloured. Servius therefore shows how Vergil takes the two ideas and makes them work together to create a polychromatic effect. This explanation prepares students to go beyond reading the passage accurately, as it demonstrates how they might shape their own descriptions in response to the literary texts they have read. Roberts (1989: 52) suggests that the ‘polychromatic’ nature of visual description is not just in colour, but also in the visual patterning to create extravagant effects. In the catalogue of warriors in Aeneid 7, Vergil describes the way some of the warriors are armed, and Servius explains: falcati comminus enses adfectavit varietatem: nam consequens fuerat ut diceret ‘dextras armant gladii falcati’. Curved swords for close combat He aimed for variety, for he’d have been copying [the previous clause] if he’d said ‘the curved swords armed their right hands’. (Serv. Aen. 7.732) Servius focuses on the idea that Vergil aims for variety in expression in this list. He shows how Vergil varies his construction to describe the warriors’ swords differently from the way he described their shields. Servius models how to rephrase the second clause so 69

A Late Antique Poetics?

that it matches the first in syntax. In this comment, Servius trains students to become more sensitive to both the variety and patterning of syntax, and shows them how to switch between the two. His rephrased example would not scan in Vergil’s verse, but it might be appropriate if a student composed in prose or fitted the phrase into a different verse.

Variatio of vocabulary and theme I have already touched on the idea of stylistic variety in the example above, since it is so central to the concept of a jeweled style. Roberts (1989: 55) suggests that in late antiquity ‘a sense for the literal meaning of variatio was reawakened’ as texts were understood as containing multicoloured flowers and jewels. There is perhaps some overlap between variatio and the polychromatic visual effects. They are closely linked in the way Roberts describes the jeweled style. Ausonius praises Attius the Elder for his linguistic skills, declaring that this man memor, disertus, lucida facundia, / canore (prof. [11.]4.17f.: he had an excellent memory, well spoken, a shining command of language, melodious). Ausonius concentrates on how well Attius used language for effect. His choice of the adjective lucida (shining) recalls the visual dazzle of the jeweled style. Ovid uses the word lucida to describe jewels (epist. 15.74). However, in his description, Ausonius also draws the reader’s attention to his own command of variatio, since he uses three synonymous expressions (disertus, lucida facundia and canor) to recall Attius’ eloquence, thus showcasing his own mastery of the jeweled style. Servius is well attuned to variatio, and he highlights the feature for students so that they can see how it is constructed. At the start of Aeneid 3, when Aeneas tries to collect wood, Servius draws the students’ attention to the way in which Vergil has varied his language in describing the vegetation under which Polydorus is buried: silvam bene variat, ut nunc silvam [Aen. 3.24], nunc vimina [3.31], nunc virgulta [3.23], nunc arborem [3.27] dicat. Woods He varies it well, as he says now ‘wood’, now ‘bushes’, now ‘undergrowth’, now ‘tree’. (Serv. Aen. 3.24) He notes that Vergil has used four synonyms within less than ten lines of verse, and praises Vergil’s skill for the variation. In this comment, Servius does not define the meaning of these words, which indicates that students were generally familiar with their meanings. Instead, his focus is on how Vergil has used the breadth of his vocabulary for effect. On another occasion, in the boat race in Aeneid 5, Servius draws attention to Vergil’s use of a different type of variatio: discrimine interstitio. et bene variat, nunc naves, nunc ductores commemorans. In the distance: After an interval. And he varies it well, speaking now of the ships, now of their captains. (Serv. Aen. 5.154) 70

Learning the Jeweled Style

Here, Servius starts by defining what Vergil means by discrimine, indicating that students were sometimes confused by what was happening. In this example, he shows how Vergil varies the focus of his descriptions between the ships and their captains, rather than by employing synonyms. On both occasions, Servius praises Vergil’s skill with the adverb bene, showing this is a feature worth emulating. On another occasion Servius is more explicit about instructing his students in composition. Towards the start of Aeneid 10 Vergil lists some of the Trojans guarding the besieged camp against the attacking Rutulians, and Servius provides an explanation about the names: Imbrasides Imbrasi filius. Hicetaonius Hicetaonis filius. et bene ex varietate syllabarum quaesivit ornatum: nam patronymica aut in ‘des’ aut in ‘ius’ exeunt, quibus et uti possumus. quae enim ‘on’ terminantur, ad nos transire non possunt. Imbrasides The son of Imbrasus. Hicetaonius The son of Hicetaon. And he looks for decoration of syllables out of variety well. For patronymics end in either ‘-des’ or ‘-ius’, which we can also use. For those which end in ‘-on’ cannot transfer to us. (Serv. Aen. 10.123) First, Servius explains the patronymics, providing the usual forms of the names Imbrasus and Hicetaon. Late antique students seem to have found many of the names in Vergil’s epic difficult, particularly those which were Greek, and so this information helps them to use these names in their own responses. However, Servius then evaluates Vergil’s use of patronymics, suggesting that he has placed these names together because he sought syllabarum . . . ornatum (decoration of syllables) to provide variety. Servius again praises the decorative effect with the adverb bene (well), holding up this element of the verse for admiration. Servius uses the noun ornatus (decoration, embellishment) to describe the effect Vergil has created through the variety of endings. What Servius admires is a very small and fine detail, but it exemplifies the art that Roberts (1989: 55) describes, whereby the skills of poets lay in their ability to ‘manipulate brilliant pieces’ which they could then throw into relief by ‘effects of contrast and juxtaposition’. Servius observes the details, but also the way the poet has contrasted them for decorative effect, and he highlights this to students as an important feature. Finally, he instructs students which patronymic endings they can use. The use of patronymics is a literary feature and particularly common in epic verse, so Servius’ instruction seems to be directed at those students who are expected to emulate this register in their own compositions. The proximity of the instruction to the comment claiming that Vergil has varied the decoration well suggests that Servius expected students to aim for similar effects when composing.

Description and structure This focus on juxtaposition and variety affects not just Servius’ readings of Vergil’s expression, but his understanding of Vergil’s descriptive structures. Roberts (1989: 56) 71

A Late Antique Poetics?

notes that late antiquity ‘preferred juxtaposition and contrast to logical interrelationship’, focusing on individual sections rather than on a coherent whole. Servius perceives Vergil’s use of contrast and structural juxtaposition by examining how he arranges the narrative at different moments. Near the start of Aeneid 1, Vergil sets the scene, referring back to the Trojan war and the anger of Achilles, Troas, reliquias Danaum atque immitis Achilli (Aen. 30: the Trojan survivors, who had been left behind by the Greeks and by savage Achilles). Servius observes: atque inmitis Achilli bene secundum Homerum segregat duces a populo. sic Aeneas in secundo post omnium casum Priami separavit interitum, ut ‘forsitan et Priami fuerint quae fata requiras’ [Aen. 2.506]. And by savage Achilles Following Homer, he separates the leaders from the people well. Thus, in Aeneid 2, after the destruction of everything, Aeneas sets the death of Priam aside, as ‘by chance you may ask what Priam’s fate was’. (Serv. Aen. 1.30) Servius points out that Vergil follows Homer in the structure of this description because he separates the leaders from the people when he says the Trojan survivors were ‘left by the Greeks and by savage Achilles’. In this clause, Achilles is split from the rest of the Greeks by the conjunction atque (and), as well as his position at the end of a verse. Servius compares this to the narrative structure of the fall of Troy in Aeneid 2, where Priam’s death forms a separate section from the destruction of the city. He praises Vergil’s choice with the adverb bene (well), indicating that this is an appropriate structure for the narrative, and one that students could imitate in their own work. He also trains students to recognize a particular kind of narrative juxtaposition since he draws the parallel between a small phrase in Aeneid 1 and an entire passage in Aeneid 2.

Conclusion Ausonius emphasizes the linguistic skills of many of his professores, and he particularly praises those who composed verse, such as Attius Tiro Delphidius, who – he says – composed a verse epic more swiftly that someone composing in prose (prof. [11.]5.10– 12). Ausonius carefully constructs his Professores in verse, varying his choice of metre and frequently displaying his own jewels in composition while describing the skills of others. He ties the profession of teaching to that of skilled poet, implying that literary composition was not incompatible with teaching, although his work gives no hint of how teachers might have used these skills in their classrooms. By contrast, Servius gives us a sense of how students learnt poetry: he tends to focus on the minutiae of Vergil’s language and individual episodes as they occur in the narrative. In part, this is likely to be the result of classroom teaching, as the aspects of studying Vergil that students still tend to find challenging today relate to the language, description and unfamiliar details, rather 72

Learning the Jeweled Style

than the overall narrative. However, the effect in Servius’ commentary resembles the way Roberts (1989: 56) describes the tastes of late antiquity, where the ‘sense of proportion’ may be difficult to find, but the ‘elaboration of the individual episode’ receives careful consideration. Thus, the minute focus of the commentary provides training in paying attention to close details. Vergil held a prominent place on the Roman curriculum since his own lifetime, but while Servius’ predecessors seem to have favoured Republican authors, Servius frequently provides comparisons and contrasts with authors from the first century ce (Kaster 1978: 208). Roberts (1989: 8) notes that the evidence of poetic imitation for the jeweled style points to ‘Vergil and the first century ad poets as being most influential in late antiquity,’ and this accords with the curriculum Servius offers. This curricular content helps to canonize the jeweled style of verse and artistic prose by late antiquity (‘the canonical style’, Roberts 1989: 62), since Servius analyses these stylistic elements and presents them as examples for students to emulate and build on, while Ausonius showcases the style. The large late Roman literary elite came from all over the empire, living and working in different places during their careers as successful mobile professionals. This group had a skilled command of literary, classical Latin, regardless of where they originated from, and their writing is exceptionally ‘classical’ in content as well as register and style. Ausonius belonged to this group of highly literate professionals, and he articulates this through his jeweled poetry. It is likely that Servius also belonged to this group, and the sensitivity he shows towards Vergil’s use of language, particularly features which may form jewels when used in the right context, suggests that Servius had a detailed familiarity with the jeweled style. In many of the examples I have given, Servius uses Vergil’s own words, and draws attention to the order, patterning and juxtaposition of those words in the verse. In these examples, he does not reorder the words to help students understand the sense, as he does with the phrase ordo est (e.g. Aen. 1.109 2.14; 3.73: the order is). This is particularly common when noun-adjective pairs in Vergil’s verse are split from each other by another word, phrase or clause, and students seem to have taken the words in the order in which they occurred, leading to a misunderstanding of the sense (Foster 2017: 277 and 2019: 64f.). Instead, in these examples he makes clear how Vergil creates patterns and contrasts with vocabulary, language and structure, where the sense is clear to the students. An appreciation of this type of patterning and contrast requires training, since it is not a feature of everyday spoken language. The analysis Servius provides enables students to examine Vergil’s style intelligently, identifying the patterns and their effects, but more importantly, it also empowers them to develop their own patterned style. Servius consistently praises stylistic features which are aspects of jeweled style, so that students see and hear the language in action, just as Ausonius praises teachers for their own skills in literary composition. Ausonius and Servius were both elite teachers who demonstrate (respectively) skills in and knowledge of elements of the jeweled style. They would have been able to instruct students in using literary language to create jeweled effects. The education that the Roman literary elite underwent, and the activities they were expected to undertake in 73

A Late Antique Poetics?

class, trained them to write in a highly stylized language, with complex patterning and juxtaposition, showing an extensive command of considerable cultural capital. This type of education served the late Roman elite, and was probably only available from teachers who, like Ausonius and Servius, were themselves elite. In his commentary, Servius does not repeatedly instruct students to imitate jeweled features in their classroom exercises. This suggests that composing in the jeweled style was reserved for only a select minority, although a wider group of educated people would have had the training to appreciate such literary writing. Elite teachers probably only encouraged the jeweled style among a minority of students, reserving it for those wishing to pursue a literary career. This selective teaching helped to canonize the jeweled style and ensure its continuity as a ‘prestigious model of literary excellence’ (Roberts 1989: 155) for many generations to come.

74

CHAPTER 5 QUANTITATIVE APPROACHES TO LATE ANTIQUE POETICS: ENUMERATION AND CONGERIES Joshua Hartman and Jacob Levernier

Introduction When Michael Roberts offered his assessment of late antique aesthetics in The Jeweled Style (1989: 9), he began by introducing readers to a passage of Cyprianus Gallus’ Heptateuchos, highlighting qualities like enumeration, highly visual verses and Christian themes that are now closely associated with late antique poetry. In particular, the passage of Cyprianus focuses on a list of precious stones adorning the armour of the biblical hero Aaron, a versification of a more prosaic biblical list. This foundational example of late Latin poetics is reproduced below, alongside the text Roberts posited as its prose inspiration: sardia prima loco, topazo adiuncta smaragdus; sapphirus hanc sequitur, cum qua carbunculus ardet, iaspisque viret fulvoque intermicat auro: tertia ligurio sedes: hic iunctus achati atque amethysto, fulgens quem purpura tingit. chrysolithus quartus, berillo adnexus onychnus. (Hept. [Cypr. Gall.] exod. 1098–1103) First in position is the carnelian, and emerald along with the topaz; then comes the sapphire, with which the carbuncle blazes, and the jasper is green and shines with tawny gold. Third place is taken by amber, and along with it the agate and amethyst, with its bright purple hue. Fourth the chrysolite, and onyx next to the beryl. (Translation by Roberts 1989: 10) Sardius, topazynus, zmaragdinus, versus unus. Et versus sequens: carbunculus, et sapphyrus, et iasphys. Et versus tertius: lygyrius, achates, ametthustus. Et versus quartus: chrysolitus, et onycynus . (Vet. Lat. exod. 28.17–20 [cod. 100 = cod. Lugd. 403])

75

A Late Antique Poetics?

Carnelian, topaz and emerald row one. The next row carbuncle, sapphire and jasper. The third row amber, agate and amethyst. The fourth row chrysolite, onyx . (Translation by Roberts 1989: 11) Roberts (1989: 11) observed that Cyprianus enhanced the biblical description with words emphasizing the colour and brightness of these stones (ardet, viret, intermicat, fulvo, auro, purpura), embellishing an object whose description had only suggested physical beauty before. While this is certainly true, it is worthwhile to remember that the biblical original would probably have been attractive to readers. As Jesus Hernández Lobato discusses in this volume, late antique prose is not without its own poetry, and this list displays two features important to late antique poetics. The first is the abundance of rare words and sounds, as numerous Greek loan words offer unique letters and combinations of letters, such as y, z, ch, chrys, zm, gd and tth (on rare words see Roberts 1989: 12 and Gualandri 2017: 131–6). Meanwhile, the second feature is the list itself. As Roberts (1989: 59–61) – and later Hernández Lobato (2012: 384–9) – observed, this kind of verbal pile-up, frequently called congeries, also typifies late Latin poetry. Their reflections on congeries prompted both Roberts and Hernández Lobato to ask a much broader question about the aesthetics of late antique poetry relative to its predecessors, noting the rarity of this poetic feature in earlier poetry. Roberts (1989: 60 n. 72) essays a list of potential loci in earlier authors that may have inspired their late antique successors, ultimately arriving at the now widely accepted conclusion that the forerunners of the jeweled style were (above all) Statius, Lucan and Ovid. Hernández Lobato (2012: 384 n. 159) likewise asserts the relevance of such predecessors, especially Statius, and cites the paucity of classical examples. Furthermore, he (2012: 385–9, esp. 385f.) argues that congeries is essential to a late antique poetics which luxuriates in individual words, but he goes beyond this to suggest that such agglomerations of words created impressionistic verbal images of much larger concepts and scenes. An example common to his study and Roberts’s is Sidonius Apollinaris’ description of the labours of Heracles: cui sus, cerva, leo, Gigas, Amazon, hospes, taurus, Eryx, aves, Lycus, fur, Nessus, Libs, iuga, poma, virgo, serpens, Oete, Thraces equi, boves Hiberae, luctator fluvius, canis triformis portatusque polus polum dederunt. (carm. 9.95–100) To whom the sow, stag, lion, Giant, Amazon, host, bull, Eryx, birds, Lycus, thief, Nessus, Libyan, peaks, apples, maiden, snake, Oete, Thracian horses, Spanish cattle, wrestling river, three-formed dog, and burden of the heaven gave the heavens. (Translation by Roberts 1989: 60) 76

Quantitative Approaches

While enumeration can be thought of as operating at the level of a passage or catalogue, late antique poetic congeries is a type of enumeration that relies upon the line as its most important unit. The passage of Sidonius is a succession of such wholly enumerative lines, but a single line can often have a similar effect. Consider the following epigraphic poem of Damasus, describing the martyrdom of Saint Lawrence: verbera, carnifices, flammas, tormenta, catenas vincere laurenti sola fides potuit. haec damasus cumulat supplex altaria donis martyris egregii suspiciens meritum. (Damas. carm. 33 = Inscr. christ. urb. Rom. [ICVR] 2.4114) Blows, executioners, flames, racks, chains – Lawrence’s faith alone was able to lay low Damasus, a suppliant, heaps this altar with gifts, honoring the merit of a distinguished martyr.1 (Translation by Trout 2015: 141f.) In spite of their diverse content and contexts, each poem relies on a similar strategy that privileges the line as a unit. When a heavily enumerative line ends, there is often little narrative progress and only the barest suggestion of syntax, and the reader’s imagination must fill in these gaps. At times, this grammatical power vacuum is quickly filled and reads as enjambment (as in Damasus’ case) and sometimes it is prolonged, while the grammatical resolution only appears within a recondite expression (as in Sidonius). There are numerous examples of such heavily enumerative lines throughout the corpus of late antique poetry and their presence has long been considered a hallmark of later Latin style, but it is now possible to confirm empirically the hypotheses of studies like Roberts’s Jeweled Style. With the aid of Natural Language Processing (NLP), computational philology can empirically demonstrate trends that might only be theorized earlier. Using NLP to automate an analysis of Latin poetry at a corpus-wide scale, we demonstrate an upward trend in such heavily enumerative lines over time, explore the role of poetic predecessors and statistically examine the late antique authors who deployed such lines most frequently.

Methods and context Studies of enumeration and congeries that are overarching (rather than moments in particular texts/authors) are rare (Galjanić 2007; Ledentu and Loriol 2020; Laemmle, Scheidegger Laemmle and Wesselmann 2021).2 To our knowledge, there are none that focus on late antiquity.3 Other digital approaches to enumeration or congeries are unknown to us at the time of this writing. Keeline and Kirby (2019) confronted a similar 77

A Late Antique Poetics?

situation in their study of Latin prose rhythm as the first to apply computational approaches, although the preceding bibliography of traditional studies on prose rhythm is vast compared to that of enumeration, especially in its late antique context (including congeries). Like that of Keeline and Kirby (2019), this study relies on an existing digital corpus (Musisque Deoque or MQDQ) and the Classical Languages Toolkit (CLTK), a Python library offering a suite of Natural Language Processing tools for Latin and other ancient languages. MQDQ is a database that collects all extant Latin poetry (including fragments) from archaic Latin until approximately 700 ce. In spite of this wide-ranging chronology, the MQDQ database uses classicizing orthography for vowels, and thereby avoids introducing further complications for automatically parsing texts (e.g. anime as a potential variant of classical animae, see Smith 2019). Texts were obtained via automatic download from the MQDQ website and saved as JSON files. Our colleagues at MQDQ have now informed us that they are distributing their files in XML format (by request), but since our code was developed with sample JSON files of the kind discussed above, we proceeded with a corpus of downloaded JSON files (see Mastandrea et al. 2005–21).4 Once this corpus was obtained, special characters and certain punctuation had to be removed to ensure compatibility with the Natural Language Processing algorithms offered by CLTK.5 Before using CLTK’s toolset to analyse the corpus, a measure of ‘enumerativeness’ that would be interpretable to a machine needed to be devised. Because congeries often operates at the line level, a method for determining heavily enumerative and wholly enumerative lines would be effective and would correspond to the following formula, referred to hereafter as an ‘enumeration score’: Number of nouns matching most frequent case in a line number of words in a line Indeed, its operation at the line level makes congeries an attractive target of analysis, as code that privileges lines as units can be developed, rather than more complex (and thus, potentially more inaccurate) code that operates at the level of a ‘passage’ (placed in quotation marks because ‘passage’ must remain arbitrary in quantitative terms).6 Therefore, we ran the cleaned text through CLTK’s tokenize and parsing functions or the processes that render Latin text into discrete words (tokens) and return the grammatical information for them. Because attributes such as case are not fully interpretable without viewing a sentence as a whole, for each work in the CLTK corpus, we tokenized each sentence (which often comprised more than one line of poetry). We then reconstructed the line divisions originally present in the text, yielding, for each line, a collection of tokens and their grammatical metadata. Using these tokens and the associated grammatical information, we applied the formula above to calculate the enumeration score for every line within an individual author’s corpus. According to this metric, a score of 1 indicates a ‘wholly enumerative’ line, or that every word in the line is a noun of the same case. Identifying such lines is useful of itself, but comparing the number of lines 78

Quantitative Approaches

that score highly against the total of lines within an author’s corpus also allows us to estimate how frequently different authors write such lines.

Methodological limitations and restrictions The code returns accurate results (Table 5.1), but the nature of the research question (and the Latin language itself) presents limitations. While we analysed all poems within an author’s corpus, we only present (as we had planned before analysing results) quantitative data from elegiac couplets or dactylic hexameter. Most other meters allow much shorter

Table 5.1. Authors, corpora and enumeration scores. Author

Date

Corpus Wholly enumerative Highly enumerative size lines (=1) lines (≥0.8)

Damasus papa

384

686

6

10

C. Sollius Modestus Apollinaris Sidonius

486

2656

20

33

Orientius, episcopus Ausciensis (?)

400

1232

8

22

Magnus Felix Ennodius

521

1408

8

14

14

540

3

4

Paulinus Pellaeus, Gallus

460

635

3

6

Prosper Tiro, Aquitanus

450

971

4

8

Blossius Aemilius Dracontius, Carthaginiensis

499

5950

24

69

Venantius Honorius Clementianus Fortunatus, episcopus Pictaviensis

601

9959

39

105

Q. Septimius Florens Tertullianus, Carthaginiensis

220

1303

5

24

Maximianus poeta

550

793

3

6

Eugenius, archiepiscopus Toletanus

657

1607

6

13

Paulinus Petricordiae

450

3727

12

18

M. Aurelius Olympius Nemesianus, Carthaginiensis

250

672

2

10

Alcimus Ecdicius Avitus, episcopus Viennensis

518

3469

10

17

Severus, episcopus Malacitanus (?)

599

710

2

4

Grattius

79

A Late Antique Poetics?

Author

Date

Corpus Wholly enumerative Highly enumerative size lines (=1) lines (≥0.8)

Claudius Claudianus

400

4546

12

34

Publilius Optatianus Porfyrius

337

761

2

7

Sextus Propertius

-15

4009

10

27

pseudo Cyprianus Carthaginiensis

450

832

2

9

D. Iunius Iuvenalis

130

3870

9

29

M. Valerius Martialis

103

6655

14

33

D. Magnus Ausonius, Burdigalensis

394

3866

8

39

Pontius Meropius Paulinus, episcopus Nolanus

431

6354

13

36

Fl. Cresconius Corippus, Afer

550

6376

13

61

Luxurius, Carthaginiensis

500

501

1

4

Q. Horatius Flaccus

-8

4087

7

20

P. Papinius Statius

96

14334

22

99

Cl. Marius Victor, orator Massiliensis

400

2020

3

8

Faltonia Betitia Proba

350

709

1

5

Petronius Arbiter

66

715

1

3

P. Vergilius Maro

-19

12879

18

90

Rutilius Claudius Namatianus

410

742

1

1

Aurelius Prudentius Clemens

405

2261

3

14

T. Calpurnius Siculus

68

759

1

2

P. Ovidius Naso

18

34915

43

169

Ti. Catius Asconius Silius Italicus

101

11497

12

58

Priscianus Caesariensis, grammaticus

500

1088

1

5

Quintus Serenus

30

1107

1

2

C. Valerius Catullus, Veronensis

55

1115

1

10

-55

7376

6

31

M. Annaeus Lucanus

65

8077

6

31

Marcus Manilius

37

4253

3

16

-19

1422

1

6

T. Lucretius Carus

Albius Tibullus

80

Quantitative Approaches

Author Sedulius presbyter

Date

Corpus Wholly enumerative Highly enumerative size lines (=1) lines (≥0.8)

400

1495

1

5

50

2306

1

12

350

3301

1

7

62

650

0

4

Arator, subdiaconus ecclesiae Romanae

550

2492

0

5

Avianus

400

794

0

1

C. Vettius Aquilinus Iuvencus, presbyter

350

3253

0

7

Commodianus

250

1135

0

6

Cyprianus Gallus poeta

400

3170

0

5

19

984

0

4

Hilarius, episcopus Pictaviensis

367

508

0

2

M. Tullius Cicero

-43

755

0

3

appendix Vergiliana Rufius Festus Avienus A. Persius Flaccus

Germanicus Caesar

lines and, therefore, their analysis could potentially introduce false positives. Similarly, authors whose total output was less than 500 lines were removed from consideration. A small or fragmentary corpus containing a few lines with high scores would give the (distorted) impression that the author preferred heavily enumerative verses generally. Similarly, any line that is less than three words has been removed, to prevent artificially inflating the scores of half-lines. In spite of our interventions, these problems cannot be solved entirely. For example, the issues discussed above make Damasus appear as the most heavily enumerative poet in history, but this proceeds only from a small corpus size and the presence of one poem composed solely of congeries.7 We intervened in the presentation of chronological data so that incidence of wholly and heavily enumerative lines could be graphed over time. Many ancient authors cannot be securely dated, and the editors of MQDQ employ conservative dating. When precise dates are absent, MQDQ will list a range, indicate an era (e.g. aetate Diocletiani, at the time of Diocletian), or specify a century (e.g. saec II a., second century bce ). We have reduced all ranges to the final year within the range. When historical figures provide relative chronology, we have listed the date as the end of that figure’s life (e.g. aetate Hadriani, at the time of Hadrian = 138 ce , amicus Ovidii, a friend of Ovid = 18 ce ). The code also presents limitations. Since the analysis targets the frequency of case use within a line, nouns and adjectives (or participles) that agree and occur within the same line will score as highly as an asyndetic instance of verbal overload like the passage of Sidonius above. For example, we list below a line of Venantius Fortunatus that scored as a 1, against a line of Ausonius that only scored as a 0.857: 81

A Late Antique Poetics?

inguine nudato, pede iacto, poplite panso. With his groin laid bare, his foot thrust out, his knee extended. (Ven. Fort. Mart. 4.493) ore manu flatu buxo fide voce canentes Singing with their expressions, hands, blowing, boxwood pipe, string, voice. (Auson. griph. [15.]23) Since Ausonius’ line, detailing the competition of Sirens and Muses, contains a string of enumerative ablatives before a nominative participle, it scores a 6/7 (= 0.875) while Venantius’ scene of a monk who warms his naked body scores a 1 because it is comprized solely of ablatives. The reader must decide if lines like those of the warming scene should be considered ‘wholly enumerative’, but the phenomenon of an entire hexameter or pentameter line occurring in only one grammatical case seems particularly late antique in our estimation. If the reader should disagree, however, they should be aware that that assumption is implicit in this study. Our methodological reliance on parsing also means that any line featuring an accusative participle that governs accusative objects will score as heavily enumerative when it is not.8 Finally, the parser cannot measure -que as a separate word, so not every line that scores as a 1 is asyndetic. Beyond the complications that arise from a parsing-focused analysis, we must address the issue of accurate parsing. The nature of Latin makes parsing with 100 per cent accuracy difficult even for human readers, let alone an automated algorithm like CLTK’s. For example, 20 lines with a score of 1 were identified within Sidonius Apollinaris corpus of elegiac couplets and hexameters, which is itself 2,656 lines. This figure may be short of the actual total, however, in light of the presence of two features common to the corpora of many individual authors: ambiguous forms and Greek words. Consider a second instance of congeries in Sidonius that details the labours of Heracles: sus, leo, cerva, Gigans, taurus, iuga, Cerberus, hydra, hospes, Nessus, Eryx, volucres, Thrax, Cacus, Amazon, Cres, fluvius, Libs, poma, Lycus, virgo, polus, Oete. (carm. 15.141–3) The boar, lion, deer, Giant, bull, the yokes, Cerberus, the Hydra, the host, Nessus, Eryx, birds, the Thracian, Cacus, the Amazon, the Cretan, the river, the Libyan, the apples, Lycus, the virgin, the heavens, Oeta. Each of these three lines should be scored as a 1, but only the second line appears as such. In the first line, the neuter noun iuga was incorrectly identified as an accusative, rather than a nominative. In the third line, however, the neuter noun poma is correctly identified as a nominative, but the algorithm struggles with the Greek loan word Oete, and, 82

Quantitative Approaches

unable to determine its case, simply lists it as a noun. Thus, both the first and third lines of the excerpt have scores of 87.5 (7/8), rather than 1. We have attempted to compensate for this by considering lines that score above 0.8 ‘heavily enumerative’. To evaluate this measure, consider carm. 9.95 above (cui sus, leo, Gigas, Amazon). This line, although parsed accurately, scores an 0.833 (5/6) due to the presence of cui. While accurately parsed here, the line also prompts discussion of Greek loan words. Greek forms sometimes lead to artificially high scores, since genitives in -es (-ης) can be misparsed as thirddeclension accusative or nominative plurals in Latin. Similarly, Greek neuter plurals or third-declension accusative singular nouns may be confused with Latin neuter plurals or feminine singular nominative nouns. Thus, a relatively small corpus with a high incidence of lines containing Greek loan words may appear far more or far less enumerative than it is. Greek words always entail parsing challenges, but their issues can be easily remedied if they are transmitted in Greek characters (unlike the examples above, which were transliterated in the MSS). The algorithms cannot process Greek characters natively, so, for these to be analysed, they would first need to be transliterated. Although we developed and implemented a solution that transliterated loan words automatically, the parsing difficulties presented by Greek loan words prompted us to exclude them from analysis, omitting all lines that transmit words in (Unicode) Greek characters. This decision seems to be corroborated by conversations held by the developers of CLTK as they evaluated the different but related issue of multilingual corpora (see Johnson et al. 2015). All corpora offer unique challenges, and the corpus of Latin poetry (whether MQDQ’s or another) is no different. Interventions have been performed to mitigate the risk of error and to offer a more comprehensive view of the data (