Seneca's Troades: A Literary Introduction with Text, Translation and Commentary 9780691197715

Table of contents :
Contents
Preface
Abbreviations
Introduction to the Text
L. Annaei Senecae Troades
Commentary
Bibliography
Indexes

Citation preview

Seneca's TROADES

Seneca's

TROADES A Literary Introduction with TEXT, TRANSLATION,

and

COMMENTARY

Elaine Fantham

PRINCETON UNIVERSITY PRESS PRINCETON, NEW JERSEY

Copyright © 1982 by Princeton University Press Published by Princeton University Press, 41 William Street, Princeton, New Jersey In the United Kingdom: Princeton University Press, Guildford, Surrey All Rights Reserved Library of Congress Cataloging in Publication Data will be found on the last printed page of this book Publication of this book has been aided by grants from the Paul Mellon Fund of Princeton University Press and Bollingen Foundation This book has been composed in Linotron Bembo Clothbound editions ofPrinceton University Press books are printed on acid-free paper, and binding materials are chosen for strength and durability

Preface Abbreviations

vu xt

INTRODUCTION TO THE TEXT

3 3 9

1. Tragedy at Rome I. Seneca's Predecessors and Contemporaries II. The Dating of Seneca's Tragedies 2. Seneca's Motives and Methods in Composing the Tragedies I. The Case for Philosophical Drama II. Seneca's Attitude toward Poetry and Po~sis lll. Aemulatio and the Role of Rhetoric in Senecan Poetry 3. The Medium of Senecan Tragedy: Stage Drama, Recitation, or Private Study? 4. Iliou Persis: The Growth of the Tradition I. The Epic Tradition II. Polyxena in Greek Tragedy Ill. The Roman Tradition IV. Did Seneca Compose from a Dramatic Model? V. Seneca and Euripides Appendix to Chapter 4: The History of the House of Priam 5. Death and the Dead in Seneca's Troades 6. Senecan Syntax and Features of Diction I. Syntax II. Morphology and Orthography 7. Seneca's Use of Dramatic Meters in the Troades I. Iambic Trimeters II. The Choral Meters 8. The Text

76 78 92 92 98 104 104 110 116

L. ANNAEI SENECAE TROADES

125

15 15 19 24 34 50 50 57 61 68 71

COMMENTARY

The Prologue (1-163) Act Two (164-408)

V

203 231

CONTENTS

Act Three (409-860) Act Four (861-1055) Act Five (1056-1179)

Bibliography

271

332 365 389

INDEXES

Index Locorum General Index

vi

397 408

PREFACE I doubt if Aristotle would have approved of Seneca's Troades; measured by the criteria of the Poetics, it is hardly a proper tragedy. There is no single line of development, depicting a hero who passes from good to bad fortune through an error of understanding or moral judgment. Instead we have a dual plot based on the enforced death of two moral innocents, and dominated by the essentially passive state of two bereaved women, again without responsibility for the events that have brought their ruin. But Seneca himself would not have conceived his play in these terms: rather the whole action is the city's tragedy, and its downfall is presented as a demonstration of the arbitrary reversals of fate, and to a lesser extent, of retribution for the national inheritance of guilt. The legend of Troy meant more to a Roman than to a Greek, since this was his own origin, and the city's destruction had made possible the birth of Rome: but this proud theme of Augustan poetry, like Aeneas himself, is absent from Troades, in which recidiva Pergama (Tro. 472) is disowned as a deceitful hope. If the play ends in the annihilation of Troy, it is not because Seneca was out of sympathy with the Aeneid, or afraid to include a prophetic element in his plays: there is prophecy in Troades, which looms into closer focus as the play draws to its end, but it is prophecy of further retribution, when Greeks will succeed Trojans as the object of fortune's anger. To read Senecan tragedy with understanding requires more than knowledge of the underlying myth, and of tragic tradition: we must share Seneca's own literary inheritance of familiarity with Roman poetry, his aesthetic principles, and the philosophical viewpoint that determined his interpretation of the established myths. We must also take into account the influence of nondramatic genres of poetry, and the implications of recitation, rather than stage performance, as the dominant medium of his time. In this introduction, I have tried to restore Troades to its full context, seeing it, as it were, three-dimensionally. The first approach follows an outline of the history of tragedy at Rome, which will give special emphasis to the tension between professional playwriting and a more dilettante dramatic poetry, to the growing independence of Romans from Greek tragedy, and to the diversity of compositions that were called tragoediae.

vii

PREFACE

Secondly, we should consider Seneca's own knowledge of earlier poetry, both Greek and Roman, and his moral and aesthetic views on poetic composition. What did he see as the function of poetry? How did he conceive the process of imitation or competition with his predecessors? The third approach can be made along a narrower front. To evaluate his relationship to earlier literary presentations of the fall of Troy, we must trace the development of the legends from Homer and his cyclic successors through Greek tragedy into Roman tragedy and epic. To know that divergent forms of a myth exist-say the death of Astyanax-and show that Seneca's version is previously found in Arctinus or Accius is not to prove that he imitated or even directly knew this particular account. Every other line of the play reflects his intimacy with the Trojan narratives of the Aeneid and Metamorphoses, but beyond this certainty we cannot link an earlier account to Seneca's work without evidence from his other writings that he knew this work, or internal clues in the form of close imitation oflanguage, narrative detail, or sequence of argument and action. But once comparison with Euripides' Hecuba or Troades has shown Seneca's familiarity with these plays, it becomes legitimate to consider the omission or change of features in the Greek tragedy as part of Seneca's literary choice. What he has chosen not to write also contributes to our understanding the play before us. Earlier generations, offended by the violence of Senecan tragedy and its preoccupation with horror and vengeance, rejected its rhetoric and condemned its divergence from classical Greek tragedy. Ours is a more violent and excessive age, but present-day appreciation of Seneca is not simply the fellow-feeling of companions in excess; it springs from a response to the sound of Seneca, and the urgent rhythm and passion of his speeches. This fundamental quality of his poetry outweighs the lapses of taste or consistency that no critic can ignore, and has led me, in both Introduction and Commentary, to recognize Seneca's weaknesses, but to concentrate on analyzing and displaying his real poetic achievement. I could not have written this commentary without the help of many people, and my thanks to some of them will be quite inadequate return for their kindness. I might begin with the gallant stranger from Tivoli whose persistence enabled me to find the elusive Tabula Iliaca on a crowded Saturday night in the Capitoline Museum: I could not have found it alone. More conventionally, let me thank my students at the University of Toronto whose intelligence and sensitivity

viii

PREFACE

led me to hope that an edition on this scale could be of use and interest to student readers. I originally hoped that I might be able to benefit from Otto Zwierlein's eagerly awaited Oxford text of the tragedies; but this Herculean achievement is still in the future. Instead Professor Zwierlein has been kindness itself in sending me material and answering my queries. Professor Alexander MacGregor of the University of Illinois, Chicago Circle, has been equally generous, providing me with his collation of an important manuscript and showing me his work to refine our understanding of the stemma codicum. But I have been most blessed in the generosity of two friends: my colleague Richard Tarrant, himself the author of the most distinguished commentary on Senecan tragedy available today, has given me all the help that I have asked for and more. I would have been lost without his guidance on textual matters but it would have been poor repayment of his kindness had I yielded to temptation and imposed the entire manuscript upon him. Indeed it would have been a surrender of responsibility to take advantage of all the help he could have given. This commentary is very different from his own, but I hope he will approve of it. Gordon Williams encouraged me to write and to stand by the unorthodoxy of some of my views. With his unfailing good humor and patience he read through my second draft, and helped me to give better form to many of my arguments, and to rethink others. Despite his wise counsel there will be some errors and eccentricities for which I alone am to blame. Dr. Brad Inwood read the manuscript before and in proof. It was a great reassurance to invoke his expertise in improving the form of my text and reducing the burden upon the staff of Princeton University Press. To them, especially to Joanna Hitchcock, I shall continue to be grateful for their care, consideration, and efficiency since the inception of this book. The publication of this introduction and commentary was assisted by grants from the Mellon and Bollingen Foundations. September 1980

ix

ELAINE

FANTHAM

ABREVIATIONS

ANCIENT AUTHORS AND TEXTS Ace. Se. (R)

Accius, Seaeniea romanorum poesis fragmenta, ed. 0. Ribbeck, 3rd ed., vol. 1, 1895. Cat. Catullus Cic. Cicero Acad. Academica Arch. Pro Archia Att. Epistulae ad Atticum Brut. Brutus de Opt. Gen. de Optimo Genere Oratorum de Orat. de Oratore de Rep. de Republica Div. . de Divinatione Fam. Epistulae ad Familiares Fin. de Finibus Place. pro Flacco Leg. de Legibus Mil. pro Milone Off. de Officiis ND de Natura Deorum Pis. in Pisonem QF Epistulae ad Quintum Fratrem Sest. pro Sestio S. Rose. pro Roscio Amerino Tusc. Tusculan Disputations V err. in Verrem Diog. Laert. Diogenes Laertius Enn. Ennius Ann. Annales, ed. Vahlen Scaenica Scaenica, ed. Vahlen (Tragic fragments not in Jocelyn, q.v.)

xi

Tragedies Eur. Ale. Andr. Ba. Hee. Heracl. Hipp. I.A. Med. Tro. Homer I/.

Od. Hor. Epist. Sat. Luer. Macr. Sat. Ov. A.A. Am. ex P. F. Her. Met. R.A. Tr. Pindar Isthm. 01. Pyth. Plautus Aul. Bacch. Capt. Mil. Trin.

Tragedies, ed. Jocelyn Euripides Alcestis Andromache Bacchae Hecuba Heracleidae Hippolytus Iphigenia Aulica Medea Troades Iliad Odyssey Horace Epistles Satires Lucretius Macrobius Saturnalia Ovid Ars Amatoria Am ores ex Ponto Fasti Heroides Metamorphoses Remedia Amoris Tristia Isthmians Olympians Pythians Aulularia Bacchides Captivi Miles Gloriosus Trinummus

ABBREVIATIONS

Pliny Ep. N. H. Plut. A/ex. Cat. Min. Cic. Them. Prop. Quint. Sal!. Cat. Jug. Seneca Ben. Brev. Clem. Const.

Epistles Natural History Plutarch Alexander . Cato Minor Cicero Themistocles Propertius Quintilian Sallust Catiline ]ugurtha de de de de

Bene.ficiis Brevitate Vitae Clementia Constantia Sapientis Epistulae Morales ad Helviam ad Marciam Natural Questions ad Polybium de Providentia de Tranquillitate Agamemnon Hercules Furens Hercules Oetaeus Medea Octavia Oedipus Phaedra Phoenissae Thyestes Troades

Ep. Helv. Marc. N.Q. Polyb. Prov. Tranqu. Ag. H.F. H.O. Med. Oct. Oed. Phae. Pho. Thy. Tro. Seneca Rhetor Contr. Controversiae Suas. Suasoriae

Soph. Ant. O.T. Phi/. Stat. Theb. Ach. Suet. Aug.

DJ. Tib. Claud. Ner. Tac. Ann. Dial. Val. Flacc. Val. Max. Vir. Aen. Eel. G.

Sophocles Antigone Oedipus Tyrannus Philoctetes Statius Thebaid Achilleid Suetonius Augustus Divus ]ulius Tiberius Claudius Nero Tacitus Annals Dialogus de Oratoribus V alerius Flaccus Argonautica Valerius Maximus Memorabilia Virgil Aeneid Eclogues Georgics

GENERAL TERMS ab. ace. codd. dat. f., ff. gen. in f. re cc. v., vv.

ablative accusative codices dative folio(s); following genitive infinitive recentiores versus, versi (line, lines)

REFERENCES AND PERIODICALS Standard reference works and journals are abbreviated according to the lists of abbreviations in the Oxford Classical Dictionary (Oxford, 1970) and the American journal of Archaeology 82 (1978).

xii

Introduction to the Text

Tabula Iliaca Capitolina, from 0. Jahn and A. Michaelis, Griechische Bilderchroniken (Bonn, 1873), plate 1.

I. SENEGA'S PREDECESSORS AND CONTEMPORARIES Among writers of tragedy Accius and Pacuvius are the most renowned of the ancients for the high seriousness of their thought, their weighty language, and their impressive characters. They lack brilliance and the final touches in the polishing of their plays-but that may be thought to have been a deficiency of their age, not of themselves. Accius however is conceded to have more power, while critics who lay claim to learning would have us believe this is where Pacuvius excels. Varius' Thyestes is comparable to any Greek tragedy. Ovid's Medea shows, in my view, what its author could have achieved if he had been ready to control his genius rather than to pander to it. Of my contemporaries far the best is Pomponius Secundus. Old men thought him not tragic enough, but they had to agree that he excelled in learning and brilliance. · Quintilian, Institutio 10. 1. 97-99, trans. Winterbottom.

Let Quintilian's brief sketch serve to illustrate how the Romans of the first century of the principate viewed their national record in tragedy. It is a rhetorician's appraisal, more concerned with style and argumentation than characterization or dramatic structure, but then Roman tragedy from the beginning was shaped as much by the study1 as by the working theatre. The first tragedy to be written in Latin was an adaptation of a Greek script by the Tarentine prisoner of war Livius Andronicus, 2 whose regular profession was that of a grammaticus, and who also provided a Latin text of the Odyssey for schoolroom use. Quintilian omits from consideration both Livius and his near-contemporary Naevius, 3 who wrote both tragedy and comedy. The plays may also have been experienced more often in the study than on stage by the educated Roman of Cicero's day, for on his evidence, the Roman tragedies were regular reading matter. See Acad. 2.10: quid enim causae est cur poefas Latinos Graecis litteris eruditi, legant, philosophos non legant? Fin. 1.4: ... idem fabellas Latinas ad verbum e Graecis expressas non inviti legant; de Opt. Gen. (a suspected interpolation): sed tamen et Pacuvium et Accium potius quam Euripidem et Sophoclem legunt. For a brief critic.iill history of Roman tragedy, see the chapter by Adrian Gratwick, in Cambridge History of Classical Literature, vol. 2 (Cambridge, 1980), pp. 128-37. 2 Performed at the Ludi Romani of240 B.C. The evidence for the chronology ofLivius, Naevius, and Ennius is presented by Cicero, Brut. 72-74 on the authority of Varro. On Livius Andronicus, see Fraenkel, s. v. Livius, PW Suppl. 5:598-607. There is a brief account in Beare, The Roman Stage, London, 1964, pp. 25-32. The tragic fragments are easily available in Warmington, Early Roman Poetry, vol. 2, London, 1936, 1-21. 3 See Fraenkel, PW Suppl. 6: 622-42; Beare (cited at n. 2), pp. 33-44; Warmington (cited at n. 2), 2: 110-37. 1

3

INTRODUCTION TO THE TEXT

He also passes over Ennius, whom Cicero regarded as the greatest tragic writer of Rome: 4 again in Ennius we have a poet, rather than a man of the theatre; he wrote not only the famous Annales incorporating myth and history in one epic, but also minor poetry in various sophisticated genres reflecting his erudition and Greek culture. However, his comedies were unsuccessful, and the greatness of his tragedies lay chiefly in the pathos of his monodies, which are extensively quoted by Cicero. Ennius shared with his predecessors an interest in themes based on the Iliad and post-Homeric tales of the Trojan cycle. Compare from Livius Achilles, Aiax Mastigophoros, Equos Trojanus, from Naevius, Aesiona (Priam's sister, Hesione), Equos Trojanus, and Hector Proficiscens. Among Ennius' tragedies are recorded an Achilles, Aiax, Alexandros (based on the first play of the Euripidean trilogy which ends with the Troades), Andromacha Aechmalotis, apparently set at Troy after the death and btJrial of Astyanax, a Hectoris Lytra, based on Priam's ransom of his son's body in Il. 24, and a Hecuba whose fragments show its adaptation from Euripides' play. Like the Augustan tragedians and Seneca he wrote a Medea (modeled closely on Euripides) and his last play was the Thyestes of 169 B.c. 5 The first poet to limit himself to tragedy, and first to be mentioned by Horace, Quintilian, and Tacitus, 6 is Ennius' nephew Pacuvius,7 famed for his often harsh innovations of diction; the constant epithet doctus applied to him by the critics (Cic. Brut. 155; Hor. Epist. 2.1.55; and Quint. above) implies Alexandrian erudition and allusion. It is perhaps also Alexandrian that although he lived to the age of ninety he wrote only thirteen plays. From the Trojan cycle he adapted an Armorum Judicium (obviously a theme rich in rhetorical appeal) and Iliona, another version of the sequel to the death of Polydorus told in Euripides' Hecuba. Like his predecessors, Pacuvius wrote Praetextae, dramas in tragic form based on episodes of Roman history. The interest of these lost plays lies in their lack of Greek models and their On Ennius, see the introduction to The Tragedies of Ennius, ed. H. D. Jocelyn, but note thatJocelyn does not comment on the unassigned fragments. For Cicero's praise, and quotations ofEnnian monody, see especially Tusc. 3.44-45 (from the Andromacha), de Drat. 3.21£, Acad. 2.88-89, and Div. 1.42 and 66-67. The best study of Cii:ero's knowledge of the tragedians is W. Zillinger, Cicero und die Altromische Dichter, Erlangen, 1911; see also H. D. Jocelyn, "Greek Poetry in Cicero's Prose Writings," Yale Classical Studies 23 (1973). 5 On the date of the Thyestes, see Cic. Brut. 78. 6 In Tacitus' Dialogus, Pacuvius and Accius are mentioned only by the modernist A per (20.5; 21. 7), who treats them without distinction as hopelessly archaic. 7 On Pacuvius, see Beare (cited at n:2), pp. 79-84; Warmington (cited at n. 2), 2: 158323; M. Valsa, M. Pacuvius Poete Tragique, Paris, 1957; and S. Mariotti, Introduzione a Pacuvio, Urbino, 1960. There is a new edition of the fragments by G. D'Anna, M. Pacuvii Fragmenta, Rome, 1967. 4

4

TRAGEDY AT ROME

use of annalistic subject matter: for although the dramatists will surely have adapted appropriate scenes and motifs from Greek patriotic plays such as Euripides' Heracleidae or Erechtheus, they will ultimately have had to impose their own dramatic techniques in order to convert epic material for the stage-in some ways a precedent for the creative process by which Seneca was later to dramatize episodes from Virgil and Ovidian Epic. Accius, 8 the last of the professional playwrights, was also very much a man of letters, writing on orthography and style as well as the history of the Roman theatre. In the list of his surviving titles, there are many Trojan themes: Achilles (or Myrmidones), Ant'enoridae, Astyanax (to be discussed below, in chapter 4, section 3), Deiphobus, Epinausimache (the Trojan attack on the Greek fleet known to us from 11. 15-16), Hecuba (probably the same as his Troades), Neoptolemus, and a Nyktegresia based on the exploits of Ulysses and Diomedes told in Il. 10. Reviewing the titles of all the dramatists, we note severalHector Proficiscens, Hectoris Lytra, Epinausimache, and Nyktegresia--which may have been based not on Greek dramas but directly on the epic narrative of Homer. Accius' Armorum Judicium, on the other hand, like that of Pacuvius, adapted Aeschylus' Hoplon Krisis. Accius reflects the tendency of Roman tragedy to rhetoric, and is most renowned for his powerfully emotive speeches. An anecdote from Quintilian9 suggests his real interests: asked why he did not become a court orator, since he showed such powers of disputation in the tragedies, he explained that in tragedy the argument followed his direction, but in court the adversary would advance arguments that he did not wish at all. His exclamatory style, rhetorical questions, sententiae, and heightened diction show that drama was above all a medium for his rhetorical skills: Accius had vires (Cic. Brut. 155) and elevation (Hor. Epist. 2.1.55). Nothing is known about his sense of the stage, but it is perhaps ominous that the gala revival of his Clytemnestra 10 in 55 B. c. paraded six hundred mules in Agamemnon's See Beare (cited at n.2), pp. 119-27; Warmington (cited at n.2), 2: 324-577; and Lucio Accio frammenti, ed. A. R. Barrile, Bologna, 1969. Leo's discussion of Accius' life and writings in Geschichte der Romischen Literatur, Berlin, 1913, pp. 392-408, has not been superseded, and a detailed study of the tragic fragments is still needed. 9 5.13.43: Aiunt Accium interrogatum cur causas non ageret cum apud eum in tragoediis tanta vis esset, hanc reddidisse rationem, quod illic ea dicerentur quae ipse vellet, in foro dicturi adversarii essent quae minime vellet. 10 Reported by Cic. Fam. 7.1. Accius' plays were greatly favored in Cicero's day, not least for their political allusions; compare the performance of Eurysaces and the Praetexta Brutus in support of the motion to recall Cicero in 57 B.c. (Sest. 119-22 and 123). His Astyanax, (or possibly another Trojan play involving Astyanax) was performed in 54 (see Att. 4.15), and his Tereus was Brutus' choice for the Ludi Apollinares of 44 B. c. (see Att. 16.2).

8

5

INTRODUCTION TO THE TEXT

retinue. Nevertheless he was an experienced playwright, and we can sympathize with his refusal to rise from his seat in the College of Poets for the young aedile Caesar Strabo, "because he knew his superiority in their common literary pursuits. " 11 Strabo was a symptom of the amateur's intrusion, which may have done more harm to tragedy than to any other literary genre at Rome. Cicero's generation is full of litterateurs; in the Brutus, after reporting the succession of serious dramatists from Livius to Accius, he is left with only dilettantes to mention, such as Strabo and the knight C. Titius, who shared Strabo's wit, but whose plays displayed wit at the expense of tragic feeling. 12 Quintus Cicero's four tragedies composed over sixteen days, possibly including a Troades, 13 were the product of isolation in the off-season of Caesar's Gallic campaigns, Caesar's own Oedipus being a youthful experiment, rightly suppressed by his heir. 14 Augustus too, for all his love of drama, recognized the weakness of his Ajax and let it die unpublished. 15 There was a great flowering of interest in poetry, as Horace protested in his letter to Augustus: besides the tragic ambitions of the younger Pisones, there were more persistent authors, such as Asinius Pollio, 16 whose tragedies are praised by Ho race and Virgil as worthy of Sophocles. They did not survive him. Servius reports that he wrote tragedies in both languages, and Tacitus' Dmlogus criticizes his tragedy and oratory alike as archaizing; the evidence suggests imitative composition without any original contribution. One play in this generation achieved enormous success: the Thyestes of Varius Rufus was commissioned by Octavian for the games of 29 B.c. to celebrate his triple triumph and rewarded with a gift of a million sesterces. 17 For Quintilian, Varius is the only Roman dram11 Val. Max. 3.7.1: quod in comparatione communium studiorum aliquanto se superiorem esse confideret. On Strabo, compare Cic. Brut. 177: sunt aliquot orationes, ex quibus, sicut ex eiusdem tragoediis lenitas sine nervis perspici potest. 12 Brut. 167: easdem argutias in tragoedias satis ille quidem acute, sed parum tragice transtulit. 13 Cf. Cic. QF 3.5(6)7: quattuor tragoedias sedecim diebus absolvisse cum scribas, tu quicquam ab alio mutuaris? et EAEOI quaeris, cum Electram et t Trodam t (Troades? Troada?) scripseris? 14 Suet. DJ. 56.7 15 Suet. Aug. 86; cf. Macr. Sat. 2.4.2. 16 The evidence ofHor. Epist. 2.1.111: scribimus indocti doctique poemata passim; the tragic ambitions of the Pisones seem to be implied by Ars Poetica 24f., 366-90. On Pollio see Vir. Eel. 8.9, with Servius ad loc., Hor. Odes 2.1.9-12, Tac. Dial. 21.7: Pacuvium certe et Accium non solum tragoediis set etiam orationibus suis expressit; adeo durus et siccus est. 17 See now Eckard Lerevre, Der Thyestes des Lucius Varius Rufos: Zehn Uberlegungen zu seiner Rekonstruktion, Akad. der Wissenschaften u. der Literatur, Mainz, 1976, n9, reviewed by R. J. Tarrant, CQ 29 (1979): 149-50, and R. J. Tarrant, "Senecan Drama

6

TRAGEDY AT ROME

atist of classic stature, fit to be measured against any Greek; according to Tacitus' speaker, he is more famous than any contemporary orator. But for all his fame only four certain lines survive, less than is preserved of Maecenas' private verses. Ovid's Medea is coupled with Varius' tragedy by both Quintilian and Tacitus, and the assessment of Quintilian suggests that this tragedy may have been more disciplined than many of his other works, but again quotations survive only in Quintilian and Ovid's contemporary Seneca the Elder. There is no evidence that it received a public performance; indeed, a later comment by Ovid seems to deny it. 18 But then without the incidental quotation of the didascalia for V arius' play, there would now be no evidence even for that showpiece production. Obviously, once recitation became a public occasion, plays would be more easily introduced to literary circles in recitation, and subsequently perpetuated by publication for readers. Yet despite the growing predominance of mime and spectacle, tragedies continued to receive public performance. Little can be inferred about the tragedy written by Mamercus Aemilius Scaurus, which llngered Tiberius because it contained abuse of Agamemnon, 19 but the sources imply that it circulated only privately and had not received any public showing. A closer parallel to Seneca as a public figure and a dramatist is the distinguished ex-consul and triumphator Pomponius Secundus, stepbrother of Caligula's wife Caesonia, who survived to enjoy the favor of Claudius. Tacitus reports that Claudius intervened on his behalf when he was abused by the theatre crowd, and Pliny (Ep. 7.17) tells a charming anecdote of Pomponius' reaction when his tragedies were criticized by his friends: he used to declare that he would "appeal to the people," and follow his. own or his friend's choice on the basis of the applause or silence of the crowd. This presupposes three stages in shaping his tragedies: first recitation, then public performance, and finally publication in a form modified for readers. Cichorius has shown us how to interpret the brief notice in Quintilian on the dispute between Pomponius and and Its Antecedents," HSCP 82 (1978): 258-61, on the common characteristics that can be predicated ofVarius' and Ovid's plays. 18 On the dating of the Medea see H. Frankel, Ov)d, a Poet between Two Worlds, Berkeley, 1945, chap. 5 with notes, pp. 193-94. In default of evidence, Frankel suggests a date for the tragedy of about 8 B. c. It is quoted at Quint. 8.5.6; Seneca Rhetor, Suas. 3. 7; Ovid's denial, Tr. 5. 7.27: nil equidem feci, ut tu scis ipse, theatris. 19 Dio 58.24; Tac. Ann. 6.29; Suet. Tib. 61. Suetonius does not name Scaurus, or the play, but quotes Tiberius as penalizing a dramatist for criticizing Agamemnon in a tragedy; Agamemnon would be out of place in an Atreus as named in the other sources. It was Macro who informed Tiberius about the content of the play, quoting the verses that could be interpreted unfavorably.

7

INTRODUCTION TO THE TEXT

Seneca in their Praefotiones about the legitimacy of archaism. 20 Quintilian's memory would go back to about A.D. 50, but in view of Seneca's exile until 49 and Pomponius' absence as governor of the Germanies in 50 and 51, the dispute must belong in or after A.D. 51. The discussion was oral, for the Praefationes were the preliminary remarks of the recitalist before presenting his new work; as Cichorius suggests, Seneca, who deprecated archaism must have been the critic, and it will have been Pomponius, the older writer, who defended the use of eliminare (a favorite word of Ennius, Pacuvius, and Accius, which was thereafter attested only in satire). Pomponius, like Seneca, wrote a play on the theme of Atreus and Thyestes, and among his fragments there is a piece of choral lyric dealing with the fall ofTroy, and spoken by a Trojan chorus. 21 But Quintilian's praise for Pomponius in 10.1. 98 recalls Caesar Strabo: "old men thought him not tragic enough . . . he excelled in learning and brilliance." Why then did Quintilian, who gives extensive personal attention to Seneca in his survey of prose writing, omit his plays from the survey of tragedy? He certainly knew them, for in 9.2.8 he quotes Med. 354 as Medea apud Senecam using the same formula of citation as his next excerpt, Sinon apud Virgilium: so the tragedies would be familiar to his readers also. Possibly he did not think of the Medea as tragedy proper? More likely the omission of Seneca sprang from his immense distaste and unwillingness to give further notice to his literary bete noire. One problem in evaluating the medium of these dramatic works is the ambiguity of the word carmina. Pomponius' Tragedies were carmina for Tacitus; Seneca too, according to Tacitus, Annals, wrote carmina, 22 and wrote them more frequently after Nero developed an interest in composition. But what were Nero's carmina? Besides the inevitable dithyrambic prize-songs with which he toured Greece, Suetonius reports 23 that he sang (that is performed) tragoediae of heroes and gods, wearing the mask, and with the masks adapted to

° C. Cichorius, Romische Studien, Leipzig, 1922,_ pp. 426-29: Quint. 8.3.31: nam memini iuvenis admodum inter Pomponium et Senecam etiam praefationibus esse tractatum, an "gradus eliminat" in tragoedia dici oportuisset. 21 Pomponius' play was called Atreus, like that of Accius; the choral fragment implying Trojan speakers is obrue nos Danaosque simul (fr.7, p. 231 [R]). 22 Carmina of Pomponius, Ann. 11.13; of Seneca, Ann. 14.52: obiciebant etiam ... carmina crebrius factitare, postquam Neroni amor eorum venisset. 23 Suet. Ner. 21: Tragoedias quoque cantavit personatus heroum deorumque, item heroidum et dearum, personis effectis ad similitudinem oris sui et feminae, prout quamque diligeret. Inter cetera cantavit Canacen Parturientem ... etc. On Nero's tragic performances see Juvenal Sat. 8.223; Dio 62.26 and 63.22; and H. Wagenvoort, "Pantomimus und Tragodie," N]bb 45 (1920): 111-14. 2

8

TRAGEDY AT ROME

Nero's features or those of his current mistress. Suetonius quotes these works by a double title, of protagonist and action: Canace Parturiens, Orestes Matricida, Oedipus Excaecatus, Hercules Insanus. But unless these were full dramas named after their most striking episode, they were most probably one-man shows, presenting a scena ed aria of Nero's own composition, and combining techniques of tragedy and pantomime. Suetonius calls both Nero's and Scaurus' works tragoediae; Tacitus refers to both as carmina. This might be ·a stylistic avoidance of the Greek form, but the contexts are too vague for us to determine the reference ofNero~s carmina, and in the Dialogus Tacitus' speakers freely mention trag6ediae (2, 3, 4). The plays so described are the recitation dramas of the moralist Curiatius Maternus. The Dialogus is set in the time of Vespasian, and Maternus has newly given a recitation ofhis Cato: there is anxiety in case he has offended authority by too passionate an identification with his hero; he is described24 as revising this work for publication and at the same time· preparing a Thyestes for recitation. The Roman and Greek plots alike are seen as vehicles for his political ideals. Obviously he is an independent composer on the Roman theme (where there is no question of a Greek model), but both works are treated as the same kind of creation; his word disposui is the equivalent of Greek diatithenai and implies the organization of his own structure of acts and scenes for each play. Both will be presented first in recitation and then in book form. Was Seneca more like Pomponius or Maternus? I see Seneca's carmina or tragoediae more in terms of Maternus' procedure and purpose, but any attempt to pronounce whether his tragedies were intended as stage plays or recitation drama must consider not only dramatic history, as we have done, but the form of the compositions themselves. First, however, it will be convenient to outline the course of Seneca's crowded life and the limited evidence from the plays themselves and from historical sources that can be used to suggest a dating for Seneca's activity as a dramatic poet. This in turn will lead us to examine his motives for turning to drama and the literary aims and values reflected in his prose writings. II. THE DATING OF SENECA'S TRAGEDIES

Seneca himself never mentions his tragedies in his prose wntmgs, and Tacitus, our main source for the public events of Seneca's life, 24 Dial. 3: Si qua omisit Cato, sequenti recitatione Thyestes dicet; hanc enim tragoediam disposui et intra me formavi.

9

INTRODUCTION TO THE TEXT

offers only one ambiguous reference to the composition of carmina. 25 Thus if we wish to assign a date or range of years to the composition of Troades or any other of the tragedies, we are reduced to various indirect approaches. The last thorough analysis of the different types of evidence available was that of Otto Herzog 26 fifty years ago. Since then scholars have if anything become more reluctant to draw conclusions from what are essentially ambivalent data, and the most recent scholarly edition of a Senecan tragedy is willing to accept only a terminus ante quem of 54 for the Hercules Furens, and the early sixties for the other tragedies. 27 While sympathetic to the idea that the tragedies are "youthful" works, Tarrant rejects many of the inferences about specific plays reported below, and ends by endorsing the non liquet of Coffey's Lustrum report twenty years back: "In general the tragedies may have belonged to any stage of Seneca's literary career." Seneca was born in or around 1 B. c., the third son of the wealthy and cultured Cordovan L. Annaeus Seneca, himself the author of a lost history of Rome and surviving memoirs of his youth in the rhetorical schools. 28 By the early years of Tiberius' principate, Seneca was a student at Rome, devoted to the philosophical school of Q. Sextius, an ascetic who combined Stoic and Pythagorean doctrine. But he was delicate, almost certainly consumptive, and after a period of ill health and convalescence in Egypt under the care of his maternal aunt, wife of the Prefect Galerius, he returned to Rome about A.D. 31 and was soon elected quaestor. His early political and rhetorical career proceeded smoothly until the accession of Gaius in 37, but his brilliance as a speaker provoked the resentment of the emperor, who according to anecdote (Cassius Dio 59.19. 7-8) would have executed him if Seneca's health had not seemed so precarious that he was not expected to live. When Claudius became emperor Seneca was seen as a dangerous figure by the emperor's first wife Messalina and exiled to Corsica, on the grounds of adultery with Gaius' sister Julia Livilla. He remained there for eight years until Messalina's downfall and Claudius' remarriage to Agrippina, who now secured his recall, and the continuation of his political career, designating him as praetor for A.D. 50. She was responsible for his Ann. 14.25, quoted and discussed below. "Datierung der Tragoedien des Seneca," Rh. Mus. 77 (1928): 51-104. v Seneca Agamemnon, ed. R. J. Tarrant: Cambridge Classical Texts and Commentaries 18, Cambridge, 1976. Coffey's verdict is quoted for his first survey of scholarship, "Seneca, Tragedies," Lustrum 2 (1957): 150. 28 For a more detailed account of his life see M. Griffin, "Imago Vitae Suae" in Seneca, ed. C. D. Costa, London, 1974, pp. 1-38, or A. L. Motto, Seneca, Twayne World Authors Series, New York, 1973. 25

26

10

TRAGEDY AT ROME

appointment as tutor to her son Domitius Ahenobarbus, adopted by Claudius in 50 as Nero Claudius Drusus Germanicus, 29 and marked out by his seniority to Claudius' child Britannicus, as the emperor's heir. Seneca was now about fifty years old, and from this time identified with the interests of Agrippina, and Nero so long as they coincided. On Claudius' death in 54 it was Seneca who composed both Nero's laudatio of the dead emperor (his least successful work30) and the brilliantly funny satire on the official ritual of deification, the Apocolocyntosis, or "Pumpkinification" of Claudius. With the succession of Nero, first Seneca's brother Gallio, then Seneca himself, became suffect consuls. Seneca held this honor in 56, but it represented a less significant claim on his time and loyalty than his continuing role as political counselor of the emperor, a responsibility he shared with the Pra~torian Prefect Afranius Burrus. At the time of Burrus' death in 62 Seneca recognized that he could not control (or survive) Nero's increasing irresponsibility, and the immorality and folly of his personal life. He went to the emperor, begging to be allowed to retire and surrender his wealth for the emperor's use. Although N.e.ro only partly acceded to this request, Seneca withdrew from court to compose his philosophical and scientific works, the Epistulae Morales and the Naturales Quaestiones. Even so, he was implicated in the Pisonian conspiracy of 65 and forced to commit suicide. 31 When would such a man have time for tragedy? It was usual for Romans of the political classes to divert themselves with poetry in their youth before passing on to inoral, political, or historical writings. We might compare Caesar's Oedipus, grouped by Suetonius (DJ. 56) among the works written a puero or ab adulescentulo, or the unidentified Greek tragedy written by Pliny at the age of fourteen. ) Again Romans tended to absorb themselves in composition away from the city, at their villas,_ in slack periods of provincial or military duty, or in exile. Thus Senecan scholars normally assign his plays to the empty years of exile, when as he reports, his mind either diverted Recall of Seneca, Tac. Ann. 12. 8. The adoption, 12.25-6; Seneca was seen as a popular figure, and a valuable ally. 12.8 is worth quoting at length: Agrippina ... veniam exilii pro Annaeo Seneca, SVnUl praeturam impetrat, laetum in publicum rata ob claritudinem studiorum eius, utque Domitii pueritia tali magistro adolesceret et consiliis eiusdem ad spem dominationis uterentur, quia Seneca fidus in Agrippinam memoria beneficii et infensus Claudio dolore iniuriae credebatur. 30 Cf. Tac. Ann. 13.3: nemo risui temperare, quamquam oratio a Seneca composita multum cultus praeferret, ut fuit illi viro ingenium amoenum et temporis eius auribus accommodatum. 31 Retirement requested, Tac. Ann. 14.52; discreetly taken, 14.56; his implication with Piso and suicide, 15.60-65.

29

11

INTRODUCTION TO THE TEXT

itself with levioribus studiis or with the more lofty enquiry into the nature of man and the universe. 32 On the other hand, Cichorius's discussion of the debate between Seneca and Pomponius over the diction of tragedy led to the conclusion that Seneca was writing tragedies after his return from exile, and Pomponius' last provincial command, in A.D. 51 or 52. 33 This is firmer evidence than the accusation made against him by unnamed opponents in 62 that he had intensified his output of poetic works since Nero had developed an enthusiasm for them (Tac. Ann. 14.52): carmina crebrius factitare postquam Neroni amor eorum venisset. The accusation may be random, the genre of poetry is unspecified, and amor eorum could equally denote love of composing his own carmina, or of reading Seneca's work. Herzog links this with Tacitus' allusion in Ann. 14.16 to Nero's new practice in 59 of composing carmina (surely lyric, if the word pangere is precise) with a group of dilettanti.34 But the carmina of14.52 could well be Seneca's epigrams, and the notice hardly justifies Herzog's inference that Seneca was writing substantial poetry, and therefore some tragedies, between 59 and 62. Internal evidence is scanty. We might consider dating on grounds of verse technique. For instance, the choral lyrics of Thyestes, Hercules Furens, and Troades are simpler than those of the other plays, being based on Horatian meters. 35 But Troades is more innovative than the other plays, since it introduces an anapaestic monody at 705f; in this it resembles Medea, which has conventional choral songs, but a complex monodic sequence in the third act (740-842). Herzog also singles out Thyestes, Hercules Furens, and Medea as early because the chorus seems to perform without participation in the soloists' actions, whereas he sees the Trojan women as more closely involved, through their kommos with Hecuba; it is true too that their third and fourth songs are in character and related to the progress of the drama. But this merely reflects the Euripidean presentation of the fall of Troy and -" Helv. 20: animus omnis occupationis expers operibus suis vacat et modo se levioribus studiis oblectat, modo ad considerandum suam universique naturam veri avidus insurgit. 33 See above, n. 20. 34 Ne tamen ludicrae tantum imperatoris artes notescerent, carminum quoque studium adfectavit, contractis quibus aliqua pangendi facultas necdum insignis erat. The collective versification described by Tacitus does not seem a suitable procedure for the composition of dramatic dialogue. 35 Herzog (cited in n.26), p. 66: but he is surely mistaken to argue that in Troades as in Phaedra and Oedipus "the chorus is always deemed to be present." See further the discussion in chapter 3 below.

12

TRAGEDY AT ROME

need not imply evolution in Seneca himself. Nor does the verse technique of the dialogue offer any leads toward a relative chronology. 36 Affinities of treatment are ambivalent as evidence. I have noted in the commentary strong resemblances of motif and phrasing between Hercules Furens and Troades; these lend support to the common assumption based on the order of plays in the Etruscus, that Troades was written next after Hercules Furens. But I have also argued elsewhere37 that the affinities between Troades and Agamemnon are partly careless echoes in Agamemnon of material from Troades, which establish the priority of the Trojan play, and partly conscious reuse in the mythological sequel of moral and architectural themes featured in the earlier play. When a poet's idiosyncrasies are so prominent in all his works, attempts to assign an order amongst them may go astray by making cross-references out of coincidence. Herzog's suggestions for a chronology based on topical allusion tend to conform with hi~ previoHs inferences, but he himself recognizes that they are speculative. 38 For Troades he suggests a dating after 49, on the ground that allusion to the lusus Troiae in 777-79 would be inspired by Nero's participation in the ritual in A.D. 47, 39 but would only date from 49, when Seneca was first concerned with young Nero. A particularly opportune time would be 53, when Nero, aged seventeen, made his first court appearance, acting on behalf of the town of Ilium. 40 But this is to be too precise. There was a long tradition of interest by the Julio-Claudians in Ilium; Tiberius had founded the sacrarium gentis Iuliae at Bovillae in which the tabula Iliaca Capitolina was found; Germanicus had been sent to visit the site early L. Strzelecki, De Senecae trimetro iambico quaestiones selectae, Krak6w, 1938, p. 4, notes that Troades is more sparing in the use of antilabai (changes of speaker within the line) than Thyestes or Medea, and closer to Phaedra in practice, but the incidence of first-foot proceleusmatics and tribrachs is higher than in any other play (ibid., pp. 78, 92) so that the statistics would be in conflict if used as an index of relative dating. 37 "Seneca's Troades and Agamemnon: Continuity and Sequence," C] 77 (1982): 118129. 38 Herzog (cited in n. 26) p. 83. See also the recent cautious attempt of P. Grimal, Seneque, ou la conscience de l'E111pire, Paris, 1977, pp. 424-28. 39 At the secular games; cf. Tac. Ann. 11.11: Sedente Claudio circensibus ludis cum pueri nobiles equis ludicrum Troiae inirent interque eos Britannicus imperatore genitus et L. Domitius adoptione mox in imperium et cognomentum Neronis adscitus, favor plebis acrior in Domitium loco praesagii acceptus est. If anything, Seneca's allusion to the lusus Troiae would be more valuable before Nero's adoption was ensured than after, when it might merely recall that he had shared the honor with the displaced Britannicus. '"'Tac. Ann. 12.58: causa Iliensium suscepta Romanum Troia demissum et Iuliae stirpis auctorem Aeneam aliaque haud procul fabulis vetera facunde executus perpetrat, ut Ilienses omni publico munere solverentur. 36

13

INTRODUCTION TO THE TEXT

in Tiberius's principate (Tac. Ann. 2.54) and Nero's interest in Troy persisted into his principate. The lusus Troiae certainly suggests a terminus ante quem (though surely of 47, not 49) but we can say little more. The question of cross-references with Seneca's philosophical works must be left aside as unprofitable, since the dating of the dialogues is itself tentative; the firmest cross-reference I have noted is between the allusions to periodic conflagration of the Universe at Tro. 38f. and the same theory advocated in Marc. 26.6, but deprecated in Polyb. 1.2-4. But there are equally close verbal echoes of this chorus in the Epistulae Morales which we know to have been written after A. D. 62, 41 while other affinities could be adduced with the undatable de Brevitate Vitae. The most reliable criteria are technical aspects of diction or versification, offering a consistent pattern of formal change from play to play. Recently Professor John Fitch has taken as an index the increasing proportion between sense-pauses in the dialogue of each play, a statistic he shows to give valid results when applied to the Sophoclean corpus and Shakespearian tragedy. 42 The percentages indicate three groups of plays: Agamemnon, Phaedra, and Oedipus, with between 32.4 percent and 36.8 percent of pauses within the line; Medea, Troades, and Hercules Furens with between 47.2 percent and 49.0 percent of internal pauses; and finally Thyestes and Phoenissae with 54.5 percent and 57.2 percent, respectively. Given the accepted dating of Hercules Furens prior to the parody of its dirge in the Apocolocyntosis in 54 B. c. the chronological implications would exclude, for example, allusions to events of Nero's principate seen by several scholars in Oedipus, or my own hypothesis that the affinities between Troades and Agamemnon arise from Seneca's imitation of elements in Troades in the other play. Yet it is also supported by Fitch's statistics for Seneca's growing license in shortening the final -o, in which Thyestes and Phoenissae clearly go well beyond the other plays. We might also note that this dating puts Seneca's boldest metrical experiments at the beginning of his career, and thus implies paradoxically that he is showing increasing confidence in his handling of the trimeter at the same time that he is becoming more cautious and restricted in his use of lyric meters. See below, chapter 5. "Sense-Pauses and Relative Dating in Seneca, Sophocles and Shakespeare," A]P 102 (1981): 289-307.

41 42

14

S

Two

ENECA'S MOTIVES AND METHODS

IN COMPOSING THE TRAGEDIES

I. THE CASE FOR PHILOSOPHICAL DRAMA

Judged by traditional criteria, the writing of tragedies and of moral treatises seem far apart, and to the Romans of Cicero's time also they must have seemed different pursuits, of a different order of seriousness; the wise and mature thinker who would compose philosophical treatises would surely have no use for tragedy. Thus up to the time ofErasmus, who edited both Seneca's tragedies and his prose corpus, it was assumed that the author of the tragedies was not the philosopher but a brother or son. 1 Scholars since the nineteenth century, who have accepted the identity of moralist and tragedian, have instead asked why the philosopher should have spent his time on writing tragedy; many, as we saw above, 2 have inferred that Seneca wrote his tragedies while still relatively young, perhaps specifically during the period of exile when he was prevented from the active life of public service that Romans saw as the role of a responsible member of the governing class. By his own admission, 3 Seneca even during his exile made the distinction between moral writing and leviora studia: are the tragedies these leviora studia, or should the student look for an ulterior moral motivation behind the literary form, a lesson concealed as entertainment that would explain the philosopher stooping to imaginative fiction? There is surely no reason why a creative writer should act from only one motive, and a literary form that serves an ideology need not be denied aesthetic purpose. Thus it seems to me that the right formulation of the question is not "Was Seneca's motive in composing tragedies literary or philosophical?" but rather "To what extent was Seneca led to write tragedy by nonliterary motives?" or "Can we determine whether his purpose was primarily moral or aesthetic?" On Erasmus' concern with Seneca see W. Trillitzsch, "Erasmus und Seneca," Phi/of. 109 (1%5): 270-93, and "Seneca Tragicus, Nachleben und Beurteilung im Lateinischen Mittelalter," Philol. 122 (1978): 133. Erasmus himself saw the moral prose writings as the work of the rhetorician Seneca the Elder, and distinguished the tragedian from the moralist; cf. Ep. 2091, 530f. (Allen): nam tragoediarum opus eruditi quidam malunt Senecae filio tribuere quam huic; sunt qui fratri Senecae adscribant. 2 Introduction, chap. 1, sec. 2, above; Herzog, Rh. Mus. 77 (1928): 51-62, arguing against Birt, N]bb 27 (1911): 352ff. 3 Helv. 20. 1

15

INTRODUCTION TO THE TEXT

The case for an ethical purpose was most powerfully expressed in the last generation by Berthe Marti in an influential paper "Seneca's Tragedies: A New Interpretation. " 4 But her thesis that Seneca composed the tragedies to provide a systematic moral program is contingent on assumptions that few scholars can now accept. She argued that the order in which the tragedies were preserved in the Codex Etruscus, Hercules Furens, Troades, Phoenissae, Medea, Phaedra, Oedipus, Agamemnon, Thyestes, and Hercules Oetaeus, was that intended by Seneca, whether or not he composed the plays in that order, to convey a systematic demonstration of Stoic moral teaching. Thus the corpus opens with the ordeal of Hercules, the great Stoic exemplum, as he learned to endure life burdened with his unwitting murder of his family, and ends in his triumphant apotheosis in the Hercules Oetaeus. She divided the other tragedies into three progressive groups; Troades and Phoenissae were designed to illustrate the implications of defeat, showing death as liberation for the passive victims of war; Medea and Phaedra demonstrated the evil consequences when passion (foror) defeated reason (ratio) within the soul, consequences that were evil for the offender, for his victims, and for the whole world surrounding him. In the next three plays, Oedipus, Agamemnon, and Thyestes, she saw dramatizations of the hazards of kingly power, and retribution for its misuse; finally, she argued, all that seemed unexplained or unjust in the outcome of these tragic actions would be resolved in the reward of apotheosis earned by Hercules when he had attained the wisdom of the Stoic sage. Even after scholars had raised objections to her interpretation, Marti persevered in attempts to vindicate the authenticity and crowning function of the Oetaeus. 5 But the formulation of her moral theory leaves too many difficulties: thus the order of the plays given in the Etruscus cannot be shown to have superior status over that of the A family, 6 and since the investigations of Axelson 7 few students now adhere to Senecan authorship of the Oetaeus. Again too little of Phoenissae survives for us to infer its projected contents, while the fact that it is incomplete jars with Marti's claim that it was deliberately ordered in the collected plays as part of an integrated moral TAPA 76 (1945): 216-45. See also Grimal, Seneque ou la Conscience de !'Empire (Paris, 1977), pp. 424-31. 5 "La Place de l'Hercule sur L'Oeta dans le corpus des Tragedies de Seneque," REL (1949): 189-210. For a criticism of Marti's thesis see N. T. Pratt, "The Stoic Basis of Senecan Drama" TAPA 79 (1948): 1-11. 6 H.F., Thy., Pho. (called Thebais), Phae. (called Hippolytus), Oed., Tro. (called Troas), Med., Ag., Oct., and H.O. See chap. 8 below. 7 In Korruptelenkult: Studien zur Textkritik der Unechten Seneca- Tragoedie "Hercules Oetaeus" (Lund, 1967).

4

16

SENECA'S MOTIVES AND METHODS

demonstration; its unfinished state belies considered publication by the author. There is no cogent reason why the tragedies of defeat should precede those dealing with the passion of the individual. Indeed, judged as vehicles for the instruction of the Stoic proficiens, the plays, especially those based on female lust and jealousy, seem inappropriate. We would have expected drama more like Corneille's Polyeucte, to glorify principled resistance to the tyrant or the mob by the sage whom wisdom has released from fear of death, the iustus et tenax propositi ofHorace Odes. 3.3.1f. .There is no doubt that the plays conform with Stoic understanding of psychology; Herington has illustrated from the Medea the close ad~erence to the description of anger in the de Ira, and E. C. Evat¥.> earlier demonstrated, chiefly from this play, the use of Stoic treatises of physiognomy to describe the effect of passion upon the facial features. 8 Both Medea and Thyestes portray the growth and ruinous triumph of personal evil within and beyond the individual soul; again several plays feature an argument between the protagonist and a subordinate, who urges the case for restraint9 and warns of evil to follow: we may note such scenes in the second act of Medea, Phaedra, Thyestes, and Agamemnon; but the warning given to Oedipus, like that given by Agamemnon in the second act of Troades, does not ultimately affect the predetermined outcome of the tragedy. Marti has stressed the concern with death that dominates Troades, which she ·sees as intended to dispel false fears and teach men to welcome death as liberation. Certainly for the reader in sympathy with Stoic values it can be exhilarating to contemplate the heroic deaths of the child and the maiden, who make a virtue of necessity and, by the manner of their death, give moral value to a martyrdom they cannot choose or reject. It is also true that the famous denial of the life after death in the second choral interlude harmonizes with views frequently asserted by Seneca in the moral writings. 10 But in this play and others there are mythological elements incompatible with Stoic teaching-the only too real ghost of Achilles, or the underworld traversed by Hercules and Theseus in the Hercules Furens. See Herington, "Senecan Tragedy," Arion 5 (1966): 455; E. C. Evans, "A Stoic Aspect ofSenecan Drama, Portraiture," TAPA 81 (1950): 169-84. 9 See Herington, ibid., p. 453f. Troades differs from the other plays because it is the subordinate who urges on the evil course of action that will duly triumph, while the superior, Agamemnon, resists it. 10 Marti, TAPA 76 (1945): 225-27. See also introduction, chap. 5, below.]. Dingel, Seneca und die Dichtung (Heidelberg, 1974), treats the tragedies as late works, written to express the pessimistic obverse of Seneca's philosophical teachings. In his study of Troades, pp. 92-94, he perversely argues that the events of the play are intended to override and invalidate the enlightened view of the second chorus.

8

17

INTRODUCTION TO THE TEXT

It is to be expected that Senecan choruses will advocate moderation and advise withdrawal from dangerous power and corrupting luxury: even the non-Stoic choruses of fifth-century Athens had traditionally counseled restraint, and mistrust of luxury typifies all genres of Roman poetry from Lucretian didactic to Horace's Odes and the Satires of Horace, Persius, and Juvenal. There are, however, passages in the Senecan tragedies that show the powerful instinct of one who writes imaginative literature to create character and let it respond to situations. Seneca shows his literary independence of moral stereotypes as much by the weakness ofhis "good" characters as by the dynamism of his passionate offenders. Moralizing roles like Agamemnon in Troades may be shown as weak or petty, yielding to the temptation to score a rhetorical point; Agamemnon descends to cheap abuse of Pyrrhus' birth; Andromache resorts to physical violence and abuse against Ulysses (668f., 750f.) and the noble Polyxena dies with a gesture of violent hatred towards her infernal suitor (1157-59). The good Stoic does not hate-nor does he hope or fear-but good drama depends on hope and fear; it is only Andromache's futile hope that gives vitality to the long central act of Troades. Seneca writes to stir pity, fear, revulsion, and indignation; two of these emotions were the recognized response of classical tragedy, but revulsion and indignation are perhaps peculiarly his own. 11 It is true that he also stirs admiration for the morally good, but alongside the respect which the audience feels for the suffering Hercules is an even stronger, amoral, fascination with the uninhibited evil of Medea or Atreus. Just as Lucan's Caesar is the most powerful character in the Bellum Civile, so Seneca's wrongdoers carry his dramas by their own momentum. It is their emotional impact that made Senecan tragedy, despite its unremitting high tension and its melodrama, so popular in the Renaissance and so influential upon the history of the European stage. Our own generation since 1945, when Marti wrote her paper, has responded to Senecan tragedy for the same reason. It is not just the lack of a conciliatory ending which differentiates Senecan tragedy from the Christian dramas of the French seventeenth century, Corneille's Polyeucte or Racine's Esther and Athalie; the plays impress through their evocation of characters beyond restraint, and through their power to seize the attention and emotions of their public. 12 All these reactions can be illustrated from Seneca's report of the spectators at the sacrifice ofPolyxena: cf. Tro. 1129, odit scelus spectatque (revulsion); 1134-36, revealing their indignation through quotation of their words; and 1147-48, tremunt mirantur et miserantur (fear, admiration, and pity). 12 Cf. Herzog, Rh. Mus. 78 (1928): 62, who draws the analogy with Pomponius. Like Pomponius, Seneca wrote from a genuine literary impulse, because he wanted to create

11

18

SENECA'S MOTIVES AND METHODS

What I have tried to convey is ultimately a literary effect, and the result of a peculiar kind of imaginative talent. The furious vitality of Seneca's characters shows, I believe, that he felt the same urge to create these roles and set them in action as a novelist who is moved to compose a fiction or interpret a historical episode. It is easier to speak generally of a literary motivation than to determine why Seneca chose to present his creation in dramatic form. The Romans traditionally thought in terms of genre, relating specific genres to certain types of material, but in the post-Augustan generation, genre itself was in the melting pot; orators sought poetic coloring for their prose, while at least one poet-Seneca's nephew Lucan-was seen as a better model for orators than for poets. 13 The problem requires us to consider the influence on Senecan tragedy of three literary forms: of tragedy itself, consisting of the Greek and Roman predecessors who dramatized the content of his plays; of rhetoric, the medium of his education, which can be studied from the reminiscences of his father, and indirectly in Seneca's own prosewritings; and finally of narrative poetry, above all the epics of Virgil and Ovid, which dominated education and literary taste in Seneca's lifetime and must have colored the conception of poetry held by all his generation. Before considering the internal evidence for these influences in the plays themselves, it will be useful to examine the prose works for traces of his interest in the different poetic genres in both Greek and Latin, and his expressed tastes and judgments about poetry. Il. SENEGA'S ATTITUDE TOWARD POETRY AND POESIS

The most recent studies 14 of Seneca's attitude to and knowledge of the poets have demonstrated the surprising narrowness of Seneca's characters in action. Thus even Eckard Lefevre, a strong advocate of Seneca's philosophical motivation for writing the tragedies, acknowledges (in "Schicksal und Selbstverschuldung in Senecas Agamemnon," Hermes 94 [1966], reprinted in Senecas Tragiidien, Wege der Forschung 310 [Darmstadt, 1972], pp. 457-76) that the "real moral theme" is not only overlaid but often contradicted by elements of popular philosophizing and traditional poetic material. He stresses the priority of the interpretatio Stoica more insistently in his later paper on the same topic, "Die Schuld des Agamemnon," Hermes 100 (1972): 64-91, esp. 65. 13 Cf. Tac. Dial. 20.5: exigitur enim iam ab oratore etiam poeticus decor, non Accii aut Pacuvii veterno inquinatus sed ex Horatii et Virgilii et Lucani sacrario prolatus; Quint. 10.1. 90: Lucanus ardens et concitatus et . . . magis oratoribus quam poetis imitandus. 14 W. S. Maguinness, "Seneca and the Poets," Hermathena 88 (1956) 7: 81-98; Giancarlo Mazzoli, Seneca e la Poesia (Milan, 1970); and J. Dingel, Seneca und die Dichtung, pp. 48-58.

19

INTRODUCTION TO THE TEXT

allusions to poetry in his prose works. There are no quotations in Greek outside the Apocolocyntosis, but it was probably a point of stylistic etiquette to avoid this, since, for example, he even composes his own verse translation of an excerpt from Cleanthes' hymn to Zeus (in the same trimeters we know from his tragedies) so that the flow of elegant Latin shall not be disturbed. Although Lucilius, the addressee of his letters, undoubtedly read Greek with ease, Seneca's Greek philosophical forbears are not quoted but paraphrased. Of special interest for us is his knowledge of Homer and the Greek tragedians. If we based our assumptions of his reading on the quotations in the Bpistulae Morales we would find only one classic allusion to Homer's comment on the mourning of Niobe (Il. 24.602) and an erudite reference to Homer's knowledge of the potter's wheelthere are no allusions to the Odyssey in all the letters. Even so, like every educated Roman Seneca must have learned to read from Homer-see N.Q. 6.23: quisquis primas litteras Graecas didicit, scit ilIum apud Homerum £vootx8ova vocari. 15 We need not infer from scholarly arguments reproduced in Ben. 1.3 about the names of the Graces in Homer and Hesiod, or the derivation of an Ennian line from Il. 5. 749 in Bp. 108.34, that Seneca himself had this academic concern with Homeric criticism; indeed he repudiates such discussions as futile in Bp. 108 and Bp. 88.6f. But his allusions to the last book of the Iliad at Tranqu. 2.12 and de Ira 2.33.5 bear out the appreciation of Homer that he shows in the essay of consolation to Polybius, who distracted his leisure by translating Homer into Latin prose and Virgil into Greek (Polyb. 11. 5): null us erit in illis scriptis liber, qui non plurima varietatis humanae incertorumque casuum et lacrimarum ex alia atque alia causa fluentium exempla tibi suggerat. 16 Tears and the uncertainty of fate were especially important to the bereaved Polybius, yet in Seneca's tragedies they are a source of exaltation; it was for these lacrimae rerum, as Mazzoli has shown, that Seneca most valued Homer, even though Virgil held fitst place in his loyalty and admiration. The Greek tragedians, on the other hand, are barely quoted; Aeschylus and Sophocles are named only once-as authorities on the flooding of the Nile in an inherited catalogue at N.Q. 4.2.17. Of the two excerpts translated from Euripides, one, from the Danae, is mis"Whoever has learned his elementary Greek knows that he is called 'Earthshaker' in Homer." 16 "There will not be any book of their poetry that fails to offer countless examples of human vicissitudes and unforeseen chances,. and tears that flow for one reason after another." 15

20

SENECA'S MOTIVES AND METHODS

attributed in Ep. 115.15, probably because it was an indirect quotation from the moralist source ofSeneca's theatrical anecdote. His tragedies do show familiarity with both plots and dialogue elements of Euripidean tragedy, and the freedom of his choice of action and thought for imitation argues for a longstanding familiarity, rather than the short-term study appropriate to a translator. 17 Seneca's citations of the Latin poets are more influenced by his sense of style than by the-nonetheless relevant-question of genre. His allusions to Ennius are the traditional excerpts available through Ciceronian quotation, from the de Re Publica (Ep. 108.30f.), from Brutus (Ep. O.C.T. 2:540 == Gellius 12.2) and from the Tusculans (Polyb. 11.2). There are no first-hand quotations from Plautus, Terence, or the republican tragedians; Gellius (12.2.10) reports that he rejected their diction and could not understand what merit it had in the eyes of Cicero and Virgil: Vergilius quoque noster non· ex alia causa duros quosdam versus et enormes et aliquid supra mensuram trahentis interposuit quam ut Ennianus populus agnosceret in novo carmine aliquid antiquitatis. 18 This quotation shows the paradox of Seneca's attitude; for it was Virgil's ac;hievement that prevented his successors from being able to accept the element of inspiration in archaic poetry which Virgil himself had understood. As it is, Virgil, poetarum maximus, dominates Seneca's literary consciousness and is quoted more than four times as often as any other poet, even Ovid, the favorite of his addressee Lucilius (N.Q. 4.2.2). Virgil was now universally known, guaranteeing to the writer who introduced a Virgilian allusion into his argument the full understanding of quotation and context by his readers. This familiarity may explain the incidence of quotations from Georgics or Aeneid used to bear out a moral statement that could equally have been supported from the evidence of tragedy. See also Herington (reviewing Tarrant, Agamemnon), Phoenix 32 (1978), p. 273; he points to the evidence of Quintilian for the study of Greek poetry by boys under the grammaticus, and Quintilian's explicit recommendation of Euripides (10.1.67). He adds that, with the exceptions of the Hercules Furens and Thyestes, the titles and themes of the Senecan tragic corpus belong to fifth-century Greek tragedies which were incorporated in the ancient school-selection of -seven plays each by Aeschylus and Sophocles, and ten by Euripides; hence the natural inference that the young educated Roman of Seneca's era had studied some classical Greek tragedies. Seneca could therefore expect a moderate acquaintance with them in his public. Tarrant himself, in "Senecan Drama and Its Antecedents," HSCP 82 (1978): 258, notes that the lack of allusion to Greek tragedy in Seneca's prose works "cannot be pressed .too hard," since these writings reveal virtually nothing about his activity as a poet. 18 "Our great Virgil worked in some harsh and irregular lines that overflowed the meter, just so that Ennius' nation could recognize a touch of archaism in the new poetry." 17

21

INTRODUCTION TO THE TEXT

Dingel offers another interpretation for the neglect of tragedy in Seneca's prose writings: that as a moralist Seneca held aloof from the entire genre of tragedy because he wanted to study only works that would advance his progress toward the beata vita; he could, Dingel argues, find this stimulus in Virgil but not in tragedy. 19 The positive claim for Virgil is true, but the contrast is false, as Marti20 has illustrated from the N.Q. 6.2.1-2. There Seneca quotes as a message for the human race Virgil's comment on the fall of Tray (Aen. 2.354): una salus victis, nullam sperare salutem. 21 He reflected this belief in the words of every Trojan captive and the exemplary deaths of his young martyrs in the Troades; he could also have found this message implied in Euripides' repeated rejection of false hope for Polyxena, Astyanax, and Troy itsel£ There is nothing antiphilosophical, as Dingel would suggest, in the genre of tragedy, and Seneca's neglect of the genre is adequately explained by his low esteem for the Roman adaptations of Greek tragedy, and the predominance of Virgilian epic. As a moralist Seneca passes over poets whose skill elsewhere earns the honor of imitation by Seneca the poet; the lesser genres are left aside, with no reference to Catullus, or Propertius, or Tibullusindeed there is virtually no allusion to Ovid outside the Metamorphoses. Except in the satiric Apocolocyntosis Seneca quotes only from hexameter poetry, from Lucretius, several times quoted in excerpts of two or more lines, from Virgil's didactic Georgics, from the Aeneid, and from Ovid's Metamorphoses, giving special preference to the cosmic first book and the Pythagorean speech of book fifteen. Horace, whose odes Seneca echoed in meter and diction in the tragedies, is represented only by quotations from the Satires. Because of its moral content, satire achieves quotation along with the more dignified genres. The only exceptions to this principle are the several excerpts from Maecenas' poetry, not confined to the letter on stylistic decadence (114) but serving elsewhere as a paradigm of art ruined by loss of moral strength-Bp. 92.35: habuit enim et grande et virile, nisi ilium secunda discinxissent. What did Seneca want from poetry? At the least it could soothe the bereaved (Polyb. 11.2) and calm the inflamed (de lra 3. 9.1: lectio illum carminum obleniat et historia fabulis detineat). It attracted him as a superior vehicle for moralizing. He can praise (Bp. 33.6, 94.27) the efficacy of praecepta compressed into verse, and singles out the pedestrian Publilius Syrus (Bp. 108. 9) for his superiority over prose 19 20

21

Dingel, Seneca und die Dichtung, p. 58. TAPA 76 (1945): 226. "There is but one hope for the conquered, to hold no hope of survival."

22

SENECA'S MOTIVES AND METHODS

in stimulating the audience to shame and moral reform, quoting Cleanthes' dictum: sensus nostros clarie>res carminis arta necessitas efficit. 22 He uses Virgil, whom he loves most, as a basis for sermons, even where Virgil intended no secondary allegorical 111eaning, as at Aen. 2. 726-29, which is turned in Ep. 56.12 into a text on courage. We can agree in Ep. 108 when he rejects the grammarians' approach to Virgil as inadequate and irrelevant, because we see their antiquarian comments had no literary valu~ but what evidence is there that he had literary appreciation in the modern sense of aesthetic, rather than moral, enthusiasm for the poets? Again and again he shows us that he is deeply aware of the power of poetry to move the spirit, and of the quality of spiritual exaltation which is close to poetry; cf. Seneca Tranqu. 1.14: ubi se animus cogitationum magnitudine levavit, ambitiosus in verba est, altiusque, ut spirare, ita eloqui gestit et ad dignitatem rerum exit oratio; oblitus turn legis pressiorisque iudicii sublimius feror et ore iam non meo. 23 He knew the Platonic tradition of the inspired poet (Ep. 58.17) and shared Plato's fear of the poet's power to provoke human emotion for evil-as in Ep. 115.12: carmina poetarum quae adfectibus nostris facem subdant 24-but escaped the consequence of Plato's condemnation of the poets because he saw the pity and fear of the auditorium as: motus . . . animorum moveri nolentium, nee adfectus sed principia proludentia adfectibus (de Ira 2.2.5). 25 Passions were evil and destructive of the soul's calm and capacity for good, but these were involuntary movements of the spirit, preliminary exercises testing control of passion. Thus Seneca sees the poet as able both to feel and to cause in others exalting emotion conducive to virtue. At the level of mere craftsmanship, we can deduce his idea of the process of composition from his advice on how the literary artist should read so as to foster his own creativity. He advocates alternation of reading and composition, but he stresses that the artist must separate absorption from creation by a lapse of time to allow for assimilating what is read. He relies on two analogies, to the skill of bees in converting nectar into honey (Ep. 84.3) and to digestion, (ibid. 6-7) in which food must not be left whole and undigested lest "The confining discipline of verse makes our perceptions more vivid." "When my spirit is uplifted by the grandeur of its thoughts, it grows ambitious in its diction and yearns to express itself more loftily, just as it breathes more loftily; then the style emerges worthy of the dignity of its subject, and forgetful of convention and more inhibiting judgment I soar aloft with a speech that is not mine." 24 "Poems that set a torch to kindle our emotions." 25 "The voluntary emotions of the soul, not passions, but the first stirrings that precede the onset of passion."

22

23

23

INTRODUCTION TO THE TEXT

it remain foreign to the body; so we must really digest intellectual food, alioqui in memoriam ibunt, non in ingenium. 26 His measure of artistic capacity is that the man of real talent will impose his own stamp on everything he has absorbed from any source, so that separate components will form a unity in which the identity of any author imitated is not perceptible. Here Seneca is talking of oratio, prose speech, and more particularly of the formation of a personal style, but we may equally apply to poetry his repudiation of piecemeal imitation, and pastiche. Stylistic imitation is to be something diffused, transforming the original into a new idiom. We are often told that the Romans felt no inhibitions about repeating familiar material, seeing a positive challenge to artistry in the rehandling of traditional themes. Seneca endorses this general principle in a letter answering Lucilius' proposal to write a description of his visit to Sicily. Seneca encourages his friend to write about Aetna, hunc sollemnem omnibus poetis locum, 27 adding (Bp. 79. 5) that Virgil had done this perfectly, but it had not discouraged Ovid, nor did either poet deter Cornelius Severus from treating the topic. He distinguishes between material that is worked over, like a plowed field (subacta), and material that is worked out like a mine. Far from being exhausted, the topic of Aetna offers increasing scope, and the advantage goes to the most recent poet; he finds the words ready-made, and when he has reorganized them they have a new appearance-aliter instructa novam faciem habent. Like Ho race, 28 Seneca sees existing literature as public property to be freely exploited. Ill. AEMULATIO AND THE ROLE OF RHETORIC IN SENECAN POETRY

Lucilius' poem on Aetna was to be a purely intellectual exercise; the goal was a new presentation of traditional narrative material and conventional responses to it. So we can expect Seneca himself, in artistic reaction to his forbearers, to practice adaptation of their language as a technique for pointing new emphases, or highlighting new emotional colors. Sometimes this adaptation is disguised, drawing ele"Otherwise it will enter the memory, not the intellect." On imitation, see Mazzoli, Seneca e la Poesia, pp. 87-91, "Imitazione ed Originalita," and pp. 91-96, "Criteri d'Imitazione;" also A. M. Guillemin, "L'Imitation dans la Literature Romaine," REL 2 (1924): 35-37. 27 "That obligatory commonplace of all poets." 28 Ars Poetica 131-32: publica materies privati iuris erit si/non circa vilem patulumque moraberis orbem, "subject matter publicly available will fall to your private right, so long as you do not stick to the dreary beaten track." 26

24

SENECA'S MOTIVES AND METHODS

ments from more than one passage in the Augustans and transferring motifs from one context to another, but often it will be explicit, aiming to be recognized as a variation, like Ovid's adaptation of the phrase plena deo allegedly from Virgil. Ovid claimed that he had used it in his play Medea, non subripiendi causa sed palam mutuandi, hoc animo ut vellet agnosci. 29 When he chooses, Seneca deliberately tests his skill by working close to an admired model, as in his imitation of Aen. 2.270£. in Troades 438f., or of Met. 12.108£. at Troades 215f. Here he would have felt that only part ofhis artistry was appreciated by those who read his work unaware of his model-the goal is overt aemulatio. The kind of artistic dispute that arose over such conspicuous imitation can be illustrated from the memoirs of Seneca's father at Contr. 7.1.27. He quotes two lines of Varro of Atax which Virgil was supposed to have reworked at Aen. 8.26-27: nox erat et terras animalia fessa per omnes/alituum pecudumque genus sopor altus habebat. In fact it is more likely that Virgil's.lines were based, like Varro's, on Apollonius Rhodius Argonautica (3. 749-50) and quite independent of the other Latin poet. Ovid had criticized Varro's line omnia noctis erant placida composta quiete, arguing that Varro should have cut away the second half and written simply omnia noctis erant. Although father Seneca saw that this changed the meaning (and syntax of noctis), he expresses no preference, treating each version as suitable to the poet's intent. Yet if we look for the motive behind Ovid's suggestion, I think it will help us to understand the strength and weakness of baroque poetry. Varro's line has a weak, because predictable, ending; the adjective placida anticipates the content of composta quiete (whereas Virgil's sopor altus habebat gives in full measure the awaited sense of envelopment, with deep sleep holding them secure). But Ovid is not merely eliminating a weak half-line, he is trading on the implications of night, requiring his readers to see in the characterizing genitive noctis a whole set of associations. A similar cult of the elliptical and epigrammatic is shown by the emphasis on characteristic behavior, the predication of a set of responses and actions for each role, which is common to the declamatory writers; we will find it in Senecan tragedy as the participants argue about types-Tyrants, Mothers, Captives, Victors-in a way that does not merely suggest thought patterns but dictates phraseology. Materque tota coniuge expulsa redit, says Medea (928); we must gloss, "all the feelings of a mother return, driving away the attitudes of a wife." Or from the great agon of Andromache and Ulysses in the third act of Troades, compare 626 Seneca Rhetor. Suas 7.3. The phrase plena deo does not occur in the extant text of Virgil. See E. K. Borthwick, Mnemosyne 25 (1972): 408-12.

29

25

INTRODUCTION TO THE TEXT

matrem timor detexit, "her fear has betrayed the mother's nature." Even names are used with this pointed intensification of reference; Ulysses invokes his guile: nunc advoca astus, anime, nunc fraudes, dolos/nunc toturn Ulixem, "all that Ulysses stands for" (613-14); or argues that Hector's greatness as a fighter dooms his son as a threat to Greek peace: et si taceret augur haec Calchas, tamen!dicebat Hector, "yet all that Hector was declared this as clearly as overt speech could have told us" (534-35). To gloss these trenchant phrases helps us to appreciate their fierce economy. But this kind of typological shorthand is just one outcome of the quest for original and effective expression. In fact it is impossible to separate discussions of aemulatio in Ovid or any subsequent poet from the ideals and demands of contemporary rhetoric. Yet the effect of the declamatory training that Seneca received in his youth is of a different order from the influence exercized by the great Augustan poets, since it does not offer a rival model for his poetic output, but directs the manner in which his competitive imitation will modify the poetic material. It was Leo who expressed most forcefully, and perhaps unfairly, the role of rhetoric in Senecan tragedy, coining the phrase tragoedia rhetorica for what he saw as a perversion of the Greek tragic form: cuius indoles breviter sic describi potest ut ~So~ in ea nullum, Jta8o~ omnia esse dicatur. nam quae ad mores spectant sententiis comprehenduntur, affectus plene et diffuse repraesentantur, oratione omnis generis coloribus sensibusque instructa, descriptionibus et narrationibus undecumque arcessitis et ubicumque inlatis. 30 Not only in specific discussion of declamatory influence but throughout the introduction to his edition of the tragedies Leo offered many insights into the unquestioned rhetorical coloring of Seneca's tragic diction, which have since been amplified by S. F. Bonner, in his study Roman Declamation. 31 But the tragedies show more coherence and consistency of outlook than Leo acknowledged, and it requires a greater love of poetry and richer creative ability to compose Agamemnon, for all its faults, or Troades, than to devise a suasoria like those of Arellius Fuscus or Cestius Pius on the theme deliberat Agamemnon an Iphigeniam inmolet, negante Calchante aliter navigari fas esse (Seneca Rhetor, Suas. 3). But if the influence of rhetoric serves to explain the limitations "Its nature can be briefly described in this way: characterization has no part in it, but the play on emotions is all-pervading. For moral judgments are compressed into aphorisms, while the passions are fully and widely presented, and the style is arrayed with nuances and sentiments of all kinds, and set descriptions and narratives are brought together from every kind of source to be inserted in every possible place," quoted from Seneca Tragoediae (Berlin, 1963), 1: 147-59. 31 (Liverpool, 1949), pp. 149-67. 30

26

SENECA'S MOTIVES AND METHODS

rather than the merits of Senecan tragedy, one must still recognize its effects in order to distinguish the general patterns imposed by Seneca's training upon his poetry from the particular imitation of his poetic models, and from his own contribution of interpretation or expression. We have seen how Seneca's predecessors practiced aemulatio; before examining his own techniques in relation to Virgil and Ovid, I would like to add to the discussions of Leo and Bonner some comments on post-Augustan taste and its reflection in Senecan tragedy. For this our best source is Quintilian, especially his views, expressed in books 8 and 9, on the problem of corrupta eloquentia and its origin. It is significant that Quintilian cites the tragedies only once, for the rhetorical question quas peti terras iubes (Med. 453), framed, as Quintilian points out (9.2.8), to evoke an emotional effect of resentment against Medea's interlocutor. The contribution of rhetoric to Senecan tragedy lay more in figures of thought, such as this, than in figures of diction. Indeed there was no figure of language that had not already been exploited to the full in Ovid's poetry, both epic and elegiac. But however versatile Ovid's inventio in devising new figures of thought, the ensuing generation went beyond him in the preference for obliquity, for allusion, irony, and innuendo. In the declamations of the schools the speaker could choose the characterization and circumstantial detail of his fictional brief more freely than any tragedian working with an established myth, but even so, once the natural interpretation of a declamatory context had been preempted by an early speaker, his successors had to resort to a more distorted or paradoxical formulation: the need for novelty and the desire to impress imposed excess and artificiality upon the latecomer. Quintilian's discussion of contemporary trends in oratory illustrates these failings in what has been called the pointed style. In the praefotio of the eighth book he reproaches the orators of the new generation, the one most influenced by the prose if not the verse, of our Seneca. "We beat around what could be said simply out of our relish for words; we repeat what we have said adequately, and say in several phrases what is apparent in one. In general we think it better to imply rather than say most things;, (8. Praefotio 24-25); this leads to condemnation of orators who borrow figures of speech from the most affected and extravagant poets (a corruptissimo quoque poetarum) and pride themselves on requiring real cleverness to be understood. Thus he regarded the cult of allusion and adianoeta (8.2.20) even in poetry as a mark ofbad artistry. We might illustrate adianoeta (ordinary words used to convey an obscure sense) or indeed the fash-

27

INTRODUCTION TO THE TEXT

ionable figure called noema (8.5.15) from this or any scene of the Troades; utrimque est pater says Andromache at 650; that is, both father and child are in some sense Hector; perdere est patriam grave/gravius timere argues Helen at 912-13, moving from a conventional sententia to the obscure patriam timere, explained by her fear of punishment by Menelaus. Both the fashion for sententiae and the high esteem given to brevity help to explain the age's obsession, in both prose and verse, with the pregnant word. The same pursuit of aphorisms also contributed to the staccato rhythm of both declamation and tragedy, for, as Quintilian pointed out, any sententia is in a sense terminal: subsistit enim omnis sententia (8.5.27). The listener is brought to a stop and has to renew his momentum afresh. At the same time the tendency to look for a secondary meaning in every phrase led both speaker and listener to burden language with a significance it could not convey. Quintilian himself gives to the loaded use of plain terms like homo, vir, vivere, the name emphasis (8.3.83), but he sees this procedure primarily as a figure of thought and so treats it more fully, when he discusses irony and its variants at 9.2.44£. Irony is the most conspicuous figure produced by play, not on the form, but on the reference of language, and Quintilian gives it an extended discussion. A favorite type in verse is the mock command, e.g., Aen. 4.381, i sequere Italiam ventis, which Dido utters contemptuously at the departing Aeneas. From Troades compare Achilles' ghost, ite, ite inertes manibus meis debitos/auftrte honores (191-92), or Hecuba's last words at 116567. There is irony of statement in Agamemnon's retort to Pyrrhus at 318, at non timebat tunc tuus, fateor, parens . . . , or Andromache's insult to Ulysses at 755, nocturne miles, fortis in pueri necem. In discussing the related figure of emphasis Quintilian attributes to it three advantages (9.2.64-66): the speaker could imply what it was unsafe to say outright, or what would have been improper, or he could merely exploit emphasis for variety and stimulus:' ipsa novitate et varietate magis quam si relatio sit recta delectat. 32 In our tragedies, as in the declamations, there is a great reliance on the use of figured speech (figurata oratio, or schema in a special sense, cf. 9.2.65, 66), and it is noteworthy that Quintilian reports this mannerism as a fashion of his own youth, that is the Neronian period of Seneca's prime (9.2. 77). It was of course especially favored in communication with stage or declamatory tyrants, as a satisfying method of implying allegations without incurring retribution, but as Quintilian is quick to "It gives more delight by its sheer novelty and variety than if the narrative had been straightforward."

32

28

SENECA'S MOTIVES AND METHODS

observe, this was often quite unrealistic; most examples of this type of diplomatic irony were only too intelligible even to an unimaginative tyrant, and the dominant motive was really the third-the speaker's quest for novelty. Such endless ingenuity led incompetent declaimers to utter innuendoes that were contrary to their own logic, or loyalties, or characterization: Itaque non solum si persona obstaret rectae orationi, quo in genere saepius modo quam figuris opus est, decurrebant ad schemata, sed faciebant illis locum etiam ubi inutiles ac nefariae essent (9.2.79). 33 I cannot illustrate precisely this from Troades, for Seneca usually maintains a dear grasp of the loyalties of his personae, and their will to die relieves him and them of the need to conceal their opinions, but there is a comparable sacrifice of consistency to ingenuity in, for example, Hecuba's conflicting comments on Priam's lack ofburial, 34 or Agamemnon's use of Achilles' ransom of Hector's body, firstly to praise Achilles' respect for the suppliant Priam, in contrast with the brutal murder committed by his son (312), but later (325-26) to reproach Achilles with the withdrawal from war that permitted Priam's safe passage. In general, however, we find in Senecan tragedy this very overworking of irony and allusion that strains and can even induce resistance in the listener, while the repetitiousness and excessive reliance on aphorism of Quintilian's contemporaries had been an affiiction, not only of Seneca's generation (in Quintilian's youth), but of the late Augustans before him. Indeed Seneca's father had tried to warn his sons, with a shrewd comment on Ovid and other Augustan declaimers, against the faults that we find in his own son's writing. He blames Montanus, "the Ovid of the schools," for spoiling his own bons mots by repetition (sententias suas repetendo corrumpit, Contr. 9.5.17) and formulates the famous criticism of Ovid that he could not leave well alone: nescit quod bene cessit relinquere. But his illustration, taken from Met. 13.503-5, is particularly relevant for students of the tragedian. The occasion is "So they resorted to figures not only if respect for persons was an obstacle to straightforward speech, a category in which there is more need of restraint than of figures, but they also gave them room even when the figures were harmful and vicious." 34 At Tro. 30, Priam has all Troy as his tomb, quem Troia toto conditum regno legit, but in 55 (criticized by Banner, Roman Declamation, p. 166, as an "unnecessary point"), although father of so many children, caret sepulchro Priamus ... /ardente Troia. But in 137f. the chorus will use his many children to make a different point: post elatos/ Hecubae partus regumque gregem/postrema pater funera cludis. R. Kassel, Untersuchungen zur Griechischen und Romischen Konsolations-Literatur (Munich, 1958), p. 43, illustrates the similar contradictory use of a single motif for both lament and consolation from the Consolatio ad Liviam (393f., opposed to 95) and Statius Silvae 2.6. I have noted such conflicting uses wherever they occur in Troades. 33

29

INTRODUCTION TO THE TEXT

Hecuba's outcry against the dead Achilles' demand for the sacrifice of Polyxena, also treated in act 4 of Seneca's Troades. Ovid had written cinis ipse sepulti!in genus hoc saevit, tumulo quoque sensimus hostem,/ Aeacidae fecunda foi; but while Seneca the Elder blames Ovid for presenting all three variations on the theme "even dead, Achilles attacks our family," his son particularly admires this very passage, and at Tro. 955-57 builds up his own variations based on lines 501-2 of Ovid's text: adhuc Achilles vivit in poenas Phrygum/adhuc rebellat? 0 manum Paridis levem./cinis ipse nostrum sanguinem ac tumulus sitit. He has echoed Ovid's double allusion to ashes and tomb, but added the vivid physical image of ashes absorbing blood and made a new point; the double verbs vivit and rebel/at balance the double nouns, and the relatively uncommon rebellat graphically suggests renewal of war; finally the whole is twisted to make a reproach against Paris, whose arrow slew Achilles, but to no effect; he is levis, ineffectual, as he was always levis, irresponsible. In fact the larger narrative of Met. 13.399-575, presenting the Fall of Troy as the setting for Hecuba's metamorphosis, was Seneca's chief inspiration for both narrative detail and stylistic or emotional coloring in Troades. Ovid interweaves the Fall of Troy and sacrifice of Polyxena with the tragedy of Polydorus (429-38, 534-75) following the order of action in Euripides' Hecuba, but he has influenced more in Seneca's play than the two messenger narratives of Achilles' apparition and Polyxena's death scene. Seneca has kept Polyxena silent, and has excluded from his death narrative any counterpart of her speech in Ovid (457-73), but he has drawn repeatedly on the long lament of Hecuba (494-532) for the characterization of both Hecuba and Andromache. I list instances of certain or probable imitation in the order of Ovid's narrative from Met. 13, indicating change of context or speaker, and quoting both authors, wherever competitive variation (aemulatio) is apparent. 408-10: Ilion ardebat, neque adhuc consederat ignis/exiguumque senis Priami Iovis ara cruorem/combiberat. Seneca stresses the continuing flames in his prologue 16-20, diripitur ardens Troia, and rephrases the hyperbolic reference to Priam's blood; 50, ensis senili siccus e iugulo

red it. 415-17: ... illis de turribus unde/pugnantem ... /saepe videre patrem monstratum a matre solebat. Seneca echoes the symmetrical antithesis of 417, but modifies the allusion, 1071f.(cf. 1074, paterna puero bella monstrabat senex). Both authors draw on Aen. 2.455-57.

30

SENECA'S MOTIVES AND METHODS

425-28: Hecuba . . . tamen unius hausit,/ inque sinu cineres secum tulit Hectoris haustos./Hectoris in tumulo . . . crinem lacrimasque reliquit. Seneca has given the hair and tears to Andromache in her ritual over the living Astyanax (799-800, 806-7) as well as the kisses of Met. 13.424; he transfers to Andromache unchanged the desperate desire for any trace of Hector's ashes (811-12). 441-43: quantus cum viveret . . . temporis illius vultum referebat Achilles/quo . . . . Seneca varies quantus by the simple ingens umbra (Tro. 181) and replaces the circumstances of Achilles' anger by three coordinate occasions of his triumphs 182-89, qualis ... aut cum ... aut cum. 445-48: Immemores ... mei disceditis ... utque meum non sit sine honore sepulchrum,/placet Achilleos mactata Polyxena manes. Seneca combines these motifs in Tro. 191, manibus meis debitos!auferte honores, solvite ingratas rates, and retains the verb mactare in 196, but he has changed the question of 445 to a scornful command, with an implied threat (per nostra ituri maria), and inserted here an allusion to his great anger, omitted above at the equivalent place to Met. 44243. The "bridal" motif is his own innovation. 449f: In 450: quam iam prope sola fovebat. Ovid, mindful of Polydorus, adds prope; Seneca's Hecuba claims Polyxena as unique, 96063, sola nunc haec est super . . . hac sola vocor/iam voce mater (but this also uses Met. 13.514f.). 451: fortis et infelix et plus quam femina virgo/ducitur. Seneca's parallel account of the sacrifice divides this thought into 1146, animus ... fortis et leto obvius, and 1151, audax virago. 455: utque Neoptolemum stantem ... vidit; a more or less standard feature of such a narrative. Tro. 1148, ut primum ... tetigit atque alte edito/iuvenis paterni vertice in busti stetit may be quite independent. 474-75: populus lacrimas quas illa tenebat/non tenet. This is transferred by Seneca to the death of Astyanax, 1099, non fiet e turba omnium/qui fietur. 35 Compare the weeping of the crowd for Polyxena, without imitation of form, 1161 Uterque fievit coetus. Bonner, Roman Declamation, p. 167 notes the play on active and passive forms of the verb in this passage (as he does in Ovid and other Silver Latin authors), but he misses the derivation of this figure from Ovid's parallel account.

35

31

INTRODUCTION TO THE TEXT

475: ipse etiam flens invitusque sacerdos. Seneca takes this detail (borrowed by Ovid from Euripides) and stresses the paradox, 1153, novumque monstrum est Pyrrhus ad caedem piger. 478: pertulit intrepidos ad fata novissima vultus. Seneca again transfers the formal echo to Astyanax, 1092-93, vultus hue et hue acres tu/it/ intrepidus animo: but Polyxena has her counterpart, 1151-52, non tu/it retro gradum,lconversa ad ictum stat truci vu/tu ferox. He introduces the new motif of her defiant stance. 479-80: Polyxena's fall in Ovid echoes Euripides' stress on decency. Seneca gives the fall a new eo/or, of hostility against the enemy occupying the ground below. 483: teque gemunt virgo, teque 0 modo regia coniunx. Perhaps the division of the line between. apostrophes has influenced Seneca's messenger, 1059, tuosne potius an tuos luctus anus. 485-87: nunc etiam praedae mala sors; quam victor Ulixes/esse suam nollet ... I ... dominum matri vix repperit Hector. Seneca draws on this in two passages: in the prologue (57, praedaque en vilis sequar, and 62, mea sors timetur); and in act 4 at 980, Ithaco obtigisti praeda nolenti brevis, and 985, matrem Hectoris, etc. 495: videoque tuum, mea vulnera, pectus. The figure of mea vulnera is imitated in Andromache's speech at 413-14, cum ferus curru incito/ mea membra raperet. (Not "mine," but "those dear to me.") 499-505: In this passage, partially quoted above, Ovid moves from Achilles' destruction of Polyxena's brothers and herself to Hecuba's hopes that Paris' fatal shot had ended Achilles' power to harm her family, and the irony that all her fertility was for the son of Aeacus. Instead, Seneca opens with the paradox of Achilles' destructive power after death, but he refers to his death only by the slighting allusion to Paris' effectiveness, and condenses Met. 13.503-5 into one clause, cinis ipse nostrum sanguinem ac tumulus sitit. Only then does he develop the idea of Hecuba's fertility, and the brood of children (958-60) before returning to Polyxena as sole comfort, a motif he found twice in Ovid; Tro. 960£. recalls both Met. 13.450 above and 514-,15: quae sola levabas/maternos luctus. 510: tumulis avulsa meorum. Seneca adapts this to Hecuba's contemptuous rejection of Ithaca, ignoring Ovid's allusion to Penelope (511-13). Cf. 991f., sterilis . .. I . .. tellus non capit tumulos meos.

32

SENECA'S MOTIVES AND METHODS

515-16: Inferias hosti peperi. This is imitated by Seneca, not in Troades, but for his Clytemnestra, at Ag. 163: lustrale classi Doricae peperi caput. 517-19: quidve moror ... quo me servas, annosa senectus/quo ... nisi uti nova funera cernam/vivacem differtis anum? Seneca has made multiple use of this motif; c( 42, quid ruinas ... gemis!vivax senectus, and 963, dura et inftlix age!elabere anima, denique hoc unum mihi/remitte fonus, and finally 1169f., ubi hanc ani/is exspuam leti moram? 519-22: quis posse putaret/felicem Priamum post diruta Pergama dici? Felix morte sua est/ ... vitam pariter regnumque reliquit. Cf Tro. 144-46, 157-58, Felix Priamus ... secum excedens sua regna tulit, and 162-64. There are weaker echoes, which may be coincidental, between 523: funeribus dotabere, and Tro. 873-74, ftlici parat dotare thalamo (construction only), 527( superest cur vivere tempus/in breve sustineam, and Tro. 419-21 (where, however, the attitude to life of Andromache is opposed to that of Hecuba); and finally 538-41: obmutuit ilia dolore . . . duroque simillima saxo/torpet, which may have influenced Seneca's portrayal of stunned grief at 417, torpens malis rigensque sine sensu, and 948f. But even where Seneca has no obvious model, he will generate variations; while some extend the meaning with each reprise, and many sweep over the listener like merciless waves, overwhelming with their insistent onslaught, they can impede the progress of argument and distract the listener from the content to the sound and rhythm of the language. Quintilian rightly said of Seneca's prose what Seneca's own father had found to criticize in Ovid. He relished the features in his writing that we reject as faulty. Si parum recte non concupisset, si non omnia sua amasset (10.1.130), makes the same point as Contr. 2.2.12, non ignoravit vitia sua sed amavit. It seems paradoxical that this willful excess, shown in repeated formulation of the same idea, should be allied with drastic compression. But within the single statement Seneca could achieve astonishing economy. This was partly imposed on him by the format of the trimeter, a line varying between % and % of the syllabic content of the hexameters used by his primary Roman models. 36 The density of The hexameter, being dactylic by nature, can contain up to five trisyllabic feet, and therefore seventeen syllables. My personal count of Aen. 2.469-505 yielded an average of 15.5 syllables to the line. In contrast, the trimeter will accept resolved feet, but rarely more than two to a line. A similar syllable count based on Tro. 1-60 (continuous speech) and 318-48 (stichomythic dialogue) yielded an average of 13.3 syllables. There is thus room for an additional disyllabic word or equivalent in each hexameter, in comparison with the trimeter.

36

33

INTRODUCTION TO THE TEXT

thought within the phrase is accentuated by the use of drastic metaphorical verbs, of personified abstractions, and more frequent epithets. To take a last illustration of this strange blend of density and expansion, consider the apostrophe to Calchas at Tro. 353f. He is invoked in terms of his attributes: (1) "he loosed the mooring of the Greek vessels and the delays of war" (morasque belli gives the conceptual equivalent of the material symbol vincla . . . rati); (2) "he can interpret the heavens" is conveyed by the metaphor arte qui reseras polum, and the double clause cui mundi fragor/et stellae ... dant signa fati, singling out the portents of thunder (again conceptualized, as the crashing of a shattered firmament) and comets; he also (3) receives signa fati from viscerum secreta-the Roman ritual of extispicy (secreta are both concealed within the living creature and hidden from any man save the expert, a double implication). However, the last parallel clause, cuius ingenti mihi/mercede constant ora is no attribute, but a pointed allusion to Iphigenia's sacrifice, on the occasion described in 353-54; this is designed to move on the action, through the link of the cost in human life (ingenti mercede) to the new occasion when Calchas will again demand a life: but this time it will be a double sacrifice. This web of allusion, anticipation, and cross-reference is an aspect of Senecan writing that cannot be assessed without some tedium, but the double reference of such dramatic irony adds density of thought to verbal compression, making as severe demands on the listener's erudition and ingenuity as on his emotions. Such tendencies gave Seneca's writing, even when he chose to rework Virgil or Ovid, a new pace and pitch of intensity fitted to the tastes of his generation.

'F:

Three

E MEDIUM OF SENECAN TRAGEDY: STAGE DRAMA, RECITATION,

OR PRIVATE STUDY? When Seneca composed his tragedies, his best-known rival, Pomponius, was writing plays both for recitation and for the public stage. 1 Indeed there is no inherent reason why the same dramatic text could not have been presented with equal success in either medium, if it was composed with the requirements of both types of communica' See Introduction, chap. 1, p. 7.

34

THE MEDIUM OF SENECAN TRAGEDY

tion kept before its author. A weli-contrived script should avoid both ambiguity if recited without being staged, and physical problems entailed by staging what was conceived for oral delivery only. In considering whether or not Seneca intended his tragedies for the stage, the modern scholar has to avoid the twin pitfalls of-at one extreme-measuring his technique against the stagecraft of fifth-century Athens, and-at the other-assuming all the flexible alternative conventions of modern staging to resolve awkward features of physical performance. Herrmann argued at length against more than one German scholar that these plays were composed for the theater, 2 and. his point of view is maintained by Calder, Steidle, and Walker. 3 The case for recitation has recently been given its most thorough advocacy by Zwierlein, 4 who illustrates both the ancient tradition of drama composed for recitation, and the features of Seneca's plays that seem to exclude composition for the stage but would enhance the effectiveness of a script written to be heard, not seen. Yet there is no feature in Zwierlein's indictment for which the advocates of staging cannot quote a parallel in successful stage plays. This discussion will use 'Zwierlein's arguments as a basis, first indicating the problems that would be entailed in presenting the tragedies according to the conventions of the Greek classical theatre. It then attempts to redress the balance and carry the debate one step further, by considering the quite different hazards of presenting these plays in recitation. Any stage director would fmd some difficulty in plotting the moves of Senecan characters, whether from insufficient cuing in the text or more seriously from conflicting stage directions. Thus at Phaedra 1154, Phaedra is reported as on the palace roof, but two lines later she is apparently at stage level, bent over Hippolytus' corpse. As for the corpse, it is treated as on stage in 1158 and 1168, but has to be brought in at Theseus' command in 1247. A producer could only present this sequence by omitting or changing the text. A strictly logistical problem arises in the fifth act of Medea. She has killed one child with the sword at 970 and will climb up to the roof, urging the living child to follow her while she drags the other son's body: tuum quoque ipsa corpus hinc mecum aveham. By 982 she must be on the roof, where she kills the second son at 1019; when the serpent chariot Le Theatre de Seneque (Paris, 1924), pp. 153-232. Cf. W. M. Calder, "Originality in Seneca's Troades," CP 65 (1970): 75-82, and "Seneca's Agamemnon," CP 71 (1976): 27-36; Wolf Steidle, "Zu Senecas Troerinnen," Philol. 94 (1941): 266-84, and Studien zum Antiken Drama (Munich, 1968), pp. 55-63; B. Walker, reviewing Zwierlein, Rezitationsdramen (seen. 4, below) in CP 64 (1969): 183-87. 4 0. Zwierlein, Die Rezitationsdramen Senecas (Meisenheim am Glan, 1966). 2

3

35

INTRODUCTION TO THE TEXT

arrives at roof level, she tosses down the bodies (recipe iam natos parens!) to Jason below. 5 There is no need to protest against the impropriety: ne pueros coram populo Medea trucidet/ aut humana palam coquat exta nefarius Atreus; 6 such rules probably meant little to a generation familiar with the imperial arena. Nor would there be any difficulty in supplying the stage mechanism for her chariot. Even the bloodshed could be easily simulated, just as the entry of the blinded Oedipus or the piecemeal reassemblage of Hippolytus' fragmented corpse could be faked for the stage. But it is unreasonable to expect an actor to haul a child's limp body up a stage staircase while propelling a second actor, or to throw down the living actors from the roof. If this scene had been designed for staging, the dramatist could have left the corpses at ground level; but he would have sacrificed· the tantalizing effect by which Medea slays the second child before his father's eyes, while out of his reach. The effect would have been difficult to stage, but was easily available to the imagination of a listening or reading public. It is true that traditional catastrophes such as the sacrifices of Polyxena and Astyanax are reported as offstage events, although in an unstaged recitation there is no physical obstacle to direct presentation. But Seneca had inherited the narrative form for these pathe and had much to gain by retaining it, since the vehicle of narration allows descriptive interpretation of the motives of the child and the girl which their direct speech or dumb show on stage could not convey so articulately. The real issue was less one of presentation versus reportage than of commentary, on the completed act versus the act in process; the messenger speech can present the pathos in summary before the detailed narrative. Because it is now past and irreversible, the emotions of actors and audience can be sublimated and controlled .. Seneca's text too often requires leaps of time or place. Thus in Tro. 352f. Calchas must be on stage when addressed at 353, although Agamemnon only summoned him in 351-52. As Zwierlein points out, a conventional dramatist could have brought him on stage swiftly enough, if Agamemnon had first announced that he had already summoned Calchas and was awaiting his arrival (Sophocles' device at For the problems in Phaedra, see Zwierlein Rezitationsdramen, p. 15£; in Medea, p. 26. But Zwierlein merely mentions the sequence without analysis. · 6 Horace Ars Poetica 185-86; c£ Herrmann Theatre de Seneque, pp. 165-66. Zwierlein is more conservative, arguing that Seneca himself condemned the spectators of horrors in the arena and would not have represented them; he uses the fact that Seneca describes events which would violate Horace's canon if staged, as a further argument that these episodes were not meant to be staged. 5

36

THE MEDIUM OF SENECAN TRAGEDY

0. T. 293f.). The Medea offers a similar problem: Medea asks her nurse to summon the children at 843, and by 845 she is addressing them and sending them on their way with gifts for Creusa. 7 A stage director could avoid absurdity by occupying time between these lines with silent business; as recitation technique, the economy of time is not only acceptable, but an asset which goes unchallenged. The plays too often show an uncertain sense of place. 8 Sometimes the action is set so that a change of scene is required but not acknowledged; in act 2 of Troades the successive scenes cannot occur in the same place, since the Trojans of the first scene have not heard and must be told of the earth tremors and apparition of Achilles, whereas the Greek leaders in the second scene enter discussing the demands of the apparition. There are other conflicts that I discuss in the separate introductions to acts 2, 3, 4, and 5. In other plays, in:terior scenes such as were displayed by the Greek eccyclema are implied but not consistently reported. But the most common criticism is that the plays fail to mark the entrances and exits of characters: 9 a producer would have to decide' whether they were on stage or not; the text leaves this unresolved. The chorus do not seem to be continuously present, as is the norm in Greek tragedy, but even their absence has to be inferred, because they seem unaware of what has passed, or are not addressed by other speakers. In Troades there are no formal announcements of entrance or exit except at 813 and 1178. Thus in act 4 the news that Polyxena is to be sacrificed comes as a shock to Hecuba and the Trojan women; yet Talthybius had reported Achilles' demand to the chorus at 166-67, and it had been confirmed by Calchas at the end of the second act. It is simple to explain the absence of the chorus from act 2 scene 2 by a change of scene, but a director will have to decide whether they left without waiting to hear Talthybius, or heard him but left without spoken reaction at the end of Calchas, Zwierlein, Rezitationsdramen, p. 29; Medea's childr~n, ibid., p. 125; cf. also ibid., p. 54. a Inadequately treated by Herrmann, Theatre de Seneque, p. 1..76. See Zwierlein, Rezitationsdramen, 38f. Although I differ in some interpretations of individual allusions in the text, I endorse his conclusion; each allusion to setting is strictly transitory, for immediate atmosphere and contradictory data would not be compared but passed unnoticed by the hearers. 9 On uncued entries and exits of the chorus see Zwierlein, Rezitationsdramen, pp. 7276; on the Talthybius scene, ibid. pp. 91-93. But Tarrant, in a definitive study of changing stage conventions, "Senecan Drama and Its Antecedents," HSCP 82 (1978): 213-63, has now shown how many ostensibly deviant features of Senecan stagecraft had already been adopted for the stage by Euripides in later plays like Phoenissae, and can be traced through Hellenistic and Roman comedy running parallel to the lost Roman tragic tradition before Seneca's time. 7

37

INTRODUCTION TO THE TEXT

his report. It will be simplest to present the inferences about changes of scene and undeclared exits in a tabular form. Directions that need not be assumed are listed with a query. Act 1/Prologue Hecuba 1-63 Hecuba, Trojan Women,• 63?/68?-163

AT HEcToR's ToMB No indication that chorus has entered; lamenta cessant suggests rather that they have been moaning softly on stage and now cease. Hecuba exits at 164, without cue?

Act 2 se. 1 Talthybius 164-202 Trojan Women

AT HEcToR's ToMB When does chorus exit? Certainly by 202, but no cue.

sc.2 Pyrrhus, Agamemnon 203f.

IN THE GREEK CAMP Neither character named; identity must be inferred. Summoned at 351-52, appears 353. All leave without cue. Not Trojan Women? Certainly not their attitude to death. Outside dramatic time and space.

Calchas 353f. Chorus 371-408

Act 3 Andromache, Old Man, Astyanax 409f.

Andromache, Astyanax, Ulysses 524f. Trojan Women 814-60 Act 4

Helen, Andromache, Hecuba, (Polyxena), 861f. (Pyrrhus) 999f.?

Trojan Women 1009-55

38

AT HECTOR's TOMB Chorus addressed by entering Andromache. Offstage? When does it exit? Before 435, surely-must be absent from agon with Ulysses. Old man's entry uncued, but he knows she is afraid-therefore has entered with her. Exit uncued. 518? Like chorus, he should not be a witness to the agon. U's entry is clearly announced, as is his exit at 813. Continuous in time and place with act 4? NEAR SIGEUM (931), BATTLEFIELD AND ACHILLES' TOMB (893-95)? No entry cues, except for Pyrrhus, who enters 999 and exits 1003. Do Helen and Polyxena leave with him or at end of scene? (They all reappear together, according to the messenger speech 1133-47). Continuous with act 4. Do Chorus leave at 1055? Can we conceive a common setting for acts 4 and 5? If so, the Chorus can be seen as continuous with both.

THE MEDIUM OF SENECAN TRAGEDY

Act 5 Messenger, Hecuba, Andromache 1056f.

AcHILLEs' ToMB Is Now OFFSTAGE Messenger enters at 1056: women already present. All leave together, cued at 1178-

79.

In all, then, we might assume four different scenes; at Hector's tomb (acts 1, 2 se. 2, and act 3) in the Greek camp (act 2 se. 2) near Sigeum on the battlefield (act 4), and perhaps by the shore (act 5). But absurdities still remain; Andromache, the old man, and the child need to be alone by Hector's tomb in act 3, yet we do not know whether Hecuba and the chorus have left the tomb since 164, or why and where they have gone, only that by act 4 they are in a different place. Conversely, Hecuba and Andromache and the chorus have no reason to move between acts 4 and 5; there is no dramatic point in changing the scene, which would in fact force the chorus-if this were stagedto leave in order to make the change of scene conceivable. We could certainly rationalize Seneca's topography and compromise with a common setting for these acts near Sigeum, but out of sight of Achilles' tomb. But this kind of loophole is futile when so many other discontinuities remain. If Seneca was at all concerned with spatial continuity he could have omitted the allusion to Sigeum in act 4 and to Hector's tomb in the reported narrative of act 5; if he was interested in psychological continuity, he could have shown clearly in his text which characters had left the stage before scenes of which they must subsequently show ignorance. The stage dramatist usually takes pains to protect his work from misinterpretation by the director; b4t if Seneca did not write for production he did not need to confine himself to a single setting, nor to identify the places of each separate scene. Above all, he did not care whether the chorus were present or absent during the dialogue: of the four choruses of Troades the· first requires their dramatic participation and presence at the tomb; the second is so incompatible with their first lyric that it is better explained as spoken, not by the Trojan women, but by voices from Seneca's own world. Yet the Troades are addressed at the beginning of act 3 as if they had been lamenting on the spot-that is, at the tomb--since 164. 10 Apparently absent from act 3, they sing choruses We could assume either that 1-202 and 202-370 were sequences taking place simultaneously, or that when the action moved away in 202 the chorus continued mourning on the spot. The technique of changing the scene to introduce a separate sequence is probably adapted from l:he normal procedure of epic (e.g. Aen. 2.298: diverso interea miscentur moenia luctu; or 9.1f.: atque ea diversa penitus dum parte geruntur ... ). See the persuasive analysis by W. H. Owen, "Time and Event in Seneca's Troades," WS 83 (1970): 118f., who suggests that 1-163, and 409 represent one action sequence,

10

39

INTRODUCTION TO THE TEXT

relevant to their own situation before act 4 and act 5, but in neither show any knowledge of the events taking place in dialogue. Both these choruses reflect a mood of self-absorbed anticipation, quite detached from the tragedy destroying their royal family. A modern producer would be well-advised to use disembodied voices for the second choral ode and physically to isolate his chorus from the actors, perhaps burying them in darkness between 202-409, 524-813, 860-1008, and 1056 following. But such techniques had no parallel in the first century A.D. Zwierlein has also criticized Seneca for his handling of the dialogue in act 4. Here there are three speakers, Helen, Andromache and Hecuba, while the central figure of the crisis is the silent Polyxena. 11 She is addressed both at 871f. and 945f., but it is Andromache who replies for her. Later in the same scene Hecuba reports the entrance (999) of Pyrrhus, and we can infer his exit with Polyxena from Hecuba's taunt abreptam trahe (1003). But he makes no answer and Polyxena utters no cry. Was Seneca concerned to follow the Horatian rule nee quarta loqui persona laboret (Ars Poetica 193)? I doubt it. But would not the violent dumb show have seemed absurd, and even the most dignified silence have become perverse? In the conditions of recitation, we might argue, it would be desirable to limit dialogue to three speaking roles so as to convey clearly the fairly complex action. Actually, Seneca could have left Andromache out of this scene and allowed Polyxena to speak for herself, as in Euripides' Hecuba. He must have chosen the silent role to register Polyxena's aloofness and withdrawal from the fears and anxieties of the living. Although Hecuba will give passionate voice to her feelings in 955f., Andromache is made to offer a kind of commentary on the gestures, first of Polyxena as she welcomes her fate in 945-48, then of Hecuba as she faints to the ground. (Zwierlein has also argued that 965-66 is a similar description by Hecuba of Polyxena's grief on behalf of her mother.) There would be no difficulty in staging this, and external description may be substituted for the sufferer's words, not to avoid a fourth speaking part, but to dignify with silence emotions too deep for expression. 12 This could be effective as a technique in stage perand 164-202, with 371-408 a second, while 202-370 is a simultaneous third. I would differ, arguing for only two sequences, with a change of place at 202, and excluding 371f. from the dramatic action as the poet's outside comment, not spoken by the Trojan women. 11 On the related problem of silent players and the three-actor rule, see Herrmann, Theatre de Seneque, p. 172f. Zwierlein sees the value of limiting the dialogue to three speakers as easing the task of the reciter and the comprehension of the audience. See also Zwierlein, Rezitationsdramen, pp. 572-62, on "Stummes Spiel." 12 Aeschylus' use of a silent protagonist was famous; cf. Aristophanes Frogs 911f., the

40

THE MEDIUM OF SENECAN TRAGEDY

formance, but in recitation it would be essential. Seneca's decision to leave Hecuba silent at the news is the dramatic equivalent of Timanthes' procedure in his famous painting of the sacrifice of Iphigenia. Cicero and Quintilian are our sources for this, which was clearly a classic exemplum used in rhetorical instruction. Timanthes showed the grief of all the participants except Agamemnon, the most bereaved: consuniptis affectibus non reperiens quo digne modo patris vultum posset exprimere, velavit eius caput et suo cuique animo dedit aestimandum (Quint. 2.13.13; cf. Cic. Orator 74). There are many instances in Senecan tragedy where the actions of characters deemed to be present are reported in this way. We can compare with the four-line commentary of Hecuba on Pyrrhus' intrusion at 999, Hippolytus' description of his own treatment ofPhaedra at Phae. 705[., or Medea's account of her murderous acts at Med. 971f., 1020f. There is also a different type of description, portraying the mood and expression of the protagonist, as in the nurse's account of Medea's condition at Med. 383f., and the chorus's report of Cassandra's prophetic seizure at Ag. 710f. In stage drama, both tragedy and comedy, it is usual to employ such descriptions to build up characterization before the entrance of a major role; we might compare with Phae. 360f. or Med. 382f. the conversation between the doctor and the waiting woman preceding the appearance of Lady Macbeth in the sleepwalking scene. But at Phae. 705f. it would be quite farcical if Phaedra were to embrace Hippolytus as he continued to describe his actions (how could he find time to breathe?) or he himself were to wrench back her head before our eyes as he declared: en impudicum crine contorto caput/laeva reflexi (707-9). In a context of reading or recitation these lines would be the sole cue to what is taking place. 13 Such descriptive passages are naturally a kind of aside, a form awkward to handle in stage dialogue where actors must be seen together. Senecan characters are particularly prone to asides, because they so often need to address themselves. The divided personality operates its own private dialogue. The third act of Troades is loaded with descriptive and hortatory asides, starting with the words of Ulysses at 607: quid agis, Ulixe? 14 His next ten lines are spoken aside, covering successively arguments and counterarguments for and against first criticism by Euripides. Compare a discussion of Cordelia's silence in the opening scene of King Lear, by J. Levenson, "What the Silence Said, Still Points in King Lear," in Shakespeare 1971, ed. Clifford Leach and]. Margeson [Toronto, 1972], pp. 215-29. 13 Zwierlein, Rezitationsdran:~en, pp. 58-62. But Owen (WS 83 [1970]: 136-37) sees the described action of this sequence as the most cogent argument that Seneca wrote the play for staging. 14 See Herrmann Theatre de Seneque, pp. 179-80; Zwierlein, Rezitationsdramen, pp. 6364, and Tarrant, HSCP 82 (1978): 242-46.

41

INTRODUCTION TO THE TEXT

believing Andromache-her loyalty as a parent, the omen implied by the tomb, her oath to him (607-12)-then his self-exhortation, nunc advoca astus anime (613), and three lines describing her conflicting behavior as she combines tears of mourning with anxious pacing. When he finally turns to her (619), her reaction is not a reply but an aside describing her paralysis of fear (623-24). Faced with the destruction of the tomb containing the living child she expresses her anguish in a self-address of twenty lines, debating whether to save her child or her husband's grave. It is in fact a false dilemma, since there is no possibility of saving her son. But even if her inner debate were supremely logical, it should not strain dramatic credibility. What is the actor playing Ulysses to do while she debates? Indeed the very demonstrative pronouns of her soliloquy would trip up the stage performer. How is the actor to gesticulate (or avoid gesture) pronouncing hinc . .. illinc (645), hie . .. ilium (655-56), illinc (658), hie (660)? For the child is in the same place as the tomb. Any producer would be baffied by such a scenario, and yet there are scholars who respect Seneca's writing and still believe he intended this to be staged. There remain the more generally undramatic features noted by Zwierlein in considering the structure of the Senecan plays. In discussing the uncertain relationship between chorus and action, I have already anticipated some of the issues to be raised here. Thus the second chorus, denying the afterlife, takes us completely out of the world of the play, going beyond an alienation effect in its denial of supernatural intervention such as we have heard described in act 2 and will meet immediately after the chorus in act 3. The very different behavior of the Trojan women implied by Andromache's entrance speech (409) merely confirms that Seneca intends no continuity with the preceding chorus. Andromache in act 3 knows nothing of the events of act 2 scene 2, 15 and Hecuba and Helen in act 4 show no knowledge of the seizure of Astyanax in act 3. As Andromache does not inform them directly, we might fairly say that the leading roles, the collective Trojan protagonists, are not allowed to grow in knowledge with events, in the way that Hecuba's knowledge of her family's sorrow grows progressively in Euripides' Troades. It may seem strange, by the standards of Greek tragedy, that Hecuba offers no reaction to the news of Astyanax's death before or after the narrative in act 5 (even missing the opportunity for such acknowledgment at 1060) until her epigrammatic concidit virgo ac puerlbellum peractum est Zwierlein, Rezitationsdramen, p. 88f. But as he notes (p. 91) v. 805, /entus et segnis iaces? /red it Achilles, shows that act 3 assumes Trojan knowledge of Achilles' apparition, though this knowledge is implicitly denied in act 4.

15

42

THE MEDIUM OF SENECAN TRAGEDY

(1167-68). But this is to be expected if Seneca is more concerned to present the climax of events (as in epic) than to portray consistent and complete human responses. Astyanax is Andromache's loss, so once she has reacted with the appropriate emotion in act 3, Hecuba need say nothing. The subject can be left aside in order to focus the audience totally on the new event-the threat to Polyxena. Thus in each act it is as if the previous act had not occurred. These. scenes (for in Seneca scene and act are usually coextensive)- exist in their own right and could almost be reordered without inconsistency. If act 3 followed act 4, only the irony of Andromache's words in 928f. would be lost. In the same way there are disproportions within the scenes, which spring from Seneca's training in development of a rhetorical theme, and perhaps from imitation of the shifting pace of Virgilian and Ovidian epic. Thus the forty-line diatribes of Pyrrhus and Agamemnon in act 2 expand to meet the rhetorical potential of the issue, while the brief ten lines at the end of the scene must include Calchas' decree and the resolution of the quarrel. In fact this, the most decisive moment of the action, falls into a silent void, leaving no trace of where or how Agamemnon has abandoned his principles. On the other hand Troades does not have long descriptive reports like that of Theseus' journey through Hades (H.F. 642f.) or the magic rites of Medea, first reported by the nurse at Med. 670f. then enacted from 740 to 848, or Manto's commentary for her blind father on the outcome of the sacrifice at Oed. 314f. The scene setting of the doublemessenger narrative is economical enough; Seneca shrewdly reduces the description of place and crowd from the eighteen lines of 106886 for the death of Astyanax, to eleven lines for the setting of Polyxena's sacrifice. In these narratives his use of simile is brief, s~rik­ ing, and opportune. 16 Astyanax is portrayed as a young lion in 109396, Polyxena as the setting sun in 1140-42. Elsewhere, in his use of similes in dialogue Seneca shows a closer affinity to epic than to drama: compare Ulysses' sequence of metaphors and similes to represent the threat posed to the Greeks by Astyanax's future growth (936-45), or Andromache's revival of this theme, interrupting her last words to her son at 794 with an extended comparison of enemy, mother, and child with lion, cow, and calf. Euripides had allowed 16 See Zwierlein, Rezitationsdramen, p. 119. In general contrast Euripides' discreet use of simile in messenger speeches, discussed by S. Barlow, The Imagery of Euripides (London, 1971), 197, p. 73f. She illustrates Seneca's undramatic excesses by contrasting the portrait of the seamonster in Phaedra 1035[. with that ofEur. Hipp. 1198f. (pp. 70-73).

43

INTRODUCTION TO THE TEXT

himself only the briefest and most evocative of images, "like a fledgling thrusting yourselfbeneath my wings" (Tro. 750-51). Seneca shows poetic artistry in binding Astyanax's doom to the first portrayal of his future through a renewed image, which also has the architectural function of framing the episode through a repetition of the motif at beginning and end, rather like the verbal echoes that create the ring composition of an independent Horatian lyric. But the device brings dramatic action to a stop, and the symmetries of ring composition interfere with the forward movement of traditional drama. Yet if the taking of Astyanax is recited and unseen, the imaginative power of the simile reinforces visualization, substituting for the acted event. Monologue is far more dominant in Seneca than in Greek tragedy, but at times it is difficult to distinguish whether a speech is monologue or address. Thus Hecuba's first sixty-two lines apostrophizing personified Troy (4), or the dead Priam (29), Hector (31), or herself (42) should be seen as monologue: she only turns to the chorus at 63, turba captivae mea. Talthybius' speech is ostensibly dialogue, since it is written as a response to the chorus's questioning of 165-66, but he loses contact with his audience. Ulysses' entrance is more clearly an address to Andromache (cf. credas in 526), but Helen's entrance speech is certainly monologue from 861 to 871. 17 At the end of this act Hecuba may be addressing Andromache or Helen, but in view of her apostrophe to the absent Ulysses (993) we can only see her speech as a self-oriented monologue. Indeed the same pattern has occurred in her speech from 955 on, which slides from rhetorical questions into self-address as 963, and it will arise in her final speech of the play, which opens with an ironic exhortation to the Greeks (embodied in the nuntius), then turns to rhetorical questions, and finally into an apostrophe to death (1171). Whoever is conceived as still present, Hecuba is really alone and thinking, not communicating. This offends less than Senecan monologue in other dramas because the role of Hecuba in Euripides too is self-absorbed, isolated by age, loss, and pain. But in stage presentation, the characters left on stage would find themselves unable to react appropriately. In other words, once again, this type of monologue in company is undramatic and succeeds only so long as it does not have to be dramatized. 17 Herrmann, TheJtre de Seni!que, pp. 179-80, sees Helen's monologue as necessitated by the need to occupy the time of her crossing the stage to make contact with the Trojan women, who are grouped in an introverted tableau center-stage. Tarrant, HSCP 82 (1978): 232-41, offers evidence of a long tradition of entrance monologues ignored by characters already on stage, in Euripides, Menander, and Plautine adaptations of Greek comedy. Such "suspension of dramatic time," was clearly a regular feature of the ancient theatre.

44

THE MEDIUM OF SENECAN TRAGEDY

It is a pity that Zwierlein's scrupulous, realistic assessment of the play's lack of dramatic continuity and unsuitability for the stage has been misunderstood in recent criticism. Thus Steidle mistakes his comment on lack of coherence in the action, for a denial of thematic unity. 18 Steidle himself (in an earlier paper) and Willy Schetter had done valuable work in analyzing Seneca's modification of the argument to provide large-scale symmetry between the Astyanax and Polyxena actions; 19 whereas his most recent discussion in rebuttal of Zwierlein's thesis relies too much on hypothesizing offstage actions, he again offers positive intefpre~ations that bring out the merit of Seneca's composition. The actions of Troades are artistically balanced to build a climax both of violence by the victor and emotional response in the audience. Polyxena's fate is raised and battled over first; then the tricks and countertrickery of act 3 carry the doom of Astyanax to a comparable point; but in this extra-long action, virtually a self-contained drama at the center of our play, violence first rises to a crescendo (at 671-84) then subsides to a closing cadence at 81314. The leading away of Astyanax is balanced by the attempted deceit of Helen and the greater br;tality of Pyrrhus' seizure of Polyxena; then in the final narrative, Astyanax, the less-developed hero figure, is first to die, while she who was first to be endangered dies last and in a more violent context. Here is certainly a symmetrical structure achieved, to the best of our knowledge, by Seneca's independent organization of formerly separate actions. Again the constant reiteration of the almost reciprocal fortune of Troy and Greeece creates an equilibrium in which future Greek suffering is expected to restore the balance of retributive justice; we see the Greeks in their triumph occupy the position held by the Trojans earlier in this cycle, and figures of authority on both sides, Hecuba and Agamemnon, share the refrain of warning that leads beyond the action of our play to a Greek tragedy-the pathos of Agamemnon himself and the collective suffering of the storm that forms the centerpiece of Seneca's own play on this theme. 20 As a poet Seneca has skillfully provided just this mesh of reference Studien zum Antiken Drama (Munich, 1968), pp. 56-62; Steidle's seventy-line note 85 criticizing Zwierlein is answered by Zwierlein in an extended review, GGA 222: 196227; see especially 197-202 on Troades and n. 4, p. 199, dealing with Steidle's arguments about continuity of place. 19 "Sulla Struttura delle Troiane di Seneca," RivFC 93 (1965), 396-429 ( = Senecas Tragodien, Wege der Forschung 310, ed. Le!evre, pp. 230-72, "Zum Aufbau van Senecas Troerinnen"). 20 I have argued elsewhere that the Agamemnon was written after Troades and in deliberate continuation of its themes. 18

45

INTRODUCTION TO THE TEXT

backward and forward that binds the narrative, presenting it as part of a continuing moralized history; he has insisted on saturating the events of these few hours with the consciousness of prior Greek and Trojan actions, shown as moral successes or failures that have activated fortune in a chain of causality. Without affirming the justice of providence, Seneca has nonetheless imposed the framework of superhuman ordinance to give shape and unity to his composition. Unlike the advocates of composition for the stage, those who believe Seneca composed his plays as recitation dramas are not usually challenged to illustrate how this dramatic poetry would have been performed; yet several of the features I have indicated as obstacles to a good stage production would have caused equal difficulty when converted into recitation. Critics have complained at the lack of formal entrances and the erratic identification of speakers in the plays; thus in Troades 164f. Talthybius, whose name we can read in the scene heading, is not named either by himself or the answering chorus; at 202f. Pyrrhus is identified only by his subject matter, until his use of the word pater in 232 shows that this advocate of Achilles must be his own son; his addressee in turn goes unnamed and is merged in a collective plural at 236 vos diruistis. The roles could be at least partly identified by stage costuming; how would they be conveyed in recitation? It seems that the act would have required an introduction identifying the speakers of both scenes; once this introduction is provided, it can also supply any other circumstantial information, specifying changes of place or lapses of time. In act 4 Helen introduces herself and addresses Polyxena. Who replies? How does the listener know it is Andromache, not Polyxena (or indeed Hecuba)? Andromache's speech does not identify her (888902) and she is first named at 925, where Helen's use of the vocative would more easily be understood as referring forward to the new addressee than back to the previous speaker. At 945 the descriptions of Polyxena and Hecuba exclude them as speakers, while 954, spirat, revixit serves as a cue for Hecuba. But throughout this triangular scene in which Hecuba and Andromache voice similar sentiments, the audience of the recitation would depend on cues in the dialogue; the absence of adequate cues cannot in this case be compensated by any preliminary announcement of the participants, the scene, or even the order of their speaking. As Mrs. Walker notes, Zwierlein and other advocates of the recitation play have not considered the exact nature of such a recital.2 1 2t

CP 64 (1969): 185.

46

THE MEDIUM OF SENECAN TRAGEDY

We should perhaps ask this kind of question, even though there is too little ancient evidence for the procedure of a Roman author at a recitatio. Neither Tacitus nor Quintilian mentions addition~! reciters besides the poet himself, and two passages in Pliny suggest that reading by the poet alone was the norm. In Bp. 9.34 Pliny is hesitant and apologetic about employing a skilled freedman to read his poetry and debates whether he should mime and gesture as the poems are recited; in Bp. 7.17 he asks why men criticize his reciting of speeches yet agree that tragedies should be recited, quae non auditorium sed scaenam et actores [se. poscunt]: cur lyrica quae non lectorem sed chorum et lyram poscunt? At horum recitatio usu iam recepta est. 22 If recitation of tragedy was given only by the poet, he would have no trouble in presenting long speeches and other relatively undramatic excerpts. The analogy of Maternus in the Dialogus of Tacitus suggests that such a one-man presentation would be a natural preliminary to publication of the text. In Dial. 3.3 he tells us that he has planned his Thyestes and is looking forward to a recitation of it; indeed he is hastening publication of his recently recited Cato in order to dismiss his concern with it and devote himself wholeheartedly to the new play. It has usually been assumed that such recitations were given by the poet alone, and that the dialogue of a tragedy could be handled by a single speaker. 23 No doubt he could effectively present long speeches, messenger narratives, and even a simple, not too swift-changing confrontation of two roles; but as Herington pointed out, it becomes absurd, if not impossible, when one performer must alternate in stichomythia;24 Tro. 318-48 would lose much of its effect if it had to be presented by a single reader. How much could be eased by the performer's expertise? The declamations excerpted by Seneca's father "Which do not require a concert hall but a. stage and actors; why is lyric poetry recited, when it needs not a reader but a chorus and lyre accompaniment? Yet the reading of these genres is now customary." Pliny is our main source for various aspects of recitation, For the recitation of poetry see Ep.: 3. 7, Silius Italicus reciting epic; 5.27, Sentius Augurinus reciting occasional poems; 5.3, Pliny himself, reciting mildly improper verse (hendecasyllables?); 7.17 (quoted above); 6.21, his own book et opusculis variis et metris; ·and 9.34. In Ep. 7.4 Pliny reports having written elegy varia metra, heroa (i.e., hexameter epic?) and most recently hendecasyllables, but none of these genres would require more than one speaker, and only the hexameters need have been long compositions. 23 We might perhaps treat the wording of Juvenal 1.4-6 as evidence permitting a different assumption. After protesting at poets' reading their nondramatic poetry, he resorts to an impersonal expression to describe tragic recitations; impune diem consumpserit ingens Telephus, aut summi plena iam margine liberlscriptum et in tergo necdum finitus Orestes. The language leaves open the possibility of recitation by more than one speaker. 24 In "Senecan Tragedy," Arion 5 (1966): 422-71 (p. 445 argues this point quoting Medea 170-71), reprinted in Essays in Classical Literature Selected from "Arion" (Cambridge, 1972), pp. 169-220. 22

47

INTRODUCTION TO THE TEXT

feature dialogue exchanges of two- or three-word phrases (often unspaced by inquit or aiebat) and pass constantly from the speaker's persona to quotation of the absent litigants: cf. Contr. 2.5 (the wife tortured under a tyrant) or 9.4 (the man struck by his son), especially 9.4.14: quid si non cecidero, inquit, quid factums es? Torquebis? Occides? Plus est quod imperas quam quod minaris ... caede, inquit: non caedo. verbera; non ferio. We may also recall readings on radio in which the skilled reader can take on all the roles in the dialogue of a novel without adapting the written text. Yet it is unlikely that Seneca intended the two notorious lines from Medea (170-71): moriere :: cupio :: profuge :: paenituit fugae. :: Medea :: fiam :: mater es :: cui sim vides

to be delivered by one performer. And a triangular dialogue, such as that of Troades act 4, could not be correctly attributed to its speakers by the listener without identifying vocatives, which are in fact lacking from the scene. The obvious alternative is a sort of concert reading with roles consistently maintained by distinct performers. Most of Seneca's plays contain at least one scene that would require three rather than two performers: but even the equivalent of the full tragic cast of three actors will not solve certain difficulties. When the Talthybius speaker reports the apparition of Achilles, does the representative of the chorus walk away, or stand listening until the end? When the news of Polyxena's fate is brought by Helen, will the speakers gesticulate into space at the words vide ut animus ingens laetus audierit necem? (945(); any approach toward dramatization makes the absence of Polyxena's dumb show, or of the savage Pyrrhus striding in to snatch his victim at 999-1005, an embarrassment. Is it possible to come to any conclusion? As works composed for the stage these dramas would contain unnecessary difficulties for actors and producers alike; in recitation, whether solo or by several readers, there would be places where the identity of the speaker would be left unclear. There is, it seems to me, only one medium in which the action of Senecan tragedy comes through with complete clarity, and that is the written text. I would suggest then, that Seneca composed with the expectation that he would himself recite chosen passages, or would give a dramatized reading in cooperation with others: but ultimately the play would be known through written copies, and only the readers would experience the plays as complete works. In this he would be subjecting his work to the same forms of publication as Virgil or Ovid had intended for their epic poems, and as

48

THE MEDIUM OF SENECAN TRAGEDY

Maternus proposes for his ideological tragedy in the generation after Seneca's death. These are not well-crafted stage plays, and their merit lies in poetic vitality rather than theatrical effectiveness. I would like to suggest as a postscript that this may explain why Seneca's powerful characterizations and thrilling portrayal of anger, revenge, and self-destruction deeply influenced the themes and tone of Renaissance drama in Italy, England, and France, but led to few actual imitations of the plays themselves. It is perhaps significant that Troades, though the first of his tragedies to be translated into English verse, has a relatively scant posterity in European drama; its French and Spanish and German adaptations have not survived in the dramatic repertoire of their countries. 25 This century, with the increasing flexibility of cinema (which thrives on calculated discontinuity) could produce an effective visual performance of our play; on the other hand, radi(') provides the ideal medium, 26 letting no trite effects of staging or photography compete with the impact that Seneca sought from the passionate and often shocking language of his mythical actions. As a prose writer he was used to achieving his effect of instruction or persuasion through words alone, and I believe he was content to let the words serve as the unaided medium of his dramatic poetry. See Les Tragedies de Seneque et le Theatre de la Renaissance, Centre Nationale de la Recherche Scientifique, ed. Lebeque/Jacquot (Paris, 1964). For Troades they list the Heywood translation (first of the Tenne Tragedies, 1581) in England; in France, Garnier's La Troade (Paris, 1579), and other versions, by Sallebrey, 1639, and Pradon, 1679; in Spain, Gonzalez de Salas, Las Troianas, 1633; in Germany, Jacob Duym's Spieghel der Hochmoets, Wesende Troiados, 1600, and Opitz's Die Troianerinnen, 1625; in the Netherlands there was Vondel's translation of 1626. Boas, University Drama in the Tudor Age (Oxford, 1914), quotes records of production of Troades, Oedipus, Medea, and Hecuba at Trinity, Cambridge, between 1559 and 1561, but these may be neoLatin tragedies, not the plays of Seneca, who certainly wrote no Hecuba. 26 See now the vivid and appreciative study by W.M. S. Russell in Papers of the Radio Literature Conference, 1977, ed. Peter Lewis (Durham, 1978), 1: 1-26. The second half of the paper, from p. 13, discusses the characteristics of the plays as sound drama (with special stress on the Hercules Furens), and argues that it was these characteristics which contributed most to the progress of Renaissance tragedy (p.24). "We owe it to Seneca's sound drama that the Elizabethans were not overwhelmed by a Greek model, and used a looser one, which they had to modify for stage performance, but which gave them essentially a new framework. Hence it seems the function of sound drama is experiment and the dissolution of old forms "to make way for new ones. Nowhere is this clearer than in the history and influence of the sound drama of L. Annaeus Seneca."

25

49

0;ERSIS: LI0~ I

THE GROWTH

OF THE TRADITION

I. THE EPIC TRADITION

This attempt to trace the evolution of the legends associated with the fall of Troy will initially report the different versions of each myth as they arise, without any special concern for Seneca's sources. For although it is valuable to know the range of accounts available to an author in order to appreciate his selection of detail, the existence of an early account is no guarantee of its influence. Similarity between an older and a newer version of the same myth is neither a necessary condition for localized imitation of detail in language or characterization, nor a sufficient condition to imply adaptation on a larger scale. The two plays of Euripides, Hecuba, and Troades, which contain the first surviving full-scale treatments of the deaths of Priam, Polyxena, and Astyanax, may be taken as presenting the myths in their mature form, evolved from the conflicting variants of post-Homeric epic, the continuing pictorial tradition, and the early lyric poets, especially Stesichorus. The cyclic epics that took up the tale of Troy after the death of Hector, drawing on brief allusions in the Iliad and Odyssey and adapting their stories to local hero cults and communal feuds, have survived only in the abridgments ofProclus and of Apollodorus' Bibliotheke (both influenced by later tragic reworkings) and in odd fragments and testimonia from scholiasts and antiquarians. Zielinski1 has shown how the post-Homeric poets used for their continuation of Andromache' s tale the fears expressed by Hector in his leave-taking from her in Il. 6; he foresees the day "when perhaps one of the bronze-clad Achaeans will take you away weeping, cutting off your life of freedom, and in Argos you will weave at another woman's loom, or perhaps carry water from the springs of Messeis or Hypereia suffering much ill-treatment, and dire necessity will weigh upon you. " 2 They also knew the forebodings of Andromache herself Zielinski, "De Andromacha Posthomerica," Eos 31 (1928):1-39. On the Cyclic poets see G. L. Huxley, Greek Epic Poetry (London, 1969) chap. 11 (including a translation of Proclus), and J. Griffin "The epic cycle and the uniqueness of Homer," JRS 57 (1977):39-53. 2 !1. 6.454-58: 1

ocrcrov creu, o-re xtv n~ 'Axm&v xa)..xoxmovwv liaxgu6wcrav UYiltaL, t)..Eu9egov ~lillQ (mouga~. xa[ 'XEV EV "AQyEL eoucra 1tQO~ UAAT]~ Lcrtov ucj>a[voL~,

50

ILIOU PERSIS

after Hector's death, "as for you, my child, either you will go with me, where you may do loathsome chores, toiling for a ruthless lord, or some Achaean will take you by the hand, and hurl you from the battlements-a dreadful death-in his anger because Hector killed his brother or father or son. " 3 From this kind of material Arctinus of Miletus built up in the late eighth or early seventh century his two books Iliou PersiJ on the sack ofTroy, starting from the device of the Wooden Horse, while Lesches ofPyrrha composed an Ilias Mikra, in four books-at least these were the ascriptions commonly if not universally accepted in our Greek sources. But by the fourth century the name Little Iliad seems to have been used comprehensively for the Ilias Mikra and Iliou Persis, and most probably for the five books of the Aethiopis too; these contained the events from the death of Hector to that of Achille~. This collective title inevitably causes some confusion in later reports of the contents of the separate epics, and hampers our understanding of the relationship between, say, the four-book Ilias Mikra and the Iliou Persis. Further, although we 'can trust Proclus' positive statements of the contents of these narratives, chance evidence for myths alludes to material omitted from his account, suggesting that he has less value as proof of what was not contained in the poems. His account of the Iliou Persis ends, "then having set fire to the city they sacrificed Polyxena at the tomb of Achilles. And when Odysseus had disposed of Astyanax, Neoptolemus took Andromache as his prize. And the rest of the spoils were distributed. Demophon and Athamas found Aethra and took her away with them. Then the Greeks sailed away and Athena brought destruction upon them by sea. " 4 Thus Arctinus seems to have placed the sacrifice of Polyxena after the firing of the city and subordinated the murder of Astyanax to Neoptolemus' claim to Andromache, mentioning no obstacle to the Greek departure. Proclus rep'orts nothing from Lesches' xa( xev MroQ «j>oQEOLS MwO"r)l6os ii 'YnEQELTJS n6f.J.' l:texa~Of.tEvTJ, XQateQ"I) 6' bnxe(oet' l:tvayxT]. 3 It. 24.732-37: cro 6' ab texos ij t110t autfi E\jiEat, £v6a xev ~Qya anx£a EQYU~OLO, ciei..Eurov 1tQO avaxtOS clf.tEtALJ(.OU, ij ttS 'A)(.atWv QL\jiEL J(.ELQOS EAWV uno 1tUQYOU AUYQOV OAE6QOV, x.ro6f.1evos,

ttJtE"tEO~ JtO"ta!loi:o/'Lm:a"t' aEtQO!lEvov); (c) 18889, his triumphal chariot-ride before the walls of Troy, dragging Hector's body, which represents the city's doom (cf. Il. 22.395f.)this association, made in 31 and 124-29, is now an identification: Hectorem et Troiam trahens, 189. But the first allusion, to the Threicia arma of 183, is obscure. The event is set before the Trojan campaigns or even the encounter with Cycnus-a mere preliminary joust (for prolusio, proludere, used of warming-up exercises in the arena, and hence in oratory, see Cic. de

236

COMMENTARY: ACT TWO

Orat. 2.325; in Seneca, H.F. 222: et tumida tenera guttura elidens manu/prolusit hydrae; Med. 907; Phae. 1061; N.Q. 3.28.3; and Ep.102.23). Kingery, Three Plays of Seneca (repr. Norman, Okla., 1966), p. 221, refers this to "Cisseus the father ofHecuba, who came to the relief of Troy and was defeated by Achilles," but I can find no evidence for this in ancient sources. A more appropriate episode would be the defeat and killing of Tenes, ruler of Tenedos and son of Apollo, on Achilles' journey from Aulis to the Troad (Apollodorus Epit. 3.2327, Plutarch Quaestiones Graecae 28). Since Thetis had warned Achilles that if he slew Tenes he would incur his own death at the hand of Apollo, the event is significant for Achilles' fate. His conquest of Tenedos (cf. I/. 11.624) is listed at Met. 12.109 in the Cycnus sequence on which Seneca has drawn for the contents of this passage and of 224f. The only eccentricity is the use of Threicius for the man of Tenedos, but the island can be seen, like its neighbor Lemnos, as part of the Thracian region. See also 215f. for the continuation of Achilles' deeds of valor, and on Seneca's errors or licenses of geography compare commentary 8-13. 191. ite, ite, inertes. Ovid's Achilles had opened his ghostly challenge with a question, Met. 13.443, inmemoresque mei disceditis ... Achivi? ... (445) ne facite! Seneca adapts the ironical command, usually to do what is harmful, wrong, or impossible, which is a popular figure of Augustan rhetoric. The phrase may have an ancestor in the derisive address of Helen to Paris at Il. 3.432-33, af...f...' Hlt vuv 1tQox6.A.wom . . . MeveA.aov/e~ailttc; ~-taxeoaoem evavt(ov, but is greatly extended in Latin; cf. Hor. Epist. 2.2. 76, Aen. 4.381, 7.425, 9.634; Prop. 2.29.22, 3.18.17; Ovid Her. 3.26, 4.127, 9.105, 12.204, etc. The awkward synizesis of meis and the disordered reading of the A family led Gronovius to conjecture that Seneca's text had been manibus iustos meis, but a gloss debitos had replaced iustos in the early, undivided tradition; hence the ancestor of E had rearranged the phrase to ease the metrical problem, whereas the A family had perpetuated the unmetrical phrasing. He may be right, for synizesis is rare in Senecan tragedy, and his conjecture improves the distribution of noun and epithet. There is no other instance of synizesis in Troades, but Leo, Ohs. 90, compares saxeo in Thy. 233, fatale sax@ pascuum. See also commentaries on 45 and 341 about problems of elision, in 341, entailing a synizesis before elision. 193-94. non parvo luitliras Achillis Graecia. As at 164, Seneca brings out the element of recurrence in the mythical narrative. The whole Iliad can be seen as the consequence of Achilles' anger, the J-tfivtc; of

237

COMMENTARY: ACT TWO

Homer's opening line, and with magno luet his ghost now threatens a second 'IAL. The phrase regum tyranne is a deliberate distortion of the Homeric ava; avbQ&v. We may guess from Cic. Att. 14.17a.2 and Livy 45.27.9 that he was usually called rex regum in Latin; this title is used by Seneca at Ag. 39, rex ille regum . ... ductor ... ducum, and by Helen at Tro. 978, regum . . . maximus rector. But Pyrrhus disputes his legitimacy as overlord, just as Achilles had blamed him for wanting absolute power at Jl. 1.287-88, JtUV't