Tragedy and Redress in Western Literature: A Philosophical Perspective 9781138498082, 9781351017039

Table of contents :
Cover
Title
Copyright
Dedication
Contents
Acknowledgements
Introduction
Part I Tragedy and Moral Redress
1 Oedipus: Hamartia, Freedom, and the Supernatural
2 Antigone’s Holy Crime
3 From Cognitive Failure to No-Fault Tragedy
Part II Tragedy and Linguistic Redress
4 Theoretical Considerations
5 Can Suffering Be Expressed in Words?
6 Tragedy and Linguistic Idealism
Bibliography
Index

Citation preview

Tragedy and Redress in Western Literature

This book offers a unique interpretation of tragic literature in the Western tradition, deploying the method and style of analytic philosophy. Richard Gaskin argues that tragic literature seeks to offer moral and linguistic redress (compensation) for suffering. Moral redress involves the balancing of a protagonist’s suffering with guilt (and vice versa): Gaskin contends that, to a much greater extent than has been recognized by recent critics, traditional tragedy represents suffering as incurred by avoidable and culpable mistakes of a cognitive nature. Moral redress operates in the first instance at the level of the individual agent. Linguistic redress, by contrast, operates at a higher level of generality, namely at the level of the community: its fundamental motor is the sheer expressibility of suffering in words. Against many writers on tragedy, Gaskin argues that language is competent to express pain and suffering, and that tragic literature has that expression as one of its principal purposes. The definition of tragic literature in this book is expanded to include more than stage drama: the treatment stretches from the Classical and Medieval periods through to the early twentieth century. There is a special focus on Sophocles, but Gaskin takes account of most other major tragic authors in the European tradition, including Homer, Aeschylus, Euripides, Virgil, Seneca, Chaucer, Marlowe, Shakespeare, Corneille, Racine, Lessing, Goethe, Schiller, Kleist, Büchner, Ibsen, Hardy, Kafka, and Mann; lesser-known areas, such as Renaissance neo-Latin tragedy, are also covered. Among theorists of tragedy, Gaskin concentrates on Aristotle and Bradley; but the contributions of numerous contemporary commentators are also assessed. Tragedy and Redress in Western Literature: A Philosophical Perspective offers a new and genuinely interdisciplinary perspective on tragedy that will be of considerable interest both to philosophers of literature and to literary critics. Richard Gaskin is Professor of Philosophy at the University of Liverpool, UK. He is the author of seven books, including Horace and Housman (2013), Language, Truth, and Literature: A Defence of Literary Humanism (2013), and The Unity of the Proposition (2008).

Routledge Research in Aesthetics

Michael Fried and Philosophy Modernism, Intention, and Theatricality Edited by Mathew Abbott The Aesthetics of Videogames Edited by Jon Robson and Grant Tavinor Tragedy and Redress in Western Literature A Philosophical Perspective Richard Gaskin

Tragedy and Redress in Western Literature A Philosophical Perspective Richard Gaskin

First published 2018 by Routledge 711 Third Avenue, New York, NY 10017 and by Routledge 2 Park Square, Milton Park, Abingdon, Oxon OX14 4RN Routledge is an imprint of the Taylor & Francis Group, an informa business © 2018 Taylor & Francis The right of Richard Gaskin to be identified as author of this work has been asserted by him in accordance with sections 77 and 78 of the Copyright, Designs and Patents Act 1988. All rights reserved. No part of this book may be reprinted or reproduced or utilised in any form or by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying and recording, or in any information storage or retrieval system, without permission in writing from the publishers. Trademark notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identification and explanation without intent to infringe. Library of Congress Cataloging-in-Publication Data A catalog record for this book has been requested ISBN: 978-1-138-49808-2 (hbk) ISBN: 978-1-351-01703-9 (ebk) Typeset in Sabon by Apex CoVantage, LLC

To the memory of Christopher Gray 1964–96 et tamen adhuc per fidem, nondum per speciem: spe enim salvi facti sumus. spes autem quae videtur non est spes. adhuc abyssus abyssum invocat, sed iam in voce cataractarum tuarum. adhuc et ille qui dicit, ‘non potui vobis loqui quasi spiritalibus, sed quasi carnalibus’, etiam ipse nondum se arbitratur comprehendisse, et quae retro oblitus, in ea quae ante sunt extenditur et ingemescit gravatus, et sitit anima eius ad deum vivum, quemadmodum cervi ad fontes aquarum, et dicit, ‘quando veniam?’, habitaculum suum, quod de caelo est, superindui cupiens, et vocat inferiorem abyssum dicens, ‘nolite conformari huic saeculo, sed reformamini in novitate mentis vestrae’. Augustine, Confessions XIII, 13.

Contents

Acknowledgementsviii Introduction

1

PART I

Tragedy and Moral Redress19 1 Oedipus: Hamartia, Freedom, and the Supernatural

21

2 Antigone’s Holy Crime

84

3 From Cognitive Failure to No-Fault Tragedy

131

PART II

Tragedy and Linguistic Redress187 4 Theoretical Considerations

189

5 Can Suffering Be Expressed in Words?

266

6 Tragedy and Linguistic Idealism

322

Bibliography358 Index404

Acknowledgements

I am most grateful to several anonymous readers for their useful criticisms of earlier drafts of part or all of this book, and especially to Malcolm Heath for his detailed and incisive comments on a late draft of the whole thing. All these commentators found much to contest in my argument; I am of course responsible for the final product, and for any errors, factual or argumentational, that it may contain. In a few places I have relayed, and responded to, some of these readers’ objections, since I imagine that others will share their worries. As I explain in the Introduction, the book is analytic and polemical in method and style. It challenges a number of settled views in the criticism of literature in general, and in the reception of tragedy in particular. Whether I have made out a satisfactory case will be for the reader to judge. I could not have undertaken and completed this project without the support and encouragement of my wife Cathrin and my sons Thomas and Markus, and to them I express my deep gratitude, as I do also to my sisters Rosemary, Hilary, and Fiona, to my brothers-in-law Chris, Raymond, and Christopher, and to my friends and colleagues Barry Dainton, Daniel Hill, and David Langslow: all of these helped me during a testing bout of illness that befell me suddenly and unexpectedly during the period of writing this book, and their support was invaluable in seeing me through that difficult time. Barry and Daniel also commented on part of an early draft of the book, as did Chris Bartley and Stephen McLeod. I am further indebted to George Gereby and the members of an audience at Budapest to whom a syncopated version of the first chapter was delivered as a talk. I dedicate the book to the memory of my friend Christopher Gray, with whom I had many discussions on topics philosophical and literary during the time of our acquaintance, which, alas, proved all too brief. We did not discuss the passage of Augustine that I quote in the dedication, but I am confident that he would have loved it; and he would have read it with the eye of faith, unlike me, who must secularize it to make it render

Acknowledgements ix up its riches. But, to anticipate a point I shall make at more length in the book, that secularizing move not only can be made, but has been available to be made since the earliest beginnings of Western literature. Richard Gaskin January 2018

Introduction

The subject matter of this book is tragic literature in the Western tradition, approached from a philosophical point of view and with a particular argumentative project in view. I shall come to the perspective and the project shortly, but first I should say something about the subject matter, because right there we face a difficult definitional problem. How are we to demarcate the subject area? What is to count as tragedy? ‘Tragedy’ is a highly contested term. Vindice in The Revenger’s Tragedy, now generally assigned to Thomas Middleton, tells us that ‘When the bad bleeds, then is the tragedy good’ (III, 5, 202), but that was not Aristotle’s conception of tragedy. Different again is the definition offered by Chaucer’s Monk, for whom tragedy tells ‘Of hym that stood in greet prosperitee, / And is yfallen out of heigh degree / Into myserie, and endeth wrecchedly’ (1975–8). One way of responding to the definitional problem would be simply to circumvent it by following Henry Kelly’s example, who in his valuable studies of medieval tragedy proceeds by ‘accepting everything that is called a tragedy as a tragedy’, and refusing to count anything as a tragedy that is not explicitly presented as such; as examples of works falling in this latter category Kelly mentions Beowulf and the Song of Roland (1997, 2, 7). In the spirit of the later Wittgenstein’s remarks on games, we might then repudiate the search for a common thread running through all works that have been or are called tragedies.1 After all, since the label has been widely and variously applied, perhaps there is no common thread running through the tragic tradition, but only, at best, a network of overlapping strands. Wittgenstein’s suggestion that there may be nothing interesting common to games actually seems to me rather implausible,2 but one might think that a related suggestion held of all the things that have been called tragedies. However, though Kelly’s descriptive and a posteriori approach may be the right strategy for historians to adopt, philosophers usually wish to add some normative and a priori elements to their investigation. For if we substitute for the democratic a more discriminatory strategy—so if we are willing to exclude some works that have been called tragedies from the category and include other works that have not been so designated—and if we apply a structure that

2  Introduction separates the resultant field into core cases and outliers, we may then be able to find certain interesting commonalities among the works that are identified as belonging to the core. And in fact there is very general, if not universal, agreement on a subset of works that, by almost anyone’s standards, must count as tragedies. Everyone would agree that Sophocles’ Oedipus Rex and Antigone, two works on which I shall particularly focus in Part I, belong to this group. One needs here to beware the risk of trivialization, that is, of adopting a definition which conveniently excludes from the extension of the term ‘tragedy’ just those works that would, if admitted, undermine whatever thesis it is that one wishes to advance about ‘genuine’ tragedies. On the other hand, circularity cannot entirely be avoided. The a posteriori elements of the definition require one to admit as tragedies what most people, or (better) most competent judges, would regard as such, and the appeal here to competence already introduces an element of circularity into the procedure.3 Nevertheless, the notion of competence does, as Hume insisted (1987, 233–4), contain some purely objective elements, and the requirement that the definition answer to a consensus does place some external constraints on what is allowed to pass muster, so that if, in following this procedure, we can identify interesting features of (what are agreed to be) central cases of tragic literature, features that are not simply spelt out in the definition, we may count this as intellectual progress. The core–outlier axis mentioned above envisages the existence not only of central and peripheral cases of tragedy, but also, by implication, of borderline cases. The existence of this latter category has often been held to make the task of definition impossible, but that is not so. All empirical concepts are vague at the edges, without detriment to our ability to make clear conceptual and definitional distinctions. Alastair Fowler quotes a passage of Lord Kames with approval, in which the vagueness of boundaries between literary genres is compared to that between colours, and it is suggested that this shows that it is futile to try to define genres.4 In fact the comparison points in the opposite direction: the distinction between red and orange, say, and the possibility of giving definitions of them, is in no way undermined by the fact that one colour shades into the other, and that in an intermediate region, itself vaguely delimited, it is simply indeterminate whether a given sample is red or orange. We can still say that the top traffic light is determinately red, not orange, and the middle light determinately orange, not red. The existence of borderline cases indeed precisely presupposes the existence of cases that are not borderline. I have spoken of commonalities attaching to central cases of tragedy, and I implied no temporal or geographical or cultural restrictions in the application of these ideas; it is, after all, the business of science—and the study of literature is a science in the broad sense—to seek generality. There is, however, a critical tradition that refuses (or pretends to refuse)

Introduction 3 to generalize in that way. For example, Hugh Grady tells us (2014, 791) that, in his book Modern Tragedy, Raymond Williams shows decisively and in succinct detail how the still commonly held idea that tragedy is a term whose meaning has remained stable over more than 2,500 years—indeed, that it is a major example for the notion of a continuous Western civilization—cannot withstand even a cursory examination of the genealogy of the word and its radically different underlying conceptions from ancient Athens to European medieval culture to several eras of modernity. It is true that tragedies and conceptions of tragedy vary across times and cultures, but Grady’s ‘radically different’ goes much too far. He himself in effect concedes this when, a few pages later, he remarks that the idea of tragedy can be used ‘to encompass a whole range of plays, from the Greeks to the current theatre, that differ greatly from one another but form a kind of genealogical, generational continuity. They are plays that take on issues of death, of suffering, of identity, of human nature, of human meaning, and more’ (ibid., 796). So it appears that there is after all a stable core to the notion of tragedy, and that this core does indeed support ‘the notion of a continuous Western civilization’. There are in fact severe theoretical limits on any attempt to find that other people do things, or think things, in a way that is radically different from our way; in the present case, if the term ‘tragedy’ did not enjoy stability of meaning across eras and cultures, we should not be able to recognize the term in the speech and practice of other ages and communities. Hence the Williams–Grady relativism undermines itself in a familiar way: if ancient conceptions of (say) tragedy were radically different from ours, we would not call them—and indeed they would not be—conceptions of tragedy. I shall return to and explore this important point in Part II of this book. In the passage from Grady just quoted, we are given an outline definition of tragedy. In Chapter 6 of the Poetics, Aristotle offered a definition of tragedy which (as usual) he arrived at by adopting a mixed a posteriori and a priori strategy. According to this definition, ‘a tragedy is a mimēsis of a high, complete action (“complete” in the sense that implies amplitude), in speech pleasurably enhanced, the different kinds [of enhancement] occurring in separate sections, in dramatic, not narrative form, effecting through pity and fear the catharsis of such emotions’ (1449b24–8);5 in Chapter 13 he added that tragedy should concern itself with a protagonist who is ‘not outstanding in virtue and justice’, and who ‘falls into misfortune, not through vice and wickedness but because of a hamartia, and who is one of those in high reputation and fortune such as Oedipus and Thyestes and the foremost men of such clans’ (1453a7–12). In these quotations the three transliterated Greek words—‘mimēsis’, ‘catharsis’, and ‘hamartia’—mean, roughly, imitation, purification, and

4  Introduction mistake, respectively. But their precise meanings, and their role in Aristotle’s theory of tragedy, are controversial. For reasons of space, I shall reserve treatment of the ideas of imitation and catharsis for another occasion, but I shall have much to say about hamartia, which will figure centrally in my argument. Etymologically, ‘tragedy’ refers to a certain genre of literature. But the word is used, and indeed was already used in antiquity and in the Middle Ages,6 for extra-literary suffering—for real-life tragedy, as we say. A philosophical study of literary tragedy cannot avoid talking about real-life tragedy as well as the genre of literature that we call ‘tragedy’, because that is what that genre is about. Most tragic literature is fictional, but in this case, too, it is about the sort of thing that can and does happen in the extra-literary world. In order to avoid confusion, where I use the simple word ‘tragedy’, and its congeners, I shall mean the literary genre; where real-life tragedy is in question, I will make that clear with appropriate modifiers, such as the one I am using here. Aristotle arrived at his account of the essence of tragedy by reflecting on the literature with which he was familiar, but he intended it as a general definition, and it has often been applied generally. I shall indeed be arguing that the doctrine of hamartia has general application to the tradition of European tragedy from Homer onwards. But, however things may stand with the case of hamartia, it seems clear that other parts of Aristotle’s formulation no longer enjoy general validity. For example, the idea that tragedy must befall a person who, in Chaucer’s words, ‘stood in greet prosperitee’, though accepted in antiquity and largely followed till the onset of modernity,7 is one that has fallen away: for us Büchner’s Woyzeck and Hauptmann’s Rose Bernd and Miller’s Death of a Salesman count as tragedies in as good a sense as Aeschylus’ Agamemnon and Sophocles’ Ajax and Euripides’ Bacchae. Tragedy, like many other things, has been brought down to earth and democratized, a process that can be nicely observed in the leap from Schiller’s Wallenstein to Brecht’s Mutter Courage und ihre Kinder. In the first part of Schiller’s trilogy, Wallensteins Lager, the playwright designs us to see his hero through the eyes of ordinary soldiers and camp followers, and major themes of the trilogy are anticipated rather in the manner of the first scene of Shakespeare’s Julius Caesar.8 Brecht saw the democratic potential of the cut-price crowd to which Schiller had given a voice, and a hardly conspicuous Marketenderin was transformed into one of the most powerful figures of the modern stage, the sutler Mother Courage.9 Indeed the process of democratization had already begun in the Renaissance with dramas such as Arden of Faversham, A Warning for Fair Women, Middleton’s Women Beware Women, and Heywood’s A Woman Killed with Kindness.10 (There is even an isolated medieval example: in John of Garland’s Parisiana Poetria we find a tragedy which had no predecessors and no immediate successors about the passion and treachery of a washerwoman.)11 These are essentially bourgeois tragedies, a genre that was picked up and developed in

Introduction 5 the eighteenth century, when sympathy and fellow-feeling came to be emphasized as universal human emotions,12 in dramas such as Lillo’s The London Merchant and Lessing’s Miss Sara Sampson and Emilia Galotti,13 and in novels like Richardson’s Clarissa.14 It has even been suggested that Racine dramatized middle-class attitudes.15 And, though the reaction was delayed, we can now see that Büchner’s Woyzeck and Hauptmann’s Die Weber ushered in an era of working-class tragedy.16 With the advent of the novel, Aristotle’s insistence on dramatic rather than narrative form became obsolete, too. Indeed it was obsolete already, for the Iliad, though not a tragedy in the strict sense of the word (‘tragōidia’) according to fifth- and fourth-century BCE usage, is nothing if not tragic.17 In what sense is the Iliad tragic? The poem contains dialogue, but it would be odd to maintain that that is what makes it tragic. It is scarcely less odd to find some readers denying that the Iliad is tragic:18 what makes it so emerges already in the preoccupation of the opening lines with, as Colin Macleod puts it, ‘human passion and blindness, which lead to suffering, death and loss of burial’ (1982, 7). It is possible to use the word ‘tragedy’ in an exclusively narrow sense, confined to stage drama, and ‘tragic’ in a wider sense, encompassing other genres, and indeed an ancient commentator does just this when, commenting on the first line of the Iliad, he remarks that the poet had ‘created a tragic proem for several tragedies’.19 (By contrast, Aristotle thought that the Iliad provided matter for only one or two tragedies, Horace that it could be turned into a single tragedy by division into five acts.)20 But for my purposes that way with the terminology would be awkward. I shall follow a more relaxed policy of allowing both words to attach to a particular kind of subject matter, rather than to a particular genre defined purely formally. This looser usage is standard and has a good pedigree. Dante has his Virgil call the Aeneid ‘my high tragedy’ (‘l’alta mia tragedìa’, Inferno XX, 113): at least, one might say, it encompasses the tragedies of Dido and Turnus. The Iliad is, as the subtitle of James Redfield’s book on the poem has it, the tragedy of Hector.21 Indeed Plato had already called Homer a master of tragedy and the first of the tragedians,22 and Aristotle himself says that the Iliad, and even the Odyssey, are analogous to tragedy strictly so called;23 he also observes that a mere narration of the story of Oedipus would itself produce the distinctively tragic effects of pity and fear.24 It is hardly going too far to say that Aristotle in the Poetics, and perhaps Horace in the Ars Poetica, regarded tragic drama and epic as constituting, respectively, actable and non-actable species of a single genus, namely tragedy in the broad sense, the sense in which I shall use the word.25 Probably the thorniest question that arises in connection with the task of defining tragedy concerns the ending: Scaliger said that tragedies have ‘exitus horribiles’, Samuel Johnson that they have ‘a calamitous conclusion’, and George Steiner that they ‘end badly’.26 Aristotle famously

6  Introduction changed his mind on this question, first (in chapter 13 of the Poetics) adopting a standard of the ideal in tragedy which privileged plays that, like Sophocles’ Oedipus Rex, end in disaster, but then (in chapter 14) preferring those that, like Euripides’ Cresphontes (reconstructed by Matthew Arnold in his Merope) and Iphigeneia in Tauris, seem to be heading for disaster but manage to deflect it in time, so that all ends happily.27 And if Aristotle vacillated on the question, Wilamowitz rejected the requirement of a bad ending outright.28 The epistle to Cangrande, traditionally ascribed to Dante, defined tragedy as moving from an opening that is ‘admirabilis et quieta’ to a conclusion that is ‘foetida et horribilis’, whereas comedy, the author suggests, traces an opposite path;29 this way of distinguishing the two genres, which goes back to the grammarians Diomedes and Donatus, has been influential in the tradition.30 Of course, not all tragedies begin well: some, such as Seneca’s, start badly and get worse.31 But perhaps the traditional definition is right to insist on a disastrous outcome? More recently, Antony Quinton has offered, by way of an ‘acceptable and intelligible residue of Aristotle’s formula’, the statement that ‘a tragedy should be the representation of a single and rationally connected series of events that involve misfortune and suffering and end in disaster’ (1960, 156). One thing that must be said immediately about Quinton’s suggested definition is that, while it may give us a necessary condition, it does not give us a sufficient condition for tragedy, since it might plausibly be argued that many comedies satisfy its conditions. Conflict, which has also been advanced as a defining feature of tragedy,32 is likewise insufficient (perhaps not even necessary), being also a feature of many comedies. And the same might be said of Grady’s characterization, quoted above, that tragedies ‘take on issues of death, of suffering, of identity, of human nature, of human meaning, and more’: the greatest comedies explore just these issues. Now one difference between the sufferings of bona fide tragic figures and the disasters that befall comic characters lies in the attitude that the audience is encouraged to take to the suffering portrayed.33 We are invited by Sophocles to watch or read his Oedipus Rex in a mood that is suffused with a high seriousness; but flip the story of the club-footed selfexposing criminal around and you find yourself ploitering through the exquisite genial surrealism of Kleist’s Der zerbrochne Krug. Lessing produced an even more vestigial version of Aristotle’s definition: a tragedy, he writes, is simply a poem that arouses pity (‘ein Gedicht . . ., welches Mitleid erreget’).34 That no doubt goes too far in the minimalist direction: it would force us to count Catullus’ elegy for Lesbia’s sparrow and Herrick’s imitation thereof as tragedies, and agree with Dante that his lyric ‘Donne ch’ avete intelletto d’amore’ is a tragedy.35 It seems to me that, for taxonomic purposes, a good case can be made for accepting Scaliger’s and Johnson’s and Steiner’s and Quinton’s insistence, based on Chapter 13 of the Poetics, that tragedies must end badly,

Introduction 7 at least for protagonists, rather than Aristotle’s second thoughts, in Chapter 14, that allows them—or even, ideally, requires them—to end well. It makes sense, in the light of the whole tradition, to distinguish tragedy not merely from comedy but also from tragicomedy,36 and one common understanding of this genre has been that it is like tragedy but with a happy ending, a conception that perhaps achieved its historical apogee in Corneille’s Le Cid.37 Putting together the points that have arisen so far, and accepting Quinton’s statement as a first move towards a definition of tragedy, I should say that what needs to be added to it are three (partly overlapping) conditions: we should require of a tragedy that it evince (i) an overall seriousness of tone and purpose in its depiction of suffering, that, as A. F. Garvie puts it, ‘tragedy should deal with suffering and take it seriously’;38 (ii) largeness of scale, or what Jacopo della Lana called ‘architectonic magnificences’;39 and (iii) what Dante called gravitas sententiae.40 Comedy and tragicomedy may then share some but not all of these features, to varying degrees. For example, comedies often end badly from someone’s point of view, as Scaliger pointed out;41 but we are not invited to adopt the perspective of the sufferer—Euripides’ Polyphemus, as it might be (for whom, as the early-modern commentator Robortello noted, the blinding was not terribly amusing),42 or Plautus’ miles gloriosus, or Jonson’s Volpone, or Kleist’s Judge Adam.43 We are not encouraged to treat the suffering of these figures with high seriousness. Contrariwise, as Strindberg pointed out in his preface to Miss Julie (1976, 92), even a tragedy that ends with as many deaths as Hamlet will have a silver lining for someone—the gravedigger, perhaps, or Fortinbras—but, again, that is not a perspective that we are asked to inhabit: the key point, as I noted above, is that Hamlet ends badly for the protagonist. Verna Foster offers a more sophisticated definition of tragicomedy than the simple one given above: according to this definition, ‘A tragicomedy is a play in which the tragic and the comic both exist but are formally and emotionally dependent on one another, each modifying and determining the nature of the other so as to produce a mixed, tragicomic response in the audience’ (2004, 11). As it stands, this definition does not sufficiently distinguish tragicomedy from tragedy that includes comic elements (for example the Bacchae and Macbeth): I shall look at this point in Chapter 3. But if we read Foster’s definition as concerning the effect of tragicomedies taken as wholes, then perhaps it works: applying it, we would say that tragicomedy lacks (i) from the definition of tragedy, since the overall mood of a tragicomedy would not be tragic. The proposed definition (Quinton plus the three extra conditions) is broad, in accordance with my terminological policy on ‘tragedy’ and ‘tragic’: for instance, it admits some epic poetry and some novels, allowing me to count Homer and Thomas Hardy as tragedians. But the definition does not let in just anything that has had the label ‘tragedy’ attached to it over the centuries. In particular, the suggested definition rules out as bona

8  Introduction fide tragedies not only plays of the Iphigeneia in Tauris stamp but also so-called Christian (or, in general, religious) drama, such as Everyman, in which a ‘fundamental logic of radical consolation undoes the tragic force of Everyman’s dying’,44 or The Castle of Perseverance, with its ‘merciful denouement’,45 or Muret’s Julius Caesar, in which Caesar undergoes apotheosis at the end and the chorus delivers a highly Christian conclusion, or Calderón’s El Mágico Prodigioso, which closes with the salvation of Cipriano and Justina; likewise martyr plays such as Buchanan’s Baptistes or Calderón’s El Príncipe Constante, in which the hero suffers in worldly terms, but is considered to earn an eternal reward.46 Many commentators from Dacier onwards have rejected the claims of religious drama to count as tragedy;47 Lessing condemned ‘martyr tragedy’, focusing his attack, as Dacier had done, on Corneille’s Polyeucte (which was subtitled ‘tragédie chrétienne’).48 If, following the above definition, we insist that tragedies must end with a genuine catastrophe—not a catastrophe that we are invited to consider as being really or ultimately a piece of good fortune—the standard rejection of the claims of Christian tragedy to count as bona fide tragedy is plausible. (This does not rule out Christian tragedy in the sense in which Doctor Faustus and Macbeth might be considered to be such, where the hero loses his soul.)49 For my purposes in this book, it will not usually be necessary to insist on the relatively precise definition of tragedy that I have just given; mostly I shall be relying on our intuitive understanding of what counts as a tragedy, and not raising definitional worries. But it is nevertheless important to my project that the definition is neutral on the twin topics of moral and linguistic redress, the subject matter of this book. As for the first of these, the definition has nothing to say about two interrelated matters, namely the questions of fault and desert, which will together compose the major concern of the first part of this book: there I shall argue that tragic heroes and heroines are typically at fault, usually in a cognitive sense, that their faults are instrumental in bringing about disaster, and that, accordingly, these agents are considered to deserve their fate. The great Shakespearean critic A. C. Bradley wrote that ‘No mere suffering of misfortune, no suffering that does not spring in great part from human agency, and in some degree from the agency of the sufferer, is tragic, however pitiful or dreadful it may be’ (1959, 81). With the qualification that we must restrict Bradley’s claim (as he probably in any case meant to do) to protagonists, I shall be agreeing with this thesis and defending it. Since the provisional definition of tragedy that I have given is, as I say, neutral on this matter, no questions should be begged: there should be no risk of circularity. But in addition, lest the reader feel that the definition of tragedy I have accepted is mistaken, I shall focus in this part of the book on works in the Western tradition which not only satisfy that definition, but are also ones that must in any case be counted as tragedies. For I take it as uncontroversial that, on any good definition of tragedy, Sophocles’

Introduction 9 Oedipus Rex and Antigone will pass the test; and it is on these two tragedies that, in arguing the case for Bradley’s thesis (in Part I), I shall mainly focus, though I shall have things to say about other works too, such as Lessing’s Miss Sara Sampson and Emilia Galotti which, again, must count as tragedies on any showing. The two Sophoclean dramas I have mentioned are nowadays standardly taken (so far as the figures of Oedipus and Antigone go) to be no-fault tragedies. If I succeed in overturning this prevalent view, I believe I shall establish something significant. But I shall also suggest, with varying detail of argumentative support, that the Bradleian thesis applies quite generally to Western tragedy—if not, perhaps, to all works that satisfy the definition I have given or that we would intuitively recognize as tragedies (it would be difficult to establish such a sweeping claim), at any rate to a much greater proportion of them than many recent critics have supposed. So it seems to me that the definition of tragedy I have accepted begs no questions as far as the first part of this book is concerned. And the same applies to my argument, in the second part of the book, that tragedy offers us a form of linguistic redress. I should warn the reader that I shall not be providing anything approaching a systematic survey either of tragedies themselves or of theories of tragedy: there are many books that perform these tasks, and readers who want a conspectus of tragedies from their ancient beginnings to modern times, or who wish to find out in detail what Hegel, Schopenhauer, Kierkegaard, Nietzsche, Bradley, and other theorists have said about tragedy, will find no difficulty locating material to instruct them. (My bibliography contains many relevant sources.) Of course I shall, where appropriate, mention and discuss both particular tragedies and classic theories of tragedy. But my main purpose in this book is not to provide a descriptive account either of tragedies or of theories of tragedy; rather, as I have said, it is to argue a case, to present my own (partial) theory of tragedy. In general I shall be contending that tragic literature offers, or purports to offer, to a much greater extent than has been appreciated, redress or compensation (I shall use these two terms interchangeably) for suffering. The sorts of redress that I shall be concerned with are two: moral and linguistic. In Part I of the book I shall, as intimated in my last paragraph, argue that tragic literature treats suffering as incurred by culpable agent error (in particular by cognitive failure), and so as deserved, to a much greater extent than has been recognized by critics, especially in recent decades. Note here that, though we normally think of redress or compensation as operative in one temporal direction only—so that the compensating thing or event comes after what is compensated— the sheer idea of balancing one thing with another is temporally neutral, and it will be important to my argument that redress or compensation can precede what is redressed or compensated. (In such cases we would ordinarily speak of compensation in advance, but I shall treat the ideas of redress and compensation as being, in themselves, temporally neutral.)

10  Introduction In Part II I shall argue that, even in cases where suffering is undeserved— and tragic literature does indeed present such cases—it is subject to a kind of linguistic redress, in the sense that its sheer expressibility brings it under our control, places it for us emotionally and intellectually, and in a certain measure celebrates it, offering us some consolation in what Bertrand Russell once memorably called our ‘long march through the night’ (1986, 18). By way of arguing for this latter thesis, it will be an important component of my project to show that pain, loss, and suffering are indeed expressible in language; for this has often been denied. The second part of the book is more theoretical than the first, and in Chapter 4 I introduce some of the considerations that are going to be relevant in this part, as well as elucidating and defending a number of assumptions that have been operative in the first part. The suggestion that tragedy offers a kind of redress for suffering is not in itself new. Most familiarly, tragedy has sometimes been regarded as performing a compensatory role for the community by presenting a tragic victim that goes proxy for the rest of that community. This idea is found in tragedy itself, particular in Euripides, some of whose ­characters—­Macaria (as she is named by the tradition) in Children of Heracles, Polyxena in Hecuba, Menoeceus in Phoenician Women, Iphigeneia in Iphigeneia at Aulis—offer themselves willingly to appease the gods or the dead.50 The motif is also found in later tragedy, being exemplified for example by Iphis in Buchanan’s Jephtha.51 Some critics, following Nietzsche’s lead in Die Geburt der Tragödie (§21),52 have generalized the approach, so that every tragic victim becomes a pharmakos, a scapegoat, or perhaps—like Buchanan’s Iphis, Corneille’s Oedipe, and Hauptmann’s Rose Bernd—a Christ-figure,53 who suffers and dies for the community. (It has been suggested that Iphis is ‘the first female type of Christ’.)54 The community hopes, by concentrating the wrath of the gods or of God onto the figure of the scapegoat, to avert it from everyone else; or the individual offers him- or herself to atone for ancestral sins. Interpretations along these lines have been explored by Gilbert Murray, René Girard, and numerous other commentators.55 Now the pharmakos notion certainly surfaces in some tragedies, including the ones just mentioned, but others too: when Euripides’ Dionysus says mockingly to his Pentheus ‘You alone undergo this trial on behalf of the city’ (Bacchae 963), the idea that Pentheus is a scapegoat seems indeed to be present;56 in the Aeneid Turnus’ death bears some resemblance to a formal devotio on the community’s behalf;57 Cicéron in Garnier’s Cornélie offers himself to atone for the sins of his compatriots (1–12);58 as does the Dauphin in Schiller’s Die Jungfrau von Orleans to expiate the transgressions of his ancestors (1024–31); there are also hints of the pharmakos idea and the related notion of carnival king or rex stultorum in Shakespeare’s Julius Caesar.59 But it seems to me that the importance of the scapegoat theme, and the extent of its occurrence in European tragedy, can easily

Introduction 11 be exaggerated, and that much of the reasoning adduced in favour of a general application of the idea to tragedy is uncogent. Girard’s suggestion (2010, 119–20) that Sophocles’ Oedipus may not have killed Laius, that the play is a search for a scapegoat, and that guilt is handed from one agent to the next in a frantic parody of pass the parcel, is quite implausible. We must take the play’s identification of Oedipus as Laius’ slayer at face value: Oedipus Rex is not The Usual Suspects. Moreover, even if it is correct that Greek tragedy had its origins in religious ritual involving sacrifice (a derivation which is consistent with, but by no means conclusively established by, the evidence),60 it does not follow that tragedy continued to be a ritual—still less that it continued to be a ritual involving a scapegoat—by the time of its blossoming in the Athens of the fifth century BCE, not to mention its later flourishings in Europe. To say that would be to succumb ‘to the fallacy inherent in Nietzsche’s genealogy: that the origins of a phenomenon fully explain its essence’.61 In general I am sceptical of the fruitfulness of regarding tragic performances as rites or religious rituals, and I shall not be further concerned with tragic redress in the scapegoat sense, which I mention here to set aside. My interest focuses rather on the idea of compensation in the different senses I have indicated, namely moral and linguistic. Moral compensation involves the balancing of a protagonist’s suffering with guilt (and vice versa): this compensation operates at the level of the individual who sufferers. Linguistic compensation, by contrast, operates at a higher level of generality, namely at the level of all language users. Many critics reject the idea that tragedy offers any kind of redress or compensation: indeed that is undoubtedly the orthodox view. So, for example, Sebastian Gardner says that ‘loss and suffering cannot be acknowledged from the moral point of view to have the kind of unconditional and hence uncompensatable reality which they are represented as having in tragedy’ (2003, 235). Bernard Williams writes of Sophocles’ Women of Trachis that ‘All the force of the play is directed to leaving in the starkest relief its extreme, undeserved, and uncompensated suffering’ (2006, 58). David Scott Kastan argues that ‘Tragedy for Shakespeare is the literary genre in which suffering is not only irreparable but is also neither compensated nor even effectively consoled’ (2003, 10); George Steiner that ‘In the norm of tragedy, there can be no compensation’, that ‘where there is compensation, there is justice, not tragedy’ (1961, 129, 4); and Rowan Williams that tragedy resists the ‘language of debt’, of ‘compensation’ and ‘recompense’ (2016, 118). The context of Kastan’s assertion makes clear that he is thinking of compensation for suffering that runs temporally from earlier to later (as is Rowan Williams), and he is mainly concerned to distinguish Shakespeare’s approach to suffering from that of Christian ‘tragedies’ such as Everyman, where prior suffering is compensated by posterior salvation. I think he is right that Shakespeare offers no compensation in that sense, and in that temporal

12  Introduction direction: the flights of angels that Horatio hopes will sing Hamlet to his rest do not turn Hamlet into a Renaissance Everyman. But that leaves the question of prior compensation—and, in particular, the compensation of suffering by desert—untouched. And what about other forms of compensation, such as linguistic? Steiner makes the claims I have quoted in the context of what seems to me a rather doubtful distinction between, on the one hand, the supposed concern of the Judaic tradition with justice, not tragedy, and, on the other hand, the supposed concern of the Greek tradition with tragedy, not justice.62 This dichotomy overlooks the possibility that an upshot may be both just and tragic. I shall be concerned in this study to rebut the idea that tragedy does not offer compensation for suffering, and I shall reject the common view that justice and tragedy exclude one another. Gardner, Steiner, Kastan, Bernard Williams, Rowan Williams—these are some of my interlocutors, and there will be many others. For the book is constructed around arguments with real or hypothetical opponents. Here I follow Lessing, who in his Hamburgische Dramaturgie (§70) wrote of the theorist that ‘he should first seek out someone with whom he may argue: that way he will gradually become familiar with the material, and the rest will look after itself’ (1996, vol. 4, 559). Solet Aristoteles, he adds, quaerere pugnam in suis libris—Aristotle likes to pick fights in his books. And Lessing himself followed Aristotle’s lead, for example in his Hamburgische Dramaturgie, where he spent a lot of time arguing against Voltaire.63 This is a work in the analytic philosophy and criticism of literature. My aim is to bring the methods of analytical philosophy to bear in the reading and understanding of a particular genre of creative literature, namely tragedy. Though analytic philosophy has for some time now regarded creative literature as a legitimate area in which to apply its characteristic techniques, I think it is probably true to say that an analytic approach specifically to the genre of tragedy, on the scale attempted in this study, is a rather new venture. What is perhaps the nearest predecessor to this book, Steiner’s The Death of Tragedy, is executed along quite different lines. This book aims to be genuinely interdisciplinary, between two methodologies that are still not found together as often as they should be. Fulfilling that aim also involves adopting, at least to a large extent, the style of analytical philosophy, which is essentially argumentative. Charles Martindale complains that analytic philosophers ‘have, in the main, ceased all constructive engagement with the rest of the humanities’ (2007, 36). I agree that analytic philosophers should engage with the other Humanities, and this book is intended as an effort in this direction. But the reverse side of that coin is that the other Humanities should engage with analytic philosophy: when they do that they may be led to some surprising results. For decades now it has been an article of faith among scholars of the Humanities that literature, and the genre of tragedy par excellence, demonstrate (to put it in broad terms) the limits of

Introduction 13 rationality, the inadequacy of language, the ineffability of suffering and emotion, the elusiveness of the self, and the incomprehensibility of the world. These consoling mantras are endlessly intoned in the echo chamber of contemporary literary-critical and -theoretical discourse. But the reason why critics and theorists believe in such limits and inadequacies and incomprehensibilities is not that they have investigated the matter and found these barriers to exist; rather, they hold themselves exempt from having to carry out any such enquiry because they have decided in advance that they do exist. A reader of a draft of this book complained that I failed simply to presuppose what everyone agrees to be the truth, namely that tragedy offers ‘a vision that resists full articulation’, and that ‘is elusive to the discursive intellect’. Of course I do not presuppose these claims: they are false, and I inveigh against them. I think it is time for the Humanities to subject some of their favourite clichés to rather careful examination, and analytic philosophy can help with that. The remark about argumentational style in the last paragraph connects importantly with a point that, in concluding this introduction, I need to make about my discussions of real-life tragedy. In my treatment of reallife tragedy I shall draw on Shoah or Holocaust literature, for two main reasons. One is the fact that the Holocaust is extremely well documented, and has given rise to important literary responses. The second reason is the unique status that it has in modern Western consciousness.64 But here a caution is in place. At the end of his superb book The Better Angels of Our Nature, in which he traces the trajectory, and remarkable decline, of violence in evolutionary and historical time, Steven Pinker comments on the fact that he has throughout adopted—as I shall be adopting here—an ‘analytic’ tone: he has done this, he says, because the subject of violence ‘has inspired too much piety and not enough understanding’ (2011, 696). He also concedes that his tone has been ‘at times irreverent’, though there I think that he perhaps does himself a mild injustice: certainly I detect no lack of sensitivity in his excellent treatment of the terrible catalogue of violence in the twentieth century. At any rate, I should like to avail myself of the same defence in respect of my adopting an analytic style in discussing real-life as well as literary tragedy. Naturally there is a risk of insensitivity in this area. One recalls the controversy occasioned by the Carmelite convent and cross at Auschwitz: just as someone might object ‘Where was Christianity when the Holocaust happened?’ so also you might raise the questions ‘Where was the analytic tradition? Where was clarity and rigour of thought?’. In my case the risk is exacerbated by the fact that I wish to argue—here finding myself in disagreement with the assertions of many Holocaust survivors and witnesses—that all suffering, including theirs, is expressible in language. Still, that is what I believe to be the truth, and it is surely possible to argue for it. Moreover, the memory must be kept alive, and one way of doing this is to engage seriously with the statements of witnesses; dissenting from their conclusions,

14  Introduction and saying why one so dissents, where one does dissent, is a way of doing that. And, to anticipate a point I shall make later in the text, it seems to me that, so far from its being the case that proper respect is shown to the victims of cruelty and persecution by maintaining that their pain is inexpressible in language, the opposite is the case: it is only because suffering is expressible in language that we can, and have the right to, condemn the wanton infliction of pain and suffering in the way that we do and expect all civilized people to do. The victims’ interests—indeed everyone’s interests—are actually not served by the ‘inexpressibility’ thesis; quite the reverse. At the beginning of her book on the problem of evil, written from a Christian perspective, Eleonore Stump states that (2010, 16) for some evils, the grief and the pain are so great that, in my view, those evils are not fit subjects for the academic exploration of the problem of evil. For me, the Holocaust is an evil of this sort. Although it is vitally important for us to remember the Holocaust and to reflect deeply on it, taking it simply as one more example or counterexample in academic disputation on the problem of evil strikes me as unspeakably awful. It is enough for me that I am a member of the species that propagated this evil. Stricken awe in the face of it seems to me the only response bearable. I must admit to finding this a little puzzling. A study of the problem of evil, and one which essays a defence of divine omnipotence and benevolence, but which announces at the outset that it is not going to consider a particular episode of human evil because it would be ‘unspeakably awful’ to do so, seems to me rather to undermine its own prospects of success. One finds oneself wondering whether a writer who is not prepared to look that horror in the eye—and in an academic work, that means discussing it and arguing about it—is best placed to offer a convincing theodicy. But that is not my concern here, because this is not a book about the problem of evil. I mention the point because it seems to me that, quite generally, Stump’s attitude, however honourable, is misguided, and that it would be mistaken for me to adopt it here. The Holocaust was a human phenomenon, and that means that, like all human phenomena, it can, and should, be understood. For that to happen it needs to be studied, like any academic area, and since it will inevitably turn out that scholars and investigators do not agree on everything, that means that the Holocaust will figure in academic debates—at least as long as universities and other relevant institutions preserve free speech. To ‘reflect deeply’ on any phenomenon, human or other, means to raise questions about it; but questions—at least, questions of any weight and significance—typically call forth a range of different answers, and so the argument has already begun. To suggest, as John McDowell does, that

Introduction 15 it is illegitimate for philosophers to (for example) hypothesize, for purposes of ethical reflection, a counterfactual scenario in which the Final Solution was differently carried out, on the grounds that ‘we bring that horror into the domain of debate, where it does not belong’ (2008, 132), seems to me quite wrong. If we close down discussion of certain topics, we risk repeating the kinds of error that authoritarian regimes themselves make. When everyone feels that something is beyond discussion, it is absolutely incumbent on philosophers and other intellectuals to start discussing it. And, quite apart from the injunction that is laid upon us to find out the truth, however uncomfortable it may be for us, we should, as I said in the last paragraph, welcome debate and controversy, because they hold the facts, disputed or not, in view; they keep the flame of memory alive.

Notes 1. Wittgenstein 1977, I, §67. Cf. A. Fowler 1982, 40–2; Silk 1996a, 5; 2000, 67; Eagleton 2003, 3; Mastronarde 2010, 47; C. Hamilton 2016, 82–3. 2. See here Rundle 2001, 186–7. 3. See Gaskin 2013a, 32–6. 4. A. Fowler 1982, 37; Kames 2005, vol. 2, 649. 5. Tr. Hubbard in Russell and Winterbottom 1972, 97. Throughout this book, where no translator is specified the translation is mine. 6. See Kelly 1993, 23–7, 78–81, 146, 170, 194, 218–22; 1997, 182–3; Halliwell 1996, 336–9; Poole 2005, 14; Goldhill 2012, 144–5. 7. Seidensticker 2008, 334 n. 3. 8. Cf. Shapiro 2005, 182. 9. See Foster 2004, 159–62. 10. See Coyle 2003, 31–5. Cf. Cooper 2010, 156–8. 11. For the text see Lawler 1974, 136–42; cf. Kelly 1993, 100–2. 12. Hoxby 2015, 106–7, 147. 13. See here Steinmetz 1987, 100; Lamport 1990, 22–3; H. Nisbet 2013, 193–206. 14. Cf. Leech 1969, 34–5; Donaldson 1982, 75. 15. Montherlant 1955, 3, noted by J. Campbell 2005, 140; cf. Paige 2017, 205–6. 16. See Hall 2014. 17. See Steiner 1961, 5; Leech 1969, 30; Macleod 1982, 1–16; Ewans 1996, 438–40; J. Gould 2001, 169; Dimock 2008, 66–76; Halliwell 2012, 60. 18. So e.g. Gelfert 1995, 25; Silk 2000, 61. 19. Macleod 1982, 7 n. 3. 20. Poetics 1459b2–4; Horace, Ars Poetica 128–30; see Scodel 2009, 181–5. 21. See here Schmitt 2015, 211–15. 22. Plato Theaetetus 152e4–5; Republic 595b10–c2; 598d7–9; 602b6–10; 605c10–d5; 607a2–3. 23. Aristotle, Poetics 1448b38–1449a2, 1449b9–20. 24. Aristotle, Poetics 1453b1–7. Cf. Lessing, Hamburgische Dramaturgie, §81 (1996, vol. 4, 604); letter to Mendelssohn, 18 Dec. 1756 (ibid., 194–5). 25. See esp. Poetics chs. 5 and 23–4; Kelly 1993, 2, 8, 140; Haas 1987, 49–53.

16  Introduction 26. Scaliger Poetices I, 6 (1964, 11); S. Johnson 1958–77, vol. 7, 68; Steiner 1961, 8; cf. Bradley 1959, 70; Kelly 1997, 261; Kappl 2015, 65. 27. At least, that is the usual interpretation of Aristotle, which I follow. But for an interesting attempt to find consistency between chs. 13 and 14, see now Heath 2017. 28. U. von Wilamowitz-Möllendorff 1935–72, vol. 1, 113. 29. Dante, Epistolae X, §10 (1966, 175). 30. See Kelly 1993, 9–15; 144–57; 1997, 21–3; Norland 2009, 21. 31. Kelly 1989, 197; 1997, 6, 21, 49. 32. Burian 1997a, 181–3. 33. Cf. Seidensticker 1982, 43. 34. Hamburgische Dramaturgie, §77 (1996, vol. 4, 588). Cf. letter to Nicolai, Nov. 1756 (ibid., 163); Lamport 1981a, 88–94; H. Nisbet 2013, 211–16, 400. 35. De Vulgari Eloquentia II, 8, 8; Kelly 1993, 145–6; Silk 2000, 61. 36. On this genre and its history, see Herrick 1955; Seidensticker 1982, 20–7, 261–71; Hirst 1984; Kelly 1993, 198–9, 202; Foster 2004; Grund 2011, xxxvii–viii; Pollard 2015. 37. See Herrick 1955, passim, e.g. 314–16; Foster 2004, 17–20, 97–8. 38. Garvie 2007, 174; cf. Morris 1991, 246. 39. See Kelly 1993, 151. 40. Dante, De Vulgari Eloquentia II, 4, 7; cf. Wilamowitz 1984, vol. 1, 113. 41. Poetices I, 5; III, 97 (1964, 11, 145); Herrick 1955, 4, 25. 42. See Seidensticker 1982, 264. 43. Cf. Mueller 1980, 123; Silk 2000, 58–9. 44. Kastan 2003, 9. 45. Farnham 1970, 193. 46. On the (very extensive) body of Reformation and Counter-Reformation (esp. Jesuit) religious drama, see Chevalier 2013; Grund 2015a, 111–15; Hoxby 2015, ch. 5. 47. Dacier 1692, 186. See on this point Richards 1930, 246; Raphael 1960, 37–8; Steiner 1961, 5–7, 31, 324, 331–41; 1984, 280 (cf. 1990, 154–6); Camus 1967, 180–3; Sutherland 1990; Morse 1993, 16; Gelfert 1995, 44–8; Poole 2005, 28. Cf. Pocock 1973, 64–5; Ridley 2003, 408–10; Dewar-­ Watson 2014, 15–22; Hoxby 2015, 200–1, 240–1; Rowan Williams 2016, 110. 48. Hamburgische Dramaturgie, §§1–2 (1996, vol. 4, 234–43); letter to Mendelssohn, 18 Dec. 1756 (ibid., 188); cf. Hoxby 2015, 246–7. 49. See Waswo 1974. 50. The fragmentary Erechtheus and Phrixus probably contained similar willing self-sacrifices. Cf. also Evadne in Suppliants and Alcestis in Alcestis. See in general Allan on Children of Heracles 169–70, 408–9, and p. 34; Mastronarde 1994, 393; 2010, 261–70; Cropp on Electra 1024–6. 51. Nyquist 2008, 337, 349–53. 52. Cf. Silk and Stern 1981, 272. 53. See Macintosh 2009, 52; Greiner 2012, 628, 634–5. 54. Shuger 1994, 156. 55. See e.g. Murray 1912; Leech 1969, 51–5; Segal 1981, 45–6, 175, 208, 213; 1986a, 50–4; Oudemans and Lardinois 1987, 196–7; Seaford 1994, 311– 18; Mitchell-Boyask 1996; Goldhill 1997c, 331–3; Vernant in Vernant and Vidal-Naquet 2001, vol. 1, 114–31; Burian 2009, 103–16; Girard 2010; Greenberg 2010, 1–25 and passim; A.-B. Renger 2013, 39–41; Dewar-Watson 2014, 10–15.

Introduction 17 6. See Dodds ad loc., and on lines 1096–8. 5 57. See Hardie 1993, 28–9; Leigh 1993; Nicoll 2001, 190–8; Putnam 2001, 102–3. 58. Garnier draws on Virgil, Lucan, and the Gospel of John: see Ternaux ad loc.; Belle and Cottegnies 2017, 185 n. 2. 59. See Stirling 1968; Berger 1997, 71, 95; Miola 2000, 104–5; Liebler 2002; Shapiro 2005, 183. 60. For discussion, from various points of view and with further references, see Vickers 1973, 33–43; Halliwell 1990b, 178–9; Friedrich 1996; Seaford 1996; Vernant in Vernant and Vidal-Naquet 2001, vol. 2, 17–24; Taplin 2003, 162; Rutherford 2012, 37–9; Rudd on Horace, Ars Poetica 220; and esp. Scullion 2002. 61. Friedrich 1996, 270. 62. He is following Schopenhauer: 1986, §51, vol. 1, 354; cf. Wittgenstein 1984, 452. 63. See H. Nisbet 2013 passim, e.g. 28, 104, 139–43, 250–62, 392, 416–20, 509, 655–6. 64. See here Nozick 1989, ch. 20.

Part I

Tragedy and Moral Redress

1 Oedipus Hamartia, Freedom, and the Supernatural

1 A Question of Blame Nothing in the study of tragedy has been the subject of so much controversy as the question of the guilt or innocence of Sophocles’ Oedipus, who, having been told by Apollo’s oracle that he would murder his father and marry his mother, shunned Corinth, where he had been brought up by Polybus and Merope. He did this because he assumed that Polybus and Merope were his real parents, despite the fact that his birth had been called in question (admittedly by a drunkard) at a feast, that, on being asked to satisfy him on the point, his supposed parents had prevaricated, and that Apollo, too, when asked by Oedipus who his real parents were, had evaded the question. Travelling from Delphi, Oedipus arrived at a crossroads where he encountered an older man, clearly a nobleman of some sort, riding in a carriage and accompanied by servants. These ordered Oedipus to get out of the way, he refused and struck out, they retaliated, and in the ensuing fracas Oedipus killed the lord and, as he thought, the entire entourage (in fact one retainer escaped). He found his way to Thebes, which was being terrorized by the grisly Sphinx; Oedipus solved the latter’s famous riddle, and so routed it. Since Thebes’ king, Laius, had recently been killed, Oedipus was offered the throne, which he accepted together with the hand of Laius’ widow, Jocasta. Thus the play’s background: during the course of Sophocles’ drama it turns out that the older man whom Oedipus killed was Laius, king of Thebes and his own father, and that Jocasta is his mother. When the facts emerge, Jocasta commits suicide, and Oedipus blinds himself. The play begins some years into Oedipus’ reign. As king of Thebes, he is called upon to cure the city of a strange plague: his attempts to ascertain its causes lead, step by agonizing step, to the discovery of his origin and past deeds; the play ends with Oedipus’ separation from his children, a scene which is ‘one of the most overwhelming moments in the Western imaginative tradition, bearing comparison with Priam’s kissing of Achilles’ hands, or Lear’s final entry with Cordelia’s body in his arms, or Wotan’s Farewell’.1

22  Tragedy and Moral Redress The question I wish to address first in this book is whether Oedipus is guilty. Is he to blame for his fate? This has often, especially since the onset of the early-modern period, been considered to be the play’s central problem.2 An influential statement was that of André Dacier, who in his commentary on Aristotle’s Poetics declared that Oedipus’ transgression was ‘the fault of a man who, consumed by anger at the insolence of a coachman who tries to move him aside against his will, kills four men two days after the oracle had warned him that he would kill his own father’, and that he was beset by ‘pride, violence and a fit of anger, temerity and imprudence’ (1692, 192). In a sense, Dacier conceded, Oedipus’ punishment was unjust, because his acts of parricide and incest were involuntary; but though he was ignorant he was also foolish, reckless, and subject to passion. His vices are just those ‘of which Sophocles wants us to rid ourselves’.3 If Dacier imagined that he had spoken the last word on the question of Oedipus’ guilt he was sadly mistaken, and the matter has been vigorously debated ever since. Sophocles’ tragedy, according to many, is not about guilt and punishment; rather, it aims to be an imago humanae vitae, and in particular to insist on the frailty of human happiness. Such critics have often held Oedipus to be morally innocent. Thus, in his influential book on Aristotle’s Poetics, first published in 1894, S. H. Butcher construed the facts of the Oedipus story rather differently from Dacier, opining that, ‘though of a hasty and impulsive temperament’, Oedipus was not, ‘broadly speaking’, brought down by ‘any striking moral defect’, for ‘his character was not the determining factor in his fortunes. He, if any man, was in a genuine sense the victim of circumstances. In slaying Laius he was probably in some degree morally culpable. But the act was certainly done after provocation, and possibly in self-defence’ (1951, 320). Others have gone further than Butcher and affirmed Oedipus to be entirely free from blame. So in 1899, two hundred years after Dacier’s déclaration, the foremost classicist of the age, Ulrich von ­Wilamowitz-Möllendorff, spoke ein großes Machtwort: ‘Oedipus hat sich nichts vorzuwerfen’, he pronounced; ‘Oedipus has nothing to reproach himself with’ (1935–72, vol. 6, 209). And therewith he certainly intended to end the dispute once and for all. But the opposition could not be so easily dismissed, and in the chapter that he contributed to his son Tycho’s book on Sophocles, Wilamowitz Vater stated that the verbiage about Oedipus’ guilt would never end, because there would always be people who den getretenen Quark weiter treten. ‘But they should at least admit that they thereby claim to understand Sophocles better than he understood himself’ (1996, 350). I shall be treading the same quark again here; do I claim to understand Sophocles better than he understood himself? Well, Sophocles may or may not have understood his own play, and what we, Sophocles’ modern readers and spectators, are interested in is not what he, the man, thought his work meant, but what it did, and does, mean.4 But if Wilamowitz’s

Oedipus 23 talk of how Sophocles understood his play is just a roundabout way of referring to the meaning of the work (as such talk often is), then clearly someone who treads the same quark again is not going simultaneously to concede that he or she misunderstands that work. Whichever way you cut the cake, then (if that is what the quark is for), Wilamowitz’s remark misfires. Still, many recent commentators could have wished that he had laid the whole matter of Oedipus’ guilt to rest. As E. R. Dodds discovered at Oxford in the early 1960s, undergraduate heads were simply chockfull of heresy. Whereupon Dodds spoke ein kleines Machtwort for the edification of the young and impressionable, in his article ‘On Misunderstanding the Oedipus Rex’, hoping therewith to consign such views as Dacier’s to the archive for good. Dodds’s essay is often regarded as a classic,5 and as having put the whole issue beyond dispute. But, alas, it was again not to be: Dacier’s approach lives on, and it has even been revived in recent years in Wilamowitz’s own land by such able commentators as Eckard Lefèvre and Arbogast Schmitt. ‘The plays of Sophocles, it is often supposed, exalt heroic individuals who surmount the worst that can be put upon them by gods indifferent or cruel’, remarks Robert Parker, continuing: ‘But there is much to be said for a more Aristotelo-Bradleian view, whereby Sophocles’ world is one marred, above all, by the disastrous flaws endemic in human personality’ (1999, 23). Against this stands a pervasive modern view that the ‘tradition of humanistic, secularized, and psychological readings’,6 championed by Bradley and in some sense also by Aristotle, is in error. I shall defend the ‘Aristotelo-Bradleian view’, with nuances, here in connection with Oedipus Rex, and in the next chapter with Antigone. Further, I shall suggest in this part that the view applies quite generally to the protagonists of Western tragedy. In brief: typically, tragic protagonists fall through their own fault. I start my defence of the ‘Aristotelo-­Bradleian view’ with Aristotle; we shall come to Bradley in due course.

2 Aristotle and the Concept of Hamartia In chapter 13 of the Poetics Aristotle argued that the best form of tragedy concerns a hero or heroine who is neither morally outstanding nor morally base, but somewhere in between, or is rather better than worse, but who falls, in such a way as to elicit the emotions of pity and fear in the audience, because of some significant (megalē) hamartia.7 The translation of this last word has vexed commentators and critics endlessly. Dodds remarks that the concept of hamartia covers both false moral judgement and intellectual error, and that ‘the average Greek did not make our sharp distinction between the two’.8 (In English the use of the term ‘ignorant’ in a social or moral sense is still found in some places, or was until recently.)9 Dodds might have gone further, and pointed out that the distinction itself is suspect. In his article ‘Hamartia in Aristotle and

24  Tragedy and Moral Redress Greek Tragedy’, still the most important discussion of hamartia to date, T. C. W. Stinton argued persuasively that the word ‘hamartia’ has a wide range of meanings, including ones that by any reckoning count as moral. For example, Aristotle characterizes the ignorance of the wicked man (mochthēros) of what he should do and what omit to do as a hamartia.10 Again, in Euripides’ Iphigeneia at Aulis, Clytemnestra describes Helen, whom she has every reason to regard as a serious transgressor, as ‘mistaken’ (examartousa, 1204), which shows how strong, in a moral sense, this word can be.11 Perhaps even more strikingly, Orestes accuses Apollo in the Orestes of hamartia (596), a bold move in Euripides’ best blasphemous style, clearly evincing a moral sense. Stinton was anticipated by Gilbert Murray, who in his preface to Bywater’s translation of the Poetics remarked, concerning Aristotle’s concept of hamartia, that ‘it is a mistake of method to argue whether he means “an intellectual error” or “a moral flaw”. The word is not so precise’.12 ‘Flaw’ here is wrongly selected, both given the contrast Murray wanted to draw and in the light of general considerations, as we shall see shortly: he should have written ‘a moral error’. But modulo that adjustment his point stands. Despite these precedents, we still regularly encounter simplistic statements such as the following: ‘For Aristotle (Chapter 13), Oedipus is an exemplary tragic figure whose fall is consequent on an intellectual error (hamartia) rather than any morally dubious action’.13 What Stinton showed is that the ‘rather than’ here is misconceived. Stinton notes that it would be fallacious to argue that contexts in which the word ‘hamartia’ is applied to a failure to hit the target, or similar, are non-moral ones if the failure in question is one that we would classify as moral.14 But he does not question the widespread modern assumption that there is a fundamental division between moral and intellectual error: he does not allow the Greek policy of using one word to cover both of these cases to suggest to us that perhaps our neat distinction between the two is facile. This emerges from his insistence that Oedipus’ error was of a factual, not of a moral, nature; what goes missing here is the thought that it might be both—that a given factual error could be a moral error, and, more generally, that moral error might be a species of factual error (morality being a domain of fact); correspondingly, that moral intelligence and moral stupidity might be genuine cognitive categories. Again, Stinton draws up a list of what he takes to be the range of hamartiai that we can presume Aristotle would have wished to allow as constituting good subject matter for tragedy. The list comprises (i) some involuntary acts (akousia), namely acts done through ignorance, that is, mistake of fact, unless the consequences are wholly unforeseeable; and (ii) voluntary acts (hekousia), including so-called mixed actions (see §5 below), acts done in ignorance due to passion, acts of injustice (adikēmata) committed without deliberation because of the onset of a passion, and acts that arise from akrasia, weakness of will (1990, 157–8). Stinton remarks

Oedipus 25 that all these involve extenuating circumstances. He does not suggest that they all have a moral dimension, but they do. Even the involuntary acts that he lists have a moral status, since here wholly unforeseeable consequences are expressly excluded: the agent could and perhaps should have foreseen the consequences, but did not. Dodds observes that ‘Since Poetics 13 is in general concerned with the moral character of the tragic hero, many scholars have thought in the past (and many undergraduates still think) that the hamartia of Oedipus must in Aristotle’s view be a moral fault’ (1973, 66). Dodds himself rejects the view that Aristotle held any hamartia—and so Oedipus’ hamartia—to be necessarily in some way moral;15 but I believe that this traditional interpretation of Aristotle is correct. That is secured, I suggest, by the fact, already noted, that the tragic hero is there represented as being intermediate between the wholly virtuous agent, the epieikēs, who by implication is morally outstanding (not the usual sense of this word, but the context seems to constrain it here),16 and the wicked or base person, the mochthēros or ponēros or kakos. Since the extremes are moral qualities, it follows that the intermediate position must also be moral in nature; and indeed that the tragic hero is morally intermediate is explicitly stated by Aristotle. Now it is true, as some critics have observed,17 that Aristotle does not explicitly say that the hero’s hamartia has to do with his morally intermediate status: one might insist that Aristotle’s wording was compatible with the hero’s being of a morally intermediate status anyway, which would leave the door open to the possibility that the hamartia was non-moral, and unconnected with the hero’s moral character.18 But that would be a disappointingly uneconomical upshot. One would expect the hero’s morally intermediate status and his hamartia to be significantly related. In fact one would expect Aristotle to think that the hero enjoyed a morally intermediate status by reason of his hamartia, or rather by reason of possessing a character issuing in, and exemplified by, that hamartia. The philosophically taut position would be that the hamartia in question measures the moral distance between the hero’s morally intermediate status and the moral status of a fully virtuous person.19 Economy of thought demands that the hamartia be, for Aristotle, the reason why—or at least organically connected with the reason(s) why—the hero does not count as morally outstanding. If this is right, it follows that, for Aristotle, the hamartia committed by a tragic hero, and because of which he falls below the high standards of the epieikēs, is a specifically moral mistake. (This also fits with Aristotle’s use of ‘hamartia’ and cognates in his discussion of the virtues in Nicomachean Ethics II–V, where ‘hamartia’ is regularly used of missing the mean and tending to excess or deficiency, so that it is clearly moral in content.) The fall of a morally flawless agent would, Aristotle informs us, occasion outrage (Poetics 1452b36). Why? The obvious reason is that there would, in accounting for the fall of a morally flawless agent, be a deficit

26  Tragedy and Moral Redress of rationalization and justification.20 In accordance with the principle that the events of the tragic plot should, given the initial parameters, unfold according to probability and necessity (Poetics 9),21 Aristotle’s view, one surmises, is that no such perspicuous and comprehensible causal sequence would be exemplified by a wholly good man struck down by simple ill luck (by an atuchēma).22 We must be able to account for the hero’s fall in terms of his actions, and these actions in turn must be grounded in the agent’s moral character: only so will the tragic plot unfold in an explanatorily satisfying way. So the hamartia must arise explicably from the agent’s character. This is a version of the point about economy. Here we should note that Aristotle is primarily concerned with tragic agency. Martha Nussbaum objects to the supposition that ‘hero-causality’ and ‘unintelligibility’ are ‘exhaustive options’: ‘They are not. There is no mystery about what happens to the Trojan women. War, rape, slavery, murder are all too easy to understand. But what mistake did they make, innocent or otherwise, to bring all this about?’ (1992, 141). I shall examine the case of the Trojan women in Chapter 3: here I simply note that one of my conclusions there will be that, to the extent that tragedy portrays suffering as undeserved, it also diminishes the agency of the sufferers. (Nussbaum seems to concede this point: 1992, 156.) So I think Aristotle’s view is best cast in terms of agency: to the extent that disaster befalls a tragic agent, it is from an explanatory perspective most economical and satisfying if that agent’s suffering is organically connected to mistakes that flow from his or her character. (I shall say more about this formulation shortly.) And, as far as the point about outrage goes, we might add the following: it would scarcely be less outrageous for a tragic agent who, as Aristotle insists (1453a16–17), though morally intermediate is rather better than worse, to be punished for a mistake that was not organically related to his moral character, than for a morally perfect protagonist to be punished for such a mistake.23 That supplies additional support to the view that the tragic hero’s hamartia serves to measure the moral distance between his morally intermediate status and a fully virtuous person’s moral status. A further, quasi-linguistic argument for the moral nature of the hero’s hamartia was given by Butcher in a famous passage where he wrote that the word ‘hamartia’ is brought by Aristotle in Poetics 13 ‘into relation with other words of purely moral significance, words moreover which describe not an isolated act, but a more permanent state’ (1951, 319). Butcher’s point is that, since Aristotle says that the tragic hero is neither outstanding in virtue (aretē) and justice (dikaiosunē) nor marked by wickedness (kakia and mochthēria), but falls because of a significant hamartia, we would expect this last word to denote, like the other key terms in its vicinity, a relatively fixed, morally assessable state of character. Now I have agreed that the context shows that Aristotle did conceive his hero’s hamartia in moral terms. However, we must exercise care over Butcher’s other inference—that ‘hamartia’ denotes a state. In the same

Oedipus 27 place Butcher influentially offered ‘flaw of character’24 as a gloss on Aristotelian hamartia, adding that the flaw should not be ‘tainted by a vicious purpose’. But, as a matter of terminological book-keeping, we should insist that a hamartia is not as such a character trait; it is a mistaken action, or a mistaken belief that gives rise to a mistaken action.25 So the word ‘hamartia’ does not mean ‘flaw of character’. If, for example, the hamartia of Virgil’s Dido was her betrayal of Sychaeus, it would be a category error to suppose that this was her tragic flaw, because betrayal is an act, not a state.26 Nevertheless, the underlying point that those persuaded by Butcher’s translation have often had in their sights is, I suggest, correct: namely that tragic heroes who commit hamartiai do so because they have flawed characters.27 Dacier may have thought that Oedipus was punished for his character, but that cannot be right just as it stands: as we shall see in due course, he is punished for his deeds. Still, his deeds arose in a comprehensible way from his character: in a play constructed according to Aristotelian principles of probability and necessity, that is just what we should expect,28 and as we shall see it is what, time after time, the tragic tradition offers us. Moreover, for an Aristotelian, an agent is at least partially responsible for his character.29 The essential point was seen by Lessing, who in a letter to Moses Mendelssohn emphasized that the hero’s misfortune must be grounded, via a hamartia, in his character.30 And it takes us to Bradley, for two of the key doctrines of Shakespearean Tragedy are that ‘The centre of tragedy . . . may be said with equal truth to lie in action issuing from character, or in character issuing in action’, and that ‘action is essentially the expression of character’ (1991, 29, 35). So the facts that hamartia is moral and that it is grounded in character are intimately connected in Aristotle’s thinking. They are connected via the idea of the tragic agent’s morally intermediate status: the hamartia arises from that status, which makes it both moral and grounded in the agent’s character. As far as the latter point is concerned, there are two opposite errors here that we need to avoid. On the one hand, there is the view encapsulated in the translation ‘flaw’, namely that hamartia just is character.31 Against this, we should insist that character (ēthos) and tragic error (hamartia) are categorially distinct: character is a standing disposition (or set of dispositions); an error is an action. On the AristoteloBradleian view, the tragic error, which is an action, flows from the relevant agent’s character.32 At the opposite extreme, the view that ēthos and hamartia have nothing to do with one another, and that a hamartia is a lapse committed by someone who is of generally good ēthos,33 forfeits any link between character and action. In between these two unattractive extremes, we have (appropriately enough) the Aristotelian and the Bradleian and, I suggest, the correct view that hamartia and ēthos are indeed ontologically distinct, but that the former arises from the latter in a fully causal and comprehensible way. Christopher Gill argues (1986, 262) that in Sophocles’ Ajax the moral character of Ajax is not a datum

28  Tragedy and Moral Redress but is itself the central problem of the play, and that this reading of it is non-Aristotelian. But there is no difficulty for Aristotle: if we accept Gill’s interpretation of the play (as I think we should), an Aristotelian can say that the problematic nature of Ajax’ hamartia leads back to the problematic status of his character, precisely because the hamartia arises naturally from his character. Peter Lamarque argues (2004, 275) that we should not think of hamartia as a ‘fatal flaw’ in character in a purely negative sense, for then the character would be somewhat less deserving of our sympathy. Rather, as Aristotle conceives it, hamartia is best thought of as a contingent by-product of otherwise admirable character traits—for example, the tendency of the courageous soldier to take especially unwise risks. But the tendency of a soldier to take especially unwise risks is rashness, and in Aristotelian terms that is a flaw ‘in a purely negative sense’.34 Lamarque is right that, for Aristotle, tragic heroes must be admirable; but they must also, in Aristotle’s view, be flawed. Tragic economy would favour a connection between their flaws and what makes heroes admirable—and indeed a necessary connection, not mere contingency, as Lamarque suggests. That is just what we find in the tradition. For example, the tragedy of Euripides’ Hippolytus consists, as W. S. Barrett remarks, ‘in the fact that his downfall springs from a defect that is the reverse side of his very virtue: his cult of purity, for all its beauty and nobility, is bound up with an intolerant rejection of an essential part of human life’ (1964, 391). The relevant defects of character in tragic protagonists do not, as such, make them wicked. It is worth stressing this point, since one sometimes meets the contrary assertion. For example, Richard Waswo argues that Aristotle’s doctrine of hamartia does not apply to Marlowe’s Faustus or Shakespeare’s Macbeth on the basis that ‘If we regard the choice as guilty, then the terrible suffering which Faustus and Macbeth subsequently undergo is [sc. on Aristotle’s view] merely the just punishment of bad men, which is not tragic’ (1974, 64). But a choice can be wrong, even deeply wrong, without rendering the chooser wicked. It would be an opposite mistake to shy away from the full implications of the judgement that the tragic agent does indeed go wrong morally. Critics sometimes worry that, if we say this, we cast ourselves in a morally superior role, wagging a homiletic finger at the hero or heroine. But a judgement can be moralizing without being moralistic, in the pejorative sense of this latter word. The moralizing stance is inevitable: if I judge that you have made a moral mistake, I necessarily censure your action; to that extent, I do take up a morally superior stance. It does not follow that I have to adopt a sanctimonious attitude towards you and your error, and indeed Aristotle’s stress on the similarity to us of the tragic protagonist, and on

Oedipus 29 the audience’s feelings of fear and pity, precludes this. There is no inconsistency between the audience’s censuring the tragic agent’s conduct, and feeling sympathy for him or her: on Aristotle’s view the spectator can, and does, do both.35 (I shall return to these points in §5.) I have argued that Aristotle intended his use of ‘hamartia’ in Poetics 13 to be taken in a moralizing sense, against interpreters who deny that a hamartia is a moral mistake. Brian Vickers, following Gerald Else, holds that ‘hamartia’ can only mean ‘ignorance of the identity of a blood relative’ (1973, 61). (Vickers is influenced by Aristotle’s examples, to which I shall come in due course.) Dodds gives a slightly less restrictive account of hamartia, writing that it is almost certain that Aristotle was using ‘hamartia’ here [in Poetics 13] as he uses ‘hamartēma’ in the Nicomachean Ethics (1135b12) and in the Rhetoric (1374b6), to mean an offence committed in ignorance of some material fact and therefore free from ponēria or kakia.36 Unfortunately, Dodds’s cross-reference is inexact. If we look at what Aristotle actually says in the first of the passages that we are referred to, NE V, 8, we find a threefold distinction among kinds of harm (blabai) between atuchēmata, hamartēmata, and adikēmata, this last species being further divided into two subspecies (1135b11–25).37 When the harm is accompanied by the ignorance of the agent, and could not have been foreseen, Aristotle tells us that it is a stroke of bad luck, an atuchēma; at the other extreme, we have acts of injustice, adikēmata, performed by unjust and wicked agents, adikoi and ponēroi. Aristotle distinguishes two intermediate cases: on the one hand there are mistakes, hamartēmata, which are like atuchēmata except that they could have been foreseen; on the other hand there are acts of injustice, adikēmata, which are unlike extreme such acts, in that the agent did the critical deed without wickedness. In this latter case the agent knows what he is doing, but he does not do what he does as a result of prior deliberation; he is therefore not wicked. He acts, Aristotle informs us, as a consequence of anger or another passion. Agents who slip up in this way, we are told, are hamartanontes (that is, they commit a hamartia); in fact they are described in very similar terms to akratics, that is, weak-willed agents.38 Of these four types of harm, the two intermediate cases are clearly both relevant to Poetics 13, whereas adikēmata performed by wicked agents are ruled out as candidates for tragic hamartiai, as are atuchēmata, which do not occur by probability or necessity.39 Examination of NE V, 8 reveals that not only what Aristotle expressly calls ‘hamartēmata’, but also adikēmata that are performed without deliberation or wickedness, and in respect of which, as we have just noted, Aristotle also uses the language of hamartia, correspond to the ‘hamartia’ of Poetics 13.40

30  Tragedy and Moral Redress Such adikēmata are moral transgressions; but it is also true, as I have observed, that what Aristotle calls ‘hamartēmata’ at NE 1135b18—that is, those blabai that are like atuchēmata except that they could have been ­foreseen—are moral failings, precisely because, though the relevant consequences were not foreseen, they could—and, it is implied, should— have been foreseen.41 So while Dodds is right that a hamartia is free from kakia, he is wrong to hold that this is because it is non-moral, limited to mere ‘ignorance of some material fact’.42 It is also misleading to say that in NE V, 8 ‘hamartia and hamartēma are sharply distinguished from flaw or defect of character’.43 They are distinguished in the sense that they are not as such flaws or defects of character; their ontological category is that of, broadly speaking, events. But they do, as we have seen, arise from such character traits. Again, when Artemis at the end of Euripides’ Hippolytus (1334–5) tells the grief-stricken and repentant Theseus that ‘ignorance acquits your hamartia of wickedness (kakē)’, Barrett’s comment on ‘hamartia’ (ad loc.), that ‘the word marks the action [of bringing about Hippolytus’ death] as wrong without either condemning it or excusing it’, is hardly coherent. Artemis allows that Theseus’ ignorance excuses him of wickedness, that is, of committing Aristotle’s extreme type of adikēma. But, overcome by passion, Theseus did commit Aristotle’s milder form of adikēma; that was a hamartia, and as such it does attract censure. It is time to return to Oedipus. Stinton, like Dodds, thinks that Oedipus’ hamartia was not of a moral nature, and that this shows that ‘hamartia’ in Poetics 13 cannot be exclusively moral in sense;44 but I shall now reject the premiss of this argument.

3 Oedipus and Cognitive Failure Dodds’s view, in agreement with Stinton, as we have just seen, is that, although he killed his father and married his mother, Oedipus is morally innocent. Many other commentators concur: Thomas Gould, opining that Aristotle’s doctrine of hamartia was ‘probably the worst blunder in the history of literary criticism’, asserts that ‘of all the plays that we have, ancient or modern, none is so specific and eloquent about the complete absence of any connection at all between the character of the protagonist and the consequences of his acts’; for William Empson, Sophocles’ play is ‘only a bad-luck story’; for John Gould the play ‘has nothing to say about responsibility’; for Uta Korzeniewski, Oedipus’ fate is ‘eine schreiende Ungerechtigkeit’; and so on.45 Dodds indeed concedes that Sophocles’ hero has faults, which are given ample airing during the stage action: he is ‘proud and over-confident; he harbours unjustified suspicions against Teiresias and Creon’, and he expresses scepticism about oracles (1973, 66). But these faults—we shall have to fill out Dodds’s rather meagre inventory—are, in his view, irrelevant to the tragedy: ‘Years before the action of the play begins, Oedipus was already an incestuous

Oedipus 31 parricide; if that was a punishment for his unkind treatment of Creon, then the punishment preceded the crime—which is surely an odd kind of justice’ (ibid.). I shall suggest (in Chapter 3) that ‘oddities’ of this kind may indeed be found in tragedy, but as far as the Oedipus Rex is concerned, we can concur with Dodds that the murder of Laius and marrying of Jocasta were not punishments for Oedipus’ misconduct during the action of the play: these earlier offences (I shall come to their exact nature in due course) are punished in the denouement of the drama, but they are not to be regarded as being themselves punishments—not unless one accepts from Hugh Lloyd-Jones that Sophocles intends the ancestral curse on the house of the Labdacids to be functional in the play.46 But, though the curse might be held to be important in Aeschylus’ Seven Against Thebes (and perhaps in the Oresteia),47 as well as in Euripides’ Phoenician Women and Orestes, I agree with numerous commentators that it is of little significance to Sophocles’ drama.48 How are Oedipus’ punishment for parricide and incest and his bad behaviour on stage related? Those who hold that he is punished for a moral hamartia must choose between (i) the bad behaviour is that hamartia; (ii) the bad behaviour shows the sort of person Oedipus is, and so was at the time of the acts of killing his father and marrying his mother. According to (ii), the bad behaviour explains how Oedipus could have been, and in fact was, responsible for those acts, and is therefore culpable. Some older moralizers might have been happy with (i), which involves the concept of poetic justice; but modern moralizers are careful to limit their position to (ii); they hold that Oedipus is punished for his deeds (of parricide and incest), that those deeds flow from his character, and that the stage action shows how Oedipus has the kind of character that might issue in such deeds.49 Of course, the acts of parricide and incest occur, as Schiller complained to Goethe,50 and as Aristotle had already observed, ‘outside the drama on stage’ (Poetics 1453b31–2), in the ‘long time’ of the play.51 But that, as Aristotle notes, is true of crucial acts in other tragedies too: indeed we find it exemplified throughout the tradition, by for instance Aeschylus’ Agamemnon, Garnier’s Hippolyte, Milton’s Samson Agonistes, Bellini’s Norma, and Büchner’s Dantons Tod.52 So prudent moralizers will insist that it is no objection to their approach that the crucial hamartiai occur before the stage action begins.53 Now Oedipus is in some respects a good man; at any rate he is not a wicked man in the Aristotelian sense. But he has faults, which emerge during the action of the play, and these are highly relevant to his earlier deeds of parricide and incest. Dodds, as we have seen, mentions his overconfidence,54 and his unjustified suspicions of Teiresias and Creon. These failings are both connected with his fabled swiftness of thought and action.55 We first meet this characteristic as an apparently positive endowment: when the priest in the opening scene hesitantly recommends

32  Tragedy and Moral Redress consulting an oracle to ascertain the cause of the plague, it turns out that Oedipus has already acted (69); by the time the chorus proposes summoning Teiresias, Oedipus has already done so (287).56 But as the play develops, we realize that this speed of reaction has another and less satisfactory side to it, namely a tendency to react too swiftly and without proper reflection:57 as the chorus puts it, ‘the quick in counsel are not sure’ (617, tr. Jebb). We see this aspect of Oedipus’ character operating in the interviews with Teiresias and Creon—for he ignores the fact that hitherto Creon has been a trustworthy friend (385) and Teiresias a good seer (300–1)58—and in the accusation of arrogance that he directs at Jocasta, when she has discerned the truth.59 But the key failing that is responsible for Oedipus’ tragedy lies in what his precipitate behaviour signals, namely cognitive deficiency—the burden carried, I suggest, by the great majority of tragic heroes and heroines in the Western tradition. Moreover, Oedipus’ intellectual deficits are not the reverse side of an intellectual strength, because, contrary to what numerous commentators have asserted, he does not have any such strength. Dodds does not mention Oedipus’ cognitive defects, because like so many others—both characters in the play and readers and spectators of it—he is dazzled by Oedipus’ reputation for intellectual prowess. Thus Maurice Bowra talks of Oedipus as ‘a man of powerful intelligence’, having ‘the gift of reaching rapidly the answer to a problem, of asking pertinent questions’; he has ‘vigorous mental equipment’, is ‘keen-witted and live-minded’; Bernard Knox tells us that Oedipus ‘has a brilliant intellect’; Brian Vickers that he is a ‘supremely active and intelligent man’; R. P. Winnington-Ingram that he is ‘the most intelligent of men’; Charles Segal that he has ‘extraordinary intelligence’ and is ‘an expert at decoding difficult messages’; Simon Goldhill that he is ‘a man of domineering intelligence and powerful scrutinizing mind’; Hans-Dieter Gelfert that Oedipus goes about his investigation ‘like a sharp-witted detective’; R. D. Dawe that he is ‘the most far-sighted of men . . . with a brilliant incisive intellect’; Vayos Liapis that he is ‘one of the finest specimens of human intelligence’; Peter Holbrook that he is ‘an intellectual prodigy’ with a ‘preternatural intelligence’.60 Similarly, many commentators have thought that the play presents Oedipus as someone dedicated to finding the truth,61 and have expatiated on his ‘epistemophiliac passion to lay bare his own origins’, as Terry Eagleton nicely puts it (2003, 233). Jean-Joseph Goux even thinks that Oedipus ‘inaugurait la conscience philosophique’ (1990, 140). The list of commentators who have been staggered by Oedipus’ alleged intellectual brilliance could be considerably extended.62 These accolades demonstrate how easy it is for critics to find what they want or expect to find, regardless of the evidence before them. In reality Sophocles’ drama shows Oedipus unintelligently avoiding the truth, which is right under his nose, for as long as possible; it shows him to

Oedipus 33 be fixated on the near in time and place, to lack the imagination and ability to entertain different and alternative possibilities, to weigh and compare probabilities, and to marshal correctly the information that he has at his disposal.63 Bowra concedes, as most critics do and as anyone indeed must, that ‘despite his acute intelligence [Oedipus] is unable to see the truth until it is forced upon him’ (1945, 191), but it does not occur to him or to like-minded critics that Oedipus’ tardiness on the uptake could imply that the characterization of his intellect as ‘acute’ might be in need of revision. Again, if you start from the assumption that Oedipus has a ‘brilliant incisive intellect’, as Dawe does, then you will inevitably find the Teiresias scene dramatically unsatisfactory: ‘the apparent failure of the highly intelligent Oedipus to grasp what has been said to him is unconvincing; and the structure of the plot suffers from premature disclosure’ (2006, 9). But Dawe does not ask himself what the dramatic point of the ‘premature disclosure’ might be, or whether the fact that he finds the Teiresias scene unconvincing might be a function of his initial assumption that Oedipus is highly intelligent rather than an indication of Sophocles’ poor draughtsmanship. He puzzles over the fact that ‘a man whose intuitive brilliance had solved the riddle of the Sphinx’ (to which I shall come in §4) is unable to put together the pieces of information before him in the obvious, and correct, way—pieces of information which ‘should have led even the least gifted intelligence to the right conclusion’ (ibid., 16). But it does not strike him that it might be worth re-examining his initial assumption that Oedipus is a man of ‘intuitive brilliance’. We find a similar effect in some versions of the Faust story, such as Marlowe’s, whose Faustus is cried up, by him and his students, as a cerebral prodigy, though what we actually see on stage, in his dialogues with Mephostophilis, is intellectual folly and a rash willingness to make a conspicuously bad bargain.64 Oedipus’ cognitive failure emerges most impressively from Sophocles’ drama in the way we find an almost exact match between the information that the hero receives, quite early on in the story, about the circumstances of Laius’ death, and what he already knows about his own past: yet he fails to make the not-very-demanding mental leap and put the pieces together in the obvious way until near the end of the play.65 Lefèvre sets out some of the correspondences:66 Jocasta tells Oedipus that Laius received an oracle that he would be killed by his son (711–14); Oedipus already knows that he received an oracle that he would kill his father (793). She tells him that Laius was killed at a crossroads (715–16); he already knows that he killed an older man at a crossroads (800–13). She tells him that Laius’ son had his ankles fastened together (718); Oedipus knows that just this fate befell him (1032–3). To Lefèvre’s list we can add that Oedipus hears from Creon and Jocasta that Laius was accompanied by a retinue, all of whom were killed except for one man who escaped and reported the incident in Thebes (106–19, 750–6); Oedipus already

34  Tragedy and Moral Redress knows that in the fracas with the elder in the carriage whom he killed, he also slew several others—all the others, as he thought (800–13). Now one does not have to have read Aristotle on reciprocal relations to put these pieces of information together in the obvious way. But instead of doing that, Oedipus grasps at straws and fantasizes about possible escape routes: he does not, it is true, make anything of the discrepancy between his own recollection of having killed all his opponents in the crossroads incident, and the fact that one man survived from the Laius incident,67 but he does dwell on the discrepancy between the survivor’s report of robbers, in the plural, and his own singular identity. Further, as Voltaire observed,68 Teiresias accuses Oedipus in precisely the terms of the Delphic oracle that Oedipus had received (457–60, 791–3). Why did Sophocles show his hand so early in the play—why did he engage in what, as we have seen, Dawe censures as ‘premature disclosure’—if not to cast an unflattering light on his hero’s intelligence? That is the point of the Teiresias episode: seeing blind man meets blind seeing man.69 When Oedipus hears from Creon that one of Laius’ retinue survived the incident in which Laius himself died, why does he not send for the survivor?70 Why has he not investigated the circumstances of Laius’ murder?71 Why, when he comes to understand that he slew Laius and married his wife, does Oedipus extract from these facts no more than the thought that he slew the Theban king and married that man’s wife (813–22)?72 For although all the bits of the puzzle are on the table, and fit together in an obvious way, Oedipus still thinks that Polybus and Merope are his parents. In line with what was said above in this section, we should insist that the canny moralizing critic will not leap to the conclusion that Oedipus is punished for the cognitive deficiencies that he displays on stage, but rather that we are supposed to infer that he evinced those deficiencies at a crucial earlier point—namely at the time when he killed his father and married his mother. Just so, in Aeschylus’ Agamemnon, the hero is not punished for treading on the tapestries but for earlier acts in the ‘long time’ of the play, and which emanated from the same hubristic mentality as Agamemnon shows in the ‘carpet scene’.73 Now Sophocles’ Oedipus tells us himself that he killed the elder at the crossroads in an access of passion,74 and it is clear from the description of the act of homicide that, as I shall argue below (§5), his response to the provocation he thought he had received was not an act of reasonable self-defence on his part, but an irrational and culpable overreaction to what had in fact been normal behaviour on the part of his victims. As a student of mine put it: ‘he killed his dad in a fit of road rage’.75 But more important than this point is the fact that Oedipus had the insouciance to get himself into the position, in the first place, of killing a man old enough to be his father and marrying a woman old enough to be his mother, not long after he had received Apollo’s prediction;76 by contrast, in Grillparzer’s Die Ahnfrau, Jaromir, who kills his own father, is given no such forewarning. (The twin facts

Oedipus 35 that Laius was old enough to be Oedipus’ father and Jocasta old enough to be his mother are not given any emphasis by Sophocles, but that is because he knew that they would be familiar to the audience as elements of the myth.)77 Oedipus’ outbursts of passion during the stage action, his tendency to react too precipitately, his short-sightedness, his failure to exercise his supposed intelligence—these things are there in the play to apprise us of the fact that they have always been there as features of Oedipus’ character, and were responsible for the acts of parricide and incest in the ‘long time’ of the drama.78 Those who assert that there is no connection between Oedipus’ behaviour on stage and his former crimes79 must answer a simple question: what, then, is the dramatic purpose of Oedipus’ bad behaviour on stage? Why has Sophocles portrayed him as the sort of person who would, unthinkingly and akratically, murder his father and marry his mother, if not to indicate that this is exactly what he did? It seems insufficient to say that the point of portraying Oedipus as short-tempered is merely to help elicit the emotions of pity and fear in the audience, as they observe someone like themselves suffering catastrophe.80 More than that must be at stake.

4 Oedipus and Enlightenment Values The prediction that Oedipus received from Apollo’s oracle was a response to his question: are Polybus and Merope my parents? Instead of answering this question directly, Apollo replied, as we noted at the beginning of the chapter, evasively, telling Oedipus that he would kill his father and sleep with his mother. Apollo’s failure, like the occasional exam candidate’s, to answer the precise question set speaks volumes to us, and should have done so to Oedipus too.81 Oedipus now has three indications that his apparent parents may not be his real parents: the drunkard’s words at the feast, calling his origins into question (779–80), Polybus’ and Merope’s failure to allay his doubts (781–6)—in the Enescu/Fleg Œdipe, Merope lies to Oedipus—and Apollo’s similar refusal to answer a straight question about his parentage (788–9).82 The significance of the first piece of evidence is no doubt hard to assess: it might be worthless (on the other hand, in vino veritas).83 But the latter two pieces of evidence have a very high value indeed. An intellectually alert person would combine these clues, infer that Polybus and Merope were in all likelihood not his real parents—or at least that this was a distinct p ­ ossibility—and then return to Corinth, at the same time making a mental note of the fact that he must, under no circumstances, kill a man old enough to be his father or marry a woman old enough to be his mother. Instead, Oedipus takes to the open road and performs the contra-indicated acts in short order. That Oedipus really knows that his parentage is doubtful is indeed shown by his anxious question to Teiresias (438): ‘Who bore me?’84 Those who insist on Oedipus’ moral innocence on the basis that

36  Tragedy and Moral Redress he did not know that the man he slew was his father and the woman he then married his mother are, I suggest, missing the point, which is that he should have known, or at least strongly suspected, that these identities held, and should have entertained these suspicions before he performed the actions of slaying and marrying. Putting the point in modern terms, we may say that he has constructive knowledge of these facts; that is, he is such, and the circumstances in which he finds himself are such, that he ought to have the relevant knowledge (and not just dunamei, but energeiai). Ignorance does not, as commentators so often think,85 guarantee guiltlessness: it does not do so if it is culpable ignorance. (Note that constructive knowledge is distinct from repressed, or subconscious, knowledge, though Oedipus may have that too.)86 It is not enough to say, as Vickers does (1973, 507), that the fact that Oedipus believed his parents to be Polybus and Merope and reasoned on that basis clears him of a charge of stupidity. We have to go further and ask whether he should have held those beliefs, whether an intelligent person would have done so; and the answer is, in both cases, negative. Vickers concedes that Oedipus ‘does not ask enough questions’ (ibid., 511); but, like so many other critics, as we have seen, he does not think to retrace his steps and revise his view that Oedipus is ‘a supremely . . . intelligent man’ (ibid., 498). By contrast with Sophocles, Aeschylus and Euripides may not have given the Oedipus of their versions any inkling of the doubtfulness of his parentage before he performed the acts of slaying Laius and marrying Jocasta.87 Before the 2003 Iraq War the US Defence Secretary Donald Rumsfeld famously observed that There are known knowns; there are things we know we know. We also know there are known unknowns; that is to say we know there are some things we do not know. But there are also unknown unknowns—the ones we don’t know we don’t know. What is striking about the remark is that, as Steven Pinker observes (2011, 514), Rumsfeld omits the fourth category, that of unknown knowns, things we do not know but should, or things knowledge of which we are suppressing. Hence a recurring theme of Rowan Williams’s recent book on tragedy, namely that ‘tragedy is about the effect upon us of what we do not know’ (2016, 97), is too simple. Oedipus is a case in point: the crucial facts are, for him, unknown knowns. So Oedipus’ offence, though in one way ‘venial, because unwitting’, as John Dover Wilson suggested, contrasting Oedipus’ incest with Gertrude’s (1964, 39), is very far from ‘innocent’, as Nussbaum calls it.88 Butcher, as we have seen, concedes that ‘in slaying Laius [Oedipus] was probably in some degree morally culpable’; but the act of marriage with Jocasta, he tells us, ‘was a purely unconscious offence to which no kind of blame attached’ (1951, 320). On the contrary, in his marrying of Jocasta as in his slaying of

Oedipus 37 Laius, Oedipus’ cognitive failure was culpable; it is wrong to suggest that Oedipus exerted himself to avoid committing those crimes.89 In fact he made very little effort, and the effort that he did make (not returning to Corinth) was misguided. Cognitive error is, or can be, morally blameworthy: we have obligations to find things out, to reason sensibly about information at our disposal, to draw conclusions that stare us in the face, to reckon with probabilities. So, to take another Sophoclean example, when Deianeira insouciantly makes use of the potion that Nessus gave her, she is cognitively at fault.90 Of course, in one way those numerous commentators who laud Oedipus’ intelligence are in a better position than I am to impute constructive knowledge to Oedipus and so hold him to account: they can say that he should have known the truth because he was clever enough to know it. But I can still make my point: Oedipus’ lack of intelligence does not exempt him from asking questions that anyone should have asked. The tendency to separate the cognitive from the moral should, as I have said, be resisted. Vickers is right that older attempts to preserve the notion of tragic flaw by cashing it out exclusively in terms of hubris fail (1973, 29–33), but the notion can be retained provided it is correctly analysed, and I suggest that ‘cognitive deficiency’ yields the desired analysis, with a correspondent understanding of hamartia as cognitive error. So Bywater’s translation of ‘hamartia’ as ‘error of judgement’ (1920, 50) was right: to speak, as I am doing, of ‘cognitive error’ does no more than modernize the terminology. An important Aristotelian point to make here, and one that in using the term ‘cognitive’ I presuppose throughout this study, is that the cognitive cannot be neatly separated from the appetitive or orectic.91 Morality concerns, broadly speaking, perception—seeing things in a particular way. And these seen things are not motivationally inert. If, for example, I see that someone is in distress and needs help, that perception already motivates me to action (which is not to say that I will necessarily act, or even that I ought to act: there may be countervailing factors). As far as the notion of hamartia is concerned, a crucial implication of this point about cognition is that those tragic mistakes which we wish to account for in terms of the notion of akrasia (weakness of will)—and there are many such cases—do not form a separate category from that of cognitive error, but fall squarely within it. Akrasia is, speaking abstractly, a failure to see the facts in the right way—that is, in a way that exactly matches the perception of the virtuous agent. If Oedipus’ reputation for intellectual brilliance is unmerited, how did he acquire it? The answer is that he alone solved the riddle of the Sphinx. At first blush this seems an impressive achievement, and not only other characters in the play but also critics of it are duly impressed. Further investigation reveals the matter to be less straightforward, however, and shows that Oedipus’ achievement against the Sphinx is compatible with the lack of intelligence that he elsewhere displays. For the Sphinx’s riddle

38  Tragedy and Moral Redress played directly into Oedipus’ hands—or perhaps we should say that it played directly to his feet. If there was one sort of conundrum that a man of Oedipus’ background, character, obsessions—and, one might add, name, for one of the etymologies of his name makes Oedipus ‘the man who knows about feet’—was going to have a chance of solving, it was a riddle about feet. (Oedipus puns on this etymology of his name and obliquely alludes to the Sphinx’s riddle at line 397. The content of the riddle is not mentioned by Sophocles, but that is because it was well known: it seems to have been given in Euripides’ Oedipus.)92 So Oedipus ‘got lucky’: regrettably, it turned out to be a unique stroke of luck, which had the unfortunate consequence of distorting his and others’ estimation of his abilities. The insight that Oedipus uncharacteristically displayed when he solved the Sphinx’s riddle was not repeated, either before or after: it was not shown when, in spite of the triple warning he had received, he slew his father and bedded his mother, and it is not shown in the play as he gropes his way towards final understanding. Perhaps this is felt most strongly when Oedipus fails to respond to Jocasta’s mentioning (718) that Laius’ son had had his feet pierced (the fact that lay behind the alternative, more famous, etymology of his name, alluded to in the play by the Corinthian messenger).93 Everyone in the audience solves that riddle, but not Oedipus. So when one critic writes that Sophocles’ ‘basic dramatic problem is to delay Oedipus’ discovery without straining credibility or allowing Oedipus, the great solver of riddles, to appear a fool’,94 we should reply that Oedipus is meant to appear a fool; that in turn has implications for what we should say about his status as ‘the great solver of riddles’. The first thing to say about this is that, contrary to what the plural ‘riddles’ insinuates, before the truth eventually dawns on him Oedipus has solved only one riddle.95 (The inaccuracy has an interesting precedent in the dramas themselves.)96 From the beginning of the play he carries around with him riddles about the scars on his feet, the identity of his parents, the identity of the man he killed and of the woman he married, his own identity—but he has solved none of them.97 The second thing to say is that his solving of that single riddle was not a feat of superior intelligence, but a stroke of good luck—or rather of bad luck, as it turned out, indeed of tragic misfortune. It follows that the play is not, as has often been held, in the business of mocking or subverting either analytical intelligence or Enlightenment values—Wolfgang Schadewaldt’s Dämonie des Wissenwollens um jeden Preis—in the name of a deus absconditus or dim religious primitivism.98 (Pascal’s idea of a deus absconditus was famously exploited by Lucien Goldmann in his interpretation of Racine;99 I shall suggest in Chapter 4 that the idea has equally little plausibility in that case, too.) The Oedipus Rex gives no support to the Nietzschean idea that Aeschylus and Sophocles promulgate a deeply pessimistic, mysterian, irrationalist Weltanschauung, supposedly set against a Socratic rationalism and

Oedipus 39 optimism popularized by Euripides.100 Nor indeed does the Antigone, to which I turn in the next chapter.101 Quite the reverse: in both these plays Sophocles precisely upholds Enlightenment values of pragmatic intelligence, by demonstrating how a failure to apply them leads to catastrophe.102 The doctrine of hamartia is itself a fundamentally secularizing and rationalizing move on Aristotle’s part,103 so that its application to both the Oedipus Rex and (as I shall argue) the Antigone shows that these plays are in the service of the Enlightenment project. (There is an irony in Nietzsche’s assertion that tragedy died of Socratism: Socrates himself, in the so-called passion dialogues, is in some ways a tragic figure,104 who is himself brought down by cognitive failure.) Goldhill notes that in the play Oedipus is called the ‘man of utmost mastery’ (46) and suggests that ‘even the man of utmost mastery cannot have mastery or control in all things. It is precisely the possibility of such total, all-embracing resourcefulness or knowledge or mastery that is challenged at the end of this drama’; Goux, developing his view of Oedipus as ‘first philosopher’ and precursor of Descartes, tells us that it is ‘l’exaltation de l’élément raisonable’ that leads to catastrophe; Richard Rutherford may be gesturing at a similar idea when he writes (in Hegelian vein) that ‘misguided heroes are often undone by their own strengths (Oedipus, Heracles)’.105 We have seen that Rutherford’s point can indeed be applied to Hippolytus. But it does not fit Oedipus: against Goldhill and Goux, we must say that Oedipus’ resourcefulness and knowledge and mastery are not total or all-embracing; they fall a long way short of that, and this is exactly the problem. So Oedipus is not, in the main, undone by his strengths, but by his weaknesses. What we need in the Oedipus Rex is not less resourcefulness, but more; it is not the exaltation of l’élément raisonable that we see in Oedipus, but its disastrous deficit. Oedipus’ misapplication of rational thought—his failure of intellectual mastery—comes out especially clearly in the way he grasps desperately at the discrepancy, mentioned above, between the report of robbers, in the plural, given by the survivor of the brawl at the crossroads, and his own singular identity. Oedipus and the chorus both hope that this discrepancy will save him, but the survivor’s account can only help if one reckons with the possibility that Laius and his entourage were killed by robbers and that, at more or less the same time and place, a similar elder, travelling with a similar entourage, was killed by Oedipus—a collocation of events that is not impossible, but hardly of sufficiently high probability to be worth taking seriously.106 Oedipus snatches at the difference between plural and singular, as though that could be used to prove that he did not kill Laius. He deploys the point about singular and plural to combat the point about probabilities, as though the certainty that one man could not be identical with many outweighed the mere probability, not certainty, that the survivor’s report and his own recollection might home in on the same incident. Ironically enough, despite his confidence in the

40  Tragedy and Moral Redress numerical point, there is, as has often been observed, a sense in which the discoveries that Oedipus goes on to make show that one can be many:107 just as Conscience assures the narrator of Piers Plowman that ‘knyghte, kynge, conqueroure may be o persone’ (XIX, 27), so too one man can be a homicide and a parricide, Thebes’ saviour and its scourge, Laius’ son and slayer, Jocasta’s son and husband, the infant exposed on Mount Cithaeron and the ‘child’ of Polybus and Merope. Not that these identities upset the logical point, which is that one man cannot be numerically the same as many men. They simply trade on the fact that one object may instantiate many properties, together with the Fregean point that something which one thinks of under one mode of presentation may turn out, perhaps to one’s surprise, to be identical with something which one thinks of under another mode of presentation. Oedipus’ confident belief that to reckon with the possibility that his slaying of an older man and entourage at a crossroads, on the one hand, and Laius’ being slain together with his entourage at a crossroads by brigands, on the other, were one and the same incident is tantamount to supposing the realization of a metaphysical impossibility, namely that he is identical with several men, rather than merely showing the survivor of the latter incident to have been an unreliable witness (recall Hume on miracles: 1975, 109–31), is another aspect of his cognitive underperformance in the drama. Moreover, the possibility that the survivor might deliberately, perhaps for good reasons, have falsified the evidence does not occur to him, any more than, in his play, a parallel possibility does to Othello. In addition, Oedipus is more concerned with what the survivor said at the time than with what he might say now—that is, he is more concerned with confirming a convenient story than with finding out the truth. Notice also the shift, in their discussion of the mêlée at the crossroads, from Creon’s talk of ‘robbers’ (at 122) to Oedipus’ reference to a single ‘robber’ (at 124), a discrepancy picked up by Seneca.108 It is almost as if Oedipus knows—really knows, that is, not just constructively knows. Or perhaps repressed knowledge is surfacing.109 Dawe denies that we have to do with a Freudian slip (2006, 7). Be that as it may, the mere fact that Oedipus represents Creon’s plural as a singular should have alerted him to the possibility that the alleged plural was a singular. The move to the singular is, I suggest, the playwright’s way of notifying us that Oedipus had an obvious thought available to him which he failed to pursue. Some commentators try to defend Oedipus on the grounds that he is entitled to reckon with what are, after all, logical possibilities—namely that he is not identical with the son whom Laius exposed, or with Laius’ slayer(s), and so on.110 But this misses the point. Of course the possibilities on which Oedipus pins his hopes are indeed bare possibilities. (A further bare possibility, not entertained by Sophocles’ Oedipus but neatly exploited by Corneille’s, is that the wayfarers whom he slew were identical with the ‘brigands’ who murdered Laius.)111

Oedipus 41 But a cognitively alert person would have reckoned with the actual facts, as these turn out to be, at a much earlier stage of the investigation than Oedipus does—indeed well before the investigation began, namely at the time when he performed the fateful actions.

5 The Murder of Laius At this point we need to take cognizance of a subtlety which all commentators I have seen miss, though one or two get quite close.112 Perhaps misled by an over-literal reading of Poetics 13 and the fact that the word ‘hamartia’ there occurs in the singular, critics ask what Oedipus’ hamartia is, and whether, once we have decided what it is, it is a moral mistake. Canny moralizers, as we have seen, think that Oedipus is punished not for his actions on stage but for the prior acts of parricide and incest, and these acts evidently constitute not one hamartia, but at least two. I imply in that rider (‘at least’) that they might consist of more than two hamartiai, and the subtlety that we now have to register is the point that Oedipus’ killing of Laius already and on its own consists of two hamartiai, before ever we get to the incest. For on the one hand what occurred between Oedipus and the elder at the crossroads was an act of homicide (indeed murder, as I shall argue below in this section), and on the other hand it was an act of parricide. In terms of the various kinds of harm that Aristotle lists at NE V, 8, the act of homicide (murder) is an adikēma that Oedipus performs with knowledge but without forethought, not out of inherent wickedness, but as a consequence of anger; the act of parricide, by contrast, counts in Aristotelian metaphysical theory as an accidental property (a sumbebēkos) of the killing,113 and as a hamartēma that is like an atuchēma except that it could have been foreseen. This last point is crucial: that the killing of the elder turned out to be an act of parricide could have been foreseen by Oedipus, since he had just received an oracle to that effect; so the act of parricide was indeed a hamartēma and not a sheer atuchēma. It would be an atuchēma if it were unpredictable by the relevant agent; but in Oedipus’ case that is not so.114 Oedipus’ tragedy is that, while he indeed intended to kill a man, in a fit of rage issuing from a flawed nature and precipitating a culpable action, he did not intend to kill his father. But the killing was nevertheless an act of parricide, and predictably so; that act of parricide polluted Oedipus, and it is what he is punished for. Does Oedipus deserve his fate? One answer to this question, building on elements of Aristotle and Bradley and on what I have just said, might go as follows. Oedipus is unlucky inasmuch as an act for which he is fully responsible and which deserves punishment, namely homicide (murder), accidentally coincided with an act, namely parricide, for which he is at least partially responsible, given his state of knowledge, but which then draws down a punishment in excess of his strict deserts. On this view,

42  Tragedy and Moral Redress Oedipus does indeed deserve punishment, but not to the extent that he receives it,115 an imbalance which is held by both Aristotle and Bradley to be an important feature of tragedy,116 and which is identified by Nietzsche as something that marks out ancient tragedy, by contrast with Christian thought.117 So, too, someone might wish to say, with Virgil’s tragic Dido, who blames herself for her fall and judges that she deserves death: on the view I am considering she would be right in the first part, but at least partially wrong in the second.118 Hera thought that Heracles deserved the terrible fate that he meets in Euripides’ Heracles, but an adherent of the view I am currently airing might argue that we, the audience, can see that, though he brings his fate on himself and is therefore culpable (see further below, §§6 and 8), the punishment goes beyond the strict terms of the offence. In the first instance, we might say, divine justice is not human justice: it is not, from our point of view, morally ideal.119 (What this would in turn come to, in the light of the naturalizing move that I shall urge in §8, is that the world is not morally ideal.) In Seneca’s Hercules Furens the protagonist thinks that Juno is punishing him (604); but though Hercules is, as John Fitch says, ‘brutal and hubristic’ (1987, 276), we might again think that the punishment exceeded the offence. In connection with Dionysus’ punishment of the whole house of Cadmus for Agave’s and Pentheus’ sins, Dodds remarks that ‘when great natural forces are outraged we can expect no nice adjustment of the punishment to the magnitude of the individual offence’.120 On the view I am considering here, though the punishment is deserved to some extent, it exceeds strict desert, and so is partially just, partially unjust. (Incidentally, we should not, as is often done,121 equate an imbalance between strict desert and punishment with absolute injustice.) The imbalance between suffering and desert can then explain why Oedipus’ fate elicits the Aristotelian emotion of pity in the audience—pity for suffering that is in excess of what was strictly deserved (Poetics 1453a4).122 I have described one possible (in fact, in its general outline, quite common) view about the relation between desert and punishment, but our earlier discussion of constructive knowledge has already shown where it goes wrong. On the Aristotelo-Bradleian approach, the tragic hero(ine) is not wicked, but he or she does have defects of character; these defects do not, as such, make the hero(ine) wicked, but they do lead to mistakes which, even when these are of a sort that we might be inclined to call ‘intellectual’, are still morally culpable; the consequences of the critical actions were both foreseeable and avoidable, and a suitably vigilant agent would have foreseen and avoided them. Oedipus, though in one sense ignorant, had, as we have said, constructive knowledge of the relevant facts, and should have acted on that knowledge. As Donald Mastronarde observes, the pejorative connotations of a word like ‘amathēs’ (‘ignorant’), which is often used in a moralizing sense, ‘derives from the implied (rationalistic and optimistic) belief that the ignorance is culpable

Oedipus 43 and could have been cured by effort’.123 This censure does not apply to all forms of ignorance, but it does apply—indeed, in morally significant cases, it applies truistically—to ignorance accompanied by constructive knowledge; and it is this combination that Oedipus exemplifies. Oedipus’ ignorance does not amount to wickedness because, put in Aristotelian terms, his ignorance is not of the major premisses of relevant practical syllogisms, but of minor premisses, which treat of particularities.124 Still, his ignorance of these particularities is not adventitious, but grounded in character defects, so that he is genuinely at fault. In general, the agent’s mistake or mistakes produce consequences predictably: they arise according to Aristotle’s constraint of probability and necessity. It follows that the agent could and should have foreseen them. Aristotelian pity is, we have said, keyed to a supposed mismatch of punishment and desert. But there is a good sense, as I have just implied, in which there is no such mismatch: the agent should have foreseen the consequences, or at least foreseen that there might be (serious and undesirable) consequences, of his action. In that sense, though the agent is not wicked, it is also the case that the punishment is not undeserved, so that there need be no moral deficit between desert and punishment, at least according to traditional conceptions of appropriate levels of punishment (which were harsher than ours). Does that mean that we cannot feel pity for the tragic hero? No. Pity need not be elicited by a feeling that the suffering is undeserved: though Aristotle at one point implies otherwise (Rhetoric 1385b13–16), there is no conflict between our feeling pity for a tragic agent and that agent’s exhibiting serious culpability.125 Rowan Williams is quite wrong to say that ‘Explanation and compensation suggest that there is something somewhere that makes mourning inappropriate’ (2016, 119). Fully deserved suffering—so suffering which is both explicable and compensated (in advance) by fault; suffering, that is to say, which is subject to full moral redress—can be, and is, mourned. Mourning has a self-regarding aspect, but then we, who are in many cases no better than we should be, have reason to fear (Aristotle’s other tragic emotion: Poetics 1453a6) that a similar moral and intellectual failure, precipitating catastrophe, will befall us. If Oedipus was lucky with the Sphinx, he was unlucky with the killing; and that is the kind of thing—at least in general terms—that could happen to anyone. Dodds’s verdict on the play, that ‘we feel both pity, for the fragile estate of man, and terror, for a world whose laws we do not understand’ (1973, 67) is too loose.126 Our feelings of pity and fear are more precise than that. And it is unclear why Dodds finds the ‘laws’ of Sophocles’ world hard to understand. Given the coincidences, which, for Aristotle (Metaphysics, Δ. 30 and E. 2), are not a matter of law, everything happens quite smoothly, according to probability and necessity. Elsewhere, Dodds writes that ‘for Homer, as for early thought in general, there is no such thing as accident’ (1951, 6). Would this apply to the

44  Tragedy and Moral Redress archaizing Sophocles? You might say that, in an absolute sense, there is no such thing as an accident, at least at the level of human affairs. For here it is plausible that accidents occur in ways that are relative to causal and explanatory chains. That was at any rate Aristotle’s view: for him accidents occur when, for example, things behave unexpectedly, given their usual characterizations, as when a cook doctors or a doctor cooks; or when two causal chains intersect, as when a creditor goes to market and unexpectedly meets his debtor, or a man in a besieged city leaves it to take a drink at an outside well and to his consternation finds enemy soldiers there, who kill him (as in some versions of Troilus’ death at the hands of Achilles).127 For someone who knew enough about the ambient conditions, these coincidences could be predicted and would come as no surprise. For the Stoics, an accident is simply an event whose causes are hidden.128 Now in Oedipus’ case, as we have in effect said, the act of parricide was not much of an accident, in Aristotle’s sense, so that it scarcely matters whether we extend Dodds’s point about early Greek thought to Sophocles or not. Given the oracle, Oedipus was in a position to predict with near certainty that any act of homicide that he subsequently committed on a significantly older man, unless he made a habit of performing such acts, was going also to be an act of parricide. It was in a relative sense an accident that he met Laius there and then, but it was not an accident in an absolute sense, and someone who knew enough about all the relevant causal chains could have predicted it, exactly as in the case of the man who leaves the besieged city and encounters his mortal enemy at the remote well. This is not to deny that there could be accidents in an absolute sense, a matter on which we need here take no stand; but, if there are such accidents, Laius’ death is not one of them. Dodds states that ‘if Oedipus had been tried before an Athenian court he would have been acquitted—of murdering his father’, and Jean-Pierre Vernant opines that Oedipus has not ‘committed any crime of his own volition (de plein gré)’, or ‘intentionally (intentionnellement)’, which could ‘be personally imputed to him from a legal point of view’.129 But here Dodds and Vernant slide over an important distinction that I have been at pains to stress. No doubt Oedipus would have been acquitted of an intention to commit parricide, but would he have been acquitted of murder—the intentional killing of a man? Well, whatever the verdict in our hypothetical case would have been, it emerges from the manner in which the fateful scene is described that Oedipus’ behaviour on that occasion involved an overreaction which went well beyond self-defence—in fact I believe it is clear that Oedipus was not in any serious sense under attack from Laius or his retainers130—and that he was guilty of murder, so that in our hypothetical scenario the jury ought not to acquit him of that charge.131 As Egon Flaig shows, Oedipus cannot plead self-defence, contrary to what his apologists have supposed.132 (Contrast the versions of the killing of Laius described by Euripides’ Jocasta in Phoenician

Oedipus 45 Women, and by Cocteau’s Œdipe in La Machine Infernale.)133 Let us, following Flaig, reconstruct the scene at Sophocles’ crossroads. Laius’ servants included a herald, who will have gone before the carriage in order to announce his master’s coming and clear the route. He will accordingly have required Oedipus to give way, but Oedipus refused, thereby forcing the carriage to brake. In so behaving, Oedipus automatically put himself in the wrong, since it was an understood rule in all premodern societies that mere pedestrians, who would normally be members of the lower orders, had to yield passage to nobility riding on horseback or in vehicles. (This point, implicit in Sophocles, emerges explicitly in Euripides’ version at Phoenician Women 39–41, where Laius’ driver abruptly tells Oedipus to make way for royalty.) Anyone who flouted this convention could justifiably be regarded as a potential highwayman and treated appropriately; indeed highwaymen caught in the act could be killed with impunity.134 Now, as it happened, this pedestrian was in fact— and exceptionally—a nobleman, but given that other travellers could not know of his exceptional status, Oedipus incurred an obligation to play by the rules. So his refusal to get out of the way was the primum malum and a deliberate act of provocation. However, in spite of their entitlement to initiate physical violence, Laius and his entourage did not do so: by his own account, it was Oedipus himself who did that, striking the driver (807).135 And, even if we ignore the blow to the driver, we should remind ourselves that when two parties engage in a quarrel that descends to physical violence, he who strikes first is not necessarily the true initiator of violence.136 Melville’s Billy Budd is a good illustration of this point; it is also understood by Goethe’s Duke Alfonso of Ferrara when Tasso draws his sword against Antonio;137 and Euripides’ Medea tries to avail herself of the same point against Jason (1366, 1372). In fact Oedipus did strike first, in a physical sense, when he lashed out at the driver, but he also originated violence in a deeper sense by blocking the entourage’s passage. Laius now retaliated, as he had by this stage more than good cause to do, hitting Oedipus on the head with his goad. This response was mild by comparison with what Laius and his retainers would have been entitled to mete out: at this point in the proceedings they would have been justified in inferring nefarious motives on the stranger’s part, and killing him. Unfortunately for them, however, they did not do so, and Oedipus replied with disproportionate (810) and illegitimate force, in his rage killing Laius and all but one of his servants. It is a measure of his violence and excess that he caused these deaths not with a sword or spear, but with a staff. Oedipus in fact committed multiple murders. And the crucial point is that, as Flaig puts it, ‘even if he did not wish deliberately to kill Laius as his father, he nevertheless did deliberately kill the man who coincidentally was his father’ (1998, 101–2). Michael Lurje’s interpretation of the play hinges on the thesis that the kind of double hamartia I have outlined is impossible: given that the

46  Tragedy and Moral Redress act of parricide was a hamartēma, the killing could not have been an adikēma, because we have to do with one and the same act, and the former description trumps and excludes the latter.138 But an Aristotelian will not be impressed by that reasoning: qua killing of his father Oedipus’ act was a hamartēma, qua killing of a man it was an adikēma; these two aspects simply coexist in balance and there is no question of one’s trumping or excluding the other. Indeed Aristotle appears to imply almost exactly this, with Oedipus in mind, when he remarks, in the context of a distinction between voluntary and involuntary acts of striking, that ‘the person struck may be the agent’s father, though the agent knows only that he is a human being or one of the people present, and is unaware that he is his father’.139 The so-called ‘mixed actions’ of NE III, 1 provide another model of how a single act can, in Aristotle’s view, partake of both the voluntary and the involuntary: the captain who throws his cargo overboard in a storm acts involuntarily simpliciter, since he would not choose to be in the circumstances in which he finds himself; but relative to those circumstances his action is voluntary, since its moving principle lies in him and the particular action of jettisoning the cargo is not externally imposed. Such relativization of the correct characterization of a situation to the way it is viewed is staple Aristotelian fare. After the revelations of Oedipus’ past, the chorus states (1213) that ‘time the all-seeing hath found thee out in thy despite’ (akonta; tr. Jebb). Oedipus is characterized here as akōn (unintending, unwilling) because, although he pursued the investigation hekōn (intentionally, willingly), he did not intend the outcome, of which he was ignorant, and the outcome caused him grief and regret.140 As comparison with Aristotle’s discussion of the voluntary and involuntary shows, Oedipus’ is a perfectly straightforward case of mixed akousion and hekousion.

6 Culpability, Tragic Heroes, and Poetic Justice Dodds remarks that ‘the theory that the tragic hero must have a grave moral flaw, and its mistaken ascription to Aristotle, has had a long and disastrous history. It was gratifying to Victorian critics, since it appeared to fit certain plays of Shakespeare’ (1973, 67). I have argued that an Aristotelian hamartia, though not identical with a moral flaw, does issue from one. Moreover, the theory does fit certain plays of Shakespeare: as Bradley noted (1991, 28–31, 37, 45), Shakespeare’s tragic heroes contribute to their own downfall, and they do so by performing (or omitting) actions of a fateful nature that express their character, so that if the action (or inaction) is flawed, that is because the underlying character is flawed. James Shapiro reads Julius Caesar as Hegel read the Antigone: ‘Shakespeare didn’t conceive of his tragedy in Aristotelian terms—that is, as a tragedy of the fall of a flawed great man—but rather as a collision of deeply held and irreconcilable principles, embodied in characters

Oedipus 47 who are destroyed when these principles collide’ (2005, 147). But Brutus is an Aristotelian hero, and his repeated misjudgements—his cognitive ­failures—are there to be seen on the surface of the work.141 Butcher thought that Sophocles’ Oedipus and Shakespeare’s Othello ‘differ [widely] in moral guilt’ (1951, 322), but it is hard to see how this can be so, for they share the same basic flaw: culpable cognitive failure. In Oedipus’ case the flaw takes the form of an inability to make proper use of information at his disposal; in Othello’s case it takes the form of suggestibility and gullibility. Othello forgets that testimony is not cognitively equivalent to perception:142 for in the case of testimony, unlike that of perception, there is an interface, an intermediary in the form of a human decision to act as a conduit of information (or misinformation), and this fact requires any sensible recipient to ask after the testimony giver’s credentials and reliability. Even perception, though it does not involve an interface, can be unreliable, and misinterpreted, as Othello illustrates. (A similar point applies to Voltaire’s Zaïre, which is partly based on Othello: Orosmane misconstrues the letter to Zaïre that his officers intercept.) That tragedians build flaws into their heroes is what you would expect; for, whether consciously or not, they prefer to conform to Aristotle’s dictum that narrated events should arise by probability and necessity, as well as to a principle of sufficient reason along Leibnizian lines. And, as we have said (§2), the most satisfying explanation of a tragic hero’s fall is that he brought it on himself. Of the six tragic heroes that Aristotle lists in Poetics 13 as being the most effective—Alcmaeon, Oedipus, Orestes, Meleager, Thyestes, and Telephus—we know from elsewhere in his œuvre that Aristotle thought at least two of them to be culpable, Alcmaeon and Orestes.143 The presence in this list of Orestes, in particular, has always been a thorn in the side of those critics who take Aristotelian hamartia to be, or involve, a mere ‘mistake of fact’—that is, for these critics, an essentially non-moral kind of ignorance.144 (On the wider, realist–cognitivist view of morality that I, like Aristotle, am following, there is no harm in saying that a hamartia involves ignorance of matters of fact, so long as one remembers that the relevant facts may be moral in nature.) As for Thyestes, it is unclear which version of the myth Aristotle is thinking of, but it could be the famous story of Thyestes’ seducing his brother Atreus’ wife Aerope, ‘for which Atreus revenged himself by pretending to forgive Thyestes and then serving him up his children’s flesh at a banquet’, as D. W. Lucas reminds us (1972, 145). But if that is what Aristotle has in mind, it is not true, as Lucas goes on to say, that the only instance of hamartia in the offing is the fact that Thyestes did not know what he was eating, for that interpretation depends on an exclusively non-moral conception of hamartia. In the Laws, Plato refers to the tragedies of Oedipus and Thyestes as involving just punishments for serious crimes, and there is a neat match between Thyestes and the figure in the Republic’s Myth of Er

48  Tragedy and Moral Redress who greedily chooses the lot of the tyrant, only to find that he is destined to eat his own children. Destined—but when he understands his fate he bewails it, forgetting ‘that his misfortunes were his own fault, blaming fate and heaven and anything but himself’.145 So four of the six cases listed by Aristotle—Alcmaeon, Oedipus, Orestes, and Thyestes—seem to involve moral hamartiai and a hero who brings trouble on himself. The final two cases, Meleager and Telephus, may follow suit, though we do not know enough about the versions of their stories that Aristotle presupposed to be sure of this. The pattern of culpable error balanced by punishment is exemplified by the majority, and the most important, of Greek tragic heroes and heroines, from Homer’s Hector, Aeschylus’ Xerxes, Eteocles, Agamemnon, Clytemnestra, and the Prometheus of the Prometheus Vinctus, through Sophocles’ Antigone and Creon (theme of my next chapter), his Ajax, Heracles, and Deianeira, to Euripides’ Medea, Hippolytus, Phaedra, and Pentheus, and the post-Euripidean Rhesus.146 Even Euripides’ mad Heracles, who in the view of some commentators fails to conform to this pattern,147 can be fitted into it without strain, as the early Wilamowitz saw (though he later seems to have changed his mind)148—Seneca’s mad Hercules still more so.149 Similarly, in the Senecan Hercules on Oeta we find a heightening of Deianeira’s guilt: not only does she initially wish to kill Hercules in revenge for his infidelity, but when she moves instead to the application of Nessus’ supposed aphrodisiac, her culpable negligence in not suspecting treachery on the centaur’s part is made even plainer than it is in Sophocles (see 716–21); though both the Nurse and Hyllas try to absolve her of guilt (884–6, 900–1, 982–3), she refuses to absolve herself. And Seneca’s Thyestes, Agamemnon, and Medea are obvious exemplars of the type,150 as are Virgil’s Dido and Lucan’s Pompey. Arguably Virgil’s Turnus also exemplifies the type; possibly even, in some measure, his Aeneas.151 After Seneca, we find the syndrome replicated throughout the entire European tragic tradition, as for instance by the Roland of the Chanson, Siegfried, Hagen, and Kriemhild in the Nibelungenlied, Chaucer’s Troilus and Criseyde and Henryson’s Cresseid, Aretino’s Celia (Orazio), Buchanan’s Jephtha, Norton and Sackville’s Gorboduc, Hughes’s Arthur and Mordred (The Misfortunes of Arthur), Wilmot’s Tancred, Gismund, and Palurin, Garnier’s Thésée, Alabaster’s Oromasdes (Roxana), Locrine in the play of that name, Marlowe’s Faustus, Shakespeare’s Hamlet, Jonson’s Sejanus, Webster’s Duchess of Malfi, Beaumont and Fletcher’s Evadne and Amintor (The Maid’s Tragedy), Middleton and Rowley’s Beatrice (The Changeling), Milton’s Samson, Corneille’s Suréna, Racine’s Phèdre, Voltaire’s Tancrède, Lessing’s Emilia Galotti, Goethe’s Götz von Berlichingen, Schiller’s Wallenstein, Kleist’s Penthesilea, Hebbel’s Agnes Bernauer, Büchner’s Danton, Grillparzer’s Rahel, George Eliot’s Lydgate, Hardy’s Michael Henchard, Tolstoy’s Anna Karenina, Flaubert’s Madame Bovary, Ibsen’s Hjalmar (The Wild Duck), Strindberg’s Miss Julie, and Synge’s Deirdre of the Sorrows. This

Oedipus 49 list does not aim at anything like completeness, but is simply meant to indicate the range of applicability of the hamartia doctrine. (For example, the model applies not just to Hamlet, but to all of Shakespeare’s tragic heroes.) I have already briefly mentioned Faustus and Samson, and shall say more about a number of other cases in the above list in due course, especially some of the more controversial ones. Dodds traces the idea he is attacking to ‘older nonsense about “poetic justice” ’, an idea which he says is ‘completely foreign to Aristotle and to the practice of the Greek dramatists’.152 The idea of poetic justice is the idea that rewards and punishments are distributed to the good and bad respectively not as a consequence of a procedure—that would be legal justice—but by chance: the charitable woman wins the lottery; the felon falls off a cliff. So ‘poetic justice’ is a technical term with a precise meaning: a case of justice in a poem is not eo ipso a case of poetic justice. Now the notion of poetic justice is arguably not foreign to Aristotle, given that the concept of the philanthrōpon—that which satisfies human feeling, an idea which is very close to the idea of poetic justice—is aired in Poetics 13, and seems to be almost as important to him as the concepts of pity and fear.153 Still, whatever we should say about the practice of the Greek dramatists in general, I agree with Dodds, and have in effect argued here, that poetic justice plays no role in the Oedipus Rex: the faults that Oedipus discloses during the dramatic action of that work are not there to justify his punishment in any poetic sense, but to reveal his character—a character that at an earlier stage in the story did produce actions that were punishable, and are now to be punished. Stinton suggests that a thesis which I argued for above, namely that hamartia serves to measure the moral distance between the intermediate tragic hero and the fully virtuous person, is ‘a fatal step, if it leads (as it has sometimes led) to crediting—or debiting—Aristotle with the notion of poetic justice’ (1990, 165). But there is no need for this undesirable consequence to ensue. It is true that the hero’s hamartia in a given tragedy could be in the service of poetic justice, because there might be no causal connection between that hamartia and the hero’s punishment: the hamartia could be there simply to make the audience feel comfortable with that punishment in a non-legalistic way. But that is not, I suggest, either Aristotle’s recommendation or Sophocles’ policy in the Oedipus Rex or what we find in the tragic tradition generally. As I have argued, dramatic economy favours an appropriate causal connection between the hero’s moral deficiencies and catastrophe, and that is what we find in the case of Oedipus and most tragic hero(in)es. On the one hand hamartia expresses character; on the other hand it precipitates punishment.

7 Oedipus and Freedom Is Oedipus free? Dodds’s position on this question is straightforwardly inconsistent.154 In the first section of ‘On Misunderstanding Oedipus

50  Tragedy and Moral Redress Rex’, where he argues against the moralizing interpretation, Dodds claims that Oedipus was not free to act otherwise in slaying his father and marrying his mother, because that was what the oracle had unconditionally predicted that he would do. But in rebutting the idea that the play is a ‘tragedy of destiny’, in the second section of his essay, he proceeds to urge the precise opposite, namely that Oedipus was free to act otherwise, because, as it is now conceded, divine foreknowledge does not amount to divine determination. Dodds’s second position is much better than his first. For the suggestion that since the oracle was unconditional Oedipus could not have avoided his fate involves a modal error. From the fact that Oedipus will do such and such—that is what the oracle predicts—it does not follow that he must do such and such. The divine being can predict human free action without undermining that freedom, because it has factored in to its prediction not only the results of the human agent’s free deliberation, but also the effect of its making the prediction and of its revealing the prediction to that human agent in advance, if it does. (Similarly, Christ’s prediction to Peter that he would deny him three times takes account of the fact that Peter hears the prediction; when the time comes, Peter’s denials are free.)155 Hence Oedipus can be free not to do what he is predicted—and predicted with absolute certainty—to do. For if he were not to do what he is predicted to do, then the oracle would not have made that prediction. The oracle merely knows that Oedipus will do such and such, not that he must do it, and knows also that divulging the prediction to the human agent will not affect—indeed will exactly contribute to the bringing about of—that outcome: in mathematical terms the prediction of that action is a fixed point. As Boethius noted, God’s seeing that I will do something no more detracts from my freedom than my seeing you doing something now detracts from yours.156 And it would be a species of category error to suggest here that Oedipus could not have acted otherwise because the myth was settled and known and Sophocles could not change it: the myth may have been fixed and familiar to all, but in that fixed myth Oedipus acted freely; it is a fixed story about freedom.157 It might be objected to my argument of the previous paragraph that it ignores an important historical and an equally important linguistic point. The historical point would be that, even if it is a modal error to confuse ‘will’ and ‘must’, it is an error that many people have committed, including most famously Aristotle. In the ninth chapter of De Interpretatione, Aristotle—at least according to the traditional interpretation of this ­passage158—feels obliged to restrict, with respect to statements about future contingencies, the principle that every statement is either true or false, on the grounds that if statements about the future were true (false), the corresponding events would thereby be rendered necessary (impossible). So, it might be said, if Aristotle can infer a ‘must’ from a ‘will’, it would hardly be surprising if Sophocles did so too. The linguistic point

Oedipus 51 is this. It is sometimes claimed that the oracle that Oedipus receives from Apollo is expressed in ‘must’ terms.159 Oedipus tells Jocasta that Apollo prophesied terrible things for him, ‘even that I was fated (chreiē) to defile my mother’s bed; and that I should show (dēlōsoim’) unto men a brood which they could not endure to behold; and that I should be (esoimēn) the slayer of the sire who begat me’ (791–3, tr. Jebb). Jebb translates the crucial impersonal verb ‘chrē’ (here occurring in the optative form ‘chreiē’) as ‘is fated’, which is a standard rendering (Lloyd-Jones has ‘was destined’).160 Note that Sophocles also uses future tenses (dēlōsoim’, esoimēn) for what Apollo says Oedipus is ‘fated’ to do. Now ‘chrē’ has a range of meanings, including one that is close to a pure future tense, often translated ‘is to be’,161 which is also a standard meaning of the verb ‘mellein’.162 In this sense fate simply represents a jumping up of the actual to the necessary. As A. F. Garvie writes on Aeschylus, Persae 908–17: ‘In his despair Xerxes makes no attempt to understand why everything has gone so wrong. It has happened, so it must have been fated to happen’. Hence the meaning of Oedipus Rex 791 might be simply ‘that I was to. . .’ rather than the modally weightier ‘that I was fated/obliged to. . .’. However, it is clear from Aristotle’s discussion of future contingency alone, not to mention many other passages, that for Greek (as indeed for modern) thought the idea of what will happen is often felt to be close to the idea of what must happen. That is especially so in epistemic contexts. If you know what will happen, it is natural to think that that can only be so because what will happen also must happen: for how else could you know? Someone who knows the future has, plausibly, tapped into a necessary process and extrapolated the outcome. Seriously entertaining the idea that the future might be metaphysically determinate, and even foreknown, but not necessary was largely an achievement of late antique and medieval thought.163 So it would be reasonable for my objector to insist that the future tenses in the quoted passage, and the use of ‘chrē’, are at least not clearly distinguished from the modal ‘must’. There is a natural transition from the idea of what will be to the idea of what is fated to be.164 My response to this historico-linguistic objection is that, even if Sophocles thinks that Oedipus, in some sense, must do what he does (these being the terms, according to the objection, in which Apollo issues the prophecy), he is nevertheless a fully free agent.165 I here presuppose compatibilism about freedom, and I rely on the simple fact that Oedipus is described as acting like a free agent. In the story he makes choices, and is under no duress when he makes them; for a compatibilist about freedom that is good enough to ensure that the actions really are free, provided they also meet a rationality constraint. This is so regardless of whatever subterranean necessities (physical or theological) may be operative. At the surface level, the level of actions and their typical characterizations, of reasons for action and agent rationality, of presence or absence of duress, and so on—and it is on that surface where, according

52  Tragedy and Moral Redress to the compatibilist, freedom exists—Oedipus does not have to do what he does. To the extent that we are inclined to say that he must do what he does, that simply reflects an ex post facto inference from the factual to the fated: he did it, so it ‘was to be’.166 Vickers distinguishes between Oedipus’ actions during the play and his critical actions beforehand, arguing that the former are free and the latter not (1973, 498–500); but I see no relevant distinction. The fact that Oedipus killed Laius and married Jocasta in ignorance of their relations to himself does not derogate from the freedom of those acts. Vickers suggests that ‘freedom means—at the very least—that I am able to act and control my actions with full knowledge of who I am’ (ibid., 499). But this sets much too high a standard for freedom. We never have that degree of knowledge: it cannot be the case that if, say, I unearth some old family documents which reveal that I have hitherto been in error about my paternity, I must judge that my actions to date have not been free. Tragedies always involve contingency—in particular free action, and often sheer coincidence as well. Relatedly, as Bradley reminded us in one of his most memorable passages, events have a habit of turning out otherwise than intended.167 True, the ingredients of catastrophe are generally all there from the beginning—the flawed characters, the latent deceits and falsehoods waiting to be exposed, the ambient hatreds, and so on. But a priori it is not a given that those ingredients will combine in just the right way for tragedy; things might just as well proceed in the same old uneventful way. For every actual tragic event, there are many such events that could occur but do not, because the necessary conditions are not assembled at the right time, in the right place, or in the right order. How easily might Hamlet have lived to old age happily married to Ophelia, as George Eliot once mused (1979, 514). So I do not think we should agree with Stanley Cavell when he writes (2003, 112): Of course if Othello had not met Iago, if Lear had not developed his plan of division, if Macbeth had not listened to his wife. . . . But could these contingencies have been prevented? If one is assured that they could have been, one is forgetting who these characters are. For if, for example, Othello hadn’t met Iago he would have created another, his magnetism would have selected him, and the magic of his union would have inspired him. So a radical necessity haunts every story of tragedy. On the contrary, we may reply, Othello’s tragedy depends on a sequence of contingencies: Iago’s high-wire act could at any moment have crashed to earth. The preconditions for tragedy—depending, as Cavell says, on character, and especially on the lack of mutual understanding between Othello and Desdemona168—were all in place, but it needed a particular and highly contingent combination of circumstances, including crucial

Oedipus 53 free decisions on the part of the main agents, for tragedy to be activated.169 So too with the plot of the Oedipus Rex.170 Euripides’ Cassandra says that if the Greeks had stayed at home, Hector’s virtues would not have shone forth (Trojan Women 395–7); indeed so, and he would also not have been tragic. Horst Steinmetz draws our attention to the bizarre array of coincidences that characterizes Lessing’s Emilia Galotti, and suggests that the play could just as well have ended as a comedy; A. D. Nuttall remarks that Romeo and Juliet moves from comic to tragic form.171 But it is true of all tragedies that in imaginable counterfactual scenarios catastrophe would have been averted: it is true of the Antigone, for instance, that if Creon had elected to release Antigone before burying Polyneices no one would have died; what if Edmund’s final message in King Lear had not arrived too late?172 Conversely, in plays of the ‘catastrophe survived’ genre, the aversion of disaster often depends on pure luck, of which Euripides made such good use in his recognition plays.173 It is important to add here that it does not follow, as Steinmetz also claims, that Emilia Galotti’s reliance on coincidence in any way lessens the guilt of the principal agents or their responsibility for catastrophe. We are responsible for the errors we commit even when (as is invariably the case) other causal factors contribute to the ensuing of undesirable consequences. No one is causally isolated when he or she performs an action. So I think that Cavell is wrong to banish contingency from the tragedies of Othello, Lear, and Macbeth, but what is right in the quoted passage is the emphasis on the importance of character. We need to hold in balance the twin points that tragedy involves contingency, but that protagonists’ fates play out their characters in ways that conform to the Aristotelian requirements of probability and necessity. Bradley saw the matter clearly: ‘The dictum that, with Shakespeare, “character is destiny” is no doubt an exaggeration, and one that may mislead (for many of his tragic personages, if they had not met with peculiar circumstances, would have escaped a tragic end, and might even have lived fairly untroubled lives); but it is the exaggeration of a vital truth’ (1991, 29–30). Free action and the accidental are both species of contingency, but they need to be kept distinct. Aristotle fails properly to include, as a feature of tragedy, contingency in the sense of the accidental, and this omission, as I have just in effect implied, is a mistake on his part, since tragedy is replete with that kind of contingency. No doubt Aristotle suppresses mention of the accidental because he thinks that it would not square with his doctrine that events in a tragedy must follow upon one another according to principles of probability and necessity.174 But what he should have done was hedge that doctrine, not ignore the very evident importance of the accidental in tragic plots. (Aristotle does allow contingency to feature ‘outside the play’, and in its initial conditions, but he excludes it from the tragic action itself.)175 Free action, by contrast, conforms to the doctrine—at least according to anyone who thinks in a

54  Tragedy and Moral Redress compatibilist way about freedom, as it is plausible that Aristotle did.176 And the compatibilist is right to insist that our central cases of freedom are rational, predictable actions—actions that are performed for good and sufficient reasons—not mad, unpredictable ones. If serving God is the most rational thing to do, then his service is indeed perfect freedom, as the Collect has it, provided my decision to enter into that service is not forced on me by others. Tragic plots revolve around rational, predictable, free actions. For tragedy ‘is deeply concerned with decision’:177 it is striking that quotations and discussions of tragic works in antiquity tend to focus on the ethical dilemmas and aporias that they present,178 and it would be hard to think of a single tragic work in the tradition that does not involve at least one structurally important decision somewhere in its plot. Even tragedies of inaction, such as (perhaps) some of Chekhov’s or Beckett’s plays, fall under this rubric: for a decision to do nothing is still a decision. A decision, such as Prometheus’, to stick to one’s resolve is still a decision.179 Free action is action that, like all contingency, ‘could have gone otherwise’, as we often put it. Aeschylus’ Agamemnon donned the strap that tied him to the yoke of necessity (218), but he did so freely. Aeschylus’ language at the crucial point in the work is focalized, reflecting Agamemnon’s own feelings of compulsion: he simply cannot bear the alternative to sacrificing his daughter, namely desertion of the fleet. But the words that express his dilemma (205–27) make clear that he has a real choice.180 Pelasgus too, in the Suppliants, feels that he is acting under necessity (440, 478), but his decisions are clearly free. Again, both Agamemnon and Pelasgus are described as acting freely: it is not that Agamemnon could not desert the expedition; it is not that Pelasgus could not ignore the Danaids’ threat; it is just that these alternatives seem comparatively unattractive.181 Similarly, in saying that Oedipus did what he did freely, I imply that he had a genuine choice in the matter, that he could have acted otherwise—at all relevant stages of the story, both before he received Apollo’s oracle and afterwards—than the way he did act.182 W. H. Auden once opined that ‘Greek tragedy is the tragedy of necessity: i.e., the feeling aroused in the spectator is “What a pity it had to be this way”; Christian tragedy is the tragedy of possibility, “What a pity it was this way when it might have been otherwise” ’;183 but Auden’s motto for Christian tragedy also applies to Greek, and indeed all, tragedy. Of course, as Kierkegaard remarked (1987, vol. 1, 143), free action is not performed in a vacuum, but in a context of substantial constraints. But the presence of such a context does not derogate from freedom: quite the reverse. Free action is essentially rational, as we have said,184 and rational action is constituted as such by being situated in a context that gives it identity, that makes demands and supplies reasons. Action in a social and rational vacuum would not be free, but mad (or would not be action at

Oedipus 55 all). A good illustration of this point is provided by Ellida’s decision at the end of Ibsen’s The Lady from the Sea. Ellida cannot resolve her impasse until she is permitted to decide in full freedom. But the freedom of her decision consists principally in the absence of duress from her husband: it does not require—and could indeed not be made in—a context devoid of operative causes and reasons. The freedom of the tragic hero or heroine is not constituted merely by the capacity to take a reflective attitude, perhaps as envisaged by the early Wittgenstein, to his or her situation;185 it is constituted by an ability to make a difference to the world, given genuine alternatives.186 Nothing in the compatibilist story about freedom requires us to deny this evident truth. There is an important logical point that needs to be mentioned here. Someone might object to my argument as follows: at any given moment, the past in respect of that moment is unalterable, so a past oracle and its content are fixed at that moment; if in addition we assume that Apollo’s oracle, say, made an infallible prediction concerning Oedipus’ future actions, then, when the time comes for him to act, how can Oedipus do other than what he does? The necessity of the past, along with the necessity of the connection between prediction and predicted outcome, import the necessity of that predicted outcome, by dint of a familiar transfer-ofnecessity principle, common to all modal systems: if the antecedent of an implication is necessary, and the implication is itself necessary, then the consequent is necessary, too. (The transfer principle was denied by Luis de Molina, founder of the important doctrine of middle knowledge; but few have followed him in this.)187 In order to escape from this bind, it looks as though we must either tolerate backwards causation, or give up the infallibility of the oracle, or both. That is, it seems that we must say either that Oedipus later has the power to bring it about that Apollo did not make the earlier prediction, or that he has the power to bring it about that the earlier prediction, though fixed, was in error, or both. The first option involves tampering with the necessity of the past, something that most thinkers have been unwilling to do ever since the tragedian Agathon, as Aristotle reports him, declared that the one thing even god cannot do is to reverse what has happened (NE 1139b9–11).188 What about the second option? In the context of medieval Christian debates on these matters, thinkers were understandably reluctant to countenance the idea of a deus fallens, of a god who makes predictions that turn out to be false.189 Whether pagan gods can make false predictions is a moot point. Plato preserves a beautiful fragment of Aeschylus in which Thetis accuses Apollo of falsely prophesying long life for her future son Achilles.190 But, as Robert Parker points out, it may be that ‘Thetis had misunderstood the god’s riddling words (in which, certainly, she deserves all our sympathy)’ (1999, 21). Thetis construed his words in one way, and understood in that way they turned out to be false, but there was an

56  Tragedy and Moral Redress alternative interpretation, perhaps, which rendered them true. As Chaucer’s Criseyde puts it, For goddes speken in amphibologies, And for o soth they tellen twenty lyes.191 That is certainly one way of saving Apollo’s credibility in the face of Thetis’ accusation: his words were ambiguous. But there is another way. The Christian God, being omniscient, must be presumed to know the meaning of his own predictions. But in a pagan context the parallel assumption that divine oracles know the meaning of their own predictions generates a crucial subtlety that eases the logical bind I mentioned in the last paragraph. For although pagan oracles know in general terms what their predictions mean, they do not always know their precise meaning; and, by contrast with the Christian God, pagan oracles often do not know in detail how their predictions will be realized. Parker again notes that ‘there are strict limits to what is actually revealed even through the especially authoritative medium of prophecy. You will kill your father and marry your mother, Oedipus is warned; neither he nor we learn why’ (1999, 14). Nor—and this is the point here—do we learn how. Indeed it is usual with ancient oracles that the manner and circumstances of their fulfilment are left unstated. For one thing, it will normally be indeterminate when the oracle is to be realized:192 so Darius, in Aeschylus’ Persae, pins his hopes on the possibility that the prophesied disaster may lie in the distant future (739–41). More importantly in the present context, oracular predictions may leave open the possibility of a metaphorical implementation.193 No doubt it is fixed that Oedipus will in some sense kill his father and sleep with his mother, but in what sense will he do these things? ‘Father’ and ‘mother’ often carry the metaphorical meaning of ‘homeland’;194 so killing one’s father and sleeping with one’s mother might in the event amount to no more than (say) plunging a sword or spear into the earth of one’s native land and then lying down to rest on one’s native soil.195 Rutherford remarks, in line with Parker’s suggestion above about Aeschylus’ Thetis: ‘Those who have heard the god’s pronouncement often misunderstand it, misremember it, or hope that it will remain unfulfilled. But in tragedy oracular predictions are always fulfilled’.196 True, but there is another possibility, which is that the oracle is indeterminate in meaning, so that although tragic agents know, or should know, that oracular predictions are always fulfilled, they often do not know, and sometimes cannot know—because it cannot be known; it is simply not yet fixed—in what sense the oracle will be fulfilled. It is not merely that the recipients cannot imagine how the oracle will be fulfilled:197 there may be literally nothing there, as yet, to imagine. Until it is fulfilled, it may just be indeterminate what the meaning of the prediction

Oedipus 57 was. This indeterminacy is more radical than the standard indeterminacy of the Croesus kind of oracle.198 We can go further. Oedipus could have decided to perform the actions I have mentioned precisely in order to defuse the oracle, by making it come true harmlessly. When the harpy Celaeno predicts to Virgil’s Aeneas and his men that they will end up devouring the very tables in their hunger, it is fixed, we might say, that they will indeed eat their tables, in some sense, but not fixed in what sense they will do so. And in fact the prediction is fulfilled in the most banal way, when Ascanius interprets the bannocks they eat on reaching Italy as their ‘tables’.199 Ascanius thereby lays the curse to rest: it would be absurd for Aeneas (or the reader) to worry that the Trojans might still, even after eating their tables metaphorically, have to consume them in a literal sense, on the grounds that that was what Celaeno meant. Celaeno did not mean anything so definite: the metaphorical realization of the prediction is quite sufficient to satisfy the metaphysical demands of the poem and its oracles. Admittedly Celaeno did mean to predict something unpleasant for Aeneas and his men, and in that she turns out to be wrong; but that is a mere matter of detail, and classical—by contrast with Christian—theology has no difficulty with it. So we might say that the classical position is that a certain amount of anti-realism about past meaning is allowed: that is, while in general terms the content of an oracle is fixed at the time of its pronouncement, its precise meaning remains to be determined by subsequent events and by the relevant agent(s). (There is all the difference in the world between literally and metaphorically eating tables: that is matter of meaning.) Hence, having received the Delphic oracle, it remained open to Oedipus to undertake not to kill his father and marry his mother in the literal sense; he might have avoided these actions, but if he had done so then one day it would no doubt have become clear to him that there was a metaphorical sense in which he had fulfilled the prediction, perhaps quite harmlessly. And, as I have said, he might have deliberately undertaken to realize the prediction innocuously, in order to defuse it. That deflecting manoeuvre, had Oedipus engaged in it, might to be sure have been unsuccessful; then again, it might have worked. Suppose, however, that my objector refuses all this, and insists on both the necessity of the past and (implausibly, I suggest) the determinate meaning of Apollo’s oracle; in that case, and modulo those assumptions, the compatibilist about freedom must agree that, in a metaphysical sense, Oedipus’ foretold actions are necessary. But then the compatibilist’s whole point is that such underlying necessities, whether physical, theological, or (as we have just been exploring) temporal, are irrelevant to real freedom. Real freedom is social or political: it involves having the power to do as one chooses, provided that one’s choice meets a rationality constraint, and having the power to act otherwise than one does in the

58  Tragedy and Moral Redress sense that one is subject to no duress from any agent, animate or inanimate, forcing one to do what one chooses to do. Hence, even if we think that there is a sense in which Oedipus must do what he does—on the grounds that Sophocles, like Aristotle, thinks that what will happen must happen, and because it is true in advance, and Apollo knows, that Oedipus will (literally) kill his father and marry his mother—these facts still do not derogate from Oedipus’ (real) freedom. If Oedipus must do what he does in a metaphysical sense, then that, as we have said, is no more than the reflection of an ex post facto inference from the factual to the necessary. It matters, for freedom, how this necessity comes about: if it comes about merely because that is what Oedipus will choose to do, there is no conflict with freedom, because his actions may still be performed freely in the socio-political sense. The future will be that particular way—and so must be that way, if we accept the modal inference—because Oedipus will freely perform such and such actions. By contrast, he would not be free if he acted under duress. But acting within the causal order—this was one of Hume’s great insights—does not as such put agents under duress: causes do not enforce their effects; they merely precede them in a lawlike manner. Force, duress, compulsion, constraint, to the extent that these things deprive one of freedom, necessarily involve human or non-human agency; they are not, just as such, what causes do to effects. The same goes for predictions: predictions do not enforce the predicted outcomes. Hence, even if we think that Oedipus acts under a kind of temporal necessity, because he cannot change the past and Apollo’s prediction is assumed to be both infallible and determinate in meaning, so that by dint of the transfer-of-necessity principle Oedipus must now do what he was predicted to do, still he is free in the genuine sense of the word, and it was only because Apollo foresaw what Oedipus would freely (in the genuine sense) do that he was able to make the prediction in the first place.200 Apollo does not exercise any kind of duress on Oedipus; he does not actively bring about Oedipus’ fall. As numerous commentators have observed, he merely foretells it.201 However, the view that Apollo not only foresees but also determines Oedipus’ actions and fate persists;202 it is implied in the common statement that Oedipus is the ‘plaything’ of the gods.203 Klaus Nickau writes: ‘Der Gott Apollon stellt die Aufgabe, der Mensch Ödipus löst sie’ (‘The god Apollo sets the problem, the man Oedipus solves it’: 1994, 10). It is true that in the Aeschylean and Euripidean versions of the Laius story, Apollo does set a task or problem, since there his prophecy is conditional, so that Laius has to decide whether to refrain from having a son, or proceed and take the consequences.204 But in Sophocles’ Oedipus Rex, Apollo does not figure in that way. And in any case, as we have said, it is plausible that the meaning of Apollo’s oracle is not so fixed that Oedipus cannot fulfil it harmlessly.

Oedipus 59

8 Operatur deus in unoquoque secundum eius proprietatem I have argued that Oedipus is free in his crucial actions. In fact, even if Sophocles had represented Apollo as determining the acts of parricide and incest—even if you think that that is how Sophocles does represent things—that would still not detract from Oedipus’ freedom. And the same goes, mutatis mutandis, for other tragic heroes. This is so because of a fundamental principle governing divine intervention in traditional literature, namely that operatur deus in unoquoque secundum eius proprietatem: the god acts in each person according to that person’s character.205 Labels such as ‘overdetermination’ (Dodds), or ‘the “double determination” of cause’ (Willink), or ‘the unity of two aspects’ (Lesky),206 do not accurately register the true nature of the explanatory connection between divine prompting and human response. These accounts do not reflect the fact that there is an explanatory asymmetry between human and divine planes. The explanatory relation runs from former to latter, and not from latter to former: that is, it is the human plane that explains the divine, and not vice versa. Lesky criticizes Dodds on the basis that ‘overdetermination’ fails to capture the mutual connectedness of divine and human planes (1961, 29–30, 44), but it seems to me that he in turn misses the fact that this mutual connectedness is asymmetric—that the human plane enjoys explanatory priority, so that in ‘the unity of the two aspects’ the two aspects do not contribute equally or independently. This asymmetry is also missed by Jean-Pierre Vernant in his discussion of the Heraclitean dictum ‘ēthos anthrōpōi daimōn’.207 This aphorism is an identity statement, and identity statements can be read in two different directions, with correspondingly different emphases. Is Heraclitus saying ‘You might have thought that a human being’s character, ēthos anthrōpōi, was something natural, but actually it is daimōn, which is (as we know) supernatural’? Or is he saying, contrariwise, ‘You might have thought that daimōn was something supernatural, but actually it is just a human being’s character, which is (as we know) natural’? In other words, is he reducing the natural to the supernatural or vice versa? The former option seems to me unlikely, since it is the natural that is given, the onus being on the supernatural to earn its metaphysical keep; but, whichever way Heraclitus himself intended his aphorism, it is only when it is construed in the latter way that it speaks truly about the role of the supernatural in classical and classicizing tragic literature. For in this genre it is the natural that is in the driving seat, the supernatural that tags along behind. Lesky discusses (1961, 43–4) the Pandaros episode in Homer’s Iliad (IV, 73–148), in which Athene prompts Pandaros to break the truce that has been established between Trojans and Greeks. Why does Athene approach Pandaros in particular? She does so because he has the right combination of

60  Tragedy and Moral Redress ambition, bravery, greed, and folly to give ear to temptation.208 As we learn later, Pandaros is keen to prove himself as an archer, having ignored his father’s advice to take chariot and horses to Troy (V, 180–216). Similarly, Virgil’s Allecto appears to Turnus in order to rouse him to violence against the Trojan interlopers precisely because he is the kind of man who will listen to her.209 He initially spurns her approaches, then knuckles under; we have a representation of Turnus’ inherent violence overcoming temporary resistance, as also in Seneca’s Thyestes when Tantalus succumbs to the Fury.210 When, in the same play, Atreus devises his revenge, he repeats the word ‘rapior’ (261–2)—‘I am snatched away, borne along’—but ‘his vagueness about the force that is possessing him is appropriate, since it ultimately proceeds from within him’.211 Downing’s comment on Lord Monchensey in T. S. Eliot’s The Family Reunion (1969, 346)— We most of us seem to live according to circumstance, But with people like him, there’s something inside them That accounts for what happens to them— encapsulates the fate of almost all tragic heroes and heroines. Aristotle has often been criticized for marginalizing the role of the supernatural in his discussion of tragedy in the Poetics, but in one good sense he was right: it is the natural that does all the work.212 Stephen Halliwell’s view is that Aristotle omits the gods because their involvement in tragedy is, from a human point of view, fortuitous (governed by tuchē), so that mentioning them would upset his emphasis on probability and necessity in tragedy.213 On my approach, he incurs praise rather than blame: Aristotle omits the gods because he thinks that the supernatural is, in the deep structure of tragedy, irrelevant, and he is right about that. ‘Supernatural’ causation is, au fond, natural. As Darius puts it in Aeschylus’ Persae, ‘when a man is in a hurry himself, the god will lend him a hand’ (742; tr. Sommerstein): that captures the ex post facto status of divine motivation of human action.214 Likewise, too, with the positive version of the same idea: God helps those who help themselves.215 In the messenger speech of the Persae, and in the words of its main agents, the idea of divine causation is repeatedly invoked,216 but the narrated and portrayed events are nevertheless given purely naturalistic explanations, and it is made clear that the calamities of Salamis and Psyttaleia result from Xerxes’ own culpable misjudgements.217 The point of divine intervention in traditional literature is not to undermine the autonomy and sufficiency of human motivation but to represent it in a particular way—in fact to enlarge and dignify it,218 rather in the manner of the sublime moment in Bellini’s Norma when the heroine senses the presence of the druid god Irminsul (II, 7): A mirare il trionfo de’ figli ecco il Dio sovra un raggio di sol, s’un raggio di sol.

Oedipus 61 (‘See the god marvelling at his sons’ triumph, upon a ray of sunshine’: 2013, 160.) But Irminsul takes no part in the action, which is motivated entirely naturalistically. What the divine says to the mortal, the mortal has already said, or is about to say, to itself. From Marlowe to Calderón to Goethe to Berlioz to Gounod to Busoni to Mann, the devil appears to Faust because he is primed, or has even expressly abjured heaven and adjured hell, as in Marlowe’s version. It is true that Marlowe’s Mephostophilis arrogates the initiative to himself (B-text: V, 2, 96–9): ’Twas I that, when thou wert i’ the way to heaven, Damned up thy passage. When thou took’st the book To view the Scriptures, then I turned the leaves And led thine eye. Mark Burnett thinks that these lines communicate the author’s view that Faustus ‘is damned all along: he has no freedom of choice’ (2010, 172). But they are part of Mephistophilis’ characterization, and what they convey is just bravado: Faustus is from first to last responsible and free, and the contest between the good and bad angels merely gives external expression to a debate that is going on in Faustus’s soul.219 Like Faustus, Milton’s Eve is ready to listen to Satan’s temptations: there is a sense in which she has already fallen.220 The psychologizing of Aphrodite’s role in Euripides’ Hippolytus is rendered easy by the fact that she speaks the prologue and then takes no further part in the drama: she seems to sink into the action as a psychic force,221 which is the way that the Nurse characterizes her (443–50).222 (Similar remarks apply to the roles of Cupid and Megaera in The Tragedy of Tancred and Gismund by Robert Wilmot and others.) In this respect she contrasts with Dionysus in the Bacchae, who speaks a similar prologue to Aphrodite’s in the Hippolytus, but then does take part in the play, in human form.223 But this difference, though dramatically important, is from our point of view theoretically superficial: it does not signal a deep difference in the extent to which the supernatural can be naturalized. As Dodds remarks in his commentary on the Bacchae: ‘In the maddening of Pentheus, as in the maddening of Heracles . . ., the poet shows us the supernatural attacking the victim’s personality at its weakest point—working upon and through nature, not against it. The god wins because he has an ally in the enemy’s camp: the persecutor is betrayed by what he would persecute—the Dionysiac longing in himself’.224 In the Helen, the contest between Hera and Aphrodite has to be resolved on the day of action, but the prophetess Theonoe makes clear that the decision between them is a matter of whether she decides to tell Theoclymenos of Menelaus’ presence in Egypt, so that the divine level is in effect anthropomorphized.225 Athene drives Sophocles’ Ajax mad, but the madness is readily naturalizable when we recall that it supervened

62  Tragedy and Moral Redress on Ajax’ decision, which he never later regrets, to launch a treacherous attack on the Greek leaders. Ajax’ madness emerges organically from his character and dispositions, in the context of the slight he has received and his impulse to avenge it: ‘the seeds of madness were already in him, evidenced by an abnormal megalomaniac pride’.226 As Gisbert Ter-Nedden notes (2007, 345–6), Athene’s apparent inducing of Ajax’ madness is just a dramatic device to represent his Verblendung, which Athene does not induce, but rather reveals. Athene remarks to Odysseus: ‘Do you see, Odysseus, how great is the power of the gods? What man was found to be more farsighted than this one, or better at doing what the occasion required?’ (118–20; tr. Lloyd-Jones). The play—and indeed the tradition more generally—show these assessments of Ajax’ character to be wrong,227 which suggests a debunking reading of the ‘power of the gods’ too. What we really see in the Ajax is not the power of the gods, but the power of human nature: this is demonstrated in spectacular fashion when Athene fails in her attempts to make Odysseus triumph over Ajax and Ajax sympathize with Odysseus (79, 111). Before his single combat with Paris, Homer’s Menelaus prays to Zeus for aid, and he asks the god to ‘subdue him beneath my hands’ (Iliad III, 352). If Menelaus had killed Paris, he would have thanked Zeus for answering his prayer, but would have taken credit for victory no whit the less.228 As Garvie remarks concerning the divine aid to the Athenians in Aeschylus’ Persae (342), the audience would not feel that this help detracted from their achievement: ‘they would remember the Iliad, in which divine support did not normally diminish the glory which a hero earned by his exploits; if anything, it enhanced it by showing that he was heroic enough to deserve that support’.229 Conversely, when humans blame gods for their actions, they do not think that this lessens their own responsibility, as we learn from the case of Agamemnon’s apology in Iliad XIX.230 Some commentators take Agamemnon’s words at face value,231 but the sense at 137–8, filling out the subtext, is ‘But since I was blinded by atē and Zeus robbed me of my wits I wish to make amends and to give abundant recompense’. (­Earlier—XIX, 86–90—Agamemnon had apparently denied responsibility and placed it on the gods: but there he is fudging the issue and trying to save face.)232 And, in Aeschylus’ Agamemnon, the fact that in sacrificing his daughter he is carrying out the will of Zeus does not absolve Agamemnon from responsibility.233 Ancient defendants were indeed ill-advised to plead divine instigation of their actions, for any such supernatural intervention would naturally be construed as a sign of voluntariness and so of guilt.234 A similar point applies to apparent intrusions on the agent’s autonomy that might otherwise be regarded as diminishing responsibility: if another mortal makes me commit a crime, I may be able to plead duress in extenuation of the offence;235 but if a god ‘makes’ me do it, that renders me all the more responsible for it. The chorus of Aeschylus’ Agamemnon

Oedipus 63 remarks that Cassandra goes to her death like a beast approaching the altar ‘driven by a god’ (1297–8), that is to say, voluntarily. For it was important in sacrifices that the beast should not have to be forced—that is, not forced by the priests. If it is forced ‘by a god’, then it is not forced at all.236 Dodds suggested that, whereas for Plato’s priest, speaking on behalf of Lachesis in the Myth of Er, ‘the responsibility is the chooser’s; the god is exempt from responsibility (theos anaitios)’ (Republic 617e4– 5), Aeschylus might have preferred ‘the responsibility is the chooser’s; the god is responsible for everything (theos panaitios)’ (1973, 56–7). But the two formulations are equivalent. In Matthew Arnold’s ‘Sohrab and Rustum’, which inverts the motif of the son slaying the father, Sohrab blames fate for his death (708–15). But Sohrab’s death is quite clearly the outcome of human decisions and human mistakes. Its fatefulness does not undermine that human causality, but is a way of representing it, of celebrating it, of coming to terms with it. Even in A Midsummer Night’s Dream, where the fairies’ interventions in the human action might seem essential to the resolution of the love plot, Shakespeare implies a naturalistic construal of events. ‘The spell is never taken off Demetrius’ eyes; without its lasting effect the lovetangle could not be resolved’, comments Harold Brooks (on III, 2, 452). Well, but recall Helena’s words (I, 1, 242–3): For ere Demetrius looked on Hermia’s eyne, He hailed down oaths that he was only mine. Perhaps Oberon’s lasting spell on Demetrius can be regarded naturalistically as the removal of an obsession for Hermia that had artificially interfered with his true love for Helena.237 In that case, though in a relatively superficial sense ‘in the Dream, the mortals are unaware how far they are puppets of the gods, and they never realize how their own sight has been tampered with’,238 at a deeper level we have human autonomy. Sophocles’ Oedipus says ‘These things were Apollo’, but he adds ‘None but I struck myself’ (1329–32).239 (Creon in the Antigone expresses himself similarly.)240 Oedipus is not making a distinction between those events for which Apollo was responsible, and a distinct set of events for which he was responsible.241 He takes responsibility for everything: in invoking Apollo, Oedipus elevates, without disavowing, his own part in his demise. There was nothing untypically free about Oedipus’ self-­blinding: the freedom that he exercised when he performed that desperate act was one that had attached to all his actions. Again, when Oedipus says (at 1382–3) that the gods have disclosed his guilt, he immediately adds (1384–5) that he has done so himself. It is true that, once gods are on the scene, storytellers can represent them as full and independent agents, doing things that have no immediate or straightforward allegorical reduct:242 an artist may even take a

64  Tragedy and Moral Redress secular tale and inject a pantheon into it, as Wagner (borrowing from Norse legend) did with the Nibelungenlied. Above we contrasted Euripides’ Aphrodite and his Dionysus, after they have spoken their prologues: the former takes no part in the main action of the Hippolytus, whereas the latter plays a decisive role in the main action of the Bacchae. Or consider the appearance of the Furies in the Eumenides. They first appear to the mad Orestes at the end of the Choephori, but not to the chorus, and in that manifestation they can be readily and immediately psychologized; but they then appear to the Pythia at the beginning of the Eumenides, so must have objective reality, and in the third part of the trilogy they emerge as autonomous agents, and continue to be such after Orestes has recovered sanity and been purified of blood-guilt and acquitted by the Areopagite court.243 Similarly, although the Furies in Seneca’s Agamemnon (359–64) and Medea (958–65) have no objective reality, a Fury appears as a stage figure in the prologue to the Thyestes.244 In Sophocles’ Ajax Athene’s role in driving Ajax mad can be straightforwardly psychologized, as we have said, but in her exchanges with Odysseus and Ajax she figures as an independent agent, no matter how we settle the controversial question of the extent of her visibility to characters and audience.245 Again, as Dodds implied in a passage cited above, the role of Euripides’ Iris and Lyssa in driving Heracles mad can be psychologized in an obvious way: the strain of Heracles’ labours, and particularly of the last one, the harrowing of hell, has turned his wits, so that he has become a serial murderer, and thinks of the killing of his own children as his thirteenth labour; Heracles is a man of violence who needs violent tasks to occupy him;246 against some critics,247 I believe that Wilamowitz was right to discern signs of incipient madness in Heracles before Iris and Lyssa enter.248 But Hera’s two lackeys are hardly on stage before they quarrel, since Lyssa, who is the apotheosis of madness, paradoxically does not want to madden Heracles. So Iris and Lyssa, though to a large extent psychologizable, are also put on stage by the playwright as independent agents. It follows that we need to adopt a split-level approach to naturalizing the divine in traditional literature: some uses of the supernatural are immediately psychologizable; others are not. But it makes sense to treat the former as primary, or basic, and the latter as secondary, or derived, deployments of the supernatural. In the first instance, we might say, the divine in traditional literature does no more than body forth the human in a way that is immediately reducible to human psychology. But once divine agents on the scene, they can be given a life of their own, and can interact with mortal agents in a way which has no immediate allegorical reduct, only a mediate one. Again, once divine agents are on the scene, the whole question of their relation to mortal affairs can be raised. Thus Euripides sometimes permits his characters to discuss explicitly the role of the gods and whether it undermines human freedom: in the Heracles, Theseus and Heracles

Oedipus 65 debate this matter at length; Heracles’ famously rationalistic response to Theseus’ naïve belief in the tales of the poets stands in contradiction with his own sense of himself as a victim of Hera.249 In the Orestes, Helen seeks to excuse herself as the victim of a ‘god-sent madness’ (79), and in the Trojan Women, the question whether Helen acted freely in deserting Menelaus or was driven by Aphrodite becomes part of the central debate between her and Hecuba, so that here again Euripides is allowing himself to play with the issue of the connection between divine intervention and human freedom.250 In the Aeneid, Nisus raises the same question: ‘Are the gods responsible for the ardour that I feel, or do we make gods of our desires?’ (IX, 184–5). The answer is the latter, but once the gods physically interfere as agents in the action, more thoughtful characters will inevitably pose Nisus’ question. Even so, as I have suggested, the relative independence of divine agents, if that literary gambit is present, though it may block an immediate psychologizing move, does not derogate from the fundamental legitimacy of the naturalistic strategy, seeing that the relative independence of the divine is a derived status. And poets can work at both primary and secondary levels, as we have seen. Virgil’s Allecto has an easy time of it turning Amata against Aeneas, but she encounters resistance when she visits Turnus, even though she has been careful to present a persuasive appearance by disguising herself as a priestess (VII, 435–44). But the fact that Turnus, unlike Amata, initially rebuffs Allecto no more detracts from Allecto’s essentially allegorical status than the fact that Palinurus, Aeneas’ helmsman, initially resists Sleep (also in disguise) detracts from the latter’s similar status (V, 847–53).251 (The vigour of Turnus’ response to Allecto suggests in any case that her words have hit home.)252 Tellingly, in his 1614 tragedy Turne, Jean Prévost entirely omitted the part of Allecto. And Lucan had already dispensed with the gods. It remains the case, then, that divine interactions with human beings have the asymmetric structure I have identified. Divine action in traditional literature is—immediately or mediately—parasitic on, and deeply driven by, facts of human psychology. Aeschylus’ Clytemnestra provides an interesting case in this connection. Initially, Clytemnestra is happy to take responsibility for her husband’s murder, indeed glories in her deed, and defends her action to the chorus as though it were entirely her responsibility (1372–1447). But when the chorus responds (1470) that the spirit of the house is wielding power through women, and that Zeus is responsible for everything (1486), Clytemnestra seizes the opportunity to urge that an avenging spirit committed the murder in her person (1501). The chorus rejects this pretext and affirms that Clytemnestra did the deed (1505–6), but then concedes that an avenging spirit may have been an accessory (1508). In his commentary Eduard Fraenkel says that Clytemnestra ‘is not making excuses for herself’ (1978, vol. 3, 711); but that is exactly what she is doing, though in due course (1551–3) she drops the strategy and reverts

66  Tragedy and Moral Redress to her initial acknowledgement that she killed Agamemnon.253 I suggest that the back-and-forth movement illustrates the essential identity of natural and supernatural glosses on Clytemnestra’s action, which has the effect of giving priority to the natural. She tries to shift the blame for her deed onto the daimōn of the house, but, though it was the chorus that first suggested the excuse, it now disallows the move:254 even if an avenging spirit is an accessory, that does not absolve Clytemnestra from guilt. Despite overt mention of an ancestral curse, Aeschylus’ agents are still free and incur their guilt freely,255 in the Oresteia no less than in Eugene O’Neill’s powerful adaptation of it, transposed from the Greek heroic age to the immediate aftermath of the American Civil War, Mourning Becomes Electra. Garvie writes that ‘In Aeschylus it seems that the son who inherits the family curse is never an innocent sufferer. He inherits not just guilt but a propensity to incur fresh guilt himself, and he is thus always in some degree responsible for his suffering’.256 The point applies to Clytemnestra too.257 Attempts to blame the gods in Attic tragedy are indeed generally implausible. Aeschylus’ Eteocles, like his Clytemnestra, tries to transfer responsibility for his actions to the divine (Seven Against Thebes, 689–91); but the chorus insists that it is his own fatal desire that prompts him to meet his brother in single combat (692–4).258 The excuse fails from Homer, whose Zeus roundly rejects the stratagem, to Garnier’s Cleopatre, who dismisses the theological escape that Charmion offers her, to Hughes’s Mordred, who blames fate but is aware of his own guilt, to Goethe’s Egmont, who likewise inveighs against his Schicksal but admits that he freely ignored the warnings he received, to Thomas Hardy’s Napoléon, who pretends to be ‘Ruled by the pitiless Planet of Destiny’ in a world that is clearly governed by human decisions; recall too the epiphenomenal close of Tess of the D’Ubervilles.259 In Sophocles’ Ajax it is reported by the messenger that Ajax once boasted that even a nobody could be successful with the help of the gods, but that he believed he could win glory without them (766–9). The phrase ‘without them’ (δίχα κείνων) is remarkable, as Patrick Finglass notes (ad loc.), because of the absence of a negative: ‘great feats are typically achieved “not without the help of the gods” ’.260 Ajax’ boast is unheroic, and it puts him in the same camp as notorious blasphemers like the Locrian Ajax (Odyssey IV, 504) and Capaneus (Seven Against Thebes 427–8).261 One may reconstruct the theoretical background of Ajax’ boast as follows. The presence or absence of divine help is initially a matter of success or failure: if one succeeds one infers that one had divine help, and not if not.262 So here the operatur principle is fully functional: the divine shadows the human. But once that idea is in place, it is natural to pray to the gods for their assistance in advance, so as to ensure success. By the same token, it might be thought risky not to do so, and then by a further step the boastful and the blasphemous can claim that ‘even without divine aid . . .’. So this looks like another case where the

Oedipus 67 operatur principle provides theoretical underpinning, but, once a divine apparatus is in place, it takes on a life of its own. This structure enables us to solve a puzzle noted by A. L. Brown (1983, 33): the Olympian gods are depicted as enjoying private lives as independent agents, but when they intervene in human affairs they are restricted to ‘bringing about’ events that could have happened anyway through natural causes. Why? The answer lies in the asymmetry I have been stressing. The gods are so restricted because their primary or basic function is to ennoble the natural, not to change or undermine it. Once they have been posited, a secondary or derived level comes into existence at which the gods can be portrayed as superhuman agents with private, independent lives, but the fact that their powers of intervention in mortal affairs remain severely limited betrays their naturalistic aetiology. A similar point applies to allegorical abstractions. In his Life of Milton Dr Johnson famously observed that ‘such airy beings are, for the most part, suffered only to do their natural office; and retire. Thus Fame tells a tale, and Victory hovers over a general, or perches on a standard; but Fame and Victory can do no more’ (1958–77, vol. 21, 198). So, for example, when the reckless Segramors attacks Parzifal, the latter is held in distraction by Lady Love (vrou Minne). Fortunately his steed is alert to the danger and its movements bring Parzifal to his senses—or, as Wolfram von Eschenbach puts it, Lady Reason (vrou Witze) brings him to his senses (Parzifal 288, 3–14). Johnson goes on just after the passage quoted to censure those poets who give ‘airy beings’ more to do than perform their natural office: ‘In the Prometheus of Aeschylus, we see Violence and Strength, and in the Alcestis of Euripides, we see Death brought upon the stage, all as active persons of the drama; but no precedents can justify absurdity’ (ibid.). Absurd or not, the important point for our purposes is that these developments have the articulated, split-level structure I have identified: in its conceptually original manifestation, Death can do no more than perform its natural office, but once it is brought on stage as a character, it can be given more to do. When, in Tasso’s Gerusalemme Liberata, Erminia steals Clorinda’s armour and secretly leaves Jerusalem to find Tancred, Love invigorates her failing spirits and flagging limbs, but also laughs at the ridiculous figure she cuts in her masculine attire (VI, 92–3). In his classic work The Allegory of Love, C. S. Lewis skilfully traced the process of abstraction that underlay the trend to allegory in late classical and medieval literature. What does not emerge from Lewis’s discussion with sufficient clarity, however, is that the process of abstraction is in some respects a reversion to type: the ‘allegorization of the pantheon’ (2013a, 73), by virtue of which, for example, Mars becomes a mere symbol of war (no love affairs with Venus), is in fact a rediscovery of the gods’ original function and point. The gods, Lewis tells us, ‘sink into personifications’ (ibid., 78–9); but that is what they always really were.

68  Tragedy and Moral Redress To obviate possible misunderstanding, I need to make clear that I am not suggesting that the naturalizing and debunking thoughts that I have been exploring in this section in connection with the operatur principle were actually entertained as a matter of course by ordinary people in antiquity or the medieval period. The moves were certainly available, and they were entertained by some, relatively enlightened thinkers (I shall return to this point, §20), but one presumes that for the vast majority of people divinity was, in Ruth Padel’s words, ‘part of the fabric of the world and the self’ and that for them, as Thales said, ‘all things are full of gods’ (1992, 48). Again, it may be true to say, as Blair Hoxby does, that ‘the early moderns, unlike most of us today, experienced their selves not as bounded egos but as voids open to spiritual influences’.263 In fact I think that this—and certainly ‘voids’—goes too far (I shall return to this issue in Chapter 4); but I am not denying that, until relatively recently, for most people the ordinary world was felt to be interpenetrated by an extraordinary, spiritual world with its own distinctive ontology and dynamics. The question in this section has concerned the extent to which that ontology and those dynamics are—and are portrayed in tragic literature as being—independent of the natural, and in particular independent of natural human motivation. My contention has been that, at the primary or basic level, there is no such independence, and indeed that the supernatural is dependent on the natural. Only at a secondary or derived level does space arise—assuming that a divine apparatus, generated at the basic level, is in place—for a certain measure of divine autonomy. This is a theoretical point. In effect what I have been offering is a genealogy of the role of gods in traditional literature, in the sense in which Hume and Nietzsche offered genealogies of morality.264 The story reconstructs a derivation of the phenomena in a way that is intended to make sense of them and display their underlying rationale. It would not falsify Hume’s genealogy of justice, for example, if it were found to be historically inaccurate: Hume is not trying to reconstruct actual historical events. In the same way, my account here is meant to be a philosophical heuristic rather than a genetic ‘just so’ story. It does not follow from what I have said that, if you ‘strip [Macbeth] of a supernatural stranglehold, placing responsibility for the evildoing entirely on the ambitious couple’, you are thereby forced into ‘seeing it as compensatory for some personal loss or trauma’.265 That would be too modern an inference, requiring us to find that every wrongdoer is at some deeper level a victim. In Aristotelo-Bradleian fashion the Macbeths’ characters give rise to their actions—they are responsible for their deeds— but the play does not invite us to ask how they got their characters in the first place. There is nothing to suggest deviance from the traditional, Aristotelian view that we are responsible for our characters. Recall the dictum of Plato’s priest that ‘the responsibility is the chooser’s; the god is exempt from responsibility’ (Republic 617e4–5). If we read the Myth

Oedipus 69 of Er as part of Plato’s anti-tragic polemic, then we arrive at a ‘Platonic contrast between two ultimate hypotheses about the world—the first that human lives are governed by external forces which are indifferent to, and capable of crushing, the quest for happiness; the second that the source of true happiness is located nowhere other than in the individual soul’s choice between good and evil’.266 My point here has been that the first of these hypotheses is liable to send us down a wrong path, towards a misunderstanding of tragedy; if I am right, we have the basis of a response to Plato. Tragedy is about individual responsibility, and the involvement of gods in no way undermines that. It is not the case that tragedy shows how ‘destiny is ultimately beyond human control’.267 Schelling argued that tragedy depicts a clash of necessity and freedom, that human freedom consists in struggling against a superior destiny, and that Aristotle was wrong to try to substitute hamartia for the will of destiny in the aetiology of tragic guilt.268 In his essay ‘Über die tragische Kunst’, Schiller propounded the more radical view that ancient tragic heroes were trapped in a causal order that robbed them of freedom, whereas modern such heroes manage to transcend the causal order.269 Schelling’s position influenced A. W. Schlegel and Coleridge, who in his lectures on Shakespeare held that tragedies involve ‘those Struggles of inward Free Will with outward Necessity’, and was embraced more recently by Raymond Williams.270 Schelling and Schiller were themselves influenced by the Kantian antinomy between freedom and nature, not to mention the effects, as it developed from aspiration and hope to tyranny and terror, of the French Revolution.271 And the essential view had already been formulated by Goethe in his 1771 Rede zum Shakespears Tag.272 If my argument in this chapter has been sound, these assertions, and in particular the claim that there is an opposition between freedom and necessity in ancient tragic literature, are unsustainable. As Schmitt observes, ancient tragic heroes are both free and part of the natural order.273 Indeed modern heroes are too—how could they not be? Schiller’s exemplary modern heroes are Shakespeare’s Coriolanus, Corneille’s Rodrigue and Chimène, and Wieland’s Hüon and Amanda;274 but when one examines the precise ways in which Shakespeare, Corneille, and Wieland characterize them one finds that, though these figures defy the stars, they do not escape from the causal order. That would, after all, be a rather difficult thing to do: you cannot do it merely by railing against the world, as Lear found when the thunder would not peace at his bidding. There is a fascinating passage in Werner Heisenberg’s autobiography in which he describes how he conversed with a friend as they picked their way through the ruins of Berlin after an air raid in March 1943. (Here one cannot help but recall T. S. Eliot’s conversation, in similar circumstances, with the ghost of his former self in the central panel of ‘Little Gidding’.) In that discussion Heisenberg diagnoses the nisus we have identified in Schelling and Schiller—the thought that freedom can only

70  Tragedy and Moral Redress consist in escape from the causal order—as a peculiarly German tendency of thought.275 One suspects that it is a good deal more widespread. But it is certainly associated with German tragedy, and not just by German critics. In his splendid biography of Lessing, H. B. Nisbet writes (2013, 405): In Goethe’s Götz von Berlichingen, Schiller’s Wallenstein, Büchner’s Danton’s Death and most major tragedies of the nineteenth century, there is much more emphasis on the destructive power of impersonal external forces than in Lessing’s dramatic theory and practice, where individual character and tensions within smaller human groups remain central. This rather strongly implies that Götz, Wallenstein, and Danton struggle, in their respective plays, against superior forces—that they undergo Schicksalstragödien. But it seems to me clear that all three heroes fully conform to the Aristotelian pattern I have identified, in which hamartiai, consisting of culpable cognitive failures, are punished home. Wallenstein is plainly subject to what Bernhard Greiner calls his ‘enormous self-­delusion’ (2012, 439). Götz indeed thinks he is struck down by the gods,276 but the play of which he is the hero makes clear that the vital decisions are in his hands and that he commits crucial errors: he misjudges Weislingen, and he foolishly joins the peasant rebellion, the excesses of which he then—predictably—cannot control. Büchner’s play resembles the Oedipus Rex inasmuch as Danton’s crimes lie in the past (‘durch alle Gassen schrie und zetert es: September!’, a reference to the September massacres in which Danton had had a hand),277 and it resembles the fifth act of Macbeth both in this respect and in its mood of world-weariness; Danton must now pay the price for his past felonies. It is a pity that Nisbet does not mention which other major tragedies he has in mind, but the three that he does mention fully fit the Aristotelian model, as do Sophocles’ Women of Trachis, Shakespeare’s Romeo and Juliet, and Voltaire’s Tancrède, which Schopenhauer listed (along with Oedipus Rex) as Schicksalstragödien.278

9 Conclusion Sophocles’ Oedipus Rex does not tell us what would have happened if Apollo’s oracle had been silent: would Oedipus even so have killed Laius and married Jocasta? We do not know, but presumably one possibility is that he would not (recall the discussion of Cavell above, §7), and in that case one might say that, though (as I have argued) Apollo did not make Oedipus perform those actions, he did do something which was such that, had he not done it, Oedipus would not have performed them, so that, even if we do not have the sort of causation that destroys freedom, namely duress, still we have counterfactual dependence. Even if we say,

Oedipus 71 with Kovacs (2009, 361), that Oedipus ‘has been deliberately set up’ by Apollo, still, agents who are (merely) ‘set up’ perform their actions freely: if I insult you intending that you lash out at me so that I can then threaten you with legal action, assuming the ruse is successful you might feel that you had been ‘set up’, but you could not deny that you had assaulted me freely. In a similar way, the youths of Chaucer’s Pardoner’s Tale would not have met their grim end had they not encountered Death and acted on his words, so that we have counterfactual dependence, and Death indeed ‘sets them up’, but their actions in the story are from first to last free. In the Nibelungenlied, the Meerfrauen tell Hagen that no one will survive the visit to Etzel and Kriemhild except the chaplain. Hagen tries to falsify the prediction by drowning the chaplain at the crossing of the Rhine, but he manages to escape to the river’s hither bank, and so does not take part in the fatal expedition:279 Hagen’s very attempt to falsify it contributes to ensuring the prediction’s truth. Again, we read in the Holinshed chronicles on which Macbeth was based that King Natholocus sent a gentleman to consult a witch about a revolt which was being fomented against his rule.280 The witch tells the gentleman that the king will be murdered by a close and trustworthy friend, in fact by that gentleman himself. He then reviles her for her villainous prophecy, but on his return journey he realizes that if he repeats her words to the king, or if they are divulged by another, his life will be imperilled, for Natholocus will have good reason to put him out of the way. Whereupon he resolves to strike first, and on his arrival at court he kills the king. Here too we have counterfactual dependence, but also full freedom. (The gentleman is an early victim of the ‘Hobbesian trap’.)281 The same holds of the witches’ prophecies to Macbeth in Shakespeare’s play: he would not have acted as he did if he had not heard their prediction that he would be king, but his actions are those of a free agent. Though I disagreed with the view conveyed in the quotation from Nickau given above (§7), it is interesting to note how it echoes the Tractatus, where Wittgenstein remarks: ‘Die Tatsachen gehören alle nur zur Aufgabe, nicht zur Lösung’ (‘The facts all belong only to the task set, not to its solution’: §6.4321). By ‘solution’ Wittgenstein means, as he goes on to tell us, ‘the solution to the problem of life’, and for the Wittgenstein of the Notebooks and the Tractatus that solution resides in taking up a certain attitude to the facts, which in a transcendental sense we cannot affect. For Sophocles, the relevant sense in which the facts obtain is not a transcendental but an empirical one, and as I have stressed Oedipus certainly can affect—settle, determine—the facts. He cannot change the facts; but then no one can change the facts, if by that phrase we mean, supposing that it is the case that p, having the power to bring it about that not-p as well. No one can do that; but having the power, given that p, to bring it about that not-p instead, is by contrast unproblematic (at least in many cases), and Oedipus, like any free agent, has that power.

72  Tragedy and Moral Redress (If it is a fact that p, then it follows that I do not exercise any power I might have to bring about not-p instead, but it does not follow that I do not have that power; if I had exercised it, then it would not have been a fact that p.) But there is also a sense in which Sophocles’ play is about the proper reaction to the facts as they are. Oedipus responds to his catastrophe by a wilful act of self-blinding; but there were other options. He might have left Thebes and succeeded Polybus at Corinth. It might even have been possible for Oedipus to continue to rule Thebes, as in the Homeric and Euripidean versions of the story (it is presupposed in the Oedipus at Colonus that he remained in Thebes after the revelations),282 though during the course of the play Oedipus himself shuts these options off.283 The parallel with Jocasta is instructive here. Oedipus drives her to suicide, in effect, but it is by no means clear from the play that she could not have lived with knowledge of the truth if it had been kept sufficiently submerged. It is a nice question when exactly Jocasta gains understanding. Lefèvre thinks that she is just as obtuse as Oedipus (2001, 142), but she evidently grasps the truth more quickly than he does. When Oedipus learns that Polybus has died a natural death, so, apparently, falsifying the oracle, Jocasta tries to stop the investigation (973–83; tr. Lloyd-Jones): JOCASTA:  Did I not foretell this to you long ago? OEDIPUS:  You told me: but I was led along by fear. JOCASTA:  Then let none of these things worry you any more! OEDIPUS:  And how can I not fear intercourse with my mother? JOCASTA:  But what should a man be afraid of when for him it is the event

that rules, and there is no certain foreknowledge of anything? It is best to live anyhow, as one may; do not be afraid of marriage with your mother! Many have lain with their mothers in dreams too. It is he to whom such things are nothing who puts up with life the best. It is plausible that Jocasta’s line ‘Then let none of these things worry you any more!’ (μὴ νῦν ἔτ᾽ αὐτῶν μηδὲν ἐς θυμὸν βάληις), and especially her use of the vague phrase ‘none of these things’ (αὐτῶν μηδέν), indicates that she knows the truth. Dawe merely remarks (on 976) that ‘Jocasta had inadvertently half opened the door to Oedipus’ apprehensive question by using the phrase “αὐτῶν μηδέν” instead of “τοῦτο μή” ’—which latter phrase would in context mean: ‘Do not let this particular thing (the thought about parricide) worry you’. But the point is subtler than that: Jocasta is trying to close down further inquiry into the possibility of incest by saying in effect ‘The parricide part of the prediction has turned out to be wrong, so forget all of it (including the incest)’. But in order to put the idea of incest out of Oedipus’ mind she first has to mention it, and then it sticks. She tries to refer to it as obliquely as possible, hence the deliberate vagueness of her phraseology, but the tactic fails. Once

Oedipus 73 you have mentioned something, however obliquely, you have mentioned it. (And ‘in dreams too’ looks like a Freudian slip—that is, ‘in dreams as well as in waking life’.)284 Plausibly, indeed, Jocasta already knows the facts at an earlier stage, namely when she hears Oedipus’ account of his background (771–833), for that makes the truth obvious (though not to Oedipus). And the clear irrationality of her response to his account (848–58) seems to betray an urgent impulse to divert Oedipus from what she already knows to be the reality of the situation.285 At any rate, D. W. Lucas’s suggestion that ‘Jocasta realizes that Oedipus is her son between 1026 and 1056’ (1972, 131) puts the discovery much too late. (In a natural development of Sophocles, Gide’s Jocaste knows the truth from the start: 2007, 206–11.) In the marvellous opening panel of the fourth epeisodion (911–23), Jocasta ‘provides a royal and private counterpart to the public acts of piety at the opening of the play, 3–4’, as Dawe remarks (on 912–13). She steps forward bearing wreathed branches, and in a poignant and desperate appeal begs clemency of Apollo. Oedipus, Jocasta complains, ‘does not judge the present in the light of the past as a man of sense would do’ (915–16): that is, he does judge the present in the light of the past, but not in the way in which a man of sense would.286 A man of sense would not let the past impose itself on the present; it is better, as she subsequently tells Oedipus in the passage quoted above, ‘to live anyhow, as one may’, without delving into things best left hidden; for who does not so live, as the Marschallin reminds us in Der Rosenkavalier, will be punished by life and unpitied of God.287 The atmosphere of Jocasta’s words is thick with a sense of foreboding and failure, rather like some classical ekphrases, such as Aeneas’ arrival on the shores of Italy at the beginning of the second half of the Aeneid (VII, 25–36),288 or the matrons’ prayer depicted on the mural in Virgil’s Carthage (Aeneid I, 479–82), or Clytemnestra’s prayer to Apollo in Sophocles’ Electra,289 or the locus horridus scene-setting in Seneca’s Thyestes before the description of Atreus’ appalling acts (641– 82),290 or, again, like the Zwinger scene of Goethe’s Faust (1985–99, vol. 7, 156), where Gretchen prays to the Virgin— Ach neige, Du Schmerzenreiche, Dein Antlitz gnädig meiner Not!— (‘Oh graciously incline your countenance, you who are rich in pain, to my need!’), or like the beginning of the second act of Madam Butterfly (Puccini 1984, 97), where the tragic heroine’s maid Suzuki kneels before an image of the Buddha, and implores the gods: Fate che Butterfly non pianga più, mai più, mai più, mai più!

74  Tragedy and Moral Redress (‘Grant that Butterfly may weep no more, no more, no more, no more!’). The words with their musical accompaniment insinuate in unambiguous terms that the maid’s intercession will not avail: the more we hear the incantatory repetition of the phrase ‘mai più’, the more we sense with absolute certainty that the prayer is falling on deaf ears. Shortly afterwards Sharpless arrives with Pinkerton’s letter, and the inevitability of tragedy sets in. Just so in the Oedipus Rex: Jocasta’s hopeless words of appeasement are hardly uttered when the Corinthian messenger arrives, with news that will induce the final catastrophe. ‘It is as though, by this coincidence, the gods were mocking Jocasta’s act of piety’.291 If Oedipus is morally innocent, as Dodds and others think he is, why does he blind himself? This was by no means a fixed feature of the myth: for example, we know that in Euripides’ version Oedipus is blinded by Laius’ servants.292 Bernard Williams argues that Oedipus feels ‘agent regret’ for actions that he performed unintentionally: he did not intend the parricide or the incest, but he decides to ‘own’ them nevertheless.293 Along similar lines, Hegel thought that Oedipus blinds himself because he identifies himself with his acts.294 The reason for the blinding that Dodds offers has to do with pollution: ‘Morally innocent though [Oedipus] is and knows himself to be, the objective horror of his actions remains with him and he feels that he no longer has any place in human society’ (1973, 72). Dodds cites as ‘the nearest parallel’ to Oedipus’ situation the story of Adrastus in Herodotus (I, 35–45), who ‘was the involuntary slayer of his own brother, and then of Atys, the son of his benefactor Croesus; the latter act, like the killing of Laius, fulfilled an oracle’ (ibid.). Although Croesus forgave Adrastus, on the basis that the killing of his son had been brought about by divine agency, Adrastus refused to demit responsibility, and committed suicide. In one respect Adrastus’ case is unlike Oedipus’; for Adrastus got over the fratricide, and it was the further killing—of someone who was genetically a stranger to him—that broke him. But in another respect the parallel is close, though the connection does not help the understanding of the Oedipus Rex according to which Oedipus is morally innocent. For in the case of Herodotus’ story we must ask, as we should always ask when confronted with instances of divine intervention: why did the god intervene? Why pick on Adrastus? Well, observe that Adrastus was the sort of man who would make just this mistake: ­suspicious—no? And does not half the blame lie with Croesus, who, terrified by the prophecy that his son Atys would be killed, employed to protect him a man who, he knew, had a record of accidental killing? So far from supporting Oedipus’ moral innocence, the parallel implies the opposite. But Dodds’s explanation of the self-blinding is independent of this point and can stand. Oedipus’ innocence or guilt is irrelevant to the issue of pollution. As Dodds says, ‘Oedipus mutilates himself because he can face neither the living nor the dead’ (1973, 71).295 But Oedipus also blinds himself because, like Wilmot’s Tancred and Shakespeare’s Gloucester, he stumbled when he saw.296 Then he was a seeing blind man; now,

Oedipus 75 like Teiresias, he is a blind seeing man.297 The act of self-maiming brings to its natural conclusion a running theme of the play.298

Notes 1. Buxton 1996, 45. On the problems presented by the end of the play, see Segal 1981, 208; M. Davies 1982 and 1991; Macleod 1983, 141–2; Roberts 1988, 183–4; Mastronarde 1994, 24–5; D. Fowler 2000, 6, 244–6, 302; Taplin 2001, 45–6; Budelmann 2006a; Burian 2009; Goldhill 2015, 240–2. 2. See Korzeniewski 2003, 476–505; Lurje 2004, 13–225. 3. Dacier 1692, 193. Cf. Lurje 2004, 132–3; Billings 2014, 24–5. 4. See Gaskin 2013a, ch. 7. 5. Bain 1993, 81; J. Gould 2001, 244. 6. Ewans 1996, 450. 7. On the legitimacy of the terms ‘hero’ and ‘heroine’, see D. Lucas 1972, 115, 140, and Vickers 1973, 7–8, against Jones 1962, 12–18 (duBois 2008 supports Jones); also Halliwell 1990b, 173–4; Gill 1996, 98–9; Silk 2004, 246– 7; Rose 2012, 257. 8. Dodds 1973, 66. This is esp. clear in Euripides: see e.g. Cyclops 173; Children of Heracles 458; Hippolytus 377–81, 996; Andromache 674, 1165; Hecuba 327; Electra 50–3, 294–6, 386–90, 971–2, 1035; Heracles 172, 283, 347, 1254; Iphigeneia in Tauris 385–6; Ion 374; Helen 851, 1441; Phoenician Women 86, 393, 569–70, 584, 1726–7; Orestes 417, 492–3; also [Aeschylus], Prometheus Vinctus 9 and 1039; with the commentaries of Seaford, Allan, Barrett, Stevens, J. Gregory, Denniston, Cropp, Bond, L. Parker, Lee, Mastronarde, Willink, and Griffith ad locc. Cf. Denniston 1939, xxii; Butcher 1951, 321; Dodds 1951, 16–17; D. Lucas 1972, 301; Winnington-Ingram 1980, 122; Dihle 1982, 33–4; Dover 1994, 116–24; Mastronarde 2010, 190; Rutherford 2012, 379. 9. Bond mentions Ulster in this connection (on Euripides, Heracles 283); I recall the usage from my childhood in the North-East of Scotland; Malcolm Heath tells me that he encountered it in Yorkshire. 10. Nicomachean Ethics (NE) 1110b29. Cf. D. Lucas 1972, 300. 11. Cf. Euripides, Trojan Women 1028; Ion 426; Alcestis 709–10, with L. Parker ad loc. 12. Bywater 1920, 11. 13. Macintosh 2009, 1–2. 14. Stinton 1990, 145–6. Cf. Adkins 1966, 82; Cessi 1987, 15–28, 37; Schmitt 1988a, 184. 15. Cf. D. Lucas 1972, 144–6; Halliwell 2005, 404; Conte 2007, 158; Greiner 2012, 16, 47–51, 124–6. 16. See on this point D. Lucas 1972, 140; Stinton 1990, 164; M. Nussbaum 1992, 150–4; Heath 2017, 337–9. 17. See von Fritz 1962, 3–4; Lurje 2004, 335 n. 52. 18. Cf. D. Lucas 1972, 141–4. 19. Harsh 1945, 56–7; F. Lucas 1957, 117 n. 1; Cessi 1987, 29–30; Lefèvre 2001, 7; Schmitt 2011, 438; Kappl 2015, 52. 20. Cf. Lear 1992, 329. 21. Cf. Aristotle, Metaphysics 1090b19; J. Armstrong 1998. 22. Harsh 1945, 49; F. Lucas 1957, 122–3; T. Gould 1966, 514–15; Cessi 1987, 265; Schmitt 2011, 437; Greiner 2012, 65. 23. Cf. Alexander 1955, 84–5; Halliwell 2006, 135–6. 24. This translation was adopted by e.g. Kitto 1960, 233, and T. Gould 1965, 366.

76  Tragedy and Moral Redress 25. Barnwell 1982, 243; cf. D. Lucas 1972, 300; Dewar-Watson 2014, 54. For a case where ‘hamartia’ means a mistaken belief, see Aristotle, NE 1110b29. 26. See here Rudd 1976, 34–5. 27. Cf. Ahrensdorf 2009, 169. 28. See Poetics 15, esp. 1454a33–6. 29. NE III, 5; Schmitt 1988a, 162–3; Everson 1990, 93–7. 30. Letter of 18 Dec. 1756 (1996, vol. 4, 192–4); Hatfield 1956, 289; H. Nisbet 2013, 217, 397–8. 31. Cf. Sicherl 1977, 93. 32. See e.g. Aristotle, NE 1127a27–8, 1127b2–3, 14–15; Bradley 1991, 29–30. 33. See Gill 1986, 263; cf. 270–1. 34. See Aristotle NE 1115b28–1116a9. 35. See Schmitt 1994, 331, 341–2; 2014, 5 n. 13; 2015, 227–31, 234, 237, 240; Kappl 2015, 72. 36. Dodds 1973, 67; cf. Dewar-Watson 2014, 54. 37. Cf. Demosthenes XVIII, 274–5, with Harris 2004, 30. 38. See Aristotle NE VII, 8, esp. 1151a10; Cessi 1987, passim; Lurje 2004, 337– 8, 374; Schmitt 2011, 457–8; Kappl 2015, 52–5. 39. Stinton 1990, 149–50, 156–7; Halliwell 1998, 208–10. 40. Cf. Thucydides III, 40, 1–2; Cessi 1987, 11. 41. Harsh 1945, 52. 42. Cf. Aristotle NE 1125a17–23. 43. M. Nussbaum 1986, 382; cf. 1992, 140. 44. Stinton 1990, 150, 165–6; cf. Denniston and Page 1957, xxix n. 2. 45. T. Gould 1965, 385–6; Empson 1986, 113; J. Gould 2001, 260; Korzeniewski 2003, 476. See also Schadewaldt 1960, 237; Schopenhauer 1986, §51 (vol. 1, 355); Kovacs 1987, 71–2; Vickers 1973, 497–8; Winnington-Ingram 1980, 203; Gill 1986, 264; Nicolai 1992, 63–4; M. Nussbaum 1992, 155; Sherman 1992, 178–80, 186, 189–90; White 1992, 224–5. 46. Lloyd-Jones 1983, ch. 5; 1990, 202–7. Cf. Kierkegaard 1987, vol. 1, 150; Denniston and Page 1957, xxix. On Agamemnon, see further their commentary on 386ff.; Fraenkel 1978, vol. 3, 623–5. Cf. Easterling 1993a, 23–4. 47. On the curse in Seven Against Thebes, see Hutchinson on 720–91, 723, 801f., and 842; Gill 1990, 24–7. On the Oresteia, see Lloyd-Jones 1990, 283–99; N. Hammond 1973, 395–8, 407–8. 48. See Bowra 1945, 163–4; Dodds 1973, 69, 71; Vickers 1973, 496–7; Garvie 1986, xxvii–xxviii; Kirkwood 1994, 75; Kullmann 1994, 114; M. West 1999, 41–2; Finglass 2007, 236–8; Sewell-Rutter 2007, 61–7, 112–30; A. Boyle 2011, li–lii, 278; Sommerstein on Aeschylus, Eumenides 934–7; Bond on Euripides, Heracles 1258. 49. Gillett and Hankey 2005, 276–8; cf. Edmunds 2006, 49–50. 50. Letter of 2 Oct. 1797 (1992–2004, vol. 12, 330–1); cf. Billings 2014, 116. 51. On the concept of ‘long time’, see Kirkwood 1994, 70; Honigmann 1997, 68–72; P. Hammond 2009, 152–3. 52. Aristotle NE 1101a31–3; Pocock 1973, 99; Mohamed 2005; Garvie 2009, 142; T. Gregory 2010. 53. Kaufmann 1968, 119–20. 54. Cf. Schmitt 1988b, 19, 26. 55. Dawe on Sophocles, Oedipus Rex 619. 56. See Knox 1998, 15. 57. Cf. Flaig 1998, 118–22. 58. Schmitt 1988b, 16. Contra Ahrensdorf 2009, 12–13 n. 8. 59. Cf. Schmitt 2014, 48.

Oedipus 77 60. Bowra 1945, 190; Knox 1964, 29; Vickers 1973, 498; Winnington-Ingram 1980, 203; Segal 1981, 210, 241; Goldhill 1986a, 209; Gelfert 1995, 41; Dawe 2006, 16–17, 76; Liapis 2012b, 92; Holbrook 2015, 11, 15. 61. See e.g. Bowra 1945, 190; Knox 1998, 117. 62. See e.g. Whitman 1951, ch. 7, esp. 131, 138; Segal 1986a, 66; J. Schmidt 1989, 39; Nicolai 1992, 64; Nickau 1994, 21; Ahrensdorf 2009, 10 n. 5. 63. Cf. Poole 1987, 95–6; Cessi 1987, 261, 267; Schmitt 1988b, 20, 22; 2014; Daniels 2006; Hall 2012, 306–9. 64. Cf. H. Levin 1961, 158; Waswo 1974, 85–9; M. Burnett 2010, 165–70; Holbrook 2015, 110–12. 65. See Schmitt 1997, 25. 66. Lefèvre 2001, 125. Cf. Schmitt 1988b, 13; Flaig 1998, 65–7; Dawe 2006, 16. 67. Cf. Dawe on 813. 68. See Reynaud and Thirouin 2004, 192; Mueller 1980, 110. 69. Segal 1994, 74; Lefèvre 2001, 122–3; Garvie 2016, 43. 70. Lefèvre 2001, 129. 71. Flashar 1994, 64. 72. Lefèvre 2001, 126. Cf. Dawe on 814. 73. Winnington-Ingram 1983, 78–100; cf. M. Edwards 1978, 29. 74. Cf. Pack 1939, 354; Cessi 1987, 262; Schmitt 1988b, 14–15; Flashar 1994, 68. 75. M. West got there first: 1999, 39; cf. Young 2013, 132. 76. Cf. Harsh 1945, 48; Greiner 2012, 43, 51; T. Mann 2012d, 250–1. 77. Cf. Aristophanes, Frogs 1180–95, with Bain 1993, 88–9. 78. Cf. Lefèvre 2001, 136; Scodel 2005, 240–1. 79. So e.g. von Fritz 1962, 11, 26; Stinton 1990, 163. 80. So M. Nussbaum 1986, 387–8. 81. Cf. Poole 1987, 101–3; contra Manuwald 1992, 15–17. 82. Schmitt 1997, 32–3. 83. See Horace, Ars Poetica 435, with Rudd ad loc. 84. Cf. Schmitt 1988b, 21. 85. So e.g. Vickers 1973, 499–500; Ahrensdorf 2009, 13, 21. 86. See Pucci 1992, 42–8. 87. A. Boyle 2011, lii–liii; Schmitt 2011, 481. 88. M. Nussbaum 1986, 383; cf. Sorabji 1980, 297. 89. So Ahrensdorf 2009, 42. 90. Lefèvre 2001, 17–26; pace Zeitlin 1996, 347. Cf. Hall 2009, 83–9. 91. See McDowell 1998a, Essay 3 (esp. §§3 and 6); Gaskin 2001, 164–5; Halliwell 2012, 209. 92. Euripides fr. 540a, with Collard, Cropp, and Gibert ad loc. See Benardete 1966, 106; Goldhill 1986a, 216–18; Pucci 1992, 35–8, 66–78; Rutherford 2012, 100–1; A.-B. Renger 2013, 11. 93. Sophocles, Oedipus Rex 1036; cf. 718, 1034. See also Euripides, Phoenician Women 26–7, with Mastronarde ad loc.; Dover on Aristophanes, Frogs 1192. 94. Scodel 1984, 63. 95. Pace, too, Segal 1981, 207; D. Schmidt 2001, 151, 243. 96. Sophocles, Oedipus Rex 1525; Euripides, Phoenician Women 1688, 1759. Cf. Rutherford 2012, 74 n. 17. 97. Daniels 2006, 563. 98. Schadewaldt 1960, 281. For versions of this thesis, see T. Gould 1966, 505; Goldhill 1986a, ch. 8; J. Schmidt 1989; Nicolai 1992, 18, 64–5; Segal 1994, 88 (cf. 1986b, ch. 5); Rocco 1997, ch. 2; Knox 1998, 47–8; Grady 2009, 137–9; Greiner 2012, 49–50; A.-B. Renger 2013, 12–13; Rowan Williams 2016, 97, 102–5, 143, 158.

78  Tragedy and Moral Redress 99. Goldmann 1959, 45–9; 1970, 19. 100.  Cf. Ahrensdorf 2009, 2; Young 2013, 131–2. I shall return to Nietzsche in Ch. 3. 101. Pace Burns 2002, 549. 102. Cf. Kullmann 1994, 113; Schmitt 1997, 25; 2014, 50–1. 103. Cf. Halliwell 2005, 403; 2006, 137. 104. Halliwell 2006, 124–7. 105. Goldhill 1986a, 209–10; Goux 1990, 156; Rutherford 2012, 350; cf. Waswo 1974, 81. On Hegel, see Roche 2009, 54–5. 106. Cf. Schmitt 1988b, 23; 2014, 24–5. 107. Segal 1981, 215–16; Scodel 1984, 64; Zeitlin 1986, 111; Wilson 2004, 32–5; P. Hammond 2009, 77; Liapis 2012b, 90. 108. See A. Boyle on Seneca, Oedipus 221–2. 109. Segal 1994, 74; cf. Garvie 2016, 44; Hutchinson on Aeschylus, Seven Against Thebes 756. 110. So e.g. Manuwald 1992, 26; cf. Nickau 1994, 15. 111. Corneille, Œdipe 389–90 (Reynaud and Thirouin 2004, 29–30); Edmunds 2006, 91. 112. See Dacier 1692, 191; Harsh 1945, 53; Stinton 1990, 177–8; Rutherford 2001, 291. 113. Heath 1991, 392. 114. Pace Heath 1991, 392–3, 395. 115. See Cessi 1987, 262; Schmitt 1990, 92–9; Gelfert 1995, 14; Lefèvre 2001, 7. 116. Aristotle, Poetics 1453a4; Bradley 1991, 46; cf. Moles 1984, 49; Gadamer 1990, 136. 117. See Morgenröthe, §78 (1999b, 76–7); Gadamer 1990, 136–7. 118. See here Aeneid IV, 547–52, 696; cf. Rudd 1976, 36; Hejduk 2013, 149–50. 119. See Lee 1982, 52–3. 120. Dodds 1960, 237; cf. xlv; Stevens 1971, 235; Mastronarde 2010, 188–9. 121. So e.g. Whitman 1951, 124; F. Lucas 1957, 119–20; Steiner 1961, 4–8, 127–9; von Fritz 1962, 7; Kaufmann 1968, 243–4; Lloyd-Jones 1983, 104. 122. See Seidensticker 1982, 41; G. Edwards 1999, xxii–xxiii; M. Nussbaum 2001, 311–12; Konstan 2005, 50–4; Sternberg 2005, 21–2. 123. Mastronarde 1994, 260; cf. n. 8 above. 124. Cf. Heath 2011, 168–9. 125. Heath 1987, 82–3; White 1992, 229; Allan and Kelly 2013, 102 n. 90. Contra M. Nussbaum 1986, 384; Halliwell 1998, 174; cf. 2006, 119–20. Nussbaum has since modified her position: 2008, 153. 126. Cf. his verdict on Euripides’ Bacchae in his note on 1348–9; Manuwald 1992, 43. 127. Boitani 1989a, 9–10. For discussion of these and similar examples, together with references to their sources in Aristotle and Alexander of Aphrodisias, see Gaskin 1995, ch. 14. The example from Aquinas that I quote there (205 n. 33) is rather close to the case of Oedipus. Cf. also Boethius, Consolatio Philosophiae V, pr. 1; J. Mann 2003, 98. 128. Stobaeus, Ecl. I, 7, 9; Alexander of Aphrodisias, De Fato VIII (Suppl. Arist. II, 2, 174, 1–2); cf. Coleman 1990, 63 n. 44. 129. Dodds 1973, 71; Vernant in Vernant and Vidal-Naquet 2001, vol. 1, 71. Cf. Bowra 1945, 164–5, 168; Burkert 1991, 16. 130. See Gottsched 2009, 102, with Lefèvre 2001, 119, 135. Cf. Flashar 1994, 66–7; Schmitt 1997, 39; 2014, 22–4. Contra Segal 2001, 58; Vernant in Vernant and Vidal-Naquet 2001, vol. 1, 110. 131. So too Harris 2012, 292–5. 132. Flaig 1998, 98–102, 110–12; cf. Redfield 1994, 72–5.

Oedipus 79 133. Euripides, Phoenician Women 32–45; Cocteau 2012, 90. 134. Aeschines I, 91. 135. Contra Girard 2010, 106, 114–15. 136. Aristotle, NE 1135b25–7. 137. Goethe, Torquato Tasso II, 5 (1616–22; 1985–99, vol. 5, 779); the Princess and Eleonore understand it too: III, 1 (1661–9; ibid., 781); Lamport 1990, 92. 138. Lurje 2004, 321–34, 342, 384. 139. NE 1135a28–30; tr. Crisp. 140. Cf. Aristotle, NE 1110b18–22. 141. Miola 2000, 104; Liebler 2002, 140–1; Gaskin 2013a, 97–9; cf. Shapiro 2005, 163. 142. Cf. Gaskin 2006, 77–9. 143. NE 1110a27–9, Rhetoric 1401a35–b3; cf. NE 1148a33–4; Harsh 1945, 48–9. 144. See Kappl 2015, 62–3. 145. Plato, Laws 838c1–7; Republic 619b7–c6, tr. Lee. 146. See on these figures Harsh 1945, 54–5; Barrett 1964, 391, 403, 413–14; von Fritz 1962, 11–12; Bremer 1969, 120; Winnington-Ingram 1983, 32–8, 51–4, 78–100; Stinton 1990, 162–4, 176–9; Cairns 1993, 318–19 n. 202; Segal 1993, 150; J. Gregory 1997, 70–4; Lefèvre 2001, 11–26, 32–5, 48–62; Garvie 2009, xxix; Schmitt 2011, 446–55, 489–91; 2015, 215–18; Allan and Kelly 2013, 104, 108–9; Dodds on Euripides, Bacchae 1117– 21; Hutchinson on Aeschylus, Seven Against Thebes 653–719; Griffith on [Aeschylus], Prometheus Vinctus 8–9. 147. D. Lucas 1972, 141; J. Gregory 1977, 267–8 (but see 1997, 137–9); Gellrich 1988, 116, 119; Korzeniewski 2003, 477; Mastronarde 2005, 325. 148. Wilamowitz 1984, vol. 2, 128–9; 1935–72, vol. 1, 466. Cf. A. Burnett 1971, 168; Lee 1982. 149. Fitch 1987, 28–30. 150. See Schiesaro 2009, 224–5; Lefèvre 2015, 238–9, 281, 285–6. 151. See Gaskin 1992 and [forthcoming] for argument on these points and further references. 152. Dodds 1973, 67–8. So too Lloyd-Jones 1990, 201–2; Nuttall 1996, 34 (cf. 95). 153. I follow Schadewaldt 1960, 353–4; D. Lucas 1972, 142; Flashar 1984, 11. But against this construal of ‘to philanthrōpon’, see Lessing 1996, vol. 4, 585, 591; Halliwell 1998, 219 n. 25. 154. See T. Gould 1966, 479–80; Kovacs 2009, 358–9. 155. Cf. Dodds 1973, 70–1; N. Hammond 1973, 399; Knox 1998, 39–40. 156. Boethius, Consolatio Philosophiae V, pr. 4 and 6; cf. Dante, Paradiso XVII, 37–42; Lorris and Meun, Le Roman de la Rose 17401–24; Chaucer, Troilus and Criseyde IV, 960–1082; Marston, The Tragedy of Sophonisba II, 1, 131–5; J. Mann 2003, 96–8. 157. The category mistake is surprisingly common: see e.g. Flashar 1994, 67; 2000, 115; McLuskie 2003, 396; Smith 2016, 93–5, 100. 158. I defend this interpretation, with nuances, in Gaskin 1995, chs. 1–12. See also MacFarlane 2003. 159. So e.g. Pucci 1992, 25; Manuwald on line 791. Cf. lines 854, 995. 160. Cf. Sophocles, Philoctetes 200; Euripides, Phoenician Women 1602; [Euripides], Rhesus 752. 161. So e.g. at Euripides, Children of Heracles 491; Iphigeneia in Tauris 1288; Aristophanes, Thesmophoriazusae 16. See LSJ s. v. II. 162. So e.g. at Sophocles, Ajax 925; cf. Philoctetes 1083. 163. See Gaskin 1994a; 1995, ch. 12 with Appendices 1 and 2; Schmitt 1997, 27–9.

80  Tragedy and Moral Redress 164. Cf. [Aeschylus], Prometheus Vinctus 485, 1067, with Griffith ad locc.; Euripides, Andromache 1247, with Stevens ad loc. 165. Doubts about ascribing free will to fifth-century dramatic characters that are grounded in the so-called lexical method (so e.g. Vernant in Vernant and Vidal-Naquet 2001, vol. 1, ch. 3) are well disposed of by Halliwell 1990a, 34–42, and Flaig 1998, 33–4; see also Gaskin 2001. I shall return to the inadequacies of the lexical method in Part II. 166. Cf. Hejduk 2013. 167. Bradley 1991, 41–3; cf. Montrose 1996, 15. 168. See Leavis 1952, 145; Garner 1976, 244–50; Neely 1980, 228; Gaskin 2013a, 99–101. 169. Cf. Bradley 1959, 82; Wittgenstein 1984, 466. 170. Cf. Poole 1987, 103; M. Nussbaum 1994, 64; Küpper 2014, 300. 171. Steinmetz 1987, 106–10; Nuttall 2007b, 99–119. Cf. Lamport 1981a, 163–4; H. Nisbet 2013, 491–2. 172. Nuttall 2007b, 307. 173. See Bond on Euripides, Heracles 513f. 174. Cf. Gellrich 1988, 104–25. 175. See Poetics 1450b27–8, 1454b6–8, 1460a27–32; Barnwell 1982, 219–21; Heath 1991, 394–5. 176. Everson 1990; pace Vernant in Vernant and Vidal-Naquet 2001, vol. 1, 59–61. 177. Sewell-Rutter 2007, 171. See also M. Nussbaum 1986; B. Williams 1993; Hall 1996, 296 (and 2012); Ridley 2003, 414–19; Rutherford 2012, 315– 22; Pippin 2015. 178. Most 2000, 27–8. 179. Griffith 1983, 13. 180. So N. Hammond 1973, 403–8; Fraenkel 1978, vol. 2, 99; Hutchinson 1985, 149. Contra Denniston and Page 1957, xxiii–xxv; Lloyd-Jones 1990, 289. 181. Cf. Dodds 1973, 57; Lesky 1983, 15; Peradotto 2007, 233. 182. On Philoctetes’ freedom, see Krewet 2015, esp. 132–4. 183. Quoted by Jarrett-Kerr 1965, 368–9. 184. For an excellent discussion of this point, see Wolf 1993. 185. So Holbrook 2015, 15–27. On Wittgenstein, see e.g. Appelqvist 2013a, 47; Schönbaumsfeld 2013, 66–7. 186. Pace Fitch and McElduff 2008, 177. 187. See Gaskin 1994b, 570; Hasker 2011, 50. 188. On some of the issues here, see Zagzebski 1991; Gaskin 1994a and 1998a; Hasker 2011. 189. See Gaskin 1997. 190. Aeschylus fr. 350 = Plato, Republic 383a7–b9. 191. Chaucer, Troilus and Criseyde IV, 1406–7; cf. Spearing at Chaucer 1995, 89. 192. Cf. Zeitlin 1982, 19; Garvie 2009, 293. 193. See e.g. Euripides, Medea 679, with Mastronarde ad loc. 194. See e.g. Aeschylus, Seven Against Thebes 584–5, with Hutchinson ad loc.; note also the passage of Artemidorus cited by M. Nussbaum 1994, 53–4. 195. Recall the stories of Hippias (Herodotus VI, 107–8) and Brutus (Livy I, 56, 10–12), the latter retold in Heywood’s The Rape of Lucrece; Deucalion and Pyrrha (Ovid, Met. I, 313–415); cf. T. Gould 1965, 605; Burkert 1991, 8; Flaig 1998, 25–6; Vernant in Vernant and Vidal-Naquet 2001, vol. 1, 97–8; Morgan 2015, 136–7. 196. Rutherford 2012, 349; cf. Budelmann 2000, 109; Garvie 2009, 101.

Oedipus 81 97. Cf. Schein on Sophocles, Philoctetes 610–13; Budelmann 2000, 123–4. 1 198. See Aristotle, Rhetoric 1407a39; Garvie on Aeschylus, Persae 864–7. 199. See Aeneid III, 255–7; VII, 112–19; XII, 834–40; T. Gould 1965, 606; Panoussi 2010, 59; Reed 2010, 68–9. 200. Cf. Garvie 2016, 48. 201. Whitman 1951, 140–1; Knox 1960, 22; 1998, 12–16, 33–41; Kitto 1961, 395; Vickers 1973, 497; Kirkwood 1994, 73, 76, 173; Flaig 1998, 27–8; Lefèvre 2001, 146; Griffith 2005a, 345; Greiner 2012, 43, 51. 202. T. Gould 1965 and 1966; Lloyd-Jones 1983, 217–18 n. 112; Nicolai 1992, 58–67; Peradotto 1992, 8; Nuttall 2007b, 286, 289; Liapis 2012b, 95; Hall 2012, 304, 309; Rose 2012, 256–7, 266. Cf. Winnington-Ingram 1980, 159–60, 178, 202. 203. See e.g. Ahrensdorf 2009, 42. 204. Aeschylus, Seven Against Thebes 750–1, 801–2; Euripides, Phoenician Women 17–20; cf. M. West 1999, 39–40. Kovacs thinks (2009, 366) that the oracle is conditional in Sophocles too. 205. See Aquinas, Summa Theologica Ia q. 83 a1 (1894, vol. 1, 634); Schmitt 1982, 22–3; 1988a, 164–6; 1990, 91–9; Lefèvre 2015, 421–44. 206. Dodds 1951, 7, 16; Willink on Euripides, Orestes 974–5; Lesky 1961, 40–1, 1983. Cf. Segal 1981, 200; Halliwell 1990b; Easterling 1993a, 15; Mastronarde on Euripides, Phoenician Women 621. 207. Heraclitus DK B119; Vernant in Vernant and Vidal-Naquet 2001, vol. 1, 21–40, 66–70. Cf. Winnington-Ingram 1980, 177. 208. Schmitt 1982, 21–2; 1988a, 164–6; 1990, 82–4; 1997, 15, 31. 209. See Heinze 1957, 306; R. D. Williams 1990, 31; Gaskin 1992, 304; Otis 1995, 323–5. 210. See Tarrant 1985, 85. 211. Tarrant on Seneca, Thyestes 261; cf. Boyle ad loc.; Miola 1992, 25; Schiesaro 2009, 225–6; Lefèvre 2015, 340–1, 393–400. 212. See e.g.  Poetics 1454a37–b6; cf. Gillett and Hankey 2005, 274; Görner 2015, 167. Pace Halliwell 1995, 90; Hall 1996, 296. 213. Halliwell 2006, 136–7; 2012, 215. 214. See Garvie ad loc.; 2009, 288; Peradotto 2007, 229–31. 215. See Euripides, Iphigeneia in Tauris 910–11, with L. Parker ad loc. 216. Messenger: 353–4, 362, 373, 454–5, 495–9, 502–3, 513–14; Atossa: 472–3, 724; Darius: 725; Xerxes: 941–3, 909–16; Chorus: 515–16, 904–6, 1005–7. 217. Garvie 2009, 185–6, 192, 205, 210, 215–16, 223, 238–9, 272–3, 287–8; Bowen on Suppliants 511. 218. See here de Mourgues 1967, 116; Kirkwood 1994, 80. 219. Cf. Rozett 1984, ch. 7; Belsey 1985, 43–4; Cooper 2010, 86. 220. Newlyn 1993, 75–80, 155–6; Zamir 2018, 46, 138–9. 221. Gaskin 2013a, 108. Cf. Fitch 1987, 32, on Seneca’s Thyestes and Hercules Furens. 222. See Barrett ad loc., and on 542–4. 223. See Dodds 1960, 62. 224. Dodds 1960, 172; cf. his note on Bacchae 360–3; L. Parker 2007, 252–3. 225. Pace Allan on 887–91, 1017. 226. Winnington-Ingram 1980, 41. Cf. ibid., 27; Segal 1981, 128–9; Cairns 1993, 230; Finglass 2011, 313–14. 227. Scodel 2005, 237–8. 228. Cf. Mastronarde on Euripides, Medea 528, 1013–14. 229. Garvie 2009, 178. Cf. Dover on Thucydides VI, 17, 1. 230. See Adkins 1960, 51–2; Lesky 1961, 40–2; Willcock 1976, 127; LloydJones 1983, 23; Schmitt 1990, 85–9, 95; Redfield 1994, 97; Gaskin 2001,

82  Tragedy and Moral Redress 155; Cairns 2016, 71–3 with n. 42; Dawe on Sophocles, Oedipus Rex 1329; Willink on Euripides, Orestes 3. 231. So e.g. Dodds 1951, 3; Fraenkel 1978, vol. 2, 374. 232. Taplin 1990, 75–6. 233. Lloyd-Jones 1990, 289–90. 234. See Flaig 1998, 36–7. 235. Dover 1987, 136–8. 236. See Fraenkel ad loc.; J. Gregory 1997, 97. 237. See also I, 1, 105–12; III, 2, 171; IV, 1, 170–5; Nuttall 2007b, 132. 238. Cooper 2010, 218–19. 239. Cf. Budelmann 2000, 171–5; Lurje 2004, 244–9. 240. Antigone 1272–4 with 1317–20. See Linforth 1961, 244, 257. 241. Pace Dawe on Oedipus Rex 1329. 242. Cf. A. Brown 1983, 32; R. Parker 1999, 11–14. 243. See A. Brown 1983, 13–27; Garvie 1986, 316–18; Sommerstein on Eumenides 46–59. 244. See Tarrant on Agamemnon 759ff. 245. See on this Buxton 1980, 22; Goldhill 1986a, 183; 2012, 41; Heath 1987, 165–6; Pucci 1994, 18–22; Finglass 2011, 137–8. 246. See Heracles 1229, 1279, with Bond’s notes ad locc.; Kamerbeek 1966; Wilamowitz 1984, vol. 2, 128; Girard 2010, 65–8; Fitch 1987, 32 (on Seneca’s Hercules). 247. See e.g. J. Gregory 1977, 264–8. 248. See n. 148 above, and Wilamowitz’s notes on Heracles 566, 569. Cf. Bond on 562–82; Dodds 1973, 83. 249. See Wilson 2004, 77–80; Pucci 2016, 84. 250. On this debate, see Lloyd 1984; Goldhill 1997b, 145–50; J. Gregory 1997, 170–6; Lee 1997, xxiii, and notes on 987–8, 1038, 1043; Mossman 2005, 357–63; Rutherford 2012, 159; Pucci 2016, 34–49. On Gorgias’ encomium to Helen, which is relevant here (esp. §6), see Goldhill 1986a, 234–5; Rutherford 2012, 54. 251. See W. Kühn 1971, 92, 111–12; O’Hara 1990, 67–8; Mackie 1991. 252. C. Renger 1985, 41. 253. See Denniston and Page ad loc.; Denniston 1939, xvii; Bremer 1969, 120; Garvie on Aeschylus, Persae 354; Finglass on Sophocles, Electra 199, 528. 254. Cf. Euripides, Children of Heracles 989–90, with Allan ad loc. 255. See N. Hammond 1973, 397–9; Dodds 1973, 56. Contra Denniston 1939, xvii. 256. Garvie 1986, xxviii. See also xxix–xxxiv, and note on 435–7. 257. On the terminology of responsibility, see Dodds 1973, 60 n. 2; Fraenkel on Aeschylus, Agamemnon 811 (though he misses the vital asymmetry, on which I have been laying emphasis); Garvie on Choephori 910, 911; Stevens on Euripides, Andromache 1005–6; Willink on Orestes 3, 31. 258. See Hutchinson 1985, 148 and notes on lines 689, 690; Winnington-Ingram 1983, 37, 51; R. Parker 1997, 153. Cf. Sophocles, Philoctetes 1118–20, with Schein ad loc. 259. Homer, Odyssey I, 32–4; Garnier, Marc Antoine 466–82 (cf. 1136–69); Hughes, The Misfortunes of Arthur (in Cunliffe 1912) I, 4, 110–12; II, 3, 127–30; (cf. III, 1, 211–20; III, 4, 18–23; V, 1, 121); Goethe 1985–99, vol. 5, 548; Hardy, The Dynasts III, 6, 3 (1965, 468; cf. 505, 519); 1967, 446. 260. See e.g. Aeschylus, Persae 164 with Garvie ad loc. Contra Rose 2012, 258. 261. See Finglass’s note ad loc. for further references; Winnington-Ingram 1980, 18. 262. Cf. Garvie on Aeschylus, Persae 950–3. 263. Hoxby 2015, 80; cf. Orgel 2002, 162; McLuskie 2003, 399–400; Shapiro 2015, ch. 10.

Oedipus 83 64. See B. Williams 2002, ch. 2; Gaskin 2006, 39–40. 2 265. Shapiro 2015, 225. 266. Halliwell 1996, 347. 267. Hall 1996, 296. 268. Schelling 1856–61, I, vol. 1, 336; vol. 5, 693–711. Cf. Silk and Stern 1981, 305–12; Steiner 1984, 2; Most 1993, 160–1; Flaig 1998, 17; Goldhill 2012, 148, 175–6; 2014, 310; Young 2013, 80–5; Hoxby 2015, 14–26; Leonard 2015, 50–6. 269. Schiller 1992–2004, vol. 8, 260–1. Cf. ibid., 395–422; Goethe 1985–99, vol. 19, 642–5. 270. See Coleridge 1987, vol. 1, 458; Raymond Williams 1966, 20–1, 26, 30–1; Billings 2014, 98–100; Hoxby 2015, 250. 271. Cf. Lamport 1990, 98–9; D. Schmidt 2001, 75; Lambropoulos 2006, 30–41; Billings 2014, 75–6, 80–97, 103; Hoxby 2015, 14–15; Leonard 2015, 13–27. 272. Goethe 1985–99, vol. 18, 11; cf. Lamport 1990, 36. 273. Schmitt 1988a, 187; see also his 1992. 274. Schiller 1992–2004, vol. 8, 242–3, 260. 275. Heisenberg 1996, 218; cf. Schmitt 2015, 205. 276. Goethe, Götz von Berlichingen V (1985–99, vol. 4, 382, 386). 277. Büchner 2009, 99. On Büchner’s Danton see Szondi 2011, vol. 1, 256–7. 278. Schopenhauer 1986, §51, vol. 1, 355. So too Hughes’s Misfortunes of Arthur and Gager’s Dido, which Norland thinks are tragedies of fate (2009, 96, 98, 180). 279. Schulze 2015, 456–61. 280. See Bullough 1973, vol. 7, 478; Clark and Mason 2015, 87. 281. Pinker 2002, 322. 282. See Lloyd-Jones 1983, 118. 283. See Schmitt 1997, 34–5. 284. Though the literal meaning may be ‘in dreams as well as in the oracles’: cf. Pucci 1992, 96–7; R. Armstrong 2012, 488–9. 285. See Nickau 1994, 19. 286. See Dawe on Oedipus Rex 915. 287. Hofmannsthal 1979, vol. 5, 41. 288. See Reckford 1961. 289. See Finglass 2007, 288; Goldhill 2012, 17–18, 58. Cf. also Antigone 1183–5. 290. See the commentaries of Tarrant and Boyle ad loc.; Schiesaro 2003, 85–98. 291. Dawe, note on 924; cf. Winnington-Ingram 1980, 180–3; Segal 1981, 236; Goldhill 2012, 13; Rutherford 2012, 347; Rowan Williams 2016, 22. 292. Euripides fr. 541; Burian 2009, 100 n. 8. 293. B. Williams 1993, 68–72, 133–5; cf. Gill 1996, 91. 294. See Greiner 2012, 142–3. 295. Cf. Cairns 1993, 216–19; Kullmann 1994, 114. 296. Sophocles, Oedipus Rex 1334–5; Wilmot, Tragedy of Tancred and Gismund, H3v; Shakespeare, King Lear IV, 1, 21. Cf. Nuttall 1996, 103; Lefèvre 2001, 140; contra Manuwald 1992, 30. 297. Garvie 2016, 43. 298. See Buxton 1980, 24–5; Lyne 1987, 194 with n. 62.

2 Antigone’s Holy Crime

10 Saint Antigone? I suggested in the previous chapter that the majority of Greek tragic protagonists exhibit moral flaws, usually of a cognitive nature, contrary to a standard modern view that they are morally innocent. Typically, tragic protagonists commit hamartiai which flow in a natural and comprehensible way from character defects that these agents possess, and they are punished for their errors in the denouements of their respective tragedies. The relevant errors, and moral flaws, can be thought of as redressing, compensating for, the punishments they receive (recall that I am employing a temporally neutral concept of redress). In the list that I gave in §6 to illustrate my claim, I included the two main protagonists of Sophocles’ Antigone, namely Antigone herself and Creon. In this chapter I want to justify the application of that claim to these figures, and particularly to Antigone. For though most spectators and readers will grant that Creon is in some sense at fault, and is punished for his transgressions, Antigone has been widely held to be without sin, a saint avant la lettre.1 So S. H. Butcher, though remarking, in line with what I have just stated to be the general tendency of Greek tragedy, that Aristotle’s ‘reluctance to admit a perfect character to the place of the protagonist has been almost justified by the history of the tragic drama. Such a character has been rarely chosen, and still more rarely has been successful’, nevertheless cashes in that ‘almost’ and those ‘rarely’s by making an exception of Antigone: ‘Nothing but a misplaced ingenuity, or a resolve at all costs to import a moral lesson into the drama, can discover in Antigone any fault or failing which entailed on her suffering as its due penalty’. And this has been the dominant view of Antigone in the tradition. A. W. Schlegel lauded ‘the pious sanctity of her mind’; for Hegel ‘the heavenly Antigone’ was ‘the most glorious figure ever to appear on earth’; Thomas De Quincey called her a ‘holy heathen’ and ‘Christian lady’; Richard Jebb, affirming that ‘the right is wholly with her, and the wrong wholly with her judge’, agreed that she bore comparison to ‘a Christian martyr under the Roman Empire’, and even averred that ‘Nowhere else has the

Antigone’s Holy Crime 85 poetry of the ancient world embodied so lofty or so beautiful an ideal of woman’s love and devotion’; Hugo von Hofmannsthal wrote that ‘this radiant creature belongs to no age’ (‘Dies strahlende Geschöpf ist keines Tages’).2 Sophocles’ Antigone does not have a monopoly of sanctity: we are told that Garnier’s Antigone, too, is ‘a saint who suffers martyrdom for her special virtue of piété’.3 The acute early-nineteenth-century critic and theorist August Böckh indeed objected to this tendency, stating that ‘the poet was very far from glorifying her unconditionally’, but his remained, and still remains, a minority voice. For in our own time, accolades have continued to pour in from the critics: Cedric Whitman tells us that ‘It is useless to speak of the defects of Antigone’s qualities; there are no defects’; Walter Kaufmann holds that Antigone ‘has no blemish, and our sympathies are not divided between her and Creon’; D. W. Lucas insists that ‘There can of course be no question of a hamartia on the part of Antigone’; Brian Vickers holds that ‘Antigone is presented as an admirable, committed character who is never criticized’; Ruth Scodel agrees that ‘Antigone is entirely right’; for Walter Nicolai, too, Antigone ‘hat das Recht . . . vollkommen auf ihrer Seite’; ‘Who but a bigoted nationalist,’ demands Hugh Lloyd-Jones, ‘and one grossly deficient in aesthetic sensibility, would have argued that Creon and Antigone represented moral viewpoints of equal validity?’, and he entrusts us with the (possibly unwelcome) confidence that ‘The attitude of those who contend that Antigone is hardly, if at all, less to blame than Creon makes me even angrier than the attitude of those who contend that Oedipus is punished for his own crimes’; Patricia Mills writes that ‘Against Hegel’s interpretation, Sophocles does not create Antigone and Creon as ethical equals. Antigone alone is the ultimate defender of the good’; Dennis Schmidt, while conceding that her strange motives render Antigone ‘a mystery to us’, still asserts that she is ‘very much an ideal by virtue of her steadfast commitment to her own nature’.4 Interestingly, there have been some dissenting views among philosophers: Heidegger translated the word ‘deinon’ of the second choral ode as ‘unheimlich’ (‘uncanny’), and explicitly associated the adjective with Antigone5—as perhaps the chorus intends, and as Antigone herself does in Aeschylus’ Seven Against Thebes, at 1031–2, part of a passage which was almost certainly interpolated and which was probably influenced by Sophocles.6 Lacan, though shying away from calling Antigone ‘monstrous’, noted that the chorus does call her this, and stated that she ‘goes outside human limits’ and is ‘beyond Atè’. Lacan’s debunking reading of Antigone is picked up by Slavoj Žižek, who speaks of her stance as ‘no longer human’ and ‘effectively “mad”, disruptive, evil’; and Žižek does not scruple to refer to Antigone’s ‘monstrosity’. Against this incipient trend, some recent feminist writers have come to Antigone’s rescue, arguing that she is a role model in the struggle against the patriarchy.7

86  Tragedy and Moral Redress

11 Questions of Right and Wrong The action of the Antigone takes its cue from the dispute between Eteocles and Polyneices, sons of Oedipus, who fight over the throne of Thebes. Both are killed, but Eteocles is killed defending Thebes while Polyneices was the aggressor: so Creon, who succeeds to the throne, decides to honour the former’s patriotism by burying him in state, but make an example of the latter’s treason by denying him burial. The play begins with a dialogue between Antigone and Ismene, daughters of Oedipus, in which Antigone reports Creon’s edict refusing burial to Polyneices, and condemning to death anyone who contravenes it. Polyneices’ body is to be left as prey to the dogs and birds. Antigone resolves to bury her dishonoured brother and accept the consequent penalty. Accordingly, the audience is at once confronted with the question what to make of Creon’s edict. Several commentators have argued, I think persuasively, that in issuing his edict Creon, as king of Thebes, acts within his rights.8 He appears initially to be supported by the chorus, which expresses relief that the threat of Thebes’ destruction which Polyneices had held out has been lifted (100–61, 332–75).9 In attacking Thebes, Polyneices became a traitor.10 The denial of burial to traitors was quite normal in archaic and classical Greece,11 though the precedent of Sophocles’ own Ajax, in which Odysseus sues for and secures Ajax’ burial, despite the latter’s treachery, shows that a more humane attitude was possible.12 In fact, not only treason but also sacrilege and tyranny could be punished by exposure of the corpse, and there is a good case (which Creon makes) for indicting Polyneices on all three counts.13 It has been suggested by Robert Parker that Polyneices is presented in the Antigone as a mere enemy of Thebes rather than as a traitor, so that the right of burial standardly permitted to enemies14 should apply in his case, and therefore that Creon’s decree is to be condemned.15 But Parker’s statement that nothing in the play prompts us to view Polyneices as a traitor, rather than as ‘a hero who has led an army to defeat’ (1983, 48) seems to me a clear error. It is true that Polyneices is not called a traitor, even by Creon; but if you lead an army against your native land you are eo ipso a traitor, and it is not necessary for anyone to point this out.16 (In the tradition, Polyneices’ expedition had been condemned by the seer Amphiaraus.)17 Polyneices is a traitor, then, but does Creon’s ban go beyond what is reasonable? It appears that strict Attic law forbade the burial of traitors only within Attic territory, not outside it,18 so that in issuing a general ban on Polyneices’ burial Creon might seem to exceed his authority: perhaps he should have arranged for the body to be cast outside the bounds of Thebes, thereby permitting its burial in a remote location?19 Nothing is made of this point in the play, however, as one would expect had it been relevant.20 (Contrast the stress laid on the remoteness of Antigone’s crypt.)21 The play is after all set in prehistoric

Antigone’s Holy Crime 87 Thebes, not in historical Athens, so that the conventions obtaining in the latter cannot without more ado be applied to the former;22 and in the heroic world leaving even an enemy’s (let alone a traitor’s) body to the dogs and the birds is normal practice, as is implied by the opening lines of the Iliad and many other passages.23 It is true that burying the dead was a panhellenic nomos,24 and that the Iliad records not merely the phenomenon of the exposure of dead bodies to the dogs and birds, but also the Greek horror of that practice.25 When, in Aeschylus’ Suppliants, the Danaids welcome that prospect in preference to marriage with their cousins, it is made plain that they are going to an unnatural extreme,26 and it appears that in the third part of the trilogy Aphrodite preached marriage and reconciliation.27 It must also be conceded that, in the Iliad, although there is much talk of leaving bodies to the dogs and the birds, the only case of it that we actually see is that of Achilles’ refusal to allow Hector’s burial, which elicits strong disapproval from the gods.28 (In Gregorio Correr’s humanist tragedy Procne, the horror of the Thyestian feast that Procne prepares is made vivid by the stress on how much worse it is than exposure of the dead body of her son Itys to wild beasts, and in general the dreadfulness of lack of burial is a theme echoed throughout the tradition.)29 Consequently some critics take the view that Creon’s decree is meant to disturb us from the first moment of the play, on the grounds that prolonged exposure of the dead was considered by the Greeks to be impious.30 But the supernatural dimension here, as elsewhere (§8), was driven by facts of human morality and psychology, which is why it was possible to make exceptions for traitors and certain other classes of criminal.31 Exposure of the dead was permitted for these classes of criminal, and if we were supposed to think that Creon ought to have imposed only a restricted ban, we should, as I have said, expect something to be made of this in the play: it could have been pointed out by any of Antigone, Haemon, Teiresias, the chorus, even by the sentry— perhaps jocularly, but that would have been enough. Yet no one makes anything of it. (The reader will have noticed that, in this and the previous paragraph, I have taken opposite lines on the significance of two omissions. Polyneices is not explicitly called a traitor in the Antigone; but that, I said, is because it is obvious that he is one, so that, although the fact is highly relevant, it does not need to be mentioned. By contrast, I have maintained that Creon’s refusal to cast Polyneices’ body beyond the borders of Thebes is both not mentioned and irrelevant. There is no algorithm for determining the significance, in general, of omission: sometimes we are meant to presuppose omitted matter, sometimes ignore it. I shall return to this point in Chapter 4 below.) It would seem, then, that the sheer issuing of the ban is defensible, at least at the time. But it becomes clear as the play unfolds, and especially after Teiresias’ intervention, that Creon’s stubborn refusal to modify his edict, is, like the similar intransigence of Menelaus and Agamemnon in

88  Tragedy and Moral Redress the Ajax, an error—with the Ajax in mind, we will find Creon’s statement that an enemy is never a friend, even when dead (522; cf. 643–4) to be already indicative of his cognitive failure—for which he is punished at the end of the drama by the death not only of his niece Antigone but of his wife Eurydice and son Haemon too. Creon cramps himself in embattled and wrong-headed obduracy. He also makes a crucial tactical error in electing to bury Polyneices first, before trying to rescue Antigone.32 Had he taken these tasks in reverse order, Sophocles’ Antigone would, as we noted earlier (§7), have been a play in the ‘catastrophe survived’ genre. Ironically enough, Creon resembles Antigone (as we shall see) in placing the dead before the living.33 His mistakes are accordingly, from the interpreter’s point of view, fairly straightforward in nature—they are of the usual cognitive cast—and this has been widely recognized.34 His intransigence in the face of opposition to his edict arises because he is obsessed with his own authority and is unable to tolerate challenges to it. Initially, when he speaks of nomos he means the law of the state (177, 191), but the word soon comes to be identified in his mind with his edict (287, 449, 481, 663), and this tendency is exemplified by his statement to Haemon that the city is the ruler’s property (738). Later, he learns better (1113–14).35 Creon’s fixation on his own supremacy is, at least to begin with, understandable, given his precarious position as the new ruler of a state that has just survived the threat of destruction. It is natural in this situation for a king to be unwilling to make compromises with potential political rivals who might well have an interest in dislodging him. What Creon fails to see, however, is that the challenge represented by Antigone’s urge to bury her brother is not a move in a game of realpolitik,36 and that, contrary to his intemperate assertion that he has been nurturing ‘two disasters and revolutions against my throne’ (533), the sisters do not offer a serious threat to his government. Creon overestimates the extent to which compromise on the question of Polyneices’ burial would really undermine his authority or the stability of the state. So Antigone is not part of any plot to upset Creon’s rule: her reasons for action, which we shall come to shortly, are of a purely personal nature. In effect, it is Creon himself who politicizes Antigone’s act.37 Though at a superficial level the play comes across as a contest between polis and oikos (and this is how many critics, from Hegel onwards, have taken it),38 at a deeper level, as we shall see, Antigone is not trying to shift the balance of power between family and city in favour of the family.39 Nor is she a political dissident, standing up against a Gewaltherrschaft.40 (This latter construal is a common one, especially in the area of adaptation and performance: see §14 below.) Nor is there a simple correlation, such as Bonnie Honig proposes,41 between Creon and democracy on the one side, and Antigone and aristocracy on the other.42 Indeed if you were connecting up Creon and Antigone one-to-one with democracy and aristocracy, you might as well do it oppositely from the way Honig suggests: after

Antigone’s Holy Crime 89 all, Creon’s method of proceeding is notably anti-democratic, by contrast with Pelasgus’ in Aeschylus’ Suppliants (607), or Theseus’ in Euripides’ Suppliants (404–8),43 and Antigone initially claims the support of the people of Thebes (504–5, 509)—though whether one should believe her is another matter, and she retreats from that initial position later (907), significantly echoing Ismene’s earlier objection to her (79). In the (probably spurious) ending of Aeschylus’ Seven Against Thebes, Antigone states explicitly that her action will involve anarchia and defiance of the city.44 Sophocles has Haemon claim (692–700 and 733) that Antigone enjoys popular backing, and many commentators believe him;45 but others have a more cautious assessment of the likely trustworthiness of his testimony on this point.46 If Creon overestimates the risk to his authority of accepting a compromise, he also underestimates the extent to which he himself depends emotionally upon familial relations: he makes the mistake of identifying himself too closely with his political role.47 James Redfield notes that, in the Iliad, ‘Achilles (who in this point also resembles Coriolanus) brings the catastrophe upon himself because he is not, after all, as detached from his community as he had supposed’ (1994, 106), and a similar point applies to Creon, whose devotion to city in preference to family is reversed during the course of the play.48 One may compare Euripides’ Phoenician Women, in which the loss of his son Menoeceus puts Creon in a posture of private grief, parallel to his loss of Haemon in the Antigone;49 but in Euripides the private grief precedes the fateful edict against the burial of Polyneices. Sophocles’ Creon, then, fits the classic Aristotelian pattern of a tragic hero who commits errors, and ones which, like Oedipus’ hamartiai, are of a basically cognitive nature, so that his faults, as he himself finally recognizes (1265), amount to a kind of intellectual failure—his failure to appreciate the politically unthreatening nature of Antigone’s protest, his failure to understand the value to himself of his own domestic concerns, and his failure to recognize and reverse the wrongs he has committed quickly enough (and in the right order)—for which he is duly punished. Creon is not evil, as Schopenhauer, who aligned him with Phaedra, Richard III, Iago, Shylock, and Franz Moor (1986, §51, vol. 1, 355), thought; but the catastrophe and the punishment arise from his offences in an entirely comprehensible, Aristotelian way. J. M. Bremer denies that the notion of hamartia applies to either Antigone or Creon (1969, 140, 145); but that is because he gives too narrow a reading of hamartia, restricting its range to a small area of factual ignorance (ibid., 99). If, following my exposition of the doctrine of hamartia in Chapter 1, we have a broader conception of it, according to which it incorporates both moral and intellectual error—and recall that for me moral error is a (proper) subcategory of intellectual error—then that concept does apply to Creon. And, as we shall now see, it also applies to Antigone. Rowan Williams suggests (2016, 66) that ‘the great figures

90  Tragedy and Moral Redress of classical tragedy know (all too well) how to be themselves’—and he instances Creon and Antigone. This claim, which is a cliché of Sophoclean hermeneutics, is wrong about Creon, as we have noted; I shall now attempt to show that it is also wrong about Antigone.

12 Antigone’s Real Motivation Why does Antigone defy Creon’s ban and bury Polyneices? Her ostensible reason is that this action is required of her by ‘the unwritten and unfailing statutes of heaven’ (454–5, tr. Jebb). For as Polyneices’ sister she claims that she has an inevitable obligation to bury him. I noted in the last section that it would be a mistake to draw simple correlations between Creon and Antigone, on the one hand, and democracy and aristocracy, on the other, either way round. And it would, I suggest, be equally naïve to think, as some commentators do,50 that we have to do here with a simple conflict between law (nomos), represented by Creon, and nature (phusis), represented by Antigone. For again, if we were dividing up nomos and phusis one-to-one between Creon and Antigone, we might as well do it the other way round: the principle that Antigone appeals to is that of divine law, whereas Creon, in devising his edict, bases it upon the unnaturalness of bestowing the same degree of honour on the traitor Polyneices as on the patriot Eteocles. (The unnaturalness of Polyneices’ assault on his native city is powerfully conveyed in Jocasta’s great speech in Seneca’s Phoenician Women, at 480–585, and in Iocaste’s appeal to Polynice in Garnier’s Antigone, at 810–41—this passage being modelled in part on the Senecan one.) But to set up an exclusive one-to-one correspondence either way round would again be erroneous, since both Creon and Antigone appeal to both law and nature, just as in a sense they are both right and both wrong too, as Hegel observed.51 Equally—and this is also a point we can extract from Hegel52—it is not a simple matter of reasons of state on the one side confronting divine reason on the other, for in honouring Eteocles above Polyneices Creon sees himself as favouring the brother who defended the local shrines of the gods against the brother who would have sacked and despoiled them (110–47, 199–200, 284–7): that he was a potential hierosulos is indeed, as we have mentioned, an important aspect of Polyneices’ criminality.53 In any case, in promoting the interests of civic law Creon might be said to be defending the prerogatives of divine law, given that civic law was held to be grounded in a divine dispensation.54 So Creon as well as Antigone acts on principles that ‘can lay legitimate claim to divine sanction’.55 Antigone is in some sense vindicated by the denouement of the play, and particularly by Teiresias’ intervention; but it is too simple to state that ‘Antigone is entirely right’,56 or infer from the fact that ‘Creon was wrong’ that ‘Antigone must have been right’,57 and leave matters there. On the question whether or not Polyneices should be buried, Antigone

Antigone’s Holy Crime 91 turns out to have been right, but two general points need to be made by way of extenuation of this fact. First, although Antigone’s position on the sheer question whether Polyneices should be buried is in the event vindicated, that was not inevitable. If things had not gone badly in the way Teiresias describes, Creon would have been right on the issue, or at least not wrong. His starting position—the decree—was, as we said, not unreasonable, and Wouter Oudemans and André Lardinois are gesturing towards a truth when they say that Creon is ‘overruled rather than refuted’ (1987, 199). As it stands this assertion is not quite satisfactory: we do not need to, and should not, embrace any kind of anti-realism about past moral judgements; that is, we do not want to say that Creon’s decree was not wrong at the time, but only became wrong later. It was wrong at the time—so it is refuted, not just overturned—but the point is that its wrongness did not become clear until Teiresias’ intervention. Until that juncture Creon’s decree was morally defensible: it was in fact wrong but justifiable in its context, or not unjustified. Creon’s hamartia does not consist so much in the promulgation of his decree as in the way he behaves subsequently—his intransigence, his failure to grasp his own domesticity—and if he had changed tack at a suitable point, for example in the Haemon scene, or even, as we have noted, if he had rescued Antigone before burying Polyneices, tragedy would have been averted.58 Secondly, and more importantly for our purposes here, though Antigone is vindicated by the event, she is wrong to oppose Creon, and her deep motivation for her actions is misguided. I take these two points in order. Antigone is wrong to oppose Creon, for although the play supports her view that Polyneices should be buried, it does not follow that she should bury him.59 When Patrick Finglass writes that Antigone ‘does not see her act as morally ambivalent; and in the event, the gods, via Teiresias, endorse it’ (2007, 175), we need to ask what ‘it’ is: the gods endorse Polyneices’ being buried, to be sure, but do they endorse Antigone’s burying him? That is a different question, and the answer, I suggest, is negative. It was not up to Antigone to bury Polyneices.60 As a young unmarried woman, under the legal guardianship of Creon,61 Antigone has no status to challenge his authority, and no status to purport to speak on behalf of divine ordinances against the authority of the polis, as embodied in its ruler: the city and its ruler are the legitimate interpreter of religious obligations, not an unmarried girl who is, as R. P. Winnington-Ingram says, ‘little more than a child’ (1980, 241).62 To be precise, she is an adolescent. This is an important point, often missed: Antigone is repeatedly called a pais in the play; but youth means folly.63 Antigone has no status, either, to bury Polyneices, for burial is a male prerogative—that is why Creon automatically assumes that a man has done the deed (248)—and she has no status even to lament his body on her own, for this action, though a female prerogative (along with bathing, anointing, and dressing the corpse), should be undertaken communally.64 These points tend to be ignored or even

92  Tragedy and Moral Redress denied by commentators.65 Edward Harris states that ‘The Athenians who watched the play would not have found Antigone’s disobedience to Creon’s order dangerous or unusual conduct’ (2012, 290), which is the reverse of the truth. (He goes on to cite possible disobedience of the Ephebic Oath, and Socrates’ disobedience of the Thirty, which are hardly models for Antigone’s case.) So far from addressing Creon humbly, as etiquette would demand, she treats him as a subordinate, which is entirely inappropriate to her status. Antigone’s proper career is to marry, as soon as she comes of age,66 and procreate, not to interfere in male matters of public policy, still less to martyr herself.67 As an aristocrat, she is expected to take a suitable husband, have children, and organize a large domestic outfit—to do precisely the things that Sophocles’ Electra regrets that she is unable to do.68 (In the Hofmannsthal/Strauss Elektra, these regrets are expressed by Chrysothemis; in the Euripidean Antigone, Antigone apparently survived, married Haemon, and bore him a son.)69 Putting it in Aristotelian terms, that is her ergon, her telos, fulfilling which would alone mean eudaimonia for her.70 Antigone is brave and clever, ‘but it is not suitable for a woman to be brave or clever in this way’.71 Are we are alerted early in the play to Antigone’s transgression by the fact that she meets Ismene outside the courtyard (18–19), where the two sisters have no business to be? This is a notorious crux, on which social and stage convention might seem to be in tension. For my purposes no stand need be taken on the issue: the play provides quite enough evidence for a ‘bad woman’ construal of Antigone, without our having to invoke this point.72 But mention of the relation between social and stage conventions raises an important issue that I do need to address. For it is sometimes objected to arguments of the sort that I propounded in the last paragraph that they presuppose too crude an understanding of the relation between social mores and dramatic reality. In connection with the Antigone, for example, Douglas Cairns (2016) suggests that we should resist finding a neat correlation between fifth-century BCE social mores and what happens in the depicted world. And it has often been observed that, in general, women are more outspoken and more self-assertive on the Attic stage than contemporary mores permitted in real life.73 This latter point must be agreed. Nevertheless, it seems to me that Antigone’s defiance goes too far, even granted the relative freedoms of the stage. Cairns writes that ‘it is one thing for a woman to rebel against the head of her household in normal circumstances or where that person acts as he should, but another to do so in abnormal circumstances in which the head of the household is demonstrably wrong’ (ibid., 42). Does Cairns mean that the circumstances are abnormal anyway, or does he think that their abnormality is constituted by the fact that Creon is ‘demonstrably wrong’? If the former, his point has, in my view, no weight: heads of households exercise their authority in all circumstances; indeed you might say that abnormal circumstances are precisely those in which

Antigone’s Holy Crime 93 their authority matters most. If the latter, then Cairns is ignoring the fact that, at the point in the drama when Antigone defies him, Creon is not demonstrably wrong. Creon is not demonstrably wrong until the Teiresias scene, by which time Antigone has left the play. The burden of Teiresias’ intervention is indeed skilfully anticipated by Sophocles in the Haemon scene, but that anticipation does not reach as far as showing that Creon is wrong, though it strongly suggests it. It cannot show it, however, because Haemon is, as I have implied (§11), not an unbiased witness: we do not yet know whether to believe his claim that the citizens are on Antigone’s side. Only when Teiresias speaks do we come to know that Creon is wrong. Given all this, why does Antigone act as she does? What is her motivation? Tina Chanter states that ‘a long tradition has neglected to take seriously Antigone’s own explanation of her action’ (2010, 21 n. 9), but my impression is that the opposite is true. It is rather the case that critics have too readily fallen for Antigone’s own representation of her motives for action, as though that were the final word on the matter. D. A. Hester asks ‘Why is it so important to her to bury Polyneices?’ and replies, innocently, ‘We may allow her to tell us’, continuing: ‘the refusal of burial is the last of a long series of insults cast upon her family (1–10)’ (1971, 21). Helene Foley opines that ‘Antigone’s motive for action is throughout a deeply felt personal responsibility both to bury blood relatives in the same household, and to honour the gods below with their due’ (2001, 173). G. M. Kirkwood, although he correctly descries that ‘Antigone’s first concern is not for religious duty, which looms so large in her scene with Creon’, nevertheless continues: ‘Her first reaction is a personal one; the matter is one of family loyalty, where, she feels, Creon has no right to intrude’ (1994, 119). Shortly after he adds: ‘To Creon’s clumsy interference with her duty to her family, she responds with instinctive hostility. She is furious that Creon should seek to legislate to her in a matter so personal to her: “Such conditions they say the worthy Creon has proclaimed for you and me—yes, even for me!” (32–3)’. Against this one may object that, if family honour really were Antigone’s motive, there would be no justification for her ‘yes, even for me’—for family honour affects Ismene as much as it does Antigone—and this should have alerted Kirkwood to the possibility that Antigone’s professed motive might not be her real one. Matthew Santirocco is cannier: he first lists Antigone’s various statements of her motives—these are ‘the gods’ displeasure at an unburied corpse, the righteousness of Polyneices’ cause, his unique position as her brother and thus more special than a husband or son, her own desire to be with the dead, her hope of glory, her loving nature’—and he correctly notes that ‘it is in the apparent confusion of motives that the real reason for her behaviour is revealed as something other than a concern for the gods and their justice’ (1980, 187). But, having started so well, even he succumbs to the orthodox line and concludes that the ‘real reason’ (ibid.,

94  Tragedy and Moral Redress 188) for her actions is her love for her brother.74 He also agrees with Lloyd-Jones that the family curse is operative in the background.75 The family curse, and in general the idea of atē, does figure prominently in the play, but, as in the cases of Sophocles’ Oedipus and Homer’s Agamemnon, not in such a way as to diminish agent responsibility.76 Arbogast Schmitt accepts that Antigone is not ultimately concerned with religious precepts concerning the burial of the dead, but still concludes, again accepting Antigone’s own words at face value, that it is the dishonouring of a kinsman that affronts her.77 There are, I suggest, two principal reasons why the explanations that are, following Antigone’s own statements, most often proffered as accounting for her action cannot be right—that is, cannot give us a deep understanding of her behaviour. First, if she were really motivated by respect for divine ordinances, family dishonour, or her love of Polyneices (with or without his uniquely irreplaceable status), that is, selecting some of the items from Santirocco’s list, if one or more of ‘the gods’ displeasure at an unburied corpse, the righteousness of Polyneices’ cause, his unique position as her brother and thus more special than a husband or son, . . . her loving nature’—if, I say, one or more of these things really explained Antigone’s actions, then she would not, in the opening scene, tell Ismene to publicize her deed (86–7): for obviously if word of her intention spreads, that is likely to stymie her project. If Antigone were really interested simply in ensuring her brother’s burial, she would agree to Ismene’s recommendation of secrecy. As Seth Benardete comments: ‘Had Ismene taken her at her word [in 86–7], Antigone would have failed at her first attempt. . . . Antigone seems to regard it as essential that she be caught and as inessential that she succeed’ (1999, 16). This remark points us in the right direction: for the truth is that Antigone does not want to bury Polyneices simpliciter; it is rather the case that she wants to be caught burying him.78 Secondly, if the motives listed above had been Antigone’s real motives, she would not refuse (538–9), and refuse with such arrogant superiority, in such ‘disconcertingly selfish and hurtful terms’,79 Ismene’s offer of complicity (which initially she had sought, 41–3) and the sharing of guilt; nor would she reject with such scorn and bitterness Ismene’s plea to share her sister’s death (544–60). If these items really had been some or all of Antigone’s motives, we should have expected Ismene’s offer of solidarity in death to be welcomed: Men seyn, ‘to wrecche is consolacioun To have another felawe in hys peyne’.80 Ismene’s offers of partial responsibility for the burial and companionship in death should at worst be a matter of indifference to Antigone, not things to be rebuffed so radically.

Antigone’s Holy Crime 95 Mark Griffith suggests that Antigone’s motives in rejecting Ismene’s request are (i) a desire to save her sister from death, and (ii) a wish to preserve full credit for herself.81 But the former motive is hardly present: it is at best marginally detectable in lines 546–7, where Antigone says ‘Do not try to share my death, and do not claim as your own something you never put a hand to! My death will be enough!’ (tr. Lloyd-Jones), and which Griffith cites in support of (i); in line 553 (Antigone: ‘Save yourself! I do not grudge you your escape’; tr. Lloyd-Jones), which he also cites, Antigone’s instruction to Ismene to save herself is clearly sarcastic, not loving. In response to the even more mocking line 549, where Antigone suggests, absurdly and callously, that the person Ismene really cares about is Creon, her sister replies: ‘Why do you give me such pain, when it does you no good?’ (tr. Lloyd-Jones). That is indeed a pertinent question: why does Antigone consistently treat Ismene with such hostility? An extraordinary mixture of solipsism and zealotry is displayed in Antigone’s refusal of Ismene’s assistance and even presence in the enactment of her high calling,82 and the question is why she refuses these offers, if all that matters to her is that Polyneices be buried, or that family dishonour be expunged, or that the gods’ commandments be obeyed. The question answers itself: Antigone’s attitude to Ismene shows that, in point of fact, these things are not what matter to her.83 And, as Eckard Lefèvre observes, if one had been hoping that the speech about the unwritten laws would be delivered ‘in celebratory seriousness’, one would be disappointed: even these words are ‘decked out with derision, bitterness, contempt, and irony’ (2001, 101). Again, why? Antigone claims to be acting out of love for her brother, but, as just noted, she treats her sister as an enemy (e.g. 86, 93). Thereby she undercuts her own most famous line (523): ‘It is not my nature to join in hating (sunechthein), but in loving (sumphilein)’ (tr. Jebb, adapted).84 Kirkwood permits himself a nagging doubt concerning whether the line is ‘quite consistent with what we have seen elsewhere’, but the doubt is promptly quashed: ‘Essentially it is so’ (1994, 124–5). Commentators easily fall into awe when they hear this line—‘Antigone is remarkable for her love and lack of hatred’, we are told85—overlooking the way in which its content is undermined by Antigone’s own conduct toward those, including Ismene, whom she identifies as ‘enemies’. Honig’s suggestion that Antigone refuses ‘to take seriously the friend–enemy distinction’ (2014, 30) could hardly be further from the truth. For, like nearly everyone else in Greek tragedy,86 Antigone has an ‘us and them’ (better: a ‘me and them’) mentality:87 her final wish (927–8) is that her enemies may suffer as much as she does. Paul Hammond writes of our line: ‘The two verbs, sunechthein and sumphilein, are unique, found nowhere else in classical Greek, so poignantly Antigone has to coin for herself the words with which she expresses the bonds which link her to another: this is her own lexis, not the vocabulary of any community’.88 This would be convenient for me if

96  Tragedy and Moral Redress it were right, for the uniqueness of the key verbs in Antigone’s sentiment might be held to hint at its self-refutation. In fact, however, I cannot avail myself of this argument, for the philological point on which it is based— the suggestion that the words ‘sunechthein’ and ‘sumphilein’ are ‘not the vocabulary of any community’—is unsound. These two verbs, whether or not attested elsewhere, are formed in a quite unexceptionable way.89 But maybe self-refutation is in any case the wrong way to construe line 523. Earlier, in her opening exchange with Ismene, Antigone had said ‘I shall lie with him, loving him and loved by him, having committed a holy crime’ (73–4). Richard Rutherford remarks that line 523 should be construed with these lines: ‘Antigone’s loyalties are with the dead’.90 It is indeed Antigone’s nature to join in loving: only, the scope of her loving does not include the living; in the case of the living, it is in her nature to join in hating.91 The figure conspicuous by his absence in lines 73–4 is Haemon, the proper object of the expressed sentiment and, ironically, the one whom she will lie with in death.92 Returning, now, to Santirocco’s original list of candidate explanations for Antigone’s actions, the exclusion of pretenders still leaves Antigone’s ‘own desire to be with the dead’ and ‘her hope of glory’ in the field as possible motives for her behaviour, and these suggestions are now coming closer to the mark. They start to answer the question: wherein lies Antigone’s hamartia?93 Böckh, despite being one of the few earlier critics to have perceived that Antigone as well as Creon is at fault, locates her mistake in her breach of human (as opposed to divine) law (1843, 174). But this approach is disappointingly unpsychological. For Antigone pushes the urge to bury her brother to an unnatural extreme, and the question is why she does this. The reason, to put it simply and telegraphically, is that, as Lacan saw, Antigone has a death-wish; she is a rebel looking for a cause to die for.94 A statement like ‘Antigone chooses to act to right what she sees as an intolerable wrong, even if it means her death’95 is back to front: it is her desire for death that comes first and the pretext second. Her death-drive precedes Creon’s decree, which gives her the perfect opportunity to step into the coveted role and achieve what she desires.96 As she tells Creon (460–6): I knew that I would die, of course I knew, even if you had made no proclamation. But if I die before my time, I count that a gain. For does not whoever lives among many troubles, as I do, gain by death? So it is in no way painful for me to meet with this death. (tr. Lloyd-Jones) The chorus says that no one is so foolish as to desire his own death (220), but Antigone proves the contrary. She seeks the glory of a unique martyrdom. That is why she is put out when Creon threatens to involve Ismene (497), and it is also why she urges him to hasten her own end (499).

Antigone’s Holy Crime 97 Creon remarks of the two sisters (561–2) that ‘one of these girls [Ismene] has only now been revealed as mad, but the other [Antigone] has been so from birth’ (tr. Lloyd-Jones).97 This statement, which one might initially be tempted to discount as mere slander, is confirmed by the way it differentiates between Ismene’s and Antigone’s respective follies: for if Creon had simply wished to insult them he would not have troubled to be so precise. And it is also confirmed by the chorus’s characterization of Antigone’s ‘savage breeding from her savage father’ (471–2). ‘Savage’ (ōmos) is a very strong word to apply to Antigone, and it implies that something is seriously amiss with her character.98 Elsewhere the word is used, for example, to characterize Eteocles, Ajax, Clytemnestra, and Death.99 In the Antigone it recurs only in the compound word ‘ōmēstōn’ (697), applied to the dogs which might eat the dead Polyneices’ flesh. ‘Why this appalling cross-reference?’, asks George Steiner, suggesting that Antigone’s ‘obsession with Polyneices’ corpse’ is ‘not wholly innocent of a primal, nocturnal instinct distantly analogous to that of the beasts of prey and of carrion’ (1984, 245). Antigone has long wished to die (559–60), and she has now found a cause to die for. That it offers such an opportunity is more important to her than the nature of the cause itself: this is the reason why her self-justifications are so diffuse and muddled (in arguing with Creon she twice shifts her ground: 519, 523), and it is the reason why she rejects Ismene’s advances so roundly, for to share the distinction of martyrdom would be, in her eyes, to diminish its value. There is something very frightening in Antigone’s fanaticism: Steiner likens her to Saint-Just (1984, 81); Žižek calls her ‘proto-­totalitarian’ (2005, 306). Even if these soubriquets exaggerate, it is certainly fair to draw attention to an ironic resemblance between Antigone’s treatment of Ismene and Creon’s of Polyneices.100 Chanter’s suggestion that ‘Antigone proves herself in death to have had the potential to be a more effective political leader than Creon could ever be’ strikes me as rather optimistic:101 Antigone in charge (if we can imagine such a thing in heroic Thebes) would have been just as extreme, just as intransigent, as Creon, if not worse.102 She and he are matched in temperament like Prometheus and Zeus in Prometheus Vinctus.103 As Charles Freeland says, ‘Antigone’s beauty does not leave either the thought or the language of European traditions of humanism intact’ (2013, 153). In that respect the Antigone seems to me to be closer in spirit to the Electra than has sometimes been held.104 The opening lines of the Antigone mark one of the most powerful moments in the Western tragic tradition:105 ὦ κοινὸν αὐτάδελφον Ἰσμήνης κάρα . . . O my sister Ismene, my own dear sister, Do you know what evil, of all those that come from Oedipus, Zeus will not accomplish for us two while we live?

98  Tragedy and Moral Redress The first line, translated literally, means something like ‘O my own common sisterly head of Ismene’. Steiner describes the designation of Ismene by means of her head as a ‘vehement anomaly’ (1984, 209), but in fact it is a standard synecdoche, common in classical and classicizing poetry (which was why Housman was able to parody it in his fragment of a Greek tragedy), found even as recently as Goethe’s Iphigenie.106 Steiner is on firmer ground when he comments that the line’s beginning points to Antigone’s and Ismene’s monstrous origins as both daughters and sisters of Oedipus (1984, 208–9): as Benardete remarks, ‘In κοινόν [koinon, ‘common’] lurks the incest of Oedipus’.107 From the off, Antigone’s appeal to kinship is undermined by the perverted nature of the family on whose behalf she purportedly acts.108 Jebb has a perceptive note here: ‘The pathetic emphasis of this first line gives the key-note of the drama. The origin which connects the sisters also isolates them. If Ismene is not with her, Antigone stands alone’. Perhaps this is what Steiner means when he writes: It is a crucial paradox or duality of the human condition that kinship is, in one respect, the most universal, ordinary of biological–social facts, yet in another the most irreducibly singular and individually specific. In the mouth of Antigone, as Kierkegaard sensed, κοινόν is fatally charged.109 Fatally charged with what? With, I suggest, negativity (‘She rides into the play on a torrent of negatives’),110 with refusal, with rejection, with bitter, twisted, sardonic zealotry. ‘Hers is a nihilistic courage that serves the dead and her own death’.111 The irony is that, while Antigone indeed addresses Ismene as though she (Antigone) were, as Goethe calls her, ‘the most sisterly of souls’ (‘die schwesterlichste der Seelen’),112 she is in fact going to treat Ismene in a way that is not at all sisterly. This irony is present in ‘The evils proper to our enemies, directed against friends’ (10), of which she informs her sister: that is exactly how Antigone is about to behave towards Ismene.113 Steiner endorses Goethe’s characterization of Antigone (1984, 12, 209), as does Kirkwood, though the latter concedes that the description ‘is right, provided we do not equate sisterliness with gentleness’ (1994, 119 n. 19), a qualification which points to the truth, namely that it is Ismene who is sisterly in Sophocles’ play, not Antigone. For Jebb’s ‘If Ismene is not with her, Antigone stands alone’ is not quite right; it does not capture the point that Antigone wants to stand alone. It is not, as Steiner suggests, that Antigone ‘is . . . made the most solitary, individual, anarchically egotistical of agents’ (1984, 213; my emphasis); rather, that is what she makes herself. That is the significance of the chorus’s description of her as ‘autonomos’ (821), and of her disposition as ‘autognōtos’ (both words in effect meaning ‘self-willed’), adjectives which are, in their application to Antigone and their suggestion

Antigone’s Holy Crime 99 of solipsism, negatively charged.114 Her use of the word ‘koinon’ was never seriously intended, something that perhaps comes across from the ponderous literalness of Hölderlin’s translation of the play’s first line in his Antigonae: ‘Gemeinsamschwesterliches! o Ismenes Haupt!’ (1992–4, vol. 2, 861). If Antigone’s rejection of Creon’s edict is, as Jean Bollack observes, ‘antérieur même à la trame de la tragédie qu’il lance’ (1999, 58; cf. 86), her rejection of Ismene is equally so.

13 Antigone’s Last Speech The reason, we are led to believe, why Antigone starts out with the disturbed psyche that we have analysed lies, as she herself half-recognizes, in her unnatural background as both daughter and sister of Oedipus (857– 66): as Griffith puts it, ‘this is a daughter who belongs in all respects to a uniquely inward-turned, self-fixated family, a family characterized in this play as hopelessly involved in atē (“ruin, doom, madness, fixation”)’ (2010, 113–14).115 In mentioning ‘my ill-starred mother’s incestuous sleeping with my father’ (865), Antigone is put in mind both of Polyneices’ equally ill-fated marriage (870) and of her own unwedded status (869). Jocasta’s union with Oedipus was unnatural, and Antigone comes to think that her own unmarried position is equally so. Indeed the unnaturalness of her chosen course of action is brought out in the many alphaprivative epithets that are attached to Antigone and her actions during the course of the play:116 her fate is unwept (adakruton: 881), she dies friendless (aphilos: 876), unwed (agamos, anumenaios, alektros: 867, 876, 917), and unmourned (aklautos: 847, 876); even the ‘unwritten and unfailing statutes of heaven’ (ἄγραπτα κἀσφαλῆ θεῶν νόμιμα: 454–5) to which she appeals in her famous speech of 450–70 contribute indirectly to the feeling of misguided negativity that surrounds Antigone’s nisus to suicide.117 Griffith notes that the speech consists largely of negative statements (1999, 32 n. 98, 200), but I think that he fails to appreciate the significance of this fact. So far from representing Sophocles’ own views, as Griffith speculates (and as many have supposed), the speech points, rather, to the profound unnaturalness of Antigone’s decision. It is borne in on Antigone herself as death approaches that she is making a terrible error. That is why she thinks the chorus is mocking her when they stress the great honour of her sacrifice and embellish her comparison of herself to Niobe (834–40). The chorus quite understandably thinks: ‘An honourable death was what she wanted, was it not? Well, now she has it; let us praise her success’. But no, it turns out that it was not what she wanted. Jebb accounts for Antigone’s hostile reaction to the chorus with the simple observation (on 834–8) that ‘She had looked for present pity. They had comforted her with the hope of posthumous fame’; but he does not notice the extraordinary volte-face that this new attitude represents: for sympathy had, hitherto, been the very last thing that

100  Tragedy and Moral Redress Antigone sought. Yet now she is mourning herself as she never mourned Polyneices. What Antigone does not regret, however, as in consistency she should, is the fact that the death that Creon has determined for her— living entombment—will involve a transgression of divine law. It is not so much the manner of her punishment—an anomalous in-between state, like Polyneices’ (he is dead but not buried; initially she will be buried but not dead)118—as the sheer fact of her death that she mourns.119 She does not now view her death purely objectively, as an act of sacrilege on Creon’s part, but as her death. The thought ‘this is real, and is happening to me now’ has at last dawned.120 The reservations about dying that are so memorably expressed by Shakespeare’s Claudio and Shelley’s Beatrice are now coming home to Antigone.121 The inchoate realization that she has made a dreadful mistake also accounts for the remarkable retrospective rationalizations that Antigone offers in her long speech at 891–928, which have offended so many of her admirers. Audiences have been right to find something distinctly odd about these lines.122 Antigone tells us that she would not have acted as she has done on behalf of a husband or child, for the latter are replaceable, but a brother, once the parents are dead, is not. But, in the first place, as A. M. Dale remarked, ‘the only point of preferring a brother on the strength of his irreplaceability would be if it were a question of keeping him alive’ (1969, 154). Secondly, even if the new nomos can claim the authority of tradition,123 Antigone’s embracing of it now is flatly inconsistent with her avowed motives for action up until this point. In particular her assertion that she would not have acted as she has done on behalf of a dead husband or child clashes with what we had thought she was championing—the exceptionless right of the dead to find burial at the hands of their relatives.124 There is also a tension between the sentiments expressed in these lines and Antigone’s unconscionable treatment of Ismene. As Michelle Gellrich remarks (1988, 52): ‘with both parents gone her sister is irreplaceable, and so she might very well choose to save and protect a live sibling rather than bury a dead one’. And she adds that the denigration of kin, which many attribute to the civic-minded Creon, is already played out in Antigone’s initial confrontation with her sister. A Hegelian conflict between Antigone and Creon is decentered, since the opposition held to be between the antagonists is contained within one side of the binary structure. In fact, ‘her most deliberately reasoned explanation for her act nullifies all three groundings in principle earlier adduced: divine law (450ff.), familial loyalty (45f.) and love (523)’.125 In a famous passage, Hegel offered his own version of Antigone’s rationalization of her action at 891–928, arriving, like her, at the conclusion that ‘the loss of the brother accordingly cannot be made good to the sister, and her obligation to him is the

Antigone’s Holy Crime 101 highest’ (1986, vol. 3, 338), and involving, along the way, the curious assertion that ‘the female, as sister, has the highest intuition of ethical being’ (ibid., 336). As Dennis Schmidt notes, Hegel, like Antigone herself, in privileging the sibling relation considers only a sister’s obligations to her brother, not to her sister.126 Matt Neuburg asks whether we are entitled to expect psychological consistency in a work of ancient literature (1990, 64). I think the answer is that we do have that entitlement: we can be sure of this because ancient literature nearly always does portray agents as psychologically consistent, so that departures from this norm are there to be noticed (at least for bona fide agents: the dictum does not apply so straightforwardly to the chorus).127 Not everyone agrees that Antigone is portrayed by Sophocles as inconsistent. Foley, for example, repudiates the claim of inconsistency, arguing that Antigone does not deny that a husband’s or child’s unburied status would equally be an affront to their honour and to the prerogatives of the nether gods: ‘She simply wants to make the chorus understand that, as the last survivor of the doomed Labdacids, she would only have acted to bury her brother’ (2001, 177). But Antigone twice appeals to a nomos in justifying her new position (908, 914), which indicates that she is stating a general principle. And the new general principle plainly is in conflict with the old. Foley ends her argument by remarking that Antigone ‘would not have challenged the state for a set of relations that are hypothetical to a virgin’ (ibid., 178; cf. 1995, 138). But this involves a confusion: in actuality she is a virgin, but in the counterfactual scenario in which we imagine that she is not a virgin and does have a husband or child—and it is that counterfactual scenario that she considers in her speech—consistency with her former principles would demand that, hypothetically, she act on their behalf just as she has in actuality acted on Polyneices’ behalf. And not only consistency with her former principles, but also consistency with fifth-century BCE domestic theory: as Christiane Sourvinou-Inwood remarks, ‘The fact that in Athenian eyes Antigone’s was a perverted choice is made explicit and intensified at 904–12, where she states that she would not have acted in this way for a husband or a son, while in the ideology of the time her familial duty was very much more compelling in those cases’.128 Honig criticizes this remark as ‘moralistic’ (2013, 128): the remark is certainly moralizing (recall the distinction between these two concepts), but the point is precisely that fifthcentury audiences would have reacted to Antigone’s character in that moralizing way. In William Gager’s 1582 neo-Latin tragedy Meleager, Althaea’s urge to murder her son Meleager in revenge for his slaughter of her brothers is countered by the Nurse’s reminder that a mother’s obligation outweighs a sister’s (‘at minor matre est soror’).129 Goethe famously hoped that a philologist would prove the contested lines to have been interpolated,130 and many others have concurred.131 Even those who accept the lines as genuine do so, it has been said, with

102  Tragedy and Moral Redress misgivings.132 Speaking for myself, I have always felt the opposite: from my first encounter with the Antigone I have been struck by the essential aptness of this passage, and I say of it what I would also say of the Helen episode in the second book of the Aeneid, a passage that in some respects poses comparable problems, namely that if the lines are interpolated they are an inspired interpolation, made by someone who understood his author.133 Goethe’s reason for harbouring a hope that the lines are spurious, and the reason that moves those who share this hope, is that he and they are dazzled by the myth of the divine Antigone: he and they have taken Antigone’s earlier idealistic professions at face value, and so are understandably disappointed when those professions crumble under the onset of reality. But the truth is that the idealism (her ostensible motivation) and the death-wish (her real motivation) were just so much bravado, no more than adolescent fantasies that are swept away by the reality of approaching death. As Lardinois remarks (2012, 63): When she is led to her tomb and given what she asked for, she laments that she has to die ‘before the term of my life is spent’ (896). She has never cared much for the city, denying it the right over the corpse of her brother (45–6), but in her last scene she calls on the chorus as citizens of Thebes (806; cf. 843, 937, 940). Now she is willing to admit that she acted ‘in defiance of the citizens’ (βίαι πολιτῶν, 907; cf. Ismene’s words at 79), and for the first time she refers to Creon’s edict as a law (847). Whereas she had formerly gloried in her act of self-sacrifice on behalf of her brother, Antigone now regrets that he has, in effect, killed her: ‘in dying you have slain me while I was still alive’, she remarks (871). The pleonasm throws a strong emphasis on the phrase ‘while I was still alive’. Previously she had wanted to be dead. Now the fact that her brother is robbing her of life and youth prompts despair. Now she regrets that she must die before her term of life is spent. Now the unnaturalness of her absoluteness for death comes home to her. Critics have been made uncomfortable by Antigone’s last speech because in effect they were expecting either no reasoning at all at this juncture or, if reasoning, then a coherent piece of moral philosophy— the presentation of a precise and cogent thesis which could be weighed in a theoretical balance with its opposing thesis, and preferably found superior. Instead we are given a desperate and ill-conceived piece of rationalization—‘unpersuasive’ (apiston), as Aristotle says (Rhetoric 1417a28)—which falls far short of presenting a compelling case. The crystalline purity of the eternal laws of the gods has degenerated to a set of arbitrary, homemade, factitious pretexts.134 Her ‘love’ for her brother gives way to a purely structural argument turning on a brother’s irreplaceability if the parents are dead, and that argument is irrelevant to

Antigone’s Holy Crime 103 Antigone’s situation, because her brothers are both dead, so that even if she had an overriding obligation to them while they were still alive— something that even fifth-century, let alone heroic, morality would by no means have accepted—her obligations now to marriage, domesticity, and procreation are clear.135 But, though Antigone’s sophistical speech does not pass muster when viewed abstractly as a piece of argumentation, it does do what we expect of dramatic discourse, namely disclose character: it reveals someone who is inchoately discovering that the ideology she has hitherto served does not make sense, and is not, in the event, right for her. She scurries to shore it up, and the hopelessness of her reasoning is indicative of its having lost any genuine attraction for her. After all, it is unnatural to follow a cause so far that one ends up negating one’s own yearnings and aspirations, which turn out, in Antigone’s case, to be utterly normal ones.136 So I think A. J. Waldock is wrong to say that ‘The lines do not fit Antigone’ (1951, 142). They do not fit the polemical case that she earlier presented, but for that very reason they do fit Antigone: they fit the character that the dramatist has given her. The earlier polemics were out of step with her underlying nature. As the end approaches, desperation forces her genuine self into the light of day. Not altogether dissimilarly, the sweet, innocent Ophelia, whose nature we thought so alien from ‘country matters’, ‘reveals in her breakdown just how naturally familiar she is with the world of sexual innuendo and speculation’.137 J. M. Bernstein remarks that if the obvious interpretation of the vexed passage in the Antigone held, ‘grand tragedy would devolve into personal pathology’ (2010, 115). Well, what do we expect of ‘grand tragedy’? Is personal pathology not good enough for it? That would rule out quite a bit of what we think of as great writing in the European tragic tradition, from numerous versions of the Medea and Hippolytus myths to Arthur Schnitzler’s Fräulein Else. On my interpretation of it, the Antigone is not a Hegelian tract about high politics or higher metaphysics; it is an altogether more elemental and more workaday thing, about the wrong choices of a precipitate, nervous, arrogant man and a fanatical, messedup, adolescent girl. Their choices conspire to produce disaster, but it did not have to be that way: ‘There is a sense that none of this had to happen, that both sides backed themselves into a corner, and that the idea that this is somehow part of the divine order can seem strained’.138 If this reading disqualifies the play from earning the label ‘grand tragedy’, so be it. Cairns writes of Hegel’s interpretation of the Antigone that it ‘operates at the level of principle, not of the conduct of the two characters as individuals. It is no argument against it to say that it glosses over the motivations of Creon and Antigone as they are actually represented in the play’ (2016, 125). But quite evidently this is an argument against Hegel’s reading: the play is not about principles; it is about ‘the conduct of the two characters as individuals’. And it is no argument against that point

104  Tragedy and Moral Redress to protest that those two individuals talk about principles, that indeed for much of the play they talk about little else. Just so, but it would be a fallacy to suppose that a play must itself be about what its characters talk about. Shakespeare’s Brutus is always prating on about principles; but Julius Caesar is interested in him, not them. The play finesses the principles to focus on the man—a man who is very good at talking about principles, but rather less good at converting them. Similarly, the Antigone is not ultimately about the unwritten laws of the gods, or the burial of the dead: it is about two lonely people.139 Hegel is of course right that we sympathize deeply with both. As with the Antigone, so with the Oedipus Rex: Oedipus is not ‘the first philosopher’, who ‘differentiates himself from the world and orders it accordingly, internalizing the subject/object separation’.140 The play is a much more ordinary, more particularist, less metaphysical, less pretentious tragedy about one man’s cognitive failure and its catastrophic consequences. (‘Metaphysical’ is a stock epithet used—wrongly—to distinguish tragedy from comedy.)141 This one man is, as Dodds says (1973, 77), in a sense Everyman: but that does not make him a philosopher. Not but what Oedipus does differentiate himself from the world and internalize the subject–object distinction: however these are not peculiarly philosophical attainments; they are things we all do automatically, usually by the age of eighteen months (at the latest); they have nothing especially to do with Oedipus Rex or with any other tragedy.

14 Anouilh’s Antigone George Eliot rightly discerned sensuality in Antigone (1965, 221). For like Ophelia, and like Cordelia and Desdemona (see §17 below), Antigone is not the ethereal creature of many spectators’ and readers’ imaginings. Despite the coyness of some commentators on the point,142 a sexual interest in her dead brother is implied: for while she expresses no affection for Haemon,143 she yearns, as we have seen (§12), to lie with Polyneices (73), a point picked up and exploited later by Rotrou and Gide.144 As Rutherford notes, ‘lie with’, though ‘ostensibly referring to the shared grave she intends to provide, cannot wholly be purged of the suggestion of a different kind of intimacy’ (2012, 75).145 Benardete adds that ‘it cannot be accidental that in her case the language of incest should coincide with the language of the grave’ (1999, 13). There may also be a hint of the contemporary medical view that prolonged sexual inactivity in young women causes mental instability.146 Furthermore, in Antigone’s case affection towards a brother, whether sexual or not, is inevitably tainted by the fact that her father Oedipus is also her (half-)brother.147 Like Hippolytus and Hamlet, Antigone is sexually confused. As Martha Nussbaum observes, she loves no living being ‘with the sort of love that Haemon feels and Ismene praises’.148 The ideals represented by Hamlet

Antigone’s Holy Crime 105 and Claudius are—so G. Wilson Knight famously argued against what had been the accepted nineteenth-century reading of Hamlet—unhealthy and healthy respectively;149 one might contend in similar Nietzschean terms that Antigone’s reaction to Creon’s edict is unhealthy, whereas Ismene’s (initial) reaction is healthy. Antigone’s desire for death is a perversion of the natural order,150 just as her preference for a dead brother over a living sister and a living fiancé is pathological: her death-drive is, I have suggested, her real motivation, but it is abetted by a perverted sexuality. Ismene tells Antigone (88): ‘Thou hast a hot heart for chilling deeds (psuchroisi)’. So Jebb translates, taking ‘psuchroisi’ as an abstract neuter plural (likewise Lloyd-Jones: ‘Your heart is fiery in a matter that is chilling’). But it could be a generalizing masculine plural, and then the sense would be ‘you have a hot heart for those who are cold and dead’. Despite her level-headed assessment of Antigone’s stance in their opening exchange, Ismene is subsequently swept away by its grandeur and charisma, so that she purposes to die with her sister. But it is significant that her reason for this new urge is her unwillingness to live without Antigone (548, 554, 566); her motivation has nothing to do with Polyneices’ burial. In that respect she is curiously like her sister, whose real and ostensible motivations pass one another by like ships in the night. It follows from my argument so far in this chapter that the many adaptations of the Antigone that treat the heroine as a political dissident, as a freedom fighter, as an anti-authoritarian or even anti-capitalist icon, are wide of the mark as readings of Sophocles.151 That is not just so far a criticism: adaptations have always engaged in creative misconstrual. But there is one adaptation that seems to me—and against several commentators—not to misread Sophocles, but on the contrary to capture the essence of his play. For the meaning of Sophocles’ Antigone is, I think, well brought out in Jean Anouilh’s dramatization of the story. Against Luce Irigaray, who holds that Anouilh ‘does not understand a great deal about Antigone’ (2013, 124–5),152 I suggest that he renders explicit what is implicit in Sophocles. As his Créon says: ‘Antigone était faite pour être morte. Elle-même ne le savait peut-être pas, mais Polynice n’était qu’un prétexte . . . Ce qui importait pour elle, c’était de refuser et de mourir’ (1954, 86). That is exactly right not only of Anouilh’s Antigone, but also of Sophocles’.153 This fact has escaped some critics of Anouilh’s play, who have been misled by mere surface differences between the two versions of the myth, in particular by the fact that, whereas Sophocles’ Antigone appeals to divine law and the nether gods, Anouilh’s Antigone does not. Thus W. M. Landers comments that ‘[Anouilh’s] Antigone has no thought of changing the world [sc. unlike Sophocles’ Antigone]; she merely contracts out. The universal values, the principles of piety and justice on which the Greek play is based, are lost beyond recall in Anouilh’s version, and we are left with a conflict of a very different character’.154 And Jedd Deppman argues that Sophocles’ Antigone ‘is doctrinal in her

106  Tragedy and Moral Redress duty, relying on articulable principles in order to reject the emotional appeals emanating from Ismene, Haemon, and Creon. Anouilh’s is a more emotional, histrionic, and divided being’ (2012, 527). But what both critics fail to see is that these apparent differences have no depth: the appeals to piety and justice that are absent from the French play are only superficially present in the Greek; in fact what is going on is that Anouilh is managing to strip away inessentials and penetrate to the core of Sophocles’ drama. Anouilh makes his Antigone very explicitly an adolescent, but, as Lacan pointed out, and as we have already noted, Sophocles’ Antigone is also an adolescent, though this fact has been widely missed.155 There is again an interesting connection with Hamlet, who, like Antigone, thinks in absolutist and polarizing terms; and ‘Hamlet’s thinking’, as Marilyn French observes, ‘is very young thinking’ (1992, 99). Like Lessing’s Philotas and Goethe’s Torquato Tasso, and to some extent also like Chaucer’s Troilus and Criseyde, Sophocles’ Antigone is about adolescence, and the difficulty of integrating it into an adult world:156 in Goethe tragedy is averted; in Sophocles and Chaucer and Lessing it is not. (I shall return to Philotas in §25.) Anouilh’s dramatization brings out the fact that, in Sophocles too, the adult–adolescent axis is much more important than the male–female. So we do not have ‘a conflict of a very different character’ in Anouilh, but rather the quintessence of the Sophoclean conflict. When Anouilh’s Antigone is asked by Créon why she has done what she has done, she replies ‘Pour personne. Pour moi’. And she agrees with him that her action is absurd (1954, 72). With a bit of insight into her own motives, Sophocles’ Antigone might have said the same. Rowan Williams remarks that, ‘in the end, Anouilh’s Antigone [finds] the prospect of death terrible and no longer knows what she is dying for’ (2016, 68); he omits to mention that exactly the same holds of Sophocles’ Antigone. Anouilh well brings out the Sophoclean theme that such a solipsistic death-wish in an adolescent girl (or boy: compare again Lessing’s Philotas) is deeply unnatural, and even has something fascistic about it, as Lacan remarked.157 When Créon tries to persuade her to recant by alluding to the life force in simple, natural things, Landers comments that his appeal ‘begs the question entirely, since neither animals nor trees are faced with the necessity of making moral judgments’.158 But this misses the point: Créon’s argument is that the impulse to live must come first in any animate being. We too are animals, and our first business is to survive, at (if necessary) any cost. There is, ultimately, no deeper imperative. Anouilh’s Créon does not beg any questions; there is no dramatic petitio principii. Similar remarks apply to Sartre’s depiction of Cassandra in his Les Troyennes. Nicole Loraux suggests (2002, 5) that Sartre psychologizes and secularizes Euripides’ Cassandra. But, given the operatur principle (§8), we can see that Sartre is simply bringing out what is implicit in Euripides. So the ‘translation of inspired bacchism into a clinical insanity

Antigone’s Holy Crime 107 worthy of nineteenth-century psychiatry’ (ibid.) is less of a distortion, if indeed it is one at all (see further Part II below), than Loraux’s phraseology implies. Again, Peter Burian thinks that Eugene O’Neill’s approach to the myth of the house of Atreus in Mourning Becomes Electra ‘is inevitably reductive and restrictive’, because ‘the passions of the characters are reduced to an endlessly repeated, implausibly symmetrical set of attractions and repulsions, self-consciously and relentlessly enacted. And the world of the play is very largely restricted to this psychopathology’ (1997b, 255). To this the obvious response is that the Aeschylean trilogy is already an exploration of psychopathology. And, as in connection with Bernstein’s comment above (§13), we should resist the implied dichotomy between grand tragedy and mere psychopathology: the latter is very often the shape that the former takes. In general the operatur principle implies that modern adaptations of classical drama which work by secularizing their originals are less liable to distort those originals than is often supposed. The same point applies to Sartre’s Les Mouches, faulted by Burian on the basis that ‘its design denies reality to the order beyond human control and understanding’ (1997b, 261). In Les Mouches Jupiter comes on stage as an ordinary agent, just as he does in Plautus’ Amphitruo: there is no sense of divine mystery. But the operatur principle invites us to exercise caution before we throw around such inflated phrases as ‘the order beyond human control and understanding’, which sound impressive but are in reality unsupported. As I have been urging so far in this part, in its deep structure classical tragedy presents us with no such order.

15 Does Antigone Have a ‘Gethsemane Moment’? Antigone’s folly emerges in a systematic and comprehensible way from her character, so that, as in the cases of Sophocles’ Oedipus, Ajax, Deianeira, Heracles, and other major tragic protagonists, there is no question of punishment for a casual error that bears no significant relation to character;159 on the contrary, it is correct in all these cases to speak of a tragic flaw issuing in misguided action. So far, I have analysed Antigone’s flaw in terms of a perverted death-wish; but there is a further twist that needs to be added to the story of her error. As commentators have noted, Antigone resembles Ajax (as well as the pseudo-Aeschylean Prometheus) in her recklessness and her contempt for enemies, as well as in her selfcentredness: as we have seen, she complains, in going to her death, that she has no friends or mourners, but she spares no thought for Ismene or Haemon, just as Ajax abandons Tecmessa and Eurysaces.160 She speaks of herself as the last of her line (895, 941): Ismene does not count for her.161 She falls because of her ‘self-willed temper’ (autognōtos orga). There is something peculiarly horrifying, on the lips of a young girl, in her words to Creon at 460–6 (cf. 559–60), quoted above (§12). She has not—at

108  Tragedy and Moral Redress this stage of the drama—thought through the real implications for herself of her impulse towards death, and she will not start to do so until it is too late. She does not realize until it is too late that the things that anyone—in particular, the things that Sophocles’ Athenian audience— would expect to matter to a girl of her age and status—namely, as we have said (§12), finding an aristocratic husband, in due course having children, organizing a large domestic outfit: in general, participating appropriately in the life of the oikos and polis—do matter to her, contrary to her own self-assessment. They are what she deeply wants and needs. (It has been suggested that the cry Antigone emits when she sees Polyneices’ uncovered body, and which is compared to that of a mother bird espying her nest robbed of its nestlings, anticipates Antigone’s subsequent regret at her own childlessness.)162 In the headiness of her assertion that she wishes to die by burying her brother (72), she does not realize until it is too late that she does not wish to die,163 that she does not wish to ‘marry Acheron’ (815–16; cf. 804–5, 891), any more than Polyxena wished to ‘marry’ the dead Achilles.164 In a famous passage Lucretius lamented the sacrifice of Iphigeneia as a ghastly inversion of the wedding that had been her due.165 ‘Both in funeral epitaphs and in tragedy, the death of a young woman before marriage is commonly likened to a union with Hades’, notes Justina Gregory in her commentary on Euripides’ Hecuba (on line 612), where the idea recurs in connection with Polyxena.166 A death that is a wedding is no more natural than a wedding that turns into a death, such as Evadne’s in Euripides’ Suppliants,167 or than—what Tacitus tells us was the youthful Octavia’s fate in her union with Nero,168 and what the ghost of Agrippina threatens for Poppaea in the Octavia169—a wedding that resembles a funeral. (Antigone’s maidenhood is also an important and recurrent motif in Euripides’ Phoenician Women.)170 Antigone is naturally unnatural: not only are her genuine, underlying desires entirely normal, but her fate is also a common one, and the sorrow of her premature death does not differ in its essentials from that of any real or literary girl who dies before marriage.171 ‘Antigone, being led to her death, naturally thinks of the very different procession in which she, a young girl, would have expected to be led’.172 Instead of (ostensibly) worrying about the nomima in respect of the unburied, she should have been thinking about the nomima of her marriage. ‘It is a lamentable thing that young girls should be plucked unripe and should travel to the end a hateful journey from their homes before the time has come for the customary rites of marriage’, sings the chorus of Aeschylus’ Seven Against Thebes (333–5, tr. Hutchinson); they are thinking about exile and captivity. And they add (335–6) that this is a fate worse than death. But death is the journey that Antigone has chosen, against nature. Like Creon, as we have seen, Antigone does not u ­ nderstand—or begin to understand—her own nature until it is too late. That, I suggest, is her ultimate cognitive failure, the real nature of her error.

Antigone’s Holy Crime 109 But does Antigone achieve self-understanding? Is she, like Creon (1270–2) and so many other tragic figures, a late-learner?173 Does she achieve that Euripidean arti manthanō moment? Does she think ‘now— now when it is too late—I understand,’174 words that ‘might be made the burden of many tragedies, ancient and modern—a truly tragic anagnōrisis [recognition] in a sense not included in Aristotle’s definitions’, as Dale remarked?175 Does she even attain to the inchoate recognition of the ‘où suis-je?’ of Racine’s Hermione and Athalie?176 As far as understanding her own nature is concerned, it seems clear that in the moving kommos beginning at line 806 Antigone comes to see the unnaturalness of her sacrifice, and that it is just this realization that prompts the rationalizations of her ensuing speech, which have caused audiences such heartache. (Compare the unnaturalness of the Danaids’ resistance to marriage as that emerges in Aeschylus’ Suppliants.)177 What about the issue of Polyneices’ burial? Here I have argued that, although Antigone turns out to have been right that Polyneices’ body ought to be buried, she was wrong to undertake it herself, in defiance of Creon. How does she regard this transgression? One view is that she acknowledges her own ill counsel from the start, and she does indeed appear to do so quite early on in the drama (95–6); but is this a real concession on her part,178 or does she only mean that this is the opinion that Ismene and others will form of her conduct (and indeed Ismene immediately agrees: 99; cf. 67–8)?179 This latter interpretation may be right, though if it is we are not precluded from detecting a moment of dramatic irony, after Sophocles’ familiar manner, in which his heroine stumbles unwittingly (or half-knowingly) on the truth. Similar remarks apply to Antigone’s description of herself as committing a holy crime, quoted above (§12). (So too in Corneille’s Polyeucte Pauline tells us that she is ‘saintement rebelle’—holily rebellious—to the laws of her birth: 1739.) And Antigone ends her controversial speech with what looks like a tentative admission that she has been making some kind of mistake in defying Creon (925–6): ‘Well, if this is approved among the gods, I should forgive them for what I have suffered, since I have done wrong’ (tr. LloydJones). We have seen (§§11, 13) that, as death nears, Antigone shows a new awareness of the citizens of Thebes (806, 842–3, 937)—citizens in whose despite she has acted (907), and whom Polyneices had sought to slaughter and enslave (201–2), and in so doing she echoes earlier sentiments of her sister (59–60, 79). But it would be going too far to say, with Nussbaum, that ‘Antigone comes to see that the service of the dead requires the city, that her own religious aims cannot be fulfilled without civic institutions’ (1986, 66). This, to be sure, is something that the spectator and the reader will, ideally, come to see—for it is indeed a fundamental part of the work’s cognitive content180—but Antigone does not get there. When she invokes the gods of her native land (839– 40, 938), we remember that it was the shrines of these very gods that

110  Tragedy and Moral Redress Polyneices had threatened with destruction (199–201); but that thought is not entertained by Antigone.181 Just as Antigone appears briefly to recognize her own folly early on in the drama, so near the end of her life she seems momentarily to contemplate the possibility that she has been guilty of an impiety. Immediately before the words just quoted in the last paragraph, Antigone accuses the gods of betraying her (921–4): ‘What manner of divine justice have I transgressed? Why must I still look to the gods, unhappy one? Whom can I call on to protect me? For by acting piously I have earned for myself impiety (τὴν δυσσέβειαν ἐκτησάμην)’.182 In their notes on this last phrase, Jebb and Griffith interpret it as denoting that Antigone has earned the reputation of impiety, rather than the thing itself. But Sophocles is going further than that: Antigone has lost confidence in her ideals and, as the end approaches, is raising the possibility that she might have committed an impiety, that Creon might be right after all. For one thing, the gods have not saved her. This thought must also strike the spectator: Antigone goes to her death; Creon survives. He suffers, but he learns from his mistakes, and stays alive. Was Creon, in a deeper sense, right all along? Antigone, one might say, comes within an ace of anagnōrisis (in Dale’s, not Aristotle’s, sense)—that is, of recognizing her real nature, and her infidelity towards that nature—but she does not quite achieve understanding, and goes to her death in a state of emotional and intellectual turmoil, as she reverts in her last words to the claim that she has been pious after all (943). She no longer believes that what she is doing is right,183 but nor has she grasped why she has lost that belief. Steiner speaks of a ‘last flinching before a willed, accepted self-sacrifice’, and he compares Christ’s ‘Gethsemane moment’: ‘Without this flinching, there would not be the self-knowledge . . . which gives to self-sacrifice its lucidity and meaning’ (1984, 279–80). But it is the mental turmoil that is to the fore in Antigone’s final moments rather than any such lucidity or meaning. Her sense that the gods have betrayed her seems so consequential that it is easy to take in the words without pausing for thought, but the question that one should be asking at this point is: why does Antigone feel that, in allowing her to go to her death, the gods have betrayed her? After all, piety is not guaranteed to keep you alive, nor even, in a dangerous world, likely to do so, as she formerly well recognized. She knew that in burying her brother she would incur death, and she went into that prospect with her eyes open. That was the deal, so to speak, the terms of which she formerly welcomed. ‘Antigone had chosen death (555), counted herself already a dead soul (559f.), told Ismene that death was kalon [beautiful] (72) and Creon that it was kerdos [a gain] (462, 464)’.184 In seeming now to disavow those terms, and in mourning her own death as the passing of a young, unmarried girl (813, 869, 876, 917–18), Antigone begins to shift in her understanding from the unnaturalness of her former, defiant position to a more natural posture, one that involves the emotional

Antigone’s Holy Crime 111 rejection of her impending death.185 But there is no Gethsemane calmness and clarity of purpose. There is no resigned communion with the gods, no hint of a si possibile est, transeat a me calix iste: verumtamen non sicut ego volo, sed sicut vos. Eduard Fraenkel notes of Aeschylus’ Cassandra that ‘like Antigone she might say ἡ δ᾽ ἐμὴ ψυχὴ πάλαι τέθνηκεν’ [my soul has long been dead],186 and the comparison well brings out the unnaturalness of Antigone’s stance. Earlier I compared Antigone to Ajax, but there is an important disanalogy between them: for Ajax goes resolutely to his death, whereas Antigone breaks down and nearly flunks the moment of supreme trial.187 The difference arises not, as D. W. Lucas suggests in connection with Aristotle, Poetics 1454a23–4 (quoted above, §12), because he is a man and she is a woman—‘Antigone is brave, but it is appropriate that she should not face death with the same unbending resolution as Ajax’ (1972, 158)—but because she changes her mind about what she really wants, whereas Ajax, having achieved clarity about his position, and seeing that it implies his own death, embraces the consequence unswervingly. Ajax descries what would be required of him were he to adapt to modern mores, but he does not treat this possibility seriously as an option for him.188 Antigone, by contrast, is starting, albeit uncertainly, to realize towards the end of her life that marriage and children and normality were, after all, sensible options for her. She worshipped the wrong images: she chose to be a nun, but she could have been a mother. Bernard Williams writes that ‘Antigone gets what she wants’ (1993, 86). Well, but by the time she gets it, she does not want it any more.189 Inscribed on the propylaion of the temple of Leto at Delphi was the motto that the best thing in life is to get what you want,190 but Antigone is a counterexample: in her we see ‘the tragic irony of answered prayers, as St Teresa has it, over which more tears are shed than unanswered ones’.191 ‘All get what they want; they do not always like it’, says C. S. Lewis’s Aslan (1998b, 179). Lewis would have recalled the words of Chaucer’s Arcite when he is released from prison: We witen nat what thing we preyen heere; We faren as he that dronke is as a mous.192 ‘One enjoys less what one obtains than what one hopes for, and one is happy only before happiness is achieved’, remarks the heroine of La Nouvelle Héloïse (1967, 528; tr. 569). Like, it is implied, Abelard’s Heloise, deprived of her lover by main force, Rousseau’s Julie does not get what she wants, and it is what she wants. Like Chaucer’s Arcite, released from a captivity that gave him the sight but not the possession of Emelye, and like Lewis’s Jadis, regenerate queen of the dead city of Charn who joys in the theft of a life-giving apple, Sophocles’ Antigone gets what she wants, and it is not what she wants. Getting what you want is indeed,

112  Tragedy and Moral Redress as St Teresa tells us in the above quotation, a classic tragic motif. Ibsen’s Hedda Gabler gets what she wants (1950, 304). The Faust of the tradition gets what he wants.193 The Theban women who rend Pentheus limb from limb get what they want and, as Dodds comments (on Bacchae 1161–2), ‘their victory is also their punishment’. Juliet Dusinberre notes that ‘Lydgate, in Middlemarch, destroyed by a wife who adorned and entertained in the best tradition of feminine light relief, might cherish the bitter thought that, like many husbands, he had got what he wanted’ (2003, 114). Steven Pinker describes the insight that ‘you should be careful about what you want because when you get it, you may not enjoy it’ as ‘the major realization of adulthood’ (2011, 500). That reminds us that Antigone is tragically not yet an adult.

16 Antigone and the Gods Rather than saying that Antigone’s tragedy consists in that fact that she ‘has done what she thought right in reliance on the gods, has been deserted by them, and is now leaving a world in whose justice (either on the human or divine level) she no longer believes’,194 it would be more accurate to say that, though this is indeed what Antigone herself in part thinks her tragedy consists in, she also in part perceives the deeper fact that, at the end, she has lost faith in the reasonableness of her martyrdom, and has been betrayed not by the gods, but by her own fanaticism and misjudgement. The gods, in fact, are strikingly irrelevant to the drama as it unfolds: everything is understandable in a purely human terms (‘Alles geht rein menschlich zu’, as Böckh aptly said).195 The gods are not ‘playing’ with men in the Antigone,196 any more than they are in the Oedipus Rex, or in Sophocles’ other tragedies.197 Creon is not ‘the victim of the gods’;198 and the same is true of Antigone. Teiresias mentions the Furies as punishers of Creon but, as Lefèvre observes (2001, 113), if Teiresias had not mentioned them no one would have missed them. ‘The transcendent absolutes to which Antigone appeals in her debate with Creon are, in a radical sense, secular’, says Steiner, adding that ‘Antigone draws about herself an ethical solitude, a lucid dryness which seem to prefigure the stringencies of Kant’ (1984, 271). Creon, like Antigone, has a moment when he accuses the gods of bringing him down (1272–6); but he has told us that he takes full responsibility for his folly (1262–5), and the allusion to divine responsibility, rather than exonerating him, points—in a manner that, as we have seen, recalls Oedipus and many other tragic figures, such as Aeschylus’ Clytemnestra and Eteocles, Euripides’ Helen, Hughes’s Mordred, Goethe’s Götz and Egmont, Hardy’s Napoléon—to Creon’s responsibility for his own fate. Some commentators have suggested that the gods are responsible for the first burial of Polyneices’ body.199 This cannot be right, because gods in ancient works do not themselves perform acts of burial: that requires

Antigone’s Holy Crime 113 human agency.200 Further, Antigone’s response to the uncovering of the corpse when she returns a second time shows that she performed the first burial, as does the fact that she accepts responsibility for both burials when charged with them by the sentries (434–5).201 Marsh McCall suggests that the description of Antigone’s reaction to the uncovering of the corpse is just the sentry’s own interpretation of the event, and that the reason why Antigone does not deny having performed the first burial is not that she did perform it, but rather that ‘She is willing, even eager, to accept responsibility for all that has to do with her brother’s burial’ (1972, 116). But if the sentry’s report were unreliable, we would expect an indication of that fact to be given in the text: there is no such indication, and it is rather the case that the report has the status of a messenger speech, and is to be taken at face value. (Moreover, if Antigone herself thought that the gods were responsible for the first burial, she would not regret their desertion of her, as we have seen that she does.) Another view is that the first ‘burial’ is not, strictly speaking, a burial, but the mere preservation of a corpse.202 Again, this view cannot be right, first because, though not a complete burial, the sprinkling of dust on Polyneices’ corpse is part of a complete burial, and secondly because, somewhat obviously, covering a corpse with dust is not a good way to preserve it. The gods’ treatment of Hector’s body in the Iliad shows that they have more effective methods at their disposal.203 Richard Minadeo asks: if Antigone performed the first burial, why did she slip away? After all, as we know from her initial dialogue with Ismene, she wants the deed to be known (1985, 143 n. 22). Minadeo is right that she wants her deed to be known: indeed it is even the case, as I suggested above (§12), that she wants to be caught, for she seeks martyrdom. So why does she not stay with the body, the first time she visits it, until she is caught? The answer, as J. E. G. Whitehorne sees (1983, 129–31), is that Greek burials were performed in several distinct stages, so that Antigone returns to the body in order to carry out the next stage: she has covered the body with dust in her first visit, a largely symbolic beginning, and now she intends to pour libations and lament. She will continue with the process, if necessary raising a mound (80–1, 91), until she is caught. (I shall return to this issue, in the context of examining the proposal that Ismene performed the first burial, in §18.) Shakespeare’s Lear and Gloucester exemplify the topos we have mentioned, of blaming the gods for what are plainly their own failings, a style of excuse memorably caricatured by Edmund in his speech ‘This is the excellent foppery of the world’;204 and in general, despite the presence of supernatural elements, Shakespeare’s tragedies manifest the same spirit of secularism that Böckh diagnosed in the Antigone.205 In King Lear, especially, the involvement of religion is purely epiphenomenal, as has often been observed:206 Steiner’s assertion (apparently in tension with his remarks on Antigone quoted above), that ‘tragedy is that form of art which requires the intolerable burden of God’s presence’, is mistaken.207 What

114  Tragedy and Moral Redress makes the presence of God, or of gods, in traditional tragedy the reverse of intolerable is that, as we have seen (§8), the divine is not generally represented as independent of human motivation. Alison Shell may be right that Shakespeare’s audiences would have felt that King Lear exemplified the inscrutability of God rather than his non-existence (2010, 186–96, 217), but it does not follow, as she also suggests, that modern nihilistic readings of the play are anachronistic. The play does not need God—or gods: recall that its theology is inconsistent208—to motivate the action; and, in view of the bleakness of the play’s mood, that invites the nihilistic interpretation. This is an application of the traditional point (found in one version of the Ontological Argument) that God, if he exists, exists necessarily; so if he does not have to exist, he does not exist. Actually, to be fair to Steiner, his position on the Antigone is not quite as contradictory as it might seem from the two passages of his that I have cited above in this section. His view is that the divine is present in the Antigone, just not in the scenes involving the protagonists: it is the choral odes which supply the preternatural element, which ‘burst open the secular gates’.209 I would prefer to say, in keeping with my general line on the involvement of supernatural elements in tragedy, that the chorus transposes, stylizes, and celebrates, but does not undercut or compromise, the naturalism of the play’s episodes. The supernatural makes sense, and only makes sense, when treated as subserving a fundamentally naturalistic agenda. Similarly too in Sophocles’ Women of Trachis, when the chorus tells us (860–1) that Aphrodite is responsible for the tragic events of the drama, its words can be accepted as true, but only if interpreted metaphorically, for nothing in the play requires the explicit presence of the goddess as an external autonomous agent; as with Euripides’ Hippolytus, where Aphrodite actually appears on stage, the action of the drama makes excellent sense when explained in purely human terms. And Hyllus’ famous final words, if they are rightly attributed to him, in which he attributes responsibility for everything that has unfolded to Zeus (1278), are to be received in the same spirit.210

17 Cordelia and Desdemona If Sophocles’ Antigone commits a ‘holy crime’, Shakespeare’s Cordelia is a ‘holy fool’ in the tradition of St Paul and Erasmus.211 Like Antigone, Cordelia has long attracted the most ecstatic and extreme accolades. Schopenhauer identified her and Coriolanus as two of the noblest characters in Shakespeare.212 A. C. Bradley’s remark that ‘she is a thing enskyed and sainted’ (1991, 291) will suffice to indicate an orthodoxy which is by no means dead.213 Feminist critics have responded to this sentimentalism by noting that Cordelia is a sterner and more masculine figure than is often assumed.214 But for my purposes the important point is that even Cordelia, for all her spiritual beauty and integrity, can plausibly be held

Antigone’s Holy Crime 115 to conform to what I have suggested is the standard tragic model, according to which heroes and heroines taste the cup of their deservings, and can say, in Hegel’s formulation, ‘because we suffer we recognize that we have fallen short’ (‘weil wir leiden, anerkennen wir, daß wir gefehlt’).215 In the opening scene of King Lear, the refusal of Lear’s youngest daughter to join her elder sisters in entering into the spirit of the pantomime which the old king has so absurdly set up is a kind of cognitive failure. Cordelia finds herself in a formal situation that requires formal behaviour: it is not a time for the ‘here I stand, I can do no other’ gesture. Fathers expect obsequious behaviour from their daughters: that is the relevant cultural norm, which Cordelia outrageously flouts.216 She could and should have imitated her sisters; she should have followed the advice that Volumnia gives to Coriolanus (III, 2). After all, did Lear deserve better?217 Cordelia in effect turns the situation into a contest with her father that he then feels politically obliged to win, in order to avoid public humiliation;218 they are mimicking moves already made by Antigone and Creon. But there was an alternative. Colin McGinn conveys the necessary point when he writes (2006, 128) that Cordelia is not ‘entirely blameless, despite her manifest integrity’: She refuses to accept the superior virtue of the white lie, as if not wanting to taint herself. That is, to be sure, a failing that flows from a virtue, but it is a failing nonetheless. She is more concerned with preserving her integrity than in protecting her father’s feelings. By the play’s end, when she replies ‘No cause’ to her father’s admission that she had cause to be resentful, she is no longer sticking to the literal truth come what may—for she did have cause. Now she understands that kind words can sometimes be better than true words. If Cordelia can relax the truth-telling at this later point, why could she not have done so earlier, before it was too late?219 Peter Holbrook’s view is that, just as Antigone had no choice but to bury Polyneices, so Cordelia has no choice but to speak the truth (2015, 24). That is certainly what both agents think. But they are deluded: they do have a choice. Interestingly, Bradley denies that Cordelia does tell the truth in the opening scene of the play, in response to her father’s plea. He writes (initially in similar terms to the passage from McGinn just quoted) that truth is not the only good in the world, nor is the obligation to tell truth the only obligation. The matter here was to keep it inviolate, but also to preserve a father. And even if truth were the one and only obligation, to tell much less than truth is not to tell it. And Cordelia’s speech not only tells much less than truth about her love, it actually perverts the truth when it implies that to give love to a husband is to take it from a father. There surely never was a more unhappy speech.220

116  Tragedy and Moral Redress Lear says to Cordelia ‘Thy truth then be thy dower’ (I, 1, 109), but unlike Lady Elizabeth Grey, who says as much of herself in the third part of Henry VI (III, 2, 72), Cordelia, according to Bradley, is not being truthful in her responses to her father. It seems to me that this is mistaken. Against what Bradley says in the quoted passage, it is the case that to give love to a husband is to take it from a father. There is a difference between the state of childhood, in which one’s love for one’s parents is (ideally) untrammelled by other emotional commitments, and the state of adulthood, in which (again, ideally) one’s love for one’s parents can no longer be so unstinting because one now has new emotional commitments—to a spouse, to children. That is the natural order of things, and there is nothing in it that anyone needs to apologize for. Parents who expect anything different are simply being unreasonable, and forgetting that they themselves, once, had to cut the moorings to their own sires. So Cordelia is not perverting the truth; she is speaking the plain truth. But this was not the place to say it, and it is part of her tragedy that she did not see that. (Another case where the principal hamartia seems to be tactlessness is afforded by Celia in Aretino’s Orazia: Publio advises her to take her grief elsewhere (III, 41–6), but she refuses, and publicly mourns the death of her fiancé Curazio, so making herself appear to be an enemy of Rome (III, 254–6); the Nurse tells her to go home (III, 271–4), but she ignores this sensible advice. A similar point applies to Camille in Corneille’s Horace, who goes out of her way to offend Horace, so deliberately provoking her own death. Sabine tries to do the same but fails: Horace’s anger is spent.) Interpreting Cordelia’s failure as cognitive, according to what I am suggesting is the standard model, seems to me more plausible than Maureen Quilligan’s idea that Cordelia is punished for incestuous feelings towards her father (2005, 234–5), but Quilligan is right that we should not shrink from applying the sin–punishment schema even to Cordelia: ‘Our twenty-first-century sensibilities do not allow us to blame someone who is clearly a victim, but we differ from the Renaissance in this’ (ibid., 215).221 Desdemona, like Cordelia, has been held to be divine ever since Cassio called her so. Bradley waxes lyrical: she is ‘the “eternal womanly” in its most lovely and adorable form, simple and innocent as a child, ardent with the courage and idealism of a saint, radiant with that heavenly purity of heart which. . .’,222 and so on. He tiptoes in gingerly fashion round the question of her sexual attraction to Othello, preferring to hide under the safety of a ‘nobility of soul meets nobility of soul’ explanation of their congress (1991, 182–3, 190), ignoring the fact that she was ‘half the wooer’ of Othello.223 Again the soubriquet of divinity is a misnomer. Like Cordelia, Desdemona lacks judgement and a sense of what is appropriate to the occasion.224 Like Cordelia, Desdemona tells a lie: she absolves her husband of murder.225 Like Cordelia—and indeed like Ophelia too, who lacks insight into her situation, unwisely allowing

Antigone’s Holy Crime 117 herself to be manipulated by Claudius and Polonius226—Desdemona is punished for her failings.227 As the nineteenth-century critic Gervinus remarked, Desdemona lacks ‘wide circumspection, quick dexterity and mobility of mind, insight and knowledge of human nature’ (1872, vol. 2, 54). She is betrayed by her complete Arglosigkeit (ibid., 55), in which she is marked out from other Shakespearean heroines: her fate could not have befallen Beatrice or Rosalind. And in the course of an excellent discussion Gervinus identifies what might have come to Desdemona’s and Othello’s rescue (1872, vol. 2, 83): If anything might save them both, it could only have been a shrewdness and mental dexterity on Desdemona’s part, an ability to see through his condition and understand how it might, gently but surely, be cured, and how the bewitched Moor, by dint of counteracting, beneficent bewitchments, might be led back to the truth. The cunning of an Isolde, the cleverness of an adviser like Brangäne, combined with Desdemona’s innocence—that was what might have exorcized the evil spirit in Othello. A splendid passage; but how remarkable that Gervinus, in reaching for an example of the sort of feminine intelligence that he has in mind, goes outside the work and thereby overlooks the obvious contrast with Emilia. For one reason why Shakespeare lays such stress on the maidservant’s worldliness is to intimate to us how things might have been otherwise, if Desdemona had only attended Emilia’s school.228 (And yet, ironically, it turns out that Emilia is not worldly enough: she fails to understand Iago, or suspect him of foul play in the handkerchief business.) Gervinus places much emphasis on Desdemona’s disobedience to her father and her disregard of his ill-will (1872, vol. 2, 58–60): he appears to regard her undaughterly behaviour in Venice as the primal sin from which all else follows; like Webster’s Duchess of Malfi and Hebbel’s Agnes Bernauer, Desdemona in effect ignores the advice of the chorus in the Prometheus Vinctus, not to marry outside one’s station.229 In thus censuring Desdemona’s deception of her father, it may be that Gervinus is closer to the mentality of Shakespeare’s contemporary audience than we are today.230 At any rate, he is right that Desdemona’s innocence is a partial innocence, mixed with tragic flaw,231 though I should prefer, even taking into account the traditional hallowing of duties towards one’s parents, to locate that flaw in the cognitive failure that marks her dealings with Othello rather than in the brash precocity that she displays towards her father Brabantio.232 In a certain sense, and despite the manifest differences, one might liken Desdemona to Herman Melville’s Billy Budd.233 Clifford Leech offers Budd as a counterexample to Aristotle’s thesis that the tragic hero should not be wholly good (1969, 38–9). Budd, suggests Leech, is wholly good;

118  Tragedy and Moral Redress and similar remarks have been made about Desdemona. Now certainly Budd is, one might say, wholly innocent: in Captain Vere’s words, he is ‘innocent before God’ (1970, 387). But is he wholly good? Goodness— and this is a central Aristotelian thought—is a cognitively richer state than mere childlike naïveté, lack of guile, and the inability to think ill of anyone. The morally good person does think ill of some people, is generally on the lookout for nastiness, and is realistic about human nature. I think we must dissent from Tzachi Zamir’s judgement that ‘An Edgar cannot be prepared for the appearance of an Edmund, Emilia cannot suspect that Iago is who he is, a Gloucester cannot grasp that Lear’s daughters can manifest utter heartlessness. Shakespeare is associating goodness with a restricted world-view, which renders it vulnerable but also infuses it with a particular dignity’ (2016, 85). Zamir is right about the dignity, but wrong, in my view, to suggest that Shakespeare equates goodness with insouciance. The natures of figures like Desdemona and Billy Budd are pure and in a clear sense aesthetically beautiful; such natures can be deeply loved, even worshipped; but moral goodness, in this imperfect world, is—and must be, if it is to mean anything—an altogether more sophisticated business than sheer innocence. One recalls Milton’s words in the Areopagitica: ‘I cannot praise a fugitive and cloister’d virtue, unexercis’d and unbreath’d, that never sallies out and sees her adversary, but slinks out of the race, where that immortall garland is to be run for not without dust and heat’ (1954, 18). Kleist’s Prince Friedrich of Homburg does sally out to meet his adversary, but because he lives in a romantic dreamworld he disobeys orders and makes the attack at the wrong time, an error which almost costs him his life. His innocence is in a sense vindicated in the denouement of the drama, but not without dark hints of how it all nearly went terribly wrong.

18 Antigone and the Patriarchy The old orthodoxy that Shakespearean heroines such as Cordelia and Desdemona are perfect creatures has been challenged in recent decades, especially in feminist criticism. By contrast, in their treatment of the figure of Antigone, feminist writers have often seen their task as one of rebutting her critics and maintaining the myth of her perfection, albeit in a new guise.234 In particular, and against the kind of debunking interpretation of Antigone’s motives such as I have offered, feminist writers frequently regard her as a role model in the struggle against the patriarchy.235 But this reaction to Antigone strikes me as misconceived. It is true that, as Patricia Mills says, ‘Sophocles shows Creon to be a misogynist and a tyrant’ (1996, 75), and that Antigone is in some sense a brave woman who defies the patriarchy. But just resisting the patriarchy is not enough to make someone a feminist: it needs to be done for feminist reasons, and in Antigone’s case I have argued that it is not. Feminist critics, as

Antigone’s Holy Crime 119 I implied earlier, have focused too narrowly on the male–female dynamic of the Antigone, ignoring the more important adult–adolescent one. They have rather prematurely jumped on the bare facts of the play in order to maintain the idolizing tradition in another form, without sufficiently examining the detail of the clash between Sophocles’ two protagonists. For Tina Chanter, Antigone ‘contests [Creon’s] reasoning, the absolute nature of his claims, and the restriction of Theban royal, political lineage to one sex only’ (2010, 26; cf. 45), and she suggests (ibid., 24) that in refusing the authority of [Creon’s] decision, Antigone not only recalls the polis to its own proper origin, an origin from which it had deviated—by requiring the symbolic preservation, rather than the repudiation, of communal bonds; in doing so, she exposes the partiality of a political system that has decided in advance to exclude women’s voices from its purview. This conveniently ignores—as Antigone herself ignores—the fact that Polyneices had participated in an assault on Thebes whose aim was to destroy those very communal bonds, as well as the point that Antigone’s own ‘communal bonds’ are communal in a perverted way, which unfits her to stand as representative of the polis. The idea that Antigone is challenging the whole concept of patrilineal succession to the throne of Thebes gives her a general political motive that she patently does not have: as I have argued in this chapter, Antigone’s focus is much narrower and more personal than that. Nor do I think Mills is right to say that Antigone’s suicide is ‘a form of defiance against patriarchal domination’ (1996, 74): rather, it is an act of terrible despair. For Mills, Antigone ‘is the precursor of the women who, in the recent past, proclaimed the personal as political’ (ibid., 77): but, as I have suggested, Antigone’s defiance is deeply personal in its motivation, and Creon is mistaken to read her as aiming to subvert the state. In a wide sense of the word ‘political’, according to which it has to do with interpersonal power relations, it is true that the personal is (at least often) the political; but in a narrower sense of the term, according to which it has to do with particular forms of government of a polity, the personal is not the political. Pace Mills (and Virginia Woolf),236 it falsifies the play to recruit Antigone as an ancient suffragette.237 Again, Jean Elshtain has as an ideal ‘to see ourselves as Antigone’s daughters, determined, should it be necessary, to chasten arrogant public power and resist the claims of political necessity’,238 seemingly oblivious to the irony of the soubriquet, given that Antigone—unnaturally, as I have suggested, from the viewpoint of Sophocles’ contemporary audience—had no children. To enlist Antigone in support of ‘maternal thinking’ and ‘family values’, as Elshtain does, looks like quite a stretch (and is already in tension with Antigone’s significant name, which means

120  Tragedy and Moral Redress ‘against/instead of generation’).239 Equally bizarre is Luce Irigaray’s suggestion that Antigone represents ‘respect for the order of the living universe and living beings, respect for the order of generation’ (2013, 118). It is really quite odd to find a feminist arguing, as Irigaray does (ibid., 124), that Antigone’s passion for her dead brother is both natural and a duty to life. As Mary Rawlinson puts it, ‘One might wonder why abandoning the living sister in order to die for the dead brother should command respect [sc. among feminist critics]’ (2014, 103). Antigone’s fraternalism is, if anything, anti-feminist, and her treatment of Ismene the very reverse of sorority.240 (Compare Electra’s attitude to the sacrifice of her sister Iphigeneia in the three Electra plays.)241 It does not exemplify the chorus’s injunction in Euripides’ Helen: ‘women must help one another’ (329).242 But is this reading perhaps too crude? After all, sorority can take unexpected forms. The interpretation I have offered would fail if a suggestion of Bonnie Honig’s were right, reviving an earlier theory of W. H. D. Rouse, according to which the first of the three burials of Polyneices in the Antigone was performed by Ismene, not by Antigone as has generally been held, and the two sisters are in a conspiracy against Creon.243 Unfortunately there is no evidence for this interpretation and plenty of evidence against it. Honig is partly misled by a mistranslation of 434–5, where the sentry tells Creon that, in Honig’s rendering, ‘We interrogated her, charging her with offences past and present—she stood up to it all, denied nothing’ (2013, 159). But what Sophocles’ sentry actually says is not that he and his fellows charged Antigone with indefinite offences, past and present, but that they charged her with the past and the present offences (καὶ τάς τε πρόσθεν τάς τε νῦν ἠλέγχομεν / πράξεις); here the inclusion of definite articles plainly ties the offences down to the two burials; so what Antigone does not deny in the immediate sequel (435) is that she performed those two burials. The Rouse–Honig line requires us to say that at this point Antigone is covering for her sister; but what is the evidence for that reading? One problem with Honig’s presentation of her argument is that it is not made clear whether she thinks that Antigone is plotting with Ismene’s collusion and consent, or rather plotting in her sister’s interests but on her own and without Ismene’s knowledge. Honig’s advance notices of her view imply the former (2013, 2–3, 90, 151–4, 182, 195); when the defence of the view comes (ibid., ch. 6), however, it appears to be the latter that is maintained. But it is hard to see that there are any grounds in the text for either interpretation, and Antigone’s treatment of Ismene is evidence against Honig’s approach. Why, for example, does Antigone speak to Ismene with such cruelty in the presence of Creon? If this is to be a ruse designed to put Creon off the scent,244 as Honig’s reading requires us to say, there would need to be a hint of that in the text: but, as I mentioned (§12) in connection with Griffith’s (and Jebb’s) similar view, there is none.245 Antigone exculpates Ismene not in order to protect her, but because she does not want her sister to share in her glory.246 ‘I did the deed—if only she consents’

Antigone’s Holy Crime 121 (536), says Ismene when she is brought before Creon. As Honig herself concedes (2013, 164), if Ismene really did perform the first burial, she does not need Antigone’s consent. The reason why she seeks that consent is evidently that she did not do the deed, but now wishes to identify herself with Antigone’s defiance of Creon; and in order to do that, given what has passed between them, she needs Antigone’s agreement. Honig’s line is that Antigone’s martyrdom is really for her living sister, not her dead brother. In response to the claim that I made at the end of the last paragraph, namely that Antigone’s treatment of Ismene evinces the very reverse of sorority, Honig will say that what Antigone is aiming for is a subtle sorority (2013, 189); she does not, after all, have a death-drive but rather wishes to promote life—Ismene’s life (ibid., 30, 165). This would certainly vindicate one standard feminist reading of the play if it were right; but I think we have to admit that the interpretation is groundless.247 One welcome effect of feminist approaches to literature in general has been to dismantle the old goddess/saint/virgin vs whore/vixen dichotomies that formerly structured much thinking about the feminine in literature.248 On the other hand, feminist critics sometimes operate with stereotyping categories that are nearly as unsatisfactory as those they are seeking to replace, substituting their own crude binaries for the discarded ones, as when for example the critic, in Diane Purkiss’s words, ‘always sees scopophilia as bad, and theatrically problematized identities as good’ (1996, 208).249 In general criticism of older literature that is strongly motivated by distinctively modern political concerns risks a charge of anachronism. A small but sufficient example, relevant here, is afforded by a textual problem at Euripides, Medea 384–5, where Medea is wondering how to proceed against her enemies: ‘It would be best to proceed by the direct method, in which we (women) are especially skilled (sophai), and kill them with poison’. I here translate Gilbert Murray’s text. Murray follows the manuscripts, which read ‘sophai’ (the feminine plural of ‘sophos’, here meaning ‘skilled’) at the beginning of the line 385, but some recent editors, including James Diggle and David Kovacs, change ‘sophai’ to ‘sophoi’ (the masculine plural), which yields a reference to Medea alone (Kovacs translates: ‘. . . in which I am the most skilled’), by dint of the standard idiom according to which an individual woman who speaks of herself in the plural uses the masculine gender.250 Donald Mastronarde supports the change in his commentary. The transmitted text, he implies (ad loc.), is misogynistic, and he states that it makes Medea say something—namely that women are particularly skilled in magic, drugs, and skulduggery generally—that is not true. But characters in fiction say many things that are, objectively speaking, not true (as well as being misogynistic); further, the fact is that the view, whatever we think of it now, used to be widespread.251 And it is a familiar literary topos, from Sophocles’ and Pseudo-Seneca’s Deianeira to Euripides’ Creusa to Tacitus’ Agrippina to John of Garland’s lotrix to Guarini’s Corisca (Il Pastor Fido) to Shakespeare’s Goneril to Corneille’s Cléopâtre (Rodogune) to

122  Tragedy and Moral Redress Lessing’s Marwood (Miss Sara Sampson) and Orsina (Emilia Galotti) and to Wedekind’s (and Berg’s) Lulu.252 Most importantly—this is the decisive textual argument—Medea repeats the very same ‘charge’ against herself and her sex shortly afterwards, at 408–9, where no one has suggested that the use of ‘sophōtatai’ (superlative of ‘sophai’), paralleling that of ‘sophai’ at 385, is faulty; that indicates that we should not tamper with the vulgate reading at 385, either. A more specific problem with feminist criticism, in the context of the present discussion, is this. Feminist critics frequently shy away from overt censure of the character or conduct of literary, and in particular Shakespearean, heroines. Partly they do this because they feel that they must defend such characters against male criticism.253 Another reason is that these critics are often influenced by a literary-theoretical view according to which such criticism involves a category mistake, namely the supposed error of ‘treating fictional characters as though they were real people’: but this common view is itself based on a category mistake, as I shall explore in Part II below. A third reason why feminist critics are often unwilling to criticize Shakespearean heroines, in particular, is that they do not wish to risk association with those features of traditional criticism that evince a political leaning we would no longer endorse, such as Thomas Rymer’s notorious verdict on the fate of Desdemona’s marriage with Othello that it was ‘a caution to all Maidens of Quality how, without their Parents consent, they run away with Blackamoors’.254 Modern politically aware literary criticism wishes to distance itself as far as possible from ‘the racist, sexist, and colonialist discourses of his time’ to which, it has to be admitted, ‘Shakespeare was certainly subject’.255 But it must be possible to discern a flaw in Desdemona without adopting a politically suspect posture; correlatively, it is by no means always the case that feminist commentators on Shakespeare steer clear of disparaging his heroines.256 That is unsurprising: after all, if Cordelia and Desdemona are no longer being placed on pedestals as goddesses, it is thereby implied that they share the faults of ordinary mortals, and it should not then be invidious to spell out what those faults are. It is interesting that, just as in Antigone’s case a dismantling of her divinity was anticipated by Böckh in the early nineteenth century, so also the dethroning of Cordelia and Desdemona was thoroughly anticipated by Böckh’s compatriot Gervinus at around the same time. If Bradley had paid more attention to Gervinus’s writings, some elements of the feminist revolution that has recently taken place in Shakespearean criticism might, given his enormous influence on Shakespeare studies, have gained currency much earlier.

Notes 1. Cf. Hester 1971, 13; Bollack 1999, 76. 2. Butcher 1951, 310, 309; Schlegel 1966, vol. 1, 95; Hegel 1986, vol. 18, 509; De Quincey 2000–4, vol. 15, 315, 326; Jebb 1883–1908, vol. 3, xxii, xxv,

Antigone’s Holy Crime 123 xxxiv; Hofmannsthal 1979, vol. 3, 484. For the effusions of Shelley, Wagner, and Kingsley, see Poole 1987, 186–7; cf. Oudemans and Lardinois 1987, 107–10; Honig 2013, 155–6. On Jebb’s Antigone, see Winnington-Ingram 1980, 130, 139; McCoskey and Corbett 2012. 3. Mueller 1980, 31. 4. Böckh 1843, 167; Whitman 1951, 90; Kaufmann 1968, 253; D. Lucas 1972, 306; Vickers 1973, 527; Scodel 1984, 48; Nicolai 1992, 43; Lloyd-Jones 1990, 202 (cf. 204); 1983, 116; Mills 1996, 69; D. Schmidt 2001, 272. Cf. Raphael 1960, 19, 22; Trapp 1996, 83 n. 4. 5. Heidegger 1987, 112; 1993, 129. 6. See Hutchinson on 1005–78. 7. Lacan 1986, 306; Žižek 2004, 52; 2005, 306, 309; cf. Benardete 1999, 50. On Lacan’s Antigone, see Leonard 2006; Chanter 2010; Buchan 2012; Freeland 2013, ch. 5. I shall examine some feminist approaches in §18 below. 8. MacKay 1962, 166–70; Vickers 1973, 527–8; Sourvinou-Inwood 1989, 138–9; Holt 1999, 667–8; Lefèvre 2001, 84. 9. Griffith 1999, 139; Ahrensdorf 2009, 114–15. 10. Rosivach 1983, 207–8; Heath 1987, 73; Lardinois 2012, 59–60. Cf. Winnington-Ingram 1983, 32–3, 53 (on Aeschylus’ Polyneices). 11. See Rosivach 1983, 191–2; Oudemans and Lardinois 1987, 101; SourvinouInwood 1990, 23; Griffith 1999, 30; Osborne 2012, 274. 12. Sophocles, Ajax 1332–45; Linforth 1961, 191–3; Foley 1995, 140. 13. Rosivach 1983, 208; Whitehorne 1983, 137; Holt 1999, 663; L. Parker 2016, 340–1. 14. Cf. Denniston on Euripides, Electra 896–8. 15. R. Parker 1983, 43–8; cf. Osborne 2012, 277–8. 16. Sourvinou-Inwood 1989, 139 with n. 25; 1990, 23–4; Hutchinson on Aeschylus, Seven Against Thebes 1014. Note esp. Seven 1005–25, 1042–53; Seneca, Phoenician Women 540, with Frank ad loc. 17. See Aeschylus, Seven Against Thebes 580–3, with Hutchinson on 568–630. 18. See Xenophon, Hellenica I, 7, 22; Thucydides I, 138, 6. 19. Hester 1971, 19–20; Sourvinou-Inwood 1990, 27–8; Rösler 1993, 85–7; Benardete 1999, 27. Cf. Hutchinson on Aeschylus, Seven Against Thebes 1014, 1017; Mastronarde on Euripides, Phoenician Women 1630. 20. Hester 1971, 20–1; Oudemans and Lardinois 1987, 162; Easterling 1997b, 27–8; Holt 1999, 665–6; Houlgate 2007, 152; Lardinois 2012, 60; Osborne 2012, 277; Cairns 2016, 39–40; Griffith on Antigone 26–36. Contra Whitehorne 1983, 138; Heath 1987, 75; Nicolai 1992, 45; Harris 2004, 35–8; Burian 2010, 276. 21. Antigone 773. Cf. Sophocles, Electra 382 with Finglass ad loc. 22. Lefèvre 2001, 83; Allan 2008, 6. This point is ignored by Harris 2004 and 2012. 23. Rosivach 1983, 196–8; cf. Osborne 2012, 275. See e.g. Euripides, Helen 852–4. 24. See e.g. Euripides, Suppliants 526, 671; Rosivach 1983, 201. 25. Redfield 1994, 167–9; Harris 2004, 37–8; Cairns 2016, 37–9; Finglass on Sophocles, Ajax 829–30, 1363; Dunbar on Aristophanes, Birds 393–4. Cf. Suetonius, Augustus 13, 2; [Seneca], Octavia 515–16, with A. Boyle ad loc. 26. See Bowen on 802–3. 27. Aeschylus fr. 44; cf. Putnam 2001, 99–100. 28. Whitehorne 1983, 133–4. 29. Correr, Procne 926–30 (Grund 2011, 176). Cf. Seneca, Thyestes 747–53, 1032–4; La Chanson de Roland 1746–52; Wilmot, The Tragedy of Tancred and Gismund, G2v–3r; Muret, Julius Caesar 458 (Hagmaier 2006, 50); Alabaster, Roxana 1195–8, 1270–3; Shakespeare, Titus Andronicus V, 3, 194–9, with Crawforth, Dustagheer and Young 2015, 43.

124  Tragedy and Moral Redress 30. Bennett and Tyrrell 1990, 456; 1998, 18, 22, 27; Foley 1995, 135–6, 142; 2001, 185 n. 40. 31. Cf. Rosivach 1983, 199. 32. Cf. MacKay 1962, 169; Zeitlin 1986, 124. 33. Segal 1981, 176; Minadeo 1985, 154. 34. Bowra 1945, 72–8; Linforth 1961, 196, 248; Hogan 1972, 97; WinningtonIngram 1980, 125; Schmitt 1988c, 4–11; Blundell 1989, 123–6; Griffith 1999, 33, 48, 156; Foley 2001, 183–6; Lefèvre 2001, 79–80, 82–98; Garvie 2007, 183; Ahrensdorf 2009, 147; Greiner 2012, 61–81; Cairns 2016, 41–2. 35. Santirocco 1980, 183–4. Cf. Vickers 1973, 529, 542; Burian 2010, 278. 36. Scodel 2005, 246. 37. Benardete 1999, 70; Greiner 2012, 67. 38. Hegel 1986, vol. 14, 60; Knox 1964, 76–9; Segal 1981, 192; Friedrich 1996, 275–7; Butler 2000, 5–6; Vidal-Naquet in Vernant and Vidal-Naquet 2001, vol. 2, 160–1; Boedeker and Raaflaub 2009, 119–22; Lardinois 2012, 61; Osborne 2012, 277; Rose 2012, 261–5. 39. See Easterling 1997b, 28; Carter 2007, 111; Cairns 2016, 110–13. 40. So Nicolai 1992, 48–9. 41. Honig 2013, 95, 97; 2014. Cf. MacKay 1962; Foley 1995, 135–7, 141–2. 42. See Rhodes 2003, 114–15. 43. See S. West 1999, 124–9; Burian 2010, 279; Fletcher 2010, 168–70. 44. Aeschylus, Seven Against Thebes 1029–30; cf. 1042, 1060; Rösler 1993, 83–5. 45. Winnington-Ingram 1980, 138; Rösler 1993, 93; Holt 1999, 682–3; Carter 2007, 112; Finglass 2007, 403; Garvie 2007, 184; Harris 2012, 291; Rose 2012, 264. 46. Oudemans and Lardinois 1987, 183; Sourvinou-Inwood 1989, 144–6; 1990, 15–16. 47. Minadeo 1985, 153–4; Ahrensdorf 2009, 134–48; C. Holland 2010, 37; Cairns 2016, 7–8, 35–6, 99. 48. See Hegel 1986, vol. 15, 549–50; Oudemans and Lardinois 1987, 114. 49. Cf. Mastronarde 1994, 511. 50. So e.g. P. Hammond 2009, 101, relying on Griffith 1999, 209–10. 51. Hegel 1986, vol. 3, 539; vol. 17, 133; vol. 19, 60; vol. 20, 523, 549. So too Böckh 1843, 159–64. Cf. Dodds 1960, xliii; Trapp 1996, 79–80; Lefèvre 2001, 74–5; Houlgate 2007, 149–56; Burian 2010, 256–9; Greiner 2012, 132–47; Billings 2014, 182–3. 52. Silk and Stern 1981, 319; cf. Gellrich 1988, 48–9; Rowan Williams 2016, 56–66. 53. See Griffith on Antigone 199–201; Lefèvre 2001, 83, 115; cf. Fraenkel on Agamemnon 525ff. Much is made of this point in Aeschylus’ Seven Against Thebes, at 582–3 and 1009–10, as well as in Euripides’ Phoenician Women, e.g. in the exchange between Eteocles and Polyneices at 588–637. It is noteworthy that Mastronarde’s commentary on this play shows a curious bias against Eteocles in favour of Polyneices, parallel to the tendency among commentators on the Antigone to downplay Polyneices’ treachery. For example, when Eteocles issues his challenge to single combat (1226–45), assuming that these lines are genuine, he is criticized by Mastronarde for issuing it after lives have been lost, rather than before (note on 1217–63). That he ‘acted first, took the initiative’ (glossing 1223) in seeking to forestall further loss of life, albeit at this late stage, unlike his brother, apparently earns him no credit, and is passed over in silence. (See also 1994, pp. 27–8, and notes on lines 443–637 passim, 1365, 1376.) For a balanced assessment of Aeschylus’ treatment of the two brothers in Seven, see Hutchinson on 631–52. 54. Cf. Goldhill 1986a, 96; Oudemans and Lardinois 1987, 102, 161, 166; Harris 2004, 26–9; Lardinois 2012, 60; Austin and Olson on Aristophanes, Thesmophoriazusae 361–2. Contra Fletcher 2010, 179.

Antigone’s Holy Crime 125 55. Griffith on Antigone 304–5. Cf. Knox 1964, 75, 91, 101–2; Butler 2000, 51; Vernant in Vernant and Vidal-Naquet 2001, vol. 1, 34. 56. Scodel 1984, 48. Cf. Winnington-Ingram 1980, 121, 124. 57. Garvie 2007, 182. 58. See Holt 1999, 672–87. 59. Cairns 2016, 42, 46–8; cf. Garvie 2016, 34–5. 60. A. Brown 1987, 9. 61. Lardinois 2012, 60. 62. Cf. Sourvinou-Inwood 1990, 21–6; Bowen on Aeschylus, Suppliants 176. 63. Antigone 378, 423, 472, 561, 654, 693, 948, 987. See Dover on Thucydides VI, 12, 2, and 1994, 102; Hutchinson on Aeschylus, Seven Against Thebes, 686; Garvie on Persae 744. 64. Whitehorne 1983, 137; Oudemans and Lardinois 1987, 113; SourvinouInwood 1990, 30–1; Finglass on Sophocles, Electra 285; Dunbar on Aristophanes, Birds 474. 65. See e.g. Lane and Lane 1986, 167; Griffith 2001, 130. 66. Cf. Euripides, Helen 282–3, 1355–7, with Allan ad locc., and pp. 51–4 of his introduction. 67. Oudemans and Lardinois 1987, 167–8. 68. Winnington-Ingram 1980, 231. 69. See Burian 1997a, 184–5. 70. Oudemans and Lardinois 1987, 172, 192. 71. Aristotle, Poetics 1454a23–4, tr. Hubbard. Cf. Herrick 1955, 65; Finglass on Sophocles, Electra 983 (with further references). 72. For discussion see Easterling 1987, 17–22; Sourvinou-Inwood 1989, esp. 138; Dover 1994, 95–8; Foley 1995, esp. 145 n. 9; Hall 1997, 124–6; Bennett and Tyrrell 1998, 30; McClure 1999, 24–5; Mastronarde 2010, 248– 54; Dewar-Watson 2014, 61–2; Garvie 2016, 31–2. A similar question arises at the beginning of the Electra: see Finglass on 155, 312–13, 516–18, 617– 18; Wheeler 2003, 377–80. Cf. Aeschylus, Seven Against Thebes 232, with Hutchinson ad loc.; Seneca, Agamemnon 954. 73. See e.g. Stevens on Euripides, Andromache 85; Foley and Howard 2014, 627. Women are also remarkably prominent in English neoclassical tragedy: see Norland 2009, 12–13. 74. So too Knox 1964, 116; cf. Butler 2000, 76. 75. Santirocco 1980, 187–8; cf. Lloyd-Jones 1983, 113–17. 76. See here Cairns 2016, 66–81, 88–92. But, though Cairns accepts that infatuation does not diminish Creon’s responsibility (77, 90), he is much more tentative about Antigone (72, 88, 92, 113), and is even tempted by the idea that she is without fault (91). 77. Schmitt 1988c, 12–13. Rowan Williams (2016, 56–66) also takes Antigone’s words at face value. 78. Cf. McCall 1972, 105; Garvie 2007, 181; 2016, 33. 79. Griffith 1999, 212; cf. Garvie 2007, 181. In general on Antigone’s treatment of Ismene, see Blundell 1989, 111–15; Goldhill 2012, ch. 9. 80. Chaucer, Troilus and Criseyde I, 708–9. Cf. Euripides, Andromache 1041–2, with Stevens ad loc.; Phoenician Women 894–5; Aristotle NE 1171a29–30; Seneca, Trojan Women 1009, with A. Boyle ad loc.; Marlowe, Doctor Faustus II, 1, 42; Shakespeare, The Rape of Lucrece 789–91. 81. On Antigone 538–9. On (i) Griffith is following Jebb 1883–1908, vol. 3, xxix. See also Easterling 1990, 96, and Honig 2013, whose views I shall discuss below (§18). 82. Cf. Linforth 1961, 186. 83. Cf. P. Johnson 1997, 370, 390–1; Cairns 2016, 99.

126  Tragedy and Moral Redress 84. Cf. Böckh 1843, 155–6, 163, 165–6, 175; Winnington-Ingram 1980, 134– 5; Minadeo 1985, 148; Lacan 1986, 306; Blundell 1989, 113; Woolf 1992, 272; Lefèvre 2001, 103; Greiner 2012, 73; Cairns 2016, 94–5, 99. 85. Vickers 1973, 543. So too Winnington-Ingram 1980, 241; Cavarero 2010, 59. 86. See Goldhill 1986a, 83–8; Blundell 1989, 26–59; Dover 1994, 180–4; Aeschylus, Choephori 123; Sophocles, Ajax 79; Euripides, Medea 165, 809; Children of Heracles 881–2, 939–40; Andromache 438; Hecuba 749, 903–4; Heracles 585–6, 731–3; Orestes 544–601, 1101–2; Aristophanes, Frogs 1012, and 1994; [Euripides], Rhesus 483; with the notes of Garvie, Finglass, Mastronarde, Allan, Stevens, J. Gregory, Bond, Willink, Dover, and Liapis ad locc.; Seneca, Trojan Women 771–4. 87. Knox 1964, 28–34, 80–90; Griffith 1999, 122–3; Goldhill 2012, 31–2, 241. 88. P. Hammond 2009, 98; cf. Goldhill 1986a, 98. 89. See Aeschylus, Agamemnon 586; Sophocles, Philoctetes 93, 320; Euripides, Heracles 832; Orestes 1089–91; Children of Heracles 26–7, 826; Phoenician Women 394; with the commentaries of Fraenkel, Schein, Wilamowitz, Bond, Willink, Allan, and Mastronarde ad locc. 90. Rutherford 2012, 185 n. 41; cf. Segal 1981, 190. 91. Cf. Oudemans and Lardinois 1987, 180; Griffith 2001, 129; 2010, 116, 119; Lardinois 2012, 62. With line 523 cf. Euripides, Phoenician Women 1681. 92. Minadeo 1985, 141–2. 93. Hester 1971, 41; cf. 17. 94. Lacan 1986, 327; Cf. Rosivach 1979, 19–20; Oudemans and Lardinois 1987, 174; B. Williams 1993, 86; Benardete 1999, 61, 73; Eagleton 2010, 104–5; Freeland 2013, 163–4; Garvie 2016, 33–4. 95. Burian 2010, 283. 96. Cf. Cropp 1997, 154; Griffith 2005b, 113; Miller 2007, 1; Ahrensdorf 2009, 109. 97. Cf. Griffith 1999, 34 with n. 103. 98. Cf. Segal 1981, 34–5; Lacan 1986, 306; Benardete 1999, 63; Lardinois 2012, 63. 99. Aeschylus, Seven Against Thebes 691; Sophocles, Ajax 205, 885, 930; Euripides, Electra 27; Alcestis 64. 100. Oudemans and Lardinois 1987, 172; J. Gould 2001, 268. 101. Chanter 2010, 45; cf. Murnaghan 2009, 235. 102. Cf. Minadeo 1985, 140; J. Gould 2001, 268; Lardinois 2012, 64. 103. Cf. Griffith 1983, 10. 104. So e.g. Segal 1966, 543–4; 1981, 252–3. 105. See the analysis at Rutherford 2012, 71–2. 106. Housman 1997, 244. See Gaskin 2013a, 320, adding Aeschylus, Agamemnon 905; Euripides, Children of Heracles 539; Helen 835, with Allan ad locc.; Iphigeneia in Tauris 983, with L. Parker ad loc.; Seneca, Oedipus 291; Phaedra 677, with Coffey and Mayer ad loc. 107. Benardete 1999, 1 n. 2; cf. Goldhill 2012, 235–6. 108. See Oudemans and Lardinois 1987, 167; Butler 2000, 5–6; Žižek 2004, 54; Scodel 2005, 236; Cairns 2016, 112. 109. Steiner 1984, 208. Cf. Cavarero 2010, 50. 110. Winnington-Ingram 1980, 128. 111. Rawlinson 2014, 109. 112. ‘Euphrosyne’ 135 (1985–9, vol. 2, 193). 113. Minadeo 1985, 139; cf. Goldhill 2012, 241. 114. See Minadeo 1985, 146; Goldhill 2012, 110–11 with n. 5. 115. See also Sourvinou-Inwood 1990, 29; P. Johnson 1997, 375; Griffith 2001, 130; 2005b, 97, 113; Žižek 2004, 54; Burian 2010, 280; Mader 2014, 125–9.

Antigone’s Holy Crime 127 116. Such epithets are a tragic speciality: see Antigone 1071, Fraenkel on Aeschylus, Agamemnon 412; Aristophanes Frogs 204, with Stanford ad loc.; Griffin on Homer, Iliad IX, 63; Rutherford 2012, 70; Shakespeare, Hamlet I, 5, 77; Milton, Paradise Lost II, 185. 117. Cf. Electra at Sophocles, Electra 164–5 with Finglass ad loc.; Wackernagel 2009, vol. 2, 293. A reader queries my point about the negativity of alphaprivative compounds, asking ‘Do (e.g.) “unspoilt” or “immaculate” or “unsurpassed” give you a feeling of negativity?’. Well, they all refer to what is not such and such. Descartes famously attempted (in the third Meditation) to treat infinity as a positive concept, prior to finitude, but he was fighting forlornly against meaning. No matter how favourably disposed we may be towards the unspoilt or the infinite, these are still negative concepts. 118. Scodel 2005, 243. 119. Pace M. Nussbaum 1986, 66; Bennett and Tyrrell 1990, 109. 120. Cf. [Aeschylus], Prometheus Vinctus 1080, with Griffith ad loc. 121. Shakespeare, Measure for Measure III, 1, 118–32; Shelley, The Cenci V, 4, 48–75. 122. See Linforth 1961, 227. 123. Reinhardt 1947, 92–3. 124. See Knox 1964, 105–7; L. Parker on Euripides, Alcestis 293. 125. Minadeo 1985, 135. 126. Schmidt 2001, 97; cf. Billings 2014, 170–1. On Hegel’s Antigone see esp. Cairns 2016, 124–7. 127. Finglass on Sophocles, Electra 1058–97. I discuss the psychology of ancient characters in Part II. 128. Sourvinou-Inwood 1989, 140. Cf. 1990, 29, 32; Oudemans and Lardinois 1987, 112–13. 129. Gager, Meleager 1264; cf. 90, Norland 2009, 159–60. 130. See Eckermann 1981, 562 (28 March 1827). 131. So Jebb 1883–1908, vol. 3, 164, 257–61; Whitman 1951, 92–3; Linforth 1961, 227; Winnington-Ingram 1980, 145. See also Santirocco 1980, 189– 90; Rösler 1993, 88–99; Griffith on 904–15. 132. Murnaghan 1986, 194. 133. Cf. Putnam 2011, 61–5. 134. Cf. Böckh 1843, 168; Ahrensdorf 2009, 126–7; Burian 2010, 280–1. 135. Sourvinou-Inwood 1990, 18–20. 136. Cf. Neuburg 1990, 67–8; Ahrensdorf 2009, 132–3. 137. Bayley 1981, 173. 138. Pinkard 2015, 146. 139. Though not in quite the same way, as a reader points out to me: Antigone starts lonely; Creon becomes lonely. Even so, they both create their lonelinesses. 140. Taxidou 2004, 44, reporting the views of Jean-Joseph Goux. 141. Cf. Silk 2000, 94–7. 142. Linforth 1961, 249–50; Winnington-Ingram 1980, 130; Lane and Lane 1986, 174; Griffith 1999, 33, and 2010, 115–19, 123 (though contrast 2005b, esp. 95–6, 110). 143. Assuming, with most modern editors, that lines 572 and 574 should be assigned to Ismene. For discussion of this crux see Dawe 1978, vol. 3, 106– 8; M. Davies 1986; Lloyd-Jones and Wilson 1990, 127–8; Benardete 1999, 74–5; Griffith ad loc.; 2001, 123; 2010, 115 n. 14. 144. See Mueller 1980, 36–7; Burian 1997b, 233–5. 145. Cf. Steiner 1984, 160; Oudemans and Lardinois 1987, 112–13, 172; Butler 2000, 6, 17; Torrance 2010, 245–6; Cairns 2016, 104–6. Contra Vernant in Vernant and Vidal-Naquet 2001, vol. 1, 89–90; Irigaray 2013, 118.

128  Tragedy and Moral Redress 146. See Oudemans and Lardinois 1987, 97–8; Hall 1997, 109–10; Wheeler 2003, 381; Cairns 2016, 110; Cropp on Euripides, Electra 44. 147. Butler 2000, 79–80; Žižek 2004, 54; Griffith 2005b, 95–6; 2010, 115, 119–20; Miller 2007, 8; Randhawa 2014, 299–300. See also P. Johnson 1997, 391–3. 148. M. Nussbaum 1986, 65. Cf. Bollack 1999, 4. 149. Wilson Knight 1986, 39–42, 299, 319–20; cf. P. Edwards 1992, 20–2. 150. Cf. Goldhill 1986a, 102–3, 117–19; Söderbäck 2010, 76. 151. See Cairns 2016, 132–54. 152. So too Silk and Stern 1981, 318; D. Schmidt 2001, 14; Vidal-Naquet in Vernant and Vidal-Naquet 2001, vol. 2, 161; cf. Rowan Williams 2016, 67–8. 153. Cf. J. Fleming 2006, 172; Cairns 2016, 136. 154. Introduction to Anouilh 1954, 22. So too, in effect, Steiner 1984, 193; J. Fleming 2006, 178. 155. Lacan 1986, 293; cf. J. Fleming 2006, 174–5; Cairns 2016, 111. 156. On Philotas, see Pütz 1986, 110–18; on Troilus, see Boitani 1989b, 290. 157. Lacan 1986, 293; cf. J. Fleming 2006, 168, 179–82; Deppman 2012, 524–5, 532. 158. Introduction to Anouilh 1954, 23. 159. Lefèvre 2001, 105. 160. Knox 1964, 65; Griffith 1983, 9–10; Minadeo 1985, 137, 146; Kirkwood 1994, 120; Lefèvre 2001, 101. 161. Winnington-Ingram 1980, 144; Foley 2001, 31. 162. Bennett and Tyrrell 1998, 67. Cf. Dué 2012, 247. 163. See Oudemans and Lardinois 1987, 189–90; Schmitt 2011, 493; Lardinois 2012, 64. 164. See Seneca, Trojan Women 861–87, with A. Boyle 1994, 24–5, 27, 208. 165. Lucretius I, 79–101; cf. Seneca’s Clytemnestra at Agamemnon 163–8 (see Tarrant on 163). 166. On this motif in the Antigone, see Griffith 1999, 52, and notes on lines 653–4, 806–16, 893–4. See also Euripides, Medea 985; Children of Heracles 579–80; Electra 247; Iphigeneia in Tauris 369–71; Orestes 1109; with the commentaries of Mastronarde, Allan, Cropp, L. Parker, and Willink ad locc.; Segal 1981, 179–83; Loraux 1985, 51–3, 68–75; Seaford 1987, 107– 8; Sourvinou-Inwood 1990, 17; Goldhill 1990a, 103–4; Rehm 1994, ch. 4; Seidensticker 1995, 157; Bennett and Tyrrell 1998, 93–121; M. Davies 1999, 38; Rutherford 2012, 144–6; Lefèvre 2015, 339; Cairns 2016, 107–10. 167. See Morwood 2007, 219–20, and note on line 1022. 168. Annals XIV, 63; cf. A. Boyle 1994, 29. 169. [Seneca], Octavia 593–7; cf. also 23–4, 721–3, 750–1, 899–900, with A. Boyle’s notes ad locc. (cf. his 2006, 207). 170. See Mastronarde on 1487. 171. Cf. Euripides, Orestes 205–6; Helen 282–3, 689–90, 933, 1476–7; Iphigeneia in Tauris 220; with the commentaries of Willink, Allan, and L. Parker ad locc. 172. L. Parker on Euripides, Alcestis 915–17. 173. Cf. Rutherford 2001, 264–70; Peters 2009. 174. Euripides, Alcestis 940; Bacchae 1296; Hippolytus 1401, 1403; cf. Hecuba 687. 175. Dale 1954, xxii. On the ‘too late’ motif generally in tragedy, see Silk 1996b, 469–70. 176. Racine, Andromaque 1393; Athalie 1731; cf. J. Campbell 2005, 192. 177. See e.g. 392–3, 1059, with Bowen ad locc.

Antigone’s Holy Crime 129 78. So Oudemans and Lardinois 1987, 173, 189. 1 179. So Linforth 1961, 186. Cf. Electra at Sophocles, Electra 345–6. 180. Cf. Friedrich 1996, 273. 181. Cf. Knox 1964, 114. 182. Tr. Lloyd-Jones, adapted. On the meaning of line 921 see Oudemans and Lardinois 1987, 192. 183. Pace Steiner 1984, 282; Nicolai 1992, 49. 184. Winnington-Ingram 1980, 138. Cf. Griffith on [Aeschylus], Prometheus Vinctus 747–8. 185. See Minadeo 1985, 136; Oudemans and Lardinois 1987, 188, 190. 186. Antigone 559–60; Fraenkel 1978, vol. 3, 585. 187. Pace Knox 1964, 62, 66–7 (but compare 103), 73; Kirkwood 1994, 52. 188. Cf. Cairns 2006, 112–13; Rutherford 2012, 392–8. 189. The same applies to Creon: Lardinois 2012, 64. 190. See Theognis 255–6; Sophocles, fr. 356; Aristotle, NE 1099a24–8; Euripides, Andromache 368–9, with Stevens ad loc. 191. Morse 1993, 12. 192. Chaucer, The Knight’s Tale 1260–1. 193. Cf. N. Boyle 2010, 1083. 194. Hester 1971, 38. 195. Böckh 1843, 159. See also Lefèvre 2001, 112–13; Sewell-Rutter 2007, 119–20. Contra Halliwell 1990b, 177. 196. So Reinhardt 1947, 80. 197. Lefèvre 2001, 38–9. 198. So Dawe 1968, 113. 199. S. Adams 1931; McCall 1972; Minadeo 1985, 143. Cf. Winnington-Ingram 1980, 125 n. 31; Segal 1981, 159–61. 200. Cf. Bennett and Tyrrell 1998, 57. 201. See Whitehorne 1983, 132. 202. Bennett and Tyrrell 1998, 59–60; cf. McCall 1972, 112. 203. See Iliad XXIII, 184–91; XXIV, 18–21, with Macleod ad loc. 204. Shakespeare, King Lear I, 2, 118. Throughout this book I cite this play in Foakes’s text, which is a compilation of Q and F. 205. Bradley 1991, 40. Cf. Miola 1992, 24–5; Bate 1995, 22–3. 206. See e.g. Wilson Knight 1986, 186–95, 204–6; Cooper 2010, 164, 226–7; Marcus 2016, 431–4. 207. Steiner 1961, 353. See Kaufmann 1968, 225. 208. Cf. Shell 2010, 194, 205. 209. Steiner 1984, 273. Cf. Halliwell 1990b, 177–8. 210. See Budelmann 2000, 169–71; Goldhill 2012, 159–61; Rutherford 2012, 376–7. Cf. Sophocles Philoctetes 451–2, and Electra 823–5, with Finglass ad loc. 211. See Cooper 2010, 137–8. 212. Schopenhauer 1986, vol. 2, 561; cf. §51, vol. 1, 354. 213. Raphael 1960, 19; Kaufmann 1968, 323; Wilson Knight 1986, 177 (but cf. 198); Miola 1992, 173; Nuttall 1996, 88–90; 2007b, 300–12; Cavell 2003, 73; Medcalf 2007, 103–11. Cf. Foakes 2000, 34–5; on Bradley’s Lear, see Lobsien 2014. 214. See e.g. Werner 2009. 215. Hegel 1986, vol. 3, 348. Hegel refers to (but mistranslates) Antigone 926: cf. Billings 2014, 173–4. 216. See here Foakes 2000, 38–9. 217. Cf. Bayley 1981, 27. 218. Cf. Bradley 1991, 230–1. 219. See here also Gellrich 1988, 220–1.

130  Tragedy and Moral Redress 220. Bradley 1991, 295. Cf. Bayley 1981, 27; Nuttall 2007b, 304; Eagleton 2015, 115–17. 221. See further on Cordelia Gervinus 1872, vol. 2, 208–11; H. Berger 1997, 41–6. 222. Bradley 1991, 190. Cf. McGinn 2006, 86–8. 223. Cf. Dusinberre 2003, 73. 224. Cf. Bradley 1991, 193–4, 292. 225. Gervinus 1872, vol. 2, 54; cf. Duncan-Jones and Woudhuysen 2007, 45 (on Lucrece). 226. See Heller 2002, 49–51; S. Hamilton 2003, 85. Cf. Wilson Knight 1986, 39. 227. For a sophisticated and persuasive exploration of this theme, see H. Berger 2004. 228. Or indeed Bianca’s: see here Jardine 1996, 30–1. 229. [Aeschylus], Prometheus Vinctus 887–93; cf. Tillyard 1956, 17–18. 230. Cf. Cook 1980, 188–9. 231. Gervinus 1872, vol. 2, 93. Cf. Gaskin 2013a, 100–2. 232. See further Rosenberg 1993, 6–7, 207–9. 233. Cf. Lamport 1990, 175. 234. See e.g. Foley 2001, 196–200. Cf. Rawlinson 2014, 101–3. 235. So e.g. Wohl 2009, 158–9; Cavarero 2010, 55–6, 60. 236. Woolf 1992, 395; Mills 1996, 88 n. 25. 237. Cf. Allan 2008, 53–5. 238. Elshtain 1998, 375; cf. Wohl 2009, 159; C. Holland 2010, 29. 239. Benardete 1999, 55, 61, 99, 111–12; Griffith on [Aeschylus], Prometheus Vinctus 85–6. 240. Rawlinson 2014, 103–4, 106, 112–16; cf. Butler 2001, 39. 241. See Aeschylus, Choephori 242; Sophocles, Electra 566–76; Euripides, Electra 1086; with the commentaries of Garvie, Finglass, and Denniston ad locc. 242. Tr. Kovacs. Cf. Euripides, Iphigeneia in Tauris 1060–2, with L. Parker ad loc. 243. Rouse 1911; Honig 2013, ch. 6. Cf. Calder 2005, 81. 244. As originally proposed by Simpson and Miller 1948. Cf. Scodel 2005, 243. 245. Cf. Winnington-Ingram 1980, 134–5. 246. Calder 2005, 85. 247. Cf. Goldhill 2012, 247. 248. See e.g. Garner 1976, 233–5; McLuskie 1994, 97; Boose 2004, 38. Cf. Walcot 1984. 249. See further Vickers 1994, ch. 6; L. Parker on Euripides, Alcestis 285. 250. Kühner–Gerth 1898, vol. 1, 83; Wackernagel 2009, vol. 1, 99–100. 251. As Mastronarde himself notes ad loc. and on 408–9, 419–20; cf. 2010, 271–9; Denniston 1939, xxviii–xxix; Easterling 1987, 15; Kovacs 1987, 51; Seidensticker 1995, 159–64; Zeitlin 1996, 356–61; McClure 1999, 26–9; Griffith 2001, 124–5; Schiesaro 2003, 73; Garvie 2007, 184–5; Rutherford 2012, 306; Pucci 2016, 25. See e.g. Aeschylus, Seven Against Thebes 187–95; Euripides, Andromache 85; Iphigeneia in Tauris 1032–3; Aristophanes, Thesmophoriazusae 430; Seneca, Agamemnon 116–22; [Seneca] Octavia 31, 164–5, 868–9, with the commentaries of Hutchinson, Stevens, L. Parker, Austin and Olson, Tarrant, and A. Boyle ad locc. 252. For John of Garland, see Lawler 1974, 138, 140 (lines 72–85, 103–33). Cf. Stefonio, Crispus I, 1, 125; Hughes, Misfortunes of Arthur I, 2, 10–16 (Cunliffe 1912, 229). 253. See e.g. Woodbridge 1977; Neely 1981, 6. 254. Rymer 1956, 132; cf. Honigmann 1997, 29. 255. K. Newman 1991, 93; see also Bartels 2004. 256. See e.g. Garner 1976; Cook 1980; Neely 1980.

3 From Cognitive Failure to No-Fault Tragedy

19 The Operatur Principle Revisited In the last chapter (§16) I observed that nothing in the Women of Trachis requires the explicit presence of Aphrodite as an external and autonomous agent; the action of the drama makes excellent sense when explained in purely human terms. Similarly, Aphrodite speaks the prologue in Euripides’ Hippolytus, but if her part were excised it would not be missed. Recall the principle (§8) that operatur deus in unoquoque secundum eius proprietatem: the god acts in each person according to his or her (that person’s) character. This principle applies to Shakespeare as much as it does to classical tragedy.1 As Gervinus noted, echoing the Heraclitean saying that I discussed in §8, ‘in Shakespeare fate is nothing other than a person’s own nature’ (1872, vol. 2, 179); recall also A. C. Bradley’s remark (§7) that the dictum ‘character is destiny’ contains a vital truth. Against this naturalizing approach Steiner writes: ‘To ask of the gods why Oedipus should have been chosen for his agony or why Macbeth should have met the Witches on his path, is to ask for reason and justification from the voiceless night. There is no answer. Why should there be?’ (1961, 128–9). It may be true that there is no a priori reason why any tragedy has to represent its agents as deserving the suffering they incur, and it is ungainsayable that some tragic suffering is represented, in the European tradition, as undeserved, a point which will occupy us more extensively below in this chapter. It remains the case, however, that most tragic suffering, and especially that of protagonists, is represented, in that tradition, as flowing from agent error, itself seen as grounded in character; and that, I have suggested, provides us with a kind of moral redress.2 As I remarked in Chapter 1, the notion that tragic agents bring suffering upon themselves by making mistakes that flow from their characters and dispositions is derived from, on the one hand, the Aristotelian idea that tragic action should not be random, but unfold according to principles of probability and necessity, and, on the other hand, the Leibnizian principle of sufficient reason, which is essentially the thesis that there are

132  Tragedy and Moral Redress no causal or explanatory deficits in nature: ex nihilo nihil fit, nothing comes of nothing.3 Certainly the two cases that Steiner mentions, Oedipus and Macbeth, conform to this traditional pattern, Oedipus in ways that we have already explored in detail, and Macbeth because there is, pace Steiner’s assertion quoted above, a reason why the witches appear to him and speak to him in the way that they do, and that reason lies in Macbeth’s own dispositions.4 Hugh Lloyd-Jones, like Steiner, rejects the idea that the fate of a tragic protagonist lies in his or her character, asking ‘Is this any truer of Macbeth than it is, say, of Sophocles’ Ajax?’ (1990, 202). No, it is equally true of both. There is a reason why the dagger appears to Macbeth, namely: ‘Thou marshall’st me the way that I was going’ (II, 1, 42). ‘Macbeth did not murder Duncan because he saw a dagger in the air: he saw the dagger because he was about to murder Duncan’.5 Again, Banquo’s ghost appears to him, and not to the lords who are guests at the feast, because Macbeth’s guilty conscience knows that ‘It will have blood, they say. Blood will have blood’ (III, 4, 121). Most strikingly of all, Macbeth starts when he hears the witches’ prophecy, because it echoes thoughts he has already had.6 In a remark that also recalls Heraclitus’ dictum, David Bevington notes that in Shakespeare’s Troilus and Cressida ‘Character is fate in Hector, as in the greatest of tragic heroes. He chooses what is destined. He chooses freely and yet is hemmed in by being who he is’.7 That does not simply apply to the greatest tragic heroes; it applies to all of us. We are all hemmed in by being who we are, without (save in pathological cases) derogation from our freedom.8 Rowan Williams accepts that, in tragedy, character is destiny—he speaks of ‘an appropriateness in the character’s fate to the nature of that character’—but he thinks that this dictum does not ‘remotely’ apply ‘in the sense of punishment for misdeeds’ (2016, 72–3). That is because he also thinks that ‘classical tragic narrative is not about flawed choices made by individuals who might or should have known better; it is about characters who in a significant sense do not choose at all’ (ibid., 58). I have argued that tragic protagonists make free choices; they do so even as—in fact precisely because (assuming that certain other conditions, in particular a rationality constraint, are met)—they play out their natures; they make mistakes they should not have made; they could and should have known better (at least constructively, and often actually, they did know better); and they are punished. That is true not merely of tragic heroes and heroines; it is true of real people. Is Williams setting an impractical and incoherent (perhaps incompatibilist) standard for freedom? Or does he think that no one is ever free? Given the immense range of decisions portrayed by tragic writers, it is hard to see how, if their protagonists are never free, any real person could ever be, either. Because Renaissance drama was written after the Reformation, there is a temptation to read into it the (notably, but not exclusively) Calvinist doctrine of predestination, this being the form, according to the

From Cognitive Failure to No-Fault Tragedy 133 tempting line of thought, that ancient fate now took.9 But the temptation should be resisted. Alison Shell thinks that The Two Noble Kinsmen portrays ‘man’s lack of agency before the divine’ (2010, 221), and is ‘a virtuoso exercise in predestinarianism—nothing at all to do with repentance, scarcely to do with character, everything to do with divine decretal’ (ibid., 222). Scarcely to do with character? Earlier, Shell has conceded that ‘the play’s quasi-tragic denouement can certainly be put down to human causes as well as divine ones—Palamon and Arcite’s stubborn rivalry, Theseus’ imposition of the death penalty for the loser of the tournament, Emilia’s reluctance to choose between her two suitors’ (ibid., 219), but she still wants to say that ‘the notion of divine immanence is pervasive throughout, preparing the audience for an ending precipitated by external event rather than character’ (ibid.). Well, what role do the gods really play? They show signs—thunder, doves, a rose tree—and the spark that frights Arcite’s horse and causes him to fall fatally may have a supernatural cause (Pirithous reserves judgement on the matter),10 but apart from that—nothing. To be sure, the main agents appeal to the gods, as Lear and Gloucester and so many tragic protagonists do, but, as in the case of King Lear, the action proceeds, and is explicable, in almost entirely naturalistic terms, or perhaps we should say: in entirely naturalistic terms. For what do the signs really amount to? Are they more than epiphenomena? After Arcite’s death Theseus tells Palamon (V, 6, 105–9): The powerful Venus well hath graced her altar And given you your love; our master, Mars, Hath vouched his oracle and to Arcite gave The grace of the contention. So the deities Have showed due justice. In other words, Arcite won the duel but Palamon got the woman: we have to do, yet again, with the ex post facto significance of divine intervention. Of predestinarianism there is not a hint in the play: the fact that it was preached in the pulpit does not mean that we should automatically detect the doctrine on the stage. There is actually a better case for finding it in Shakespeare’s Chaucerian model, The Knight’s Tale, for there the gods intervene much more actively in the plot: in particular, Chaucer has Arcite’s horse rear because Saturn sends a fury to terrify it. But in Chaucer also the operatur principle and its concomitant naturalism reign. As for the rearing horse, ‘perhaps too the ground was slippery’, suggests A. C. Spearing;11 or perhaps it was just that the ground was slippery. So I think that Steiner and others underestimate the extent to which, in Shakespeare, as also in Racine and Greek tragedy, ‘the supernatural is always placed in the closest relation with character’, as Bradley put it in respect of Shakespeare; ‘It gives a confirmation and a distinct form to inward movements already present and exerting an influence’ (1991, 30).

134  Tragedy and Moral Redress Steiner concedes the point for Goethe’s Faust who, he says, ‘assumes his role voluntarily’ (1961, 199); but the point holds much more generally, starting with many of the other depictions of Faust in the tradition (§8). As for Racine, Odette de Mourgues well remarks that The actions of the gods do not interfere with [the] dramatic action because they coincide with it. In Phèdre, for instance, the human element, Phèdre’s passion, and Vénus’ decision to destroy her are completely fused. . . . The furies who drive Oreste to madness [in Andromaque] come from his own self-torturing mind. . . . The intervention of Neptune [in Phèdre] enhances the dramatic horror and has a poetic value but it does not alter the mechanism of psychological reactions which lead Hippolyte to his doom.12 This is exactly right about Euripides as well as Racine. Malcolm Heath suggests that psychologizing Aphrodite in Hippolytus would only be plausible ‘if it were a natural or psychological law that women of impeccable integrity fall in love with their stepsons whenever the latter are sexually abstemious; but if that absurd premise is not granted, then to direct attention in any degree away from the figure of Aphrodite as seen in the prologue is to render the play unintelligible’ (1987, 53). But why should naturalizing Euripides’ plot require Heath’s suggested psychological law? The action can be explicable in purely naturalistic terms without our needing to have recourse to any such—to be sure, absurd—law: it will be explicable in terms of the detail of the plot and ­characterization—­explicable, that is, in subtler terms than is envisaged in Heath’s law. He continues: ‘Unless there is a causal connection between Hippolytus’ abstention from sexual intercourse and his fate, the whole action is turned into an odd, implausible and meaningless coincidence; and such a connection can only be provided by the machinations of the person we see in the prologue, an individual possessed of an intelligence which responds ruthlessly to slights on her personal honour’. Heath is right that the causal connection he mentions needs to be there, but it does not follow that this connection can only be furnished by an undemythologized Aphrodite. The connections can be—and, I suggest, are—natural. It is true, as we have said (§8), that once a divine apparatus is up and running, artists are free to use it in ways that cannot be immediately and straightforwardly psychologized: Euripides does so in the Hippolytus in the figure of Artemis, and generally in his deus ex machina scenes. (Aristotle’s worry about such scenes is that they will not be organically connected with the main plot; Horace by contrast thinks that they are the only place where use of divine intervention can be justified.)13 There will be a continuum of cases from pure allegory at one extreme (Venus and Mars in the passage from The Two Noble Kinsmen quoted above) to the integral involvement in the plot of autonomous divine agents (Jupiter

From Cognitive Failure to No-Fault Tragedy 135 and Mercury in Plautus’ Amphitruo) at the other. But the asymmetry to which I have drawn attention remains in force: natural processes and dispositions are what, fundamentally, drive the supernatural, not vice versa. We should distinguish, as I have said, between a primary or basic level, at which supernatural elements can simply and immediately be factored out by the spectator or reader in favour of natural explanations, and a secondary or derived level at which they cannot, because the supernatural is now being treated by the artist as having independent reality. That autonomy, to the extent that the supernatural elements in any given work enjoy it, is essentially mediated by dependence on the natural, taken as conceptually prior. There is no conflict between this naturalism and the evident fact that agents of tragedy often believe in the reality of divine prompting, and often think that divine impulses leave them no choice. What goes on in the work, and in particular what its agents say, is quite distinct from the things we say about it, and from the things that the work itself says. ‘Impitoyable Dieu, toi seul a tout conduit’, is the cry of Racine’s Athalie (at line 1774). ‘But is the God of Athalie necessarily the God of Athalie?’, asks John Campbell, pertinently. And: ‘It is only natural for Thésée to cry “Inexorables Dieux, qui m’avaient trop servi!” ([Phèdre], 1572), but the rhetoric cannot hide the simple fact that he has sprung a trap of his own making’.14 H. B. Nisbet notes that Lessing’s Miss Sara Sampson ‘is full of religious images and terminology. But they are employed in a secular and naturalistic sense, and the supernatural element inherent, for example, in the Lutheran theology of justification and grace is conspicuously absent’; he adds that the ‘use of religious language to reinforce and intensify secular emotions (as when applied by young men to attractive women) is a common feature of literary sensibility in Germany, from Klopstock’s love poetry to Goethe’s Werther’ (2013, 202). Stephen Greenblatt suggests that ‘the psychological in [Hamlet] is constructed almost entirely out of the theological’ (2001, 229); but with Bradley’s apt remark in mind (‘the supernatural is always placed in the closest relation with character’) I think we will find this claim the wrong way round, for it is rather that the theological is constructed out of the psychological, both in respect of Hamlet and generally. Even if the ghost of old Hamlet is not simply a figment of young Hamlet’s imagination— and it would be hard to maintain that, though W. W. Greg famously tried (1917)—still the father tells the son what he is ready to hear (‘O my prophetic soul!’), and lays on him a task he is primed to undertake.15 And, in general, when we find applications of the Senecan technique of placing a ghost at the start of a drama, calling for revenge, as in the Agamemnon and Thyestes, we also find that the motif can be allegorized quite straightforwardly: so, for example, Diomedes in Correr’s Procne; Gorlois in Hughes’s Misfortunes of Arthur; Cupid in Wilmot’s Tancred and Gismund;16 Mégère in Garnier’s Porcie; Moleon, Death, and Suspicion in Alabaster’s Roxana; Revenge in Kyd’s The Spanish Tragedy; Sulla

136  Tragedy and Moral Redress in Jonson’s Cataline His Conspiracy.17 In Webster’s The White Devil, Francisco conjures up an image of his dead sister (IV, 1, 97–101): To fashion my revenge more seriously, Let me remember my dead sister’s face: Call for her picture: no; I’ll close my eyes, And in a melancholic thought I’ll frame   [Enter Isabella’s Ghost] Her figure ’fore me. Or has he succeeded in raising the dead sister herself, as is suggested by the stage direction? The scene is, as Christina Luckyj notes (ad loc.), ‘highly ambiguous’. She adds that ‘Francisco’s view of the ghost as a figment of his own imagination is in tension with a long stage tradition of unquestioned presentation of ghosts’. But Webster is only exploiting an implicit feature of that tradition, namely that ghosts, whether they are mere figments of the imagination or not, appear to characters who are disposed, for good psychological reasons, to see them. In the Aeneid Anchises appears to Aeneas (IV, 351–3), in Locrine Albanact appears to Humber (III, 3),18 in Shakespeare’s Julius Caesar Caesar appears to Brutus (IV, 3), in Webster’s The White Devil Brachiano appears to Flamineo (V, 4), and in Middleton and Rowley’s The Changeling Alonzo appears to De Flores (IV, 1; V, 1), in each case as a projection of the visited agent’s own fear and guilt.19 In Lope de Vega’s El Caballero de Olmedo the ‘masked shadowy figure’ that appears to Don Alfonso represents, as Alfonso himself recognizes, a premonition of his own death (2257–302). In Hardy’s The Dynasts, ‘a vision passes before Napoléon as he lies, comprising hundreds of thousands of skeletons and corpses in various stages of decay. They rise from his various battlefields, the flesh dropping from them, and gaze reproachfully at him’.20 This on the eve of Waterloo. In Aeschylus’ Persae, Darius’ mood fits that of his addressees Atossa and the chorus: as A. F. Garvie comments, ‘Darius is not a happy ghost’.21 (Very few literary ghosts are.)22 In general, ghosts and demonic visitants from a world beyond prompt the willing. They reflect and give concrete form to the mental condition of the visited, and do not convey an authorial metaphysical statement: thus Troilus’ appearance to Hecuba in Loschi’s Achilles;23 her mother’s to Althaea in Gager’s Meleager;24 Erictho’s to Syphax in Marston’s The Tragedy of Sophonisba (IV, 1), and Andrugio’s to Antonio in the same author’s Antonio’s Revenge (III, 1); Satan’s to the Pope in Milton’s In Quintum Novembris. Again, Phaedra in Stefonio’s Crispus is told by a demon to throw a torch into Fausta’s chest, as Allecto does to Turnus; but Fausta, like Turnus, is primed (I, 1, 189).25 Steiner actually contradicts the claim about Oedipus and Macbeth that I quoted at the start of this section when he states that ‘Before Kleist,

From Cognitive Failure to No-Fault Tragedy 137 tragedy embodies the notion of moral responsibility. There is a concordance between the moral character of the tragic personage and his destiny,’ and he goes on to apply this point to Oedipus, Lear, Othello, and Macbeth. In all of these cases ‘we assume some measure of causal and rational dependence between the character of the man and the quality of the event’ (1961, 222). That now seems to me to be almost exactly right, as against Steiner’s first quoted view. Oedipus we have dealt with fully, Lear and Macbeth in much less detail, but they can be accepted good illustrations of Steiner’s point. Othello has sometimes been regarded as an innocent victim, but that view was decisively refuted by F. R. Leavis in a famous essay. Iago succeeds in part because ‘the essential traitor is within the gates’, and in the ‘testing [to which Iago subjects him], Othello’s inner timbers begin to part at once’.26 I would not disagree with Steiner’s implication that the standard model breaks down in the modern period, but it seems to me that Kleist still fully conforms to it. Steiner mentions Käthchen of Heilbronn and the Prince of Homburg as examples of heroes who escape the traditional insistence on the protagonist’s moral responsibility for his or her fate, since these figures are subject to visions which govern their actions. But the causal–rational dependence of event on character is just as firmly in place in their stories as it is in pre-Kleistian literature. Their visions, like the divine interventions and the appearances of ghosts and demons in earlier tragedy, are carefully targeted: the visions happen to Käthchen and the Prince because of the sort of people they are. So too with Penthesilea, whose ‘foolish heart’ (‘töricht Herz’) is her fate (1281).27 Visionary Kleistian heroes and heroines are fully responsible for their actions.

20 Rationalism and ‘the Death of Tragedy’ Steiner’s position on the death of tragedy is a modification of Nietzsche’s in Die Geburt der Tragödie.28 There Nietzsche, taking cues from Aristophanes and the Schlegel brothers,29 had argued (in §§11–14) that the tragic spirit, still strong in the irrationalistic Aeschylus and Sophocles, was given its quietus by Euripides’ Socratic optimism and rationalism. Steiner agrees that rationalism killed tragedy (1961, 23), but he locates the moment of demise not in the advent of Euripides on the Athenian stage, but much later, in the flowering of the scientific spirit in the late seventeenth and early eighteenth centuries. Racine still ‘stands on the far side of the chasm’ (looking back from our vantage point), but only just. Après Racine, le déluge. Shakespeare is firmly on the distant side of the divide; he ‘is closer to Sophocles than he is to Pope and Voltaire’, who are on our side. ‘The modes of the imagination implicit in Athenian tragedy,’ we are told, ‘continued to shape the life of the mind until the age of Descartes and Newton,’ but ‘with the Discours de la méthode and the

138  Tragedy and Moral Redress Principia the things undreamt of in Horatio’s philosophy seem to pass from the world’: In Greek tragedy as in Shakespeare, mortal actions are encompassed by forces which transcend man. The reality of Orestes entails that of the Furies; the Weird Sisters wait for the soul of Macbeth. We cannot conceive of Oedipus without a Sphinx, nor of Hamlet without a Ghost. The shadows cast by the personages of Greek and Shakespearean drama lengthen into a greater darkness. And the entirety of the natural world is party to the action. The thunderclaps over the sacred wood at Colonus and the storms in King Lear are caused by more than weather. In tragedy, lightning is a messenger. But it can no longer be so once Benjamin Franklin (the incarnation of the new rational man) has flown a kite to it.30 But this, like Nietzsche’s story, is an oversimplification and a distortion of the reality. Consider Nietzsche first. In Zukunftsphilologie!, his polemical response to Nietzsche, Wilamowitz argued that the idea that Euripides is more optimistic than Aeschylus or Sophocles hardly bears examination of the evidence: compare the endings of the Hippolytus or the Bacchae with those of the Eumenides or the Philoctetes.31 I do not think that this is quite right either: the matter is more nuanced, as we shall see in a moment. But we can start by agreeing with Wilamowitz, against Nietzsche and Steiner, that it is wrong to say that either Aeschylus or Sophocles is irrationalistic: in §4 I rejected the common view that Sophocles’ Oedipus Rex is opposed to rationalism, and I have at many points so far in this study been implicitly generalizing that claim to the European tradition. Tragedy is a rationalistic genre; to that extent, and in that sense, it is inherently optimistic. Tragic catastrophe is precipitated by error, usually of a cognitive cast, and by free actions; but freely performed mistakes based on cognitive failure are corrigible.32 When Cresseid speaks at the end of Henryson’s Testament of Cresseid, ‘implicit in her warning to other women . . . is the belief that a reorientation of human character is possible, that the tragic figure potentially and paradoxically has the god-like power to help prevent tragedies’.33 Stephen Halliwell notes that the doctrine of hamartia offers ‘a means of making catastrophe intelligible’ (2006, 137), and faults it on this ground, namely that it yields, in effect, an infantilizing travesty of the incomprehensibility of Greek tragedy. But it is, rather, infantile to give up on the possibility of comprehension too quickly, when tragic literature in fact shows that much that befalls us is all too comprehensible, and open to influence and amelioration. This does not mean that cognitive failures will, in any given case, be corrected, and the genre of tragedy charts a range of scenarios where they are not. To that extent it is pessimistic: there will always be fresh

From Cognitive Failure to No-Fault Tragedy 139 mistakes, and there will always be failure to correct the old mistakes.34 But where mistakes are not corrected, we, the audience (and sometimes some of the characters), can survey the reasons why not: in the Oedipus Rex, for instance, Teiresias, and we the audience of the play, understand everything.35 Even if you wish to insist (against my argument of §§8 and 19) that Teiresias’ divinely inspired knowledge is not naturalizable, the fact is that we the audience come by the same knowledge without supernatural assistance, simply by putting available bits of information together; and, as I argued in Chapter 1, Oedipus could and should have done as much himself—or rather, should have done it sooner. But this is a typical case. Hence the springs of tragedy are comprehensible; they are also, as I shall explore in Part II, expressible in language.36 An ability both to understand the world even where we cannot change it—or (as usually in European tragedy) can change it but fail to do so—and to convey how things are in language gives us a certain measure of at least intellectual and often practical control as well over our sublunary lot, so that in supporting and promoting that control tragedy has a broadly optimistic nisus. (This is actually a Nietzschean thought: see §50 below.)37 Far from destroying tragedy, Socratic confidence in the comprehensibility of the world lies at the very base of tragic literature. Even Euripides’ Bacchae, which might seem at first glance to be a pessimistic and irrationalistic play, as Wilamowitz and many others have construed it, conforms to this pattern: we learn that we must reach an accommodation with our irrational urges, that they cannot simply be quelled by brute force, that there are depths in the psyche beyond a crude or superficial rationalism. These points are part of what the play insinuates, but they are themselves rationalistic thoughts. They involve a second-order rationality, for they concern the deployment, and necessary limitations, of first-order rationality. To understand these limitations, as the play helps us to do, is to be not less but more rational.38 The Bacchae is in fact a rationalistic drama. So Nietzsche turns out to have been right that a Socratic optimism underlay Euripidean tragedy, but wrong that it did not also underlie Aeschylean and Sophoclean tragedy, and indeed most tragedy written before his time; and wrong, too, that that optimism destroys tragedy—on the contrary, it constitutes its essence. Moving to Steiner’s case we need, as I have argued, to set against such assertions as those contained in the passage quoted at the beginning of this section the operatur principle, which governs so much of classical and classicizing tragedy. It is a consequence of the operatur principle that any view of tragedy such as Steiner’s—and Rowan Williams’s: ‘tragedy is a deeply unsecular matter’ (2016, 103)—that links it in any irreducible way to the sacred or the divine, is wrong. Insofar as the sacred and the divine feature in tragedy, they are a superficial crust concealing a molten core of human motivation. Steiner may be right that the Furies, the Witches, the Sphinx, and the Ghost are integral parts of the Oresteia,

140  Tragedy and Moral Redress Macbeth, Oedipus Rex, and Hamlet; but their role, as I have argued, is to externalize internal, psychic forces, motives, and character traits. It is true that the blood on Oedipus’ hands unleashes the plague in Thebes,39 true that the thunderclaps over the wood at Colonus occur there because that is where Oedipus is, and true that the tempest on the heath occurs there because that is where Lear is. We might add that it is the furor of Seneca’s Atreus which makes the sun vanish;40 that it is Roland’s impending death which brings about earthquakes and tempests (1428–9) De Seint Michel del Peril josqu’as Seinz, Des Besentun tresqu’as porz de Guitsand; that it is Schiller’s Johanna who by her presence draws down thunder and lightning on Rheims;41 that the storm at the climax of Kleist’s Penthesilea is induced by the heroine’s decision to meet and kill Achilles in single combat.42 Steiner is right that these happenings are not just meteorological events, but the explanatory slack is taken up by facts about human character and motivation: great tragic heroes and heroines make their own weather. Steiner’s point about lightning also oversimplifies: naturalistic explanations of lightning go back to antiquity, and were by no means the invention of the seventeenth century. One has only to recall Lucretius’ scepticism about the ‘divine messenger’ theory of lightning (VI, 96–101, 246–54, 400–5) in order to assure oneself of that. It is easy to fall into the fallacy of thinking of rationalism as a relatively recent phenomenon in human intellectual development. In fact many of the essentials of modern rationalism were already firmly embedded in ancient thought. The point applies not just to the way one regards lightning and other physical phenomena, but to psychology too. The ancient world from at least the late fifth century BCE onwards was familiar with the allegorizing or psychologizing strategy that I have argued should be applied to divine interventions in classical and classicizing literature. Electra in Euripides’ Orestes tells Orestes that he does not actually see the Furies, but only thinks he does, effectively disposing of Steiner’s assertion that ‘The reality of Orestes entails that of the Furies’; Cicero freely naturalizes the Furies as expressions of human guilt; Longinus remarks that Orestes only ‘sees’ the Furies because he is mad; Cicero, Seneca, and the author of the Octavia rationalize and naturalize Cupid, as does Wilmot too in his Tancred and Gismund.43 Rationalism in antiquity did not put a stop to tragic art, and it has not done so in the modern age, either. On the contrary: tragedy has burgeoned in modern times, having colonized new forms of representation, such as the novel and the film, as I mentioned in my Introduction. And indeed Steiner is forced to concede that La Nouvelle Héloïse and Die Leiden des jungen Werthers are tragic in just as good a sense as anything penned by Sophocles or Shakespeare (1961, 195). It follows that tragedy did not die at the beginning of the

From Cognitive Failure to No-Fault Tragedy 141 eighteenth century,44 and that Steiner’s seminal and (despite its manifold contradictions, to some of which I have drawn attention) brilliant work The Death of Tragedy was misnamed.

21 Absolute Tragedy So far in this book I have been concentrating on cases where tragic suffering arises in a comprehensible and explicable way from character and choice. This is essentially, as we have seen, the Aristotelian and Bradleian approach. For Aristotle and Bradley, tragic heroes are responsible for their fates; they are not mere passive victims.45 And that is indeed, as I have suggested, the rule in tragedy. In his discussion of Othello, Gervinus remarks that, as I have already noted, Iago exploits opportunities that others freely offer, rather than simply manipulating his victims as though they were clay in his hands: if Shakespeare had elected to depict Iago’s victims as merely passive, ‘he would have missed the first and highest aim of tragic literature, which should always make clear how the man himself is author of his own destiny’ (1872, vol. 2, 68). Edgar’s words in King Lear (V, 3, 168–9), The gods are just and of our pleasant vices Make instruments to plague us, perhaps as ‘emended’ by Niall Rudd, who suggests that ‘sins’ would make better sense than ‘vices’ (1976, 35)—except that, as we have seen, the orthodox view is that sins flow from vices—could stand as a motto of European tragedy, as could the lament of Gager’s faithless maidservant Melantho—‘sua quemque perdunt vitia moresque improbi’ (‘Each of us is destroyed by his own vices and evil ways’, tr. Sutton)46—or the lines with which Schiller concludes Die Braut von Messina (2837–9), Dies Eine fühl ich und erkenn es klar, Das Leben ist der Güter höchstes nicht, Der Übel größtes aber ist die Schuld. ‘This one thing I feel and know with clarity, that life is not the highest good, but guilt is the greatest evil’. Joshua Billings suggests that ‘Schiller’s conception of guilt (Schuld) internalizes the ancient concept of fate, making the individuals of the play morally responsible for the disastrous outcome’ (2014, 121). I have argued that the internalization of which Billings speaks was already operational in ancient tragedy, where there is no deep contrast between fate and guilt—not, at any rate, in respect of protagonists. But European tragic literature also recognizes that some suffering is undeserved and not compensated, at least not morally. Steiner calls this

142  Tragedy and Moral Redress ‘absolute tragedy’, or ‘tragedy pure and simple’. A few passages of his early work The Death of Tragedy suggest that he then thought all genuine tragedy to be of this kind (1961, 4–8, 127–9)—though as we have seen his position in that book on the involvement of moral responsibility in the hero’s downfall is inconsistent—but a quarter of a century later we find that he has considerably modified his position. In Antigones (1984, 280) we are told that absolute tragedy is ‘exceedingly rare’, and only ‘comprises a handful of Greek tragedies, Marlowe’s Faustus, Shakespeare’s Timon of Athens (there are ambiguities of compensation at the close of Lear), Racine’s theatre of Jansenist retribution’. And a few years later (in his 1990 and 1996), Steiner arrives at what is surely the correct view, namely that uncompensated, no-fault tragedy is merely a species of the genus, and that there is plenty of tragedy that is not, in his sense, ‘absolute’, or ‘pure and simple’. Still, as a species of the genus, absolute tragedy deserves notice: ‘Euripides’ Hecuba and Trojan Women (is there really a puzzle about Aristotle’s tragikōtatos?) come near to a degré zero of existential vision: an approach underlined by Sartre’s adaptations of Euripides, in times which were again those of systematic torture and massacre’ (1996, 538). Steiner now gives us a fuller and more explicit list of what he takes to be instances of the species: apart from the works already mentioned, he itemizes Sophocles’ Oedipus Rex and Antigone, Euripides’ Medea and Bacchae, Racine’s Bérénice and Phèdre, Shelley’s The Cenci and Artaud’s adaptation thereof, Büchner’s Woyzeck, Alban Berg’s Lulu, and some of Beckett.47 What should we make of this? At least if we are considering whole works of art, as Steiner is, and asking whether they count, taken in the round, as absolute tragedies, then I think we must strike a good few candidates from Steiner’s proposed list. Oedipus Rex and Antigone I have argued in detail are not, in Steiner’s sense, absolute tragedies: they contain not ambiguities but explicities of compensation—punishment and suffering redressed by fault. (Recall that I am treating redress and compensation as temporally neutral.) Nor, I have suggested (§8), is Marlowe’s Faustus an absolute tragedy. Shakespeare’s Timon of Athens is also, like the cases just mentioned, a tragedy of cognitive failure, in which the hero is punished for his errors. Timon ricochets misguidedly from excessive generosity and naïveté to excessive misanthropy, and is responsible for his own downfall; he is condemned for his lack of realism.48 It has been proverbial since Theognis that friendship is unreliable: when things take a downward turn, as Petronius put it (38, 13), ‘amici de medio’.49 The intense bitterness of Timon—and of Lear, which is why Steiner is tempted to include it in his list—does not as such make for absolute tragedy. In absolute tragedy, as it has been defined, there is no hamartia to redress the pain and make us think that it could have been avoided. Again, Racine’s Phèdre fully conforms to the standard model: all of its main figures—Hippolyte, Thésée, and Phèdre herself—make identifiable and serious mistakes (including moral

From Cognitive Failure to No-Fault Tragedy 143 mistakes): like Euripides’ Hippolytus (and like Stefonio’s Crispus, who is also based on Hippolytus),50 Hippolyte fails to defend himself when confronted with a false accusation; Thésée, like Euripides’ Theseus, is too precipitate; Phèdre fatally turns her judgement over to Oenone, as Euripides’ Phaedra does to her Nurse, and there is a good case for saying that she is a fantasist along the lines of Don Quixote or Madame Bovary.51 (Aricie is perhaps cognitively faultless,52 but she is not a significant enough figure to determine the overall nature of the play.) It is true that Racine’s Hippolyte is much more sympathetically portrayed than Euripides’ Hippolytus, whose abjuration of love makes him a remote and unattractive figure and compounds his fault,53 but he makes some of the same mistakes.54 Continuing with Steiner’s list, Bérénice ends sadly, but not even tragically simpliciter, let alone tragically in Steiner’s absolute sense. It is a play in the ‘catastrophe survived’ genre: as Richard Parish notes, when Bérénice’s last speech begins we expect Titus, Antiochus, and Bérénice herself all to die, but the speech obviates that catastrophic outcome.55 It is true that Racine himself argued that the absence of death at the end of the play did not derogate from its tragic status, asserting in his preface that the drama ‘breathes that majestic sadness (tristesse majestueuse) which is what makes the whole pleasure of tragedy’ (1982–3, vol. 1, 374), and some modern critics have agreed.56 T. S. Eliot famously said that the play ‘represents about the summit of civilization in tragedy’ (1964, 41–2): Bérénice is a civilized work, but Eliot added that ‘it is, in a way, a Christian tragedy, with devotion to the State substituted for devotion to divine law’, which (given my comments in the Introduction on the dubious status of Christian tragedy) ought to make us suspicious about its candidature for genuine tragedy. Many, including Voltaire, have held that Bérénice is not a tragedy but an elegy, and that seems to me the right view: the renunciatory ending, like that of Eugene Onegin, is depressed, not catastrophic.57 It is perhaps surprising that Steiner does not include Britannicus in his list of absolute tragedies, for this has often been regarded as Racine’s bleakest drama:58 but I shall suggest below (§24) that it does not yield an exception to the standard model. Euripides’ Bacchae can hardly feature in a list of absolute tragedies, for Pentheus—though a deeply sympathetic figure to any enlightened age—is clearly presented by the play as being at fault; his mistake, as I implied in the last section, is (like Oedipus’: §4) not that he is too rational, but that he is not rational enough. Similarly Medea’s candidacy must also be judged to fail. The Trojan Women and Hecuba are getting closer to being absolute tragedies, but in their cases, as in that of The Cenci, there are subtleties which, I think, act against attaching the label in any simple way: I shall expand on this point in §24. Neither Wedekind’s nor Berg’s Lulu seems to me a plausible candidate for absolute tragedy, nor does Büchner’s Woyzeck, or Berg’s operatic version thereof, seem a good fit:

144  Tragedy and Moral Redress in all these cases we have, in a sense, a ‘tragedy of circumstances’, so that Woyzeck, for example, is as much a victim as he is a transgressor, but that is true, in general terms, of all tragedy, and it does not, I suggest, detract from the freedom of protagonists. Woyzeck did not have to murder Marie, any more than Macbeth had to assassinate Duncan or Brutus Caesar. All tragic protagonists are placed in circumstances that make right action difficult and present wrong action, in one way or another, as extremely tempting: but that does not exculpate them; they have a choice.59 (Recall that taking this line does not preclude our showing understanding toward tragic protagonists. Responding sympathetically to a tragic hero does not require us to absolve him of responsibility for his fate. The agony of Woyzeck’s tragedy arises, one might say, precisely because he is free.)60 Beckett’s Waiting for Godot is not a tragedy but a tragicomedy (see §22 below); so indeed Beckett subtitled the English version.61 In any case, with Beckett we enter modernity, when we expect traditional models to lapse. Another period in which we expect traditional—at least, ­Aristotelian— rules to be put into abeyance is the Middle Ages,62 and accordingly it makes sense to ask whether there are medieval examples of absolute tragedy. Before Aristotle’s Poetics became widely available around the turn of the sixteenth century—Valla’s Latin translation was published in 1498 and the Aldine edition of the Greek text in 1508 (earlier versions by Herman Alemannus and William of Moerbeke seem to have had little impact)63—the main doctrinal sources on the nature of tragedy were few, scattered, and undeveloped: they included, in particular, Lady Philosophy’s remark in Boethius’ Consolatio that tragedy bewails ‘fortune overturning blessed kingdoms with indiscriminate stroke’.64 This definition has nothing to say about the question of fault, and insofar as there can be said to have been a medieval consensus on the nature of tragedy, it might seem that it focused not on any theory of hamartia but on the transition from good fortune to bad befalling men and women of high estate.65 For example, though many of the stories collected in Boccaccio’s De Casibus Virorum Illustrium and in John Lydgate’s English version of it, The Fall of Princes, as well as some of those featuring in Chaucer’s Monk’s Tale and in the various editions of the influential Mirror for Magistrates, clearly exemplify the moral that ‘There is no destinie, but is deseru’de’,66 some appear to involve no fault on the part of the hero or heroine, but rather to present innocent victims of bad fortune.67 I express this latter point tentatively, however, because the matter is unclear: fortune has a suspicious habit of picking on the rich and powerful, a tendency immortalized in Horace’s Licinius ode (II, 10). Now Horace implies in that ode that if fortune brings you down, it is your fault: you sailed too far into the deep, or hugged the shoreline too closely; in general, you lacked the bene praeparatum pectus. Fortune’s malice, in other words, is successful with the rich and powerful because they are disposed to sin.68 That is also

From Cognitive Failure to No-Fault Tragedy 145 implied in the fact that Boccaccio officially presented his book as a moral guide for the powerful, and so by implication as illustrating punishment for avoidable sin: a moral guide would not be of much use if there were nothing one could do to ward off fortune’s caprice. Boccaccio did not call his stories tragedies, but both Chaucer and Lydgate clearly regarded them as such, and that label is explicitly applied to the various narrative components of the Monk’s Tale, as well as to the considerably more substantial Troilus and Criseyde. Boccaccio’s and the Monk’s stories concern any kind of fall from prosperity, whether deserved or not. Many of these tales are too laconic and undeveloped to count as tragedies, in the sense of that word that I established in the Introduction. One might be tempted to deduce from some of the Monk’s stories that Chaucer is not concerned with tragedy in any sense which involves the precise tracing of cause and effect, but simply with the fickleness of fortune; but that, as I have suggested, would be to misconstrue the way in which fortune figures in these stories. The Monk’s Tale begins programmatically with three exemplary tales of error and punishment: Lucifer falls for pride, Adam for misgovernance, Samson for failing to keep his counsel.69 These figures all commit Aristotelian hamartiai. Given my account of hamartia in Chapter 1, I reject Helen Cooper’s view (2010, 147) that that doctrine and the medieval conception of tragedy as involving ‘just retribution for . . . sins’ are ‘substantially different’. On the contrary, they are substantially the same: here we may note that ‘hamartia’ and cognates are regularly used for ‘sin’ in the New Testament.70 Troilus and Criseyde certainly does not suffer from the defect that it lacks scale. Does it fit the Aristotelo-Bradleian model? Henry Kelly suggests that, though Criseyde is at fault, Troilus is not.71 But other critics have located errors and flaws in Troilus’ conduct and character, and it seems to me that they are right: in particular, I would identify Troilus’ key mistakes as his failure to support Hector’s position of defiance—had he done so, the Trojans might have kept Criseyde—and his failure push through his plan to flee Troy with Criseyde; when she raises objections, he caves in too readily. These mistakes are grounded in his character: he is timid.72 (Once again I stress the point that there is no conflict between the reader’s holding Troilus responsible for his fate and feeling sympathy for him.)73 So I think that Chaucer, in spite of his unfamiliarity with Aristotle’s Poetics and its doctrine of hamartia, on the one hand, and his familiarity with a Boethian conception of tragedy that permitted it to treat of undeserved suffering, on the other, gravitated towards an Aristotelian approach in his only substantial tragedy. Kelly’s contrary view is partly based on his (surely questionable) judgement that tragedy in the wider Boethian sense is superior to tragedy in the narrower Aristotelian sense, and that we nowadays prefer it.74 Kelly also thinks that the non-Aristotelian approach to tragedy that he finds in Chaucer was the one that informed Elizabethan and Jacobean

146  Tragedy and Moral Redress tragedy (1997, 141, 261); but I have rejected that view, in agreement with Bradley.75 Kelly follows the same line in his discussion of several further medieval tragedies. For example, he implies that the protagonist of Loschi’s Achilles is an innocent victim (1993, 187). This play is certainly Boethian rather than Aristotelian in its pursuit of the theme ‘o fortuna, velut luna’, but it nevertheless conforms to what I am calling the standard (Aristotelian) model, inasmuch as Achilles exemplifies the usual cognitive failure of the tragic hero: his infatuation with Polyxena blinds him to the risks of going to meet her unaccompanied and virtually unarmed; that insouciance leads to his downfall.76 Kelly offers similar interpretations of Zacchia’s De Captivitate Ducis Jacobi and Fabriano’s De Casu Cesene (1993, 193–6). But these pieces are hardly substantial enough to count as literary tragedies (though the latter might count as a tragedy in the extended sense of the term). In the former, Jacobus is portrayed as an innocent victim, but he is a very minor figure, and the play really has no protagonist, nor indeed any plot or characters to speak of. Reflection on this case, and on other possible medieval tragic heroes, suggests that while innocent suffering is found in medieval tragedy, as one would expect given the Boethian starting point, insofar as the sufferer is made the protagonist of a substantial tragedy, he or she tends to conform to the standard model, in which punishment tracks desert. Apart from Chaucer’s Troilus, the Roland of La Chanson de Roland is a good case in point: his arrogant contempt of Ganelon precipitates the tragedy, and his failure to heed Oliver’s advice and call help in his moment of need is clearly marked as an error. In this he resembles Homer’s tragic Hector, who ignores Poulydamas’ advice and by his folly destroys his people.77 So my list of absolute tragedies would be a good deal shorter (and more tentative) than Steiner’s. But at this point we must introduce a complication. Steiner’s list is an attempt to identify works of literature (or, more generally, works of art) that, as we have said, count as absolute tragedies when considered as wholes. But if we are talking about the depiction of undeserved suffering, we should consider adopting a more nuanced approach than this. When we judge that tragedies like Medea and Macbeth do not function as absolute tragedies in toto, we are thinking of their protagonists, and of the fact that Medea, Jason, and the Macbeths conform to the standard model according to which punishment tracks cognitive failure. But what should we say about Medea’s children, or Lady Macduff’s children? Their suffering surely is undeserved. If we are to acknowledge this, while continuing to operate with the category of absolute tragedy, we should allow this category to be applied not merely to works of literature (in general: works of art) taken as wholes, but also to parts or aspects of such works. And if we follow up on this latter idea, we may find that absolute tragedy is much commoner than Steiner makes out, even though, as I have suggested, the list of works that function as absolute tragedies when regarded as wholes is significantly shorter than

From Cognitive Failure to No-Fault Tragedy 147 Steiner’s already quite short inventory. I shall consider in more detail below in this chapter (§24) how we can apply the category of absolute tragedy to smaller units than works considered as wholes.

22 Tragedy and Tragicomedy Steiner’s early impulse to say that only absolute tragedy counts as genuine tragedy is matched by a related tendency to demote what usually passes for tragedy to the status of tragicomedy. Thus he argues that Shakespeare ‘is not, or is only very exceptionally, a “tragedian” ’, his reason for this statement being—here he repeats the gist of a famous passage in Samuel Johnson—that Shakespeare balances hope and despair in an essentially tragicomic way: Shakespeare knows, in every fibre of his compendious being, that a child is being born next door, a birthday celebrated below stairs, in the very instant of the murder of Agamemnon or the blinding of Oedipus. He knows . . . that the facts of the world are hybrid, that desolation and joy, destruction and generation are simultaneous . . . Fortinbras will be a sounder king than Hamlet could ever have been. Scotland will flower after the slaying of Macbeth. Cassio will govern justly in Cyprus.78 Herder thought Shakespeare’s plays ‘History im weitsten Verstande’, and A. W. Schlegel held that the plays of both Shakespeare and Calderón were in reality neither tragedies nor comedies but romantic dramas.79 Coleridge agreed, and in fact the view that Shakespeare’s plays (indeed the products of the English stage in general) are neither comedies nor tragedies can be traced back to Sir Philip Sidney and Elizabeth I’s Italian tutor John Florio.80 Steiner is right that the facts of the world are hybrid, and that Shakespeare’s dramas convey this powerfully. Shakespeare is by no means unique here. Recall the treacherous murder of Siegfried,81 or the polonaise that follows Lensky’s death in Tchaikovsky’s Eugene Onegin. Homer and the Greek tragedians, especially Aeschylus and Euripides, know that ‘desolation and joy, destruction and generation are simultaneous’.82 The third stasimon of the latter’s Suppliants, when the mothers see the return, at last, for burial of their sons’ bodies, is a supreme bittersweet moment.83 Or, in terrible exemplification of the simultaneity of destruction and generation, think how Aeschylus’ Clytemnestra exults in the murder of Agamemnon, who ‘breathing out a sharp slaughter of blood, strikes me with a dark shower of bloody dew, while I rejoiced no less than the crop rejoices in the rich blessing of the rain of Zeus when the sheath is in labour with the ear’.84 Eduard Fraenkel notes (ad loc.) that Homer lies in the background of this Aeschylean passage. In one of

148  Tragedy and Moral Redress his finest similes, the poet of the Iliad describes evening descending on the field of battle: As long as it was early morning and the sacred daylight increased, so long the thrown weapons of both sides took hold and men [dropped under them. But at that time when the forester makes ready his supper in the wooded glens of the mountains, when his arms and hands [have grown weary from cutting down the tall trees, and his heart has had enough of it, and longing for food and for sweet wine takes hold of his senses; at that time the Danaans by their manhood broke the battalions calling across the ranks to each other.85 Our attention is diverted from the mêlée before Troy to a lonely mountain scene, where a woodman, spent from the day’s labour, prepares his supper. We are taken from war to peace, from the heat and dust of battle to an alpine coolth and clarity of air, and the auditor or reader experiences a palpable sense of relief. We expect, in line with what we are often told is the purpose of Homeric similes, ‘a self-contained reflection from the world without’.86 But here the relevant domains are not so neatly compartmentalized. For the word used by Homer and translated above as ‘makes ready (his supper)’ is ‘hōplissato’, from the verb ‘hoplizein’, which commonly means to arm for combat: there is a particularly harrowing use of the verb by Euripides’ Medea—‘but come now, arm thyself, my heart’ (1242: ἀλλ᾽ εἶ᾽ ὁπλίζου, καρδία)—as she steels herself to kill her own children.87 We cannot help entertaining this meaning of the verb in the Homeric passage, so poignantly does the simile of the woodman return us, before we expected it, to the arena of conflict: in the midst of war there is peace; in the midst of peace there is war.88 Homer, Aeschylus, Euripides, Shakespeare—these exemplify the Johnsonian point in a way that perhaps few other artists have achieved (though it is tantalizing to speculate on what Büchner might have done had he lived);89 but we cannot conclude from this that they were not tragedians, or only wrote tragicomedy. In making the suggestion that Homer and Shakespeare deal in tragicomedy rather than tragedy Steiner is following not only Johnson but also Aldous Huxley, who, in his 1931 essay ‘Tragedy and the Whole Truth’, argued that the poet of the Odyssey and the author of Tom Jones, though scarcely any other artists (and not the poet of the Iliad), give us what he calls the ‘whole truth’, in the sense that they cover the entire gamut of human experience. The first words of Huxley’s essay remind us of a famous scene in the Odyssey (1938, 3): There were six of them, the best and bravest of the hero’s companions. Turning back from his post in the bows, Odysseus was in time

From Cognitive Failure to No-Fault Tragedy 149 to see them lifted, struggling, into the air, to hear their screams, the desperate repetition of his own name. The survivors could only look on, helplessly, while Scylla ‘at the mouth of her cave devoured them, still screaming, still stretching out their hands to me in the frightful struggle’. And Odysseus adds that it was the most dreadful and lamentable sight he ever saw in all his ‘explorings of the passes of the sea’. But shortly after this ordeal Odysseus and his men are to be found beaching their ship and ‘expertly’ making their evening meal. The ‘whole truth’ that Homer gives us thus includes the fact that, even after enduring such a loss, Odysseus and his men had to eat, so had to prepare a meal, which naturally they did with their usual skill. And then they had to sleep. Grieving has an end, and normal life resumes. Genuine tragedy, by contrast, such as Racine’s Phèdre, gives us, in Huxley’s view, ‘chemically pure’ suffering, isolating a few terrible elements from the totality of our experience, and so to that extent distorting the reality of our lives. As for Shakespeare, Huxley is unsure. ‘The tragedies of Shakespeare are veined, it is true, with irony and an often terrifying cynicism’ (ibid., 10). But he thinks that, even so, everything subserves an ultimately tragic vision (ibid., 11): Shakespeare’s ironies and cynicisms serve to deepen his tragic world, but not to widen it. If they had widened it, as the Homeric irrelevancies widened out the universe of the Odyssey—why, then, the world of Shakespearean tragedy would automatically have ceased to exist. For example, a scene showing the bereaved Macduff eating his supper, growing melancholy, over the whisky, with thoughts of his murdered wife and children, and then, with lashes still wet, dropping off to sleep, would be true enough to life; but it would not be true to tragic art. But it is clear by now that something has gone wrong, and we can assure ourselves of that by remembering that the thought that life must continue through suffering, that even after the death of those close to them the survivors must eat, is central to one of the profoundest books of the Iliad, namely the last, where, in the extraordinary confrontation between Achilles and Priam, Hector’s slayer expresses precisely that thought to Hector’s father, after he has granted the latter’s request to have his son’s body returned to him: Your son is given back to you, aged sir, as you asked it, and he lies on a bier. When dawn shows you yourself will see him as you take him away. Now let us remember our supper. For even Niobe, she of the lovely tresses, remembered

150  Tragedy and Moral Redress to eat, though her twelve children were killed in her palace, six daughters, and six sons in the pride of their youth, whom Apollo slew with arrows from his silver bow, angry with Niobe, and shaft-showering Artemis slew the daughters, because Niobe had likened herself to Leto of the fair complexion, saying that Leto had borne only two, while she herself had borne many. But these two, though only two, destroyed all those others. Nine days long they lay in their blood, nor was there anyone to bury them, for the son of Cronos made stones of the people. But on the tenth day the Uranian gods buried them indeed. But she remembered to eat when she was worn out with weeping.90 Then, after they had wept together, and eaten, Homer’s Achilles and Priam took their rest.91 So, too, in Matthew Arnold’s ‘Sohrab and Rustum’, after Rustum has slain Sohrab ‘Both armies mov’d to camp, and took their meal’ (871). The point is this. Contrary to Huxley’s insinuation, tragedy does not exclude the idea, or the representation, of eating and sleeping after grief. In fact I believe we should go so far as to say that it does not exclude any subject matter. Shakespeare (or Middleton) could have put a scene into Macbeth along the lines of Huxley’s suggestion, without undermining that play’s status as a tragedy. Eating and drinking—even downing a dram of whisky—after grief are not inherently either comic or tragic ideas; they are handled comically or tragically. When Ajax awakes from his madness he sits among the slaughtered cattle asitos and apotos, as Tecmessa tells the chorus, without food and drink.92 Falling asleep after grief is quite often mentioned or displayed in serious contexts without upsetting the seriousness of those contexts: Lear is a case in point.93 The very play that Huxley mentions, Macbeth, shows that the limits of tragedy are set wide, for the inclusion in that play of the comic porter scene does not derogate from its status as a tragedy,94 or turn it into a tragicomedy in the sense in which Marcellino Verardi’s Fernandus Servatus, perhaps the first genuine tragicomedy to be explicitly labelled as such,95 is one, or in the sense in which Euripides’ Cyclops, Alcestis, Iphigeneia in Tauris, Ion, and Helen, Shakespeare’s so-called problem plays (Troilus and Cressida, Measure for Measure, All’s Well That Ends Well) and late romances (Pericles, The Winter’s Tale, Cymbeline, The Tempest), and his collaboration with Fletcher, The Two Noble Kinsmen, Fletcher’s The Faithful Shepherdess (and its original, Guarini’s Il Pastor Fido), The Island Princess, and some of his collaborations with Beaumont such as Philaster and A King and No King, perhaps Jonson’s Volpone, many of the products of the Spanish school (Cervantes, Calderón, Lope de Vega), and Corneille’s Le Cid and Cinna, can reasonably be classified as tragicomedies.96 In a discussion of Don Quixote, Charles Lamb argued that those who laugh at the Don are ‘mistaking his author’s purport, which was—tears’

From Cognitive Failure to No-Fault Tragedy 151 (1903, vol. 2, 233). But we are meant to laugh and cry. Tragedy and comedy are, as I have said, a matter of how the material is treated, not what it is. A man performing a somersault might be a professional acrobat entertaining an audience, or he might be a mortally wounded warrior tumbling headlong from his chariot—‘to the spectator these can look much the same’97—like the Trojan Cebriones, who then ‘in the turning dust lay / mightily in his might, his horsemanship all forgotten’,98 a tragic end that connects him with Achilles, of whom some of the same words are used, while his body is fought over by Hector and Patroclus, the latter having mocked him for the manner of his fall. Ancient Greek critics regarded tragedy and comedy as utterly distinct genres, but Socrates is reported in the Symposium to have argued (though we do not learn the details of his argument, because Aristodemus was too drowsy at the time to remember them) that one and the same author can write both (222d3–6), and he was vindicated by later practice, from the Roman period onwards.99 Sean O’Casey called Juno and the Paycock a tragedy, but if we bear in mind the Foster definition of tragicomedy quoted in the Introduction, according to which ‘a tragicomedy is a play in which the tragic and comic elements both exist but are formally and emotionally dependent on one another, each modifying and determining the nature of the other so as to produce a mixed, tragicomic response in the audience’, we might judge the Boyle–Joxer comic subplot sufficiently integral to the main plot to make the play a tragicomedy. The issue in each case will turn on the extent to which comic elements subserve a basically tragic response in the audience or contribute to the production of a distinct, tragicomic, response. Dürrenmatt labels his Der Besuch der alten Dame a tragicomedy, but I respond to it as I do to Macbeth, as a tragedy with comedic elements. Fielding, mentioned by Huxley, gives us the whole gamut of human experience, and he does so in an essentially comic way; Shakespeare also gives us the whole gamut of human experience, and he does so sometimes in a tragic way, sometimes in a comic way, sometimes in a mixture of the two. But the fact that he has this range, both of subject matter and of style, does not mean that if we take any given play, such as Othello or The Merry Wives of Windsor, that play must really be a tragicomedy, and cannot be determinately a tragedy, or determinately a comedy. I pick those two plays purposely: I pointed out in the Introduction that Kleist’s Der Zerbrochne Krug—whose Judge Adam is a self-exposing criminal with a club foot—is a comic reflection of the Oedipus Rex,100 and we are familiar, too, with the idea that Shakespeare’s Ford, who comes close to spoiling the party in The Merry Wives of Windsor, is the obverse side of a coin whose reverse is the tragic hero of Othello.101 Romeo and Juliet have their comic doublet in Pyramus and Thisbe; Jaques anticipates Hamlet; Gloucester’s mock death is itself mocked by the mock deaths of Bottom and Falstaff; Lear’s fivefold repetition of ‘never’ is parodied by Bottom’s

152  Tragedy and Moral Redress fivefold repetition of ‘die’.102 Odysseus’ confrontation with the repentant Neoptolemus in Sophocles’ Philoctetes (1254–8) has tinges of the comic, as does his sudden entrance (1293) and forced exit when Philoctetes threatens him with Heracles’ bow (1302):103 Philoctetes is what later ages would call a tragicomedy, but that is because of its happy ending, not these quasi-comic scenes. Again, in Lessing’s limpid and joyous Minna von Barnhelm, the sun passes briefly behind a cloud when the play’s eponymous heroine accuses her lover, the jealous Major von Tellheim, of having rather too much in common with the notorious Moor of Venice.104 In Plautus’ Amphitruo, the scene in which Amphitryon accuses Alcmena of adultery (II, 2) would be deadly serious and painful were it not undercut by clear comic signals, and by our knowledge, based on generic expectations and specifically on what Mercury tells us (at 474–5), that Jupiter will smooth over everything in the end.105 But such shades and undertones and allusions do not transpose the genres of their respective plays. It needs to be stressed, given the current popularity of the idea of permeation between generic boundaries, that such permeation is only possible if there are boundaries: generic transgression and interference do not undermine the existence of those boundaries, but precisely presuppose them.106 And, as I mentioned in the Introduction, the existence of borderline cases presupposes the existence of cases that are not borderline. Tragedies may contain unintended comic elements. Emilia Galotti, we said, has the underlying form of a comedy: the Emperor Joseph II liked it, but remarked, after seeing a performance, that he had never in his life laughed so much at a tragedy.107 Although we must beware of finding comedy in Euripides anachronistically, it seems fair to say that he specializes in the inclusion of comic cross-currents in tragic or at least serious settings: the recognition scene in the Electra has a comic undertow, as do several moments in the Orestes—the banter between Orestes and the Phrygian, Apollo’s command to Orestes to marry the woman at whose throat he is holding a sword.108 If, as I said above, the work in question really is a tragedy rather than a tragicomedy, these elements will subserve a fundamentally tragic purpose. In Children of Heracles, the humorous arming scene ‘ultimately enhances rather than subverts the tragic tone of the play’.109 The terrible moment in the Bacchae when Dionysus makes Pentheus don the attire of a maenad is, as E. R. Dodds notes, ‘as gruesome as anything in literature’ (1960, 192). The scene is a hair’s breadth away from the curative exuberance of Aristophanic cross-dressing, and in the same play the togging up as maenads of the two old men, Teiresias and Cadmus, actually does traverse the boundary into comedy.110 But these elements of farce serve to deepen the tragedy, in the manner of Iago’s black humour, or the comic elements of Titus Andronicus,111 or the soldiers’ banter in the York Crucifixion,112 rather than undermine it.113 Compare too the glorious tenor aria (‘Di rigori armato il seno’) in the first act of Der Rosenkavalier, sung at the Marschallin’s levée. Strauss

From Cognitive Failure to No-Fault Tragedy 153 is parodying Italian opera, but the words anticipate Octavian’s meeting with Sophie, and so gesture towards the high pathos of the close of Act 1, when the Marschallin knows that Octavian will desert her for a younger woman. Elsewhere in Euripides, of course, we find a nearer approach to genuine tragicomedy, as I have said.114 But in general it would be a mistake to say that the genre which commonly goes under the name ‘tragedy’ divides without remainder into absolute tragedy, in Steiner’s sense, on the one hand, and tragicomedy on the other, leaving nothing in between to count as tragedy simpliciter: that would not be a sensible terminological policy. We need a category for the tragic in something like Aristotle’s sense in chapter 13 of the Poetics. Macbeth, for example, is a genuine tragedy. It is not a case of absolute tragedy—at least, not when viewed as a whole—because the punishment of the protagonists is balanced by desert, and it is not a tragicomedy, because, among other things, it ends badly for its protagonists and because the porter scene, as we have said, is not significant enough to affect the overall genre of the drama. If we assimilate plays like Macbeth to genuine tragicomedies like The Two Noble Kinsmen, we blur an important distinction. Steiner’s second assertion in the passage quoted at the beginning of this section is that Shakespeare’s tragedies often end on a note of hope. This is obviously a distinct point from the observation that Shakespeare has a panoramic vision of life that few artists have achieved. In itself, and considered rather narrowly, it is no doubt correct; for it is plausible that Fortinbras will be a better king than Hamlet would have been—ironically enough, in view of Fortinbras’ statement of Hamlet that ‘he was likely, had he been put on, / To have proved most royally’ (V, 2, 351–2)—and so on for Steiner’s other examples.115 But not only is this a different thesis from Steiner’s first one, the one about tragicomedy; the two points actually work against one another. Even if the tragedies which Steiner mentions end optimistically, it is an implication of his first, Johnsonian claim about Shakespeare that that optimism can be no better than partial and temporary. The palpable relief felt at the end of Macbeth is tempered both by a sense that good and bad exercises of violence are, to some extent, subject to an effect of moral flattening, and by a sense that the restoration of peace will in any case be succeeded, as it has been preceded, by troubles: for one thing, as many commentators have noted, it is Malcolm, not Fleance, who is crowned at the end, but we know that Banquo’s descendants must in due course occupy the throne.116 Growth and decay alternate in an unending cycle. War and peace are the night and day of the human world. Hope is balanced by despair elsewhere and elsewhen. We are condemned to an endless interchange of state. Buchanan’s Jephtha has won a great victory, but at a terrible cost: tristia secundis et secunda tristibus vicissitudo acerba sortis temperat.

154  Tragedy and Moral Redress ‘A bitter alternation of fate tempers sadnesses with successes and successes with sadnesses’.117 In any case, even if we think, against the trend of recent criticism, that Macbeth ends optimistically, it would seem faintly absurd to say the same of Hamlet.118 And, one might ask, in what sense does, or should, the future trump the past? The writings of, say, Jean Améry and Primo Levi are inspirational, but would anyone think that the price these men paid in suffering for their literary achievement was therefore worth paying? That takes me to a theme of my next section.

23 Tragedy and Secular Remedy In discussing Ibsen’s Pillars of Society, A Doll’s House, Ghosts, and An Enemy of the People in his The Death of Tragedy—at which stage in the development of his thought he was inclined to insist that genuine tragedy must be ‘absolute’—Steiner remarks (1961, 291) that these tracts, enduring as they may prove to be by virtue of their theatrical vigour, are not tragedies. In tragedy, there are no temporal remedies. The point cannot be stressed too often. Tragedy speaks not of secular dilemmas which may be resolved by rational innovation, but of the unaltering bias toward inhumanity and destruction in the drift of the world. But in these plays of Ibsen’s radical period, such is not the issue. There are specific remedies to the disasters which befall the characters, and it is Ibsen’s purpose to make us see these remedies and bring them about. A Doll’s House and Ghosts are founded on the belief that society can move toward a sane, adult conception of sexual life and that woman can and must be raised to the dignity of man. Pillars of Society and An Enemy of the People are denunciations of the hypocrisies and oppressions concealed behind the mask of middle-class gentility. They tell us of the way in which money interests poison the springs of emotional life and intellectual integrity. They cry out for explicit radicalism and reform. And earlier in the book Steiner had made the same point: More pliant divorce laws could not alter the fate of Agamemnon; social psychiatry is no answer to Oedipus. But saner economic relations or better plumbing can resolve some of the grave crises in the dramas of Ibsen. The distinction should be borne sharply in mind. Tragedy is irreparable.119 In one obvious respect, the argument of this book so far has been that Steiner is wrong about this. Tragedy, I have contended, is usually precipitated by cognitive failure, and there are temporal remedies to cognitive failures—often quite simple, if not easy, such remedies: think; find out

From Cognitive Failure to No-Fault Tragedy 155 the relevant facts; use the facts intelligently when you have found them; work out what you really want, while remembering that what you want is not always in your real interests; work out what other people are up to; assess their real motives; be alert about human nature and reality in general; and so on. Further, given that one’s perceptions are affected by desire, and that in tragedy much error stems from the clouding of the agent’s perception (including moral perception) by desire, a remedy to tragedy that is often implied is that the agent needs to work at the second-order level on his first-order desires. That is not to say, as I noted above, that any of this is easy. As Heraclitus said: ‘it is hard to fight against the heart’s desire; it pays for what it wants with soul’ (DK B85). Tragedy arises because the in-principle solutions that it suggests are difficult for particular agents (and we are all particular agents) to convert. But the common view that tragedy makes us ‘confront problems that typically have no solutions’120 is, I have maintained, mistaken: on the contrary, typically there are solutions. Tragic heroes deliberate badly; but it is presupposed in that very statement that they could have deliberated well.121 All this is, as I have said (§§4, 20), an important aspect of tragedy’s profound rationalism.122 So tragedy does offer—or at least point in the direction of—secular remedy; tragic works are full of such remedies, expressed or implied. And secular remedy is a form of compensation, though different from the moral compensation—the balancing of punishment by desert—on which I have been focusing so far. But if, in one sense, Steiner is wrong to deny secular remedy to tragedy, there is a deeper sense in which he is gesturing towards an insight. We can see this by reflecting on, and indeed dissenting from, his remarks about Ibsen. The problems that precipitate tragedy in, for example, An Enemy of the People can, in one sense, be solved by sensible economic relations and good plumbing, but it would be an inattentive audience that did not discern that these problems mask deeper conflicts between human beings that are not to be resolved by merely laying down fresh water pipes. (To that extent a Marxist approach to tragedy,123 according to which the genre is concerned with remediable social ills, seems to me off target. Remediable social ills do play a role in some tragedies, but that is not where the main emphasis lies; as I have argued, it lies rather on remediable individual error.) There are profound tensions between people which will out in one way or another, the ways in which the tensions get expressed being often mere pretexts, like the trivial episode that sparks the full-scale domestic row. And how easy is it, really, to move towards a ‘sane’ conception of sexual life? That is a deep problem, one which we are a long way from solving, and which has exercised other writers than Ibsen, including Shakespeare and Fletcher in their tragicomedies,124 Kleist in Penthesilea, Hebbel in Judith, and Strindberg in Miss Julie. So I think Steiner’s contrast between a tragic Aeschylus and Sophocles on the one hand, and an untragic Ibsen on the

156  Tragedy and Moral Redress other, is unfair to Ibsen, given the concomitant insinuation of relative superficiality.125 But that correction actually strengthens Steiner’s main claim, namely that in genuine tragedy there is no secular remedy. Putting it differently, and factoring in our revaluation of Ibsen, we might say that, in Ibsen as much as in Aeschylus or Sophocles, secular remedy in tragedy is at best only part of the point. (Or rather, more simply: remedy is only part of the point. I am not going to suggest that the slack should be taken up by religion.) The solution offered at the end of the Eumenides, say, does not cancel the loss and despair of the preceding tragedies. For one thing, it may be that the presentation of that legal solution contains parodic and undercutting elements, precisely in order to hint at the shallow nature of the solution.126 But even if we take the ending of the Eumenides at optimistic face value,127 the point remains that the remedies that tragedies explicitly or implicitly offer do not and cannot cancel the loss and the suffering undergone, which were there, were real, and matter abstractly as much as the act of healing, if present.128 And the same reservation affects the rationalism and optimism that I have argued are implied in the genre of tragedy generally. Cognitive failure is indeed corrigible, but against this we must set not only the points that it is often difficult to correct and often not corrected, but also the fact that ameliorative measures do not wipe out the tragedies that it has caused. We live in one temporal direction only, and so are inclined to think that a situation that moves from bad to good in the apparent direction of travel is better than one that moves from good to bad. We do not think that a situation that moves from bad to good is no better than one that moves from good to bad on the grounds that the one is just the temporal mirror-image of the other, so that, from a timeless, god’s-eye perspective, there is nothing to choose between them. But if that is how we function in our daily lives, tragedy helps us to a more abstract and objective perspective, viewed from which these situations are indeed seen to be symmetrical. From that more abstract perspective there is no real sense in which the future is more important than the past. Wittgenstein once remarked (1984, 455): If someone believes that he has found the solution to the problem of life, and would like to say to himself that everything is easy from now on, he only needs to remember, in order to refute himself, that there was once a time when this ‘solution’ had not been found. But at that time, too, it must have been possible for a human being to live, and in respect of that time the solution that has been found looks like an accident. (The remark was made in 1930: at this stage in his thinking Wittgenstein still adhered to his Tractarian view that nothing accidental can belong

From Cognitive Failure to No-Fault Tragedy 157 either to the deep structure of the world or to our engagement with it.) So: regarded from a sufficiently lofty standpoint, the hopeful mood at the end of Macbeth (if that is how we are construing the play) does not outweigh or compensate for the bleak reign of terror that it supersedes. The fact that the one comes after the other is of no ultimate importance: if the one comes after t’other, then t’other comes before the one—so what? But if hope is balanced by despair, it follows by the symmetry of the relation that despair is balanced by hope. Just as there is no good reason to give undue weight to the messages of hope with which some tragedies conclude, so also there is no good reason to focus on the desolation at the expense of the joy. You might say, embellishing Steiner’s point, that from the perspective of tragedy, it is not good enough for a remedy to work in one temporal direction only—from past to future, say. The only remedy that tragedy could accept would have to work in both directions. But that makes no logical sense; a remedy necessarily works in just one temporal direction. A ‘remedy’ that worked in both directions would not be a remedy. Though its components were not written together, or in narrative order, Sophocles’ so-called Theban plays well illustrate the balance of hope and despair of which I have been speaking. Oedipus Rex ends bleakly, but already there are signs of the hero’s future strength, so poignantly explored in Oedipus at Colonus. When Jocasta, having failed to suppress the truth, rushes away to commit suicide, Oedipus speculates, absurdly, that she fears he will be revealed to be of low birth. The prospect does not worry him: ‘I, who hold myself son of Fortune that gives good, will not be dishonoured. She is the mother from whom I spring; and the months, my kinsmen, have marked me sometimes lowly, sometimes great’ (1080–3; tr. Jebb). In an obvious dramatic irony, Oedipus stands revealed as cognitively challenged yet again, as he continues to misidentify his mother. But if we bear Sophocles’ later Oedipus at Colonus in mind, we may find here a further, deeper irony, reversing the direction of the first irony. For in that later play, in which Oedipus is transfigured into something like a cosmic force, it will turn out to be true, in a sense, that he is a child of fortune and that the months are his kinsmen. Indeed at the very beginning of that play Oedipus speaks not of his old age, but of ‘χὠ χρόνος ξυνὼν μακρός’ (7–8), ‘the years in our long fellowship’, as Jebb translates, though the sense is bolder than that, closer to ‘long time merging with me’.129 But if Oedipus at Colonus ends well (in some sense) for Oedipus himself, it ends ominously for his children, as the conditions for the next stage of the saga, the plot of the Antigone, are set out. In view of these points, there is considerable plausibility in F. L. Lucas’s observation that it hardly matters whether a tragedy ends well or badly: ‘If there is an unhappy ending, we may call it a tragedy; but if the play is a serious attempt to represent life, it makes no great difference whether or not good fortune intervenes in the last scene’ (1957, 33). In

158  Tragedy and Moral Redress the Poetics, Aristotle clearly envisaged both sorts of ending, and not just in Chapter 14, either, where happy endings are explicitly preferred to the sad endings presupposed in Chapter 13.130 Perhaps, then, there is no deep difference between tragedy and tragicomedy. There might also be a sense in which it is quite sufficient for the purposes of tragedy on stage to depict horror in prospect, in retrospect, or merely counterfactually: it is not necessary to go as far as displaying horror in actuality and in the here and now. After all, horrifying events do happen, we know, and if we escape them today we will incur them tomorrow, or incurred them yesterday; or if we escape them in the actual world, there are other possible worlds where we incur them. It might be thought enough to remind the spectators of the contingency of these things, without sending them away in a passion of tears. (With Wittgenstein’s remark and my gloss on it in mind, one might suggest that tragedy goes deeper than mere contingency.) Perhaps Ibsen went far enough in Pillars of Society: was it necessary for Miller, taking the same idea, to ratchet it up to the terrible denouement of All My Sons? It must be of significance that some of the greatest tragedians in our tradition end their works with the avoidance of catastrophe as often as with its incidence—Aeschylus, Euripides, Shakespeare, Goethe, Kleist, Ibsen. The counterfactual is but a heartbeat away from the actual. As Stanley Cavell remarks of King Lear, ‘Edgar’s is the most Christian sensibility of the play, as Edmund’s is the most Machiavellian. If the Machiavellian fails in the end, he very nearly succeeds; and if the Christian succeeds, his success is deeply compromised’.131 Tragedians sometimes impress this point on us by starkly juxtaposing the expression of hope with its immediate frustration, as we mentioned in connection with Jocasta’s prayer (§9), or as when the chorus of Sophocles’ Ajax rejoices at the hero’s apparent change of mind, only to learn soon after of its error, or the Hecuba of Euripides’ Trojan Women voices her hopes for Astyanax and his progeny just before Talthybius enters to announce the boy’s imminent death.132 So the fact that tragedies sometimes end on a note of hope does not undercut the despair that we have travelled through, or glimpsed off stage, or sensed in another possible world, in order to get to that hopeful end, and works that manage to avoid catastrophe hint at its non-avoidance in other times, places, and worlds. In that sense one might say that tragedy has a nisus to temporal and modal realism: it moves towards supposing that other times and other possible worlds exist in just as good a sense as the present time and the actual world. As a matter of taxonomy, I have confined the term ‘tragedy’ to works that end badly for the protagonists (Introduction)—that is, end badly now (at the end of the work), and in the actual world (the world of the work). So for taxonomic convenience I have simply built into the definition of tragedy both the priority of the present (over the past and future) and the priority of the actual (over the counterfactual). That seems to

From Cognitive Failure to No-Fault Tragedy 159 me, for reasons that would take us too far afield to investigate here, to reflect an inherent metaphysical bias in our ordinary thought in favour of the present and the actual.133 However, once that species of tragedy has been taken as definitionally basic, we can then acknowledge that, in a derivative sense, there is such a thing as past, future, and counterfactual tragedy. We can speak of primary and secondary tragedy. And, following that split-level approach (reminiscent of, but not the same as, the splitlevel approach I have taken to the involvement of the supernatural in tragedy), it is then possible and coherent to say that though Kleist’s plays, for example, mostly avoid final catastrophe, they are more profoundly tragic than previous German drama;134 in that sense, it is intelligible to say that Prinz Friedrich von Homburg, though technically a tragicomedy since it ends well for its protagonist, is more deeply tragic, because of the terrible despair of its counterfactual vision, than many plays that end badly for their protagonists. Here it is relevant to recall the change that Goethe made to end of the first part of Faust. In the Urfaust, Mephistopheles pronounces a grim verdict on Gretchen: ‘Sie ist gerichtet!’ (‘She is judged!’); in the later Faust I, Goethe added a heavenly voice that corrects this to ‘ist gerettet!’ (‘is saved!’).135 How substantial is this alteration? Given the involvement of theological elements and the resurrection motif at the end of Faust II, it is true that the whole work’s status as tragedy is doubtful, at least in respect of Faust himself.136 But suppose one asks how the change bears on Gretchen: does it deflect what would otherwise be her tragedy—and perhaps is her tragedy in the Urfaust— into tragicomedy? Possibly so; but then one might feel that the change is too superficial to mend the damage, that the grief of Gretchen’s fate is not assuaged by that supervenient ethereal voice, that the descent into hell is not undone by the resurrection. Rowan Williams seeks to extend a version of this point from tragedy to religion, or at least to Christianity: ‘Christian theology—contrary to some caricatures of it—does not require us to suppose that suffering is cancelled or even compensated by the hope of ultimate reconciliation’.137 If that were right, it would open the door to a reinstating of Christian (in general, religious) tragedy as a bona fide species of the genus, against the line that I took (in agreement with many others) in my Introduction. But it is not right. What Williams here represents as a caricature of Christianity is in fact orthodoxy: the believer holds that the tears of the saved will be wiped forever from their eyes. For the orthodox, heaven is not merely something that happens after the travails of this life; for those who earn it, it cancels them.

24 Tragedy and Undeserved Suffering In the previous section I suggested, in connection with the idea of secular remedy, that the only genuine remedy that tragedy could accept would have to work in both temporal directions, which makes no logical sense;

160  Tragedy and Moral Redress for a remedy necessarily works in just one temporal direction, and a ‘remedy’ that worked in both directions would not be a remedy. One might initially suppose that exactly the same point applied to redress; but that is not so. Remedy is temporally asymmetric, but redress in my usage is identified with compensation, which is symmetric: crime is compensated by punishment in one temporal direction, punishment by crime in the other; etymologically, the idea of compensation is just the idea of one thing’s balancing and being balanced by another; as I noted in the Introduction, no temporal bias is built into the meaning of the word. Normally crime will precede punishment; however, I shall suggest shortly that tragic literature does admit exceptions to that rule. But, whichever way it goes, the point is that tragic redress works in both temporal directions: for popular morality, it would be a scandal if crime were not compensated by punishment, and equally scandalous if punishment were not compensated by crime, and tragedy by and large conforms to these tenets of popular morality. I have argued that, in tragedy, where there is disaster it is standardly brought about in a comprehensible way by protagonists’ choices that are grounded in their characters, so that they are at fault. I have rejected the common view that tragic protagonists do not deserve their fates at all. An equally common view is that the balancing of desert and punishment, while present in tragedy to some extent, is not nice, so that the punishment exceeds the protagonist’s strict deserts (it is usually thought to go that way round); but I have suggested (§5) that when we factor in the point that the untoward consequences of the fateful choice were predictable, because they arose by probability or even necessity, it turns out that the protagonist is to blame for his failure to foresee the results of his action. Protagonists have knowledge, or constructive knowledge (that is, they should have knowledge), of the way the world works, and in particular of the fact that the consequences of an action can rarely be contained by the agent. Tragic protagonists should have known better; in that sense, and to put it bluntly, they get what they deserve. But not all tragic suffering in the Western literary tradition tracks desert: there is undeserved suffering; there is such a thing as what Steiner calls ‘absolute tragedy’ or ‘tragedy pure and simple’. As we observed earlier (§21), however, it is not easy to find tragic works that count as absolute tragedies when judged as a whole. Such works would be ones where the protagonists are depicted as suffering in an undeserved way. Are there any such works? Euripides’ Hecuba and Trojan Women, mentioned as we have seen by Steiner in this connection, perhaps exemplify the type; but the matter is not straightforward in either case. In Hecuba the ghost of Hecuba’s son Polydorus speculates that one of the gods is destroying her in compensation for (antisēkōsas) her previous good fortune (57–8). That might look like a kind of compensation not matched by desert (though recall my remarks on Horace’s Licinius ode). But it

From Cognitive Failure to No-Fault Tragedy 161 could be said that Hecuba’s innocence in this play is compromised by the gruesome nature of the revenge she exacts for her suffering, not to mention the pleasure which she takes in exacting it (1256–8). This might be a case where we have that non-standard phenomenon that I mentioned in the last paragraph—the punishment preceding the crime. You might object to that possibility on legal grounds: surely, you might say, suffering cannot be merited by what you go on to do, perhaps as a consequence of that suffering (as in Hecuba’s case), but only by what you did before the suffering was incurred, and independently of it. But that thought would, I suggest, be out of line with marketplace morality, as it has always been; it is that popular morality, not the morality of the courtroom or the philosophy seminar, that tragedians presuppose. Human beings naturally feel a sense of satisfaction at retribution, and they do not always worry about the temporal order of events. Brian Vickers thinks that Hecuba is innocent (1973, 4–5, 8), and he notes that, in the Hecuba, ‘Hecuba’s fate symbolizes that of so many characters in Euripides: suffering neither purifies nor ennobles but degrades, brutalizes, for she was no longer a human being in the full sense of the word when she grovelled before Agamemnon seeking revenge, and she has declined progressively since that point’.138 Sophocles’ Electra, Euripides’ Medea and Alcmene, Ovid’s and Correr’s Philomela provide similar cases of the brutalizing power of suffering, as do Shakespeare’s Lavinia and Titus Andronicus, the former assisting the latter in his vicious treatment of Tamora’s sons Demetrius and Chiron.139 (Some feminist critics approve of the blow that Procne and Philomela strike against the patriarchy: for example, Jane Newman compares their action favourably with Lucrece’s allegedly feeble and complicit suicide.)140 In the case of Hecuba as she figures in Euripides’ Trojan Women, there is a sense in which her suffering is compensated not morally but linguistically, by her rhetorical performance in the set-piece debate with Helen, in which Helen is roundly defeated; Hecuba speaks second and is not answered. (At least, Helen is defeated in dramatic terms, though she remains unanswered on points of substance.)141 Steiner, we recall, also listed Shelley’s The Cenci as an absolute tragedy, but here again it seems to me we have a structure in which punishment precedes crime: the play begins with Beatrice’s incestuous rape, but her suffering is compensated in retrospect by the crimes that she goes on to commit. In Artaud’s rewrite of Shelley, Lucrétia tells Béatrice that ‘un crime n’existe que quand il est fait’:142 her immediate purpose is to calm Béatrice’s fears about further rape, but the line also alerts us—by suggestio oppositi—to the possibility of the inverted structure I have been discussing. The list of instances where the status of undeserved suffering is mistakenly claimed for protagonists can be considerably extended. Corneille thought that his Polyeucte, Nicomède, and Héraclius suffered ‘without any fault on their part’ (1999, 100–1), but when we examine

162  Tragedy and Moral Redress their conduct we find that the playwright’s judgement is awry.143 Polyeucte intolerantly disrupts a pagan service of worship, his lyrical soliloquy in the second scene of the fourth act shows disturbing signs of fanaticism, and he deliberately provokes his own martyrdom without regard for the interests and feelings of either Félix or Pauline.144 Nicomède is arrogant and insolent in his behaviour towards his brother Attale (for example in the scene where he does not disclose his identity until his brother has made a fool of himself); but Attale forgives him, saves him, and indeed falls not far short of stealing the limelight from Nicomède.145 The latter, unlike Polyeucte, survives catastrophe, as does Héraclius. Of the three figures Héraclius comes closest to being a tragic hero without fault, but since he survives—and, unlike Sophocles’ Creon, does so intact—we hardly have a case of the undergoing of catastrophe. Most of Corneille’s tragic heroes are venal, even if (usually) the denouements of their plays are not catastrophic for them: one thinks, for example, of Horace, Sertorius, Cinna, and Suréna.146 Suréna is, unusually for Corneille, a genuine tragedy. The hero finds himself in a position under his king Orode which is parallel to Macbeth’s under Duncan. Macbeth, who has no good reason to act against Duncan (as he himself recognizes), does so; Suréna, who by contrast has every reason to act against Orode (as he himself recognizes: 1637–54), fails to do so. (Melantius and Amintor in Beaumont and Fletcher’s The Maid’s Tragedy evince a similar contrast.) Suréna meekly trusts Orode (1663–7), as Duncan does Macbeth, and as Goethe’s Egmont does Philip II: all three are punished for their ingenuousness. Both Duncan and Banquo fit the standard model, as Gervinus well saw,147 so that, though Duncan is not the tragic hero of his play, the option is there to transform him into one, and that is in effect what Corneille did in Suréna. If Suréna and Macbeth, or alternatively Orode and Duncan, were to swap plays— the same point is often made about Hamlet and Othello—­catastrophe would be averted. All of these heroes illustrate the important tragic truth that it is not action or inaction just as such that matters, but action or inaction in a context.148 Innocence is sometimes claimed for Racine’s Britannicus,149 but his naïveté precludes that. Consider in this connection Aeschylus’ Agamemnon, who is forced by Artemis into a dilemma from which his only reasonable exit, as he sees the matter, lies in sacrificing his daughter Iphigeneia, since the alternative— desertion of the fleet—is for him unthinkable (212–13). (By contrast, in Seneca’s Agamemnon there is no dilemma and no moral unclarity.)150 Artemis has becalmed the fleet at Aulis, demanding the sacrifice of Iphigeneia in return for favourable winds. Why does she do this? The poet is notoriously cryptic, referring vaguely to her sympathy for slaughtered animals (123–38), and in particular to the fact that eagles sent by Zeus as a favourable portent have devoured a pregnant hare. A prior offence by Agamemnon against Artemis, found in traditional versions of the

From Cognitive Failure to No-Fault Tragedy 163 saga, seems not to be envisaged by Aeschylus.151 ‘An omen is nothing apart from that of which it is ominous’,152 and one obvious construal of Artemis’ motivation is that, as a Trojan partisan, or as protectress of the young, or both, she is punishing Agamemnon for the havoc he will wreak in Troy, and in particular for the deaths to come of innocent women and children.153 (This generates divine inconsistency, for Zeus ordered the expedition; but such inconsistency is standard fare in traditional myths, as the Hippolytus illustrates.154 Indeed, it is not merely that different gods—Artemis and Zeus—are ‘at odds’;155 it is even the case that individual gods act inconsistently: Artemis both reacts with anger to the slaughter of innocents, and orders the slaughter of an innocent in compensation; Zeus both orders the expedition against Troy and punishes it.)156 But the expedition will reach Troy and commit those atrocities only if it receives favourable winds, and it will obtain those winds only if Agamemnon performs the sacrifice, so that not only does the punishment precede the offence,157 it is also the case that there is a logical circle connecting them. That is to say, it is both the case that Agamemnon incurs the punishment, here and now, only if he commits—goes on to c­ ommit— the crime, and the case that Agamemnon will commit the crime in the future only if he, here and now, incurs the punishment. An even tighter circle can be drawn. Perhaps Artemis is punishing Agamemnon, with the sacrifice of Iphigeneia, not (or not only) for the murderous sack of Troy, but for the very act of sacrificing his daughter.158 The goddess is punishing him with and for the very same thing. The tighter circle reflects the moral oddity of the situation.159 Why does Artemis punish one awful act by demanding another awful act? The tighter circle models the situation by making the awful acts one and the same. How exactly does it work? Perhaps in this way: Artemis knows that if she becalmed the fleet and demanded the sacrifice from him, Agamemnon would (freely) perform it rather than desert his men (recall the discussion of Agamemnon’s freedom in §7), and she punishes him for that hypothetical transgression by realizing the conditions for it, so that when he freely goes on to perform the sacrifice, that act functions simultaneously as the transgression for which he is punished and as the punishment itself. It is important here that Artemis’ knowledge of what Agamemnon would do in hypothetical circumstances (so-called middle knowledge) is quite compatible with the full freedom of Agamemnon’s act.160 In effect Artemis punishes Agamemnon for his hypothetical readiness to sacrifice his daughter—for the kind of person he is, for his ēthos.161 She acts in such a way as to justify her punishment, and since the chain of events crucially depends upon Agamemnon’s free act, the punishment is not unjust. Martha Nussbaum’s characterization of Agamemnon as ‘a previously guiltless man’ (1986, 34) would then miss the point: he may be previously guiltless (at least in Aeschylus), but he is not counterfactually guiltless. In Racine’s Iphigénie, too, divine commands reflect Agamemnon’s own

164  Tragedy and Moral Redress passions.162 Of course, in a court of law previous guilt is the only kind of guilt that matters—no one is interested in future or counterfactual guilt. But tragedy is, as we have said, painting on a larger canvas. Now the usual model for the relation between crime and punishment in the Oresteia is a linear one:163 a crime calls forth punishment, but the punishment is a new crime, which itself calls forth punishment. One speaks in this sort of situation of a ‘cycle’ of violence, but my model of the Agamemnon is more strictly circular, inasmuch as, on my model, the punishment for a crime is not simply itself a crime, but a logical pre­requisite of the crime, and possibly (if we go for the ‘even tighter circle’) the same act as the crime.164 As I have implied at several points in earlier discussion, the clearest cases of undeserved suffering in tragic literature are supplied by characters who are not the protagonists of the works in which they figure. They are often women or children, matching their subject status in civil life:165 one thinks, for example, of Niobe’s children in the Homeric passage quoted above (§22), Aeschylus’ and Seneca’s Cassandra;166 Sophocles’ Tecmessa, Jocasta, and Eurydice; Heracles’ wife and children in Euripides’ Heracles and in Seneca’s Hercules Furens; Iphigeneia in Aeschylus’ Agamemnon, in Euripides’ Iphigeneia at Aulis, and in Lucretius’ description of her sacrifice (§15); likewise Jephtha’s daughter Iphis, who is partly modelled on Iphigeneia, in Buchanan’s version of the biblical story, and Idomeneus’ son in the classical version of the same folk-tale pattern;167 Evadne in Euripides’ Suppliants, Macaria in Children of Heracles, Jocasta in Phoenician Women, Polymestor’s children and Polyxena in Hecuba; Medea’s children (and her female rival) in numerous versions of the myth, especially those of Euripides, Seneca, Corneille, Grillparzer, and Anouilh (but Lessing’s Miss Sara Sampson, which is based on the same story, takes a different path, as we shall see shortly), as well as Roxana and her children in Alabaster’s Roxana; Cassandra, Polyxena, Andromache, and Astyanax in the Iliad and in the various versions of the Trojan Women (especially those of Euripides, Seneca, and Sartre); Tantalus, Plisthenes, and the third, unnamed son in Seneca’s Thyestes;168 the pseudo-Senecan Octavia, who is compared in her eponymous fabula praetexta to Aeschylus’ and Seneca’s Cassandra and Iphigeneia;169 Livy’s and Ovid’s Lucretia, and Chaucer’s in the Legend of Good Women;170 Procne’s son Itys in Ovid and Correr;171 Virginia in Livy and in the Tudor Apius and Virginia by one R. B. (possibly Richard Bower),172 and in Webster; Sabren in The Lamentable Tragedy of Locrine; Alarbus, Mutius, Quintus, Martius, and Lavinia in Shakespeare’s Titus Andronicus,173 the young princes in Richard III, Cinna the poet in Julius Caesar,174 and Lady Macduff and her children in Macbeth;175 Julio, young son of Piero Sforza, in Marston’s Antonio’s Revenge; Sejanus’ children in Jonson’s Sejanus His Fall; Séleucus in Corneille’s Rodogune; Ugolino’s children in Heinrich Wilhelm von Gerstenberg’s dramatization of an episode taken

From Cognitive Failure to No-Fault Tragedy 165 from Dante’s Inferno; Gretchen’s child, mother, and brother in Goethe’s Faust (but not Gretchen herself: see below), and Maria in his Clavigo; Hedvig in Ibsen’s The Wild Duck; Marty South in Hardy’s The Woodlanders, and Jude’s children in Jude the Obscure. I have included a few self-sacrificers in this list—Euripides’ Evadne, Macaria, Polyxena, Iphigeneia (in the Iphigeneia at Aulis), and Buchanan’s Iphis—though their entitlement to feature in it is dubious, since it is unclear whether their suffering is really uncompensated. Certainly, I have suggested, by the time we get to Christian and other martyrdom dramas, where the victim earns a heavenly crown, we no longer have suffering without redress. Many of these figures are not even fully autonomous moral agents in the Aristotelian sense,176 let alone the protagonists of their stories; hence, I suggest, the availability to traditional tragedy of the possibility of their undeserved suffering. For where tragic characters are depicted as genuine agents, and particularly if they are made the protagonists of their stories, I propose that they fit the model according to which suffering is matched to desert. We might set as a rule: the more a tragic figure is depicted as an agent, and the more he or she approximates to being a protagonist, the more, if suffering is incurred, that figure brings it on him- or herself. Racine’s La Thébaïde well illustrates this rule: Jocaste, Antigone, and Hémon all die guiltless deaths (unless the suicides of the two former are to be censured: see §25 below), but, while their roles are important, they are not the protagonists of the drama; those are Étéocle, Polynice, and especially Créon, whose fates all fit the standard model of punishment balanced by fault, for though Jocaste blames the gods (III, 2 and 3), the operatur principle clearly applies. Racine’s Eriphile is a protagonist of his Iphigénie, but she is not guiltless. As Racine puts it in his preface, she meets ‘the misfortune which, in her jealousy, she wished to bring upon her rival’, and so ‘in a way deserves to be punished, without being, however, altogether unworthy of compassion’ (an excellent summary of the standard model).177 Io in the Prometheus Vinctus is an undeserving sufferer, but her agency is correspondingly diminished (and in any case Zeus will make all good in the end).178 Rahel in Grillparzer’s Die Jüdin von Toledo is a Cassandra figure converted into a protagonist, who carries responsibility for her actions and her tragedy after the standard model. (Her entering the garden, and her mimicking of the queen, are acts of hubris and hamartiai.) She has something of Antigone about her, as her sister Esther has something of Ismene. Poulenc’s Dialogues des Carmélites is also a telling case in this regard: most of the nuns are innocent victims of the Terror, but the figure on whom the opera focuses, and whom it makes its protagonist, namely Blanche, is compromised by her motives for joining the convent, and by her impulse, as it seems, to seek martyrdom. Goethe’s Gretchen is a sufficiently significant figure to test the thesis I am advancing; but, again, it seems to me that her hamartiai are clear to see, in her killing of her mother and her own child. As in

166  Tragedy and Moral Redress the case of Kostelnička’s act of infanticide in Janáček’s Jenufa, we are not inclined to be moralistic, in the pejorative sense of the word, but that does not mean that we should not moralize. Gisbert Ter-Nedden has suggested (1986, 234) that Gretchen resembles Lessing’s Emilia Galotti (whom I shall discuss in §25), Schiller’s Luise (Kabale und Liebe), and Hebbel’s Klara (Maria Magdalena), in standing accused of excessive piety and primness. That seems right: Luise keeps her oath, like Hippolytus, when she should break it; Klara does not have the courage to tell her father of her pregnancy. In this connection it is relevant to consider Shakespeare’s Lucrece.179 One might initially think that she should be listed above with Livy’s, Ovid’s, and Chaucer’s Lucretia as an undeserving sufferer, and then, since she is the protagonist of Shakespeare’s poem, she would constitute a problem for my thesis, namely that tragic protagonists bring their fate upon themselves, undeserved suffering being earmarked for non-protagonists. But close inspection of Shakespeare’s poem shows, I believe, that she conforms to what I am calling the standard model, the schema of punishment redressed by error. For Lucrece’s innocence goes too far, morally speaking. Tarquin’s lascivious expression gives him away (99–105). But she that never coped with stranger eyes Could pick no meaning from their parling looks, Nor read the subtle shining secrecies Writ in the glassy margents of such books. She touched no unknown baits nor feared no hooks,    Nor could she moralize his wanton sight    More than his eyes were opened to the light. In other words, his lust was plain as day, but she failed to see it. When she later looks at a painted figure of Sinon, who deceived Priam with a fair outward form, just as Tarquin has deceived her, Lucrece realizes her ingenuousness (1499–1568); she tears at the picture with her nails and chides her own folly in so doing (1568), emblematic of her earlier tragic folly. Brutus suggests that Lucrece should not have committed suicide but struck back at her assailant (1826–7), but it is made clear in the poem that this was not a realistic possibility for her.180 That is not her hamartia, but rather her trusting insouciance, like Priam’s in respect of Sinon. The Countess of Salisbury in the Shakespearean King Edward III is a Lucretia type, but she is able successfully to repel the king’s advances, and tragedy is averted.181 Richard Levin thinks that the Lucretia of Heywood’s The Rape of Lucrece is ‘wholly innocent and wholly admirable’ (1981, 89). But Heywood stresses a point (V, 1) which is underplayed by Shakespeare but which will be important in the next section, namely that her act of suicide is a fault.

From Cognitive Failure to No-Fault Tragedy 167 Such figures as the ones I have listed as undeserving sufferers are often described as ‘minor’ characters, and in a technical sense they are, but viewed from a larger perspective they are often not. The murder of Cinna the poet in Shakespeare’s Julius Caesar is vital to its thematic structure.182 Cassandra’s role in Aeschylus’ Agamemnon, say, or Eurydice’s in Sophocles’ Antigone, or Tecmessa’s in his Ajax, seem to be anything but minor. At the end of her long scene in the Agamemnon, Cassandra exits to her death and the chorus sings a short anapaestic dirge (1331–42) in which it mourns and moralizes on Agamemnon’s death, but not a word is wasted on Cassandra’s fate, which she had foretold: this is powerfully felt as an absence.183 Adrian Poole well remarks of Eurydice in the Antigone: ‘Like Deianeira, Eurydice exits without saying a word in reply, but her significance is out of all proportion to the number of lines which she speaks’.184 In Tecmessa’s case a kind of ‘double take’ operates in respect of her status. Initially the reader or spectator might assume that she is Ajax’ wife, but we are reminded that she is his slave, a ‘spear-bride’, not his wife,185 and so well below enjoying a status as Ajax’ social equal; but then we observe from the play how she actually interacts with Ajax, and this has the effect of blurring the categories of slave and wife in our minds: there is a sense in which Tecmessa is both. For, as several critics have noted, Tecmessa appeals to Ajax as her husband, in the same terms, in fact, as Andromache uses to Hector in the sixth book of the Iliad, and their relationship seems to be viewed by both in distinctively marital terms.186 Ajax clearly regards their son Eurysaces as his legitimate heir.187 So categories like major and minor, and even, as we see, free and enslaved, can be blurred in tragedy. Hecuba, too, is a slave in the Euripidean play named after her, but she has the face to argue openly with Agamemnon and she even gets her way.188 Andromache in her eponymous play argues with Menelaus and is even somewhat abusive to him; she is then rescued from death by Peleus after a wrangling scene between Peleus and Menelaus rather like the quarrel between Lear and Cornwall over the stocked Kent.189 (Of course, Tecmessa and Hecuba and Andromache are, one might say, a special kind of slave—prisoners of war who once enjoyed high social status.) Women and children are, then, prime victims of undeserved suffering in tragedy—and animals: in his third Georgic Virgil describes the plague that brings death to the land in unmistakeably tragic terms,190 and animals are, though not agents in the Aristotelian sense (for they do not share in praxis),191 hardly minor characters in that poem. In the third Georgic we find a remarkable merging of human and animal worlds, rather in the manner of Janáček’s Cunning Little Vixen. Animals are repeatedly described in distinctively human terms,192 so that it is not a solecism to think of their fates as being, sometimes, tragic. The frequent use of hunting imagery in tragedy often bridges the gap between human and animal worlds, especially when sacrificial victims are compared to beasts.193

168  Tragedy and Moral Redress In spite of Servius’ suggestion that the fourth book of Virgil’s Aeneid has an almost comic style (because of the love interest),194 Dido follows the standard tragic pattern, as does Turnus in the second half of the Aeneid,195 which is why some of Virgil’s artistic successors have used their stories to construct their own tragedies. (One thinks of Gager, Marlowe, Purcell, and Berlioz in Dido’s case, and in Turnus’ of Jean Prévost’s Turne.)196 But while Laocoon may be guilty of hubris, his sons have nothing to answer for. In a moving fragment of his Bellum Punicum (preserved by Servius Auctus on Aeneid III, 10), Naevius reports the departure from Troy into exile of the wives of Anchises and Aeneas: ‘Amborum uxores noctu Troiade exibant capitibus opertis, flentes ambae abeuntes lacrimis cum multis’ (‘The wives of both were passing out of Troy by night; their heads were veiled, and both were weeping many tears, as they went away’; tr. Warmington). In the Aeneid Virgil offers several cases of undeserved suffering, such as the capsized sailors in the storm of the first book, who (I, 118) apparent rari nantes in gurgite vasto appear, widely dispersed, swimming in the vast waters, or his innocent victims of the sack of Troy. When the fateful wooden horse is being dragged with hawsers up into the city (II, 238–9), pueri circum innuptaeque puellae sacra canunt funemque manu contingere gaudent, all around boys and unwed girls sing hymns and rejoice to touch the rope with their hands. Virgil, as Roland Austin remarks (ad loc.), ‘was thinking with a certain tender realism of the youth and happiness of these simple creatures, so soon to perish, or to be seized as slaves’; he was also, no doubt, recalling similar descriptions in the first stasimon of Aeschylus’ Seven Against Thebes, and in the first stasimon of Euripides’ Trojan Women, as well as his own fourth Georgic (476), later echoed again in the underworld of the Aeneid (VI, 307).197 The lines just quoted are echoed after the city has fallen, when (II, 766–7) pueri et pavidae longo ordine matres stant circum. There, all around, stand endless lines of frightened mothers with their children. I give Austin’s translation (ad loc.); he observes that ‘The unfinished line stant circum is one of those that no one would wish completed’.198

From Cognitive Failure to No-Fault Tragedy 169 These cases demonstrate, I think, that it would be going too far to say, as Richard Jenkyns does (1998, 657), that Those things which in our everyday discourse we readily call tragic— earthquake and famine, the torture and murder of innocents—are not, within literature, the matter of tragedy at all: they are too ghastly. Tragedy requires a kind of greatness, and that greatness resides not merely in the literary form, nor in the quantity of suffering depicted, but in its quality; tragedy is concerned not with the most pain that can be imagined but with the manner in which it is experienced and understood. The examples we have given show that the first sentence of this quotation is questionable. But Jenkyns is right in what he goes on to say, and indeed it has been an important part of my argument so far in this book that deserved suffering is more prominent in tragedy than has been recognized in recent criticism. The urge felt by the tragic tradition to justify the sufferings of protagonists in terms of their actions and characters pushes undeserved suffering to a penumbra in which the fates of characters are played out who either are not, in the fullest sense, agents, or at any rate are not protagonists. As Donald Mastronarde notes, in connection with the presence of the Euripidean Medea’s children on stage while she delivers her famous monologue (1021–80), ‘as child characters they do not, by convention, have full dramatic status’.199 Children are indeed usually anonymous in Greek tragedy and Seneca, unless they have a significant name, like ‘Astyanax’ or ‘Eurysaces’.200 By contrast with children, women can be agents in tragedy, and are often protagonists; but then, as we have noted, they exemplify the rule that suffering is compensated by desert. Still, as I have been exploring in this section, tragedians do recognize the phenomenon of undeserved suffering, and sometimes in the very act of apparently marginalizing such suffering they draw attention to it. In the Ajax Tecmessa laments her husband’s death, then falls silent: her last words—‘he has gone away leaving me grief and tears’ (972–3: ἀλλ᾽ ἐμοὶ λιπὼν ἀνίας καὶ γόους διοίχεται)—echo eerily through the ill-tempered debate between Teucer, Menelaus, and Agamemnon that follows.201 In the Antigone Eurydice’s silent exit to the palace is a moment of high pathos.202 Silence in public was, except in a few special (religious and ritual) contexts, socially enjoined on high-class women in classical Athens;203 because of this, silence on the Attic stage is particularly associated with women.204 When one reflects on such cases one might be tempted to agree with an assertion that Sebastian Gardner makes, in a passage that I quoted in the Introduction, that tragedy presents loss and suffering as having ‘unconditional and hence uncompensatable reality’ (2003, 235). But that would be, as I hope I have shown in this part, an overreaction.

170  Tragedy and Moral Redress

25 Tragedy and Morality Gardner argues for his thesis in the context of a discussion of Schiller’s quite different view that ‘the point of tragedy lies in the final realisation that tragic loss and suffering are, in comparison with moral evil, of no true, ultimate import, and in that sense it affirms a total redemption’, and that ‘to the extent that the tragic victim has avoided moral evil, he has in reality lost nothing’ (ibid., 233, 234). Schiller’s is the Socratic stance— that the virtuous person cannot be harmed by the exterior contingencies of life, only by an inner impulse that diminishes his or her virtue—and it contrasts with an Aristotelian approach according to which eudaimonia is subject to luck, and can be undermined by external and contingent events.205 The Socratic view is beautifully expressed by Milton in his Comus, where ‘the sage / and serious doctrine of Virginity’ (786–7) is expounded: we are assured that (589–90) Virtue may be assailed, but never hurt, Surprised by unjust force, but not enthralled; and the Lady defies her captor with the words (663–5): Thou canst not touch the freedom of my mind With all thy charms, although this corporal rind Thou hast immanacled, while Heav’n sees good.206 Castiza in The Revenger’s Tragedy tells us that (IV, 4, 152–4) A virgin honour is a crystal tower, Which, being weak, is guarded with good spirits; Until she basely yields, no ill inherits.207 So too Livy’s Lucretia after she has been raped by Tarquin (‘my body alone has been violated; my mind is free of guilt’: I, 58, 7), Shakespeare’s Lucrece likewise (1653–9),208 and Richardson’s Clarissa after she has been raped by Lovelace.209 The opposite, Aristotelian view is equally powerfully expressed by Chaucer’s Black Knight in The Book of the Duchess, which is based, as so often in Chaucer, on a dream motif.210 The narrator hears the knight’s sad story in his dream, and proceeds (710–20): And whan I herde hym tel thys tale Thus pitously, as I yow telle, Unnethe might y lenger dwelle, Hyt dyde myn herte so moche woo. ‘A, goode sir,’ quod I, ‘say not soo!

From Cognitive Failure to No-Fault Tragedy 171 Have som pitee on your nature That formed yow to creature. Remembre yow of Socrates, For he ne counted nat thre strees Of noght that Fortune koude doo.’ ‘No,’ quod he, ‘I kan not soo’. Note the omission of the initial unstressed syllable of the iambic tetrameter in the last line, throwing extra weight on the simple ‘No’. The Black Knight here roundly rejects the narrator’s attempted Socratic consolation. Lessing’s Miss Sara Sampson and Emilia Galotti might seem to yield exceptions to the rule that I have enunciated, but in fact they exemplify it. The former of these plays updates the story of Jason and Medea, and innovates by foregrounding the figure of the Medea’s female rival,211 who is now the heroine, Sara, and she, like Richardson’s Clarissa on whom she is so closely based, and in accordance with a suggestion that I aired in the last section—namely that traditional tragedy permits minor characters (in the technical sense) to undergo undeserved suffering, but that as soon as they become major characters they adhere to the standard model, in which suffering tracks desert—makes crucial moral and intellectual mistakes:212 she flouts conventional morality by disobeying her father and eloping with Mellefont, in her dealings with whom, as with Marwood, she displays naïveté. It is wrong to say, as H. B. Nisbet does, that ‘Sara is too innocent to fit easily into a traditional tragic framework’ (2013, 198): as he himself puts it, ‘the hour of the self-assertive woman has not yet come—not, at least, in German drama’ (ibid., 203). To be sure, the play is moving towards a more enlightened morality: Sara’s father Sir William forgives both his daughter and Mellefont, buries them together in, it is implied, consecrated ground, despite the fact that Mellefont is a suicide, and he frees his servant Waitwell.213 But the old sexual morality still dominates the play, and frames the tragedy: Sara herself regards her death as a punishment for sin.214 So we have the standard framework consisting of hamartia causing catastrophe.215 The major figures of the drama, as so often in tragedy, see themselves as victims of fate, but their actions are, as usual, fully naturalistically motivated.216 Emilia Galotti adopts the Socratic posture of Milton’s Lady, but the play of which she is the heroine expresses an Aristotelian outlook in the sense that Emilia’s choice of death over seduction is presented as a genuine loss. Does Emilia deserve her fate? One might initially think that she is brought down by sheer bad luck.217 She is first indecently accosted at Mass by the Prince of Guastalla, and then ambushed by his henchmen and kidnapped as she travels to her wedding; her fiancé is fatally wounded in the attack, and Emilia is forcibly escorted to the prince’s Lustschloss, where only her willing death at the hands of her father frustrates the prince’s designs on her. The outline of the story goes back to Livy’s tale

172  Tragedy and Moral Redress of Appius and Virginia (III, 44–58), which was widely exploited in the medieval tradition, emerging for example in the Roman de la Rose and Chaucer’s Physician’s Tale,218 as later in Shakespeare’s Measure for Measure and Grillparzer’s Ein treuer Diener seines Herrn. But in fact, despite superficial appearances, Emilia’s fate is not a matter of bad luck, and she herself takes a more realistic view of her position than Milton’s Lady does of hers, partly because her tempter is a handsome prince, not a bucolic monster. When her father Odoardo suggests that her innocence is über alle Gewalt erhaben (‘transcends any act of violence’), Emilia replies: But not any act of seduction.—Violence! Violence! Who cannot frustrate violence? What violence means, is nothing: seduction is the true violence.—I have blood, father, youthful, warm blood as any other. My senses too are senses. I guarantee nothing. I am good for nothing.219 Emilia knows that she will not resist the prince; perhaps, as Goethe thought, she already loves him, or is at least sexually attracted to him.220 He has murdered her husband-to-be; but she sees and fears in herself a reincarnation of Petronius’ Widow of Ephesus.221 It is plausible that the quoted passage marks Emilia’s first realization of these depths in her own nature.222 A literary descendant of Emilia is Grillparzer’s Erny, whose sexual attraction to Otto is more evident and better motivated than Emilia’s to the prince, but Grillparzer is filling in details of an outline that Lessing gave him. Mistakes are made by all the main characters of the play, Emilia included. For a start, she goes on her own to church, where the prince easily traps her; her fiancé Appiani foolishly fails to inform the prince of his intention to marry Emilia; Emilia’s parents have brought her up too strictly and unimaginatively, so that she is not properly armed, mentally and physically, against the prince’s sexual overtures; Emilia’s mother, bedazzled by fashionable court life, insists that she not tell Appiani of the prince’s indecent advances; and Emilia obeys her; Odoardo passes up an opportunity to kill the prince in favour of killing his daughter. These are all tragic errors.223 The errors are grounded in character, as they should be according to Aristotelian theory, endorsed as we have seen by Lessing (§2): Emilia is weak and naïve, her mother silly and vain, her father erratic and impulsive.224 The errors rest on decisions that could have been made differently, so that it is not correct to say that ‘the causes of the tragedy are ultimately to be found not in the psychology of the characters concerned, but . . . in the nature of the society they live in’.225 Decisions are made in a social context, to be sure, and, as we have said (§7), contexts place constraints on what can be decided, indeed constitute the very possibility of making decisions in the first place; but the decisions must still actively be made by people—contexts do not make decisions—and the business of deciding can be done in a better or a worse way. The point that

From Cognitive Failure to No-Fault Tragedy 173 catastrophe in Emilia Galotti stems from human, freely chosen errors of judgement is well noted by Horst Steinmetz (1987, 100), but he thinks that the grounding of disaster in culpable and avoidable cognitive failure is ‘untragic’, whereas I have argued in this part that it is of the essence of European tragedy. Steinmetz also argues that the catastrophe is not sufficiently motivated by agent error; he thinks that there is too much coincidence in the play, and (as we noted in §7) in his view the involvement of coincidence diminishes agent responsibility. But this view is too exacting: the plot of Othello, as I observed (§7), is shot through with accident and contingency; but we do not think, I suggest, that this fact undermines the responsibility of the main agents. It would be absurd to claim that because Iago’s stratagem depends so extensively on luck he is therefore off the moral hook. That idea, generalized to all human action, would destroy the entire applicability of any notion of agent responsibility. I have listed some of the tragic errors that we find in Emilia Galotti. But perhaps none of these errors is the real tragic mistake. Aquinas, following Augustine, states that a woman may not commit suicide in order to avoid violation, since suicide is the greatest of sins, whereas violation is no sin at all if there is no consent.226 Perhaps Emilia’s suicide at the end of the play—and it is in effect an act of suicide, for although her father deals the death blow, she begs him to do it—is her principal hamartia. That thought might initially seem perverse: her death is her punishment; how can it also be her transgression? The suggestion would involve a logical circle reminiscent of, but not the same as, the circle that, earlier in this chapter, I argued was operative in the case of Aeschylus’ Agamemnon. Now, one argument for discerning a pattern in a particular work is that we can find the same pattern exemplified in another work by the same author. I take the logical circle that I detect in Emilia Galotti to be present in a very similar form in Lessing’s Philotas, which is based on Plautus’ Captivi. (Perhaps the circle is present, too, in some versions of the Sophonisba story, on which Philotas is also, in part, based.)227 Philotas, a young prince, is taken hostage by King Aridäus, who once was on friendly terms with Philotas’ royal father but is now at war with him. Philotas learns that Aridäus’ son has also been captured by his (Philotas’) father, so that the situation is now symmetrical: each king has the other’s son in his power. To break the stalemate—and the chess analogy naturally suggests itself228—Aridäus plans an exchange of hostages, and he tells Philotas that he hopes such a gesture will restore peace. At this point, in order to frustrate Aridäus’ irenic initiative and give his own father a military and diplomatic advantage, Philotas commits suicide. Suicide in tragedy is problematic, from Sophocles’ Ajax to Garnier’s Porcie to Addison’s Cato. Augustine thought Lucretia blameworthy for her suicide,229 and it seems clear that we are meant to censure Philotas’ conduct.230 In the play Philotas is criticized on grounds of immaturity and ­irrationality—Peter Pütz calls him ‘autistic’, and the drama ‘a tragedy

174  Tragedy and Moral Redress of youth’ (1986, 102, 115)—so that his hamartia is quite close to Antigone’s, as I diagnosed that in Chapter 2, and allusions to Ajax’ and Heracles’ madnesses suggest that the drama is hostile to an unreflective and excessive militarism.231 That fits with what we independently know of Lessing himself, who strongly disapproved of suicide, and took a disparaging attitude to martyrdom plays.232 There can be no doubt that, in Philotas and Emilia Galotti too, as elsewhere, Lessing would have preferred to see a kompromissbereiter Humanismus triumph, as it does in Minna von Barnhelm and Nathan der Weise, rather than the moralistic rigorism that he in fact depicted in his chief characters.233 If this reading is right then once again we have a coincidence between transgression and punishment. A similar point seems to me to apply both to Calderón’s Constant Prince, who brings his fate on himself,234 and to Johann Elias Schlegel’s Ulfo in his Canut, whose uncompromising Ruhmbegier at the expense of Menschenliebe—the last word of the play—leads him to a death that is in effect suicide, condemned by Canut as Raserei. Indeed we might say the same of Sara Sampson, for ‘Sara dies like a martyr, forgiving her enemies, but her martyrdom is self-inflicted’.235 Karl Guthke argues (2012–13), cogently, that Lessing’s three suicide plays show a humane attitude towards Mellefont, Emilia, and Philotas; but I do not think that we are therefore invited not to judge their conduct:236 we may pass judgement, without supposing that we would have done any better, and without losing sympathy for the tragic agents. The Aristotelian structure is fully in place: the suicides of Emilia and Philotas, in particular, are forgivable, but they are still errors;237 if they were not, the question of forgiveness could not arise. We have uncovered logical circles in the plots of Aeschylus’ Agamemnon, and Lessing’s Emilia Galotti, Philotas, and perhaps Miss Sara Sampson, in which the punishment either precedes, or coincides with, the crime; I suggest that the phenomenon, though non-standard, is not particularly uncommon in classical and classicizing tragedy. In Seneca’s Hercules Furens Juno has ordered Hercules to harrow hell (604), yet she also complains of his hubris in doing just that (46–63). The harrowing of hell is itself both an occasion for, and a punishment of, hubris; the imposition of that labour on Hercules is a punishment for (among other things) carrying it out. In tragedy, a single act may be both a duty and a crime.238 The tragic Medea kills her own children as a way of punishing Jason’s infidelity and treachery, but—at least in some treatments of the story—it is implied that the infanticide is not only a terrible act of moral folly but also constitutes its own punishment. This is particularly clear in the versions that we have from Euripides and Seneca; in Euripides, for example, the idea that the murder of her children is also tragic for Medea is anticipated by the chorus (996–9), before it is set out by the heroine herself in her great dilemmatic monologue, and reaches its apogee in her despairing cry as she does the deed: δυστυχὴς δ᾽ ἐγὼ γυνή (‘Wretched woman that

From Cognitive Failure to No-Fault Tragedy 175 I am’, 1250). (By contrast, in the treatments of the story by Corneille, Grillparzer, and Anouilh that feature of their model is not reproduced.) Agamemnon answers Orestes’ prayer in Aeschylus’ Choephori before he has uttered it, by sending Clytemnestra’s dream, which prompts her to dispatch the libation bearers, which is in turn a causal factor leading to Orestes’ prayer.239 As a last example of the logical phenomenon to which I am drawing attention, consider again Sophocles’ Antigone. Antigone chooses her own death, and that choice is both a mistake, a hamartia, and a punishment for that very mistake itself. Initially, as we have seen, she thinks of her impending death not as a punishment but as a consummation devoutly to be wished; but finally she reverts to the normal and natural view that her virginity was made for marriage and childbearing and domesticity, not for premature death. ‘If there is, in the symmetries of mortality, any counterpoise to a tomb, it is the bridal bed and the bed of childbearing’, writes Steiner (1984, 241). Yes indeed, and Antigone selects the wrong option: she is punished for the choice she makes, the tomb, by the choice she makes, the tomb. As I have said, I do not think, despite surface appearances, that Lessing’s Emilia is an exception to the rule that the fate of tragic protagonists is morally redressed by desert. And that upshot is in line with Lessing’s own theory of tragedy:240 The thought that there could be people who are unhappy without any fault is in and of itself hideous. The pagans tried to remove this hideous thought as far away as possible; do we wish to nurture it? Do we want to take pleasure in plays that confirm it? We? Whom religion and reason should have convinced that it [that thought] is as false as it is blasphemous?241 ‘Did this man,’ F. L. Lucas asks, ‘who lived through the horrors of the Seven Years’ War, fondly console himself with the thought that all its agonies were deserved?’ (1957, 117). A good question, no doubt, but the grip of tradition is strong, especially if it is supported by theory. And the truth is that, despite the occasional flickering of doubt, Lessing did think that the answer to Lucas’s question—which he indeed took seriously, and to which he responded—was affirmative.242 A supporter of Leibniz’s theodicy, Lessing argued that everything is for the best, and that the world is full of wisdom and goodness (Weisheit und Güte), even if we cannot always see the connections. Further, he held that artists should restrict themselves to cases where the connections are morally clear, where punishment and suffering are matched by desert. The inclusion of undeserved suffering in a work of art, he said, is to be avoided because that would not help us understand the whole.243 So while Livy’s Virginia, on whom Lessing’s Emilia is based (Lessing himself called Emilia a ‘bürgerliche Virginia’),244 is a mere victim, in Lessing’s hands the story undergoes

176  Tragedy and Moral Redress the treatment standardly applied to tragic protagonists, and becomes a moral tale with the usual concatenation of desert and punishment. And, although it cannot be argued here, it seems to me that the same is true of that other great classical figure of eighteenth-century fiction, modelled this time on Lucretia, Richardson’s Clarissa.245 Lessing’s has been the orthodox position among theorists of tragedy, at least until recently, and it also fits, as I have argued in this part, a wide range of tragedies; indeed I have suggested that, in the European tradition, it is the rule for protagonists. Cases which might at first glance look like exceptions, or which have been held to be exceptions by critics, such as those of Sophocles’ Oedipus and Antigone and Lessing’s Sara Sampson and Emilia Galotti, show on closer inspection not to be so. Even Shakespeare’s Cordelia and Desdemona and Lucrece are, as we have seen, not without fault. Melville’s Billy Budd reminds us that it involves a cognitive error for an adult in an adult world to behave with childlike innocence and ignorance. Brecht’s Mother Courage insouciantly lives off the war— the Thirty Years’ War, this time; who can be surprised if it turns round and bites her?246 As I mentioned in the Introduction, Mutter Courage democratizes Schiller’s Wallenstein; the moral that Brecht brings down to earth is already there for, though not discerned by, Schiller’s cognitively deluded warlord.247 It is common to dismiss Aristotle’s doctrine of hamartia as having no application to tragedy,248 but I have argued that the reverse is true. There may, as I conceded at the beginning of this chapter, be no a priori reason why the sufferings of protagonists had to be represented in the European tradition as deserved; it may be the case that, as Heath says, ‘it is not self-evident that the suffering of a man of outstanding moral integrity would be untragic without moral redress’ (1987, 86; my emphasis). Nevertheless, self-evident or not, the Aristotelian approach is what we find in the tradition. And the fact that so many different tragic authors, including some such as Chaucer who had no knowledge of Aristotle’s doctrine of hamartia, have even so conformed to that doctrine very strongly suggests that, in formulating the idea of hamartia as the motor of tragedy, Aristotle hit on something of a natural kind. The present-day popularity of the idea that tragic suffering is undeserved is in part driven by a modern ideology which says that all suffering is, ultimately, undeserved. This ideology may be right or it may be wrong; I hope it is clear that I have taken no stand on that question. The point here has been that the tragic tradition does not consider its protagonists to suffer innocently, that tragedy does offer moral redress. Hence, while Gardner’s view that tragedy represents loss and suffering as having ‘unconditional and hence uncompensatable reality’ seems initially attractive when one considers clear cases of undeserved suffering, such as Cassandra’s, Tecmessa’s, or Eurydice’s, it turns out to be less secure under the impact of a wider and deeper survey, and in particular when we focus on the suffering of protagonists. Here, so far from its

From Cognitive Failure to No-Fault Tragedy 177 being the case that, in Terry Pinkard’s words, ‘human sinfulness cannot be an appropriate subject of tragedy’ (2015, 152), we find the precise opposite: sin—in general, error, including and especially cognitive and moral error (recall once more that these are not exclusive categories)—is tragedy’s subject par excellence. Kafka’s K., who is arrested suddenly one day ‘without having done anything wrong’ (‘ohne dass er etwas Böses getan hätte’), is a distinctively modern type.249 Before the onset of modernity, if you are the protagonist of your story and bad things happen to you, they do so because you are at fault. It is, then, incorrect to say, as Gardner does, that ‘Tragedy offers a total perspective on the world which does not accord with the way in which moral reflection conceives it’, or that ‘tragedy contests the supremacy of morality by intimating a more fundamental kind of value which is indifferent to moral concerns and which cannot be brought into coherent relation with moral value without ascending to a speculative philosophical level’ (2003, 245). Most tragedy, most of the time, is in harmony with morality—that is, with ordinary, popular morality. Having started my defence of the Aristotelo-Bradleian view with Aristotle, let me finish it with Bradley. Gardner’s assertions stand against Bradley’s famous claim that ‘the ultimate power in the tragic world is a moral order’ (1991, 46–7). In this part I have been defending a slightly weaker claim than Bradley’s, namely that tragedy, in an important sense (that is, so far as the fate of its protagonists is concerned), conforms to the moral order. But the stronger claim is plausible. Of course Bradley does not intend the plainly false thesis that tragedies always end well from an ordinary moral point of view, that they always end with a restoration of a morally good order.250 That would be poetic justice, and Bradley makes clear that he repudiates the application of that notion to tragedy (ibid., 45). He means that the underlying drive and ethos of tragedy is a moral one. (Paul Kottman rejects Bradley’s view, quoting Johnson’s famous assertion that Shakespeare ‘seems to write without any moral purpose’; but, in context, it is clear that what Johnson means is that Shakespeare is not sufficiently morally didactic, a claim that is not inconsistent with Bradley’s approach. Few people would now think that their lack of didacticism was a weakness in Shakespeare’s works, and it is interesting that elsewhere Johnson appears to except Othello from the complaint, for he told Boswell that it ‘has more moral than almost any play’.)251 Again, Bradley is careful to note that his view does not imply that we should be moralistic, in the pejorative sense (§2), towards tragic protagonists. He adds the claim that the moral order ‘shows itself akin to good and alien from evil’ (ibid., 47), and the context of this assertion indicates that he takes the idea of evil very seriously. There I think he goes too far, and I have said nothing about evil in my own exposition, except to note, following Aristotle, that tragic heroes and heroines are typically not

178  Tragedy and Moral Redress evil. But Bradley’s underlying idea can be satisfactorily expressed without appealing to such an extreme and dubious notion as that of evil: it is that, in tragedy, what we see is the body of the moral order labouring, more or less successfully but at great cost to its integrity, to purge itself of harm. Bradley writes that the moral order ‘appears to engender this evil within itself, and in its effort to overcome and expel it, it is agonized with pain, and driven to mutilate its own substance and to lose not only evil but priceless good’ (ibid., 50). As I say, I have not been concerned with the workings of the moral order in this sense, but I believe that something close to Bradley’s view is right, provided that his appeal to the concept of evil is suitably secularized and toned down, and provided also that we read his claim in a sufficiently figurative way—in particular, in such a way as not to undermine the basic point, on which I have laid consistent emphasis, that tragic heroes and heroines act freely, and are themselves responsible for their actions.

Notes 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12.

See P. Hammond 2009, 28; Gaskin 2013a, 191. Cf. Heath 1987, 87. Cf. Boethius, Consolatio Philosophiae V, pr. 1; Sontag 2003, 36. Bradley 1991, 316, 320, 353–4. Bradley 1991, 30. Cf. Gervinus 1872, vol. 2, 149, 151. Nuttall 2007b, 284–9. Bevington 1998, 33; cf. Miola 1992, 99; Ewans 1996, 438–40. Cf. L. B. Campbell 1938, 171; Farnham 1970, 287–8. So Shell 2010, 184–5, 221–2. See Potter’s note on V, 4, 65. Chaucer 1995, 88. De Mourgues 1967, 113–14; cf. Pocock 1973, 227–8; Sussman 1975, 250– 1; Gossip 1981, 151–4. 13. Aristotle, Poetics 1454a37–b2; Horace, Ars Poetica 191–2, with Rudd ad loc. 14. J. Campbell 2005, 180, 238; cf. 143, 158, 222–38, and esp. 180–6. 15. Cf. Orgel 2002, 167. 16. Pace Norland 2009, 98, 116–17. 17. See Miola 1992, 23–4, 34, 39; Cooper 2010, 150; A. Boyle 2017, cxix–cxxiv. 18. Gooch 1981, 91–3. 19. See Greenblatt 2001, 169–74, 185–95; Halpern 2016, 28–30. 20. The Dynasts III, 6, 3 (1965, 467; cf. 501). 21. Garvie 2009, 321; see further ibid., 258–61; Tarrant 1976, 157–9; A. Brown 1983, 30–1. 22. In a remarkable scene in the Octavia—remarkable because it folds one ghostly appearance within another—Claudius’ ghost appears to Agrippina’s ghost, demanding vengeance for his own and Britannicus’ murder (614–17). On Agrippina’s ghost and Nero, see A. Boyle 2008, 218–19, and note on line 618; Kragelund 2016, 246–8. 23. Loschi, Achilles 50–86; given Paris’ scepticism about ghosts and the underworld, which follows hard upon (87–121), it is suggested that Hecuba saw Troilus’ ghost because she wanted to (118–19): see Grund 2011, 52–4.

From Cognitive Failure to No-Fault Tragedy 179 24. Oeneus scoffs at Althaea’s dream, suggesting that it simply bodies forth Althaea’s waking fears: Meleager 603–36 (1994, vol. 1, 82). Note that in classical thought a figure seen in a dream is a ghost: L. Parker on Euripides, Alcestis 354–5. 25. Hoxby 2015, 217. Cf. also Megaera in Wilmot’s Tancred and Gismund D4v; Norland 2009, 119. 26. Leavis 1952, 141, 144. Cf. Nuttall 2007b, 279–80. 27. Mueller 1980, 92–104. 28. On this work see Silk and Stern 1981; Henrichs 2005. 29. See Aristophanes, Frogs 1491–2; Henrichs 1986; J. Gregory 2005, 255; Mastronarde 2010, 12; Halliwell 2012, 94, 152. 30. Steiner 1961, 193–4; cf. 1990, 152–3; 1996, 543; 2008, 33–4; Alexander 1955, 29–30; Dover Wilson 1964, 52. 31. Wilamowitz 1872, 24–32. See also Kaufmann 1968, 191–4; Dodds 1973, ch. 5; Vickers 1973, 48–9 n. 35; Silk and Stern 1981, 251–65. 32. Cf. Ahrensdorf 2009, 176–7. 33. McKenna 1991, 30. 34. Cf. Thucydides III, 45, 3; Cairns 2016, 81–2, 89. 35. Garvie 2016, 43. 36. Cf. Halliwell 2012, 92. 37. See also Halliwell 2005, 409–11 (cf. 2012, 353), on Longinus. 38. Cf. Eagleton 2005, 13, 17. 39. Sophocles, Oedipus Rex 101; cf. Oudemans and Lardinois 1987, 91. 40. Seneca, Thyestes 637–8, 776–7; see A. Boyle 2017, cxxv, and ad locc. 41. Schiller, Die Jungfrau von Orleans IV, 11; V, 1; V, 4, 3113–16. 42. Kleist, Penthesilea 2400, 2405–6, 2445. 43. Euripides, Orestes 259, 268–74, with Willink ad locc.; Cicero, Tusculan Disputations IV, 68–76; Pro Sexto Roscio, 66–7; In Pisonem, 46–7, with R. Nisbet ad loc.; Longinus, De Sublimitate 15, 8; Seneca, Phaedra 184– 207; [Seneca], Octavia 554–65. See also Aeschines I, 190–1; Lucretius III, 977–1023; IV, 1037–72; Horace, Ars Poetica 391–6; Padel 1992, 177–8; Dover 1994, 146; Staley 2010, 35; Porter 2016, 406. 44. Pace, too, Gelfert 1995, 7, 23. See F. Nussbaum 2014. 45. Cf. Halliwell 1998, 146–8. 46. Gager, Ulysses Redux 1826 (1994, vol. 2, 136). 47. Steiner 1990, 148; cf. 2008. 48. See Grady 2003, 438–9, 445–6; Dawson and Minton 2008, 31, 45. 49. See Nisbet and Hubbard on Horace, Odes 1, 35, 26; Dunbar on Aristophanes, Birds 134. 50. See Hoxby 2015, 214–30. 51. See here Mueller 1980, at 49–57, 91; Küpper 2014. Pace Rowan Williams 2016, 131–2. 52. Though see Pocock 1973, 255–7. 53. Cf. Burian 1997a, 203–4. Contra Heath 1987, 84–7; Kovacs 1987, 22–38, 55–8. 54. Contra Greenberg 2010, 200. 55. Parish 1993, 18; cf. 24–6. 56. See e.g. Pocock 1973, 214–15; Stierle 2014, 79–82. 57. Biard 1965; Hawcroft 1992, 142; J. Campbell 2005, 45; Greenberg 2010, 124–5; Hoxby 2015, 136–7. 58. See J. Campbell 2005, 126–33. 59. Cf. Schmitt 2015, 237; Zamir 2018, 108 (on Milton’s Satan). 60. This implies that, in cases where a tragic outcome is averted but only just, the audience experiences a similar agony: see my remarks on counterfactual tragedy in §§23 and 44 below. When I say that Woyzeck did not have to murder

180  Tragedy and Moral Redress Marie, a reader objects to me that ‘this is precisely the issue that Büchner is raising. He wrote the play in order to explore just the question to which you think there is a clear and obvious answer’. My view is that Büchner’s play, and tragedy in general, simply takes freedom for granted, as we do in our ordinary daily lives: the question is not whether the tragic protagonist is free, but what he or she should freely do. At least that is one question, but I also think that tragedy is working at a deeper level than at that of the question ‘What should I do?’. I shall say more about what I take that deeper level to be in Chapter 6. 61. See Foster 2004, 2, 32, 166–76, 117. 62. See Cooper 2010, 43–6. 63. See in general Kelly 1993; Grund 2015b, 112. 64. Boethius, Consolatio Philosophiae II, 2 (1978, 182). Also important were the definitions of Diomedes and Donatus-Evanthius: for details see, apart from Kelly 1989, 1993, and 1997; Dewar-Watson 2007, 17–19; 2013, 50; Cooper 2010, 139–40; Hoxby 2015, 58–69. 65. See Kelly 1993, 128, 171; 1997, 50–4; Budra 2000, 44. 66. Quoted from the 1574 edition of the Mirror for Magistrates by Farnham 1970, 298. 67. See Farnham 1970, 69–136, 160–70, 277–303; Ruggiers 1973; Kelly 1993, 170–5; 1997, 1–38, 65–79, 91, 140, 149–75. 68. See Macleod 1983, 143 (on Thucydides’ Athens); Lefèvre 2015, 275–97. Cf. Garnier, Marc Antoine 1196–9; Daniel, Tragedie of Cleopatra 860–1. 69. See Budra 2000, 45. 70. Morgan 2015, 229–34, 367–70, and passim. 71. Kelly 1989, 201; 1993, 173; 1997, 92–130. Cf. Boitani 1989b, 283–9; C. S. Lewis 2013a, 228–36. 72. See Robertson 1961; Shanley 1961; A. Taylor 1979; Kearney and Schraer 1988, 189–90. 73. Pace Brewer 1989, 101. 74. Kelly 1997, 140; cf. 1989, 200. 75. See also Farnham 1970, 419. 76. Loschi, Achilles 620–703 (Grund 2011, 88–94). 77. Compare Chanson 1726, ‘Franceis sunt morz par vostre legerie’ (Brault 1984, 106) with Iliad XXII, 104, ‘νῦν δ᾽ ἐπεὶ ὤλεσα λαὸν ἀτασθαλίηισιν ἐμῆισιν . . .’. 78. Steiner 1996, 540. Cf. 1984, 235–6; 1990, 148–50; 1997, 35; S. Johnson 1968, vol. 1, 66; Lyne and Mukherji 2007, 13–14. 79. Herder 1985–2000, vol. 2, 520; Schlegel 1966, vol. 2, 110–11; cf. Hoxby 2015, 23, 258. 80. Coleridge 1987, vol. 1, 466; Sidney 2002, 112, 1–2; Cooper 2010, 45–6; Pechter 2016, 56–7. 81. Nibelungenlied 985 (Schulze 2015, 288): ‘Dô viel in di bluomen der Kriemhilde man’. 82. Cf. Poole 1987, 133–5, 148–9. 83. See esp. line 783, with Morwood ad loc. 84. Agamemnon 1389–92, tr. Fraenkel, adapted. See Segal 1981, 16; Seidensticker 1995, 160; Pelling 2005, 98; Rutherford 2012, 304–5; Euripides, Phoenician Women 1570–6, with Mastronarde ad loc. 85. Iliad XI, 84–91, tr. Lattimore, adapted. 86. Fraenkel on Agamemnon 59. 87. See Mastronarde on 403, 1051, 1242–6; Bond on Heracles 833; Finglass on Sophocles, Electra 995–6; Rutherford 2012, 321 n. 107. 88. Cf. Virgil, Aeneid XII, 525, with Tarrant ad loc.; Ovid, Metamorphoses I, 689–723; Boitani 1989a, 10. 89. Cf. Auerbach 1946, 421; Gelfert 1995, 114.

From Cognitive Failure to No-Fault Tragedy 181 90. Homer, Iliad XXIV, 599–613, tr. Lattimore, adapted. Note, in connection with our discussion of Polyneices’ burial in Ch. 2, that the penultimate line does not mean that the gods themselves buried Niobe’s children (pace Richardson ad loc.). The gods stopped making stones of the people so that they buried them, following a standard causative idiom, common in the classical and indeed modern languages, according to which ‘I do x’ can mean ‘I get someone else to do x’: cf. Euripides, Helen 1124–5, with Allan ad loc.; Dover on Aristophanes, Frogs 504–11; Dunbar on Birds 474. 91. See Macleod ad loc. Cf. Finglass on Sophocles, Electra 145–52; Honig 2013, 28–9. 92. See Sophocles, Ajax 324, with Finglass ad loc. 93. Cf. Milton, Paradise Lost XII, 613–14; Wordsworth, The Excursion I, 769–70. 94. So too the Strumbo comic subplot of Locrine. The best discussion of the Macbeth porter scene that I know (apart from De Quincey’s classic essay) is Seidensticker 1982, 29–32. 95. See Kelly 1993, 198–9; Grund 2011, xxxvii–xxxviii. Plautus’ Amphitruo called itself a tragicomedy (59, 63), but that was just because it mixed high and low characters: see Christenson 2000 ad locc. and p. 43; DewarWatson 2007, 17; Mastronarde 2010, 59. 96. See Herrick 1955, esp. 249–76, 313; Farnham 1970, 193–4; Hirst 1984, esp. 13, 25, 32–4; Bevington 1998, 1–6; Cooper 2010, 227–34; Mastronarde 2010, 54–62; H. Nisbet 2013, 395, 403–4; Pollard 2015, 426; Potter 2015, 2–6. 97. Morwood on Euripides, Suppliants 692. Cf. Phoenician Women 1149–52, with Mastronarde ad loc. 98. Homer, Iliad XVI, 775–6, tr. Lattimore. 99. See Seidensticker 1982, 14–20, 249–60. 100. See Mueller 1980, 115–28; Lamport 1990, 171; Burian 1997b, 247–8. Cf. Euripides, Cyclops 696, and Aristophanes, Knights 1229–52, with Rutherford 2012, 348–9. 101. See Cooper 2010, 186–7; Burrow 2013, 158–61; Cutrofello 2014, 40. 102. On these cases see Bate 1993, 172–3; Wilson 2004, 115; Shapiro 2005, 367; Nuttall 2007b, 119; Burrow 2013, 122–3, 184. 103. See Schein on 1254–8 and 1293–1307. 104. Lessing 1996, vol. 1, 679. See Herrick 1955, 245; Steinmetz 1987, 113–14; H. Nisbet 2013, 350–5. 105. See Christenson 2000, 37–45, 254–5, and note on line 639. 106. Taplin 1996, 188–9. 107. See Steinmetz 1987, 95, 111–14; Fick 2010, 395–6; H. Nisbet 2013, 492. 108. See Electra 544, 671–83, with the commentaries of Denniston and Cropp ad loc.; Seidensticker 1982, 101–14; Willink 1986, xxii, lvi–lvii; Goff 2000; J. Gregory 2000, 61, and 2005, 269–70 n. 5; Foley 2008, 28–33; Mastronarde 2010, 194–5; Rutherford 2012, 21–2, 61–4. 109. J. Gregory 2005, 267; cf. Seidensticker 1982, 92–100; Allan on 680–747. 110. Cf. Segal 1981, 51; Seidensticker 1982, 116–27; Goldhill 1986a, 262–3; 2006, 157–8; Silk 2004, 247–8; Schiesaro 2003, 136–8 (on Seneca’s Thyestes). 111. Bate 1995, 11–15, 22. 112. Foster 2004, 39–40. 113. Cf. Seidensticker 1978; 2006, 119–20; Taplin 1996, 189–91; Silk 2000, 59. 114. See Zeitlin 2003, 341; Allan 2008, 198–9. 115. Cf. Jenkins 1982, 132–40; Bloom 1999, 422. At least, that is probably the standard understanding of the end of Hamlet; for a darker view of Fortinbras’s succession, see Shapiro 2005, ch. 15.

182  Tragedy and Moral Redress 116. See Sinfield 1992, ch. 5; H. Berger 1997, 70–84; Kastan 1999, ch. 9, esp. 174; 2003, 16–19; Orgel 2002, 172; Clark and Mason 2015, 34, 81–2; Shapiro 2015, 231–2; Kottman 2016, 13. 117. Buchanan, Jephtha 656–7, tr. Sharratt and Walsh, adapted. 118. Griffin 1998, 54. Cf. Kastan 2003, 14, for a similar point about the end of Romeo and Juliet. 119. Steiner 1961, 8; cf. 128; 1984, 276–7; 1990, 155. Cf. Goldmann 1970, 14–15; Gelfert 1995, 136. 120. Easterling 1997a, 171; cf. Burian 1997b, 247. 121. See Hall 2009, 94–5. 122. Anyone who is inclined to agree with Steiner about ‘the unaltering bias toward inhumanity and destruction in the drift of the world’ should read Pinker 2011. 123. Such as we find in e.g. Holbrook 2015 and Ryan 2015. 124. See Foster 2004, chs. 3 and 4. 125. See here Silk 2000, 80–2. 126. Gellrich 1988, 149–55. Cf. Winnington-Ingram 1983, 78–131; Goldhill 1984a, 279–83; Rocco 1997, ch. 5. See also J. Campbell 2005, 145–50. 127. As Taplin does: 1996, 197–8. 128. Cf. Rowan Williams 2016, 118, 157–8. 129. On this conceit, see Fraenkel on Aeschylus, Agamemnon 105f.; Griffith on [Aeschylus], Prometheus Vinctus 981; Dawe on Sophocles, Oedipus Rex 1082–3. Cf. Zeitlin 1986, 138–41. 130. See e.g.  Poetics 1451a11–15, 1452a31–2, 1452b2–3. 131. Cavell 2003, 55. Cf. Gaskin 2013a, 112–17. 132. Sophocles, Ajax 693–718; Euripides, Trojan Women 703–5, with Lee ad loc. 133. See e.g. R. Adams 1981. 134. So Lamport 1990, 169, 174. 135. Goethe 1985–99, vol. 7, 199, 539. 136. But see Lamport 1990, 151–2. 137. Rowan Williams 2016, 124; cf. Eagleton 2015, 36–7. 138. Vickers 1973, 83; cf. 68. But for a different reading of Hecuba’s ‘degeneration’, see J. Gregory 1997, 107–11. 139. See Finglass 2007, 510–12, 525–8, 539–40; Allan 2001, 28–9, 206–7; Ovid, Metamorphoses VI, 642–5, 655–60; Correr, Procne 934–6 (Grund 2011, 178). 140. J. Newman 1994; cf. Kahn 1997, 63; Belsey 2001, 335; Pucci 2016, 26–8 (on Medea). 141. See Allan and Kelly 2013, 103 n. 92. On Hecuba’s speaking second, see Lloyd 1984, 304, and Lee on 912–13; cf. Morwood on Suppliants 566–80. Linguistic compensation will be the theme of Part II. 142. Artaud, Les Cenci II, 1 (2011, 73). 143. Cf. Gossip 1981, 159–60. 144. On this play see Walker 1988, 82–4; Auchincloss 1996, 67–72; Hoxby 2015, 234–40. 145. For a different reading, see Auchincloss 1996, 19–20. 146. See Barnwell 1982, 228, 238–43, 250; Auchincloss 1996, 27; Oster 2014, 252–61. 147. Gervinus 1872, vol. 2, 154, 167–8. Cf. Poole 1987, 45; H. Berger 1997, 84–97; Dusinberre 2003, 63. 148. F. Lucas’s comparison between Hamlet and Brutus (1957, 119–20) ignores this point. 149. So Greenberg 2010, 200. 150. Lefèvre 2015, 285–6. 151. Fraenkel 1978, vol. 2, 96–9.

From Cognitive Failure to No-Fault Tragedy 183 52. Winnington-Ingram 1983, 99. 1 153. For proponents of this interpretation, see Fraenkel 1978, vol. 2, 97. 154. Cf. also Euripides, Helen 1139–43, with Allan ad loc.; L. Parker on Iphigeneia in Tauris 1414–19. 155. So Winnington-Ingram 1983, 99. 156. See N. Hammond 1973, 402–3; Dover 1987, 144–6; Lloyd-Jones 1990, 285–7; 1991, ch. 24; Peradotto 2007, 216–26. 157. Cf. Hammond 2009, 44; Scodel 2009, 184. Contra Fraenkel 1978, vol. 2, 97 n. 3. 158. Cf. Peradotto 2007, 226–7. 159. We get a similar moral oddity in the motivation of Guarini’s Il Pastor Fido (I, 2; 1999, 94–5). 160. See Gaskin 1998b and 2000. 161. Peradotto 2007, 235–7. 162. See de Mourgues 1967, 114. 163. See e.g. Garvie on Aeschylus, Choephori 650–2. 164. Cf. Padel 1992, 118, 178–9. 165. See here Seidensticker 1995, 156–8; cf. Hoxby 2015, 243. 166. See Macleod 1983, 38. 167. See Buchanan, Jephtha, 508–9, 557–9, 784–841, 1223–31, 1264–8, 1302, 1318, 1328–9, 1348; Servius on Aeneid III, 121 and XI, 264; Shuger 1994, ch. 4; Nyquist 2008; L. Parker 2016, xxii. 168. See Seneca, Thyestes 1100. Cf. Seneca, Medea 932–6; Panoussi 2005, 423–5 (on Polyxena). 169. See [Seneca], Octavia 914–23, 961–2, 972–82, with A. Boyle ad locc. 170. Livy, I, 57–9; Ovid, Fasti II, 721–852; cf. Donaldson 1982, 11–12. 171. Ovid, Metamorphoses VI, 424–674; Correr, Procne, esp. 911–17, 949–51 (Grund 2011, 176, 178). 172. Livy, III, 44–58; Happé 1972, 271–317. 173. See Miola 1992, 28–30. 174. On which see G. Taylor 2002. 175. Heller 2002, 29; contra Rackin 2005, 134–5. 176. Cf. M. Nussbaum 1992, 156; Foley 2001, 117–18. 177. Racine 1982–3, vol. 2, 202 (tr. Cairncross). Cf. J. Campbell 2005, 148. 178. See Griffith 1983, 8, 12. 179. See R. Levin 1981; Donaldson 1982, esp. 46, 52–4; Miola 1992, 18–41. 180. Pace Maus 1986, 67; cf. Kahn 1997, 36–7. 181. See Melchiori 1998, 36–9. 182. Pace Zamir 2016, 75. 183. See Fraenkel on Agamemnon 1330 and 1444f. 184. Poole 1987, 179. Cf. J. Gregory on Euripides, Hecuba 48. 185. Ajax 368, 485, 489, 585; Easterling 1984, 3; Blundell 1989, 74; Burian 2012, 75. Cf. Horace, Odes II, 4, 15–16. 186. See Blundell 1989, 74–5; Foley 2001, 90–2; Burian 2012, 75–6; J. Gregory on Euripides, Hecuba 834. On Tecmessa, see further Kirkwood 1994, 103–7; Easterling 1997b, 25–6; J. Gregory 1997, 106–7; Finglass on Ajax 263–81, 288–94, 368, 485–524, 485–6, 496–9, 500–4, 504–5, 514–19, 879–973, 972–3. 187. Winnington-Ingram 1980, 30–1 with n. 57. 188. See Trojan Women 343–7, with Lee ad loc. 189. See Stevens on Andromache 576. 190. See A. Boyle 1986, 64. 191. Aristotle, NE 1139a20. 192. See A. Boyle 1986, 59; Thomas on Georgics III, 60, 66–8, 96, 102, 125, 128, 165, 215–41.

184  Tragedy and Moral Redress 193. So e.g. Cassandra at Agamemnon 1066–7; Segal 1981, 129–31; Rutherford 2012, 132–3. 194. Servius 1881–1902, vol. 1, 459; Kelly 1993, 29–30. 195. See von Albrecht 1970; Wlosok 1990, 320–43; Gaskin 1992; Hardie 1997; Lefèvre 2015, 201–19. 196. On Turne’s tragic character see Kantor’s introduction to Prévost 1985, 11–17. 197. See Aeschylus, Seven Against Thebes 321–68, with Hutchinson ad loc.; Euripides, Trojan Women 527, 557–9, with Lee ad locc.; O’Sullivan 2009, 451–4; Hejduk 2013, 154. 198. See Hardie 1997, 320; cf. Aeschylus, Agamemnon 326–9 with Fraenkel ad loc. 199. Mastronarde 2002, 333, and note on 1053. 200. See Dale on Euripides, Alcestis 393; Stevens on Andromache 27, 56–90, 504ff.; L. Parker 2007, 131–2; A. Boyle on Seneca, Thyestes 716–43. 201. She leaves the stage at 989, but returns at 1168: see Finglass on 972–3. 202. See Santirocco 1980, 194. 203. McClure 1999, 19–24. 204. See Griffith 2001, 123–4. 205. See Aristotle, NE 1100a5–9, 1101a6–8; cf. Boethius, Consolatio Philosophiae IV, pr. 3. 206. See Dusinberre 2003, 30–1. 207. Cf. Marston, The Tragedy of Sophonisba II, 1, 145–8; III, 1, 15. 208. See Maus 1986, 69–71; Belsey 2001, 331–2. 209. Cf. Donaldson 1982, 7, 30–1, 69–70; Seneca, Phaedra 892, with Coffey and Mayer ad loc. 210. See Phillips’s introduction to Chaucer 1997, 13–16; C. S. Lewis 2013a, 208–11. 211. Lamport 1981a, 67–9; 1990, 17; Ter-Nedden 1986, 13–43, 87–98. 212. See Ter-Nedden 1986, 29, 45, 51, 59, 88, 92–9, 103–7, 110–11. 213. See H. Nisbet 2013, 197–8, 204–5. 214. Cf. H. Nisbet 2013, 201. 215. Pace H. Nisbet 2013, 197, 217. 216. Fick 2010, 158. 217. H. Nisbet 2013, 495, 503. 218. See Farnham 1970, 251–2. 219. Emilia Galotti V, 7 (Lessing 1996, vol. 2, 202). See further Hatfield 1956, 294. 220. Cf. G. Wells 1984; J.-D. Müller 1993, 78–9; Fick 2010, 388–90; Greiner 2012, 328. 221. Fick 2010, 398. 222. Lamport 1981a, 179. 223. See Hatfield 1956, 290, 296; Lamport 1981a, 167–8; Pütz 1986, 177; TerNedden 1986, 164, 200; Steinmetz 1987, 98–105; Fick 2010, 398; Greiner 2012, 321–34; H. Nisbet 2013, 498–9. 224. Cf. G. Wells 1984, 163, 172. 225. H. Nisbet 2013, 497. 226. Aquinas, Summa Theologica IIa–IIae q. 64 a. 5 ad 3 (1894 vol. 3, 476); see Walker 1988, 79. 227. See e.g. the versions by Trissino (1975) and Marston (1986); there are many others. Cf. Burian 1997b, 231–2; Hoxby 2015, 122–6; Kragelund 2016, 379–90. 228. See here Ter-Nedden 2007, 323. This article contains an interesting application of the prisoner’s dilemma to the Philotas. I do not agree with Ter-­ Nedden, however, that the position Philotas is in forces suicide on him

From Cognitive Failure to No-Fault Tragedy 185 (337–8), so that his suicide is the ‘fault’ of the situation, rather than his own: on the contrary, Philotas is offered a co-operative strategy, which is a recognized way of solving prisoner’s-dilemma-type situations, and his refusal of it is culpable. 229. De Civitate Dei I, 19. See Donaldson 1982, 29; Maus 1986, 69; Miola 1992, 36–7; Hendricks 2000, 110–11; Duncan-Jones and Woudhuysen 2007, 44; Garvie 2016, 14. 230. See Lamport 1981a, 98–105; Pütz 1986, 100–17. 231. See Ter-Nedden 1986, 139; 2007, 336; Fick 2010, 184–90; H. Nisbet 2013, 240–4. 232. Lamport 1981a, 167–8; Ter-Nedden 2007, 365; H. Nisbet 2013, 299, 496. 233. See Lamport 1981b, 15–16; Steinmetz 1987, 131; Gelfert 1995, 88–91; Ter-Nedden 1986, 165; 2007, 365–6; Fick 2010, 393–8; Greiner 2012, 306–10; H. Nisbet 2013, 235–47. 234. Cf. Hoxby 2015, 231–2. 235. Lamport 1990, 17. 236. Pace Lamport 1981a, 172; cf. 176. 237. Ter-Nedden 1986, 110–11, 230–4. 238. See Lloyd-Jones 1990, 197–8, on Eteocles and Agamemnon. 239. See Garvie 1986, 174, 193–4. 240. Pace H. Nisbet 2013, 397, 503. 241. Hamburgische Dramaturgie, §82 (1996, vol. 4, 613). 242. See Korzeniewski 2003, 424–6; H. Nisbet 2013, 109–10, 174–81, 355, 580, 612–13. 243. Hamburgische Dramaturgie, §79 (1996, vol. 4, 598–9). 244. Letter to Nicolai, 21 Jan. 1758 (1996, vol. 2, 702). 245. See Donaldson 1982, ch. 4, for an excellent discussion. Emilia is based on Clarissa as well as on Virginia: H. Nisbet 2013, 487. 246. See Steiner 1961, 346–9. 247. Cf. Greiner 2012, 434–5. 248. So e.g. Burian 1997a, 181; Silk 2004, 246. 249. Kafka 1987, 7; cf. Krieger 1960, 12–13. 250. Extraordinarily, he is sometimes so misinterpreted: see e.g. Grady 2009, 133–5. 251. Kottmann 2016, 11; S. Johnson 1958–77, vol. 7, 71; Boswell 2008, 539.

Part II

Tragedy and Linguistic Redress

4 Theoretical Considerations

26 Grammar, Concepts, and Disagreement At the end of the last chapter I took issue with Sebastian Gardner’s view that tragedy represents loss and suffering as having ‘unconditional and hence uncompensatable reality’, and that ‘tragedy contests the supremacy of morality by intimating a more fundamental kind of value which is indifferent to moral concerns and which cannot be brought into coherent relation with moral value without ascending to a speculative philosophical level’ (2003, 235, 245), though I conceded to Gardner that his analysis does at least appear to fit representations of undeserved suffering in tragic literature. But the burden of my argument in Part I of this book was that, while the Western tradition certainly does offer examples of such representations, tragic literature prefers to represent its protagonists as deserving their fate. The key examples so far—Oedipus and Antigone, with others playing supporting roles—have been advanced to illustrate my claim that moral innocence is somewhat thinner on the ground in the Western tragic tradition than has been generally supposed; for these figures have been widely held to present cases par excellence of absolute innocence. The hamartiai of tragic heroes and heroines are usually of a cognitive nature; but the cognitive should not be conceptually separated from the moral, for we have intellectual as well as practical obligations, and the two sorts of obligation are indeed intimately intertwined. Still, as I have agreed, there are depictions of undeserved suffering in tragic literature, and it will be a large part of my aim in this part to argue that, while this kind of suffering cannot, by definition, be compensated morally, there may be other ways in which undeserved (and indeed deserved) suffering can be compensated. In particular, there is a sense in which—as I suggested in connection with the figure of Hecuba in Euripides’ Trojan Women (§24)—suffering is linguistically compensatable. I will come to the core of this idea in Chapters 5 and 6. In this chapter I wish to lay some groundwork for those later reflections, and also provide a theoretical underpinning for a number of the more significant assumptions of the study as a whole. Both these tasks can, I think, be discharged at the

190  Tragedy and Linguistic Redress same time and are appropriately taken here, involving as they do similar, mainly language-theoretical, considerations. I start with a simple point, but one that is of fundamental importance. The mere fact that tragic literature is or was written in a public language shows that it is not beyond the reach of the ordinary functioning of that language. Some tragic literature is written to be performed on stage, or to be recited before an audience, though not everything so written is actually performed; some tragic literature is not written for performance of any kind, whether because, though written in dramatic form, that form is just a rhetorical device, or because it is not written in that form, but rather as, say, prose narrative. Because of the current dominance of performance studies we tend to forget that the consumption of tragedy through individual reading has a long tradition; in fact probably most engagement with tragic literature, including literature written for the stage, has taken that form.1 I am inclined to agree with a number of recent commentators that the evidence favours the view that Seneca’s tragedies were written to be staged, rather than merely recited or privately read; we also know that they have been frequently and successfully staged from the Renaissance onwards.2 Nevertheless, they have more often been privately read. In the Poetics Aristotle made a significant place for the reading of tragedy (1462a11–18). Recent writers sometimes disparage the critical value of reading tragic drama as opposed to attending a performance,3 even though it is the way in which they themselves approach tragic literature most of the time. But the key point here is that all tragedy, whether accessed via performance or private reading, is linguistic; it is all (or virtually all: I shall come to possible exceptions in the next chapter) constructed from meaningful words, put into grammatical sentences. These simple facts have, I shall argue, profound consequences, including ones that used to be accepted as platitudes but have been ignored or rejected by a number of recent critics, who often miss the real implications of tragedy’s linguistic nature, and of its status as part of the ordinary public language. It is even not uncommon to find the grammaticality of tragic language denied. For example, Page duBois suggests that the communication between Creon and Antigone is vitiated by ‘contiguity disorder’, according to which ‘the syntactical rules organizing words into higher units are lost; this loss . . . causes the degeneration of the sentence into a mere “word heap” ’ (1986, 378; she is quoting Roman Jakobson). Again, Paul Hammond holds that ‘In tragedy the expected grammar—the sequence of cause and effect, the distinction between subject and object, the link between agent and verb—is persistently thwarted’ (2009, 29). One only needs to consider such claims for a moment in order to see that they are egregiously false. The Antigone, like the vast majority of works of literature, is composed of grammatical sentences. There is no syntactic breakdown of communication between Antigone and Creon such as is

Theoretical Considerations 191 posited by duBois: Creon is not, as she says at one point (1986, 379), ‘unable to control syntax’; at least he is not so unable if by that is meant that he is unable to get himself, try as he might, to speak syntactically correct Greek sentences. (There is a larger sense in which he is unable to control—determine—the syntax of Greek; but then that goes for every individual speaker of a public language.) And generally, tragedy—any genre of literature—upsets none of the three things Hammond mentions, at least not ‘persistently’. Concerning the sequence of cause and effect, one of the main claims of Part I was that that sequence—and especially agent causality—is firmly in place in some of the principal tragedies of the Western tradition, which depend for their impact on the comprehensibility of the sequences of events—and in particular the sequence of hamartia and catastrophe—that they depict. In that sense most tragedy, I argued, is Aristotelian. As for the rest of Hammond’s assertion, subject, object, agent, and verb are in this context all grammatical categories and, as you would expect, they figure in tragic and all other literature in an entirely standard and predictable way. This is so whether we use the word ‘grammar’ in the strict sense or in a certain extended sense. In the strict sense the vast majority of the sentences found in tragic literature, as we have noted, obey the grammatical rules of the relevant language, which are accordingly in no sense thwarted. There is perhaps an extended sense in which we can talk of mental states, or situations, as having agents or objects, say, but in that sense too it is hard to see how tragedy, or any other kind of literature, ‘thwarts’ the relevant connections. We may gather more exactly what Hammond has in mind by looking at his continuation of the passage just quoted (ibid., 29): Sometimes tragic figures thwart connections by the way in which they use language not as a means to communicate with others but as a means of isolating themselves into a world which they can control. Leontes and Othello both build narratives of their wives’ adultery, and in so doing destroy the relationship between language and the external world, as every principal signifier points to the same signified, which is always adultery. Hammond then supplies examples from The Winter’s Tale of Leontes’ twisting of his interlocutors’ words, innocently intended, so that they bear a lascivious double entendre, suggesting infidelity on Hermione’s part. He concludes that ‘Leontes destroys the referential capacity of language as each signifier signifies his wife’s transgression’ (ibid., 30). Now it is true that tragic agents (not to mention ordinary people) sometimes use language not as a means of communication but as a way of isolating themselves (or for some other self-regarding purpose). But in what way does that use of language derogate from its referentiality? How do the narratives that Leontes and Othello construct for themselves undermine

192  Tragedy and Linguistic Redress referentiality? The innuendos that Leontes plays on are extra meanings: they do not subvert, or replace, the literal meanings of the words in question. David Christenson suggests that ‘Language in Plautus’ hands becomes unrealistically prolix, and profuse alliteration, assonance and other kinds of parallelism combine to create discourse that is largely nonreferential’ (2000, 12). But prolixity does not undermine reference, if that is what Christenson wishes to imply; it precisely depends on it. If we did not know what the words meant, we could not know that the speaker was being prolix. Prolixity, like innuendo, gives us more meaning, not less: in both cases, but for different reasons, the superfluity may be unwanted. Now innuendo, as a use of language, has nothing essentially to do with tragedy. It is found in tragedy as in many other places, but the coincidence is an adventitious one. Perhaps the most familiar form that it takes in tragic literature is so-called dramatic or Sophoclean irony, which exploits the fact that the words which a tragic character utters may have one meaning for him or her, and another meaning for the audience.4 But that further meaning, where it is present, does not destroy the literal meaning relied on by the speaker; it is an extra layer of significance which the audience may pick up, alerting them perhaps to future events, but not cancelling the ordinary meanings of the relevant words. Hammond gives a further example of the sense in which he thinks tragedy subverts referentiality in a discussion of a line taken from Seneca’s Thyestes: ‘When Atreus says fas est in illo quidquid in fratre est nefas [220] (“All that is wrong in dealing with a brother is right in dealing with him”), this almost palindromic sentence shows that he assumes the capacity to reverse the meaning of fas, which the translation cannot quite bring out’.5 But this claim unravels in a rather obvious way. If Atreus were reversing the meaning of ‘fas’, his sentence would mean ‘All that is wrong in dealing with a brother is wrong in dealing with him’—a tautology that is entirely without interest. Naturally Atreus does not intend that jejune thought, but the more radical and shocking thought that all that is wrong in dealing with a brother is right in dealing with him. That is Hammond’s translation of the Latin line, and, so far from failing to bring out its real thought, it captures it exactly. Atreus would miss his bold thought if ‘fas’ did not go on meaning, on his lips, exactly what it usually means, namely right. Similarly, if you say, as Hardy’s Beck Knibbs does in Tess of the D’Urbervilles, ‘All’s fair in love and war’ (1967, 207), it would be mistaken for me to register my disapproval by responding ‘If you say that you are reversing the meaning of “fair” ’. The entire point of that old saw—the basis of its capacity to unsettle us—would be lost if the word ‘fair’, and all the sentence’s other component words, did not go on meaning just what they ordinarily mean. The shock value of Atreus’ line comes not from its doing something strange with meaning: exactly not. It comes from its being unsettlingly false, and it can only be false in

Theoretical Considerations 193 the required way if the utterer does not tamper with the meanings of the component words. So when Hammond asks, concerning the Senecan line quoted above, ‘In such a milieu, what could fas and nefas mean?’ (2009, 117), the answer is that they mean what they usually mean. The Senecan example does not show how ‘Linguistic and conceptual boundaries are troubled, “solicited” in Derrida’s sense of the term’ (ibid.): and that is a good thing from the point of view of achieving the tragedian’s purpose because, to repeat, if Atreus’ line did show linguistic and conceptual boundaries being troubled, that would diminish its power to shock. Seneca is heavily dependent on Ovid, and one source for Thyestes’ line is a passage in the Metamorphoses (VI, 585–6) where Procne, having learnt of her husband Tereus’ treatment of her sister Philomela, fasque nefasque confusura ruit, ‘rushes to confound right and wrong’, by murdering their son Itys and serving his flesh to Tereus cooked, as in the Thyestes story.6 In what sense is Procne confounding right and wrong? Again, the meanings of the words are not confounded: rather, Procne performs an act which in one sense is right (it is justified revenge on Tereus) and in another wrong (it is murder of their child). We have the same structure as was noted in §24 in connection with the house of Atreus, a pattern of crime and punishment that is usually described in linear terms, according to which the punishment of a crime is itself a crime calling for punishment, but which may also be considered in circular terms, according to which a crime is its own punishment. (But in Ovid’s ebullient narrative the cycle is short-circuited by metamorphosis, as Procne is changed into a swallow, Philomela into a nightingale, and Tereus into a hoopoe.) To take another example, when Aeschylus’ Clytemnestra boasts of her fidelity on her husband’s return from Troy (Agamemnon 611–12), Simon Goldhill argues that ‘The model of language suggested by this phrase implies a gap at the heart of communication between signifiers and what they signify, as if in a message form could be separated from content’ (1986a, 8–9). But the message is not separated from its content; it has its normal content. It is just that that content is false. There is no ‘gap’ between signifier and signified in deceit. Deceit does not undermine the fixity of meaning; on the contrary, it precisely depends on it. (Elsewhere Goldhill suggests that Clytemnestra’s use of language in the carpet scene is both deceitful and indeterminate in meaning; as John Moles points out, you cannot have both.)7 Consistently enough, Hammond claims that the witches of Macbeth undermine the meaning of ‘fair’ in their maxim ‘fair is foul and foul is fair’ (2009, 127). Again, the reply is that the very paradoxicality of the witches’ utterance precisely depends on its words’ retaining their usual sense. And when Seneca’s Atreus wonders which of Thyestes’ sons he will sacrifice first, and Seneca comments ‘nec interest—sed dubitat et saevum scelus / iuvat ordinare’ (715–16), Hammond translates as ‘It makes

194  Tragedy and Linguistic Redress no difference, yet he hesitates and takes pleasure in ordering the savage crime’, adding: the crucial word here is ordinare: ‘to set out in due order’. What could be due or orderly about such a perversion of sacrifice into— . . . the antithesis hangs incomplete, for there seems to be no term which could describe a sacrifice when it is performed with such deliberate intent to destroy. The actual formula of civilization is employed specifically to make the atrocity more awful.8 Of course, all sacrifices are performed with a ‘deliberate intent to destroy’—nothing unusual there. Still, Hammond is right that there is a peculiar horror in the use of the civilized word ‘ordinare’—more at home, one might think, in the worlds of farming and commerce—to apply to such a barbaric act as Atreus is preparing. But the point here is that Seneca’s use of ‘ordinare’ could only ‘make the atrocity more awful’, as Hammond insightfully observes that it does, if ‘ordinare’ were being deployed in its workaday sense. The very horror of the use of ‘ordinare’ in this context depends on its being such a humdrum word denoting what is at its most basic level a humdrum activity—namely putting one thing before another. Putting one thing before another is exactly what Atreus is doing; but that means, contrary to what Hammond here says, that nothing is unexpressed, or under-expressed, by Seneca. There is no problem finding ‘a term which could describe a sacrifice when it is performed with such deliberate intent to destroy’: Hammond himself, in telling us that there is no such term, expresses what is going on entirely satisfactorily. Atreus commits a deliberate act of murder. The mundane word ‘ordinare’ is used to describe an act which is in reality hideous and felonious: that is why it is so effective, for we feel the gap between the banality of the word’s meaning and what, at another level of description, is actually going on. But this gap is not a gap between language and world: it is a gap between some bits of language and other bits of language—between the humdrum, neutral language that Seneca uses to describe Atreus’ deed, as though there were nothing more to say, and the morally charged language which Seneca does not provide but which the audience inevitably entertains, and which also gives a true and accurate description of it. I stress the word ‘also’, because it is not as though ‘ordinare’ gave a false description of Atreus’ action: he is indeed sacrificing one boy before another. It is just that this fact, which the verb ‘ordinare’ perfectly conveys, is fraught with horror. There is a recollection, perhaps, of the choice of Intaphernes’ wife, described by Herodotus in a passage that probably formed the basis of Antigone’s controversial defence of her actions at lines 904–15 of Sophocles’ play.9 For we naturally feel that ‘ordinare’ implies some principle of ordering—that, in putting one act of murder before the other, Atreus does so on some rational basis. But there

Theoretical Considerations 195 is no such basis: the choice is an arbitrary one, and when the matter concerns life and death the juxtaposition of that arbitrariness with the moral weight of the action is peculiarly horrible. ‘Justice in King Lear is a shifting concept according to what authority is in power’, says R. A. Foakes; ‘A word such as “justice” may mean such different things to different people that their differences can only be settled by death’, writes Adrian Poole.10 As I hope the foregoing discussion has shown, we have to be cautious with such assertions. Let us consider the matter. What some protagonists of tragedy—Antigone and Creon, as it might be—disagree about is what counts as, say, justice, here and now. The first point to note is that this is not a merely verbal disagreement. If all they disagreed about were the application of the word ‘justice’, they could settle their difference by relativization: ‘you are talking about justice-for-Creon; I am talking about justice-for-Antigone: these are different concepts, so we are not really disagreeing’. But that is not how it is between Antigone and Creon.11 They mean to be talking about the same thing, the concept justice, and they are disagreeing about how that concept bears on the present situation. So they agree on the meaning of the word: it means the concept justice. If they differed about which concept were in question, and if they were in fact talking about two concepts, there would be no disagreement, or at any rate no disagreement that could not be resolved by the relativizing move mentioned above; before that relativization, they would simply be talking past one another. But the concept is there as the meaning of the word ‘justice’, and that is what they both have in their sights; their disagreement concerns its extension— that is, its range of application. They agree about the concept, and the disagreement concerns which of Creon’s decree and Antigone’s disobedient act is a case of justice, given that they cannot both be. There is an important proviso that has to be registered here, concerning my distinction between disagreement about a concept, on the one hand, and disagreement about its extension, on the other. Clearly the two things cannot be kept entirely separate. For example, it would not be possible for two people to share a concept while disagreeing entirely, or even substantially, about its extension: if they disagree sufficiently about its extension—that is, if they disagree sufficiently about which objects the concept is true of—then they cannot be said to share that concept. If, to take an extreme example, two speakers agree on none of the objects that merit the appellation ‘red’, then they associate different meanings with that word, and so do not share the concept red—at least, not in virtue of using that word. (Note here, by way of terminological book-keeping, that I think of a concept, following Frege, as the meaning—in particular, the referent—of a concept-expression.)12 It would be absurd to say, in the situation I have just described, that the two speakers agreed about the concept red but disagreed merely about its extension. If disagreement about the application of a concept-expression becomes too widespread

196  Tragedy and Linguistic Redress or radical, then communication—the sharing of meanings—is undermined. To put it the other way round: there has to be a wide measure of agreement in the way two people apply a given term for them to count as sharing the concept to which it refers, and so as communicating with the relevant concept-expression.13 If disagreement becomes too radical then communication is undermined, in which case disagreement itself becomes impossible, because it presupposes communication. So disagreement about the extension of a concept (the correct application of a concept-expression) is an essentially secondary phenomenon; it depends on a prior, and indeed considerable, measure of linguistic and conceptual agreement. Once that prior agreement is in place, then disagreement ‘around the edges’ of the concept’s application becomes a theoretical possibility, but not otherwise. This is a point that Donald Davidson, following Wittgenstein, has made familiar, and it is encapsulated in his so-called ‘principle of charity’ governing interpersonal linguistic communication.14 As Davidson puts it: ‘widespread agreement is the only possible background against which disputes and mistakes can be interpreted. Making sense of the utterances and behaviour of others, even their most aberrant behaviour, requires us to find a great deal of reason and truth in them’ (1984, 153). (Here Davidson should have written not ‘even’, but ‘especially’.) It follows that, if Antigone and Creon are to disagree about whether it is just to bury Polyneices, and just for her to bury him (recall that these issues are distinct), they must largely agree on the application of that concept, the concept justice. Hence for differences to arise over the concept justice (Poole), or for us to recognize that it is indeed justice that is undergoing a shift (Foakes)—where what is meant in these cases is that there is a disagreement or shift between speakers concerning the application of the word ‘justice’—there must be a high degree of commonality of word meaning, in particular of the word ‘justice’, or its equivalent in the relevant language. As it is often put, ethical (or any) disagreement can only take place against a background of shared beliefs and values. With these points in mind, consider again Foakes’s assertion that ‘Justice in King Lear is a shifting concept according to what authority is in power’. As the context makes clear, Foakes is thinking of the scene in which Lear excoriates the official representatives of justice in civil society. His charge against the Law’s agents is one of hypocrisy (IV, 6, 156–63): Thou, rascal beadle, hold thy bloody hand; Why dost thou lash that whore? Strip thine own back, Thou hotly lusts to use her in that kind For which thou whipp’st her. The usurer hangs the cozener. Through tattered clothes great vices do appear; Robes and furred gowns hide all. Plate sin with gold, And the strong lance of justice hurtless breaks; Arm it in rags, a pigmy’s straw does pierce it.15

Theoretical Considerations 197 And now the point against Foakes is this: for the accusation of hypocrisy to stick, the concept of justice must remain fixed. If that concept shifted according to who was in authority, Lear’s criticisms would hurtless break. If the usurer can define what justice is, then he will be entitled to hang the cozener. That will simply be justice—or at least justice for him, and no one will have the right to challenge or upset that relativization. But Lear’s point is that this is not how it is: the usurer’s conduct can and should be assessed against a fixed, objective concept of justice that he is powerless to change or tailor to his own actions, a concept of justice that is the same for him and everyone else. And when that standard is applied, his actions are found to be base. Barbaric regimes do not change the concept of justice, and that is why we are allowed to censure them; they simply do unjust things. As I have just implied, the point I am making has nothing especially to do with literature as opposed to life. George Steiner writes: ‘It is now thought, in sober analysis, that approximately 100,000 men, women, and children were buried alive—should one write, let alone try to attach concrete meaning to such a sentence?—in the killing-fields of Cambodia by the Khmer Rouge’ (1997, 106–7). I suggest that the answer to Steiner’s parenthetical question is: yes, one should write that sentence, at least if it is true, and attach concrete meaning to it. For it is only if we do attach concrete meaning to that sentence—and, in particular, attach its ordinary meaning to it—that we will know what we are talking about and be warranted in reacting as we do. The fact from which Steiner recoils is indeed a monstrous one, and his hesitancy is understandable: but the horror before which we bow our heads when we are brought up against such realities as the one he mentions is a horror of the truth; but a truth is expressed by a declarative sentence, and for such a sentence to be true it must have a definite meaning. No one’s interests are served—certainly not the victims’ interests, but not the perpetrators’ interests either, at least not their real interests, which as Socrates maintained cannot diverge from justice and truth—by supposing that Steiner’s sentence lacks a determinate meaning, or is even too appalling to be written. Uttering such a sentence is a necessary step towards addressing the impulses that have made it true.

27 Historicism and Semantic Objectivity So far in this book I have been working from an assumed critical position which has two important features that I need now to make explicit and defend. In the first place, my approach to literature is objectivist about meaning, in the sense that the words and sentences of literary works are taken to have objective meanings which are shared between communicators, and which are also able to remain relatively stable across space and time, and so to be accessible to audiences other than their original ones; thus I have not doubted that we can translate and understand Sophocles

198  Tragedy and Linguistic Redress and Seneca, for example. In the second place, my position is transhistoricist and humanist, in its presupposition of a relatively fixed human essence over time: that was presupposed to my discussions of the characters of Oedipus and Antigone and Emilia Galotti, among others. Though these positions were once taken for granted, they have been strongly challenged in the literary theory and criticism of the last few decades. In particular, they set me at variance with the (largely American) doctrine of new historicism, together with its (largely British) congener cultural materialism, which are ideologically diverse movements, but which are marked, in what one might think of as their central manifestations, by (i) the assertion that the critic’s own historical embeddedness makes objective interpretation of older works impossible, and (ii) the denial that there is a fixed human essence that is constant across cultures and over time. Both (i) and (ii) are dominant features of current academic discussion of older literature.16 In connection with (i), my approach also sets me in opposition to a deconstructionist view, according to which meaning is constantly shifting, always deferred, indeterminate, unstable. In the ensuing discussion I shall examine these various oppositions. Taking the first strand of new historicism first, we find that the idea that an interpreter’s own historical embeddedness makes objective interpretation of older works of literature impossible is widespread; Louis Montrose, for example, writes that ‘a historical criticism that seeks to recover meanings that are in any final or absolute sense authentic, correct, and complete is in pursuit of an illusion’.17 But why? A physicist is physically embedded in the world: does that mean that objective physical science is impossible?18 Or consider the phenomenon of perception. Perception is a physical process: if I see someone walking across a field, there occurs an elaborate physical process, which we are now well able to describe, in which I the perceiver am thoroughly embedded. Does that mean that objective, perception-based knowledge—my knowledge, based on perception, that there is someone over there walking over that field—becomes impossible? Plainly it does not. The fact that perspective is involved in my perceptual act does not fight against objectivity, as Nietzsche thought:19 quite the contrary, for a perspective on the world is precisely a perspective on the world.20 As Simon Blackburn puts it (2014, 198), ‘it is a mistake (one to which Nietzsche was peculiarly prone) to run together the idea of perspective and the idea of illusion’. Perspectivalism implies not irrealism, but realism. In fact, when one looks at the practice of new historicists and other theorists, one always finds that they presuppose that authentic, objective historical knowledge is possible. New historicist essays, as has often been noted, frequently start with a historical anecdote, usually involving some shocking act of violence, before treating the account as background and context for literary interpretation;21 and we are meant to regard the historical context as fact.22 Self-scrutiny, in the practice of the new historicists, appears to have its limits.23

Theoretical Considerations 199 I have argued at length against the deconstructionist account of meaning elsewhere;24 but there is more to be said in the present context. In the first place, we may note that a set of Davidsonian considerations, connected with those rehearsed in the previous section, tells strongly against any doctrine of untranslatability of language from culture to culture, or across time.25 Our sheer ability to recognize something as a language presupposes a great deal: it presupposes, for a start, that we are able, at least in principle, to translate the thing. Language is inherently meaningful, but meaning is, as Wittgenstein taught us, publicly accessible, and so must be available from more than one point of view: if we could not, at least in principle, translate something, we could not recognize it as a language in the first place. (There may be practical difficulties in the way of translating a given language: for example, there may not be enough evidence, as is the case with Linear A. But our ability nevertheless to recognize Linear A as a language depends on our having access to its context, and in particular on our ability to translate other ancient languages.) The second thing that the process of translation presupposes is (as I indicated in the last section) a very large measure of agreement in factual judgements between the relevant parties, the language users and their translators. For, as Davidson puts it, ‘If we cannot find a way to interpret the utterances and other behaviour of a creature as revealing a set of beliefs largely consistent and true by our own standards, we have no reason to count that creature as rational, as having beliefs, or as saying anything’ (1984, 137). It is this large measure of agreement that, as we said, makes disagreement possible. I can only think that something you say is false if I think that most of what you say (and think) is true; without that shared basis, I could not identify your supposedly false belief as a belief, or isolate its content. Davidson again: ‘Before some object in, or aspect of, the world can become part of the subject matter of a belief (true or false) there must be endless true beliefs about the subject matter. False beliefs tend to undermine the identification of the subject matter; to undermine, therefore, the validity of a description of the belief as being about that subject’ (ibid., 168)—and ultimately, if the falsity we think we find is sufficiently pervasive, to undermine the ascription of beliefs to the relevant speakers, who may then turn out not to be speakers, or rational beings, at all. After the passage I have just quoted, Davidson gives a particular, and rather unfortunate, example. He implies that the ancients could not have believed that the earth was flat—that is, that they could not have had a belief with exactly that content. So we would be wrong to translate the ancients as having had that belief. This goes too far,26 but the bad example does not invalidate the general claim, that translation rests on widespread agreement between speaker and translator. In order to defuse the threat posed by that example, we need simply recall that, while the ancients held false astronomical beliefs, those beliefs formed a tiny and

200  Tragedy and Linguistic Redress relatively insignificant subset of the total mass of their beliefs about the earth, most of these too trivial to attract notice. This is a point that Davidson himself stresses: ‘If the vast amount of agreement on plain matters that is assumed in communication escapes notice, it’s because the shared truths are too many and too dull to bear mentioning. What we want to talk about is what’s new, surprising, or disputed’ (ibid., 153; cf. 200). So the ancients can certainly have held the false belief that the earth was flat, but their entitlement to hold that belief, with exactly that content—a belief about the earth, that it was flat—is fixed by the fact that the vast majority of their beliefs about the earth, and about flat things, were, by our lights, true. The agreement in factual judgements that is presupposed in all linguistic communication and understanding includes, crucially for our purposes, agreement in moral, and in general evaluative, judgements: we must, by and large, agree with those we are translating about what things are good, just, and true; making linguistic and psychological sense of someone else involves finding him or her ‘consistent, a believer of truths, and a lover of the good’ (Davidson 1982, 222). Moral disagreements, where they arise, either within or across cultures, are and must be insignificant by comparison with the large areas of agreement.27 (Obviously we cannot specify particular beliefs that communicators must agree on: the requirement of agreement is much looser and more flexible than that, but far from being vacuous.)28 It is certainly possible for us to interpret a culture as believing that slavery is just, for example, provided that not too many of their beliefs about justice (or slavery) differ from ours.29 Joshua Billings suggests (2014, 20) that ‘Antiquity can no longer be measured against modernity; there are no firm grounds for evaluation’. But there are: in translating ancient languages, we are given such grounds; the very possibility of translation presupposes them. Billings’s claim unpicks itself: if antiquity could not be ‘measured against modernity’ or ‘evaluated’, there would be nothing there to measure or evaluate, or nothing relevantly there; we could no longer descry in antiquity anything that amounted to a culture in the first place. What we call ‘antiquity’ would be assimilated to that long prehistoric period in the earth’s evolutionary story before anything like human mindedness came on the scene. But look at what we do say about the ancient world; look, in particular, at moral judgements that we pass when we engage with ancient literature. We say, for example, that in Euripides’ Medea, ‘Jason’s dishonesty and his shameful abandonment of his family are made absolutely clear’, that Jason is ‘a bad husband’, and that Medea ‘shamefully’—as she herself says (166–7)—‘betrayed her father and murdered her brother’.30 For ‘she does what, by the standards of any age and culture, is clearly wrong’.31 ‘Shamefully’ translates ‘aischrōs’, and the point here is that we could not translate words like this word at all, let alone translate them using anything as rich and precise as English moral vocabulary, if we did not,

Theoretical Considerations 201 by and large, agree that acts described by the ancient Greeks as shameful are shameful. I include a proviso in the last sentence because there must be exceptions: we can translate a word from another language as ‘witch’, for example, without agreeing that there are witches, just as we can translate an ancient text as stating, falsely by our lights, that the earth is flat or that slavery is just. But the Davidsonian point is that such cases are necessarily exceptional: disagreement on a few points depends on an enormous penumbra of agreement—agreement which concerns moral and non-moral facts that are in most cases too trivial to attract interest, but which are there, and must be there, if translation is to be possible. The Davidsonian point applies also where little, or no, translation is involved: any kind of mutual linguistic understanding, with or without translation, presupposes massive agreement in factual, including moral, judgements. Catherine Belsey writes that ‘Lucrece’s culture, whether Roman or early modern, is not our own, and her struggle to arrive at a remedy for her loss is conducted in other terms than ours (which is to say, in terms of other meanings)’ (2001, 332). But if the meanings really were other, we could not understand Livy’s history or Shakespeare’s poem, and then, unwinding, we should be disinclined to suppose that the relevant artefacts were texts composed of meaningful words in the first place, as opposed, say, to objects d’art. But we can understand these texts, and so are happy to conclude that they really are texts composed of meaningful words. From a theoretical perspective, we need to reflect on how much this ready accessibility of old texts presupposes. So it is not ‘miraculous’, as Belsey thinks (1992, 34), that there should be a significant level of moral agreement between us and Chaucer. Even if Chaucer’s language were more remote from us than Middle English is (which is not very remote), there would be no miracle. Again, our ability to translate Homer depends on a wide measure of agreement in judgements between us and the culture of the Homeric age. Homer portrays a world in which (for example) a balance has to be struck between the exercises of competitive and co-operative virtues; but it is not merely the case that, as John Gould has stressed (2001, 167), this was just as true of fifth-century Athens as of the Homeric age; it has always and everywhere been true of human culture, and it is every bit as true today.32 Even points on which we might think we discern significant differences between Homer’s age and ours turn out, on closer examination, to lose much of their cogency. The aristocratic equation of birth and wealth with excellence came under increasing pressure in the fifth century BCE, particularly in the plays of Euripides.33 But the equation has never lost all its power—even today traces of it remain—and the tension in the aristocratic code between ideals of good breeding and moral virtue pervades the tragic tradition.34 The old idea that Homeric society is a ‘shame culture’ and modern society a ‘guilt culture’ has long been discredited—that is, insofar as the notions of shame and guilt can in any case be coherently

202  Tragedy and Linguistic Redress separated: one might say, for instance, that shame has reference to the whole personality and guilt to specific acts, or that shame implies an external and guilt an internal tribunal. But on any reasonable way of making the distinction, one will find that both ancient and modern societies exemplify both concepts: perhaps there has been a shift in emphasis, but the difference, if any, is one of degree, not kind. Homer’s Agamemnon apologizes not, or not merely, because he feels shame, but because he feels guilt; on the other hand, shame still plays a significant role in modern Western culture—and not only in modern Greece, as David Kovacs implausibly suggests.35 In Attic tragedy, the concepts of retributive and legal justice constantly recur, and we have no difficulty grasping these concepts: both are still very much with us. In Euripides’ Orestes, for instance, Tyndareus criticizes Orestes for not invoking legal process against Clytemnestra, and argues rationalistically against the morality of vendetta killing;36 both arguments are fully comprehensible, indeed ‘modern’. In all three Electra plays, there is a questioning of the morality of vendetta killing, in a way which is immediately accessible to us today. (At least that is evident for Aeschylus’ and Euripides’ versions; Sophocles’ is perhaps more controversial.)37 The genetic argument for the primacy of the father over the mother, given by Apollo at Aeschylus, Eumenides 658–9, and repeated by Orestes at Euripides, Orestes 551–6, is ‘offensive to present-day ideas’,38 but that is a difference which does not bear on the concept of justice as such: the older idea is derived from empirically false beliefs about the relative importance of the roles of male and female in procreation, and when those beliefs are corrected the concept of justice naturally adjusts. One view, as we have noted (§23), is that the audience is intended to regard Apollo’s and Athene’s reasoning in the trial scene of the Eumenides as flawed: this interpretation obviously depends on a transhistorical application of the concept of justice. But even if the interpretation is wrong, and we are meant to take the end of that play at face value, we still have to do with an application—flawed, in our terms—of the normal concept of justice, not with a different concept which could only ambiguously be called one of ‘justice’. Once again, the Davidsonian point is that our ability to recognize what is going on in the trial scene of the Eumenides as a (possibly flawed) case of the administration of justice, as opposed to the administration of something else, or no kind of administration or ratiocination at all, guarantees the sameness of concept. It is exactly this identity of meaning—and identity of concept, concepts being meanings, as we have said—between the ancient and the modern that enables us to see ourselves in Oedipus (§13).39 Only on the basis of this transhistorical identity are we entitled to say, as most spectators and readers of ancient tragedy (including those influenced by recent theoretical developments) still want to say, that, in the Thebes of the Attic stage, ‘the most serious questions can be raised concerning the fundamental relations of man to his universe’.40

Theoretical Considerations 203 Much recent work on Greek tragedy aims, in Edith Hall’s words, ‘to locate the plays within the historical conditions of their production’ (1997, 94), and there can be no quarrel with that project: I have presupposed it myself, in my discussions of the Oedipus Rex and the Antigone. But Hall, like many others, sees a tension between this task and the universalism promoted by Aristotle, the Romantics, and Kierkegaard, which presupposed (ibid.) ‘an immutable human condition whose teleological imperative was suffering, and which somehow transcended transhistorical changes, and differences in culture and language. The “eternal verities” contingent upon it were assumed to have been mysteriously encoded, by the “genius” of the playwrights, in the tragic drama of the Greeks’. But why could the Greek dramatists not both respond to the specificities of their age and produce drama that has universal appeal? Is that not exactly what they did? There is no tension between the contextual specificity of the plays and the universality of their import precisely because the language of the Greek tragedians, though a product of their spatio-temporal location, is translatable into other languages— languages embedded in quite different times and places. So there is no mystery about their ability to ‘encode’—if we wish to use that word, though ‘express’ would be more accurate41—‘eternal verities’. And there is no call to scare-quote this latter phrase: we are talking about eternal verities, or at least persistent verities. The Davidsonian point that our ability to translate the Greek tragedians forces us to recognize them (and the culture they sprang from) as (by and large) expressing truths ensures that, in the case of as profound a genre as tragedy, we are warranted in replacing the neutral term ‘truths’ with the grander ‘verities’. But here one may ask: does this argument not presuppose, or import, a relatively fixed human essence? Yes, it does. But this is a feature of it that we should not baulk at: this brings me to the second strand of new historicism as I defined that doctrine above.

28 Historicism and Human Nature According to the second strand of new historicism as I have defined that there is no such thing as a human essence. This position is common to new historicists, cultural materialists, and adherents of some other recent theoretical trends. On these approaches, human nature is held to be a social construct (a ‘social text’) which is culturally determined (‘inscribed’), and so inevitably varies from culture to culture, and from era to era. These views are so commonly expressed that it is hardly necessary to provide citation, but here are a few examples. Richard Wilson and Richard Dutton write that ‘Where criticism traditionally assumes a “human condition” and refers Renaissance drama to the supposedly transhistorical and universal concept of “human nature”, Cultural Materialism begins from the thesis that there is no essential human core since individuals are

204  Tragedy and Linguistic Redress products of societies in history’. For Stephen Greenblatt, ‘the very idea of a “defining human essence” is precisely what new historicists find vacuous and untenable’. Simon Goldhill informs us that ‘ “human nature” is not a cross-cultural “essential truth” . . . To rewrite Foucault, human nature is an invention’. Jonathan Dollimore rejects, on behalf of cultural materialism and feminism, a view ‘which sees Shakespeare’s women as exemplifying the transhistorical (universal) qualities of “woman”, with Shakespeare’s ability to represent these being another aspect of a genius who transcends not only his time but also his sex’. Catherine Belsey tells us that ‘sexuality is not given but socially produced’; Alan Sinfield agrees that sexualities are ‘ideological formations’; subjectivity, he tells us, is ‘an effect of cultural production’. And so on.42 The thesis that human nature is a cultural construct is an empirical one, and you would therefore expect those theorists who advance it to produce some evidence in its favour. But not a shred of empirical evidence is ever forthcoming from the literary theorists who promulgate and parrot the thesis. The truth is that those who believe it do so because they want to believe it; they want to believe it because they think it serves the ends of political equality. In that they are wrong, as I shall explain shortly. The point for now is that, considered (as it should be) from a biological and evolutionary point of view, the claim that human nature is a cultural construct is hopelessly adrift from the facts.43 Human biological nature has changed negligibly in historical time, it varies negligibly across cultures, and it is our biological make-up that determines many aspects, and the most basic aspects, of our physical and psychological lives. There has simply not been enough evolutionary time since Homer, say, for fundamental changes to have occurred in the human psyche. And, crucially, some of the most substantial elements of human life, such as birth, maturation, cognition, sense-experience, pleasure and pain, sexuality and reproduction, metabolism and dietary requirements, hunger and thirst, physical and emotional vulnerability, illness, ageing, finitude, and death, long antedate, in our evolutionary history, the emergence on the scene of anything resembling human culture, and so cannot be mere products of culture—or of cultures, if you prefer, but the pluralizing of this word cannot do away with the evident fact that there are necessary fixtures from culture to culture, keyed to fundamental human needs and to our biological constitution. Common to different cultures, too, are many other aspects of our lives beyond the basic ones just listed, including the rites of passage from youth to adulthood, love and hate between individuals and groups, co-operation and conflict, egoism and altruism, deliberation and dialogue and argument and persuasion, the pursuit of individual and group success (however success is defined), a sense of group identity and belonging, humour and play, the development of skills for solving problems, and the exchange of goods and services. The items I have listed so far in this paragraph are just a tiny sample of the human universals

Theoretical Considerations 205 that ethnologists have uncovered in recent decades: the list is staggering, comprising properties found not merely in most but in all human societies ever documented. Many things that one might have been happy to concede to the relativists while insisting on some of the basic items mentioned above turn out to be cross-cultural features of human nature, from fear of snakes to the use of logical operators, from ownership of property to the proscription of rape. The great ‘Ode to Man’ in Sophocles’ Antigone, which meditates so profoundly and beautifully on human capacities and limitations, speaks as directly to us today as it must have done to its first audience, and its final prayer, that the wrongdoer ‘may never share my hearth or my thoughts’ needs only the addition of the rider ‘as posted on Facebook’ to be brought fully up to date.44 Staying with that play, we might note that the proper treatment of the dead, whether by earth burial, cremation, or otherwise, is a human universal that stretches from the earliest societies to the present day: it is as important to us as it was to Sophocles’ audience,45 and the only mental adjustment that we have to make in coming to understand its central role in the Antigone, or in the Ajax, or in Euripides’ Suppliants, is the thought that, in these plays, it is not to be taken for granted. But there are situations today in which it cannot be taken for granted, and, even when it is, there is still an underlying counterfactual possibility of failure. Matthew Arnold was quite wrong to say, in the preface to his 1853 Poems, that the issue of burial in the Antigone ‘is no longer one in which it is possible that we should feel a deep interest’ (1909, 12). Here I dissent also from Martha Nussbaum’s view that ‘When we read Sophocles’ Antigone, we see a good deal that seems strange to us; and we have not read the play well if we do not notice how far its conceptions of death, womanhood, and so on differ from our own’ (1993, 261). I would say that we find very little in the Antigone that is strange to us, concerning death, womanhood, or anything else. Culture itself is a fixture, for, as Aristotle said, man is by nature a political animal;46 there is no need for an essentialist humanism to deny this in favour of an extreme individualism, as Dollimore implies (2010, 250). So when Belsey writes of birth, reproduction, and death that ‘these events are undergone in culture’ (1999, 8), that is correct; but when she adds that ‘if we experience them as natural, that only goes to show how thoroughly we have internalized the meanings culture prescribes’, that clearly misfires. We experience birth, reproduction, and death as natural because they are natural: the best explanation of our belief goes through the fact, and so vindicates it.47 (Contrast what an atheist will say about belief in a god or gods, namely that the best explanation of the existence of that belief is purely naturalistic—sociological and psychological, so that there is no explanatory need to posit the existence of a god. The atheist gives a debunking account of religious belief, as Nietzsche’s account of the origin of morality was debunking; Hume’s account of that origin

206  Tragedy and Linguistic Redress is vindicatory.)48 Few of the thinkers who subscribe to the dogma that sexuality, for example, is a ‘social construct’ ever stop to reflect on the evolutionary facts of the matter. Freud explored the continuity between our sexuality and that of very simple organisms (2013, 52–77); more evidently, we share our sexuality with much of the rest of the higher-animal kingdom, and it would be implausible to hold that canine sexuality, or the sexuality of the great apes, which is so close to our own, were ‘social constructs’, so that—given how deeply sexuality is embedded in our biological nature—it is hardly likely that ours is one, either.49 Tragedy is a genre (though not the only one) that engages with these fundamental aspects of our lives,50 and it is for that reason that it has always been held to be universal. Martin Coyle opines that ‘we should not confuse the ability of later ages to respond to texts with the notion that there is something of great matter in tragedy, something outlasting its moment’ (2003, 25); but how is this a confusion? You might say that the ability of later ages to respond to older tragedy is what its universality consists in. The one thing is constituted by the other. I noted at the end of the last section that much of what recent critics wish to claim about older literature in fact presupposes t­ranshistoricism— ironically enough, given the hostility of many of these critics to that doctrine. The point applies particularly to politically motivated criticism, much of which is not only moralizing, but highly moralistic (in the pejorative sense). If transhistoricism were false, feminist treatments of the Antigone, for example, would be quite straightforwardly anachronistic, as would attempts to view the play as an early study in civil disobedience.51 (Sometimes such criticism is anachronistic, as I have said: but that is not an inevitable consequence of the method; it depends on the details.) So we find a strange antinomy in modern radical criticism: on the one hand, the urge to read older literature as speaking to modern political concerns; on the other, a denial of the continuity in human nature which alone would enable such a confluence of interests between past and present to occur. And it is worth observing in this context that, quite apart from its falsity, the thesis that our sexuality is a social construct fails to achieve politically what its proponents plainly want it to achieve. It does not serve the ends of sexual equality. Just as anyone who is interested in the preservation and integrity of minority cultures is ill-advised to espouse cultural relativism52—for that doctrine in effect permits stronger cultures to oppress weaker ones if it suits them—so too anyone who supports the cause of sexual equality is ill-advised to maintain that sexuality (or gender, if that is being distinguished from sexuality) is a social construct. That thesis plays directly into the hands of the opponents of equality: for what is constructed is artificial, and can be changed; those who have been constructed in the ‘wrong’ way can be deconstructed (forcibly, if necessary) and then reconstructed in the ‘right’ way. It is wiser, as well as truer, to argue the case for equality on the basis of an essentialist humanism.53

Theoretical Considerations 207 Dollimore and others are hostile to essentialist humanism because they think that the very idea of a human essence rules out alternative ways of being human, but that is an elementary technical error: an essence can be disjunctive.54 (Which is not to say that any old disjunction of properties can constitute a natural essence, that is, one that has a role to play in scientific explanation.) In work on Greek tragedy, the new historicist rejection of the idea of human essence finds an echo in the structuralism of Jean-Pierre Vernant and his followers.55 Vernant maintains (1996, 10) that ‘the writings that Ancient Greece produced are sufficiently “different” from those that form our spiritual universe to disorient us from ourselves and to give us, along with the feeling of historical distance, an awareness of a change in human kind’. (I am not sure why ‘different’ is scare-quoted:56 surely Vernant wants to say that they are different.) And certainly, when one reads such books as Ruth Padel’s In and Out of the Mind (1992), which, following in the tradition of Onians (1951), charts minute aspects of Greeks’ beliefs about mind and body, it is easy to feel that what Vernant says must be right. Consider the practices of animal sacrifice and haruspicy. Padel discusses the meanings that ancient diviners extracted from the condition of entrails, and the connections they drew with mentality (1992, 17–18): To us, these sacrificial anatomic overtones seem alien and irrelevant to questions of ‘mind’. How many of us hold a calf’s entrails in our hands, realize the liver lobe is missing or how markings vary on the ‘portal vein’, believe this matters, and apply words for what we are holding to the inner equipment with which we imagine we feel and think? But tragedy and its audiences were familiar, in intense, mystery-surrounded, physical experience absent from our own lives, with the stuff to which they attribute activity within themselves. True, but such beliefs are adventitious epiphenomena as far as the language of mind goes. We know this on the basis of the Davidsonian linguistic considerations already rehearsed. In order to get as far as we do in our understanding of Greek beliefs—in order to be able to say, for example, that it was widely believed that the layout and condition of calves’ entrails were so much as relevant to human mentality—it needs to be the case that we have a good grip on the Greek language of mentality, and of much else. (Otherwise, what entitles us to say that it was believed that calves’ entrails had a bearing on mentality?) But that presupposes that the vast majority of things that the Greeks said or would have said about the faculties of mind are also things that we would say. Davidson, as we saw, reminds us that we tend to focus on what is strange in other cultures, because that is more interesting, ignoring the commonalities. Many of these commonalities are too trivial to attract

208  Tragedy and Linguistic Redress notice: they are not as striking as the facts to which Padel draws our attention. What is familiar is hard to see.57 But they are there nevertheless, and provide a basis of shared mentality which, empirically speaking, is what grounds our understanding of Greek culture.58 For example, before we can think about the various senses and associations of a word like ‘thumos’ we need to know such basic things as that, when a Homeric hero ‘speaks to his thumos’, as the phrase has it, he is not speaking to his comrade-in-arms, his horse, his sword, his native land, his favourite Olympian deity, or the little finger of his left hand. But in ruling out these candidates, which we do automatically when we read Homer, we should not forget that our ability to make these exclusions depends on our sharing with Homer and his audience the vast majority of our beliefs about the human mind. We isolate what is strange about the Homeric thumos by embedding it in a context of the familiar; and that is the only way we can get at, or even make sense of, the idea of the strange. Padel herself effortlessly demythologizes the unfamiliar idioms in a way that displays their adventitious and epiphenomenal nature: ‘ “I measured your phrenes and realized how great a bitterness you are to me, as my enemy,” says Ion to his would-be murderer ([Euripides, Ion] 1271). He thinks he has seen into, as we would say, her heart or mind’ (1992, 21). ‘As we would say’: and we would be right, and know that we were right, to say that. Now how did we manage that? The idioms are not so strange after all. There is a very simple argument against any kind of extreme historicism, such as is (sometimes) embraced by Vernant. At one point, Vernant implies that fifth-century BCE tragedy is so ancient and remote that not even Aristotle, let alone Freud, could understand it.59 But such a view is straightforwardly self-refuting: for in order to be in a position to say that Aristotle no longer understood Greek tragedy, we must ourselves understand that tragedy—and understand Aristotle (which ought to be just as problematic). So we have cut off the branch on which we were sitting.60 And indeed there are numerous places in his writings where Vernant is forced to recognize the transhistoricity and universality of Greek tragedy.61 Arguments such as the one I have just given, however, are of rather limited value: they do demonstrate, conclusively, that the view in question must be wrong, but they do not show us in any deep sense how it goes wrong. So let me attempt this latter task: here I can appeal to points already made, and present the argument in summary form. The view that the ancient Greek approach to mindedness is fundamentally different from ours is based on a methodology that starts too late in the heuristic process. It simply takes for granted all the (mostly trivial) commonalities between ancient culture and ours, without noticing them, and then, noticing the differences (which there undoubtedly are), exaggerates their importance. ‘Faced with an ancient foreign language whose speakers are inaccessible, our business is to reach for as many implications as we can in the words they used for feeling and how it is felt. These

Theoretical Considerations 209 implications must be grounded in their associations, not ours’, writes Padel (ibid., 36). No: our first business is to translate (that is, interpret) this ancient foreign language,62 before ever we get as far as looking for ‘implications’ in the ways speakers of that language talked about feelings so as to invoke their associations, not ours. Ab initio, we do not yet know that they had any feelings; we do not yet know that they had minds at all; we do not yet know that this bunch of curious markings on papyrus or stone is a language. So strictly speaking our first business is to recognize Homeric Greek (say) as a language; and we can only do that if we can, at least in principle, translate it. But we can translate Homeric Greek,63 not only in principle but in practice, and we can do so extremely well. (There are a few words whose meanings are unknown to us, the evidence being insufficient to constrain a meaning, but they do not seriously impair our understanding; in any case, their existence reinforces my point, since it is only possible to isolate them as words whose meanings we do not know because we understand nearly everything else in Homer.) And translation, as we have seen, presupposes significant agreement in judgements: moreover, it is only on the basis of that wide-ranging agreement that we can so much as give sense to the idea that ancient thinkers might differ from us in some of their beliefs and associations, concerning the mind or anything else, let alone interpret them as actually doing so.

29 Ancient Psychology and the Question of Anachronism In spite of my affinity with Böckh’s general approach to the Antigone, I criticized him in Chapter 2 for offering a reading of the heroine that was ‘disappointingly unpsychological’. Some readers might have felt at the time that this criticism was anachronistic, in that it ignored differences between the ways we think and theorize about the human psyche, and perhaps also the ways in which we represent mindedness on stage, on the one hand, and the ways in which Sophocles and his original audience might have treated these matters, on the other.64 If my argument so far in this chapter has been along the right lines, this worry ought to seem exaggerated: for it underestimates the extent of our continuing closeness to ancient ways of thought and representation of thought. And, even setting aside the relatively a priori Davidsonian considerations I have rehearsed, I believe that our study of the ancient texts themselves shows how close we are psychologically to the Greeks of the fifth century BCE and indeed to Homer. In a sense I am reluctant to give examples here, for the thesis seems so self-evident—it speaks to us so clearly from every page of ancient literature that we read—that to provide a handful of examples might appear to limit and so to weaken the case. But lest the claim fail to persuade for want of support, let me briefly illustrate it. In Aeschylus’ Persae, when news of the disaster at Salamis first reaches Atossa and her court, the messenger and the chorus engage in a grieving

210  Tragedy and Linguistic Redress lyric exchange, but the queen is silent. At length she comes to utterance: ‘I have been silent all this while because I am wretched and struck dumb by the disaster: this catastrophe quells my power to speak or ask about the sufferings’ (290–2). In the next chapter I shall be interested in the bearing of passages like this on the question of language’s competence to express suffering, but here the point is that these lines are, as Ann Michelini says (1982, 30), ‘psychologically persuasive’. Atossa’s reaction to the bad news fits the way people often react in such circumstances, then and now. In Sophocles’ Ajax, when Teucer, standing over his brother’s dead body, imagines their father Telamon blaming him for Ajax’ death (1012–16), Patrick Finglass comments (ad loc.) that ‘Telamon’s reaction, seizing on someone to blame, is psychologically plausible’. Psychologically plausible to whom, one may ask? To Athenians of the fifth century, or to us in the twenty-first century? And which ‘us’ are we talking about? Psychologically plausible to everyone, is the clear and correct answer. That people look round for someone to blame when catastrophe strikes is a human fixture that neither geographical distance nor the passage of time does anything to alter. (It indeed underlay the main thesis of Part I.) When the sentry in the Antigone expresses ambivalent feelings about his arrest of the heroine (436–40), Mark Griffith comments that ‘These are, in effect, the audience’s own reactions, albeit exaggerated and even parodied; in these expressions of mixed sympathy and dismissal, affection and rejection, voiced by the Guard, we may find our own confused responses to Antigone’s message conveniently articulated’ (2005b, 120). Griffith here moves effortlessly from the reactions of the fifth-century BCE audience to our reactions. He suggests that the incestuous interpretation of Antigone, which I favoured in Chapter 2, engages with the unconscious in a way that connects with Freud, and comments that ‘this masked, speaking character [Antigone] does give voice to desires that are at one and the same time, I suggest, deeply felt by almost every member of the theatre audience and yet barely acknowledged (not consciously recognized)’ (ibid., 116). That applies equally to Sophocles’ audience and to modern audiences. When Euripides’ Creusa answers Ion’s question about the age of her lost son with ‘About your age’ (354), Kevin Lee remarks (ad loc.) that ‘The irony here is both natural and moving. Parents who have lost children are continually reminded of what might have been when they see those of the same age’. True then, true now. On the scene in the Hippolytus where the Nurse overbears Phaedra—letting her do so is Phaedra’s crucial hamartia—Christopher Pelling comments that ‘psychology matters here, but more the psychology of the character who is listening and persuaded than the character of the person who speaks’ (2005, 92). Psychology matters—to the participants in the dialogue, and to us interpreters of it. Justina Gregory argues that the striking changes of register that we find at Aeschylus, Agamemnon 1072–1197, at Sophocles, Antigone 806–928, and at Euripides, Alcestis 252–325, are

Theoretical Considerations 211 ‘psychologically implausible’.65 Actually I disagree about the implausibility (and in effect argued against it in the case of the Antigone: we have to do with two different stages in Antigone’s breakdown). But the point here is that Gregory’s comment makes good sense, whether it is right or wrong: we demand not only consistency of ancient works (§13), but also psychological ­plausibility—by our lights, which, as I am suggesting, are not importantly different from ancient Greek lights—and we criticize departures from it.66 The application of modern psychological methods and ideas, particularly Freudian ones, has proved illuminating in the study of ancient and Renaissance literature. Freud adopted a universalist approach to human nature based on what he took to be the general validity of the psychoanalytic method.67 My universalist strategy is based, rather, on ­linguistic— Davidsonian translational—considerations, but I accept that the Freudian approach has some traction. Notoriously, of course, Sophocles’ Oedipus does not have an Oedipus complex;68 but that does not mean that Freudian readings have no bearing on the play; and their relevance to Seneca’s Oedipus is even clearer.69 It has been standard critical practice since E. R. Dodds published his commentary on Euripides’ Bacchae to interpret Pentheus’ voyeuristic interest in the maenads and their activities in Freudian terms.70 In the same playwright’s Alcestis, Admetus’ response to the strange woman that Heracles tries to foist on him is suggestive of a subconscious recognition of her on his part.71 And in the Hippolytus it is plain—though nowhere expressly said—that Hippolytus’ bastardy strongly influences his attitude to sex in a way that invites explanation in terms of the subconscious.72 So too Phaedra’s delirious desire to go to the meadows is naturally taken as giving vent to a subconscious desire to be with Hippolytus.73 When Hippolytus complains of his bastardy (1082–3), W. S. Barrett comments (ad loc.) that ‘Euripides’ purpose is to throw subtle light on Hippolytus’ psychology for its own sake, to suggest [a] feeling of inferiority, of otherness, as what lies behind his urge to establish himself in compensation as a paragon of virtues that common man cannot share’. Some commentators reject this kind of approach entirely;74 but that is usually because they overlook the operatur principle (§§8, 19). I think Barrett’s comment undershoots: Hippolytus’ outburst signals not so much an inferiority complex as a disgust of sexuality (compare line 1455), grounded in his consciousness of his origins.75 It is attractive to read Shakespeare in Venus and Adonis as pursuing a similar thought, for Shakespeare’s Adonis, unlike the Ovidian original, is a Hippolytus figure who rejects sex: perhaps the explanation lies in his nonstandard parentage and birth, conceived as he was in Myrrha’s act of incest with her father, and delivered into the world from the trunk of the tree into which his mother had been metamorphosed. There is no explicit mention in Venus and Adonis of Adonis’ origin, but that the reader is meant to recall the pertinent facts is suggested by the very striking change

212  Tragedy and Linguistic Redress in Adonis from Ovid’s compliant figure to the Hippolytus redivivus of Shakespeare’s poem: for the weird background, as in the case of Euripides’ (and Seneca’s) Hippolytus, explains the moral rigorism and puritanism.76 That is perhaps a peculiarly modern (Freudian) thought, but it can nevertheless be true of the Greek and Roman tragedies and the Shakespearean poem.77 Conversely, it may be that ancient ideas of the unconscious have more relevance to us today than has been generally recognized.78 An example which illustrates both psychological and moral continuity (to resume a theme of §27) is afforded by Euripides’ Electra and her insults to the dead body of Aegisthus. Martin Cropp comments: ‘Insulting a dead man is morally questionable in itself (902 n.), and this display of spite shows the obsessive nature of Electra’s hatred’.79 Again we can ask: morally questionable for whom? If we consult Cropp’s note on line 902, to which he refers us, we find that he there cites a passage in the Odyssey (XXII, 412) where Odysseus condemns insulting the dead. But Cropp’s statement looks like a general one. So it appears that he is saying that insulting the dead is wrong for Homer, wrong for Euripides, and wrong absolutely. And he is right to generalize in this way, because our understanding of Homer and Euripides is underpinned by a common morality that, among many other things, disapproves of insulting the dead. (In Homer, for example, it is made clear that Achilles’ desecration of Hector’s corpse goes too far.)80 We also notice that Cropp makes a psychological judgement about Electra that employs what one might regard as a distinctively modern concept, namely the concept of obsession. Again, the work bears him out, because it is simply the case that we can see that Electra evinces obsession, which is a psychological universal. This obsessive aspect of the Electra is well brought out in the Hofmannsthal/Strauss version, especially at its famous close: again, as with Anouilh’s Antigone, we have a case where a modern adaptation, though apparently departing from the original, captures its essence. When the Iphigeneia of Euripides’ Iphigeneia in Tauris relates the events at Aulis (359–71), where she narrowly escaped being sacrificed to Artemis, we can see that she has been, as L. P. E. Parker says (ad loc.), ‘deeply traumatized’ by the experience. Parker adds that Euripides’ portrayal of Iphigeneia ‘justifies’ application of ‘the modern concept’ of trauma. But here we should note that, in terms of my Fregean way with the relevant terminology (§26), and as I have just hinted in connection with the concept of obsession, this statement comes out as inaccurate: for me, as for Frege, concepts are reference-level entities, not sense-level entities, that is, they are objective components of the world, not aspects of our ways of thinking about the world. So for me the concept of trauma has been around for at least as long as it has been possible for human beings to be traumatized, so at least since human culture reached an appropriate stage of emotional sophistication; it is just that that concept has not been

Theoretical Considerations 213 named until recently. Hence it is not the concept, but the name, that is modern. Hence, too, a concept may be (and often is) applicable to behaviour even if the culture in which the behaviour is embedded not only does not itself make the application but cannot make the application in the sense that it does not dispose of a name for that concept. (That ‘cannot’ is relativized to its actual lexicon: in a wider modal sense, the culture certainly can make the application, by introducing a name for the concept.)81 So, too, it makes good sense to impute constructive knowledge to Oedipus, as I did in Chapter 1, even though the legal recognition of constructive phenomena is distinctively modern. All this does not upset Parker’s basic and correct point, which is that Iphigeneia exemplifies the concept of trauma—to which I simply add that the concept was ontologically present in ancient culture, though as yet unnamed.

30 Common Concepts In several places Goldhill inveighs against the transhistorical and objectivist humanism that I have been defending. In one passage, he attacks Brian Vickers’s comment that ‘Greek tragedy is about people . . . Human behaviour in these plays . . . concerns those fundamental human passions which are reflected to a greater or lesser degree in the literature of all nations at all periods. In Greek tragedy, people love and hate as we do . . .’.82 Goldhill’s objection to this statement is that it ‘begs the question . . . by the repression of the cultural specificity of the fifth-century Athenian construction of “love” and “hate” ’, and he adds: How can philein or erōs be simply translated as ‘love’, as if the history of courtly love, Romanticism, not to mention Christianity, makes no difference to a modern reader’s approach to such a term; as if, indeed, the interrelations of the sexes in the ancient world could simply be mapped onto a modern, post-Freudian emotional topography?83 But does this mean that Vickers is wrong to say that ‘in Greek tragedy, people love and hate as we do’? Note that ‘as’ here does not mean ‘in exactly the same way as’. Vickers is saying: ‘there is something in common between them and us: we love and hate, and they do too’. This, if not already clear, emerges from the continuation of the passage: ‘. . . they protect or destroy as we do; like us, they deceive each other, abuse language or beliefs to suit their own ends. They are no less concerned than we—or our descendants, hopefully—with personal integrity, justice, and political health’. Apart, obviously (given what I said in §26), from wishing to enter a reservation about the extent to which Greek tragic characters ‘abuse language’, I am in accord with this statement; and indeed (modulo that qualm, to which I shall return) I cannot see how it could be seriously disputed. The correct understanding of Vickers’s ‘as’ is

214  Tragedy and Linguistic Redress actually quite an important point: Goldhill rehearses some of the differences between the conceptualizations of virginity under Greek, Roman, and Christian mores and concludes, rhetorically, ‘in what sense can either [Haemon or Antigone] be said simply “to love as we do”?’ (1990a, 104). Perhaps not exactly as we do: but there is a significant core of commonality. It would need a good argument to show that the differences are more than superficial. And indeed, as I have already argued and shall explore further here, they cannot be more than superficial. Are we to suppose that it is just a solecism to use the English words ‘love’ and ‘hate’ in talking about Greek tragedy, or ancient culture more widely? When Ruth Scodel writes that ‘Haemon in Antigone is in love, but Antigone is not. Women of Trachis is a tragedy of erōs, about a woman who is in love with her husband’ (2005, 234), has she uttered nonsense? That, generalized to all words—and there would be nothing to stop the generalization, since there is nothing special about the word ‘love’—would make discussion of the ancient world impossible, including Goldhill’s own: for he uses English words to characterize what is going on in Greek tragedy, and he clearly takes himself to be uttering not merely meaningful but even true sentences. For example, elsewhere we are told that ‘unlike any Judaeo-Christian notion of loving one’s neighbour or turning the other cheek, perhaps the most basic and generally agreed position with regard to correct behaviour in the ancient world was “to love one’s friend and to hate one’s enemy” ’ (1986a, 83). I take it that the reason why this latter phrase is placed in quotation marks is not to discourage one from dropping the quotation marks and uttering the phrase transparently, but merely because it is a standard phrase. That must, in consistency, be right—at least, it is the charitable r­ eading— because Goldhill has just permitted himself to talk transparently (that is, without employing quote marks) about loving one’s neighbour and turning the other cheek, phrases that translate words of New Testament Greek, and it would hardly be plausible to maintain that, though Greek of the fifth century BCE was untranslatable into modern English, it had become translatable by the time the New Testament was written. Goldhill is quite happy to translate ‘philein’ as love; and indeed he must go so far because, as stated, if he did not his discussion could not get off the ground. As I have observed, the objection he raises, if sound, affects all words: we, and he, would not even be able to translate ‘kai’ as and. It is true that ‘kai’ has a range of nuances that do not exactly match those of the English ‘and’, such as its use in final clauses.84 But we are still right to translate ‘kai’ as and in many of its uses, and (a point to which I shall return) we can express those different nuances (as Denniston and others do) in our language.85 In an article on the politics of Aeschylean tragedy, having told us that the term ‘emotion’ was an invention of nineteenth-century science—in fact that post-dates the introduction of the word by about two centuries,

Theoretical Considerations 215 as a glance at OED shows86—Goldhill is quite happy to talk a few pages later about the ‘emotional . . . possibilities’ explored by Euripides’ Bacchae, and the ‘emotionally powerful choral odes’ and ‘speeches about emotional reaction’ of the Oresteia, opining that this trilogy contains ‘some of the most emotionally charged high points in the corpus of extant tragedy’ (2000, 36, 41). What entitles him to say all this, if the language of ‘emotion’ was a post-classical invention? The answer lies in separating word and thing (a point that will be a recurring theme of this part). The term ‘emotion’, in the psychological sense, was an early-modern coinage, but the thing that the term refers to, emotion, has been around since creatures with the requisite degree of mental sophistication came on the scene.87 Consider ‘psychology’ itself. Marjorie Nicolson notes that ‘it is curious that the seventeenth century, which invented so many new words for new ideas, did not happen upon the word psychology’ (1959, 285); but, she observes, they were doing psychology in the seventeenth century (and long before that, I add), despite the absence of the word. Stephen Halliwell speaks of the ‘psychological drama’ of Odyssey VIII (2012, 84), and it would be absurd to object that there could be no such thing until the word ‘psychology’ entered the language. Or take that workhorse of current criticism, ‘ideology’. This term was invented in modern times (the 1790s), but it is widely applied to older literature.88 The same goes, incidentally, for ‘literature’ itself, which is not an ancient term,89 but still subsumes ancient literature. Again, ‘optimism’ and ‘pessimism’ were first used in 1737 and 1759 respectively (in their French forms), but that does not mean that we cannot legitimately discuss the optimism or pessimism of Sophocles’ Electra, Virgil’s Aeneid, or Racine’s Britannicus.90 As I noted in the last section, it is a fallacy to suppose that concepts have no application to the world unless and until they come to be named. Goldhill tells us that German Idealism treated the tragic as ‘a way of exploring central questions of human freedom, political autonomy, self-consciousness, and ethical action’, and as ‘a means to comprehend the self as a political, psychological, and religious subject’ (2014, 310). But if his remarks on Vickers were right, not only could German Idealism do none of these things—for there would be no commonality between its concepts of freedom, etc., and ancient concepts—but Goldhill himself could not make these claims about German Idealism, either, because there would be no commonality between his, and German Idealism’s, concepts of freedom, etc. If indeed German Idealism had a concept of freedom: for how could we possibly know? How could we even make sense of the supposition that it did? So Goldhill’s remarks quoted at the beginning of this section, though purporting to undermine Vickers’s project, in fact undermine his own. What, then, is the upshot? Goldhill is right that Christianity, medieval chivalrous tales, Romanticism, Freud, and so on (the list could be considerably extended) have all had things to say about love, and that these

216  Tragedy and Linguistic Redress things are part of our cultural inheritance today. And he is also right that these conceptions of love differ, at least to some extent, from one another. However, in the first place, it is easy to exaggerate the differences, and Goldhill is by no means the only commentator to do so. For example, Malcolm Heath suggests that ‘philia is not, at root, a subjective bond of affection and emotional warmth, but the entirely objective bond of reciprocal obligation’ (1987, 74). But it is both, just as ‘love’ is both in modern English: one can, and many people do, love members of one’s family in Heath’s ‘objective’ sense, in the sense in which blood is thicker than water, but without affection. Secondly, as I have indicated, whatever the differences there is a common core of meaning—a common concept—that has remained constant over time and across cultures. Page duBois castigates conservatives for their belief in ‘an unchanging, fixed “human nature” ’ (2001, 58; note the usual scare quotes), but a few pages later she tells us that Helen ‘disregarded family values, abandoning husband and child, following . . . the one whom she loved’ (ibid., 69), which sounds like something of a human fixture to me. For evere it was, and evere it shal bifalle, That Love is he that alle thing may bynde, For may no man fordon the lawe of kynde.91 So Vickers is talking at a higher level of abstraction—or at a more fundamental conceptual level—than Goldhill in his objection. And there must be this higher (or more fundamental) level: that is guaranteed by the Davidsonian considerations of translatability. If there were no common concepts between us and the ancient Greeks, we could not translate their language. But then we should have no reason to call their ‘language’ a language at all, or to attribute rationality to them; nor should we be in a position to say all the many things about them that Goldhill wants to say, for example that they employed a concept philia that does not exactly map onto our concept love. The crux lay in my expression at the beginning of this paragraph of what was right in Goldhill’s position: I said that Christianity, medieval chivalrous tales, Romanticism, Freud, and so on, have all had things to say about love; and I added, in agreement with Goldhill, that their conceptions of love differed (at least to some extent) from one another. In other words, and focusing now on the italicized phrases of the last sentence, there is a common core to their several conceptions, and that core is correctly captured by our word ‘love’. Otherwise what Goldhill and I agree on would be false, or worse.92 The word ‘love’ refers, then, to a concept which is substantially common to Greek, Roman, Christian, and post-Christian cultures. When they sang their ballads medieval troubadours were not changing the subject, so that for us now to speak of their songs as being about (courtly) love would be just wrong, a mistranslation of the relevant words—‘amour’,

Theoretical Considerations 217 ‘Minne’, and so on. When Chaucer’s Criseyde declares ‘Ne love a man ne kan I naught ne may / Ayeins my wyl’, her words sing to us across the ages.93 That does not necessarily mean that what she says is true; and the sombre fact is that Criseyde will find that she can abandon Troilus.94 Against the idealism of Criseyde’s words we may set not just her author’s irony, picked up and accentuated by Shakespeare,95 but the cynicism (or perhaps just realism) of Mozart’s Così fan tutte, and of the thought, fed by the writings of La Rochefoucauld and Mandeville and Nietzsche, that love involves a good deal of egoism.96 But that thought, though perhaps a modern one, applies to the life and literature of all times: it applies directly to Dido’s love for Aeneas. And the idealism of love is still with us. One of the greatest products of the age of medieval courtly love, The Book of the Duchess, remains as fresh and vital as on the day it was written. The term ‘courtly love’ (‘amour cortois’) was an artefact of nineteenth-century scholarship,97 but, again, it can apply to any period: it makes perfect sense—whether or not it is true, though I think it is true—to say with Mitchell Greenberg that Racine’s Titus and Bérénice are ‘paragons of the most rarefied ideals of abnegation and courtly love’ (2010, 126). And we should not think that, at the end of the Duchess, ‘the gate to the garden of the Rose is closed forever’:98 it is open, in a new sense, at the beginning of T. S. Eliot’s Four Quartets, and it is always in some sense open, because, as C. S. Lewis remarked (2013a, 142), ‘we are still living, to some extent, under the influence of courtly love’. Vickers can, then, simply concede that there are differences between courtly love and other conceptions of love—and we can express these differences in (our) words, a crucial point noted above in the case of ‘kai’—but insist that courtly love is still (a kind of) love; ‘philia’ and ‘erōs’ may indeed be appropriately given different translations,99 but the concepts to which they refer are still species of love. After all, the English word ‘love’ itself is a term with a wide semantic range: it is not restricted to the ‘Judaeo-Christian notion of loving one’s neighbour or turning the other cheek’. For example, we may say that Alcestis’ servants love her,100 or that ‘the companions of Socrates are moved by a genuine love of him’.101 It is an extraordinary fact—to which I shall return in my last chapter—that we can say meaningful and even true things, in English, about a time and a place when and where the English language did not exist. But it is a fact. We note that Goldhill uses—and is right to use: it is a perfectly adequate term—the word ‘love’ in the phrase ‘courtly love’, when he writes about ‘the history of courtly love’: this already and on its own shows that his entire objection is misconceived. If his objection were sound, it would not be possible for him or anyone else to use the expression ‘courtly love’ meaningfully. But then, generalizing—legitimately, since there is nothing special about the words ‘courtly love’—it would not be possible to talk meaningfully about the past at all. And then it would not even be possible to say, meaningfully, that we cannot use our

218  Tragedy and Linguistic Redress word ‘love’ to talk about medieval courtly love. The position undoes itself in the by-now familiar way. Douglas Cairns argues that, though in the case of some concepts, such as, he suggests, the concept guilt, there may be no significant difference from culture to culture, since ‘it seems clear that the structure of guilt need not differ as a direct result of difference in the values which establish its proper eliciting conditions’ (1993, 9), there are other concepts of which this is not so. The Greek concept aidōs, often matched with shame, respect, or shyness (among other things), is, in his view, a case in point.102 He writes (ibid., 9 n. 21): ‘To recognize that we feel respect where a Greek might feel aidōs provides no warrant for the claim that we experience aidōs, since shame, embarrassment, shyness, and respect are many, their associations and their identificatory beliefs different, whereas aidōs is one, with one closely related set of identificatory beliefs’. But if the Greek word ‘aidōs’ really can be correctly translated in all these different ways (shame, embarrassment, and so on), depending on context, does that not imply that the concept aidōs is after all not one? If ‘shame’ and ‘respect’, say, with their allegedly different sets of associations and identificatory beliefs, can both be good translations, in suitable contexts, of ‘aidōs’, then that simply shows that the supposition that the concept denoted by this latter word has ‘one closely related set of identificatory beliefs’ is false. If, on the other hand, you wanted to hold on to the compactness and uniformity of the concept aidōs, then you would have to say that neither of these English words translated the word ‘aidōs’, in any context; and then, plausibly, no other English word would do so, either. But that, presumably, is not something that we would wish to say. It therefore seems much more satisfactory to accept that the concept aidōs is not uniform, that it embodies a multiplicity of sub-concepts, each keyed to different (though perhaps overlapping) sets of associations and identificatory beliefs. And then it will be possible for modern English speakers to feel aidōs, just as it was possible for them to feel Schadenfreude before that word was imported into English, even though there was then no single English term denoting that emotion.103 Indeed it will happen quite commonly that modern English speakers feel aidōs;104 only, we do not use a corresponding word because we make finer, or at any rate different, verbal distinctions between different varieties of aidōs, and we use words corresponding to these nuances. So aidōs maps onto several of our concepts (‘ours’ only in the sense that we have names for them); but, as I have already stressed in connection with the word ‘kai’, we can express all the necessary distinctions in English. We can talk about Greek aidōs even though none of our correlative terms precisely matches ‘aidōs’ in its semantic range. And the fundamental point is that we could not even recognize ‘aidōs’ as a word if we could not, at least in principle, translate it, and for that to be possible there must be commonality of

Theoretical Considerations 219 concepts: Greek aidōs must be something that we feel; and our shame, embarrassment, etc. must be things that the ancient Greeks felt.105 It may be helpful here, before closing this particular stretch of argument, to mention an objection that a reader has put to me, which I suspect may trouble others. This reader says: ‘Your tendency to polarize (or appear to polarize) the debate between “nothing in common” and “only superficial differences” misses the difficult middle ground which is where I spend a lot of my professional time’. The point that I am allegedly missing, I take it, is that there are differences between (say) ancient Greek culture and our own that are not merely superficial, but also not so deep as to block any understanding of that culture. My response is that this objection, in turn, misses the point. It mistakes the degree of abstraction at which I am operating; alternatively, it is merely terminological. ‘He spoke to his thumos’: does Homer’s use of this locution mark a superficial or a deep difference from us? Say what you like, so long as you accommodate the Davidsonian considerations I have emphasized. I say the former, and I would say the same of all the differences in that ‘difficult middle ground’ where my objector is professionally occupied. I do not, of course, add that superficial differences are not interesting or worth studying. But according to my way of deploying the surface– depth distinction, there are no deep differences between ancient Greek culture and ours: that is what the Davidsonian considerations secure. If, per impossibile, there were such differences, we would not, and could not, speak of ancient Greek culture in the first place—and there would be no such thing. So superficial differences between cultures are the only differences there are, and naturally they are worth studying, if anything is. Now, if you wish to deploy the terminology differently, and speak of deep differences between cultures, or of differences that inhabit a ‘difficult middle ground’, that is fine, so long as you do not permit this way of talking to tip over into assertions about incomprehensibility, radical otherness, etc. Within the category of what I am ranking as superficial, you can if you wish distinguish different levels of depth: so you might hold, for instance, that speaking to one’s thumos marks a deeper difference between Greek culture and ours than driving a chariot as opposed to a car. (I think it is probably, if anything, the other way round.) But the Davidsonian point is that what you are calling a deep difference exists, and can only exist, within a context of massive similarity and commonality. And I must admit to entertaining the suspicion that the reason why you are using the terminology in your way rather than mine is that you are overlooking this point, as it is so easy to do and as many critics do. The Davidsonian considerations invite us to take a step back and adopt a more abstract perspective—the perspective from which one asks how it is so much as possible for us to translate the word ‘kai’, possible for us to identify an ancient Greek culture at all. From that perspective, the

220  Tragedy and Linguistic Redress differences between ancient and modern culture that scholars investigate are (by no means uninteresting, but) vanishingly small.

31 Historicism and Relativism It follows from what has been said so far in this chapter that there are severe restrictions on the prospects for any kind of relativism about human nature, mind, and morality. In our engagement with ancient tragedy, for example, we are inevitably subject to what Goldhill calls ‘the drive towards transhistorical truth’ (2012, 165) in the plays’ discourse and reception. And not only inevitably, but also rightly subject. Goldhill no doubt intends his remark about transhistorical truth disparagingly, but he himself exemplifies the drive he mentions towards transhistorical truth by discussing in purely objectivist terms the meaning of the end of Sophocles’ Women of Trachis, a few pages before the above quotation (ibid., 159–61). Shortly before that he had spoken, without apparent irony, of ‘the whole and steady gaze of Sophocles’ (ibid., 156), a phrase that could have been penned by a Victorian critic, as though that were a transhistorical truth about Sophocles. Goldhill nevertheless stresses that he, and his relativism, are themselves products of his own culture, and he thinks that all critics should be aware of their historical locatedness (ibid., 227, 256–7). The Martindales make the usual genuflection to critics’ historical locatedness and the need for ‘accommodations’ in representing the past; but a few pages later they are quite sure that James Redfield is just ‘wrong’ to treat Hector as the hero of the Iliad (1990, 94–5, 99). Again, Christopher Norris accuses Leavis of ‘simply substituting one interpretation for another’ in his criticism of Bradley (1985, 64). And what is Norris himself doing? Obviously Leavis is substituting one interpretation for another: but he believes his interpretation to be better than Bradley’s, and Norris cannot in consistency deny that Leavis might be right, because in lodging his own claims Norris thinks that he, Norris, is right. These discomfitures illustrate the perennial problem with relativism, namely that in espousing it one immediately finds that the stance adopted is an absolute one. If relativism is true, that is not a relativistic truth, but an absolute one. The thesis of relativism has a general and a priori status: if it is advanced, it is advanced absolutely; and if it is true, it is true absolutely.106 So it is false. Bernard Williams once called this self-immolating absolutist relativism ‘vulgar relativism’.107 For example, if one reacts to moral diversity across cultures by saying that one should not criticize alternative moralities, one finds that the ‘should’ of this statement is an absolute one. Otherwise—that is, if the ‘should’ is merely r­ elative—one runs the risk that the moral outlook of my culture may be such as to permit, or even compel, me to make moral judgements about anyone and everyone. To get beyond this—to get into a position where relativists can censure their opponents on grounds of cultural imperialism, vel

Theoretical Considerations 221 sim.—they have to try to climb out of particular moral systems and adopt a purportedly absolute stance, that of the ‘cosmic exile’. But this stance is not only an absolute one, so inconsistent with relativism, but independently incoherent, for the cosmic exile, being outside all moral systems, would not have the wherewithal to understand, let alone make judgements, relativistic or other, about these systems.108 (In some later publications, Williams tried to outline a sophisticated form of relativism that would not be subject to his own refutation of vulgar relativism. But the attempted rescue does not work.)109 Goldhill writes about ‘the naïvety of the scholar who, in full denial of the historicity of reading and the anxieties of epistemology, merely asserts an objective, pure, and accurate knowledge of the past’ (2012, 258). I shall come to the supposed ‘anxieties of epistemology’ later on in this chapter; here I am concerned with the topic of historicity. Against Goldhill, I say that no one has ever been under any illusion about his historical locatedness; the recent stress on ‘the historicity of reading’ does not, in my view, convey an insight that eluded earlier ages, but embodies a philosophical mistake. Goldhill and many others are inclined to suppose that when scholars write, or imply, ‘This is how it was; what I’m telling you now is the truth about the past’, they are in denial about their own cultural embeddedness. But that is a simple error, and is inconsistent, as we have noted, with Goldhill’s own practice as a critic: for he, like everyone else (including the most relativistic of the relativists), is constantly staking such claims to truth. Of course the scholar who claims to know the truth about the past may be wrong; but that—like the fact that Romans build roads—goes without saying. Naturally scholars who make truth-claims do not think they are wrong: if they did, they would not make those claims. However, they know perfectly well that it is epistemically possible that they are wrong. But now, if you can be wrong about the past, you can also be right about it, and that is what we all think we are when we make our claims about how it was in the past, or about the correct interpretation of literature, ancient and modern. This is what Goldhill thinks too when he describes the pre-play rituals of Attic tragic performances, in an article I shall return to below in this chapter, and he does not there suppose that his own historical embeddedness invalidates his claim to be telling the truth. And there is, in fact, no tension between being historically embedded and making claims about the past that one takes to be, and that are, true simpliciter. Does that mean that there can, contra Goldhill’s assertion above, be an ‘objective, pure, and accurate knowledge of the past’? Well, if ‘pure’ here has the sense of unmediated, then clearly there cannot be pure knowledge of the past: all such knowledge is mediated by memories, reports, documents, and so on. But no one has ever doubted that or suggested otherwise. As far as objective and accurate knowledge of the past go, if these were stymied by the scholar’s historical embeddedness, then, as I said

222  Tragedy and Linguistic Redress above (§27), knowledge in the natural sciences would be impossible too: for natural scientists are culturally and historically embedded also. But we know that objective and accurate knowledge of the physical world is possible, and that we often have it—including knowledge of the state of the physical world at vastly different times and places from our own—so that any general attempt to infer lack of objectivity and accuracy from cultural and historical embeddedness must fail. And history and literary criticism are, after all, sciences in the broad sense: they are subject to what Goldhill calls ‘a commitment to certain rules of evidence, to shared standards of objectivity, to shared criteria of argumentation’ (ibid., 258). Goldhill suggests (ibid.) that this ideal is in thrall to ‘the notion of a direct or even unmediated engagement with the object of study’; but there is no such thralldom, nor has any serious investigator ever supposed our engagement with the past to be unmediated. It would be natural to object to my argument so far as follows. Surely, it might be said, when we read the scholars and critics of the past we are often struck by the thought that they are ‘of their time’, that their views are simply functions of their historical location—that, to put it bluntly, they are simply playing out the prejudices of their age, with maybe a few of their own into the bargain. Are not many of the judgements of past critics culturally specific, so that when their culture passes away the validity of their criticism is also a casualty? And then, once we have had that thought, we cannot help but apply it to ourselves. (At least, I say we cannot help but do so; yet it is extraordinary how many relativistic critics fail to notice the inference, and blithely proceed as though relativism applied to everyone but them.) I am burdened with the views and prejudices of my age, plus no doubt a few extra ones of my own. Does that specificity consist with the objectivism of my assertions about the past? Goldhill himself, in the discussion we have been tracking, in effect gives up on this problem: claims of objectivity and the historical embeddedness of the critic are both there, stand opposed to one another, and there is nothing to be done about it. Instead he turns aside and tries to resolve what he sees as a tension between historical embeddedness on the one hand, and claims, not of truth, but of value, on the other (ibid., 259–63). But this is a red herring. There is no even apparent inconsistency between the historical embeddedness of the critic and claims of value; no reconciliation is called for. Goldhill gets pleasure from reading the classics and finds them valuable. So do I and so do many others. That pleasure and that value are in no way even apparently threatened either by the historical embeddedness of the classical authors, or by our own. In contrast, there might at least appear to be a tension between critics’ historical embeddedness and their claims to tell the absolute truth; and that is what my objector was giving expression to at the beginning of this paragraph. The worry, I have said, is unfounded: if historical embeddedness implied the unattainability of objectivity in the case of literary criticism, or of

Theoretical Considerations 223 cultural studies more generally, it would also do so, by parity of reasoning, in the case of physics. But it does not do so in the latter case, so not in the former either. Newton was very much a man of his age—for example, he spent as much of his time on alchemy as he did on what we would call proper science—but that did not stop him getting some things right as well as other things wrong. What, then, should we say about the feeling that we often have when we read older literary critics, namely that their historical locatedness somehow invalidates their criticism? Well, we say that sometimes these critics’ being of their time does invalidate their criticism, but sometimes it does not. We say exactly what I have just said about Newton. But an illustration may help here. Why does Aeschylus’ Agamemnon do as Clytemnestra bids him and tread on the tapestries? Eduard Fraenkel famously offered two reasons: Agamemnon is ‘a great gentleman’, but he is also ‘worn out by the unceasing struggle’ (1978, vol. 2, 441–2). Critics have been harsh on him ever since he wrote the first of these remarks: R. P. Winnington-Ingram’s ‘Fraenkel’s “great gentleman” can perhaps be dismissed as the amiable eccentricity of a distinguished scholar’ (1983, 88) is one of the milder censures his view has attracted. Goldhill tells us that the interpretation ‘says far more about Fraenkel’s ideas of social interaction than about Greek ideas of gender or persuasion’ (1997c, 327). (As an aside, I note that Goldhill’s statement implies that he himself knows about Greek ideas of gender and persuasion, and knows that these English words fit ancient experience; otherwise he would not be in a position to tell us that Fraenkel’s ideas about these things were unsatisfactory.) Hugh Lloyd-Jones comes to the point: the first reason fails because ‘chivalry towards a lady . . . seems to be a medieval and a modern rather than an ancient concept’; and the second because ‘the psychological explanation that the King sees through his wife but is too weary to oppose her has a decidedly modern ring. It is a far cry from Aeschylus’ Agamemnon to Mann’s Aschenbach’ (1990, 294). So we are back with the charge of anachronism. Gentleman and world-weariness are apparently not ancient concepts. Fraenkel erred, it would seem, because he applied concepts that were of his (or a medieval) age and culture to another age and another culture, and that, we are told, cannot be done. Pelling’s tentative explanation of Agamemnon’s defeat by his wife might superficially appear to be less culturally loaded than Fraenkel’s: ‘Perhaps the force of Clytemnestra’s own personality is important, and Agamemnon is bullied’ (2005, 98). But are force of personality and bullying ancient concepts? How would you translate these expressions into classical Greek? Kovacs asks the same question about smugness (1987, 24). But using our concepts to talk about earlier ages is what we are all doing all the time: so how is any criticism possible? On this basis, we will be silenced before we can open our mouths. My answer to this quandary has been well anticipated earlier in this chapter. All these concepts—gentleman, world-weariness, force of

224  Tragedy and Linguistic Redress personality, bullying, smugness—do in point of fact apply to the ancient world. Lloyd-Jones’s claim that the notions of chivalry and ennui do not cross cultural and temporal boundaries is, like Goldhill’s similar claim about love, simply false. That does not mean that we have to endorse Fraenkel’s reading of the carpet scene in the Agamemnon: speaking for myself, I disagree with him about Agamemnon’s first motive (chivalry) but agree about the second (ennui). But the chivalry point is not mistaken for the reason that the ancient world knew nothing of chivalry, so that the concept just fails to have application. Of course, being a gentleman does not have the same superficial marks that that concept had in Victorian or Edwardian England, but there is a core of commonality, consisting of (among other things) ideals of good breeding, honour, integrity, courage, responsibility, and nobility which have remained essentially unchanged from the Homeric age to ours. It is true that birth and ancestry have become less significant over time (§27), but then there are many passages in older literature where genuine nobility is said to be a matter of behaviour, not breeding, so that either gentlemanly behaviour in those of low estate is recognized and praised—as, for example, in Euripides’ Electra Orestes praises the nobility of Electra’s yeoman husband (262)—or ignoble behaviour in those of high estate is disparaged. For ‘he is gentil that dooth gentil dedis’.110 There is nothing essentially Victorian, or for that matter medieval, about the ideal of chivalry. Nor is there anything essentially English: Fraenkel, an international scholar, was appealing to an international value, celebrated for instance in Le Roman de la Rose (18, 593–638), and Chaucer’s ‘verray, parfit gentil knyght’, though English, was an international type who told an international tale. Once again, the fact that ancient Greek does not have words corresponding to our ‘chivalry’ or ‘gentleman’ (if it is a fact)111 does not mean that the concept does not apply to ancient Greek culture, any more than the fact that the term ‘emotion’ was an early-modern invention means that before that point people did not have emotions. Martha Nussbaum mentions a common criticism of Aristotle’s deployment of the concept of megalopsuchia, to the effect that it has an excessively narrow application, to ‘a certain sort of Greek gentleman’:112 if the criticism is to make any kind of sense (whether or not it is fair), the concept gentleman must have application to Aristotle’s milieu. In Sophocles’ Ajax, as Gisbert Ter-Nedden remarks, Odysseus ‘possesses the aristocratic chivalry (aristokratische Ritterlichkeit) to acknowledge the personal heroic greatness of an enemy’ (2007, 350). When the same tragedian’s Neoptolemus stands frozen in his dilemma, and says ‘αἰσχρὸς φανοῦμαι’ (Philoctetes, 906), ‘I shall appear base’, he is precisely wondering what the noble—the knightly—thing to do is. What would a gentleman do? After all, It is contrair the lawis of nature A gentill man to be degenerate,

Theoretical Considerations 225 Noucht folowing of his progenitoure The worthy reule and the lordly estate.113 In Euripides’ Alcestis, when Pheres and Admetus wrangle, they ‘deteriorate from gentlemen to brutes’;114 Admetus’ butler does not initially think that Heracles is a gentleman, but it turns out that he is, and these are matters of Heracles’ behaviour.115 So when we read Fraenkel on Aeschylus there is no need for a parenthetical and apologetic ‘He was of his time’. Indeed he was of his time, as I am of mine and you are of yours,116 but in the first place being of one’s time does not stop one from finding out the truth about the meaning of past literature any more than it stopped Newton and Einstein from theorizing about gravity as a real phenomenon; and in the second place these two times—and indeed all historical times—are much closer to one another—culturally, mentally, morally—than the historicizing and relativistic theorists would have us believe. And we know this, because if it were not the case we should be unable to understand the languages of these other times, or even recognize what we find inscribed on (say) ancient papyri and medieval manuscripts as pieces of language. But we can so recognize them, and we can translate and understand them.

32 Character and Self in Literature and Life It follows from some of the considerations that we have rehearsed so far in this chapter that traditionalist critics are not wrong to talk in terms of the characters of tragic protagonists.117 Goldhill, echoing John Jones, thinks that the fact that Greek actors wore masks prevents a ‘focus on idiosyncrasies of personality’, unlike, say, ‘the dramas of Ibsen or even Chekhov’.118 And he suggests, in connection with Greek tragedies, that ‘criticism whose aim or method of explication is based on inclusive modern notions of “character” as a person’s whole personality seems to ignore the tensions both in the notion of the self, and between “character” and “discourse” in these plays. Indeed, there does not seem to be any notion in fifth-century Greek that quite reflects the inclusive idea of “character” ’ (1986a, 174). I shall postpone the issue of masking to the next section. By ‘tensions between character and discourse’, Goldhill refers, following Roland Barthes, to the fact that plays operate under internal and external constraints. On the one hand, action must be internally motivated, but on the other it must fit certain external parameters, provided by the relevant tradition (there is room for manoeuvre here, but it is restricted). There certainly is such a tension as the one which Goldhill mentions, but it is common to all literature (Barthes’s own example is a story of Balzac’s), and it has nothing to do with the extent to which the notion of character can be realized dramatically, or is ‘inclusive’. As for the ‘tensions in the notion of the self’, it is not clear what Goldhill has in mind, but he may

226  Tragedy and Linguistic Redress be alluding to Bruno Snell’s notorious claim that since Homer has no word for the self, he also lacks the concept: the Homeric ‘self’, for Snell, is a kind of Humean aggregate of diverse and possibly conflicting urges, motives, and desires. (The idea has been recycled by new historicists in respect of Renaissance culture.)119 I have here, by implication, rejected Snell’s position, on the basis that a culture may dispose of a concept for which it lacks a word; I argue against it in more detail elsewhere.120 So the fact, if it is one, that the Greek of the fifth century BCE lacked a word with the semantic range of the modern English word ‘character’ does not mean that that concept was unavailable to fifth-century Greeks, or is unavailable to modern consumers of fifth-century works. It may be that that concept is needed to account for what is going on in those works, even if the language that produced them lacked a corresponding label. In that case, the relevant culture did have the concept, only not a name for it.121 Both the literary-critical idea of characterization and the application of the English word ‘character’ to drama are relatively recent developments—they go back to the seventeenth and eighteenth centuries122—but that does not mean that they cannot be retrospectively applied: like ‘emotion’ and ‘obsession’ and ‘trauma’ and a host of other neologisms, they can. Further, we must ask: is it really the case that Aristotle’s word ‘ēthos’ does not correspond to our concept of character? Jones thought that ‘character’ in Aristotle’s Poetics meant ‘ethical colouring’ (1962, 32), and in her translation of the Poetics (Russell and Winterbottom 1972) Margaret Hubbard rendered ‘ēthos’ as ‘moral character’. But the epithet does not mark much of a restriction, if indeed any at all: in both literature and life all of character is morally relevant. Moreover, the remarks on character in Poetics 15 are consistent with a modern conception of character, and show that Aristotle did not limit the idea of character to moral character in any very narrow sense. Nor does he limit it to ‘explicit declarations of a man’s will and what he has decided to do or maxims that reveal his settled convictions’, as Blair Hoxby suggests (2015, 70): that confuses ēthos with dianoia.123 Hoxby also suggests that, in Shakespeare’s Antony and Cleopatra, ‘Antony’s tragedy, as he experiences it, is his inability to hold onto the carefully cultivated ēthos that he once projected to the entire world’ (ibid., 267), and he adds that ‘If we can speak of Shakespeare as inventing the modern self, or creating literary character avant la lettre, then moments like this in which a dramatis persona tries to make sense of himself using the classical conception of ēthos and yet finds it inadequate must be central to that achievement’ (ibid., 268). But we cannot speak of Shakespeare as inventing the modern self, or as creating—if by that is meant inventing—literary character avant la lettre. There is nothing new in the idea of an agent who loses his sense of self, who feels that he is not acting in character; that is exactly what befalls Sophocles’ Neoptolemus. Scholars often think that their age of specialization ‘invented’ the modern conception of the self—‘The

Theoretical Considerations 227 internalization of the self is one of Shakespeare’s greatest inventions’, writes Harold Bloom in Shakespeare: The Invention of the Human (409)124—but that conception has been in place at least since Homer, and probably long before that, as has the conception of the human. When I say that the modern conception of the self has been in place at least since Homer, I mean our ordinary conception of it, as a unified bearer of mental states and locus of decision-making. I do not mean ‘modern’ in the sense of any particular recent philosophical theory of selfhood, such as Descartes’s or Kant’s. On the contrary, I reject subjectivist and internalist accounts of the self and mind in the Cartesian tradition, according to which consciousness always involves self­consciousness, and the distinctive mark of consciousness is enjoying mental states that, as the phrase has it, are like something for the subject.125 So it is wrong to argue, as Joseph Russo does (2012, 23), that Snell can escape the criticisms of his position that I along with Arbogast Schmitt, Bernard Williams, and others have made of it by explicitly relativizing his critique of Homeric man to a ‘post-Cartesian and post-Kantian’ standard, and then finding an absence of that conception of the self in the Homeric poems. Agreed, Homer’s conception of the self is not that one; but then neither is ours. The ‘post-Cartesian and post-Kantian’ account of the self misdescribes it. Snell simply assumed the correctness of that description: my objections to him (as set out in my 2001) are directed as much against that assumption as they are against his interpretation of Homeric psychology. As I shall explore in a moment, Homer and his literary successors portray real people, and so naturally they deploy their—that is, our—ordinary conception of the self, namely as the seat of mentation and decision. Descartes did not discover, let alone invent, either the unity of the self or our conception of that unity, and it is the ordinary unified self that is presupposed to the literature of all periods: it is absurd to say that Corneille, in his Horace, was ‘espousing the triumph of the unitary Cartesian “ego” ’;126 Corneille, like every other tragedian mentioned in this study, depicts ordinary selves. Greek tragic characters are, as we know, to a large extent traditional figures, but the fact that ‘As literary inventions, . . . the characters of Greek drama draw on, define themselves through, and develop in relation to other texts’127 does not affect the issue; for exactly the same is true of the most intricately executed character in a novel by George Eliot or Thomas Hardy or Thomas Mann. Indeed an analogue of the same point holds of real people: we use language that other people have used, express hackneyed opinions, and so on:128 if intertextuality detracted from characterization in tragic literature, it would have to do the same in real life. In a similar way, to maintain that the large role played by rhetoric in tragic literature undermines the possibility of characterization would prove too much: for exactly the same holds of real people. Gould objects to Barrett’s psychologizing interpretation of Hippolytus’ rhetoric

228  Tragedy and Linguistic Redress (at Hippolytus 986–7) on the grounds that Hippolytus’ arguments ‘are precisely in the manner of the fictitious persons of Antiphon’s Tetralogies’ (2001, 102). So they are; but Antiphon’s fictitious persons are imitations of real people and the arguments that they make. Again, the fact that characters in a drama should not be considered separately from each other or from the plot, but regarded as part of an organic whole,129 also has its analogue in real life; for it is true of real people as well that their lives are not sharply isolated from their environment, yet we are able to work with a notion of individual character notwithstanding. When R. P. Winnington-Ingram asserts that, in Aeschylus, ‘It is not an individual as such, but an individual in society, that is characterized’,130 we may reply that that is scarcely worth saying, seeing that it applies to all character, real as well as literary. Individuals are indeed contextualized, in all sorts of ways: once that has been agreed on all sides—who needed ­reminding?—we can go back to discussing individual characters in literature (and life). It is not the case that ‘The essentialist-humanist approach to literature . . . depends upon the belief that the individual is the probable, indeed necessary, source of truth and meaning’.131 Meaning, insofar as it is a human product at all—a matter to which I shall return in Chapter 6—is a communal, not an individual, achievement. (Truth is quite different: it is, at least in general, settled neither by individuals nor by communities.) Marxist critics like to stress individuals’ social contextualization because they wish to play down the role of the individual in traditional literature: in particular they reject the Aristotelian focus on individual— and, as I have argued, corrigible—error, hoping instead to find that traditional tragedies are really about social injustice requiring to be rectified by collective action. So, for example, Peter Holbrook reads Hamlet’s melancholy as rooted in his awareness of political oppression; Denmark is a prison because it robs the people of basic liberties; Brutus and Cassius are the people’s heroes; Lear’s kingdom is disordered because it is full of real Poor Toms; and in general tragedy was inevitable in King Lear, because ‘Lear’s Britain was poorly governed’.132 Holbrook reads Lear’s and Gloucester’s remarks on redistribution of wealth as though they were the message of the play (2015, 52), not words spoken by individual characters in a dramatic context, and to be heard in that light. For Holbrook, Renaissance tragedy in general subserves a Marxist agenda (ibid., 82): Doctor Faustus is about triumphing over social disadvantage (ibid., 107), and so on. Similarly, in his book Shakespeare’s Universality, Kiernan Ryan goes through the plays extracting politically correct sentiments and imputing them to Shakespeare, ignoring which characters utter them and why, disregarding what the playwright’s purpose might have been in giving precisely those sentiments to precisely those characters at precisely that point in the play, and carrying on for all the world as though he were a nineteenth-century gentleman radical compiling what one might, with

Theoretical Considerations 229 a nod to the shade of Charles Lamb, dub ‘Marxist beauties of Shakespeare’. Shakespeare, it turns out, was opposed to capitalism (2015, 127); he was against globalization (ibid., 99, 127); one half-expects to be told that he deplored the creeping privatization of the NHS. Ryan’s reading of Shakespeare reaches its acme of distortion in his interpretation of Timon of Athens, where we are told that the play endorses Timon’s early naïve generosity (ibid., 124), whereas in fact, as we have said (§21), it brands it as folly. Characters in literature are indeed literary constructs, but when we state this obvious truth it is crucial that we go on to observe that, as such, they are representations (as) of human beings—real people—which guarantees that our response to them is, at least in part, a response (as) to real people.133 We can, and do, talk of Oedipus and Antigone, as I did in Chapters 1 and 2, in just the terms that we would use of real people, and that procedure is not only legitimate but quite unavoidable; if we did not do that, we would be treating the Oedipus Rex and the Antigone as an assemblage of sound waves in the air or black marks on the page. But literary works are more than that: they comprise words with meaning. Catherine Belsey tells us that ‘If psychoanalysis is right, the signifier is a good deal more than the instrument of a thematic content. Instead, it can wrench, seduce and unnerve us, make us cry, or motivate action’ (2003b, 30). But signifiers can only do these things if they are meaningful: bare sound waves or ink marks have, in themselves, no such power. So to say that we are moved by signifiers is only to say that we are moved by meaningful words, which is something we have known at least since Gorgias and Plato, and we did not need psychoanalysis to enlighten us. We can dissect the dilemmas faced by, for example, Senecan tragic figures in terms that are entirely comprehensible to us from real life (another application of transhistoricism).134 Real people can vacillate, and be inconsistent; so can literary characters. There is nothing in the idea of literary character which warrants Belsey’s remark that ‘an understanding of what it is to be human . . . is perhaps not best explained by the stabilizing and unifying idea of character’ (2003a, 86). That is a category error: the concept character, in one sense, is ‘stabilizing and unifying’; all concepts are, for they are abstract universals that collect instantiations. But those ­instantiations—the things that fall under a given concept—can be as unstable and disunified as the concept permits. So it is not a contradiction to talk about a disunified character in a literary work. Similarly, the concept of change is, qua concept, stable and unified; what the concept applies to—actual changes—are not. In his book The Authentic Shakespeare, Stephen Orgel describes a Damascene moment in his education (2002, 8): When I was a student learning how to deal with literary texts, William Empson vanquished the sentimental hordes at one blow by pointing

230  Tragedy and Linguistic Redress out that we cannot dispose of ‘As flies to wanton boys are we to the gods: / They kill us for their sport’ ([King Lear] 4.1.36–7) by remarking that it represents the view of the crude and prosaic Gloucester, rather than that of the sensitive and perceptive Cordelia or Lear. Plays, Empson observed, have linguistic and poetic structures, and characters are not independent of those structures. The notion that the gods make fools of us is one that runs throughout King Lear, and is not dependent on the view of any particular character. Characters, that is, are not people, they are elements of a linguistic structure, lines in a drama, and more basically, words on a page. But particular words are given to particular characters rather than to others. In some cases this assignment may be arbitrary, but in most cases it is not. If that were not so, there would no difference between the characters of Cordelia and Regan, or Edgar and Oswald. Ideally, words are assigned to speakers fittingly, and that means that there is such a thing as character, that characters are not just ‘words on a page’: the words have meaning, and ideally those meanings will add up to a coherent c­ haracter—even if that character’s nature is to be incoherent—distinct from other people represented in the drama. (Interestingly, the words spoken by Euripidean figures to gods and goddesses who appear ex machina—and this is a place, if any, where one might expect mere formality—are always spoken in character.)135 It is only on the basis of these and similar considerations that textual emendation is possible. Miranda’s words in The Tempest ‘Abhorrèd slave. . .’ (I, 2, 353ff.) are often reassigned to Prospero; Orgel rejects the reassignment on the basis that ‘This is a very clear case of the character being considered both prior to and independent of her lines’ (ibid., 12). But the character is independent of any particular lines; the text as a whole is independent of particular lines, even though it is made up of particular lines. That is what enables textual criticism, as A. E. Housman pointed out (1989, 335–6). In emending a single word or line of an author, one holds the rest of the author’s work fixed. You could not simultaneously emend everything an author wrote, just as you could not simultaneously reassign all of Miranda’s lines on the basis of an assessment of her character. But it makes good sense to reassign any given line, holding the rest fixed. A hermeneutic circle operates, but once you are inside the circle you can stand on some parts of the structure in order to refashion other parts.136 Miranda’s character is a construct of the words given to her, but in acknowledging this point we must, as often, beware of the compositional fallacy, which lurks round so many intellectual corners: we must remember that the whole is greater than, and—more fundamentally—simply different from, the parts taken one by one, or even in aggregate. The words that are given to Miranda form a whole—a representation of a real person with a particular personality. That enables

Theoretical Considerations 231 us to say, for example, that a particular line that the received text assigns to Miranda is not in character, and then reassign it to another speaker. After the passage quoted above, Orgel concedes that ‘It is, of course, very difficult to think of character [as just ‘words on a page’], to release character from the requirements of psychology, consistency and credibility’ (2002, 8). It is very difficult, because it is wrong: it involves overlooking the meaning of the work. The muddle that accounts for the mistakes in Orgel’s thinking about character then emerges clearly from his subsequent remark that the difficulty he has admitted ‘is arguably a difficulty that drama itself accepts, . . . When Coriolanus angrily rejects Rome with the words “There is a world elsewhere” (3.3.136), he imagines a space outside his play, a world he can control. . . . The remainder of the play is in every sense designed to prove him wrong, designed to prove that no character can escape his play’. But the remainder of the play is designed to show no such thing, and we can make that point vivid by reflecting that Coriolanus could not escape his play whatever happened in it. Even if Coriolanus had found his words to be true (as in an obvious sense they are),137 and the story had ended well for him, he would still not have escaped his play. Everyone in a play is in a sense trapped within its parameters, but that is irrelevant to the question whether, as represented people, they are trapped within their represented characters and existence. The point is that Coriolanus is trapped within his personality, as most of us are. There is indeed a world elsewhere, but he is not able to exploit its possibilities and avoid tragedy, because ‘character is fate’, because he carries his personality with him wherever he goes. The fact that he, like every literary character, is trapped within the work in which he figures is a point that operates at a quite different level of discourse, and it is a category mistake to confuse the two levels. (Orgel would have seen this if he had noticed that his final point had nothing at all to do with Coriolanus as opposed to Falstaff or any other literary figure.) The fundamental error is the confusion of representing medium, which in the case of drama is composed of words, and what is represented, which is a reality that includes people with characteristic dispositions and mentalities. Drama is a portrayal of mindedness, and mindedness is something real. In its general form the confusion to which I am drawing attention pervades our intellectual tradition: the sense-datum theory of perception is a good case of it.138 Or take early-modern discussion of the dramatic unities. Insistence on the unities of time and place—even, sometimes, to the extent of requiring the represented time to last no longer than the representing time, and the represented place to occupy exactly the same amount of space as the stage—has often been grounded in the very confusion in question.139 What of Goldhill’s claim, with which I began this section, that the modern notion of character is, unlike the Greek, ‘inclusive’? His idea is that the word ‘ēthos’, unlike ‘character’, does not refer to ‘a whole

232  Tragedy and Linguistic Redress personality or the make-up of a psyche, but rather a particular disposition or set of attitudes that can be seen to be instantiated in a particular course of action’ (1990a, 102).140 This question connects with the moral nature of character, for the idea that ‘ēthos’ in Greek tragedy is expressed at moments of actions that realize dispositions is very close to the idea that ‘ēthos’ is moral character. My view is that focusing on a purported ‘inclusiveness’ in the modern notion of character—set against a purportedly narrower concern in ancient thought with dispositions instantiated in action—exaggerates the distance between ancient and modern conceptions of character, and establishes a false dichotomy: for us, too, character, in real life as in literature, is a matter of dispositions to behaviour, and is revealed in action.141 Moreover, even if we concede that our word ‘character’ expresses an ‘inclusive’ notion of character which the word ‘ēthos’ does not, the inclusive concept is a mere generalization of a concept that was available to fifth-century BCE dramatists and is contained at least virtually in their plays. This is another application of the point that the absence of a word for a concept in the language of a community does not necessarily signal the absence of the concept; it may be there implicitly, and we can name it, even if the members of the community did not. (In principle, as I have said, they could have named it too.) We may grant that ancient dramatists are not interested in all aspects of (inclusive) character: they are interested in character as expressed at crucial moments of decision and action, the ones they highlight in their plays.142 But that is true of modern dramatists too. It certainly does not mean, as some critics have held and as is implied to be Empson’s view in the Orgel passage cited above, that we should retreat to the position promulgated by Tycho von Wilamowitz in his book on Sophocles (‘Tychoism’),143 namely that what is said by a figure on the tragic stage is determined not by considerations of consistent characterization, but solely by the exigencies of the plot at that particular point in the drama. The fact that ancient dramatists are interested in character as expressed at crucial moments of decision means that A. M. Dale was wrong to argue, as she famously did, that tragic speeches illuminate the ‘rhetoric of the situation’, rather than character.144 They do both: choices that tragic figures make to (say) argue in one way and not another are revealing of character.145 Nor does it mean, as Goldhill implies at one point, that we can never ‘arrive at the truth of a character’s motivations in a literary text’ without ‘treating the character as a “real person” ’ (1986b, 165). No doubt there are difficult cases, and in some examples, such as perhaps Clytemnestra’s reaction to the news of Orestes’ death in the Choephori (the case Goldhill is considering), a character’s motivation may be indeterminate: the work in question may not yield enough for us to arrive at clear verdict on the question of his or her motivation. But, equally, there are many cases where this is not so, where we do have enough to make secure judgements about characters’ motivations; there are also

Theoretical Considerations 233 cases where the truth is determinate, but we have not yet determined it. Nothing in the process of arriving at such truths requires us to treat characters as real persons in any objectionable sense: that is, we are not required to confuse the categories of real life and stage drama.146 We are, as I have said, required to treat literary representations as representations (as) of real people, but that is unobjectionable and indeed mandatory, since representations of real people is what literary representations are.

33 Masking and the Mind John Jones famously argued that the conventions of the Greek tragic stage, in particular the wearing of masks, meant that there was little room for subtle characterization, and he placed heavy emphasis on Aristotle’s dictum in the sixth chapter of the Poetics that characterization exists for the sake of plot, not vice versa (1450a20–5).147 In consequence, following Hegel and Kierkegaard he maintained—in a comparison that has been often repeated—that Greek tragic figures have, unlike Hamlet, no inner life.148 (Schiller remarked in a letter to Goethe that ancient tragic characters were ‘ideal masks and not real individuals’.)149 Quoting (actually slightly misquoting) a familiar passage from Macbeth, Jones wrote (1962, 45–6): The mask, as a casual survey of masking cultures makes plain, can present all manner of versions of the human self; it is almost inexhaustibly rich in its presentational modes. But it is vulnerable at one point. It cannot maintain itself against the thought that all presentational modes are inadequate to the truth. The situation in which Duncan reflects: There’s no art Can find the mind’s construction in the face destroys it, for the mind’s construction must be found there if it is to be found at all. And when it cannot be found there, masking becomes pointless. Now normally the mind’s construction is shown in the face (and, generally, in behaviour): that was a commonplace of ancient (especially Stoic) theorizing about the emotions.150 It was forgotten in the early-modern period under the influence of Cartesianism, and had to be recovered by Wittgenstein in his later philosophy. So far from its being the case that ‘all presentational modes are inadequate to the truth’, as Jones suggests, it is rather, and necessarily, the case that normally presentational modes are adequate to the truth (I shall return to this point shortly). What is correct is that the mind’s construction is not always shown in the face, which is

234  Tragedy and Linguistic Redress enough for Shakespeare’s purposes; deception is possible. (A nice example is afforded by the picture in The Rape of Lucrece: compare the depictions of Ajax and Ulysses with that of Sinon: 1396, 1506–7.) But deceit only works, as we have said (§26), because it is exceptional. Duncan can only be fooled by Cawdor, and then by Macbeth, because normally he would be right to take facial and other behavioural indicators at (as we aptly say) face value. The implication of Jones’s point seems to be that masking would only make sense if facial deception were an impossibility, or were at any rate not acknowledged to be a possibility. But ancient tragedy does exhibit scenes in which one character deceives another, and several ancient stage characters might have echoed Duncan’s very words. The actor playing a deceitful character did not wear a mask that represented deception: obviously he did not do that, for that would have undermined the character’s strategy; the whole point of deception is that it should not be visible in behaviour. Equally, it would be misguided to say that the idea of a deceitful face makes no sense on the ancient tragic stage. Does it then make no sense, either, for the spectator of a modern drama who is sitting right at the back of the auditorium and cannot see the actors’ faces? And if we equip this spectator with opera glasses, does the idea suddenly make sense again?151 The truth is that masks make no difference to interiority one way of the other; you could, for example, play Hamlet—or indeed any drama—in masks without detriment to the meaning of the work. For the interiority, where it is present, is present in the words. Jones’s point ignores the extent to which classical drama, like all drama, is—as I am stressing throughout this chapter—linguistic. And, as Bernd Seidensticker says, ‘language as the means of unfolding in great detail and subtlety what it is that drives a person to act in the way he or she acts is not seriously limited by the special conditions and conventions of the ancient theatre’.152 So masking presents no bar to the depiction of interiority in ancient theatre. And it is depicted: one thinks of Sophocles’ Electra (as well as his Ajax, Philoctetes, and Neoptolemus), and of Seneca’s Clytemnestra.153 If there is a difference here between ancient and modern drama—which I am very far from allowing, but if there is a difference—it is one of degree, not of kind.154 In the previous section I placed some emphasis on the distinction between the representing medium and what is represented. Even in modernist plays that seem to trade on blurring the boundary—as for example in Anouilh’s Antigone or Pirandello’s Sei Personaggi in Cerca d’Autore— the distinction remains: it is just that the actors play characters for whom the distinction is blurred.155 What this means in the present context is that, for example, Electra’s investigation of Orestes’ footprints in Aeschylus’ Choephori can be perfectly satisfactorily portrayed by shod actors: it is the usual category mistake to say, as A. F. Garvie does (on 209), that ‘The passage makes sense only if Orestes, Pylades, and presumably

Theoretical Considerations 235 Electra, wear no boots’. Booted, even buskined, actors can perfectly well represent bare-footed characters in the fiction. The same point applies to Agamemnon’s command to a maidservant, in the carpet scene of the Agamemnon, to remove his boots. Fraenkel writes (on 944) that ‘There can be no doubt that the king’s command is immediately obeyed and that the untying of his shoes actually takes place on the stage’;156 but how does he know? Shoe- (and later buskin-) wearing conventions no more preclude the portrayal of bootlessness than did the fact that Cleopatra was played by a squeaking boy on Shakespeare’s stage preclude his portraying her as a sexually active woman.157 Similarly, to return to the phenomenon of masking, when Aeschylus’ Electra says to Orestes ‘O face [literary: eye] delightful to me!’ (Choephori, 238–9), the fact that ‘it is on her brother’s beloved face that Electra dwells’, as Garvie comments (ad loc.), does not show that the actor who played Orestes was not wearing a tragic mask, or took it off at that point. In Euripides’ Phoenician Women Jocasta tells Eteocles to stop glaring fiercely (454), and it would be silly to object that there is nothing he can do about the expression of his mask.158 Sophocles’ Electra is told by Orestes, when they go into the house to carry out the murders (Electra 1296–7), to quell her joy and compose her features into a suitable expression of grief; and in general reference to facial expressions is relatively common in ancient tragedy.159 When Sophocles’ Oedipus and Euripides’ Polyphemus are blinded, their altered appearance is noticed by the chorus (Oedipus Rex 1297–1306; Cyclops 670), but, as Richard Seaford observes (on Cyclops 663), the chorus’ remarks ‘do not prove that tragic masks were altered or changed, because the words may have been designed to stimulate the imagination of the audience’. As we have said, it is all in the words: the cataclysm at the end of the Prometheus Vinctus, and the storms of Lear and The Tempest, are in the words; they do not depend on stage machinery and are not necessarily enhanced by it (or indeed by the latest audio-visual technology). Again, the fact that on the Greek tragic stage everything is played outside, in front of the skēnē, does not mean that Greek tragic characters enjoy no privacy, and that everything that takes place before the skēnē is represented to occur in a public space.160 This point holds quite independently of the vexed question to what extent the ekkuklēma—and the exostra, if that was d ­ istinct—were employed in Greek and Roman theatre to represent indoor scenes.161 Sometimes this space is meant to be public, sometimes not: we have to look at each case individually. Conversations with, or in the presence of, the chorus or confidant(e)s, do not necessarily diminish privacy:162 again, it depends on the particularities. In Lope de Vega’s El castigo sin venganza (I, 237–46, 347–50), Frederico insists on being alone, but being alone apparently does not preclude the presence of his friend Batín and several servants as well. We all slip into the category mistake on occasion when taking about theatre, often in quite trivial ways, as when Horace tells Peleus and

236  Tragedy and Linguistic Redress Telephus—as opposed to the actors playing them—that ‘if you wish me to weep, you must first weep yourself’,163 or when Patrick Finglass (on Ajax 915–16) says of the grief-stricken Tecmessa, standing over the body of her husband Ajax, that ‘she could remove [her veil] and cast it on the dummy’, where he should say either that the actor could cast the veil on the dummy, or that Tecmessa could cast it on Ajax’ body (the veil is a constant between representing medium and represented reality: it represents itself). Finglass makes a similar mistake in commenting on Electra 1475, where Aegisthus, tricked into removing the shroud from Clytemnestra’s body, says ‘Alas, what do I see?’: Finglass quotes Lewis Campbell’s note (ad loc.), where we are told that ‘after a glance of horrified recognition at the corpse, Aegisthus looks strangely on Orestes’, and Finglass adds that ‘ “strangely” should not be pressed, as Aegisthus is wearing a mask’. It is not Aegisthus who is wearing a mask, but the actor playing him. Aegisthus does indeed look strangely on Orestes: that is contained in his words, in particular ‘Who are the men into whose trap I have fallen, unluckily?’ (1476–7, tr. Lloyd-Jones). It is not Imogen who ‘smears herself in stage blood’,164 when she awakes next to Cloten’s body, but the actress playing her. Nor is it Imogen who ‘grieves in anguished poetry’,165 for the poetry is Shakespeare’s, not Imogen’s. Likewise, we might say, there is no paradox when Hamlet, whose soliloquies are made of the sublimest poetry, appends to the verses that he sends to Ophelia ‘I am ill at these numbers’ (II, 2, 119). The sublime poetry of the soliloquies is Shakespeare’s, not Hamlet’s: Shakespeare is a great poet, Hamlet a mediocre one. Sophocles’ Agamemnon complains that he cannot understand Teucer’s barbarian tongue; Garvie counters that Teucer ‘has been speaking normal Greek’ (2016, 14). The actor has, to be sure—but Teucer? Mario Erasmo remarks that ‘The relationship between onstage and offstage reality is one of coexisting and competing realities: onstage reality is the reality of the actors on the stage, which does not refer to or acknowledge the reality of the audience’ (2004, 4). Here again we encounter the mistake, for the reality of the actors is exactly the same as the audience’s reality; it is the represented reality, which concerns neither stage nor actors—unless we have to do with a play about a play—that is different. Metatheatre, which Erasmo goes on to discuss, bridges the gap between the realities; but theatre is not, as such, metatheatre. (Note that a play about putting on a play, such as Sheridan’s The Critic, is not, as such, metatheatrical: a play can be about putting on a play without showing awareness that it is itself a play.)166 But the slips to which I have drawn attention in this paragraph do not matter in themselves. What matters is when critics build theory on a foundation that incorporates the category confusion. What I have in mind with this last point is well exemplified by the constant reminders to which we are subject, and which I mentioned in the last section, that we must not ‘treat the characters as real people’ or ‘forget that they are merely literary constructs’.167 As I insisted there,

Theoretical Considerations 237 though it is true that dramatic characters are literary constructs (and I am not sure that anyone needed reminding of this fact), they are not merely that: they are representations (as) of real people, so that our engagement with them will not be a mere response to literary constructs, either, but also to the content of those constructs; that requires us to react to literary characters, at least to some extent, as we would to real people.168 So to say, with Russo, that our reply to the question ‘Can Homeric heroes make real decisions?’ must be that ‘fictional characters can only make fictional decisions’ (2012, 27) is to miss the point: the decisions made by fictional characters are, to be sure, in one sense fictional—everything in fiction is fictional in the sense that it is part of the fiction (which does not necessarily mean that it is contrary to fact).169 But fictional characters are representations of real people, and their fictional decisions representations of real decisions. Again, to complain, with L. C. Knights, that ‘the bulk of Shakespeare criticism is concerned with his characters, his heroines, his love of Nature, or his “philosophy”—with everything, in short, except with the words on the page, which it is the main business of the critic to examine’ (1964, 18), is to engage in false dichotomizing: the words on the page are endowed, as we have said, with meaning, and when we understand that meaning a whole world opens out in which there are characters, heroines, philosophy, and much besides. (Note that Knights later resiled from his earlier and more famous view that we are not interested in the question of Lady Macbeth’s children.)170 In fact what we find is that critics and theorists who speak of dramatic characters as ‘constructs of language and gesture’, or who subordinate characterization to the exigencies of rhetorical tropes, then inevitably psychologize the characters whose status they officially wish to downgrade, and they do this because there is simply no other way to understand the stage action.171 Alternatively, and looking at the matter from another angle, having agreed that tragic characters are ‘constructs of language and gesture’, we can ask: what of it? Critics and theorists always present this as a discovery, as though it took us somewhere interesting; but it does not, first because, as I have been arguing, we must insist that, as constructs of language and gesture fictional characters are representations of real people; and secondly because a parallel point applies to real people. Real mindedness consists of language, gesture, and other sorts of behaviour. When Felix Budelmann and Pat Easterling write that ‘the question of a person’s inner nature—what he or she is “really like”—can only be answered experientially’ (2010, 299), that is correct, though if the ‘only’ is meant to insinuate that the experiential route is an epistemic pis aller we should reply that an experiential route to mindedness is not only as good as it gets but also as good as it (logically) could get. Mindedness lies in behaviour; it is not hidden behind it.172 As I have just implied, the mistake that is so often made in discussions of literary and real character is to fall into Cartesianism about the mind.

238  Tragedy and Linguistic Redress This view is remarkably prevalent among literary critics and theorists,173 who tend to think of behaviour as a kind of façade or intermediary that disguises the subject’s mind from observers. In the case of stage characters, it is supposed that there is nothing behind that façade—or nothing relevant, for we are not interested in the actor’s mental states, at least not in this context—whereas in the case of actual people like you and me the idea is that there is a mind, conceived as a private arena to which the subject has privileged, and infallible, access. Contrariwise, when critics reject this picture of mindedness, they often conclude (holding the Cartesian assumptions in place) that there is no such thing as the self,174 instead of concluding, as one should, that it is the Cartesian picture of that self and its mindedness which is mistaken.175 So long as that picture is in place, it seems intelligible to say that, whereas with ordinary people we may guess at what is behind the façade (and guessing is all that, on the Cartesian approach, we can do), in the case of stage drama there is no point in making such guesses, for there is nothing (relevant) there, and no work for an interior mindedness to do (for we are not interested in predicting the characters’ off-stage behaviour). But it should go without saying, nowadays, in the wake of Wittgenstein’s later work on mind (‘the human body is the best picture of the human soul’: 1977, II, §iv), that this Cartesian approach to mindedness is awry. As I have said, mindedness lies in behaviour, not concealed behind it: and that applies just as much to stage representations as to real life. A tragic character is a representation (as) of a real, minded person, with all that that entails. Moreover, treating the self, against Descartes, as essentially embodied does not upset the Bradleian conception of plays as portraying agents who act out their characters.176 The fact that, as Knights wrote, Shakespeare’s dramas employ words and tropes and conventions ‘to which, without some training, we are no longer accustomed to respond’ (1964, 16)—and the point obviously applies quite generally in the tragic tradition—makes no difference here; for, once again, a parallel point holds of real life, since all behaviour (and especially linguistic behaviour) requires training if we are to respond to it. The language of tragedy is indeed a Kunstsprache, and demands acclimatization;177 but that is true of all language. The prevalent descent by critics and theorists into Cartesianism is typically abetted by a misconstrual of the phenomenon of deceit—the point to which Shakespeare’s Duncan alludes in the passage from Macbeth cited above. This is a traditional epistemological syndrome: the existence of deceit, illusion, hallucination, and so on—au fond, we are talking about the existence of falsehood—is thought to push us towards a scepticism about other minds and the socalled ‘external’ world. In fact, as I have hinted a couple of times, the opposite is the case: the existence of deceit and illusion precisely depend on the fact that we indeed have direct access to other minds, and to the world. In general, if you get it wrong it must be possible to get it right, and (Davidson’s point again) getting it wrong in any particular case rests

Theoretical Considerations 239 on a bedrock of, by and large, getting it right, because if you do not by and large get things right your thoughts will not enjoy the determinacy of content necessary for you to go wrong (for us to interpret you as going wrong) on isolated occasions: sufficiency of success secures content, and it is only when content is in place that failure (falsehood) becomes a live (but necessarily sporadic) option.178 Critics constantly repeat the undergraduate error of supposing that the fallibility of a given method of acquiring knowledge is in tension with the objectivity of its subject matter,179 whereas in fact it presupposes that.

34 Political Context and Audience Response All this does not mean that we are entitled to speculate about the lives of dramatic characters in a way that goes beyond what is licensed by the play, in the way that Bradley has often been accused of doing with Shakespeare.180 But everything depends on the question what is licensed by the play, and here we must judge each case on its merits: there is nothing of a general nature to guide us. We must beware of reaching for facile principles that might seem to offer us guidance but in reality do not. Consider an example. Could Sophocles’ Oedipus have escaped his doom if he had been more careful? After he had received the Delphic oracle, should he not have formed an unwavering intention to kill no man old enough to be his father and marry no woman old enough to be his mother? I implied in my discussion of Chapter 1 that the answer to these questions is both affirmative and highly relevant to our assessment of Oedipus’ role in bringing about his tragedy. Dodds, however, dismisses them as a misplaced attempt to judge the action of the play by inappropriate standards taken from the law courts and real life. We are not entitled, he says, to blame Oedipus for failing to be more careful after he received the Delphic oracle, for failing to form the above intention, and for failing to exercise self-control, since no such possibilities are mentioned in the play, or even hinted at; and it is an essential critical principle that what is not mentioned in the play does not exist. These considerations would be in place if we were examining the conduct of a real person. But we are not: we are examining the intentions of a dramatist, and we are not entitled to ask questions that the dramatist did not intend us to ask.181 We need not be distracted by Dodds’s allusion to the dramatist’s intentions: let us just remind ourselves that all that can legitimately be meant by talk of those intentions (and, in fact, all that is usually meant) is the meaning of the play before us.182 So the question is this: is Dodds right that what is not mentioned in the play does not exist, and what does this principle actually amount to? Here we might start with an ad hominem

240  Tragedy and Linguistic Redress point: if we look at what Dodds himself tells us is part of the meaning of the Oedipus Rex we find that it includes not the identity of the chorus with Theban elders, as the play seems to indicate, but their extra-­ theatrical identity with Athenian elders. For in the second stasimon they ask why they should dance if men lose all respect for the gods, and on that passage Dodds remarks: ‘If by this they merely mean “Why should I, a Theban elder, dance?” the question is irrelevant and even slightly ludicrous; the meaning is surely “Why should I, an Athenian citizen, continue to serve in a chorus?” ’ (1973, 75). But the identity of the chorus with contemporary Athenians is not mentioned in the play. That seems to indicate that Dodds would have to concede that his principle is, at the very least, not exceptionless. Dodds would reply by saying, as indeed he does, that there is a convention that the chorus may ‘step out of the play into the contemporary world’.183 So he might say that the chorus’s identity with Athenian elders, while not mentioned in the play, is at any rate (by convention) present. But if the chorus’s Athenian identity can be present in the play, though not explicitly alluded to at any point, then it hardly seems a stretch to say that Oedipus’ failure to think through his position in the light of the oracle and exercise self-control is equally present in the play in such a way as to affect our view of his culpability. For the play is, as I have said, portraying real people in a real world, a world in which there is law and justice and blame and guilt. And in portraying that world the artist perforce exploits many features of the actual world that he does not—indeed cannot—mention, but which he expects his audience simply to read over into the dramatic action.184 Does Oedipus eat, drink, and sleep? That is to say: is Oedipus portrayed as someone who eats, drinks, and sleeps? Certainly he is: he is represented as a human being, and human beings have to do these things. They are not mentioned in the play, but they are presupposed, for otherwise Oedipus would not be represented as human. (There remain indeterminacies—extra-textual matters that the work does not settle, even by implication.)185 Oedipus’ killing of Laius in a rage shortly after he had received the Delphic oracle speaks for itself: we do not need to be told what to make of it.186 I am not asserting that Dodds’s principle never applies. That would go too far in the other direction. There are occasions when we are right to make claims having the general form ‘Such and such is not mentioned in the play and it is just irrelevant to it’. I have myself employed something like Dodds’s principle on several occasions already, for example against Hugh Lloyd-Jones’s view that the ancestral curse is functional in Sophocles’ Oedipus Rex (§3). In Chapter 2 I argued that Polyneices’ treason, though not mentioned in the Antigone in just those terms, is highly relevant to that play, while the possibility that Polyneices might be buried outside Thebes is both not mentioned and irrelevant. The point is simply that, like all other maxims of literary criticism, Dodds’s principle

Theoretical Considerations 241 has to be used with judgement; it cannot be erected into an exceptionless rule, or deployed as some kind of conversation-stopping algorithm. For one thing, it may be unclear what exactly is mentioned in the play; for another, even if we can agree that something is not mentioned, there may still be room for argument over the reason for the omission. Perhaps it is omitted because, although it is important, it is so familiar to the audience that it does not need to be mentioned: I suggested in my discussion of Oedipus Rex (§4) that this would apply to the omission of the Sphinx’s riddle. Or something may be omitted because it is simply obvious, as with (I have argued) Oedipus’ folly and Polyneices’ treason. (We might add Agamemnon’s sexual interest in Cassandra, which so discomfited Fraenkel.)187 Or a point may be omitted precisely in order to draw attention to it:188 maybe the audience has been trained by the tradition into expecting mention of a particular fact, so that when it fails to materialize they inevitably ask why the author has omitted it; and there may be a good literary purpose in having them ask just that question. Shakespeare’s Venus and Adonis perhaps supplies a case of this, in the highly suggestive omission of any mention of Adonis’ birth and parentage.189 Lessing’s Emilia Galotti is modelled, as we have seen (§25), on Livy’s story of Virginia, but in Livy (and Webster) the rape of Virginia precipitates political revolution, whereas Lessing ends his tragedy with Emilia’s death, and politically everything appears to revert to the status quo ante, as it does also in Grillparzer’s Ein treuer Diener seines Herrn (where there is an uprising, but not against the king). But is the omission of anything paralleling Livy’s revolution significant—subtly subversive, perhaps?190 Then again, it must be possible for an omission or deviation from what is expected to be genuinely without significance. Each case, as I have said, must be judged on its merits: you cannot stop scholars from arguing about the point of the double burial in the Antigone by saying that ‘because no reason is given in the text, no reason exists, and the question is an irrelevant one that ought not to be posed and cannot be answered’,191 for it may be the case that there is a good reason for the double burial which Sophocles’ original audience would have recognized immediately, without needing to be told, but which is dark to us. Silence may be just that, silence, with no further significance; but it may, instead, speak louder than words (a point to which I shall return in Chapter 5). In fact it has been satisfactorily shown, by Scott Scullion, that Dodds was wrong to claim that the ‘Why should I dance?’ of the chorus in the Oedipus Rex is extra-theatrical. The question is by no means irrelevant in the context of the drama, as Dodds opines: in the cultural context presupposed by the play, old men do dance, and dancing is conceived as a response to divine prosperity.192 This issue connects with the question, which has been much discussed in recent criticism of Greek tragedy, to what extent the dramas should be understood politically, say as expressions of Athenian democracy. I cannot enter this debate fully here, but

242  Tragedy and Linguistic Redress a few remarks are in order. If, with Colin Macleod, we define a political interest as ‘a concern with human beings as part of a community’ (1983, 28), then evidently Attic tragedy—indeed virtually all literature—is political.193 The dispute concerns the extent to which politics more narrowly conceived is relevant. In some contexts it surely is. For example, it has been suggested that, in the Oedipus Rex, the chorus’s identification of the plague-sender with Ares rather than Apollo has a contemporary Athenian significance;194 an allusion to contemporary Athens has been detected in the characterization of Oedipus;195 and it is common to see in the Theban plague a reference to the plague of Athens in 429 BCE. Again, it is plausible that Euripides’ Suppliants contains covert attacks on contemporary Sparta and praise of contemporary Athens.196 Not that these examples are uncontroversial; but most instances one could cite are contested, and reputable critics disagree. Is there a reference to the Sicilian Expedition at Euripides, Electra 1347–8? J. D. Denniston says there is; Martin Cropp says there is not.197 It is not always obvious that noticing political allusions enhances our understanding of a play; such facts often seem to be mere curiosities, like the fact that, as Aristotle tells us (NE 1123a23–4), the chorus at Megara wore purple. To take an example from later drama, Bérénice’s speech ‘De cette nuit, Phénice, as-tu vu la splendeur?’, which I shall cite in part in Chapter 6, no doubt contains a passing homage to the magnificence of Louis XIV’s court, but does that fact have a bearing on our understanding of the central dilemma of the play?198 Recent critics who have argued for a more politicized construal of Attic tragedy are usually not interested in whether this or that line contains an allusion to contemporary events: they are after something much more significant. At least, that appears to be what they are after, but their claims are often so nebulous that it is hard to make out what the implications really are. Assertions that we should not lose sight of the political context of Attic tragedy often lead nowhere. (They are reminiscent of the literary theorist’s tired assertion that ‘everything is ideological’: yes—so what?)199 Some think that the pre-play rituals of the City Dionysia, for example, must be relevant to the plays themselves,200 and Aristotle is criticized for ignoring the political and ritual context of tragic performance;201 but what exactly is the relevance? In an article on civic ideology (2000), Goldhill describes these rituals, but the rest of the article does nothing to indicate how they affect our understanding of the plays. Undoubtedly some preplay rituals are relevant to some plays: for example, the parade of orphans has a bearing on the theme of Euripides’ Suppliants. But were all the pre-play rituals always relevant?202 Here I agree with Scullion that a large degree of autonomy was granted the tragic poets, and that, in general, the immediate cultic context of performance was irrelevant to the content of Greek tragedies.203 The same applies to Athens’s democratic constitution. Certainly democracy is relevant to some tragedies, such as Aeschylus’ and Euripides’ Suppliants (though as P. E. Easterling notes, the former play is

Theoretical Considerations 243 still predominantly heroic in outlook);204 but is it always relevant? Tragedy, as several critics have shown, is located in the community as such, at least in general, rather than specifically in the democratic polis.205 It is true that the polity of the heroic world is in some sense a Gegenmodell of democracy: in Edith Hall’s words, the portrayal of ancient Argos and Thebes ‘offered a confirmation of the propriety of the political system that had replaced monarchy and tyranny in democratic Athens’.206 But this point can be overpressed: the locating of plots in heroic Argos or Thebes could have a telescoping rather than distancing effect; when something is presented explicitly as other, we see it clearly, and then we may realize that it is actually not particularly other after all.207 Further, as I have said, the democratic frame of Attic tragedy does not, at least in general, have detailed political repercussions in the plays. It is too readily assumed that if the frame of Greek tragic performance had certain properties, the content must also have had those properties; but this inference is a familiar fallacy.208 Is tragedy an ‘eminently democratic form’, as Page duBois thinks (2014, 305)? (Interestingly, the claim that great literature is associated with democracy has ancient roots.)209 No: how can it be? Tragedy has developed in very diverse political contexts, including under the notably undemocratic governments of Elizabeth I and James I, not to mention the still more authoritarian regime of Louis XIV, the various princedoms of the Holy Roman Empire, and nineteenth-century Germany and Austria. This point is especially relevant when we are confronted with the argument that the parrhēsia of tragedy embodied specifically the ideals of the democratic polis.210 Tragedy represents deliberation, which is, as I have noted, a human universal:211 it has nothing to do with any particular kind of polity, and deliberation is the subject matter of tragedy in Homer, in Virgil, in Seneca, in Chaucer, in Garnier, in Marlowe and Shakespeare and their contemporaries, in Corneille and Racine, in Lessing and Goethe and Schiller and Kleist and Hebbel and Grillparzer, none of whom wrote under remotely democratic conditions. And, conversely, not all democracies produce tragedy. Democracy is neither necessary nor sufficient for tragedy. William Allan and Adrian Kelly state that, in Greek tragedy, ‘the heroes’ ultimate ruin is shown to be a result of their own errors and their flawed society’ (2013, 98), but, as I have already implied more than once, there is very little emphasis in the plays on the latter. The audience is meant to go away thinking not that democracy is better than autocracy, but rather that intelligence is better than stupidity. Neil Croally suggests that a transhistoricist approach to Greek tragedy misses its detail and specificity (2005, 68–9). This is indeed a risk; but there is also an opposite risk.212 Critics who study only the Greek corpus and pay no attention to later tragedy are liable to miss the commonalities, and to interpret as specific what is in fact universal. The claim that the Greek tragic chorus represented the democratic polis213 is especially weak. The groups they portray are too marginal for

244  Tragedy and Linguistic Redress that: it is striking how, in the surviving plays, the composition of the chorus is different from the presumptive make-up of the audience, who would be mostly, if not exclusively, adult male citizens, many of military age, whereas the chorus regularly represents women, slaves, and old men.214 (The question whether women were present in Athenian audiences has been much discussed.215 Respectable women, we may surmise, did not attend.)216 The fictive identity of the chorus is settled by the poet, and it is their fictive identity that matters to the drama, not the status of the actors playing the members of the chorus.217 It is true that, for all their marginality, the chorus members sometimes speak with authority, but the authority with which they speak, when they do, is the general authority of tradition;218 it is nothing so precise and localized as the authority of the fifth-century democratic polis.219 Compare here Racine, who was raised in the shadow of the Port-Royal, and critics, especially Lucien Goldmann and George Steiner, have accordingly hastened to dub him a Jansenist and Augustinian tragedian (§21); but many of the features to which these critics point are in fact part of the common inheritance of the Western tragic tradition; if you wanted an exposition of the doctrine of original sin, you would go to Augustine, not Racine.220 Similar remarks apply to the alleged Athenocentrism that many critics have discerned in Greek tragedy.221 Some plays are focused on Athens (the Eumenides being the most obvious example); but the surviving tragedies are, in general, as little Athenocentric as Shakespeare’s major tragedies are centred on London,222 Racine’s on Paris, or Schiller’s on Weimar.223 Aristotle does not talk about Athens, or democracy, or the polis in the Poetics, because he realized that these things are not of the essence of tragedy.224 He gave us the doctrine of hamartia, and he also downplayed the roles of both the gods and the polis. In all these respects, I believe, his treatment of tragedy was not only accurate to what he knew, but also prescient and profound. Greek tragedy did not have to be internationalized;225 it was through and through international already; it did not have to be universalized or transhistoricized, for it traded in the universal and the transhistorical already. The basic problem for those who wish to insist on the democratic Athenocentrism of Greek tragedy is its extraordinary Nachleben in so many different cultural and political contexts: by their lights, this should not have been possible.226 But do all these different audiences react in the same way? Do all the members of a single audience react in the same way? Aristotle tells us that the members of the tragic audience feel pity and fear, but do they all do so? It has become fashionable in recent decades to stress the heterogeneity of audiences, and the heterogeneity of their responses to tragic (and other) performances.227 We know next to nothing about the reactions of individual audience members until we reach modern times, but critics are nevertheless quite sure that Athenian audiences, say, would have exhibited a range of different responses to tragedy. At least, that is the official

Theoretical Considerations 245 position of many writers, though in practice they often inconsistently revert to speaking in terms of uniform audience response. So, for example, Robin Osborne first tells us that ‘No two members of an audience will have heard exactly the same play. Any attempt to stress the political relevance of one aspect of a tragedy’s plot will inevitably prove reductive’ (2012, 276), but a couple of pages later we are assured that ‘The issue in Antigone was an issue in political theory, but no one in the audience can have thought it an issue in theory alone’ (ibid., 278). So at least on that one point it appears that there was a uniform audience response. Gold­ hill inveighs against ‘the trivial rhetoric with which so many critics have continued to use the imagined audience as a bastion for their own opinions’, and the ‘naïve and univocal idealization of the audience as a single and instant body’ (2012, 40–1), but a few pages later, discussing Sophocles’ Electra, he tells us that ‘As Clytemnestra turns the [Paidagogus’] words into an omen, she is quite unaware of what sort of an omen she is ­ratifying—and for the audience, recognizing the frisson of such a twist on the power of words, the irony produces a self-conscious awareness of the doubleness of language, an awareness which is the defining characteristic of an audience’ (ibid., 48). So it appears that the audience uniformly sways to the rhythm of Goldhill’s ‘doubleness of language’—which here, as often, just boils down to the fact that some sentences are true and some are false: we have to do with the usual Sophoclean dramatic irony, for the omen (‘Orestes is dead’) will turn out to be false. Indeed, in the quoted passage Goldhill tells us that the audience not only does react uniformly to ‘doubleness of language’, but that it is defined by its so reacting. If defining the audience to be just those playgoers who agree with your interpretation of the play is not using the imagined audience as a bastion for your own opinions, I am not sure what is. A page later Goldhill has recalled his official position and reminds us not to ‘assume audience response to be homogeneous in the theatre’ (ibid., 49). But then he forgets it again: we are subsequently told that ‘The image of the armed woman is not a comfortable one for the Athenian imagination’ (ibid., 243), the Athenian imagination being, presumably, something that all Athenians have and bring with them to performances. What are we to make of this? Clearly no one is going to deny that theatre audiences from Thespis’ day to ours have been composed of different people with different backgrounds and different interests, both across time and instantaneously. But what bearing does this fact have on our interpretation of drama? Critics like to stress the heterogeneity of audiences as though that point in itself got us somewhere, but it does not (an instance of a general syndrome that we have observed before): everything turns on what you do with it. Well, what do critics do with it? So far as I can see, not much. For example, in the introduction to his book The Language of Sophocles, and in the main text, Felix Budelmann repeatedly invokes the heterogeneity of audiences,228 but what he

246  Tragedy and Linguistic Redress actually wants to argue, like everyone else, is that Sophocles’ text carries a certain meaning as an objective feature of it that is there for everyone. Budelmann is talking, therefore, about how audiences ought to react to the plays, about how the ideal spectator or reader, of any age or place, reacts to them. His discussion of Clytemnestra’s defence in the Electra, for instance (2000, 66–71), proceeds entirely in terms of what the text means, and so in terms of an ideal response to it. The recurrent insistence on audience heterogeneity turns out to have no functionality in the argument; it is just window dressing. The fact that, as Budelmann insists, the plays contain indeterminacies of meaning makes no difference here: ideal spectators simply factor those indeterminacies into their understanding. Similarly, the fact that a work contains undercutting elements makes no difference to the point at issue: it does not connect organically to audience heterogeneity. In this connection Pelling mentions the figure of Euripides’ Medea: her feminism is unnerving but is undercut by her own problematic status (1997, 220–1). She asserts that she would rather stand in the line of battle than give birth once (Medea 250–1): ‘how could every male spectator’, Pelling asks (ibid., 221), ‘respond [to this line] in the same way?’ But the fact (if it is one) that a text contains undercutting elements is an objective fact about that text, so that audience members should respond in the same way, namely by registering those undercutting elements. It is not as though the work were addressing two (or more) different claques in the audience, feminists and misogynists as it might be, and saying one thing to the one group and quite different things to the other, like the guest speaker at a school prize-giving who spends half the speech addressing the parents and half of it addressing the pupils. If the Medea is a feminist work, then it is a feminist work; if its apparent feminism is, rather, compromised, then it is compromised. These facts are there for everyone, and so an audience member who fails to pick up on, say, ironic or undercutting elements in the play that really are there has not fully understood it. The same point applies to the Thebes-as-other motif that I explored earlier in this section. Pelling’s idea (ibid., 228–9) is that some audience members might react in one way (‘This is ancient Thebes, not modern Athens, so has nothing to do with me’), and others in the opposite way (‘This is really about me’). Pelling is right that actual audiences may have reacted, and might still react, in such diverse ways, but that is not the end of the enquiry: we have to ask what sort of reaction to a play set in ancient Thebes is the one that the work warrants. And, as I implied in my earlier discussion, the thought that heroic Thebes is in some respects a Gegenmodell of contemporary Athens—even the patriotic thought that we modern Athenians do things much better in our glorious democratic city than they did in the old days in that awful autocratic Thebes—is perfectly consistent with the thought that really, profoundly, the Antigone (say) is still, despite its remote setting, a play

Theoretical Considerations 247 about me, because Sophocles is not ultimately concerned with this or that polity, but with human relations and human nature, and these things are a fixture. In sum, of the various possible responses that one might envisage in an actual audience, those that are good responses will be factored into an overall understanding of the play by an ideal audience, which might say to itself, adapting the two responses offered above: ‘It is true that this is a play set in ancient Thebes, not modern Athens, and there are differences to reflect on in that regard—maybe things that we Athenians can learn, maybe other things on which we can compliment ourselves— but the playwright is really talking about commonalities of human nature across space and time; inflexible Creons and transgressive Antigones and bluff sentries are found everywhere, not just in primeval autocracies; so it is about me’.229 Accordingly, to return to Aristotle, his claim that audiences will react with pity and fear to tragic performances is a claim about the ideal audience. And talk about the ideal audience is, as I have indicated, just another way of talking about the meaning of the work before us.230 Aristotle’s view is that the tragic work and the tragic performance are such as to warrant, to make appropriate, responses of pity and fear. The twin ideas of the meaning of a play and of an ideal audience are normative ones that point at one another in the way specified: that is, the meaning of a play just is what an ideal spectator gets out of it, and vice versa. It is in this way—and it is only in this way—that Goldhill’s appeal to an ‘awareness which is the defining characteristic of an audience’ can be accepted. For actual audiences, being (as we are constantly and superfluously reminded) heterogeneous, cannot be so defined: actual audiences include all sorts, those who understand the play and those who do not, those who are paying attention and those who are checking their smartphones or (in fifth-century Athens) worrying about the price of corn. Only the ideal audience is defined by its grasp of the meaning of the work. And, to stress the point again, different features of the work— including its ironies and self-undercuttings—are not discrete components that are separately targeted at distinct factions in the audience, but are part of its overall meaning; and the audience is invited to grasp, and respond to, that whole meaning.

35 Communication and Deconstruction Following, in particular, the work of Vernant and his associates, a cultural meme has established itself in recent decades in the criticism of Greek tragedy, which holds that its agents and actions are presented, in Vernant’s words, ‘as problems that have no solution, enigmas whose double meanings always remain to be decoded’.231 Goldhill, for one, has promoted a doctrine of pervasive ambiguity in tragic language; and in his hands it becomes a general deconstructionist thesis of the instability and

248  Tragedy and Linguistic Redress indeterminacy of meaning in Greek tragedy.232 Consider some of Goldhill’s comments on justice (dikē) in the Oresteia. He first states that ‘different characters at different times appeal to dikē as a criterion, support, or reason for action. Clytemnestra, Agamemnon, Electra, Orestes, the Erinyes, Apollo, Athene and the various choruses . . . each claims dikē to be on his, her or their side. Each character appropriates dikē to his or her rhetoric’ (1986a, 46). So far so unexceptionable. But Goldhill then approvingly cites Vernant, who says: ‘On stage, the different heroes of the drama make use of the same words in their debates, but these words in the mouths of different characters assume opposed meanings’.233 It is an implication of what I have already said in this chapter that this thesis is mistaken. It is true that there is, as has often been observed, a general movement during the course of the Oresteia from a governing concept of retributive justice to one of legal, procedural justice, and that the word ‘dikē’ serves both of these purposes.234 But it is wrong to say that, as between the contesting parties that Goldhill lists, there are shifts in the sense of this word matching the dramatic oppositions: this commits an error that we have already diagnosed (§26). If you and I have a disagreement and I say ‘This action is just’ and you contradict me with ‘No, this action is not just’, the word ‘just’ does not undergo a shift in sense from my utterance to yours. Indeed it cannot do so, as we have said; otherwise there would be no disagreement. For disagreement to get off the ground, we must agree about the meanings of the words we use, and the disagreement then concerns a difference over the truth of our respective utterances. Misunderstanding of this point stretches back through Leibniz—who thought that if everyone could agree on definitions (if they spoke his projected characteristica universalis, in which the meanings of terms would be perspicuous), all political and military strife would cease—to antiquity. In the Phoenician Women, Euripides has Eteocles say: ‘If everyone defined justice and wisdom in the same way, there would be no quarrelling or strife among men. As things stand, the only similarity or equality mortals show is in their use of words (onomata): the reality to which these refer is not the same’ (499–502, tr. Kovacs). But if we only had the words in common, there would be nothing for us to quarrel about: for disputes about justice and wisdom to arise, we must agree, at least broadly, on the meanings of the words ‘just’ and ‘wise’—the concepts to which they apply—and the disagreement can then concern the extensions of these concepts (though, as we have said, only around the edges). So the Leibnizian ambition, if politically laudable, was philosophically back to front: it is only because we agree about the meanings of words that disputes arise. There is no urgent project to get us all to agree on definitions: we already do that, and still fight wars. And Eteocles’ words are in fact refuted by the scene in which they occur (446–637), where he and Polyneices argue about the justice of Polyneices’ claim. The scene

Theoretical Considerations 249 comprises skilful repartee, with frequent cappings of the interlocutor’s last statement. It is a model of successful, gripping communication: each participant knows exactly what the other is saying. In the case of the concept of justice, as between the contesting parties of the Oresteia there is broad agreement that justice is fairness—a matter of treating people as they deserve and of getting what you deserve.235 Even the acknowledged shift from retributive to legal justice preserves this core of meaning in the world ‘dikē’:236 the issue is still one of fairness and desert. (Even the surprising metaphor in the claim of Euripides’ Theonoe to have ‘a great temple to justice in my nature’ preserves it.)237 The disagreement over, say, the justice of Clytemnestra’s killing of her husband concerns what Agamemnon deserved, which is a real dispute, not a verbal one, and one that can be sensibly debated.238 When in Sophocles’ Electra Clytemnestra says that justice killed Agamemnon (528), she means that he got what he deserved; in that play she and Electra argue vigorously about the justice of Agamemnon’s killing, but ‘the principle of justice to which they both appeal is the same, while their applications of it are irreconcilable’.239 Similarly, the debate between Odysseus and Agamemnon at the end of the Ajax over the burial of Ajax concerns the question whether it would be just to bury him, and that is again the question what Ajax deserves (cf. 1342): do his recent treacherous acts outweigh or cancel his past services and his acknowledged stature as a hero?240 On Goldhill’s approach, as on Vernant’s, the contest between Clytem­ nestra and Orestes in the Choephori essentially involves a problem of communication. But these two protagonists understand one another well enough—Goldhill’s own analysis of their crucial exchange (1984a, 178– 83) indeed amply illustrates this point—and their dispute concerns truth, not meaning. Vickers also accepts Vernant’s approach to tragic ambiguity, and draws relativistic conclusions from the existence of moral disputes: ‘Evidently claims of “law” or “justice” are to be evaluated according to who makes them’ (1973, 27). In that case, one wonders, what is the problem? If Vickers is right about moral dispute in tragedy, then, as we have said (§26), it is no real dispute at all, so why all the angst? This action is just-for-me, which is perfectly consistent with its being unjust-for-you. Goldhill refers us (1986a, 47) to Vernant’s claim that ‘What the tragic message communicates, when it is understood, is precisely that there exist in the words exchanged by human beings zones of opacity and incommunicability’.241 Apparently tragedy as a genre is able to get its message across (at least to Vernant) even when its characters, and presumably the rest of us too (apart perhaps from Vernant and his readers), are stuck in zones of incommunicability. But, waiving the self-contradictoriness of Vernant’s claim, it is mistaken, at least in general, to diagnose failure of communication in tragedy. In the Oresteia, for example, there is no ‘undercutting of the security of communication’ or ‘questioning of the ideal of Justice’, as Goldhill wants (1986a, 77). All the parties sign up to the ideal of justice

250  Tragedy and Linguistic Redress (there is no need to capitalize the word: that is mere mystification) in the basic sense of ‘justice’ sketched above: what they dispute is which actions and states of affairs are just. There are, in literature as in life, particular cases of communicative breakdown; but in both areas they are, as I have said, necessarily exceptional. There are cases in tragedy as in other kinds of literature and as in real life where a crucial moment of decision is affected by the ambiguity of key terms employed, although, once again, they are not the usual case. Goldhill writes (1986a, 134–5) of Phaedra’s reflections on the meaning of the word ‘aidōs’ in Euripides’ Hippolytus (385–7): Phaedra . . . discusses the ambivalence of another key term, aidōs. Now I do not intend to look at the various arguments as to what she might mean by this ambivalence, but I do want to consider the vocabulary by which she expresses her sense of the doubling of language. I use my own translation (385–7): . . . There are two sorts of aidōs; the one is not evil, the other a burden on the oikos. If the right measure were clear, then there would not be two things having the same letters. The ambiguity in moral language is expressed as the ambiguity in the signifying form of the letters: one set of letters for two ideas. This is what hinders the clarity of the ‘right measure’, of ‘hitting the target’. Goldhill refuses to enter the debate about the meaning of ‘aidōs’ on Phaedra’s lips. But unless we work out what Phaedra means by ‘aidōs’, how can we be sure that this is a case of ‘doubling of language’? Well, Phaedra herself says it is such a case, perhaps. But if her words are subject to ‘ambivalence’, can we understand what she says about ‘aidōs’ in the first place? How did Goldhill know how to translate the Greek? If, now, we assume for the sake of argument that we do know the meaning of Phaedra’s words, and that Goldhill’s translation is correct, the point for Phaedra, as for Euripides, is precisely that we do not have to do with a general linguistic phenomenon of ‘doubling’, but, rather, that ‘aidōs’ is a special case. If a general phenomenon were in question, there would be no reason to pick on the word ‘aidōs’, and Phaedra’s remarks would lose their force and interest; but the case of ‘aidōs’, while no doubt not unique, is highly distinctive, and that is what prompts Phaedra’s reflections on it and elicits our engagement. More fundamentally, if a general phenomenon were in question, it would not even be possible ab initio to interpret Phaedra’s words as saying something determinate about ‘aidōs’, such as that it is ambiguous.242 Goldhill applies his ‘doubleness’ idea also to Phaedra’s accusing letter, which brings about Hippolytus’ death: ‘The assumed certainty of the written word leads Theseus to utter the irreversible curse, the spoken

Theoretical Considerations 251 word which bears no retraction. And it will be the very uncertainty, the very doubleness of letters as a signifying medium, which will lead him to wish back his curse in vain’ (1986a, 135). But there is no such doubleness. The problem with Phaedra’s message to Theseus is not that the words she employs are ambiguous or opaque. On the contrary, her message is only too clear: she accuses Hippolytus of sexual assault. The problem is that this charge is false (as Goldhill himself notes at 1986a, 234). Just that, and no more. And in order to be false, the words must be determinate in meaning. There are, to be sure, places in the play where we might say that communication breaks down, for example at 1032–5, where Hippolytus tells his father: ‘What the fear was that made [Phaedra] take her life I do not know, for it would not be right for me to speak further. But she showed chastity, though she could not be chaste, while I, who could, have used it to my hurt’ (tr. Kovacs). Hippolytus’ words are inevitably obscure to Theseus.243 But communication only ‘breaks down’ here in an attenuated sense. Hippolytus is allusive, he implies that he knows more than he says, and Theseus lacks the background knowledge, possessed by the spectator, which would put him in a position to see what his son hints at. But the words are comprehensible, and on one level Theseus understands them well enough. He simply does not know what the point of saying them might be, and in what way they might be true. To take a further example, Ajax’ ‘deception speech’ does not, as Goldhill avers (1986a, 197), point towards the deceptiveness of language per se. Language is not in general deceptive—as we have seen, there are theoretical reasons why that could not be so: a generally deceptive language would undercut its own status as a language, and generally deceptive uses of language would undermine themselves as uses of language, since it is an a priori constraint on language use that most declarative sentences we utter are true, and that deceptive uses of language are parasitic on non-deceptive such uses. If it were in the nature of words to equivocate,244 there would be no meaningful words—no language—in the first place. Equivocation, like deception, is necessarily an exceptional status. Further, as in Phaedra’s and Hippolytus’ cases, we rob Ajax’ particular speech of its interest and power if we try to make its special nature the condition of all language. What exactly is going on in the deception speech is a notorious crux, which I cannot handle in detail here: my own view is that Ajax intends no deception, that the agenda is suicide throughout (this is all but explicit at the beginning and especially at the end of the speech: see 687–9), but that he is so absorbed in his own world and thoughts that he fails to realize that some of his words could be taken in a different way from the way he intends.245 Consider, too, a Shakespearean example. A natural use of the expression ‘doubling of language’ would make it refer to hendiadys, where a single idea is expressed by two words or phrases. The doubling is in one sense redundant, but in literature it often has a function. An outstanding example is afforded

252  Tragedy and Linguistic Redress by Hamlet, which contains an unusual preponderance of such doublings. Why? It has been plausibly suggested that the frequent use of hendiadys underscores the theme of delay.246 It is as if the play were stammering,247 as if there were ‘a kind of collective desperation to all the hendiadys in Hamlet’.248 But, again, we would spoil the special effect that Shakespeare achieved if we held that language as such stammered, or despaired of ever quite reaching its goal. (Another such special effect is the doubling of language in Plautus’ Amphitruo, where it subserves the doubling of characters in the plot.)249 Similar remarks apply to the role played by riddles in the Oedipus Rex,250 where Goldhill characteristically banalizes that play’s dramatic irony by generalizing it: ‘It is always all too late that we learn fully the significance of the language we use and hear’ (2012, 35). No: it is hardly ever too late. That is what makes the few cases where it is too late interesting. The inconcinnity between Goldhill’s remarkable ability to translate Euripides’ Greek, on the one hand, and the alleged elusiveness and instability of meaning, on the other, is replicated across the board in deconstructionist and postmodernist writings. Peter Rose reports enthusiastically Derrida’s ‘insistence that there is no access to a “real” outside the process and play of signification’.251 Now there is a positive way of reading this Derridean assertion, which I endorse, and shall argue for in Chapter 6. On that reading, and jettisoning the scare-quotes round the word ‘real’, the claim is one to the effect that there is access to the real— the really real—but only through language (the real is constitutively what can be talked about in language). But in so taking the claim we must leave aside the usual Derridean implications of the instability and unreliability of language, and the insinuation in the scare-quotes that the ‘real’ to which language gives us access is not really real. Rose himself clearly intends a more negative reading of Derrida than the one I endorse, for shortly after the cited passage he is found associating ‘the play of signifiers’ with ‘the elusiveness of fixed meanings’ (ibid., 62). A page after that, and as though Derrida’s name had never been dropped into the conversation, we are treated to this description of the background to the Ajax (ibid., 63): The award of the arms of Achilles was presumably the central narrative focus of the lesser epic treatments and, as far as we can tell, of the first play of Aeschylus’ lost trilogy. Sophocles begins his play after the contest has been decided and after Ajax’s mad attack on the cattle. Thus there is no opportunity for direct debate or a full dramatic juxtaposition between opposing political, physical, military, or psychological types. Sophocles chooses the version of the award decision that attributes it to a vote of the Greek army. This makes the decision a direct judgement by the community about its own leadership. By the same token such a decision is open to charges of corruption,

Theoretical Considerations 253 envy, or stupidity. In Sophocles’ version the charge of bias is made and denied, but the question is left open. Given the elusiveness and instability of meaning, one can only marvel at Rose’s ability to extract these facts from our ancient texts. (They are meant to be facts, are they not? And Rose is claiming to know them, is he not?) To take the floating signifiers served up by our sources and extract such a rich yield of determinate and known truth is a quite stunning achievement, on a par with Goldhill’s serendipity in hitting on the word ‘love’, out of all possible finite permutations of letters, to translate ‘philia’. Irony aside, we must invert the whole procedure. The facts that Rose recites are facts, we know them, and so, unwinding, there must be fixed meanings after all; otherwise we could not have construed the ancient documents on which Rose bases his assertions. Rose shows by his practice that he accepts this. The giveaway in the above passage is that seemingly inconsequential phrase ‘as far as we can tell’: by this Rose does not mean, as you might expect from his official Derridean posture, ‘as far as we can tell, given the elusiveness of meaning’; he means what any conservative critic would mean by it, namely ‘as far as we can tell, on the basis of the evidence available to us, and given the fixity and stability of meaning’.252 The points I have made in this chapter against Hammond, Foakes, Vernant, Goldhill, and others might seem merely verbal. Surely, an objector might say, if literary critics want to talk about instability of meaning in cases like the use of ‘dikē’ in the Oresteia, or ‘justice’ in King Lear, that is a perfectly licit way of proceeding. After all, we do in ordinary parlance speak of a ‘breakdown in communication’ in cases where disputants understand each other well enough—so there is no breakdown in the literal sense—but differ profoundly over the disputed issue. In reply to this objection I say that if these critics meant the phrase ‘breakdown in communication’ in that anodyne sense, then I would have no argument with them. If they glossed ‘breakdown in communication’ as, for example, Richard Rutherford does—‘one character will not listen to the other, or cannot understand the other’s point of view, or cannot bear to remain in the other’s presence’ (2012, 15–16)—there would be no problem, and we could all agree that, in that sense, ‘communication frequently breaks down’ (ibid.) in the tragic corpus. In that sense, no doubt, it is right to speak of a breakdown in communication between Antigone and Creon, or among the main agents of the Philoctetes, or between Hippolytus and Theseus.253 And in the idiomatic sense of such phrases as ‘We were just speaking different languages’, according to which it signalizes something like breakdown of communication in Rutherford’s sense, perhaps we can even say that Oedipus and Teiresias, Creon and Antigone, Ajax and Tecmessa, and Heracles and Deianeira speak different languages.254 But critics are constantly tempted to overstretch the idea. These latter examples of ‘speaking different languages’ are dubious by my lights, and

254  Tragedy and Linguistic Redress it certainly seems to me to go too far to speak of a ‘breakdown in communication’ in the murder scene of Sophocles’ Electra.255 Electra hears Clytemnestra, we do not know whether either Clytemnestra or Orestes (who are both indoors, while Electra is outside) hears Electra, but in any case it is not really a case of attempted communication between Clytemnestra and Electra in the first place: Clytemnestra is calling for help, for Aegisthus, and Electra is glorying in the murder; Electra is speaking to the chorus, not to Clytemnestra. Rutherford cites Euripides’ Phoenician Women 446–637 as ‘an excellent example’ (2012, 15–16 with n. 41) of communicative breakdown, and in his sense of that phrase it perhaps is. But we saw earlier in this section that, in the strict and literal sense, Eteocles and Polyneices are communicating, and indeed doing so all too successfully—tragically successfully.256 At Hippolytus 601–69 we face the difficult question whether Phaedra remains onstage to hear Hippolytus’ dialogue with the Nurse and his ensuing tirade, and, if so, how much she really takes in. Afterwards, she is sure that Hippolytus will reveal her secret to Theseus (689–92), even though Hippolytus has expressly said that he will keep his oath to the Nurse (656–60). Donald Mastronarde argues that Phaedra remains onstage, that she hears and comprehends Hippolytus’ statement that he will keep his oath, but that she does not believe it, for ‘it is essential to the tragedy of Hippolytus and Phaedra that neither is able to come to an understanding of the goodness of the other’.257 And he calls this a ‘failure of communication’. In a loose sense of that phrase it maybe is, but strictly speaking it is not. If I hear what you say you will do, but do not believe you and act accordingly, I may or may not have made a mistake, but communication has not failed. More objectionable than the overstretching of the idea of breakdown of communication is its generalization to all linguistic transactions. For, observing alleged instances of communicative failure, critics frequently wish to draw, and then rely on, the deconstructionist conclusion that language is everywhere unstable, that there are no fixed meanings, that communication is impossible, even in the literal sense, that ‘rather than a medium of communication, language becomes a source of blockage’,258 and that secure interpretation of literary works is unavailable. And it is this theoretical move that is, I suggest, illegitimate.259 Not only does engaging in the move have the flattening effect I have commented on—that is, it destroys the specialness of cases like Phaedra’s reflections on ‘aidōs’; it is also vacuous, from a practical point of view. If all language were unstable in the Derridean way, that would apply not just to the words of the Oresteia, but also to the critics’ words, Derrida’s and Goldhill’s, as it might be, and yours and mine. But then the Derridean move would simply nullify itself from the critical and practical point of view: as a gambit in the philosophy of language it might be right or it might be wrong, but it carries no implications for critical practice. Suppose we agree that meaning is endlessly deferred, what then? We are left

Theoretical Considerations 255 exactly where we were. For consider: we confront a work consisting of words whose meanings are, allegedly, endlessly deferred; I interpret the work using words whose meanings are, by the same token, also endlessly deferred; you criticize my interpretation using words whose meanings are endlessly deferred, and so on. Nothing in the theory or practice of criticism is changed, or even remotely affected, by the Derridean thesis; the endless deferral of meaning simply cancels out across the board.260 It is true that meanings can be indeterminate. I argued in Chapter 1 (§7), in connection with Apollo’s oracle to Oedipus, that we may need to accommodate a certain anti-realism about past meaning in the language of oracles. Of course, although the standard type of oracle, as exemplified by the case of Croesus, is in one sense ambiguous, as I noted in that earlier discussion,261 it is in another sense perfectly determinate in meaning, just as when I hear you mutter darkly ‘Someone will pay for this’ your utterance is perfectly determinate in meaning, though it gives no information on a point of possibly vital interest. But then every statement leaves a mass of things unsaid about its truth-conditions: the philosophers’ favourite ‘the cat is on the mat’ leaves a good deal unsaid, about the cat, the mat, their precise relative locations, not to mention a host of other things; but we do not therefore classify that statement as ambiguous. Again, when in Sophocles’ Philoctetes the False Merchant says to Neoptolemus and Philoctetes ‘may the god aid you both as best he can’ (627), Philoctetes will think of such divine aid as securing his return home, whereas Neoptolemus no doubt thinks (at least at this stage of the drama) that divine aid is required to bring Philoctetes to Troy. So they read different things into the utterance, but that is not because the utterance says two different things. It is just that, meaning determinately what it does mean, it leaves open two ways (at least) in which it could come true.262 (The same holds of the double entendres in the scenes of Euripides’ Helen where Helen and Menelaus deceive Theoclymenos.)263 So we must not exaggerate the indeterminacy of oracles or, in general, of wishes and predictions about the future. But, allowing that semantic indeterminacy is nevertheless possible, the point here again is that it is to be regarded as a special case, not the general condition of all language. Similarly, in literature you often get one word where you expect another, and the missing word is heard in an undertone. Michael Silk explores this phenomenon, which he calls ‘semantic diversion’ (2009). He gives some examples, and there are numerous others. When Virgil’s Aeneas tells of the night Troy fell, he exclaims (Aeneid II, 361–2): quis cladem illius noctis, quis funera fando explicet aut possit lacrimis aequare labores? ‘Who could explain the disaster of that night, who the deaths, by talking about it, or equal with his tears its griefs?’ In place of the expected ‘dictis

256  Tragedy and Linguistic Redress aequare’ (‘equal with his words’) we get ‘lacrimis aequare’ (‘equal with his tears’).264 But, as Silk recognizes, ‘most classical poetry—in style as in substance—operates largely in terms of expectation and fulfilment’ (ibid., 147). We can go further: all language, all meaning, operate in terms of expectation and fulfilment. The cases where you expect one word and get another are necessarily exceptional—indeed necessarily highly exceptional. Silk wishes to exclude modernist literature and Aristophanic comedy from his observation just quoted (ibid.), but the point applies as much to these genres as to other literary (and non-literary) uses of language. Comedy depends for its effect on the unexpected, including the verbally unexpected; but even in comedy the verbally unexpected cannot become the norm, on pain of a collapse of meaning. If there were just surprise there would be no surprise, because there would be no understanding. The ontological correlate of the thesis that language is everywhere unstable is the thesis that the world is everywhere indeterminate. And that is a connection that critics who are attracted by the linguistic thesis often make. Charles Segal, for example, tells us (1981, 58–9) that the saviour-king is the despised outcast and pollution (Oedipus Tyrannus); the unmovable, stalwart warrior of the Greek battle line is the treacherous night killer and the hated enemy (Ajax); the civilizing beast tamer is the bestial destroyer (Trachiniae); the daughter of the royal house is the criminal (Antigone, Electra); the future hero of the Trojan war is savage like a beast (Philoctetes); or, in the reverse direction, the unclean pariah holds mysterious blessings for the city (Oedipus at Colonus). The world itself, divided by polarities that no longer find mediation, breaks through human differentiation and threatens to revert to a prelogical oneness, absolutely Other, unknown and momentarily unknowable, the face of chaos, and yet the face of God. But the paradoxical situations that Segal lists in the first part of this passage do not imply that the world—even momentarily—threatens to degenerate into a Kantian noumenal realm, an undifferentiated ooze of Dinge an sich, as Segal wants to say in the second part. On the contrary, the paradoxes exactly rely on conceptual differentiation. Ajax’ world, for example, is not a ‘prelogical oneness’, a uniform non-conceptual sludge, but a highly conceptually differentiated world—our ordinary world, in fact—in which we can say such precisely nuanced things as that ‘the unmovable stalwart warrior of the Greek battle line is the treacherous night killer and the hated enemy’. Just savour, if you will, the conceptual opulence of that characterization of Ajax, and reflect for a moment on its detailed implications and ramifications. Segal presumably wants his claim about Ajax to be true, but, as so often with critics who wish to sound theoretical and profound, he has simply not thought through what the truth of even quite simple statements—let alone anything as rich and

Theoretical Considerations 257 involved as his statements about the Ajax and other Sophoclean plays— brings in its train. One thing it certainly does not import is a world of Dinge an sich, or ‘the face of chaos’.

Notes 1. Kelly 1993, 23; Halliwell 2012, 113 n. 35; Munteanu 2012, 80–90; Porter 2016, 409. 2. See here A. Boyle 2017, xl–xlii. 3. So e.g. Hall 1996, 297. 4. See Liapis 2012b, 91–2. 5. P. Hammond 2009, 117. Cf. Schiesaro 2003, 81: ‘the very notion of fas becomes blurred beyond recognition’. 6. See A. Boyle on Seneca, Thyestes 220–4; Schiesaro 2003, ch. 3. Ovid’s expression found many later imitators, e.g. in John of Garland’s tragedy in the Parisiana Poetria (113; Lawler 1974, 140); in Mussato’s Ecerinis (327; Grund 2011, 24); Buchanan’s Baptistes (1204–5; 1983, 129); and in Shakespeare’s Titus Andronicus II, 1, 134 (= I, 1, 633 in Bate’s edn., and see Bate ad loc.). 7. Goldhill 1984a, 66–81; Moles 1986, 61. 8. P. Hammond 2009, 121. Cf. Schiesaro 2003, 91; A. Boyle 2017, lxxxvi–lxxxix. 9. Herodotus III, 119; cf. Euripides, Andromache 384–6. See S. West 1999; L. Parker 2007, xiii; Griffith on Antigone 904–15. 10. Foakes 2000, 74; Poole 1987, 11; cf. Kerrigan 1998, 20–1; Halliwell 2012, 15. 11. Cf. A. Moore 1997, 33 (on the ‘second hazard’ of relativization); Williamson 2013, 420–1. 12. There must be a corresponding reference-level entity, distinct both from the sense of the concept-expression and from its extension, as Frege saw: see Gaskin 2008, ch. 2. If you use the word ‘concept’ to mean the sense of a ­concept-expression, as some do, you will need to make adjustments. Substitute for my ‘concept’ whatever word you use to designate the conceptual referent of a concept-expression. 13. On disagreement about the extension of a concept, see further MacFarlane 2007. 14. Wittgenstein 1977, I, §242; 1984, 152; Davidson 1982, 220–3, 236–7; 1984, 137, 152–3, 159, 168–9, 192, 196–201; 2001, esp. essays 10–14. See in general on disagreement and charity D. Lewis 1983, 112–13; B. Williams 1985, esp. chs. 8 and 9; Williamson 1987–8; Dasenbrock 1993; Gill 1996, 177; Lepore and Ludwig 2005, 198–207; Glüer 2006; Pritchard 2011, 279; Pagin 2013. 15. Shakespeare, King Lear IV, 6, 156–63. 16. See Parvini 2012, 2 (on Shakespeare). 17. Montrose 1996, 16; cf. 5–6; 1986, 305; Howard 1986, 18–23. 18. See Nuttall 2007a, 28. Returning a negative answer to my question does not require one to buy into the idea of an absolute conception of the world in Bernard Williams’s sense: see B. Williams 1978, 64–8, 211–12, 239, 245–9, 300–3; 1985, 138–40; Moore 1997, ch. 4. 19. Cf. Gray 2007, 242–4. 20. A. Moore 1997, 103–7. 21. See e.g. Greenblatt 1980, 193–221. 22. Cf. Gellrich 1995, 50; P. Hamilton 2003, 133–7.

258  Tragedy and Linguistic Redress 23. Cf. Howard 1986, 31; P. Hamilton 2003, 142–4; Parvini 2012, 107–11. See further Pechter 1987, 293; R. Levin 2003, 178. 24. Gaskin 2013a, esp. chs. 5–9. 25. See Davidson 1984, 136–7, 152–3, 168–9, 185–98, 199–201. 26. Glock 2008, 43–4. 27. See Jackson 1998, 132; Balaguer 2011, 377. 28. Williamson misses this point: 2007, 121–6. 29. See Jackson 2001, 656, answering Williamson 2001. 30. Allan and Kelly 2013, 108 with nn. 105 and 106. 31. Garvie 2007, 188. Cf. Pinker 2002, 418–19. 32. On our moral affinity with Homer see Taplin 2001; Crisp 2013; Gaskin 2013a, 138–9. 33. J. Gregory 1997, 123–4. See e.g. Euripides, Electra 253, 367–400, with the commentaries of Denniston and Cropp ad locc.; Cropp 2013, 161. 34. For a clear illustration of this, see e.g. Lope de Vega, El Caballero de Olmedo 2289–303. 35. Kovacs 1987, 25. See on these points Cairns 1993, 14–47; I return to them in §§30–1 below. 36. Euripides, Orestes 491–511. See Willink on 496–506, 512–17. 37. See e.g. Aeschylus, Choephori 119–20, with Garvie ad loc.; Denniston 1939, xx; Kitto 1960, 40–3; Euben 1982, 28; Euripides Electra 1093–6, with Cropp ad loc., and his note on 747–858. On Sophocles see WinningtonIngram 1980, ch. 10; Cairns 1993, 244–5 with n. 107; Finglass 2007, 140, 175–6, 253, 265, 271–2, 277–8, 483–4, 521, 525–8, 539–40, 541–2. 38. Willink on Euripides, Orestes 551–6. 39. See Gadamer 1990, 137–8; P. Hamilton 2003, 78–9. 40. Zeitlin 1986, 102. 41. Against talk of codes in this context, see Raymond Williams 1977, 168–9; Gaskin 2013a, §5. A code translates one language into another, but the Greek dramatists were not engaged in translation. Language does not translate reality, because reality is not itself a language: see further Ch. 6 below. 42. Wilson and Dutton 1992, 181; Greenblatt 1990, 222; Goldhill 1986a, 197– 8; cf. 55; Dollimore 1994, 11; cf. 1990, passim; 2010, 17–19, 250; Belsey 1992, 43; cf. 35; 1990, 452; Sinfield 1992, 128, 36. See also (in support or criticism of the doctrine) Raymond Williams 1966, 45; Howard 1986, 20; Neely 1988, 7; Holstun 1989, 199; Easterling 1990, 88–9; R. Levin 1990, 436–7; Martindale and Martindale 1990, 202; duBois 2001, 75–6; Grady 2009, 130–6; various essays in Haslanger 2012. 43. Cf. R. Levin 1990, 469; Dawkins 1998, 12–13; M. Nussbaum 1993, esp. §6; 2001, 151–2, 308; J. Gregory 1997, 20–2; Budelmann 2006b, 128–9; Ter-Nedden 2007, 336–7; Coyne 2009, 30; Parvini 2012, 4, 54, 175–9; Holbrook 2015, 7–9. In general see D. Brown 1991, and Pinker’s groundbreaking 2002 passim, esp. 55, 142–3, 435–9. 44. See Cairns 2016, 59–66. 45. Winnington-Ingram 1980, 57; cf. Pelling 1997, 227–8. 46. Aristotle, NE 1097b11, 1169b18–19. See Segal 1981, 13; M. Nussbaum 1995; B. Williams 1995, 194–202. 47. Cf. A. Moore 1997, 34–7. 48. See here again B. Williams 2002, ch. 2; Gaskin 2006, 39–40. 49. See Bristol 2009, 23. 50. See Pinker 2002, 241. 51. See Burian 2010, 286–7. 52. O’Grady 2002, 162–3; Dustin 2014, 129. 53. See here again Pinker 2002, 145, 156–7, 164, 202, 421, 426–7. 54. M. Nussbaum 1993, 256.

Theoretical Considerations 259 5. 5 56. 57. 58. 59. 60. 61. 62. 63. 64. 65.

Cf. Billings 2014, 226; Billings and Leonard 2015, 4–5. As also at Vernant 1991, 263. Cooper 2010, 3. O’Grady 2002, 155–6. Vernant in Vernant and Vidal-Naquet 2001, vol. 1, 21, 80–1. Leonard 2012, 150; 2015, 7. See e.g. Vernant and Vidal-Naquet 2001, vol. 2, ch. 4. Cf. Sternberg 2005, 16. Pace, apparently, duBois 2014, 308. Cf. Bennett and Tyrrell 1998; Taxidou 2004, 84; Hoxby 2015, 75. J. Gregory 2005, 261; cf. 1997, 31–2; the point derives from Dale on Alcestis 280ff. 66. As e.g. at Euripides, Iphigeneia in Tauris 538: see L. Parker’s note ad loc., her note on 674–722, and her remarks at pp. xxxi–xxxii. 67. See Leonard 2015, 108–22. 68. See Empson 1986, 112; Vernant in Vernant and Vidal-Naquet 2001, vol. 1, 90–8; Griffith 2005b, 109. 69. See Fitch and McElduff 2008, 161–2. 70. See Dodds 1960, 172–3, and his notes on lines 222–3, 343–4, 800–1, 810– 12, 821–38. Cf. Goldhill 1997c, 342–3. 71. See L. Parker on 1055–6. 72. Cf. Goldhill 1986a, 127, 131. 73. See Barrett on 208ff. 74. So e.g. Heath 1987, 121–3; J. Gould 2001, 103. 75. So too, perhaps, with Eteocles in Aeschylus, Seven Against Thebes: see Zeitlin 1982, 18–19. 76. For discussion of Shakespeare’s adaptations of Ovid in this poem, see Martindale and Martindale 1990, 58–61; Bate 1993, 48–65; Miola 2000, 20–6; Duncan-Jones and Woudhuysen 2007, 57–68. On Freudian elements in Titus Andronicus, see Bate 1995, 7–9, 36–7. 77. On psychoanalytic readings of older literature in general, see Segal 1994, 88–9; Griffith 2005b, 98–122; Armstrong 2012. On Hamlet, see Empson 1986, 112–14; on Macbeth, see Nuttall 2007b, 287; on Racine, see Greenberg 2010, 246–7 and passim. 78. This is suggested by M. Nussbaum 1994. 79. Cropp 2013, 2. See also Finglass on Sophocles, Electra 277, with further references, and Tarrant on Seneca, Agamemnon 905. 80. Cairns 1993, 235. 81. D. Lewis 1986, 77. 82. Vickers, 1973, 3; cf. Winnington-Ingram 1980, 6–7. 83. Goldhill 1990a, 101; cf. 1986b, 163; 2000, 36–7; 2012, 24; Padel 1992, 10–11, 35; J. Gould 2001, 157. 84. See L. Parker on Euripides, Alcestis 779; Denniston 1950, 298. 85. Cf. Cairns 1993, 1–4. 86. Cf. Rowe 2003, 48–9. 87. On pistis and fides as ancient emotions, see Morgan 2015, 447–54. Cf. Halliwell 2012, 223; Porter 2016, 124–30. 88. See here Raymond Williams 1977, ch. 4; Nuttall 2007b, 171. 89. Raymond Williams 1977, 46–7; Laird 2006, 3; cf. Halliwell 2012, 10–12 (on ‘fiction’). 90. On Sophocles, see n. 37 above; on Virgil, Gaskin [forthcoming]; on Racine, J. Campbell 2005, 125–32. Similar remarks apply to ‘equivocate’ (on which see Shapiro 2015, 196), and many other terms. 91. Chaucer, Troilus and Criseyde I, 236–8. 92. See here Heath 2006, 262; Nuttall 2007b, 10.

260  Tragedy and Linguistic Redress 93. Chaucer, Troilus and Criseyde II, 478–9. Cf. Reichl 1989, 152; Kelly 1997, 98–9; C. S. Lewis 2013a, 219; 2013b, 43–4. 94. On Criseyde’s character see Dunning 1962, 180–1. 95. See Andrew 1989, 87–8; J. Mann 1989. 96. See C. Hamilton 2016, 128–35. 97. See Phillips’s introduction to Chaucer 1997, 33–5. 98. Lawlor 1961, 257. 99. On their different meanings see e.g. L. Parker on Euripides, Alcestis 1080– 1, and on Iphigeneia in Tauris 220. 100. See Euripides, Alcestis 193–5 (with L. Parker ad loc., and p. lii), 767–71. 101. Halliwell 2006, 127. 102. See also Craik 1993; Konstan 2006, ch. 1; Rusten on Thucydides II, 43, 1. 103. As Cairns agrees: 1993, 9 with n. 19. 104. Cf. Griffin 2007, 190. 105. See here again Cairns 1993, 14–15. 106. This point is missed by Kölbel in his otherwise helpful discussion of relativism at 2011, 26. 107. See B. Williams 1972, 34–9; O’Grady 2002, 146–9; Dustin 2014, 115–16. 108. Cf. Quine 2008, 242–3; Dustin 2014, 117–19; Lockie 2014, 163. 109. See e.g. B. Williams 1981, ch. 11; 1985, passim; refuted by Dustin 2014, 120–31; cf. also Lear 1988, 89–94. 110. Chaucer, The Wife of Bath’s Tale 1170; cf. 1109–24; The Franklin’s Tale 1543–4; The Romaunt of the Rose 2189–92; ‘Gentilesse’ (1987, 654). 111. See n. 115 below. 112. M. Nussbaum 1993, 244; for the criticism, see e.g. C. Hamilton 2016, 97–9. 113. Henryson, ‘Orpheus and Eurydice’, 7–10 (1981, 132). 114. J. Gregory 1997, 36. 115. See L. Parker on Alcestis 746–860, 860, and p. xxxiv; on line 1097 (‘Then receive this lady generously [gennaiōs] within your halls’), she comments that the force of ‘gennaiōs’ is that Admetus should receive the lady (in fact Alcestis) ‘like the true gentleman that you are’. Cf. Aristophanes, Frogs 737–40, with Dover 1993, 46–7; Thesmophoriazusae 220, with Austin and Olson ad loc. 116. Cf. P. Berger and T. Luckmann 1966, 41–2. 117. See Willink 1986, xlvii. 118. Goldhill 1986a, 174; 1990a, 101–2. 119. See e.g. Belsey 1992, esp. 43; 2003a. For a good response, see R. Levin 1990, 436–7, 443. 120. Gaskin 2001. See further against Snell: Schmitt 1990; B. Williams 1993. 121. See further Halliwell 1990a, 34–6; Gill 1996, chs. 1 and 3; Gaskin 2001; Day 2013, 36. 122. See Hoxby 2015, 48–51, 264. 123. See Poetics 1450a6–7; Halliwell 2012, 233. 124. Cf. Greenblatt 1980, 2–3; Engle 2013, 203; Bickley and Stevens 2016, 1–6; Thompson and Taylor 2016, 20. 125. See Gaskin 2018. 126. Greenberg 2009, 397. 127. Goldhill 1986a, 188. Cf. 1990a, 105–11; Hoxby 2015, 75. 128. See Martindale and Martindale 1990, 10. 129. J. Gould 2001, 109–10; cf. Mitchell-Boyask 1996, 429–30. 130. Winnington-Ingram 1983, 96; cf. Mossman 2005, 353; Macleod 1983, 34–5. 131. Sinfield 1992, 37; cf. Jardine 1996, 37–8. 132. Holbrook 2015, 34–5, 43, 36–7, 51, 53.

Theoretical Considerations 261 133. Cf. Vickers 1973, 57; Easterling 1990, 87; Nuttall 2007b, 357–9; Bristol 2009, 21. Note on ‘as’: I shall often omit this word, taking its presence as understood in relevant contexts. What is its import? If I say I am painting a picture of someone, I might mean this extensionally—there is some actual, non-fictional person out there, my eldest sister as it might be, whom I am painting—or I might mean it intensionally, in which case there need be no such actually existent person, but the painting is nevertheless as of a human being, and in that sense it is of a real person—as opposed to its being of a cat, a Martian, a landscape, an arrangement of fruit, etc. In my discussion, by ‘representation of’ I mean ‘representation as of’; that and similar locutions are being used intensionally. Sometimes I shall insert a parenthetical ‘as’ in relevant locutions as a reminder. 134. See Gill 2009. 135. See L. Parker on Euripides, Iphigeneia in Tauris 1475–6 (with further examples). 136. See further on this point Gaskin 2013a, §48. 137. Schalkwyk 2016, 483. 138. See Harman 1990, 35. 139. Young 2013, 43–7. 140. Cf. Jones 1962, 32–8; Pelling 1990, 255–6; Halliwell 1998, 151; Garvie 2007, 144. 141. See here Seidensticker 2008, 337–9. 142. Pelling 1990, 251. See e.g. Finglass’s comments on Sophocles’ Electra in his commentary, esp. on lines 86–120, 110–20, 254–309, 516–633, 603–9, 764, 822, 804–22, 823–70, 911–12, 920–46, 947–89, 950, 958–66, 1017, 1126–70, 1339–63, 1398–441, 1483–90. Cf. Winnington-Ingram 1983, 96; Hutchinson 1985, xxxiv–xl; Seneca, Agamemnon 301, with Tarrant ad loc. 143. See Lloyd-Jones 1990, 401–18. 144. Dale 1954, xxv, xxvii; 1969, 139–55, 272–80. 145. Cf. Pelling 2005, 100. 146. See Moles 1986, 60. 147. Jones 1962, 43–6, 59–60, 232–5; cf. Heath 1987, 115–16; Seidensticker 2008, 333–4. 148. Jones 1962, 37. See e.g. J. Gould 2001, 84; D. Schmidt 2001, 15, 54 with n. 12, 79; Rutherford 2012, 317; Young 2013, 125, 139–45. 149. Schiller 1992–2004, vol. 12, 261–2. 150. See Cicero, In Pisonem 1, with R. Nisbet ad loc.; Seneca, Agamemnon 128, with Tarrant ad loc. 151. Cf. Seidensticker 2008, 340 n. 27. 152. Seidensticker 2008, 341; cf. Vickers 1973, 55; A. Brown 1983, 20–2. 153. On these see Dunn 2012, 105–8; Tarrant on Agamemnon 131–44. Cf. Virgil’s treatment of Turnus at the moment of his peripeteia (Aeneid XII, 665–71), with Gaskin 1992, 305–6. 154. Seidensticker 2008, 341; cf. Rutherford 2012, 33. 155. Cf. Deppman 2012, 530–1. 156. Cf. Denniston 1939, 113 n. 1. 157. Vickers 1973, 54; Easterling 1993b, 58; cf. Austin and Olson on Aristophanes, Thesmophoriazusae 266–8. We find the same category mistake at Schopenhauer 1986, §46, vol. 1, 320–1. On buskins, see Rudd on Horace, Ars Poetica 80–1. 158. See Mastronarde ad loc. 159. See Finglass ad loc. and on 1209–10. See also Aeschylus, Suppliants 197–9; Euripides, Alcestis 773, with L. Parker ad loc.; Helen 456, 632–5, with Allan ad locc.; Tarrant on Seneca, Agamemnon 128; Padel 1992, 59–60.

262  Tragedy and Linguistic Redress 160. Easterling 1987, 17–18; Seidensticker 1995, 155. Cf. J. Gould 2001, 84–5, 89, 94–6. 161. On the exostra see A. Boyle 2008, 95, 117, 169, 217. 162. See Euripides, Ion 237–400, with Seidensticker 2008, 340; L. Parker on Iphigeneia in Tauris 714–15; Rutherford 2012, 33. 163. Horace, Ars Poetica 102–3; see Rudd ad loc. 164. Lyne and Mukherji 2007, 5. 165. Ibid. 166. A. Boyle shows confusion on this point in his otherwise excellent discussion of Seneca’s theatricality at 2006, 208–18 (cf. also Erasmo 2004, ch. 5). The category mistake is a staple of comedy, being exploited for instance in The Critic. 167. So e.g. Goldhill 1984a, 73; 1986a, 163–4; 1990a, 111–14; Easterling 1987, 16–17; 1990, 83–4; Burkert 1991, 7–8; Pineau 1991, 48; T. von Wilamowitz-Möllendorff 1996, 162; Gregory 1997, 30; Gould 2001, ­ 78–80; Taxidou 2004, 33–4; Seidensticker 2008, 344; Griffith 2010, 112– 13, 123–4; Russo 2012, 14, 22; Esposito 2016, 471. 168. This point is argued with typical incisiveness and humanity by Nuttall in his 1984. 169. Gaskin 2013a, §10. 170. Knights 1965, 196–204. 171. See e.g. Dale 1954, xxii–xxix; Goldhill 1984a, passim (to be read with Moles 1986); Easterling 1993a; J. Gould 2001, 79, 100–1, 104. 172. See McDowell 1998b, passim, esp. essays 8, 13, 15, and 17; Scruton 2004, 3. 173. So e.g. Davis 1986, 153–4; Easterling 1990, 99; Padel 1992, 34; Griffith 1999, 37–8; J. Gould 2001, 79–80, 89; Marshall 2002, 179; Seidensticker 2008, 340. 174. So e.g. Bristol 2009, 26. 175. See further Gaskin 2013a, §40. 176. Pace Smith 2016, 96–8. 177. Allan 2008, 42–5; Seidensticker 2008, 340. 178. Cf. Gaskin 2013a, 263, 289. 179. Professional philosophers too: see e.g. Haslanger 2012, 118. 180. See e.g. J. Gould 2001, 86; cf. Garton 1972, 15; Gill 1996, 123; MitchellBoyask 1996, 429; Parvini 2012, 11–12; Rutherford 2012, 285. 181. Dodds 1973, 68. Cf. 1960, xlix, 148, 170; Waldock 1975, 23; Kirkwood 1994, 69–70. 182. Gaskin 2013a, ch. 7. 183. Dodds 1973, 75; cf. Aristophanes, Thesmophoriazusae 972, with Austin and Olson ad loc.; Birds 685–722, with Dunbar ad loc. 184. See Gaskin 2013a, 44–51. 185. Cf. Rutherford 2012, 306 with n. 69; Gaskin 2013a, 51. 186. Against Dodds’s principle see usefully Seaford 2000, 37–9. 187. See his note on Aeschylus, Agamemnon 954f. Cf. Nuttall 2007b, 203. 188. See Thomas 1988a, 109; Poulsen 2013, 105. 189. Cf. Nuttall 2007b, 190–1. 190. See Steinmetz 1987, 116–29; Lamport 1981a, 182–92; 1990, 24–5; Fick 2010, 381–5; H. Nisbet 2013, 501. 191. Calder 2005, 83–4. 192. Scullion 2002, esp. 118–21, 131–4; see also Henrichs 1994, 65–73. 193. See Friedrich 1996, 274–7; Heath 2006, 253. 194. Dawe 2006, 89. 195. Knox 1998, 67–8.

Theoretical Considerations 263 196. See Morwood 2007, 5–6, and his notes on lines 187, 426–62, 485, 1123– 64, 1191–5. 197. See their commentaries ad loc. For discussion of further interesting cases, see e.g. Dodds 1973, 45–53, 61–3, and Macleod 1983 (Oresteia); Garvie 2009, xvi–xxii (Persae); Woodruff 2012, 133–4, and Schein 2013, 10–12, 20–3 (Philoctetes); in general see Osborne 2012. 198. See Pocock 1973, 212–14. 199. So e.g. Kennedy 1992; Montrose 1996, 11–12. See Gaskin 2013a, §64. 200. See e.g. Goldhill 1990b; 2000; 2012, 154–5; Burian 2011. 201. So e.g. Hall 1996, 297–9, 301–4; Segal 1996, 157. 202. On Ajax, see Goldhill 1990b, 115–18, refuted by Friedrich 1996, 264–8, and Seaford 1996, 291–2. 203. Scullion 2002, 134; cf. Mastronarde 2010, 17–18, 45. 204. Easterling 1985, 2–3; cf. Macleod 1983, 147–9; Euripides, Helen 395–6, with Allan ad loc. 205. See Griffin 1998; 1999, 151–9; 2007, 203; J. Gould 2001, ch. 17; Rhodes 2003; Carter 2007, 35–43; Garvie 2007; Boedeker and Raaflaub 2009, 115–18, 124–5; Heath 2006; 2009; 2011. 206. Hall 2014, 780. See also Zeitlin 1986, 102; Rösler 1993, 81; Burian 2011, 114–15; Allan and Kelly 2013, 99–100, 107, 115. 207. R. Parker 1997, 149; Pelling 1997, 227–35. 208. See Heath 1987, 48; 2006, 258–61. 209. See e.g. Longinus, De Sublimitate 44; Potkay 2012, 210–11. 210. Burian 2011, esp. 97–102. 211. See §28 above; Heath 2011, 170. 212. Rowan Williams 2016, 141–2. 213. So Vernant in Vernant and Vidal-Naquet 2001, vol. 2, 158–9. 214. Taplin 1996, 193–4; J. Gould 2001, 378–89. 215. A good, brief survey of the arguments is Murnaghan 2009, 245–7. 216. Austin and Olson on Aristophanes, Thesmophoriazusae 395–7; Dunbar on Birds 30–1. 217. See J. Gould 2001, 382, 394–5. 218. Cf. Goldhill 1996, 252–3. 219. On the range and variety of Greek tragic choruses see esp. Foley 2003. 220. J. Campbell 2005, 151–76; Küpper 2014, 295–8. 221. So e.g. Hall 1996, 1997, 100–3. 222. Notwithstanding Crawforth, Dustagheer and Young 2015. The similarities that they find between Verona and London in the essay on Romeo and Juliet are particularly arbitrary: a score of other early-modern European towns would have borne comparison to Verona just as well, if not better. 223. See Taplin 1999. 224. See here Heath 2009. 225. Pace Hall 1996, 304–6; Macintosh 2009, 31. 226. See Hume 1987, 233; Taplin 1999, 56; Heath 2011; Pinker 2002, 408. 227. On Elizabethan audiences, see e.g. Bickley and Stevens 2016, 137. 228. Budelmann 2000: see e.g. 14–15, 64–5, 73, 94, 137–8. 229. Cf. Halliwell 2012, 90–1 with n. 110; Bickley and Stevens 2016, 179 (on The Spanish Tragedy). 230. See Gaskin 2013a, §§18, 31. 231. Vernant in Vernant and Vidal-Naquet 2001, vol. 2, 85. On the meme, see e.g. Seaford 1995, 202; 1996, 290; Friedrich 1996, 261–8; Allan and Kelly 2013. 232. See e.g. Goldhill 1984a; 1984b; 1986a; 2012, 41, 79; 2015, 235–6; Introduction to Goldhill and Hall 2009, 15. Cf. Segal 1981, 52–9; 1986a, 46–8,

264  Tragedy and Linguistic Redress 72–3; 1993, ch. 6; duBois 1986, 376; Taxidou 2004, 83, 92. In criticism, see Moles 1986 (to which Goldhill’s reply, 1986b, is quite unsatisfactory). 233. Vernant in Vernant and Vidal-Naquet 2001, vol. 1, 35; cf. Vickers 1973, 29; Goldhill 1984a, 239, 245; Rocco 1997, 25. 234. Note the anticipation of this development at Aeschylus, Choephori 119–20, with Garvie ad loc., and Havelock 1978, 289, 293–5. See also Choephori 306–14 and 461, with Garvie ad locc.; Macleod 1983, 29–34. 235. Cf. the definition of justice attributed to Simonides at Plato, Republic 331e3–4. 236. Pace Goldhill 1986a, 50. On the concept of dikē, and the history of the word, see LSJ s. v.; Hutchinson on Aeschylus, Seven Against Thebes 171, 645–8; Bowen on Suppliants 408–9; Young 2013, 102; L. Parker 2016, xxvi n. 43 (with further references). 237. Euripides, Helen 1002–4, with Allan ad loc. 238. So e.g. Denniston 1939, xiv–xvii; M. Nussbaum 1986, ch. 2. 239. Winnington-Ingram 1980, 222. Cf. Finglass 2007, 280; Schein on Philoctetes 1035–44. 240. Winnington-Ingram 1980, 66; cf. Goldhill 2012, 163. 241. Vernant in Vernant and Vidal-Naquet 2001, vol. 1, 36. Cf. Gellrich 1995, 50. 242. For an excellent discussion of this passage, with references to further literature, see Cairns 1993, 321–40. A fairly standard interpretation of the meaning of Phaedra’s use of ‘aidōs’ is offered by Segal 1970, but I believe that Craik is right (1993; 1997) that in this context the word is a euphemism for sex. See also McClure 1999, 130–1. 243. Rutherford 2012, 195–6. 244. So Belsey 2003a, 83. 245. On the deception speech, see J. Moore 1977; Sicherl 1977; WinningtonIngram 1980, 46–56; Segal 1981, 113–15; Davis 1986, 153–6; Heath 1987, ch. 5; Gill 1990, 19–21; 1996, 204–16; Pelling 2005, 93–5; Cairns 2006, 111–13; Lardinois 2006; Finglass 2011, passim, esp. 16, 55, 379, and ad loc. A key question concerns whether lines 652–3 are ambiguous. My view is that the right reading is Blaydes’s ‘lipōn’, and this has the effect of removing any ambiguity. I also think that at crucial points (at 666–7 and at 677) there are suppressed conditions (‘ I shall have to. . .’), whose omission is what occasions Tecmessa’s and the chorus’s mistake: cf. Winnington-Ingram 1980, 55. 246. Kermode 2001, 102. 247. Cf. Wittgenstein 1977, I, §433; Nuttall 2007b, 201. 248. Shapiro 2005, 322. 249. See Christenson 2000, 15–16. 250. Cf. Goldhill 1986a, 211; D. Schmidt 2001, 66; Rutherford 2012, 353 (on Pentheus and Dionysus). 251. Rose 1995, 60; Derrida 1967, 409–28; 1972, 29–31. Cf. Goldhill 1984a, 26–7. 252. For a similar gap between official stance and actual practice, cf. Montrose 1996 prologue and ch. 1. 253. See Cairns 1993, 219; Podlecki 1966, 233; Pelling 2005, 91. 254. Bushnell 1981, 198; duBois 1986, 376; Segal 1981, 134. 255. So Goldhill 2012, 77. 256. Similarly with Clytemnestra and Electra at Sophocles, Electra 764–822: see Finglass ad loc.

Theoretical Considerations 265 257. Mastronarde 1979, 81; cf. 2010, 224–5. Kovacs (1987, 54) rejects the view that Phaedra remains onstage. 258. Segal 1981, 53. 259. Cf. Silk 1996b, 463–4. 260. Cf. Gaskin 2013a, 265. 261. Cf. Budelmann 2000, 50–1. 262. See Schein ad loc.; note also 779–81 and 812, with Schein ad locc. 263. See Allan on 1165–1300 and 1369–1450. 264. Cf. Putnam 2010, 85; on Sophocles, see Budelmann 2000, 40–50.

5 Can Suffering Be Expressed in Words?

36 The Idea of Linguistic Theodicy Recall from Chapter 1 my discussion of Schiller’s view of tragedy. As Sebastian Gardner says, in a passage partially cited already (2003, 233–4): Schiller’s account entails the justifiability, in some sense, of tragic suffering: Schillerian tragedy enacts a kind of moral theodicy. To be sure, Schiller’s view does not entail the grotesquery that tragic loss and suffering are themselves good—that what happens in tragedy is what ought to happen—or that they are nevertheless justified as means to moral ends: but it does entail that the point of tragedy lies in the final realisation that tragic loss and suffering are, in comparison with moral evil, of no true, ultimate import, and in that sense it affirms a total redemption . . . [T]o the extent that the tragic victim has avoided moral evil, he has in reality lost nothing. But, as we have seen in earlier discussions, there is a good sense in which tragic literature does illustrate the thesis, grotesque or not, that ‘what happens in tragedy is what ought to happen’. For tragic heroes are generally punished for faults that they exemplify. In respect of its protagonists, tragedy does present us—perhaps not exceptionlessly, but to a much greater extent than has been acknowledged in the commentaries—with a justification of suffering. So, as I have indicated, I believe that Gardner is wrong to say, as he does shortly after the quoted passage, that tragedy represents loss and suffering as having ‘unconditional and hence uncompensatable reality’ (ibid., 235). By and large the loss and suffering are compensated in the sense that the hero deserves his fall; the punishment is often harsh, but then its extent was itself not unpredictable. That is a form of moral redress, and its importance to tragic literature was the theme of the first part of this book. Still, the view that tragic loss and suffering are of no ultimate significance compared with moral evil, though embraced by a number of literary figures themselves, such as Milton’s Lady in Comus and Lessing’s

Can Suffering Be Expressed in Words? 267 Emilia Galotti (§25), is not the usual motto of the great tragic works of the Western tradition, and I think Gardner is right to reject it. The Socratic position that the good man or woman cannot be harmed by anything external—only by an inner impulse to resile from his or her ­goodness—is not what we find in Sophocles and Euripides and Shakespeare and Racine, or indeed in Schiller’s own greatest plays, Don Carlos and the Wallenstein trilogy. But if the somewhat oxymoronic phrase ‘moral theodicy’ is to have a use at all, it must mean something along the lines of ‘a moral justification of the human condition’, or, in plain terms: you get what you deserve. And in that sense there is a kind of moral theodicy in most Western tragedy, I have argued, because that genre deals not with terrible events like genocides, where large numbers of ordinary people are caught up, through no fault of their own, in events that destroy their dignity and their lives, but rather with small-scale, personal disasters in princely families, where the catastrophe is represented as traceable to some fault or flaw in one or more of the main figures. (There is indeed a recognition in tragedy that princely catastrophe can bring disaster for ordinary people in its train.)1 What I propose now is that we should subject the idea of moral theodicy to the familiar postWittgensteinian linguistic turn, and talk of linguistic theodicy, by which I shall mean the encompassing of suffering in language, in such a way as to generate a feeling that, ultimately, this is how things have to be and that it is somehow right, or all right. This encompassing has, I suggest, a celebratory aspect: in particular, tragedy celebrates the sheer fact that suffering can be expressed in language. There is a two-way dependence operative here: we value and take pleasure in tragic language because it presents profound tragic emotions, and we value and take pleasure in those emotions because they are presented in a language that encompasses and expresses them, thereby not only taming and domesticating them, but also—simultaneously and paradoxically—exalting and ennobling them. That is a kind of linguistic redress.2 Now, in saying that tragedy has a celebratory aspect, I need to be cautious. Many thinkers have objected, surely rightly, to the idea that tragic literature offers us an easy redemption from suffering. It cannot do so if the suffering is too close to home. As Hume observed (1987, 223), if we are too close to real-life tragedy, we dislike hearing it dressed up in ornate language. The beautification that language confers on suffering may be inappropriate for other reasons, too. For instance, it may be too overtly rhetorical—as, for instance, the long speech that Marcus makes in Titus Andronicus (II, 4, 11–57) on encountering his mutilated niece Lavinia offends because of its rhetorical excess.3 The speech, with its learned invocation of Ovid’s Philomela, seems to be an almost frivolous response to Lavinia’s terrible maiming. Against this natural thought, Jonathan Bate defends Marcus’ rhetoric, resting his case on the power of a good performance, as opposed to a mere reading of the play: ‘one might

268  Tragedy and Linguistic Redress say: if Chiron and Demetrius had seen a dramatization of the Philomel story, instead of reading it cold-bloodedly in the classroom, they would have wept for her instead of re-enacting her rape’ (1993, 112). This remark recalls a comment in Curtius’s commentary on the Poetics, voicing a sentiment shared by Lessing.4 One suspects that it is rather too sanguine, at least in the short run.5 But, even if Marcus’ rhetoric is not a good illustration, there is such a thing as the celebration of pain and grief, as of all other aspects of human life, in poetic language. I share the view so powerfully expressed by Euripides’ Bellerophon,6 and by Seneca’s Jason,7 that a theological theodicy—that is, a theodicy in the etymologically correct sense of the word, as opposed to the extended senses I have been considering—is out of the question. As Hume so characteristically and eloquently put it, Epicurus’ ‘old questions’ about God ‘are yet unanswered. Is he willing to prevent evil, but not able? then is he impotent. Is he able, but not willing? then is he malevolent. Is he both able and willing? Whence then is evil?’.8 Primo Levi’s ‘There is Auschwitz, so there cannot be God’ seems to me unanswerable.9 The Holocaust is not in se inexplicable; it becomes inexplicable in the context of a theism that makes the usual assumptions about God’s nature.10 But it would be an overreaction to suppose that, because redemption cannot be easily won, and should not be superficially sought, it is therefore not to be had at all, or that the traditional idea that tragedy overcomes and transcends and in some measure hallows suffering is mere pietistic nonsense.11 Moral redress operates at the level of the individual protagonist; linguistic redress is, rather, an aspect of audience response. Language does not, at least not usually, in any way compensate the individual who suffers.12 Nor does suffering necessarily help the tragic individual to knowledge. Insight comes to some, but by no means to all. As David Scott Kastan says, Cordelia’s death ‘must mock the idea that tragic suffering leads to knowledge’ (2003, 13)—at all events it does not lead to compensatory knowledge for Cordelia and Lear. As far as the suffering individual is concerned, I think we have to agree with the substance of Bernard Williams’s assertion that ‘The idea that meaning, or purpose, or understanding, or even, perhaps, a true philosophy could make all suffering bearable is a lie, whether it is told by recruiting sergeants or by ancient sages’ (2006, 334). But Williams’s claim does not distinguish between different possible candidates for whom suffering is to be made bearable. My suggestion is that the tragic portrayal of suffering can compensate us, the users of language, for our painful lot in this world: it can give us linguistic redress.13 (In his recent book on tragedy, Rowan Williams comes close to endorsing this idea at several points, though officially he is opposed to finding an explanatory and compensatory structure in tragedy:14 the result is obfuscation, for you cannot have it both ways.) When in September 1796 Charles Lamb’s sister Mary, overcome by insanity, stabbed their mother to death, Charles’s life fell apart. In due

Can Suffering Be Expressed in Words? 269 course he recovered, and resumed his literary avocations, but in the first onset of grief he begged Coleridge to write him ‘as religious a letter as possible’, and to ‘mention nothing of poetry’, for ‘I have destroyed every vestige of past vanities of that kind’ (1975–8, vol. 1, 44–5). This does not mean that poetry cannot speak to Lamb’s situation; only, that it cannot—or does not—speak to Lamb while he is in that situation. But there is nothing privileged about the sufferer’s perspective on his suffering at the time when he is undergoing it; it is not more authentic than an onlooker’s perspective, or than the view of it that the sufferer himself takes at other times. The possibility of literature’s losing its relevance to those who are in extremis—the thought that, as Jean Améry memorably put it, no bridge led from death in Auschwitz to Death in Venice (2008, vol. 2, 47)—does not mean that literature is irrelevant to suffering simpliciter, because there are more perspectives on pain and suffering than that of the sufferer at the moment of felt pain, and indeed not all sufferers react in the same way to their pain, nor does a given individual sufferer necessarily react always in the same way. Améry said of his experience at Auschwitz that the life of the mind was of no assistance (ibid., 45), but that was not Primo Levi’s experience, as the famous account of his recollection of Dante’s Canto of Ulysses illustrates (1958, 98–103). Nor was it Ernst Wiechert’s experience, when he stood under the oak tree— its stump may still be seen—whose shadow, so it was said, had once fallen on Goethe and Charlotte, and which now fell within the perimeter of Buchenwald concentration camp.15 Unfortunately, the story about Goethe and Charlotte was only a camp legend. But what is not legend is that some Buchenwald prisoners were tasked with making a replica of Schiller’s writing desk, which was accordingly removed to the camp from his house in Weimar.16 The remarkable Buchenwaldlied could be sung openly, but only because its ambiguity went over the heads of the guards.17 More explicit subversion had to be concealed. So it came about that, while Schiller’s desk served time in Buchenwald, the great laus libertatis from his Don Carlos was recited and cherished secretly and in lowered tones: Ein Federzug von dieser Hand, und neu Erschaffen wird die Erde. Geben Sie Gedankenfreiheit. ‘A stroke of your quill, and the world will be renewed. Grant us freedom of thought’.18 That the life of the mind was of no assistance was not actually Améry’s experience at all times during the period of his incarceration at Auschwitz: he tells us how, on one occasion, he was unexpectedly given extra food by an orderly from the sick barrack, and started thinking about the phenomenon of human goodness, about the brave Joachim

270  Tragedy and Linguistic Redress Ziemßen of Der Zauberberg, about other books and pieces of music, and he even found himself entertaining original philosophical thoughts (2008, vol. 2, 35–6). When Walter Poller, in Buchenwald, witnessed the release of prisoners after a period of confinement in darkness, he thought he could catch faint strains of Fidelio (1947, 137–8). In the Lager, as Levi remarked, culture could be as precious as a precious stone (1986, 111). Again, after a sufficient interval had elapsed, Améry came to adopt a different perspective on his suffering from the one that he had had under the immediate impact of the pain and discomfort, a perspective, namely, that was infused with his rediscovered faith in the humanist tradition. When he received the Lessing prize in 1977, in his acceptance speech he passionately affirmed his commitment, still maintained in spite of everything he had undergone, to the ideals of the Enlightenment. As he there pointed out, Enlightenment rationality carries with it its own corrective—an inherent ability to adapt and adjust to all eventualities, to absorb and overcome all criticism—so that it is idle to try to dislodge it dialectically (2008, vol. 6, 550). Such attempts would themselves just be enlightened thought, even as the attempt to demonstrate the pointlessness of philosophy would itself be philosophy. Accordingly, I see the idea of linguistic redress as a move within the humanist and Enlightenment tradition. In a discussion of King Lear, Jonathan Dollimore distinguishes between a Christian and a humanist interpretation of that tragedy, and of tragedy in general. The Christian interpretation, he says, ‘locates man centrally in a providential universe’. The humanist view, he continues, ‘likewise centralizes man but now he is in a condition of tragic dislocation . . . If he is to be redeemed at all he must redeem himself’. The humanist rejects providence, and holds that if ‘suffering is to be justified at all it is because of what it reveals about man’s intrinsic nature—his courage and integrity. By heroically enduring a fate he is powerless to alter, by insisting, moreover, upon knowing it, man grows in stature even as he is being destroyed’ (2010, 189). The Christian view is rejected by Dollimore, in my view rightly: the simple fact is that the universe is not providential, and most tragedy does not present it as such. But Dollimore also rejects the humanist view, both absolutely and as an interpretative stance towards tragedy, and that seems to me mistaken. His reason is that, to take the case of King Lear, the humanist view ‘mystifies suffering and invests man with a quasi-transcendent identity whereas the play does neither of these things’ (2010, 190). Now, I agree that in talking about the celebratory and ennobling aspect of tragedy there is a risk of mystification. On the other hand, though the risk is there, it is one that we need to take. For in my view tragedy does make a kind of transcendentalist move in the face of suffering, and this transcendentalist move is consolatory and compensatory. However my suggestion, in accordance with my adoption of the linguistic turn, has been that we should construe the transcendentalist move as a linguistic one, not

Can Suffering Be Expressed in Words? 271 moral or theological. Knowledge is important, as Dollimore’s humanist insists; but more important, because more fundamental, is expressibility in language. Language compensates us for suffering because it gives us knowledge and understanding and, by dint of these two things, a certain measure of control; but it does these things by virtue of our ability to capture what is going on in words. We do not acquire understanding of the why in the sense of a genuine theodicy. That was the Christian approach, which we have left behind us, because we no longer believe, and neither do the great exemplars of Western tragedy, that there is a why in that sense. But we do, I have argued, acquire understanding of why the tragedy came about, in terms of the cognitive failures of the protagonists. (And of course it hardly needs repeating that I do not accept Dollimore’s statement that humankind is presented by tragedy as being ‘powerless to alter’ its fate.) More importantly in the present context, we acquire understanding of the what.19 By putting suffering into words, we are enabled to express what is going on in our lives. This is indeed a humanist position, for language is a human invention. It is also a transcendentalist position, because what is here in question is not expressibility in any particular human (or other) language, but expressibility in language simpliciter. (I shall return to this point, and say more about the transcendentalism of my position, in the next chapter.) Although linguistic compensation is directed at the audience rather than at tragic agents, such agents sometimes show awareness of the idea of linguistic redress in an essentially metadramatic way. The prototype is Helen in the Iliad, who tells Hector that ‘hereafter we shall be made into things of song for the men of the future’.20 So too Artemis, in the Hippolytus, promises her protégé that he will be celebrated in ritual and song:   age after age A rich harvest of tears and mourning shall be yours, And maidens’ skill in music flow perpetually To tell your story. Phaedra too shall give her name To memory, and songs recall her love for you.21 And Hecuba, in the Trojan Women, seems to console herself and her fellow sufferers with a similar thought: Our sacrifices and our prayers have all been vain. Yet, had not heaven cast down our greatness and engulfed All in the earth’s depth, Troy would be a name unknown, Our agony unrecorded, and those songs unsung Which we shall give to poets of a future age.22 Whether these things can be consolations to the suffering individual remains doubtful, as I have said,23 and Nicole Loraux may be right to

272  Tragedy and Linguistic Redress detect ‘bitter irony’ in Hecuba’s words (2002, 69). It is also true that Hecuba and the other Trojan women come to question the survival of Troy’s name.24 (But, as Pietro Pucci points out, the play itself proves them wrong: 2016, 81.) Artemis’ words at the end of the Hippolytus are hardly consolatory for Hippolytus (or Phaedra); the same goes for the promise of cult status or apotheosis envisaged in the denouements of a number of classical tragedies.25 Hardy’s Tess enjoyed a brief notoriety: ‘There are counterpoises and compensations in life; and the event which had made of her a social warning had also for the moment made her the most interesting personage in the village to many’ (1967, 111). But the cost was high. Still, however things may be for the individual, we, the tragic audience, can find consolation in the verbalization of suffering, because we see both how tragic suffering can be made the subject of great art, and how it can be expressed in language. We view the tragedy of the Aeneid ‘with the purer emotions of artistic detachment, so that we are given a higher consolation, and sorrow itself becomes a thing to be desired’.26 (The pleasure of sorrow is a familiar topos in the tradition.)27 The mere fact that someone has gone to the trouble of putting these things into words shows that the artistic expression of an unrelieved pessimism unpicks itself.28 Even if the words are mere words, lacking aesthetic merit, still someone thought them worth singing or writing down, and worth communicating, and these were themselves optimistic acts; after all, we are fighting against the end, against the inevitable annihilation, and we do so with words that give voice to our pain—words that go out Through the dark cold and the empty desolation, The wave cry, the wind cry, the vast waters Of the petrel and porpoise.29 When we so give utterance to our pain we may achieve, as F. L. Lucas says, ‘the consolation of perfect language’.30

37 Indescribability and the Authors In the last section I claimed that language can express suffering: this was important for my proposal that tragedy offers linguistic redress. But it is a large claim, which I must now defend. Indeed I will spend the bulk of this chapter defending it, because I take it to be the most controversial aspect of my proposal. Many people will find it counterintuitive: it is often said that language cannot express such scourges of human existence as suffering, pain, and cruelty. Hölderlin remarked that Oedipus’ sufferings ‘seem indescribable, unspeakable, inexpressible’ (‘unbeschreiblich, unausprechlich, unausdrücklich’: 1992–4, vol. 1, 480), and the critic Helen Cooper suggests that, in the scene of Gloucester’s blinding, ‘the terribleness of what is happening is conveyed precisely by the inadequacy

Can Suffering Be Expressed in Words? 273 of the language’ (2010, 67). Language’s alleged inadequacy in this regard is not just something that poets and critics assert: it is affirmed regularly in ordinary discourse. A severe migraine is, one says, ‘simply indescribable’; or when one contemplates an act of monstrous inhumanity, one might protest that ‘language falters before’ the reality. In Claude Lanzmann’s documentary film about the Holocaust, Shoah, Simon Srebnik, who survived a massacre at Chelmno, says, when he returns with the film crew to the site where the bodies were burnt, ‘Das kann man nicht erzählen. Niemand kann erinnern, was da war. Keiner kann das verstehen’ (‘One cannot recount it. No one can recall what happened here. No one can understand it’). ‘There are moments in Waiting for Godot,’ writes George Steiner, ‘that proclaim with painful vividness the infirmity of our moral condition: the incapacity of speech or gesture to countenance the abyss and horror of the times’ (1961, 350). We naturally talk in such terms, and we do so not only when we contemplate pain and suffering, but also their opposites, states of extreme pleasure and elation. Indeed we speak in this way when we think quite generally about many aspects of our mental lives: for example, we say that ‘words cannot express’ our emotions. I believe that such locutions should not be forced; they should not be taken literally, but sympathetically expounded. For, as I shall argue in this and the next chapter, the truth is that language can express everything— or at least can do so in a sense which I shall try to make precise. Yehuda Bauer writes (2001, 7): True, the depth of pain and suffering of Holocaust victims is difficult to describe, and writers, artists, poets, dramatists, and philosophers will forever grapple with the problem of articulating it—and as far as this is concerned, the Holocaust is certainly not unique, because ‘indescribable’ human suffering is forever there and forever being described. In principle, then, the Holocaust is a human event, so it can be explained, because it was perpetrated for what were unfortunately human reasons. This does not mean that the explanation is easy. On the contrary. This seems to me exactly right. In what follows I shall primarily be concerned not with the question of understanding (though that will play a subsidiary role), but with expression. It may seem impertinent of me to take issue with a witness like Srebnik, given everything that he went through, and the fact that my own life has been mercifully free of such horrors, but, though this is a natural thought to have and one that in some moods I have myself, I think, as I explained in the Introduction, that further reflection shows it to be mistaken. If we simply nod through such statements as Srebnik’s without engaging with them—partly perhaps because we are afraid of engaging with them, partly too, no doubt,

274  Tragedy and Linguistic Redress because we wish to suppress the uncomfortable thoughts which they prompt in us—we thereby take them less than seriously. We thereby avoid paying the kind of attention to the witnesses of the past that their statements deserve. By contrast, to pick up another point from the Introduction, in disagreeing with these statements—not all of them, but some—and in saying why we disagree with them, we show a respect for their makers, and for truth, that helps keep the flame of recollection alive. It is a remarkable fact, which anyone who explores our literary tradition is bound to notice, that authors—who do after all, one would have thought, have a professional interest and a considerable intellectual investment in the expressibility of reality in language—frequently complain about the supposed inadequacy of language to convey the real. Well, you might retort, is that so surprising? Perhaps it is precisely because they are in the business of expressing reality that creative writers are peculiarly sensitive to the difficulties of doing so; perhaps they are always running up against the limits of language and feel these limits more than the rest of us. Perhaps: but another striking fact is that, when writers do complain about the inadequacies of language, they often refute themselves out of their own mouths. Virgil’s Polyphemus is ‘shapeless’ and (in some versions of the text) ‘indescribable’, but he is described; the shield of Aeneas presents a non enarrabile textum, but the scenes that Vulcan has fashioned on it are conveyed in detail.31 In the Nibelungenlied we are told (2230–1) that, when Rüdiger’s body was laid before Kriemhild and Etzel, no poet could describe the many signs of grief displayed by women and by men, but then one such sign is described—Etzel’s roaring like a lion. Buchanan’s Herod complains that his anxiety cannot be expressed (Baptistes 524–6), but then goes on to express it very well. Chaucer tells us that, after a night of love-making, Troilus is so elated that This joie may nought writen be with inke; This passeth al that herte may bythynke.32 But he writes these words at the end of a book of the poem that has been given over to little else. Again, Chaucer’s Black Knight in the Book of the Duchess says that he cannot describe his lady’s face for lack of English and wit (896–901), but the sequel shows that he was too modest. When Thomas Mann narrates, in Der Zauberberg, how one of the patients of his sanatorium had an epileptic fit, Mann calls the consequent uproar ‘indescribable’ (‘unbeschreiblich’), but has no trouble describing it; when another inmate talks about an operation that he underwent without anaesthetic, he calls the experience indescribable (‘es erging mir ganz unbeschreiblich’), but seems to have no difficulty describing it.33 Schopenhauer calls ‘die Leiden der Menschheit’ namenlos, but seems to concede that Dante does name and describe them, in his hell, being able to do so since his materials are drawn from the sufferings of this world.34

Can Suffering Be Expressed in Words? 275 Herman Melville calls the whiteness of Moby Dick ‘well nigh ineffable’, but he takes up a whole chapter discoursing on it; Hugo von Hofmannsthal’s Chandos letter expatiates on the alleged failures of language in eloquent, even opulent terms; the ‘unspeakable’ reality alluded to by Ellida at the end of Act II of Ibsen’s The Lady from the Sea could not be spoken before now, because it would have been emotionally destructive to do so, but in the rest of the play it is spoken, and catastrophe is averted; the ‘well-beloved’ sought by Thomas Hardy’s Jocelyn Pierston ‘was indescribable’, but she is described.35 When the same author’s Clym Yeobright and Eustacia Vye meet at the Rainbarrow under an eclipse of the moon, the first moments of their Liebesrausch pass in breathless ecstasy (1974, 218): ‘They remained long without a single utterance for no language could reach the level of their condition: words were as the rusty implements of a by-gone barbarous epoch, and only to be occasionally tolerated’. But, happily for the reader, Clym and Eustacia now treat their rusty verbal implements with remarkable forbearance, enabling their author to give them over four pages of uninterrupted dialogue. Perhaps the tools are not so useless after all. You might reply that the ominous way in which the lunar eclipse presides over their conversation and fate remains unexpressed by Hardy. Here perhaps it is; but not elsewhere, not in his novel Two on a Tower, for example, or in his poem ‘At a Lunar Eclipse’ (2001, 116). Thy shadow, Earth, from Pole to Central Sea, Now steals along upon the Moon’s meek shine In even monochrome and curving line Of imperturbable serenity. How shall I link such sun-cast symmetry With the torn troubled form I know as thine, That profile, placid as a brow divine, With continents of moil and misery? And can immense Mortality but throw So small a shade, and Heaven’s high human scheme Be hemmed within the coasts yon arc implies? Is such the stellar gauge of earthly show, Nation at war with nation, brains that teem, Heroes, and women fairer than the skies? Clym, hero redux whose brain teems with philanthropic visions, and Eustacia, woman fairer than the skies, are, we sense, doomed to moil and misery. The precedents of a teeming brain are in any case not good, for Hardy is recalling Keats’s sonnet ‘When I have fears that I may cease

276  Tragedy and Linguistic Redress to be’. The contradiction that I am drawing attention to runs through our tradition from the very origins of Western literature.36 The sorrow that Hector, dying, has brought on his parents is called ‘unspeakable’ (arrēton) by Andromache in Homer’s Iliad (XXIV, 741), but how much of the last three books of that poem is given over to expressing just that grief!37 At the beginning of her influential study, The Body in Pain, Elaine Scarry cites Virginia Woolf’s remark that ‘English, which can express the thoughts of Hamlet and the tragedy of Lear, has no words for the shiver or the headache’.38 Scarry comments on pain’s ‘resistance to language’, and agrees that it is ‘nearly impossible to express’, adding that ‘physical pain is monolithically consistent in its assault on language’.39 But she herself spends several pages describing how medical science has developed, and continues to develop, a technical vocabulary for describing different types of pain, based on descriptions that patients themselves give. Later she refers to the work of Amnesty International and its attempts ‘to restore to each person tortured his or her voice, to use language to let pain give an accurate account of itself’; and her remark that ‘ordinarily there is no language for pain’ (my emphasis) perhaps concedes that language, if not adequate as it stands, is nevertheless capable of being resourced in such a way as to enable it to describe pain.40 But if it can be given those resources, then it follows that pain is expressible, if not in the language as it stands, at least in an embellished language. (Frequently the language of pain is metaphorical,41 though, like language in general, it neither is nor could be, as Nietzsche thought, exclusively metaphorical.)42 In that case language, as such, can express pain. Similar arguments work in other areas of discourse, but here I shall defend the thesis of the adequacy of language in just the area that concerns me in this study, namely tragedy and its distinctive subject matter—such phenomena as pain, loss, suffering, violence, cruelty, and inhumanity.43 Woolf’s assertion that English has no words for the shiver or the headache trips up before it gets going: how about ‘shiver’ and ‘headache’? Cynthia Marshall says that tragedy aims ‘to express the inexpressible’:44 presumably then it must fail; otherwise the inexpressible would turn out to be expressible. Colin Burrow writes, of Ovid, that ‘Repeatedly his clear-edged style is used to describe realms of experience and of pain which are too great for human words’ (2013, 106). This is a simple contradiction: the best we can do with it, from Ovid’s point of view, is to read it as crediting him with artistic success, in which case the ‘realms of experience and of pain’ are not too great for human words, but rather, perhaps, hard to express, requiring artistic skill or genius. A. J. Boyle tells us that the Schreirede of Seneca’s Thyestes ‘expresses the most inexpressible pain’ and that Seneca’s rhetoric is able to ‘articulate’ Thyestes’ ‘ineffable pain’45—again these are straight contradictions. I shall maintain

Can Suffering Be Expressed in Words? 277 that the so-called inexpressible is in fact expressible, by tragic literature among other things, that, as John Bayley puts it, ‘The language of tragedy has words for any situation’.46 The kind of thing I have in mind when I say that talk about indescribability should be sympathetically expounded rather than taken literally can be illustrated from a scene in Sophocles’ Women of Trachis. It is the crucial moment when Deianeira tries to win back Heracles’ love by smearing a garment, a gift for him to wear, with a substance that she supposes to be a love potion. The concoction was given to her by the dying centaur Nessus, and in reality it is not a love charm but a corrosive poison which acts when it comes into contact with the light. Nessus was her enemy, but Deianeira negligently fails to entertain the possibility that she might be falling into a trap. She gains a first inkling of the awful truth, and of her fatal hamartia, when she sees the effect of the potion on the tuft of wool that she had used to anoint the gift when the tuft is exposed to sunlight. As she tells the chorus (693–4): ‘When I was going out I saw a thing too strange for words (phatin aphraston), beyond human understanding’. The passage continues (695–704, tr. Lloyd-Jones): I happened to have thrown the piece of sheep’s wool into the sun’s ray; and when it grew warm, it melted away into nothing and crumbled on the ground, looking most like the sawdust you see when somebody cuts wood. So there it lay, where it had fallen; and from the ground where it was lying clotted foam boiled up, as when the rich liquid from the blue-green fruit is poured upon the ground from the vine of Bacchus. Deianeira tells the chorus that she beheld a phatin aphraston—a thing too strange for words. But she goes on to describe this thing minutely, and ‘too strange for words’ turns out to mean something like ‘amazing’, not literally ‘indescribable’. Similarly, in Aeschylus’ Persiae, Atossa tells us that she is plagued by a ‘twofold indescribable anxiety’ (διπλῆ μέριμν’ ἄφραστος, 165). ‘But,’ as A. F. Garvie says, ‘Atossa proceeds to express her double anxiety’—it is ‘for loss of wealth and men’.47 In context, Deianeira’s expression ‘a thing too strange for words’ seems to be glossed by the next phrase, ‘beyond human understanding’. The expression ‘phatin aphraston’ is a nice oxymoron, since it literally means unspeakable speech, but what Deianeira intends is not that the marvel cannot be spoken about, but that it cannot be explained. (In fact it can: the explanation comes in due course.) In her commentary, P. E. Easterling notes (ad loc.) that ‘ἄφραστος hovers between “inconceivable” and “unutterable” ’, and she compares Virgil’s ‘pecudesque locutae; / infandum!’ (‘and cattle uttered; unspeakable!’: Georgics I, 478–9) in his description of the portents that followed the death of Julius Caesar. Boyle

278  Tragedy and Linguistic Redress notes that ‘the quintessentially “tragic” word, infandus, . . . crystallize[s] the paradox of tragedy’s speaking the unspeakable’, and that ‘the request or necessity to speak the unspeakable is regular Senecan paradox’.48 Here, by contrast with the passage I quoted above, Boyle gets it right: the idea of speaking the unspeakable is a paradox, to be resolved, I am going to argue, by denying that the so-called unspeakable really is unspeakable. In Latin, there may be an etymological connection between ‘fari’, the verb meaning to speak that is present in ‘infandum’ (‘unspeakable’), and the words ‘fas’ and ‘nefas’, which we have already met in the discussion of Seneca’s Thyestes (§26), meaning right and wrong respectively.49 ‘Nefas’ would then literally mean something like unspeakable, but it would mean that in the sense not of what cannot be spoken about—of what it is literally not possible to speak about—but of what should not be spoken about because it would be blasphemous to do so, or similar. That is all very well, an objector might interpose: language may be adequate to describe the shrivelling of a tuft of wool, but what about the phenomena we started with in this chapter, and in particular what about suffering? When Heracles puts on the fatal robe and performs a sacrifice in full sunlight, the poison eats into his flesh and devours his body with excruciating pain. He cries out (1046–57; tr. Lloyd-Jones): Many and savage, evil even to relate, have been the labours of my arms and my back! And never yet has the wife of Zeus or hateful Eurystheus set such a thing upon me as the woven covering of the Erinyes which the daughter of Oeneus with beguiling face has put upon my shoulders, by which I am perishing! It has clung to my sides and eaten away my inmost flesh, and lives with me to devour the channels of my lungs. Already it has drunk my fresh blood, and my whole body is ruined, now that I am mastered by this unspeakable bondage. The bonds are called ‘unspeakable’, ‘aphrastos’, the same word as we had before. But once again, although he tells the chorus that the bonds cannot be described, Heracles seems, in the above passage and its continuation (1046–111), to describe the pain that ‘the intolerable shirt of flame’,50 the original tunica molesta, causes in him, quite well. (Still more so in the Senecan Hercules on Oeta, at 1218–78.)51 At another point in Women of Trachis, Heracles declares that the pain is so great that he wishes to have his head severed from his body (1015–16), just as Philoctetes, another classic tragic sufferer (to whom I shall return), wishes to have his painful foot cut off (Philoctetes 747–9), and later asks for a weapon with which to dismember himself (1204–9). In the Oedipus at Colonus, the ‘unspeakable’ thunder of Zeus (1464) may be ‘critical hypallage’ for Oedipus’ mysterious end;52 at any rate, it is described.

Can Suffering Be Expressed in Words? 279

38 Language and Barbarity The attempt to say that some horrific event is indescribable constantly runs into what at least appears to be self-refutation. So, for example, the Holocaust historian Saul Friedländer relates (2007, 122) that he has been reading a book about SS doctors and their child victims, ‘which describes how about twenty Jewish children, aged from five to twelve years and from all over Europe, were brought together for the purpose of medical experiments. At the end of the tests the children were hanged in the cellar of the Bullenhuser Damm school in Hamburg. At this level of horror, it is my view that descriptions are impossible’. Impossible? Or too painful to give? That descriptions are not impossible is shown by Friedländer’s own words. What is described is a horror that is almost too great to be confronted, exacerbated as it is by such details as the name of the school where the atrocity occurred. But the school can be named, and what happened in its cellars can be stated and described. One feels as though the event should have occurred in some unreachable hell, somewhere discontinuous with our space, and not in a place you can get to by taking a bus or walking. But these things do occur in our ordinary space, and our ordinary space is connected. The really upsetting fact is not that the event that Friedländer mentions cannot be expressed, but that it can be stated so briefly and so simply. ‘You have spoken so many ills (kaka) in so few words’, says Creon to Teiresias in Euripides’ Phoenician Women (917), when the latter states that Thebes can only be saved by the sacrifice of Creon’s son Menoeceus.53 Creon understandably refuses to accept the seer’s prophecy, to which refusal Teiresias harshly responds: ‘Is the truth no more because you are to undergo misfortune?’ (922). Menoeceus is duly taken away to the place of sacrifice. I said just now that one feels the event Friedländer describes should have occurred in a merely possible world, not in the actual world: but I have also said (§23) that tragic literature invites us to adopt a larger, modally (and temporally) neutral perspective. In that sense, it is quite bad enough that these things can happen. Friedländer’s example recalls a terrible scene in Jorge Semprun’s Le grand voyage (1963, 192–7), in which some children who have survived the freezing conditions of a transport to Buchenwald (all the adults have died on the train) are literally hounded to death on their arrival at the camp. Semprun’s is a fictionalized account, but it describes the kind of thing that can, and did, happen. The event is, after Semprun’s narrative manner, anticipated several times before it is related, and when the relation comes it is given in dry, factual prose that is so much better able to do justice to the grief than simply declaring that ‘words cannot convey. . .’. Words can convey, and in the name of those who are no longer around to use those words, they should be made to do so. As I said above (§26),

280  Tragedy and Linguistic Redress no one’s interests are served by the indescribability thesis. Walter Poller, an inmate at Buchenwald who acted as a medical assistant, recollects the conditions in a Sonderlager, inside the main Lager, where hundreds of Jews were confined, and subjected to slow death through malnutrition: As I enter the Lager, about 400 inmates are already dead. Even so, the remaining inmates are still in their tents—which they may only leave to collect food and go to the toilet—literally lying on top of one another. The living and the dead, the healthy and the mortally sick, the elderly and the young, the terrified and the fatalistic. Unbelievable miasma, indescribable filth, human beings whose still living bodies are rotting, the crazy and the insane, cramped and contorted, comatose—an apocalypse, such as no brain could conceive and no pen describe.54 Poller at least begins to describe the reality. Can the description go no further? Or, no matter how far it goes, will it always fall short? It will not of course be the reality, but will it fall short as description? (I shall return to this crucial point shortly.) Primo Levi wrote of his internment at Auschwitz, ‘Then for the first time we became aware that our language lacks words to express this offence, the demolition of man’.55 Again, has Levi not expressed the offence? Does the phrase ‘the demolition of man’ not express the offence Levi means? Or not really express it, not adequately? In that case, could we find another form of words to express it? Or is there simply no form of words that will do the job? As Lawrence Langer observed, Levi, perhaps more than any other writer, brought the experience of the camps to those who have no natural connection with the Holocaust and have not undergone anything remotely comparable to it: for Levi ‘found the language to confirm its deep and permanent impact on our time’.56 Langer quotes part of a passage from the notes that Rolf Hochhut wrote to preface a scene of his drama about the Holocaust, Der Stellvertreter, in which it is said that (1998, 296) ‘What the most significant events and discoveries of our time have in common is that they place too great demands on human powers of imagination. No imagination is able to put before our eyes Auschwitz or the destruction of Dresden or of Hiroshima or exploratory flights in space or even industrial capacity and speed records’. This list of allegedly unimaginable things gets less plausible as it goes on, until we arrive at the frankly ridiculous suggestion that there is something unimaginable about industrial capacity and speed records. As far as Auschwitz is concerned, Langer comments that ‘So much commentary currently exists on the camp . . . that such an attitude [as Hochhut’s] now appears naïve and even antiquated’ (1995, 95). And, as Bauer remarks, ‘The history of the Holocaust tells us of horrors and brutalities that are “indescribable”, by which we mean that we

Can Suffering Be Expressed in Words? 281 view them with total revulsion—but they have of course been described, which is how we know about them’ (2001, 18). Richard Bernstein tells us that Auschwitz ‘is the type of evil that resists what Kant calls “synthesis”. It defies what Kant took to be essential for experience—that it can be synthesized, conceptualized, and categorized’ (2002, 228). This is a characteristic example of postmodernist excess, which (whether or not that is the intention) blocks the progress of science in the name of a primitive mysterianism; but then it was a primitive mysterianism that, at least in part, led to Nazism in the first place. As Giorgio Agamben points out (2012, 29–30), talk of the unsayability of Auschwitz risks playing the Nazis’ game—elevating the thing into an object of almost religious significance, even reverence. There is no difficulty about Kantian synthesis in the present case: Auschwitz can be—and has been by the many records and meditations on it that exist—conceptualized and categorized. If a room that serves as a gas chamber is described as such in a piece of discourse about it, what exactly remains unsynthesized, or unsaid? Kantian synthesis involves the bringing of experience under concepts. We may not agree with Kant that experience needs to be actively brought under concepts, as though the experiencing subject started with a sensual Given that was in itself non-conceptual, and then conceptualized it; we may agree with John McDowell (1994) that the idea of the Given is a myth, and that experience is fully conceptually structured by the time (or rather: by the logical stage at which) anything like experience comes into view. But, whichever way one leans on this philosophical question, it is hard to see that there is anything about Auschwitz that resists the conceptual. That a room is a gas chamber is an appalling fact about it, but that appallingness does not reside in any kind of failure to bring, or difficulty about bringing, an object (that very object) under a concept (that very concept). On the contrary, it is precisely and only because the object does fall squarely under the concept that we are horrified, and have the right to be horrified. If the object resisted categorization, we would not have that right. Ironically, if Bernstein were correct, he would not be entitled to call Auschwitz evil, or indeed anything else: there would in fact be nothing there to call anything. So it is a mistake to contest the categorizability of Auschwitz and its chambers of horror in the name of impressive-sounding but incoherent mystery-mongering.57 It is terrible to learn that the building where the Lebensborn programme was carried out, Schloss Hartheim, was converted into ordinary dwellings after the War. But it is terrible not because of what that sentence does not say, still less because of what it cannot say, but because of what it can and does say. Filip Müller’s extraordinary account of the three years that he passed as a member of the Sonderkommando in the crematoria at Auschwitz makes for sombre, indeed devastating reading, but not, I suggest, because one has the feeling that anything is left unexpressed or under-expressed:

282  Tragedy and Linguistic Redress quite the reverse. It is because the experience of hell is so minutely and accurately described that the reader is so profoundly shocked.58 One of the most upsetting incidents described in the book is Müller’s despairing impulse to take his own life, by joining his fellow countrymen from the ‘Family Camp’ in the gas chamber. His attempt to hide behind a pillar so that he will not be recognized fails and a group of girls, ‘naked and in the full bloom of youth’, approach him: ‘They stood in front of me without a word, gazing at me deep in thought and shaking their heads uncomprehendingly. At last one of them plucked up courage and spoke to me: “We understand that you have chosen to die with us of your own free will, and we have come to tell you that we think your decision pointless: for it helps no one” ’ (1979, 113). The speaker, Yana, tells him to take her necklace, when she is dead, and give it to her boyfriend Sasha. The girls then grab Müller and propel him over the threshold and out of the room of death, where he is immediately seized by the SS men guarding the doors and forced to return to the cremation hall upstairs. Later Müller finds Yana’s body, recovers her necklace and gives it to Sasha as she had desired. One thing that emerges powerfully from Müller’s narration is the significance of the threshold to the gas chamber, as the line separating life from death. This is a significance with which, again, tragic literature can perhaps assist us, for the threshold to the house is a fundamental symbol of the Greek tragic stage: it is crossed by Aeschylus’ Agamemnon, for example, who does not realize its significance, and by his Cassandra and Clytemnestra, who do.59 I suggest that, so far from its being the case that language is inadequate to capture the reality of the horror that Müller underwent, it both can do so and is the only thing that can do so. If you visit Auschwitz today, you find a vast complex of buildings and ruins, overrun by visitors and surrounded by the usual accoutrements of major tourist attractions—shops, restaurants, parking areas for cars and (especially) coaches, and so on. The original camp, Auschwitz I, has largely survived intact, including its gas chamber and crematorium, but at the much larger Auschwitz II, the Birkenau camp, the SS tried to destroy the traces of their crimes by detonating crematoria II, III, IV, and V, before abandoning the site under the threat of the approaching Red Army. Just beyond the ruins of these crematoria, with their now roofless gas chambers and undressing rooms, lie thick woods—planted by the SS to mask their monstrous deeds—which, when I visited in mid-May, were full of the sound of cuckoos, golden orioles, nightingales, and the usual springtime songsters. I surprised a fox-cub nosing among the moonscape slabs of the undressing room of crematorium III. Nature is close—that nature to which, as former inmate Tadeusz Sobolewicz tells us, ‘we no longer belonged’ (2011, 95). An eerie effect—or perhaps not, for what could be more natural than that nature should return? When you stand by these ruins, or even when you stand, as you may do, in the gas chamber of Auschwitz I, what is there to think,

Can Suffering Be Expressed in Words? 283 other than that this is where it all happened, and that, when it did, the place less resembled a tourist destination or a nature reserve than Dante’s Inferno? To know what happened there and then, you have to read the eyewitness accounts and historical reports. And these are linguistic. It is language, not place, which illumines the facts and gives them significance. Against the indifference of the place we may set the zeal of the word. When Jean Améry was arrested by the Gestapo in wartime Belgium for his participation in the Resistance, he was taken to Breendonk prison where he was tortured. In the bunker there hung from the vaulted ceiling a chain that above ran into a roll. At its bottom end it bore a heavy, broadly curved iron hook. I was led to the instrument. The hook gripped into the shackle that held my hands together behind my back. Then I was raised with the chain until I hung about a meter over the floor. In such a position, or rather, when hanging this way, with your hands behind your back, for a short time you can hold at a half-oblique through muscular force. During these few minutes, when you are already expending your utmost strength, when sweat has already appeared on your forehead and lips, and you are breathing in gasps, you will not answer any questions. Accomplices? Addresses? Meeting places? You hardly hear it. All your life is gathered in a single, limited area of the body, the shoulder joints, and it does not react; for it exhausts itself completely in the expenditure of energy. But this cannot last long, even with people who have a strong physical constitution. As for me, I had to give up rather quickly. And now there was a crackling and splintering in my shoulders that my body has not forgotten until this hour. The balls sprang from their sockets. My own body weight caused luxation; I fell into a void and now hung by my dislocated arms, which had been torn high from behind and were now twisted over my head.60 The torture is described: but can language express the pain to which Améry was subjected? He himself answers our question, shortly after the passage just quoted, as follows: It would be totally senseless to try and describe here the pain that was inflicted on me. Was it ‘like a red-hot iron in my shoulders’, and was another ‘like a dull wooden stake that had been driven into the back of my head’? One comparison would only stand for the other, and in the end we would be hoaxed by turn on the hopeless merrygo-round of figurative speech. The pain was what it was. Beyond that there is nothing to say. Qualities of feeling are as incomparable as they are indescribable. They mark the limit of the capacity of language to communicate. If someone wanted to impart his physical

284  Tragedy and Linguistic Redress pain, he would be forced to inflict it and thereby become a torturer himself.61 The view that Améry here states, namely that language cannot describe physical pain, and that the only way to convey or impart or express a painful feeling to someone else would be to inflict the pain on that person, is one that would, I think, be quite widely shared. Nevertheless, I believe that it is mistaken. Comparisons such as the ones Améry offers do indeed tell us what the pain was like, in the only legitimate (namely the dyadic and comparative) sense of that phrase. They liken one thing to another; they tell us what it resembles.62 That is something one can do with, plausibly, any subject matter. So I think Améry is wrong to object to the comparisons he gives: at least as far as they go, they do what comparisons are meant to do. Of course they are no more than comparisons, but it does not follow that they are useless, or that the only correct thing to say is that ‘The pain was what it was’. If you want to know what Améry’s pain was like, you want to know what it resembled, and then comparisons can indeed answer that question for you. Hence Améry is also wrong, in my view, to say that ‘Qualities of feeling are as incomparable as they are indescribable’: the comparisons he himself offers refute this assertion, for they both compare the pain and describe it. But here I suspect that many people would be inclined to protest that this argument misses the point. The point that it is allegedly missing can be expressed in various ways, but one common way is to say that these comparisons do not really, or adequately, describe the pain, and that to convey that information you would need to subject your auditor to the pain, as indeed Améry says at the end of our second passage. The trouble with language, Améry complains, is that all it can give you are words and more words: it cannot give you the thing, which in this case is a pain. But putting the matter in this way points immediately to an obvious response. It is quite true that language can only give you words; words are its business. And it is also true that it cannot give you the pain, if by that is meant: transfer the pain from me to you, or evoke in you a qualitatively similar pain to the one I feel. But these agreed facts do not entail that language cannot express the pain, and do so really and adequately. That it can do so adequately is shown by the incoherence of the idea that language as such might at best partially express a stretch of reality. A given item of language—a particular sentence or collection of sentences in a particular language—might of course express something inadequately: this, as well as being possible, is an extremely common occurrence. But in that very possibility is comprised the further possibility that another item of language improves on the inadequacy of the first attempt. A shot can be wide of the mark, but only if another shot can be on target. Partial adequacy of expression is

Can Suffering Be Expressed in Words? 285 not an option for language as such (that is, including possible as well as actual languages); if it were, language could not express anything at all. A given language (or piece of language) can be merely partially successful; but for language as such the only alternative to complete success is complete failure. And—as philosophers, asking how language works and what it is—we cannot countenance the latter option, so we must select the former. (I shall return to this important point in the next chapter.) Language can express pain, and everything else. ‘Express’ in what sense? Name and describe. There is a name for the torture than Améry underwent, the ‘strappado’: it is an old name for an old torture. Scarry quotes, from the Amnesty International archives, an account—similarly harrowing to Améry’s—of a prisoner in Paraguay who was repeatedly submerged in water and resuscitated each time just before he drowned (1985, 43). She comments: ‘To attach any name, any word to the wilful infliction of this bodily agony is to make language and civilization participate in their own destruction’ (ibid.). As the sequel makes clear, she principally has in mind the use of euphemistic names, such as ‘the submarine’ for the Paraguayan torture, but her claim is a general one and, considered as a general claim, it is patently mistaken. Euphemistic names are, we may agree, objectionable on several grounds, one of which is that they make it easier for the perpetrators to commit their crimes, laughing and joking as they do so.63 But naming a form of torture is not in itself objectionable, and it certainly does not ‘make language and civilization participate in their own destruction’. Quite the reverse. As I have said, naming a thing is a crucial preliminary to confronting it. Child abuse is a case in point.64 This has always existed and, where not named, has always been nameable, but the fact that it has, so often and in so many places, not been named has assisted its perpetration. What is not spoken about can be thought of as not really occurring. Using language to name and describe abuse and torture is an important step on the road to confronting and combating them. Language does not work by destroying the pain, as Scarry goes on to assert: ‘by objectifying it in language . . . the pain itself is diminished and destroyed’ (ibid., 51; cf. 54, 56). It would be convenient if language did have the magical power to destroy pain, but regrettably it is not so. The pain goes on being there, whether it is named or not. And there are many kinds of pain that arise organically and are not inflicted by one human being on another: maybe this natural pain will one day be eliminated by medical science, but in the meantime you cannot get rid of a migraine, for example, by naming it, describing it, or by any other linguistic stratagem. Naming has no physical effect on pain. It has no metaphysical effect, either: even before it is named, the pain is, as I have said, nameable. But naming and describing pain are good and important things to do, especially in the service of eliminating, or seeking to eliminate, pain that is wantonly inflicted.65

286  Tragedy and Linguistic Redress

39 Diagnosis of the Syndrome In Shakespeare’s Richard II, Bolingbroke says that he ‘sighed my English breath in foreign clouds, / Eating the bitter bread of banishment’ (III, 1, 20–1). In Euripides’ Phoenician Women, we find the following exchange between Jocasta and her son Polyneices, who has returned from exile to claim the Theban throne (388–9; tr. Kovacs): JOCASTA: What

is it like to be deprived of your country? Is it a great calamity? POLYNEICES:  The greatest: the reality far surpasses the description. Shortly after this the point is repeated (406–7; tr. Kovacs):66 JOCASTA:  Dearest POLYNEICES:  Your

to men, it seems, is native soil. words could not describe how dear it is.67

In the context of these exchanges we have a sense that Polyneices is led into hyperbole by a blindness to his own real nature and motives,68 the violence of which emerges from Apollo’s oracle to Adrastus that he would marry his daughters to two beasts (Polyneices and Tydeus are meant). But, whatever his words show about his own character, the claims Polyneices makes are ones that are of a general form and are commonly encountered, and so may be examined as such. Reflection on this case will lead us naturally on to a diagnosis of the reason why it is so commonly held that suffering cannot be expressed in words. Consider the first of the two quoted exchanges. Polyneices’ reply to Jocasta’s question is, literally, that exile is greater in reality than in word.69 What is going on here? What, one wants to ask, is the common dimension of assessment, the common currency, in terms of which we are supposed to compare reality and description and find the reality ‘greater’—that is, worse—than the description? Surely that makes no sense; is it not indeed a kind of category mistake? Consider the possibilities. Perhaps Polyneices’ expression means that undergoing the reality of exile is worse than hearing a description of it. But that cannot be right. Are we supposed to think that undergoing the reality of exile is worse than hearing a description of it come what may? Hardly: it cannot be worse than undergoing exile and hearing a description of it while one is in exile (unless hearing the description lightens the load: but it might have the opposite effect). Or is the sense that undergoing the reality of exile is worse than hearing a description of it when one is not in exile? But that cannot be right either: it might be true, but it is not worth saying. Presumably being in exile is worse than not being in exile, period. At least, that will usually be so, though not invariably: exile can be enjoyable; it can even, as in Aeneas’ case, be in a manner a coming home.70 For that reason we cannot accept an interpretation of Polyneices’ words along the

Can Suffering Be Expressed in Words? 287 lines of: ‘the experienced reality of exile is worse than the reality that any description can convey’. Compare a fulfilling and successful exile with the description of an awful one. Donald Mastronarde translates the crucial phrase in our text as ‘worse in fact than people say it is’, remarking (ad loc.) that the contrast is ‘rather colourless’. Worse in fact than some people say it is? That would be intolerably weak. Worse than everyone says it is? That would be false: the bitterness of exile was a common topos of ancient literature, as we know particularly from the cases of Cicero and Ovid and Seneca and Boethius, all of whom had to endure it, from Euripides, who thematized it in Children of Heracles, and from Virgil’s first and ninth Eclogues.71 We are not doing very well in our attempt to make sense of the passage quoted from Euripides, but let us keep trying. How about: ‘the reality of exile is worse than can be conveyed in language’? But to say ‘actual exile is worse than can be conveyed in language’ or ‘. . . is worse than the content of any description’ just invites someone to concoct a particularly nasty description—a description of Polyneices’ actual experience of exile and then some. And to say that exile is worse than the description itself looks absurd. For the description is just a bunch of words, and a bunch of words cannot as such be good or bad, at least not in the sense in which exile can be good or bad. So I think we are forced to return to the suggestion initially proffered above, that there is a category mistake here. Words are one thing, the reality they are about quite another, and a direct comparison between them—though in principle possible, because in principle you can compare anything with anything—in practice makes little sense. Category mistakes are often the subject of jokes, and we indeed find Shakespeare amusing himself with the error in question by making the procurer Boult, in Pericles, say ‘There was a Spaniard’s mouth watered’ when the latter heard a description of Marina, ‘as [i.e. as if] he went to bed to her very description’ (16, 95–7). So I suggest that all Polyneices’ assertion can reasonably come to is that (hearing) a description of exile cannot be, or substitute for, the experience of it. Unless something like this is meant, his meaning is categorially faulty. It makes sense to say ‘Being in exile is worse than being impoverished in one’s native land’, for example, or ‘Being tortured is worse than being exiled’, but it would be absurd to say ‘Being exiled is worse than a description of being exiled’, unless this claim is glossed in the suggested way. Similarly, you can intelligibly say ‘These apples are tastier than those pears’, or ‘These apples are less tasty than those oranges’, and it might also make sense to say ‘This description of apples is tastier than that description’—where you might mean, for instance, that the one description was an aesthetically more pleasing use of language than the other, or even, at a stretch, that the one description described tastier apples than the other description—but it would make no sense, or at any rate no literal sense, for you to say ‘These apples are tastier than that description’, or again ‘These apples are tastier than any description of them can

288  Tragedy and Linguistic Redress convey’. To this latter assertion one could only respond: ‘How do you mean? Tastier even than would be conveyed by a description of apples taken from the Garden of Paradise?’ What will you reply? ‘Well no, not as tasty as those apples, obviously’, or ‘Yes, tastier even than that description can convey, because the description is just words, which have no taste, whereas these things here are apples’? These reflections put us in a position, now, to give a diagnosis of the linguistic syndrome that we are discussing, that is, the tendency to assert that language is inadequate to express, for example, pain. When people complain that language is inadequate to express pain, they are, in effect, complaining that the words we use to talk about pains are not themselves pains. ‘In effect’ here registers an application of the principle of charity: the only coherent thing to mean by talk of language’s inadequacy to express pain is that words for pains are not themselves pains. But now that is agreed on all sides, and does not imply any inadequacy in language. For similarly, it has to be said, the words we use to talk about tables are not themselves tables; but no one complains about them on that basis. In that case, it seems good enough that the word ‘table’ merely refers to tablehood. Surely no one would say ‘A word cannot really express tablehood unless it is itself a table’? In fact some philosophers have said that, or something rather similar. Sextus Empiricus reports Gorgias as arguing that, because language is different from the ‘external’ objects it is about, it cannot express these objects.72 And sometimes philosophers have objected to the abstractive nature of language, that is, to the fact that language is selective, that on any occasion of use it treats only of certain aspects of reality: to give a simple example, if I say ‘the cup has coffee in it’, I tell you nothing about the shape of the cup or its colour.73 I can, in a supplementary statement, tell you about these other things, but when I do so there will necessarily be further properties of the cup that I miss out. To circumvent this feature of language—its digital as opposed to analogue method of conveying information (contrast a photograph)—it has sometimes been held that although we use words to talk about things, we could instead use the things themselves, and even that it might be better to do so. The early Wittgenstein’s picture theory of meaning is based on the thesis that words go proxy for things in the sense that they do duty for things as a mere technical convenience: in principle we could employ the things themselves, as in Swift’s famous pastiche of the Royal Society.74 I have criticized this approach to language elsewhere.75 Its basic problem, I believe, is that it ignores the syntactic nature of language. Words are inflected for case (either morphologically, or by position in the sentence); they have syntactic properties. Non-verbal things, such as tables and chairs, are not, and cannot conveniently be, inflected. I put this denial in a qualified way because we may be able to envisage situations or practices in which certain physical objects—it is hard to see how it could

Can Suffering Be Expressed in Words? 289 work with abstract objects—are inflected, in such a way as to mimic syntax. If that could be achieved, then to that extent we might be able to use non-verbal things to communicate, instead of words. But clearly that would be a highly artificial scenario, parasitic on language as we actually have it, which consists of syntactically propertied words. There is no possibility that words might be superseded by non-verbal things as nonverbal things are—that is, by syntactically uninflected non-verbal things. And the important point is that, even if we could, by dint of a bit of clever carpentry, endow a table with syntactic properties, and then use it to make utterances about tables, we would not be improving on language as we have it; we would not be repairing a deficiency in our ordinary language. We would not say ‘Now we can dispense with that cumbersome word “table” and use real tables instead’. Anyone who attempted that would find himself in trouble long before he tried to make an assertion about the Himalayas, or the moon, or the number two.

40 Describing and Picturing In his epic poem Childe Harold’s Pilgrimage, Byron devotes some stanzas to Rousseau, in one of which he praises Rousseau’s skill in putting his emotions into words (III, 77, 725–33). But when he comes to the ‘memorable kiss’ which he received every morning from the Comtesse d’Houdetot (III, 79, 746–7), Byron tells us in a footnote that Rousseau’s own words describing the event ‘must be felt, from their very force, to be inadequate to the delineation: a painting can give no sufficient idea of the ocean’.76 This passage contains something like the mistake I have identified, of confusing words with things. Byron here relies on the Horatian ut pictura poesis tradition, which has been so influential and (I suggest) damaging in aesthetic theory.77 For although it would make good sense to say that a painting can give no sufficient idea of the ocean, that is only because a painting is a thing of the same general sort as the ocean. A painting, like the ocean, is a material object that presents a certain visual appearance. And, though as a sheer physical object a painting is (more or less) two-dimensional, unlike the ocean, as a representation it has depth, and so achieves three-dimensionality, like the ocean. We can accordingly ask whether a painting, qua representation, resembles the ocean, and this is the question whether the painting, qua representation, and the ocean have certain interesting and salient properties in common. We might then conclude of a given picture that it indeed resembled the ocean, or, with Byron, that it did not, perhaps because it failed to capture a sense of the ocean’s vastness, or constant restless motion, or similar. But it makes no sense to raise the question whether a sentence resembles the ocean. Or rather, in one way it does make sense to ask that question, for, as I noted above, it can be asked (even if only, sometimes, in a degenerate sense) of anything at all whether it is like anything at all—you can ask

290  Tragedy and Linguistic Redress whether the current rate of inflation is like the planet Mercury, or whether Julius Caesar is like the complex number 3 + 4i—but the answer in the case of sentence and ocean will be a clear negative, regardless of the sentence’s content. For a sentence qua type is an abstract object, whereas the ocean is a concrete one, and a given abstract object and a given concrete object will typically have almost no interesting and substantial properties in common, only gerrymandered and uninteresting ones such as being thought about by the Pope, distinct from pi, self-identical, and so on. Similarly, to recur to an example I used in the last section, there is a degenerate sense in which an apple is tastier than any description, because a description, being an abstract object, has no taste at all: but in that sense the most insipid apple, so long as it has some taste, is tastier than any description. Even if we transfer our attention from type to token sentences, which are concrete objects, the question whether they are like the ocean is not one that would exercise any non-philosopher, or any philosopher with a sensible view of meaning. It is only if one is captivated by something like the picture theory, according to which words stand proxy for objects, that one will wonder whether a sentence is like the reality that it is about. Only so will one contemplate the question whether the sentence ‘The ocean is a vast, blue object in constant restless motion’, for example, is like the oceanic reality it describes, and conclude either that it is or perhaps, with the early Wittgenstein, that it will be after analysis, or, more likely, that it is not like the ocean, either before or after analysis, given that there is, one would suppose, nothing vast, blue, or restlessly moving about that sentence (whether type or token), or any sentence that accurately analyses it. It is another way of putting the same point to say that, while the painting has representational content, the sentence does not; it has semantic content.78 These are two quite different things. The painting has representational content by virtue of having a visual appearance that shares or purports to share at least some interesting visual properties with the represented object. (It may not share them sufficiently, in which case we might criticize it.) The sentence means something; it does not, qua abstract type, look like or purport to look like anything. (Qua token it does have a visual appearance if it is a written token, or an aural appearance if it is a spoken token, but those appearances are—at least in general—irrelevant to its content.) So how is it that, in Byron’s view, Rousseau’s words cannot supply an adequate delineation of his feelings? If all that Byron is really saying is ‘After all, the words are not the feelings: talking or reading about being in love is one thing, actually being in love quite another’, then everyone can agree; but that is no criticism of language, seeing that the word ‘love’ purports merely to refer to love, and has no ambitions to be love, to be in love, to resemble love, to conjure up a loving feeling in your mind whenever you use it, to be performative (so that, for example, whenever you sincerely utter the words ‘I love you’ they are guaranteed to be true), or anything of the sort. But if that is what Byron means, then the painting

Can Suffering Be Expressed in Words? 291 analogy is simply irrelevant, indeed positively misleading, since the reason, in any particular case, why we criticize a picture for giving ‘no sufficient idea of the ocean’ is not that we aim to make the observation that the picture hanging on my wall, say, is not the same thing as the ocean, but rather that we want to say that it is not a very good representation of the ocean—for example, it is not very realistic—in a number of crucial oceanic respects. In one of his sonnets (103), Shakespeare seems to treat quite seriously the idea that his verbal description of his subject is in competition with the image produced by a mirror. When one sees an image in a mirror, the image can be thought of as a representation of the original, or one can consider that one actually sees the original, via a deviant causal route. Either way, one is presented with a visual appearance, and mirrors can distort the appearance of the original, so that it makes sense to compare what one sees in a mirror with what one would see if one looked directly at the original, and to criticize the mirror on grounds of accuracy. What does not make sense is the idea that the original might be compared with a verbal description of it, and the latter, no matter how skilfully crafted, then judged wanting on the basis that it does not look like the original. The ut pictura poesis tradition was famously challenged by Lessing in his Laokoon, but his way of doing so, ironically, had the effect of reinforcing the idea that came to fruition in Wittgenstein’s picture theory. For Lessing repudiated the assimilation of poetry to painting not on the grounds that the whole idea is misconceived—which I take to be the right response—but on the much tamer, and surely mistaken, grounds that they represent in different ways: a painting, being a spatial array of (say) brush strokes on canvas, represents static, spatially arranged bodies, whereas a piece of poetry, being a succession of sounds or marks on paper (Lessing has the acts of recitation and reading in mind), represents bodies in time, and especially actions.79 The obvious difficulty with this distinction is that poetry, whether recited or read, can represent static bodies spatially arranged. Lessing deals with this objection by arguing that when poetry tries to describe a complex spatial object, such as the shield of Achilles, it does so by going through the parts successively, and that this means that it cannot give you the object in its simultaneity and wholeness.80 But the basis of this view is his assertion, which anticipates the spirit of the Wittgensteinian picture theory, that ‘signs that are arrayed next to one another [in space] can only express (ausdrücken) objects (or their parts) that exist next to one another [in space], whereas signs that follow one another [in time] can only express objects (or their parts) that follow one another in time’81—and that is manifestly wrong in both its clauses.

41 Explanations of the Error Referring to the second passage from Améry that I quoted in §38 above, Robert Eaglestone comments: ‘It is not, or not only, the pain and suffering

292  Tragedy and Linguistic Redress that is incomprehensible. This is not special to Holocaust testimony: any representation of physical pain or other sensations is not those sensations. Pain can be described but, through description, it cannot be experienced by another’ (2004, 18). I have agreed that a description of pain is not the same as an experience of it, and does not give one an experience of it, either, just as a description of love does not give one an experience of love. So the agreed fact is that the word ‘pain’ does not give you (an experience of) pain. But now what does that show? The fact in question is not peculiar to pain: a description of anything is just a description of that thing, and cannot give you either that thing or an experience of it. In the case of so-called ‘inner’ experience, such as feeling a pain, we may wish to identify pain, say, with the experience of pain, and in that case there will be just one thing that we are agreeing a verbal description does not give you, namely a pain (= an experience of pain). In the case of ‘outer’ experience, such as seeing a table, we will not make the identification I have just mooted in respect of ‘inner’ experience: that is, everyone will agree that a table is quite distinct from the experience of a table. So there we will say that a description of a table does not give you either a table or the experience of a table. But modulo that subtlety, the point in both cases—‘inner’ and ‘outer’ experience—is the same. To take the ‘outer’ case: a description of a table is made up of words, words are not tables, or pictures of tables, or any other things that interestingly resemble tables, and the description does not, just as such, convey an experience of tables. (Uttering, or hearing, the word ‘table’ might in some subjects causally induce an experience with tablehood content; but then so might a blow on the head. In either case the connection would be, from a semantic point of view, purely adventitious. That was a point that the later Wittgenstein constantly stressed, in opposition to the empiricists’ theory of meaning.) We are back here with something like the observation that the picture of the ocean hanging on my wall is not the same thing as the ocean itself, and Eaglestone indeed, in the middle of the quoted passage, makes precisely that observation, mutatis mutandis, about pain and its representation (though, as I have noted, it is inaccurate to talk of representation where what is in question is language, which is a vehicle of semantic meaning, not a representation). In the first sentence of the quoted passage, Eaglestone probably means that ‘it is not only the pain and the suffering of the Holocaust that is incomprehensible’, but even with this gloss the sentence does not connect with the sequel, for from the fact that ‘any representation of physical pain or other sensations is not those sensations’ it does not follow that those sensations are in any respect incomprehensible. You cannot get such a substantial and portentous conclusion out of the mere fact that pain, the sensation, is not identical with ‘pain’, the word. When people talk of the inadequacy of language, they are, I have suggested (§39), in effect expressing the point that (to put it compendiously)

Can Suffering Be Expressed in Words? 293 language is not the world. Language is indeed not the world, and it does not go proxy for the world, either. Words are one thing, the reality that they are about is quite another—at least usually: words can be about words, but that is not the normal case. The confusing of language with what it is about, of words (and other linguistic items) with their referents, runs like an Ariadne’s thread through our intellectual tradition, from Gorgias (§39) to Cora Diamond’s influential but bogus idea of the ‘difficulty of reality’.82 Why is the mistake so prevalent? Why are we quite happy, raising no objection, that the word ‘table’ should not be (or produce an experience of) a table, but when it comes to the word ‘pain’ there is a widespread tendency to complain that (in effect) the word is not (and does not produce an experience of) a pain, and then to conclude that pain is unnameable, unspeakable, indescribable, or even incomprehensible? Why are we tempted to impose standards on the ‘inner’ that we do not impose on the ‘outer’—to say things like ‘I was horrified at what she said to me, but in uttering these words I come nowhere near to conveying the emotion that I then felt’, but not to say things like ‘The cat was sitting there on the mat, but really you cannot express their relative locations in words: you might think that “superposition” did it, but actually that word conveys only a pale shadow of the full-blooded cat-on-the-mat reality’? I think I can identify five reasons for the tendency; there may be others. My first reason has been anticipated earlier in this chapter. It is often felt that we can only do justice to the sheer awfulness of something by calling it ‘indescribable’.83 In the only words to survive of Varius’ Thyestes, his Atreus says ‘iam fero infandissima, iam facere cogor’ (‘Now I endure, and now I am compelled to perform, most unspeakable things’); the altars and swords that Virgil’s Sinon in his lying tale claims to have fled are ‘unspeakable’ (‘nefandi’); the documented facts of his father’s crimes against Stefonio’s Crispus are ‘unspeakable, dreadful, bloody’ (‘infanda, horrida, fera’).84 What has been done to Atreus and what he does in turn, Sinon’s imaginary altars and swords, and the crimes of Crispus’s father can all be described, but they are terrible. It is in my first suggested sense, perhaps, that Iago’s motivation is ‘unspeakable’.85 In this case, too, and regardless whether or not we accept his own statements of motive, Iago’s reasons for his heinous offence can be described. Coleridge may be right that in giving his reasons Iago was ‘motive hunting’, but the idea that he was prompted by ‘motiveless malignity’ is not satisfactory: there are specifiable grounds of Iago’s hatred.86 Still, we shudder at what he does, and perhaps feel that calling it unspeakable is the only way we can do our revulsion justice. When the chorus of Sophocles’ Ajax talks about ‘the unspeakable deeds of the two ruthless Atridae’, again these deeds can be described, but it is as though the chorus fears that description would somehow lessen or cheapen their awfulness.87 In fact it is the other way round, as I have several times indicated in my earlier discussions. Simple description of the facts, as for example in Filip Müller’s eyewitness

294  Tragedy and Linguistic Redress account of his activity in the Birkenau Sonderkommando, or in the deadpan narrative of Tadeusz Borowski’s Bei uns in Auschwitz, imparts the awfulness as nothing else can; throwing up one’s hands and proclaiming ‘words cannot convey. . .’ does not. Conversely, if something is especially good, we talk of it as being beyond words because we feel that words would somehow diminish its goodness: recall Lamb’s ‘unspeakable rural solitudes’, or Goneril’s ‘Sir, I do love you more than word can wield the matter’.88 A second reason for the mistake was also identified earlier in this chapter (§37) when it was noted, in connection with the Latin word ‘nefas’ and Sophocles’ description of Zeus’ thunder as unspeakable, that what should not be named, or spoken, or described—perhaps because it is anathema, or taboo89—is sometimes conflated with what cannot be named, or spoken, or described. In Aeschylus’ Seven Against Thebes, the mutual fratricide of Eteocles and Polyneices ‘cannot be spoken about, only bewailed’ (846–7).90 In Sophocles’ Oedipus Rex, the chorus tells us that the horrors wrought by the slayer of Laius are ‘utterly unspeakable’ (465; cf. 1289). But it is not that they cannot be spoken about, as many passages in the play show; it is rather that they in some sense ought not to be spoken about. Similarly with the charge that Theseus tells us Phaedra brought against Hippolytus.91 The pictures with which Hölderlin opens his poetic ‘In lieblicher Bläue’ are so holy, he says, that he is afraid to describe them (1992–4, vol. 1, 479). When St Paul was snatched up to Paradise, he heard arcana verba that might not lawfully be uttered to men (2 Cor. 12, 4). In Hemingway’s For Whom the Bell Tolls, the terrible massacres narrated by Pilan take place in a village that, significantly, is not named, as if it were too terrible to name. But it has a name. In ancient society, as in modern, there are certain words which can be uttered but which, in particular contexts, may not be uttered.92 It is even the case that a language may lack a word to designate something which its speakers regard as too terrible—or, as I suggested above in connection with the phenomenon of child abuse, which they are too ashamed—to name, as for example there is no generic word for ‘incest’ in classical Greek. In Shelley’s The Cenci, the incest of Beatrice and her father is ‘unutterable’; it is an act ‘without a name’; it is ranged under those ‘deeds / Which have no form, sufferings which have no tongue’.93 But incest can be named, and was indeed explicitly named in many early modern plays, as for instance in Shakespeare’s Hamlet, Beaumont and Fletcher’s A King and No King, and Dryden’s Aureng-Zebe. When Euripides’ Hecuba in her eponymous drama characterizes her son Polydorus’ murder, she says that these things are (714–15; cf. 197–201) ἄρρητ᾽ ἀνωνόμαστα, θαυμάτων πέρα, οὐχ ὅσι᾽ οὐδ᾽ ἀνεκτά, unspeakable, unnameable, beyond the merely miraculous, unholy and not to be borne.

Can Suffering Be Expressed in Words? 295 The second line glosses and gives the content of the first. It is not that the deeds in question cannot be spoken or named, but that it would be impious to do so.94 A third reason for the common censure of language on grounds of expressive inadequacy is this. It may be that certain kinds of description of experience can only be understood by those who have had similar experiences themselves. This may be part of what Améry is getting at in the passages of his quoted above. If it is right that only those who have had certain experiences can understand—or fully understand—words describing those experiences, there would be nothing strange or unusual in this. For it is quite generally the case that many words denoting concrete objects or perceptible properties can only be understood by those who have relevant experiences to call upon. Colour words are a case in point. You cannot understand the word ‘red’ unless you at some stage experience the colour. This is not a matter of your having to know ‘what red is like’ if you are really to understand the word ‘red’. In the legitimate sense of that phrase, according to which (adapting an example of Locke’s) one might know that red is like the sound of a trumpet,95 such knowledge is irrelevant to (at least a basic) understanding the word ‘red’, and in the illegitimate sense of ‘what red is like’—according to which it refers to the phenomenology of seeing red, or something of the sort—it is not, in the first instance, a question of knowing anything, but of having the experience. Even more clearly than in the ‘red’ case, you cannot understand certain uses of demonstrative expressions unless you have a relevant perception or can call up a relevant memory of the designated object.96 But note that this fact does not detract in any way from the expressibility in language of colours, pains, emotions, or anything else— including demonstrated concrete objects—to which the point applies:97 the reference to the world is routed via experience, but it still runs from language to the world. Even the fact that I may need to make use of bits of the world, suitably demonstrated, in order to say what I want to say does not undermine the expressibility of the world in language. For example, if there is no English word for a specific shade of blue, I may nevertheless talk about that shade with the help of a sample and a demonstrative term—‘I mean that shade’.98 Here the phrase ‘that shade’ is essentially world-involving, and it cannot be understood unless the addressee has perceptual access to the shade in question. The demonstrated shade is, as Wittgenstein said, part of the language.99 A fourth reason why there is a temptation to criticize language because of its supposed expressive inadequacy is that pain often cannot be talked about by the sufferer since the experience itself inhibits speech, as a purely physical or psychological effect. This applies, for example, to Philoctetes’ pain.100 In the current sense Scarry is right to say that pain ‘destroys language’.101 But it does not follow that pain cannot be described, simpliciter, by anyone ever; only that it may be difficult or even impossible for the sufferer to describe his or her own pain at the time of suffering.

296  Tragedy and Linguistic Redress This was a commonplace in Renaissance tragedy. Influenced by the Senecan ‘curae leves loquuntur, ingentes stupent’ (‘light cares speak, great ones are silent’)102—and recall the words of Aeschylus’ Atossa quoted in §29 (Persae 290–2),103—playwrights produced many variations on the theme. For example, Webster’s Isabella quotes the words of a common proverb: ‘Unkindness do thy office, poor heart break, Those are the killing griefs which dare not speak’;104 Even if we think that Marcus’ attempt to express Lavinia’s suffering fails, still it remains one of tragedy’s major functions to give voice to suffering, to put it into words, attempting like Titus to ‘wrest an alphabet’ from Lavinia’s pain,105 as if inventing language afresh.106 That the expression of feelings is, in general, subject to contingent limitations is a point well made by Adrian Moore (1997, 148): Most of us have said something like, ‘Words cannot express how I feel at the moment,’ and have known a certain anguish in saying it. But still, ultimately, we have been talking about contingent limitations. It is a familiar enough fact that contingent limitations can make it a practical, social, or indeed physical impossibility for one to express things in a certain way, even sometimes to express them at all. One’s vocabulary might be limited. The vocabulary of the language one speaks might be limited. One might lack the skills to use non-linguistic means of expression: gestures, looks, deeds, music. One might be paralysed. Even if a given individual faces such contingent limitations, the feelings can be expressed in words—by someone else, perhaps. This point emerges most strikingly from the case of non-human animals and human infants. Here the limitation of the conscious subject’s powers of expression may, given certain background conditions and suitable relativization, be recast as a necessary limitation. But, absolutely speaking, it is still only contingent; it is still only a limitation on the powers of that particular subject—or subjects of that species, or at that stage of d ­ evelopment— not a restriction on the powers of language as such. Language has the wherewithal to express the feelings of a sentient being, whether or not that being itself has that wherewithal.107 Moore’s suggestion, towards the end of the quoted passage, that there might be non-linguistic means of expressing feelings is one to which I shall return: we shall have to consider in what sense, if any, non-linguistic techniques can go proxy for language as a way of expressing feelings. A fifth and final reason why, I suggest, people are inveigled into making the mistake that I have identified is that, as my talk of ‘inner’ and

Can Suffering Be Expressed in Words? 297 ‘outer’ has hinted, they commit the fallacy that the later Wittgenstein eloquently exposed and refuted in the sections of the Philosophical Investigations that deal with the so-called private language argument.108 The essential error of Wittgenstein’s would-be private linguist, as I have argued elsewhere, in agreement with McDowell, is that he thinks of the sensation as a Kantian Ding an sich, and so as something that escapes the net of public language.109 Instead of concluding that it cannot be spoken of at all—not as a sensation, not even as a thing, for a thing is necessarily a thing of a certain sort, that has, or can be given, a label in the public language110—so that the whole hypothesis collapses, the would-be private linguist tries to rescue the situation by supposing that he has private linguistic access to this ‘thing’, and so can talk about it in a private language, that is, a language that is logically incomprehensible to everyone else.111 But the private linguist did not invent the relevant names—‘language’, ‘sensation’, ‘thing’, ‘access’, and so on—that he is relying on, these names come with price tags, and we speakers of the public language, who did devise the names, settle the prices. That is, we determine how these words are to be (correctly) applied. Further, there can be no object to which, logically, only one person has epistemic access. You cannot use logic to set up barriers between people; that is not what logic is for.112 Once a language, of any kind, is up and running, it is potentially there for everyone. There is no object to which, as a matter of logic, only I can have access; and there are no Dinge an sich. So it is a mistake to think of pain—or anything else—as belonging to a Kantian noumenal realm. The Cartesianism so sedulously unpicked by Wittgenstein is still rife, not only among literary critics and theorists, as we have seen (§33), but also among writers on pain. So, for example, Scarry tells us that ‘to have great pain is to have certainty; to hear that another person has pain is to have doubt’ (1985, 7; cf. 13). Against which we may set: ‘Just try—in a real case—to doubt someone else’s fear or pain’.113 Scarry remarks that she wishes to ask the question ‘How is it that one person can be in the presence of another person in pain and not know it—not know it to the point where he himself inflicts it, and goes on inflicting it?’ (1985, 12). This is indeed a question that can be asked about certain types of paininfliction, particularly, perhaps, the infliction of psychological pain: such pain can be inflicted unwittingly. But it is not a question that arises in standard cases of physical torture. Here the torturer knows—or at any rate believes, but normally he will know—that his victim is in pain: that is the whole point of the torture. In standard cases of infliction of pain, the inflictor knows that the victim is in pain with all the certainty that knowledge ever achieves. (That is, he knows it short of the realization of an outlandish sceptical possibility.) And he knows it because the victim behaves as one in pain; mind is revealed in behaviour, not hidden behind it in a private, Cartesian inner sanctum.

298  Tragedy and Linguistic Redress

42 Pain and the Noumenal The characterization of the private linguist’s would-be sensation as a Kantian thing-in-itself is suggestive in this context, because we do in fact find that pain, in particular, is often viewed in such Kantian terms.114 And that is perhaps unsurprising: since pain is able to destroy the sufferer’s ability to express and describe the pain itself, it is easy to see how one might be tempted into affirming that pain destroys language in a deeper sense, by destroying its own conceptuality, so that the pain becomes a mere thing-in-itself, thereby also destroying its expressibility in language. In his discussion of Lacan’s reading of the Antigone, Charles Freeland suggests that it is ‘at the limits of language’ that ‘a desire attains its identification with pain’ (2013, 153). In the continuation of the passages quoted above from Jenseits von Schuld und Sühne (§38) Améry describes how pain reduces the sufferer to pure bodily existence (das Körperhafte), to mere thinghood, almost as though he were experiencing his own death (2008, vol. 2, 74–5). Human and animal bodies, whether alive or dead, are still things of certain sorts, not Kantian things-in-themselves, but what Améry charts seems rather like a movement towards the Kantian noumenal state, if only that existed, as though it were an ideal which we could not quite, but in pain strove to, reach. The Kantian noumenal realm is supposed to be the most real thing there is; everything else is phenomenal, mere appearance. And it is natural to think that pain is somehow more real than everything else.115 Büchner’s Payne remarks that ‘one can deny the existence of evil, but not of pain’;116 ‘Quid non possit superare dolor?’, asks the chorus in the Senecan Hercules on Oeta (1279): ‘What cannot pain overcome?’. And Hercules later all but admits that he has been conquered by pain (1446–7).117 As Tolstoy’s Ivan Ilyich dies, he comes to a point where his pain seemed to be the only reality.118 Kafka wrote in his diary: ‘Looked at with a primitive eye, the real, incontestable truth, a truth marred by no external circumstance (martyrdom, sacrifice of oneself for the sake of another), is only physical pain’.119 Repetition seems to be a way of, as it were, trying to approach the noumenal.120 The greatest modern exponent of this technique is perhaps Beckett, whose prose so often has an incantatory, mesmeric effect (2006, 401): I say the floor here, now bare, this strip of floor, once was carpeted, a deep pile. Till one night, while still little more than a child, she called her mother and said, Mother, this is not enough. The mother: Not enough? May—the child’s given name—May: Not enough. The mother: What do you mean, May, not enough, what can you possibly mean, May, not enough? May: I mean, Mother, that I must hear the feet, however faint they fall. The mother: The motion alone is not enough? May: No, Mother, the motion alone is not enough, I must hear the feet, however faint they fall.

Can Suffering Be Expressed in Words? 299 These spell-binding words are taken from Footfalls, which begins with a dialogue between May (M) and the voice (V) of her mother (ibid., 399): M:  Mother.

[Pause. No louder.] Mother. [Pause.] V:  Yes, May. M:  Were you asleep? V:  Deep asleep. [Pause.] I heard you in my deep sleep. [Pause.] There is no sleep so deep I would not hear you there. These lines remind us of Nietzsche’s words ‘Ich schlief, ich schlief— / Aus tiefem Traum bin ich erwacht’ (1983, 312), so powerfully set to music in the fourth movement of Mahler’s third symphony. Beckett’s ‘There is no sleep so deep I would not hear you there’ is a line of gnomic intensity and beauty worthy of Euripides. Indeed there is a haunting echo, in the dislocated dialogue between May and her mother in Footfalls, of the similarly ghostly exchange between Iphis and his daughter Evadne in Euripides’ Suppliants—one of the most stunning scenes in extant Greek tragedy, comparable in its effect to the finale of Bellini’s Norma—Iphis below at ground level, pleading with his daughter, Evadne above on a cliff edge insanely got up in wedding weeds and poised to throw herself onto her husband Capaneus’ funeral pyre in an act of suttee.121 In the second kommos of this play we find mourning repetitions marking a grief that spans three generations—the slain warriors, their mothers, and their orphaned children.122 (The pathetic repetitions have a darker side that we have already explored, in general terms, for they usher in the linear or cyclic repetition of violence, since the boys when they grow up will in turn exact revenge.)123 In Euripides’ Phoenician Women, the chorus uses repetition as a way of recalling Thebes’ grief—the repetitive keening of the mothers and maids—over the loss of its young men to the Sphinx (1030–42).124 When Hecuba sees her dead son Polydorus, she stutters in disbelief: ἄπιστ᾽ ἄπιστα, καινὰ καινὰ δέρκομαι. ἕτερα δ᾽ ἀφ᾽ ἑτέρων κακὰ κακῶν κυρεῖ unbelievable, unbelievable is what I see, unheard of, unheard of, evils upon evils, evils upon evils strike me.125 Martha Nussbaum comments that, in these words, language nearly fails (1986, 408). Perhaps, it might be suggested, that is because it is trying to reach beyond itself? Recall Lear: And my poor fool is hanged. No, no, no life! Why should a dog, a horse, a rat, have life And thou no breath at all? O thou’lt come no more, Never, never, never, never, never.126

300  Tragedy and Linguistic Redress The extraordinary final line, an iambic pentameter with pure heterodynes (clashes of ictus and accent), and consisting of a single word repeated five times, is unique in the English canon. In Shakespeare’s line it is almost as though Lear, in repeating the word ‘never’, were trying to attain to sheer neverness, neverness-in-itself. But there is no such thing—there is no such thing as anything in itself, in the Kantian sense—and if that is his aim then Lear is making an intellectual mistake: ‘thou’lt come no more’ says it all, and the rest is just excess baggage, superfluous if superlative rhetoric. It is rhetoric that makes us seem to feel a gap between language and world of a kind that is not really there: for the more Lear repeats the word ‘never’, as if he were grasping desperately at the noumenal—the (supposedly) ultimately real—the further he seems to travel from reality, even as Cordelia’s life recedes from him.127 Perhaps if repetition of a word cannot reach the noumenal, repetition of a mere noise can? That is the thought to which Wittgenstein’s private linguist retreats when he is disallowed from using words like ‘pain’, ‘sensation’, and even ‘something’, because these words, like all words, have already been earmarked by the public language and are accordingly associated with public criteria of correct application—‘So in the end when one is doing philosophy one gets to the point where one would like just to emit an inarticulate sound’.128 But merely uttering a noise is of no help to the private linguist, either, for ‘such a sound is an expression only as it occurs in a particular language-game, which should now be described’.129 If Shakespeare’s repeated ‘never’ is a bold stroke, it is matched in audacity by the way in which Sophocles tries to give dramatic expression to Philoctetes’ pain, by the repetition of mere noises, as for example ‘attatai’, and ‘apappapai papa papa papa papai’.130 Philoctetes’ monumental pain has been an inspiration to writers on language and suffering since Cicero.131 Lessing exploited Winkelmann’s remarks on it at the beginning of Laokoon, and he has an excellent discussion of Philoctetes’ (and Heracles’) cries of anguish in the body of that work; Herder opened his Abhandlung über den Ursprung der Sprache with Philoctetes’ agony, as Scarry does her book The Body in Pain.132 And commentators have noted how the expression of Philoctetes’ suffering, which initially seems a sign of weakness, gains in power in the progress of the drama, as authority drains away from Odysseus.133 Philoctetes tells Neoptolemus that his pain is ‘unspeakable’ (oude rhēton: 756). And, correlatively, we are often told by commentators that his cries are untranslatable.134 Nussbaum remarks that, though the long cry just quoted (746) is metrical (it is a regular iambic trimeter), ‘it has lost syntax and morphology, the hallmarks of human language’ (2008, 160). It is true, in general (and I have relied on the point), that syntax and case (which languages sometimes signal morphologically) are essential to human language, but perhaps it is not so straightforward a matter to label Philoctetes’ cries as non-linguistic. For the word ‘papai’, which

Can Suffering Be Expressed in Words? 301 Sophocles uses in repeated and varied form to convey Philoctetes’ cries of anguish, is plausibly a regular word of classical Greek conveying exclamation or surprise.135 Seth Schein notes (on 742–50) that the long cry at 746 suggests the Greek word ‘pais’, child (as Philoctetes calls Neoptolemus at 776), and ‘pappas’, papa, and of course these words have regular syntax and morphology. It is also noteworthy that Neoptolemus uses the word ‘papai’ later in the play to express his moral dilemma (895), where Jebb translates it as ‘Alack!’ and Lloyd-Jones as ‘Ah!’.136 Shortly after that Neoptolemus expresses the pain of his dilemma in quite ordinary words, when he says (in a line partially quoted earlier: §31) that, if he forsakes his true nature, ‘I shall appear base; this is what has pained me (tout’ aniōmai) for a long time’ (906). Nothing inexpressible about pain here: Neoptolemus puts it admirably into words.137 To drive the point home, he repeats the phrase: ‘this is what has pained me for a long time’ (913).138 Here repetition is not attempting to get beyond language to something supposedly inexpressible in words, but is precisely reinforcing the point that language is adequate to express the pain of (in this case) dishonour. The verb ‘aniasthai’ is used not only of Neoptolemus’ mental pain, but also of Philoctetes’ physical pain (283), and in general we find a blurring in the Philoctetes between the categories of physical and mental pain: but that favours the thesis of the expressibility of pain in language.139

43 Language and Silence I have in effect been starting to suggest that Philoctetes’ screams are part of the language: if it could be shown that they were, that would support the expressibility thesis. What the question whether his screams are part of the language means is whether the noises he produces consist of words that not only have syntax and case, as Nussbaum says, but also have sense and reference.140 Reflecting on Philoctetes’ expressions of pain already quoted, one will, I think, be inclined to answer the question affirmatively, since, as we have seen, Sophocles constructs those cries from ‘papai’, which, as I have suggested, is a regular word. But here it might be objected that, even if ‘papai’ counts as a word, there will be plenty of other sounds which would not yield to the same treatment. Sophocles gives Philoctetes a variety of cries, such as ‘ā ā ā ā’,141 sounds that are clearly meant to serve as groans of pain; his Ajax repeats ‘iō’ and ‘aiai’, and his Heracles ‘aiai’ and ‘eé’, which are again onomatopoeic cries of anguish.142 Aeschylus gives similar noises to Cassandra, ‘like that of a sea bird, wild and without meaning’, as Steiner nicely puts it, in part following Fraenkel, who noted that Cassandra ‘breaks, not indeed into speech or song, but into something between a song and those wild notes of lamentation which were familiar to the Athenians from the ritual performances of the barbarian mourning-women from the East’.143

302  Tragedy and Linguistic Redress Consider Aristophanes’ representations of frogs’ croakings and bird calls. What about Don Giovanni’s ‘ohimé’—sung over an augmented fourth, the diabolus in musica—when he grasps the Commendatore’s proffered hand?144 Or, to go further in the direction of mere noise, recall Kattrin’s raw, wordless cries when the recruiting officer of Brecht’s Mutter Courage takes her brother Eilif away,145 and Mimi’s consumptive cough hacking through the sublime music of La Bohème. In other words, even if we grant that Philoctetes’ ‘papai’ is a regular word, it will not be difficult to find dramatic examples of mere noises, such as groans, yelps, gasps, chokings, coughs, and so on. And surely, one might think, these will not be part of the language, since they stand in no syntactic relation, as words must do, to other words in sentences. One way of registering this thought would be to say that mere noises do not fall under the Fregean context principle.146 But is this right? Certainly, the mere fact that the noises in question often stand alone—in the sense that they are not constructed with (other) words in a sentence—is no argument against their linguistic status. For exactly the same applies to many uses of bona fide words, such as single words uttered in command. Sentences can consist of a single word, at least at the level of surface grammar. It is true that, at the level of logical form, analysis of a one-word sentence will discern predicative structure and so complexity—syntactic or semantic complexity, depending on how we regiment the notion of logical form.147 But why should not the same be true of the ‘mere’ noises we are talking about? Thus, for example, ancient Greek shepherds’ cries to their sheep, like ‘psutta’ and ‘sitta’, are in effect commands, which have complexity at the level of logical form.148 So it is at least not obvious that the context principle does not apply to the sounds in question. Here it is relevant to note that Greek cries of anguish are usually stylized,149 and are often transformed into what anyone must recognize to be regular words: for example, the sound ‘aiai’ yields an ordinary verb, ‘aiazō’;150 the cry ‘iē’ was converted into the noun ‘iēlemos’, lament.151 One derivation of the word ‘elegos’ (‘elegy’), was from ‘e e legein’ (‘to say e e’).152 It is also important that, as A. J. Bowen notes, ‘Greek playwrights prescribe a particular noise and do not leave it to the actors’.153 So perhaps, as in the case of ‘papai’, these noises have a proto-verbal form. Further, it may be that sheer cries do conform to the context principle, even if not in a narrow, at least in a broad sense: one might say that noises have dramatic significance if they are inserted into a linguistic context which makes sense of them, which gives them semantic significance. That will not apply to linguistic representations of all noises—not, for example, to Aristophanes’ representations of the frogs’ croakings,154 though it may well do so to some of his representations of bird calls.155 In those cases where semantic significance can be descried, even if the relevant noises bear no strict syntactic relation to their contexts, they nevertheless have a semantic relation to that context, in virtue of which

Can Suffering Be Expressed in Words? 303 they mean something; they acquire sense and reference. They are, as I have just hinted, if not full words, at least proto-words. Dionysus’ ‘a!’ at Euripides, Bacchae 810 is the turning point in the drama.156 The special status of this ‘a!’ is reinforced by its location extra versum. But that the sound has semantic meaning seems clear, even if it is difficult for us now to identify what that meaning is. In his note on the line, Dodds suggests that Pentheus has started to leave the stage and that Dionysus’ ‘a!’ recalls him, so that it means ‘Stop!’. Elsewhere (on 644) Dodds says that ‘ea’ is ‘a noise, not a word’, but it seems to me that the matter is not so straightforward: sounds (especially indeed ‘ea!’) are commonly used by Euripides to express a crisis, or in general something unexpected, which indicates that they have some kind of semantic meaning.157 Commenting on Philoctetes’ cry of pain (line 746 of the play) that I quoted above (§42), Bernard Knox writes (1964, 130–1): This long rhythmic cry does not mean anything, and cannot be translated . . . Jebb does not translate it at all, and he is right. For English, unlike Greek, has no formulaic sounds for grief and agony. There is a pitch of physical suffering which words are inadequate to express, and when human beings reach it they make sounds, like animals, which convey nothing but the extremity of their pain. But if Greek has formulaic sounds for grief and agony, then, whether or not they can be translated into English, does that not suggest that at least some languages are capable of expressing the ‘pitch of physical suffering’ that Knox here says ‘words are inadequate to express’? After all, the noise you make when you pronounce ‘table’ is equally a formulaic sound. What is formulaic about it is that it has a conventional meaning; that suggests that a similar kind of meaning attaches to the formulaic sounds that Knox is talking about. And once that possibility has been mooted, I think that one will want to go back and examine Knox’s assertion that English cannot express what apparently can be expressed in Greek. That contention looks suspicious, since it conflicts with the natural Davidsonian thesis of the intertranslatability of languages. In general, I have argued that the assertion that ‘There is a pitch of physical suffering which words are inadequate to express’ rests on a confusion, namely the confusion of supposing that it is the business of words to mimic, or induce, the pain they seek to express. But Knox is quite right that there are situations of extreme pain in which people do not, and perhaps could not even if they wanted to, describe their pain, but simply make noises like animals. Indeed the point applies more generally. Human beings make all sorts of noises in all sorts of situations, and not just when they are in extremis. In his essay ‘Grace Before Meat’, Lamb wrote of Dr Johnson that ‘The author of the Rambler used to make inarticulate animal noises over a favourite food’ (1903, vol. 2, 95). Even if such vocalizations are not in

304  Tragedy and Linguistic Redress the strict sense part of the language, they may, and often do, have semantic meaning. When Mozart’s Papageno has a curb put on his mouth by the serving-maids of the Queen of the Night, and can only sing ‘Hm hm hm . . .’, Tamino tells us that the poor fellow can talk about punishment, because he has been deprived of speech (‘Der Arme kann von Strafe sagen, denn seine Sprache ist dahin’).158 So one might ask the question: when do sheer vocalizations have semantic meaning, and when not? My answer, broached in the last paragraph, is that such noises will have semantic meaning just if they are placed in a linguistic context which confers that meaning on them. There will then be a sense in which they conform to the context principle. Provided a suitable verbal context is in place, nonor proto-verbal elements can become signs with semantic meaning. They will then have sense and reference. In effect we confer on them a kind of honorary and derivative linguistic status, a status which may, with use, be converted into full membership of the club of words. And, whether their membership is honorary or full, such noises will, as Davidsonian principles require, be translatable from one language to another.159 In fact, the point applies not just to noises but to non-verbal gestures of many sorts, such as Emma’s offering Knightley her hand at a crucial point in Jane Austen’s novel, or Phèdre’s and Bérénice’s and Athalie’s famously stage-directed acts of sitting in the notoriously stage-­directionless Racine.160 These gestures acquire meaning from their contexts: Emma loves Knightley;161 Phèdre, Bérénice, and Athalie are momentarily paralyzed by conflicting emotions.162 The point also applies to silences, such as Davies’s silence at the end of Pinter’s The Caretaker,163 Gloucester’s silence in the battle scene of Lear between Edgar’s exit and re-entrance,164 or Dido’s proud refusal to answer Aeneas when he addresses her in Virgil’s underworld, a scene imitated from the Odyssey. Many have shared Longinus’ admiration for Ajax’ haughty silence when Odysseus accosts him in Homer’s underworld, begging him to put his enmity aside at last: So I spoke. But he gave me no answer, but went off after the other souls of the perished dead men, into the darkness.165 Hume endorsed Longinus’ praise of this passage, remarking that Ajax’ silence ‘expresses more noble disdain and resolute indignation than any language can convey’.166 As I noted above, the capacity of a mere noise to convey meaning depends upon its linguistic context, and the same applies to silence. Given an appropriate verbal context, it is clear both what Dido’s and Ajax’ silences mean—and they do mean something, in the semantic sense—and that they are, as Hume (following Longinus) claims for Ajax’ case, more powerful than words would have been. The sublimity of Ajax’ silence depends not only on Homer’s but also on Odysseus’ verbal narration, so that it is doubly verbally contextualized.167

Can Suffering Be Expressed in Words? 305 The power of silence to convey meaning is a familiar topos.168 ‘Those who have the power of reproaching in silence may find it a means more effective than words’, as Hardy put it in Far from the Madding Crowd (1972, 225). Thomas Cromwell, in Bolt’s A Man for All Seasons, argues in court that, although silence just as such—the silence of a dead man, say—is meaningless, Thomas More’s silence (his refusal to take the oath of supremacy, and refusal to say why he will not take it) speaks volumes, and is so understood from one end of the country to the other (1960, 91–2). The silence that is invoked at the end of the Tractatus is supposed to have mystical meaning, but that special meaning is provided by a frame of sentences that have regular meaning, according to that work, that is, by sentences of natural science (not, notoriously, by the sentences of the Tractatus itself), which are about arrangements of objects in the world. And the final clause of the Tractatus is not simply an injunction to silence, as is often supposed:169 it is also a tautology, telling us that we logically cannot speak about what we cannot speak about. Try as we might, the only meaningful sentences that we can—according to Wittgenstein’s Tractarian position—come out with are sentences of natural science: the rest is just nonsense. (Wherein it is seen how the Tractatus undoes itself, for this last sentence does not belong to natural science.) Silent departures from the stage can be exceptionally powerful: here one thinks of Tecmessa’s wordless exit in Sophocles’ Ajax, Eurydice’s in the Antigone, Deianeira’s in the Women of Trachis, Jocasta’s in Euripides’ Phoenician Women.170 Seneca’s Trojan Women revolves around the sacrifices of Astyanax and Polyxena. The former death is accompanied by much talk from Ulysses, and Astyanax is even allowed to plead briefly for his life. By contrast, when the time comes for Polyxena to be led away to death, Pyrrhus, flanked by soldiers, enters silently and removes her; she, like him, says nothing.171 As pregnant as silent exit can be the silent presence, while others talk, of a significant figure, such as Atossa’s silence during the lyric exchange between Messenger and chorus in Aeschylus’ Persae (249–89), which we have mentioned already (§29);172 the silence that follows Electra’s ‘pheu’ (‘alas’), at Aeschylus, Choephori 194, which Fraenkel describes as ‘perhaps the most beautiful of the many aposiōpēseis in Aeschylus’ (1978, vol. 3, 819); Ajax’ silence at Sophocles, Ajax 311, after he has come to his senses, and Tecmessa’s silent presence during his ‘deception speech’ and before her mute departure,173 as well as her silent presence in the last 250 lines of the play;174 Orestes’ silence in the closing lines of Sophocles’ Electra;175 Jocasta’s silence at Oedipus Rex 988–1059, as she comes to realize the truth;176 Neoptolemus’ silence during the row between Philoctetes and Odysseus in the Philoctetes (974–1003);177 the silent presence of the chorus in general;178 the silence of Virgil’s Lavinia, the causa mali tanti who says nothing.179 (By contrast, in his Turne Prévost has her lead a lament for the dead Aimée: 1555–662.)180 Not that all stage silences are significant. Richard Tarrant doubts whether ‘the silent

306  Tragedy and Linguistic Redress presence of a character has clear dramatic significance’ (1976, 318) anywhere in Seneca. Be that as it may, the point here is that, where silence has semantic significance, it acquires that significance from its linguistic frame. The silence of Sophocles’ Tecmessa during the deception speech means that she does not understand it; Aeschylus’ Electra and Virgil’s Lavinia, one feels, understand the situations they are in only too well. We have seen Hume remark, concerning Ajax’ silence in the Odyssey, that it ‘expresses more noble disdain and resolute indignation than any language can convey’. It will be clear from the discussion so far that I think this assertion needs to be handled carefully. Consider another example of tragic silence. Brecht’s Mother Courage lives off the war, as we have said (§25), but it punishes her by taking her children, one by one. Perhaps the most agonizing moment in the drama occurs when the dead body of her son Schweizerkas, who has been shot by the enemy for attempting to return the regimental pay-chest to his commander, is brought on stage.181 The enemy soldiers suspect that Courage was implicated in the theft and knows where the money is hidden; they show her the dead body to see if she gives herself away. If she and her daughter Kattrin are to survive, it is vital that she betray no sign of recognition. The soldiers prompt her twice to acknowledge that she knows who the dead boy is, but she stands in stony silence without flinching; they give up, and carry the body off to be cast into a pit. Steiner saw Helene Weigel act the scene (1961, 353–4): As the body of her son was laid before her, she merely shook her head in mute denial. The soldiers compelled her to look again. Again she gave no sign of recognition, only a dead stare. As the body was carried off, Weigel looked the other way and tore her mouth wide open. The shape of the gesture was that of the screaming horse in Picasso’s Guernica. The sound that came out was raw and terrible beyond any description that I could give of it. But, in fact, there was no sound. Nothing. The sound was total silence. It was silence which screamed and screamed through the whole theatre so that the audience lowered its head as before a gust of wind. Steiner writes that he cannot describe the terrible sound of Weigel’s gesture, but he does so very successfully, and the reader has no difficulty understanding and indeed imagining how the scene went in the performance Steiner saw. There was no sound, but the absence of sound, in that context, was more effective and harrowing than sound would have been, and Steiner indeed describes its effect. Weigel’s gesture depended on the phenomenon of real screaming for its impact, and that dependence was crucial to its meaning, just as Picasso’s horse imitates real screaming, and that imitation is crucial to its meaning. Picasso’s horse recalls (from the same war) the big grey horse that proved

Can Suffering Be Expressed in Words? 307 Robert Jordan’s undoing in For Whom the Bell Tolls. Hemingway does not describe ‘the sound the horse was making’ after it had been wounded by shrapnel (2004, 479). In his great epic poem, ‘Sohrab and Rustum’, Matthew Arnold says more. At the moment of darkness, before Rustum delivers the death blow to his son Sohrab, the horse Ruksh screams (499–508): And Rustum bow’d his head; but then the gloom Grew blacker: thunder rumbled in the air, And lightnings rent the cloud; and Ruksh, the horse, Who stood at hand, utter’d a dreadful cry: No horse’s cry was that, most like the roar Of some pain’d desert lion, who all day Has trail’d the hunter’s javelin in his side, And comes at night to die upon the sand:— The two hosts heard that cry, and quak’d for fear, And Oxus curdled as it cross’d his stream. The poet compares Ruksh’s cry to a lion’s roar; alternatively, one might think that horses’ screams are so unnerving because they are almost human; a horse’s scream is, as C. S. Lewis said, ‘one of the most terrible noises in the world’ (1998a, 154). In fact Arnold cleverly exploits both comparisons, because in the background of his simile is Virgil’s tragic Turnus, likened to a wounded lion in the grand opening of the final act of his tragedy (Aeneid XII, 4–9). And these are all things that can be put into words. What Steiner meant when he wrote that the ‘sound’ Weigel made was ‘raw and terrible beyond any description that I could give of it’—or, at any rate, what it would be correct to say—is not that he cannot describe the actress’s gesture and its effect on the audience, but that any description he might give could not possibly substitute for that gesture and its effect. As we say, ‘You needed to be there’. The fact that a description cannot substitute for a gesture, such as Emma’s giving her hand to Knightley, or for non-verbal noise, such as Mimi’s coughing, or for silence, such as Mother Courage’s silence in Weigel’s interpretation, points us in a new and important direction. My diagnosis of what I take to be the error of complaining about language’s inadequacy was that, when people talk of the inadequacy of language, they are simply drawing attention to the fact that language is not the world, and does not go proxy for the world, either. Words are one thing, the reality that they are about quite another. This claim applies in the present context, too. So one purpose that we may have when we say things like ‘You needed to be there’ is to indicate that language and reality are different things. It is not that words cannot describe reality; rather, they cannot be the reality they describe (at least in the usual case, so ignoring for these purposes such phenomena as self-reference). But we may have

308  Tragedy and Linguistic Redress a further thought in mind, namely that language, though it is indeed a route to reality, is a different route from gesture, noise, or silence, and that while language may be able to do what these other things do, it cannot do it in the way that they do it. Hume’s point about Ajax’ silence is that words could not achieve what silence achieves. Ajax’ silence is an expression of contempt: he could have expressed that contempt in words, and we can describe that same contempt, that emotional reality, in words. It is also possible for us to describe in words the way in which either language or silence conveys that contempt—that is, we can describe Ajax’ hypothetical words’ and his actual silence’s differing modes of presentation of that reality. But what language—either our language or Ajax’—cannot do is express that contempt in the same way as his silence expresses it. To do that it would have to be that silence, which is logically impossible. As the reader will have gathered, I am locating this point within a familiar dialectic: the sense–reference distinction. The aspect of that distinction which is important here is this. Two senses may home in on the same object of reference, but, when they do so, they do not do it in the same way. Logically, they cannot do it in the same way, for then they would be one and the same sense. The usual deployment of this point lies within the pale of language: we say, for example, that two coreferential words that differ in sense necessarily present their common referent in different ways. That yields a restricted kind of ineffability: what one sense does—present its referent in just that way, the way of the sense in question—cannot be done by another sense.182 But in the case under consideration we are applying the point between language on the one hand, and something non-linguistic, such as a gesture, sheer noise, or silence, on the other. Exactly the same phenomenon occurs here, however: a piece of language cannot duplicate the very way in which silence, for example, expresses (say) contempt, though language may express that same contempt in another, necessarily linguistic, way. If, as I urged above, it is right to think of these various non-linguistic methods of expressing things as having semantic meaning, then we may infer that, although language may very well be able to express the same what (the same referent) as some non-linguistic sign, it cannot express the how (the sense)—it cannot duplicate the exact way in which that non-linguistic sign expresses the common reality. To that extent Hume is vindicated. But once we have generalized the ineffability point in this way, we see that it represents no real derogation from language’s competence and expressive power. That is because, as we said to begin with, exactly the same stricture—one sense cannot, logically, be duplicated by another sense—applies within language. Take the familiar example involving ‘Hesperus’ and ‘Phosphorus’. ‘Hesperus’ and ‘Phosphorus’ are both names of the planet Venus, and in the standard story they are assumed to differ in sense, ‘Hesperus’ presenting Venus as an evening appearance, and ‘Phosphorus’ presenting it as a morning appearance. Now let

Can Suffering Be Expressed in Words? 309 us suppose, for the sake of the argument, that no other word shares the sense of ‘Hesperus’. We have assumed that ‘Phosphorus’ does not share it, and we are now supposing that no other word does so, either. That then means that there is a way in which sentences that say things about Venus using the name ‘Hesperus’ have a content that can be conveyed in no other way, because no other word shares the precise sense of ‘Hesperus’. The sense (in Frege’s terminology, the thought) conveyed by ‘Hesperus is bright’ is one that could be conveyed by no other sentence. But, although you might regard that as a restriction on the expressive power of language, it does not seem to be a terribly serious one, because there are other ways of expressing, if not the sense-content of ‘Hesperus is bright’, at least its referential content (‘Venus is bright’, ‘Phosphorus is bright’, etc.). And, as I have argued elsewhere, referential content is the key concept that explains, or models, communicative success.183 Similarly in the cases we started with: gestures, noises, silences. If these have a referential content that can be captured linguistically, and they sometimes do, then the fact that the ways in which they present that referential content (that is, their senses) cannot be duplicated linguistically seems an insignificant derogation from language’s expressive power. And it is an important point, to which I shall return, that these senses can themselves be referred to in further pieces of language. But before then we need to consider language’s shortcomings as a communicative tool.

44 The (In)appropriateness of Language The silence that Helene Weigel put in place of a cry was able to carry a semantic (Fregean) sense that no other non-linguistic gesture, and no piece of language either, could have borne. It was also, so we gather, more effective than any sound, linguistic or non-linguistic, that she might have made instead. That points to an important restriction on the usefulness of language, at least in some contexts. For on occasions when we are giving expression to pain and suffering, language can be inappropriate; or it can be appropriate if employed in some ways but not in other ways. For example, language may be inappropriate because, given the situation in which it might be produced, it could only be misunderstood: so, perhaps, Virgil’s Aeneas insists on brevity in his reply to Dido after her accusation of betrayal because he fears that, if he said more, she would only misconstrue it. He feels ‘that words are powerless here, that nothing can be achieved with words’.184 Or language may be too clever, as Berowne in Love’s Labour’s Lost, and Beatrice and Benedick in Much Ado About Nothing, illustrate. These latter two figures ‘have to be delivered by a stratagem of their friends, from wit to love’:185 they have talked themselves into a corner where they cannot express their mutual love because they cannot think of a sufficiently urbane way of doing so, as the rules of the game they have set up demand. Language can be dangerous: the

310  Tragedy and Linguistic Redress ‘something unspoken’ of Tennessee Williams’s drama could be named, but it is better not, because it would be destructive of the central relationship to call the thing by its right name, to capture that emotional reality in words. Something similar applies to the stutterings and silences of Racine’s Bérénice, which Barthes called a ‘tragédie de l’aphasie’.186 (And recall Ibsen’s Ellida from §37.) Language can be used without responsibility, as the ungainly prince in Thomas Mann’s Königliche Hoheit, Klaus Heinrich, finds when he woos the disconcertingly eloquent Imma Spoelmann (2004, 260–1), or as the otherwise rather timorous hero of Der Zauberberg, Hans Castorp, demonstrates when he speaks French to Madame Chauchat on the evening before her departure, in an attempt— with what success is not made entirely clear—to seduce her (‘parler français, c’est parler sans parler, en quelque manière,—sans responsibilité, ou comme nous parlons en rêve’: 2012a, 511). Language can be glib, or tactless, or cheap. We have already noted (§36) Marcus’ arguably inapposite rhetoric in Titus Andronicus, and F. T. Prince suggested something similar of Shakespeare’s Lucrece.187 Shelley’s treatment of Prometheus’ towering agony in the Prometheus Unbound, and Byron’s of torture in The Two Foscari, may also be cases in point. Perhaps, too, although he is well aware of the danger, a related stricture applies to Mann. There is a splendid passage in Der Zauberberg where the humanist Ludovico Settembrini deplores the widespread tendency to think of illness as a kind of supplement to health (2012a, 679–80), as though health were a highest common factor underlying both well-being and sickness: in the case of the well person the extra element, over and above that common factor, would presumably be null (or perhaps more health), whereas for the ill person it would be sickness. This is quite the wrong way to think about illness, as Settembrini so skilfully shows, though one cannot help occasionally entertaining the unfriendly thought that the author himself falls foul of his brilliant character’s point, in portraying the inmates of his clinic as though they were really, underneath, quite fit, and merely wore their illnesses as they wore clothes: until they actually die in their rooms, Mann’s sick patients seem rather wohlauf. That deficit of realism would not be surprising, for it is quite generally the case that artists portray ill-health unrealistically: Sophocles gives the tormented Heracles a long and highly mannered speech about his past labours; the Prometheus of the Prometheus Vinctus, though in acute pain, manages to tell Io her life’s story; a Roman tragedian’s Patroclus, instead of tending to the wounded Eurypylus, questions him about the battle;188 in Chaucer’s Miller’s Tale, ‘after Nicholas receives his third-degree burn and calls out for water, he and Alison seem to have taken up their coital posts in bed once again’, and in the Knight’s Tale Arcite and Palamon ‘fight so long and fiercely that they are ankle-deep in blood, and immediately thereafter they clearly have little or nothing wrong with them’.189 Mann’s Hans Castorp decides, despite having a strong headache, to pay a visit

Can Suffering Be Expressed in Words? 311 to the flamboyant Dutchman Mynheer Peeperkorn, and even accepts a glass of wine from him (2012a, 870–2); in fact his behaviour, as Mann describes it, seems hardly to be affected by his supposed pain. (It is only fair to add that the portrayal of Adrian Leverkühn’s migraines in Doktor Faustus is much more convincing.) Steiner sometimes worries that it is possible for people to be sensitive to suffering acted on stage but insensitive to real suffering.190 With this I implicitly agreed in my discussion earlier of Titus Andronicus (§36). ‘Language, in Kafka, overlays the action like an anaesthetic, allaying suspicion and inhibiting pure emotional response’, says Langer (1995, 116). This is not a view of Kafka that I share, but Langer is right that the risk is real. In the essay from which this quotation is taken, ‘Kafka as Holocaust Prophet’, Langer seeks to rebut Steiner’s claim that Kafka was a Vorbote of Nazism. Langer is correct to note that Kafka did not anticipate the policies of the Nazis in detail, but that seems to me rather a superficial point to make against Steiner, who at a deeper and more abstract level is surely right. And indeed, oddly enough, it is even the case that some of the detail in Kafka is astonishingly prophetic. When, for example, we read in a collection of Holocaust documentation that, on 5 June 1942, Dr August Becker, inspector of gas vans, wrote to his SS superior that ‘Since December 1941, 3 lorries, for example, have dealt with (‘verarbeitet’) 97,000, without developing any faults’191—note the suppression of a noun to go with the numeral—we are reminded of the officer of Kafka’s In der Strafkolonie, obsessed with the precise operation of his gruesome killing machine, to which finally he himself succumbs. Towards the end of Der Prozess, after the priest has told the protagonist K. a story, he remarks that the doorkeeper of his tale represents the Law, adding that to question his authority would therefore be to question the authority of the Law. K. is not persuaded (1987, 188): ‘I don’t agree with that view,’ said K., shaking his head, ‘for if one adheres to it, then one must take everything that the doorkeeper says to be true. But you yourself have shown in detail that that is not possible.’ ‘No,’ said the priest, ‘one does not have to take everything to be true; one has only to take it to be necessary.’ ‘That is a bleak view [Trübselige Meinung],’ said K. ‘It makes lying part of the world order’. For a philosopher, these words are instinct with a dark fascination: how can discourse be necessary—that is, its content express a necessity—if it is not true? How, if one is not obliged to take everything the doorkeeper says to be true, can one only be obliged to take it to be necessary, as though necessity amounted to less than truth? Steiner writes of ‘the mendacities of language’, which he contrasts with the honesty of music (‘Is there a lie, anywhere, in Mozart?’).192 But language is not inherently

312  Tragedy and Linguistic Redress mendacious: indeed the reverse is the case, as I remarked in Chapter 4 in connection with Davidson. Language is—language users are—inherently truth-telling, and the use of language to tell the truth is metaphysically speaking more fundamental than its use to purvey falsehoods, partly because you could have the former without the latter, though not vice versa, but more importantly because it is the truth-telling use of language that fixes content. And although music does not tell lies it does not tell truths either, because it has no semantic content (a point to which I shall return in the next chapter). Nevertheless, despite the difficulties of the Kafka passage, the strange combination of necessity and lying that it propounds does in a way anticipate the Holocaust: the lies that were told to the victims of extermination, along with the iron constraint under which they were held from the moment of arrest to the moment of death. Langer’s view is that Kafka is too gentlemanly and refined to be a prophet of the Holocaust: the victims of the Holocaust would have been all too happy to substitute Kafka’s dreamlike world for their own much grimmer reality (1995, 116). That observation is right, but it still misses the point. Kafka is not writing documentary before the fact (though, as I have suggested, there are some uncanny correspondences), and his characters do not inhabit a literal hell on earth, but he does miraculously capture a sense of the irrationality and unpredictability of the methods of power, of the hopelessness and helplessness of being ausgeliefert, and perhaps too he captures something of the unconscionable horror of what was to come.193 In a somewhat similar way, though in one sense Racine seems remote from the terrible events of the mid-twentieth century, no one can now engage with his Esther without thinking of some of those events, a poignancy heightened by the fact that the play was written for performance by young girls. (One thinks, perhaps, of the girls in Müller’s report quoted above, §38.) The piece belongs to the ‘catastrophe survived’ genre of tragedy; it ends with a paean to God for delivering the Jews from a threatened massacre. Steiner comments that ‘One has difficulty in seeing why Racine should entitle Esther “a tragedy” ’ (1961, 98). But, if we bear in mind what has been said (§23) about the closeness of the counterfactual to the actual, and about tragedy’s relative indifference to temporal direction, one will not find it so strange. The paean was, after all, destined to turn to dust and ashes: the play evinces a subterranean sense of this, by insisting on the contingency of the happy ending. As Racine’s other Jewish drama, Athalie, draws to its close, the foiled heroine cries out in despair ‘Dieu des Juifs, tu l’emportes!’ (1768). But her defeat records only one passing event; as Schiller’s Gordon vainly hints to Wallenstein,194 one should not praise the day before evening comes. ‘Far from recognizing divine omnipotence, [Athalie’s] final predictions indicate only that a battle has been lost, but not the war (1784–90)’.195

Can Suffering Be Expressed in Words? 313 Primo Levi tells how, after he has survived a ‘selection’, he overhears Kuhn, another inmate who has escaped death, thanking God: Kuhn is out of his senses. Does he not see Beppo the Greek in the bunk next to him, Beppo who is twenty years old and is going to the gas chamber the day after tomorrow and knows it and lies there looking fixedly at the light without saying anything and without even thinking any more? Can Kuhn fail to realize that next time it will be his turn? Does Kuhn not understand that what happened today is an abomination, which no propitiatory prayer, no pardon, no expiation by the guilty, which nothing at all that is in the power of man can ever cleanse again?196 Just as it is not possible, now, to see or read Lessing’s Nathan der Weise without reflecting on the terrible betrayal of its vision in the land of its genesis,197 so I cannot confront Racine’s Esther without thinking of Levi’s censure of Kuhn’s disgraceful prayer.198 I do not say that my experience is part of the meaning of the play.199 But I do say (recall again §23) that other endings than the survival story Racine gives us—plots in which catastrophe is incurred—are part of its meaning, and part of the meaning of any tragic drama in the ‘catastrophe survived’ mode.200 The Alcestis ends well in Euripides’ telling, but Chaucer’s Troilus envisages the obvious alternative.201 I have been speaking so far in this section and in the last about some of the ways in which language could be considered inadequate, or may on occasion be judged inappropriate. But it is time to redress the balance. One way in which language might be thought to be inadequate, but where, instead, we can discern its competence, lies in its ability to express lack or absence. We saw in the cases of Oedipus, Antigone, and others (Part I) that language is able to express cognitive failure; but only a cognitively successful artist can do that. It can also express inarticulacy; but only an articulate artist can do that. Perhaps Woyzeck cannot express his agony, but, if not, Büchner can. Steiner again (1961, 280): Woyzeck’s anguish crowds to the surface of speech, and there it is somehow arrested; only nervous, strident flashes break through. So in black dreams the shout is turned back in our throats. The words that would save us remain just beyond our grasp. That is Woyzeck’s tragedy, and it was an audacious thought to make a spoken drama of it. It is as if a man had composed a great opera on the theme of deafness. This, like much of what Steiner writes, is highly suggestive. An opera on the theme of deafness would naturally—unless it happened to be a composition in the manner of John Cage—be an experience of musical

314  Tragedy and Linguistic Redress sound: it would map silence to musical sound, just as opera in general is an art form that maps all aspects of life, including and especially the nonmusical aspects, to music. (We should not think of deafness as silence in the head, just as we should not think of blindness as darkness in the head: silence and darkness are absences that get their significance from a contrast with presence.)202 So in a way it seems, against Steiner, that there is no audacity: such a mapping, from what is not music to music, is what sung drama engages in all the time; an opera on the theme of deafness would be no more audacious than an opera on the theme of the electrification of the Soviet Union.203 When Dvořák turned Andersen’s story of the little mermaid into an opera, he and his librettist faced a similar problem, since the mermaid loses her voice in return for acquiring human form: the problem was solved by setting up the terms of the pact with the witch Ježibaba in such a way that Rusalka retains the power of speech towards the non-human (at the end, when she is a will-o’-thewisp, she can again communicate with the human). But another device that the composer could have used would be to make Rusalka sing of her voicelessness. There would have been no inconsistency—though certainly a metatheatrical frisson—in so proceeding. Steiner criticizes Berg’s operatic version of Woyzeck on the grounds that ‘it distorts Büchner’s principal device. The music makes Woyzeck eloquent; a cunning orchestration gives speech to his soul. In the play, that soul is nearly mute and it is the lameness of Woyzeck’s words which conveys his suffering’ (ibid., 275–6). The scoring of Berg’s opera is certainly clever and sophisticated, with effective use (as also in Britten’s Turn of the Screw) of that most urbane of instruments, the coruscating and tinselly celeste, but I think Steiner exaggerates Büchner’s Woyzeck’s inarticulacy.204 It is true that, if he were right about that, then Berg’s orchestration might indeed count as a distortion. If the orchestration is too cunning, that is certainly a basis on which one can censure the musical version of the story. But if you wanted to say that Berg was distorting Büchner merely by dint of giving music to Woyzeck’s inarticulate soul, that would be a different and less cogent point; and you would have to add that Büchner was already distorting the reality by turning it into stage drama, which is after all a highly conventional and stylized art form. In fact neither artist distorts his subject matter, any more than, in geometry, the representation of transformations using matrices distorts space, or, in logic, the arithmetization of syntax distorts language and referentiality. It would, for example, betray a fundamental misunderstanding to object to the arithmetization of syntax by saying ‘But language is not just about numbers!’. All these transactions map one thing to another; they model a domain in a co-domain. They would only be guilty of distortion if some significant feature of the domain—something that the mapping was intended, or ought, to capture—were given no representation in the co-domain: so, for example, if an important feature of deafness found

Can Suffering Be Expressed in Words? 315 no musical analogue in our conjectured opera on the subject, that would be a basis on which we could criticize the opera for distorting reality. The general point here is one that I have already made several times, which is that we should not confuse the representing medium with what is represented. The two domains can—indeed must—have quite distinct properties. A soundless world can consistently and even interestingly be represented through the medium of sound. There is a further point to be made here, already stated briefly above, which is that only articulacy has a perspective on inarticulacy: inarticulacy does not have a perspective on itself, or on articulacy. Only non-­suffering has a perspective on suffering; suffering does not have a perspective on itself, or on non-suffering. Only painlessness can think about pain. (That would be so a fortiori if pain really were a Ding an sich.) Only health can think about health and sickness; sickness cannot think at all. Ill-health can be reality (though no more so than health), but it cannot understand or describe it. The Nietzschean idea that ill-health gives us insight into the true nature of reality and that health serves up mere illusions is the absolute reverse of the truth.205 Only the full stomach can think about the empty stomach. Recall Jean Améry’s account (§36) of the occasion at Auschwitz when he was unexpectedly given extra food by an orderly from the sick barrack, and found himself thinking about human goodness, literature, music, and philosophy. Only the saved—this is what Giorgio Agamben calls Levi’s paradox (2012, 76)—can describe the experience of the drowned: the Muselmann of the camps who has experienced complete dehumanization cannot speak about it; the survivor who did not descend to those depths can. Only adulthood can understand childhood, maturity adolescence, sanity insanity, the thinker the vacant, the speaker the dumb, the living the dead. The tale told by an idiot signifies nothing in the judgement of the sound of mind, and that verdict is definitive. Only the language user can say ‘what it is like’ to be a creature that has no language, and we can indeed say that, in the only sense that makes sense, namely the comparative sense. Only the optimistic civilization can produce tragedy; only the happy man can write sad poetry; only the freest mind can compose, or listen to, music that fathoms the pit of our travail. ‘I’m afraid there are moments in life,’ says Henry James’s Madame Merle to Isabel, ‘when even Schubert has nothing to say to us. We must admit, however, that they are our worst’ (1947, 185). What is there? How are things disposed? What causes what? What is the significance of what? These matters are settled by how we think of them in our best—our healthiest, cleverest, most articulate, most lingualized— moments. That brings me to the theme of linguistic idealism.

Notes 1. See e.g. Gascoigne and Kinwelmersh, Jocasta I, 2, chorus 41–2 (Cunliffe 1912, 85).

316  Tragedy and Linguistic Redress 2. Cf. Pöschl 1977, 133–5. 3. See Kermode 2001, 7–11, responding to Bate 1995, 62–3. Cf. Rowe 2003, 57; Burrow 2013, 106, 110–12; M. Neill 2016, 349–50. 4. See H. Nisbet 2013, 215; Billings 2014, 36–7. 5. For a different defence of Marcus’ rhetoric, see Burrow 2013, 112. Cf. also Martindale and Martindale 1990, 52. 6. Diggle 1998, 99. On Euripides’ blasphemy see also Andromache 1161–5; Hecuba 488–500; Electra 583–4, 737–46, 971–2, 1190–3, 1245–6, 1301– 2; Heracles 347, 655–6; Iphigeneia in Tauris 570–1; Helen 744–57, 851; Phoenician Women 86–7, 1726–7; cf. Seneca, Oedipus 1045; with the commentaries of Stevens, J. Gregory, Denniston, Cropp, Bond, Allan, Mastronarde, A. Boyle ad locc. See also Denniston 1939, xxii–xxiii, xxvi; R. Parker 1999, 21–2; Mastronarde 2010, 169–74; Rutherford 2012, 378–82; Gaskin 2013b, 68–9; Pucci 2016, 1–4 and passim. 7. Seneca, Medea 1027; cf. [Seneca], Octavia 911–13, with A. Boyle ad loc., and 2006, 218; T. S. Eliot 1951, 73; Miola 1992, 104–5. 8. Hume 2007, 74. See Lactantius, De Ira Dei, ch. 13 (PL, vol. 7, 121). 9. Camon 1997, 72. Cf. Wiechert 2008, 84, 92, 118, 137; C. Hamilton 2016, 13. 10. Rowan Williams skirts rather disingenuously round this point at 2016, 155. 11. So Eagleton 2003, 31. 12. Cf. C. Hamilton 2016, 89; Rowan Williams 2016, 16–17, 44, 70–1. 13. Cf. F. Lucas 1957, 180; Leech 1969, 51. 14. See Rowan Williams 2016, esp. 14–18, 71, 86–7, 143–4, 158. 15. Wiechert 2008, 65, 85–7, 105, 109–16, 137. 16. See D. Kühn 2007. 17. Poller 1947, 131–2. 18. Don Carlos III, 10, 3214–16; D. Kühn 2007, 208–15. 19. Cf. Schopenhauer 1986, §15, vol. 1, 119, 135. 20. Iliad VI, 357–8, tr. Lattimore; cf. Odyssey VIII, 579–80; Macleod 1982, 6; Redfield 1994, 35–6. 21. Hippolytus 1426–30, tr. Vellacott. Cf. the end of Hamlet, with Segal 1993, 24. 22. Trojan Women 1242–5, tr. Vellacott. Cf. Euripides, Trojan Women 394–9; Suppliants 1224–5; Alcestis 445–54; Seneca, Medea 423, with A. Boyle ad loc.; Agamemnon 659, with Tarrant ad loc. (though he oddly states that the motif is rare in Greek and Senecan tragedy); F. Lucas 1957, 77; Poole 1987, 140–1, 149; J. Gregory 1997, 176–9; Pucci 2016, 79–80. For the anti-type of the figure, see Shakespeare, The Rape of Lucrece 806–19. 23. See further Gaskin 2013a, 109. 24. Halleran 1986, 180–1; Segal 1993, 31. 25. See Segal 1996, 161–2; Allan 2008, 64–5. 26. Parry 1966, 111; cf. Goff 1990, 112–13, 120; Halliwell 2012, 66, 68, 76. 27. See e.g. Homer, Iliad XXIV, 513–14; Aeschylus, Persae 41–2, 135, 541, 1072, with Garvie ad locc.; Euripides, Suppliants 78–9; Trojan Women 608– 9, with Lee ad loc.; Andromache 93–5, with Stevens ad loc.; Electra 125–6, with Denniston and Cropp ad loc.; Aristotle, Rhetoric 1370b25–6; Burke, Enquiry I, §5; Shakespeare, King John III, 3, 92; Nuttall 1996, 17; Bristol 2009, 36–7; Halliwell 2012, 46–7, 88, with nn. 14, 18, 103; Hoxby 2015 112, 136–7. The topos connects with the so-called ‘paradox of tragedy’, which I hope to examine elsewhere. 28. Cf. Brewer 1989, 109; Steiner 1996, 544; Gaskin 2013b, 163–4; Rowan Williams 2016, 16–17, 86–7 (but also, inconsistently, 151, 156–7). 29. T. S. Eliot, ‘East Coker’ 207–9. 30. F. Lucas 1957, 77; cf. Bate 1995, 32–3.

Can Suffering Be Expressed in Words? 317 31. Virgil, Aeneid III, 621, 658; VIII, 625. See on these passages Hardie 2009, 96–7, 102, 123; Horsfall ad locc. 32. Troilus and Criseyde III, 1693–4; cf. V, 267–73, 1562–5. 33. T. Mann 2012a, 453–4, 470–1; cf. 483. See also his 2012b, 157; 2013, 516. 34. Schopenhauer 1986, §59, vol. 1, 445–7; cf. Nuttall 1967, 20, 29. 35. Melville 1998, 168; Hofmannsthal 1979, vol. 7, 461–72; Ibsen 1965, 273; Hardy 1997, 184 (cf. 219). See also Hardy 1971, 233; 1998, 187. 36. See A. Boyle on Seneca, Thyestes 633–8, 682–4, 1035–7. 37. See too Thucydides II, 50, 1; VII, 71, 2, with the commentaries of Rusten and Dover ad locc. The topos is parodied by Aristophanes at Birds 1706–17; see Dunbar ad loc. And once it is up and running, there is a natural temptation for poets to employ the topos for exaggerated and rhetorical effect, as in the familiar ‘Even if I had a hundred tongues, I could not express. . .’ figure found from Homer onwards: see on this figure Hinds 1998, 34–47. 38. Scarry 1985, 4; Woolf 2009, 102. 39. Scarry 1985, 5, 12, 13; cf. Morris 1991, 10. 40. Scarry 1985, 50, 12. Cf. Rowan Williams 2016, 41–2. 41. Budelmann 2006b, 137–8. 42. See Gaskin 2013a, §55. 43. Similar questions to ones I shall be addressing arise in connection with the sublime, but I must reserve consideration of that topic for another occasion. For discussion of the sublime in relation to tragedy, see Most 2000, 29–31; Trigg 2004; Beiser 2005, 257–62; A. Neill 2012; Shapsay 2012; Porter 2016, 334–50. 44. Marshall 2002, 179. This is a usual view of poetry (e.g. Wray 2009, 245), and music (Ch. 6). 45. A. Boyle 2006, 213; 2017, xliii, cxi with n. 286. 46. Bayley 1981, 10; cf. Bate 1995, 32; Nooter 2012, 21–2; Zamir 2016, 79–80. 47. Garvie 2009, 109. He compares our passage in the Women of Trachis, and also mentions Sophocles, Ajax 214; Oedipus at Colonus 1001; Euripides, Hippolytus 602; Ion 782–3. 48. Notes on Seneca, Oedipus 15, and on Seneca, Trojan Women 1065f. See also Agamemnon 981, with Tarrant ad loc. 49. Boyle assumes the connection (on Seneca, Oedipus 15); the Oxford Latin Dictionary is more tentative, s. v. ‘fas’. 50. T. S. Eliot, ‘Little Gidding’ 210. 51. Budelmann 2007, 448–9. 52. Porter 2016, 302 with n. 56. 53. On the text, see Mastronarde ad loc. 54. Poller 1947, 195–6. Cf. his description of the ‘Dunkelbaracke’ at 158–60. 55. Levi 1958, 23; tr. S. Woolf 1989, 32. 56. Langer 1995, 95. Cf. Bauer’s remarks on Elie Wiesel: 2001, 15. 57. A reader objects: ‘But the question you set out to address was expressibility, not categorizability’. It is important to see—and more will be said about this in the course of this and the next chapter—that these are the same question. 58. So too in respect of the eyewitness accounts collected by Gideon Greif in his 2011. 59. Cf. Vernant in Vernant and Vidal-Naquet 2001, vol. 1, 104; Murnaghan 2009, 237; Rutherford 2012, 299. 60. Améry 2008, vol. 2, 72–3; tr. Sidney and Stella Rosenfeld, 32. 61. Améry 2008, vol. 2, 73–4; tr. 33. 62. See further on this point Gaskin 2018.

318  Tragedy and Linguistic Redress 3. See here Glover 2001, 36–7; Pinker 2002, 274, 320–1. 6 64. See Hacking 1999, ch. 5. 65. Cf. Haslanger 2012, 127–8. 66. Cf. Mastronarde ad loc.; Euripides, Suppliants 844; Iphigeneia in Tauris 837–40. 67. Gascoigne and Kinwelmersh, Jocasta II, 1, 145: ‘So heavie happe as toung can not expresse’ (Cunliffe 1912, 91). 68. See Mastronarde on 413. 69. Cf. T. Mann 2012d, 29. 70. Winterbottom 1993, 26. See also Aeneid III, 94–6; Gransden on Aeneid VIII, 37–8, 134ff., and p. 11 of his introduction. 71. See Blundell 1989, 43–4; Winterbottom 1993, 17; Hall 1997, 97–8; Putnam 2006, 409; 2010; Griffin 2007, 197–201; Harrison 2007, 129–34; A. Boyle on Seneca, Medea 19–23; Finglass on Sophocles, Ajax 1019–20, and on Electra 602. The Octavia paints an idealistic picture of Seneca’s exile in Corsica, at 381–90, but we know that the reality was different: see A. Boyle ad loc. 72. Sextus, Against the Logicians I (= Adv. Math. VII), 83–4. 73. See on this point Candlish 1989, 347; Dodd 2000, 168, 173. 74. Gulliver’s Travels III, ch. 5 (1985, 230). 75. Gaskin 2008, 244–5; 2013a, 301–2. 76. Byron 1980–93, vol. 2, 309. 77. Horace, Ars Poetica 361. Simonides famously stated that painting is silent poetry and poetry speaking painting: see Russell and Winterbottom 1972, 5. Aristotle gave a significant impulse to the view by beginning the Poetics (as Horace begins the Ars Poetica) with a comparison between poetry and painting; in so doing he was picking up an important feature of Plato’s philosophy: see Schmidt 2001, 26. Cf. Quintilian XI, 3, 67; Plutarch, Moralia, 17f–18a. See further D. Lucas 1972, xix, 56, 269; Zeitlin 1995, 185; Grady 2003, 436; Tarrant on Virgil, Aeneid XII, 397. 78. See S. Davies 1994, 12–14, 19–24, 51–79; Gaskin 2009, 55–7. 79. Laokoon, §16 (1996, vol. 6, 102–9). 80. Laokoon, §17 (1996, vol. 6, 112–13). 81. Laokoon, §16 (1996, vol. 6, 103). 82. Diamond 2008, 58; cf. Hacking 2008, 169. 83. Cf. Seneca, Thyestes 684: ‘quis queat digne eloqui?’. 84. Quintilian III, 8, 45; cf. A. Boyle 2017, lxxvi–lxxvii; Virgil, Aeneid II, 155; Stefonio, Crispus V, 4, 657–8. 85. Braden 1985, 215. 86. Coleridge 1987, vol. 2, 315. Cf. Halpern 2016, 31; Nuttall 2007b, 281–4. 87. Sophocles, Ajax 946–8, tr. Finglass. See Finglass ad loc.; Fraenkel on Aeschylus, Agamemnon 238. 88. Lamb 1903, vol. 2, 29; Shakespeare, King Lear I, 1, 55; cf. Aristophanes, Birds 427 (with Dunbar ad loc.), 1189; Sidney 1999, 162; Smollett 1998, 153, 314. On Goneril, see Nuttall 2007b, 304. 89. Girard 2010, 87–8. 90. Cf. Aeschylus, Suppliants 908. 91. Euripides, Hippolytus 846, 875; cf. 602; McClure 1999, 126. 92. See Clay 1982. 93. Shelley, The Cenci V, 3, 81; III, 1, 116; III, 1, 141–2. 94. Cf. Tecmessa at Sophocles, Ajax 214, with Finglass ad loc; Seneca, Phoenician Women 540, with Frank ad loc.; Virgil, Aeneid II, 3; VIII, 489; Mussato, Ecerinis 443, 447 (Grund 2011, 32, 34; cf. xix–xx); Grund 2015b, 105–6, 108–10; Holbrook 2015, 62. 95. Locke, Essay III, 4, 11.

Can Suffering Be Expressed in Words? 319 96. Evans 1982; see e.g. 307–8. 97. Confusion on this point is common: see e.g. Hellie 2004, 342. 98. Cf. McDowell 1994, 57–8. 99. Wittgenstein 1977, I, §§16, 50. 100. Podlecki 1966, 235. 101. Scarry 1985, 19, 35, 54, 172; see also Poole 2005, 66. 102. Seneca, Phaedra 607; cf. Ovid, Metamorphoses VI, 583–5; Garnier, Hippolyte 1345–6. 103. Note too Electra’s silence at Sophocles, Electra 764–87: see Finglass 2007, 336. 104. Webster, The White Devil II, 1, 276–7; see Luckyj ad loc. See also Alabaster, Roxana II, 1, 364–6; Shakespeare, Macbeth IV, 3, 210–1, with Clark and Mason ad loc.; Miola 1992, 93–4; Hamlet I, 2, 85, with Jenkins ad loc. 105. Shakespeare, Titus Andronicus III, 2, 44; cf. Bate 1993, 117. 106. M. Neill 2016, 356. See further Hughes, The Misfortunes of Arthur IV, 2, 14; Shakespeare, Sir Thomas More 13, 195, with Jowett ad loc. (2011, 299); Middleton, The Revenger’s Tragedy I, 4, 23 (2007, 557); Gager, Dido 279–80 (1994, vol. 1, 278); Marston, The Tragedy of Sophohisba V, 4, 54–7. 107. See here Gaskin 2006, ch. 4. 108. Cf. S. Davies 1994, 158–9. 109. See McDowell 1998a, Essay 13; Gaskin 2006, 91; 2013a, 207. 110. Wittgenstein 1977, I, §261. 111. Wittgenstein 1977, I, §243. 112. Cf. Pinker 2011, 647–8. 113. Wittgenstein 1977, I, §303 (tr. Anscombe). 114. Kant himself associated the sublime with the supersensible: see Kritik der Urteilskraft, §§25–8. Cf. Crowther 1989, 110–11; Young 2013, 88. 115. Cf. Scarry 1985, 54–5; Honig 2013, 18. 116. Büchner, Dantons Tod III, 1 (2009, 107). 117. On Seneca’s treatment of pain more generally, see C. Edwards 1999. 118. Tolstoy 1960, 138; see Morris 1991, 32. 119. Kafka 1990, 899; tr. 410. Cf. Heidegger 1990, 85. 120. Cf. Gaskin 2013a, 206–7. 121. See Pelling 1997, 230–2; Mastronarde 2010, 268–9; Rutherford 2012, 259–60. 122. See Morwood on 1123. 123. Cf. Morwood on 1146. 124. See Mastronarde on 1034–5. 125. Euripides, Hecuba 689–90. See too 96–7, with J. Gregory ad loc. 126. Shakespeare, King Lear V, 3, 304–7. On the end of Lear, see Shapiro 2015, 349–57. 127. See Bradley 1991, 270; Silk 2004, 243; Gaskin 2013a, 78. 128. Wittgenstein 1977, I, §261, tr. Anscombe. Cf. Waismann 1976, 309; Gaskin 1996, 96–8. 129. Wittgenstein 1977, I, §261, tr. Anscombe, adapted; see von Savigny ad loc. 130. Sophocles, Philoctetes 743, 746, 754, 790; see Schein 2013, 236. 131. Cicero, Tusculan Disputations II, 7, 19, 23. See Schein 2013, 43–58; Hoxby 2015, 152–61. 132. Lessing, Laokoon, §§1, 4 (1996, vol. 6, 13, 28–40); Herder 1985–2000, vol. 1, 697; Scarry 1985, 5; cf. Morris 1991, 248–55. 133. See esp. Nooter 2012, ch. 4. 134. So e.g. Knox 1964, 130 (quoted below, §43); Hoxby 2015, 118.

320  Tragedy and Linguistic Redress 135. A parallel point seems to hold of ‘attatai’: see Dover on Aristophanes, Frogs 649 (and on 1264f.); cf. Thesmophoriazusae 223, 945, 1005, 1191, with Austin and Olson ad locc. 136. Cf. M. Nussbaum 1986, 45; Seaford on Euripides, Cyclops 110, 503; Dover on Aristophanes, Frogs 57. See also Aeschylus, Persae 1031–2. The word was Latinized (‘papae’), and used in comedy as an expression of surprise, but is found in a serious context again in e.g. Leonardo Dati’s Hiempsal, at 339 (Grund 2011, 216). 137. See Cairns 2005, 311–12. 138. On ‘for a long time’ (‘palai’), see Goldhill 2012, 68–9. Cf. also Philoctetes 806 and 966, with Schein on 965–6. 139. See Schein on Philoctetes 283, 732, 1169–70; Weissberg 1989, 557, 565; Rutherford 2012, 123, 311–12, 334. 140. See Gaskin 2008, esp. §§9, 15, 38, 40, 53. 141. Philoctetes 732, 739, 782. Cf. Euripides, Andromache 1076, with Stevens ad loc. 142. Cf. Aeschylus, Persae 550–81; Seven Against Thebes 150, 158; [Aeschylus], Prometheus Vinctus 66, 566; Euripides, Andromache 1187–8; Hippolytus 594; Trojan Women 105, 164, 168, and passim. See also Budelmann 2007, 450; Garvie 2009, 342; Aristophanes, Frogs 1029, with Dover ad loc.; Thesmophoriazusae 1042, with Austin and Olson ad loc. 143. Steiner 1961, 281; Fraenkel 1978, vol. 3, 539, and see on line 1305. See also Segal 1981, 57–8; Loraux 2002, 75–80; Poole 2005, 93–4. 144. Cf. Gaskin 1989, 333–4. 145. Brecht, Mutter Courage und ihre Kinder, Scene 1 (1977, vol. 4, 1359). 146. See here again Gaskin 2008, §§38, 40. 147. See Gaskin 2008, 29–33. 148. Cf. Seaford on Euripides, Cyclops 49–50. 149. Budelmann 2006b, 140–1. 150. Note the connection between ‘aiai’ and ‘aiete’ at Euripides, Suppliants 819– 20, with Morwood ad loc. 151. See Aeschylus, Suppliants 114–15, where the move is esp. clear, with Bowen ad loc. See also Halliwell 2012, 64–5; Austin and Olson on Aristophanes, Thesmophoriazusae 173–4. 152. Page 1936, 209–10; Rudd on Horace, Ars Poetica 75. 153. Bowen 2013, 171; cf. Budelmann 2007, 445. 154. See Dover 1993, 219. 155. See Aristophanes, Birds 193, 227, 228, 235–7, 242, 259–60, 310–12, 857, 889, 1122, 1170, 1343, 1510, 1764–5, with Dunbar ad locc. 156. Taplin 2003, 120–1. 157. See Fraenkel 1978, vol. 3, 580 n. 4; Seaford on Euripides, Cyclops 157; Page on Medea 1004; Stevens on Andromache 896; Denniston on Electra 341; Willink on Orestes 277; Liapis on [Euripides], Rhesus 574; Austin and Olson on Aristophanes, Thesmophoriazusae 699; Dunbar on Birds 327. 158. Mozart, Die Zauberflöte I, 7 (2011, 73). 159. Pace Scarry 1985, 5; Weissberg 1989, 558. 160. Austen 2003, 303; Racine, Phèdre 156; Bérénice 1362; Athalie 438. 161. Gaskin 1989, 330–1. 162. Pocock 1973, 208. 163. Cf. Leech 1969, 80. 164. See Kermode 2001, 11. 165. Homer, Odyssey XI, 563–4, tr. Lattimore, adapted; Longinus, De Sublimitate 9, 2.

Can Suffering Be Expressed in Words? 321 166. An Enquiry Concerning the Principles of Morals, §VII (1975, 252). 167. See here Halliwell 2012, 357; Porter 2016, 94–7, 401; Doran 2017, 61. 168. Cf. Loraux 1985, 48–9; Taplin 2003, 102. 169. See e.g. Steiner 1998, 89. 170. See Santirocco 1980, 194; Griffith 1999, 340; Poole 1987, 57; 2005, 86; Mastronarde on Euripides, Phoenician Women 635. 171. See 999–1003 with A. Boyle on 999 and 1118–64. 172. See Garvie 2009, 143–5. 173. Cf. Taplin 2003, 109. 174. Cf. Stevens 1971, 218–19, 243. 175. See Goldhill 2012, 97–8. Here there is a question whether we have the Sophoclean ending: see Finglass ad loc. for further discussion. 176. See Goldhill 2012, 47. 177. See Podlecki 1966, 241; Goldhill 2012, 42–7. 178. J. Gould 2001, 400–1. 179. See Tarrant 2012, 83. 180. See further on dramatic silences Mastronarde on Euripides, Phoenician Women 1090–2; Taplin 1972; 2003, ch. 7; Parish 1993, 191–2; Griffith 2001, 123–4; Rutherford 2012, 16. 181. Mutter Courage, Scene 3 (1977, vol. 4, 1391). 182. Gaskin 2013a, §17. 183. Gaskin 2008, §14; cf. Evans 1982, 400. 184. Pöschl 1970, 156. 185. Nuttall 1996, 86. 186. Barthes 1963, 91; cf. Hoxby 2015, 132–4. 187. Prince 1960, xxxvi; cf. Bate 1993, 81; Duncan-Jones and Woudhuysen 2007, 34. 188. Cicero, Tusculan Disputations II, 38–9; cf. Garton 1972, 195–6. 189. Kelly 1997, 223. 190. Steiner 1998, 5–7, 50. 191. Klee 1988, 72. 192. Steiner 1971, 122–3; cf. 115. 193. For related thoughts on Richard III, see Greenblatt 2001, 164–7, 179–80. 194. Schiller, Wallensteins Tod V, 4 (3577). 195. J. Campbell 2005, 194; cf. 145–50, 194–204. Cf. Harth 1972, 391–5. 196. Levi 1958, 116; tr. S. Woolf 1989, 135–6, adapted. 197. H. Nisbet 2013, 621. 198. See J. Campbell 2005, 199. 199. See here Gaskin 2013a, ch. 5. 200. Cf. J. Mann 2003, 100. 201. Chaucer, Troilus and Criseyde V, 1527–33; cf. Kelly 1997, 128–9. 202. Cf. Wittgenstein 1977, I, §424. 203. Cf. De Quincey 2000–4, vol. 15, 318, on ‘whistling Waterloo’. 204. Cf. Gelfert 1995, 116–19. 205. See here C. Hamilton 2016, 53–4.

6 Tragedy and Linguistic Idealism

45 Language and World On the one hand there are words, such as ‘pain’; on the other hand there are things of certain sorts, such as pains. Words are themselves things of a certain sort, but they are not usually the things to which they refer. We have a tendency to feel let down by language, I suggested in the last chapter, because words are not (usually) the things of certain sorts to which they refer, forgetting that they are not supposed to be; alternatively, we are inclined to feel disappointment because words cannot express thingsin-themselves, forgetting that the whole idea of Dinge an sich is incoherent. Again, we in some sort resent the fact that language does not, as such, give us experience of the things it talks about. In ‘outer’ e­ xperience— experience of tables and chairs—we do not notice these facts, and they do not trouble us; but when we think about ‘inner’ experience—(­experience of) sensations—we do notice them and they do trouble us. We are tricked by the (apparent) proximity to the self of ‘inner’ experience into supposing that ‘inner’ experience gives us a more direct access to objects than ‘outer’ experience does: we suppose that, in effect, when we feel a pain we are confronted with a thing-in-itself, whereas when we see a table we stand over against a thing of a certain sort. In truth a pain is just as much a thing of a certain sort as a table is, but we fail to notice this, and are then disappointed that the word ‘pain’ is not, and cannot give us an experience of, a pain, whereas we never expected the word ‘table’ to be, or to give us an experience of, a table, and so in that case do not feel let down. It is as though we supposed that we came closer to reality in our experience of pains than in our experience of tables, so that when the inevitable gap between language and reality looms it strikes us as more unconscionable in the ‘inner’ case than in the ‘outer’. But the incoherence of so much of our thinking about the ‘inner’ is well brought out by Wittgenstein, who proceeds from the hypothesis, set up for the purpose of reductio, that if an individual really were confronted with ‘inner’ objects—­sensations—that were somehow things-in-themselves, it would make sense to suppose that that individual could talk about those

Tragedy and Linguistic Idealism 323 objects, objects to which he, and he alone, had access, in a language which he, and he alone, understood. But these things do not make sense, for the simple reason that even to posit that there is a thing, purportedly featuring in an individual’s experience and talked about in that individual’s purportedly private language, is already to have located the whole transaction firmly within reach of the public language. The thesis of linguistic idealism, in the version I favour, is the thesis that the world is essentially expressible in language, that it is not constituted independently of the possibility of language, that it is a transcendental precipitate of language, that it is the internal accusative of language, that it is composed of entities which are in some sense meanings. I take all these formulations to be variations on a single theme—the idea that what there is, is constitutively sayable.1 Consider one of these formulations. When I say that the world is a transcendental precipitate of language, I am using the word ‘transcendental’ in its Kantian sense, according to which it introduces necessary conditions: the idea is that the world is necessarily fitted up to be expressed in language, that it is essential to things, of whatever sort—and taking ‘things’ here in the widest possible sense, so including both concrete and abstract things, which latter category includes, for example, properties, relations, modes, operations, and functions—that they can be expressed in language as such, that is, in some language or other (actual or possible). This is a necessary condition of their sheer existence. In the version of linguistic idealism that I favour, it is claimed the world is made up, in the (metaphysically) first instance, of propositions, which are constitutively the referents of declarative sentences.2 But propositions are themselves composed of objects and various kinds of abstracta, such as properties, so that in the second instance the world is composed of the familiar objects and abstracta that we ordinarily take it to be composed of: we are right to think that tables and chairs and other concrete objects, as well as justice and inflation and sets and numbers and negation and Fregean senses and other abstract objects, are in the world, but their existence in the world is derivative of their existence as components of propositions, these propositions being, from a metaphysical point of view, what the world initially comprises. Goethe’s Faust brazenly corrected ‘In the beginning was the word’ to ‘In the beginning was the deed’ (I, 1237), and the later Wittgenstein approved of the change.3 But the deed is only a deed in the recognition of language; the deed owes its status as such to the word. We should stick with the vulgate, suitably secularized. There might seem to be a tension between linguistic idealism, as just expounded, and the way I defended the adequacy of language in the last chapter (and recapitulated in the first paragraph of this chapter). My argument for language’s adequacy, against the widely claimed inexpressibility in language of suffering and other feelings, has been based on the observation that words are (usually) different from the things that they

324  Tragedy and Linguistic Redress refer to: for example, words for sensations are not themselves sensations; the aetiology of the common complaint about the alleged inadequacy of language, in the area of sensations, resides in a forgetting that we do not normally expect words to be what they refer to, or to convey an experience of what they refer to; if we apply these latter points to the language of sensations, as we do ungrudgingly to the language of tables, our sense of disappointment will (I hope and predict) evaporate. Does this stress on the difference between language and what language is about square with linguistic idealism, as I have outlined that? Yes. My linguistic idealist is not saying that the world is made up of (bits of) language. It is made up (in the first instance) of propositions, and propositions, though they are essentially the referents of bits of language (declarative sentences), are not themselves bits of language. So a correct understanding of the referential relation is the key to dissolving the appearance of antinomy. The world is constitutively the referent of language. This sentence can be read with different emphases. For example: wearing my linguistic-idealist hat I read this sentence as saying that the world is constitutively the referent of language, as opposed to being something else—that is, that the world is that particular thing, the referent of language, as opposed to its being something that has nothing inherently to do with language; by contrast, what I argued in the last chapter, in defending the adequacy of language to express feelings, was that the world is constitutively the referent of language, as opposed to having some other relation to language, and in particular as opposed to being identical with language itself.4 Putting the matter the other way round, we might say that the referential relation running between language and the world—between an item of language and a worldly entity, whether that referent be a proposition, a concrete object, a property, a function, or some other kind of abstract entity— (normally) connects its relata as two different things; but it really does connect them, and that fact is constitutive of both the world and language. World and language are distinct things, but the latter really does express the former, and that fact is essential to both. In a discussion of changes that language has undergone since the onset of the early-modern period, Steiner writes that until the seventeenth century, the sphere of language encompassed nearly the whole of experience and reality; today, it comprises a narrower domain. It no longer articulates, or is relevant to, all major modes of action, thought, and sensibility. Large areas of meaning and praxis now belong to such non-verbal languages as mathematics, symbolic logic, and formulas of chemical or electronic relation. . . . The world of words has shrunk. One cannot talk of transfinite numbers except mathematically; one should not, suggests Wittgenstein, talk of ethics or aesthetics within the [currently] available categories of discourse.5

Tragedy and Linguistic Idealism 325 This can hardly be right as it stands. We are alerted to the difficulty by the tension between (i) Steiner’s initial statement that the range of language has narrowed and (ii) his passing concession, further on, that mathematics and so on, though they are (Steiner thinks) non-verbal, nevertheless are languages. If mathematics and other ‘non-verbal’ languages really are languages, then the range of language as such has not narrowed; all that has happened is that there has been an adjustment of the territories within its borders—the non-verbal has expanded at the expense of the verbal. But in fact mathematics, logic, chemistry, and electronics are not nonverbal. What is correct is that the scope of technical languages has greatly increased in modern times, but technical languages are, deeply, verbal. Indeed they are simply extensions of natural language, not replacements for it, and there is no competition between the two, as Steiner implies. (Recall Wittgenstein’s analogy between language and a city: the old parts of language are like the medieval city centre, the new parts like the city’s outlying, recently built suburbs: 1977, I, §18.) Everything that is expressed in (say) the symbolism of some formalization of logic could be expressed in words of English, or German, or. . .; it is just that it would be extremely tedious to read and write everything that we want to express by means of that symbolism in a natural-language translation of it—and not only tedious, but also detrimental to comprehension, for to those who know the dialect it is much easier to read even a relatively complicated formula than it would be to construe its natural-language translation. But the natural-language version is there in the background—it has to be there—and can be wheeled out as and when necessary. There is, then, no problem about talking of the transfinite numbers in natural language: indeed our ability to talk about them mathematically depends on that possibility. The continuum hypothesis, for example, can be expressed in special symbols or in ordinary words, and there is nothing in the symbolic expression of it that is not in the natural-language version except concision and ease of comprehension. All technical discourse that enriches natural language by the use of special symbolism is obliged to define that symbolism, and though the process of definition need not be fixed to natural language immediately, it must eventually be so. This fact shows that the idea, often implied or asserted, that symbolic languages are more precise than natural language is and must be false. All special symbolism is tied down, sooner or later, to natural-language definitions, so that if one is prepared to be long-winded enough any symbolic formulation can be translated into a natural-language version which is every bit as precise as the original; as I have just said, the only advantages of special symbolism are concision and ease of comprehension. And, as far as the final sentence of the quoted passage goes, Steiner omits to mention that Wittgenstein’s Tractarian thesis of the ineffability of ethics and aesthetics depends upon his picture theory of meaning, according to which sentences, at least in their fully analysed state, are configurations of names

326  Tragedy and Linguistic Redress mirroring in their structural arrangement configurations of objects on the ground, so that in the absence of ethical or aesthetic objects, there is nothing ethical or aesthetic for these sentences to express; but that theory is highly contentious, to put it mildly, and in fact it is arguably false, at least in the version expounded by Wittgenstein (in which names go proxy for objects), so far as we are able to reconstruct that.6 It is a mistake to think that the Tractarian Wittgenstein has successfully shown that the idea of the limits of language makes sense, and that beyond those limits lies the unsayable.7 He has done no such thing. Accordingly, technical languages do not in any fundamental way shrink the domain or the competence of natural language: they cannot do that, for they depend on natural language, not only genetically, as a matter of their historical provenance, but also constitutively, as a matter of their syntactic and semantic essence. Steiner argues that natural languages can, in a sense, make up for the invasion of their sovereignty that he mistakenly posits by stretching their competence in the direction of music. Thus he writes of the final section of Hermann Broch’s Der Tod des Vergil (1998, 29): The words literally flow in sustained polyphony. Strands of argument interweave exactly as in a string quartet; there are fugal developments in which images are repeated at governed intervals; and, at the last, language gathers to a dim, sensuous rush as remembrance, present awareness, and prophetic intimation join in a single great chord. The entire novel, in fact, is an attempt to transcend language towards more delicate and precise conveyances of meaning. We may agree that there are ways in which words, and works of literature, can imitate music. But words cannot ‘literally flow in sustained polyphony’. They could only do that if we read several parallel lines of type at once, or listened to several distinct recitations simultaneously. That would be polyphony, and it would make possible the musical effect of counterpoint (perhaps even harmony). Again, if strands of argument interwove ‘exactly as in a string quartet’ they would have to be read or heard simultaneously. But most of us are incapable of following such polyphonic effects in the domain of words: for most people, words have to be read or heard serially, not in parallel, if they are to be understood. That is indeed the point: the difference between words and notes of music lies in the fact that you do not have to understand notes of music; you just have to hear them.8 That means that you can hear several musical notes at once, without getting confused—for no activity of understanding is involved, at least not in the (semantic) sense in which words are understood9—whereas you have to understand each word you read or hear with the whole of your being. The glorious ‘respice’ moment of Thomas Tallis’s Spem in alium, where eight choirs, each comprising five

Tragedy and Linguistic Idealism 327 voices, merge in a wash of heavenly sound, would not be possible as a purely verbal effect; at least it would not be possible in any language that was meant to be understood by us. Certainly you could arrange for forty people to perform Spem in alium as a spoken text, and the result might be in some sense music. But, although there could no doubt in principle be creatures capable of simultaneously listening to and understanding forty distinct philosophy lectures, or simultaneously reading and comprehending forty distinct mathematical proofs, or forty different Russian novels, these are not skills that we possess. So I think that Steiner’s suggestion that ‘wherever literary structure strives toward new potentialities, wherever the old categories are challenged by genuine compulsion, the writer will reach out to one of the other principal grammars of human perception—art, music, or more recently, mathematics’ (1998, 87) is mistaken. This involves a misuse of the word ‘grammar’: art, music, and mathematics do not have their own grammars. Art and music do not have their own grammars because they do not have grammars at all:10 they are rule-governed,11 but that is not enough to make them linguistic. Mathematics, by contrast, does have a grammar, but its grammar is just the grammar of natural language (subject and predicate, punctuation and scope-indication, and so on), so it does not have its own grammar. The same goes for the grammar of the Tractatus: Steiner’s suggestion that this work ‘makes its own syntax’ (1998, 89) is, at least in one obvious sense, wrong, since the Tractatus is written in ordinary language, using ordinary syntax, and is perfectly comprehensible as such. But there is another sense in which Steiner is right, because in that ordinarily written work Wittgenstein propounds a doctrine of analysis which commits him to the view that sentences of ordinary language can be analysed and that, at the end point of this analysis, we will be left with sentences that, syntactically, look nothing like what we are used to (they may not even be linear). And it is hard to see how it could be otherwise, given that ordinary language does not conform to the constraints that Wittgenstein places on fully perspicuous language, in particular the constraint that its sentences must have the same logical multiplicity as the reality of which they treat (Tractatus, §4.04).12 It is an implication of the doctrine of the Tractatus—one that Wittgenstein himself explicitly embraces (§6.54)—that its own sentences are nonsense, because they do not conform to the criteria of meaningfulness set out in that work: they are not statements about arrangements of objects. But, as I have implied here and earlier (§43), this upshot constitutes a reductio of Wittgenstein’s Tractarian position, because the sentences of the Tractatus are not nonsense. (Many readers of the Tractatus seem to me rather too ready to take Wittgenstein’s own pronouncement on this point at face value.)13 They are written in ordinary German, and on that basis they can be, and are, understood. Naturally commentators argue about their precise import, but that is true of any serious literary or philosophical work: no one

328  Tragedy and Linguistic Redress would suggest that the sentences that compose Shakespeare’s Hamlet or Descartes’s Meditations are nonsense on the grounds that reputable critics do not all agree on their correct interpretation.

46 The Competence of Language Steiner’s view just stated is a popular one, especially among practising artists in non-verbal media, such as musicians. Debussy told his friends ‘that he wanted to create a work in which “music would take over at the point at which words became powerless, with the one and only object of expressing that which nothing but music could express” ’.14 (Theoretically, the approach can be derived from Schopenhauer’s thesis that music expresses the world-in-itself: 1986, §52, vol. 1, 365–8.) In what sense, if any, can music express what words cannot? You might say that it was simply guaranteed by the semantic range of the verb ‘express’, and its congeners in other languages (‘ausdrücken’, ‘exprimer’, etc.), that music can indeed express something, or things: for we talk in that way all the time. Let us accept this. But can music express things that words cannot express? Well, an obvious point to make here is that, while music can express things, it does not express things in a semantic sense. Correlatively, music can be grasped or understood,15 but that is not a matter of grasping the meanings of the components of a linguistic system—a system for which a syntax and semantics can be specified, and which employs its syntax and semantics to convey propositions with truth-values.16 Only language can convey meaning in that sense, and the linguistic idealist claims that language can indeed express everything in its proprietary semantic sense. At first blush, that claim might look like mere tautology: it might look as though what we had said was just tantamount to ‘language can express everything that can be expressed by language’. What would make the position non-trivial, and interesting, would be the detail about what language can indeed express, and why we mistakenly think that there are things, such as pains and emotions, that it cannot express. In my discussion of the last chapter I suggested that we are mistakenly inclined to fault language for not simply giving us a pain when the word ‘pain’ is uttered. A similar point applies to emotions: the word ‘sad’ does not, as such, make us sad, and because of that we sometimes feel, mistakenly, that this fact derogates from language’s expressive adequacy. Now it may be that music gives us emotions. Perhaps also music conveys emotions by exemplifying them.17 And, though the word ‘sad’ does not work by making us sad—that is not its semantic role—nevertheless language can in certain contexts give us emotions, and perhaps also it conveys emotions, when it does so, by exemplifying them. We have sad poetry as we have sad music. In addition, poetry’s ability to give us emotions is not a purely musical effect, divorced from the referential function of language: it is intimately bound up with meaning.18 But this fact does not

Tragedy and Linguistic Idealism 329 palliate the specialness of music, or refute Debussy’s point, because a sad poem, for example, cannot substitute for a sad piece of music: each is sad in its own way. Those ways are a mix in each case of individual and general features, but even the general features will include some that are special to the art form, musical or verbal, in question. So, though there might, and presumably would, be commonalities between what a sad poem and a sad piece of music each expressed, it would be mistaken to say that the one expressed just what the other expressed. The emotion that the sad poem gives us, or exemplifies, is not, and cannot be, the same as the emotion that the sad piece of music gives us, or exemplifies. The modalities in question are quite different, and the emotions produced are at least in part proprietary to their modes of production, literary and musical. So the emotions produced are logically inseparable from the works that produce them. This was a point that Wittgenstein, with characteristic flair, implied when he wrote that, if you heard that someone was painting a picture titled ‘Beethoven writing the ninth symphony’, you could easily imagine what such a picture might portray, but ‘suppose someone wanted to represent what Goethe would have looked like writing the ninth symphony. Here I could imagine nothing that would not be embarrassing and ridiculous’.19 But in lodging the thesis that language can express everything, the linguistic idealist is not interested in the power of words to give us, or exemplify, emotions, so that the fact that music gives us, or exemplifies, emotions that words cannot give us, or exemplify (and vice versa), does not constitute a problem for the thesis. We can acknowledge that, in that sense, music does things that language does not, and cannot, do. Perhaps language can do similar things (it can be sad, like music), but it cannot do just what music does: no stretch of language can give us the proprietary sadness of a particular piece of music. But language’s inability to give us the proprietary sadness of a particular piece of music represents no failure on its part, because language is not in the business of giving us, or of exemplifying, the things it talks about, but rather in the business of exactly talking about those things. (Music, as I have said, cannot do this—talk about these things—because it is not linguistic.) And it is here that the linguistic idealist intervenes to assert that, in that sense, language can express everything, including sensations and emotions. The omnicompetence of language consists in its ability to talk about—name and describe—everything. Suppose you and I are discussing the expressive adequacy of language, and when I assert ‘Language can express everything’ you counter with ‘No it can’t: for example, it can’t express toothpaste from a tube’. Yours is a perfectly legitimate use of the verb ‘express’: indeed it is the original, non-metaphorical use. So too for ‘ausdrücken’ and ‘erpressen’ in German, and ‘exprimer’ in French. In the sacrifice that forms the centrepiece of Seneca’s Oedipus, the slaughtered bull exprimit, ‘expresses’ (344), its

330  Tragedy and Linguistic Redress struggling soul;20 in the same author’s Hercules Furens the Clashing Rocks ‘will express’ (expriment, 1213) the sea upwards when they converge.21 Now language not only does not express everything in that original, literal sense of the verb; it does not express anything in that sense, either. Nor does music. So the assertion that language can express everything needs to be limited in an appropriate way—a way that (the linguistic idealist hopes) makes the claim true, but not at the cost of trivializing it. In order to steer this middle path between falsity and triviality, I suggest we say (avoiding the falsity) that language can express everything that is fitted up to be expressed by words, in a way that is appropriate to verbal expression, and (avoiding the triviality) that everything can indeed be so expressed, that is, can be named and described. When we reflect on what language cannot express, such as toothpaste from a tube, or the precise emotions conveyed by Sibelius’s Sixth Symphony, we see that this ‘cannot’ does not represent any genuine failure on the part of language: for it can talk about, that is name and describe these things, and that is all it is supposed to do. The emotions that Sibelius’s symphony gives us include emotions that already have names in the language, and are logically independent of the experience of listening to that work, together with emotions that can only be adequately characterized in terms of that very symphony. Only the music, heard or recollected, can actually give us precisely those emotions, just as only Milton’s ‘Lycidas’ can give us precisely the emotions we feel when we read it, but those emotions, in each case, can be expressed by language in general in the sense that it can talk about them: language can name and describe them. Actually, both language and music can express toothpaste from a tube: the reader may have thought of this point. There would be no difficulty, even with fairly old technology, in rigging up some Heath-Robinson contraption that received words or music as causal input and produced the expressing of toothpaste from a tube as output; and it is quite possible that one day soon, when we come to surround ourselves with ‘the internet of things’, we will have tubes of toothpaste that respond to spoken commands. But I refrained from mentioning this point in the last paragraph, since I presume that the sense in which words or music express toothpaste from a tube, in these imagined scenarios, is irrelevant to the discussion: for spoken words, to take that case, would not, in the envisaged scenario, be expressing the toothpaste qua meaningful language, but simply qua noise of such and such a form. But my claim is that language can express everything qua meaningful language, and in order to defend that claim I need to limit it in the way specified, that is, gloss it as the assertion that language can talk about—name and describe—­ everything. What we here have to avoid is the temptation to present matters as though language, in not doing what it does not do, were a failure, when in fact it is a success, because in fact it does everything that it is meant to do.22 You cannot put your cup of coffee on the word ‘table’,

Tragedy and Linguistic Idealism 331 only on a table, but we do not for that reason think that language is any kind of failure. In his discussion of the ‘Dover Cliff’ scene in King Lear, Jonathan Goldberg writes that ‘The refusal to allow the word Dover to arrive at the place it (apparently) names, the failure, in other words, for signifier to reach signified—the failure of the sign—establishes the place that Dover occupies in the text’ (1993, 147). But the word ‘Dover’ does name the place—it does not merely appear to do so—and that is the whole of its raison d’être. It is not the word’s job to ‘arrive at the place’: given that the word is an abstract object, that idea makes no sense in any case. And the word would not be of much use to us if it could only be successfully employed by someone located at Dover. (For a start, that very sentence would be nonsense.) Goldberg’s claim is a good example of ways in which critics find inadequacies in language on the basis of a wrong conception of what language is for. In this case the confusion is the old one of mixing up words and things. It is Lear and Gloucester who want to reach Dover; the word ‘Dover’ has no such ambitions. In the discussion of §43 I conceded that, in the context of the Fregean sense–reference distinction, there is a way in which senses can reasonably be thought of as ineffable. Senses are modes of presentation of referents; they are routes to their referents. The claim that I lodged in that earlier discussion was that these routes are unique, in the sense that, although we can refer to and talk about them, such language is not a substitute for actually travelling the routes in question. This point applied both to items of language themselves—so, for example, the sense of the word ‘Hesperus’ may be the sense of no other item of language—and to nonlinguistic, but semantically significant, items such as dramatic gestures, non-verbal noises, and silences, which may do what no words could do in the sense that, though again we can use words to name and describe the effects of these devices, that use of words cannot duplicate the effects that these devices achieve in their own right. Words can talk about and around Ajax’ silence in the underworld, and they can do this more or less helpfully, but they cannot substitute for that silence; they cannot achieve just what the silence itself, in context, achieves. It is far more effective for Ajax to say nothing to Odysseus than for him to utter the words ‘I have nothing to say to you, Odysseus’, and—this is the crux—irrespective of the point about relative effectiveness, his silence achieves something distinct from what those words achieve. That was the upshot of the earlier discussion, but my concern here is slightly different, or rather differently emphasized: here I am interested not in the main, negative, assertion of that earlier discussion, but in the parenthetical, positive concession made along the way, namely that words can talk about senses. Routes can themselves be destinations of other routes. So the claim that language can talk about everything is not compromised by our insistence on the uniqueness of senses. Senses are themselves things, so if language is to be omnicompetent it must be able to talk about those things; and it can. In

332  Tragedy and Linguistic Redress talking about a sense it does not become that sense; in driving down the feeder road to a motorway, you do not travel along the motorway. In that respect, as also by comparison with the case of music, there are things that language cannot do. But, just as we should not consider language’s competence to be impugned by the fact that the word ‘sad’ does not make you sad, so we should not think that the uniqueness of sense, as I have glossed this, derogates from the ability of language to express everything, in its own proprietary way.

47 Linguistic Idealism on Stage? Commenting on Racine’s dramatic technique, Steiner suggests that ‘As nothing of the content of Phèdre is exterior to the expressive form, to the language, the words come very near the condition of music, where content and form are identical’ (1961, 96; cf. 104). I have been stressing that words have semantic meaning, whereas music does not; that applies just as much to Racine’s words as to other words. Still, there is something right about the way Steiner characterizes French classical drama as peculiarly linguistic. Rousseau’s St Preux, unimpressed by the Parisian theatre, remarks in La Nouvelle Héloïse that ‘In general there is much talk and little action on the French stage’ (1967, 180; tr. 207). T. S. Eliot said something similar of Seneca: ‘the drama is all in the word, and the word has no further reality behind it’.23 One might say that Rousseau and Eliot are criticizing Seneca and Racine on the grounds that they embody a crude form of linguistic idealism, one that says ‘There is nothing but language’. A more favourable assessment of Seneca and Racine might credit them with expressing a more sophisticated version of the doctrine, along the lines I have explored, and according to which reality was in the words, as their internal accusative. Dramatic words should not be thought of as a façade, behind which either there is some reality, as in Shakespeare’s case perhaps, or there is not, as in Seneca’s perhaps. (And, after all, Seneca’s tragedies contain some finely realized characterization.)24 The words bring reality with them. One might then hazard the idea that some of Seneca’s, Corneille’s, and Racine’s plays are linguistic idealism in action: perhaps they are exactly making the point that the world is a precipitate of language. By presenting us with something which is both intensely linguistic and precipitates a reality, they are in effect saying ‘This is how things anyway are; the world is a deposit of language’. Steiner does not put it quite like that, but he does note that ‘All that happens, happens inside language. This is the special narrowness and grandeur of the French classic manner. With nothing but words—and formal, ceremonious words—at his disposal, Racine fills the stage with the uttermost of action’,25 and there follows the sentence that I quoted at the beginning of this section. Hippolyte’s death, for example, is not shown, but narrated; in Euripides, his death is shown, though his receiving the

Tragedy and Linguistic Idealism 333 fatal injury is narrated. In Greek tragic theatre, violence is usually narrated rather than shown.26 Narrated off-stage violence can take appalling forms, and Dodds suggests in his commentary on the Bacchae that the description of the off-stage pelting of the elevated Pentheus will strike modern spectators and readers as having a grotesqueness ‘below the dignity of tragedy’.27 But tragedy is more tolerant of the terrible and the grotesque than Dodds’s comment implies, and in any case the denouement of that particular scene in the Bacchae is considerably more harrowing than the earlier narration. Stage violence is, as Aristotle implied, commoner than is sometimes supposed,28 and we often see the consequences of violence on stage,29 but avoidance of overt violence on stage remains the rule, at least in extant ancient tragedy, until we come to Seneca, who breaks the Horatian injunction ne pueros coram populo Medea trucidet by having his Medea commit filicide on stage.30 In Euripides’ Hecuba, Polymestor wants to get his hands on Hecuba to tear her limb from limb after she has blinded him (off stage), but Agamemnon tells him to master himself and say it in words (1129–31). This is not just because, as Justina Gregory notes (ad loc.), ‘Polymestor must renounce force and submit to Greek judicial procedure’, but also because, metapoetically, he must submit to Greek tragic procedure. The Racinian stage not only lacks overt violence, but evinces a remarkable absence of ‘business’ of any kind, which is why Phèdre’s and Athalie’s and Bérénice’s acts of sitting that I mentioned earlier (§43) are so striking.31 But, though much that is allowed on the Greek stage and would also be allowed on the Elizabethan, Jacobean, and Romantic (not to mention more recent) stages is disallowed on the Racinian stage, because to show it would contravene the bienséances to which he so closely adheres—if Phèdre and Athalie and Bérénice are permitted to sit, like Shakespeare’s Duchess of York and Queen Elizabeth, or his Katherine of Aragon, you could not imagine a character in Racine groaning in pain like Sophocles’ Heracles and Philoctetes,32 or exclaiming ‘A plague upon it, I have forgot the map!’ like Shakespeare’s Hotspur, or smashing his reflection in the mirror like Klinger’s Guelfo, or uttering an obscenity like Goethe’s Götz, or asking for a glass of lemonade like Schiller’s Ferdinand von Walter, or killing himself onstage like Schiller’s Mortimer, not to mention the extravagances of Die Räuber or the bloody excesses of Shakespeare and Webster,33 and indeed the tolerance of the French theatre itself in the period between Garnier and Corneille34—still, everything is present in the words, in which we perceive, as Steiner finely observes, ‘intense force being channelled through narrow, complex apertures’.35 Corneille is somewhat less of a purist than Racine: his Le Cid and Médée, for instance, incorporate important stage business; he follows Euripidean and Senecan tradition by having his Médée appear at the end ‘en l’air dans un char tiré par deux dragons’ (1565), a ‘figure who is her own dea ex machina’.36 And some stage violence is permitted on

334  Tragedy and Linguistic Redress the Cornelian stage: we have not only the famous box on the ear in Le Cid,37 but the onstage deaths of Créon and Créuse in his Médée,38 not to mention the final suicide of Jason (and even Racine permits Atalide to commit suicide onstage at the end of Bajazet).39 But La Mort de Pompée is as austerely verbal as anything that Racine wrote.40 One might find the contrast between the intense verbality of La Mort de Pompée and its gruesome subject matter paradoxical; but would that not be to miss the point of linguistic idealism in action? Phèdre’s and Athalie’s and Bérénice’s surprising acts of sitting derive their significance, I suggested earlier, from their linguistic context. In that sense a kind of ‘double take’ operates. Initially we are impressed by the playwright’s apparent audacity in including a non-verbal act in an otherwise language-obsessed drama. But, though the act of sitting is non-verbal, the fact that its significance is supplied by the surrounding linguistic context softens the edges of the departure from convention. And all stage business, even in the most action-packed Shakespearean history play—the first part of Henry VI, as it might be—gets its significance from its linguistic context. So the tentative thought that I have offered, namely that the French classical stage offers a sort of theatrical realization of linguistic idealism, is perhaps too narrow. There is, to be sure, a dramatic difference between Racine and Shakespeare, along the lines we have mentioned, but it would be a mistake to cast this as a difference of philosophical significance. Fundamentally, Shakespeare is just as obsessed with language as Racine, and the fact that his characters are allowed to do things on stage that Racinian figures could, at best, describe, is of no consequence from a theoretical point of view. Racine, one might say, is trying to find out what can be achieved with minimum stage resources; this is an interesting experiment, but Racine’s success in it does not show that busier playwrights are less dependent on language than he is—only that they are differently dependent. One might as well say that Greek tragedy was linguistic idealism in action: the Athenian audiences were, after all, as Thucydides’ Cleon famously said, ‘spectators of words’ (III, 38, 4). But then we might as well say that language is linguistic idealism in action, which is just another way of saying that linguistic idealism is true. Steiner writes that in the great quarrel scene of Shakespeare’s Julius Caesar, Act IV, ‘the actual words have around them the resonance of the unspoken’, whereas in the parallel scene in Corneille’s Sertorius ‘all is said’ (1961, 69). And he generalizes this to: ‘In Shakespeare, the words in their complex groupings accumulate meanings in excess of the actual statement. In Corneille as in Dryden, they signify exactly what they say, but they signify the whole of it’ (ibid., 70). I suspect that Steiner has something like this point in mind when he dubs Racine ‘Cartesian’ (1961, 81, 84, 106; cf. 124): perhaps the thought is that Racinian words are associated with ‘clear and distinct ideas’. (I hope it is not the misbegotten

Tragedy and Linguistic Idealism 335 idea that Racine ‘invented the subject’: §32.) John Bayley writes that ‘A Racine tragedy is perfectly successful in the way it matches what happens with the right way of describing and reacting to what happens’, and he remarks of ‘such an exact beauty of fitting and responding’ that it ‘does not exist in King Lear’ (1981, 24–5). Blair Hoxby may be gesturing towards a similar point to Steiner’s when he says that ‘Poets such as Corneille, Racine, and Dryden often strive to make darkness transpicuous and the unutterable articulate’ (2015, 268). What Steiner, Bayley, and Hoxby all seem to be aiming at is the thought that Corneille, Racine, and Dryden try to lingualize everything in their plays, unlike Shakespeare; and then either one thinks that they are successful, perhaps because they draw in their horns and limit their ambitions, or one thinks that they are diminished by the comparison with Shakespeare, perhaps because they, unlike him, fail to convey a sense of the unutterable realm beyond words. But all this strikes me as fanciful, for two reasons: in the first place, there are unspoken resonances in Corneille and Racine as there are in Shakespeare, and what Steiner specifically has in mind in connection with the Shakespearean scene—the fact that Brutus is distracted from arguing properly with Cassius because he is grieving for Portia41—is made explicit in the words. A supposed contrast between cut-and-dried, fully explicit, resonance-free French drama (and Dryden), on the one hand, and messy, inexplicit, resonance-rich Shakespearean drama, on the other, will not stand up. In what way does Shakespeare not match what happens with the right way of describing and reacting to what happens? Or Schiller, for that matter, whose coup de théâtre in the meeting of the two queens in Maria Stuart is, as Steiner notes (1961, 183), comparable in its effect to the great quarrel scene of Julius Caesar. So the first point is that there are unspoken resonances in Corneille, Racine, and Dryden, as well as in Shakespeare. The second point is that while these resonances may be unspoken they are not unspeakable. The unspoken resonances do not emanate from a realm of the unutterable. That idea, as my linguistic idealism implies and as I shall explore further below (§49), is incoherent. Nor does the fact that French dramatists observe proprieties that English dramatists do not amount to as much as critics since the eighteenth century have wanted to make of it.42 Steiner suggests that Racine found the unities of time, place, and action congenial because they allowed him to exclude materiality. ‘The disorder of life, the material grossness of things, cannot be excluded from human affairs for more than twentyfour hours at a stretch. Even a Bérénice or a Phèdre must surrender to the vulgarity of sleep’ (ibid., 78). Actually, you cannot exclude the material grossness of things even for so much as twenty-four hours, at least not entirely. But the point of the unities is not to capture a serendipitous period during which none of the agents eat, drink, sleep, wash, excrete, and so on through all the ordinary biological functions, but only talk, and then preserve that miraculously in amber. Rather, as we have said, Racine

336  Tragedy and Linguistic Redress is simply following a traditional conception of propriety, according to which the performance of various necessary biological functions not only must not be shown on stage, but should not be mentioned either.43 The Greek tragedians too are, as Denys Page puts it, ‘remarkably fastidious in their use of words to denote parts of the human body’.44 Further, Steiner’s point about the absence of materiality in Racine can be overplayed.45 Even if biological functions are not explicitly mentioned, they are there in the background. In Bérénice, for example, although Titus and Bérénice do not consummate their love—Racine notes this point in his preface as being a crucial departure from the Virgilian model—the possibility of their doing so lies in the background: the drama, which is about ordinary human beings, depends on that fact. Titus and Bérénice are alive and well, physically speaking, so they must eat and drink, and so on; most crucially, they are attracted to one another, and Racine’s work makes sense on no other terms. Mitchell Greenberg’s view, that ‘Having done away with sexuality and death, the play presents us with purely ethereal heroes’ (2010, 126), seems to me quite mistaken. In the same way, although sexuality is not explicitly mentioned in Jane Austen’s novels, it is part of what, fundamentally, they concern: it is not as though Elizabeth and Darcy just want to be good friends. Racine’s characters, like Austen’s, are representations (as) of real people (§32). Samuel Hughes suggests that one reason for tragedy’s lack of emphasis on ordinary biological functions (he also thinks that tragedy ­de-emphasizes social role) might be to throw a stronger light on agents’ freedom and responsibility: insofar as ‘the viewer’s imaginative sense of the tragic hero is of a creature of physical needs and appetites, or one reducible to a social role, it precludes apprehending his freedom as a rational person’ (2015, 427). On the contrary, as I have already stressed (§7), physical and social contextuality not only does not undermine freedom; it constitutes its very possibility. Human physical needs and social roles are, in tragedy as in ordinary life, the necessary context in which freedom and responsibility are exercised. We tend to get distracted here by the philosopher’s fantasy of a locus of pure thought: were such an entity per impossibile to exist, it would not be free, because it could not act at all—could not even think. For freedom you need friction—­ physical and social.

48 Les nuits de Racine In §28 I talked about human universals. One universal human biological need is for sleep. In a production of Bérénice that I directed as a student I retained the unities, but set the dramatic time so that the stage action began on (what was represented to be) the evening of one day and ran through the night into the next day—rather in the manner of Ibsen’s Ghosts and indeed some ancient tragedies46—partly in order to hint at the physical importunities of sleep forgone. John Campbell describes

Tragedy and Linguistic Idealism 337 Bérénice as ‘in every sense a tragedy of time’,47 and we are familiar with the crucial role played in Racine by such phrases as ‘dès ce jour’, ‘avant la fin du jour’, and ‘jusqu’ à ce jour’. These phrases are drawn from the Sophoclean and Euripidean ‘tēide en hēmerai’, ‘on this very day’, and similar,48 and they indicate the critical time when a decision has to be made, or a debt has to be called in—like the clock at the end of Marlowe’s Doctor Faustus,49 or the ship’s bell at the end of Ibsen’s The Lady from the Sea, or the increasingly insistent refrain of the pub landlord in Part II of T. S. Eliot’s The Waste Land. The mention of day in these phrases is generally unmarked, so that they simply amount to ‘at this point in time’, ‘from this time forth’, or ‘during this period of twenty-four hours’, and so on:50 we are not usually talking in the first instance about day as opposed to night, but rather about this day as opposed to an earlier or later day (tomorrow being, so often in tragedy, too late).51 Still, the fact is that day is mentioned, which suggests the contrasting idea of night. And night is indeed important in Racine, especially as a time when sleep can be displaced. At first, in Bérénice, everything seems to go well with this displacement (301–6): De cette nuit, Phénice, as-tu vu la splendeur? Tes yeux ne sont-ils pas tout pleins de sa grandeur? Ces flambeaux, ce bûcher, cette nuit enflammée, Ces aigles, ces faisceaux, ce peuple, cette armée, Cette foule de rois, ces consuls, ce sénat, Qui tous de mon amant empruntaient leur éclat. (‘The splendour of that night did you behold? / Are not your eyes full of his majesty? / This pyre, these torches, this inflamèd night, / These eagles, fasces, ­ orrowing people, army, and / This host of kings, consuls and senators, / B their radiance from my beloved’; tr. Cairncross.)52 But nights in Racine are ominous. Night misleads Jocaste in La Thébaïde (679). The night of Iphigénie’s opening is full of foreboding; Agamemnon has endured nightly anguish since his rejection of Calchas’ behest to sacrifice Iphigénie (83–8); Achille arrives at night (110). Night is when, in Britannicus, Néron first sees Julie (386), and it is when Julie fears that Néron will strike against her and Britannicus (1543–4). Night is when, in Andromaque, Oreste proposes that he carry out Hermione’s demand that he kill Pyrrhus (1213); she insists that he do it at once, by day, but when she turns on him later, disavowing her instigation, the night of madness descends on Oreste (1625), and he fears the onset of eternal night (1640), just as Phèdre is terrified of fleeing into ‘la nuit infernale’ where her father Minos judges the dead (1277–80). Andromaque recollects how Troy fell at night (997–8): Songe, songe, Céphise, à cette nuit cruelle Qui fut pour tout un peuple une nuit éternelle.

338  Tragedy and Linguistic Redress (‘Cephisa, call to mind that cruel night / Which was perpetual night for all our race’; tr. Cairncross.)53 The plot of Andromaque is based on Aeneas’ tale to Dido in the third book of the Aeneid, but the quoted lines recall an earlier part of that narrative, from the second book, which told of the last night of Troy, whose destruction is ‘the very heart of a fire that has burned throughout history’, ‘stuck in our collective memory’, ‘perhaps the greatest secular story of the Western world’, ‘the archetypical World War’.54 It is the founding myth of European humanist culture. And that Troy fell at night is a fundamental element of the tale.55 Notice the superb irony, in the passage just quoted, of the repeated use of ‘songer’, which literally means dream: for on their last night the Trojans slept and dreamt, but should not have done. Night is the right time for sleep and dreams, at least usually, for most people, but not that night—‘that last, terrible night’56—for them. Night in Senecan tragedy is (paradoxically, since this word is normally applied to day) alma—kindly, cherishing— because it brings relief from toil.57 The deep refreshing daytime sleep of the maenads in Euripides’ Bacchae is sinister (692, where ‘thaleros hupnos’ equates to ‘alma quies’):58 we are meant to think that this is not the right time for sleep. Sleep is ominous in the Oresteia, until its ominousness is finally put to rest, as it were, in the Eumenides (705).59 ‘A curse on that night and its fate!’ sings the chorus of Euripides’ Trojan Women;60 and in the wonderful third stasimon of the Hecuba (905–51), a captive Trojan woman relates how the city met its end at midnight, as she was arranging her hair, staring into the unfathomable depths of a golden mirror while her husband slept on their bed, his lance hanging inertly from its peg (914–27).61 In his account of Troy’s last night, Virgil’s Aeneas admits that he shared the insouciance of his countrymen (II, 250–3): Vertitur interea caelum et ruit Oceano nox involvens umbra magna terramque polumque Myrmidonumque dolos; fusi per moenia Teucri conticuere; sopor fessos complectitur artus. (‘Meanwhile the heaven turns and night rushes up from Ocean, wreathing in its enormous shade earth and sky and the treacheries of the Myrmidons. Scattered through the city the Trojans fall silent; sleep embraces their weary limbs’.) In his ill-taken sleep Aeneas dreams of Hector, Andromache’s dead husband and Troy’s lost military hope; as he dreams he seems to accost the slain hero, but in his thoughtlessness and forgetfulness, emblematic of the Trojans’ folly and hamartia, Aeneas speaks to Hector as though he were still alive (281–3): o lux Dardaniae, spes o fidissima Teucrum, quae tantae tenuere morae? quibus Hector ab oris expectate venis?

Tragedy and Linguistic Idealism 339 (‘O light of Dardania, O most faithful hope of the Trojans, what has kept you from us for so long? From what shores, Hector, long awaited, do you come?’) For it is that time ‘between midnight and dawn, when the past is all deception’62— Deep night, dark night, the silent of the night, The time of night when Troy was set on fire, The time when screech-owls cry, and ban-dogs howl, And spirits walk, and ghosts break up their graves63— that time of night when, as the prophetic Cassandra knows, the murdered Agamemnon will be buried in hugger-mugger,64 that time when Clytemnestra awakes screaming, distraught from her nightmare vision of a serpent, and calling for braziers to be lit:65 ‘The lights blaze up for the sake of the mistress, but darkness is about to come upon her. It is for Orestes that the light is now to shine’.66 In the passage quoted above from Bérénice we sense that there is something wrong, something unnatural, in the night that Bérénice found so electrifying: she thinks she is secure in Titus’ love, as Dido thought she was secure in Aeneas’ love (Aeneid IV, 291–2), but all the while she knows nothing of his agonies and doubts. The splendid night has betrayed her— as it betrayed Dido, as it betrayed the Trojans. The Trojans are betrayed while they sleep, Dido and Bérénice while they watch. When day comes to the sacked and smoking ruins of Troy, its people, those that survive, are betrayed again: they had forfeited one of their sons, Tithonus, to be consort of Aurora, goddess of the dawn; she eagerly snatched him aloft in a chariot of stars, but gives the city nothing in return, looking on its fate with indifference.67 For gods desert defeated heroes and cities.68 In the transition from the second book of the Aeneid to the fourth—perhaps in the depths of the ‘timeless night’ (nox intempesta: III, 587)69 that the Trojans pass beneath Etna’s troubled mass as Enceladus turns his weary side—the nocturnally betrayed become nocturnal betrayers, and there is a sense in which the two nights merge.70 Aeneas’ tale to Dido of nighttime betrayal is itself told at night, when speaker and hearer should be asleep,71 and it is then that she fatally falls in love with him. In Macbeth, the night that betrays Duncan, who sleeps, also, less immediately but no less really, betrays the regicides, who are watchers. Marlowe’s Faustus summons Mephostophilis at night, and is betrayed by the midnight bell.72 I have been meditating on the palpability of night, in Racine and other tragedians. But there is also a sense in which Racine’s language heightens and throws emphasis on the material grossness of things, just by virtue of not mentioning it explicitly. This is not, I think, because, as Steiner suggests, we suspect that, as soon as they are safely off stage, the characters will ‘release their pent-up agony’ (1961, 78)—as though we were to imagine operatic characters saying to each other, in the wings, ‘Thank

340  Tragedy and Linguistic Redress God that awful singing is over; now we can talk like normal people’. The performers might say that, but the characters do not; there is no ‘off stage’ for the characters (unless the characters are themselves actors, and it is an opera about putting on an opera, like Richard Strauss’s Ariadne auf Naxos). In opera, all life is, as I have said (§44), mapped into music, so that there is no scope for the thought that, in other circumstances, one might be released from one’s confinement to the musical element. Rather, everything—even silence and deafness, as we mentioned earlier, even an existence without music—is expressed in musical terms in opera: an operatic character who complained, truly, that his life lacked music would sing that complaint to orchestral accompaniment, and there would be no paradox (though there would, as we noted, be a metatheatrical frisson), because in the represented world—as opposed to the representing medium—his life would lack music. There is, in opera, no question of confinement to the musical element. Similarly, in Racine actions are represented in words, rather than performed on stage. It is not that Bérénice leaves the stage at the end of her play for post-dramatic ‘screaming or weeping’, as Steiner suggests (ibid., 78). Her screaming and weeping take place on stage, but they are mapped, as we have said, into the proprieties of the neoclassical drama, where they show in purely verbal, and polite, form. The exceptions to this procedure are all the more impressive when they occur, just as Mimi’s consumptive stage coughing, being a breach of operatic convention, according to which coughing should be sung, or sung about, is all the more affecting.73 The intensely linguistic nature of Racine’s drama throws the material grossness of things into strong relief because we feel its absence. Here one may be reminded of a famous remark that Wittgenstein made (1980, 78) in a letter to Engelmann in 1917: ‘If one does not attempt to express the inexpressible, then nothing gets lost. Rather, the inexpressible is— ­inexpressibly—contained in what is expressed’. If one restricts the domain of the modality in the word ‘expressible’ (‘aussprechlich’) so that it means, in effect, ‘expressible within the parameters set by the classical doctrine of propriety’, and correspondingly for ‘inexpressible’, then Wittgenstein’s remark applies very neatly to Racine. Even so, in an absolute sense there is nothing inexpressible about the things that are not expressed by the language of Racine’s dramas. Other pieces of language can express those things perfectly well, though naturally they do so in their own way. When he made the quoted remark Wittgenstein had his picture theory of meaning in mind, along with the saying–showing distinction of the Tractatus, which propounds the doctrine that there are things of a mystical nature that scientific language—which for the early Wittgenstein is the only kind of language that makes sense—shows but does not say. But, abstracting from the peculiar metaphysics of the Tractatus, we may affirm that when language does not express something, a thing that is then felt as an absence, this thing can be expressed in language; it is not that the thing is

Tragedy and Linguistic Idealism 341 inexpressible, just that it has not on this occasion been expressed, though as I have agreed the way of expressing the thing in language is different from the manner of its expression by silence or omission. Steiner is right that syntax itself bears witness to the fact that ‘there is a there there’ (1997, 64), for words are the child of syntax, and objects are just what words refer to—so that, as we said at the beginning of this chapter, the world itself is just a precipitate of ­language—but he is wrong to add that this there is beyond language in any sense that signals a ‘falling short’ of language. Unless he really does just mean that language is not the reality it talks about; mystics might not be happy with that anodyne formulation, but it is all that can be coherently extracted from talk about the limits of language.

49 Saying the ‘Unsayable’ When Bottom recovers from his transformation in Shakespeare’s A Midsummer Night’s Dream, he tells us (IV, 1, 200ff.): I have had a most rare vision. I have had a dream past the wit of man to say what dream it was. Man is but an ass if he go about t’expound this dream. Methought I was—there is no man can tell what. Methought I was, and methought I had—but man is but a patched fool if he will offer to say what methought I had. The eye of man hath not heard, the ear of man hath not seen, man’s hand is not able to taste, his tongue to conceive, nor his heart to report what my dream was. I will get Peter Quince to write a ballad of this dream. In his book The Philosophy of the Unsayable, William Franke remarks that, ‘As is typical of those who speak about what cannot be said, Bottom cannot keep it short. He stammers on. He says over and over again . . . what he cannot say’, adding that ‘There is endlessly much to say about this experience of inadequacy vis-à-vis the unsayable and miraculous, and precisely this verbiage constitutes perhaps its only possible expression’.74 How so? In fact the point is that Bottom is trying and failing to express what is not inexpressible at all, but clunkingly and comically expressible. Bottom was given a donkey’s head and acquired some of a donkey’s tastes, while—like many classical subjects of metamorphosis, such as Circe’s victims75—retaining much of his old mind, and indeed most of his body. ‘All that Robin has done, as modern productions seem inclined to forget, is to put an ass’s head on Bottom; the rest of him stays irredeemably human’.76 It is all perfectly sayable, and the satire is on the hempen homespun for failing to say it. ‘Methought I was— . . . methought I had—’: it is easy enough to complete Bottom’s false starts. His continuation (‘The eye of man hath not heard. . .’), with its parody of St Paul,77 might suggest ineffability, but only because he is mixing

342  Tragedy and Linguistic Redress up sense modalities, in a way which parodies his erstwhile confusion of human and donkey natures, not because there is something there which we feel language ought to be able to express but is unable. Of course you cannot express what the eye of man hears, because that collocation of concepts makes no sense. It is no derogation from language that it cannot express the inexpressible in that sense. The joke is on Bottom, but the audience’s laughter ‘is curiously tender and even loving’.78 Shortly after the remarks quoted above, Franke argues that the long description of Moby Dick’s whiteness is indicative of its ineffability, whereas I take it, as I have said (§37), to point to the opposite conclusion. It is interesting that, on the front cover and spine of Franke’s book, the word ‘unsayable’ is crossed out, Derrida-fashion, but on the title page inside it is not: when you look into it, I suggest, the ‘unsayable’ turns out to be sayable. Peter Quince can write a ballad about it—especially if, as Helen Cooper speculates (2010, 214), Quince is a front for Shakespeare himself. Indeed, Franke has written a whole book about the unsayable, an irony of which he seems unaware.79 Franke’s reason for thinking that there is such a thing as the unsayable is based on his understanding of the difference between individual and universal, concrete and abstract. Here is a representative passage (2014, 22): The concrete as such is infinitely dense and is never adequately expressed. Only its relatively abstract form can be stated in language. And such distinctions are all conceptual creations that purport to be based on realities. The real, however, stripped of all conceptual determination, cannot as such be grasped or said. In the end we make discriminations, such as that between abstract and concrete, on the basis of a judgement that can never be fully justified in words. We see here the connection, which has already surfaced in our discussion (especially in §42), between the idea of the inexpressible and the Kantian supersensible—the idea of a thing-in-itself, a thing shorn of all its properties. The idea of the unsayable, for Franke, is ultimately the idea of the real so stripped down. But why should one suppose that anything at all— even something unsayable—answers to the idea of the real when it is so conceived? What you obtain when you deprive an object of its properties is not a something, not even (or perhaps I should write: especially not) an unsayable something, but just nothing.80 And it is no inadequacy on language’s part if it does not, and cannot, say anything about that nothing. Franke agrees that the unsayable ‘thing’, the object stripped of all its properties, is not an object: ‘The unsayable is not an object at all. To be an object is to be within the purview of some framework for perception or conception’ (ibid., 60; cf. 74). But if it is not an object, how does it still get to be a thing—a thing that we can (apparently) talk about and

Tragedy and Linguistic Idealism 343 describe truly in words, for example when we say that it is an unsayable thing? Things are objects. And anyway as soon as we have called the thing an unsayable thing, we have made it sayable. Notwithstanding these difficulties, Franke is confident that ‘literary forms can . . . enable us somehow to “see” [the unsayable], to give it a form’ (ibid., 60), and that despite the very evident fact that literature traffics in words and concepts. The motivation for Franke’s (as for Nietzsche’s) animus against the adequacy of language is largely derived from thoroughly traditional empiricist and nominalist sympathies. The idea is that, metaphysically, we start with the concrete thing in all its individual and material glory, and then, in order to talk about it, engage in a process of conceptualization that involves an abstraction that supposedly elides the thing’s reality: ‘Real individuals are absolutely singular and other’, and so ‘cannot be expressed in their uniqueness by language, which inevitably generalizes’ (ibid., 134; cf. 99). So the suggestion is that in abstracting we misrepresent the original object: a distorting rift is alleged to open up between the individual object, endowed with particularity, and our conceptual way of approaching it; the unique thing itself escapes the net of our generalities. But this idea is entirely fantastic: the object is already, at the metaphysically most basic level, a thing of a certain sort, equipped with at least essential and usually also accidental properties. So it already exemplifies general properties—it already falls under universal concepts—and language therefore has no task of abstraction to carry out. Language merely has to notice (refer to) those properties. In talking about an object in our conceptual language, we do not distort its nature, but convey its nature, if we speak truth. As that last proviso hints, the only possibility of a distortion’s opening up between our conceptual language and the world resides in the threat of falsity. But false language is still, necessarily, meaningful language: it misses the truth, but it does not miss the world. This point was made by Wittgenstein in one of his most beautiful and profound aphorisms: The agreement, the harmony, of thought and reality consists in this: if I say falsely that something is red, then, for all that, it isn’t red [es doch immerhin nicht rot ist]. And when I want to explain the word ‘red’ to someone, in the sentence ‘That is not red’, I do it by pointing to something red.81 If I say falsely that an object is red, I have still achieved something substantial: my sentence is merely false, not nonsense. The concept I have wrongly predicated of the object is a bona fide concept. And the object itself is not red: it has got that far. The property it does not instantiate is redness, nothing less than that. A highly traditional form of empiricism survives in many recent theorists of literature: Franke is very far from being alone. Dennis Schmidt

344  Tragedy and Linguistic Redress tells us that ‘Aristotle is not troubled by the simple fact that his analysis of mimetic practices is carried on in a thoroughly conceptual language. In other words, he is not troubled by his own translation of the language of art into the language of philosophy, and there is no discussion of what might be lost in that translation’.82 Certainly Aristotle’s language is thoroughly conceptual: all language is. But the world is thoroughly conceptual, too. So what is it that gets ‘lost in translation’? Here the m ­ etaphor—for it is a metaphor—of translation is quite misleading. Strictly, we translate one language into another: in other words, the relation between translated and translating languages is horizontal; it joins things at the same ontological and categorial level. So verbal appraisal of graphic art, say, is not translation: for graphic art is not, as such, linguistic. Even if we confine the discussion to literature, literary works are not translated by critics and interpreters of them; they are translated by translators (from English into French, as it might be). The relation between critical discourse and the work of art that it is about is not horizontal but vertical: it is a matter of language’s being about the world; it is just that in this case the bit of the world that we are interested in is itself linguistic in form. Schmidt’s confusion, as I take it to be, emerges at a later point where he precisifies his criticism of Aristotle: ‘After Aristotle, the development of philosophy leads increasingly to the view that the truth of language is found in its conceptual possibilities, not in its capacity to present what we suffer’ (2001, 274–5). Again we encounter the thought that language could only ‘present what we suffer’ by giving us the suffering, by making us suffer, by closing the gap between language and world which is opened by the aboutness relation and becoming an episode of suffering itself. To ask that of language, as I have in effect argued, is to ask that it stop being language and be something else—a piece of the world. But it is language’s business not to be a piece of the world, but to be about pieces of the world. (Or rather, putting in the necessary qualification, since language is itself a piece of the world, the point is that it is not language’s business, at least not usually, to be the piece of the world which it is about.) And language has to stay conceptual—it has no choice about that in any case—in order to talk about suffering, because suffering is itself, like everything else in the world, conceptual—that is, it is a thing (or many things) of a certain sort (of certain sorts), and sorts are properties, which is exactly to say that they are conceptual in nature. In Die Geburt der Tragödie (§16), Nietzsche expressed the idea that, just as ordinary conceptual language involves an abstraction from concrete things, so, at the next level up, as it were, music effects a further abstraction from the conceptuality of language (1999a, 105). ‘Looked at in this way,’ Schmidt says, ‘music appears to be a sort of superlanguage, . . . a superior language which is more universal even than the universality which belongs to concepts’ (2001, 213). And Nietzsche also seems to have thought that this super-universality of music fitted it to express pain, in a way that ordinary language could not.83 Pain is here

Tragedy and Linguistic Idealism 345 thought of as a thing-in-itself, which language has to go beyond conceptuality to express. But I have already rejected both the idea that pain is a thing-in-itself and the idea that music is a language; the suggestion that music might be a kind of superlanguage, generalizing beyond the generalities of concepts, makes as little sense. If you take a concept and perform a generalizing or abstracting operation on it, you simply move up Porphyry’s tree: what you get is a more general concept, one that stands to the original concept in the relation genus–species; you do not move beyond the conceptual to something non-conceptual. I have agreed that there is a sense in which music is able to express some things that language cannot: that is the sense in which a particular piece of music may give you a feeling which language can talk about, but not give you. (Language can give you other feelings.) This difference, I have suggested, should not be registered as an inadequacy on the part of language, because it would be a confusion to suppose that language has as its purpose to transmit all the feelings there are; language does what it is supposed to do, which is to say things about all the feelings there are, and everything else. Let me here resume and embellish a point that I broached in §38. I have found that a common reaction to the line of argument that I have been giving in this chapter and the last is say something like the following: ‘Yes, there is a sense in which language can express reality; it can indeed talk about, refer to, and describe even the most recalcitrant parts of reality, such as pain and suffering. But it cannot do so adequately. There is always (and especially in the case of pain) something missing’. If one says that language is not adequate to reality, one might mean one of two things. One might mean that language fails to express reality tout court. This is not the direction of my imagined objection, and it is clearly a non-starter as a philosophical position. Quite apart from the fact that such a claim would be self-refuting—because the statement is advanced as being true, and all true statements express reality—there is the point, which I mentioned above, that falsity is the worst that can befall most ordinary uses of language, and falsity is a kind of success. If I say ‘the cat is on the mat’, in what way(s) can my statement fail? So long as the statement is meaningful—so long as it is composed of recognized lexical components of the language and constructed according to that language’s grammar, so long as its indexical components are suitably grounded, and so on—the worst outcome one can expect is that the statement is false. And if it is false it is about reality; it could not be false (let alone true) if it were not. So false statements, as well as true, express reality: that was the point of the Wittgenstein aphorism quoted above. Alternatively, one might mean, as my objector seems to mean, that language does express reality, but not ‘really’, or only ‘partially’, or something of the sort. It was this idea that Wittgenstein attacked when he wrote, again gnomically: When . . . we disapprove of the expressions of ordinary language (which are after all performing their office), we have got a picture

346  Tragedy and Linguistic Redress in our heads which conflicts with the picture of our ordinary way of speaking. And then we are tempted to say that our way of speaking does not describe the facts as they really are. As if, for example, the sentence ‘he has pains’ could be false in some other way than by that person’s not having pains. As if the form of expression were saying something false even when the statement faute de mieux asserted something true.84 As if the sentence ‘He is in pain’ could not, even if true, be an adequate description of reality. Here I repeat the crucial point from my earlier discussion: the complaint that ‘He is in pain’ is, even if true, an inadequate description of reality only makes sense if another form of words would do the job better. What does not make sense is to refuse any possible form of words as adequate. For against what standard are you judging that ‘He is in pain’ is inadequate? To call something inadequate is to have a conception of what would count as adequacy; to call something a bad shot is to be able to answer the question what a good shot would be. But those who complain about the adequacy of language as such—of any piece of any (actual or possible) language—to express pain (or anything else) have no place to stand in making their criticism. They have taken criticism beyond the territory within which alone criticism makes sense. When Franke talks of ‘the radical inadequacy of all our language vis-à-vis the real’ (2014, 65), where does he suppose he is located? The answer can only be that, absurdly, he is attempting to stand at a point located outside both reality and language, and to view them sideways on: here is reality over on my left, and there over on my right we have all of language; now let us compare them to see how good a fit the latter is for the former. But it is obvious that there is no such external location. For the linguistic idealist, the essential point can be expressed as follows. Since the world is the metaphysical product or precipitate of language, reality just is what language produces. Language cannot do its job of creating the world badly, for there is no standard against which to measure relative success or failure. We are stuck with what language gives us: we have—can have—no conception of anything better (or worse). That yields a direct route to the conclusion that if one form of words is inadequate, there must be another form of words that is not. As I have said, I diagnose any residual feeling that language is inadequate as evincing a disappointment that language is different from reality, that the word ‘pain’ is not painful and does not convey an experience of pain, or the word ‘sad’ of sadness, and so on. But this feature of language is actually a strength, not a weakness: language has to be different from reality in order to be about it: without that gap, there would be no such things as truth and falsity, or meaningful discourse, at all.85 Although there is this gap between language and reality, the fact that reality is essentially expressible in language entails that the

Tragedy and Linguistic Idealism 347 gap can always be crossed. Crossed: not closed. The referential relation is always there to be had, but it is a relation between two things which (usually) subsist at different ontological levels; it is a relation between something linguistic and (usually) something non-linguistic. It might be—has been—objected to me: how can the mistake that I take myself to have uncovered reside in a forgetting of the fact that the word ‘pain’ and the phenomenon of pain are two different things—that language is one thing and the reality it is about another? That a word and its referent are two different things is, one might think, too obvious, too superficial a point to account for the deep feeling that people have that words are inadequate to express reality and especially the reality of suffering. Surely, you might say, if people are making a mistake here, it cannot be simply that they are confusing language and reality in the way I propose? Does that not trivialize the error? Well, I think that this is indeed the mistake: but so far from its being trivial or superficial, it is actually a deep mistake, and the linguistic idealist can explain that depth. (Note that deep philosophical mistakes often look trivial, even stupid, when they are laid bare: relevantly in the present context, the perennially popular idea that names go proxy for things is a case in point.) According to the linguistic idealist the world is essentially expressible in language, and this in turn means that the world is propositionally structured, indeed is composed, in the first instance, of propositions. There are, as I have agreed, ordinary things in the world, things of all sorts— tables and chairs and properties and events and numbers and so on—but these things get into the world by virtue of their figuring in propositions, true and false. So the world, though not itself linguistic—for propositions are not sentences, but meanings (referents) of sentences—is structured in a way that reflects linguistic structure. It is a mirror image of language, and that is a transcendentally necessary truth. There is a gap between language and world, between words and sentences, on the one hand, and their referents, namely objects (of all sorts) and propositions, on the other; but the propositions that make up the world (and a fortiori the objects that compose these propositions) fall out of, are the precipitate of, language. The first of these points dissolves the confusion; for we see why the gap, though it may be mistaken for an inadequacy, is in fact no such thing. And the second point explains why the confusion arises in the first place: the world is not language, but it is close to language, and that is why the mistake of confusing the one with the other is so easily made, and not a facile one.

50 The Liberating Word There are two species of tragic outcome. In the first place we have the kind of plot that presents a ‘moral theodicy’ in the sense that the heroes are portrayed as getting their just deserts. Their fate arises in a

348  Tragedy and Linguistic Redress comprehensible, explicable, predictable, and justifiable way from their character and actions, and in particular from cognitive failure. In the first part of this book I argued that recent critics, unlike older critics, have underestimated the extent to which classical tragedies, and tragedies influenced by the classics, are formed on this model. Such great exemplars of the genre as the Oedipus Rex and the Antigone fall firmly within it. And I have considered many other cases, with different degrees of argumentative detail. One case I have not considered, though it figured in the list of examples I gave early on (§6) of exemplary tragic heroes exemplifying Aristotle’s doctrine of hamartia, is Shakespeare’s Hamlet. In the context of feminist attempts to rescue the character of Gertrude, Lisa Jardine asks why we seem to feel the need, ‘in order to keep our sympathy for the hero (to sustain the tragedy)’, to ‘shift the burden of blame from Hamlet [sc. onto Gertrude] and make him blameless’ (1996, 149). I agree with Jardine that Hamlet is not blameless, but it does not follow, for me, that Gertrude is therefore blameless. As I have stressed, we should avoid any such simplistic search for a single hamartia in a given tragedy: tragedies exhibit multiple hamartiai, often committed by more than one person (both Antigone and Creon, for example, are at fault), and it is even possible (as I suggested in the case of Oedipus), that a single act may comprise more than one hamartia. But Jardine is right that audiences feel a need to apportion blame on the principle that, if things go wrong, someone must be responsible. Whatever we think of this principle in ­general—obviously it is, stated baldly like that, false—it is, I have argued, a principle that guides most traditional tragic literature most of the time. One commonly finds the view expressed that, although audiences feel the need for moral redress, the greatest tragedians rise above that primitive attitude and show us uncompensated suffering, that is, suffering not balanced by any kind of desert.86 I argued in the first part of this book that, at least as far as the protagonists of tragedy go, this widespread view is mistaken. Morally speaking, tragedians and their audiences are at one; and that is, after all, exactly what you would expect. Recent critics have overemphasized the importance of what Steiner calls ‘absolute tragedy’ or ‘tragedy pure and simple’, that is to say, of tragic action in which a fate that is undeserved befalls an agent. But tragedy does exemplify undeserved suffering, and this is the second kind of tragic outcome. Many tragedies illustrate both sorts of denouement: thus Aeschylus’ Agamemnon is punished with death for the sacrifice of his daughter Iphigeneia, an act which he undertook freely and with full responsibility, whereas Cassandra is slaughtered because she is Agamemnon’s concubine, a status which the sack of Troy conferred on her in her innocence and powerlessness; or Sophocles’ Creon is punished for his intransigence towards Antigone, but his wife Eurydice, who is without sin, takes her own life in grief. But I have suggested that, where undeserved suffering does overtake tragic

Tragedy and Linguistic Idealism 349 figures, such figures are not protagonists of their dramas, and in some cases cannot even be regarded as agents. To the extent that tragic figures exhibit genuine agency, we find that, like Blanche in Poulenc’s Dialogues des Carmélites, they conform to my standard model, which has it that suffering tracks hamartia. Discussing the possibility of Athalie’s eventual redemption, Richard Parish argues that any such redemption would be ‘incompatible with the later Racinian understanding of tragedy’ on the grounds that ‘redemption makes sense of human suffering; tragedy does not’ (1993, 41). Perhaps the point is that religious redemption (which is what is in question here) would be too easy, too superficial. That is, indeed, the reason why, as I observed in the Introduction, so many commentators have taken exception to the idea of Christian (or, in general, religious) tragedy: for tragedy, it is felt, is a deep genre, and so cannot tolerate a denouement of divine forgiveness. I have agreed with that. Nevertheless, I have argued that tragedy is a redemptive genre in a secular sense, namely morally and linguistically. In genuine tragedy—tragedy as I defined it in the ­Introduction—there is no final deus ex machina, no solving divine intervention, that makes everything all right, but there is a deeper sense of recompense, of redress, for suffering, conveyed by our understanding of cause and effect and so of moral responsibility, and conveyed by the nature of language as adequate to the world and celebratory of human existence; these things give us a measure of cognitive control over our lot. Bonnie Honig distinguishes between what she calls ‘humanist’ and ‘anti-humanist’ readings of tragedy, arguing that (2013, 151–2) For humanists, tragedy performs the paradoxically impossible when the art form makes meaning out of man’s insignificance. For antihumanists, tragedy is the non-redemptive genre that explores human ambition, desire, or compulsion but then confronts the protagonists (and the audience) with the inevitable demise that destroys human illusions of grandiosity. For me this mixes up the categories. I have defended a version of the redemptive view of tragedy, but my humanist takes a strongly moralizing line (which, as I have noted, need not be a moralistic one, in the pejorative sense of the word): the sufferings of protagonists are caused by their own mistakes and especially by cognitive failure. To that extent tragedy, as Nietzsche said,87 celebrates man’s power, not his insignificance: the same idea was also forcefully advanced by Bertrand Russell in ‘A Free Man’s Worship’ (1986, 9–19). Tragedy says, in effect: here are the mistakes that we can and do make; but mistakes, though we seem to keep on making them with depressing consistency, are not actually inevitable; in particular, there is a remedy for cognitive failure—learn more, know more, think more; know more about other people and how they work;

350  Tragedy and Linguistic Redress know more about yourself and how you work. The moralizing approach to tragedy also sets me in opposition to Sebastian Gardner, who writes that (2003, 236): The conflict [of tragedy] with morality consists . . . in the opposition between the necessary tendency of morality to reshape the world in a manner that consolidates the reality of morality, and the implication of tragedy that such a moral theodicy is impossible. This conflict emerges from the ‘negative’ moment of tragedy—its representation of loss and suffering as the deepest facts about the world. Against this, I have argued that tragedy, by and large, goes along with the doctrine of moral theodicy rather than resisting it. That suggests to me that, as far as what Gardner calls the ‘negative’ moment of tragedy is concerned—that is, as far as tragedy ‘pure and simple’ is concerned, where the suffering is represented as wholly undeserved—he may also be wrong to claim that tragedy represents such loss and suffering as ‘the deepest facts about the world’. For if that were so, would we not expect the representation of the absence of moral desert to be much more prominent and pervasive in the Western tragic tradition than it is? In particular, would we not expect the great protagonists of tragedy to be represented as exemplifying ‘tragedy pure and simple’—punishment without desert? But that is not how things are. The deepest fact about the world that we find expounded in tragic literature is, I propose, not a moral, but a linguistic one, and it is that everything, including suffering, can be presented in language. Language has the last word; it speaks das erlösende Wort—the liberating word.88 In §36 I called the encompassing of tragic suffering in language a kind of linguistic theodicy, and I spoke, though I hope with due reserve and caution, of the celebratory aspect of this theodicy. Reserve and caution are necessary, because of a fact that I have noted in §38, namely that insisting, as I do insist, on the expressibility of real-life and stage tragedy in language does not diminish the misery of human suffering. But the expressibility of cruelty (for example) in language is, as I have said, part of what entitles us to repudiate that cruelty. Now there might seem to be a tension between this point and the thesis that putting things, including suffering, into words has a celebratory aspect. But in fact there is no conflict: the latter point operates at a more abstract and general level than the former. Particular descriptions of suffering and cruelty enhance, rather than diminish, our horror of these things, and justify our moral response; but it is the fact that these things can be put into words at all that is liberating. For tragic suffering can indeed be put into adequate words. These are not words, as I have stressed, that themselves induce in the spectator or reader anything like the experience of undergoing that suffering. But they do what words are there to do: they talk about things,

Tragedy and Linguistic Idealism 351 including suffering, and they may do that in a way which, in the hands of a great artist, leaves little or nothing more to be said. Everything can be said, and sometimes, in some demarcated domains, everything—or as good as everything, or everything that matters—is said. Charles Segal suggests that ‘we have tragic art that we may not forget the dimensions of life that our structures cannot encompass. Tragedy pushes back the structures and reopens the painful possibility of seeing life as chaos’ (1981, 42). I believe, and hope to have demonstrated in this book, that this is the reverse of the truth. Tragedy does not show us ‘dimensions of life that our structures cannot encompass’. On the contrary, in depicting guilt and pain, it encompasses these things. It tries, so far as possible, to redress suffering by making it a consequence of guilt, and it also gives us a measure of control over and understanding of suffering by putting grief into words. Tragedy does not leave life in a chaotic state, but orders it; it makes it cognitively and linguistically accessible. Language as such is ordered, and gives pleasure: ‘Epic song tells stories of violent action; but when the stories are told, a tranquillity descends upon the folk. The bard sings of sorrow and death, but his songs give pleasure. The themes of heroic songs are themes of ruin and disorder, but the song itself is an ordered thing, and heard as such’.89 Even the most harrowing account of man’s inhumanity to man, even Müller’s eyewitness report of his activities in the Sonderkommando at Auschwitz-Birkenau, or Améry’s account of his torture in the bunker at Breendonk, even such things as these give pleasure of a sort: there is a narrative, we learn things from it, we sense that the insight it gives us into human nature is absolutely indispensable, and the sheer casting of the material in linguistic form is satisfying to reader as to writer.90 But for a linguistic idealist this seeming paradox—that is, the paradox residing in the undeniable fact that reading Müller’s book is pleasurable—is at least partially dissolved by the reminder that there is a trap lurking in phrases like ‘the sheer casting of the material in linguistic form’, namely the trap of supposing that to put something in linguistic form is in some way to distort it or wrench it away from its natural bias. For the linguistic idealist will insist that that is not the case: the material is already in propositional form, waiting—­ constitutively disposed—to be captured in language. We find ourselves cast into a world where there are objects waiting to be referred to by proper names and demonstratives, properties waiting to be referred to by general terms, functions and operations waiting to be referred to by connectives, and so on. And these objects, properties, functions, and operations do not hang loose from one another like so many Kantian things-in-themselves: they are already arranged in propositional structures that are waiting to be referred to by sentences, both true and false. Hence there is a sense in which the world is already fitted up to give linguistic pleasure and, if there is a paradox in the offing, it is not (or not just) a paradox of language, but a paradox of the world.

352  Tragedy and Linguistic Redress The linguistic idealist holds, as I said at the beginning of this chapter, that language is transcendentally responsible for the existence and general constitution of the world: that is to say, it is a necessary condition of the existence of anything at all that it can be expressed in language. In that sense language creates the world. Language does not, of course, create the world in an empirical sense. In this latter sense the relation of dependence is the other way round: the world was already there long before historical languages came on the scene, and will no doubt be there long after they have vanished. That fact is indeed what makes the linguistic idealist’s position so extraordinary. Human language is an empirical phenomenon; it arose in historical time; it came into being and will likely pass away again in what, from the point of view of the universe, will have been a relatively short period.91 Yet the universe itself is transcendentally dependent on human language, because (it is a necessary condition of its existence that) it and its entire history are expressible in that very language. The progress of the cosmos will no doubt discard human language as it will discard human beings themselves, and yet these very acts of dismissal can, like everything else, be described by language. Language, trounced as an empirical phenomenon, rises like a phoenix from the ashes, transcendentally triumphant. The essence of linguistic idealism is encapsulated in the almost ungraspable fact that language can talk about a time and a place when and where it was not, is not, or will not be. In saying that language is not in those times and places, what I mean is that it is not there in any empirical sense. But the linguistic idealist’s point is that, in some sense (a transcendental one), it really is there, in those empirically languageless places and at those empirically languageless times, after all. This generalizes a point that I made in §30, namely that the English word ‘love’, for example, expresses a concept which existed and was exemplified in ancient Greek culture, long before the English language came on the scene. English was not around in fifth-century BCE Athens, but its meanings were. So, in general, meaning pervades space and time: it even exists where and when empirical language does not exist. ‘Now I am tempted to say that the right expression in language for the miracle of the existence of the world . . . is the existence of language itself’.92 In the conclusion to his book Shame and Necessity, Bernard Williams wrote, initially echoing Russell: We know that the world was not made for us, or we for the world, that our history tells no purposive story, and that there is no position outside the world or outside history from which we might hope to authenticate our activities. We have to acknowledge the hideous cost of many human achievements that we value, including this reflective sense itself, and recognize that there is no redemptive Hegelian history or universal Leibnizian cost-benefit analysis to show that it will come out well enough in the end.93

Tragedy and Linguistic Idealism 353 Trübselige Meinung: and yet you might say that, if anything, Williams’s words are not bleak enough. Not only do we not know that the life of our human species will turn out well enough in the end; we strongly surmise that, in the end, it will all come to nothing.94 As well as the sadness of this vision, there is also the sadness of the thought that our very sadness will itself die. So Shelley: Alas! that all we loved of him should be, But for our grief, as if it had not been, And grief itself be mortal!95 Or perhaps we feel relief with Brecht’s Nanna: Gott sei Dank geht alles schnell vorüber, Auch die Liebe und der Kummer sogar. Wo sind die Tränen von gestern abend? Wo ist der Schnee vom vergangenen Jahr?96 (‘Thank God, everything passes quickly, love too and even pain. Where are the tears of the evening gone by? Where are the snows of yesteryear?’) Or we feel amusement with Chaucer’s Pandarus when Troilus puts his trust in Criseyde: Pandare answerde, ‘It may be, wel ynough,’ And held with hym of al that evere he seyde. But in his herte he thoughte, and softe lough, And to hymself ful sobreliche he seyde, ‘From haselwode, there joly Robyn pleyde, Shal come al that that thow abidest heere. Ye, fare wel al the snow of ferne yere!’97 But from a linguistic point of view there is perhaps a slightly more optimistic conclusion to be had than the desolation of Williams’s message. For the sense in which, say, sadness will die is that there will probably one day no longer be any human beings around, and perhaps no conscious beings of any sort, to think sad thoughts and read sad poetry. Mortalia facta peribunt, sings the poet, nedum sermonum stet honos et gratia vivax: ‘Mortal works shall perish, and do not imagine that the glory and charm of speech can survive’.98 But the concept sadness, having once existed, will always exist, and that concept is something that is essentially expressible in words. Propositions, true and false, containing the concept of sadness, will, having once existed, always exist, and propositions are essentially expressible in sentences. And some propositions must exist, come what may, no matter what else does or does not exist: even in the so-called ‘empty universe’ there are propositions.99

354  Tragedy and Linguistic Redress Even so refined and extraordinary a thing as tragic literature once existed. Empirical language will pass away, as Timon hoped it would,100 but meaning will carry on, as Shakespeare cunningly hints by continuing to talk after Timon has spoken. When the young, vagrant Thomas De Quincey collapsed one evening in Oxford Street, his companion Ann dashed off to get spiced wine, which revived him.101 Ann later failed a vital rendezvous with Thomas, and was grievously lost forever in the maze of Soho. But first she struck an attitude of selflessness and grace that will not fade. Lear and Cordelia must die, but the scene of their reconciliation has intervened, searing the darkness like a ‘blinding shaft of transcendent light’: ‘This is the justification of the agony, the sufferance, the gloom. Though once more the shadows close, it has existed, immortal, in its own right, bending to no natural law’.102 Rusalka’s handsome prince, too, must die, and she is condemned to wander the earth as a phantom; but her song to the moon before she makes her fateful bargain with the witch Ježibaba is a supreme moment of liberation from time, an infinitely precious gleam from the abode where the eternal are. For these salvific facts—like the gentler past that Ernst Wiechert recollected in his harsh present when he stood under Goethe’s oak at Buchenwald—are themselves things that will always have been there, and are never to be annulled.

Notes 1. I repeat these formulations from Gaskin 2013a, 14. 2. See Gaskin 2008, esp. chs. 1 and 2. 3. Über Gewissheit, §402 (1984, 199; cf. 493). 4. I argue against McDowell’s view that the world is the (Fregean) sense of language at Gaskin 2006, ch. 5. 5. Steiner 1998, 24–5. Cf. 1971, 111–12. 6. For the argument, see Gaskin 2009. 7. For the mistake, see e.g. Franke 2014, 21. On the incoherence of the Tractarian conception of the limits of language, see Gaskin 2013a, 209–10; A. Moore 2013. 8. Cf. Longinus, De Sublimitate 39, 3; Halliwell 2012, 338. 9. See Scruton 2004, 4–5. 10. Scruton 2004, 2; Peacocke 2009, 268. 11. Appelqvist 2013b. 12. See Gaskin 2009, 57–9. 13. See e.g.  A. Moore 1997, 149–55. 14. Holmes 1989, 36; Jankélévitch 1983, 92; cf. T. Mann 2012c, 234; C. Hamilton 2016, 31. 15. Cf. Budd 1985, 151–2. 16. See here S. Davies 1994, chs. 1 and 3; 1999, 285–7; cf. Budd 1985, 161–72; Perrett 1999. 17. See Budd 1985 and 2005; Scruton 1988; S. Davies 1994, esp. chs. 4–6; 1999; Stecker 1999; Peacocke 2009. 18. Gaskin 2013a, §15. 19. Wittgenstein 1977, II, vi, 291, tr. Anscome (adapted).

Tragedy and Linguistic Idealism 355 0. On this use of physical vocabulary for the mental, see Padel 1992, esp. ch. 4. 2 21. Cf. Seneca, Trojan Women 1116; Medea 340–6. In Stefonio’s Crispus, the hero refuses to ‘express’—force—his salvation by use of arms: ‘non exprimenda viribus nostra est salus’ (IV, 397). See also Jonson, Cataline His Conspiracy III, 1, 209–10 (2012, vol. 4, 87). 22. Cf. Wittgenstein 1977, I, §374; S. Davies 1994, 156. 23. Eliot 1951, 68. The abuse of Senecan style reached its apogee in F. Lucas 1922, ch. 3. Recently there has been a rehabilitation: see A. Boyle 2006, chs. 7 and 8; Lefèvre 2015. 24. Garton 1972, 189–202. 25. Steiner 1961, 96; cf. 101; de Mourgues 1967, 139; Braden 1985, 181–2. 26. Easterling 1997a, 154–5. 27. Dodds 1960, 206, 214; but contrast his note on lines 1133–6. 28. See Poetics 1452b12–13, with D. Lucas 1972, 134; Allan on Euripides, Children of Heracles 63ff.; Silk and Stern 1981, 300. 29. M. Davies 1999, 40–1; Goldhill 2006, 154–6. On decorum in Greek tragedy more generally, see Scodel 2009, 189–91. 30. Horace, Ars Poetica 185; A. Boyle 2006, 217; 2014, xlix–l; 2017, l–liv. See also Coffey and Mayer on Seneca, Phaedra 1197–8; Seidensticker 2006, 103–4; Hoxby 2015, 245. Against the view that Sophocles’ Ajax commits suicide on stage, see Finglass 2011, 375–9; Garvie 2016, 13. 31. Cf. Steiner 1961, 49, though he overlooks Bérénice (ibid., 88); Parish 1993, 19 n. 1. 32. Voltaire held Sophocles’ Philoctetes to be unsuitable for the modern stage, though he was rebutted by Diderot, Lessing, and others: see Budelmann 2007, 452–5; Hoxby 2015, 152–3. 33. Shakespeare, Richard III IV, 4, 29 and 34; King Henry VIII IV, 2, 3; 1 Henry IV III, 1, 5; Klinger, Die Zwillinge IV, 4; Goethe, Götz von Berlichingen III, ‘Jaxthaussen’; Schiller, Kabale und Liebe V, 2; Maria Stuart IV, 4; cf. Mueller 1980, 64–7; Lamport 1990, 51–5. 34. Gossip 1981, 141–3. 35. Steiner 1961, 80. Cf. 1997, 27–37; Gossip 1981, ch. 11. 36. Zeitlin 2003, 339. On the end of Seneca’s Medea, see A. Boyle 2006, 217– 18; 2014, cxiv–xviii, and notes on lines 991–4 and 1019–22. 37. See Lessing, Hamburgische Dramaturgie, §§54, 55 (1996, vol. 4, 483, 487–9). 38. Here going beyond Seneca: A. Boyle 2014, cxxx. 39. On the problematic end of Horace, see Gossip 1981, 143–4; Goodkin 2009, 382; Oster 2014, 258. 40. See Steiner 1961, 53. 41. Cf. Miola 1992, 60. 42. See here Hoxby 2015, 258–61. 43. Cf. de Mourgues 1967, 35, 44–6. 44. Note on Euripides, Medea 30–1, with further references. 45. See Pocock 1973, 249–52, on the importance of the motifs of blood and excrement in Phèdre. 46. For details see Fitch 1987, 117. 47. J. Campbell 2005, 65; cf. 56–65. 48. See e.g. Sophocles, Ajax 131–2, 753–7, 778, 801–2, 1362, with Finglass ad locc.; Electra 17–19, 674, 783, 1149, 1363, with Finglass ad locc.; Oedipus Rex 438, 1283; Antigone 14, 55, 170–1; Women of Trachis 740; Oedipus at Colonus 1612; Euripides, Alcestis 20, 232, 513; Medea 1231, 1247–8; Hippolytus 22, 726, 889–90, 1003; Andromache 803; Hecuba 44, 285; Ion

356  Tragedy and Linguistic Redress 420–1; Helen 879, 1420; Phoenician Women 425, 1085, 1197, 1263, 1579, 1689 (with Mastronarde ad locc.); Orestes 48–9, 656, 858, 948; [Euripides], Rhesus 447, 600–4, with Liapis ad locc. Cf. Aeschylus, Seven Against Thebes 21; Agamemnon 320, 575, 1301; Aristophanes, Birds 1045, with Dunbar ad loc.; Thesmophoriazusae 71, 76, 83, 181, 729, with Austin and Olson on 729; Taplin 1977, 292; Scodel 2009, 188; Rutherford 2012, 188. 49. See M. Burnett 2010, 170. 50. See J. Gregory on Euripides, Hecuba 285. 51. Henrichs 1994, 72. 52. See Barthes 1963, 88–9. 53. Cf. Seneca, Trojan Women 280, with A. Boyle ad loc. 54. Yourcenar, quoted by J. Campbell 2005, 169; Dimock 2008, 73; Martindale and Martindale 1990, 93; Boitani 1989a, 5. 55. Recall the words of Naevius quoted in §24. See Fraenkel on Aeschylus, Agamemnon 826; Heinze 1957, 24–5; Putnam 2010, 83–5. 56. Lessing, Laokoon, §6 (1996, vol. 6, 55). 57. See Seneca, Medea 876, with the commentaries of Costa and A. Boyle ad loc.; Trojan Women 438; Agamemnon 74, with Tarrant ad loc.; Alabaster, Roxana II, 3, 473. 58. See Dodds ad loc. 59. See Sommerstein ad loc. 60. Euripides, Trojan Women 204, tr. Kovacs. 61. Cf. the first stasimon of Trojan Women, 511–67, and the third chorus of Seneca, Agamemnon, 589–658. 62. T. S. Eliot, ‘The Dry Salvages’ 43. 63. Shakespeare, 2 Henry VI I, 4, 16–19. 64. Euripides, Trojan Women 446; cf. Aeschylus, Agamemnon 1541–50; Choephori 8, 429–33, with Garvie ad locc. (and on 22–83). Cf. the death and burial of Britannicus, as reported by Tacitus, Annals XIII, 17, and Suetonius, Nero XXXIII, 3. 65. Aeschylus, Choephori 535–7. 66. Garvie on Choephori 536–7. 67. Euripides, Trojan Women 845–59. 68. An idea that lends itself to ready naturalization in the manner of §8. See Pelling on Plutarch, Antony 75, 4; Tarrant on Virgil, Aeneid XII, 875; R. Parker 1997, 154–5; Gaskin 2013b, 85 with n. 26. 69. Cf. Aeschylus, Choephori 34. 70. Putnam 1965, 20. 71. See Aeneid I, 726–7, 748; II, 9, with Austin ad loc.; cf. Seneca, Hercules Furens 843, with Fitch ad loc. 72. See Wilson Knight 1986, 126–7, 144–6; Martindale and Martindale 1990, 17; Bradley 1991, 307. 73. Cf. Euripides, Suppliants 842–3, 914–17, with Morwood ad locc. 74. Franke 2014, 13–14; cf. S. Wells 2016, 155–6. 75. Cf. Boethius, Consolatio Philosophiae IV, metr. 3. 76. P. Holland 1994, 150–1. Cf. Martindale and Martindale 1990, 64–6. 77. See Brooks ad loc.; Montrose 1996, 192–3; Nuttall 2007b, 127. 78. Greenblatt 2005, 52; cf. Montrose 1996, 198. 79. I think it is an irony (cf. Nozick 1983, 155), but for a different view see A. Moore 1997, 156. 80. Cf. Gaskin 1996. 81. Wittgenstein 1977, I, §429 (tr. Anscombe). Cf. I, §§95, 402, 443, 447; 1973, 135; McDowell 1994, 27; Gaskin 2006, passim, e.g. 53–4. 82. D. Schmidt 2001, 72; cf. 73, 283.

Tragedy and Linguistic Idealism 357 83. Nietzsche 1999a, 51. Cf. Schmidt 2001, 196–7; Dewar-Watson 2013, 93. 84. Wittgenstein 1977, I, §402 (tr. Anscombe, adapted). 85. Cf. Nuttall 2007a, 53–4. 86. See e.g. B. Williams 2006, 56–8; cf. C. Hamilton 2016, 119–21. 87. Nietzsche 1996, §§851–2; cf. Kaufmann 1968, 194; Silk and Stern 1981, 267; Ahrensdorf 2009, 155. 88. Wittgenstein 1980, 72; 1993a, 77; cf. Goethe, Torquato Tasso V, 5, 3427–33. 89. Redfield 1994, 39; cf. Brewer 1989, 98. 90. Cf. Evers and Deng 2016, 5. 91. Cf. Raymond Williams 1977, 21. 92. Wittgenstein 1993b, 43–4: in the part I have omitted Wittgenstein relies (unnecessarily) on the Tractarian picture theory. 93. B. Williams 1993, 166; Russell 1986, 14. 94. Cf. Goethe Faust II, 11600–1 (cf. I, 1335–58). 95. Shelley, Adonais XXI, 181–3. 96. Brecht, Die Rundköpfe und die Spitzköpfe, ‘Nannas Lied’ (1977, vol. 3, 931–2). 97. Chaucer, Troilus and Criseyde V, 1170–6. 98. Horace, Ars Poetica 68–9; see Rudd ad loc. 99. Cf. Rundle 2004, 112. 100. Shakespeare, Timon of Athens V, 2, 105. 101. De Quincey 2000–4, vol. 2, 203–4. 102. Wilson Knight 1986, 203.

Bibliography

Classical and other primary works not included in this list are cited in standard editions. Adams, R. 1981: ‘Actualism and Thisness’, Synthese 49, 3–41. Adams, S. 1931: ‘The Burial of Polyneices’, Classical Review 45, 110–11. Adkins, A. 1960: Merit and Responsibility: A Study in Greek Values (Oxford: Clarendon). ———. 1966: ‘Aristotle and the Best Kind of Tragedy’, Classical Quarterly 16, 78–102. Aeschylus. 1972: Aeschyli Septem Quae Supersunt Tragoediae, ed. D. Page (Oxford: Clarendon). ———. 1998: Tragoediae, Cum Incerti Poetae Prometheo, ed. M. L. West (Stuttgart/Leipzig: Teubner). ———. 2008: Aeschylus: Fragments, ed. A. Sommerstein (Cambridge, MA: Harvard University Press). Agamben, G. 2012: Quel che resta di Auschwitz: L’archivo e il testimone (Turin: Bollati Boringhieri). Ahrensdorf, P. 2009: Greek Tragedy and Political Philosophy: Rationalism and Religion in Sophocles’ Theban Plays (Cambridge: Cambridge University Press). Alabaster, W. 1632: Roxana (London: Jones), repr. Olms 1987. Albrecht, M. von. 1970: ‘Zur Tragik von Vergils Turnusgestalt: Aristotelisches in der Schlußszene der Aeneis’, in M. von Albrecht and E. Heck (eds.), Silvae: Festschrift E. Zinn (Tübingen: Niemeyer), 1–5. Alexander, P. 1955: Hamlet: Father and Son (Oxford: Clarendon). Allan, W. 2001: Euripides: The Children of Heracles (Warminster: Aris and Phillips). ———. 2008: Euripides: Helen (Cambridge: Cambridge University Press). Allan, W. and Kelly, A. 2013: ‘Listening to Many Voices: Athenian Tragedy as Popular Art’, in A. Marmodoro and J. Hill (eds.), The Author’s Voice in Classical and Late Antiquity (Oxford: Oxford University Press), 77–122. Améry, J. 1988: Jenseits von Schuld und Sühne: Bewältigungsversuche eines Überwältigten (Munich: Deutscher Taschenbuch Verlag), tr. S. Rosenfeld and S. P. Rosenfeld, At the Mind’s Limits: Compilations by a Survivor of Auschwitz and Its Realities (New York: Schocken Books, 1986). ———. 2008: Werke, 9 vols., ed. I. Heidelberger-Leonard (Stuttgart: Klett-Cotta).

Bibliography 359 Andrew, M. 1989: ‘The Fall of Troy in Sir Gawain and the Green Knight and Troilus and Criseyde’, in P. Boitani (ed.), The European Tragedy of Troilus (Oxford: Clarendon), 75–93. Anouilh, J. 1954: Antigone, ed. W. M. Landers (London: George Harrap). Appelqvist, H. 2013a: ‘Why Does Wittgenstein Say That Ethics and Aesthetics Are One and the Same?’ in P. Sullivan and M. Potter (eds.), Wittgenstein’s Tractatus: History and Interpretation (Oxford: Oxford University Press), 40–58. ———. 2013b: ‘Wittgenstein and the Limits of Musical Grammar’, British Journal of Aesthetics 53, 299–319. Aquinas, T. 1894: Summa Theologica, 5 vols. (Rome: Forzani). Aretino, P. 1971: Teatro, ed. G. Petrocchi (Rome: Arnoldo Mondadori Editore). Aristotle. 1965: De Arte Poetica Liber, ed. R. Kassel (Oxford: Clarendon). Armstrong, J. 1998: ‘Aristotle on the Philosophical Nature of Poetry’, Classical Quarterly 48, 447–55. Armstrong, R. 2012: ‘Freud and the Drama of Oedipal Truth’, in K. Ormand (ed.), A Companion to Sophocles (Oxford: Wiley-Blackwell), 477–91. Arnold, M. 1909: The Poems, 1840–1867, ed. A. Quiller-Couch (London: Oxford University Press). Artaud, A. 2011: Les Cenci, ed. M. Corvin (Paris: Gallimard). Auchincloss, L. 1996: La Gloire: The Empire of Corneille and Racine (Columbia: University of South Carolina Press). Auerbach, E. 1946: Mimesis: Dargestellte Wirklichkeit in der abendländischen Literatur (Tübingen: Francke). Augustine. 1981: De Civitate Dei, 5th edn., 2 vols., ed. B. Dombart et al. (Darmstadt: WBG). Austen, J. 2003: Emma, ed. J. Kinsley (Oxford: Oxford University Press). Austin, C. and Olson, S. D. 2004: Aristophanes: Thesmophoriazusae (Oxford: Oxford University Press). Austin, R. G. 1964: P. Vergili Maronis Aeneidos Liber Secundus (Oxford: Oxford University Press). ———. 1971: P. Vergili Maronis Aeneidos Liber Primus (Oxford: Oxford University Press). ———. 1973: P. Vergili Maronis Aeneidos Liber Quartus (Oxford: Oxford University Press). Bain, D. 1993: ‘A Misunderstood Scene in Sophokles, Oidipous (O. T. 300– 462)’, in I. McAuslan and P. Walcott (eds.), Greek Tragedy (Oxford: Oxford University Press), 81–94. Balaguer, M. 2011: ‘Bare-Bones Moral Realism and the Objections from Relativism’, in S. Hales (ed.), A Companion to Relativism (Oxford: Wiley-Blackwell), 268–90. Barnwell, H. 1982: The Tragic Drama of Corneille and Racine: An Old Parallel Revisited (Oxford: Clarendon). Barrett, W. S. 1964: Euripides: Hippolytos (Oxford: Clarendon). Bartels, E. 2004: ‘Othello on Trial’, in L. Orlin (ed.), New Casebooks: Othello (Basingstoke: Palgrave Macmillan), 148–70. Barthes, R. 1963: Sur Racine (Paris: Seuil). Bate, J. 1993: Shakespeare and Ovid (Oxford: Clarendon). ———. 1995: Shakespeare: Titus Andronicus (London: Routledge).

360  Bibliography Bauer, Y. 2001: Rethinking the Holocaust (New Haven: Yale University Press). Bayley, J. 1981: Shakespeare and Tragedy (London: Routledge & Kegan Paul). Beckett, S. 2006: The Complete Dramatic Works (London: Faber and Faber). Behrens, R. 2014: ‘Racines Britannicus: Der Trug der Zeichen, das Theater und die Macht’, in N. Bender et al. (eds.), Ethos und Form der Tragödie (Heidelberg: Winter), 269–94. Beiser, F. 2005: Schiller as Philosopher (Oxford: Clarendon). Belle, M-A. and Cottegnies, L. 2017: Robert Garnier in Elizabethan England: Mary Sidney Herbert’s Antonius and Thomas Kyd’s Cornelia (Cambridge: MHRA). Bellini, V. 2013: Norma (London: Decca Music Group). Belsey, C. 1985: The Subject of Tragedy: Identity and Difference in Renaissance Drama (London: Routledge). ———. 1990: ‘Richard Levin and in-Different Reading’, New Literary History 21, 449–56. ———. 1992: ‘Literature, History, Politics’, in R. Wilson and R. Dutton (eds.), New Historicism and Renaissance Drama (London: Longman), 33–44. ———. 1999: Shakespeare and the Loss of Eden (New Brunswick, NJ: Rutgers University Press). ———. 2001: ‘Tarquin Dispossessed: Expropriation and Consent in The Rape of Lucrece’, Shakespeare Quarterly 52, 315–35. ———. 2003a: ‘The Divided Tragic Hero’, in R. Dutton and J. Howard (eds.), A Companion to Shakespeare’s Works: The Tragedies (Oxford: Wiley-Blackwell), 73–94. ———. 2003b: ‘Reading and Critical Practice’, Critical Quarterly 45, 22–31. Benardete, S. 1966: ‘Sophocles’ Oedipus Tyrannus’, in T. Woodward (ed.), Sophocles: A Collection of Critical Essays (Englewood Cliffs, NJ: PrenticeHall), 105–21. ———. 1999: Sacred Transgressions: A Reading of Sophocles’ Antigone (South Bend, IN: St. Augustine’s Press). Bennett, L. and Tyrrell, W. 1990: ‘Sophocles’ Antigone and Funeral Oratory’, American Journal of Philology 111, 441–56. ———. 1998: Recapturing Sophocles’ Antigone (Lanham: Rowman & Littlefield). Berger, H. 1997: Making Trifles of Terrors: Redistributing Complicities in Shakespeare (Stanford: Stanford University Press). ———. 2004: ‘Impertinent Trifling: Desdemona’s Handkerchief’, in L. Orlin (ed.), New Casebooks: Othello (Basingstoke: Palgrave Macmillan), 103–24. Berger, P. and Luckmann, T. 1966: The Social Construction of Reality: A Treatise in the Sociology of Knowledge (London: Penguin). Bernstein, J. M. 2010: ‘ “The Celestial Antigone, the Most Resplendent Figure Ever to Have Appeared on Earth”: Hegel’s Feminism’, in F. Söderbäck (ed.), Feminist Readings of the Antigone (New York: SUNY Press), 111–30. Bernstein, R. 2002: Radical Evil: A Philosophical Investigation (Cambridge: Polity Press). Bevington, D. 1998: Shakespeare: Troilus and Cressida (London: Cengage). Biard, J. 1965: ‘Le ton élégiaque dans Bérénice’, French Studies 1, 1–15. Bickley, P. and Stevens, J. 2016: Shakespeare and Early Modern Drama: Text and Performance (London: Bloomsbury).

Bibliography 361 Billings, J. 2014: Genealogy of the Tragic: Greek Tragedy and German Philosophy (Princeton: Princeton University Press). Billings, J. and Leonard, M. 2015: Tragedy and the Idea of Modernity (Oxford: Oxford University Press). Blackburn, S. 2014: ‘Relativism and the Abolition of the Other’, in M. Baghramian (ed.), The Many Faces of Relativism (London: Routledge), 186–99. Bloom, H. 1999: Shakespeare: The Invention of the Human (London: Fourth Estate). Blundell, M. 1989: Helping Friends and Harming Enemies: A Study in Sophoclean and Greek Ethics (Cambridge: Cambridge University Press). Boccaccio, G. 1970: Tutte le opere, 12 vols., ed. V. Branca (Milan: Arnoldo Mondadori). Böckh, A. 1843: Des Sophokles Antigone (Berlin: Veit and Co.). Boedeker, D. and Raaflaub, K. 2009: ‘Tragedy and the City’, in R. Bushnell (ed.), A Companion to Tragedy (Oxford: Wiley-Blackwell), 109–27. Boethius. 1978: ‘Consolatio Philosophiae’, in H. Stewart et al. (eds.), Boethius: Tractates; De Consolatione Philosophiae (Cambridge, MA: Harvard University Press). Boitani, P. 1989a: ‘Antiquity and Beyond: The Death of Troilus’, in P. Boitani (ed.), The European Tragedy of Troilus (Oxford: Clarendon), 1–19. ———. 1989b: ‘Eros and Thanatos: Cressida, Troilus, and the Modern Age’, in P. Boitani (ed.), The European Tragedy of Troilus (Oxford: Clarendon), 281–305. Bollack, J. 1999: La Mort d’Antigone; La tragédie de Créon (Paris: Presses Universitaires de France). Bolt, R. 1960: A Man for All Seasons (London: Heinemann). Bond, G. 1988: Euripides: Heracles (Oxford: Clarendon). Boose, L. 2004: ‘ “Let it be Hid”: The Pornographic Aesthetic of Shakespeare’s Othello’, in L. Orlin (ed.), Othello: Contemporary Critical Essays (Basingstoke: Palgrave Macmillan), 22–48. Borowski, T. 2008: Bei uns in Auschwitz, tr. F. Griese (Munich: btb). Boswell, J. 2008: The Life of Samuel Johnson, ed. D. Womersley (London: Penguin). Bowen, A. J. 2013: Aeschylus: Suppliant Women (Oxford: Oxbow). Bowra, M. 1945: Sophoclean Tragedy (Oxford: Clarendon). Boyle, A. J. 1986: The Chaonian Dove: Studies in the Eclogues, Georgics, and Aeneid of Virgil (Leiden: Brill). ———. 1994: Seneca’s Troades (Leeds: Francis Cairns). ———. 2006: An Introduction to Roman Tragedy (London: Routledge). ———. 2008: Octavia: A Play Attributed to Seneca (Oxford: Oxford University Press). ———. 2011: Seneca: Oedipus (Oxford: Oxford University Press). ———. 2014: Seneca: Medea (Oxford: Oxford University Press). ———. 2017: Seneca: Thyestes (Oxford: Oxford University Press). Boyle, N. 2010: ‘Goethe’s Theory of Tragedy’, Modern Language Review 105, 1072–86. Braden, G. 1985: Renaissance Tragedy and the Senecan Tradition: Anger’s Privilege (New Haven: Yale University Press). Bradley, A. C. 1959: Oxford Lectures on Poetry (London: Palgrave Macmillan).

362  Bibliography ———. 1991: Shakespearean Tragedy: Lectures on Hamlet, Othello, King Lear, Macbeth (London: Penguin). Brault, G. 1984: La Chanson de Roland (Pennsylvania: Pennsylvania State University Press). Brecht, B. 1977: Werke, 20 vols. (Frankfurt am Main: Suhrkamp). Bremer, J. 1969: Hamartia: Tragic Error in the Poetics of Aristotle and in Greek Tragedy (Amsterdam: Hakkert). Brewer, D. 1989: ‘Comedy and Tragedy in Troilus and Criseyde’, in P. Boitani (ed.), The European Tragedy of Troilus (Oxford: Clarendon), 95–109. Bristol, M. 2009: ‘Confusing Shakespeare’s Characters with Real People: Reflections on Reading in Four Questions’, in P. Yachnin and J. Slights (eds.), Shakespeare and Character: Theatre, History, Performance, and Theatrical Persons (Basingstoke: Palgrave Macmillan), 21–40. Brooks, H. 1979: Shakespeare: A Midsummer Night’s Dream (London: Methuen). Brown, A. 1983: ‘The Erinyes in the Oresteia: Real Life, the Supernatural, and the Stage’, Journal of Hellenic Studies 103, 13–34. ———. 1987: Sophocles: Antigone (Warminster: Aris and Phillips). Brown, D. 1991: Human Universals (Boston: McGraw-Hill). Buchan, M. 2012: ‘Sophocles with Lacan’, in K. Ormand (ed.), A Companion to Sophocles (Oxford: Wiley-Blackwell), 492–504. Buchanan, G. 1983: George Buchanan’s Tragedies, eds. P. Sharratt and P. G. Walsh (Edinburgh: Scottish Academic Press). Büchner, G. 2009: Werke und Briefe, eds. K. Pörnbacher et al. (Munich: Deutscher Taschenbuch Verlag). Budd, M. 1985: Music and the Emotions: The Philosophical Theories (London: Routledge). ———. 2005: ‘Aesthetic Realism and Emotional Qualities of Music’, British Journal of Aesthetics 45, 111–22. Budelmann, F. 2000: The Language of Sophocles: Communality, Communication, and Involvement (Cambridge: Cambridge University Press). ———. 2006a: ‘The Meditated Ending of Sophocles’ Oedipus Tyrannus’, Materiali e discussioni per l’analisi dei testi classici 57, 43–61. ———. 2006b: ‘Körper und Geist in tragischen Schmerz-Szenen’, in B. Seidensticker and M. Vöhler (eds.), Gewalt und Ästhetik: Zur Gewalt and ihrer Darstellung in der griechischen Klassik (Berlin: de Gruyter), 123–48. ———. 2007: ‘The Reception of Sophocles’ Representation of Physical Pain’, American Journal of Philology 128, 443–67. Budelmann, F. and Easterling, P. E. 2010: ‘Reading Minds in Greek Tragedy’, Greece and Rome 57, 289–303. Budra, P. 2000: A Mirror for Magistrates and the De Casibus Tradition (Toronto: University of Toronto Press). Bullough, G. 1973: Narrative and Dramatic Sources of Shakespeare, 7 vols. (London: Routledge & Kegan Paul). Burian, P. 1997a: ‘Myth into Muthos: The Shaping of the Tragic Plot’, in P. E. Easterling (ed.), The Cambridge Companion to Greek Tragedy (Cambridge: Cambridge University Press), 178–208. ———. 1997b: ‘Tragedy Adapted for Stages and Screens: the Renaissance to the Present’, in P. E. Easterling (ed.), The Cambridge Companion to Greek Tragedy (Cambridge: Cambridge University Press), 228–83.

Bibliography 363 ———. 2009: ‘Inconclusive Conclusion: The Ending(s) of Oedipus Tyrannus’, in S. Goldhill and E. Hall (eds.), Sophocles and the Greek Tragic Tradition (Cambridge: Cambridge University Press), 99–118. ———. 2010: ‘Gender and the City: Antigone from Hegel to Butler and Back’, in K. Bassi and J. P. Euben (eds.), When Worlds Elide: Classics, Politics, Culture (Lanham: Rowman & Littlefield), 255–99. ———. 2011: ‘Athenian Tragedy as Democratic Discourse’, in D. M. Carter (ed.), Why Athens? A Reappraisal of Tragic Politics (Oxford: Oxford University Press), 95–117. ———. 2012: ‘Polyphonic Ajax’, in K. Ormand (ed.), A Companion to Sophocles (Oxford: Wiley-Blackwell), 69–83. Burke, E. 1990: A Philosophical Enquiry into the Origin of Our Ideas of the Sublime and Beautiful, ed. A. Phillips (Oxford: Oxford University Press). Burkert, W. 1991: Oedipus, Oracles, and Meaning: From Sophocles to Umberto Eco (Toronto: University of Toronto). Burnett, A. 1971: Catastrophe Survived: Euripides’ Plays of Mixed Reversal (Oxford: Clarendon). Burnett, M. 2010: ‘Doctor Faustus: Dramaturgy and Disturbance’, in E. Smith and G. Sullivan (eds.), English Renaissance Tragedy (Cambridge: Cambridge University Press), 163–73. Burns, T. 2002: ‘Sophocles’ Antigone and the History of the Concept of Natural Law’, Political Studies 50, 545–57. Burrow, C. 2013: Shakespeare and Classical Antiquity (Oxford: Oxford University Press). Bushnell, R. 1981: ‘Oracular Silence in Oedipus the King and Macbeth’, Classical and Modern Literature 2, 195–204. Butcher, S. H. 1951: Aristotle: Poetics (New York: Dover). Butler, J. 2000: Antigone’s Claim: Kinship Between Life and Death (New York: Columbia University Press). ———. 2001: ‘Antigone, Speech, Performance, Power: Dialogue with Paul Rabinow’, in S. Salamensky (ed.), Talk Talk Talk: The Cultural Life of Everyday Conversation (London: Routledge), 37–48. Buxton, R. 1980: ‘Blindness and Limits: Sophokles and the Logic of Myth’, Journal of Hellenic Studies 100, 22–37. ———. 1996: ‘What Can You Rely on in Oedipus Rex? Response to Calame’, in M. Silk (ed.), Tragedy and the Tragic: Greek Theatre and Beyond (Oxford: Oxford University Press), 38–48. Byron [George Gordon]. 1980–93: The Complete Poetical Works, 7 vols., ed. J. McGann (Oxford: Clarendon). Bywater, I. 1920: Aristotle on the Art of Poetry (Oxford: Clarendon). Cairns, D. 1993: Aidōs: The Psychology and Ethics of Honour and Shame in Ancient Greek Literature (Oxford: Clarendon). ———. 2005: ‘Values’, in J. Gregory (ed.), A Companion to Greek Tragedy (Oxford: Wiley-Blackwell), 305–20. ———. 2006: ‘Virtue and Vicissitude: The Paradoxes of the Ajax’, in D. Cairns and V. Liapis (eds.), Dionysalexandros: Essays on Aeschylus and his Fellow Tragedians in Honour of Alexander F. Garvie (Swansea: Classical Press of Wales), 99–131. ———. 2016: Sophocles: Antigone (London: Bloomsbury).

364  Bibliography Calder, W. 2005: Theatrokratia: Collected Papers on the Politics and Staging of Greco-Roman Tragedy, ed. R. Scott Smith (Hildesheim: Olms). Camon, F. 1997: Conversazione con Primo Levi (Parma: Ugo Guanda Editore). Campbell, J. 2005: Questioning Racinian Tragedy (Chapel Hill: University of North Carolina). Campbell, L. 1879: Sophocles, 2 vols. (Oxford: Clarendon). Campbell, L. B. 1938: The Mirror for Magistrates (Cambridge: Cambridge University Press). Camus, A. 1967: Lyrical and Critical, ed. P. Thody (London: Hamish Hamilton). Candlish, S. 1989: ‘The Truth about F. H. Bradley’, Mind 98, 331–48. Carter, D. 2007: The Politics of Greek Tragedy (Exeter: Exeter Press). Cavarero, A. 2010: ‘On the Body of Antigone’, in F. Söderbäck (ed.), Feminist Readings of the Antigone (New York: SUNY Press), 45–63. Cavell, S. 2003: Disowning Knowledge in Seven Plays of Shakespeare, updated edn. (Cambridge: Cambridge University Press). Cessi, V. 1987: Erkennen und Handeln in der Theorie des Tragischen bei Aristoteles (Frankfurt am Main: Athenäum). Chanter, T. 2010: ‘Antigone’s Political Legacies: Abjection in Defiance of Mourning’, in S. Wilmer and A. Žukauskaitė (eds.), Interrogating Antigone in Postmodern Philosophy and Criticism (Oxford: Oxford University Press), 19–47. Chaucer, G. 1987: The Riverside Chaucer, ed. L. Benson (Oxford: Oxford University Press). ———. 1995: The Knight’s Tale, ed. A. Spearing (Cambridge: Cambridge University Press). ———. 1997: The Book of the Duchess, 3rd edn., ed. H. Phillips (Durham: Durham Medieval Texts). Chevalier, J-F. 2013: ‘Neo-Latin Theatre in Italy’, in J. Bloemendal and H. Norland (eds.), Neo-Latin Drama and Theatre in Early Modern Europe (Leiden: Brill), 27–87. Christenson, D. 2000: Plautus: Amphitruo (Cambridge: Cambridge University Press). Clark, S. and Mason, P. 2015: Shakespeare: Macbeth (London: Bloomsbury). Clay, D. 1982: ‘Unspeakable Words in Greek Tragedy’, American Journal of Philology 103, 277–98. Cocteau, J. 2012: La machine infernale (Paris: Livre de Poche). Coffey, M. and Mayer, R. 1990: Seneca: Phaedra (Cambridge: Cambridge University Press). Coleman, R. 1990: ‘The Gods in the Aeneid’, in I. McAuslan and P. Walcot (eds.), Virgil (Oxford: Oxford University Press), 39–64. Coleridge, S. T. 1987: Lectures 1808–1819 on Literature, 2 vols., ed. R. Foakes (Princeton: Princeton University Press). Conte, G. B. 2007: The Poetry of Pathos: Studies in Virgilian Epic, ed. S. J. Harrison (Oxford: Oxford University Press). Cook, A. 1980: ‘The Design of Desdemona: Doubt Raised and Resolved’, Shakespeare Studies 13, 187–96. Cooper, H. 2010: Shakespeare and the Medieval World (London: Bloomsbury). Corbin, P. and Sedge, D. 1986: Three Jacobean Witchcraft Plays (Manchester: Manchester University Press). Corneille, P. 1999: Trois Discours sur le poème dramatique (Paris: Flammarion).

Bibliography 365 ———. 2006: Théâtre, 3 vols., ed. J. Maurens (Paris: Flammarion). Costa, C. 1973: Seneca: Medea (Oxford: Clarendon). Coyle, M. 2003: ‘The Tragedies of Shakespeare’s Contemporaries’, in R. Dutton and J. Howard (eds.), A Companion to Shakespeare’s Works: The Tragedies (Oxford: Wiley-Blackwell), 23–46. Coyne, J. 2009: Why Evolution is True (Oxford: Oxford University Press). Craik, E. 1993: ‘AIΔΩΣ in Euripides’ Hippolytus 373–430: Review and Reinterpretation’, Journal of Hellenic Studies 113, 45–59. ———. 1997: ‘Phaedra’s Aidōs Again’, Classical Quarterly 47, 567–9. Crawforth, H., Dustagheer, S. and Young, J. 2015: Shakespeare’s London (London: Bloomsbury). Crisp, R. 2000: Aristotle: Nicomachean Ethics (Cambridge: Cambridge University Press). ———. 2013: ‘Homeric Ethics’, in R. Crisp (ed.), The Oxford Handbook of the History of Ethics (Oxford: Oxford University Press), 1–20. Croally, N. 2005: ‘Tragedy’s Teaching’, in J. Gregory (ed.), A Companion to Greek Tragedy (Oxford: Wiley-Blackwell), 55–70. Cropp, M. J. 1997: ‘Antigone’s Final Speech (Sophocles, Antigone 891–928)’, Greece and Rome 44, 137–60. ———. 2013: Euripides: Electra, 2nd edn. (Oxford: Aris and Phillips). Crowther, P. 1989: The Kantian Sublime: From Morality to Art (Oxford: Oxford University Press). Cunliffe, J. 1912: Early English Classical Tragedies (Oxford: Clarendon). Cutrofello, A. 2014: All for Nothing: Hamlet’s Negativity (Cambridge, MA: MIT Press). Dacier, A. 1692: La Poetique d’Aristote (Paris: Barbin; repr. Hildesheim: Olms, 1976). Dale, A. M. 1954: Euripides: Alcestis (Oxford: Clarendon). ———. 1969: Collected Papers (Cambridge: Cambridge University Press). Daniel, S. 1911: The Tragedie of Cleopatra, ed. M. Lederer (Louvain: Uystpruyst). Daniels, C. 2006: ‘Is Oedipus Smart?’ Philosophy and Literature 30, 562–6. Dante Alighieri. 1966: Epistolae, 2nd edn., ed. P. Toynbee (Oxford: Clarendon). ———. 1968: De Vulgari Eloquentia, 3rd edn., eds. A. Marigo and P. G. Ricci (Florence: Felice Le Monnier). ———. 1981: The Divine Comedy, text and tr. G. Bickersteth (Oxford: WileyBlackwell). Dasenbrock, R. 1993: ‘Introduction: Davidson and Literary Theory’, in R. Dasenbrock (ed.), Literary Theory After Davidson (Pennsylvania: Pennsylvania State University Press), 1–17. Davidson, D. 1982: Essays on Actions and Events (Oxford: Clarendon). ———. 1984: Inquiries into Truth and Interpretation (Oxford: Clarendon). ———. 2001: Subjective, Intersubjective, Objective (Oxford: Clarendon). Davies, M. 1982: ‘The End of Sophocles’ O. T.’, Hermes 110, 268–78. ———. 1986: ‘Who Speaks at Sophocles Antigone 572?’ Prometheus 12, 19–24. ———. 1991: ‘The End of Sophocles’ O. T. Revisited’, Prometheus 17, 1–18. ———. 1999: ‘The Three Electras: Strauss, Hofmannsthal, Sophocles, and the Tragic Vision’, Antike und Abendland 45, 36–65. Davies, S. 1994: Musical Meaning and Expression (Ithaca: Cornell University Press).

366  Bibliography ———. 1999: ‘Response to Robert Stecker’, British Journal of Aesthetics 39, 282–7. Davis, M. 1986: ‘Politics and Madness’, in J. P. Euben (ed.), Greek Tragedy and Political Theory (Berkeley: University of California Press), 142–61. Dawe, R. D. 1968: ‘Some Reflections on Ate and Hamartia’, Harvard Studies in Classical Philology 72, 89–123. ———. 1978: Studies on the Text of Sophocles, vol. 3 (Leiden: Brill). ———. 2006: Sophocles: Oedipus Rex, rev. edn. (Cambridge: Cambridge University Press). Dawkins, R. 1998: Unweaving the Rainbow: Science, Delusion, and the Appetite for Wonder (London: Penguin). Dawson, A. and Minton, G. 2008: Shakespeare: Timon of Athens (London: Cengage). Day, H. 2013: Lucan and the Sublime (Cambridge: Cambridge University Press). de Lorris, Guillaume and de Meun, Jean. 1992: Le Roman de la Rose (Paris: Livres de poche). De Quincey, T. 2000–4: The Works of Thomas De Quincey, 21 vols., eds. G. Lindop et al. (London: Pickering and Chatto). Denniston, J. 1939: Euripides: Electra (Oxford: Clarendon). ———. 1950: The Greek Particles, 2nd edn. rev. by K. Dover (Oxford: Oxford University Press). Denniston, J. and Page, D. 1957: Aeschylus: Agamemnon (Oxford: Clarendon). Deppman, J. 2012: ‘Jean Anouilh’s Antigone’, in K. Ormand (ed.), A Companion to Sophocles (Oxford: Wiley-Blackwell), 523–37. Derrida, J. 1967: L’Écriture et la Difference (Paris: Seuil). –——. 1972: Positions (Paris: Minuit). Dewar-Watson, S. 2007: ‘Aristotle and Tragicomedy’, in R. Lyne and S. Mukherji (eds.), Early Modern Tragicomedy (Cambridge: Brewer), 15–27. ———. 2013: ‘Marlowe’s Dramatic Form’, in E. Bartels and E. Smith (eds.), Christopher Marlowe in Context (Cambridge: Cambridge University Press), 49–56. –——. 2014: Tragedy: A Reader’s Guide to Essential Criticism (Basingstoke: Palgrave Macmillan). Diamond, C. 2008: ‘The Difficulty of Reality and the Difficulty of Philosophy’, in S. Cavell et al., Philosophy and Animal Life (New York: Columbia University Press), 43–89. Diggle, J. 1998: Tragicorum Graecorum Fragmenta Selecta (Oxford: Clarendon). Dihle, A. 1982: The Theory of the Will in Classical Antiquity (Berkeley: University of California Press). Dimock, W. 2008: ‘After Troy: Homer, Euripides, Total War’, in R. Felski (ed.), Rethinking Tragedy (Baltimore: Johns Hopkins University Press), 66–81. Dodd, J. 2000: An Identity Theory of Truth (Basingstoke: Palgrave Macmillan). Dodds, E. R. 1951: The Greeks and the Irrational (Berkeley: University of California Press). ———. 1960: Euripides: Bacchae, 2nd edn. (Oxford: Clarendon). ———. 1973: The Ancient Concept of Progress and Other Essays on Greek Literature and Belief (Oxford: Clarendon). Dollimore, J. 1990: ‘Shakespeare, Cultural Materialism, Feminism, and Marxist Humanism’, New Literary History 21, 471–93.

Bibliography 367 ———. 1994: ‘Shakespeare, Cultural Materialism, and the New Historicism’, in J. Dollimore and A. Sinfield (eds.), Political Shakespeare: Essays in Cultural Materialism, 2nd edn. (Manchester: Manchester University Press), 2–17. ———. 2010: Radical Tragedy: Religion, Ideology and Power in the Drama of Shakespeare and his Contemporaries, 3rd edn. (Basingstoke: Palgrave Macmillan). Donaldson, I. 1982: The Rapes of Lucretia: A Myth and Its Transformations (Oxford: Clarendon). Doran, R. 2017: The Theory of the Sublime from Longinus to Kant (Cambridge: Cambridge University Press). Dover, K. J. 1965a: Thucydides Book VI (Oxford: Oxford University Press). ———. 1965b: Thucydides Book VII (Oxford: Oxford University Press). ———. 1987: Greek and the Greeks: Collected Papers vol. 1, Language, Poetry, Drama (Oxford: Wiley-Blackwell). ———. 1993: Aristophanes: Frogs (Oxford: Clarendon). ———. 1994: Greek Popular Morality in the Time of Plato and Aristotle (Indianapolis: Hackett). Dover Wilson, J. 1964: What Happens in Hamlet (Cambridge: Cambridge University Press). duBois, P. 1986: ‘Antigone and the Feminist Critic’, Genre 19, 371–83. ———. 2001: Trojan Horses: Saving the Classics from Conservatives (New York: New York University Press). ———. 2008: ‘Toppling the Hero: Polyphony in the Tragic City’, in R. Felski (ed.), Rethinking Tragedy (Baltimore: Johns Hopkins University Press), 127–47. ———. 2014: ‘The Death of Character’, International Journal of the Classical Tradition 21, 301–8. Dué, C. 2012: ‘Lament as Speech Act in Sophocles’, in K. Ormand (ed.), A Companion to Sophocles (Oxford: Wiley-Blackwell), 236–50. Dunbar, N. 1995: Aristophanes: Birds (Oxford: Clarendon). Duncan-Jones, K. and Woudhuysen, H. 2007: Shakespeare’s Poems (London: Bloomsbury). Dunn, F. 2012: ‘Electra’, in K. Ormand (ed.), A Companion to Sophocles (Oxford: Wiley-Blackwell), 98–110. Dunning, T. 1962: ‘God and Man in Troilus and Criseyde’, in N. Davis and C. Wrenn (eds.), English and Medieval Studies: Essays Presented to J. R. R. Tolkien (London: Allen and Unwin), 164–82. Dusinberre, J. 2003: Shakespeare and the Nature of Women, 3rd edn. (Basingstoke: Palgrave Macmillan). Dustin, C. 2014: ‘The Untruth in Relativism’, in M. Baghramian (ed.), The Many Faces of Relativism (London: Routledge), 110–46. Eaglestone, R. 2004: The Holocaust and the Postmodern (Oxford: Oxford University Press). Eagleton, T. 2003: Sweet Violence: the Idea of the Tragic (Oxford: WileyBlackwell). ———. 2005: Holy Terror (Oxford: Oxford University Press). ———. 2010: ‘Lacan’s Antigone’, in S. Wilmer and A. Žukauskaitė (eds.), Interrogating Antigone in Postmodern Philosophy and Criticism (Oxford: Oxford University Press), 101–9. ———. 2015: Hope Without Optimism (New Haven: Yale University Press).

368  Bibliography Easterling, P. E. 1982: Sophocles: Trachiniae (Cambridge: Cambridge University Press). ———. 1984: ‘The Tragic Homer’, Bulletin of the Institute of Classical Studies (London University) 31, 1–8. ———. 1985: ‘Anachronism in Greek Tragedy’, Journal of Hellenic Studies 105, 1–10. ———. 1987: ‘Women in Tragic Space’, Bulletin of the Institute of Classical Studies 34, 15–26. ———. 1990: ‘Constructing Character in Greek Tragedy’, in C. Pelling (ed.), Characterization and Individuality in Greek Literature (Oxford: Clarendon), 83–99. ———. 1993a: ‘Presentation of Character in Aeschylus’, in I. McAuslan and P. Walcott (eds.), Greek Tragedy (Oxford: Oxford University Press), 12–28. ———. 1993b: ‘Character in Sophocles’, in I. McAuslan and P. Walcott (eds.), Greek Tragedy (Oxford: Oxford University Press), 58–65. ———. 1997a: ‘Form and Performance’, in P. E. Easterling (ed.), The Cambridge Companion to Greek Tragedy (Cambridge: Cambridge University Press), 151–77. ———. 1997b: ‘Constructing the Heroic’, in C. Pelling (ed.), Greek Tragedy and the Historian (Oxford: Clarendon), 21–37. Eckermann, J. P. 1981: Gespräche mit Goethe in den letzten Jahren seines Lebens (Frankfurt am Main: Insel). Edmunds, L. 2006: Oedipus (London: Routledge). Edwards, C. 1999: ‘The Suffering Body: Philosophy and Pain in Seneca’s Letters’, in J. Porter (ed.), Constructions of the Classical Body (Ann Arbor: University of Michigan Press), 252–68. Edwards, G. 1999: Lope de Vega: Three Major Plays (Oxford: Oxford University Press). Edwards, M. 1978: ‘Agamemnon’s Decision: Freedom and Folly in Aeschylus’, California Studies in Classical Antiquity 10, 17–38. Edwards, P. 1992: ‘Tragic Balance in Hamlet’, in M. Coyle (ed.), New Casebooks: Hamlet (Basingstoke: Palgrave Macmillan), 19–36. Eliot, G. 1963: Essays of George Eliot, ed. T. Pinney (London: Routledge & Kegan Paul). ———. 1965: Middlemarch, ed. W. J. Harvey (Harmondsworth: Penguin). ———. 1979: The Mill on the Floss, ed. A. S. Byatt (London: Penguin). Eliot, T. S. 1951: Selected Essays, 3rd edn. (London: Faber and Faber). ———. 1964: The Use of Poetry and the Use of Criticism (London: Faber and Faber). ———. 1969: The Complete Poems and Plays of T. S. Eliot (London: Faber and Faber). Elshtain, J. B. 1998: ‘Antigone’s Daughters’, in A. Phillips (ed.), Feminism and Politics (Oxford: Oxford University Press), 363–77. Empson, W. 1961: Milton’s God (Westport, CT: Greenwood Press). ———. 1986: Essays on Shakespeare (Cambridge: Cambridge University Press). ———. 2004: Seven Types of Ambiguity (London: Pimlico). Engle, L. 2013: ‘Marlowe and the Self’, in E. Bartels and E. Smith (eds.), Christopher Marlowe in Context (Cambridge: Cambridge University Press), 202–11. Erasmo, M. 2004: Roman Tragedy: Theatre to Theatricality (Austin: University of Texas Press).

Bibliography 369 Eschenbach, W. von. 2014: Parzifal, 2 vols., ed. K. Lachmann (Stuttgart: Reclam). Esposito, J. 2016: ‘Who Kills Turnus? “Pallas” and What Aeneas Sees, Says, and Does in Aeneid 12.939–52’, Classical Journal 111, 463–81. Euben, J. P. 1982: ‘Justice and the Oresteia’, American Political Science Review 76, 22–33. Euripides. 1913: Euripidis Fabulae, 3 vols., ed. G. Murray (Oxford: Clarendon). ———. 1984–94: Euripidis Fabulae, 3 vols., ed. J. Diggle (Oxford: Clarendon). ———. 1997/2004: Selected Fragmentary Plays, 2 vols., eds. C. Collard, M. Cropp, K. Lee, and J. Gibert (Warminster/Oxford: Aris and Phillips). Evans, G. 1982: The Varieties of Reference, ed. J. McDowell (Oxford: Clarendon). Evers, D. and Deng, N. 2016: ‘Acknowledgement and the Paradox of Tragedy’, Philosophical Studies 173, 337–50. Everson, S. 1990: ‘Aristotle’s Compatibilism in the Nicomachean Ethics’, Ancient Philosophy 10, 81–103. Ewans, M. 1996: ‘Patterns of Tragedy in Sophokles and Shakespeare’, in M. Silk (ed.), Tragedy and the Tragic: Greek Theatre and Beyond (Oxford: Oxford University Press), 438–57. Fabriano, L. da. 1969: ‘De Casu Cesene’, in G. Schizzerotto (ed.), Teatro e cultura in Romagna dal Medioevo al Rinascimento (Ravenna: Edizioni della Rotonda). Farnham, W. 1970: The Medieval Heritage of Elizabethan Tragedy (Oxford: Wiley-Blackwell). Fick, M. 2010: Lessing-Handbuch: Leben—Werk—Wirkung, 3rd edn. (Stuttgart: Metzler). Finglass, P. J. 2007: Sophocles: Electra (Cambridge: Cambridge University Press). ———. 2011: Sophocles: Ajax (Cambridge: Cambridge University Press). Fitch, J. 1987: Seneca’s Hercules Furens (Ithaca: Cornell University Press). Fitch, J. and McElduff, S. 2008: ‘Construction of the Self in Senecan Drama’, in J. Fitch (ed.), Oxford Readings in Seneca (Oxford: Oxford University Press), 157–80. Flaig, E. 1998: Ödipus: Tragischer Vatermord im klassischen Athen (Munich: Beck). Flashar, H. 1984: ‘Die Poetik des Aristoteles und die griechische Tragödie’, Poetica 16, 1–23. ———. 1994: ‘Familie, Mythos, Drama am Beispiel des Oedipus’, Colloquium Helveticum 19, 51–74. ———. 2000: Sophokles: Dichter im demokratischen Athen (Munich: Beck). Fleming, K. 2006: ‘Fascism on Stage: Jean Anouilh’s Antigone’, in V. Zajko and M. Leonard (eds.), Laughing With Medusa: Classical Myth and Feminist Thought (Oxford: Oxford University Press), 163–86. Fletcher, J. 2010: ‘Sophocles’ Antigone and the Democratic Voice’, in S. Wilmer and A. Žukauskaitė (eds.), Interrogating Antigone in Postmodern Philosophy and Criticism (Oxford: Oxford University Press), 168–84. Foakes, R. A. 2000: Shakespeare: King Lear (London: Thomson). Foley, H. 1995: ‘Tragedy and Democratic Ideology: The Case of Sophocles’ Antigone’, in B. Goff (ed.), History, Tragedy, Theory: Dialogues on Athenian Drama (Austin: University of Texas Press), 131–50. ———. 2001: Female Acts in Greek Tragedy (Princeton: Princeton University Press). ———. 2003: ‘Choral Identity in Greek Tragedy’, Classical Philology 98, 1–30.

370  Bibliography ———. 2008: ‘Generic Boundaries in Late Fifth-Century Athens’, in M. Revermann and P. Wilson (eds.), Performance, Iconography, Reception: Studies in Honour of Oliver Taplin (Oxford: Oxford University Press), 15–36. Foley, H. and Howard, J. 2014: ‘The Urgency of Tragedy Now’, PMLA 129, 617–33. Foster, V. 2004: The Name and Nature of Tragicomedy (Aldershot: Ashgate). Fowler, A. 1982: Kinds of Literature: An Introduction to the Theory of Genres and Modes (Oxford: Clarendon). Fowler, D. 2000: Roman Constructions: Readings in Postmodern Latin (Oxford: Oxford University Press). Fraenkel, E. 1957: Horace (Oxford: Clarendon). ———. 1978: Aeschylus: Agamemnon, 3 vols. (Oxford: Clarendon). Frank, M. 1995: Seneca’s Phoenissae (Leiden: Brill). Franke, W. 2014: A Philosophy of the Unsayable (Notre Dame: University of Notre Dame Press). Freeland, C. 2013: Antigone, in her Unbearable Splendor: New Essays on Jacques Lacan’s The Ethics of Psychoanalysis (New York: SUNY Press). French, M. 1992: ‘Chaste Constancy in Hamlet’, in M. Coyle (ed.), New Casebooks: Hamlet (Basingstoke: Palgrave Macmillan), 96–112. Freud, S. 2013: Jenseits des Lustprinzips (Stuttgart: Reclam). Friedländer, S. 2007: Nachdenken über den Holocaust (Munich: C. H. Beck). Friedrich, R. 1996: ‘Everything to Do with Dionysos? Ritualism, the Dionysiac, and the Tragic’, in M. Silk (ed.), Tragedy and the Tragic: Greek Theatre and Beyond (Oxford: Oxford University Press), 256–83. Fritz, K. von. 1962: Antike und Moderne Tragödie: Neue Abhandlungen (Berlin: de Gruyter). Gadamer, H-G. 1990: Wahrheit und Methode: Grundzüge einer philosophischen Hermeneutik (Tübingen: Mohr). Gager, W. 1994: The Complete Works, 2 vols., ed. D. Sutton (New York: Garland). Gardner, S. 2003: ‘Tragedy, Morality, and Metaphysics’, in J. Bermúdez and S. Gardner (eds.), Art and Morality (London: Routledge), 218–59. Garner, S. 1976: ‘Shakespeare’s Desdemona’, Shakespeare Studies 9, 233–52. Garnier, R. 1975: Two Tragedies: Hippolyte and Marc Antoine, eds. C. Hill and M. Morrison (London: Athlone Press). ———. 1997: Antigone ou la Pieté, ed. J-D. Beaudin (Paris: Honoré Champion). ———. 1999: Porcie: Tragédie, ed. J-C. Ternaux (Paris: Honoré Champion). ———. 2002: Cornélie: Tragédie, ed. J-C. Ternaux (Paris: Honoré Champion). Garton, C. 1972: Personal Aspects of the Roman Theatre (Toronto: Hakkert). Garvie, A. 1986: Aeschylus: Choephori (Oxford: Clarendon). ———. 2007: ‘Greek Tragedy: Text and Context’, in P. Finglass, C. Collard and N. Richardson (eds.), Hesperos: Studies in Ancient Greek Poetry Presented to M. L. West on his Seventieth Birthday (Oxford: Oxford University Press), 170–88. ———. 2009: Aeschylus: Persae (Oxford: Oxford University Press). ———. 2016: The Plays of Sophocles, 2nd edn. (London: Bloomsbury). Gaskin, R. 1989: ‘Can Aesthetic Value Be Explained?’ British Journal of Aesthetics 29, 329–40. ———. 1992: ‘Turnus, Mezentius, and the Complexity of Virgil’s Aeneid’, Latomus 217, 295–316.

Bibliography 371 ———. 1994a: ‘Fatalism, Foreknowledge, and the Reality of the Future’, The Modern Schoolman 71, 83–113. ———. 1994b: ‘Molina on Divine Foreknowledge and the Principle of Bivalence’, Journal of the History of Philosophy 32, 27–47. ———. 1995: The Sea Battle and the Master Argument: Aristotle and Diodorus Cronus on the Metaphysics of the Future (Berlin: de Gruyter). ———. 1996: ‘ “Kein Etwas, aber auch nicht ein Nichts!”: Kann die Grammatik tatsächlich täuschen?’ Grazer Philosophische Studien 51, 85–104. ———. 1997: ‘Peter of Ailly and other Fourteenth-Century Thinkers on Divine Power and the Necessity of the Past’, Archiv für Geschichte der Philosophie 79, 273–91. ———. 1998a: ‘Fatalism, Bivalence, and the Past’, Philosophical Quarterly 48, 83–8. ———. 1998b: ‘Middle Knowledge, Fatalism, and Comparative Similarity of Worlds’, Religious Studies 34, 189–203. ———. 2000: ‘Conditionals of Freedom and Middle Knowledge’, in E. Dekker et al. (eds.), Middle Knowledge (Frankfurt am Main: Lang), 137–56. ———. 2001: ‘Do Homeric Heroes Make Real Decisions?’ in D. Cairns (ed.), Oxford Readings in Homer’s Iliad (Oxford: Oxford University Press), 147–69. ———. 2006: Experience and the World’s Own Language: A Critique of John McDowell’s Empiricism (Oxford: Clarendon). ———. 2008: The Unity of the Proposition (Oxford: Oxford University Press). ———. 2009: ‘Realism and the Picture Theory of Meaning’, Philosophical Topics 37, 49–62. ———. 2013a: Language, Truth, and Literature: A Defence of Literary Humanism (Oxford: Oxford University Press). ———. 2013b: Horace and Housman (New York: Palgrave Macmillan). ———. 2018: ‘A Defence of the Resemblance Meaning of “What It’s Like” ’, Mind. ———. [forthcoming]: ‘On Being Pessimistic About the End of the Aeneid’, Harvard Studies in Classical Philology. Gelfert, H-D. 1995: Die Tragödie: Theorie und Geschichte (Göttingen: Vandenhoeck and Ruprecht). Gellrich, M. 1988: Tragedy and Theory: The Problem of Conflict Since Aristotle (Princeton: Princeton University Press). ———. 1995: ‘Interpreting Greek Tragedy: History, Theory, and the New Philology’, in B. Goff (ed.), History, Tragedy, Theory: Dialogues on Athenian Drama (Austin: University of Texas Press), 38–58. Gerstenberg, H. W. von. 2001: Ugolino (Stuttgart: Reclam). Gervinus, G. 1872: Shakespeare, 4th edn., 2 vols. (Leipzig: Engelmann). Gide, A. 2007: Œdipe, ed. C. Debard (Paris: Honoré Champion). Gill, C. 1986: ‘The Question of Character and Personality in Greek Tragedy’, Poetics Today 7, 251–73. ———. 1990: ‘The Character–Personality Distinction’, in C. Pelling (ed.), Characterization and Individuality in Greek Literature (Oxford: Clarendon), 1–31. ———. 1996: Personality in Greek Epic, Tragedy, and Philosophy: The Self in Dialogue (Oxford: Clarendon). ———. 2009: ‘Seneca and Selfhood: Interpretation and Disintegration’, in S. Bartsch and D. Wray (eds.), Seneca and the Self (Cambridge: Cambridge University Press), 65–83.

372  Bibliography Gillett, G. and Hankey, R. 2005: ‘Oedipus the King: Temperament, Character, and Virtue’, Philosophy and Literature 29, 269–85. Girard, R. 2002: ‘Collective Violence and Sacrifice in Julius Caesar’, in R. Wilson (ed.), New Casebooks: Julius Caesar (Basingstoke: Palgrave Macmillan), 108–27. ———. 2010: La violence et le sacré (Paris: Pluriel). Glock, H-J. 2008: ‘Relativism, Commensurability, and Translatability’, in J. Preston (ed.), Wittgenstein and Reason (Oxford: Wiley-Blackwell), 21–46. Glover, J. 2001: Humanity: A Moral History of the Twentieth Century (London: Pimlico). Glüer, K. 2006: ‘The Status of Charity I: Conceptual Truth or a Posteriori Necessity?’ International Journal of Philosophical Studies 14, 337–59. Goethe, J. W. 1985–99: Sämtliche Werke, Briefe, Tagebücher und Gespräche, 40 vols., eds. F. Apel et al. (Frankfurt am Main: Deutscher Klassiker Verlag). Goff, B. 1990: The Noose of Words: Readings of Desire, Violence, and Language in Euripides’ Hippolytus (Cambridge: Cambridge University Press). ———. 2000: ‘Try to Make it Real Compared to What? Euripides’ Electra and the Play of Genres’, in M. Cropp, K. Lee and D. Sansone (eds.), Euripides and Tragic Theatre in the Late Fifth Century (Illinois: Stipes), 93–105. Goldberg, J. 1993: ‘Perspectives: Dover Cliff and the Conditions of Representation’, in K. Ryan (ed.), New Casebooks: King Lear (Basingstoke: Palgrave Macmillan), 145–57. Goldhill, S. 1984a: Language, Sexuality, Narrative: The Oresteia (Cambridge: Cambridge University Press). ———. 1984b: ‘Exegesis: Oedipus (R)ex’, Arethusa 17, 177–200. ———. 1986a: Reading Greek Tragedy (Cambridge: Cambridge University Press). ———. 1986b: ‘Goldhill on Molehills’, Liverpool Classical Monthly 11.10, 163–7. ———. 1990a: ‘Character and Action, Representation and Reading: Greek Tragedy and Its Critics’, in C. Pelling (ed.), Characterization and Individuality in Greek Literature (Oxford: Clarendon), 100–27. ———. 1990b: ‘The Great Dionysia and Civic Ideology’, in J. Winkler and F. Zeitlin (eds.), Nothing to Do with Dionysus? Athenian Drama and Its Social Context (Princeton: Princeton University Press), 97–129. ———. 1996: ‘Collectivity and Otherness—The Authority of the Tragic Chorus: Response to Gould’, in M. Silk (ed.), Tragedy and the Tragic: Greek Theatre and Beyond (Oxford: Oxford University Press), 244–56. ———. 1997a: ‘The Audience of Athenian Tragedy’, in P. E. Easterling (ed.), The Cambridge Companion to Greek Tragedy (Cambridge: Cambridge University Press), 54–68. ———. 1997b: ‘The Language of Tragedy: Rhetoric and Communication’, in P. E. Easterling (ed.), The Cambridge Companion to Greek Tragedy (Cambridge: Cambridge University Press), 127–50. ———. 1997c: ‘Modern Critical Approaches to Greek Tragedy’, in P. E. Easterling (ed.), The Cambridge Companion to Greek Tragedy (Cambridge: Cambridge University Press), 324–47. ———. 2000: ‘Civic Ideology and the Problem of Difference: The Politics of Aeschylean Tragedy, Once Again’, Journal of Hellenic Studies 120, 34–56. ———. 2006: ‘Der Ort der Gewalt: Was sehen wir auf der Bühne?’, in B. Seidensticker and M. Vöhler (eds.), Gewalt und Ästhetik: Zur Gewalt and ihrer Darstellung in der griechischen Klassik (Berlin: de Gruyter), 149–68.

Bibliography 373 ———. 2012: Sophocles and the Language of Tragedy (Oxford: Oxford University Press). ———. 2014: ‘Antigone: An Interruption Between Feminism and Christianity’, International Journal of the Classical Tradition 21, 309–16. ———. 2015: ‘The Ends of Tragedy: Schelling, Hegel, and Oedipus’, in J. Billings and M. Leonard (eds.), Tragedy and the Idea of Modernity (Oxford: Oxford University Press), 231–50. Goldhill, S., and Hall, E. (eds.) 2009: Sophocles and the Greek Tragic Tradition (Cambridge: Cambridge University Press). Goldmann, L. 1959: Le dieu caché: Étude sur la vision tragique dans les Pensées de Pascal et dans le théâtre de Racine (Paris: Gallimard). ———. 1970: Racine: Essai (Paris: L’Arche). Gooch, J. 1981: The Lamentable Tragedy of Locrine: a Critical Edition (New York: Garland). Goodkin, R. 2009: ‘Neoclassical Dramatic Theory in Seventeenth-Century France’, in R. Bushnell (ed.), A Companion to Tragedy (Oxford: Wiley-Blackwell), 373–92. Görner, R. 2015: ‘The (Operatic) Tragedy of Culture: Notes on a Theme in Kierkegaard, Hebbel, and Wagner’, in J. Billings and M. Leonard (eds.), Tragedy and the Idea of Modernity (Oxford: Oxford University Press), 159–77. Gossip, C. 1981: An Introduction to French Classical Tragedy (London: Palgrave Macmillan). Gottsched, J. C. 2009: Sterbender Cato, ed. H. Steinmetz (Stuttgart: Reclam). Gould, J. 2001: Myth, Ritual, Memory, and Exchange: Essays in Greek Literature and Culture (Oxford: Oxford University Press). Gould, T. 1965/1966: ‘The Innocence of Oedipus: Philosophers on Oedipus the King’, Arion 4, 363–86; 582–611; 1966, 478–525. Goux, J-J. 1990: Œdipe Philosophe (Paris: Aubier). Gower, J. 1857: Confessio Amantis, ed. R. Pauli (London: Bell and Daldy). Grady, H. 2003: ‘Timon of Athens: The Dialectic of Usury, Nihilism, and Art’, in R. Dutton and J. Howard (eds.), A Companion to Shakespeare’s Works: The Tragedies (Oxford: Wiley-Blackwell), 430–51. ———. 2009: ‘Tragedy and Materialist Thought’, in R. Bushnell (ed.), A Companion to Tragedy (Oxford: Wiley-Blackwell), 128–44. ———. 2014: ‘The Modernity of Western Tragedy: Genealogy of a Developing Anachronism’, PMLA 129, 790–8. Gransden, K. 1976: Virgil: Aeneid Book VIII (Cambridge: Cambridge University Press). Gray, J. 2007: Enlightenment’s Wake: Politics and Culture at the Close of the Modern Age (London: Routledge). Greenberg, M. 2009: ‘French Neoclassical Tragedy: Corneille/Racine’, in R. Bushnell (ed.), A Companion to Tragedy (Oxford: Wiley-Blackwell), 393–410. ———. 2010: Racine: From Ancient Myth to Tragic Modernity (Minneapolis: University of Minnesota Press). Greenblatt, S. 1980: Renaissance Self-Fashioning: From More to Shakespeare (Chicago: Chicago University Press). ———. 1990: Learning to Curse: Essays in Early Modern Culture (London: Routledge). ———. 2001: Hamlet in Purgatory (Princeton: Princeton University Press).

374  Bibliography ———. 2005: Will in the World: How Shakespeare Became Shakespeare (London: Pimlico). Greg, W. W. 1917: ‘Hamlet’s Hallucination’, Modern Language Review 12, 393–421. Gregory, J. 1977: ‘Euripides’ Heracles’, Yale Classical Studies 25, 259–75. ———. 1997: Euripides and the Instruction of the Athenians (Ann Arbor: University of Michigan Press). ———. 1999: Euripides: Hecuba (Atlanta, Georgia: Scholars Press). ———. 2000: ‘Comic Elements in Euripides’, in M. Cropp, K. Lee and D. Sansone (eds.), Euripides and Tragic Theatre in the Late Fifth Century (Champaign, IL: Stipes), 59–74. ———. 2005: ‘Euripidean Tragedy’, in J. Gregory (ed.), A Companion to Greek Tragedy (Oxford: Wiley-Blackwell), 251–70. Gregory, T. 2010: ‘The Political Message of Samson Agonistes’, Studies in English Literature 50, 175–203. Greif, G. 2011: ‘Wir weinten tränenlos. . . ’: Augenzeugenberichte des jüdischen ‘Sonderkommandos’ in Auschwitz, tr. M. Schmidt (Frankfurt am Main: Fischer). Greiner, B. 2012: Die Tragödie: Eine Literaturgeschichte des aufrechten Ganges, Grundlagen und Interpretationen (Stuttgart: Kröner). Griffin, J. 1995: Homer: Iliad IX (Oxford: Clarendon). ———. 1998: ‘The Social Function of Attic Tragedy’, Classical Quarterly 48, 39–61. ———. 1999: ‘Sophocles and the Democratic City’, in J. Griffin (ed.), Sophocles Revisited: Essays Presented to Sir Hugh Lloyd-Jones (Oxford: Oxford University Press), 73–94. ———. 2007: ‘Desperate Straits and the Tragic Stage’, in P. Finglass, C. Collard and N. Richardson (eds.), Hesperos: Studies in Ancient Greek Poetry Presented to M. L. West on His Seventieth Birthday (Oxford: Oxford University Press), 189–203. Griffith, M. 1983: Aeschylus: Prometheus Bound (Cambridge: Cambridge University Press). ———. 1999: Sophocles: Antigone (Cambridge: Cambridge University Press). ———. 2001: ‘Antigone and Her Sister(s): Embodying Women in Greek Tragedy’, in A. Lardinois and L. McClure (eds.), Making Silence Speak: Women’s Voices in Greek Literature and Society (Princeton: Princeton University Press), 117–36. ———. 2005a: ‘Authority Figures’, in J. Gregory (ed.), A Companion to Greek Tragedy (Oxford: Wiley-Blackwell), 333–51. ———. 2005b: ‘The Subject of Desire in Sophocles’ Antigone’, in V. Pedrick and S. Oberhelman (eds.), The Soul of Tragedy: Essays on Athenian Drama (Chicago: University of Chicago Press), 91–135. ———. 2010: ‘Psychoanalysing Antigone’, in S. Wilmer and A. Žukauskaitė (eds.), Interrogating Antigone in Postmodern Philosophy and Criticism (Oxford: Oxford University Press), 110–34. Grund, G. 2011: Humanist Tragedies (Cambridge, MA: Harvard University Press). ———. 2015a: ‘Tragedy’, in P. Cheney and P. Hardie (eds.), The Oxford History of Classical Reception in English Literature, vol. 2 (Oxford: Oxford University Press), 373–94.

Bibliography 375 ———. 2015b: ‘Tragedy’, in S. Knight and S. Tilg (eds.), The Oxford Handbook of Neo-Latin (Oxford: Oxford University Press), 103–17. Guarini, B. 1999: Il Pastor Fido, ed. E. Selmi (Venice: Marsilio). Guthke, K. 2012–13: ‘ “Der Güter höchstes nicht”? Lessing und Mendelssohn im Gespräch über den Selbstmord’, Lessing Yearbook 40, 31–52. Haas, R. 1987: ‘Chaucer’s Monk’s Tale: An Ingenious Criticism of Early Humanist Conceptions of Tragedy’, Humanistica Lovaniensia 36, 44–70. Hacking, I. 1999: The Social Construction of What? (Cambridge, MA: Harvard University Press). ———. 2008: ‘Conclusion: Deflections’, in S. Cavell et al., Philosophy and Animal Life (New York: Columbia University Press), 139–72. Hagmaier, A. 2006: M. A. Muret, Iulius Caesar; M. Virdung, Brutus: Zwei neulateinische Tragödien (Leipzig: Saur). Hall, E. 1996: ‘Is There a Polis in Aristotle’s Poetics?’ in M. Silk (ed.), Tragedy and the Tragic: Greek Theatre and Beyond (Oxford: Oxford University Press), 295–309. ———. 1997: ‘The Sociology of Athenian Tragedy’, in P. E. Easterling (ed.), The Cambridge Companion to Greek Tragedy (Cambridge: Cambridge University Press), 93–126. ———. 2009: ‘Deianeira Deliberates: Precipitate Decision-Making and Trachiniae’, in S. Goldhill and E. Hall (eds.), Sophocles and the Greek Tragic Tradition (Cambridge: Cambridge University Press), 69–96. ———. 2012: ‘The Necessity and Limits of Deliberation in Sophocles’ Theban Plays’, in K. Ormand (ed.), A Companion to Sophocles (Oxford: WileyBlackwell), 301–15. ———. 2014: ‘To Fall from High or Low Estate? Tragedy and Social Class in Historical Perspective’, PMLA 129, 773–82. Halleran, M. 1986: ‘Rhetoric, Irony, and the Ending of Euripides’ Herakles’, Classical Antiquity 5, 171–81. Halliwell, S. 1990a: ‘Traditional Greek Conceptions of Character’, in C. Pelling (ed.), Characterization and Individuality in Greek Literature (Oxford: Clarendon), 32–59. ———. 1990b: ‘Human Limits and the Religion of Greek Tragedy’, Journal of Literature and Theology 4, 169–80. ———. 1995: ‘Tragedy, Reason, and Pity: A Reply to Jonathan Lear’, in R. Heinaman (ed.), Aristotle and Moral Realism (Boulder, CO: Westview), 85–95. ———. 1996: ‘Plato’s Repudiation of the Tragic’, in M. Silk (ed.), Tragedy and the Tragic: Greek Theatre and Beyond (Oxford: Oxford University Press), 332–49. ———. 1998: Aristotle’s Poetics, 2nd edn. (London: Duckworth). ———. 2005: ‘Learning from Suffering: Ancient Responses to Tragedy’, in J. Gregory (ed.), A Companion to Greek Tragedy (Oxford: Wiley-Blackwell), 394–412. ———. 2006: ‘Plato and Aristotle on the Denial of Tragedy’, in A. Laird (ed.), Oxford Readings in Ancient Literary Criticism (Oxford: Oxford University Press), 115–41. ———. 2012: Between Ecstasy and Truth: Interpretations of Greek Poetics from Homer to Longinus (Oxford: Oxford University Press).

376  Bibliography Halpern, R. 2016: ‘The Classical Inheritance’, in M. Neill and D. Schalkwyk (eds.), The Oxford Handbook of Shakespearean Tragedy (Oxford: Oxford University Press), 19–34. Hamilton, C. 2016: A Philosophy of Tragedy (London: Reaktion). Hamilton, P. 2003: Historicism, 2nd edn. (London: Routledge). Hamilton, S. 2003: Shakespeare’s Daughters (Jefferson/London: McFarland). Hammond, N. 1973: Studies in Greek History (Oxford: Clarendon). Hammond, P. 2009: The Strangeness of Tragedy (Oxford: Oxford University Press). Happé, P. 1972: Tudor Interludes (Harmondsworth: Penguin). Hardie, P. 1993: The Epic Successors of Virgil: A Study in the Dynamics of a Tradition (Cambridge: Cambridge University Press). ———. 1997: ‘Virgil and Tragedy’, in C. Martindale (ed.), The Cambridge Companion to Virgil (Cambridge: Cambridge University Press), 312–26. ———. 2009: Lucretian Receptions: History, The Sublime, Knowledge (Cambridge: Cambridge University Press). Hardy, T. 1965: The Dynasts: An Epic Drama of the War with Napoleon (London: Palgrave Macmillan). ———. 1967: Tess of the D’Urbervilles (London: Palgrave Macmillan). ———. 1971: Jude the Obscure (London: Palgrave Macmillan). ———. 1972: Far from the Madding Crowd (London: Palgrave Macmillan). ———. 1974: The Return of the Native (London: Palgrave Macmillan). ———. 1997: The Pursuit of the Well-Beloved and The Well-Beloved (London: Penguin). ———. 1998: The Woodlanders (London: Penguin). ———. 2001: The Complete Poems (Basingstoke: Palgrave Macmillan). Harman, G. 1990: ‘The Intrinsic Quality of Experience’, Philosophical Perspectives 4, 31–52. Harris, E. 2004: ‘Antigone the Lawyer, or the Ambiguities of Nomos’, in E. Harris and L. Rubinstein (eds.), The Law and the Courts in Ancient Greece (London: Duckworth), 19–56. ———. 2012: ‘Sophocles and Athenian Law’, in K. Ormand (ed.), A Companion to Sophocles (Oxford: Wiley-Blackwell), 287–300. Harrison, S. 2007: ‘Exile in Latin Epic’, in J. F. Gaertner (ed.), Writing Exile: The Discourse of Displacement in Greco-Roman Antiquity and Beyond (Leiden: Brill), 129–54. Harsh, P. 1945: ‘Hamartia Again’, Transactions of the American Philological Association 76, 47–58. Harth, E. 1972: ‘The Tragic Moment in Athalie’, Modern Language Quarterly 33, 382–95. Hasker, W. 2011: ‘Divine Knowledge and Human Freedom’, in R. Kane (ed.), The Oxford Handbook of Free Will, 2nd edn. (Oxford: Oxford University Press), 39–54. Haslanger, S. 2012: Resisting Reality: Social Construction and Social Critique (Oxford: Oxford University Press). Hatfield, H. 1956: ‘Emilia’s Guilt Once More’, Modern Language Notes 71, 287–96. Havelock, E. 1978: The Greek Concept of Justice from its Shadow in Homer to Its Substance in Plato (Cambridge, MA: Harvard University Press).

Bibliography 377 Hawcroft, M. 1992: Word as Action: Racine, Rhetoric, and Theatrical Language (Oxford: Clarendon). Heath, M. 1987: The Poetics of Greek Tragedy (London: Duckworth). ———. 1991: ‘The Universality of Poetry in Aristotle’s Poetics’, Classical Quarterly 41, 389–402. ———. 2001: ‘Aristotle and the Pleasures of Tragedy’, in Ø. Andersen and J. Haarberg (eds.), Making Sense of Aristotle: Essays in Poetics (London: Duckworth), 7–23. ———. 2006: ‘The “Social Function” of Tragedy: Clarifications and Questions’, in D. Cairns and V. Liapis (eds.), Dionysalexandros: Essays on Aeschylus and His Fellow Tragedians in Honour of Alexander F. Garvie (Swansea: Classical Press of Wales), 253–81. ———. 2009: ‘Should There Have Been a Polis in Aristotle’s Poetics?’ Classical Quarterly 59, 468–85. ———. 2011: ‘Response to Burian, Hesk, and Barker’, in D. M. Carter (ed.), Why Athens? A Reappraisal of Tragic Politics (Oxford: Oxford University Press), 163–71. ———. 2017: ‘Aristotle on the Best Kind of Tragic Plot: Re-reading Poetics 13–14’, in R. Polansky and W. Wians (eds.), Reading Aristotle: Argument and Exposition (Leiden: Brill), 334–51. Hegel, G. W. F. 1986: Werke, 20 vols., eds. E. Moldenhauer and K. Markus (Frankfurt am Main: Suhrkamp). Heidegger, M. 1987: Einführung in die Metaphysik (Tübingen: Niemeyer). ———. 1990: Martin Heidegger/Elisabeth Brockmann: Briefwechsel 1918–1969, ed. J. W. Storck (Marbach am Neckar: Deutsches Literaturarchiv). ———. 1993: Hölderlins Hymne ‘Der Ister’ (Frankfurt am Main: Klostermann). Heinze, R. 1957: Vergils Epische Technik (Stuttgart: Teubner). Heisenberg, W. 1996: Der Teil und das Ganze: Gespräche im Umkreis der Atomphysik (Munich: Piper). Hejduk, J. 2013: ‘The Bough and the Lock: Fighting Fate in the Aeneid’, Illinois Classical Studies 38, 149–57. Heller, A. 2002: The Time Is Out of Joint: Shakespeare as Philosopher of History (Lanham: Rowman & Littlefield). Hellie, B. 2004: ‘Inexpressible Truths and the Allure of the Knowledge Argument’, in P. Ludlow et al. (eds.), There’s Something About Mary: Essays on Phenomenal Consciousness and Frank Jackson’s Knowledge Argument (Cambridge, MA: MIT Press), 333–64. Hemingway, E. 2004: For Whom the Bell Tolls (London: Arrow). Hendricks, M. 2000: ‘ “A Word, Sweet Lucrece”: Confession, Feminism, and The Rape of Lucrece’, in D. Callaghan (ed.), A Feminist Companion to Shakespeare (Oxford: Wiley-Blackwell), 103–18. Henrichs, A. 1986: ‘The Last of the Detractors: Friedrich Nietzsche’s Condemnation of Euripides’, Greek, Roman, and Byzantine Studies 27, 369–97. ———. 1994: ‘ “Why Should I Dance?”: Choral Self-Referentiality in Greek Tragedy’, Arion 3, 56–111. ———. 2005: ‘Nietzsche on Greek Tragedy and the Tragic’, in J. Gregory (ed.), A Companion to Greek Tragedy (Oxford: Wiley-Blackwell), 444–58. Henryson, R. 1981: The Poems of Robert Henryson, ed. D. Fox (Oxford: Clarendon).

378  Bibliography Herder, J. G. 1985–2000: Werke, 10 vols., eds. M. Bollacher et al. (Frankfurt am Main: Deutscher Klassiker Verlag). Herrick, M. 1955: Tragicomedy: Its Origin and Development in Italy, France, and England (Urbana: University of Illinois Press). Hester, D. A. 1971: ‘Sophocles the Unphilosophical: A Study in the Antigone’, Mnemosyne 24, 11–59. Heywood, T. 1888: The Best Plays of the Old Dramatists: Thomas Heywood, ed. A. Verity (London: Vizetelly). Hinds, S. 1998: Allusion and Intertext: Dynamics of Appropriation in Roman Poetry (Cambridge: Cambridge University Press). Hirst, D. 1984: Tragicomedy (London: Methuen). Hochhut, R. 1998: Der Stellvertreter: ein christliches Trauerspiel (Hamburg: Rowohlt). Hofmannsthal, H. von. 1979: Gesammelte Werke, 10 vols., ed. B. Schoeller (Frankfurt am Main: Fischer). Hogan, J. 1972: ‘The Protagonists of the Antigone’, Arethusa 5, 93–100. Holbrook, P. 2015: English Renaissance Tragedy (London: Bloomsbury). Hölderlin, F. 1992–4: Sämtliche Werke und Briefe, 3 vols., ed. J. Schmidt (Frankfurt am Main: Deutscher Klassiker Verlag). Holland, C. 2010: ‘After Antigone: Women, the Past, and the Future of Feminist Political Thought’, in F. Söderbäck (ed.), Feminist Readings of the Antigone (New York: SUNY Press), 27–43. Holland, P. 1994: ‘Theseus’ Shadows in A Midsummer Night Dream’, Shakespeare Survey 47, 139–51. Holmes, P. 1989: Debussy (London: Omnibus). Holstun, J. 1989: ‘Ranting at the New Historicism’, English Literary Renaissance 19, 189–225. Holt, P. 1999: ‘Polis and Tragedy in the Antigone’, Mnemosyne 52, 658–90. Honig, B. 2013: Antigone Interrupted (Cambridge: Cambridge University Press). ———. 2014: ‘Antigone’s Laments, Creon’s Grief’, in T. Chanter and S. Kirkland (eds.), The Returns of Antigone: Interdisciplinary Essays (New York: SUNY Press), 25–49. Honigmann, E. A. J. 1997: Shakespeare: Othello (London: Cengage). Horsfall, N. 2005: Virgil, Aeneid 3: A Commentary (Leiden: Brill). Houlgate, S. 2007: ‘Hegel’s Theory of Tragedy’, in S. Houlgate (ed.), Hegel and the Arts (Evanston: Northwestern University Press), 146–78. Housman, A. E. 1989: Collected Poems and Selected Prose, ed. C. Ricks (London: Penguin). ———. 1997: The Poems of A. E. Housman, ed. A. Burnett (Oxford: Clarendon). Howard, J. 1986: ‘The New Historicism in Renaissance Studies’, English Literary Renaissance 16, 13–43. Hoxby, B. 2015: What Was Tragedy? Theory and the Early Modern Canon (Oxford: Oxford University Press). Hughes, S. 2015: ‘Schiller on the Pleasure of Tragedy’, British Journal of Aesthetics 55, 417–32. Hume, D. 1975: An Enquiry Concerning Human Understanding, and An Enquiry Concerning the Principles of Morals, 3rd edn., eds. L. A. Selby-Bigge and P. H. Nidditch (Oxford: Clarendon).

Bibliography 379 ———. 1987: Essays Moral, Political, and Literary, ed. E. Miller (Indianapolis: Liberty Fund). ———. 2007: Dialogues Concerning Natural Religion and Other Writings, ed. D. Coleman (Cambridge: Cambridge University Press). Hutchinson, G. 1985: Aeschylus: Septem Contra Thebas (Oxford: Clarendon). Huxley, A. 1938: Music At Night and Other Essays (London: Chatto and Windus). Ibsen, H. 1950: Hedda Gabler and Other Plays, tr. U. Ellis-Fermor (Harmondsworth: Penguin). ———. 1965: The League of Youth, A Doll’s House, The Lady from the Sea, tr. P. Watts (Harmondsworth: Penguin). Irigaray, L. 2013: In the Beginning She Was (London: Bloomsbury). Jackson, F. 1998: From Metaphysics to Ethics: A Defence of Conceptual Analysis (Oxford: Clarendon). ———. 2001: ‘Responses’, Philosophy and Phenomenological Research 62, 653–64. James, H. 1947: The Portrait of a Lady (Oxford: Oxford University Press). Jankélévitch, V. 1983: La Musique et l’Ineffable (Paris: Seuil). Jardine, L. 1996: Reading Shakespeare Historically (London: Routledge). Jarrett-Kerr, C. 1965: ‘The Conditions of Tragedy’, Comparative Literary Studies 2, 363–74. Jebb, R. 1883–1908: Sophocles: The Plays and Fragments, 7 vols. (Cambridge: Cambridge University Press). Jenkins, H. 1982: Shakespeare: Hamlet (London: Methuen). Jenkyns, R. 1998: Virgil’s Experience: Nature and History: Times, Names, and Places (Oxford: Clarendon). Johnson, P. 1997: ‘Woman’s Third Face: A Psycho/Social Reconsideration of Sophocles’ Antigone’, Arethusa 30, 369–98. Johnson, S. 1958–77: The Yale Edition of the Works of Samuel Johnson, 23 vols., ed. R. DeMaria (New Haven: Yale University Press). Jones, J. 1962: On Aristotle and Greek Tragedy (London: Chatto and Windus). Jonson, B. 2012: The Cambridge Edition of the Works of Ben Jonson, 7 vols., ed. D. Bevington (Cambridge: Cambridge University Press). Kafka, F. 1987: Der Prozess (Frankfurt am Main: Fischer). ———. 1990: Tagebücher, eds. H-G. Koch et al. (Frankfurt am Main: Fischer); tr. as The Diaries of Franz Kafka, ed. M. Brod (London: Penguin, 1972). Kahn, C. 1997: Roman Shakespeare (London: Routledge). Kamerbeek, J. 1966: ‘Unity and Meaning of Euripides’ Heracles’, Mnemosyne 19, 1–16. Kames [Henry Home]. 2005: Elements of Criticism, 2 vols., ed. P. Jones (Indianapolis: Liberty Fund). Kant, I. 2006: Kritik der Urteilskraft, ed. H. Klemme (Hamburg: Meiner). Kappl, B. 2015: ‘Hamartia und Error—Das Konzept Tragischer Verfehlung bei Aristoteles und seinen Kommentatoren im Cinquecento’, in R. Toepfer and G. Radke-Uhlmann (eds.), Tragik vor der Moderne (Heidelberg: Winter), 47–74. Kastan, D. S. 1999: Shakespeare After Theory (London: Routledge). ———. 2003: ‘ “A Rarity Most Beloved”: Shakespeare and the Idea of Tragedy’, in R. Dutton and J. Howard (eds.), A Companion to Shakespeare’s Works: The Tragedies (Oxford: Wiley-Blackwell), 4–22.

380  Bibliography Kaufmann, W. 1968: Tragedy and Philosophy (Princeton: Princeton University Press). Kearney, M. and Schraer, M. 1988: ‘The Flaw in Troilus’, Chaucer Review 22, 185–91. Kelly, H. 1989: ‘Chaucer and Shakespeare on Tragedy’, Leeds Studies in English 20, 191–206. ———. 1993: Ideas and Forms of Tragedy from Aristotle to the Middle Ages (Cambridge: Cambridge University Press). ———. 1997: Chaucerian Tragedy (Cambridge: D. S. Brewer). Kennedy, D. 1992: ‘ “Augustan” and “Anti-Augustan”: Reflections on Terms of Reference’, in A. Powell (ed.), Roman Poetry and Propaganda in the Age of Augustus (Bristol: Bristol Classical Press), 26–58. Kermode, F. 2001: Shakespeare’s Language (London: Penguin). Kerrigan, J. 1998: Revenge Tragedy: Aeschylus to Armageddon (Oxford: Oxford University Press). Kierkegaard, S. 1987: Either/Or, 2 vols., eds. H. and E. Hong (Princeton: Princeton University Press). Kirkwood, G. M. 1994: A Study of Sophoclean Drama (Ithaca: Cornell University Press). Kitto, H. D. F. 1960: Form and Meaning in Drama (London: Methuen). ———. 1961: Greek Tragedy: a Literary Study (New York: Barnes and Noble). Klee, E., Dreßen, W., and Rieß, V. 1988: ‘Schöne Zeiten’: Judenmord aus der Sicht der Täter und Gaffer, 3rd edn. (Frankfurt am Main: Fischer). Kleist, H. von. 1987–97: Sämtliche Werke und Briefe, 4 vols., ed. I-M. Barth et al. (Frankfurt am Main: Deutscher Klassiker Verlag). Klinger, F. 1972: Die Zwillinger (Stuttgart: Reclam). Knights, L. C. 1964: Explorations: Essays in Criticism Mainly on the Literature of the Seventeenth Century (Harmondsworth: Penguin). ———. 1965: Further Explorations: Essays in Criticism (London: Chatto and Windus). Knox, B. 1960: ‘Sophocles’ Oedipus’, in C. Brooks (ed.), Tragic Themes in Western Literature (New Haven: Yale University Press), 7–29. ———. 1964: The Heroic Temper: Studies in Sophoclean Tragedy (Berkeley: University of California Press). ———. 1998: Oedipus at Thebes: Sophocles’ Tragic Hero and His Time (New Haven: Yale University Press). Kölbel, M. 2011: ‘Global Relativism and Self-Refutation’, in S. Hales (ed.), A Companion to Relativism (Oxford: Wiley-Blackwell), 11–30. Konstan, D. 2005: ‘Pity and Politics’, in R. Sternberg (ed.), Pity and Power in Ancient Athens (Cambridge: Cambridge University Press), 48–66. ———. 2006: The Emotions of the Ancient Greeks: Studies in Aristotle and Classical Literature (Toronto: University of Toronto Press). Korzeniewski, U. 2003: ‘Sophokles! Die Alten! Philoktet!’: Lessing und die antiken Dramatiker (Konstanz: Universitätsverlag). Kottman, P. 2016: ‘What Is Shakespearean Tragedy?’ in M. Neill and D. Schalkwyk (eds.), The Oxford Handbook of Shakespearean Tragedy (Oxford: Oxford University Press), 3–18. Kovacs, D. 1987: The Heroic Muse: Studies in the Hippolytus and Hecuba of Euripides (Baltimore: Johns Hopkins University Press).

Bibliography 381 ———. 1999: Euripides: Trojan Women, Iphigenia Among the Taurians, Ion (Cambridge, MA: Harvard University Press). ———. 2001: Cyclops, Alcestis, Medea (Cambridge, MA: Harvard University Press). ———. 2002a: Euripides: Helen, Phoenician Women, Orestes (Cambridge, MA: Harvard University Press). ———. 2002b: Euripides: Bacchae, Iphigeneia at Aulis, Rhesus (Cambridge, MA: Harvard University Press). ———. 2005: Euripides: Children of Heracles, Hippolytus, Andromache, Hecuba (Cambridge, MA: Harvard University Press). ———. 2009: ‘The Role of Apollo in Oedipus Tyrannus’, in J. Cousland and J. Hume (eds.), The Play of Texts and Fragments: Essays in Honour of Martin Cropp (Leiden: Brill), 357–68. Kragelund, P. 2016: Roman Historical Drama: The Octavia in Antiquity and Beyond (Oxford: Oxford University Press). Krewet, M. 2015: ‘Das Tragische im Handeln des Sophokleischen Philoktet’, in R. Toepfer and G. Radke-Uhlmann (eds.), Tragik vor der Moderne (Heidelberg: Winter), 103–42. Krieger, M. 1960: The Tragic Vision: The Confrontation of Extremity (Baltimore: Johns Hopkins University Press). Kühn, D. 2007: Schillers Schreibtisch in Buchenwald (Frankfurt am Main: Fischer). Kühn, W. 1971: Götterszenen bei Vergil (Heidelberg: Winter). Kühner, R. and Gerth, R. 1898/1904: Ausführliche Grammatik der Griechischen Sprache, Teil II, 2 vols. (Hanover: Hannsche Buchhandlung). Kullmann, W. 1994: ‘Die Reaktionen auf die Orakel und ihre Erfüllung im König Ödipus des Sophokles’, in A. Bierl and P. von Möllendorff (eds.), Orchestra: Drama, Mythos, Bühne (Stuttgart/Leipzig: Teubner), 105–18. Küpper, J. 2014: ‘Das Tragische der Racineschen Phèdre’, in N. Bender et al. (eds.), Ethos und Form der Tragödie (Heidelberg: Winter), 295–313. Lacan, J. 1986: Le Séminaire Livre VII: L’éthique de psychanalyse (Paris: Seuil). Laird, A. 2006: ‘The Value of Ancient Literary Criticism’, in A. Laird (ed.), Oxford Readings in Ancient Literary Criticism (Oxford: Oxford University Press), 1–36. Lamarque, P. 2004: ‘Tragedy and Moral Value’, in P. Lamarque and S. H. Olsen (eds.), Aesthetics and the Philosophy of Art: the Analytic Tradition (Oxford: Wiley-Blackwell), 274–82. Lamb, C. 1903: The Works of Charles and Mary Lamb, 6 vols., ed. E. V. Lucas (London: Methuen). ———. 1975–8: The Letters of Charles and Mary Lamb, 3 vols., ed. E. Marrs (Ithaca: Cornell University Press). Lambropoulos, V. 2006: The Tragic Idea (London: Duckworth). Lamport, F. 1981a: Lessing and the Drama (Oxford: Clarendon). ———. 1981b: ‘ “Direction, Which Thou Canst Not See”: Chance, Providence, and Faith in Lessing’, Oxford German Studies 12, 1–17. ———. 1990: German Classical Drama (Cambridge: Cambridge University Press). Lane, W. and Lane, A. 1986: ‘The Politics of Antigone’, in J. P. Euben (ed.), Greek Tragedy and Political Theory (Berkeley: University of California Press), 162–82.

382  Bibliography Langer, L. 1995: Admitting the Holocaust: Collected Essays (Oxford: Oxford University Press). Langland, W. 2006: Piers Plowman, eds. E. Robertson and S. Shepherd (New York: Norton). Lardinois, A. 2006: ‘The Polysemy of Gnomic Expressions and Ajax’ Deception Speech’, in I. J. F. De Jong and A. Rijksbaron (eds.), Sophocles and the Greek Language: Aspects of Diction, Syntax, and Pragmatics (Leiden: Brill), 213–23. ———. 2012: ‘Antigone’, in K. Ormand (ed.), A Companion to Sophocles (Oxford: Wiley-Blackwell), 55–68. Lattimore, R. 1951: The Iliad of Homer (Chicago: University of Chicago Press). ———. 1967: The Odyssey of Homer (New York: Harper and Row). Lawler, T. 1974: The Parisiana Poetria of John of Garland (New Haven: Yale University Press). Lawlor, J. 1961: ‘The Pattern of Consolation in The Book of the Duchess’, in R. Schoek and J. Taylor (eds.), Chaucer Criticism (Notre Dame: University of Notre Dame Press), 232–60. Lear, J. 1988: ‘Ethics, Mathematics, and Relativism’, in G. Sayre-McCord (ed.), Essays on Moral Realism (Ithaca: Cornell University Press), 76–94. ———. 1992: ‘Katharsis’, in A. Rorty (ed.), Essays on Aristotle’s Poetics (Princeton: Princeton University Press), 315–40. Leavis, F. R. 1952: The Common Pursuit (Harmondsworth: Penguin). Lee, K. 1982: ‘The Iris-Lyssa Scene in Euripides’ Heracles’, Antichthon 16, 44–53. ———. 1997: Euripides: Troades (Bristol: Bristol Classical Press). ———. 2007: Euripides: Ion (Oxford: Oxbow). Leech, C. 1969: Tragedy (London: Routledge). Lefèvre, E. 2001: Die Unfähigkeit, sich zu erkennen: Sophokles’ Tragödien (Leiden: Brill). ———. 2015: Studien zur Originalität der römischen Tragödie (Berlin: de Gruyter). Leigh, M. 1993: ‘Hopelessly Devoted to You: Traces of the Decii in Virgil’s Aeneid’, Proceedings of the Virgil Society 21, 89–110. Leonard, M. 2006: ‘Lacan, Irigaray, and Beyond: Antigones and the Politics of Psychoanalysis’, in V. Zajko and M. Leonard (eds.), Laughing With Medusa: Classical Myth and Feminist Thought (Oxford: Oxford University Press), 121–39. ———. 2012: ‘Tragedy and the Seductions of Philosophy’, The Cambridge Classical Journal 58, 145–65. ———. 2015: Tragic Modernities (Cambridge, Mass.: Harvard University Press). Lepore, E. and Ludwig, K. 2005: Donald Davidson: Meaning, Truth, Language, and Reality (Oxford: Oxford University Press). Lesky, A. 1961: Göttliche und menschliche Motivation im homerischen Epos (Heidelberg: Carl Winter). ———. 1983: ‘Decision and Responsibility in the Tragedy of Aeschylus’, in E. Segal (ed.), Oxford Readings in Greek Tragedy (Oxford: Oxford University Press), 13–23. Lessing, G. E. 1996: Werke, 8 vols., eds. H. G. Göpfert et al. (Darmstadt: WBG). Levi, P. 1958: Se questo è un uomo (Turin: Einaudi), tr. S. Woolf, If This is a Man (London: Penguin, 1989). ———. 1986: I sommersi e i salvati (Turin: Einaudi). Levin, H. 1961: Christopher Marlowe: The Overreacher (London: Faber and Faber).

Bibliography 383 Levin, R. 1981: ‘The Ironic Reading of “The Rape of Lucrece” and the Problem of External Evidence’, Shakespeare Survey 34, 85–92. ———. 1990: ‘Unthinkable Thoughts in the New Historicizing of English Renaissance Drama’, and ‘Reply to Catherine Belsey and Jonathan Goldberg’, New Literary History 21, 433–47, 463–70. ———. 2003: Looking for an Argument: Critical Encounters with the New Approaches to the Criticism of Shakespeare and His Contemporaries (London: Associated University Press). Lewis, C. S. 1998a: The Horse and his Boy (London: Collins). ———. 1998b: The Magician’s Nephew (London: Collins). ———. 2013a: The Allegory of Love: A Study in Medieval Tradition (Cambridge: Cambridge University Press). ———. 2013b: Selected Literary Essays, ed. W. Hooper (Cambridge: Cambridge University Press). Lewis, D. 1983: Philosophical Papers, vol. 1 (New York: Oxford University Press). ———. 1986: Philosophical Papers, vol. 2 (New York: Oxford University Press). Liapis, V. 2012a: A Commentary on the Rhesus Attributed to Euripides (Oxford: Oxford University Press). ———. 2012b: ‘Oedipus Tyrannus’, in K. Ormand (ed.), A Companion to Sophocles (Oxford: Wiley-Blackwell), 84–97. Liebler, N. 2002: ‘ “Thou Bleeding Piece of Earth”: The Ritual Ground of Julius Caesar’, in R. Wilson (ed.), New Casebooks: Julius Caesar (Basingstoke: Palgrave Macmillan), 128–48. Linforth, I. 1961: Antigone and Creon (Berkeley: University of California Press). Lloyd, M. 1984: ‘The Helen Scene in Euripides’ Troades’, Classical Quarterly 34, 303–13. Lloyd-Jones, H. 1983: The Justice of Zeus, 2nd edn. (Berkeley: University of California Press). ———. 1990: Greek Epic, Lyric, and Tragedy (Oxford: Clarendon). ———. 1991: Greek Comedy, Hellenistic Literature, Greek Religion, and Miscellanea (Oxford: Clarendon). ———. 1997: Sophocles: Ajax, Electra, Oedipus Tyrannus (Cambridge, MA: Harvard University Press). ———. 1998: Sophocles: Antigone, Women of Trachis, Philoctetes, Oedipus at Colonus (Cambridge, MA: Harvard University Press). Lloyd-Jones, H. and Wilson, N. 1990: Sophoclea: Studies on the Text of Sophocles (Oxford: Clarendon). Lobsien, V. 2014: ‘Vergeudung: Zur Affektpolitik in Shakespeares Tragödien’, in N. Bender et al. (eds.), Ethos und Form der Tragödie (Heidelberg: Winter), 165–83. Locke, J. 1975: An Essay Concerning Human Understanding, ed. P. Nidditch (Oxford: Clarendon). Lockie, R. 2014: ‘Relativism and Reflexivity’, in M. Baghramian (ed.), The Many Faces of Relativism (London: Routledge), 147–67. Lope de Vega, F. 2014a: El castigo sin venganza, ed. A. Carreño (Madrid: Cátedra). ———. 2014b: El Caballero de Olmedo, ed. F. Rico (Madrid: Cátedra). Loraux, N. 1985: Façons tragiques de tuer une femme (Paris: Hachette).

384  Bibliography ———. 2002: The Mourning Voice: An Essay on Greek Tragedy, tr. E. Rawlings (Ithaca: Cornell University Press). Lucas, D. W. 1972: Aristotle: Poetics (Oxford: Clarendon). Lucas, F. L. 1922: Seneca and Elizabethan Tragedy (Cambridge: Cambridge University Press). ———. 1957: Tragedy: Serious Drama in Relation to Aristotle’s Poetics (London: Chatto and Windus). Lurje, M. 2004: Die Suche nach der Schuld: Sophokles’ Oedipus Rex, Aristoteles’ Poetik und das Tragödienverständnis der Neuzeit (Munich: Saur). Lydgate, J. 1924–7: The Fall of Princes, 4 vols., ed. H. Bergen (London: Oxford University Press). Lyne, O. 1987: Further Voices in Vergil’s Aeneid (Oxford: Clarendon). Lyne, R. and Mukherji, S. 2007: ‘Introduction’, in R. Lyne and S. Mukherji (eds.), Early Modern Tragicomedy (Cambridge: Brewer), 1–14. MacFarlane, J. 2003: ‘Future Contingents and Relative Truth’, Philosophical Quarterly 53, 321–36. ———. 2007: ‘Relativism and Disagreement’, Philosophical Studies 132, 17–31. Macintosh, F. 2009: Oedipus Tyrannus (Cambridge: Cambridge University Press). MacKay, L. 1962: ‘Antigone, Coriolanus, and Hegel’, Transactions of the American Philological Association 93, 166–74. Mackie, C. 1991: ‘ “Nox Erat . . .”: Sleep and Visions in the Aeneid’, Greece and Rome 38, 59–61. Macleod, C. 1982: Homer: Iliad XXIV (Cambridge: Cambridge University Press). ———. 1983: Collected Essays (Oxford: Clarendon). Mader, M. B. 2014: ‘Being Genealogical: Tragic Necessity in Sophocles’ Antigone’, in T. Chanter and S. Kirkland (eds.), The Returns of Antigone: Interdisciplinary Essays (New York: SUNY Press), 125–36. Mann, J. 1989: ‘Shakespeare and Chaucer: What Is Criseyde Worth?’ in P. Boitani (ed.), The European Tragedy of Troilus (Oxford: Clarendon), 219–42. ———. 2003: ‘Chance and Destiny in Troilus and Criseyde and the Knight’s Tale’, in P. Boitani and J. Mann (eds.), The Cambridge Companion to Chaucer, 2nd edn. (Cambridge: Cambridge University Press), 93–111. Mann, T. 2004: Königliche Hoheit (Frankfurt am Main: Fischer). ———. 2012a: Der Zauberberg (Frankfurt am Main: Fischer). ———. 2012b: Lotte in Weimar (Frankfurt am Main: Fischer). ———. 2012c: Doktor Faustus (Frankfurt am Main: Fischer). ———. 2012d: Der Erwählte (Frankfurt am Main: Fischer). ———. 2013: Buddenbrooks: Verfall einer Familie (Frankfurt am Main: Fischer). Manuwald, B. 1992: ‘Oidipus und Adrastos: Bemerkungen zur neueren Diskussion um die Schuldfrage in Sophokles’ “König Oidipus” ’, Rheinisches Museum 135, 1–43. ———. 2012: Sophokles: König Ödipus (Berlin: de Gruyter). Marcus, L. 2016: ‘King Lear and the Death of the World’, in M. Neill and D. Schalkwyk (eds.), The Oxford Handbook of Shakespearean Tragedy (Oxford: Oxford University Press), 421–36. Marlowe, C. 2014: Doctor Faustus: The A- and B-texts (1604, 1616), a ParallelText Edition, eds. D. Bevington and E. Rasmussen (Manchester: Manchester University Press).

Bibliography 385 Marshall, C. 2002: ‘Portia’s Wound, Calphurnia’s Dream: Reading Character in Julius Caesar’, in R. Wilson (ed.), New Casebooks: Julius Caesar (Basingstoke: Palgrave Macmillan), 170–87. Marston, J. 1986: ‘The Tragedy of Sophonisba’, in P. Corbin and D. Sedge (eds.), Three Jacobean Witchcraft Plays (Manchester: Manchester University Press). Martindale, C. 2007: ‘Shakespeare Philosophus’, in W. Poole and R. Scholar (eds.), Thinking with Shakespeare: Comparative and Interdisciplinary Essays for A. D. Nuttall (London: Legenda), 33–50. Martindale, C. and Martindale, M. 1990: Shakespeare and the Uses of Antiquity (London: Routledge). Mastronarde, D. 1979: Continuity and Discontinuity: Some Conventions of Speech and Action on the Greek Tragic Stage (Berkeley: University of California Press). ———. 1994: Euripides: Phoenissae (Cambridge: Cambridge University Press). ———. 2002: Euripides: Medea (Cambridge: Cambridge University Press). ———. 2005: ‘The Gods’, in J. Gregory (ed.), A Companion to Greek Tragedy (Oxford: Wiley-Blackwell), 321–32. ———. 2010: The Art of Euripides: Dramatic Technique and Social Context (Cambridge: Cambridge University Press). Maus, K. 1986: ‘Taking Tropes Seriously: Language and Violence in Shakespeare’s Rape of Lucrece’, Shakespeare Quarterly 37, 66–82. McCall, M. 1972: ‘Divine and Human Action in Sophocles: the Two Burials of the Antigone’, Yale Classical Studies 22, 103–17. McClure, L. 1999: Spoken Like a Woman: Speech and Gender in Athenian Drama (Princeton: Princeton University Press). McCoskey, D. and Corbett, M. 2012: ‘Virginia Woolf, Richard Jebb, and Sophocles’ Antigone’, in K. Ormand (ed.), A Companion to Sophocles (Oxford: Wiley-Blackwell), 462–76. McDowell, J. 1994: Mind and World (Cambridge, MA: Harvard University Press). ———. 1998a: Mind, Value, and Reality (Cambridge, MA: Harvard University Press). ———. 1998b: Meaning, Knowledge, and Reality (Cambridge, MA: Harvard University Press). ———. 2008: ‘Comment on Stanley Cavell’s “Companionable Thinking” ’, in S. Cavell et al., Philosophy and Animal Life (New York: Columbia University Press), 127–38. McGinn, C. 2006: Shakespeare’s Philosophy: Discovering the Meaning Behind the Plays (New York: Harper Perennial). McKenna, S. 1991: ‘Henryson’s “Tragedie” of Cresseid’, Scottish Literary Journal 18, 26–36. McLuskie, K. 1994: ‘The Patriarchal Bard: Feminist Criticism and Shakespeare: King Lear and Measure for Measure’, in J. Dollimore and A. Sinfield (eds.), Political Shakespeare: Essays in Cultural Materialism, 2nd edn. (Manchester: Manchester University Press), 88–108. ———. 2003: ‘Macbeth, the Present, and the Past’, in R. Dutton and J. Howard (eds.), A Companion to Shakespeare’s Works: The Tragedies (Oxford: WileyBlackwell), 393–410. Medcalf, S. 2007: ‘Dreaming, Looking, and Seeing: Shakespeare and a Myth of Resurrection’, in W. Poole and R. Scholar (eds.), Thinking with Shakespeare:

386  Bibliography Comparative and Interdisciplinary Essays for A. D. Nuttall (London: Legenda), 93–112. Melchiori, G. 1998: King Edward III (Cambridge: Cambridge University Press). Melville, H. 1998: Moby Dick, ed. T. Tanner (Oxford: Oxford University Press). Michelini, A. 1982: Tradition and Dramatic Form in the Persians of Aeschylus (Leiden: Brill). Middleton, T. 2007: The Collected Works, eds. G. Taylor and J. Lavagnino (Oxford: Clarendon). Miller, P. 2007: ‘Lacan’s Antigone: The Sublime Object and the Ethics of Interpretation’, Phoenix 61, 1–14. Mills, P. 1996: ‘Hegel’s Antigone’, in P. Mills (ed.), Feminist Interpretations of G. W. F. Hegel (Pennsylvania: Pennsylvania State University Press), 59–88. Milton, J. 1954: Areopagitica, ed. J. Hales (London: Oxford University Press). ———. 1966: Poetical Works, ed. D. Bush (London: Oxford University Press). Minadeo, R. 1985: ‘Characterization and Theme in the Antigone’, Arethusa 18, 133–54. Miola, R. 1992: Shakespeare and Classical Tragedy (Oxford: Clarendon). ———. 2000: Shakespeare’s Reading (Oxford: Oxford University Press). Mitchell-Boyask, R. 1996: ‘Dramatic Scapegoating: On the Uses and Abuses of Girard and Shakespearean Criticism’, in M. Silk (ed.), Tragedy and the Tragic: Greek Theatre and Beyond (Oxford: Oxford University Press), 426–37. Mohamed, F. 2005: ‘Confronting Religious Violence: Milton’s Samson Agonistes’, PMLA 120, 327–40. Moles, J. 1984: ‘Aristotle and Dido’s “Hamartia” ’, Greece and Rome 31, 48–54. ———. 1986: ‘Review discussion of S. Goldhill, Language, Sexuality, Narrative: The Oresteia’, Liverpool Classical Monthly 11.4, 55–64. Montherlant, H. de. 1955: ‘Racine langouste’, Cahiers de la Compagnie Madeleine Renaud–Jean-Claude Barrault 8, 3–6. Montrose, L. 1986: ‘The Elizabethan Subject and the Spenserian Text’, in P. Parker and D. Quint (eds.), Literary Theory/Renaissance Texts (Baltimore: Johns Hopkins University Press), 303–40. ———. 1996: The Purpose of Playing: Shakespeare and the Cultural Politics of the Elizabethan Theatre (Chicago: University of Chicago Press). Moore, A. 1997: Points of View (Oxford: Oxford University Press). ———. 2013: ‘Was the Author of the Tractatus a Transcendental Idealist?’ in P. Sullivan and M. Potter (eds.), Wittgenstein’s Tractatus: History and Interpretation (Oxford: Oxford University Press), 239–55. Moore, J. 1977: ‘The Dissembling-Speech of Ajax’, Yale Classical Studies 25, 47–66. Morgan, T. 2015: Roman Faith and Christian Faith: Pistis and Fides in the Early Roman Empire and Early Churches (Oxford: Oxford University Press). Morris, D. 1991: The Culture of Pain (Berkeley: University of California Press). Morse, R. 1993: ‘Absolute Tragedy: Allusions and Avoidances’, Poetica (Tokyo) 38, 1–17. Morwood, J. 2007: Euripides: Suppliant Women (Oxford: Aris and Phillips). Mossman, J. 2005: ‘Women’s Voices’, in J. Gregory (ed.), A Companion to Greek Tragedy (Oxford: Wiley-Blackwell), 352–65. Most, G. 1993: ‘Schlegel, Schlegel und die Geburt eines Tragödienparadigmas’, Poetica 25, 155–75.

Bibliography 387 ———. 2000: ‘Generating Genres: the Idea of the Tragic’, in M. Depew and D. Obbink (eds.), Matrices of Genre: Authors, Canons, and Society (Cambridge, MA: Harvard University Press), 15–35. Mourgues, O. de. 1967: Racine, or the Triumph of Relevance (Cambridge: Cambridge University Press). Mozart, W. A. 2011: Die Zauberflöte, ed. N. John (London: Overture). Mueller, M. 1980: Children of Oedipus and Other Essays on the Imitation of Greek Tragedy 1550–1800 (Toronto: University of Toronto Press). Müller, F. 1979: Eyewitness Auschwitz: Three Years in the Gas Chambers (Chicago: Dee). Müller, J-D. 1993: Gotthold Ephraim Lessing: Emilia Galotti (Stuttgart: Reclam). Munteanu, D. 2012: Tragic Pathos: Pity and Fear in Greek Philosophy and Tragedy (Cambridge: Cambridge University Press). Murnaghan, S. 1986: ‘Antigone 904–20 and the Institution of Marriage’, American Journal of Philology 107, 192–207. ———. 2009: ‘Women in Greek Tragedy’, in R. Bushnell (ed.), A Companion to Tragedy (Oxford: Wiley-Blackwell), 234–50. Murray, G. 1912: ‘Greek and English Tragedy: A Contrast’, in G. Gordon (ed.), English Literature and the Classics (Oxford: Clarendon), 7–24. Neely, C. 1980: ‘Women and Men in Othello: “What Should Such a Fool/ Do with So Good a Woman?” ’ in C. Lenz, G. Greene and C. Neely (eds.), The Woman’s Part: Feminist Criticism of Shakespeare (Urbana/Chicago: University of Illinois Press), 211–39. ———. 1981: ‘Feminist Modes of Shakespeare Criticism: Compensatory, Justificatory, Transformational’, Women’s Studies 9, 3–15. ———. 1988: ‘Constructing the Subject: Feminist Practice and the New Renaissance Discourses’, English Literary Renaissance 18, 5–18. Neill, A. 2012: ‘Schopenhauer on Tragedy and the Sublime’, in B. Vandenabeele (ed.), A Companion to Schopenhauer (Oxford: Wiley-Blackwell), 206–18. Neill, M. 2016: ‘ “Romaine Tragedie”: The Designs of Titus Andronicus’, in M. Neill and D. Schalkwyk (eds.), The Oxford Handbook of Shakespearean Tragedy (Oxford: Oxford University Press), 339–57. Neuburg, M. 1990: ‘How Like a Woman: Antigone’s “Inconsistency” ’, Classical Quarterly 40, 54–70. Newlyn, L. 1993: Paradise Lost and the Romantic Reader (Oxford: Oxford University Press). Newman, J. 1994: ‘ “And Let Mild Women to Him Lose Their Mildness”: Philomela, Female Violence, and Shakespeare’s The Rape of Lucrece’, Shakespeare Quarterly 45, 304–26. Newman, K. 1991: Fashioning Femininity and English Renaissance Drama (Chicago: University of Chicago Press). Nickau, K. 1994: ‘Sophokles: König Ödipus’, in W. Barner (ed.), Ein Text und ein Leser: Weltliteratur für Liebhaber (Göttingen: Wallstein), 9–28. Nicolai, W. 1992: Zu Sophokles’ Wirkungsabsichten (Heidelberg: Winter). Nicoll, W. 2001: ‘The Death of Turnus’, Classical Quarterly 51, 190–200. Nicolson, M. 1959: Mountain Gloom and Mountain Glory: The Development of the Aesthetics of the Infinite (New York: Norton). Nietzsche, F. 1983: Also sprach Zarathustra (Stuttgart: Reclam).

388  Bibliography ———. 1996: Der Wille zur Macht: Versuch einer Umwertung aller Werte (Stuttgart: Kröner). ———. 1999a: Die Geburt Der Tragödie, eds. G. Colli and M. Montinari (Berlin: de Gruyter). ———. 1999b: Morgenröthe, Idyllen aus Messina, Die fröhliche Wissenschaft, eds. G. Colli and M. Montinari (Berlin: de Gruyter). Nisbet, H. B. 2013: Gotthold Ephraim Lessing: His Life, Works, and Thought (Oxford: Oxford University Press). Nisbet, R. G. M. 1961: Cicero: In L. Calpurnium Pisonem (Oxford: Clarendon). Nisbet, R. G. M. and Hubbard, M. 1970: A Commentary on Horace Odes, Book I (Oxford: Clarendon). Nooter, S. 2012: When Heroes Sing: Sophocles and the Shifting Soundscape of Tragedy (Cambridge: Cambridge University Press). Norland, H. 2009: Neoclassical Tragedy in Elizabethan England (Newark: University of Delaware Press). Norris, C. 1985: ‘Post-Structuralist Shakespeare: Text and Ideology’, in J. Drakakis (ed.), Alternative Shakespeares (London: Routledge), 47–66. Nozick, R. 1983: Philosophical Explanations (Cambridge, MA: Harvard University Press). ———. 1989: The Examined Life: Philosophical Meditations (New York: Simon and Schuster). Nussbaum, F. 2014: ‘The Unaccountable Pleasure of Eighteenth-Century Tragedy’, PMLA 129, 688–707. Nussbaum, M. 1986: The Fragility of Goodness: Luck and Ethics in Greek Tragedy and Philosophy (Cambridge: Cambridge University Press). ———. 1992: ‘Tragedy and Self-Sufficiency: Plato and Aristotle on Fear and Pity’, Oxford Studies in Ancient Philosophy 10, 107–59. ———. 1993: ‘Non-Relative Virtues: An Aristotelian Approach’, in M. Nussbaum and A. Sen (eds.), The Quality of Life (Oxford: Oxford University Press), 242–67. ———. 1994: ‘The Oedipus Rex and the Ancient Unconscious’, in P. Rudnytsky and E. Spitz (eds.), Freud and Forbidden Knowledge (New York: New York University Press), 42–71. ———. 1995: ‘Aristotle on Human Nature and the Foundations of Ethics’, in J. Altham and R. Harrison (eds.), World, Mind, and Ethics: Essays on the Ethical Philosophy of Bernard Williams (Cambridge: Cambridge University Press), 86–131. ———. 2001: Upheavals of Thought: The Intelligence of Emotions (Cambridge: Cambridge University Press). ———. 2008: ‘The “Morality of Pity”: Sophocles’ Philoctetes’, in R. Felski (ed.), Rethinking Tragedy (Baltimore: Johns Hopkins University Press), 148–69. Nuttall, A. D. 1967: Two Concepts of Allegory: A Study of Shakespeare’s The Tempest and the Logic of Allegorical Expression (London: RKP). ———. 1984: ‘The Argument about Shakespeare’s Characters’, in C. Cox and D. Palmer (eds.), Shakespeare’s Wide and Universal Stage (Manchester: Manchester University Press), 18–31. ———. 1996: Why Does Tragedy Give Pleasure? (Oxford: Clarendon). ———. 2007a: A New Mimesis: Shakespeare and the Representation of Reality (New Haven: Yale University Press).

Bibliography 389 ———. 2007b: Shakespeare the Thinker (New Haven: Yale University Press). Nyquist, M. 2008: ‘The Plight of Buchanan’s Jephtha: Sacrifice, Sovereignty, and Paternal Power’, Comparative Literature 60, 331–54. O’Grady, P. 2002: Relativism (Chesham: Acumen). O’Hara, J. 1990: Death and Optimistic Prophecy in Vergil’s Aeneid (Princeton: Princeton University Press). Onians, R. B. 1951: The Origins of European Thought About the Body, the Mind, the Soul, the World, Time, and Fate (Cambridge: Cambridge University Press). Orgel, S. 2002: Authentic Shakespeare, and Other Problems of the Early Modern Stage (London: Routledge). Osborne, R. 2012: ‘Sophocles and Contemporary Politics’, in K. Ormand (ed.), A Companion to Sophocles (Oxford: Wiley-Blackwell), 270–86. Oster, P. 2014: ‘ “Que vouliez-vous qu’il fît contre trois? Qu’il mourût”: Ethos und Form in den Tragödien Corneilles’, in N. Bender et al. (eds.), Ethos und Form der Tragödie (Heidelberg: Winter), 165–83. O’Sullivan, T. 2009: ‘Death ante ora patrum in Virgil’s Aeneid’, Transactions of the American Philological Association 139, 447–86. Otis, B. 1995: Virgil: A Study in Civilized Poetry (Norman: University of Oklahoma Press). Oudemans, T. C. and Lardinois, A. 1987: Tragic Ambiguity: Anthropology, Philosophy, and Sophocles’ Antigone (Leiden: Brill). Pack, R. 1939: ‘Fate, Chance, and Tragic Error’, American Journal of Philology 60, 350–6. Padel, R. 1992: In and Out of the Mind: Greek Images of the Tragic Self (Princeton: Princeton University Press). Page, D. L. 1936: ‘The Elegiacs in Euripides’ Andromache’, in C. Bailey et al. (eds.), Greek Poetry and Life: Essays Presented to Gilbert Murray (Oxford: Clarendon), 206–30. ———. 1976: Euripides: Medea (Oxford: Clarendon). Pagin, P. 2013: ‘Radical Interpretation and the Principle of Charity’, in E. Lepore and K. Ludwig (eds.), A Companion to Donald Davidson (Oxford: WileyBlackwell), 225–46. Paige, N. 2017: ‘Racine, Phèdre, and the French Classical Stage’, in C. Prendergast (ed.), A History of Modern French Literature: From the Sixteenth to the Twentieth Century (Princeton: Princeton University Press), 190–211. Panoussi, V. 2005: ‘Polis and Empire: Greek Tragedy in Rome’, in J. Gregory (ed.), A Companion to Greek Tragedy (Oxford: Wiley-Blackwell), 413–27. ———. 2010: ‘Aeneas’ Sacral Authority’, in J. Farrell and M. Putnam (eds.), A Companion to Vergil’s Aeneid and Its Tradition (Oxford: Wiley-Blackwell), 53–65. Parish, R. 1993: Racine: The Limits of Tragedy (Tübingen: Romanisches Seminar). Parker, L. 2007: Euripides: Alcestis (Oxford: Clarendon). ———. 2016: Euripides: Iphigeneia in Tauris (Oxford: Clarendon). Parker, R. 1983: Miasma: Pollution and Purification in Early Greek Religion (Oxford: Clarendon). ———. 1997: ‘Gods Cruel and Kind: Tragic and Civic Theology’, in C. Pelling (ed.), Greek Tragedy and the Historian (Oxford: Clarendon), 143–60.

390  Bibliography ———. 1999: ‘Through a Glass Darkly: Sophocles and the Divine’, in J. Griffin (ed.), Sophocles Revisited: Essays Presented to Sir Hugh Lloyd-Jones (Oxford: Oxford University Press), 11–30. Parry, A. 1966: ‘The Two Voices of Virgil’s Aeneid’, in S. Commager (ed.), Virgil: A Collection of Critical Essays (Englewood Cliffs, NJ: Prentice-Hall), 107–23. Parvini, N. 2012: Shakespeare and Contemporary Theory: New Historicism and Cultural Materialism (London: Bloomsbury). Peacocke, C. 2009: ‘The Perception of Music: Sources of Significance’, British Journal of Aesthetics 49, 257–75. Pechter, E. 1987: ‘The New Historicism and Its Discontents: Politicizing Renaissance Drama’, PMLA 102, 292–303. ———. 2016: ‘The Romantic Inheritance’, in M. Neill and D. Schalkwyk (eds.), The Oxford Handbook of Shakespearean Tragedy (Oxford: Oxford University Press), 54–70. Pelling, C. 1988: Plutarch: Life of Antony (Cambridge: Cambridge University Press). ———. 1990: ‘Conclusion’, in C. Pelling (ed.), Characterization and Individuality in Greek Literature (Oxford: Clarendon), 245–62. ———. 1997: ‘Conclusion’, in C. Pelling (ed.), Greek Tragedy and the Historian (Oxford: Clarendon), 213–35. ———. 2005: ‘Tragedy, Rhetoric, and Performance Culture’, in J. Gregory (ed.), A Companion to Greek Tragedy (Oxford: Wiley-Blackwell), 83–102. Peradotto, J. 1992: ‘Disauthorizing Prophecy: The Ideological Mapping of Oedipus Tyrannus’, Transactions of the American Philological Association 122, 1–15. ———. 2007: ‘The Omen of the Eagles and the Ethos of Agamemnon’, in M. Lloyd (ed.), Oxford Readings in Classical Studies: Aeschylus (Oxford: Oxford University Press), 211–44. Perrett, R. 1999: ‘Musical Unity and Sentential Unity’, British Journal of Aesthetics 39, 97–111. Peters, J. 2009: ‘A Theory of Tragic Experience According to Hegel’, European Journal of Philosophy 19, 85–106. Pineau, J. 1991: ‘Sur la culpabilité de la Phèdre de Racine: augustinisme et poésie’, La Licorne (Poitiers) 20, 41–50. Pinkard, T. 2015: ‘Tragedy With and Without Religion: Hegelian Thoughts’, in J. Billings and M. Leonard (eds.), Tragedy and the Idea of Modernity (Oxford: Oxford University Press), 137–58. Pinker, S. 2002: The Blank Slate: The Modern Denial of Human Nature (London: Penguin). ———. 2011: The Better Angels of Our Nature: The Decline of Violence in History and Its Causes (London: Allen Lane). Pippin, R. 2015: ‘Williams on Nietzsche and the Greeks’, in J. Billings and M. Leonard (eds.), Tragedy and the Idea of Modernity (Oxford: Oxford University Press), 285–305. Plato. 2007: The Republic, tr. D. Lee (London: Penguin). Pocock, G. 1973: Corneille and Racine: Problems of Tragic Form (Cambridge: Cambridge University Press) Podlecki, A. 1966: ‘The Power of the Word in Sophocles’ Philoctetes’, Greek, Roman, and Byzantine Studies 7, 233–50.

Bibliography 391 Pollard, T. 2015: ‘Tragicomedy’, in P. Cheney and P. Hardie (eds.), The Oxford History of Classical Reception in English Literature, vol. 2 (Oxford: Oxford University Press), 419–32. Poller, W. 1947: Arztschreiber in Buchenwald (Hamburg: Phönix-Verlag). Poole, A. 1987: Tragedy: Shakespeare and the Greek Example (Oxford: Wiley-Blackwell). ———. 2005: Tragedy: a Very Short Introduction (Oxford: Oxford University Press). Porter, J. 2016: The Sublime in Antiquity (Cambridge: Cambridge University Press). Pöschl, V. 1970: ‘Dido und Aeneas’, in D. Ableitinger and H. Gugel (eds.), Festschrift Karl Vretska (Heidelberg: Winter), 148–73. ———. 1977: Die Dichtkunst Virgils: Bild und Symbol in der Äneis, 3rd edn. (Berlin: de Gruyter). Potkay, A. 2012: ‘The British Romantic Sublime’, in T. Costelloe (ed.), The Sublime: From Antiquity to the Present (Cambridge: Cambridge University Press), 203–16. Potter, L. 2015: Shakespeare: The Two Noble Kinsmen, rev. edn. (London: Bloomsbury). Poulsen, A. D. 2013: ‘Why No Mercy? A Study of Aeneas’ Missing Virtue’, Symbolae Osloenses 87, 95–133. Prévost, J. 1985: Turne: Tragédie, ed. F. Kantor (Poitiers: Société des Antiquaires de L’Ouest). Prince, F. T. 1960: Shakespeare: the Poems (London: Methuen). Pritchard, D. 2011: ‘Epistemic Relativism, Epistemic Incommensurability, and Wittgensteinian Epistemology’, in S. Hales (ed.), A Companion to Relativism (Oxford: Wiley-Blackwell), 266–85. Pucci, P. 1992: Oedipus and the Fabrication of the Father: Oedipus Tyrannus in Modern Criticism and Philosophy (Baltimore: Johns Hopkins University Press). ———. 1994: ‘Gods’ Intervention and Epiphany in Sophocles’, American Journal of Philology 115, 15–46. ———. 2016: Euripides’ Revolution Under Cover: An Essay (Ithaca: Cornell University Press). Puccini, G. 1984: Madama Butterfly, ed. N. John (London: John Calder). Purkiss, D. 1996: The Witch in History: Early Modern and Twentieth-Century Representations (London: Routledge). Putnam, M. 1965: The Poetry of the Aeneid: Four Studies in Imaginative Unity and Design (Cambridge, MA: Harvard University Press). ———. 1995: Virgil’s Aeneid: Interpretation and Influence (Chapel Hill: University of North Carolina Press). ———. 2001: ‘Vergil’s Aeneid: The Final Lines’, in S. Spence (ed.), Poets and Critics Read Vergil (New Haven: Yale University Press), 86–104. ———. 2006: ‘The Aeneid and Paradise Lost: Ends and Conclusions’, Literary Imagination 8, 387–410. ———. 2010: ‘Vergil, Ovid, and the Poetry of Exile’, in J. Farrell and M. Putnam (eds.), A Companion to Vergil’s Aeneid and its Tradition (Oxford: WileyBlackwell), 80–95. ———. 2011: The Humanness of Heroes: Studies in the Conclusion of Virgil’s Aeneid (Amsterdam: Amsterdam University Press).

392  Bibliography Pütz, P. 1986: Die Leistung der Form: Lessings Dramen (Frankfurt am Main: Suhrkamp). Quilligan, M. 2005: Incest and Agency in Elizabeth’s England (Philadelphia: University of Pennsylvania Press). Quine, W. V. 2008: Confessions of a Confirmed Extensionalist, eds. D. Føllesdal and D. Quine (Cambridge, MA: Harvard University Press). Quinton, A. 1960: ‘Tragedy’, Proceedings of the Aristotelian Society 34 (supp), 145–64. Racine, J. 1967: Andromache and Other Plays, tr. J. Cairncross (Harmondsworth: Penguin). ———. 1982–3: Théâtre Complet, 2 vols., ed. J-P. Collinet (Paris: Gallimard). ———. 2004: Iphigenia, Phaedra, and Athaliah, tr. J. Cairncross (London: Penguin). Rackin, P. 2005: Shakespeare and Women (Oxford: Oxford University Press). Randhawa, B. 2014: ‘The Unblest Room: Kristeva’s Chora in Sophocles’ Antigone’, Philosophy and Literature 38, 293–313. Raphael, D. D. 1960: The Paradox of Tragedy (Bloomington: Indiana University Press). Rawlinson, M. 2014: ‘Beyond Antigone: Ismene, Gender, and the Right to Life’, in T. Chanter and S. Kirkland (eds.), The Returns of Antigone: Interdisciplinary Essays (New York: SUNY Press), 101–21. Reckford, K. 1961: ‘Latent Tragedy in Aeneid VII, 1–285’, American Journal of Philology 82, 252–61. Redfield, J. 1994: Nature and Culture in the Iliad: The Tragedy of Hector, expanded edn. (Durham/London: Duke University Press). Reed, J. 2010: ‘Vergil’s Roman’, in J. Farrell and M. Putnam (eds.), A Companion to Vergil’s Aeneid and its Tradition (Oxford: Wiley-Blackwell), 66–79. Rehm, R. 1994: Marriage to Death: The Conflation of Wedding and Funeral Rituals in Greek Tragedy (Princeton: Princeton University Press). Reichl, K. 1989: ‘Chaucer’s Troilus: Philosophy and Language’, in P. Boitani (ed.), The European Tragedy of Troilus (Oxford: Clarendon), 133–52. Reinhardt, K. 1947: Sophokles, 3rd edn. (Frankfurt am Main: Klostermann). Renger, A-B. 2013: Oedipus and the Sphinx: The Threshold Myth from Sophocles Through Freud to Cocteau (Chicago: University of Chicago Press). Renger, C. 1985: Aeneas und Turnus: Analyse einer Feindschaft (Frankfurt am Main: Lang). Reynaud, D. and Thirouin, L. 2004: Corneille et Voltaire: Œdipe (Lyon: Université de Saint-Étienne). Rhodes, P. J. 2003: ‘Nothing to Do with Democracy: Athenian Drama and the Polis’, Journal of Hellenic Studies 123, 104–19. Richards, I. A. 1930: Principles of Literary Criticism, 4th edn. (London: Routledge & Kegan Paul). Richardson, N. 1993: The Iliad: A Commentary, vol. 6 (Cambridge: Cambridge University Press). Ridley, A. 2003: ‘Tragedy’, in J. Levinson (ed.), The Oxford Handbook of Aesthetics (Oxford: Oxford University Press), 408–20. Roberts, D. 1988: ‘Sophoclean Endings: Another Story’, Arethusa 21, 177–96. Robertson, D. 1961: ‘Chaucerian Tragedy’, in R. Schoek and J. Taylor (eds.), Chaucer Criticism (Notre Dame: University of Notre Dame Press), 86–121.

Bibliography 393 Rocco, C. 1997: Tragedy and Enlightenment: Athenian Political Thought and the Dilemmas of Modernity (Berkeley: University of California Press). Roche, M. 2009: ‘The Greatness and Limits of Hegel’s Theory of Tragedy’, in R. Bushnell (ed.), A Companion to Tragedy (Oxford: Wiley-Blackwell), 51–67. Rose, P. 1995: ‘Historicizing Sophocles’ Ajax’, in B. Goff (ed.), History, Tragedy, Theory: Dialogues on Athenian Drama (Austin: University of Texas Press), 59–90. ———. 2012: ‘Sophocles and Class’, in K. Ormand (ed.), A Companion to Sophocles (Oxford: Wiley-Blackwell), 253–69. Rosenberg, M. 1993: The Masks of Othello: The Search for the Identity of Othello, Iago, and Desdemona by Three Centuries of Actors and Critics (London/ Toronto: Associated University Presses). Rosivach, V. 1979: ‘The Two Worlds of the Antigone’, Illinois Classical Studies 4, 16–36. ———. 1983: ‘On Creon, Antigone, and not Burying the Dead’, Rheinisches Museum 126, 193–211. Rösler, W. 1993: ‘Die Frage der Echtheit von Sophokles, Antigone 904–20 und die politische Funktion der attischen Tragödie’, in A. Sommerstein et al. (eds.), Tragedy, Comedy, and the Polis (Bari: Levante), 81–99. Rouse, W. D. 1911: ‘The Two Burials in Antigone’, Classical Review 25, 40–2. Rousseau, J-J. 1967: Julie, ou la Nouvelle Héloïse, ed. M. Launay (Paris: Flammarion); tr. P. Stewart and J. Vaché, Julie, or the New Heloise (Hanover: University Press of New England, 1997). Rowe, K. 2003: ‘Minds in Company: Shakespearean Tragic Emotions’, in R. Dutton and J. Howard (eds.), A Companion to Shakespeare’s Works: The Tragedies (Oxford: Wiley-Blackwell), 47–72. Rozett, M. 1984: The Doctrine of Election and the Emergence of Elizabethan Tragedy (Princeton: Princeton University Press). Rudd, N. 1976: Lines of Enquiry: Studies in Latin Poetry (Cambridge: Cambridge University Press). ———. 1989: Horace: Epistles Book II and Epistle to the Pisones (‘Ars Poetica’) (Cambridge: Cambridge University Press). Ruggiers, P. 1973: ‘Notes Towards a Theory of Tragedy in Chaucer’, The Chaucer Review 8, 89–99. Rundle, B. 2001: ‘Wittgenstein on Grammar, Meaning, and Essence’, in R. Gaskin (ed.), Grammar in Early Twentieth-Century Philosophy (London: Routledge), 182–97. ———. 2004: Why There Is Something Rather Than Nothing (Oxford: Clarendon). Russell, B. 1986: Mysticism and Logic (London: Unwin). Russell, D. and Winterbottom, M. 1972: Ancient Literary Criticism: The Principal Texts in New Translations (Oxford: Clarendon). Russo, J. 2012: ‘Rethinking Homeric Psychology: Snell, Dodds, and Their Critics’, Quaderni Urbinati di Cultura Classica 101, 11–28. Rusten, J. 1989: Thucydides: The Peloponnesian War Book II (Cambridge: Cambridge University Press). Rutherford, R. 2001: ‘Tragic Form and Feeling in the Iliad’, in D. Cairns (ed.), Oxford Readings in Homer’s Iliad (Oxford: Oxford University Press), 260–93. ———. 2012: Greek Tragic Style (Cambridge: Cambridge University Press). Ryan, K. 2015: Shakespeare’s Universality: Here’s Fine Revolution (London: Bloomsbury).

394  Bibliography Rymer, T. 1956: The Critical Works of Thomas Rymer, ed. C. Zimansky (New Haven: Yale University Press). Santirocco, M. 1980: ‘Justice in Sophocles’ Antigone’, Philosophy and Literature 4, 180–98. Savigny, E. von. 1994: Wittgensteins Philosophische Untersuchungen, 2 vols. (Frankfurt am Main: Klostermann). Scaliger, J. C. 1964: Poetices Libri Septem (Stuttgart: Frommann-Holzboog; facsimile of 1561 Lyon edition). Scarry, E. 1985: The Body in Pain: The Making and Unmaking of the World (Oxford: Oxford University Press). Schadewaldt, W. 1960: Hellas und Hesperien: Gesammelte Schriften zur Antike und zur neueren Literatur (Zurich/Stuttgart: Artemis). Schalkwyk, D. 2016: ‘Coriolanus: A Tragedy of Language’, in M. Neill and D. Schalkwyk (eds.), The Oxford Handbook of Shakespearean Tragedy (Oxford: Oxford University Press), 468–85. Schein, S. 2013: Sophocles: Philoctetes (Cambridge: Cambridge University Press). Schelling, F. 1856–61: Sämtliche Werke, 14 vols. (Stuttgart/Augsburg: Cotta). Schiesaro, A. 2003: The Passions in Play: Thyestes and the Dynamics of Senecan Drama (Cambridge: Cambridge University Press). ———. 2009: ‘Seneca and the Denial of the Self’, in S. Bartsch and D. Wray (eds.), Seneca and the Self (Cambridge: Cambridge University Press), 221–35. Schiller, F. 1992–2004: Werke und Briefe, 12 vols., eds. O. Dann et al. (Frankfurt am Main: Deutscher Klassiker Verlag). Schlegel, A. W. 1966: Vorlesungen über dramatische Kunst und Literatur, 2 vols., ed. E. Lohner (Stuttgart: Kohlhammer). Schlegel, J. E. 1967: Canut, ed. H. Steinmetz (Stuttgart: Reclam). Schmidt, D. 2001: On Germans and Other Greeks: Tragedy and Ethical Life (Bloomington: Indiana University Press). Schmidt, J. 1989: ‘Sophokles, König Ödipus. Das Scheitern des Aufklärers an der alten Religion’, in J. Schmidt (ed.), Aufklärung und Gegenaufklärung in der Europäischen Literatur, Philosophie und Politik von der Antike bis zur Gegenwart (Darmstadt: WBG), 33–55. Schmitt, A. 1982: ‘Athenes Umgang mit den Menschen bei Homer: zu einem Aspekt homerischer Theologie’, Die Alten Sprachen im Unterricht 31, 6–23. ———. 1988a: ‘Tragische Schuld in der griechischen Antike’, in G. Eifler and O. Saame (eds.), Die Frage nach der Schuld (Mainz: Studium Generale der Johannes Gutenberg-Universität), 157–92. ———. 1988b: ‘Menschliches Fehlen und Tragisches Scheitern: Zur Handlungsmotivation im Sophokleischen “König Ödipus” ’, Rheinisches Museum 131, 8–30. ———. 1988c: ‘Bemerkungen zu Charakter und Schicksal der tragischen Hauptpersonen in der Antigone’, Antike und Abendland 34, 1–16. ———. 1990: Selbständigkeit und Abhängigkeit menschlichen Handelns bei Homer (Stuttgart: Steiner). ———. 1992: ‘Zur Aristoteles-Rezeption in Schiller’s Theorie des Tragischen’, in B. Zimmermann (ed.), Antike Dramentheorie und ihre Rezeption (Stuttgart: Metzler), 191–213. ———. 1994: ‘Aristoteles und die Moral der Tragödie’, in A. Bierl and P. von Möllendorff (eds.), Orchestra: Drama, Mythos, Bühne (Stuttgart: Teubner), 331–43.

Bibliography 395 ———. 1997: ‘Wesenszüge der griechischen Tragödie: Schicksal, Schuld, Tragik’, in H. Flashar (ed.), Tragödie: Idee und Transformation (Stuttgart: Teubner), 7–49. ———. 2011: Aristoteles: Poetik (Berlin: Akademie Verlag). ———. 2014: ‘Innere und Äußere Wahrscheinlichkeit im König Ödipus; oder: Über Ödipus in uns’, in N. Bender et al. (eds.), Ethos und Form der Tragödie (Heidelberg: Winter), 1–62. ———. 2015: ‘Tragik vor der Tragödie? Scheiterndes Handeln im Homerischen Epos—und ein kurzer Vergleich mit dem Scheitern des Handelns im mittelalterlichen Nibelungenepos’, in R. Toepfer and G. Radke-Uhlmann (eds.), Tragik vor der Moderne (Heidelberg: Winter), 201–44. Schönbaumsfeld, G. 2013: ‘Kierkegaard and the Tractatus’, in P. Sullivan and M. Potter (eds.), Wittgenstein’s Tractatus: History and Interpretation (Oxford: Oxford University Press), 59–75. Schopenhauer, A. 1986: Die Welt als Wille und Vorstellung, 2 vols., ed. W. Freiherr von Löhneysen (Frankfurt am Main: Suhrkamp). Schulze, U. 2015: Das Nibelungenlied (Stuttgart: Reclam). Scodel, R. 1984: Sophocles (Boston: Twayne). ———. 2005: ‘Sophoclean Tragedy’, in J. Gregory (ed.), A Companion to Greek Tragedy (Oxford: Wiley-Blackwell), 233–50. ———. 2009: ‘Tragedy and Epic’, in R. Bushnell (ed.), A Companion to Tragedy (Oxford: Wiley-Blackwell), 181–97. Scruton, R. 1988: ‘Analytical Philosophy and the Meaning of Music’, Journal of Aesthetics and Art Criticism 46, 169–76. ———. 2004: ‘Wittgenstein and the Understanding of Music’, British Journal of Aesthetics 44, 1–9. Scullion, S. 2002: ‘ “Nothing to Do with Dionysus”: Tragedy Misconceived as Ritual’, Classical Quarterly 52, 102–37. Seaford, R. 1984: Cyclops of Euripides (Oxford: Oxford University Press). ———. 1987: ‘The Tragic Wedding’, Journal of Hellenic Studies 107, 106–30. ———. 1994: Reciprocity and Ritual: Homer and Tragedy in the Developing City-State (Oxford: Clarendon). ———. 1995: ‘Historicizing Tragic Ambivalence: The Vote of Athena’, in B. Goff (ed.), History, Tragedy, Theory: Dialogues on Athenian Drama (Austin: University of Texas Press), 202–21. ———. 1996: ‘Something to Do with Dionysos—Tragedy and the Dionysiac: Response to Friedrich’, in M. Silk (ed.), Tragedy and the Tragic: Greek Theatre and Beyond (Oxford: Oxford University Press), 284–94. ———. 2000: ‘The Social Function of Attic Tragedy: A Response to Jasper Griffin’, Classical Quarterly 50, 30–44. Segal, C. 1966: ‘The Electra of Sophocles’, Transactions of the American Philological Association 97, 473–545. ———. 1970: ‘Shame and Purity in Euripides’ Hippolytus’, Hermes 98, 278–99. ———. 1981: Tragedy and Civilization: An Interpretation of Sophocles (Cambridge, MA: Harvard University Press). ———. 1986a: ‘Greek Tragedy and Society: A Structuralist Perspective’, in J. P. Euben (ed.), Greek Tragedy and Political Theory (Berkeley: University of California Press), 43–75. ———. 1986b: Interpreting Greek Tragedy: Myth, Poetry, Text (Ithaca: Cornell University Press).

396  Bibliography ———. 1993: Euripides and the Poetics of Sorrow (Durham: Duke University Press). ———. 1994: ‘Sophocles’ Oedipus Tyrannus: Freud, Language, and the Unconscious’, in P. Rudnytsky and E. Spitz (eds.), Freud and Forbidden Knowledge (New York: New York University Press), 72–95. ———. 1996: ‘Catharsis, Audience, and Closure in Greek Tragedy’, in M. Silk (ed.), Tragedy and the Tragic: Greek Theatre and Beyond (Oxford: Oxford University Press), 149–72. ———. 2001: Oedipus Tyrannus: Tragic Heroism and the Limits of Knowledge, 2nd edn. (New York: Oxford University Press). Seidensticker, B. 1978: ‘Comic Elements in Euripides’ Bacchae’, American Journal of Philology 99, 303–20. ———. 1982: Palintonos Harmonia: Studien zu komischen Elementen in der griechischen Tragödie (Göttingen: Vandenhoeck und Ruprecht). ———. 1995: ‘Women on the Tragic Stage’, in B. Goff (ed.), History, Tragedy, Theory: Dialogues on Athenian Drama (Austin: University of Texas Press), 151–73. ———. 2006: ‘Distanz und Nähe: Zur Darstellung von Gewalt in der griechischen Tragödie’, in B. Seidensticker and M. Vöhler (eds.), Gewalt und Ästhetik: Zur Gewalt and ihrer Darstellung in der griechischen Klassik (Berlin: de Gruyter), 91–122. ———. 2008: ‘Character and Characterization in Greek Tragedy’, in M. Revermann and P. Wilson (eds.), Performance, Iconography, Reception: Studies in Honour of Oliver Taplin (Oxford: Oxford University Press), 333–46. Semprun, J. 1963: Le grand voyage (Paris: Gallimard). Servius. 1881–1902: Servii Grammatici Qui Feruntur in Vergilii Carmina Commentarii, 3 vols., eds. G. Thilo and H. Hagen (Leipzig: Teubner). Sewell-Rutter, N. J. 2007: Guilt by Descent: Moral Inheritance and Decision Making in Greek Tragedy (Oxford: Oxford University Press). Shakespeare, W. 2005: The Complete Works, 2nd edn., eds. J. Jowett, W. Montgomery, G. Taylor, and S. Wells (Oxford: Clarendon). ———. 2011: Sir Thomas More, ed. J. Jowett (London: Bloomsbury). Shanley, J. 1961: ‘The Troilus and Christian Love’, in R. Schoek and J. Taylor (eds.), Chaucer Criticism (Notre Dame: University of Notre Dame Press), 136–46. Shapiro, J. 2005: 1599: A Year in the Life of William Shakespeare (London: Faber and Faber). ———. 2015: 1606: William Shakespeare and the Year of Lear (London: Faber and Faber). Shapsay, S. 2012: ‘The Problem with the Problem of Tragedy: Schopenhauer’s Solution Revisited’, British Journal of Aesthetics 52, 17–32. Shell, A. 2010: Shakespeare and Religion (London: Bloomsbury). Shelley, P. B. 1935: The Complete Poetical Works, ed. T. Hutchinson (London: Oxford University Press). ———. 2003: The Major Works, eds. Z. Leader and M. O’Neill (Oxford: Oxford University Press). Sherman, N. 1992: ‘Hamartia and Virtue’, in A. Rorty (ed.), Essays on Aristotle’s Poetics (Princeton: Princeton University Press), 177–96.

Bibliography 397 Shuger, D. 1994: The Renaissance Bible: Scholarship, Sacrifice, and Subjectivity (Berkeley: University of California Press). Sicherl, M. 1977: ‘The Tragic Issue in Sophocles’ Ajax’, Yale Classical Studies 25, 67–98. Sidney, P. 1999: The Countess of Pembroke’s Arcadia (The Old Arcadia) (Oxford: Oxford University Press). ———. 2002: An Apology for Poetry, eds. G. Shepherd and R. Maslen (Manchester: Manchester University Press). Silk, M. 1996a: ‘General Introduction’, in M. Silk (ed.), Tragedy and the Tragic: Greek Theatre and Beyond (Oxford: Oxford University Press), 1–11. ———. 1996b: ‘Tragic Language: The Greek Tragedians and Shakespeare’, in M. Silk (ed.), Tragedy and the Tragic: Greek Theatre and Beyond (Oxford: Oxford University Press), 458–96. ———. 2000: Aristophanes and the Definition of Comedy (Oxford: Oxford University Press). ———. 2004: ‘Shakespeare and Greek Tragedy: Strange Relationship’, in C. Martindale and A. Taylor (eds.), Shakespeare and the Classics (Cambridge: Cambridge University Press), 241–57. ———. 2009: ‘The Logic of the Unexpected: Semantic Diversion in Sophocles, Yeats (and Virgil)’, in S. Goldhill and E. Hall (eds.), Sophocles and the Greek Tragic Tradition (Cambridge: Cambridge University Press), 134–57. Silk, M. and Stern, J. 1981: Nietzsche on Tragedy (Cambridge: Cambridge University Press). Simpson, A. and Miller, C. 1948: ‘A Note on Sophocles’ Antigone, Lines 531– 81’, Greece and Rome 17, 78–81. Sinfield, A. 1992: Faultlines: Cultural Materialism and the Politics of Dissident Reading (Berkeley: University of California Press). Smith, E. 2016: ‘Character in Shakespearean Tragedy’, in M. Neill and D. Schalkwyk (eds.), The Oxford Handbook of Shakespearean Tragedy (Oxford: Oxford University Press), 89–103. Smollett, T. 1998: The Expedition of Humphry Clinker (Oxford: Oxford University Press). Sobolewicz, T. 2011: Aus der Hölle zurück: von der Willkür des Überlebens im Konzentrationslager, tr. S. Schmidt (Frankfurt am Main: Fischer). Söderbäck, F. 2010: ‘Impossible Mourning: Sophocles Reversed’, in F. Söderbäck (ed.), Feminist Readings of the Antigone (New York: SUNY Press), 65–82. Sommerstein, A. 1989: Aeschylus: Eumenides (Cambridge: Cambridge University Press). ———. 2008: Aeschylus: Persians, Seven Against Thebes, Suppliants, Prometheus Bound (Cambridge, MA: Harvard University Press). Sontag, S. 2003: Regarding the Pain of Others (London: Penguin). Sophocles. 1928: Sophoclis Fabulae, ed. A. C. Pearson (Oxford: Clarendon). ———. 1990: Sophoclis Fabulae, eds. H. Lloyd-Jones and N. Wilson (Oxford: Clarendon). ———. 1996: Sophocles: Antigone, ed. R. D. Dawe, 3rd edn. (Stuttgart/Leipzig: Teubner). Sorabji, R. 1980: Necessity, Cause, and Blame: Perspectives on Aristotle’s Theory (London: Duckworth).

398  Bibliography Sourvinou-Inwood, C. 1989: ‘Assumptions and the Creation of Meaning: Reading Sophocles’ Antigone’, Journal of Hellenic Studies 109, 134–48. ———. 1990: ‘Sophocles’ Antigone as a “Bad Woman” ’, in F. Dieteren and E. Kloek (eds.), Writing Women into History (Amsterdam: Grüner), 11–38. Staley, G. 2010: Seneca and the Idea of Tragedy (Oxford: Oxford University Press). Stanford, W. 1963: Aristophanes: The Frogs, 2nd edn. (London: Palgrave Macmillan). Stecker, R. 1999: ‘Davies on the Musical Expression of Emotion’, British Journal of Aesthetics 39, 273–81. Stefonio, B. 2007: Crispus: Tragoedia, ed. A. Torino (Rome: Bardi Editore). Steiner, G. 1961: The Death of Tragedy (London: Faber and Faber). ———. 1971: In Bluebeard’s Castle: Some Notes Towards the Redefinition of Culture (New Haven: Yale University Press). ———. 1984: Antigones: The Antigone Myth in Western Literature, Art, and Thought (Oxford: Clarendon). ———. 1990: ‘A Note on Absolute Tragedy’, Journal of Literature and Theology 4, 147–56. ———. 1996: ‘Tragedy, Pure and Simple’, in M. Silk (ed.), Tragedy and the Tragic: Greek Theatre and Beyond (Oxford: Oxford University Press), 534–46. ———. 1997: Errata: an Examined Life (London: Weidenfeld and Nicolson). ———. 1998: Language and Silence: Essays on Language, Literature, and the Inhuman (New Haven: Yale University Press). ———. 2008: ‘ “Tragedy,” Reconsidered’, in R. Felski (ed.), Rethinking Tragedy (Baltimore: Johns Hopkins University Press), 29–44. Steinmetz, H. 1987: ‘Emilia Galotti’, in Lessings Dramen: Interpretationen (Stuttgart: Reclam), 87–137. Sternberg, R. 2005: ‘The Nature of Pity’, in R. Sternberg (ed.), Pity and Power in Ancient Athens (Cambridge: Cambridge University Press), 15–47. Stevens, P. T. 1971: Euripides: Andromache (Oxford: Clarendon). Stierle, K. 2014: ‘Katharsis: Ein Webfehler in der Aristotelischen Poetik’, in N. Bender et al. (eds.), Ethos und Form der Tragödie (Heidelberg: Winter), 63–84. Stinton, T. C. W. 1990: Collected Papers on Greek Tragedy (Oxford: Clarendon). Stirling, B. 1968: ‘ “Or Else Were This a Savage Spectacle” ’, in L. Dean (ed.), Twentieth-Century Interpretations of Julius Caesar: a Collection of Critical Essays (Englewood Cliffs, NJ: Prentice-Hall), 39–49. Strindberg, A. 1976: The Father, Miss Julie, and The Ghost Sonata, tr. M. Meyer (London: Methuen). Stump, E. 2010: Wandering in Darkness: Narrative and the Problem of Suffering (Oxford: Clarendon). Suetonius. 1997: Lives of the Caesars, tr. J. Rolfe (Cambridge, MA: Harvard University Press). Sussman, R. 1975: ‘Bérénice and the Tragic Moment’, L’esprit créateur 15, 241–51. Sutherland, S. 1990: ‘Christianity and Tragedy’, Journal of Literature and Theology 4, 157–68. Swift, J. 1985: Gulliver’s Travels, eds. P. Dixon and J. Chalker (Harmondsworth: Penguin). Szondi, P. 2011: Schriften, 2 vols., eds. J. Bollack et al. (Berlin: Suhrkamp).

Bibliography 399 Taplin, O. 1972: ‘Aeschylean Silences and Silences in Aeschylus’, Harvard Studies in Classical Philology 76, 57–97. ———. 1977: The Stagecraft of Aeschylus: The Dramatic Use of Exits and Entrances in Greek Tragedy (Oxford: Clarendon). ———. 1990: ‘Agamemnon’s Role in the Iliad’, in C. Pelling (ed.), Characterization and Individuality in Greek Literature (Oxford: Clarendon), 60–82. ———. 1996: ‘Comedy and the Tragic’, in M. Silk (ed.), Tragedy and the Tragic: Greek Theatre and Beyond (Oxford: Oxford University Press), 188–202. ———. 1999: ‘Spreading the Word Through Performance’, in S. Goldhill and R. Osborne (eds.), Performance Culture and Athenian Democracy (Cambridge: Cambridge University Press), 33–57. ———. 2001: ‘The Shield of Achilles Within the Iliad’, in D. Cairns (ed.), Oxford Readings in Homer’s Iliad (Oxford: Oxford University Press), 342–64. ———. 2003: Greek Tragedy in Action, 2nd edn. (London: Routledge). Tarrant, R. 1976: Seneca: Agamemnon (Cambridge: Cambridge University Press). ———. 1985: Seneca’s Thyestes (Atlanta, Georgia: Scholars). ———. 2012: Virgil: Aeneid XII (Cambridge: Cambridge University Press). Tasso, T. 2009: Gerusalemme Liberata, ed. F. Tomasi (Milan: RCS). Taxidou, O. 2004: Tragedy, Modernity, and Mourning (Edinburgh: Edinburgh University Press). Taylor, A. 1979: ‘Troilus’ Rhetorical Failure (4: 1440–1526)’, Papers on Language and Literature 15, 357–69. Taylor, G. 2002: ‘Bardicide’, in R. Wilson (ed.), New Casebooks: Julius Caesar (Basingstoke: Palgrave Macmillan), 188–209. Ter-Nedden, G. 1986: Lessings Trauerspiele: Der Ursprung des modernen Dramas aus dem Geist der Kritik (Stuttgart: Metzler). ———. 2007: ‘Philotas und Aias, oder Der Kriegsheld im Gefangenendilemma: Lessings Sophokles-Modernisierung and ihre Lektüre durch Gleim, Bodmer und die Germanistik’, in W. Adam and D. Holger (eds.), “Krieg ist mein Lied”: Der Siebenjährige Krieg in den zeitgenössischen Medien (Göttingen: Wallstein), 317–78. Thomas, R. 1988a: Virgil: Georgics I and II (Cambridge: Cambridge University Press). ———. 1988b: Virgil: Georgics III and IV (Cambridge: Cambridge University Press). ———. 1998: ‘The Isolation of Turnus: Aeneid Book 12’, in H-P. Stahl (ed.), Vergil’s Aeneid: Augustan Epic and Political Context (Swansea: Classical Press of Wales), 271–302. Thompson, A. and Taylor, N. 2016: Shakespeare: Hamlet, rev. edn. (London: Bloomsbury). Tillyard, E. 1956: The Elizabethan World Picture (London: Chatto and Windus). Tolstoy, L. 1960: The Cossacks, Happy Ever After, the Death of Ivan Ilyich, tr. R. Edmonds (Harmondsworth: Penguin). Torrance, I. 2010: ‘Antigone and Her Brother: What Sort of Special Relationship?’ in S. Wilmer and A. Žukauskaitė (eds.), Interrogating Antigone in Postmodern Philosophy and Criticism (Oxford: Oxford University Press), 240–53. Trapp, M. 1996: ‘Tragedy and the Fragility of Moral Reasoning: Response to Foley’, in M. Silk (ed.), Tragedy and the Tragic: Greek Theatre and Beyond (Oxford: Oxford University Press), 74–84.

400  Bibliography Trigg, D. 2004: ‘Schopenhauer and the Sublime Pleasure of Tragedy’, Philosophy and Literature 28, 165–79. Trissino, G. 1975: ‘Sofonisba’, in B. Corrigan (ed.), Two Renaissance Plays (Manchester: Manchester University Press). Vellacott, P. 1974a: Euripides: The Bacchae and Other Plays (Harmondsworth: Penguin). ———. 1974b: Euripides: Alcestis/Hippolytus/Iphigenia in Tauris (Harmondsworth: Penguin). Vernant, J-P. 1991: Mortals and Immortals: Collected Essays, ed. F. Zeitlin (Princeton: Princeton University Press). ———. 1996: Mythe et pensée chez les Grecs: Études de psychologie historique (Paris: La Découverte). Vernant, J-P. and Vidal-Naquet, P. 2001: Mythe et tragédie en Grèce ancienne, 2 vols. (Paris: La Découverte). Vickers, B. 1973: Towards Greek Tragedy: Drama, Myth, Society (London: Longman). ———. 1994: Appropriating Shakespeare: Contemporary Critical Quarrels (New Haven: Yale University Press). Wackernagel, J. 2009: Vorlesungen über Syntax, mit besonderer Berücksichtigung von Griechisch, Lateinisch and Deutsch, 2 vols. (Cambridge: Cambridge University Press, repr. of Basel 1920/4). Waismann, F. 1976: Logik, Sprache, Philosophie (Stuttgart: Reclam). Walcot, P. 1984: ‘Greek Attitudes Towards Women: The Mythological Evidence’, Greece and Rome 31, 37–47. Waldock, A. J. 1951: Sophocles the Dramatist (Cambridge: Cambridge University Press). ———. 1975: Hamlet: a Study in Critical Method (Sydney: Sydney University Press). Walker, C. 1988: ‘ “So tief liess mich die Gnade nicht sinken”: On the Absence of Divine Grace in Emilia Galotti’, Publications of the English Goethe Society 57, 75–84. Warmington, E. 1936: Remains of Old Latin: Livius Andronicus, Naevius, Pacuvius, Accius (Cambridge, MA: Harvard University Press). Waswo, R. 1974: ‘Damnation, Protestant Style: Macbeth, Faustus, and Christian Tragedy’, Journal of Medieval and Renaissance Studies 4, 63–99. Webster, J. 1996: The White Devil, ed. C. Luckyj (London: Bloomsbury). Weissberg, L. 1989: ‘Language’s Wound: Herder, Philoctetes, and the Origin of Speech’, Modern Language Notes 104, 548–79. Wells, G. 1984: ‘What Is Wrong with Emilia Galotti?’ German Life and Letters 37, 163–73. Wells, S. 2016: Shakespeare on Page and Stage: Selected Essays, ed. P. Edmonson (Oxford: Oxford University Press). Werner, S. 2009: ‘Arming Cordelia: Character and Performance’, in P. Yachnin and J. Slights (eds.), Shakespeare and Character: Theory, History, Performance, and Theatrical Persons (Basingstoke: Palgrave Macmillan), 232–49. West, M. 1999: ‘Ancestral Curses’, in J. Griffin (ed.), Sophocles Revisited: Essays Presented to Sir Hugh Lloyd-Jones (Oxford: Oxford University Press), 31–45. West, S. 1999: ‘Sophocles’ Antigone and Herodotus Book Three’, in J. Griffin (ed.), Sophocles Revisited: Essays Presented to Sir Hugh Lloyd-Jones (Oxford: Oxford University Press), 109–36.

Bibliography 401 Wheeler, G. 2003: ‘Gender and Transgression in Sophocles’ Electra’, Classical Quarterly 53, 377–88. White, S. 1992: ‘Aristotle’s Favorite Tragedies’, in A. Rorty (ed.), Essays on Aristotle’s Poetics (Princeton: Princeton University Press), 221–40. Whitehorne, J. 1983: ‘The Background to Polyneices’ Disinterment and Reburial’, Greece and Rome 30, 129–42. Whitman, C. 1951: Sophocles: a Study of Heroic Humanism (Cambridge, MA: Harvard University Press). Wiechert, E. 2008: Der Totenwald (Frankfurt am Main: Suhrkamp). Wilamowitz-Möllendorff, T. von. 1996: Die dramatische Technik des Sophokles, 4th edn., ed. E. Kapp (Hildesheim: Weidmann). Wilamowitz-Möllendorff, U. von. 1872: Zukunftsphilologie! (Berlin: Borntraeger). ———. 1935–72: Kleine Schriften, 6 vols. (Berlin: Akademie-Verlag). ———. 1984: Euripides: Herakles, 3 vols. (Darmstadt: WBG). Willcock, M. 1976: A Companion to the Iliad (Chicago: University of Chicago Press). Williams, B. 1972: Morality: An Introduction to Ethics (Cambridge: Cambridge University Press). ———. 1978: Descartes: The Project of Pure Enquiry (Harmondsworth: Penguin). ———. 1981: Moral Luck: Philosophical Papers 1973–1980 (Cambridge: Cambridge University Press). ———. 1985: Ethics and the Limits of Philosophy (London: Fontana/Collins). ———. 1993: Shame and Necessity (Berkeley: University of California Press). ———. 1995: ‘Replies’, in J. Altham and R. Harrison (eds.), World, Mind, and Ethics: Essays on the Ethical Philosophy of Bernard Williams (Cambridge: Cambridge University Press), 185–224. ———. 2002: Truth and Truthfulness: an Essay in Genealogy (Princeton: Princeton University Press). ———. 2006: The Sense of the Past: Essays in the History of Philosophy (Princeton: Princeton University Press). Williams, R. D. 1990: ‘The Purpose of the Aeneid’, in S. Harrison (ed.), Oxford Readings in Virgil’s Aeneid (Oxford: Oxford University Press), 21–36. Williams, Raymond 1966: Modern Tragedy (London: Chatto and Windus). ———. 1977: Marxism and Literature (Oxford: Oxford University Press). Williams, Rowan 2016: The Tragic Imagination (Oxford: Oxford University Press). Williamson, T. 1987–8: ‘Equivocation and Existence’, Proceedings of the Aristotelian Society 88, 109–27. ———. 2001: ‘Ethics, Supervenience, and Ramsey Sentences’, Philosophy and Phenomenological Research 62, 625–30. ———. 2007: The Philosophy of Philosophy (Oxford: Wiley-Blackwell). ———. 2013: Modal Logic as Metaphysics (Oxford: Oxford University Press). Willink, C. 1986: Euripides: Orestes (Oxford: Clarendon). Wilmot, R. et al. 1912: The Tragedy of Tancred and Gismund, ed. J. Farmer (London: Tudor Facsimile Texts). Wilson, E. 2004: Mocked with Death: Tragic Overliving from Sophocles to Milton (Baltimore: Johns Hopkins University Press). Wilson Knight, G. 1986: The Wheel of Fire: Interpretations of Shakespearean Tragedy (London: Methuen).

402  Bibliography Wilson, R. and Dutton, R. 1992: New Historicism and Renaissance Drama (London: Longman). Winnington-Ingram, R. P. 1980: Sophocles: an Interpretation (Cambridge: Cambridge University Press). ———. 1983: Studies in Aeschylus (Cambridge: Cambridge University Press). Winterbottom, M. 1993: ‘Aeneas and the Idea of Troy’, Proceedings of the Virgil Society 21, 17–34. Wittgenstein, L. 1922: Tractatus Logico-Philosophicus (London: Routledge). ———. 1973: Philosophische Grammatik (Frankfurt am Main: Suhrkamp). ———. 1977: Philosophische Untersuchungen (Frankfurt am Main: Suhrkamp), tr. as Philosophical Investigations, ed. G. E. M. Anscombe (Oxford: WileyBlackwell, 1958). ———. 1980: Briefe, eds. B. McGuinness and G. H. von Wright (Frankfurt am Main: Suhrkamp). ———. 1984: Bemerkungen über die Farben; Über Gewissheit; Zettel; Vermischte Bemerkungen, eds. G. E. M. Anscombe et al. (Frankfurt am Main: Suhrkamp). ———. 1993a: Wittgenstein und der Wiener Kreis (Frankfurt am Main: Suhrkamp). ———. 1993b: Philosophical Occasions: 1912–1951, eds. J. Klagge and A. Nordmann (Indianapolis: Hackett). Wlosok, A. 1990: Res Humanae—Res Divinae: Kleine Schriften, eds. E. Heck and E. A. Schmidt (Heidelberg: Winter). Wohl, V. 2009: ‘Tragedy and Feminism’, in R. Bushnell (ed.), A Companion to Tragedy (Oxford: Wiley-Blackwell), 145–60. Wolf, S. 1993: Freedom Within Reason (Oxford: Oxford University Press). Woodbridge, L. 1977: ‘Egyptian Queens and Male Reviewers: Sexist Attitudes in Antony and Cleopatra Criticism’, Shakespeare Quarterly 28, 297–316. Woodruff, P. 2012: ‘The Philoctetes of Sophocles’, in K. Ormand (ed.), A Companion to Sophocles (Oxford: Wiley-Blackwell), 126–40. Woolf, V. 1992: A Room of One’s Own, and Three Guineas (Oxford: Oxford University Press). ———. 2009: Selected Essays, ed. D. Bradshaw (Oxford: Oxford University Press). Wordsworth, W. 1932: The Poetical Works of Wordsworth, ed. T. Hutchinson (London: Oxford University Press). Wray, D. 2009: ‘Seneca and Tragedy’s Reason’, in S. Bartsch and D. Wray (eds.), Seneca and the Self (Cambridge: Cambridge University Press), 237–54. Young, J. 2013: The Philosophy of Tragedy: From Plato to Žižek (Cambridge: Cambridge University Press). Zacchia da Vezzano, L. 1884: ‘De Captivitate ducis Jacobi tragoedia’, ed. C. Braggio, Giornale ligustico di archeologia 11, 50–76, 111–32. Zagzebski, L. 1991: The Dilemma of Freedom and Foreknowledge (Oxford: Oxford University Press). Zamir, T. 2016: ‘Ethics and Shakespearean Tragedy’, in M. Neill and D. Schalkwyk (eds.), The Oxford Handbook of Shakespearean Tragedy (Oxford: Oxford University Press), 71–88. ———. 2018: Ascent: Philosophy and Paradise Lost (Oxford: Oxford University Press).

Bibliography 403 Zeitlin, F. 1982: Under the Sign of the Shield: Semiotics and Aeschylus’ Seven Against Thebes (Rome: Edizioni dell’ Ateneo). ———. 1986: ‘Thebes: Theater of Self and Society in Athenian Drama’, in J. P. Euben (ed.), Greek Tragedy and Political Theory (Berkeley: University of California Press), 101–41. ———. 1995: ‘Art, Memory, and Kleos in Euripides’ Iphigeneia in Aulis’, in B. Goff (ed.), History, Tragedy, Theory: Dialogues on Athenian Drama (Austin: University of Texas Press), 174–201. ———. 1996: Playing the Other: Gender and Society in Classical Greek Literature (Chicago: University of Chicago Press). ———. 2003: ‘The Closet of Masks: Role-Playing and Myth-Making in the Orestes of Euripides’, in J. Mossman (ed.), Oxford Readings in Classical Studies: Euripides (Oxford: Oxford University Press), 309–41. Žižek, S. 2004: ‘From Antigone to Joan of Arc’, Helios 31, 51–62. ———. 2005: Interrogating the Real, eds. R. Butler and S. Stephens (London: Bloomsbury).

Index

abstract(ion) 67, 156, 216, 219, 229, 288 – 90, 323, 331, 340 – 6, 350 accident(al) 41 – 4, 49 – 58, 74, 134, 156, 173, 343 adequacy of language see expressibility/inexpressibility adolescence 91, 102, 106, 116, 119, 173 – 4, 204, 315 Aeschylus 31, 36, 38 – 9, 48, 55, 58, 123n10, 137 – 40, 147 – 8, 155 – 8, 214 – 15, 228, 248 – 50, 252 – 3, 338; Agamemnon 4, 34, 48, 54, 62 – 3, 65 – 6, 111 – 12, 147 – 8, 154, 162 – 5, 167, 173, 176, 193, 210, 223 – 5, 240, 248 – 50, 282, 301, 348, 355 – 6n48, 356n64; Choephori 64, 120, 137 – 40, 175, 202, 232, 234 – 5, 248 – 50, 264n234, 282, 305 – 6, 339, 356n64; Eumenides 64, 138, 156, 202, 244, 248 – 50, 338; Persae 48, 51, 56, 60, 62, 136, 209 – 10, 263n197, 277, 296, 305, 320n142; Seven Against Thebes 31, 48, 66, 85, 89, 97, 108, 112, 124n53, 168, 294, 320n142, 355 – 6n48; Suppliants 54, 87, 89, 109, 242, 320n151 [Aeschylus], Prometheus Vinctus 48, 54, 67, 97, 107, 117, 165, 235, 310, 320n142 agreement/disagreement 195 – 7, 199 – 203, 248 – 57, 274 akrasia 24, 29, 35, 37 Alabaster, W. 48, 135, 164, 319n104, 356n57 allegory 63 – 70, 134 – 5, 140 ambiguity 56, 136, 142, 202, 247 – 57, 264n245, 269

Améry, J. 154, 269 – 70, 283 – 5, 291, 295, 298, 315, 351 anachronism 114, 121, 152, 206, 209 – 13, 223 – 5 Anouilh, J. 104 – 7, 164, 175, 212, 234 Aquinas, T. 78n127, 81n205, 173 Aretino, P. 48, 116 Aristophanes 137, 152, 256, 260n115, 302, 317n37, 320n135, 320n142, 320n155, 355 – 6n48 Aristotle 1, 3 – 7, 12, 21, 23 – 31, 34, 37, 39, 41 – 58, 60, 68 – 70, 84, 89, 92, 102, 109 – 11, 118, 131, 134, 141 – 2, 144 – 6, 153, 157 – 8, 165, 167, 170 – 8, 190 – 1, 203, 205, 208, 224, 226, 228, 233, 242, 244 – 7, 268, 318n77, 333, 344 Arnold, M. 6, 63, 150, 205, 307 Artaud, A. 142, 161 Athens 3, 11, 44, 62, 87, 92, 101, 108, 137, 142, 169, 180n68, 201, 210, 213, 239 – 47, 301, 334, 352 Augustine viii, 173, 244 Auschwitz 13, 268 – 70, 280 – 3, 293 – 4, 315, 351 Austen, J. 304, 307, 336 Bayley, J. 103, 277, 335 Beckett, S. 54, 142, 273, 298 – 9 Bellini, V. 31, 60, 299 Belsey, C. 201, 204 – 5, 229 Berg, A. 122, 142 – 3, 314 Billings, J. 141, 200, 259n55 blame see culpability Boccacio, G. 144 – 5 Böckh, A. 85, 96, 112 – 13, 122, 209 Boethius 50, 79n156, 144 – 6, 287, 356n75

Index  405 Boyle, A. 262n166, 276 – 8, 317n36, 317n49, 318n71, 355n23, 355n36, 355n38 Bradley, A. C. 8 – 9, 23, 27, 41 – 2, 46, 52 – 3, 68, 114 – 16, 122, 131 – 5, 141, 145 – 6, 177 – 8, 220, 238 – 9 Brecht, B. 4, 176, 302, 306 – 7, 353 Buchanan, G. 8, 10, 48, 153 – 4, 164 – 5, 274 Buchenwald 269 – 70, 279 – 80, 354 Büchner, G. 4 – 5, 31, 48, 70, 142 – 4, 148, 179 – 80n60, 298, 313 – 14 Budelmann, F. 237, 245 – 6, 263n228, 265n264, 317n41, 317n51, 320n149, 355n32 burial 86 – 114, 118 – 21, 147, 181n90, 205, 240 – 1, 249 Butcher, S. 22, 26 – 7, 36, 47, 84 Cairns, D. 92 – 3, 103, 125n76, 218 – 19, 264n242 Calderón, P. 8, 61, 147, 150, 174 Campbell, J. 15n15, 135, 178n14, 179n58, 182n126, 259n90, 263n220, 321n195, 321n198, 336 – 7, 356n54 ‘catastrophe survived’ see tragicomedy category distinction/mistake 27, 50, 79n157, 122, 229, 231, 234 – 5, 261n157, 262n166, 286 – 7, 344 causation 21, 26 – 7, 32, 43 – 4, 49 – 59, 60, 63, 67 – 71, 131 – 4, 137, 145, 172, 190 – 1, 292, 349 Cavell, S. 52 – 3, 70, 158 Cervantes, M. de 143, 150 – 1 Chanson de Roland 1, 48, 140, 146 character 25 – 43, 46 – 9, 52 – 3, 59 – 70, 74, 84, 103 – 4, 107, 131 – 41, 145, 160, 169, 172, 225 – 39, 347 – 50 Chaucer, G. 1, 4, 48, 56, 71, 79n156, 94, 106, 111, 133, 144 – 6, 164, 166, 170 – 2, 176, 201, 216 – 17, 224, 243, 260n110, 274, 310, 313, 353 children/childhood 21, 47 – 8, 64, 91 – 2, 100 – 1, 108, 111, 116 – 19, 146, 148, 157, 163 – 9, 173 – 6, 204, 279, 285, 301, 315 Christ/Christianity 8, 10 – 11, 13, 14, 42, 50, 54 – 8, 84 – 5, 110, 143, 158 – 9, 165, 175, 213 – 18, 270 – 1, 349 Cicero 10, 140, 179n43, 261n150, 287, 300, 319n131, 321n188 cognitive failure 8 – 9, 24, 30 – 7, 42 – 3, 46 – 9, 70, 84, 88 – 9, 104, 108,

115 – 17, 138, 142 – 3, 146, 154 – 7, 160, 173 – 8, 189, 243, 271, 313, 347 – 50 coincidence see accident(al) Coleridge, S. 69, 147, 269, 293 comedy 6 – 7, 53, 104, 150 – 4, 256 communication 190 – 203, 247 – 57, 283, 289, 309 compatibilism see freedom compensation see redress concept/conceptual 195 – 203, 212 – 33, 248 – 57, 281, 297 – 8, 309, 342 – 6, 353 – 4 constructive knowledge 35 – 7, 40, 42 – 3, 132, 160, 213 contingency 49 – 58, 71 – 2, 158, 170, 173, 296, 312 – 13 Cooper, H. 63, 145, 272 – 3, 342 Corneille, P. 161 – 2, 227, 243, 332 – 6; Cinna 150, 162; Héraclius 161 – 2; Horace 116, 162, 227, 355n39; La Mort de Pompée 334; Le Cid 7, 69, 150, 333 – 4; Médée 164, 175, 333 – 4; Nicomède 161 – 2; Polyeucte 8, 109, 161 – 2; Rodogune 121, 164; Sertorius 162, 334 – 5; Suréna 48, 162 Correr, G. 87, 135, 161, 164 counterfactuality 15, 53, 70 – 1, 101, 158 – 9, 163 – 4, 179n60, 205, 279, 312 – 13 culpability 8 – 9, 11 – 12, 21 – 37, 41 – 4, 46 – 9, 53, 60, 62 – 6, 69 – 70, 74, 84 – 5, 89, 94, 96, 110, 114 – 15, 122, 125n76, 131 – 2, 136, 140 – 6, 153, 155, 160 – 9, 171 – 8, 201 – 2, 218 – 19, 240, 249, 266 – 7, 347 – 51 cultural materialism see new historicism culture 2 – 3, 115, 198 – 226, 233, 241 – 7, 270, 338, 352 Dacier, A. 8, 22, 23, 27 Dale, A. M. 100, 109 – 10, 232, 259n65 Dante 5 – 7, 79n156, 164 – 5, 269, 274, 283 Davidson, D. 196, 199 – 203, 207 – 9, 211, 216 – 20, 238, 251, 303 – 4, 311 – 12 Dawe, R. 32 – 4, 40, 72 – 4, 112 deceit 193, 213, 233 – 5, 238 – 9, 251 – 7, 311 – 12 deconstruction 198 – 9, 247 – 57 defect see flaw/defect

406 Index deliberation 24, 29, 50, 155, 204, 243 democracy 88 – 90, 241 – 7 De Quincey, T. 84, 181n94, 321n203, 354 Derrida, J. 193, 252 – 5, 342 Descartes, R. 39, 127n117, 137, 227, 233, 237 – 8, 297, 328, 334 – 5 describing see expressibility/ inexpressibility desert see culpability; undeserved suffering destiny 48, 50 – 3, 58, 63, 66, 69 – 70, 131 – 7, 141, 144 – 6, 171 Dinge an sich see noumena disagreement see agreement/ disagreement Dodds, E. R. 23 – 5, 29 – 32, 42 – 4, 46, 49 – 50, 59, 61, 63 – 4, 74, 104, 112, 152, 211, 239 – 41, 303, 333 Dollimore, J. 204 – 7, 270 – 1 dramatic irony 109, 157, 192, 245, 252 Dryden, J. 294, 334 – 5 duBois, P. 190 – 1, 216, 243 Dvořák, A. 314, 354 Eliot, G. 48, 52, 104, 112, 227 Eliot, T. S. 60, 69, 143, 217, 272, 278, 332, 337, 339 emotion 3, 5, 7, 10, 13, 23, 35, 42 – 3, 135, 151, 204, 212 – 26, 233, 267, 272 – 3, 275, 289 – 91, 295, 308, 310 – 11, 328 – 32 empirical/empiricist 2, 71, 202, 204, 208, 292, 343, 352 – 4 Enlightenment 35 – 41, 68, 143, 171, 270 – 1 error see cognitive failure; culpability; hamartia essentialism 197 – 8, 203 – 9, 228 Euripides 6, 8, 39, 53, 58, 92, 109, 137 – 8, 147 – 53, 158, 201, 212, 230, 267 – 8, 337; Alcestis 16, 67, 97, 109, 150, 210 – 11, 217, 225, 313, 355 – 6n48; Andromache 167, 320n142, 355 – 6n48; Bacchae 4, 7, 10, 48, 61, 64, 109, 112, 138 – 9, 142 – 3, 152, 211, 215, 303, 333, 338; Children of Heracles 10, 152, 161, 164 – 5, 287; Cyclops 7, 150, 235; Electra 97, 120, 152, 202, 212, 224, 242, 258n33; Hecuba 10, 108 – 9, 142 – 3, 160 – 1, 164, 167, 294, 299, 333, 338, 355 – 6n48; Helen 61, 120, 150, 249, 255, 355 – 6n48; Heracles 39, 42, 48, 61,

64 – 5, 164, 174; Hippolytus 28, 30, 39, 48, 61, 64, 89, 104, 109, 114, 131, 134, 138, 143, 163, 166, 210 – 12, 227 – 8, 250 – 4, 271 – 2, 294, 317n47, 320n142, 355 – 6n48; Ion 121, 150, 208, 210, 317n47, 355 – 6n48; Iphigeneia at Aulis 10, 24, 164 – 5; Iphigeneia in Tauris 6, 8, 150, 212 – 13, 259n66; Medea 45, 48, 121 – 2, 142 – 3, 146, 148, 161, 164, 169, 174 – 5, 200, 246, 333, 355 – 6n48; Oedipus 36, 38, 72, 74; Orestes 24, 31, 65, 112, 140, 152, 202; Phoenician Women 10, 31, 44 – 5, 89, 108, 124n53, 164, 235, 248 – 9, 254, 279, 286 – 7, 299, 305, 355 – 6n48; Suppliants 16, 89, 108, 147, 164 – 5, 205, 242, 299, 320n150; Trojan Women 53, 65, 106, 142 – 3, 158, 160 – 1, 164 – 5, 168, 189, 271 – 2, 320n142, 338 – 9 [Euripides], Rhesus 48, 355 – 6n48 evil 3, 14, 24 – 31, 41 – 3, 69, 85, 89, 97 – 8, 170, 177 – 8, 250, 266 – 8, 281, 298 exile 108, 168, 286 – 7 expressibility/inexpressibility 194, 197, 212 – 15, 217 – 20, 226, 232, 267 – 315, 323 – 6, 328 – 32, 340 – 54 extension(al) 195 – 6, 248, 261n133 falsity see truth/falsilty fate see destiny fault see culpability fear 5, 23, 29, 35, 43, 49, 244 – 7, 297 feminism 85, 98, 114, 118 – 22, 204, 206, 246, 348 fiction 122, 235, 237, 244, 279 Finglass, P. 66, 91, 210, 236, 321n175, 355n30 Flaig, E. 44 – 5, 80n165 Flaubert, G. 48, 143 flaw/defect 23 – 37, 41 – 2, 46 – 9, 52, 84 – 5, 107, 115 – 17, 122, 131 – 2, 138, 141 – 5, 243, 267 Fletcher, J. 48, 150, 155, 162, 294 Foakes, R. 129n204, 195 – 7, 253 foreknowledge 49 – 58 fortune 3, 6, 8, 27, 38, 144 – 6, 157 Fraenkel, E. 65, 111, 147 – 8, 223 – 5, 235, 240, 301, 305 Franke, W. 341 – 3, 346, 354n7 freedom 49 – 59, 65 – 70, 71 – 2, 98, 115, 132, 138, 141, 144, 163,

Index  407 170 – 4, 178, 179 – 80n60, 215, 336, 348 Frege, G. 40, 195, 212, 257n12, 302, 307 – 9, 323, 331 Freud, S. 40, 73, 206, 208, 210 – 16, 259n76 Gager, W. 83n278, 101, 136, 141, 168, 319n104 Gardner, S. 11 – 12, 169 – 70, 176 – 7, 189, 266 – 7, 350 Garnier, R. 10, 31, 48, 66, 85, 90, 135, 173, 243, 319n102, 333 Garvie, A. F. 7, 31, 51, 62, 66, 90, 136, 200, 234 – 6, 263n197, 264n234, 277, 317n47, 339 Gervinus, G. 117, 122, 131, 141, 162 gesture 237, 273, 296, 304 – 9, 331 Gide, A. 73, 104 Goethe, J. 69, 98, 101 – 2, 135, 140, 158, 172, 233, 243, 269, 329, 354; Clavigo 165; Egmont 66, 112, 162; Faust 60, 73, 134, 159, 165 – 6, 323, 357n94; Götz von Berlichingen 48, 70, 112, 333; Iphigenie 98; Torquato Tasso 45, 106, 357n88 Goldhill, S. 32, 39, 193, 204, 213 – 18, 220 – 27, 231 – 3, 242 – 54 Goux, J-J. 32, 39, 104 grammar/grammaticality 189 – 97, 302, 327, 345 Greenberg, M. 217, 227, 259n77, 336 Greenblatt, S. 135, 204, 321n193, 342 Gregory, J. 108, 182n138, 210 – 11, 225, 333 Griffith, M. 90, 94 – 5, 99, 110, 120, 127n142, 210 Grillparzer, F. 34, 48, 164 – 5, 172, 175, 241, 243 guilt see culpability Hall, E. 69, 203, 243, 263n221 Halliwell, S. 60, 69, 80n165, 138, 215, 217, 259n89 hamartia 3, 23 – 39, 41 – 9, 69 – 70, 84 – 5, 89, 91, 96, 116, 138, 142, 144 – 5, 165 – 6, 171 – 6, 189, 191, 210, 228, 244, 277, 338, 348 – 9 Hammond, P. 95 – 6, 190 – 4, 253 Hardy, T. 7, 48, 66, 112, 136, 165, 192, 227, 272, 275, 305 Hauptmann, G. 4, 5, 10 health 104 – 5, 204 Heath, M. viii, 16n27, 75n9, 134, 176, 216, 310 – 11, 315

Hebbel, F. 48, 117, 155, 166, 243 Hegel, G. 9, 39, 46, 74, 84 – 5, 88, 90, 100 – 1, 103 – 4, 115, 127n26, 233, 352 Heisenberg, W. 69 – 70 Hemingway, E. 294, 306 – 7 Henryson, R. 48, 138, 224 – 5 Heraclitus 59, 132, 155 Herodotus 74, 80n195, 194 Heywood, T. 4, 80n195, 166 Hofmannsthal, H. von 73, 85, 92, 152 – 3, 212, 275 Holbrook, P. 115, 228 Hölderlin, F. 99, 272, 294 Holocaust 13 – 15, 268 – 70, 273 – 4, 279 – 85, 291, 311 – 13 Homer 4, 5, 7, 21, 43, 48, 59 – 60, 62, 66, 72, 87, 89, 94, 113, 146 – 51, 164, 167, 201 – 2, 204, 208 – 9, 212, 215, 220, 224, 226 – 7, 237, 243, 271, 276, 304, 306, 308, 317n37, 331 Honig, B. 88, 95, 101, 120 – 1, 349 Horace 5, 134, 144, 160, 235 – 6, 289 – 91, 318n77, 333, 353 Housman, A. E. 98, 230 Hoxby, B. 68, 226, 335, 355n32 Hughes, T. 48, 66, 83n278, 112, 135, 319n104 humanism 23, 86 – 7, 97, 174, 198, 205, 213 – 18, 228, 270 – 2, 338, 349 – 50 Hume, D. 2, 40, 58, 68, 205 – 6, 226, 267 – 8, 304, 306, 308 Huxley, A. 148 – 51 Ibsen, H. 154 – 8, 225; An Enemy of the People 154 – 5; Ghosts 154, 336; Hedda Gabler 112; The Lady from the Sea 55, 275, 310, 337; Pillars of Society 154, 158; The Wild Duck 48, 165 incest 22, 30 – 1, 35 – 6, 41, 59, 72, 74, 98 – 9, 104, 116, 161, 210, 294 ineffability see expressibility/ inexpressibility inexpressibility see expressibility/ inexpressibility involuntary see voluntary/involuntary Janáček, L. 166 – 7 Jebb, R. 51, 84 – 5, 98 – 9, 105, 110, 120, 157, 301, 303 Johnson, S. 5 – 6, 67, 147 – 8, 153, 177, 303 Jones, J. 75n7, 225 – 6, 233 – 9

408 Index Jonson, B. 7, 48, 135 – 6, 150, 164, 355n21 justice 3, 11 – 12, 22, 24, 26, 28 – 9, 31, 42, 49, 68, 93, 105 – 6, 110, 112, 133, 195 – 7, 200, 202, 213, 240, 248 – 50, 253 Kafka, F. 177, 298, 311 – 12 Kant, I. 69, 112, 227, 256, 281, 297 – 300, 323, 342 – 3, 351 Kastan, D. 11 – 12, 182n118, 268 Kelly, H. 1, 144 – 6 Kierkegaard, S. 9, 54, 98, 203, 233 Kleist, H. von 136 – 7, 158 – 9, 243; Der Zerbrochne Krug 6 – 7, 151; Käthchen von Heilbronn 137; Penthesilea 48, 137, 140, 155; Prinz Friedrich von Homburg 118, 137, 159 Kovacs, D. 71, 121, 202, 223, 265n257 Lacan, J. 85, 96, 106, 298 Lamb, C. 150 – 1, 229, 268 – 9, 294, 303 Lefèvre, E. 23, 33, 72, 81n205, 95, 112, 355n23 Leibniz, G. 47, 131, 175, 248, 352 Lesky, A. 59 Lessing, G. 6, 8, 12, 27, 70, 171 – 6, 243, 268, 270, 291, 300, 338, 355n32, 355n37; Emilia Galotti 5, 9, 48, 53, 121, 152, 166, 171 – 6, 198, 241, 266 – 7; Minna von Barnhelm 152, 174; Miss Sara Sampson 5, 9, 121, 135, 164, 171 – 6; Nathan der Weise 174, 313; Philotas 106, 173 – 4 Levi, P. 154, 268 – 9, 313, 315 Lewis, C. S. 67, 111, 217, 307 limits of language 274, 298, 326, 335, 341, 354n7 linguistic idealism 252, 315, 322 – 5, 327 – 35, 337, 339, 341 – 53 Livy 80n195, 164, 166, 170 – 2, 175, 201, 241 Lloyd-Jones, H. 31, 51, 85, 94, 105, 132, 223 – 4, 240, 301 Locrine 48, 136, 164, 181n94 Longinus 140, 179n37, 179n43, 263n209, 304, 320n165, 354n8 Lope de Vega, F. 136, 150, 235, 258n34 Loschi, A. 136, 146, 178n23 Lucan 17n58, 48, 65

Lucas, D. W. 47, 73, 85, 111 Lucas, F. L. 157, 175, 182n148, 272, 355n23 Lucretius 108, 140, 164 Macleod, C. 5, 242, 263n197 Mann, T. 61, 223, 227, 269 – 70, 274, 310 – 11 Marlowe, C. 8, 28, 33, 48 – 9, 61, 112, 142, 168, 228, 243, 337, 339 Marston, J. 79n156, 136, 164, 184n207, 319n104 Marxism 155, 228 – 9 masking 136, 210, 225, 233 – 9 Mastronarde, D. 42, 121, 124n53, 169, 254, 287 McDowell, J. 14 – 15, 262n172, 281, 297, 319n98, 354n4 meaning 3, 23 – 4, 49, 51, 56 – 8, 127n117, 190 – 209, 216 – 17, 228 – 39, 245 – 57, 289 – 97, 323 – 32, 334, 340 – 54 meaning indeterminacy 56 – 8, 193, 197 – 8, 240, 246 – 57 Melville, H. 45, 117 – 18, 176, 275, 342 middle knowledge 55, 163 Middleton, T. 1, 4, 48, 136, 150, 170, 319n104 Miller, A. 4, 158 Milton, J. 31, 48 – 9, 61, 118, 136, 170 – 2, 179n59, 266, 330 mind/mindedness see psychology modality 49 – 58, 158, 213, 279, 329, 340, 342; see also counterfactuality Moore, A. 257n11, 296, 356n79 moralism/moralizing 28 – 31, 50, 101, 166 – 7, 174, 177, 206, 349 – 50 Mozart, W. 217, 301, 304, 311 music 296, 311 – 15, 317n44, 326 – 32, 340, 344 – 5 naming see expressibility/ inexpressibility naturalism 42, 59 – 70, 87, 112 – 14, 131 – 41, 171, 205, 356n68 necessity see probability/necessity new historicism 198, 203 – 9, 220 – 6 Newton, I. 137 – 8, 223 – 5 Nibelungenlied 48, 64, 71, 147, 274 Nietzsche, F. 9 – 11, 38 – 9, 42, 68, 105, 137 – 9, 198, 205, 217, 276, 299, 315, 343 – 5, 349 Nisbet, H. B. 17n63, 70, 135, 171 – 2, 184n213, 184n214, 184n215,

Index  409 184n217, 184n225, 185n240, 316n4, 321n197 noise see silence/noise nonsense see sense noumena 256 – 7, 297 – 309, 315, 322, 328, 342 – 6, 351 novel 5, 7, 140, 227, 304, 326 – 7, 336 Nussbaum, M. 26, 36, 77n80, 77n88, 78n125, 80n194, 104, 109, 127n199, 128n148, 163, 205, 224, 258n43, 259n78, 260n112, 299 – 301 Nuttall, A. D. 53, 80n172, 257n18, 259n77, 262n168, 309, 318n88 objectivity/objectivism 2, 64, 156, 197 – 8, 212 – 25, 239, 246 – 7 omission 24, 46, 53, 60, 65, 87, 239 – 42, 339 – 41 O’Neill, E. 66, 107 operatur principle 59 – 70, 107, 131 – 41, 165, 211 optimism/pessimism 39, 42, 137 – 41, 156, 215, 272, 315 Orgel, S. 229 – 32 Ovid 80n195, 161, 164, 166, 193, 211 – 12, 259n76, 267, 276, 287, 319n102 Padel, R. 68, 207 – 9, 355n20 pain/suffering 3 – 14, 26, 42 – 3, 84, 95, 110, 115, 131, 141 – 2, 145 – 6, 149, 154, 156, 159 – 78, 189, 204, 266 – 315, 322 – 3, 328, 333, 344 – 5, 351 – 4; see also undeserved suffering painting 166, 261n133, 289 – 91, 318n77, 329 Parker, L. 108, 179n24, 212 – 13, 259n66, 260n99, 260n100, 260n115 Parker, R. 23, 55 – 6, 86 parricide 22, 31, 35, 40 – 6, 59, 72, 74 past see time Pelling, C. 210, 223, 246 perception 37, 47, 50, 64, 153, 198, 204, 289 – 91, 295, 327, 342 – 3 pessimism see optimism/pessimism picture theory 288 – 93, 325 – 6, 340 – 1, 357n92 Pinker, S. 13, 36, 83n281, 112, 258n43

pity 5, 23, 29, 35, 42 – 3, 49, 244 – 7 Plato 5, 38 – 9, 47 – 8, 55, 63, 68 – 9, 92, 137, 139, 151, 170 – 1, 197, 217, 229, 267, 318n77 Plautus 7, 107, 134 – 5, 152, 173, 181n95, 192, 252 poetic justice 31, 49, 177 politics/political 57 – 8, 88 – 9, 103, 105, 115, 119, 122, 204 – 6, 214 – 15, 228, 241 – 8, 252 Poole, A. 123n2, 167, 195 – 6 possible worlds see counterfactuality Poulenc, F. 165, 349 Prévost, J. 65, 168, 305 probability/necessity 6, 26 – 9, 33, 37, 39, 43, 50 – 8, 60 – 70, 106, 131, 160, 311 – 12 propositions 323 – 4, 328, 347, 351, 353 psychologizing/psychologism 23, 59 – 70, 87, 106 – 7, 131 – 41, 237 psychology 23, 64 – 5, 87, 96, 101, 137, 140, 172, 200, 204 – 39, 252, 295 – 7 Puccini, G. 73 – 4, 302, 307, 340 Racine, J. 5, 38, 133 – 4, 137, 142, 243 – 4, 259n77, 267, 332 – 41, 349; Andromaque 109, 134, 337 – 8; Athalie 109, 135, 304, 312, 333 – 4, 349; Bajazet 334; Bérénice 142 – 3, 217, 242, 304, 310, 332 – 41; Britannicus 143, 162, 215, 337; Esther 312 – 13; Iphigénie 163 – 5, 337; La Thébaïde 165, 337; Phèdre 48, 134 – 5, 142 – 3, 149, 304, 332 – 6, 337, 355n45 rationalism 38 – 9, 42, 65, 137 – 41, 155 – 9, 202 rationality 6, 13, 26, 34, 38 – 9, 51 – 8, 73, 132, 137 – 41, 143, 153 – 9, 173 – 4, 194 – 5, 199, 216, 270, 336 reading 190, 221 – 2, 267 – 8, 291, 351 redemption see redress Redfield, J. 5, 89, 123n25, 220, 351 redress 68, 170, 211, 270, 313, 349, 352; linguistic 8 – 11, 161, 189 – 90, 267 – 72, 349 – 54; moral 8 – 11, 41 – 9, 84, 107, 131, 141 – 2, 153 – 69, 171, 175 – 8, 189 – 90, 266 – 8, 347 – 51 reference/referentiality 191 – 7, 212 – 18, 288, 293, 295, 301 – 9, 322 – 4, 328 – 32, 341 – 7, 351

410 Index relativism/relativity/relativization 3, 195 – 7, 205 – 6, 213, 220 – 25, 249, 296 remedy 153 – 60, 201, 349 representation 60 – 3, 140, 150, 209, 229 – 39, 261n133, 289 – 93, 301, 314 – 15, 329, 336, 340, 343 – 4 responsibility 27, 30 – 5, 41, 53, 61 – 9, 82n257, 93 – 4, 112 – 14, 125n76, 137, 141 – 5, 165, 173, 178, 224, 310, 336, 348 – 9, 352 Richardson, S. 5, 170 – 1, 176 Roman de la Rose 79n156, 172, 224 Rousseau, J-J. 111, 140, 289 – 90, 332 Rutherford, R. 39, 56, 96, 104, 126n105, 253 – 4 Sartre, J-P. 106 – 7, 142, 164 Scaliger, J. 5 – 7 scapegoat 10 – 11 Scarry, E. 276, 285, 295, 297, 300 Schiller, F. 31, 69, 170, 233, 243 – 4, 266 – 7, 269; Die Braut von Messina 141; Die Jungfrau von Orleans 10, 140; Die Räuber 89, 333; Don Carlos 267, 269; Kabale und Liebe 166, 333; Maria Stuart 333, 335; Wallenstein 4, 48, 70, 176, 267, 312 Schlegel, A. W. 69, 84, 137, 147 Schmidt, D. 85, 101, 343 – 4 Schmitt, A. 23, 69, 76n35, 81n205, 81n208, 94, 227 Schopenhauer, A. 9, 17, 70, 89, 114, 261n157, 274, 316n19, 328 science 2, 137 – 8, 198, 204, 207, 214 – 15, 220 – 5, 276, 281, 285, 305, 324 – 5, 340 – 1 secularism viii – ix, 23, 39, 106 – 7, 112 – 14, 135, 139, 153 – 9, 178, 323, 338, 349 Segal, C. 254, 256 – 7, 264n242, 316n21, 351 self/subject(ivity) 68, 104, 190 – 1, 204, 215 – 16, 225 – 7, 238 – 9, 281, 296, 322, 335 semantic 197, 217 – 18, 226, 255, 290 – 2, 302 – 12, 326, 328, 331 – 2 Seneca 6, 40, 140, 169, 190, 198, 229, 243, 278, 297, 296, 316n22, 318n71, 332 – 6, 338; Agamemnon 48, 64, 135, 162, 164, 234, 317n48, 356n61; Hercules Furens 42, 48, 81n221, 82n246, 164, 174, 330;

Hercules on Oeta 48, 121, 278, 298; Medea 48, 64, 164, 174, 268, 333 – 4, 355n38; Oedipus 40, 211, 317n48, 329 – 30; Phaedra 140, 212, 319n102; Phoenician Women 90; Thyestes 48, 60, 64, 73, 81n221, 135, 140, 164, 192 – 5, 276, 278, 317n36, 318n83; Trojan Women 108, 164, 305, 317n48 [Seneca], Octavia 108, 140, 164, 178n22, 318n71 sense 212 – 14, 248 – 57, 257n12, 301 – 9, 323, 327 – 32, 343, 354n54 sentences 190 – 2, 197, 245, 251, 281, 284, 288 – 91, 302 – 9, 323 – 8, 331 – 2, 343, 346 – 7, 351 – 3 sexuality 103 – 5, 116, 134, 154 – 5, 171 – 2, 204 – 6, 211, 213 – 18, 235, 240, 251, 264n242, 335 – 6 Shakespeare, W. 46, 69, 100, 131, 137 – 8, 147 – 55, 158, 177, 204, 226 – 9, 237 – 8, 239, 257n16, 243 – 4, 267, 291, 309, 319n104, 332 – 6, 339 – 42; All’s Well That Ends Well 150; Antony and Cleopatra 226, 235; Coriolanus 69, 89, 114 – 15, 231; Cymbeline 150, 236; Hamlet 7, 11 – 12, 36, 48 – 9, 52, 103 – 6, 116 – 17, 135, 137 – 40, 147, 151 – 4, 162, 181n115, 228, 233 – 4, 236, 251 – 2, 259n77, 276, 294, 316n21, 319n104, 328, 348; Julius Caesar 4, 10, 46, 104, 136, 144, 164, 167, 228, 334 – 5; King Lear 21, 52 – 3, 69, 74, 104, 113 – 18, 121 – 2, 133, 137 – 8, 140 – 2, 150 – 1, 158, 167, 176, 195 – 7, 228, 230, 235, 253, 268, 270, 272 – 3, 276, 294, 299 – 300, 304, 331, 335, 354; Macbeth 7 – 8, 28, 52 – 3, 68, 70 – 1, 131 – 2, 136 – 40, 144, 146 – 54, 157, 162, 164, 193, 233 – 4, 237 – 8, 259n77, 319n104, 339; Measure for Measure 100, 150, 172; The Merchant of Venice 89; A Midsummer Night’s Dream 63, 151 – 2, 341 – 2; Othello 40, 47, 52 – 3, 89, 104, 116 – 18, 122, 137, 141, 147, 151 – 2, 162, 173, 176 – 7, 191 – 2, 293; Pericles 150, 287; The Rape of Lucrece 161, 166, 170, 176, 201, 234, 310, 316n22; Richard II 286; Richard III 89, 164, 321n193, 333; Romeo

Index  411 and Juliet 53, 70, 151, 182n118, 263n222; The Tempest 150, 230 – 1, 235; Timon of Athens 142, 229, 354; Titus Andronicus 152, 161, 164, 267 – 8, 296, 310 – 11; Troilus and Cressida 132, 150, 217; The Two Noble Kinsmen 133 – 4, 150, 153; Venus and Adonis 211 – 12, 241; The Winter’s Tale 150, 191 – 2 Shapiro, J. 46, 68, 252, 259n90 Shell, A. 114, 133 Shelley, P. 100, 123n2, 142 – 3, 161, 294, 310, 353 silence/noise 169, 210, 241, 300 – 9, 330 – 2, 340 – 1 sin 10, 42, 116 – 17, 141, 145, 171, 173, 177, 196, 244, 348 sleep 65, 149 – 50, 240, 335 – 41 Socrates see Plato Sophocles 137 – 9, 155 – 6, 192, 197, 245 – 7, 267, 337; Ajax 4, 27 – 8, 48, 61 – 2, 64, 66, 86 – 8, 97, 107, 111, 132, 150, 158, 164, 167, 169, 173 – 6, 205, 210, 224, 234, 236, 249, 251 – 3, 256 – 7, 264n245, 293, 301, 305 – 6, 317n47, 355n30, 355 – 6n48; Antigone 2, 8 – 9, 23, 39, 46, 48, 53, 83n289, 84 – 130, 142, 157, 162, 164 – 5, 181n90, 167, 169, 174 – 6, 189 – 91, 194 – 6, 198, 203, 205 – 6, 209 – 11, 214, 229, 240 – 1, 245 – 7, 253, 256, 298, 305, 313, 348, 355 – 6n48; Electra 73, 92, 97, 120, 125n72, 161, 202, 215, 234 – 6, 245 – 6, 249, 254, 256, 305, 319n103, 355 – 6n48; Oedipus at Colonus 72, 138, 140, 157, 256, 278, 294, 317n47, 355 – 6n48; Oedipus Rex 2, 3, 5 – 6, 8 – 11, 21 – 73, 89, 94, 98 – 9, 104, 107, 112, 131 – 2, 136 – 40, 142 – 3, 147, 151, 154, 157, 164, 176, 189, 198, 202 – 3, 211, 213, 229, 235, 239 – 42, 252 – 3, 255 – 6, 272, 294, 305, 313, 348, 355 – 6n48; Philoctetes 138, 152, 224, 226, 234, 253, 255 – 6, 263n197, 278, 295, 300 – 5, 333, 355n32; Women of Trachis 11, 37, 48, 70, 107, 114, 131, 167, 214, 220, 253, 256, 277 – 8, 300 – 1, 305, 310, 333, 355 – 6n48 standard model 115 – 16, 137, 142 – 3, 146, 162 – 9, 171 – 8

Stefonio, B. 136, 143, 293, 355n21 Steiner, G. 5 – 6, 11, 97 – 8, 110, 112 – 14, 131 – 4, 136 – 47, 153 – 6, 160 – 1, 175, 197, 244, 273, 301, 306 – 7, 311 – 15, 324 – 8, 332 – 6, 339 – 41, 348 Stinton, T. 23 – 30, 49 Stoicism 44, 233 Strauss, R. 92, 152 – 3, 212, 340 Strindberg, A. 7, 48, 155 subject see self/subject(ivity) suffering see pain/suffering suicide 21, 72, 74, 99, 119, 157, 165 – 6, 171, 173 – 4, 251, 333 – 4 supernatural 59 – 70, 74, 87, 112 – 14, 131 – 41, 159, 171 syntax 190 – 1, 288 – 9, 300 – 2, 314, 326 – 8, 341 Tacitus 108, 121, 356n64 temporal direction see time Ter-Nedden, G. 62, 166, 184 – 5n228, 224, 258n43 theodicy 14, 266 – 72, 247 – 54 Thucydides 76n40, 123n18, 179n34, 180n68, 317n37, 334 time 9, 31, 33 – 5, 46, 55 – 8, 84, 91, 142, 153 – 64, 197 – 225, 245 – 7, 279, 291, 312, 335 – 41, 352 – 4 tragedy: absolute 141 – 7, 153 – 4, 160 – 1, 348; definition of 1 – 9, 145, 158 – 9; grand 103, 107; medieval 144 – 7; paradox of 316n27 tragicomedy 7, 53, 88, 143, 147 – 54, 158, 162, 312 – 13 transcendent(al)(ism) 69, 71, 112, 138, 203 – 4, 268 – 72, 323, 326, 347, 352 – 4 transhistoricism 2 – 3, 198 – 225, 229, 243 – 7, 336 translation 23 – 7, 37, 51, 85, 88 – 9, 120 – 1, 192 – 3, 197 – 211, 213 – 20, 223, 225, 250 – 3, 258n41, 300 – 4, 325, 342 – 4 treason 86 – 90, 124n53, 249, 256 truth/falsity 192 – 7, 199 – 203, 206, 214, 217, 220 – 25, 228, 233 – 9, 245, 248 – 57, 274, 310 – 15, 328 – 32, 341 – 54 undeserved suffering 10 – 11, 26, 30 – 1, 84, 131, 141 – 2, 145 – 6, 159 – 69, 171 – 8, 189, 348 – 9 universalism see transhistoricism

412 Index Vernant, J-P. 44, 59, 207 – 8, 247 – 9, 263n213 Vickers, B. 29, 32, 36 – 7, 52, 85, 95, 161, 213 – 17, 249 Virgil 17n58, 65, 73, 102, 167 – 8, 215, 243, 272, 274, 276, 287, 293, 305 – 6, 336 – 9; Aeneas 48, 57, 65, 73, 136, 217, 255 – 6, 274, 286, 304, 309, 338 – 9; Dido 5, 27, 42, 48, 168, 217, 304, 309, 338 – 9; Turnus 5, 10, 48, 60, 65, 136, 168, 261n153, 307 virtue 3, 25 – 8, 37, 49, 85, 115, 118, 170 – 8, 201, 211 Voltaire 12, 34, 47 – 8, 70, 137, 143, 355n32 voluntary/involuntary 24, 46, 62 – 3 Wagner, R. 21, 64, 123n2 Webster, J. 48, 117, 136, 164, 241, 296, 333 Wedekind, F. 122, 143 wickedness see evil

Wiechert, E. 269, 354 Wilamowitz, U. 6, 22 – 3, 48, 64, 138 – 9 Williams, B. 11 – 12, 74, 111, 220 – 1, 227, 257n18, 268, 352 – 3, 357n86 Williams, Raymond 3, 69, 259n88, 259n89, 357n91 Williams, Rowan 11 – 12, 36, 43, 89 – 90, 106, 125n77, 132, 139, 159, 268, 316n10 Wilmot, R. 48, 61, 74, 135, 140, 179n25 Wilson Knight, G. 105, 354 Winnington-Ingram, R. 62, 91, 98, 110, 163, 223, 228, 249 Wittgenstein, L. 1, 55, 71, 156 – 8, 196, 199, 233, 238, 267, 288, 290 – 2, 295 – 7, 305, 314, 322 – 9, 340 – 1, 343 – 6, 350, 352 Woolf, V. 119, 276 words vs things 195 – 6, 213 – 26, 231 – 9, 248, 284 – 309, 322 – 54 Žižek, S. 85, 97