NARRATE AND DESCRIBE: T H E PROBLEM OF EKPHRASIS*

By D. P. FOWLER

T h e subject of ekphrasis, and in particular of the ekphrasis of works of art, has recently begun to receive a great deal of attention from classical scholars.' As will become clear, I believe that the reason for this is that many of the theoretical issues that are most pressing in classical studies - and indeed in cultural studies in general2 - are raised by the study of ekphrasis. T h e purpose of this note on the other hand is modest: I want to say a little about the narratological issues that are raised by set-piece description (I), and to look at one example in the Aeneid (11). But even so I have found it impossible not to offer some thoughts of a frighteningly general nature (111). I shall concentrate on the ekphrasis of works of art for reasons that will again become clear, but some at least of what I shall say will also be relevant mutatis mutandis to the ekphrasis of natural features and events.

Set-piece description is regularly seen by narratologists as the paradigm example of narrative pause, in the semi-technical sense of a passage at the level of narration to which

* A first version of this paper was delivered to the University of Bristol Classics Research Seminar in October 1990. I am most grateful for the invitation to Charles Martindale and Duncan Kennedy, and to all those who offered comments, especially Christopher Gill, John Gould, and Malcolm Heath. Subsequent versions were read to the Cambridge Literature Seminar, at the University of Wisconsin-Madison during an all-too-brief two weeks as Brittingham Visiting Professor, at the University of Pittsburgh, and at Northwestern University: again I owe many thanks to m y hosts on these occasions, respectively Richard Hunter, Alessandro Schiesaro, Hans-Peter Stahl, and Francis Dunn, and to all those who offered comments, especially John Henderson and Ian DuQuesnay in Cambridge, Barry Powell, Fanny Lemoine, J i m McKeown, and J e f fWills in Madison, a loyal pupil of Eckard Lefkvre in Pittsburgh, and Bernadette Fort, Daniel Garrison, and Jean Hagstrum in Evanston. In addition, a number o f scholars have been kind enough to comment on written drafts: in particular Alessandro Barchiesi, Irene de Jong, John Elsner, Andrew Laird, Oliver Lyne, Robin Osborne, theeditorial board o f J R S , and the anonymous reader who correctly divined that I want to be loved. T h e usual disclaimers apply, but I hope to be able to take more account o f the criticisms I have received in an expanded version o f this paper to appear in the collection edited by J . R. Elsner (see n. I below). I am thinking especially of E. W . Leach, TheRheton'c of Space: Literary and Artistic Representations of Landscape in Republican and Augustan Rome (1988) and S . Bartsch, Decoding the Ancient Novel, The Reader and the Role of Descnption in Heliodoms and Achilles Tatius (1989),but there are currently two volumes of essays on the subject in preparation from Cambridge, on Greek texts edited by Simon Goldhill and Robin Osborne and on Latin edited by J . R. Elsner. Bartsch offersagood general bibliography: see also those in M. Fusillo, I1 Romanzo Greco (1989), 83-90 and S. Richardson, The Homeric Narrator ( ~ g g o )5c-69, , and add amongst very recent works N . W . Slater, Reading Petronius ( ~ g g o )213-30, , and D. Rosand, 'Ekphrasis and the Generation of 61-105. , T h e standard work Images', Ation NS I ( ~ g g o ) remains that o f P. Friedlander, Johannes von Gaza und Paulus Silentiarius ( 1 9 1 2 ) ,1-103; in Latin, there is an excellent survey by G . Ravenna, 'L'ekphrasis poetica di opere di arte in Latino: temi e problemi', Quad. Ist. Fil. Lat. Padova 3 ( 1 9 7 4 )1-52 ~ (see also his article 'Ekphrasis' in the Enciclopedia Virgiliana). T h e most suggestive discussion is that of A. Pemtelli, 'L'inversione speculare. Per una retorica dell'ecphrasis', MD I (1978),87-98 = La Narrazwne Commentata (1979),ch. z. I am not concerned

here with ekphrasis as an independent genre, as in the ekphrastic epigram: on this see o f course Friedlander, op. cit., withthe furtherbibliography in S. T . Stevens, Image and Insight: Ekphrastic Epigrams in the Latin Anthology (Diss. Wisconsin-Madison, 1983),and Simon Goldhill's 'Reading, seeing, meaning: the poetics of Hellenistic ek hrasis', forthcoming in the GoldhilVOsborne collection. I shall refer to some o f the modern bibliography on description below: but note the special issues of Yale French Poe'tique 43 (1980)and 51 (1982),and Studies 61 (1981), Litte'rature 28 (1980),and the collections by J . Bessikre, L'ordre du descnptif(1988),P. Bonnefis, La Description: Nodier, Sue, Flaubert, Hugo, Verne, Zola, Alexis, Finion (2nd edn., 1980), and Y . Wert-Daoust, Descnption e'criture -peintre (1987).Apart from Genette, the most important theoretician has been P. Hamon: see 'Qu'est-ce qu'une description?'Poe'tique3 1 1 9 7 2 )465-85 ~ (trans. R. Canter in T . Todorov (ed.), French Laterary Theory Today: A Reader ( 1 9 8 2 ) ) and , especially Introduction a I'analyse du descriptif (1981 : an excerpt is translated as 'Rhetorical status of the descriptive' in Yale French Studies 61 ( 1 9 8 1 ) ,1-26). Note also J . Hagstrum, The J . Pelc, ' O n the concept o f narration', SisterArts (1953); Semiotics 3 ( 1 9 7 1 ) 1-19; , M. Barchiesi, I1 tempo e il testo, studi su Dante e Flaubert (1987) ; and J . van Appeldoorn, Pratiques de la description ( I 982). There is a large mass of comparative material, which often touches on issues of theory: see e.g. E. L . Bergmann,Att Inscribed: Essays on Ekphrasis in Spanish Golden Age Poetry (1979); A. Corbineau-Hoffmann,Beschreibung als Verfahren. Dze Asthetik des Objekts im Werk Marcel Prousts (1980); P. Dubois, History, Rheton'cal Descnption and the Epic: from Homer to Spenser (1982); R. L. Flaxman, Victorian Word-Painting and Narrative: Toward the Blending of Genres (1987); P. Imbert, Se'miotique et description balzacienne (1978); J . Kurman, 'Ecphrases in epic poetry', Comparative Literature 26 ( 1 9 7 4 ) ~1-13; L. Perrone-MoisCs, 'Balzac et les fleurs de l'Ccritoirel, Poe'tique 1 1 (1980), 305-23; W . H. Race, Classical Genres and English Poetry (1988),56-85; J . Ricardou, Problemes du nouveau roman (1967);L . Spitzer, ' T h e "Ode on a Grecian Urn", or content vs. metagrammar', 203-55; M. Van Comparative Literature 7 ( 1 9 5 5 ) ~ Buuren, 'L'essence des choses', Poe'tique 1 1 (1980),326-46 (on Claude Simon); B. Vannier, L'inscription du corps, pour une s~miotique du portrait balzacien (1972); M. Zink, 'Les toiles Agamanor et les fresques de Lancelot', Litte'rature 38 (1980), 43-61. T h e Tenth International Colloquium on Poetics held at Columbia University in 1986 was devoted to ' T h e Poetics of Ekphrasis' (see Rosand, above n. I ) but the proceedings have not been published.

26

D . P . FOWLER

nothing corresponds at the level of story.3 T h e plot does not advance, but something is described. There is an obvious sense in which description is more basic - one could theoretically imagine a narrative with only names in it, and no referring expressions, but it is practically impossible for any narrative of length not to contain description. In a deeper sense, however, as Genette noted in his article on the 'Boundaries of N a r r a t i ~ e ' , ~ description in general is secondary, is 'ancilla narrationis, the ever-necessary, ever-submissive, neveremancipated slave'. Set-piece description is not even in any real sense necessa y. Hence the controversial nature of description and the strong antipathy to it which critics from Lessing to Lukacs have often shown.5 This is connected with the issue of human interest, put in its crudest form by Lukacs in 'Narrate or Describe' with its epigraph from Marx, 'To be radical is to grasp things by the roots. T h e root of humanity, however, is man himself', and its declaration that 'objects come to life poetically only to the extent that they are related to men's life, that is why the real epic poet does not describe objects but exposes their function in the mesh of human destinies, introducing things only as they play a part in the destinies, actions and passions of men'.6 Narrative is about people, description deals with things. I t is not difficult to challenge this Aristotelian opposition of 'narrative' and 'description', and the exaltation of the telling of stories about human beings over description of things. I t may be that what we are interested in in narratives is neither plot nor pictures but ideology, the values inscribed in the work through theme and imagery as much as by story and description. I n essence I think this is correct, and this essay could perhaps stop here. But I want to go on talking of theproblem of description because this primacy of plot and almost moral distaste for description7 has been very deeply engrained in the Western tradition. Historically, description has tended to make people nervous. As Riffaterre makes clear,8 this lies at the heart of the traditionally problematic status of didactic (especially scientific didactic) poetry. T o allow a place for a poem about the world like the D e rerum natura, we must explain why we are to be interested in the blind motions of atoms in the void, to answer Aristotle's exclusion of . ~ of problems' (in Wittgenstein's phrase) can Empedocles as not imitating human a ~ t i o n'Loss be dangerous, because it can blind us to how lingering are the traces of the beliefs that we dismiss. So I ask the reader for the moment to join with me in worrying about description and its relation to narrative. If we accept for the moment the traditional opposition of narration and description, we can isolate three approaches which have tried to deal with the problem of their relation. T h e first is to stress the role of set-piece description in 'bringing the scene before our eyes' as traditional accounts of enargeia put it,'' or to say with Barthes'' that what we have is 'the effect of the real', that what details in a description signify is reality itself: 'Flaubert's barometer, Michelet's little door finally say nothing but this: we are the real'.12 Description is admitted to be narratively (or indeed thematically) redundant, but this redundancy increases our sense of the reality of the scene before us. I t is just as if we were there ourselves. Now, however, one defines these 'reality' functions of description, they are undeniable. But they do not get us very far. It was an early lesson of old-fashioned structuralism that cultural productions participate in systems of meaning independent of the conscious intentions of their creators or users. I may buy a Nissan car because of its reliability ( I did) but the significations
See e.g. M. Bal, Narratology, Introduction to the Theory ofNarrative (1985), 7 0 , 7 6 7 . G . Genette, 'Frontiers of narrative', in Figures of Literary Discourse trans. A. Sheridan (1982), 127-44, at 134 (from Figures11 (1969): see also New Literary History 8 (1976), 1-13). See especially H . C. Buch, Ut Pictura Poesis. Die Beschreibungsliteratur und ihre Kritiker von Lessing bis Lukacs (1972). I shall return to Lessing, who has been much discussed in recent years: see the introduction to Laocwn trans. A. E. McCormick (1984) with D. E. Wellberg, kssing's Laocoon, Semiotics and Aesthetics in the Age ofReason (1984); G. Gebauer (ed.),DasLaokwn Projekt (1984: collection ofessays, some translated); and especially T. Todorov, 'Esthetique et simiotique au XVIIIe sikcle', Cn'tique 308 (1973)~ 26-39 = Thioriesdu S.vmbole (1977), 161-78 (German translation in Gebauer). G. Lukacs, 'Narrate or describe?' in Writer and Critic andotherEssays, trans. A. Kahn (1978), 11-48, at 137. Not always 'almost': as Fanny Lemoine reminds me, some of the antipathy to description may be more explicitly motivated by a contempt for the things of this world, whether from a Platonic or a Christian standpoint, just as the growth of non-allegorical descriptive poetry in modern times is bound up with Romantic pantheism. M. Riffaterre, 'Systkme d'un genre descriptif ', Poe'tique 3 (197z), 15-30. cf. Tasso, Discourses on the Heroic Poem, trans. M . Cavalchini and I. Samuel (1973). 7-8 lo cf. H. Lausberg, Handbuch d:r literarischen Rhetorik (1960), 399-407; G. Zanker, Enargeia in the ancient criticism of poetry', RhM I 24 (1981), 297-3 I I . l1 R. Barthes, 'The reality effect' in The Rustle 3 f Language, trans. R. Howard (1986), 141-8, or trans. R. Carter in T. Todorov (ed.), French Literary Theory Today (1982), 11-17. l2 idem, 148.

27 'boring' and 'unstylish' which it bears when contrasted with a Porsche are independent of what I thought I was doing. Of course an interpreter can take this public signification and mess it around, but she cannot ignore it. T h e ekphrasis of the villain's car in a narrative may certainly bring the scene before the reader's eyes, but we can still ask what this car means against the matrix of alternative possibilities. I n fact, it is striking how often set-piece description in narrative is of things that participate particularly obviously in social systems of meaning: clothing or armour, furniture, architecture, the cultivated landscape. These social systems do not determine or limit the meaning of an ekphrasis, but they already take the reader beyond the reality effect, however specified. There is therefore no real 'solution' to our 'problem' here. T h e other two approaches are more important. First, we can attempt to deal with the problem of description by integrating it with the narrative. A relatively crude form of this can be found in Lessing's defence of the Homeric shield description on the grounds that not the shield but its manufacture is described:13 only slightly less crude is Lukacs' defence of the horse-race in Tolstoy's Anna Karenina against that in Zola's N a n a on the grounds that it is 'no mere tableau but a series of intensely dramatic scenes which provide a turning point for the plot'.14 Clearly narration often continues through a description (there is rarely a complete pause) and the description may occasion reactions in participants important for the plot. But the needs of the plot can usually be satisfied by a much more exiguous account than we are offered in ekphrasis, and it is difficult to see that Lessing's device of describing the manufacture of an object rather than the object itself is other than a trick. We could easily turn any description of an object into an account of its making, but would this really get to the heart of the problem? Nevertheless, it is an important point that description is rarely 'pure', because the way that narrative impurity is introduced is often through the figure of an observer. I t is my third tactic however which is most commonly found in critical writing and which raises the most important questions, that is the relation of description to narrative on a psychological level. This can be done in several different ways. T h e objects described may be causally linked with a character or an action: people choose their own furniture or wallpaper. We may have an explicit or implicit observer, through whose eyes the description is instantiated. We may have an instance of pathetic fallacy, however we wish to define that: the storm outside reflects the storm inside. Or we may have much looser metaphoric or especially metonymic links with the plot, particularly ones of prefiguration:15 the flower plucked in chapter one becomes the maidenhood lost in chapter four. Much modern critical reading of ekphrasis in classical literature takes the form of an attempt to show that what earlier critics had seen as 'merely' decorative description can in fact be integrated with narrative, indeed demands to be so integrated. Precisely because ekphrasis represents a pause at the level of narration and cannot be read functionally, the reader is possessed by a strong need to interpret. As will be seen, I believe this is the right tactic. But it is obviously vulnerable to attack, and has come under attack from post-modern theorists, for being organicist and totalizing. If we take the dominant trends in current criticism with Murray Krieger to be those which 'celebrate margins rather than centers, the aporia rather than the filled gap, the arbitrary or even the random rather than the necessary',16 then it is clear that from those points of view the way in which the classical criticism that I have been describing integrates ekphrasis may be seen as a minus not a plus. T o relate description in this way to narration is to accept its poor relation status but to give it a limited form of social mobility: the more radical move is to free description from the chains of slavery and to give it true autonomy. T h e vanguard of this approach was the nouveau roman, particularly in the theorizing of Robbe-Grillet, with its cry that 'instead of this universe of "significations" (psychological, social, functional), one must try to construct a world more solid, more immediate'" in which objects are given a role outside of any metaphorical or metonymic system of reference. Now whether the nouveau roman in fact achieves that independence of the object, and whether even if it did it would be relevant to
l 3 Lessing, h c w n , ch. 16, e.g. trans. McCormick, op. cit. (n. s ) , 84, 'we see in the poet's work the origin and formation of that which in the picture we can only behold as completed and formed'. l4 op. cit. (n. 5 ) , 111. l5 cf. R. G. M. Nisbet, 'The Oak and the Axe: symbolism in Hercules Oetaeus 1618ff.', in M. Whitby,

NARRATE AND DESCRIBE: THE PROBLEM O F EKPHRASIS

P. Hardie, and M. Whitby (eds), Homo Viator (1987), z4?-51. M. Krieger, A Reopening of Closure (1989), 3. " A. Halsall, "'La Transition", descript~ons et ambiguitis narrativo-discursives dans "Victoire" de William Faulkner', in Bessikre, op. cit. (n. 2), 27.

28

D . P . FOWLER

classical criticism are questions I want to avoid, but this anti-organicist, anti-closural movement in late-modern and post-modern criticism should make us be at least a little selfconscious about the stress on closure and integration which dominates classical criticism. A generation whose motto is 'the revenge of the ~rYstal'l8 is hardly likely to warm to a criticism which celebrates an integrating focus on human subjectivity. Moreover. the modernist classical critic is likelv to find herself stabbed in the back bv reception theoiists like Malcolm Heath, who would deny that ancient readers would have feft this need to interpret which is the standard starting-point for accounts of ekphrasis.19 I do not myself believe that Heath is right. Basic to his approach is the Hirschian opposition between 'meaning' (identified with conscious intention) and wider 'significance' as read in by modern interpreters, and the exaltation of the former characterized as historical over the latter characterized as 'the forceful imposition of alien preconceptions on ancient literature'." I t is an easy challenge to celebrate rather the latter process, and to dramatize one's critical practice as Socratic guerilla warfare. But it is more important to deconstruct the o p p o ~ i t i o nOn the one .~~ hand, the belief that one can ever free oneself from contemporary concerns is a delusion that critics have alwavs to resist if thev are to avoid self-dece~tion: it is not difficult to find elements in the present pisition of classic2 studies out of which plausible account of the factors which encourage a cool historicism can be constructed. On the other, Heath himself admits that a basic problem with hi's approach is that a culture's 'primary poetic' instantiated in practice may only very imperfectly be captured by the 'secondary poetic' which is represented in the conscious theorizing of rhetoricians and commentator^.^^ If one takes a wider view of ancient semiotics, it is not difficult to find evidence of a strong hermeneutic imperative at all periods in phenomena like divination; and indeed even in literary studies Heath underplays elements such as the allegorical tradition, which begins in the fifth century B . c . , not the fifth century oreover, over, it is most important not to accept the characterization of 'reading against the grain'24as necessarily unhistorical; to accept the conscious formulation of its own values by a culture (or some members of it) as authoritative looks more like a denial of history. One must both resist the simplification of ancient attitudes and accept that the critic may at times stress elements that members of a culture neglected precisely in the name of history. But Heath's attack is an important reminder that there is nothing necessarily natural or inevitable about modernist integrationism: it is an aesthetic that has to be defended. We have another reason not simply to take as an obvious given that the first thing one does with the description of an apple is to find a young virgin to whom it might correspond. I want to return to these themes at the end, but I hope it is already clear how the way in which we approach ekphrasis is paradigmatic of our attitudes to much wider issues of interpretation. Let me, however, turn to a question that is constantly raised with respect to ekphrasis, that of 'point of view'.25 I mentioned above that one way in which description is often related to narrative psychologically is through the figure, explicit or implicit, of the observer. This is put most strongly by J. K i t t a ~ : ' ~

When we read a representation, we also read, or read in, the account of the perception of that representation . . . There is always a choice of percipients. We can read an act as perceived by the character who carries it out (our hero, for example), by the character who is or will be its object, or by any character who might react to it, appreciate it, or be confused by it (e.g. an onlooker, a confidant, a chorus) . . . There are no autonomous limits on this power of inference and construction, this reading-in of subjectivity, as there must be at least the possibility of access (imaginative as well as provided) to a subject-based reading of the represented. Empathy is available, of one character for another, of the narrator for any character, and of the reader for anyone.
l8 J. Baudrillard, k s stratigies fatales (1983); cf. D . Kellner, Jean Baudrillard, Fmm Marxism to Postmodernism and Beyond (1989), I 54-62. l9 M. Heath, Unity in Greek Poetics (1989): see also The Poetics of Greek Tragedy (1987), 98-1 I I . O ' idem, 1 5 5 . cf. D. Kennedy, rev. S. J. Harrison, Oxford Studies in Veqil's Aeneid, Hennathena (forthcoming). 22 op. cit. (n. 19), 10. 23 cf. J . Whitman, Allegory, the Dynamics of an Ancient

and Medieval Technique (1987), with further bibiography: on Theagenes of Rhegium, conventionally made thepmtos heuretes of allegory, see R. Pfeiffer, History of Classical Scholarship I ( I968), g-II . 24 J . Winkler, The Constraints ofDesire ( ~ g g o )126. , 25 For my use here of Genette's 'focalization' (and some of the problems with the concept), see 'Deviant focaliza, tion in Vergil's Aeneid', PCPhS 216 ( ~ g g o )$2-63. 26 'Descriptive limits', Yale French Studtes 61 (1981), 225-43, at 234.

29 and obvious force by T h e question of focalization, of 'who sees?', is raised with description. Basic to Lessing's famous distinction of literary and plastic art were two notions: first, that 'succession of time is the province of the poet just as space is that of the ~ a i n t e r ' , ~ ' and second, that linguistic signs are arbitrary, visual ones iconic. A picture of a cat looks like a cat, the word 'cat' does not. Again these are distinctions which invite and have received a great deal of deconstruction, but which I want for the moment to accept. Both of these aspects of the literary description reinforce our feeling that a literary description necessarily inscribes a point of view more strongly than a plastic one. On the one hand, there is the phenomenon of what the linguist W. J. M. Levelt has referred to as the speaker's linearization problem:28 when we describe in words a scene, we have to decide the order in which we are to present the details and the duration- which may be zero- of the description of each of them. Narratologically, that is, the visual scene described functions as story to the narration of the verbal description. There is no neutral, zero-focalized way of linearizing a visual scene: a point of view is necessarily inscribed, though there may be accepted ways in a particular culture of ordering the elements -asked to describe their house, for instance, most people will give a mental tour starting at the front door and climaxing according to predilection in the kitchen, the bedroom or the study. T h e focalization does not, of course, have to be that of the actual observer: as Kittay remarks, empathy is available, and I can describe my house to a friend in a way that represents neither the way I see it nor the way she does, but the peculiar interests of my aunt. But the speaker's solution of the linearization problem necessan'ly imposes a point of view. Similarly, the non-iconic nature of the linguistic sign means that there is a much wider matrix of choice against which a particular element is seen. I n his work 1912 I , Sciascia describes a photograph of an Arab being shot amongst the dunes during the Italian imperialist war in Libya, and implicitly highlights the contrast between the emotional reaction of a modern to the scene with the presumed contemporary reading:29 In Cirenaica la guerriglia pungeva. Tribunali di guerra assiduamente sedevano per giudicare i ribelli: e cio& per passarli ai plotoni d'esecuzione. Agli italiani ne arriva qualche immagine: schizzi, fotografie. Con qua1sentimento& stato allora guardata questa fotografia che ho sotto gli occhi, della fucilazione di un arabo tra le dune? I1 plotone schierato su due file, l'ufficiale che sta per dare il segnale del fuoco, il condannato che sembra lontanissimo dal plotone, come sperduto tra l'ondulazione della duna. Agosto 1913. 1912 + I . In Cyrenaica the guerrilla war was causing irritation. Courts martin1 were in constant session to pass judgement on the rebels- that is, to hand them over to the firing squads. A few images of this reached the Italian public: sketches, photographs. What was their feeling then when they looked at this photograph that I have before me, depicting the shooting of an arab in the midst of the dunes? The platoon lined up in two lines, the officer waiting to give the order to fire, the condemned man looking a long way away from the firing squad, almost lost amongst the undulating dunes. August

NARRATE A N D

DESCRIBE:

THE PROBLEM OF EKPHRASIS

Sciascia attempts to describe the scene neutrally, like a camera with the shutter open. But his choice of 'fucilazione', like my choice of 'being shot' must be contrasted with alternatives like 'being executed', 'being murdered', 'being martyred'. T h e same photograph can be read as a sign of triumph or an indictment of crime, but verbal description has to take a stand, however 'objective' it attempts to be. Again, there is an obvious sense in which description in language inscribes a point of view more forcefully and more unambiguously than plastic art. Now I stress again that these oppositions are in actual fact far more complex: deconstruction is here as easy and as necessary as with the other oppositions that I have already discussed. There are of course various ways in which narrative art can exist, through conventions of placement, the use of panels and frames, the representation of more than one moment of time in one picture.30 I n an architectural setting viewpoint may be 'controlled', as Robin Osborne has argued with regard to the Parthenon frieze.31T h e converse of the speaker's linearization problem is the artist's non-linearization problem, how to represent time through
Trans. McCormick, op. cit. (n. s ) , 9 1 . W. J. M. Levelt, 'The speaker's linearization problem', in H. C. Longuet-Higgins, J. Lyons, and D. E. Broadbent (eds), The Psychological Mechanisms of Language (1981),305-15. 29 L . Sciascia, 1912 I (1986), 17. cf. R. Brilliant, Visual Narratives: Stoly-telling in
27

Etruscan and Roman Art (1984), with bibliography; on some theoretical problems, see N . Goodman, 'Twisted tales; or, story, study, and symphony', in W. J. T. Mitchell (ed.), On Narrative (1981),9 ~ I S1 . " 'The viewing and obscuring of the Parthenon frieze', JHS 107 ( 1 9 8 7 ) 98-105. ~

D . P . FOWLER 30 simultaneity, and there are various ways of solving this. In the literary description of a work of art. we mav find traces of both linearization and non-linearization: the cleverest e x a m ~ l e I know is thit of the ekphrasis in Moschus, where the story of Argos and 10 which was i o n linearized by the artist in spatial terms is then relinearized by Moschus in a way which allows both Drocesses to be seen.32 Then too. Lessing " has often been criticized for taking " as his paradigm sculpture in the round: there is an even more obvious sense of point of view inscribed in a panel painting. My example from Sciascia of a photograph might be said to be cheating: figurative painting is not simply the accurate representation of reality. Nor even is photography. Art however realist is no more 'naturally' iconic than literature: in both the signs are read according to systems of meaning that are cultural constructs. I n the end, visual art is not si~nificantlv different from literature. Nevertheless, Lessing's oppositions can be used to stress a paradox about visual and literary description which seems to me of great importance for contemporary classical cultural criticism. T h e tendency of the traditional view of art that I have been outlining is that visual art must be more oDen. less tied to a ~ o i nof t view. less fixed in its inter~retation than literarv art in which a point o'f view is constanGy imposed by the medium. Yet Ge are used to literaricritics being addicted to ambiguity and polysemy, while art critics are often rigidly historicist. This is ~articularlv true of the criticism of ancient art. where the revolt against the 'connoisseur' iradition irktituted by Beazley in the footsteps of Giovanni ~ o r e l l i ~ ~ vused h a s as its principal weapon a strictly functionalist methodology. There are of course exceptions to this: indeed some of the critics of ancient art whose methodology has been most rigorous (and successful) in its functionalism have also taken pains to stress the polysemy of the artistic scenes they have discussed.34But the confident historicism of much ancient art criticism is in striking contrast to the (by now cliched) rhetoric of 'crisis' that has gripped art history more generally.35I n the study of Roman art especially, the stress in recent criticism has been on remorselessly showing how artistic production serves the dominant ideology. A good example is E. Lefkvre's recent discussion of the Portico of the Danaids in Augustus' temple-complex on the Palatine.36T h e intervretation of this monument has alwavs been ~roblematic: Lefkvre argues that the ~ a n a ' i dwho s murdered the sons of Aegyptusare to be'seen as a symbol of the ~ z m a n in s their recent triumph over Cleopatra, theAegyptia coniunx. Often of course the Danaids are seen as sinners. but for Lefkvre it is axiomatic that such an intervretation is im~ossible.~' But how could aktist or patron stop ancient readers taking a differAt view of the 'Danaids, especially when the alternative view is so strongly represented in Augustan l i t e r a t ~ r e ?Lefkvre's ~' view of the monument leads him to trv to read the famous scene on the balteus of Pallas in the Aeneid in similar terms, with Pallas i s the worthy Danaids and Turnus the dastardly E g y p t i a n ~ . ~ ~ Once the argument moves into literature, its deficiency is clear: one could not say Lefkvre's view of the balteus was impossible, but it is easy to show that very different views are (at least) equally plausible and that no amount of evidence for an Augustan reading could remove the
0

Moschus, Eumpa, 37-62 Three scenes are described: 10 crossing the sea (44-9), Zeus turning her back into a woman (5-4), and the phoenix arising from the blood of the dead Argos ( 5 5 4 1 ) That is, the order of the scenes chronologically is ACB (or conceivably ABB if the second and third scenes were taken to be contemporaneous). In the last scene, first Hermes is described; then 'nearby' Argos with sleepless eyes; then the blrd arising (described in the imperfect, exanetellen) ; then its outspread wings. The temporal sequence in the last scene represents the spatial arrangement on the cup, but in such a way that the temporal sequence in the story that the visual representation supposedly had to delinearize shows through. Reading narrative art is relinearization. 33 On the origins qf Morelli's methods, see Carlo Ginzburg's brilliant Clues: roots of an evidential paradigm', in Myths, Emblems, Clues, trans. J. and A. C. Tedeschi (1990). 34 See especially the work of Christiane SourvinouInwood, e.e. 'Menace and ~ u r s u i t :differentiation and the creatiolof meaning', in'^. Birard, C. Bron and A. Pomari (eds), Images et socie'ti en Grece ancienne. L'iconographie comrne me'thode d'analvse, Actes du Colloque International, Lausanne 8-11 ftvrier 1984

'*

(1987), 41-58, at 42: 'no sign has a fixed meaning. . . signs are polysemic .. . not all the meanings produced by the signifying elements in a signifier contribute to the production of one unified coherent meaning. Some can produce different perspectives, warring discourses, which deconstruct the dominant one.' 35 cf. D. Preziosi, RethinkingArt History (1989). E. Lefilvre, Das Bild-Pmgramm des Apallo-Tempels auf dem Palatin, Xenia Heft 24 (1989). '' ibid., 15: 'Es ware zudem absolut widersinniggewesen, wenn Oktavian im Siegesjubel uber die Agypter an dem Siegesmonument par excellence eine Darstellung zugelassen hatte in den Agypter als Ehrentrager erschienen.' " 'f. P. Hardie reviewing Lefevre in CR NS 40 g . ( ~. .g o ) , 520. 39 op. cit. (n. 36), 16, 'Turnus ist ein Angreifer, dessen Tat negativ zu bewerten ist: fur sie hat er zu bussen. Er kann also nur mit den angreifenden Aegyptus-Sohnen vergleichen werden, deren schandliches Handeln (nefas) ein schmahliches Ende (foede) gefunden hat. Und Pallas ist mit den Danaiden verzlichen. insofern er freventlich angegriffen wird. Auf ihnvbezogdn muss die Aussage der balteus lauten: Pallas wird geraht werden. Man beachte auch die Anspielung: Pallas kommt vom Palatin!'

3' alternative traces left by the complexity of the tradition. My argument is not that art critics are wrong to argue like this, or that we should simply switch to seeing art as more open: I think there is a genuine paradox in our attitude to art as both more open and more closed than literature that is not to be resolved by coming down unambiguously on either side. But there is a strong case for a greater awareness of this paradox - which, as we shall see, I believe was recognized in antiquity. Our paradoxical intuition (if by now we have acquired one) that art is both more closed and more open than literature makes the phenomenon of ekphrasis where they meet of peculiar interest. Whatever position we adopt as to the degree to which plastic and literary art inscribe points of view, in literary ekphrasis the presence of the intermediary - usually fictional visual artist introduces another potential focalizer. This is particularly the case where there is an underlying narrative element in the visual representation being described. Even leaving out the more extreme possibilities that Kittay mentioned, we then have a complex hierarchy of potential points of view, which can be summarized in the following diagram:

NARRATE A N D DESCRIBE: THE PROBLEM OF EKPHRASIS

ARTIST

(A)

Visual Representation AUDIENCE

I (B)

NARRATOR/VIEWER/
TOPIC (c) AUTHOR

(D)

Description
AUDIENCE

AUDIENCE 2

(D)

3 (E)

That is, of any element in a description we can ask whether the focalization is that of the artist who made the original work of art, or his audience, or the observer, or his audience, or the author, or his audience: and we have still not brought in the observer's brother-in-law whom chapter four will reveal to be the hero of the novel.

Let me try to make this clearer with my example, the famous ekphrasis in Book One of the Aeneid where Aeneas looks at the depiction of events from the Trojan War in the Temple of Juno in Carthage."O Aeneas is here an explicit observer, whose reactions to what he sees are also explicitly stated: he weeps and groans because of what he saw, namque videbat . . . Moreover, within episodes, as Eleanor Leach observes,41 'the order of presentation creates confusion between the visual image and Aeneas' thoughts'. This is clearest in the Troilus panel, where Troilus is depicted both fugiens (474) and being dragged by his chariot (476) :
parte alia fugiens amissis Troilus armis, infelix puer atque impar congressus Achilli, fertur equis curruque haeret resupinus inani, lora tenens tamen; huic cervix comaeque trahuntur per terram, et versa pulvis inscribitur hasta. In another part of the picture poor Troilus, a mere boy and no match for Achilles, had lost his armour and was in full flight. His horses had run away with the chariot and he was being dragged along helpless on his back behind it, still holding on to the reins. His neck and hair were trailing along the ground and the end of his spear was scoring the dust behind him.42
Aeneid 1.441-93. The bibliography is predictably large: see especially R. D . Williams, 'The pictures on CQ NS 10 (1960), Dido's temple (Aeneid 1.45-3)', 145-53; K. Stanley, 'Irony and foreshadowing in Aeneid I, 462',AJP 86 (1965), 276-77; A. Szantyr, 'Bemerkungen zum Aufbau der virgilianischen Ekphrasis', MH 27 (1970), 28-40; N. Horsfall, 'Dido in the light of history', PVS 13 (1973-4), 1-13 = S. J. Harrison (ed.), Oxford Readings in Vergil's Aeneid ( ~ g g o ) , 127-40; W. R. !ohnson, Darkness Visible (1976), 99-1 14; C. P. Segal, Art and the hero: narrative point of view in Aeneid 1', h e t h u s a 14 (1981), 67-84; R. F. Thomas, 'Virgil's

"

ekphrastic centrepieces', HSCP 87 (1983), 175-84; D. Clay, 'The Archaeology of the temple to Juno in Carthage (Aen. I, 44&93)', CP 83 (1988), 195-205; Leach, op. cit. (n. I ) ; J. J . O'Hara, Death and the Optimistic Prophecy in V e ~ i l 'Aeneid s ( ~ g g o )35-9. , 4' op. cit. (n. 32j, 314; 42 Trans. D. West (Penguin, 1990). I use West's new version here and below because his attemat to make the description unambiguous for the reader inderlines how Vergil leans the other way. For some criticism of this as a method of translation, see my forthcoming 'Brief Notice' in G N R 1991 (2).

32

D. P . FOWLER

Similarly the scene between Priam and Achilles from Iliad x x ~ v is , introduced with a reminiscence of what had happened" before the scene depicted (11.483-7) : ter circum Iliacos raptaverat Hectora muros exanimumque auro corpus vendebat Achilles. tum vero ingentem gemitum dat pectore ab imo, ut spolia, ut currus, utque ipsum corpus amici tendentemque manus Priamum conspexit inermis. There too was Achilles. He had dragged Hector three times round the walls of Troy, and now was selling his dead body for gold. Aeneas groaned from the depths of his heart to see the armour stripped off him, the chariot, the corpse of his dear friend and Priam stretching out his feeble hands. the ,~~ order and duration of the Although we are told that Aeneas sees Iliacas ex o ~ d i n e p u g n a s scenes has often been seen to represent a peculiar choice, which throws emphasis on the tragic elements like the death of Troilus, the first of so many dying youths in the Aeneid. T h e strongest version of this view indeed sees Aeneas as misreading the scenes. Nicholas Horsfall and others45have suggested that Aeneas got it all wrong. On this view, Just as one would expect in a temple of Juno, the choice of pictures illustrates the success of her favourites: while Aeneas is delighted to see that Troy is not forgotten, he quite fails to observe, as we must do, that the attitude to Troy shown in these pictures is neither friendly nor sympathetic. They illustrate just those qualities which Carthaginians might admire in the victorious Greeksgreed and brutality, for which they themselves had such a fine r e p ~ t a t i o n . ~ ~ On this interpretation, we can see peeping out, especially in the final scene of Penthesilea, which clearly anticipates Dido's entry, hints of a different way of reading these scenes, a way far removed from sunt lacrimae rerum. There are good grounds for rejecting the crude form of this thesis as it is put by Horsfall, who is intent on showing that Aeneas did the right thing in Ieaving Dido. In her opening words to Aeneas she tells of how Teucer put in at Sidon and told her of the 'casus . . . urbis ITroianae', and how he 'ipse hostis Teucros insigni laude ferebat 1 seque ortum antiqua Teucrorum a stirpe volebat' (623, 625-6). She associates herself with the tragic interpretation of Trojan history, and to view her as lying in so doing would be an extreme subversion to which the text gives no encouragement. Moreover, there is clearly a battle of paradigms in the depiction of Carthage; is it Phaeacia or the land of the Cyclops? T h e presence of art is in itself part of the evidence pushing us towards Scherie; and Diskin Clay points out that when Odysseus arrives in the Cyclops' cave, stress is laid on the fact that Polyphemus has not heard of him, whereas the fame of Troy has reached P h a e a ~ i aAnd . ~ ~ finally, as Clay also stresses, one model for the scene in the temple and Aeneas' reaction is the story told variously of Aristippus and Plato of the shipwrecked philosopher coming upon geometric figures in the sand: there too what is discovered is indisputable evidence of ci~ilization.~' But while the view that has Aeneas deceiving himself totally cannot be right, the question of the focalization of the ekphrasis remains of importance. Take, for instance, the detail of the description of Achilles 'selling' Hector's body, 'auro . . . vendebat' (484). K. Stanley49 pointed out that the reader here naturally thinks not of the scene in the Iliad but of the common scene in art and elsewhere in literature of Priam weighing out gold, and Stanley, like Horsfall, saw the presence of such a brutal scene in the temple as evidence that Aeneas' reading is fatally optimistic, though he drew very different implications, that 'in Vergil's literary and historical perspective, Achilles and Aeneas, Greek and Trojan, Roman and Tyrian are bound to that realm where the roles of
43 On the use of the pluperfect, see Szantyr, op. cit. (n. 32) and especially Ravenna, op. cit. (n. I ) , 34-46, quoting Servius on 1.484: 'ingenti arte utitur verbis: nam hoc loco, quia pingi potuit, praesens tempus posuit, superius, quia pingi non potuit, sed referri, p;rfecto exsecutus est tempore dicendo "raptaverat non "ra tabat".' 49 On ex ordine and similar expressions, see Ravenna, op. cit. (n. I), 1 6 1 7 .
45 Horsfall, Stanle Johnson, Clay, Leach, O'Hara, o p g citt. (n. 32) (witrvery different emphases!).
Horsfall, op. cit. (n. 32), 138.
Clay, op. cit. (n. 32), 197.
idem, 195-6. The story is told of Aristippus in
Vitmvius v1.1, and of Plato in Cicero, Rep. 1.29: see
G. Giannantoni, I Cirenaici (1958), 213 (Aristippus frr.
42;3). op. cit. (n. 32), 2 7 6 7 .

33 the slayer and the slayed are inevitably united by the reversals of time - where, indeed, "sorrow is implicit in the affairs of men"'. T h e depiction of Achilles as cruel and mercenary is of course anyway ambiguous: that could be a Trojan view as much as a Greek one. But can we be certain that the scene before Aeneas was the 'brutal' version? vendebat could represent Aeneas' interpretation of the scene; one might say that auro pushes the reader towards seeing the scene as genuinely one of ransom because it looks like one of those details in ekphrasis which are simultaneously about the painterly surface and an object in the story being depi~ted,~ ' the existence of the golden ransom is scarcely underplayed in the Iliad version but of the meeting.'l In this ekphrasis where so much is clearly 'read-in' rather than 'seen' insofar as such a distinction holds - it is not impossible that this too is Aeneas' view rather than what is actually 'there'. As often with questions of focalization, there is more than one story that we can tell here of whose points of view the pictures and their descriptions represent.j2 Like other scenes of ekphrasis, the scene is often -and surely rightly in some degree -taken as paradigmatic for the interpretation of art, both literary and visual.j3 Eleanor Leach,j4 for instance, comments that the interaction between Aeneas as reader and the work of art 'will not appear foreign to the contemporary reader who understands that meaning is not the inherent property of a text but is instead created in variant forms through variant experiences of reconstructing the work as text', and that Aeneas' 'deeply sentimental misreading of the frieze shows the process of perception as one of selection, amplification, and reordering, and thus it casts doubt upon the reliability of factual communication through pictorial narrative'. But one might take the scene as more normative and less aporetic, as enjoining upon the reader like Aeneas to read tragically rather than triumphantly, whatever the picture that is offered. This has obvious relevance to the interpretation of Augustan art. As I mentioned above, contemporary criticism of Augustan art is dominated by functionalism and historicism: it attempts to show how subjects that are apparently aesthetically neutral actually serve Augustan ideology. If we take the scene in Aeneid I as paradigmatic, however, it suggests that more allowance should be made even within a historicist framework for more than one way of reading the symbols: that after all an observer might be able to deconstruct Roman art as well as Roman literature. T h e suggestion would not perhaps be unparalleled. I n a well-known article on Philostratus and Homer whose importance Bartsch has recently stressed, Lesky had suggested that some of the passages in the Imagines were clear 'misreadings' of the underlying picture, not through misunderstanding but as a tour-de-force of 'sophistischer Deutungskunst'.j5 Misreading and cross-reading are not necessarily modern critical inventions: and the ekphrasis in Aeneid I with its stress on the complexity of interpretation cannot be entirely isolated from its time. Too New Historicist a reading of Augustan art begins to look unhistorical.

NARRATE A N D DESCRIBE: THE PROBLEM O F EKPHRASIS

I n conclusion, I want to return to the issues with which I began in the light of the complexities of focalization which have emerged in the example from the Aeneid. I said that my sympathies were still very much with the organicist New Critical approach which would seek links between ekphrasis and the narrative of which it is part, but I also implied that the challenge of post-modern dislike of this as totalizing and authoritarian needed to be taken on board. T h e political metaphors are of course basic to the assault of post-modern theorists like
cf. Ravenna, op. cit. (n. I ) , 14-16, R. DebrayGenette 'La Pierre descriptive', Poe'tique I I (1980), 293333 on Heliodoms v. 14, which self-consciously plays with the convention. 51 cf. Iliad x x ~ v 76, . I 19, 137, 1 4 6 7 , 175-6, 195-6, 228-37, 367, $1-2,435-6, 502, 555, 579, 594, 685-6. I owe this point to Alessandro Barchiesi, who comments: 'It is too easy to forget that ransom, and gold, plays a role in Homer's narrative too. If Aeneas was a reader of the Iliad (and in a sense he is) he could still point out exactly the same points: cmelt , golden ransom, the gesture of a father. This wouldy be a selective, and tendentious, reading, but understandable from a Trojan point of view;
Priam in Aeneid 2 provides a counterbalance.'
52 ~ fPCPhS . 2 1 6 ( 1 9 9 0 ) 42-63. ~
53 This point might be strengthened by Richard Thomas' suggestion, op. cit. (n. 32), that the presence of the peplos at the centre of the ekphrasis (479-82)
constitutes a sort of mise en abyme in the light of the
tradition of ekphrasticpeploi.
op. cit. (n. 32), 323-3.
55 Bartsch, op. cit. (n. I ) ; A. Lesky, 'Bildwerk und
Deutung bei Philostrat und Homer', Hennes 75 ( 1 9 4 0 ) ~
38-53, at 45 = GesammelteSchriften (1966), I 1-25, at 17.

"

D. P . FOWLER 34 Lyotard on theories of interpretation which aim for a fixed overall truth: that way fascism lies. A carnivalesque dialogy looks radical against that. But it is not simply residual Stalinism that makes Marxist critics in particulars6feel unhappy about the assumption that stress on plurality is always radical. It can be a way of evading the difficult task of formulating a properly complex account of the relationship of the individual and society. If we move back from politics to the text, we can similarly try to formulate an account of the relationship of a description to its narrative which takes adequate account of complexity but does not simply liberate the ekphrasis to meaninglessness. A way forward is perhaps to be found in a neglected contribution by Alessandro P e r ~ t e l l iHe .~~ set up an opposition between the total subordination of description to narrative that he saw in the shield descriptions in the Seven against Thebess8with the total independence of description and narrative represented by the Aspis of Hesiod, and contrasted both with the relationship to be seen in Moschus' Europa, where there is an 'inversione speculare': 10 of the ekphrasis corresponds not to Europa but to Zeus as the bull. This relationship he termed neither narrative (where description is subordinated to narrative) nor descriptive (where it is set free) but rhetorical, conferring on the ekphrasis the status of afigure. Although his examples are limited ones, this seems a very important insight. The most interesting recent work on all types of 'digression' or narrative pause has been that which views the relationship with the main narrative as a figured one, in which elements shift and are transformed as we move from detail to whole. Perutelli's 'specular inversion', for instance, is clearly related to the way that Colin Macleod treated the mythological example in his well-known article, 'A use of myth in ancient poetry'.s9 Discussing Catullus 68, he noted that in the exemplum of Laodamia and Protesilaus both Catullus and Lesbia are compared and contrasted with both figures:

T h e myth, then, of Catullus 68 is neither a decorative and learned irrelevance, nor does it simply mirror the situation in which it is set; for the analogies between the two are qualified by no less significant contrasts. But the result of such a complexity is not mere confusion; the myth, by indicating an area of feeling beyond the direct statements of the poem, helps to express a significant conflict of attitudes. It thus makes a distinct and comprehensible contribution to the whole.

Similarly, in discussing similes, Oliver Lyne has recently tried to move beyond the alternatives of multiple-correspondence and decorative independence :60
There is thus in most similes a visible point of contact with the narrative and an illustrative function tied to it which is often advertised; in many similes further points of contact and illustrative functions can be discerned. But this sort of function is not I maintain the important or main function of a developed simile in the hands of a master. The main function of a simile is not to illustrate something already mentioned in the narrative, but to add things which are not mentioned, in a different medium: imagery. The poet is switching modes, switching from direct narrative to 'narrative' in the suggestive medium of imagery; and he capitalizes on the fact that he is now operating in a suggestive, not an explicit medium. An advertised illustrative function and concomitant point of contact with the narrative may often be seen as a means to an end, as little more than a formal device to effect the switch from direct narrative to 'narrative' in imagery.'jl

Similes and even more exempla bear, of course, different relationships to their contexts from that we might wish to posit for set-piece description. But both these formulations reflect a similar desire to Perutelli's for an account of the relation of part to whole which is significant in a non-reductive way.
56 See e.g. some of the pieces in the collection edited by M. Krieger, The Aimsof representation (1987), especially D . LaCapra, 'Criticism Today'. 57 op. cit. (n. I ) . 58 The choice is ironic in the light of Froma Zeitlinh Under the Sign of the Shield: Semiotics and Aeschylus Seven against Thebes ( I982). 59 CQ NS 24 (1974). 82-93, at 88 = Collected Essays (1983)~ 159-70, at 165. Words and the Poet (1989), 68. 61 It is interesting to observe how Lyne the 'English

empiricist' comes close to the terms in which a modern French critic, L. Perrone-MoisCs, has discussed description in Balzac, op. cit. (n. I ) . Distinguishing between 'static' description, 'a fonction redondante ualificante, explicative ou emphatique' and 'dynami>,q'?i fonction de dkplacement, de compensation, de dkfoulement', she comments that whereas the first 'renvoie circulairement a un dCja-dit du rCcit', the second produces another level of narrative: 'la description apparait ici non comme un arr&t du rkcit (pour renseigner, reposer, distraire ou convaincre le lecteur), mais comme le suit du rkcit l u n autre niveau'.

35 Perutelli's use here of the concept of 'figure' suggests an obvious comparison with Gian Biagio Conte's similar use to explain the phenomenon of allusion:62
Thus allusion works in just the same way, and in the same semantic area, as a rhetorical figure. T h e gap in figurative language that opens between 'letter' and 'sense' is also created in allusion between that which is said (as it first appears), a letter, and the thought evoked, the sense. And just as no figure exists until the reader becomes aware of the twofold nature of figurative language, so too allusion only comes into being when the reader grasps that there is a gap between the immediate meaning ('after I have sailed through many peoples and on many seas"j3)and the image that is its corollary ('as Odysseus sailed'). In the art of allusion, as in every rhetorical figure, the poetry lies in the simultaneous presence of two different realities that try to indicate a single reality . . . T h e poetry lies in the area carved out between the letter and the sense. It exists by refusing to be only one or the other. This still unknown area, this tension between meanings, can be described only by referring to the two known limits that demarcate it.

NARRATE AND

DESCRIBE:

THE PROBLEM OF EKPHRASIS

Just as with allusion, with any passage where in any sense we for a moment 'stand back' from the narrative we have the presence of two realities: the passage taken in isolation and its wider context. While as with allusion the extremes set the limits of meaning, what matters in the significance of the simile or ekphrasis or exemplum is that which in Conte's words 'exists by refusing to be only one or the other'. Any relationship we posit is inevitably an uneasy one. Precisely because the correspondences and contrasts are figured, the interpretation of them cannot be simple or clear: there is room for disagreement. We are conscious not only of a desire for integration but of a resistance to it. There is more than a whiff of the Zeitgeist here. I have excused my use of political terminology on the grounds that it is so used by post-modern critics, but this is an evasion: I believe that they are correct to see a connection between textual and political integration. And it is tempting to suggest that the troubled integration of the ekphrasis or example or simile or intertextual reference asfigure represents the same attempt at apalintonos harmonia as the critics of Lyotard and other post-modernists have suggested as an alternative to the simple celebration of individualism in the political sphere. We have moved a long way from the formalist narratology with which I began: perhaps too far. But it is important, I believe, to see that this movement is inescapable. It is a common criticism of narratology that it is merely another twentieth-century formalism, a way of talking about texts without bringing in ideology. In fact, however, if the issues raised by formal analysis are pursued, we find we cannot escape the movement towards politics. The relationship between the aesthetic and the political is not a simple one: the analogies and contrasts drawn are precisely themselvesfigured in the way that I have suggested are those between ekphrasis and narrative. And as figured, they can of course be interpreted in different ways. But the relationship is no more to be denied than ekphraseis are to be separated from their contexts -or reduced to them. Jesus College, Oxford

62

G. B. Conte, The Rhetoncoflmitation (1986), 38-9.

63 Conte is discussing the allusion to the opening of the Odyssey in Catullus 101.