I. The reader’s immersion, first fissures

What, to the uninitiated, may seem to be direct speech stemming from a narration of an encounter of a man, or a woman, with hunters before ravaged fields and a wild boar is in fact an ecphrasis, the beginning of one of the Imagines of Philostratus the Elder (1.28.1). Fully immersed in the painted scene, the speaker, a sophist explaining the pictures in a gallery to a boy and some adolescents, appears to believe that he sees the fields themselves and hears the grinding of the boar’s tusks: he even addresses the hunters directly.Footnote ²

This illusion does not last long, however. After a series of questions directed at the hunters, the rhetor exclaims (1.28.2):

The sophist notes that what he sees is only a picture, his immersion dissolves and through the emerging distance he can start to elaborate on the painting and its quality. The rhetor’s comment on his own deception illustrates that not only his response to the painting is at stake in this passage. It has become a staple of Philostratean scholarship that graphē and its cognates signify both painting and writing.Footnote ³ Philostratus exploits this lexical ambiguity to align the sophist’s description with the picture and the response of the sophist to the latter with our response to the former: ‘Just as the speaker was fooled by the painted graphē, we readers are being continually hoodwinked, drawn in by the fiction of the mediating graphē of the text’.Footnote ⁴ This alignment of Philostratus’ readers with the internal viewers is encapsulated in another ambiguity, namely that of the second-person address. The speaker chides the boy and adolescents who form his audience for not alerting him to his delusion, but, as commentators have pointed out, the second person also reaches out to Philostratus’ readers, who did not intervene (indeed, how could they?).Footnote ⁵ Mere words, it seems, have the power to make us see a painting and access the scene depicted.

The beginning of Hunters showcases a pervasive tension in the Imagines. In a much-cited paper, Zahra Newby observes: ‘In response to the enticing naturalism of the image, the viewer exerts his own erudition, undermining mimetic or wonderful effects by explaining them. Yet at other times he allows himself to be overcome by the image’s power, lured in by images of erotic beauty, he is content simply to stand and stare …’.Footnote ⁶ Absorption, or what many scholars now call ‘immersion’,Footnote ⁷ on the one hand, and, on the other, erudition, a highly sophisticated reflection on the mode of representation, form two poles between which the response of the speaker and, as suggested by the pun on graphē, Philostratus’ readers oscillate.Footnote ⁸

As intriguingly as this dynamics of reception is presented in the Imagines, its display in Hunters conflicts with a basic insight of phenomenology. The speaker is first fully immersed in the picture and then starts to meditate on its facture. However, when we see pictures, we see simultaneously the representation and the represented object. We may focus on the pictorial scene and neglect the carrier and the marks that constitute it, but in the process of ‘seeing-in’ pictures, to use the term introduced by Richard Wollheim, the representation and what it represents are concurrent in our consciousness.Footnote ⁹ Even when we are strongly absorbed and have the feeling that the picture comes alive, we keep a residual awareness of attending to a representation. Likewise, words may transport us to places and actions they describe, but we remain aware of their mediation. In our response to verbal as well as visual representations, immersion is balanced with reflection.

Philostratus is a master: his Imagines are a mesmerizing account of the engagement of a highly sophisticated teacher with pictures. Nonetheless, the phenomenological inconsistency just noted should caution us against taking his reflections on the dynamics of response at face value. Philostratus stages a scintillating game with visual and verbal representations and their reception. Scholars trained in poststructuralist theory delight in tracing the mimetic recessions from the text that we read into the speech the rhetor delivers and ultimately to the pictorial scenes he describes and the images embedded in them; several representations are nested one inside the other, the medial boundaries being simultaneously flagged and erased. Precisely because the Imagines’ games with text and image are so brilliant and their hermeneutic operations so vertiginous, however, we should beware of the Philostratus trap and not mistake his rhetoric for an adequate description of the phenomenology of representation.

In this paper, I would like to challenge Philostratus’ claim to immersiveness. This claim, mostly implicit in the Imagines, becomes explicit in scholarship. I have already quoted Michael Squire’s claim that ‘just as the speaker was fooled by the painted graphē, we readers are being continually hoodwinked, drawn in by the fiction of the mediating graphē of the text’. In his essay on ecphrasis in the Oxford Handbook, Squire notes along similar lines: ‘The mimetic make-believe of the gallery’s tableaux—so believable as to move, speak, smell, touch and even whet our appetites …—serves to figure the simulation of Philostratus’ own descriptive project’.Footnote ¹⁰ Discussing the beginning of Hunters, Jas Elsner writes, with characteristic elegance: ‘This remarkable passage begins by pole-vaulting us, whether viewers or readers, directly into the image … The very rhetoric of description (all those questions directly addressed) is contrived to carry us into the image …’.Footnote ¹¹ In his monograph on Philostratus, Graeme Miles points out that ‘the vividness (enargeia) which the sophist aims to convey by his ekphrasis is much more than simply a clear mental image. He attempts, rather, to convey the experience of the painting through words, inviting readers to enter into it emotively as he does himself’.Footnote ¹²

While Elsner, Squire and Miles make explicit Philostratus’ implicit claim to immersiveness without probing into it, Ruth Webb goes one step further; she not only states ‘that we so often feel that we can perceive something of the painting is a tribute to the power of ekphrasis’,Footnote ¹³ but tries to identify the means employed by Philostratus: ‘This power derives partly from the concentration of sensuous detail and partly from the sense of immediacy created by the casual sentence structure with its careful impression of spontaneity. A large contribution is also made by the presence of the internal audience: the Boy’.Footnote ¹⁴ If we consider the points mentioned by Webb, a first important distinction emerges that tends to be elided in the recurring claims that Philostratus invites us to step into the paintings. Whereas the sensuous details qualify the scenes depicted in the painting, the casual sentence structure and addresses to the boy concern the narrative frame, the speaker’s act of describing these scenes. The former may help us to immerse ourselves in the scenes of the paintings and the latter jolt us to the villa in Naples, where the speaker is explaining the pictures to his young audience. Even without further theoretical reflection, it is hard to imagine that a reader delves simultaneously into the speaker’s conversation with the internal audience and into the scenes recounted by him. The immersion of the reader is not as easy as it appears at first sight.

More fundamentally, we have to ask: is it actually true that Philostratus makes us see the paintings and what they depict? The last few years have seen a growing interest in the question of what renders narratives immersive. Literary scholars have drawn on phenomenology and cognitive research to identify narrative features that trigger the imagination and instil in readers the feeling of witnessing what is narrated. Admittedly, the application of findings based on experiments with modern readers to ancient texts and their reception is open to challenge; such an approach, in fact, is likely to appear highly suspicious to classicists, who devote much energy to contextualizing their material. At the same time, it is, I think, not only justifiable for classicists to draw on the insights of cognitive studies, but it would also be unforgivable not to use this opportunity to gain sharper analytical tools than such notions as ‘sensuous details’ or unqualified claims about immersion. Cognition is certainly shaped by sociocultural factors, and yet it also relies on a physiological apparatus that seems very slow to change. Most importantly, the rich critical tradition in antiquity provides us with ample material to prove whether or not the results of cognitive studies may be applied to ancient readers.

In a first step, I will briefly introduce an embodied and enactive approach to literature and illustrate its heuristic value for the study of ecphrasis through a passage from the urtext of ecphrasis, the Shield of Achilles (section II). I will then draw on this approach to argue that, although Philostratus lays claim to enargeia, his Imagines contain various elements that make it hard for the reader to enter into the world of the paintings described (section III). While I will refer to different pieces from the Imagines to make this case, I finally return to Hunters. It is in fact the key Imago for Philostratus’ engagement with representation and immersion. In order to capture fully its significance, however, we cannot discuss solely the beginning, as most scholars have done, but need to consider the entire text and look particularly into its final part. My argument will point to a strong tension between the ecphrastic description of artworks and the idea of enargeia (section IV).

II. An embodied and enactive approach to ecphrasis

Edmund Husserl and his disciples scrutinized not only our perception of the world, but also our responses to representations. In the past few decades, their phenomenological meditations have been complemented by cognitive research. Cognitive scientists have used questionnaires and various physiological reactions to investigate the effects of texts and images on recipients. Literary scholars have started to use the results of these experiments to give substance to their discussions of reader response.Footnote ¹⁵ This is not the place for a survey of this line of inquiry, but a brief look at the argument of an earlier contribution to JHS, my article with Luuk Huitink on ‘Homer’s vividness’, may help alleviate Classicists’ anxieties about cognitive approaches as well as introducing the theoretical framework of the present inquiry and a first test case.Footnote ¹⁶

Ancient and modern critics have stressed how good Homer is at immersing his audiences and readers into the epic world. Homer’s vividness, however, is a conundrum: detailed descriptions are rare in his epics: how does he nonetheless succeed in transporting us to the battlefield of Troy and Odysseus’ palace? In our paper, Huitink and I argued that an enactive model of perception could help explain the immersive effect ascribed to Homeric narrative. Our perception, as this cognitive approach contends, is not photographic; instead, it concentrates on aspects that are relevant to actual and potential interactions.Footnote ¹⁷ When, for example, we see a hammer, we do not form a complete internal image of it, we attend to aspects that pertain to its use. Some of the same brain areas are activated as when we hold and use a hammer. Consequently, if our imagination works along similar lines to our perception, exhaustive descriptions of objects and situations are not cognitively realist and therefore unlikely to be immersive. The enactive model of perception and imagination suggests other features that render narrative immersive. The immersive features listed in the article on Homer’s vividness can be summarized in two points:

1. 1. Since our perception is closely tied to action, the account of action plays an important role in immersing readers. The narration of simple, intentional bodily movements in particular seems to have a strong echo in our sensorimotor system. The narration of actions should be dynamically veracious. This means that the time that a passage takes to read should be commensurate with the duration of the described action in the real world. While it is hard to pinpoint when exactly narrative time and narrated time fully coincide, except in direct speech, the notions of pause and summary illustrate ex negativo the salience of dynamic veracity.

2. 2. Descriptions ought to be ‘just in time’, focus on affordances and appeal to the senses. Just as our perception attends to things that are relevant to possible interactions, a narrative should introduce objects as and when they become relevant and focus on affordances, that is on aspects of things that pertain to potential and actual ways of interacting with them. Likewise, references to features that appeal to our senses make the description resonate with our sensorimotor system.

Needless to say, ancient audiences and readers are not available for MRIs or other tests which could prove that they responded to these features as modern readers do. And yet, as a reading of the chariot race in Iliad 23 reveals, the immersive features identified by phenomenologists and cognitive scientists are prominent in Homer and provide at least a possible explanation for the enthralling effect flagged by ancient critics. What is more, many passages singled out as examples of enargeia in ancient criticism do not provide detailed pictorialist accounts but conform nicely to the enactive model.

At least two points, however, need to be added to complement the enactive features that Huitink and I introduced:Footnote ¹⁸

1. 1. A prerequisite for the reader’s immersion in the narrated world is transparency. An overtly present narrator and strong self-referential elements highlight the act of mediation. Only an unobtrusive narrator and the absence of self-reference make it possible for the reader to focus on and, potentially, immerse themself in the represented scene.

2. 2. A shift of the deictic centre to the narrated scene can reinforce the reader’s immersion. Supported by adverbs (e.g. ‘now’ and ‘here’), tenses, prefixes and prepositions create temporally and spatially an internal viewpoint that helps to transport the reader. In focalization, this viewpoint is even tied to a specific character.

The degree of immersion in a specific text will of course vary from reader to reader. Whether or not a reader gets absorbed in a narrative depends on many factors, such as their own willingness to engage with the narrated world, its appeal to their interests and expectations, and also on their current mood and the specific circumstances of the reception.Footnote ¹⁹ It has also been noted that ‘for immersion to retain its intensity, it needs a contrast of narrative modes, a constantly renegotiated distance from the narrative scene, a profile made of peaks and valleys’.Footnote ²⁰ We cannot therefore expect immersion to be a constant state throughout the reading of a text; it rather goes up and down and, even at its strongest, is balanced by a residual awareness of the representation.

A brief look at a passage from the Homeric Shield of Achilles may illustrate how an embodied and enactive approach to narrative can be made fruitful for the interpretation of ecphrasis.Footnote ²¹ Though it stands at the beginning of the history of ecphrasis, the Shield of Achilles is quite complex in its entwinement of different levels that command the audience’s attention. As emphasized by Gotthold Ephraim Lessing, Homer describes the making of the shield.Footnote ²² The description of the individual scenes on the shield is marked off by references to Hephaestus forging it. There are also references to the material of the shield, some of them straightforward, others ambiguously blending together the representation with the represented. In the account of the shield and its production, Homer is, on the whole, a covert narrator, and yet his presence can be felt in comparisons and evaluations, when, for example, he calls Hephaestus’ work a thauma (Il. 18.549).Footnote ²³

Although Homer does not conceal his narratorial mediation and highlights the production as well as the material of the shield, several passages of his ecphrasis are designed to immerse the reader in the depicted scenes. To give a random example (Il. 18.550–60):

After the introductory verse, neither the production nor the material of the shield is mentioned; Homer gives his audience direct access to the depicted scene. The passage abounds in actions, which echo in the recipient’s sensorimotor system and thereby draw them in. The mention of the body parts engaged in these actions reinforces this resonance: the hands hold the sickles, the arms carry the ‘handfuls’ (δράγματα) of sheaves. Objects such as the sickles, the band and the sceptre are not mentioned for their own sake, but as and when they are used. The imperfect tense of most verbs qualifies the actions as unfinished and thereby creates an internal viewpoint. The deictic centre is also spatially anchored in the scene through prefixes and prepositions that carefully locate the individual actions: the sheaves fall next to each other, the children stand behind the binder of sheaves, the king stands among them, the heralds are far away under an oak tree. The description appeals not only to the sense of sight, but also to haptic perception (the sickles held in the hands are sharp) and acoustic perception (the king is silent).

While the passage concentrates on action, describes only things deployed by agents and firmly plants the deictic centre in the scene, it is more difficult to make a case for dynamic veracity. In other parts of the Shield, Homer freely narrates sequences of action, but in the account of the royal precinct he confines himself to a description of parallel scenes that could be represented in a mono-phase picture. However, the immersive appeal of the account is increased by the offer of a specific anthropocentric viewpoint. Both the position of the king (he stands amidst the men and children working) and his reflexive attitude (he is silent and joyful), suggest that not only the distance of the heralds refers to his position, but that the entire scene may be focalized through him. Affording a specific lens, the internal audience of the king makes it easier for the Homeric audience to relocate themselves in the narrated world.

Some verses in the Shield seem to gesture knowingly to the immersion of readers through the blending together of distinct representational levels. Material and colour in particular highlight the mediating artwork but, if they characterize the represented object as well as the representation, they can also blur the boundary between them. The black colouring of the soil overturned by the plough, for example, simultaneously describes the colour of the material to which the following verse draws attention: ‘The earth darkened behind them and looked like earth that has been ploughed / though it was gold. Such was the wonder of the shield’s forging’ (ἣ δὲ μϵλαίνϵτ’ ὄπισθϵν, ἀρηρομένῃ δὲ ἐῴκϵι | χρυσϵίη πϵρ ἐοῦσα· τὸ δὴ πϵρὶ θαῦμα τέτυκτο, Il. 18.548–49). As Becker puts it: ‘The physical similarity between depiction and depicted serves to enhance the audience’s respect for the ability of the visual image to reproduce significant aspects of the world’.Footnote ²⁴ Here, the Shield of Achilles anticipates some of the reflexive games with representation that we find in Imperial authors.

As the brief look at the precinct of the king in the Shield has shown, an embodied and enactive approach can help us assess the claim to immersion that inheres in the ecphrastic erasure of medial boundaries. Before we return to Philostratus, however, a possible objection to my approach needs to be considered. In the introductory section, I touched on the question of whether a model of cognition based on experiments conducted today also holds true for ancient audiences. More pointedly, we have to ask: does the emphasis on vision in ancient philosophy, criticism and literature not indicate that cognition in antiquity was pictorialist and that texts with detailed descriptions were in fact cognitively realist?Footnote ²⁵ After all, the key term for narrative and rhetorical vividness, enargeia, was frequently defined as the ability to make an audience see through mere words.Footnote ²⁶

To start with, the salience of the pictorialist model among laypeople today demonstrates that widely held views can conflict with cognitive realities. Most of the people who participate in the studies on which the embodied and enactive view of cognition is based would claim that they fully ‘picture’ narrated worlds and that immersive accounts require detailed descriptions. Just as their self-description is wrong, pictorialist models in ancient texts need not reflect the actual workings of cognition. Moreover, a closer look at ancient discussions reveals that enargeia is defined not only as the ability to transform listeners into viewers, but encompasses a wide range of devices, including onomatopoeia and the temporal organization of narrative. Its effects are described not solely in visual but also in multimodal sensory, cognitive and emotional terms.Footnote ²⁷ Strikingly, numerous passages singled out and praised for their enargeia by ancient critics are not pictorialist, but conform to an embodied and enactive model of cognition. In fact, if cognition in antiquity had been pictorialist, how is it that the poet who received the most lavish praise for his vividness, Homer, fails to provide detailed descriptions?Footnote ²⁸

III. Enargeia of the Imagines?

The idea that the reader is enabled to see the paintings and the scenes they depict is vital to the intermedial rollercoaster ride on which the Imagines take us. The claim of direct access is thrown into relief by references to multiple instances of mediation. Philostratus’ exploitation of the lexical ambiguity of graphē is only one of numerous ways in which this claim is pronounced, sometimes straightforwardly, sometimes in sophisticated ways that seem to enact the claim. In the last sentence of the proem, Philostratus unabashedly marshals a key term of ecphrastic theory when he asks the adolescents to ask questions ‘if anything I say is not clear’ (ϵἴ τι μὴ σαφῶς φράζοιμι).Footnote ²⁹ The second prominent term of ecphrastic theory besides saphēneia, enargeia, is invoked in several Imagines.Footnote ³⁰ In a characteristic inversion, it is applied to the paintings: ‘Critical languages for theorizing ekphrastic vision are here projected into the very gallery described: the trope of purportedly visualizing through verbalization is itself said to have been visualized, ready for Philostratus to verbalize in turn’.Footnote ³¹ The application to pictures of a term that praises the representational capacity of words drives home the idea that the reader gains direct access to the painted scenes through Philostratus’ descriptions.

In a metaleptic address to Olympus, the speaker even boasts that his verbal description trumps the image of Olympus in the water: ‘If it is beauty you are investigating, pay no heed to the water; for we are better able to tell all your charms’ (ϵἰ δὲ τὸ κάλλος ἀνακρίνϵις, τοῦ ὕδατος ἀμέλϵι· ἡμϵῖς γὰρ ἱκανώτϵροι λέξαι τὰ ἐν σοὶ ἅπαντα, 1.21.1).Footnote ³² Just as the sophist effortlessly interacts with the painted figures, it is implied, his descriptions, despite being verbal, let us see these figures, too. Words, at least those strung together by the rhetor, not only seem to outshine visual representations in their capacity to make us see something; at the beginning of the Erotes Imago, the rhetor even boasts that his words are able to conjure up smell: ‘Did you catch any of the fragrance hovering over the garden, or is this slow [in reaching] you? But listen carefully, for along with my description the fragrance of the apples also will come to you’ (μῶν ἐπῄσθου τι τῆς ἀνὰ τὸν κῆπον ϵὐωδίας ἢ βραδύνϵι σοι τοῦτο; ἀλλὰ προθύμως ἄκουϵ· προσβαλϵῖ γάρ σϵ μϵτὰ τοῦ λόγου καὶ τὰ μῆλα, 1.6.1).

At first sight, it may seem that the Imagines are in fact geared to deliver on their mimetic claims. To the frustration of earlier scholarship that turned to the Imagines chiefly for insights into the nature of Imperial tableaux,Footnote ³³ Philostratus says little about painterly technique. An important source of disruption of the reader’s immersion is thus minimized.Footnote ³⁴ Moreover, it has been noted that at least some of the paintings described are ‘unmögliche Bilder’.Footnote ³⁵ The multisensory impressions figuring in the rhetor’s descriptions exceed the representational capacity of paintings and are able to intensify the effect on the reader.Footnote ³⁶ Perhaps even more importantly, the sequences of action recounted by the rhetor would be impossible to represent in mono-phase pictures; at the same time, the narration of action can be a powerful means of transporting readers to represented worlds. The ecphrastic liberties that Philostratus takes and his neglect of painterly technique are both conducive to the immersion of the reader.

However, there are striking elements in the Imagines that are non-immersive. The rhetor is fond of explications that permit him to parade his erudition but lead away from the scene he describes. In the Imago of Hermes’ birth, for example, the mention of the place, Mount Olympus, is a welcome opportunity for the sophist to expound on its remoteness with the help of a reference to Homer (1.26.1). The programmatic first ecphrasis prominently introduces the key role of literary references in the Imagines, as the rhetor starts by asking the boy ‘to turn your eyes away from the painting itself so as to look only at the events on which it is based’ (σὺ δὲ ἀπόβλϵψον αὐτῶν, ὅσον ἐκϵῖνα ἰδϵῖν, ἀφ’ ὧν ἡ γραφή, 1.1.1.) and first discusses a passage from the Iliad.Footnote ³⁷ However, not all the literary references are as explicit as this one, and not all digressions are based on literary sources. In the ecphrasis subsequent to Hermes’ birth, the two-horse chariot used by Amphiaraus prompts the rhetor to explain that ‘the four-horse chariot was not yet in use by the heroes except by Hector the Bold’ (τὸ γὰρ ἐπὶ τϵττάρων οὔπω τοῖς ἥρωσι διὰ χϵιρὸς ἦν, ϵἰ μὴ ἄρα Ἕκτορι τῷ θρασϵῖ, 1.27.1). Homer does in fact have Hector use a four-horse chariot, but the rhetor does not mention Homer, let alone quote his work. When the rhetor comments on the origin of the Centaurs (2.3.1), it is doubtful that he refers to a specific text, and yet, like the quotations from literature, this and other explanations direct the reader’s attention away from the scenes represented in the pictures.

These digressions are in tune with the sophist’s announcements in the proem. There he emphasizes that he will teach his young audience how to ‘interpret’ paintings (ἑρμηνϵύσουσι, 3; cf. 5). The professed goal of his descriptions is hermeneutic instruction.Footnote ³⁸ That being said, digressions do not rule out readerly immersion. As briefly mentioned above, immersive passages gain force from alternating with less experiential passages. It is, however, difficult to find descriptions in the Imagines that are embodied and enactive. There is, for sure, no scarcity of descriptions. Time and again, Philostratus depicts characters, settings and objects in painstaking detail. To give an example, he devotes considerable attention to the dress and appearance of Rhodogune (2.5.2–5):Footnote ³⁹

We learn that the pressure of the brooches makes Rhodogune’s upper arm rise slightly, that her left hand extends beyond the handle of the shield, that part of her hair is dishevelled and part fastened and that her brows emerge from the root of her nose.Footnote ⁴⁰ In their elaboration on details, the descriptions in the Imagines easily stand up to comparison with the Hellenistic ecphraseis analysed by Graham Zanker.Footnote ⁴¹ Zanker is certainly right to emphasize the pictorial dimension of many Hellenistic descriptions of artwork. However, ‘visual richness’ in itself does not render descriptions ‘vivid and immediate to us’.Footnote ⁴² ‘The attention to precise, minute and even insignificant detail’Footnote ⁴³ is not cognitively realist. As pointed out above, our perception does not produce photo-like pictures of our environment; instead, it selectively attends to aspects that are relevant to actual and potential interaction. Like most of Zanker’s Hellenistic material, the pictorialist descriptions of the Imagines hardly immerse readers in the scenes described.

The portrayal of Rhodogune features other aspects besides an excessive amount of detail that are an obstacle to the reader’s immersion in the depicted world. The address to the boy alerts us to the fact that we are attending to a description of a painting. The speaker’s presence is also felt when he tells his audience what to admire. Some of his descriptions have an interpretive dimension that is hard to ignore, for example when he derives conclusions about Rhodogune’s character from her physical features. The speaker’s desire to interact with Rhodogune and notably his expectation that he may hear her speak Greek indicate his own absorption in the picture, but, together with the pictorialist description, the highlighting of his mediating instance powerfully prevents the reader from sharing in his experience.

The representational boundaries are not elided, as Philostratus suggests and scholars continue to assume; rather the attention that the rhetor directs to himself hinders the reader’s immersion in the scenes depicted. Even accounts that avoid effusive descriptions and instead concentrate on action tend to be punctured by comments and references that put the spotlight on the sophist. In the Arrichion Imago, for example, he postpones a description of the scene in order to ‘view the deed of Arrichion before it is finished’ (τὸ δὲ ἔργον τοῦ Ἀρριχίωνος, πρὶν ἢ παύσασθαι αὐτό, σκοπῶμϵν, 2.6.2). The rhetor seems to deliver on his promise of immediacy when he first turns to the audience at Olympia (2.6.2):

The internal audience helps the reader to relocate themself at the spot of the action, and the account of their response is enactive and likely to resonate in the reader’s sensorimotor system. That being said, not only do the final meditation on Arrichion’s previous successes and his preparation lead us out of the depicted scene, but the account of the scene is also interrupted by the reflection on the strong effect of spectacles. While the narration of the reactions of the internal audience works towards drawing the reader in, the general reflection forcefully reminds them of the mediating instance.

In other cases, it is not so much the speaker’s reflection as his comments on the image that establish a barrier between the reader and the scene described. The Amphiaraus Imago contains a description of horses that is enactive rather than pictorialist (1.27.2):

The rhetor describes the horses as they run. The movement is likely to trigger a sensorimotor echo in the reader, and yet the claim that the dust makes the horses look less beautiful and more truthful puts the spotlight on the painting and its mediation.

Philostratus does not forego the opportunity that massacres offer for a display of rhetorical brilliance. His Cassandra and Heracles furens lavishly depict scenes of carnage. Even if the sophist presents the result of the massacre rather than the fighting itself, verbs in the perfect tense evoke the action and spatial markers establish a deictic centre in the scene. At the same time, the descriptions are punctured by references that flag the speaker’s mediation, as here in Cassandra (2.10.3):

Together with the explanation of why the dead are not pale, the learned mention of the poetic expression distances the reader from the scene described.

In the Imago of Heracles furens, the qualification ‘evidently’ (ὡς δῆλα) and the generalizing comment on children’s tears, addressed to the boy in the second person, prevent the reader from moving more fully into the scene described (2.23.2):

The subsequent account of the slaves’ attempts to overpower Heracles is vivid, but the rhetor again addresses the boy in the second person when he points out that the Erinys cannot be seen because she has entered Heracles. In addition, the tragic intertexts that Elsner detects in both Cassandra and Hercules furens alert the reader to the artful composition.Footnote ⁴⁴

Less obviously, but nonetheless effectively, the rhetor’s highly wrought style undermines the immersive tug of his descriptions. Consider for example a sentence from the Menoiceus: ‘There is strength both in the promise of his shoulders and in his supple neck; he has long hair also, but not the long hair of luxury’ (ἔρρωται καὶ ὤμων ἐπαγγϵλίᾳ καὶ οὐκ ἀτρέπτῳ τένοντι, μϵτέχϵι δὲ καὶ κόμης, ὅσον μὴ κομᾶν, 1.4.3). ὤμων ἐπαγγϵλίᾳ is a striking periphrasis; together with the litotes οὐκ ἀτρέπτῳ τένοντι, it directs our attention to the rhetorical flourish instead of channelling it to the object described. In order to be immersive, language must be as transparent as possible.Footnote ⁴⁵ The less attention the mediation commands, the easier it is for the reader to delve into what is mediated. The exclamations, questions and anacolutha that make the rhetor’s mode of expression look so casual also draw our attention away from the paintings and onto the sophist’s speech act or Philostratus’ brilliance. When the form of the representation attracts the reader’s attention, their immersion in the scene described is necessarily diminished.

The μὴ κομᾶν illustrates Philostratus’ general fondness for negations in the Imagines (2.7.4):

Poseidon ‘is painted, not dark blue nor yet as a god of the sea, but as a god of the mainland’ (γέγραπται δὲ οὐ κυάνϵος οὐδὲ θαλάττιος, ἀλλ’ ἠπϵιρώτης, 2.14.2). Evadne ‘has determined to die for love of him (i.e., Capaneus), not by drawing a knife against her throat nor by hanging herself from a noose, modes of death often chosen by women in honour of their husbands, but she throws herself into the fire itself’ (Εὐάδνη γὰρ ἡ γυνὴ ἀποθανϵῖν ἐπ’ αὐτῷ ὥρμηκϵν οὔτϵ ξίφος τι ἐπὶ τὴν δέρην ἕλκουσα οὔτϵ βρόχου τινὸς ἑαυτὴν ἀπαρτῶσα, οἷα ἠσπάσαντο γυναῖκϵς ἐπ’ ἀνδράσιν, ἀλλ’ ἐς αὐτὸ τὸ πῦρ ἵϵται, 2.30.2). Here, as in other cases, the rhetor uses negations to gesture to other modes of presentation, mythical traditions and narrative possibilities. In reminding the reader of alternative representational options, this ‘side-shadowing’ spotlights the level of mediation and undercuts the reader’s absorption.

Numerous other examples could be adduced; however, the above should suffice to illustrate that the rhetor makes his presence strongly felt through his rhetorical artistry as well as through his explanations, interpretations and evaluations. Even in accounts that avoid pictorialist description and focus on action, the visibility of the sophist directs the reader’s consciousness away from the scene described and onto the process of mediation. The claim that, just as the speaker and his audience are absorbed in the paintings, their ecphrastic description allows us as readers to see the paintings and to step into their scenes is just that, a claim. Not least, the emphasis on the immersion of the sophist and his audience directs the reader’s attention to the framing narrative and work against their immersion in the painted scenes. Put provocatively, Philostratus is too busy showcasing his rhetorical brilliance and reflecting on the immersive capacities of verbal and visual presentation to deliver ecphraseis that are actually immersive. This need not be a flaw; I consciously avoid the conclusion that Philostratus ‘fails to’ draw us into the world represented. His project is, I believe, a different one, as we will see more clearly now as we return to the Imago with which my argument started, Hunters.

IV. Hunters, encore

Not unjustifiably considered a ‘Gipfelpunkt in den Eikones’, Hunters is a key reference in most discussions of Philostratean ecphrasis and immersion.Footnote ⁴⁶ However, these discussions tend to be confined to the first two paragraphs, the sophist’s illusion of being part of the scene and its subsequent correction. Admittedly, the beginning of Hunters is arresting, and yet only an exploration of the piece in its entirety will allow us to grasp fully its engagement with immersion and help us characterize Philostratean ecphrasis more sharply.

Hunters consists of four different parts of various length. The first part sets out the rhetor’s immersion (§ 1), the second his reflection (§ 2). Introduced by ‘So let us look at the details of the painting; for it is a painting before which we stand’ (σκοπῶμϵν οὖν τὰ γϵγραμμένα· γραφῇ γὰρ παρϵστήκαμϵν, 1.28.2), the sophist describes the painting (§§ 3–6): first the hunters (§ 3), then the boy (§ 4), followed by the equipment and dogs (§ 5) and a temple to Artemis (§6). The fourth and final part narrates the hunting of the boar and its killing by the boy (§§ 7–8).

While the second and third are clearly marked off from the preceding parts, the transition into the final part, the narration, is less conspicuous. The description of the temple to Artemis that concludes the third part starts to dynamize the ecphrasis temporally; its beginning gestures towards a sequence of action and thereby prepares the narration that is to follow: ‘And the hunters as they advance will hymn Artemis Agrotera … After a prayer the hunters continue the hunt’ (καὶ τὴν Ἀγροτέραν προιόντϵς ᾄσονται … ἔχονται μϵτὰ τὴν ϵὐχὴν τῆς θήρας, 1.28.6). Likewise, at the end of the Imago the narration of the hunt discreetly returns to the static scene of the painting. After the hunting and killing of the boar and the response of the other hunters are recounted in the present tense, the perfect tense of ‘one is thrown from his horse which he excited beyond control instead of holding it in check’ (καὶ πέπτωκέ τις ἀπὸ τοῦ ἵππου μὴ κατασχών, ἀλλ’ ἐκθορυβήσας τὸν ἵππον, 1.28.8) lays the ground for the frozen scene of the final sentence: ‘The lad is still in the pool, still in the pose in which he hurled his javelin, while the youths stand in astonishment and gaze at him as though he were a picture’ (ἔτι ἐν τῇ λίμνῃ τὸ μϵιράκιον, ἔτι ἐπὶ τοῦ σχήματος, ᾧ τὸ παλτὸν ἀφῆκϵν, οἱ δὲ ἐκπϵπλήγασι καὶ θϵωροῦσιν αὐτὸ οἷον γραφέν, 1.28.8). Nonetheless, although it commences and fades out less abruptly than the others, the final part is markedly different from the one preceding; it replaces the static description of the details of the painting with a narration of action that brings together the individual groups previously described separately.

If we contemplate the four parts of Hunters, we note that they form two corresponding pairs. The description of part 3 and the narration of part 4 form a dichotomy just as the immersion of part 1 is juxtaposed with the reflection of part 2. The first pair brings together two different modes of response, the second two different modes of presentation. While not fully mapping onto each other, the two pairs are linked, more specifically part 3 to part 2 and part 4 to part 1: the awareness of having a painting in front of him leads the sophist to its description. The narration of the action is highly immersive and thus harks back to the beginning of the ecphrasis.

Let us take a closer look at the final part and its correspondence with the first. The closing remark that ‘the youths stand in astonishment and gaze at him as though he were a picture’ (1.28.8) takes up and inverts the initial immersion of the sophist: just as he treats a painting as if it were reality, the hunters regard a real person as if he were only painted.Footnote ⁴⁷ However, this time it is not the sophist who is immersed, it is Philostratus’ readers who are invited to step into the world represented. In the account of the hunt, the sophist refrains from reflections and comments on the painting; he directly accesses the scene, which he renders as an account of sequential events: the boar ‘leaps from the thicket, then rushes at the horsemen’ (ἐκπηδᾷ τῆς λόχμης, ϵἶτα ἐμπίπτϵι τοῖς ἱππϵῦσι, 1.28.7); the hunters fail to hit the boar lethally and wound it only superficially. Weakened by this wound, however, the boar ‘runs through the woods till it finds refuge in a deep marsh and a pool adjoining the marsh’ (φϵύγϵι διὰ τῆς ὕλης, ἐκδέχϵται δὲ αὐτὸν ἕλος βαθὺ καὶ λίμνη πρὸς τῷ ἕλϵι, 1.28.7). The hunters follow it, but only the youth rides into the water: ‘the creature tries to wound his horse, but bending well over on his horse and leaning to the right he delivers with the full force of his arm a blow that hits the boar just where the shoulder blade joins the neck’ (καὶ τὸ μὲν θηρίον ἵϵται τρῶσαι τὸν ἵππον, ἀπονϵῦσαν δὲ τοῦ ἵππου τὸ μϵιράκιον καὶ ἐς τὰ δϵξιὰ μϵτακλῖναν ἀφίησι τῇ χϵιρὶ πάσῃ καὶ βάλλϵι τὸν σῦν κατ’ αὐτὸ μάλιστα τὸ συνάπτον τὴν πλάτην τῇ δέρῃ, 1.28.8). The dogs carry the boar to the bank; the onlookers shout at the top of their lungs.

Obviously, the narration of the hunt is highly enactive and embodied. The sophist steps back and refrains from explanations and rhetorical epideixis. He disappears behind his narrative, which concentrates on simple actions and movements, most of them intentional and therefore particularly strong triggers of the reader’s sensorimotor system. Prepositions and prefixes shift the deictic centre of the account to the scene. The setting is described as and when it becomes relevant to the action; the marsh and pool are only mentioned when the boar runs into them. Unlike the pictorialist descriptions of part 3, this description is cognitively realist in that it does justice to the entanglement of our perception with action. It is notoriously difficult to evaluate dynamic veracity, but the simple chronological narration of the hunt leaves the reader with the impression that the verbal rendering closely follows the sequence of action.

If we look for a passage that ‘is contrived to carry us into the image’, that is ‘pole-vaulting us … directly into the image’,Footnote ⁴⁸ then we find it here. It is not the striking first paragraph but the final narration of the hunt that cues the reader to step into the painted scene.Footnote ⁴⁹ The beginning spectacularly showcases the power of representations to draw their recipients in, but it does not throw the reader into the scene of the painting. Contrary to what many scholars believe, there is no seamless recession of representational levels that allows the reader to enter the scene of the painting together with the sophist. The focus on the absorption of the sophist channels the reader’s attention to the scene in the villa at Naples instead of the painted scene. The hyperbolic nature of the sophist’s response and the Homeric allusion in χλούνην σῦν even make it unlikely that the reader is transported to the villa.Footnote ⁵⁰ The ending, on the other hand, elides the mediation of the rhetor and gives a highly experiential account of what is pictorially represented. Here, transparency is not boasted but generated.

Considering the Imago as a whole, not only its beginning, we can have another go at its significance. As my reading shows, Hunters offers a highly dissective analysis of representation and response: Philostratus first focuses on the response of the sophist, then on the representation. He has him separately describe the painting and narrate the action it represents, just as he balances an instance of reflection against a moment of immersion. Thereby Philostratus neatly breaks down something that itself is closely entwined. Whereas the sophist starts with the effect of the painting on the viewer and then moves on to its description, it is the features of the painting that elicit the viewer’s response. The immersion and reflection that Philostratus has follow upon each other are concomitant in our response. By the same token, our comprehension of the action behind the represented scene is not dissolved by our perception of the painting, as it appears in Hunters.

Hunters is unique in its precise anatomy of pictorial representation and its effect on viewers, and yet similar tendencies can be noted in other Imagines. In Gyraean Rocks, for example, the sophist first narrates the shipwreck of Ajax and then explains the painting: ‘Such is the story of the painting, but what is shown to the eye is this’ (ὁ μὲν δὴ λόγος τῆς γραφῆς οὗτος, τὸ δὲ ἐναργές, 2.13.2). In Cassandra, the sophist deploys a similar juxtaposition to score a point in the paragone with the arguably most famous presentation of the scene in another medium: ‘If we examine this scene as a drama, my boy, a great tragedy has been enacted in a brief space of time, but if as a painting, you will see more in it than a drama’ (καὶ ϵἰ μὲν ὡς δρᾶμα ἐξϵτάζομϵν, ὦ παῖ, ταῦτα, τϵτραγῴδηται μϵγάλα ἐν σμικρῷ, ϵἰ δ’ ὡς γραφήν, πλϵίω ἐν αὐτοῖς ὄψϵι, 2.10.1). As in Hunters, the description of the painted scene and the narration of the depicted action are separated. Likewise, as Newby has demonstrated, the shifting from moments of immersion to cool and detached reflection is characteristic of Philostratean ecphrasis. The sharp juxtaposition of immersion with reflection at the beginning of Hunters exacerbates a tension that pervades the Imagines more broadly.

Hunters, I contend, crystallizes the general character of the Imagines. Its anatomy of representation and response, which carefully dissects concurrent aspects and dimensions, is emblematic of the reflexivity that is at the heart of Philostratus’ ecphrastic work. The Imagines are not so much an ecphrasis in the sense of the Progymnasmata: Philostratus does not primarily try to make his readers see the scenes that he describes. Rather, he engages in a meditation on the ecphrastic practice described in the Progymnasmata. The Imagines offer, to use a term coined by Elsner, ‘meta-literary ekphrastic reflections’.Footnote ⁵¹ These reflections are alluring, in fact so alluring that scholars have fully accepted them. Webb is representative of a larger corpus of scholarship when she states: ‘il s’agit d’un texte qui se sert de l’idée de l’art non seulement pour réfléchir sur la nature de l’immersion fictionnelle mais aussi pour faire vivre cette immersion au lecteur, qui est ainsi invité à réagir aux scènes mythologiques comme si elles se passaient réellement sous ces yeux, et aussi à entrer dans le monde fictif de la galerie’.Footnote ⁵² However, an embodied and enactive reading of the Imagines cautions us against stepping into the Philostratus trap and mistaking the reflection on and claim to transparency with transparency itself.

This does not mean that we should deny the Imagines the label of ecphrasis. To limit ecphrasis to attempts to actually make the reader see artworks (in addition to all sorts of other objects), would be a cruel mutilation. Philostratus takes up and lifts to a new level a rich Hellenistic and Imperial tradition of ecphrastic epigrams that stage responses to art rather than describe it.Footnote ⁵³ The epigrams on Myron’s cow, for instance, tell us little if anything about the statue’s features; instead, they praise its lifelikeness and the technē that makes this possible. Michael Squire has brilliantly teased out the intriguing reflection on verbal and visual representation encapsulated in these epigrams.Footnote ⁵⁴ This reflexive dimension of ecphrasis is already present in the Shield of Achilles, which playfully blends together different levels of presentation. Ultimately, it is the reference to another representation that endows the ecphrasis of artwork with a reflexivity that quickly pushes the goal of making the hearer see something into the background. If, as in the Imagines, the immersive capacity of the artwork is illustrated by the response of a speaker or internal audience, then the attention drawn to their immersion goes against the immersion of the reader in the scene represented by the artwork. It is perhaps not incidental that in our records artworks are not mentioned explicitly as the object of ecphrasis until the fifth century CE.Footnote ⁵⁵ Beginning with Homer and masterfully exploited by Imperial authors, above all Philostratus, the reflexive dimension of representing representations grinds against the simple goal of transforming an audience into witnesses.