My investigation into the cognitive aspects of landscape description takes as its focus the landscapes that the poet evokes on the Shield of Achilles (Il. 18.478–608). Drawing on studies in cognitive psychology I note the extent to which an audience might derive a ‘spatial mental model’ from the topographical or ‘locative’ indicators that the Homeric poet offers. Then I consider the ‘non-locative’ information that the poet conveys about the landscapes of the Shield. In this connection I develop Barbara Tversky's notion of landscape representation as a ‘cognitive collage’. What makes the scenes on the Shield vivid, however, is human presence. Finally, therefore, I draw on enactivist theories of cognition, recently introduced into Classics, which offer a valuable supplementary approach to ‘reading’ and enjoying landscape descriptions in Homer.

For some years there has been considerable scholarly interest in landscape, particularly in the location of myth in its landscape and the interaction of memory over time with those landscapes. This paper too is landscape-focussed: it seeks to understand the means by which landscape is described, especially, but not only, in oral contexts, and how landscape descriptions, by contrast with purely narrative text, are processed by their audiences. The approach adopted here is a cognitive approach. The testing ground for this investigation is that cluster of landscapes that adorn the Homeric Shield of Achilles, a creation that, Hephaestus claims, will inspire wonder in all who behold it.Footnote ¹ The kinds of questions I ask are: what strategies do oral poets use to evoke a landscape? What, from the poet's viewpoint, is salient in the landscape and what is not? What aspects of this landscape do audience members (or subsequent readers) conjure up in their own minds? How do they do this? And, finally, how is it that the poet's descriptions of the landscapes of the Shield achieve their effects? I refer here to the delight and pleasure that the poet's audiences find in these scenes.

The discussion will, appropriately, follow a circuitous path. I propose to trace out a search for a cognitively credible—cognitively realistic—approach to landscape description in a poetic context. I then apply my findings to a selection from the ten separate and distinct static scenes—a skyscape, some cityscapes and some rural landscapes—rendered in bronze, tin, gold and silver, that the god Hephaestus works on the Shield that he creates for Achilles at Thetis’ request (18.478–608).Footnote ² Taken together these individual representations of landscape (the term is used here in its most general sense) make up what we might call the topography of the Shield; they are elements of a remarkable imagined ekphrasis that, in narratological terms, precedes a crucial moment in the epic: Achilles’ return to the fighting.Footnote ³ What has attracted my interest in this cluster of landscapes is that the poet tells us that these scenes should be envisaged in two dimensions, worked in precious metals across the Shield's surface; but what he offers us, and what we in turn reconstruct, is a landscape in three dimensions. And what is remarkable is that, although the scenes the poet describes are generic scenes of everyday living, they are, each one, vivid—and memorable.Footnote ⁴

One outcome of this descriptive vividness has been that industrious scholars in past generations attempted to reconstruct the Shield on paper as though the poet had described a splendid creation in metal that he had himself seen or that he could envisage in every detail: it was assumed that the poet had a ‘picture’ in his mind that he strove to reproduce in words.Footnote ⁵ This belief generated a number of detailed plans, which demonstrated the futility of the exercise.Footnote ⁶ The poet was more interested in emphasizing the wonder of this object, and its capacity to evoke the world as he knew it (or wanted to know it), than in detailing precise dimensions and precise details of its decorative topography—a point to which I return below. The Shield is, as Alex Purves notes, ‘protocartographic’.Footnote ⁷ Such an artefact was not to be found in the real world; it was to be conjured up in the mind.

So let us consider some important cognitive approaches to landscape description and assess their appropriateness to the task in hand. I begin with a purely spatial approach. In the 1990s, as studies in cognition expanded their scope, there was considerable interest in how our cognitive apparatus responds when we encounter discourse that includes spatial terminology—information about direction, landmarks, spatial relationships, distances. Until then it had been generally supposed that, when we read or listen to a description of a spatial environment, we each develop a ‘mental map’ that represents the landscape described. But Barbara Tversky and Holly Taylor's studies demonstrated that this was not the case.

Tversky and Taylor have shown that, as we follow a description of a landscape, we do not create a two-dimensional cognitive map; rather, we adopt a more constructivist approach. If there is a generous quantity of what they call ‘locative’ information—the kind of information identified above, such as spatial terminology, topographical details, directions—we construct, step by step, a three-dimensional ‘spatial mental model’.Footnote ⁸ That locative information may take one of two forms: it may be presented as on-the-ground ‘route’ information, which takes readers or listeners on a mental tour of the landscape, moving them through the environment, with directional pointers at every step, as Pausanias so often does; or it may be presented from a ‘survey’, or bird's eye, perspective, describing landmarks relative to one another, and capturing spatial relations.Footnote ⁹ Whether we are considering a route or a survey perspective, locative information is integral to Taylor and Tversky's spatial model.

A second category of information that may assist us in constructing a three-dimensional model is not topographical but elaborative. Taylor and Tversky use the term ‘non-locative information’ to describe this material, which conveys further details about landscapes: this material may tell us more about landmarks and their appearance, their significance, or their history, for example.Footnote ¹⁰ Information of this kind supplements purely spatial (or locative) information; it will enrich the mental model that we develop.

To illustrate these points here is an example from Pausanias, on the road from Mycenae to Argos; describing this landscape from a route perspective he presents a combination of locative (underlined) and non-locative (italicized) information (Paus. 2.18.3, transl. P. Levi):

For contrast, consider another passage from Pausanias. We are now in Phokis. This time we are not, as it were, travelling with Pausanias across the landscape; rather, evoking a landscape from afar, he presents a survey perspective (Paus. 10.34.1, transl. P. Levi):

Now we turn to the landscape of the Iliadic story itself—the Trojan landscape. Thanks to the topographical markers that the poet indicates, we can begin to build a coherent spatial model of the action on the Plain of Troy, as Jenny Strauss Clay demonstrates.Footnote ¹¹ Sometimes the poet takes us along a route, as Priam makes his way to Achilles’ domain in Iliad Book 24: from the citadel, past the tomb of Ilus, across the river at its ford, towards and into the camp of the Achaeans, with their ships on the shore (24.322–467). At other times the poet offers us a survey perspective: we look down on the plain from Olympus, with Zeus and the other gods (24.23–4), or from Mount Ida (8.47–52) or Samothrace (13.10–14), and watch the fighting from above.

But when we move from Homeric narrative to the poet's account of the Shield of Achilles, to this long Homeric ekphrasis (18.478–608), there is little chance that we can construct such a cognitive model for the Shield as an entity, for the poet has provided us with limited spatial information. Only the final element, the River of the Ocean, is explicitly located in its structure (18.607–8):

What is contained within that outer rim is presented by the poet in narrative form, and in temporal order, as a catalogue of separate scenes, in each case introduced by the verb ‘he made’ (ἔτευξ’ and its equivalents, at lines 483, 490, 541, 550, 561, 573, 587, 590, 607).Footnote ¹³ We are never allowed to forget that we are witnessing a shield under construction.Footnote ¹⁴ But the poet offers us no specific information about spatial (by contrast with temporal) relationships—about the location of one component scene relative to another.Footnote ¹⁵ It is simply not possible for audience members to construct for the Shield as an entity either a map or a coherent mental model of the kind that Taylor and Tversky propose.Footnote ¹⁶

If we consider the individual scenes on the Shield, we find only one that includes identifiable points of spatial reference: the first scene that the poet describes offers a global view of the poet's world, with special attention devoted to its skyscape (18.483–9):

The poet here conjures up a three-dimensional model from an earth-based survey perspective, looking up, not looking down—a perspective that includes the earth, the ocean, the sun, the moon and constellations visible to audiences in the northern hemisphere from May to November. The spatial markers that he offers allow us to construct a spatial mental model along Taylor and Tversky's lines. There are, however, two details worthy of note at this point. First, this is no inanimate image. These bodies have been arrested in motion: the moon is waxing (πλήθουσαν, 484); the Bear is turning about (στρέφεται, 488). And there is, besides, a modest degree of personification, which holds our attention: the Bear is keeping an eye on Orion (δοκεύει, 488).Footnote ¹⁷

In the subsequent scenes from the Shield, however, we find little precise spatial terminology: there are few spatial or directional pointers; and topographical detail is sparse. The poet's landscapes, such as the elements of two cityscapes evoked in the long passage 490–540, the field described at lines 541–9, and the precinct of a king described at lines 550–60, are presented impressionistically, with only occasional specific allusion to spaces and places. And yet, as introspection tells us, despite these limitations, audience members are able to construct a three-dimensional representation of each individual scene. What are the features of these brief passages that enable us to do this in the absence of explicit spatial terminology? And what mind-based resources do we draw on to help us in this project?

To respond to these questions let us set aside for the moment Taylor and Tversky's spatial model. Although their concept of a spatial mental model, and their terminology, will continue to be useful, it is clear that route and survey perspectives alone are insufficient to account for our ability to imagine landscapes like those of the Shield. I now turn to some further work by Tversky, who has proposed that in the absence of topographical information it is possible to make use of the non-locative information described above to assist in the creation of something akin to a mental model. She proposes the metaphor of a ‘cognitive collage’ to describe the result.Footnote ¹⁸ Like the elements of a collage in the world of art and design, the elements that we assemble may be disparate and fragmentary. Such a collage may comprise, as Tversky sees it, occasional elements of route and survey perspectives; it may include epithets, discursive commentary and, as George Miller has proposed, items supplied from our own memory store: a proposal that is in line with Quintilian's observations.Footnote ¹⁹ Tversky describes this accumulation of unrelated information as a sequence of multimedia ‘overlays’ that represent different points of view.Footnote ²⁰ As she notes, however, it is unlikely that this collation of fragmentary information could be organized into a single coherent cognitive structure like the route and survey accounts of landscape that Pausanias offers. The ‘model’ obtained on the basis of this collation of information will always be less precise, more impressionistic.

Clearly, this approach holds promise for an exploration of the ways in which audiences respond to the landscapes of the Shield. But, whereas Tversky uses the expression ‘collage’, a term that seems, to me at least, to emphasize two-dimensionality, I prefer the expression ‘cognitive assemblage’ to describe what we construct in our minds, in three dimensions, as we process largely non-topographical material.

But there is one element that Tversky and Taylor have not yet integrated into their proposals: this is the human element. Landscape in the viewer's eye may be more than an assemblage of topographical features. Landscape may also represent a realm of experience. It is notable indeed that almost all the landscapes evoked on the Shield include the activities of men and women, whose presence sparks our interest. As an important study in perception indicates, when people view scenes of any kind, whether in the real world or in a range of media, they pay special and sustained attention when human figures are included.Footnote ²¹ Even that first scene on the Shield, the skyscape, featured the movement of the moon and the humanlike behaviour of one of the constellations. In all but one of the subsequent scenes we observe people going about their daily activities, marking important events in the life-cycle, in conflict and at peace—and each cameo-like action-sequence is set in its own appropriate environment. These landscapes, which are often only roughly sketched in, are nevertheless so very vivid to us, not so much because of their topographical form but because of human presence and, more important, human activity within each one.

In order to trace the function and the effect of human participation in the scenes of the Shield, we should now bring to bear enactivist theories of cognition, which are currently being absorbed into classical studies. These theories, which draw on contributions from the broad field of psychology, and from studies of perception and neuropsychology in particular, offer valuable insights into the ways in which storytellers create sequences of narrative action and in which we audience members reconstruct these sequences in our minds as we follow them.Footnote ²² Whereas others have discussed enactivist theories in relation to Homer's narrative passages, I shall test the idea that, in Homer at least, we can fruitfully apply these theories also to a sample of his non-narrative passages—specifically to the evocation of landscapes on the Shield.Footnote ²³ I argue that it is the incorporation of the activities of men and women, older and younger, into each small scene that both encourages and enables the audience to trace out and to visualize the landscapes in which the actions take place. The poet's account of their activities, therefore, plays a vital role in our conceptualization of each of the landscapes of the Shield.

As a preliminary exercise, however, let us take a sample of Homer's Iliadic narrative and consider—very briefly—the neuropsychological evidence. These lines recount Hector's actions as he breaks through the gates in the Achaean wall (12.457–9, 12.462–3):

What is our experience as we read (or hear) these lines? Recent work in psychology has shown that, when we conjure up in our imagination a scene that depicts human action, we do not achieve this by simply visualizing that action as though we were passive third-person observers.Footnote ²⁴ Behavioural studies and experiments using neuroimaging have demonstrated that, as we process passages of this kind, we experience in our cerebral cortex sensory traces related both to the movement in question and to the object it is directed at.Footnote ²⁵ This experience is called ‘motor simulation’ or ‘motor resonance’:Footnote ²⁶ the hand area of the cortex responds in the moment to hand-related actions described in the text and the foot area responds to foot-related actions noted there; we have evidence too of vicarious visual perception as we trace the movement of objects.Footnote ²⁷ Furthermore, regions involved in the navigation of spatial environments are activated when we encounter indicators of movement through space. Thus, as movement through space is registered cognitively, a so-called ‘multimodal image’ is registered of the environment in which that movement occurs.Footnote ²⁸ The spatial imagery that has been prompted may be fragmentary, instantaneous and extremely short-lived, but it provides an intense but momentary sense of having physically entered a tangible environment—this is that sense of ‘presence’, of ‘being there’ that we all can recognize from our own experiences of listening and reading, as, for example, along with Hector, we burst through the gates in the Achaeans’ wall.Footnote ²⁹

On the basis of extensive support in experimental psychology for the covert movement that readers or listeners experience,Footnote ³⁰ these sensorimotor responses, taken together, have been described as ‘enactment’;Footnote ³¹ and it is to these momentary experiences of enactment, along with the associated experience of ‘presence’, of ‘being there’, that we may attribute much of a narrative's intensity, its impact and its memorability.Footnote ³²

Let us return now to the scenes on the Shield. My aim is to demonstrate that here too, as we process the poet's account of human actions and human movement, these actions and these movements create the environment in which they occur. The first example is a townscape (18.490–6):

In her study of motor enactment Anežka Kuzmičová tells us that, if an enactive sense of presence is to emerge in a passage of literary narrative, it is essential that the poet should have interspersed obviously dynamic indicators of human movement throughout that passage.Footnote ³³ This is precisely what we find as we read lines 492–6 above. Although the default verb tense used to describe human movement throughout this series of landscapes is past progressive (or imperfect), which might represent the static nature of the scene worked on the Shield, the scene itself is presented by the poet as cinematic, as a video clip of action.Footnote ³⁴ The effect of this representation of movement and action on us audience members is immediate; it stirs motor resonance. At lines 492–3 we leave home along with the brides and we move along the streets; we see the flaming torches and hear the wedding song: there is a suggestion of excitement in the air. We dance (494) with the young men; and we hear the sound of flute and lyre. And now changing perspective, just as Tversky suggested we might, we look out from our doorways to watch approvingly as the merry band passes by (496).

In following the movement of the brides and their escorts, we make our way through a notional three-dimensional space, a generic space, in which all this action occurs.Footnote ³⁵ What I construct in my mind is a street in a town (although the poet does not mention a street); this street is filled with activity; there is movement and dancing. And along the street I locate houses, and in their porches I see women, like us, watching the parade.Footnote ³⁶

As we process the poet's description we do not consciously build up a picture, or a map, in our heads, as Taylor and Tversky suggest; rather, through the language cues that we receive, we gradually acquaint ourselves with a three-dimensional environment, factoring in as we proceed the landscape elements that we need. In the absence of further information from the poet, we draw on our own memory store to supply generic non-locative data: my memory store, for example, provides rough paving in the street and stone facades for the houses.Footnote ³⁷

It must be the case, therefore, that any one audience member's representation of the setting in which this action takes place will not be exactly the same as another's; nor will it be what the poet himself has envisaged.Footnote ³⁸ Our imaginative responses will certainly not have the accuracy and the completeness of a detailed picture; such an outcome would be cognitively unrealistic—unrealistic for the poet to present and unrealistic for his audience to process.Footnote ³⁹ In this first city-scene, the street and the houses that line it are represented vaguely. They are shadowy forms; but they are adequate for the purpose. What we listeners or readers bring to mind will be just enough to allow us to make sense of the action represented: a scene of joyful celebration in a small community.

It is remarkable that, although the poet is describing a representation in metal, he invites us to observe the sounds that accompany the bridal procession that he describes: we register the bridal song (ὑμέναιος, 493), the sound of flutes and lyres and the shouting (αὐλοὶ φόρμιγγές τε βοὴν ἔχον, 495). This aural information has nothing to do with landscape, if we are thinking only of visual elements, but it contributes another dimension to the cognitive assemblage that represents this landscape.Footnote ⁴⁰

Finally, before we leave this example, we should pause over the image of the spectator: the women, that is, who offer a survey perspective on the merry bridal procession as they trace out their route. The poet has used this same device for the same purpose on three other occasions in this series of cameos: the constellation of the Bear which keeps an eye on Orion (18.487–8), the king who quietly surveys his reapers (18.556–7), and the crowd that watches the dancers (18.603–4). In each case the watcher takes pleasure in the scene—and frames it.Footnote ⁴¹ And yet, in using this device in this context, the poet is playing with his audience. We believe initially that we are witnessing the construction of a series of images on a shield; and yet, as we study each image while it is being constructed, we find ourselves embedded in the scene both by our enactive responses to human movement of all kinds and by the poet's insistence that we take the cue from his silent observers within the landscape, and enjoy the scene he presents.Footnote ⁴²

After the celebration of the wedding the poet takes us to a judgement-scene in this well-ordered community and from that scene he proceeds to his second town, which, like Troy itself, feels the pressures of war; life is chaotic (18.509–40).Footnote ⁴³ From there he moves to landscape proper, my second example. This is a rich agricultural landscape, a scene of productive order that, simply by its co-location in the narrative of the Shield's creation, provides vivid contrast to that city at war. The poet takes us now to a field dotted with human figures (18.541–6):

The length of this field is marked out by the movement of the ploughmen, its breadth by that of the man who approaches them at each furrow's end with a refreshing drink. The evocation of this agricultural landscape through our enactive experience is supplemented by a number of epithets that elaborate on the richness of the land and on a bountiful lifestyle for those who reap its harvests: we find a soft fertile field that is wide and triple-ploughed (νειὸν μαλακὴν πίειραν ἄρουραν | εὐρεῖαν τρίπολον, 541–2); and those who plough it are offered wine that is μελιηδής, honey-sweet, 545.

Later again, and this is my third example, we are in a vineyard surrounded by a fence (18.561–72). The landscape is introduced through some sustained visual detail (18.561–5):

As he does throughout his account of the creation of the Shield, the poet's description of the vineyard emerges through a representation of the god at work, as narrative:Footnote ⁴⁴ the interventions of the god at line 561 (ἐν δ’ ἐτίθει) and again at line 564 (περὶ . . . ἔλασσε) ensure that this scene holds our attention. And with the introduction of a path in the following line, at 565, the poet introduces human presence (18.566–72):

The grape-pickers run along the track, carrying off the fruit, in the company of a lad playing on his lyre, whose song sets the pace as they skip along. The abundance of indicators of movement and sound (the notes of the lyre, the singing, whistling, and thump of feet on the ground) brings this landscape to life. The relevant areas of our brain respond. And, along with this enactive experience, there is the poet's elaborative material to factor in: the fresh youth of the participants and the pleasing image of baskets of sweet fruit.

There is an interesting contrast between the two components of this scene. What we focus on is not so much the vineyard itself but the path that leads away from it. Although the path that each of us envisages may be different from the path envisaged by others, we all bring to mind a generic track, from a route perspective, and we each develop an individual cognitive assemblage, drawing on the effects of motor resonance as we follow the action, and factoring in the elaborative information that the poet has offered us.