In our everyday discourse we make frequent reference to pictorial narratives. We exclaim on the hunt scene in the cave painting, mention the frenzy unfolding amidst a human catastrophe in the graffiti, discuss the adventure of the baby in the illustration, and whisper about the humorous rendition of a devastated a marriage in the oil painting.
Yet a more precise question concerning pictorial narrators remains almost entirely ignored. Narratologists are still primarily focused on literary rather than pictorial narratives, the latter remaining relatively neglected along with the putative class of pictorial narrators.1
The philosopher Kendall Walton is an exception. In his paper ‘Points of View in Narrative and Depictive Representation’ he compares literary to pictorial narratives and concludes, sceptically, that pictures do not instantiate narrators. He argues that picture viewers directly access depicted content. By contrast, readers can only represent literary narratives along with a narrator. The literary narrator is crucial because he or she ‘mediates the reader’s access to the rest of the fictional world’. As a result, ‘we know what happens in the fictional world only from his reports about it’.2 Since pictures are accessed directly, they cannot sustain narrators, nor do we require them to apprehend the content.
However, this is at odds with our experience of, say, William Hogarth’s celebrated paintings of failing marriages and impecunious fools. In apprehending these pictures adequately, it seems apt to say the viewer is told a story. But what tells the viewer the story? Why think that the viewer spontaneously chooses to ‘tell it to herself’? And why would she tell herself a story when looking at these pictures and not others?
This paper has two main aims. The first is to clarify what is meant by ‘pictorial narrators’ by pointing to and discussing some positions in the literature. Since the issue has garnered almost no attention, the groundwork needs to be laid down in this exploratory paper. The second is to argue in favour of the claim that there are pictorial narrators.
I will proceed as follows. In Section I, the standard view of narrative is revisited and outlined along with an account of the narrator familiar from novels. The literary case is then roughly mapped onto pictorial cases. This serves to establish the strength of the preliminary claim that pictures manifest something akin to literary narratives. The exercise produces a rudimentary but testable definition of a ‘narrator’. In Section II, I review sceptical and friendly responses to the existential claim, that is, that pictures can or may manifest narrators (akin to literary ones). In Section III, I undermine a key premise in the sceptics’ argument by providing a precise articulation of the role that perspectives play in our engagement with pictures, specifically arising from the way pictures fix their vantage points. The analysis shows that, despite obvious differences in the representational mechanisms used to access pictorial and literary narratives, there are important similarities. In the final Section IV, the claim about pictorial perspectives is developed to provide independent grounds for a claim about pictorial narrators. This is shown to be resilient to a range of objections.
I. Narratives and Narrators
The standard version of narrative is articulated by Gerald Prince, who says that ‘narrative is the representation of at least two real or fictive events or situations in a time sequence, neither of which presupposes each other’.3
Note that, although Prince uses the term ‘fictive events’ to describe a necessary component of narrative, we could equally well use the more general terms ‘events’ or ‘depicted events’. Broad consensus on Prince’s formulation has not prevented further discussion and attempts at refinement by philosophers and narratologists, for example that narrative is a particular form of explanation or that narratives string together events in a way that recounts and renders them intelligible. Of course, improved intelligibility and understanding are not sufficient for turning data into narrative. After all, an antigen test conveys information with understanding.
Emma Kafalenos, a contemporary narratologist, has argued that they are constructions of chronologically and causally ordered sequences of events.4 This comports well with work by philosopher J. D. Velleman5 and the linguistic historian Tzetan Todorov,6 who argue that narrative emerges from a represented trajectory of equilibrium to imbalance, to new equilibrium. In other words, narrative amounts to the presence of an organizing principle that provides the sense of events moving forward to an ending or closure.7 Velleman has argued that these events can be presented to an audience as prose, or in other formats, such as through musical refrains. Whether or not understanding narratives involves a distinctive role for the imagination is controversial. Walton has argued that imagination plays a distinctive role.8 Describing imagination’s role in picture-seeing, he says that in the actual world a person looks at a flat piece of canvas covered with pigment, while ‘one imagines of one’s seeing of the canvas to be a seeing of a mill’.9 I pick up on this later in the paper and explore an alternative construal of the way imagination is implicated in depictive seeing.
There is interdisciplinary agreement that all literary narratives are created by someone, that is, a flesh-and-blood author.10 Call this the ‘actual author’. Some theorists hold that, while actual authors are metaphysically unproblematic, they are irrelevant to the way readers access the narrative.11 These theorists posit a further implied author or narrator who provides access to the narrative. As early as Book Three of The Republic,12 Plato argues that any narrative requires an intermediary to provide access to the story. More recently, this intermediary has been identified as the narrating first-person ‘I’. The reader is a second-person ‘you’ and the being or object narrated about a third-person ‘they’ or ‘it’.13 The narrator ‘recounts’ the fabula or story (the chronologically ordered sequences of events)14 with an emotional lean that intimates the causes of the unfolding sequence.15 In other words, the fabula is the temporal order in which events occur, and the sjuzhet or version is the specific way in which the sequence is told to the reader. That fabula and sjuzhet can differ considerably is exemplified by novels such as Time’s Arrow and films such as Memento, where the narrative unfolds or is told backwards, yet time is understood to move forwards within the narrative world as it does in the real world. Just as two witnesses may recount two different versions of events, so different narrators may elicit different versions of a fabula.16
Literary narrators are identified using deictic terms (‘now’, ‘here’, ‘yesterday’, ‘tomorrow’, and so on) that fix their spatio-temporal location. Meanwhile, modal terms (‘perhaps’, ‘unfortunately’, ‘clearly’, and so forth) indicate their psychological outlook. This psychologically infused spatio-temporal tracking of the narrator ensures the version’s intelligibility. The relevance of this to the pictorial case is made clear in Sections III and IV.
Although many narratives support multiple narrators, as it is in Orman Pamuk’s novel Red, there can be just one who tells us the tale from beginning to end, as it is with Dr Watson narrating the Holmes mysteries. When there is no explicit reference to a narrator, experts tend to agree that the narrator persists as an ‘effaced’ or implicated ‘I’.17 Philosophers debate whether effaced narrators are (1) implied by the nature of our engagement with fictional narratives,18 (2) required to distinguish narrated worlds from the actual world,19 or (3) to bridge an ontological gap between perceiver and narrated world by providing access for the former to the latter.20 We can remain neutral regarding these finer-grained worries. Noël Carroll, unusually, has rejected the notion of the explicit narrator, preferring to argue that the reader simply follows instructions reconstructed from the intentions of the actual author.21 This solution does not entirely eradicate a consciousness of narratorial voice. Instead, actual or implied authors stand as effaced narrators when there are no explicit character narrators. In Section II, I consider an argument from Robinson for pictorial narrators along similar lines.
The interaction between narrators and characters is relevant to the ensuing discussion and so worth mentioning here. In Mimesis and Make Believe, Walton analyses the complex way in which a character’s voice and mood are represented in tandem with a narrator’s.22 His analysis speaks to Gérard Genette’s distinction between a character who is causally efficacious in the version and the narrator’s meta-outlook, which accommodates and elevates character so that it temporarily appears to dominate the recounting. They agree that the narrator’s point of view is always dominant, whether or not it appears to be so.23 George Wilson agrees but puts things this way: versions always subtend to the narrator’s location since we must go through that in order to access the narrative.24 From this we can say that when there is no explicit narrator the question as to the narrator’s particular identity will simply not arise for the reader. All this is pertinent to the arguments given in Section IV.
With the literary model in place, we can now ask: does this model smoothly extend to pictures? Experts deny that this is self-evident. Narratives require sequence and duration, and static pictures are not obvious candidates for manifesting such temporality. I shall not aim to give a comprehensive survey of recent work on the metaphysics of the debate. Instead, I will sketch out the two main positions so that those ideas that I take to be of fundamental importance to the argument over narrators can be introduced. This will allow us to briefly consider, and then put to one side, the related question of temporality in pictures.
Compatibilists hold that depictions are compatible with narratives. This is the prevailing view among narratologists, but disagreements arise regarding the strength of the claim.25 Strong compatibilists hold that a single (monochronic) picture can imply a narrative by a single depiction of a scene.26 Roughly, the depicted content is understood in the light of a procession of events. For example, in Jan van Eyck’s The Arnolfini Portrait, the viewer understands the man to be raising his hand to greet the couple entering the room, that is, as a duration of experience. But weak compatibilists complain that only single pictures depicting disparate events and persons in the same composition can manifest continuous narratives, such as in Sandro Botticelli’s Three Miracles of Saint Zenobius.27 Some theorists remain unconvinced that single pictures can imply duration, and insist that narrative is manifested from a series of connected compositions, as it is with Hogarth’s Marriage A-la-Mode.28 The outright rejection of pictorial narrative is relatively rare, although not entirely unknown.29 Paul Barolsky’s scepticism is informed by his belief that, since pictures lack temporal ordering, any suggested similitude with literary narratives is misleading.30 A convincing rebuttal to Barolsky’s radical view is found in Kafalenos’s Implications of Narrative in Painting and Photography. In the light of this, I will bracket out further debate on the issue of temporal ordering in pictures for discussion elsewhere.
Obviously, the account of compatibilism used as a springboard for the investigation in this paper must be properly neutral regarding the existence of a ‘pictorial narrator’. Unfortunately, this makes the suggestion from Bence Nanay, one of only a handful of philosophers to address the point in print, problematic. He argues that ‘[a] picture is a narrative picture if and only if a suitably informed spectator is supposed to undergo an experience of “engaging with narrative”’.31 Relying on a symmetry claim along the lines Nanay suggests will do more than cover both pictorial and literary cases. By explaining away the apparent asymmetry in the way the verbal and visual mediums are time-sliced and time-sequenced, we end up getting pictorial narrators ‘for free’.
For this reason, we should prefer Klaus Speidel’s ordinary language argument, which comports well with the paradigmatic literary formulation and makes no assumptions about the ontology of pictorial narrators.32 Speidel takes ordinary language to be a valuable indicator of the extension of concepts and useful for dissolving philosophical problems, specifically in situations when ‘use in the language’ is essential for the meaning of a word. Since the term ‘narrative’ is not used in primarily theoretical domains (such as ‘quark’ is in particle physics), we should be confident that the word refers to whatever it picks out in ordinary discourse. This more neutral presentation of the case for pictorial narratives enables us to confidently adjudicate between compatibilism and incompatibilism. We can assume the premise that there are pictorial narratives and turn our attention to the claim about pictorial narrators.
II. The Arguments against and for Pictorial Narrators
Are there pictorial narrators akin to literary narrators? Kendall Walton denies that there is a ‘counterpart to a narrator in depiction’ since ‘one’s access to the fictional world is not mediated by another (fictional) person’.33 For example, when paintings convey the dementing of souls in the bowels of hell (Hieronymus Bosch), or a fight between two suitors (Hogarth), the viewer does not go through a narratorial ‘voice’ to see events as unfurling. The painting hangs there, we gaze into it, and narrative implicatures follow.
Walton’s scepticism is double-faceted. The first facet is phenomenal: we do not have a sense of going through another fictional person in order to see for example, the chaotic and endless dementing of souls. The second facet is epistemic: there is no logical need to posit a narrator to ‘mediate’ and enable access to the pictorial world, and through whose perspective the chaotic and endless dementing of souls is (in a sense to be worked out) recounted or reported to us the viewer. Reconstructed, the argument is as follows:
Scepticism
The only reason for positing a narrator is if it would be impossible to access the narrative directly, that is, without a narrator.
In the case of depiction, it is always possible to access the narrative directly, that is, without a narrator.
Therefore, there is no reason to posit pictorial narrators.
Some theorists are friendlier to the possibility of pictorial narrators. Erie Watkins has pointed out that there is a sustained pictorial tradition of depicting characters in narrative pictures who ‘break the fourth wall’.34 Watkins continues to say that these depicted characters function in similar and significant ways to an explicit literary narrator. For example, in Pieter Bruegel’s The Peasant and the Birdnester a figure is placed between the scene and the viewer. This figure establishes eye contact with, and directs, the viewer’s attention toward a human figure climbing the tree in the background. Further examples of fourth-wall-breaker narrators include Nicolaes Maes’s An Eavesdropper with a Woman Scolding and the frank gaze of Manet’s titular figure in Olympia, who has a glint of amusement in her eyes at the viewer’s presupposed discomfort.35
Walton can respond to this by pointing out that we do not ‘go through’ these depicted narrator’s eyes to see the action. Rather, these characters function like arrows in a PowerPoint presentation, ensuring the viewer attends to the relevant part of the picture. So, we should treat these as depicted characters whose ‘voices’ are temporarily dominant.
Watkins can insist that these depicted figures do qualify as pictorial narrators since we miss elements of the story when we do not pay attention to what these characters endeavour to show us. Furthermore, while we may not see the pictorial world ‘through the eyes’ of the depicted figure, we do see the pictorial world afresh by the depicted character indicating what we should prioritize or make salient in our grasp of the work.
The problem is that Watkins overclaims on this point by saying that ‘the location of those persons seems to coincide with our own [and so] we are left with the illusion that we see the depicted world from their point of view’.36
The phenomenology as described by Watkins is false if the suggestion is that we have a trompe-l’oeil kind of experience with these fourth wall breakers. We do not see the Bruegel painting or Édouard Manet’s Olympia from the spatial perspective of the peasant or the courtesan. This leaves his claim vulnerable to the accusation that he has merely identified elevated and dominant characters.
To justify the claim about narrators, what needs to be shown is that pictures can and do instantiate the ‘narrating I’ recalled from the earlier section. These are distinguished from elevated characters and belong, however peripherally, to the world of the fabula, regardless of how infrequent the instances of self-reference.
Jenefer Robinson’s approach shows more promise in this respect, although we need to reconstruct her view since it sits primarily in the debate concerning pictorial expressiveness rather than narration. Robinson has argued for a class of implied artists one step removed from the actual artist psychologically and conceptually.37 Implied artists are said to cohere and unify what a picture conveys, leaving a psychological trace of the actual artist who created them in the pictorial world.38 Viewers are said to experience the picture from their point of view or perspective.
There have been several criticisms of Robinson’s notion in print and Robinson has conceded the point on several of them.39 These criticisms damage the model independently, but a further concern is that Robinson’s implied painter ends up in the wrong location, that is, in the actual world rather than the picture world. This means that the pictorial narrator qua implied painter is merely augmenting rather than providing access to the narrative. It would require further argumentation to secure the claim that the expression of personality, mood, or psychological perspective in a picture is sufficient for the presence of a narrator, and it is not possible to reconstruct this from Robinson’s papers.
One way to advance the central claim is to challenge premise three of Scepticism, and to show that it is not possible to access pictorial content directly. This will establish the possibility of a mediating gateway or object that enables a viewer to adequately apprehend a picture. This is the work of the final two sections.
III. Pictorial Perspectives
In this section I clarify how perspectives play their role in depictive seeing. This clarification is put to work in the final section to defend the claim about pictorial narrators. So far, we have compared literary narrative to pictorial narrative to elucidate the representational structure of the latter. Here I begin by asking a more basic question: how do we see?
Suppose I look out of the window and see my neighbours in the street. It is uncontroversial that whatever x I am seeing, I am seeing x from my actual perspective. This is because, when I see, what I see is seen from the spatio-temporal location I currently occupy. Perspectives, while contentless in themselves, structure the way content (x) is represented. My actual perspective is from ‘here and now’ and it conduces to represent the neighbours as ‘over there, now’ in relation to ‘me’. A perspective determines what aspect or which bit of x is most prominent or salient within the visual field. For instance, I only see the facing parts of the neighbours not currently obscured by other parts of them, or other objects in the street. And, unless the neighbours move, I will need to change my actual perspective on them, perhaps by moving to another window, to see different bits of them. In summary, when we see, we do so from the actual perspective that we occupy.
But, when we see depictions, what we see is seen from a spatio-temporal location implied by the facing parts of the objects-as-depicted. We see the people from above, the chair from the right, and the café from the street to the left. One might say that seeing the depicted contents in pictorial space is like peering through a hole of a box, which acts as a tiny window onto the depicted scenario. But this misses out a significant fact about the way the perspective plays its role. Typically, we are looking at the painting about a metre or so back from it. We are not peering through a pinhole to see the picture. How then do we align the actual perspective we have on the painting (qua canvas hanging on the wall) with the pictorial perspective on the depicted content? We must be simultaneously co-ordinating the two types of perspectives, actual and imagined. Because, unlike the actual perspective, which is spatial (here) and temporal (now) and sensitive to changes in one’s location, the pictorial perspective can be from the left, the right, above, or below, as implied by the facing parts of the depicted objects. It is insensitive to small changes in the viewing subject’s location and often turns out to be non-coincident with their actual perspective. How does the viewer ‘see’ from this perspective? The answer is that they must represent the perspective in their imaginations. They imagine a spatio-temporal perspective from which the marks they see on the canvas makes sense.
This clarification about the seeing appropriate to pictures (and paintings in particular) is consistent with the phenomenon of so-called ‘twofoldness’.40 That is, of attending to both the configuration of the surface (from the actual perspective) and the recognition of objects (from the pictorial perspective). These two aspects, or folds, are said to permit simultaneous perception.41 The viewer manages to sustain seeing the brushstrokes as brushstrokes, while at once having a sense of the pictorial depth they create.42
All this provides grounds for challenging the consensus in the literature that we ‘directly’ see depictions, that is, that we clap eyes on depicted objects just as we clap eyes on the objects they depict. The consensus ignores how the perspectives involved in seeing and pictorial seeing play their roles differently, and that, while depictive seeing is perspectival, it is also essentially indirect. In other words, all depictive seeing requires the viewer to represent an imagined perspective that mediates access to the depicted object.
Interestingly, Walton has also commented on this:
What is it for a depiction to depict things from a certain point of view? One sense is obvious. The point of view consists in the perspective from which, fictionally, we perceive when we examine the depiction. Fictionally we see Hobbema’s red-roofed mill from a point a couple of hundred yards downstream on the left bank of the river.43
But he overlooks the implications of acknowledging non-actual perspectives and so wrongly claims that we directly see pictorial contents. Pictorial seeing is like visualizing in respect of perspective-taking, since we need to enter the fictional world from a vantage point prescribed by the marks, rather than our actual spatio-temporal location. Depicted objects are stubbornly presented from a prescribed perspective we must shift into imaginatively.44
This only goes as far as to show that seeing depictions requires us to represent a pictorial perspective. To be clear, I am emphatically not saying that this mediating pictorial perspective entails pictorial narrators. I do not think that when we represent a pictorial perspective we must also represent someone occupying it. We can represent events from a so-called ‘bare pair of eyes’ perspective, which is disassociated or vacant.
This is how we see an IKEA drawing. When visualizing the neighbours seen through the window, one does not also have to imagine oneself seeing as part of the visualized scene. In just the same way, seeing a depicted scene does not imply imagining an inhabitant occupying the pictorial perspective.
In the final section I show how this clarification regarding the pictorial perspective can help mitigate the puzzle of pictorial narrators.
IV. Pictorial Narrators
The existential claim about pictorial narrators needs to show that some paintings sustain a consciousness of narratorial voice. A more ambitious formulation will insist that a narrator (be it effaced, unreliable, explicit, or so on) is required for every pictorial narrative. Building on the account of pictorial perspectives outlined in Section III, I will defend their possibility. In the final paragraph I sketch out what is needed to satisfy an ambitious formulation.
The sceptic argues that we do not need pictorial narrators for the same reason we do not need narrators when we look out of the window. When I look out of the window and happen to see my neighbours locked in a romantic clinch, I do not need an interpreter to recount events for me. Likewise, when I look at Gustav Klimt’s The Kiss, I see the passion and intimacy between the lovers directly.
We can reply to the sceptic’s ‘null response’ by weighing the earlier point about accessing depictions. Looking out of a window frame is not like looking into a picture frame. To adequately model pictorial space we need a mediating pictorial perspective.
The sceptic can grant this point but persist with a null response. He can say that there are always cases where the depicted facts will underdetermine how one interprets the picture. We must ‘fill in’ what we can see with extra-pictorial information. This could include instructions about perspective or psychological framing that we apply to what we see. So, we are not impelled to represent narrators just to get ‘versions’ of the pictured narrative. We can fill in facts about the narratorial outlook by adverting to the ‘viewer’s own take’ or ‘the artist’s intention’. In this regard, seeing a picture that expresses more than depictive facts, such as a psychological perspective on events, is relevantly like imagining the neighbours romantically entangled from a vacant perspective. This is how a street camera would capture events and it would be peculiar to think that that picture was narrated.
This way of arguing ignores a significant phenomenal aspect of pictorially seeing that comes about because of the viewer’s awareness that they are looking at a picture. This is a phenomenon that illusion lacks. Suppose I take myself to be looking through a window frame when in fact I am looking at a sophisticated trompe-l’oeil that shifts with each small incremental movement of my eyes. I am looking at a picture, but I mistakenly believe I am actually seeing. Because of this, I do not have the distinctive twofolded split perception. Instead, the prescribed pictorial perspective remains coincident with the actual perspective, resulting in an illusory experience of seeing. The sceptic is conflating this kind of bad illusion case with a ‘good’ case of pictorial seeing. But only the good case reveals a location the narrator can inhabit.
It seems reasonable to predict that narrative paintings will have a strong emotional lean. This presents an opportunity to test a theory of occupied perspectives without having to deal with the full complexity of narratives. We can use paintings that merely have the strong emotional content, such as a sentimental painting by John Everett Millais or a menacing Caravaggio. When it comes to seeing the emotional ‘air’ of the picture (as opposed to a depicted expressing figure), viewers are looking at an expressive take on the pictorial world. This emotional lean makes additional demands on the viewer. To meet the demand the viewer represents both a perspective and an expressing ‘mind’ from whose perspective events are taken to be sentimental, menacing, and so on. This representation of a mediating ‘mind’ is crucial to the way the viewer also models more complex expressiveness found in narratives. For this reason, we need to spend a bit more time unpacking the phenomenon.
Expressive pictorial seeing requires the viewer to represent some additional perspectival features. The difference is grasped by thinking about what a bored and lonely person looks like from the outside and then comparing this to what it is like to experience boredom and loneliness from the inside. In the first case, a figure may appear bored or lonely by being depicted as slumped in a chair, or disengaged from other people nearby, and so on. The viewer apprehends this boredom ‘from the outside’ or from a third-person perspective. Contrast this to a second case, where a landscape or scene may appear desolate or unfriendly. Here the painting appears analogous, or in some way continuous with undergoing the emotion, from the inside. The two different perspectives can of course be combined in a single experience of a painting. For instance, Edward Hopper’s Nighthawks depicts bored and lonely people from a melancholy pictorial perspective. This second perspective pervades, stains, or guilds the pictorial world with melancholy, which means that the viewer represents the depicted content as congruent with a first-person subjective experience of melancholy. Unlike the perspective on the depicted figures, which provides clues as to what is happening (the people are sitting in the café), the perspective ‘from the inside’ is not so much a clue by which we go on to tell what is expressed. Rather, by constituting this perspective using our capacity for sensory imagination, the melancholy is revealed. In other words, by representing melancholy from the inside the melancholy perspective becomes an ineliminable part of the expressive telling.
How does the viewer represent a melancholy perspective ‘from the inside’? The viewer represents a self-reflexive perspective, which is just to say that the viewer represents a pictorial perspective and in addition represents it as occupied by something feeling melancholy.
A robust realist will disagree. They will say that Nighthawks represents a melancholy take on the café but insist this is explained away by a thicker conception of facts. That is, the overall expressive air of a picture arises from facts about sitting alone in cafés at twilight, or some facts about associations of certain colours to certain feelings. Thus, the robust realist denies that we need to represent additional ontological objects (minded emoting occupants) to do the telling of the melancholy. The pictorial facts alone supply all we need to interpret the picture.
However, such an expansive conception of facts about emotions is highly controversial. To my mind, I am worse off with a commitment to thick realism than I am with a thin conception of facts that draws on what we already know about pictorial perspectives and sensorily imagining. Sensorily imagining is egocentrically or self-reflexively seeing in the mind’s eye. In other words, to visualize a scene is to see it from an inside perspective.45
Despite this, a proponent of the ‘pictorial perspective plus sensory imagining’ may still question the need for the viewer to represent an emoting occupant, who stands as a consciousness of emotional voice, generating the emotional lean in the painting. Some sensory imagining, they can point out, occurs from a bare pair of eyes or vacant perspective. For instance, I can close my eyes and sensorily imagine a red building, without having to also imagine someone seeing the red building. Since we do not posit ‘minds’ to apprehend subjective qualities like colours, we do not need them to posit properties like melancholy.
Here is a reply to that concern. There is a difference in what it is like to represent a perspective on a cube, a red building, and a sad street. The difference is found in the kinds of properties these objects instantiate. A cube instantiates a ‘mind-independent property’ – for example, spatial extension. Whether or not an object is cuboid is not determined by how it looks to an observer. The conditions on something being cuboid are not anthropocentric. In contrast, the red building instantiates a ‘response-dependent property’; that is, the story about the redness of the building is going to have to mention the world and human perceptual systems. This, however, does not entail that whether the building is in fact red is determined by one’s perceptual experience. The conditions for an object being red do not depend on how the object appears to Smith or Jones. Mental states (such as sadness), however, are essentially mind-dependent. That is, they simply are properties of mind. While there is something it is like to represent a perspective on ‘external’ mind-independent, response-dependent, and ‘internal’ mind-dependent properties, what it is like varies. When the viewer represents an emotional perspective, they represent the pictorial contents as if contained within the mind of an expressing thing. For this reason, the viewer must represent the expressing thing at the origin of the pictorial perspective containing the pictorial world. More simply, the ‘melancholy’ thing mediates the viewers access to melancholy Nighthawks.
This model satisfies the conditions for a self-locating perspective. The perspective points forward to the visual scene and backwards (on itself) to the psychological outlook. It is in play whenever we experience a painting as expressive. It explains how viewers emotionally ‘stain and guild’ the pictured state of affairs. Given that the expresser is represented as occupying the pictorial perspective, and so prescribed by the picture, it is also part of the pictorial world.46
Of course, mere emotionally ambient works are not narratives. The veil of melancholy that hangs over Hopper’s Nighthawks does not automatically qualify it as a narrative picture, since the melancholy does not effect the representation of causality in the narrative. That is, the expression is of a single emotional note, and what we need to satisfy narrative is a represented trajectory of equilibrium to imbalance to new equilibrium.47 Without this, we do not have pictorial narrators but pictorial expressers. A consciousness of narratorial voice emerges with a version of at least two events that are intimated to be this sort of relationship with each other. However, pictorial expressers provide a strong link to narrators. This is because narrative can arise from organizing emotional refrains that move forward to the new equilibrium or closure.48 Importantly, these refrains do not have to be stored in the technology of writing, or even language-based modes of thought. They can be stored ‘in proprioceptive, and kinaesthetic memory […] in the muscle-memory of the heart’.49 As we have seen, they can also be stored in expressive pictures. To reveal them the viewer represents an inhabited perspective. Inhabited that is, by the kind of thing that has the same affective capacities as we do and through whose emotional outlook the pictorial cadence is revealed. Where there is only one emotional perspective that pervades and endures in a picture, there is an expression. But where there are two or more that are experienced as a sequence with events moving forward, there is a narrative. And in these circumstances the expresser acts as a narrator.
The sceptic still has one further objection to push: the author response. They can argue that we do not need a distinct ontological kind to account for the way we see the picture through the eyes of another. Rather, the viewer simply adverts back to the artist and imputes the emotional perspective to them. One can interpret Marcel Proust as advocating something like this when he encourages the reader to think of their experience of Jean Siméon Chardin’s still life as a ‘journey of initiation into the unknown life within the still life, which each of us can make if we let Chardin be our guide, as Dante was guided to Virgil’.50
What Proust intimates is that we might explain our ability to engage with a picture without representing an effaced narrator so much as the flesh-and-blood artist who is brought to mind while looking at the work.
The problem here is that the expressive or quasi-narratorial outlooks are imputed to a location that is, in many senses, more difficult to account for than an imagined occupant, for example when the flesh-and-blood artist is deceased, or paints something that is inconsistent with their actual outlook. The author response objection seems to accept that we need a psychological intermediary to convey the affective hue of ‘things in their most profound aspect’.51 Yet it resists the reasonable idea that its intermediary is not actually Chardin but merely our notion of a Chardinian type that effects the way things seem to be in the picture.
The idea of Chardin-like occupants also comports well with Proust’s elaboration of the point that ‘you yourself will be a Chardin’.52 Proust does not say you will be Chardin – merely that you must represent a Chardin-like perspective.
This model for basic expression can now be applied to three established types of narratorial pictures. Single monoscenic pictures such as Giorgio de Chirico’s Mystery and Melancholy of a Street, single pictures that deconstruct the implicature such as Botticelli’s Saint Zenobius, and serial pictures, such as Hogarth’s Marriage A-La-Mode. The mediating pictorial narrator model is best placed to explain how it is that we can refer to some overarching perspective that harbours and embeds whatever explicit character or figure expressions make up the content of the narrative. For instance, in De Chirico’s painting, a young girl is depicted happily playing with a hoop, but despite her cheerfulness the atmosphere is foreboding. A viewer can note this while having their own independent emotional response to this narrative. They might giggle inappropriately. Their giggling does not obstruct them from seeing the foreboding narrative. When one accepts the notion of pictorial narrators, the possibility of making sense of this kind of emotional layering is accommodated. The model meets the demands of this complex pictorial space even when it might be unclear to the viewer themselves which perspective has priority, or how it attains priority.
Can this formulation be used to justify the more ambitious claim that, whenever pictures convey narratives, they do so with a narrator? In other words, is it a condition of the medium? Or do different styles or traditions of depiction lay out narratives differently? If it can be so used, then I have not done enough here to show it. The answer will depend on whether we think the model for pictorial seeing comports smoothly across all styles and traditions of depiction. I leave that work for another time.
Pictorial narratives do more than convey information. They can recount events in a way that renders them dramatically intelligible, for example by conveying psychologically infused perspectives. I have been arguing that they may do this by prescribing the viewer to represent pictorial perspectives inhabited by the kind of psychological beings that can express emotions. When the viewer represents this kind of perspective, the perspective acts as a gateway to content concerning ‘what it is like’ to feel a particular way towards events. In other words, viewers represent a perspective on the picture, and additionally represent something that has the psychological point of view occupying that perspective. Given the analysis of the expressive case, it was concluded that pictures could also prescribe viewers to represent a narrating consciousness, who mediates and frames how the marks are seen.
Notes
[1] Michael Ranta, ‘Stories in Pictures (and Non-pictorial Objects): A Narratological and Cognitive Psychological Approach’, Contemporary Aesthetics 9 (2011), http://hdl.handle.net/2027/spo.7523862.0009.006.
[2] Kendall Walton, ‘Points of View in Narrative and Depictive Representation’, Noûs 10 (1976): 50.
[3] Gerald Prince, Narratology: The Form and Function of Narrative (Berlin: De Gruyter, 1982), 4.
[4] Emma Kafalenos, ‘Implications of Narrative in Painting and Photography’, New Novel Review 3 (1996): 54.
[5] John David Velleman, ‘Narrative Explanation’, Philosophical Review 112 (2003): 1–25.
[6] Tzetan Todorov, ‘La grammaire du récit’, Langages 3 (1968): 94–102.
[8] Kendall Walton, Mimesis as Make-Believe: On the Foundations of the Representational Arts (Cambridge, MA: Harvard University Press, 1990).
[10] Walton maintains that something can be a representation regardless of the intentions of its maker, but his view is an exception (ibid., 52).
[11] See, for example, Jenefer Robinson, Deeper than Reason: Emotion and Its Role in Literature, Music, and Art (Oxford: Oxford University Press, 2005) and for criticisms see Vanessa Brassey, ‘The Implied Painter’, Debates in Aesthetics 14 (2019): 15–29; ‘The Expression of Emotions in Pictures’, Philosophy Compass 16 (2021), https://doi.org/10.1111/phc3.12767.
[15] Manfred Jahn, Narratology 2.3: A Guide to the Theory of Narrative Narratology (Cologne: University of Cologne, 2021), http://www.uni-koeln.de/~ame02/pppn.pdf.
[18] Gregory Currie, Narratives and Narrators: A Philosophy of Stories (Oxford: Oxford University Press, 2010).
[19] David Lewis, ‘Truth in Fiction’, American Philosophical Quarterly 15 (1978): 37–46.
[20] See Jerrold Levinson, Contemplating Art: Essays in Aesthetics (Oxford: Oxford University Press, 1996); Seymour Chatman, Coming to Terms: The Rhetoric of Narrative in Fiction and Film (Ithaca, NY: Cornell University Press, 1990).
[21] Noël Carroll, ‘Introduction’, Journal of Aesthetics and Art Criticism 67 (2009): 1–3.
[23] Gérard Genette, Narrative Discourse: An Essay in Method, trans. Jane E. Lewin (Ithaca, NY: Cornell University Press, 1983), 146.
[24] George Wilson, Seeing Fictions in Film: The Epistemology of Movies (Oxford: Oxford University Press, 2011).
[25] See Kafalenos, ‘Implications of Narrative’; Wendy Steiner, Pictures of Romance: Form against Context in Painting and Literature (Chicago: University of Chicago Press, 1988); ‘Pictorial Narrativity’, in Narrative across Media: The Languages of Storytelling, ed. Marie-Laure Ryan (Lincoln: University of Nebraska Press, 2004), 145–77; Michael Ranta, ‘(Re-)Creating Order: Narrativity and Implied World Views in Pictures’, Storyworlds 5 (2013): 1–30; Klaus Speidel, ‘Can a Single Still Picture Tell a Story? Definitions of Narrative and the Alleged Problem of Time with Single Still Pictures’, Diegesis 2 (2013): 173–94.
[26] See Kafalenos, ‘Implications of Narrativity’, Ranta, ‘(Re-)Creating Order’, and Speidel, ‘Can a Single Still Picture Tell a Story?’
[27] Werner Wolf, ‘Narrative and Narrativity: A Narratological Reconceptualization and Its Applicability to the Visual Arts’, Word and Image 19 (2003): 180–97.
[29] See Paul Barolsky, ‘There Is No Such Thing as Narrative Art’, Arion 18 (2010): 49–62, and Paul Harrison, ‘The Limits of Twofoldness’ in Richard Wollheim on the Art of Painting: Art as Representation and Expression, ed. Rob van Gerwen (Cambridge: Cambridge University Press, 2000), 39–59.
[31] Bence Nanay, ‘Narrative Pictures’, Journal of Aesthetics and Art Criticism 67 (2009): 127.
[34] Eric Watkins, ‘Point of View in Depictive Representation’, Noûs 13 (1979): 379–84.
[38] I am conflating a conceptual question concerning pictorial narrators with the debate in respect of implied or hypothetical painters here. Robinson’s model has been extensively criticized; see Dominic McIver Lopes, ‘The “Air” of Pictures’, in Sight and Sensibility: Evaluating Pictures (Oxford: Clarendon, 2005), 49–90; Brassey, ‘Implied Painter’.
[39] See Lopes, ‘“Air” of Pictures’;, ‘Implied Painter’; and Jenefer Robinson, ‘Response to Critics’, Debates in Aesthetics 14 (2018), https://debatesinaesthetics.org/response-to-critics.
[40] ‘Twofoldness’ is a term of art due to Richard Wollheim, Art and Its Objects: With Six Supplementary Essays (Cambridge: Cambridge University Press, 1980).
[42] Richard Wollheim, Painting as an Art (London: Thames and Hudson, 1987), 62.
[45] Robert Hopkins, Picture, Image and Experience: A Philosophical Enquiry (Cambridge: Cambridge University Press, 1988).
[46] There are further questions concerning the identity of the occupant, for instance whether the model endorses particularity or specificity. But I take these issues to be beyond the scope of this paper to sort out.
[50] Marcel Proust, Chardin and Rembrandt, trans. Jenny Feldman (New York: Zwirner, 2016), 22.
Competing Interests
The author has no competing interests to declare.