- Split View
-
Views
-
Cite
Cite
Alon Chasid, Pictorial Experience and Intentionalism, The Journal of Aesthetics and Art Criticism, Volume 72, Issue 4, November 2014, Pages 405–416, https://doi.org/10.1111/jaac.12106
- Share Icon Share
I. INTENTIONALISM
Inasmuch as intentionalism is popular in the philosophy of perception, its underlying motivations merit thorough discussion, as do its criticisms of other views. Here, however, I focus on three tenets of intentionalism that are of particular relevance to the pictorial experience thesis (PET).
First, intentionalists claim that visual experiences are intentional mental states, or mental representations. Much like beliefs or thoughts, visual experiences have, they argue, correctness conditions (or content) and can be either veridical or nonveridical. Intentionalists thus hold that illusions, hallucinations, and other nonveridical experiences are the same kind of mental representations as veridical experiences: they are characterized by representational and phenomenal properties.1 Specifically, the representational properties of an experience are the properties it represents an object as having—whether or not the represented object exists or really has those properties. As Tim Crane claims, perceptual experience “is not really openness to the world, in the sense that it does not essentially involve a relation to its objects.”2 So, for instance, when we look at a hologram, the correctness conditions of the experience may not refer to the existence of a hologram, but, rather, they may refer to propositions describing what the hologram depicts: that there is a tropical forest, a unicorn, and so on.
A second point is that intentionalists, unlike doxastic (or belief acquisition) theorists of perception, hold that visual experiences and beliefs are different kinds of representations.3 We do not always believe what we visually experience: we can believe, and persist in believing, that what we are visually experiencing does not exist, or exists but is unlike what we experience. Famous cases of cognitive impenetrability illustrate this point. In the Müller Lyer illusion or in everyday experiences that do not correctly represent certain shades and shapes, we may know very well that our experience is nonveridical, though this knowledge does not thwart our experience.
A third intentionalist thesis is that the phenomenology of a visual experience supervenes on the experience's representational content. That is to say, any difference in what it is like to experience an object necessarily entails a difference in the properties that the object is experienced as having. The supervenience thesis (henceforth ST) guarantees that the representational content of experience determines its phenomenal character or that no phenomenal properties are nonintentional.
Various counterexamples to ST have been suggested. I will show that the case of indeterminate pictures, as interpreted by PET compatible accounts of depiction, is also a counterexample to ST. Hence, if ST is theoretically essential for intentionalists, they will have to abandon PET; those who insist on retaining PET will have to adopt a nonintentionalist notion of perceptual experience.
II. VISUAL ACCOUNTS OF DEPICTION
Looking at a picture of, say, a horse, we may visually experience a surface covered with marks. Yet on certain accounts of depiction—accounts I will call visual accounts of depiction—we may also visually experience a horse. These accounts accept PET, namely, the thesis that pictures can generate a visual experience of what they depict.
Visual accounts of depiction are not unanimous as to whether the perceiver can visually experience the picture's marked surface and the depictum at the same time. They can thus be divided into two groups: accounts on which the experience of the depictum and the experience of the marked surface cannot occur simultaneously (that is, the two experiences are exclusive, henceforth exclusionary accounts) and accounts on which the two experiences can occur simultaneously (that is, there is a twofold experience, henceforth twofoldness accounts).4
Famously, Ernst Gombrich argued that in looking at a picture, we visually experience the depictum, but we cannot experience the picture's surface at the same time (the experiences are exclusive, much like the two experiences in Jastrow's duck–rabbit figure).5 Eddy Zemach similarly defended the thesis that the two experiences cannot occur simultaneously.6 Kendall Walton's theory of depiction can, I contend, also be interpreted as holding that the two experiences are exclusive.7 Although Walton himself mentioned that his theory “goes some way toward showing how two different intentional contents [that is, surface and depictum] can be combined,” scrutiny of his account seems to lead to a different conclusion.8 Walton argues that pictorial experience consists in imagining seeing the depicted object, where imagining seeing is explained on the basis of cognitive penetration: the imagination, which somehow involves the depicted object, penetrates perceptual experience and colors the actual act of seeing. Thus, the cognitive process that starts in seeing a marked surface culminates in a (nonveridical) mental representation, namely, a visual experience of the depictum.9
Walton's understanding of photographs can also be considered an exclusionary account. Walton claims that photographs are transparent in the sense that we literally see an object through a photograph of that object.10 It could, perhaps, be argued that when we experience the pictured object, we also see the photograph itself—that is, in some externally characterized sense of seeing. Recall, however, that intentionalists take visual experiences to be intentional states and characterize them by their phenomenal and representational character (that is, not by the actual relation that holds between the perceiver and the object being looked at). If the intentionalist account is true and if photographs are indeed transparent, Walton's position might be that the picture's surface is not visually represented when the photographed object is represented.
Accounts of twofoldness have been put forward by Richard Wollheim and, more recently, Dominic Lopes, Bence Nanay, Michael Newall, and others.11 Wollheim argues that in looking at a picture, we undergo an experience with two different aspects, one configurational and the other recognitional: we are simultaneously “visually aware” of both the picture's configuration and the depictum. Wollheim initially claimed that twofoldness involves two distinct experiences but subsequently changed his mind and took pictorial experience to be “a single experience with two aspects.”12 The difference between Wollheim's early and later accounts of twofoldness is not important here. Given that both versions entail that we visually experience the depicted object and the surface simultaneously, both are relevant to my argument.
Lopes, too, puts forward a twofoldness account. He argues that in at least some cases—cases that, as I will explain later, are particularly relevant to my argument—we are able “to see a picture's content and its design simultaneously.”13 Specifically, Lopes characterizes “design seeing” as a “visual experience of … marks, colours, and textures in virtue of which the surface depicts a scene.”14 Indeed, being part of the twofold experience, design seeing is what makes pictorial experience different from seeing the depicted object face to face, as pictorial experience is “inflected” by seeing the picture's design.15
Elaborating on Lopes's idea of design seeing, Nanay claims that when seeing in is inflected, we visually experience the picture's design and the depicted scene simultaneously by consciously attending to a relational property, a property that “cannot be fully characterized without reference to both the picture's design and the depicted object.”16 However, he also argues that “design shows up only in the ‘depicted object’ fold of our twofold experience, not in the ‘surface’ fold.”17 Nanay's full account of twofoldness invokes the subpersonal distinction between dorsal and ventral perception. What matters vis à vis my focus here—namely, the compatibility of PET and intentionalism—is what he takes to be represented ventrally, that is, what figures in our conscious visual experience. I will not discuss Nanay's account in detail, but it might be that an intentionalist interpretation of his account should consider it exclusionary, and not a twofoldness account, as both design seeing and seeing the depicted object belong to the depicted object fold of experience, whereas the other fold is characterized only dorsally, hence subpersonally.18
Relying on theories from cognitive science and psychology, Newall argues that “pictures occasion nonveridical seeing of their subject matter.”19 Seeing a picture typically involves the experience of the picture's surface, which is “somehow integrated” with the experience of seeing the depictum.20 Newall, then, accepts PET and holds that, ordinarily, we simultaneously experience the marked surface as well.
Some accounts of depiction seem to reject PET. John Hyman suggests principles of pictorial art that explain how pictorial experience is visual in a way that experiencing texts is not but does not maintain that we visually experience the depictum.21 Christopher Peacocke and Robert Hopkins contend that pictures depict by being experienced as resembling the depictum in certain respects (the specific differences between these experienced resemblance accounts are not important here).22 Hopkins does attempt to explain seeing in and the phenomenon of twofoldness. However, he does not seem to maintain that we have a visual awareness or experience of the depictum; hence his account may be compatible with ST.23
Nelson Goodman and John Kulvicki also deny that looking at a picture induces a visual experience of the depictum and hold that the marked surface only refers to the depictum, much like a verbal description refers to its object.24 These accounts are therefore compatible with rejection of PET.
To sum up, some accounts of depiction accept that in looking at a picture, we have a visual mental representation—that is, a visual experience—of the depicted object. These accounts disagree as to the specific structure of pictorial experience, its relation to the experience of the surface, whether and how it is inflected by experiencing the surface, and so on. The differences between these accounts may be significant. However, what matters for the following discussion is whether they espouse PET and, if so, whether the account is exclusionary or upholds twofoldness.
In what follows, I show that any account that accepts PET is incompatible with intentionalism, though the factors that generate the incompatibility may differ.
III. INDETERMINACY
At first sight, both kinds of visual account of depiction seem to go hand in hand with intentionalism. Intentionalists have no problem accepting that, in general, we can visually experience y by looking at x. Visual accounts of depiction seem to hold that, ordinarily, experiencing a depicted object is not thwarted by our knowledge that we are looking at a picture, nor does this experience delude us into believing that the depictum is in front of us or even that it exists at all. PET seems to be a straightforward consequence of intentionalism.
What about ST? On visual accounts, it too should apply to pictorial experience. Both the twofoldness and the exclusionary accounts of depiction must hold that any difference in what it is like to visually experience, say, a depicted garden entails a difference in the properties that the garden is experienced as having. However, I will now show that on both twofoldness and exclusionary accounts, pictorial indeterminacy constitutes a counterexample to ST.
Pictorial indeterminacy can be characterized in a number of ways. To facilitate examination of the compatibility of ST and PET, I will limit myself to pictorial indeterminacy that arises when we compare configurational differences between pictures to differences in what the pictures depict.25
At first sight, we might assume that any change in a picture's configuration necessarily entails a change in what the picture depicts: were the marks on the picture's surface different, the depicted object would have different properties or would not be depicted at all. Yet this is not always the case. For instance, a detailed line drawing does not necessarily differ in what it depicts from a line drawing done in a different color or line thickness or on paper of a different texture. A stippled drawing and a hatched drawing can depict the same content when both the stippling and hatching are used to depict the same shaded part of an object. A somewhat sketchy picture might differ from another with respect to certain determinate shapes that appear on their surfaces without there being any difference in the shapes the pictures depict: the two pictures may both depict only a determinable shape property despite the different determinate shapes that appear on the pictures’ surfaces. Even when it is difficult to articulate the specific configurational feature in question, we generally realize that various elements in a picture's configuration—a particular curve or angle, a degree of brightness or line thickness, or some other relatively specific configurational feature—could change without entailing a change in the properties an object is depicted as having.
Black and white pictures illustrate this point clearly: a black and white picture differs in configuration from its sepia, bluish, or reddish counterparts, yet all these pictures depict the same object as having the same properties. To put it schematically, a configurational feature F of a picture p1 is nondepictive just in case there can be a picture p2 that does not differ from p1 in what it depicts yet differs from it in not having F; otherwise, F is depictive. A picture is indeterminate just in case it has nondepictive configurational features.
It is important to note that according to this definition, nondepictive features are not properties that are included in what Lopes calls “design seeing.”26 The whole point of my characterization of pictorial indeterminacy is that we do not experience nondepictive properties as “properties in virtue of which the surface depicts a scene.”27 Design seeing encompasses only depictive properties. Thus, insofar as certain (relatively specific) properties are nondepictive, they are not taken to inflect the experience of the depictum in any important respect. Likewise, if Nanay is right, nondepictive properties are not properties that are incorporated into the experience of the depicted object.
If PET is true, the case of indeterminate pictures has serious consequences with regard to visual experience of the depictum. If we can indeed visually experience an object by looking at a picture of that object, then in looking at two indeterminate pictures that differ only in configuration, we visually experience the same depicted object as having the same properties despite the visible differences between the pictures’ respective marked surfaces. Obviously, we can visually experience the difference between the surfaces, but the point I am making here is that we also visually experience, either simultaneously or not, the same depicta—the same depicted objects and properties.
Suppose you are looking at a black and white picture of a garden. According to PET, you visually experience a garden, though ordinarily, your experience of the garden will not represent it as having gray hues: it will represent the garden as being colored or as having certain degrees of brightness without specifying any determinate hue. The same is true of the counterpart sepia picture of the garden: you will experience the garden as having some hue, but your experience of the garden will be silent about what specific hues the garden has. The marked surface's hues can change without entailing any change in the properties your experience represents the garden as having.
Likewise, the ability to visually experience a face by looking at, say, a detailed line drawing of a face shows that a visual experience can represent an object without specifying its color or even whether it is colored at all. If the face is experienced determinately, it will be experienced as transparent (that is, as having only an outline). Though we may have such an experience, we usually experience objects depicted in line drawings indeterminately: our experience does not specify the depictum's color or texture. Similarly, a face depicted in a sketch may sometimes be experienced as having a determinable round shape property. It will not necessarily be experienced as having the determinate and somewhat swiveled roundness that is featured in the sketch. Indeed, in such cases, some determinate configurational features are nondepictive. They are clearly not properties that the depicted object is experienced as having.28
These and similar examples also demonstrate that pictorial indeterminacy has significant implications for aesthetics. A picture's depictive style or system is, in part, a function of the kind of indeterminacy the picture manifests. Lopes, for instance, asserts that “systems of depiction” are individuated by being “non committal” vis à vis the types of properties that pictures depict their subjects as having.29 More specifically, a picture can be “inexplicitly non committal” with regard to a property F, namely, when the picture “does not go into the matter of F ness.”30 Pictorial indeterminacy, as defined here, is clearly a case of inexplicit noncommitment and thus is a criterion by which systems can be individuated. For instance, “drawings and paintings that follow Alberti's rules belong to different systems, because paintings are committal about color and drawings are inexplicitly non committal about color.”31
Similar consequences of pictorial indeterminacy are mentioned by Wollheim, who distinguishes between two “hows” of representation. The “representational how” refers to what is represented: “When Degas painted the Duc and Duchesse de Morbilli in an estranged pose, he represented an estranged couple, which is what he probably thought all couples to be.”32 By contrast, the “presentational how” does not qualify what is being represented, but
may reflect a range of things from the expressive vision of the artist, through the artistic pressures of the day, to the artist's technical limitations. … [W]hen West Arnhem Land aboriginees painted stick figures, they were not representing humans who were as thin as sticks. When Parmigianino painted the Madonna with a long neck, the Madonna whom he represented is not, despite the title given to his picture, a long necked Madonna. When Matisse painted a stroke of green down his wife's face, he was not representing a woman who had a green line down her face.33
Wollheim indicates that not every configurational feature is depictive of some property, and so a picture should be appreciated not just for the properties it depicts its subject as having (the “representational how”) but also for its nondepictive configurational features (the “presentational how”). The line thickness of a stick figure (as well as its color, texture, and so on) and likewise the green stroke in Matisse's painting are configurational features that do not say anything about the depictum's properties: indeed, the depictum is represented indeterminately in these respects. But being nondepictive, these specific configurational features are crucial to appreciating the pictures, as they do impart something important regarding expressiveness, technique, and so on.
Parmigianino's Madonna is an interesting example of indeterminacy. It is invoked to show how the standard of correctness—what counts as the correct way of understanding the picture (which, according to Wollheim, follows from the artist's intention)—can dictate pictorial indeterminacy, that is, dictate that the “correct” visual experience is indeterminate.34 With no knowledge of the correct way to experience the Madonna, we are likely to visually experience her as having a long neck. Yet familiarity with the artist's intention, Wollheim maintains, renders our experience of the Madonna indeterminate in this respect: the “correct” visual experience will represent the neck with no specification of length. This claim is rather radical, and not everyone accepts it.35 But even if we reject the claim that this specific picture is indeterminate, indeterminate pictures abound. In general, pictorial indeterminacies are important factors in appreciating pictures, classifying them by style, learning about technique, and so on.
Visual accounts of depiction are committed to explaining pictorial indeterminacy by invoking the premise that we visually experience the depictum indeterminately, a constraint that follows straightforwardly from their endorsement of PET. Clearly, the overall experience we have in looking at a picture includes more than just our visual experience of the depictum: it encompasses our visual experience of the picture's surface (whether or not simultaneous with our experience of the depictum); judgments and beliefs regarding which configurational features are depictive and which are not; assessment of the manner of depiction; appreciation of the artist's technique; and so on. But what is crucial to the PET–ST incompatibility is not the overall experience we have in looking at a picture but the visual experience of the depictum—the specific visual mental representation on which our overall experience of a picture depends.
In what follows, I focus on uncontroversial examples of indeterminacy, such as line drawings and black and white pictures. These examples suffice to show that visual accounts of depiction must deny that every phenomenal difference in a visual experience of the depictum necessarily entails a difference in the properties the depictum is experienced as having. Indeed, if visual accounts seek to accept this principle (ST), they must deny both pictorial indeterminacy and its implications for the classification and appreciation of pictures.
IV. INCOMPATIBILITY
Suppose you are looking at a black and white picture of a garden. The phenomenology of your experience of the garden does not supervene on the properties the garden is experienced as having. The phenomenology features determinate hues—probably various shades of gray—whereas your experience does not represent the garden as having any determinate hue. The phenomenology can easily change—for instance, if you look at the sepia or bluish counterpart of the black and white picture—but the phenomenally different experience of the garden will represent the garden as having the same indeterminate properties.
The same is true of experiences of shape properties. Looking at a sketch of a human head, you do not experience the head as having the specific, determinate shape that the phenomenology of your experience suggests. You experience the head as having a determinable round shape. Here, again, the phenomenology of your experience of the head could be different (if you looked at a slightly different sketch), but the indeterminate shape that you experience the head as having remains the same.
Similar examples of pictorial indeterminacy show, in like manner, that assuming PET renders ST false. If a picture's marked surface can change in configuration without entailing a change in what the picture depicts, then the alleged visual experience of the depictum has nonintentional phenomenal properties. In general, an indeterminate picture has nondepictive configurational features that, despite being manifestly visible and experienced (simultaneously or not) as properties of the surface, engender a nonintentional phenomenology of our experience of the depicted object.
Recall that ST should also apply to nonveridical experiences; in particular, an experience of a depicted object, even if nonveridical, must be fully representational. ST should apply, for instance, to the nonveridical experience of a unicorn that is generated by a black and white picture of a unicorn. And if a phenomenally different experience of a unicorn—one that is generated, say, by a sepia picture—represents the unicorn as having the same (indeterminate) properties, ST must be false.
V. FURTHER CLARIFICATION
I will now refine the argument by addressing three possible objections.
V.A. Other Counterexamples
Pictorial indeterminacy, as interpreted by visual accounts of depiction, is not the only phenomenon that is, arguably, incompatible with intentionalism. A more familiar such phenomenon is blurry vision.36 A blurred experience is also indeterminate: it does not specify, for instance, the experienced object's determinate shape or the specific location of the object's boundaries. This phenomenon has been adduced in arguing, contra intentionalism, that blurred experiences have phenomenal properties that do not supervene on their representational properties. As Boghossian and Velleman state, a blurred experience cannot be “adequately described solely in terms of [its] intentional content. [Its] description requires reference to areas of colour in the visual field, areas that … become blurry without anything's being represented to you as [blurred].”37
Michael Tye, a leading intentionalist, has responded to this objection by explaining blurry vision in a way that, allegedly, renders it compatible with ST. It might be claimed, then, that my argument can be addressed in the same way.
Let us examine Tye's response. Tye argues that a given blurred experience does differ in representational content from other experiences. In having a blurred experience, “one's visual experience … makes no comment on where exactly the boundaries [of the object] lie”; it says only that they exist between spatial regions A and B.38 Other experiences either specify the experienced object's boundaries or represent the object as being fuzzy. An object can thus be experienced in three ways: one can have a blurred (that is, indeterminate) experience of it, one can have a clear experience of it, or one can experience the object as actually being fuzzy (the latter experience can also be characterized as a clear experience of a fuzzy object). These different kinds of experience indeed differ in representational content.
Tye does not mention, however, the kind of representational difference that must exist between two blurred experiences of the same object. He might argue that the two blurred experiences somehow differ in representational content, most likely in specifying different ranges within which the object's boundaries are located. This response seems somewhat plausible but suffers from vagueness. Indeed, the problem in using blurry vision as a counterexample to ST is that it is difficult to articulate the specific phenomenal difference between two blurred experiences, that is, the difference that would not entail a difference in representational content.
For this reason, at least, the case of pictorial indeterminacy is a better counterexample. The indeterminate experience that is induced, according to PET, by pictures is not characterized in terms of a relation between indeterminate experiences and determinate ones but rather in terms of a relation to other indeterminate experiences. Suppose you are looking at a black and white picture and its sepia counterpart. In both cases, you experience the same depictum as having the same indeterminate properties: neither experience specifies the depictum's determinate hues despite the manifest visual difference between the pictures’ hues. There is a phenomenal difference between the two indeterminate experiences of the depictum, but they do not differ in the properties they represent the depictum as having. Therefore, if PET is true, ST is false.
V.B. Representational Difference
A different objection is that, although on exclusionary accounts ST is indeed refuted by pictorial indeterminacy, ST is still compatible with twofoldness accounts. Given that we visually experience both the depictum and the marked surface at the same time, this overall experience should be taken into consideration in examining ST. Specifically, the said phenomenal difference between the indeterminate experiences does entail a representational difference: it entails a difference in the properties that the picture's surface is experienced as having. A black and white picture and its sepia counterpart, for instance, induce two phenomenally different experiences, though both experiences represent the depictum as having the same properties. Yet while the phenomenal difference does not entail a difference in the depictum's represented properties, it does entail a difference in the surface's properties: one surface is experienced as being gray, the other as sepia. Twofoldness accounts are, then, compatible with ST, given that the overall experience is being considered.
This objection, however, oversimplifies ST. ST is meant to guarantee that the representational content of an experience determines its phenomenology. More precisely, the properties x is experienced as having are supposed to determine the phenomenology of experiencing x. The fact that the properties x is experienced as having determine not the phenomenology of experiencing x but rather the phenomenology of experiencing a different object, y, is clearly insufficient for the intentionalist because the phenomenology of experiencing x is left undetermined. If we cast this requirement in terms of supervenience, for every experience that represents x as having certain properties, any change in the phenomenology of experiencing x necessarily entails a difference in these properties, namely, the properties x is experienced as having. If the said phenomenal change entails only a change in the properties that y is represented as having, ST is false.
In relying on the fact that visual experience can simultaneously represent different objects and properties, the objection dismisses the incompatibility between ST and PET too quickly. To clarify this point, consider a visual experience that represents two objects simultaneously, for instance, an experience of seeing a car parked under a tree. Suppose the weather suddenly turns windy, and consequently, the experience changes phenomenally. Without further assumptions, the required change in representational content may be related to the tree: due to the wind, the experience now represents the tree as having different properties (for example, swaying branches). However, let us further assume that there is a phenomenal change in our experience of the car. The question now is whether the fact that the tree is experienced as having different properties suffices for ST to be true.
The answer, I assert, is no. If, as we just assumed, a phenomenal difference occurred vis à vis experiencing the car, the car should be experienced as having different properties, whatever the representational change in the tree. If a phenomenal difference in our visual experience of the car does not entail any difference in the properties the car is experienced as having, this hypothetical case is a counterexample to ST.
Going back to indeterminate pictures, ST requires that any difference in the phenomenology of experiencing the depictum necessarily entails a difference in the properties the depictum is experienced as having whatever the difference in the surface's properties. As we saw, differently configured pictures induce phenomenally different experiences of the same depictum, which is nevertheless experienced as having the same (indeterminate) properties.
In this respect, then, twofoldness accounts are in no better a position than exclusionary accounts. If we experience the depictum and the surface simultaneously, it follows that ST should apply twice, that is, to two pairs of phenomenology and representational content. As we saw, it applies to the experience of the surface but not to that of the depicted object, as the phenomenon of pictorial indeterminacy demonstrates.
Recall that much depends on acknowledging that many, perhaps most, pictures depict their subject indeterminately. Our overall experience of pictures, an experience that features recognition of style, technique, limitations, and so on, depends first and foremost on our visual experience of the depicted object. When indeterminacies exist, this entails that we visually experience the depictum indeterminately. Hence, some of the phenomenal properties of our visual experience of the depictum must be nonintentional. This requirement, however, contradicts ST.
V.C. No Phenomenal Difference
A more radical objection is the claim that there is no difference at all between two indeterminate experiences of the depictum: they are identical not only in representational content but also in phenomenology. The only phenomenal and representational differences are differences between the experiences of the two differently configured surfaces. The experiences of a garden depicted in sepia and in black and white, for instance, have—so it might be argued—the same phenomenology. And since there is no phenomenal difference between these experiences, there need not be a difference in the properties they represent the garden as having.
This claim, however, is problematic. How can the difference between the surfaces’ hues not be constitutive of the experience of the garden? The surfaces’ hues—the gray and the sepia—are, indeed, not experienced as properties of the garden, but given that both hues occupy the same area of the visual field as the garden, they must surely affect the phenomenology of experiencing the garden. After all, the sepia or grayness, although not represented in experience, cannot be ignored to the effect that its appearance vanishes from the experience of the garden.
Or perhaps it can be ignored. The radical objection might endorse the thesis that visual experience is, generally, affected by attention (specifically, attention makes an experience's representational content more determinate).39 If the grayness and sepia are not attended to, they may indeed not be constitutive of the experience of the depicted garden: if, while looking at the pictures, we focus on the garden, the phenomenology that the grayness or sepia generates may indeed vanish from our visual experience—returning only when we attend to the marked surface—in which case the experience also has the proper representational content.
Note, first, that if this claim is true, it supports only ST's compatibility with exclusionary accounts. If we have a simultaneous visual representation of the surface and the depictum and experiences are indeed affected by attention, it follows that the surface's properties are specifically attended to and hence constitutive of the experience's phenomenology.40
Second, even exclusionary accounts may find it difficult to accept that when we experience the depicted garden, the hues that occupy the same area of the visual field as the garden are not noticed and hence do not affect the experience's phenomenology. Note that the question is not whether the sepia and grayness are represented in experience but whether they affect the phenomenology of experiencing the garden. The claim that attention increases determinacy is relevant only to the representational content of experience. This claim, in itself, says nothing about the experience's phenomenology. In our case, even if the sepia and grayness are not represented as properties of the garden, they are nevertheless clearly present in the visual field and hence can be claimed to be part of the phenomenology of experiencing the garden.
Third, attention can usually be increased at will. Suppose, then, that in looking at the said pictures, we are asked to carefully attend to the color properties that our visual experience represents. Would that change our experience of the garden? I do not think so. Despite the attention, we would ordinarily continue to experience the garden as having certain degrees of brightness without experiencing it as having any determinate hue. But given that the sepia and grayness are, on this scenario, being attended to, they obviously affect the phenomenology of experiencing the garden.
Fourth, consider the fact that unless we believe we are looking at a marked surface, we cannot really visually experience the depicted object indeterminately. If, in viewing a black and white picture of a garden, we mistakenly believe that we are looking at the garden itself, we will experience the garden determinately, as being gray. The reason for this is simple: a visual experience generated by an indeterminate picture differs phenomenally from a face to face experience of the depicted object so that even a naïve perceiver realizes that he or she is not looking at the depicted object itself. In fact, virtually all accounts of depiction acknowledge that experiencing a depicted object differs phenomenally from experiencing it face to face, though the accounts diverge with regard to how the difference is to be explained (Lopes's design seeing; Nanay's relational properties; experienced resemblance; Walton's imagining seeing; and so on).
Whatever the correct explanation, if the phenomenology of experiencing a garden by looking at a black and white (or sepia) picture differs from the phenomenology of experiencing a garden face to face, this is obviously because of the manifest visual property at issue, namely, the picture's gray (or sepia) hue. Indeed, had the picture had the typical hues of a garden, our experience might well have been determinate. Given that the grayness (or sepia) is what makes the phenomenology of experiencing the depicted garden different from the phenomenology of experiencing a garden face to face, the grayness (or sepia) is necessarily an aspect of that phenomenology.
On both twofoldness and exclusionary accounts of depiction, then, the experience of the depicted object is characterized in a manner similar to Boghossian and Velleman's characterization of blurred experience. The latter is described by reference to blurry areas of color in the visual field, areas they claim generate a nonintentional phenomenology. Likewise, in pictorial indeterminacy, the phenomenology of experiencing the depictum should reflect the area that the depictum occupies in the visual field. A visible change in that area thus entails a change in the experience's phenomenology—whether or not it also entails a representational change. Not surprisingly, Tye replies to their objection not by denying that blurred experience has a different phenomenology from clear experience but by arguing that it also has the proper representational content.
Indeed, given that the picture's surface and the depictum overlap in our visual field, it is hard to conceive of a case in which the experience of one of these entities changes phenomenally yet the experience of the other remains phenomenally the same. Even if we consider attention a constraint on perception, attending to a specific part of the visual field does not rule out an indeterminate experience of the depictum.
To summarize, on visual accounts of depiction, indeterminate pictures induce a visual experience with nonintentional phenomenal properties: the phenomenology of experiencing a depictum in such cases is not determined by the properties the depictum is experienced as having. PET, then, is incompatible with ST.
VI. CONCLUSION
Intentionalism takes nonveridical experiences such as hallucinations, illusions, and so on to be the same kind of mental representations as veridical experiences. It thus might want to welcome the thesis that pictures can induce a visual experience of what they depict. Pictorial indeterminacy, however, shows that, given PET, the supervenience thesis held by intentionalists must be false. If pictures indeed generate visual experiences of what they depict, indeterminate pictures in particular generate partly nonintentional visual experiences of the objects they depict.
As Wollheim, Lopes, and others showed, pictorial indeterminacy is a key factor in appreciating pictures. Hence, the fact that it demonstrates that either PET or ST is false cannot be ignored. What options are left for their proponents? One way to avoid the PET–ST incompatibility is to give up ST and admit that, at least in the case of nonveridical experience, the phenomenology of experience is not fully determined by representational content. This weak version of intentionalism accepts the first two points I mentioned above: visual experiences are representations, and they differ from beliefs and thoughts.41 Yet this weak version admits that, in some circumstances, visual experience has nonintentional phenomenal properties.
Those who regard ST as theoretically essential have to argue that the visual experience we have in looking at a picture does not ordinarily involve a visual experience of the depicted object. For instance, experienced resemblance accounts of depiction, as well as reference accounts à la Goodman and Kulvicki, are compatible with ST. It might also be that a subclass of what Hopkins calls “unitary” accounts—that is, accounts that consider the representation of depicta a sort of “abstraction” from the experience of the picture's surface—accord with ST.42
Lastly, the PET–ST incompatibility highlights a methodological issue. Accounts of depiction that use the term ‘visual experience’ without committing to a general theory of perception may be misleading, as such accounts may be associated with contradictory theses. Developing an account of pictorial experience that is rooted in a general theory of perception may raise additional problems connected to pictorial experience but can also be expected to deepen our understanding of it.43
The phenomenal properties of an experience include “what it is like” to have the experience, or its “immediate subjective feel” (Michael Tye, “Representationalism and the Transparency of Experience,” Noûs 36 (2002): 137–151, at p. 137). For characterizations of intentionalism, see William Fish, Philosophy of Perception: A Contemporary Introduction (New York: Routledge, 2010), chaps. 1 and 5, and Tim Crane, “The Problem of Perception,” in Stanford Encyclopedia of Philosophy (Spring 2011 edition), ed. Edward N. Zalta, http://plato.stanford.edu/archives/spr2011/entries/perception problem/, §3.3.
Crane, “The Problem of Perception,” §3.3.3.
See Fish, Philosophy of Perception, chap. 4, and Crane, “The Problem of Perception,” §3.3.1.
The term ‘twofoldness’ was first used by Richard Wollheim, Art and Its Objects, 2nd edition (Cambridge University Press, 1980), pp. 215ff.; others subsequently used it in different theoretical contexts. The incompatibility I argue for is relevant only to accounts on which twofoldness refers to the intentional content of a visual experience. Accounts on which we have a visual mental representation of the surface alone, whereas the depicted object is represented in a different way or not consciously represented at all, may be compatible with ST, for example, Robert Hopkins, Picture, Image and Experience (Cambridge University Press, 1998); John Hyman, The Objective Eye: Color, Form and Reality in the Theory of Art (University of Chicago Press, 2006); John Kulvicki, On Images: Their Structure and Content (Oxford: Clarendon Press, 2006). Note also that I use the term ‘visual experience’ as it is commonly used in the literature; compare Robert Hopkins, “Inflected Pictorial Experience: Its Treatment and Significance,” in Philosophical Perspectives on Depiction, eds. Catharine Abell and Katerina Bantinaki (Oxford University Press, 2010), pp. 151–180, at p. 173. Hopkins seems to use the term ‘standard visual representation’ to mean ‘visual experience’ in the standard sense.
E. H. Gombrich, Art and Illusion: A Study in the Psychology of Pictorial Representation, 2nd edition (Princeton University Press, 1961), pp. 4–5.
Eddy M. Zemach, “Look, This Is Zeus!” in Interpretation, Relativism, and the Metaphysics of Culture: Themes in the Philosophy of Joseph Margolis, eds. Michael Krausz and Richard Shusterman (Amherst: Humanity Books, 1999), pp. 310–333.
Kendall L. Walton, Mimesis as Make Believe: On the Foundations of the Representational Arts (Harvard University Press, 1990), chap. 8.
Kendall L. Walton, Marvelous Images: On Values and the Arts (Oxford University Press, 2008), p. 155.
I disagree with the remarks on Walton's account in Hopkins, “Inflected Pictorial Experience.” Hopkins claims that Walton's account is “unitary,” that is, that it accepts “the distinction between the two dimensions of content in pictorial experience,” yet does not take that distinction to “correspond to any further divide” in the nature of pictorial experience (“Inflected Pictorial Experience,” p. 170). However, other cases of cognitive penetration demonstrate that in being penetrated by a cognitive state, visual experience can no longer represent the marked surface. Macpherson, for example, discusses an experiment that shows how in looking at an orange cutout of a characteristically red object (an apple or a pair of lips), people tend to experience the cutout as redder than it really is (Fiona Macpherson, “Cognitive Penetration of Colour Experience: Rethinking the Issue in Light of an Indirect Mechanism,” Philosophy and Phenomenological Research 84 [2012]: 24–62). Macpherson explains that such a nonveridical experience occurs because it is penetrated by the belief that the cutout is of a characteristically red object. In this kind of case, the perceiver cannot experience the cutout as both orange and red. In much the same way, Walton's view must preclude the twofold experience at issue: if our visual experience of a horse is indeed being penetrated, it can no longer represent the marked surface's properties.
Walton, Marvelous Images, chap. 6.
Wollheim, Art and Its Objects; Richard Wollheim, Painting as an Art (Princeton University Press, 1987); Richard Wollheim, “On Pictorial Representation,” The Journal of Aesthetics and Art Criticism 56 (1998): 217–226; Richard Wollheim, “What Makes Representational Painting Truly Visual?” Proceedings of the Aristotelian Society, supp. vol. 77 (2003): 131–147; Dominic McIver Lopes, Sight and Sensibility: Evaluating Pictures (Oxford: Clarendon Press, 2005); Bence Nanay, “Inflected and Uninflected Experience of Pictures,” in Philosophical Perspectives on Depiction, eds. Catharine Abell and Katerina Bantinaki (Oxford University Press, 2010), pp. 181–207; Michael Newall, What Is a Picture? Depiction, Realism, Abstraction (New York: Palgrave Macmillan, 2011); and John Kulvicki, “Heavenly Sight and the Nature of Seeing In,” The Journal of Aesthetics and Art Criticism 67 (2009): 387–397.
Wollheim, “On Pictorial Representation,” p. 221.
Lopes, Sight and Sensibility, p. 32.
Lopes, Sight and Sensibility, p. 28.
See the analysis of Lopes's view in Hopkins, “Inflected Pictorial Experience.” Note that seeing in is sometimes uninflected, in which case we experience only the depicted object. Cases of uninflected experience are not relevant to the PET–ST incompatibility.
Nanay, “Inflected and Uninflected Experience of Pictures,” p. 194.
Nanay, “Inflected and Uninflected Experience of Pictures,” p. 198.
See Nanay, “Inflected and Uninflected Experience of Pictures,” pp. 199ff.
Newall, What Is a Picture? p. 42.
Newall, What Is a Picture? pp. 30ff.
Hyman, The Objective Eye, chap. 5.
Christopher Peacocke, “Depiction,” Philosophical Review 96 (1987): 383–410; Hopkins, Picture, Image and Experience.
Hopkins, “Inflected Pictorial Experience,” pp. 170ff. See also Abell and Bantinaki, Philosophical Perspectives on Depiction, p. 13.
Nelson Goodman, Languages of Art, 2nd edition (Indianapolis: Hackett, 1976); Kulvicki, On Images.
The next three paragraphs summarize my argument in Section 2 of Alon Chasid, “Content Free Pictorial Realism,” Philosophical Studies 135 (2007): 375–405.
Lopes, Sight and Sensibility, pp. 28ff.
Lopes, Sight and Sensibility, p. 28.
Robert Hopkins (Picture, Image and Experience, chap. 6) advances an account of pictorial indeterminacy that distinguishes between what is seen in a picture and what the picture depicts, ascribing indeterminacy only to the latter. On this account, there is no indeterminacy in what we see in, for example, a line drawing: we indeed see a transparent object in it; we only conclude, by interpreting the picture, that the drawing depicts the object indeterminately (that is, without commenting on its color). Pictorial indeterminacy, on this account, is purely a matter of cognition, not of visual experience. Can PET supporters adopt Hopkins's “separation” account of pictorial indeterminacy and thus accommodate ST? I think not. Virtually no one accepts this sort of nonexperiential account of indeterminacy, nor is this account plausible. First, Hopkins himself mentions that not every indeterminate picture should be explained by this account, only “highly indeterminate” pictures (pp. 124–128). Some indeterminacies are definitely a matter of indeterminate experience, and therefore PET–ST incompatibility applies at least to them.
Second, recall that Hopkins rejects PET and explains “seeing in” in terms of experienced resemblance. This fact is crucial for his “separation” account of indeterminacy because experiencing x as merely resembling y (that is, not as visually identical to y) in itself explains some of y's indeterminacies—indeterminacies that Hopkins's account of indeterminacy is thus not in need of explaining further. PET supporters, however, are forced to explain even the slightest indeterminacy in terms of PET: they must ascribe every indeterminacy to the visual experience of the depictum.
In general, if PET is true, to argue that experiencing the depictum is necessarily determinate conflicts with at least two facts. For one thing, pictorial experience differs phenomenally from ordinary, face to face experience of the depictum. Some indeterminacies must characterize the former; otherwise, it would not be phenomenally deviant. For another, in certain cases, visually experiencing the depictum is indeed phenomenally identical to experiencing it face to face (as is the case with trompe l'oeil pictures, which are therefore not considered indeterminate). But note that in such cases, we are at least inclined to believe, if not actually deluded into believing, that we are looking at the depictum. In other cases, we are hardly disposed to err in this way: the phenomenology of experiencing the depictum discloses the deviant nature of these experiences and, specifically, their indeterminacy.
Dominic Lopes, Understanding Pictures (Oxford: Clarendon Press, 1996), §6.5, pp. 127–131.
Lopes, Understanding Pictures, p. 118.
Lopes, Understanding Pictures, p. 130.
Wollheim, “What Makes Representational Painting Truly Visual?” p. 143.
Wollheim, “What Makes Representational Painting Truly Visual?” p. 143.
See Wollheim, Painting as an Art, pp. 48–51. Note that disagreements about the “correct” way to experience the depictum are irrelevant to the PET–ST incompatibility, since whatever the standard of correctness, it must allow some pictorial indeterminacies, as many uncontroversial examples show.
See John H. Brown, “Seeing Things in Pictures,” in Philosophical Perspectives on Depiction, eds. Catharine Abell and Katerina Bantinaki (Oxford University Press, 2010), pp. 208–236, at p. 231, n. 30. Although Brown thinks that Wollheim is wrong in classifying Parmigianino's Madonna as a case of pictorial indeterminacy, he, too, carefully analyzes other examples of indeterminacy, as his idea of “authorized seeing in” suggests. In my opinion, Wollheim's view of Parmigianino's Madonna might, perhaps, make sense if we understand it as akin to a case of experiencing something through a twisting medium. Experiencing a face by looking, say, at a twist inducing mirror, we may experience it as being twisted; yet knowing that the mirror has a twisting effect, we may shift to experiencing the face indeterminately, in which case the twisting will be experienced as a property of the mirror.
See Paul A. Boghossian and J. David Velleman, “Colour as a Secondary Quality,” Mind 98 (1989): 81–103; Michael Pace, “Blurred Vision and the Transparency of Experience,” Pacific Philosophical Quarterly 88 (2007): 328–354; Robert Schroer, “Seeing It All Clearly: The Real Story on Blurry Vision,” American Philosophical Quarterly 39 (2002): 297–301; A.D. Smith, “Translucent Experiences,” Philosophical Studies 140 (2008): 197–212; Tye, “Representationalism and the Transparency of Experience.”
Boghossian and Velleman, “Colour as a Secondary Quality,” p. 94.
Tye, “Representationalism and the Transparency of Experience,” p. 148. See also Smith, “Translucent Experiences,” p. 207.
See Bence Nanay, “Attention and Perceptual Content,” Analysis 70 (2010): 263–270. This claim is not accepted by everyone; see, for example, Macpherson, “Cognitive Penetration of Colour Experience,” p. 43.
Note that the explanation of inflected seeing in offered in Nanay, “Inflected and Uninflected Experience of Pictures,” also invokes the thesis that experiences are affected by attentiveness: he argues that if it is correct, we attend to a relational property (see Section II). It is unclear,however, what Nanay would say about manifestly visible nondepictive properties, such as the said grayness and sepia: if we can indeed ignore them, it follows that we do not experience them; hence we have a onefold experience. Nanay might insist that these properties are relational nonetheless; in this case too, however, they participate only in the depicted object fold. As I mentioned above, in this sense Nanay's account might be considered exclusionary.
Fish, Philosophy of Perception, pp. 70ff.
Hopkins, “Inflected Pictorial Experience,” p. 170.
For example, experienced resemblance accounts claim that we experience the picture as resembling the depictum. But is resemblance a relation that can indeed be represented in visual experience, or do we only believe, on the basis of perceptual experience, that one object resembles another? The answer depends on general considerations as to the nature of perceptual experience; see Fish, Philosophy of Perception, pp. 136ff.