Philosophy Looks at the Arts
ed. Joseph Margolis
(Third Edition, 1987)
[307]Picturing and Representing
MARX W. WARTOFSKY
[orig. 1979]
The thesis I want to present in this chapter, although stated very baldly, is; human vision is a cultural and historical product of the creative activity of making pictures. To put it somewhat differently, human vision is an artifact, produced by means of other artifacts- for example, by pictures; as such, it is a historically variable mode of perception, which changes with changes in our modes of representation. What follows from this thesis, if it is true, are two radical conclusions- one, methodological, the other, epistemological.
The methodologically radical conclusion is that all theoretical attempts to construct a theory of vision, which presuppose that seeing is an essential, unchanging structure of process; or that the human eye is describable in some generic physiological way, are, if not fundamentally mistaken, then essentially incomplete. For the plasticity of the visual system, if I am right, is such that it requires a historical account of its development and not simply a biological account of its evolution, whether phylogenetically or ontogenetically.
The radical epistemological conclusion is that there is no intrinsically veridical, or "correct," mode of representation, that is, there is no criterion of veridicality that is not itself a product of the social and historical choices of norms of visual representation. There is, therefore, no canon of truth in perception that can be established by reduction to the physiology of vision, or to optics, or to some species-specific biological, or even ecologi- cal, account. That is not to say that we do not inherit the mammalian eye, nor is it to say that human vision is not based in an evolved, adaptive
[308]
structure which develops with the speciation of Homo sapiens. However, truth in perception, as I will argue, is bound to canons of the veridicality of representation; these, in turn, have a history and are rooted in our social practice and in our own activities of picturing and representing. Thus, it is we who create the very norms of veridicality by our pictorial practice. Such norms are not arbitrary, though they are conventional; they are not biological but historical.
The larger theoretical enterprise of which the thesis that I propose is a part may be characterized as a historical epistemology. I will present one specific aspect of such a historical theory of vision, which I have begun to develop elsewhere (Wartofsky 1972, 1976). At issue is the question of how representation is possible in the specific mode of representation that we call picturing-_and, more specifically, in that subcategory of pictures which we call paintings.
My argument may be summarized in five points. First, the act of representing something pictorially is a creative act. That is to say, it does not depend on some antecedent notion that something (a picture) "looks like" or "resembles" or "represents" something else for example, a scene, an object, a person, and so forth), but rather that it is we who create the similarity which counts as representational. Similarity is not given, but achieved; made, not discovered. It is invented and created.
Second, the perceptual system--and the visual system in particular-is biologically evolved to take certain things in the visual world as being like or resembling others, by virtue of the forms of life activity of a given species, that is, the means by which its individual members preserve themselves in existence and reproduce the species life. These canons of resemblance, similarity, or identity are mapped into the neural and neuromuscular structure of the species; they may be said to be coded into its genetic structure by natural selection. The human species, however, has a radically alternative mode of mapping its forms of life activity into structure, and this is by means of canons of visual representation, that is, embodied rules for taking one thing as a representation of something else. Pictures--or rather styles of pictorial representation--exemplify canons of representation, by means of which we come to see the visual world as like the picture.
Third, the rules of linear perspective in painting and drawing are not "correct" representations of the way things "look" but rather proposals to see things the way they are represented pictorially. When they come to look the way they are pictured, it is because we have adopted the rule of picturing as a rule of seeing the world--that is, we see by way of our picturing.
Fourth, the alleged paradox of pictures--namely, that the three-dimen-
[309]
sional world is represented in a two-dimensional image of the world-is not a paradox; it is dissolved when we recognize that in taking a picture as a representation of the world, we come to take the world as a two-dimen- sional picture, which is, in this respect, like what pictures or represents it. When we are fooled, then the picture is no longer, properly speaking, a picture. However, we are rarely fooled. The argument is this: We say that representational pictures (for example, paintings) "look" three dimen- sional, that is, they "look like" the three-dimensional objects or scenes which they represent. However, we take the world to "look" three dimen- sional only by contrast to the two-dimensionality of pictures. The visual concept of "three dimensionality" is thus a constructive concept, which depends upon reference to, or a relation to, the two dimensionality of pictures. We would have no distinctive notion of the three dimensionality of the visual world except for the distinction we come to draw between it and two-dimensional representations of it. In short, the making of two- dimensional representations, or pictures, is what generates the contrastive visual concept of three dimensionality. I would suggest that the geometry which defines the plane projection of a solid is likewise dependent on the more primitive notion of a picture, that is, a representation or configura- tion in the two-dimensional plane.
Fifth, and finally, modes of picturing change, with changes in form of our social, technological, and intellectual praxis; representation has a history, and thus, in coming to adopt different modes of representation, we literally change our visual world. Human vision is an artifact created and changed by the modes of picturing. Different modes of picturing have different theories. We have adopted the theory of linear perspective as our theory of veridical representation. It is not incorrect. But neither is it correct. Veridicality is not a given feature of a mode of picturing. It is defined by the theory of pictorial representation that we come to adopt.
So much for the summary of my argument. Now to the question posed for this volume: "What is a painting?" The question needs to be specified more concretely in order to be answered. It has many answers, depending on what it is that is being asked. A merely ostensive definition-"that's a painting" (pointing at one)-will not do, but it does hide a deeper answer. Namely, paintings are the sorts of things that are taken to be paintings. Thus revised, the question becomes, "What is it we take to be a painting?"
I plan to deal with only one aspect of this question here, namely, that of pictorial representation, which is one of the things paintings do (or are taken to do). In order to deal with this question, however, there is a more general characterization I want to give to paintings, apart from, but related to their representational capacities, that is, paintings are artifacts, made things, the products of intentional human activity, and insofar as they are
[310]
representational, they are intended as representations -that is, they are artifacts whose purpose it is to represent something.
The question then becomes, "How do the kinds of artifacts called pictures represent?" What is made in such a way as to come to represent something else? More specifically, in the case of painting, how do arrange- ments of lines, areas, colors, and gradients of dark and light on a two-dimensional plane surface come to represent three-dimensional ob- jects, scenes, and persons?
There are several possible answers to this question. I will examine three, but first, I would like to frame my approach so that the context of my considerations will be clear. There are several general points to be made. First, if (as is clear and undisputed) representation has a history and if (as is yet unclear and in dispute) this history is a crucial factor in the historical and cultural evolution of the human visual system, then the history of art becomes an essential component of the theory of vision: alternative modes of representation, the history of styles in painting, the theoretical analysis and reconstruction of the visual concepts of space, objects, and relations, which are characteristic of a given style or period, the phenomenological reconstruction of the modes of intentionality that identify an art-historical epoch, or a school all become essential components of any theory of vision, as crucial as the study of the psychology or physiology of vision and inseparable from these latter inquiries.
Second, since modes of representation are not simply or abstractly visual matters but involve also the larger social, technological, political, scientific, and even ideological contexts of human cognitive practice, a theory of vision is embedded in this larger framework of human social activity. Our seeing is a mode of our activity; just as vision cannot be conceived of simply as an activity of the eye, taken out of its context as part of the mammalian brain, and of the whole organism, so, too, it cannot be conceived of simply as the activity of the visual system, taken out of the context of the form of life in which this system operates, in which it develops, and which it is capable of transforming. To put this another way: it is neither the eye nor the visual system that sees. Rather, it is we who see, by means of the eye or the visual system. Similarly, it is not feet that walk. Rather, it is we using our feet. I could not walk without feet or see without eyes, but the I that sees, walks, talks, paints, argues, and gives papers at symposia is a social being, an individuated member of a life-form which is essentially historical and social. No mode of the life activity of such an individual can be adequately characterized, therefore, by abstractive reduction to the particular organs or apparatus by means of which this life activity proceeds. In this regard, the approach I am proposing is analogous to James Gibson' ecological approach to vision,
[311]
but it differs sharply from his in that the ecology I am suggesting here is not a natural or biologically defined one but a cultural, or sociohistorical, "ecology." In short, vision, or seeing, is not merely the after-product, the epiphenomenon, of a given apparatus in a given environment or simply the operation of an organ. It is a creative activity that can be transformed and which, in turn, can itself transform a given form of life.
Third, and finally, if vision is a historically variable mode of cognitive practice that changes with alterations in our modes of representation, then the evolution of our visual system, as a cultural artifact, is, in contrast with its evolution as a biological system, no longer Darwinian but Lamarckian. That is to say, in the cultural evolution of this artifact, that is, of vision, there is transmission of acquired characteristics from one generation to the next. The mode of transmission is, therefore, no longer genetic but becomes social or cultural. The visual culture of a society or of an age is not inherited by the operation of genetic transmission or by means of the biochemical structures or codes that have been selected out, preserved, and developed by natural selection but rather by means of social structures- in particular, visual artifacts and modes of picturing. To put it simply, the artifact is to cultural evolution what the gene is to biological evolution. Cultural evolution, in contrast with biological evolution, is Larmarckian and not Darwinian. Such a thesis makes it possible to account for the plasticity and also the rapidity of cultural evolution; further, it is directly opposed to the current theoretical approaches to human cognition, as well as to human sociality, which find it necessary to focus on the alleged biological or genetic constraints on, or determinants of, modes of human activity.
To summarize this introductory discussion: (1) we see by way of our picturing changes in our modes of cognition, in general, and in our modes of vision, in particular, are concomitant with changes in our modes of social practice, in general, and with our modes of pictorial representation, in particular; (2) the human visual system is therefore not simply the biological structure of our species evolution but an artifact produced by our own creative activity of picturing and seeing; thus, its plasticity cannot be defined reductively in biological, that is, physiological or genetic, terms.
Against this larger framework, let me now turn to the specific question, "How is representation possible?" The title of this chapter, "Picturing and Representing," is intended to correct a certain initial tendency to hypos- tatize: "Pictures and representations" suggests, I think, that we begin with certain entities already understood. I want to put in question what it is we understand by such entities and to emphasize that pictures picture and representations represent by virtue of the fact that they are the products of an activity that intends them as picturing and representing. That is, for
[312]
something to be a picture, it needs to be made as a picture and taken as a picture; for something to be a representation, it has to be intended to represent, and this intention has to be understood in taking it as a representation. Nothing, then, is a picture or a representation in itself apart from being made as one or taken as one. There are no entities, then, which may "objectively" be characterized in this way. Instead, things of a certain sort are constituted as pictures and representations by makers and viewers. There is a radical consequence to this view as well: if nothing is intrinsi- cally a picture or a representation, then we cannot ascribe a set of intrinsic properties to something that would identify it as a picture. In effect, anything can be a picture or a representation if it is made to be one or is taken as one. Let me make clear that I propose this in the strongest way, without qualification, that is, it is not the case that certain things are taken to be pictures or representations because they exhibit certain visual properties but, rather, that they come to exhibit certain visual properties because they are taken to be pictures or representations.
Now, this flies in the face of common sense--or seems to, for, in fact, not everything is taken to be a picture or a representation; in fact, a very narrow range of things is so taken at any given time or in any given culture. However, this narrow range itself is variable: some things not taken or made as pictures in one context are pictures in another. Moreover, insofar as pictures represent, it would appear that the relation of representing is based on some relation between some properties of the picture and some properties of what it is taken to represent, usually, a relation of resembling, or similarity, or likeness. The pictorial representa- tion is said to be a representation by virtue of being a likeness; therefore, a successful representation shares properties with its reference.
My first claim, against this common view, is not original but is shared with Goodman (1976), upon whose more systematic and elaborrated argument on thie point I will rely. It is that representation requires no relation of resemblance, or likeness, but is rather constituted as an act of reference. My second claim, however, goes beyond this, though it is in the same spirit. It is that likeness, resemblance, or similarity is itself not given in our visual perception; it is not a primitive, irreducible relation, as Mill took it to be (in Nagel 1950), but is itself an achieved relation, that is, one which is constructed and, therefore, construable (compare Mann 1971). Things come to be similar, to resemble each other, be likenesses by virtue of being made as similar, or alike, or being taken as resembling each other. Thus, even in the relation of representation in which it is alleged that the representation is like what it represents, I am arguing that it is so by virtue of our taking it to be so, in a given respect. It is we who create similarity, resemblance, or likeness in those forms of pictorial representation that are
[313]
said to be based on it. We do so, I would argue, by producing artifacts that are specifically intended to be like other things, to represent them, in this way. The classic story here, of course, is about Picasso's portrait of Gertrude Stein. When told that it did not look like her, he answered: "It will."
On the basis of one interpretation, this may be no more than to say that in making a representation, we imitate what it is that is being represented, that is, we construct something that is similar to what it is to represent--in effect, we make a copy. In this case, then, there should be no surprise that there is resemblance, likeness, or similarity, for is not the representation made expressly to be like what it represents? But this interpretation- the common one--begs precisely the question that it sets out to answer. For in order to make something that is similar to, or resembles, something else, we have to establish what the feature is which is being "copied" in this sense, and that something else in fact resembles it. Imitating, copying, or reproducing the features of a given object, scene, or relation requires therefore just that creative act of achieving the likeness, of establishing that something resembles something else, which is presup- posed as a given in the standard view. In a second interpretation, one thing may be said to re-present another, not in being like it, or resembling it, in some common properties--for example, same shape, or color, or same relation among parts--but, rather, in causally effecting the same response. Therefore, the representation comes to represent by virtue of bringing about the same physiological response or visual experience as the thing or scene represented, though it does so by means that are dissimilar. So, for example, in this view, paintings of landscapes or figures are pictorial artifacts and, thus, two-dimensional arrays: they are not themselves "like," in this sense, the three-dimensional objects or spaces that they represent. Yet, they are arranged in such a way as to deliver the "same" or a similar light flux to the retina, and thus initiate similar visual responses.
This view, perhaps the most popular psychological theory of represen- tation, retains semblance in the phenomenal experience, or in the response, though it gives up identity or even similarity in the stimulus. In this sense, it presupposes that similarity in the response is passive--a matter of equivalent causal or antecedent conditions, whose equivalence is judged by the sameness of the response.
The faults with this view are many. But let me simply point to one experimental fact: the recognition of pictures as representations of x ranges over pictures that are obviously not similar to each other in any respect one could define as "causally equivalent." Gibson would evade this point by arguing that for all their dissimilarity, such a range of alternative pictorial representations are all alike in transmitting the same higher-order
[314]
visual invariances criteria for object or scene recognition, though through the range of transformations. But such invariances, in Gibson's ecological optics, are taken to be objective features of the ambient light, and the organism—in this case, the human one, or, perhaps, the higher vertebrate in general—has so evolved its visual system as to be able to pick up these invariances directly—that is, without processing lower-order variations. The net is woven, so to speak, in order to catch fish of only a certain size (or a higher order) and to let all the smaller ones (or those of a lower order) slip through unnoticed.
This interpretation begs the question in a different way from the first, for it presupposes as invariant just those features that representations represent- for example, so-called real shape or real size in the case of perceptual constancies--and, thus, does not explain how representation is possible; in fact, it theoretically constructs an explication of one kind of representation--namely, that which is constructed in accordance with the rules of linear perspective- that it takes to be canonical. It is no surprise then-in fact, it is inevitable, because circular--that in Gibson's view, what we see is what there is, and representations succeed because they re-present what we see.
If I deny that we have direct access to the way things really are, at least as the real objects in our ecological space or the space of our species' life activity, then how else is representation possible? What links representa- tion to representandum: how do we have access to the properties of the visual world so that we know how to represent them successfully?
To take a step further, I would propose that we do not come to have visual concepts of the properties of the visual world except by such a process of creating representations of it. Nature comes to imitate art precisely because what we make of nature as an object of vision is constituted, in large part, by how we choose to represent it.
Before proceeding with the argument, it may be useful to make explicit some distinctions that have thus far been merely suggested. I have stated that human vision is itself an artifact, produced by other artifacts, namely, pictures. I call something an artifact if it is a product of human activity, in the sense that it comes into being as the result of intentional human making and is made or constructed with an end-in-view or for a use or purpose. This use or purpose defines the artifact as what it is, so that an artifact is what it is made for: it is, in short, a teleological entity. It may be odd to talk of vision or seeing in this way, but in fact, it is we who have shaped our vision to certain uses~-who have adopted, adapted, and replaced different visual modes in our pursuit of different interests and ends. We have, in effect, learned to read the visual world in different ways, depending on our interests and needs.
[315]
One objection to such a view is, obviously, that the physiology of the visual system, though it may be understood as a biological adaptation to a form of life, is not itself the product of our own deliberate design nor is it subject to our intentional manipulation. We cannot be said to "change" our visual physiology as our interests or purposes change, for the structure is a genetically determined one and its evolution to the present species-form is the product of natural and not cultural selection. Within the confines of gross physiology, this is, I believe, evidently true. The adaptive variation of the visual system, in the course of species evolution, is the work of natural selection. But even here, we may characterize this species evolution as the mapping into the organism's genetic makeup of those features of its life-world and life activity that are requisite for its species-survival. The story is different, however, for the development of vision beyond the species level that is, for the differentia- tion of visual perceptions which proceed with cultural evolution or with historical changes in the human forms of life. One may go so far as to claim (with Penfield 1966) that the physiological ontogenesis of the neural system in general (and for the visual cortex in particular) is a differentiated one, which maps into the individual's neural structure the specific modes of experience and activity that characterize the life history of that in- dividual. But one need not go so far to see that the visual system as a way of seeing is subject to the variability of modes of visual re-presentation.
What then are pictures and representations, and specifically, what is a painting? Let me make a distinction between pictures and representations. Pictures are visual artifacts, that is to say, pictures are made to be seen. Many other artifacts, are of course, also visible, but by visual artifact, I mean something expressly made for the purpose of being seen. Pictures, on the other hand, do not exhaust the class of visual artifacts. One may include, here, any marks, signs, objects, or expressions that are expressly made to be seen or any entities which come to be taken principally as objects of vision. So, for example, physiognomic expression, hand signs, sculpted objects, signs or markings of warning, direction, ownership, kinship, and so forth, may be taken as visual artifacts in this sense, that is, that they are intended to be seen and to communicate a meaning visually. Thus, modes of dress, scarification, gestures, and facial expres- sions are all such artifacts. Pictures, as a special class of such artifacts, I will take to be as those that are made upon a plane surface (thus, not gestures or sculptures for example) and which depend on line or color as their visual means. Not all pictures are representations, though all pictures may be said to have meaning. Moreover, natural objects or scenes may be taken pictorially, though they are not made things. Thus, we may see a tree, a sunset, or a cloud as a picture when we take it principally as a
[316]
visual artifact, that is, as a meaningful form or shape seen as if it were on a plane surface. The force of this particular qualification will be seen later, since I will argue that it is by means of our activity of picturing that we come to be able to see the world as a picture and that the standard theory of vision which underlies constancy theory, in perception, and which proposes linear perspective as the correct representation of the visual world is based on an interpretation of the three-dimensional visual world as a two-dimensional picture.
Pictures, then, are visual artifacts made to be seen-_and when understood in this way, seen as pictures. The ubiquitousness of the pictorial in human life leads us to forget that picture making and picture seeing are learned activities of the species. I would go further and suggest that picturing, both in making and seeing, is a fundamental form of the life activity of the human species and, in this sense, I am proposing that it is this activity that shapes human vision and develops it beyond the biological inheritance of the mammalian eye.
If all pictures are not representations, all pictures may be taken to be representations. Visual representations, briefly, are visual artifacts ex- pressly made or understood as referring to something beyond themselves. Thus, there are nonvisual representations that are not pictures, for ex- ample, vocal reference in speech or gestural ostension; so, too, there are visual modes of reference that are not pictorial. Representation is sym- bolic. That is, one thing stands for another, under an interpretation given in a symbol system of which the representation is a part.
There is much more to be said about these distinctions, and Goodman's pioneering work (1976) goes a long way in developing this analysis. I will not pursue it here. But in summary, for the purposes of this chapter, I will distinguish pictures from representations, insofar as there are pictures that do not represent and representations that are not pictorial. More generally still, I will hold that all artifacts are putative representations, insofar as the very use, or understanding, or recognition and identification of an artifact as what it is requires that we take it as a representation of the mode of activity involved in its use or in its production. The artifact "represents," literally, in being taken as an imitation of an action, that is, the embodiment of the intentionality involved in its production, reproduc- tion, or use. Thus, for example, a tool or a weapon, such as an ax or a spear, is not only something made to be used for a certain end but is also itself a representation of the action involved in its use and in its production or reproduction. An ax therefore, as a visual artifact, represents the activity of chopping and a spear, the activity of throwing in order to kill in the hunt. Moreover, the artifact also represents the mode of activity involved in its production or reproduction: it is a prototype of its replicas and a model of its own process of production.
[317]
What, then, marks off pictorial representations from the wider class of artifacts as putative representations? Again, the emphasis is on intentions, that is, what is intended in the making and taking of such visual artifacts. A spear is made for hunting and also represents the mode of action of the hunt. A spear-picture, on the other hand, is not made for hunting but is made expressly as a representation of a spear. It is principally made as a visual artifact, not a hunting artifact; its very intentionality is distinct in this sense; its purpose is different. Moreover, the representation is detached from use or from production: it functions independently of the activities or contexts of what it represents. Such a relatively autonomous act of representing is made possible by the picture, for it is in the express creation of something as a visual artifact, as something made to be seen, that the separation from other contexts of use becomes possible. It is not that representational pictures have no use but that their intended use is dif- ferent. The use-value of a picture (representational or not) is in its being seen, whatever other purposes such visual presentation may have, for example, didactic, formal, expressive, informative, and so forth. The use-value of a representational picture is in its visual representation of its referent. It is in this sense that I argued earlier that something is a representation insofar as it is made to be, or taken to be, a representation. The features that come to be called representational will then depend on what is taken to serve this function. Anything can; but not everything, in fact, is chosen to do so. What is chosen to function as representational is a complex question, not to be resolved by appeal to physiological (or ecological) optics. Rather, it demands comparative study of what in fact has been taken as representational or what, in various alternative canons, continues to be taken as representational.
Pictures that represent, then, are artifacts which guide or shape our vision of the world, leading us to take this as like that, to pick out features of the seen world that are referred to by the representations we make, where the act of reference is itself a creative act and not merely a matching of pregiven similarities or identities. One may say, then, that representa- tional pictures are heuristic and didactic artifacts. They teach us to see: they guide our vision in such a way that the seen world becomes the world scene.
References
...