by Colin Mcginn
Painting presents a subtler issue. It is two-dimensional, like cinema, and it presents the viewer with an image of an absent object, also like cinema. Unlike the stage, painting is a representation of people and things, not real people and real things. But do we look into a painting (as opposed to the eyes in a portrait)? I think not: we look at a painting. We stand before the painting and admire it—we scrutinize it, let our eyes linger over the painted canvas. In no way does the painting become invisible to us—it is the object of our visual attention. The painting is not a visual route to something else; it is the perceptual end point of the process. So the canvas is not, from a visual point of view, like a movie screen. And, of course, the two are differently constituted: the painting consists of pigments on a flat surface, which themselves reflect the light; the movie image consists of projected light on a flat surface, which does not itself reflect light. One is a chemical phenomenon; the other is a phenomenon of pure light. The two types of image are made of completely different physical materials, and this is evident to the eye of the viewer. Incident light enhances the painterly image, but it destroys the movie image, since it competes with the light on the screen (this is why movie theatres are dark and art galleries light). The way light reflects from the surface of a painting gives the painting a texture, which is evident to the viewer's eye—we know that it will feel a certain way if touched (bumpy, smooth in places). The painting is perceived to be a surface, just like the surface of a table or a carpet, and its properties are manifest in the way it reflects light. But none of this is true of the movie image: it has no texture. It is not that incident light reflects the antecedent condition of the screen (nothing is imprinted on it before the projector starts); rather, the projected light is itself the image. This image tells us nothing about the screen: the light of the movie image does not reveal what the surface of the screen is really like, as the light falling on a painting reveals its standing surface features. Accordingly, we do not admire and scrutinize the screen as a surface, but proceed visually to another plane. We look into and beyond this textureless mirage; but in the case of painting we are confronted, as it were, with a flesh-and-blood object—a solid material thing, opaque and reflective in its own right. Moving paintings, if such there could be, would not engage the eye in the same way that the light-constituted movie image does—they would still be objects of, and for, the attention. This is why we marvel at, say, the way Turner uses paint to represent light, but we don't lavish the same praise and attention on the way the light on a movie screen represents light. The medium of painting—pigments on canvas—is not an absent medium, perceptually speaking. So the way we look at a painting is not analogous to the way we visually apprehend holes, water, the sky, the eyes, and so on; and so it stands in contrast to our visual relation to the screen.
The case of photography might seem to fall on the side of the movie image, since movie images precisely are photographic images. If we look through or into cinematic photographs, must we not also look through or into the kind of photographs that are imprinted on pieces of paper? However, a regular photograph is a chemical effect imprinted onto paper: it reflects light, it has a texture, and it appears as a surface. Thus we naturally invite each other to look at photographs (“Look at this picture of me in Cancun,” etc.).
Of course, as with paintings, we also see this looked-at object as representing something, a person, say; but we do not thereby look into the photograph. We look at the photograph and see it as representing something, but we don't look at the movie image and thereby see it as representing something. The movie image is seen but not looked at; the photograph is seen and looked at (and not just when it is defective). The reason for the difference is not that one is moving and the other is not; it is that one is seen as a textured surface and the other is not. This is why looking at a photograph of X is not like seeing X through a pool of clear water, though seeing X in a screen image is. Photographs are objects of visual attention; movies exist to direct attention elsewhere (soon I shall say where).
But surely, it may be said, television is just like movies, for don't we watch movies on television? When the eye is fixated on the television screen, and the viewer is immersed in the action of the movie he or she is watching, isn't there equally an act of looking into? Isn't the television screen just as self-effacing as the movie screen? And yet, don't we prefer to see a movie at a movie theatre? (We stay home for practical, not aesthetic, reasons.) Somehow, we feel, it works better there, no matter how wide and flat our TV screen may be. The question is whether this shows that TV screens are not the kind ofthing we look into.
Now it is not that we can never in principle look into a movie-screen image in our own home (some people have private movie theatres, after all); but I do think there remains a significant point of difference between the two types of screen, arising simply from the physical nature of the TV screen. For the TV screen itself—a piece of rectangular glass sitting in front of the viewer—is an object that can all too easily become a visual surface in its own right, as when light from the window or a lamp falls across its glassy face. Then we find our attention distracted from the film we are watching to the medium of our watching it; the screen asserts its identity, its solidity, its thingness. Dust on the screen, combined with incident light, will inhibit the looking into that the screened film yearns to generate. We can never quite make the TV screen go away. We are always looking at a bulky piece of hardware that is on the brink of gaining our attention. The TV set is uncomfortably close to being a piece of furniture—not an impalpable magic window onto another world. The kind of immersion characteristic of the cinema experience never quite occurs in the domestic embrace of the TV screen. We never quite enter the world of the film that is being broadcast as we do in the movie theatre. The typical television set is just too small to escape its identity as one object in the visual field among many—as just one of the things competing for our attention. By contrast, because of its sheer size, the movie screen can hardly be singled out within the visual field as one object among many; hence its capacity to assume the dimensions of a whole world. This is why we naturally say that we look at our television screens, but we don't find ourselves saying this about the movie screen. The TV apparatus is intrusive in a way that the movie apparatus is not.
Sculpture and architecture require little comment, though I think both can in a certain way quite easily evoke the looking- into response. Suppose a sculpture (say, a Henry Moore) has holes in it—aren't these the paradigm of things that can be looked into? Now I don't want to take a stand on whether a hole in a statue is really part of the statue, but I take it this case does nothing to show that sculpture itself is of its very nature a look-into medium, since clearly the solid bits are what sculpture is all about, and they obviously resist any attempt at looking into. Sculpture is all solidity and texture, and it cries out to be looked at (the only kind of sculpture you could look through would be made of transparent materials). Architecture, another three-dimensional visual art, can also invite looking into, as when you enter a cathedral and look into the space within, or look into a room through a window. But, again, this is a contingent aspect of it, drawing upon other looked-into media; it is not constitutive of architectural objects that their appreciation involves sustained and systematic looking into. Certainly, gazing at a building from the outside is a clear case of looking at; the building does not work to conduct your visual attention elsewhere. The stone facade of a building is nothing like a movie screen, so far as perception is concerned.
I conclude, then, that only cinema, among the visual arts, constitutively requires looking into as part of its proper appreciation. (I have not denied that it is possible to look at the movie image itself, even when it is flawless; but this is not the typical stance and detracts from the power of the film.) This isn't to say that there could not be an art form that also required such looking into—say, a type of artwork that involved placing objects in a watery medium—but as things st
and, cinema is distinguished by this perceptual phenomenon. A person who lacked the look-into ability could not fully enter into the movie experience, since he could not take his attention off the screen itself and its play of light images. But we normal humans are amply gifted with this capacity. Thanks to our experience with holes, the sky, the eyes, and so on, looking into things comes naturally to us.
WHAT ARE WE LOOKING AT?
What is it that we do look at when we are watching a movie? Let us consider a particular case—say, watching Citizen Kane. There is the image on the screen, a particular pattern of light; there is the actor Orson Welles, very much more substantial than a slice of light; and there is the character of Charles Foster Kane, who is by no means to be identified with Welles himself. We are not looking at the first element in this triad (though we are seeing it); so which of the other two is the object of our looking? Actually, we need to add a third possibility: nothing, no one—perhaps we are looking at nothing whatsoever, there is no object of our looking here. Not to keep the reader in suspense, I think that we are looking at Orson Welles, not at nothing and not at the fictional character he plays. We look through the image of Welles to the man himself—he is the thing we are looking at, that flesh-and-blood actor. It can't be that we are looking at nothing, since if we are looking into one thing we must thereby be looking at another. There is no pure looking into; all acts of looking into are accompanied by acts of looking at. And if there is an act of looking at, then there has to be an object of looking at. Nor, I think, is Charles Foster Kane the thing we are looking at, for we are not seeing him. It is not Kane who stands before the camera, causing the image that confronts us on the screen; it is Welles. Kane is an imaginary character. Our relation to Kane is mediated by our imagination, not the mere act of seeing. We see the image of Welles and we imagine Kane. You could, in principle, see the image of Welles and not imagine Kane, if you had no idea that this is a fictional work about an errant newspaper tycoon. The thing we are looking at, then, is the thing we are seeing; and this is Welles, not Kane. We don't need to imagine Welles in order for him to be an object of our acquaintance, since he is present to our mind by means of the photographic process: there is a causal chain leading from him to our visual experience in the cinema. But there is no such chain from Kane to our visual experience, because Kane was not being photographed when the film was made; only Welles was. You can't photograph fictional characters, only the real people who play them. Nor can fictional characters be literally looked at.
So, the right description of this complex visual relationship is as follows: the image on the screen is seen but not looked at; the actor is seen and looked at; the fictional character is neither seen nor looked at, but imagined. Different actors can play the same character, and when this happens many individuals will be looked at by the audience, not the single character they all play—though it is a single character that is imagined by all. By the same reasoning, if a movie uses a prop to stand in for a real object—say, a model city— then the thing the audience is looking at and seeing is the actual prop model, not the full-size city they imagine in connection with it. This has the consequence that audiences don't always know what they are looking at; they can make mistakes about this—they think it's a real city when it is just a model. What you are looking at and seeing is what was before the camera at the time of shooting, and you may have completely wrong ideas about what that object was.
RECIPROCAL SEEING
What is the precise relationship between the image seen and the actor (or object) seen? You see both when you watch a movie, but these are not unrelated acts of seeing; they are intimately connected. A notion often introduced in connection with pictorial perception is Ludwig Wittgenstein's notion of seeing as.3 The idea is that you see the image as Welles or as a man of such-and-such physical appearance. You don't just see the image as a patch of light; you see it as representative of a three-dimensional human being. Alternatively, we could say that you see Welles—the man—in the image on the screen. You invest the image with this real-life reference, so that the seeing has two aspects: what it is a seeing of (the image) and what it is a seeing as(a man). Thus the seeing has a kind of inner complexity, bringing two objects into close relation: the image and what it is an image of. It isn't just that you see the image andyou see the object it represents; you see the object by means of the image.
This seems to me a perfectly adequate way to capture the nature of the cinematic visual experience, but I want to add something to it, to enrich it in a certain way. To do this, I shall introduce the notion of reciprocal seeing. The point of this notion is that the seeing of the one thing essentially involves the seeing of the other—each seeing leads into the other one. The seeings are joined at the hip, as it were. When you see the image (in normal circumstances), you also and thereby see the object—you see Welles by seeing an image of him. Indeed, seeing the image enables you to see through to Welles. But it is equally true that your seeing Welles this way involves your seeing him under the aspect of an image. It is not that you see Welles just as you would if he were standing before you, unmediated by any image of him; rather, your seeing of him comes in another mode—that of image-mediated seeing. You see the image as Welles, but you also see Welles as presented by an image. The image leads you to the object, but the object leads you back to the image— hence reciprocal seeing. In Wittgenstein's famous example of the duck/rabbit drawing, you see the drawing as a duck, say, but in seeing this duck you see it as represented by an ambiguous drawing. That's the way the duck comes to you perceptually, and it is very different from the ordinary seeing of ducks. Each episode of seeing embeds the other. Similarly, in seeing the movie image, your seeing of the image embeds a seeing of the represented object, but your seeing of the object also embeds a seeing of the image. When you see Welles by means of his image, he comes to you very differently from an ordinary seeing of him, which is why you don't take yourself to have Welles in the theatre with you. We have indissoluble double seeing here, each seeing conditioned or modified by the other. And as I have said, one of these events of seeing is accompanied by an act of looking through while the other is accompanied by an act of looking at. It is not that each seeing is a barrier to the other, as if they were in competition for our visual resources; on the contrary, each seeing points toward the other. Abstractly put, I see X as Y and i see Y as X.
The image represents an object; the object does not represent an image (Welles the man is not a symbol standing for his image). So the representation relation is not symmetrical: it is not in general true that if X represents Y, Y represents X (my name represents or refers to me, but I do not represent or refer to my name). Now some kinds of representation stand for their objects in an entirely extrinsic way—it is not possible to perceive their reference just by perceiving them. Thus words have reference, but they do so extrinsically since you can see or hear a word and not see or hear the thing it stands for. This is basically because a word is a conventional symbol: the connection between word and object is arbitrary, not natural. The reference of a proper name, like “Horatio Alger,” is clearly not embedded in the name (considered as a mark or sound); you cannot gain perceptual access to the reference, a certain man, just by perceiving the name. But in the case of movie images (and maybe photographs in general) we do have embedded reference: by seeing the representation we do gain perceptual access to the reference—we see Welles just by seeing his image on the screen (though seeing his name on the screen gives no such access). This is because photographic images are not merely conventional or arbitrary. Notice that I am defining “embedded reference” in perceptual terms: we can see the reference of a movie image by seeing the image; I am not saying that the object is physically embedded in the image, as a piece of food might be embedded between my teeth. Welles the man is very distant from his image when I see him in it—he isn't somehow stuck inside the image, like a caged animal. But it is still true that I am made able to see him by means of the pre
sence of his image.
The notion of embedded reference enables us to recognize a significant feature of the movie-watching experience: that we are perceptually connected to objects in the real world and not merely to arbitrary symbols that flash before our eyes. It is emphatically not like hearing a language to which you must attach reference: it is not that you first see the images and then have to go about interpreting them according to some conventional scheme. It is rather that the object comes with the image: in seeing the one I thereby see the other. Picture recognition works very differently from word recognition. Once a child has learned how to recognize objects—to classify them in appropriate ways—she has nothing more to learn in recognizing pictures of objects: if she can recognize an object as an apple, she can also recognize something as a picture of an apple (once she has got the general hang of pictorial representation). The converse is also true: once you can recognize pictures of things, there is nothing further you need to master in order to recognize the things that are pictured—which is why a photograph can be decisive in tracking down a fugitive. But nothing like this holds for language or any other conventional symbolic system: you can be excellent at recognizing apples but have no ability to recognize that some sound you hear means apples. If you could make this step, learning foreign Ianguages would be a piece of cake! For that, you need to learn the language as an independent entity—a skill that goes far beyond the recognition of objects themselves. The relation between pictures and things is very different from the relation between words and things. Given the notion of embedded reference, this is just what we would expect, since the image embeds the object, so that the very same skill we use for recognizing the object will be employed in recognizing the picture of the object. By contrast, when the reference is not embedded, but extrinsic, as with words, we cannot use the object-recognition skill in their interpretation. Put simply, I can see you in a photograph of you, but I cannot see you just by looking at your name.
