Consciousness, Literature and the Arts

Archive

Volume 5 Number 3, December 2004

_______________________________________________________________

 Why photographs are not more realistic than paintings

by 

Justin Good

Emerson College

Introduction

Isn’t it obvious that if I take a photograph of what I am looking at - of what I can see - that the photograph is a more realistic, more accurate representation of my visual experience than if I draw or paint it? The obviousness of the intuition is an interesting aspect of the question. In part, it is interesting because it negates our supposedly more sophisticated understanding of what realism is and of the phenomena of perceptual ‘relativity.’ This is not just true of theorists of perception: how many times have you heard the story about how many Eskimo words there are for snow? If perception is as ‘relative’ as pop psychology and sociology make it out to be, then it should not strike us as obvious that the photograph more perfectly conveys what I see. Yet it does strike us as obvious, even when we feel, for phenomenological or artistic reasons, that it must be false.

In this paper, I am interested in showing the precise sense in which photography is more realistic than painting, but only relative to a certain epistemic or informational interest we can have in the image. I begin by addressing the question about realism by examining our conceptual criteria for calling a representation ‘realistic’. I relate this analysis of realism in terms of informational interests to the worry about the machine defining the purpose of visual representing; more specifically, a worry about how our interest in visual experience and visual images is defined, and so constrained, in terms of the purpose of seeing conceived as a biologically adapted information processing activity. Finally, I’ll suggest a reason why a certain artistic interest in perception cannot be conceptualized within the evolutionary-computational perspective. 

This last point relates to the peculiar importance that the question about realism has for some who are interested in art and visual perception. I’ve heard this question raised many times, especially by artists, as if it was a matter of some importance; as if their whole sense of their art and their relationship to visual experience were shaped by the answer, but without being able to say why, without being able to do anything further with the question. Part of this urgency might be attributed to the desire (or perhaps anxiety) to not be replaced by a machine (i.e. a camera or other instrument for mechanically reproducing experience) that can perform your task for you, and even do it better than you could. What I’m interested in is a kind of conceptual element of that anxiety. It has to do with a worry about the machine not replacing an artist, but about the machine defining the task or the purpose of representation, or the aim of the artistic practice, and doing so in a way limits the expressive possibilities of art.

Now there is a theory of visual perception which explains in a precise way just why and in what sense a photograph is more realistic than a painting. This theory, formed by the synthesis of computational and evolutionary psychology, conceives of seeing as a biologically-adapted information-processing activity. According to this view, the photograph counts as more realistic because it more faithfully records high-definition information about the physical environs, specifically the identity, position and location of entities. But this concept of perception is not what makes it obvious to us that  photography is more realistic than painting. At best, the theory “explains” the feeling of obviousness. We don’t find the photograph more realistic because we’ve studied that theory. Rather, the sense of obviousness has to do with our conceptual criteria being illustrated by the example. When the correct use of a concept, like realism or a predicate like is more realistic than strikes us as obvious, it is because we are speaking from the certainty of a criterion for conceptual meaning. And to call it obvious is to reassert that that is just what we mean by ‘realism.’ In turn, to question the obviousness is not to say that the judgment is false, but that there are other ways for measuring the realism of an image and that we forget or otherwise overlook these other ways when we start to think about the question. The artistic concern with the problem of realism relates to this overlooking in the following sense: part of the task of the artist is to uncover those overlooked measuring sticks, to resist the strictly biological view of visual truth.

Transparency

So what are our criteria for calling an image ‘realistic’? The first one has to do with the notion of transparency. Consider the example of Vincent van Gogh as he sits painting sunflowers. Now imagine two expressions of his visual experience: the first, a painting by van Gogh of what he sees, something like his painting Two Cut Flowers, which he painted in Paris in August 1887; the second, a photograph taken from the same position at the same time. Since nothing pertinent turns on historical accuracy, we can imagine a very high-quality photograph, e.g. an image produced by a 20.0 mega pixel digital camera. Which image is a more accurate expression of van Gogh’s visual experience and why? 

            I choose van Gogh here because he is known for his emphatic stylization, which sometimes provokes questions like: was it madness that caused him to paint like that? Or was it a problem with his eyes or his brain? Or did he consciously stylize what he saw? These questions are interesting because they presuppose an affirmative answer to the question whether a photograph as such is more realistic than a painting, as a report of visual experience. But why assume that photography is?

For Fox Talbot, who invented photography in the 1830s, the immediate appeal of the new medium lay in its ability to do something that could not be done without considerable draftsmanship. Not being able to draw, he called photography “nature’s pencil”.[1] Photography was seen, and can still been seen, as the mechanical perfection of a representational and aesthetic task long assigned to the craft of painting. The usual story is that as photography led to the perfection of the long-standing Renaissance ideal of pictorial transparency in painting – the attempt to capture the ‘purely visual’ aspects of experience – painting could only continue by giving up on transparency by consciously distorting the image. There’s some truth to this story, but everything turns on how one understands the idea of transparency.

            One way to unpack the idea of transparency is to draw a distinction between the content of an image, or what it represents, and the style or design of the representation, or how it represents. Transparency then amounts to the condition whereby an image represents a content without employing a style, or mode of representation which distorts or comments upon the content. Now the idea of a mode of representation has, minimally, a technical as well as a psychological interpretation. I’ll address each of these aspects in order.

Speaking of the technical sense of style, Marshall McLuhan remarks that photography has no syntax, unlike drawing which receives its syntax from the human hand.[2] Unlike the complicated task of translating three-dimensional objects in space via rules of linear perspective to a two-dimensional surface, no grammatical rules are followed or need to be followed when “writing light” – the literal meaning of photography - to a photo-sensitive surface. In this sense, the distorting feature of style, from the Latin stylus, has to do with the marks left by the device used to create the image; marks which have no semantic meaning (don’t refer to anything), and so get in the way of transparency.

            The technical sense of style is related to the psychological sense. Style can also be defined in terms of the user of the medium. In contrast to the content of an image, which tells us something about the world, style has to do with the perceiver’s (artist’s) attitude towards what she is perceiving. Hence, we get a sense of the obsessive character of van Gogh’s interest in the night sky from the way his brush insistently encircles the stars in Starry Night. We might say that the style of the painting tells us more about van Gogh than it does about the sky. To use Mallarmé’s distinction, style registers the effect of what is seen on the artist, and not simply what is seen.

            As Daniel Dennett puts it, the importance of the camera for science is that cameras are ‘stupid:’

In order to ‘capture’ the data represented in its products, it does not have to understand its subject in the way a human artist or illustrator must. It thus passes along an unedited, uncontaminated, unbiased but still re-represented version of reality to the faculties that are

equipped to analyze, and ultimately understand the phenomena.[3]

This view of the greater objectivity of the photograph over the illustration or painting fits nicely into the dominant naturalistic concept of seeing as an information-processing process which has been adaptively shaped by evolutionary forces. While the purpose of seeing is fundamentally biological survival, the content of perception is defined in terms of picking-up information relevant to that aim which has been chosen for us by Mother Nature, so to speak. Photographs are more realistic, and more accurate reports of what we see because they carry higher grade information about the environment. They are causally responsive to the environment in ways that paintings are not.

For example, photography for the first time made accessible to human examination certain complicated temporal phenomena that could not even be perceived, let alone painted. For example, photography is what allowed Eadweard Muybridge to confirm Leland Stanford’s contention, which was controversial at the time, that there are moments at full gallop when all of a horse’s feet are completely off the ground.[4] The fact that photography could be used in this way to show that a long-standing pictorial convention which had painters always leaving a foot or two on the ground, illustrates a sense in which photography is more accurate or realistic than painting.

            The transparent character of photography is often illustrated with reference to the instantaneous character of ‘taking’ a photograph. Because painting is a temporally-extended activity, taking hours and even weeks or months, it cannot capture the ‘moment’ of perception in the way that photography can. While a moment in time can be painted, it can’t be painted in a moment. This gives photographs, and not paintings, a unique reference to the moment in time when they were snapped. And doesn’t that make photography a more ‘transparent’ medium?

Indiscernability

It is easy to get the wrong idea here about what the higher informational quality of photography implies for how we evaluate the truth or content of visual representations. Like any medium, photography involves informational biases that emphasize certain aspects of what is perceived and ignore others.  These biases are evident in the fact that there is no question of confusing a photographic reproduction with visual experience itself. Despite its relatively high-definition information, a photograph of a sunflower is easily distinguished from our visual experience of the sunflower. An image (or sculptural rendering) of a sunflower that was indistinguishable from the sunflower itself could not even be said to be a picture at all, but a replica of a sunflower. As Rudolf Arnheim puts it, for an image to even be a picture it must employ some degree of abstraction, otherwise the object being represented ends up getting replicated instead.[5]

The biases of photography have long been noted. For one, a photograph does not offer an exact transcription of the information encoded in the varying light intensities constituting a visual field, but rather a translation of one scale of intensities onto another. The fidelity of the photograph lies in the way it preserves the ratios between the lightest and darkest points in the visual field, not on whether the darkest point in your visual field has the same light intensity, chromatic character, tone, etc as the darkest point in the photograph.[6] Photographs also have a crisp edge where they end, making them unlike your visual field which has no crisp edge, no edge at all really.[7] And of course, a camera also has to be pointed, and decisions have to be made about the distance, angle, lighting, timing, lens, etc, that can result in endless variations concerning which aspects of the object or scene are emphasized and which are thrust into the background.

            Likewise, the instantaneous character of photography is as much a bias as it is an informational enhancement.   A photograph doesn’t just catch a moment in time, it redefines what counts as ‘the moment’. One tends to represent visual experiences that are amenable the available means of expression. For example, holding a camera can cause one to notice momentary visual compositions, like a shadow cast on a sidewalk, that one would never think to notice if one was looking for something to draw. Correlatively, the camera is blind to other ways in which moments in time are experienced, or other ways in which the world appears suspended temporally in shape of a scenario.

            There is an important sense in which the instantaneous character of photography makes the medium less realistic, and more abstract, than say a drawing in pencil or ink. James Elkins makes this point in talking about why people usually do not like to see pictures of themselves. True, vanity plays a part in our disappointment, but there’s a reason connected to the medium itself.

Photographs clip out instants in time, and since we see in overlapping moments and usually base our sense of a person on a fluid sequence of moments and motions, a single photograph can often seem wrong. (Painters blend moments, so that few oil portraits have

the weirdness of snapshots.)[8]

If time were merely a sequence of Nows, then the momentariness of photograph wouldn’t have this bias, would not distort the ‘information’ in this way. Perhaps we would prefer photographs of ourselves more, but then we would have an altogether different kind of identity and existence.

These biases, however, are only part of the reason why the objectivity of photography is not guaranteed by the medium itself. It is sometimes said that photographs do not lie, but pictorial truth is not a property of the image itself. Realism does not turn merely on the structure of an image, but on what that image is used to show. Speaking of the accuracy of a visual image presupposes a method of projection which specifies what the image is about. Now the statement that photographs do not lie is not wrong per se, since we can think of occasions when it would make sense to say so. Think of the ways that a private detective might use an incriminating photograph as evidence of adultery; here the lies of the adulterer are contrasted with the truth of the photographs. But the same picture, say, of a man and woman having sex, can only be called veridical if it is understood what the photograph is about; in this case, evidence of adultery. The photograph is not essentially a photograph of an adulterous act. If the photograph is interpreted as an image portraying the lovers at some point in the past, before they had even met, then it might very well turn out to be false. In this sense, pictures neither lie nor tell the truth, but only particular uses of pictures do. That pictures require interpretation is often overlooked just because often it is understood how a picture is meant to be taken. Like all images, photographs have no intrinsic intentionality, no essential aboutness.

The point here can be understood as a comment on the original question about the greater realism of photography. For the question asks about the realism of photography as such, not for example, a particular photograph with reference to a particular question of fact. But if images have to be interpreted, then you cannot determine what a particular medium will allow you to express about a visual experience until you look to see how it is used, or can be used, to express an experience.

            That images need to be interpreted does not mean that realism in the strict sense is impossible or incoherent. The point needs to be made because one might be tempted to dismiss the question about photography by saying that no medium can be a ‘truly’ realistic expression of my experience. This impossibility is sometimes expressed by saying that experience is ineffable. For isn’t it true that anything I could give as an expression of my experience leaves something out, or could be misconstrued? But there’s an unexpressed assumption here that needs to be questioned. To say that every visual (and for that matter, non-visual or linguistic) expression of my experience as such fails at perfect realism presupposes a questionable criterion: namely, that an image is a realistic portrayal of an experience if the image is indiscernible from the experience itself. Since any ordinary image we can point to can be distinguished easily from the experience it is meant to report, they can all be said to fail in achieving complete veracity.

But there are confusions involved with this criterion for realism. It is instructive to observe that if we were to insist on this criterion of accuracy, we would have to discount just about everything that conventionally count as - that we are ordinary justified in calling – an ‘accurate representation.’ That’s a bit like defining ‘shoe’ with reference to something only worn by an obscure clan of Buddhist monks in Tibet, such that nothing we would ordinarily call ‘shoe’ counts as one. In fact, a photograph as well as a painting can count as an accurate representation of something without perfectly resembling what it is of. To say of a beautiful line drawing that it is a remarkable likeness of someone does not mean that the person in reality looks to be made of lines or pencil marks. Just think of occasions when, looking at a beautifully crafted portrait drawing you exclaim, “That looks just like her!” If you define visual accuracy in terms of indistinguishability then such remarks are certainly wrong, but there’s truth to be said in saying that a drawn portrait looks exactly like the person herself. The point is that accuracy or realism is relative to our informational

needs or interests. As Gombrich points out in his discussion of stereotypes from Art and Illusion,

To say of a drawing that it is a correct view of Tivoli does not mean, of course, that Tivoli is bounded by wiry lines. It means that those who understand the notation will derive no false information from the drawing – whether it gives the contour in a few lines or picks out

‘every blade of grass’...[9]

This might sound a bit trite. After all, there’s nothing particularly controversial anymore about the relativity of pictorial accuracy. But the commonsense idea that photographs are more ‘objective’ than drawings rests upon a failure to acknowledge just this kind of relativity.

Informational interests

There are other assumptions in play, however, that undercut reflections on informational relativity and motivate the judgment that photographs capture more accurately or objectively our perceptual experience. Despite the fact that pictures do need to be interpreted, and that interpretation is shaped by one’s informational interests, one might still hold onto the idea that a photograph is more accurate than a drawing or painting because it is richer informationally. Thus, instead of defining an accurate picture as one which is indiscernible from the visual experience itself, one

could define it in terms of how many informational interests it potentially serves. Gombrich again:

The complete portrait might be the one which gives as much correct information about the

spot as we would obtain if we looked at it from the very spot where the artist stood.[10]

For example, we might say that a photograph of a sunflower is more accurate than a van Gogh painting of one because it can be put to more informational uses than the painting.

There are two basic problems with this however. The first is connected with the idea of a perfectly informative representation, approximation to which would serve as a measure of its accuracy. It is not clear at all what could be meant by a representation that could address any possible epistemological hunger we might bring to it. The idea of a ‘god’s eye picture’ suffers the same conceptual difficulties of a map that, because it includes every conceivable detail of the terrain it represents, ends up being the same size as the terrain, and so useless as a map.

            The second problem has to do with the fact mentioned above that all media, hence all representations produced by a medium, involve informational trade-offs. It is implausible to assume that one might be able to derive from a photograph of a sunflower the particular kind of ‘information’ that van Gogh was interested in portraying in his paintings. Without seeing his paintings, we would have no idea which visual properties of sunflowers had peeked van Gogh’s interests. A photograph may be more help to a botanist than a painting, because it gives, for example, details about the flowers that van Gogh was either not interested in or simply failed to see. But neither can the information contained in van Gogh’s painting be extracted from a photograph.

            At this point, one might raise the objection that one cannot speak of a van Gogh painting containing ‘information’ about certain visual properties of sunflowers, since information is relative to epistemic need, and paintings don’t seem to satisfy any epistemic needs, at least any we could specify in any precise sense. As the objection might go, if we can speak of paintings as expressing certain informational interests, these interests are subjective, and concern the effect of the visual appearance of the flowers on van Gogh, and not anything objective about the flowers themselves.  But what can ‘objective’ mean here? It cannot mean that the photographic medium introduces no biases, or that it more closely approximates indiscernability to the experience itself, or that it is per se more informationally relevant than the painting. If the objection is that the information contained in the photograph is more objective because it is not filtered through van Gogh’s mind and his arms and his particular interests in sunflowers, then the response is that ‘more objective’ is relative to informational interests which are not met by information which is filtered in that way. Conversely, it makes equal sense to say that the way in which the visual appearance of the sunflower is filtered through the camera is informationally irrelevant to van Gogh’s particular interests, hence ‘less objective’ than his painting of sunflowers.

So what might van Gogh’s interests in sunflowers have been such that his painting serves to capture the ‘information’ relevant to them? As the art critic Robert Hughes argues, the romanticism of van Gogh’s stormy life, especially his madness, tends to distort our appreciation of the sanity of his paintings. Far from being a projection of his psychosis onto the objects he saw, van Gogh’s style is rightly seen as a systematic device for capturing the forms that caught his visual imagination.

The notion that [van Gogh’s] paintings were ‘mad’ is the most idiotic of all impediments to understanding them. It was van Gogh’s madness that prevented him from working... As a draughtsman, van Gogh was obsessively interested in stylistic coherence. Just as one can see the very movements of his brush imitating the microforms of nature – the crawling striations of a gnarled olive trunk, the ‘Chinese’ contortions of weathered limestone – so the drawings break down the pattern of landscape and reestablish it in terms of a varied,

but still codified system of marks: dots, dash, stroke, slash.[11]

It is rather empty to characterize his interest here as an interest the visual properties of the flower that give him the particular aesthetic experience he has while looking at them. There is no more direct characterization of what van Gogh was interested in than his painting itself. The painting, we could say, is a logical criterion, which determines the kind of experience we can attribute to him.[12] If we want to say what van Gogh’s interests were, which is to say, what kinds or aspects of visual truth he wanted to render, we point to the painting and say, ‘That’s what he was interested in.’ In this sense, there is no more direct or accurate expression of what van Gogh saw than his paintings. A photograph would be less realistic.

 What is visual form?

The larger point I want to make, however, is that there is something conceptually off key in speaking of van Gogh’s ‘informational’ interests. Yes, van Gogh is interested in seeing, but that doesn’t necessarily mean that he is interested in gathering information about the environment. The idea that the purpose of seeing is to gather information about the environment in ways which have proved adaptively advantageous in the evolutionary past is of a piece with the general evolutionary-psychological view of the human mind as an adaptation, or series of modular adaptations. But the relevance of that concept of seeing for perceptual psychology or the philosophy of mind does not entail the relevance of that concept for every interest we can adopt towards our visual experience. That concept understands the function of seeing as gathering knowledge, about the identity and location of objects, relative to our cognitive interests as biologically-evolved organisms, via the computational processing of information encoded in visible light. And while this might very well be the most appropriate concept of perception for purposes of the naturalistic study of sight, we need not assume that sight had that function for van Gogh as an artist.

            Yet it is very tempting to assume just that. If one thinks about perception in a broadly naturalistic spirit, then it is a natural thought to look for ways to ground aesthetics and art theory in cognitive psychology. For example, the neurologist Semir Zeki has proposed that our understanding of the purpose of art must be grounded in the neurobiology of the brain. As he puts

it,

All visual art is expressed through the brain, whether in conception, execution or appreciation and no theory of aesthetics that is not substantially based on the activity of

the brain is ever likely to be complete, let alone profound.[13]

While I do not wish to deny that a ‘biological aesthetics’ will shed light on art or aesthetic experience, I find the presumption that aesthetics must be based on neurobiology to be a particularly silly form of scientism. Aesthetics needs to be pursued with a more nuanced concept, or set of concepts, of what it means to see.

So how do we nuance the concept of seeing? If not the information-processing definition of perception, then how should we conceive of van Gogh’s artistic interest? To say that his interest is in the aesthetics of visual form is not so much wrong as it is unhelpful. Beauty too easily suggests a subjective coloring of perceptual experience.[14] Characterizing artistic interests in perception as aesthetic interests overlooks the extent to which art has separated itself from the concept of beauty, at least since Marcel Duchamp invented the idea of ready-made sculptures back in the 1920s.[15] This is where Paul Valery’s definition of seeing as forgetting the name of the thing seen helps to indicate a conceptual difference. What this concept captures about artistic interest in perception is the idea that seeing is not something that we do automatically but rather a state which needs to be achieved.[16] It is the idea that we are for the most part blind to our visual environs precisely because we are so well adapted to them, and that because we know what we are seeing, we do not have to look at it, and hence, do not see it. In this sense, realism is not what is necessarily captured in a photograph (although it might very well be so captured). It is what manages to represent the form that served as the impetus to create the image in the first place. On Valery’s concept, visual form is not information that Mother Nature has designed our visual system to pick-up from the world, it is what emerges into our visual awareness on those often rare moments when we suddenly feel like we are seeing something about the world after having been blind to it.

The accuracy of an image in this sense turns on the power of the image to engage the viewer in the same ways that the artist herself was engaged when she saw the form. This notion connects two common sayings about successful works of art. The first is the idea that successful works are ones that have the power create that sense of seeing as an achievement. As Hughes describes it,

It is a characteristic of great painting that no matter how many times it has been cloned, reproduced and postcarded, it can restore itself as an immediate utterance with the force

of strangeness when seen in the original.[17]

The second idea is that great artworks end up shaping our very experience of the original subject-matter, as when we think of how van Gogh saw sunflowers when we ourselves see sunflowers.

            Now it is not clear whether this concept of seeing could to be of any value for perceptual psychology or the philosophy of mind. I hope it is clear by now, however, that its validity in no way depends upon its relevance for those theoretical projects. It is only a way to understand the limitations of the conceptual reconstructions employed by those endeavors for our broader understanding of what it means to see. The distinction between the art and science of perception is often posed as an unproblematic datum in the most stale of dualisms. Cognitive psychology studies perception ‘objectively’ as a natural process or ‘natural kind’, whereas artists explore the ‘subjective’ side. Or scientists and philosophers of mind study vision, whereas artists and art historians study ‘visuality,’ or the historically constructed experience of sight.[18] But these facile dualisms reflect the epistemic narrow-mindedness that C. P. Snow called the two cultures. Without some broader conceptual map of perception, we are liable to end up with an understanding of sight which has been considerably, and unnecessarily narrowed down, and a view of art which has dimmed down its expressive possibilities to ones theoretically explicable within the rubric of naturalism.

Of course, this leaves the question: why would anyone be interested in forgetting the name of the thing seen? Whatever the answer is, there is no a priori reason to think that the question can be naturalized, or more importantly, that our understanding of the aims of art requires that it must be naturalized. More likely, it is the task of artworks themselves to answer.

Bibliography

Arnheim, R. (1969), Visual Thinking. Berkeley: University of California Press.

Danto, A. (2003), The Abuse of Beauty: Aesthetics and the Concept of Art. Chicago: Open Court.

Dennett, D. (1996), Kinds of Minds: Towards an Understanding of Consciousness. New York: Basic Books.

Elkins, J. (1997), The Object Stares Back: On the Nature of Seeing. New York: Harvest Books.

Foster, H. (1988), in Vision and Visuality (ed.). Seattle: Bay Press

Gombrich, E. H. (2000), Art and Illusion: A Study in the Psychology of Pictorial Representation. Hong Kong: Princeton University Press.

Hughes, R. (1990), Nothing If Not Critical: Selected Essays on Art and Artists. New York: Alfred Knopf.

McLuhan, M. (1964), Understanding Media: The Extensions of Man. New York: Signet Books.

Pinker, S. (1997), How the Mind Works. New York: W. W. Norton & Co.

Wittgenstein, L. (1980), Remarks On the Philosophy of Psychology, vol. 1, trans G. E. M. Anscombe, ed. G. E. M. Anscombe and G. H. von Wright. Chicago: University of Chicago Press.

Zeki, S. (1999), Inner Vision: An Exploration of Art and the Brain. Oxford: Oxford University Press.