1. Authentic Photography

The idea of allowing the world to imprint its own image is central to photography's self-conception. Perhaps this idea forms just one strand in that conception. Perhaps it is not even consistent with some of the other strands. Nonetheless, the strand is present in the thinking of many photographers and theorists. Rosalind Kraus articulates it clearly:

But what is it for a picture to be an ‘imprint’? I propose the following:

By ‘content’ here I mean depictive content: the way the picture depicts the world as being. For present purposes, we can treat this as content that is distinctively visual: a picture depicts whatever features or things it shows us, looks like, or which can be seen in it. If a photograph of a white dove symbolizes peace, then whiteness, the dove, and peace all figure in its contents, broadly construed. However, it depicts only whiteness and the dove, since peace is not something it shows us or looks like, nor is it something we can see in it. By a ‘scene’ I mean whatever object, event, or array serves as input to the relevant process. In the case of photography, this is whatever is the source of the light that sets that process going. But what about a ‘chain of mind-independent sufficient causes’?

Where one thing imprints another, the two must be connected by a causal chain. That chain must be composed of links each of which is a sufficient cause of the next. And each link must be mind-independent in the following sense: the causal factors composing it do not themselves involve anyone's mental states or actions. For sure, those factors may be as they are thanks to what someone does—imprinting is consistent with intervention. However, that intervention is limited to causing the causes to be as they are. The causes themselves neither are, nor essentially involve, our actions or mental states. (Compare Kendall Walton's ‘natural counterfactual dependence’ [Walton Reference Walton1984: 262–5].)

To illustrate, consider a simple form of imprinting: taking a fingerprint. This process is often highly dependent on people's actions and intentions. Paper and ink may be present only because a police officer provides them, and my fingers may press down hard enough to produce a clean image only because I follow her instructions. Nonetheless, normally the process involves causes that are mind-independent in the relevant sense. Had the paper got there by other means or had my fingers exerted the right pressure thanks to gravity alone, each would nonetheless have played its role in forming the imprint. Agency is exhausted in causing the relevant causal factors to be as they are. Contrast the situation if the officer uses a pen to retouch the print. Even if the marks she then makes are there because they match the pattern on my fingertip, the chain between them goes through her actions and mental states. The marks are causally connected to the whorls only via her seeing the latter, wanting the print and the pattern to match, and adjusting one to the other.

While the causal chain involved in photography is much more complex, the basics are the same. Here the output is a picture with its content, how it shows the scene to be. Can that content be traced back to the scene via a chain of mind-independent sufficient causes? Of course, the photographic process is usually packed with interventions, in shooting, processing, and printing or projecting the finished picture (Snyder and Walsh Allen Reference Snyder and Allen1975). Those interventions often crucially affect content. The photographer chooses what is before the lens when the shutter opens, and so chooses which scene the photo captures. She determines how the camera is set up, and thereby which of the scene's features will show up in the picture: for example, a longer exposure will capture longer lasting events, and a narrower aperture will capture detail across a greater depth of field. Her choices in the darkroom play a similar role, dictating, for instance, whether detail visible in the finished image is from brightly lit or heavily shadowed portions of the scene. And choices in presenting also matter: printing in black and white will efface any color captured in the negative or image file; and printing rather than projecting, or printing onto one surface rather than another, may reduce the contrast between light and dark. Given the differences these and many other interventions make, it's no surprise that photographers often take a keen interest in every stage in the production of the picture. However, none of these interventions need preclude imprinting. In each, the photographer's role might reduce to causing the underlying causal process to be as it is. In photography, as in fingerprinting, a great deal of intervention is compatible with imprinting.

Not all intervention, however. If the image is touched up in the darkroom or manipulated digitally in certain ways, or combined with others in photomontage, the resulting picture (at least as a whole) is not an imprint of the world. In these cases some of what the picture depicts can be explained only by reference to the actions and intentions of the picture maker. Even if the resulting photograph captures features of the scene before the lens, it does so only because of her desire to present it accurately, her belief or experience that the scene was that way, and her intervention in the process of picture making to generate that result. The chain of sufficient causes leading back from image to scene contains links that are not mind-independent.Footnote ¹

Thus some photographic practice involves making pictures by imprinting, and some does not. Where it does, I will call it ‘authentic’. This is just a name—there is no presumption that photography ought to be authentic, that it will be better photography, or better art, if it is. The term serves only to pick out photography that is true to its self-image.

Note two last points. First, authentic photography does allow the photographer a form of control over content. Above I gave various examples of interventions compatible with imprinting that affect what the photograph depicts. Combining these may allow the photographer considerable control over her picture's content. However, the control on offer is second order: she controls content by causing the mind-independent chain of sufficient causes to be as it is. What imprinting denies her is first-order control: altering content by herself entering the chain of sufficient causes. (First-order control should not be confused with direct control: in neither photography nor painting can the artist bring about content without doing something else, such as making marks on a surface.) Second, nothing above implies that the content of authentic photography will always be accurate. Causation via a chain of mind-independent sufficient causes might easily introduce systematic distortions, for example, by watering down colors or deforming certain shapes. Photography's being authentic is no guarantee that the camera will not ‘lie’.Footnote ²

2. Communicative Representational Art

We now know what authentic photography is. What about my other key notion?

Where there is art, something engages our artistic interest. (This is not intended to explain art in other terms. I have nothing to say about what artistic interest might be, other than that it is the interest characteristically engaged by art. The claim is only that the idea of art and that of artistic interest go together.) Where there is representational art, there had better be representation, and its presence had better make a difference to what interests us. Where representational art is communicative, what interests us (and what representation furthers) is the communication of thoughts.Footnote ³

Such art communicates thought by representing things—in the case of paintings, drawings and photographs by depicting them; in the case of literature by representing them in language; in the case of sculpture by representing them in whatever way sculpture standardly does; in the case of some music, by using musical representation; in the case of film, theatre and opera by deploying a range of the above (depiction, language, the imitation of gesture, the resources of music). As these examples suggest, the notion of representation here is very broad, and that breadth is inherited by the notion of communicative representational art. The thoughts such art communicates often concern the people, objects, and events represented, but often they also concern wider themes that those represented things illustrate, embody, allude to, or reflect. Either way, representation and the communication of thought are distinct: the former is the means to the latter.Footnote ⁴ There can be representation without communication, and vice versa. Both can be present in the absence of art. But where the communication of thought by representing things engages artistic interest, we have communicative representational art.

The definition refers to the ‘resources’ of representation. What are these? Since the notion of representation is very broad, a general answer will be very schematic. Still, thinking schematically, those resources are fourfold: content, vehicle properties, relations between those two, and means of production.

Wherever there is representation, there is content. No doubt there are different forms of content, varying with the kinds of representation involved. While linguistic representation generally has propositional content, pictorial representation arguably does not: it may represent its objects as bearing certain properties, but it does not represent that they are that way. Some representation does not predicate properties at all, but represents an object simpliciter (think of a simple symbol such as the flag of a country). Whatever the differences here, to represent is always to represent something. The content of a representation is simply whatever it represents.

However, there can be no pure content. Wherever there is content, the representation that bears it has other properties. Since the representation is the vehicle for delivering this content, we may as well call these vehicle properties. The vehicle properties of a photograph include its size, the distribution of colors on its surface, and the materials out of which it is made; those of a novel include the words that compose it and the order in which they occur; those of a theatrical performance include the actors’ gestures, the sounds they make, and the appearance of the set.

Since vehicle properties are simply any property of the representation other than content, many of them are irrelevant to appreciation. The weight of a picture, the total number of letters in a novel, or the start time of a theatrical performance are not, in general, the sorts of property that affect the work's artistic interest. However, though some vehicle properties are artistically irrelevant, some are not. The words chosen to compose a poem, the working of the oil on a canvas, and the precise movement of a dancer's body can all clearly be relevant to the representation's interest as art. Thus, while not every vehicle property is a resource to be exploited in making representational art, many are. Indeed, there are perhaps no limits in principle to which can be exploited in this way.

These two resources bring a third in their wake. If content and vehicle properties are always present where representation is, so is the possibility of exploiting the relations between them. The representation that bears a given content might exhibit other properties that are interestingly related to that content. The sentences of a memoir capturing a child's first faltering steps might lengthen with the forays they describe; the crux of a series of events in a graphic novel might be presented in a frame larger than, and partly occluding, the others; the leitmotif representing a character in opera might return in ever more fragmented form as his emotional state degenerates. Representational art might communicate thought by tapping such relations in these or more subtle forms. Since this phenomenon will loom large in what's to come, it will help to have a name for it. I’ll call the relations between content and vehicle properties interplay.

The fourth resource is a little different. Representations have to be made. Some kinds of representation are made in distinctive ways. Oil paintings, for instance, are made by laying oil paint down on a flat surface; authentic photographs are made by allowing a scene to imprint its own image; carved sculpture is made by excising material from a block. Where the means of making is known by both artist and audience, it is possible for that means itself to become a resource for communication. The particular work can highlight the peculiarities of its making in a way that plays into the communication of thought.

An example illustrates all four resources. Consider Abram Arkhipov's Laundresses (c.1898, State Russian Museum, St. Petersburg). The painting shows the workers busy in the laundry. It moves from detailed realism toward the rear to a strongly impressionistic style in the foreground. (This is not well preserved in reproduction.) The water-soaked floor is conveyed in paint that still looks wet and is as smooth as liquid lying on a surface. And at the left is a swathe of red paint depicting a soaked garment. The red patch is almost completely formless. Only in the context of the rest of the picture can we can recognize it as the depiction of sodden clothing.

The painting communicates various thoughts about the scene—for instance, that the labor is hard and that it anonymizes those who perform it. It does so in part by virtue of what it depicts. The point about anonymity, for instance, is made in part by the fact that two of the three faces visible belong to girls who look very much alike: sisters, perhaps even twins. But content does not do all the communicative work. Vehicle properties also contribute. The wet appearance of the paint depicting the floor reinforces our sense of the hot and humid, perhaps even treacherous, conditions. Interplay between content and vehicle properties also makes a difference. The formlessness of the red patch provides a visual analogy for the heavy shapelessness of the sodden clothing it depicts. Wet clothing is hard to handle, as the paint is hard to resolve into something that makes visual sense, and sensitivity to this parallel reinforces our sense of the workers’ efforts. Perhaps the painting goes further still. Perhaps it invites us to see Arkhipov's own work as analogous to the labor of the laundresses. As he lays down the oils on the surface so that out of their wet mass a scene emerges, so they wring from the wet, shapeless lumps emerging from the tubs, dry, light, and precisely cut items of clothing. If so, the picture also exploits the facts of its own making to convey thoughts about that process.

The Arkhipov may be an unusually clear example of painting's exploiting the full range of representational resources, but many others could be given. (See, for instance, almost any of the examples in Podro Reference Podro1998.) What about authentic photography? (C) claims that authentic photography is more limited than other arts in the degree to which it can develop as communicative representational art. One way to investigate that idea is to ask whether in communicating thought photography and painting have access to the same range of resources.

3. The Challenge Posed

Which of the resources of representation is authentic photography able to exploit? Some have worried, in effect, that it is limited to the picture's own content (Scruton Reference Scruton1981: 590; cf. Susan Sontag's comment that ‘in photography the subject matter always pushes through’ [1977: 93]). This claim gains at least a little plausibility if we clarify what content (which here, remember, means depictive content) includes. It is not merely a matter of which things were photographed. Content also includes the features the photograph ascribes to those things, the angle from which they are presented, and the way the picture frames them. (Framing is just a matter of which slice of the world before the lens makes it into the picture.) However, even given this clarification, the suggestion that content is the only resource authentic photography exploits is plainly false. At the very least, some photographs communicate thoughts by exploiting relations between their contents and those of other pictures. Others make communicative hay by tapping the parallel relations between their vehicle properties, such as the grain and palette of the printed surface, and those of other photographs. And a good deal of work exploits photography's distinctive means of making, thereby communicating thoughts about various topics, including the nature of photography itself, the quest for evidence, and the relations between photographer and sitter.Footnote ⁵

There is, however, one resource widely exploited in painting that photography can tap only to a very limited degree: interplay between content and vehicle properties. In this section I first argue for this limitation and then explain its significance. The result is an argument with five premises, the conclusion of which is (C).

If we are to use some feature of a system of representation to communicate thought, that feature must lie under our control. For communication involves more than merely getting others to grasp the thought intended. They must also recognize your intention: they must grasp the thought by recognizing that it is the thought you want them to grasp. (This is the view of communication articulated by Paul Grice [Grice Reference Grice1989]. Scruton also appeals to it.) For them to do this, you must be able to make your intentions manifest. Your product must vary with your communicative goals, so that, recognizing that the thing is as it is because you intended it so, your audience can ask themselves why you gave it these features, and thus work out what thought you intend to convey. Applying this to interplay yields my first premise:

1. (1) If interplay is to play a role in the communication of thought, and thus in communicative representational art, it must lie under the artist's control in a way that appreciators can detect.

What might manifest control over interplay involve? Interplay is a matter of how content and vehicle properties relate. Suitable control thus requires controlling vehicle properties independently of content. If the artist can vary vehicle properties only by varying content, no one will be able to work out whether the picture combines that content with those vehicle properties because she wanted them to be related in that way, because she wanted the picture to bear that content, or because she wanted it to have those vehicle properties.Footnote ⁶ What are the properties that might vary independently of content in this way?

We can divide the vehicle properties of pictures, be they photographs or paintings, into three groups. Some vehicle properties determine what content the picture has, and some do not. The content-determining properties usually include the lines that compose the picture, what colors lie where on its surface, the boundaries between its colored areas, and the like. For instance, it is because the lines composing a drawing take the form they do that it shows a man with a cane. But not all vehicle properties fix what the picture depicts. These others themselves divide, depending on whether they vary across different parts of the picture. Local content-neutral properties so vary. The brushwork in a painting, for instance, can be finer and less prominent in some places, thicker and more visible in others, without those features making any contribution to what is depicted. Global content-neutral properties, in contrast, are common to the picture as a whole. Examples might be the size of a drawing, or the gloss or matte finish of a photographic print.

To manifest control of interplay, the artist must be able to vary properties of these three kinds independently of content. Hence:

1. (2) To have suitable control of interplay, the artist must control vehicle properties independently of content. The candidates are content-determining, local content-neutral, and global content-neutral properties.

Here the painter is fortunate. For a given content, she has the requisite control over properties from each category. This should be obvious for both kinds of content-neutral property. She can vary such local properties as the fineness of the brushstrokes without necessarily affecting what her work depicts. Similarly, she can alter global properties without affecting content: for instance, producing larger or smaller, or more or less impasto, versions of a given work. For sure, for any content-neutral property, there may be limits to how far it can vary without affecting content. If, for instance, the brushstrokes become too prominent, the lines defining the depicted object may become blurred, and the picture may no longer depict details of the object's shape. Nonetheless, within limits these properties and content will vary independently.

Although less obvious, the point also holds for content-determining properties. These determine content only in the sense that it depends on them: for content to be different, they would have to be. The relation does not run the other way round: it is not true that if the content-determining properties were different, content would have to be. Rather, various content-determining properties can serve equally well to depict precisely the same scene. There are, for instance, many ways to draw a simple stick figure of a man with a cane: in oils, in ink, with chalk on slate, or carving into wood; marking boundaries with the edges of colored areas or by drawing lines; drawing lines by incising into painted ground or by laying new paint down on top; making those lines thick or thin, rough-edged or finely drawn. To depict the man, the picture must be composed of some pattern of the right shape. Any features that make up that pattern are thus content-determining properties. But all these variations on those features serve that end equally well. The painter can choose between these different ways to determine the same content—even if by that we mean content that matches in every detail. (The examples here in part exploit the fact that, as I use the term, ‘painting’ refers to any way of making pictures by hand, regardless of medium. Note, though, that the point stands even if we restrict attention to a single medium—as shown by the different ways of using paint to depict an object's boundaries.)

In sum, the painter has extensive control over all three kinds of vehicle property, independently of content, and so enjoys extensive and manifest control over interplay:

1. (3) Painting offers the artist this control in respect of all three candidate properties.

We might expect painting to use this resource to communicate thought. The Arkhipov will not be unusual.

What of photography? Can the authentic photographer, like the painter, vary properties of any of the three kinds above without varying content?

For one of the three kinds, the answer is ‘yes’. Global content-neutral properties, such as the size of the printed photograph or the gloss or matte quality of its surface, lie within the photographer's control and are independent of the picture's content. Whether the image is printed (or projected) large or small, and whether it is printed on paper with a high reflectance or otherwise, make no difference to what it depicts. Moreover, the photographer can exploit these resources to communicative effect. Consider an example. Shomei Tomastu's Beer Bottle after the Atomic Bombing (Museum of Modern Art, New York) shows a bottle twisted and burned in the heat of the blast. The bottle displays an opaque shininess that is reminiscent of, and provides a visual analogy for, scarred flesh. The picture exploits the high gloss finish of the silver gelatin print to emphasize this shininess, thereby underscoring that analogy. A global content-neutral feature (the glossiness) thus contributes to a thought the work communicates, via the analogy between the bottle and the flesh, about the effects of the bombing on its victims. At least sometimes, then, global content-neutral features are controlled so as to create interplay, and thereby to communicate thought.

However, though genuine, this resource is limited. By definition, global properties cannot be tailored to specific aspects of the picture's content: what goes for one aspect must go for all. And prima facie the range of global properties that can be varied is itself strictly limited: size, texture of surface, and reflectance. There are other examples of artistically interesting use of these features. Nonetheless, if this is all that is available to the authentic photographer, the potential for her work to exhibit interplay will be constrained.

Unfortunately, authentic photography as we know it offers the photographer little else: there are very few vehicle properties that are local, and those there are display no variation independent of content. Consider first the properties available. There is, of course, the color (or in black and white photography, the tone) of particular parts of the picture, and there is the shape of the parts that bear those colors. But what else? In photography as practiced, there are no textures that vary across the photograph, only the texture of the supporting surface as a whole. Nor is there anything like the ‘facture’ found in painting and drawing. Obviously there is no brushwork, impasto, incision, or any sign of the rapidity with which the surface has been marked. Nor, more importantly, are there photographic analogues of these. When it comes to local vehicle properties, in photography shape and color seem exhaustive.

The limited range of locally varying vehicle properties would not matter, if such properties as there are varied independently of content. In discussing painting we saw that while the overall shape of a line or area might fix content, the details of its shape need not. Detail can matter without mattering to content. Prima facie, photography might attain interplay by similar means. In fact, however, such content-independent variation within content determiners is not found there. Any difference in the color or shape of the marks gross enough to matter at all makes a difference to the precise shape or color depicted. Where the detail of vehicle properties ceases to be a guide to content, it ceases to be a guide to anything. Putting this point together with the last, the upshot is that in photography every locally varying property of the vehicle is absorbed into the task of determining content. The authentic photograph seems to aspire to efface itself, wherever possible lacking any detectable properties other than those content requires.

Still, these are all claims about authentic photography as actually practiced. That is already enough to expose authentic photography in its current state of development to the challenge I am developing. Nonetheless, we might wonder how deep these limitations go. Could we extend the range of vehicle properties that authentic photographs exhibit and that the photographer controls, thus expanding her opportunities to create interplay? Perhaps. However, there are at least significant obstacles to any such development. Let me spend a few paragraphs at least gesturing to where they lie.

First, a preliminary. It is important to any art that its products are of interest while being appreciated for what they are. Art sometimes conceals itself, but if an art always did this, something important to appreciation would be lost. Authentic photography is no exception: the whole question of its claim to be art is whether it is of artistic interest while being appreciated as authentic photography. The idea is not merely that it is a way of making artistically interesting works, but that being so made is part of their interest. Suppose for instance, we were able to use imprinting to make pictures indistinguishable from paintings. That would not vindicate authentic photography's claims to be an art. It would reduce to a novel way to make works the interest of which is familiar, a new means of production that adds nothing interesting to the products. (In section 2 I presented the idea that one way to achieve communication is by exploiting the fact that the work is made a particular way. Here the thought is rather different: whatever the artistic ends achieved, by whatever means, a proper appreciation of those achievements involves an awareness of the art form in question.)

This consideration combines with earlier points to impose a demand on any extension of authentic photography. The demand is for transparency in its products: the source of their various features must be available to the viewer. The extension needs to result in works the content of which is plainly imprinted, on pain of giving this way of making pictures no role in our appreciation of them. And it must result in works the other features of which are plainly under the artist's control, if the relations between those features and content are to play a role in communicating thoughts. Appreciators of extended photography would have to be alert both to where imprinting stops and to where intervention by the artist begins.

Suppose we somehow enrich photography so that its products involve either content-independent variation in content-determining properties, or locally varying content-neutral properties, or both. Now we need to give the photographer control over these novel features. That control might lie at the shooting and processing end of the photographic process or at the printing and projecting end. If it lies at the former, the result will struggle to meet the demand for transparency. If lies at the latter, we won't have expanded authentic photography so much as supplemented it with a distinct art of printing or projecting.

Let's begin with the first of these twin perils. Our hypothetical new vehicle properties are controlled by the photographer at the stage of shooting and processing. The difficulty is to limit this control so that it falls, and is seen to fall, short of power over content itself. If the photographer is not to have the latter, she must not be able to determine those properties that fix the photograph's content. But changes that do not affect content lie on a continuum with those that do. Alterations in content-determining properties too small to alter content differ only in degree from those gross enough to have that effect. (Consider painting, where we have real examples. The ability to draw a line with a ragged or careful edge brings with it the ability to place that edge elsewhere and so to alter the boundaries of whatever the line depicts.) And even many content-neutral properties are such that if we alter them sufficiently, there will be consequences for content. (Another painting example, noted above: control over the coarseness of brushstrokes often leaves content untouched but brings with it the possibility of making them so coarse as to lose depicted detail.) Thus, an expanded photography would have to walk a fine line, allowing the artist enough control to make an appreciable difference to vehicle properties, without allowing her so much that she controls content. It is quite unclear what a photographic system that struck this balance would be like. Moreover, even if the balance can be struck, it is unclear how it could be apparent to viewers where it lies. The limits on what the photographer can contribute, and how those limits leave room for imprinting to contribute all of content, would have somehow to be plain to see. Given these challenges, where the photographer's control of vehicle properties is exercised at those stages of the photographic process at which content is generated, there will be a strong tendency to take her to control content as well.

No such difficulties confront us if we turn to the other end of the process. We have already seen that printing and projecting offer the photographer control at the global scale over content-neutral properties. Surely it would take only minor technological change for such features to vary from one part of the print to another. If the photographer could choose whether a given part of the picture is on paper that is gloss or matte, or rough textured or fine, she could tailor these features to the details of what is depicted. Here there is no problem meeting the demand for transparency. Tomatsu's ability to choose high gloss for his print as a whole does nothing to weaken viewers’ awareness that its content is imprinted. Why would matters be different if he had varied texture or reflectance across the photo? Of course, the properties thus placed in the photographer's gift are solely content neutral. It is far harder to see how printing could offer her content-independent variation in content-determining properties. Still, extending in this way her control to local content-neutral properties is easy to conceive and might be significant. At the limit, she might use it to make a photograph of a laundry that mimics the Arkhipov: its surface shiny where it depicts slippery floors but matte elsewhere, the image unfocussed (and so hard to resolve) in just those parts that show wet clothing. Might the result not match the painting in the range of interplay it displays?

I doubt any proponent of photography will find this prospect appealing. This way of introducing interplay into photography does indeed preserve very clearly the boundary between what imprinting contributes and what stems from the artist. But that is because the latter's intervention occurs only after the generation of content. Before she has made any choices about how to distribute reflectance, texture, and the like across the picture's surface, the nature of the image to be printed is already determined. True, without being printed or projected in some way or other, that image cannot be seen. Imprinting fixes content, along with whatever vehicle properties sustain it, but only in potentia. To realize that potential, the image must be made visible, and that is what printing or projecting are for. Perhaps it follows that strictly speaking there is no image prior to printing or projecting. Even so, its nature is determined beforehand. Further, it is hard to see how anyone can deploy printing or projecting to generate interesting interplay without knowing that nature.

In order to know what the image is like, we need only print or project it in ways that add as few detectable properties as possible—the sort of self-effacing embodiment of images that current authentic photography involves. To print or project in more flamboyant ways is to give prominence to an activity that is distinct from and, in both temporal and epistemic respects, secondary to the process of forming the image itself. But that last process, imprinting under the photographer's second-order control, is already a content-generating activity with an artistically interesting product as output. Indeed, that activity and its product just are the art of authentic photography, as we know it. Flamboyant printing amounts to a sort of commentary on or response to that activity's output, adjusting local vehicle properties so as to reflect, highlight, or make observations on, features of the imprinted image.

Such printing might be very interesting from an artistic point of view. It certainly might involve a kind of interplay. But it remains far from the phenomenon at the heart of communicative representational art. There, content, vehicle properties, and the relations between them emerge together in a single, integrated process. Here, in contrast, one form of artistically interesting activity, directed imprinting, would be supplemented by another, flamboyant printing or projecting, that responds to it. The finished photograph would be a composite formed of the products of these two activities: the imprinted image and the printing/projection of it. However interesting that composite proved to be, it would resemble pairs of art objects where one is commentary on, or response to, the other—the reproductive print and the painting that is its source, the illustration and the passage from the novel it illustrates, the ekphrastic poem and the picture it attempts to put into words. The image/print composite would be as little like a single work of communicative representational art as these pairs are.

Where does this leave the Tomatsu? Does Beer Bottle attain interplay only at the cost of opening up this divide between imprinting and commentary? In exploiting its glossiness to communicative effect it, like the flamboyant cases just discussed, eschews self-effacement. It does so, however, with relative subtlety. In choosing high gloss over low or matte, and making that choice for the photograph as a whole, Tomatsu manifests control, but in an understated way. After all, if we make an image visible by printing it, the result must have some reflectance property. The photographic practice of his time gave Tomatsu few options as to what this might be, and offered those only in global form. His choice is thus comparatively recessive, compared with the case of the hypothetical photograph that mimics Arkhipov's painting. The more a given work draws attention to printing, the greater the pressure forcing the fracture between the imprinted image and the printing that makes it visible. Tomatsu's relatively understated control allows him to tap interplay without raising the pressure unduly.

That said, there may be other ways to increase that pressure. In particular, it may be that if many works used global vehicle properties to interplay with content, then the difference between the image and its printing would loom larger in our experience. It may be that if our sense of the unity of authentic photography is to be preserved, even cases such as the Tomatsu must remain the exception.

Thus, as currently practiced, authentic photography offers very little interplay and it is not obvious that it could be developed so as to offer significantly more. In sum:

1. (4) Authentic photography offers the artist only limited scope for controlling interplay—i.e., with respect to global content-neutral properties alone.

So what? Sure, painting offers the artist forms of control that authentic photography does not. Painting can thus exploit interplay to a degree that authentic photography struggles even to approximate. But surely authentic photography offers its practitioners resources that painting does not—if in no other respect, in involving imprinting (see the last set of examples mentioned in note 5 above). Why, then, should we think painting is better placed overall, as communicative representational art? Why not simply accept that each has its resources, each its limitations? The answer is that interplay is not just another resource; its exploitation is central to the very idea of communicative representational art:

1. (5) The exploitation of interplay is central to the full development of communicative representational art.

In general, communication is of artistic interest when our interest is engaged not merely by the thought being communicated, but by the way in which it is conveyed. The means by which communication is effected themselves become part of what we attend to. This in turn expands the resources available for communication. Since our attention is now partly directed to what was previously ignored or taken in only peripherally, we may come to pick up on subtle manipulations of these materials in ways we formerly did not.

These last comments apply to all artistically interesting communication, whatever the means by which it is achieved. Now let's apply the point to representational art in particular. Here the primary means for communication is, obviously, representation itself. As we have seen, three of the resources of representation are content, vehicle properties, and interplay between them. Now, as just noted, these resources cannot be exploited to communicate thought unless audiences attend to how they are manipulated. This imposes a natural ordering on them. Unless content is attended to, we don't have anything recognizable as representational art at all. (How can art be representational unless its representing plays some role in its interest?) Thus the basic form of representational art will communicate thought through content. However, once content is attended to, further steps are possible. In the first instance, attention might expand to include vehicle properties of the representation. But only once both content and vehicle properties are objects of attention can the relation between them become so. Thus there is a specific path along which communicative representational art can evolve: from the elementary form where content does the work, to the more developed form where vehicle properties also do so, and finally to the most developed form where interplay is exploited.Footnote ⁷

Nothing said here depends on accidents of any particular form of representational art. All I have done is to unpack the basic idea of such a thing. But then exploiting interplay to communicate thought is the culmination of the very idea of representational art—it is the end of the road along which such art can travel. Since painting is in a strong position to exploit such interplay and authentic photography is not, it seems the latter falls short compared to at least one other representational art. (Nor is painting likely to be unique in this respect.) Hence we reach our conclusion:

1. (1) If interplay is to play a role in the communication of thought, and thus in communicative representational art, it must lie under the artist's control, in a way that appreciators can detect.

2. (2) To have suitable control of interplay, the artist must control vehicle properties independently of content. The candidates for such control are content-determining, local content-neutral, and global content-neutral properties.

3. (3) Painting offers the artist this control in respect of all three candidate properties.

4. (4) Authentic photography offers the artist only limited scope for controlling interplay—i.e. with respect to global content-neutral properties alone.

5. (5) The exploitation of interplay is central to the full development of communicative representational art.

Therefore,

1. (C) Authentic photography, unlike painting, is not able to develop to the full as a communicative representational art.

Photography's self-conception as imprinting is therefore in tension with a natural ambition for it, to develop to the full as communicative representational art. This is the real challenge to photography.

4. Taking the Challenge Seriously

This challenge is no doubt less extreme than that Scruton's paper sought to pose. The argument does not dispute that photography represents, that it can communicate thought, that it can do so in a way that exploits more than the photograph's content, and that in virtue of all this it might be of considerable interest as representational art. However, what it lacks in hyperbole, the challenge makes up for in seriousness.

Or so one might think. Various responses to the challenge might be offered. Defenders of photography might look for some mistake in the argument for (C). Or they might accept that the argument applies to authentic photography as actually practiced, but take this as an invitation to develop the art form so as to transcend its limitations. Or they might simply deny that the challenge matters: the argument for (C) goes through, but (C) is of no consequence. Since I have already sketched some of the obstacles that any development of photography would have to negotiate, in what remains, I consider the first and last kinds of response, beginning with one attempt to identify a gap in the argument.

Does the argument really consider all the opportunities for interplay open to authentic photography? Skeptics might complain that it overlooks at least two possibilities. One is interplay between vehicle properties and non-depictive content. The argument is confined to depictive content. That is the content by reference to which imprinting is defined, and thus control of which imprinting precludes. Nothing in the idea of authentic photography denies the photographer (first-order) control over content of other kinds. Yet depictive content, though ubiquitous and fundamental, is not the only content photography can bear. For instance, a photograph might depict one thing and thereby symbolize another (as in the example of the dove and peace), or it might depict a particular that exemplifies a type (as in Dorothea Lange's famous images of suffering in the Great Depression). Perhaps the photographer can communicate thought by exploiting interplay between vehicle properties and contents of these other kinds.

The other possibility is to exploit relations within depictive content. The authentic photographer may have very limited control over vehicle properties independently of depicted content, but she has plenty of (second-order) control over the details of that content, even taking its broad subject matter as fixed. She has, for instance, a range of possible compositions within which to present her subject, and her choice among them might be highly revealing of how she thinks of it.Footnote ⁸ Of course, such intra-content relations will not count as interplay: that is specifically restricted to relations between content and vehicle properties. But they are at least its cousins. Why should their exploitation not be just as central to communicative representational art?

The skeptics are right to think that these resources are available to authentic photography, and no doubt they can be exploited to communicate thought. However, it is interplay in the form discussed above that lies at the heart of fully developed representational art. Of course, much representational art is not pictorial, and so does not involve depictive content at all. But for every representational art there is a form of content—sometimes several—basic to it. For the pictorial arts, this content is depictive; for the purely literary arts, it is linguistic; for theatre it is that form of content constituted by the imitation of speech, action, and setting; and so on. In every art, the basic form(s) of content can be complemented by content of less basic kinds: a description or musical representation of a dove might symbolize peace as easily as a depiction of one. The development common to all the arts is, however, first and foremost through the possibilities opened up by basic content. Whatever form that basic content takes, the central developmental arc of a communicative representational art lies in tapping, first content of that basic kind, then properties of the vehicle for those contents, then interplay between the former and the latter. The exploration of interesting relations within basic content is a phase in the first stage of this development. And the exploration of relations between nonbasic content and vehicle properties lies outside that arc altogether. Such explorations might be of great interest. Art that undertakes them might be more sophisticated than art that does not. Nevertheless, they are incidental to the main plot.

I turn now to skepticism about the significance of (C). It has several possible sources. (C) is addressed to authentic photography, but how much photography is authentic? The first way to shun the challenge answers ‘not much’. In the digital age, photographers can readily manipulate their images in a wide variety of ways. Even in the analogue era serious photographers often went in for various darkroom techniques that at least threaten authenticity—adding washes to the negative, or burning or dodging in printing from it. Perhaps, then, very little photography that seeks to be of artistic interest also aims to be authentic. If not, why care if authentic photography is limited in the way (C) claims?

I think it a serious question just how much art photography is authentic. I suspect the first skeptical response underestimates the portion that is. However, since it would be difficult to argue this without reviewing a wide range of examples, I offer a different reply. (C) may explicitly target authentic photography, but the claim promises to generalize significantly.

A good deal of nonauthentic photography is made to look authentic. Many images that have been digitally manipulated, for instance, are hard to distinguish from those that have not. This is true of manipulation in the name of art as much as of manipulation for more mundane purposes. (Consider, for example, some of Jeff Wall's best known work.) But if nonauthentic photography is to look authentic, it loses many of the benefits of abandoning authenticity. What the photographer needs, remember, is not merely to control interplay, but to use that control to make her communicative intentions manifest. Abandoning authenticity offers her the requisite control. But if her manipulation is apparent, the result will not look like authentic photography. And if it is not apparent, then, while she exercises control, she does not manifest it: her audience will not pick up on her intentions.Footnote ⁹ If nonauthentic photography is to look authentic, it cannot use the greater control it offers over interplay to communicate thought.

The second way to shun the challenge rejects the idea that authentic photography should aim to develop as communicative representational art. Since such art involves a distinctive means (representation) to a distinctive end (communication), we can reject the aim by rejecting either of those. Suppose we reject the end. Perhaps ‘communicative’ art's point is not to communicate, but to demonstrate the difficulty or impossibility of genuine communication. (See Schwarte Reference Schwarte, Gronau and Lagaay2008, which traces this idea in the work of Heidegger, Wittgenstein, and Adorno.) I suspect my challenge can be reformulated to accommodate this. Whether the goal is to use representation to communicate thought or to demonstrate the difficulties of doing so, the resources for pursuing it are the same, and the same disparity will obtain between those available to authentic photography and those available to other arts, such as painting.

What, then, if we reject the means? The argument for the centrality of interplay to communicative representational art (premise 5) apparently relies on the idea that the goal of any form of art should be to make best use of the resources peculiar to it. Here the art is representational, so the resources are those of representation. This sounds like the sort of concern with medium, and with exploring the possibilities it opens up, that was characteristic of modernism. But why think artists are still interested in that? These days, artists, including those who work with photography, use whatever they can to achieve their aims (Campany Reference Campany2003: 18–20; Batchen Reference Batchen1997: ch.1). Photographers not only manipulate images (thereby abandoning authentic photography), they also tap a variety of methods that are not even photographic (thereby refusing to limit themselves to photography in any form). Purity of medium, and the single-minded exploitation of the resources it offers, is no longer their interest. The challenge relies on a conception of communicative representational art that is simply outmoded.

However, I have been working at a different level of generality from the modernist. She requires that each art—painting, theatre, sculpture, literature, etc.— does what it alone can do, or at least can do uniquely well. My notion of representational art is broad enough to encompass all these. Any ‘medium’ it makes central is simply representation itself, more specifically the resources that it offers. And the ambition I propose, to exploit that ‘medium’ to the full, is proposed equally for them all. We have seen that painting is in a position to fulfill this ambition, as authentic photography is not. Of course, in making that case I have at times done so in terms specific to the pictorial arts: e.g., interplay between depictive content and vehicle properties. Moreover, I grant that different arts will involve different forms of basic content. But for every representational art the same question arises: can it exploit the full range of resources made available by its basic form(s) of representation (content of the relevant kind(s), vehicle properties, interplay, and means of making) in the communication of thought? And the answer, in most cases, will surely be ‘yes’. Why doubt that, say, sculpture, theatre, and literature will be as well placed as painting to do this? If photography is not in that fortunate position, that is a striking fact—and one does not have to be some kind of modernist to find it so.