The Most Interesting Thing that Can be Done With Representation

by James Elkins

What is the most interesting thing that can be done with representation? The best way to begin answering that question, which is (or should be) of pressing interest to visual artists working in a variety of media, is to list some ideas that were once thought to be the most interesting things about representation:

1. It is no longer interesting that photography means “painting is dead” (Paul Delaroche’s opinion, although he kept painting). If painting is dead, it is so in a far more curious fashion than anything photography could have inflicted on it. (See Stephen Melville’s excellent catalogue As Painting.)

2. It is no longer interesting, to most people outside of science, that machines can replicate features of the world apparently without human intervention. (See Raine Daston and Peter Galison’s work on this subject.)

3. Nor is it interesting, particularly, that machines can represent the world only by means of ideas articulated by the earlier history of visual representation. (Many people have worked on this, from Peter Galassi to Daston and Galison; the best work is Joel Snyder’s.)

4. It is not interesting, because it is not true, that cubism and Cézanne ruined naturalistic representation and substituted something entirely different. (For a cogent critique see Leo Steinberg’s essay in Other Criteria.)

5. It is especially uninteresting that the “age of mechanical reproducibility” has left us, so it is said, with a nostalgic yearning for the aura of unique artworks. (Walter Benjamin, and his best expositors, including Max Pensky.)

6. It is not interesting, anymore, that the obverse of Benjamin’s statement is also true: namely, that mechanically reproduced images have their own aura, so that they project a different and equally compelling sense of representation (as in Rosalind Krauss).

7. And speaking for myself, it is not particularly interesting to demonstrate the suffocating sameness of representations (as Sherrie Levine has done),

8. Or to exult in their weightlessness (Baudrillard, Paul Virilio),

9. Or decry their terrible emptiness (Debord),

10. Or join the devil’s camp and reproduce them endlessly (Warhol, and now Erró),

11. Or even to make all those previous approaches into the object of academic study (the field called visual studies; Nicholas Mirzoeff and others).

So what is left to do with representation?

Among contemporary artists, Vik Muniz’s work is exemplary. He makes photographs of materials such as ash, sugar, chocolate, and wire. Before he photographs them, the materials themselves are made into images. And the images are often taken from previous images and occasionally even photographs. The ash picks out a painting by Caspar David Friedrich; the chocolate traces one of Namuth’s photos of Pollock painting; the dirt is scraped and molded into an imitation of Courbet’s Origin of the World.

There are a number of reasons why Muniz’s work is more interesting than almost any other at the moment. I agree with Muniz that part of the answer is to playÑthe word is crucial, but needs to be adjusted and constrainedÑin “a magical space” (magic, another essential word, which also needs to be qualified) “between realism and artifice.”

“I want to make the worst possible illusion that will still fool the eye of the average person,” Muniz says. “Something so rudimentary and simple that the viewer will think, ÔI don’t believe what I am seeing, I can’t be seeing this, my mind is too sophisticated to fall for something as silly as this.'” This is different from ordinary magic: it is postmodern magic, like one of Ricky Jay’s educational exhibitions where a lightning-fast exposition of the history and theory of magic goes hand-in-hand with the display of magic. “Illusions as bad as mine,” he says, “make people aware of the fallacies of visual information and the pleasure to be derived from such fallacies. These illusions are made to reveal the architecture of our concept of truth.”

So, provisionally, representation is at its most interesting when it playfully demonstrates its insufficiency, in a particular manner which is related to magic, and in a mode that can be described as rudimentary or simple. This seems to me largely right, and I want to try to develop it into a more formal account of the current state of representation. To do that I am going to return to the beginning of the twentieth century and consider cubism, the exemplary modernist intervention in representation. I will read cubism’s relevant achievement through the most perceptive account that it has gotten to date, T.J. Clark’s description in the book Farewell to An Idea.

First I will mine Clark’s text for particular ideas about the mechanisms of success in analytic cubism in the crucial years 1909-1912, and then (and I hope Clark will pardon the sequel, which is not at all in the spirit of his book) I am going to set some of his ideas to work building a theory that is adequate to the current state of representation — a theory adequate to Muniz’s practice.

Cubism may seem an obtuse choice, given that my subject is contemporary representation. I choose it because even today cubism is extremely poorly understood and more important than ever because of that fact. I choose Clark because out of all the hundreds of theories about analytic cubism, his is by far the most reflective and complex, and therefore the best able to be appropriated, in the inevitably impolite fashion of postmodernism, to problems at hand.

1. Clark’s sense of cubism

It needs to be said that Clark’s account of cubism is extremely analytically involved (so much that few of the book’s many reviewers except Karsten Harries and Jay Bernstein got much of the point), and also almost incredibly picky. One of the questions his chapter tries to answer is this: Does analytic or “high” cubism have some high point, some absolute pinnacle? Cubism invites that kind of question, and it has been asked since its historiography got underway in the 1920s. Most historians and critics are content to name a year (1909), or part of a year (the summer at Horta de Ebro). Some scholars prefer to point to a style (the treacherous approach to abstraction in the summer of 1910), or a genre (the three portraits of Vollard, Uhde, and Kahnweiler).

Clark pushes much harder on the idea of the highest point, and he ends up looking just at a few paintings, and even at parts of individual paintings. None of his praise is unqualified, a quality that is connected in his account to the nature of modernism itself, which has always made stumbling attempts to speak about impossible ideals, even while it was trying to achieve impossible destructions. None of the reviewers of Farewell to An Idea made much of some astonishing passages in which Clark severely critiques cubism’s ostensible high points. He thinks the portraits of Udhe and Vollard are “not very happy episodes (particularization in them seems not very distant from kitsch).” The third in the trio of canonical cubist portraits, the Portrait of Kahnweiler, is said to be “high-spirited (not to say a bit glib)” (212). The result is that cubism is reduced to a few hard kernels of interest, and those are so obdurate, so welded to their times and places, that they could be successfully emulated, disseminated, or projected as a style. “High” cubism then becomes something other than a movement that might possess a character and a more-or-less systematic method. Even Braque, who Picasso liked to advertise as a fellow-traveler, is demoted.

What is cubism then, if it is not systematic enough to be the movement or style that it is universally taken to be? It is “not a language,” Clark says: “it just has the look of one” (223). Picasso was its only native speaker, but even he did not speak the language in the way that a person can choose which language to speak. In Clark’s account, cubist painting is more akin to speaking in tongues: it came on Picasso in fits and starts, and when he emerged from one of his trances, he tried, inevitably and unsuccessfully, to continue in a logical fashion what he had just made in a less than logical fashion.

This is much stranger than it may seem at first. The idea is basically that cubism was not a movement in the ordinary sense, but a kind of production by fits and starts. Its truly successful moments were also moments of blindness to method and history, and even to the possibility of going on. It’s not just that cubism is best imagined as a narrow set of works: it’s that the set itself is a fiction, comprised of secondary works that misunderstand their few inassimilable precedents. It’s not just movements and their labels that are at stake in Clark’s chapter: it’s the coherence of modernism itself, because a movement shrunken to a few isolated half-failures ruins the smooth transitions that make art history possible.

What exactly were those half-successful moments? Take for example another canonized work, the painting Reservoir at Horta de Ebro. Clark says that the painting’s “wager” is “that it would be precisely by fastening on the aporia and undecidables of representation” — the inevitable Necker cubes, the reversible convexities and concavities of space — “that a new system of spacing and singling out the parts of a world would be generated” (203). It’s necessary to read this slowly: it is not a claim that Gestalt psychology-style illusions were a plausible road forward for cubist painting in the summer of 1909. The idea is that the painting’s “wager” is to “run the machinery of illusionism for all it is worth,” so that the machinery opens “exactly there onto its opposite or ground” (203). The painting was a moment of clarity, and for that very reason it proved to be an unworkable wager. In Cadaqués in the summer of 1910, Picasso pursued another purpose, which has been understood in retrospect as a flirtation with abstraction. It was “at last” a “Ôsolution'” (in quotation marks), but “of a profoundly inimical sort. Picasso was the last person to want, or perhaps to see how, to pursue the Cadaqués solution to its logical conclusion (Mondrian being the first)” (192).

What cubism needed and occasionally managed was something much harder. To describe it I need first to say that among the many interlocutors in Clark’s account are the modernists (inevitably represented chiefly by Clement Greenberg but also by Michael Fried and William Rubin) and “their semiotic inheritors” (which tends to encompass chiefly Rosalind Krauss, Hal Foster, and the circle around the journal October). Clark’s route between the two sides is a matter of maintaining an ordinary kind of skepticism about the cliché; modernist notion that “reference to the things of the world has simply ceased with high Cubism,” and at the same time keeping away from the “semiotic” notion that cubism was at heart about “the freer and freer play of the signifier, a set of devices discovering that simply the difference between them is enough to make a world” (181, 183). Briefly: modernists have inadequately described cubism as the first step away from illusionism and toward the purity of abstraction, and postmodern “semioticians” have seen it as the prototypical leap from semantic dependence on the world to self-referential system, to a code of signs that shift and play among themselves to create meaning.

Cubist painting becomes interesting when it avoids these two models in a very specific way, and here I come to the insight that seems so important for current work. Cubism “had better be a stream of metonymies” (a set of self-referential signs) than “a neat metaphorical fix” (a namable destruction of illusionism, for example). But it can never be neat, one way or the other (185). Cubism has to negotiate the alternatives, and it does so in at least seven ways:

1. Pretense. Works of art “differ dramatically” in their “willingness to admit, or to Ôforeground,’ the arbitrariness of the sign” (185). Some cubist works are “fiercely unwilling,” but none except formulaic paintings forget that they are “pretending” to be painting in some full and untroubled sense.

2. Insufficient. It follows that interesting work does not lose hold of “the metaphor of [its] own insufficiency.”

3. Inefficient. A painting that does lose hold is the overconfident upper portion of the unfinished painting Woman with a Mandolin (1912), where Picasso forgets that “acts of illusionism” are really just that: they must not be seen to “work too efficiently” (191).

4. Counterfeit. Cubism was “not a devising of a new description of the world,” especially not one that was a response to new ideas in physics or philosophy (213). It was a way of “counterfeiting” such a description. Note that counterfeiting is more complicated than simple mechanical reproduction: it is self-aware, and founded on the inevitable failure of perfect iteration.

5. Play. Therefore cubism at its best plays at producing a new kind of illusion adequate to painting’s new situation: it mimics painting’s old métier as system.

6. Failure on several fronts simultaneously. The most interesting moments in cubism are where both “brute or schematic likeness” and the “free play of signifiers” are partly demoted, given a new “subordinate status” (218).

7. Base materiality. Both these operations take place “on the surface” — on and with the paint — not as if they were projected from somewhere outside the painting’s materiality (221).

2. Interesting representation

That is as far as I am willing to condense Clark’s account. Here is the moral I want to draw for contemporary work in representation: It must be aware that both the happy postmodernist free play of signifiers, untethered from the world, and the happy modernist rejection of appearances in favor of some new representational regime, are at once inadvisable and impossible within the project of visual art since cubism. The two have to be posed by visual art, played as a game in which a win would constitute either a moment of misguided optimism or a moment of blindness that cannot be repeated. The work of representation is to counterfeit the conditions under which an adequate representation might once have taken place, without giving up either game (the metaphoric and metonymic, or illusionist and semiotic, or modern and postmodernÑhowever they are named).

This is the context into which I would like to read Muniz’s formulas. I have already mentioned play, which has much the same valence in Muniz’s work as it does in Clark’s description (essential but too easily unburdened). Counterfeiting can be fun, but it is uninteresting simply to make fun of illusionism. Counterfeiting, real counterfeiting, is a very serious business, not a romp as it is for Jean Baudrillard or Paul Virilio. To play and counterfeit I will add three more terms that are properly Muniz’s:

8. Magic is now clearly the state of affairs in which old-fashioned Zeuxis-style magic remains possible but can never be put in control of a picture. The new magic is beleaguered: beset by carefully plotted evidence of its own impossibility, and teased by its own continued existence in pictures.

9. Illusionism “demonstrates its insufficiency,” as Muniz says, but not easily. The best contemporary pictures, like the best cubism (the tradition is continuous, despite appearances) can be “fiercely committed” to denying their insufficiency, but in the end the denial cannot be deferred.

10. The “worst possible illusion” that Muniz mentions is actually the best possible illusion, the one where the image’s possibilities are shown as the ruins that they have become: still working, but with rusted hinges, groaning, and about to give up. Cubism was always a “worst possible illusion” (its muddy colors, its little brickwork brushmarks), and it is the precedent for all work that still needs to sink to the bottom of some barrel of illusionism.

These ten points, I think, are the beginnings of a viable theory of the contemporary state of illusion, and an answer to the question with which I started. (There is one more, an eleventh, which I will save until the end.)

Works like 6,200 Yards (Lighthouse), a photograph of a picture made out of string, works with each one of these criteria. It is certainly “magic,” and the magic is definitely in quotation marks. Muniz is extremely dexterous, a requirement for any magic: but he is also a magician bent on exposing some of the machinery–like Ricky Jay doing a trick with an enormous card, just to make sure the gestures are visible. Despite Muniz’s skill at traditional, academic drawing, string just does not make good waves, and if you look closely you can see the threads looping back, as if the water was full of eels. The string painting demonstrates its insufficiency, but reluctantly.

Materials like string, chocolate, wire, cotton, and sugar bring the works downÑthey are its “base materiality”–and play at the seriousness of high illusionism. There are several reasons why Muniz’s images are better as photographs than as a unique pieces made with string, chocolate, wire, cotton, or sugar. One reason is that the actual objects, exhibited by themselves, risk becoming campy (and playing into Baudrillard’s hands) or miraculous (and playing into Benjamin’s). But the most important reason, I think, is that the photograph introduces the theme of the counterfeit in a way appropriate for photography. The situation is not a continuous outgrowth of cubism, or an inevitable consequence of cubism, but it follows a logic inaugurated by cubism. This is what representation has become, Muniz’s experiments are among the best and most promising current work.

3. Traps for representation

The eleven points I have gathered are extremely difficult, and artists who short-circuit them tend not to be playing the game seriously enough. There are many ways of lightening the burden. I will end with a sequence of photographs that show some of the more seductive options.

Plate 1 is a reproduction of one of Rembrandt’s etchings of his friend Jan Six, and Plate 2 is an enlargement. I made these a few years ago to demonstrate some problems in teaching art using reproductions. Actually, these first two plates aren’t Rembrandt etchings, exactly. They were made from a nineteenth-century photoetching of an original Rembrandt print. Photoetchings can be remarkably close to the originals. The print departments of museums in Europe used to keep collections of photoetchings and photoengravings to bring out to their naïve patrons, to save the wear and tear on the originals. Only experts who knew better got to see originals. (In my PhD program, we were taught to tell the difference, so we wouldn’t be duped when we went abroad.) But at the level of enlargement of Plate 2, there is no appreciable difference between the photoetching and the original.

That invisible difference is fairly interesting. But it is a conceptual difference, and therefore a game the photographs are not playing. This is one of the ways that picture-making opts out of the challenging régime I have been describing–by substituting conceptual and theoretical differences for things worked out in the pictures themselves.

It gets a little more interesting if I point out that Plates 1 and 2 are not really photographs of photoetchings. The image as you see it here (in either plate) is really a print of a photograph of a photograph of a print of a photograph of a print, because the original print was photographed, made into another plate, printed, and photographed; and then I sent the photograph to the publisher, where it was rephotographed (or scanned) and printed in many copies. Schematically:

(1) Original Rembrandt etching

(2) Nineteenth-century photo of the etching

(3) That photo used to produce an etching plate

(4) The photoetching in my collection

(5) The photograph I took of that photoetching

(6) The scanned file from my photograph

(7) Plate 1 as it is reproduced here, and in all other copies.

Now this kind of iteration serves as a source of pleasure among writers who follow Rosalind Krauss in finding value in the very concept of iteration. The medievalist Michael Camille has written a wonderful essay on this subject called “The Trés riche heures in the Age of Mechanical Reproduction”: in his essay each edition of the original medieval manuscript has a different value. If you go to Chantilly outside Paris, you’ll be shown one of the reproductions, but you won’t be told it’s a reproduction. Only a very few people have seen the original. Camille hadn’t seen it when he wrote the essay, but that wasn’t a concern because Camille loves reproductions, sometimes even over originals. Another medievalist, Herbert Kessler, got a look at the original Trés riche heures, which are kept in Chantilly behind locked doors; he reports nuances that have not yet been captured by any photographic process.

This story illustrates two further dangers for art that plays with reproducibility: it either falls into Benjamin’s devotion to the aura, or it slips into the diminishing joys of endless reproduction, and the pallid comfort of the notion that in the end the world is nothing but reproductions. Both are too easy.

Plate 3 is the same artwork, as it is reproduced in a book called Rembrandt: Experimental Etcher, published in 1969. That book has one of the best reproductions of the image. In this detail you can see both the paper texture and, if you look closely, the halftone dots. The smaller marks left by the etching needle are lost: the ones around Jan Six’s chin, for example. In comparison with the first two plates, this one is worse, but it is the best most people can do when they are interested in Rembrandt’s etchings. That invidious comparison, the one I was after when I first made these pictures, points the photographs directly at the “metaphorical” ideal of illusionism that is still so common in the visual arts, and so exhausted.

Plate 4 is a photograph of a slide from the teaching collection of the University of Chicago, as it looks projected onto a screen in a darkened seminar room. It is distorted (Jan’s face is squeezed a little) because I took the photo from a seat a little off to one side of the room, like most students’ seats would be. I sat about fifteen feet back, so the texture of the screen itself is not visibleÑas it ordinarily wouldn’t be. Plate 4 is actually the end result of a different sequence of seven steps:

(1) Original Rembrandt etching

(2) Twentieth-century photo of the etching

(3) Photograph used in an (unidentified) book

(4) The university’s slide, taken from the book

(5) My photograph taken in a classroom

(6) The scanned file from my photograph

(7) Plate 4 as it is reproduced here.

The difference between Plate 2 and Plate 4 is itself interesting, and not just for what it shows about what students see. It represents a different path, with an equal number of steps, from two different “originals” to their final forms. That difference is a step in the right direction, because it exposes counterfeiting in addition to Benjaminian mechanical reproduction.

But the gain is ruined with my last example. Plate 5 is a photograph taken from a computer screen. The image was a fairly large file, and filled half of a 17 inch monitor. Its route, from “original” to final form, was different once again. But this time the friction, the “base materiality,” is wholly absent. This is the free-floating play that Baudrillard, Virilio, and so many others love so much. For representation, it is a dead end.

If work on representation (as opposed to rote representation) is to go forward, it has to go more or less in the direction that Muniz is pushing it. My last Plate comes the closest, and fails less obviously.

This is a photograph of two cockroaches. The one on the bottom is a rubber roach. Someone is holding it down, and using the broken end of a pencil to help lift one of its rubber wings to reveal the stamped inscription: “MADE IN / HONG KONG.” The one on the top is a real, live Madagascar hissing cockroach, the largest species in the world. The handler isn’t touching the real roach because its carapace is coated with a waxy substance that causes rashes, raises blood pressure, elevates the heart rate, and sometimes leads to arrhythmia.

No one pats those cockroaches, although it is safe to let them crawl over your hand if you can stomach the little pricks of their feet as they pinch you to keep their grip. This photograph is about illusionism, but self-reflexivity is not the only card it has to play. It is a bit funny, but also sour. It stages its counterfeit nature, but only reluctantly gives up its hold on illusion. It tells a story about reproduction, but without proposing there is no such thing as an original. It is playful, but not in a childish or simple way.

The problem with this image, and the reason it belongs with many other failed pictures, is that it does not work hard enough with either side of the equation metonym-metaphor, or illusionism-semiotics. If Muniz had done it, both cockroaches might have been sculpted from little bugs. That might have worked. It is not easy to do interesting work with representation, and one of the crucial rules is the one Clark insists on from the beginning, which I also find in Muniz’s work. It’s my last criterion of interesting representation.

11. Contingency. There is a deep reason why Clark searches cubism for individual passages and fleeting moments, and concludes that cubism as a whole is not a language: because in modernism representation is no longer given. It can cannot be recovered, in any form, from previous successes, and it cannot be made into a formula and repeated, or emulated, or codified into a school, style, or movement. Interesting representation has to be rediscovered in each new context. Muniz does this deliberately: a few cotton sculptures, some clay, some sugar, M&Ms, syrup… they may seem skittish or unsure, but it’s the opposite: they are faithful to the only way that representation know how to work. Anything else would be programmatic, as Braque misread Picasso, as Picasso misunderstood himself. Clark calls the absolute faithfulness to the unreproduceable context contingency, and he traces it through modernism starting with Jacques-Louis David. Cubism works by “having one’s metaphors of matter reinstate… pure contingency at every point”Ñthat is, making sure that the painting’s inventions, its “signs,” are remade in every passage and in each painting so they respond to the painting’s exact occasion (220-21). It would have been easier, at any point, for Picasso to mistake cubism for a manner, as Léger did, as Gris and Laurens and Metzinger and Kubista and Filla and so many others did.

Contingency is not quite the same as the rule of the avant-garde. There is a difference between Picasso’s ongoing innovations–very much a paradigm of the avant-garde–and his unfaithfulness to the exact requirements of the work at handÑthe inevitable slip away from rigorous contingency. The avant-garde, as Greenberg said, demands perpetual motion, but that motion can often end up being thoughtlessÑnothing much more than a mindless impetus to change manners every season. Muniz’s work is exemplary: even though his work continuously changes, it does not drift in the directionless fashion of the avant-garde: it remains fixed to the point of representation, to the material and strategy of the moment, in order to say the most interesting thing about representation that it is possible to say at each moment.

Text published in Vik Muniz: Incomplete Works, 2004