The Real and the Fake in Thomas Demand's Trompe l'oeil Photography
Fig. 1: (left) René Magritte, L'Univers Démasqué (The Universe Unmasked), 1932, Oil on canvas, 28 3/4 x 36 1/4 inches; 73 x 92 cm, (right) hung over wallpapered "draperies" by Thomas Demand, 2010, and foreground: Saadane Aiff, Strategy of Anxiety, sculptural topology, 1999. Photo by Amy Lipton. Courtesy of Matthew Marks Gallery and Huffington Post
In Thomas Demand's 2000 photograph Model, we see a room with a window and a table, upon which there is an architectural model. But as we look to the darkened window to learn why the architect is not there, we see that its glass has no sheen or transparency. Instead it, like every other surface in the space, has the same strange matte quality. We inspect more closely, perhaps spying an uneven edge or seam, until we realize that the model on the table is not the only model in this photograph, rather the entire scene is a model, a paper construction. Demand constructs life-size, three-dimensional paper models that he photographs with a large format analog camera, and then destroys, exhibiting only the resulting image, which is roughly to scale, and thus a photograph of about five to nine feet in size. Here his represented content - the architectural model - self-reflexively doubles his process of building paper models.
Because his photographs, at first glance, look to be of actual places, one's spectatorial experience in front of his work is that of being tricked. One critic describes his attempt to look through the vertical blinds of Window, another of Demand's photographs, only to discover that he had been "fooled" by its "trompe l'oeil" (or 'trick the eye') effect. This critic repeats the terms of Pliny's emblematic trompe l'oeil tale, casting himself in the role of Zeuxis who mistakes a represented curtain for a real one behind which he wants to look. Rhetorically this critic not only turns Demand's photograph into a painting - Parrhasius's painted curtain - but beyond that, he turns it into a painting capable of convincing a viewer that a representation is a reality. For another critic, this is the power and the threat of Demand's work; she writes, "If we can be convinced by a photograph that a model is a real building, we might begin to doubt the reality of documents everywhere." Both Parrhasius and Demand could "convince" viewers because their representations looked real. But the stakes of this trick seem different in the case of Demand because of photography's documentary function. This documentary function has to do with the automated and chemical process of photography, which allows the photograph to be understood as causally related to its referent. In photography theory, this causal relationship has been characterized as indexical.
But Demand's photographs are often misunderstood as having no index or referent: yet another critic describes the way that Demand empties his photographs of "any relation to the real;" she writes, "How is it possible to look at these uncertain images that have no unity, no index or referent, [no] history or origin?" I say that this is a misunderstanding because his photographs do still index a referent; they are causally related to the paper models he has built and then photographed. Mistaking that referent for the real thing, rather than a construction thereof, does not rid the photograph of indexicality; the paper construction is the referent that adheres. If we mistook the photograph of the paper model for a photograph of a real room with a paper model, it was only because of how it looked. It is photography's realism that contributes to our expectation of its truth or authenticity, so that when photography's realism is disrupted, so too is its supposedly truthful and authentic relation to the real via its indexical referentiality. Thus, when the realism of Demand's photographs is undone by a stray pencil mark or an uneven edge, critics confuse this sign of the referent's construction with a manipulation of the photograph's indexical relationship and its presumed promise of truth.
This uneven edge is not only a sign of the referent's construction; it is another index - a trace record of Demand's process, his meticulous wielding of scissors. It is our recognition of this index that disrupts the truthfulness we associated with the photograph's index. As such, we begin to see how inappropriate it is to yoke truth claims to indexicality. Indeed, as film and new media theorist D.N. Rodowick writes, "a photograph can neither lie nor tell the truth; it only denotes (automatically registers space) and designates (is causally related to a past state of affairs)." Demand's photograph might trick us about the constructed-ness of its subject, but it doesn't lie; it denotes and designates the paper construction and its resemblance to the real world. In that Demand shows us the possibility of producing "misleading images that appear spatially consistent and perceptually real," he stages the dominant fear about digital photography. At least one critic has described Demand's construction within the image as rendering it a "wholly digital image." For new media theorist Lev Manovich, it is digital intervention that disrupts photography's status as a "recording medium by allowing manual construction within the image." As a result, photography, he writes, "is no longer an indexical media technology but, rather, a subgenre of painting."
But, like Demand's paper model, painting also has an index - a sign that exhibits its cause - in the brushwork or trace of the painter's point of contact with the canvas. It is this index - the telltale sign of brushwork - that trompe l'oeil painting must disguise in order for the trick to succeed. That is, the trompe l'oeil work of art uses resemblance or perceptual realism to disguise the indexical trace of its painted process. When it succeeds, its illusionism is typically described as photographic. But this repeats the misunderstanding that a photograph is necessarily a likeness: it can be out-of-focus and over- or underexposed beyond recognition. Such photographs display their indexical relationship to, rather than their resemblance of, the physical world. Trompe l'oeil paintings cannot display their indexical trace initially, but they have to eventually. Otherwise, the trick that Parrhasius' curtain is a painting and not a real curtain would never be discovered. By describing digital intervention in photography as a 'subgenre of painting,' Manovich looks to place digital image-making in a history of hand-painted animation practice. Cell animation, insofar as it is understood as the hand drawing of sequential images, is typically seen to undo film's indexicality. But Rodowick finds that cell animation, insofar as it is understood as the photographing of hand-drawn images frame by frame to produce an illusion of movement, has "a strong indexical quality," despite its imaginative use and malleability. For Rodowick, it is crucial that each animated cell is photographed, not that it's also drawn. He writes, "Here, as in all other cases, the camera records and documents a past process that took place in the physical world." This connects, I think, to Demand's project. Like the hand drawings of cell animation, Demand's handmade constructions are photographed. Here again, the camera records the results of Demand's process all of which took place, painstakingly, in the physical world of paper, cardboard, and glue. But, what seems crucial is our ability to perceive the index of this construction within the image: the drawn penguins in Mary Poppins, for example, are perceptually distinct from the human Dick van Dyke and this distinction is maintained in the film; they are understood to be existentially distinct. But the constructed elements in Demand's photographs are not as readily perceptually distinct, at least not initially; there is no 'real-world' point of contrast, no Dick van Dyke or other human subject, within the image. Indeed, the trompe l'oeil illusion that I described can occur because we don't initially perceive the existential distinction in his photographs between his paper models and the real world. The paper models look (at first) to be 'ontologically equivalent' to the real world.
Fig. 4: Thomas Demand, Hydrokultur, 2010, C-print mounted on Plexiglas, 66 1/8 x 54 3/8 inches; 168 x 138 cm. Courtesy of Matthew Marks Gallery, http://www.matthewmarks.com/artists/thomas-demand/selected-works/#/images/10/
Similarly and increasingly, digital image synthesis presents humans and animations together in ways that are perceptually indistinct. In The Social Network (dir. David Fincher, U.S., 2010) the digitally replaced Winklevoss twin can look ontologically equivalent to the filmed Winklevoss twin. But insofar as digital capture imports the image of both human actors into the world of digital synthesis, the actors and animations don't just look 'ontologically equivalent,' for Rodowick, they are 'ontologically equivalent.' Whether captured by digital cameras or synthesized on computers, they are both numerical code. In digital photography, the moment of capture is immediately and automatically transcoded, that is, converted into code. This quantifiable and symbolic data can then take any form whatsoever (thus Manovich's 'subgenre of painting'), but it tends to take a form that looks perceptually indistinct from the real world. This is what Rodowick calls the "paradox of 'perceptual realism.'" That is, the imagined painterly freedom of digital image production is paradoxically geared toward the "realist" conventions of photography. While Demand does not use a digital camera, I think his process of building paper models literalizes this digital-transcoding process. By producing perceptually realistic images that structurally correspond to a viewer's experience of photographic space, he literalizes the realist conventions of digital image-makers. Both he and digital filmmakers build the content of their images to conform to our expectations of photographic realism. Demand's constructions are built (and lit) to be photographed; they cheat for that purpose: tops or undersides of tables are built relative to the high or low camera angle; backsides of objects do not need to be built because of their frontal orientation to the camera. But this logic of photographic realism is not Demand's only constraint. His paper models are not just constructions; they are reconstructions. They reference a source, a photographic source that he typically finds in mass media publications. Thus, they are bound to the information established by this source image. In this way, he also literalizes the way that digitialization transcodes the point of capture into another symbolic representation of that data. His source photograph literalizes the digital camera's point of capture, which he then manipulates in the form of his paper constructions; as such, his source photograph is like the "raw material" of "live-action footage" that digital filmmakers can, according to Manovich, manipulate "by hand."
Fig. 5: Thomas Demand, Ghost, 2003, C-print, 47 1/4 x 63 inches; 122 x 160 cm. Courtesy of Matthew Marks Gallery, http://www.matthewmarks.com/artists/thomas-demand/selected-works/#/images/7/
Instead of converting the source image into binary numeric symbols, he manifests the terms of the digital in entirely analogical ways. He literally 'cuts and pastes' the scene of the photograph back together, while literalizing another hallmark of the digital; its "inevitable loss of information." Demand refuses to supply any potentially individuating sign that might help us link his photograph to its source photograph. Whereas in Model, this erasure was naturalized by our assumptions about architectural models - they show no marks of use or specific identifiers - elsewhere, as in Poll of 2001, the post-it notes without reminders, the ballots without votes, the telephones without numbers refuse any (even analog) display of the symbolic coding central to digital-transcoding. Demand's literalized 'digit'-alizations draw out the goal of perceptual realism at the heart of digital-transcoding. As in The Social Network's seamless digital replacement of one actor's face with another, "[g]iven enough resolution, a digital[ly filmed image] can simulate the look of a[n] … analogical image." While digital image-making thus aims toward the terms of photographic likeness, Demand's image-making only aims for this as our initial impression. Eventually, however, spying a single pencil mark or seam is enough to let Demand's viewer, like the one watching Mary Poppins' penguins, perceive the construction within the image as ontologically and existentially distinct from the real world. In this way, his trompe l'oeil images allow the indices of their construction to become visible. Demand, in contrast to most digital image production, allows his goal of photographic realism to fail. Doing so reminds us that spatial semblance is not photography's primary power. The logic of photography's index, determined by its causal relations or 'real connections,' never had a necessary relationship of resemblance. A weathervane, for example, can index the wind's direction without resembling the wind. Rodowick defines the index as, "a present trace of a past action whose causal origins must be found through reasoned conjecture." Photography's index, in this way, implies two different kinds of interactivity: the 'touch' of the trace through light and chemical interaction, as well as the viewer's interaction through 'reasoned conjecture' to discover the cause of an image. I think Demand literalizes both versions of interactivity in his project. His touch literalizes the photographic trace, denoting the image's cause, but insofar as this touch re-touches a source image, it also literalizes his 'digital' transcoding of that image. Digital interactivity, for Manovich, has to do with the user becoming co-author of the work, interacting (by touching) a media object to choose elements to display or paths to follow. Demand's version of the source image is the only version we see; he physically chooses what to show and what to withhold. His withholdings eventually upset the perceptual realism of the final image; they become visibly absent traces of his past actions upon the paper construction, signs of his co-authorship. When we recognize the index of his intervention, we experience a trompe l'oeil rupture in our viewing experience of the photograph. He literally re-touches a source photograph and our discovery of this touches us; we have a bodily experience of confusion or surprise that prompts our own interaction with the image. Our confusion prompts our desire to know how Demand built his image to look this way, which we can discover through 'reasoned conjecture.' But the cause of Demand's image exceeds his paper intervention. It includes his selection of the source image and the initial reason someone else took that photograph.
Fig. 6: Thomas Demand, Collection, 2001, C-print, 59 1/4 x 78 3/4 inches; 150 x 200 cm. Courtesy of Matthew Marks Gallery, http://www.matthewmarks.com/artists/thomas-demand/selected-works/#/images/6/
But Demand intervenes here too, withholding this cause. His banal titles refuse any historical specificity. Instead, it is up to us to remember the source by re-constructing that which Demand has withheld from his paper reconstructions. Returning to Poll, we may or may not recall a more populated version of this scene from newspapers or magazines reporting the vote recount of the 2000 U.S. presidential election. But its identification as such is not just a name game, even though art historians and curators often treat it like one by suggesting explicit sources in wall or catalogue text accompanying exhibition of his work. In an interview, Demand regrets this one-to-one identification; "if you label [it,]" he says, "it becomes too much like Madame Tussaud's - people compare before and after." Instead, he says, " I just did something about the knowledge that you and I have about [it]." His refusal to supply the now-lost source image insists that something of the limitations of our knowledge remain; he resists letting the site of a historic or tragic event seem wholly knowable and contained. Instead, he employs a trompe l'oeil strategy, which unsettles our certainty about the reality of his photograph, to prolong our uncertainty about its identification and to insist that our re-membering and re-construction of the source remain active through our 'reasoned conjecture.' Notably, Demand is discouraging comparison of the source image and his final photograph in terms of spatial correspondence or of resemblance. I think that he is warning against our reliance on the criterion of resemblance for understanding the indexical or causal logic of the photograph. Rodowick insists that photography's primary power lies in its temporal and existential qualities: a photograph is not necessarily a likeness, but an assertion of existence in past space-time. Demand discourages comparison of his source photograph and final photograph in terms of resemblance, and notably he does so through language that situates them temporally: "before and after." His use of a photographic model to construct a paper model, even when not literally doubled as in Model becomes another opportunity to confuse the stakes of spatial correspondence with those of temporal indication. Inasmuch as the (Platonic) model has been central to a history of the mimetic tradition in the visual arts, Demand contends with its long-standing implications about originals and copies. But insofar as the concept of the model is temporally-oriented toward a future realization and simultaneously toward a past prototype, Demand uses the model for these temporal implications. The temporal layers thicken when he retrospectively returns the already-realized building site to its 'before,' in the form of the paper model, and beyond that to its 'innocent before,' by doing this after a tragic or historic event has occurred there. Demand gives us a seemingly real photograph of a strangely ordinary site - and hopes that the signs of his interventions - his literal cuts with scissors or his cut content - will startle or confuse us enough to prompt interaction - our own reconstruction. Demand's literal transcoding preys on our fears of being tricked by digitally transcoded photographs. But his photographs do trick us, and this trick unfolds in the duration of our viewing. His photograph may look real, but its realism begins to fail and in failing, it shows us the terms of its existence, the terms of its construction - a construction that was (his models are discarded) really there. As the visibility of digital intervention becomes increasingly invisible, we might not experience its trick and it may go unnoticed. But this should remind us that any image can look innocent of its construction or manipulation; the onus of discovering causal origins through reasoned conjecture is (and always has been) on us.
Richard Eoin Nash, "Follies and Falsities: Architectural Photography," PAJ: A Journal of Performance and Art 23 (May 2001), 74.
Massachusetts Museum of Contemporary Art, exhibition brochure, Supermodel (2000).
Camille Morineau, "The Empty Image," trans. C. Penwarden, Art Press 264 (January 2001): 33-40. Wilson Web (200100104392007).
D.N.Rodowick, The Virtual Life of Film (Cambridge: Harvard University Press, 2007), 143.
Rodowick, The Virtual Life of Film, 147.
Morineau, "The Empty Image"
Lev Manovich, The Language of New Media (Cambridge: The MIT Press, 2000), 295.
Rodowick, The Virtual Life of Film, 121.
Rodowick, The Virtual Life of Film, 121.
Rodowick, The Virtual Life of Film, 122.
Rodowick, The Virtual Life of Film, 122.
Rodowick, The Virtual Life of Film, 101.
Rodowick, The Virtual Life of Film, 102.
Manovich, The Language of New Media, 302.
Manovich, The Language of New Media, 66.
Rodowick, The Virtual Life of Film, 119.
Rodowick, The Virtual Life of Film, 115.
Rodowick, The Virtual Life of Film, 58.