Known and Strange Things Page 16
And all of a sudden, one day, much more recently, there was Google Search by Image (no prizes for whoever named it). Somewhere deep in the thousands of parallel machines running searches, an algorithm had emerged for matching an image to exact and inexact versions of itself across the Web. This makes it sound mystical, but in actuality there were precursors to Search by Image—sometimes called Reverse Image Search—at Google itself, and at TinEye. Research in the last decade into object recognition and into the matching of visual data (research done not only at Google but at other corporations and at universities; see, for example, the work of Abhinav Gupta and Abhinav Shrivastava) was brought to a point of refinement, incorporating findings about how to estimate the relative importance of different parts of a picture. And, in time, sophisticated code that made use of artificial intelligence was written, by a team led by Amit Singhal and Ben Gomes.
So, the science was straightforward (though not simple), but the result was mystical: faces and places and scenes were legible in a new way, shorn of their mute anonymity. A photograph—say the famous one of a child carrying bottles by Henri Cartier-Bresson (Rue Mouffetard, 1954), or a tourist’s snapshot of her family in front of the Taj Mahal—could be pinned to the wider world of data in a precise way. If not, if it were a unique image of a situation or of an unremarkable location, it could be brought into the family of images. This latter, a more involved and more interesting technical feat than mere matching, was presented as pages of “visually similar images.” This use of visual similarity had a formal and coloristic sophistication that took me aback: it seemed to demonstrate an awareness of the essence of a photograph. Light, shadow, color, intensity, and composition were integrated to unite images that were similar. These were images that shared a kind of visual DNA, the kind that is apparent at a glance. But the images were not at all related to each other in terms of content.
In Naples in 1868, Vittorio Imbriani published a pamphlet entitled “La quinta Promotrice.” In this now almost forgotten text, he advanced a peculiar theory of art that centered on the color patch or, in Italian, the macchia (the word literally means stain or spot). According to Imbriani, the macchia is “the image of the first distant impression of an object or a scene, the first and characteristic effect, to imprint itself upon the eye of the artist.” It is, in other words, the total compositional and coloristic effect of an image in the split second before the eye begins to parse it for meaning. Imbriani was writing against the academic idealists of his time, who were obsessed with categories of style and execution. “Every painting must contain an idea,” Imbriani wrote, “but a pictorial idea, not just a poetical idea.” And for him, this pictorial idea was a matter of “a particular organization of light and dark from which the work takes its character. And this organization of light and dark, this macchia, is what really moves the spectator….Equally in music it is not the attached words, the libretto, but the character of the melody that produces emotion in the listener.”
Imbriani’s was an argument for the inner life of pictorial effect, not so much about the way in which visual organization transcended subject matter but the way in which it preceded subject matter. For Imbriani, a picture lived by its idea, by its melody. However, this wonderful and cogent theory gained little traction, and with the exception of a reconsideration by Benedetto Croce in 1929 and an adaptation by Hans Sedlmayr (who misunderstood Imbriani’s main argument) in 1934, it never really entered the mainstream of art history.
—
In my years of struggle as a street photographer, I had stumbled across Imbriani’s truth without knowing it: in each picture, I was building on a pictorial rather than a poetic idea. The same, I feel, is true of some of the color photographers I admire most. I love conceptual photography, but I also like to “run around taking pictures of things.” I am now using Search by Image for my own personal color studies, for my own understanding of the color patch, and I don’t know what might come of it. I’m watching to see what other photographers and artists can create using this latest manifestation of Big Data. As usual, they will have to navigate between the Scylla of copyright issues and the Charybdis of “that’s bullshit, not art.”
When I began to put some of my photographs into Google’s Search by Image, it was like an arrival. This was work I had thought of in terms of a certain disregard for subject matter. I had a passion for light and shadow, a tendency toward ostinato patterns, and an almost fanatical interest in color distribution. In introducing images to one another (as I did in a previous project I called Who’s Got the Address?) I was working with a notion of reuniting what had been shattered in some unseen pictorial prehistory. At other times, I had even organized some of my images by color kinship: the red and black ones in one group, the blue-green ones in another. But I did not understand fully that what I was after was macchie. This is what became clear when I gave myself over to Search by Image.
I fed a few of my photographs into the machine (there’s a warning on the support page: “When you search by image, any images that you upload and any URLs that you submit will be stored by Google. Google uses those images and URLs solely to provide and improve our products and services”), and out of the “visually similar images” results, I selected the ones that best expressed the idea I had been after when I took the original picture. The photographer’s task not only becomes curatorial, it has in fact always been curatorial. In each case I ended up with an assemblage of nine images, including my own original. I had arrived at Google’s macchia and, beyond that, unanticipated narratives. An out-of-focus picture of a woman with a hijab was now surrounded by her color patch relatives: a fighter jet, Bill Maher being conspiracy weary, President Obama in front of some computers, a lone rioter, a woman wearing a bikini, a policeman confronting what seemed to be, in the distance, a group of women in black hijabs.
Another photo of mine, an adumbrated red image of the interior of the Musée d’Orsay, led to unexpected prodigies of vision: a fashion show, the new pope (seated like Don Corleone, flanked by consiglieri), North Korean acrobats, a New York City street, and a night raid. The pictorial idea of each picture told me what I knew but hadn’t articulated about the pictorial idea of my own picture, its rhetoric of red and shadow and scatter. It was like hearing a familiar tune played on unfamiliar instruments, with dramatic changes in the timbre but the pitches staying the same.
It was an uncomplicated experiment, not an attempt to create a new work of art. But I could see that there was potential there, perhaps not for me, but surely for others who might want to think through the capabilities of Google Image or Google Search by Image. Machine had once again met Mind, and Machine knew of Mind what Mind itself had not known. The Machine had its own ideas. But in the shards of light flashing off its silhouette were also, as usual, new possibilities for art, new light for the soul.
The Atlas of Affect
NEAR THE END of his life, Aby Warburg (1866–1929) found a form that began to answer his questions about images. The questions had centered on the relationship between memory and history. Somehow, the conventional practices of art history had left Warburg unsatisfied. There was a deeper logic, he believed, between certain classical images, and between classical images and the ones that came later.
This restless search led him to amass an idiosyncratic library on Renaissance scholarship (this library was transferred from Hamburg to London after his death, and became the core of the Warburg Institute). And it was also this search that propelled him to travel to the United States and live for several months in the mid-1890s among Hopi and Zuni people. The great insight happened many years later, in 1924, when Warburg set up large black cloth screens, and began to pin newspaper and magazine cuttings of paintings, prints, and photographs on them. Each screen was organized around an idea, a complex theme, and the sequence of images was a matter of reiteration as well as of imaginative leaps.
One panel was an exploration of the afterlife of classical gestures, organized around a painting b
y Ghirlandaio. Another looked at the iconological value of the figure of the hurrying nymph, and included paintings by Botticelli and Raphael, as well as a photograph of a young woman from Hamburg. The panels of Warburg’s Mnemosyne Atlas owe something to the systematic imagination inherited from the eighteenth century and to the atlases of that time. But there is something else going on in the Mnemosyne Atlas: it is neither systematic nor complete. It borrows the form of the atlas for something more surreal, more suggestive, and more affective.
Scholars in the 1980s and 1990s began to pay more attention to Warburg in part because of Walter Benjamin’s interest in him, but also because, like Benjamin’s, his talent for montage-like effects was seen as emblematic of the twentieth century. In a way, the Mnemosyne Atlas had begun to do in images what Benjamin did a few years later in words with his Arcades Project. These projects, as Benjamin wrote, sought to “develop to the highest degree the art of citing without quotation marks.” The use of images in conversation with other images (in other words, the use of dialectical images) became one of the standard gestures of the art of the twentieth century. It found powerful expression in works like Tarkovsky’s Mirror and Gerhard Richter’s incessant collection of cutouts, prints, photos, and fragments: the Atlas Micromega (1962–2013).
Richter’s Atlas Micromega runs into hundreds of sheets, of faces, snapshots, mountains, cities, paintings, candles, nudes, landscapes: in short all the material that fed his large-scale pieces. But the Atlas Micromega exists also for its own sake: as a testament to what the artist saw, and what he collected, and how he sequenced it. I see a through-line here, from Warburg to Benjamin to Richter, and, finally, to Baltimore artist Dina Kelberman, who makes a long sequence of images and short films she has found on Google and on YouTube. But if one strand of the genealogy of Kelberman’s I’m Google is indebted to this high-art lineage, another strand is about something else entirely: the kind of “atlasing” that only Google could make possible. This project is involved in the affective language of Warburg (Benjamin writes in the Arcades Project: “I needn’t say anything. Merely show”). There’s a satisfaction for the eye in wordlessly accounting for the link between one image and the one that follows it. It also contains, as in Benjamin, a critique of commodity culture. But Kelberman’s project is, in addition, a visual world-building that explicitly sidesteps not only the language of antiquity and classicism but also any suggestion of “artistic” image making. Her choices are brightly colored, plasticky, almost naïve, and straightforwardly vernacular, less Warburg than Walmart.
Kelberman’s images are related to each other by a more transparent and less obviously intellectualized logic. By contrast, the deadpan affect of Richter’s Atlas Micromega is still freighted with a modernist melancholia. Kelberman’s images are all found images, discovered by trawling through Google and YouTube, mostly by use of keyword searches, and her selection process excludes images with an intentional artistic intent. In a sense, she has arrived at a goal Richter stated in his Notes 1964–65: “I like everything that has no style: dictionaries, photographs, nature, myself and my paintings.” But Richter has plenty of style; it is Kelberman whose hand-chosen images, with their impeccable timing and straight-faced absurdity, come close to being styleless.
Indeed, viewers have sometimes assumed that I’m Google is simply the result of a very clever computer program, a bot set free on Google Image Search and directed to Tumblr, rather than the selective record of countless hours of looking and sifting. As Kelberman said in an interview, “The blog came out of my natural tendency to spend long hours obsessing over Google Image searches, collecting photos I found beautiful and sorting them by theme.” Anyone could do this. The deeper value here is in Kelberman’s notion of the beautiful: “The images that interest me are of industrial or municipal materials or everyday photo snapshots.” A more automated process—for instance, the one seen in Taryn Simon and Aaron Swartz’s conceptually interesting Google-based Image Atlas (2012)—feels more limited and less suggestive, and more susceptible to distracting mistakes, in part because of the absence of a curatorial hand. It is worth pointing out that Kelberman’s work exploits a strain in Internet art that challenges the viewer to assume it was made by a bot, not an artist: a kind of reverse Turing test.
I’m Google uses the possibilities of Google Image Search in a way quite distinct from what I considered in my essay on Google’s Search by Image (“Google’s Macchia”). Kelberman has said that she does not find Search by Image particularly helpful for this project. And yet, even in its styleless style, I’m Google does explore color patch (macchia) theory, the impact of first impressions. But it goes further, relying on visual as well as semantic similarities, in addition to very careful (and occasionally comic) timing. If classical macchia theory is about the nouns of things, and therefore just the appearances, then Kelberman builds similarities also on the basis of verbs: what sprays, what proliferates, what curls or curves or gleams or effloresces in a particular way. She holds the challenge of similarity to a more stringent standard, and yet out of this stringency, which explores static as well as kinetic interests, one feels she could end up just about anywhere.
The bright color palette, flat affect, shadowless lighting, and industrial obsessions in Kelberman’s work—she also makes comic strips, animation, and GIFs, and writes plays—emerge in part from her fascination with Looney Tunes’ Chuck Jones and the epiphenomena of The Simpsons (how clouds or furniture are rendered in the show; another Kelberman project consists of screenshots of such “nothing” moments from The Simpsons).
Kelberman’s wonderfully named comic strips, Important Comics, are pared down in the extreme, often with simple shapes in place of people, but are emotionally anything but simple. In addition, she has an interest in synesthesia and in the ways that the mind makes links and meaning out of apparently unrelated things. And perhaps, for all the disavowal of classicism, there is also a strong classical idea at the heart of the I’m Google project: that of metamorphosis. As in the myths, there’s a little jolt of pleasure (or even fascinated horror) at the moment when something begins to turn into something else. Kelberman accomplishes all this by means of a language that contains the gestures of taxonomy without being taxonomic. She builds an ark of “types.”
In its list making, it is an update of the atlas of affect at the same time that it is a catalogue of what one might call catalogue realism, a sort of Sears, Roebuck catalogue stripped of labels and released into the wild, a sequence that includes logs, fencing, stadia, spray paint, colored paper, sponges, mugs, dirt, houses on fire, geysers, hoses, phone cords, construction sites, molds, paint, tomato sauce, batter, bread, plastic lids, jungle gyms, balloons, venetian blinds, table fans, scaffolding, greenhouses, kindergarten projects, messy bedrooms, televisions, explosions, liftoff, forest fires, forests, disabled-parking signs, auditoriums, soldering, car washes, surf, sand, powdered paint, string cheese, hair, wigs, clouds, foam, yards, roller coasters, blizzards, squalls, tornadoes, lightbulbs, computer screens, windows, booths, ski lifts, zip lines, watchtowers, waterslides, volcanic lakes, sand castles, team uniforms, group photographs, factory workers, crowds, balloons, mannequins, aquaria, window nets, mosquito nets, tents, canopies, piles, packing foam, gym mattresses, wildfire planes, desert rallies, dough, Play-Doh, hearing aids, earbuds, buoys, plastic bags, Ziploc bags, plastic gloves, palm prints, prosthetic hands, balloons, holders, painted bales of hay, silage wrapping, taffy machines, and more, because the world of things never ends.
Memories of Things Unseen
AT AN EXHIBITION at the FotoMuseum in Antwerp, I walked past a large color photograph of a forest. It was an exhibition about landscape in general, organized to give the visitor the feeling of a hike through mountainous terrain. The photograph of the forest was near the entrance of the exhibition, and I had walked past it without stopping because it seemed to be simply another big photo, of which there are so many in museums these days. But after going through the exhibition,
I decided to double back. This time around, I took a good look at the large photograph, which was more than sixteen feet wide, and found that there was more to it than size. The leaves were odd, simplified. I read the caption. What the photograph (titled Clearing) showed was not a forest but a model of a forest. The German artist Thomas Demand had constructed this model from paper and set it in a steel frame fifty feet wide. Two hundred and seventy thousand leaves had been individually cut. The model was illuminated by a powerful lamp, to mimic shafts of sunlight falling through the trees.
The immense labor involved in creating Clearing was part of what made it now interesting to me. But more eerie was the knowledge that Demand had destroyed the model. All that remained was the photo. It was orphaned from its source, and that source would be remembered by only this one angle, this single point of view, under precisely these lighting conditions. The photograph gave us a memory of something we had never seen. Demand had done this intentionally, but Clearing reminded me of other photographs that were inadvertent records of artworks subsequently lost to war or fire. This was the fate, for instance, of Vincent van Gogh’s The Painter on the Road to Tarascon (1888), which is believed to have been lost to Allied attacks on Magdeburg in 1945, late in the Second World War, and Gustave Courbet’s The Stone Breakers (1849), which was incinerated during the firebombing of Dresden the same year. Each now exists only as a photograph.