FALL 2012: V.08 N.02: FOUND – SAMPLED – STOLEN – STRATEGIES OF APPROPRIATION IN NEW MEDIA
Grant Taylor, Ph.D.
Associate Professor of Art History, Lebanon Valley College, Pennsylvania.
As difficult as it is to believe, the very first examples of digital art are now over half a century old. It was in the summer of 1962 that A. Michael Noll sent a memo explaining he had generated a “series of interesting and novel patterns” on the IBM 7090, the same model NASA employed to launch the first American astronaut into space.  In an attempt to avoid debate or provoke the displeasure of his employer, Bell Telephone Laboratories, he called his creations ‘patterns’ rather than ‘art.’ This more innocuous term, however, did not hide the importance this simple memo held for the history of art. Even then, this young scientist sensed the significance of his discovery, foreseeing that a new type of artist, the “artist-programmer” as he imagined, would one day generate “true art.” Beyond defining this emergent medium, Noll viewed the computer as the ultimate research tool, an instrument with the power to explore the production and reception of art.  For other scientists and engineers at Bell Labs, Noll’s memo raised important questions about the nature of art. Could the human aesthetic response be finally decoded, or could aesthetic art objects be digitally encoded, or, with more far-reaching consequences, could synthetic forms of creativity exist independent of humans?
Though such heady technical and philosophical inquiries were only implicit in the memo, Noll did suggest a course of action that began to answer those key first questions of the digital age. For in his research, he noticed that his ‘Patterns by 7090’ held a resemblance to modernist abstraction. The machine that embodied universality like no other – the digital computer – could in fact reproduce the artwork of others. For Noll, a subjective experiment that compared computer-generated art to art made by human hands might determine how the aesthetic experience functions. Within a year, Noll would embark on his “human-or-machine” experiment, a comparative study that would ultimately pit the computer against the artist. Through a novel act of digital appropriation, the scientist would first approximate and then encode the artistic process of Piet Mondrian, allowing the machine to generate with almost exact fidelity the work of the modern master. Surprising many, the subjects of his experiment preferred the computer-generated to the actual Mondrian. The experiment seemed to put at risk notions of artistic genius and originality, tenets so essential to the mythology of modern art. Stuart Preston, the New York Times critic who provided the first review of Noll’s computer-generated art, was forlorn at a future in which “almost any kind of painting” could be computer-generated.  The art critic, sensing the power of this new mechanical appropriator, believed that the computer could undermine the ontology of art and efface the identity of the artist. Eventually, however, Noll’s digital simulacrum, reinforced by highly articulate writing from the talented scientist, would become the most widely reproduced examples of what was then called ‘computer art.’ In what was a rare example of recognition, since computer art was ostensibly marginalized by the mainstream art world, Noll’s experiment received praise from Meyer Schapiro, one of the most renowned art historians of the twentieth century. 
The appropriation impulse first came to the scientist when he closely examined a work entitled Gaussian-Quadratic (1963), a simple linear composition from his initial experiments with the IBM 7090. The angular lines reminded the scientist of one of his favorite paintings in the Museum of Modern Art, Picasso’s synthetic cubist masterpiece Ma Jolie (1912). His computer-generated arrangement appeared to resemble the vertical superstructure of Picasso’s composition. Like Sherrie Levine, the 1980s appropriator who located the source of her art in the exhibition catalogue of famed American photographer Walker Evans, Noll proceeded to scan catalogues and monographs in search of an artwork that would be amendable to the then rudimentary image construction capabilities of the computer and microfilm plotter. One artwork in particular stood out: Mondrian’sComposition in Line (1916/1917), a picture with precisely placed horizontal and vertical black bars over a white background.  To generate his digitized Mondrian, Noll utilized the latest IBM mainframe, the 7094, which was the same machine employed in the epoch defining Apollo 11 lunar landing. For Noll, transforming Mondrian’s composition into numerical data was a simple reverse-engineering process. By identifying the Cartesian coordinates for each line’s end point, Noll could turn Mondrian’s painting into a program. But a perfect copy of the original did not interest Noll. He wanted the computer to apply some of its own creative agency. As a substitute for artistic intuition, Noll employed a pseudorandom number generator which varied the bar density, lengths, and widths of Mondrian’s lines. The scientist then manipulated the program until it closely approximated the Mondrian painting. Through running the program iteratively, a composition, which Noll entitled Computer Composition with Lines, appeared. The image was strikingly similar to the Mondrian original. Viewed in isolation, the image would be immediately recognizable as an early Mondrian. To the trained eye, however, there were subtle differences. The marks overlapped more frequently and it placement of lines appeared more random than the more ordered original. For Noll, the imprint of randomness was important; after all, by more closely matching the Mondrain’s composition to the parameters of his algorithm, he could have generated a computer version that would fool the most astute Mondrian specialist.
Gaussian-Quadratic, 1963, A. Michael Noll, IBM 7094, Stromberg-Carlson S-C 4020 microfilm recorder © A. Michael Noll
IBM 7094 with IBM 7151 Console © IBM Archives.
Because the computer had simulated Mondrian’s schema so successfully, something he had foreseen two years earlier, Noll felt an experiment that compared and contrasted the two would reveal some interesting findings. The experiment that Noll was to emulate was the famed Turing Test, an experiment that provided an early benchmark for the emerging field of artificial intelligence. Invented in 1950 by the father of modern computing, Alan Turing, the test objectively verified the intelligence of a system, machine or otherwise. In the spirit of post-war behavioral psychology, the test measured success by the number of human subjects fooled by his machine. If a questioner could not distinguish between the human or machine response, the machine was said to possess intelligence. Noll’s experiment worked on the same premise, albeit to simulate or actuate human creativity. Noll’s test involved taking xerographic copies of the two artworks and presenting them to one hundred subjects, all who worked at the Bell Labs in a technical or non-technical capacity. The sample taken was representative of a scientific research laboratory, although the subjects had educational backgrounds that ranged from high school to post-doctoral. Mimicking the Turing Test, the subjects were to identify which picture they thought was human made and which was computer-generated. As an added dimension, which placed the experiment somewhere between applied visual psychology and experimental aesthetics, the questioner asked which picture they preferred. The results showed that 59% of the subjects preferred the computer-generated image and only 28% were able to identify correctly the picture produced by Mondrian.
Computer Composition with Lines, 1965, A. Michael Noll, IMB 7094 and General Dynamics SC-420 micro-film plotter © A. Michael Noll
Avoiding the flawed approximations that were so often part of artificial intelligence discourse, Noll recognized the inherent weaknesses with such a subjective experiment. Creativity, as Noll admitted, was very difficult to define in objective terms. The reduction in size of the Mondrian and its xerographic reproduction would have undoubtedly degraded the painting’s aesthetic quality. Also identified was the fact that a larger proportion of the subjects with technical training would identify the computer picture because of their fluency with computers. As Noll admitted, if artists and subjects from a non-technical environment had been similarly tested, the result might have differed. Indeed, if Noll had conducted this experiment within the realm of the art world, the anti-computer prejudice, so dominant at the time, would have undoubtedly impacted the results. Nevertheless, the results showed that the computer-generated pattern was preferred. As Noll recorded, the subjects described the computer-generated picture as being “neater” and more “varied,” “imaginative,” “soothing,” and “abstract” than the Mondrian. 
Noll’s experiment was largely framed within the ‘machine versus man’ paradigm of artificial intelligence. Through the 1950s, as the AI field began to take form, the computer began to reflect the human mind as no previous machine had. Noll’s experiment, completed in the early 1960s, extended the exploration of the computer’s capabilities in the direction of creativity, specifically the visual arts. Importantly, the experiment questioned the belief that creativity, as Noll wrote, is “the personal and somewhat mysterious domain of man.”  While the experiment promoted the computer as a complement to the artist’s powers, it implicitly framed the machine as a possible future competitor to the artist. Other scientists and technologists were more direct, believing that mathematical formalization could finally purge art of its primordial mystique. The computer would prove that fine art was no longer the domain of the “artistic genius,” or, as Emmanuel Kant suggested, “a talent for producing that for which no definite rule can be given.”  Science found in the computer the possibility of a fully mechanized art, or, as computer art pioneer and theorist Herbert Franke put it, the final “delegation of the aesthetic-creative processes to machines.” 
But did Noll’s experiment produce art? After all, an engineer-trained scientist produced this sophisticated experiment, which became a classic in early computational psychology. Computer Composition with Lineseffectively made the transition from stimuli for a science experiment to experimental art in April 1965 when it was part of the first U.S. exhibition of computer art at Howard Wise Gallery in New York City. Howard Wise, the gallery’s director, was receptive to the latest technologically-based art and approached Bela Julesz, a colleague of Noll’s at Bell Labs, to exhibit his random-dot stereograms. Julesz, who knew of Noll’s innovative visual art experiments, urged his fellow scientist to particpate in the exhibition. From the onset, however, problems besieged the exhibition. Julesz was not pleased with the use of the term ‘art’ in the title of the exhibition because his images were stimuli for psychological investigations into visual perception. On the other hand, Noll was quite comfortable in identifying his works as art because his production was made “solely for their aesthetic or artistic effects.”  While a compromise was reached by titling the exhibition “Computer-Generated Pictures,” much of the ambivalence over whether or not to call it art was associated with the initial response by AT&T, the parent company of Bell Labs. When AT&T heard about the exhibition, they wanted it cancelled, believing such important scientific work would be trivialized by its appearance in the world of art. However, Wise Gallery, who had invested in the exhibition, threatened to sue. Finally relenting, AT&T ordered the scientists to avoid publicity and to seek personal copyright of the images. However, when Noll attempted to register the copyright for Gaussian Quadratic with the Copyright Office at the Library of Congress, they refused. Their refusal was on the grounds that a “machine had generated the work.”  Noll patiently explained that a human being had written the program, which incorporated randomness and order. They again declined to register the work, stating that randomness was not acceptable. The copyright was finally accepted when Noll explained that although the numbers generated by the program “appeared ‘random’ to humans, the algorithm generating them was perfectly mathematical and not random at all.”  Thus Gaussian-Quadratic became the first example of copyrighted digital art. By registering copyright, however, Noll took the position of creative artist, a designation that he was ambivalent about.
As Noll’s attempts at gaining copyright exposed, digital randomness was an entirely unique methodology in the history of art. Randomizers had participated in ancient societies to ensure fairness and to attain divine direction. The methods for generating chance were diverse, from casting animal bones to tossing many-sided die. The use of chance to solicit divine direction – called divination – guaranteed the elimination of human interference and allowed the will of the deity to manifest itself. Since antiquity, however, chance became a subject of mathematical study. Mathematicians in this new field of probability called a series of random behaviors stochastic, from the Greek word stochos, to guess. For much of its history, the field of probability analyzed mostly numerical and statistical data; but in the nineteenth century, mathematicians began visualizing random behavior. In twentieth-century art, Dadaists attempted to overcome the ideologies of rational order by introducing arbitrary effects and chance happenings into the art-making process. However, such chance mechanisms were significantly different from the mathematically determined systems in mechanical chance. For Noll, randomness was more than a metaphor of creativity; it was the actual means of realizing digital production. In Noll’s case, it was the generative mechanism that produced endless permutations of Mondrian. Yet for Noll and those digital artists that followed, aleatory behavior was more than a simple generative instrument– it allowed the computer to become a creative actor in the process.
Noll’s Computer Composition with Lines was not the only work in the exhibition that re-appropriated or referenced the art of another. Seeing that op-art was popular and easily programmable, Noll created a version of Bridget Riley’s painting Current, instead calling his computer version, with a certain mathematical literalness,Ninety Parallel Sinusoids With Linearly Increasing Period (1964). And Noll was not the only technologist to employ the simulating power of the computer to recreate well-known artworks. The young German mathematician and computer art pioneer Frieder Nake exhibited in 1964 computer-generated artwork that was programmed with the statistical laws of the early modernist painter Paul Klee. Indeed, from the beginning, appropriating tactics were at the heart of digital expression.
Ninety Parallel Sinusoids With Linearly Increasing Period, 1964, A. Michael Noll, SC-420 micro-film plotter © A. Michael Noll
But Noll, like Nake, engaged in a unique form of appropriation, one that involved not simply reproducing an image, merely copying Mondrian’s style or replicating the image unaltered, but deeply inhabiting the original by disassembling and resembling it in coded form. Rather than embedding his own artistic process into coded language, he attempted to manifest another’s. In fact, all early digital forms of appropriation were less about surface and the economy of images –the complex interplay of cultural signs that defines much of today’s appropriation art– but were more about the dematerialization and restructuring of art and its praxis, which aligned it more broadly with the emerging conceptual art movement. Certainly Noll’s computer-generated Mondrian did question the status and role of the artist in meaning construction, even before a slew of conceptual artists attempted such reconfigurations. In reality, Noll’s experiment was the first project to scientifically demystify artistic process; it fully realized Walter Benjamin’s idea of the decay of the aura of the artist through mechanical reproduction. But the similarities between conceptual art of the 1970s and Noll’s innovative strategies end there. Noll’s appropriative practice was not overtly political, nor did it question modernist authority (remember the scientist was careful not to detract from Mondrian’s achievement).
Noll’s practice held little resemblance to avant-garde studios of the era. When art theorists reconstruct the genealogy of cultural appropriation, they typically define the 1960s in terms of Andy Warhol and his Manhattan studio Factory, a cultural hub where various modes of mechanical reproduction were employed to replicate consumer icons. But in 1962, the year that the Campbell’s Soup Cans series was exhibited, perhaps Warhol’s most famous act of appropriation, Noll was envisaging his revolutionary digital reproduction techniques across the Hudson River at Bell Labs in Murray Hills, New Jersey. Bell Labs, though not mentioned in orthodox histories of art, is the site in which today’s digital art object was conceived and developed. Bell Labs, like Warhol’s famed Factory, was –as the renowned science-fiction writer Arthur C. Clarke described– an “idea factory.”  During the 1960s, acknowledged as Bell Lab’s golden age, this industrial research laboratory produced the technologies that became the foundation of digital culture. From transistors to the UNIX operating system, Bell Lab was, as Noll described, a “powerhouse of invention and discovery.”  For a good part of the twentieth century, Bell Labs was, as Jon Gertner wrote, the “most innovative scientific organization in the world.” 
In the United States, when Noll started his experiments, artistic interest in industrial aesthetics and post-war technology was in its infancy. Soon machines would redefine the traditional art studio.  But as a place of art production, Bell Labs was wholly unique in the history of twentieth-century art. While AT&T, the parent company of Bell Labs, instituted a highly structured organization, requiring its scientists and engineers to support its expanding telecommunication network, there was room for highly speculative research projects, what Noll called “diversions.” The laboratory’s highly charged and competitive environment attracted the best scientists, those passionate individuals who worked by day on official work and by nights and weekends on their own projects. Mervin Kelly, one of the many innovative presidents of Bell Labs, saw the laboratory as more of “an institute of creative technology,” an outlook that would eventually redraw the dividing lines between art and science.  Beyond his experimentation with the still image, Noll would invent and patent stereoscopic, holographic, 3-D animation, and interactive technologies. But Noll was not alone; he was part of a larger group of imaginative technologists, who, as a result of their experiments, would shape the digital medium. During this innovative era, Edward E. Zajac, Kenneth Knowlton, along with famed animator Stanley Van Der Beek, redefined film and animation, and Max Mathews and John R. Piece revolutionized music by means of digital production. Later as the orbits of the art world and Bell Labs would align under the auspices of E.A.T (Experiments in Art and Technology), artists like Lillian Schwartz would join the research team at Bell Labs. She would eventually stay, forging a career that would see her become a pioneer of computer art and applied visualization. Building on aspects of Noll’s experiments, the artist constructed a practice that was formed largely on new appropriative methods. In her essay, Computer and Appropriation Art, Schwartz sought to understand the creative act by digitizing the work of Duchamp, Van Gogh, Picasso, Leonardo, and Hiroshige. 
However, Noll’s legacy was felt far beyond the laboratories at Bell. Noll’s research continues to resonate for a new generation of software artists. If we consider the work of Casey Reas, Joshua Davis, Mark Napier, and Benjamin Fry we find the same generative structures driving their coded art forms. But it is the practice of John F. Simon Jr. where the Noll’s impact is most evident. Simon, who found Noll’s work inspiring as a student, reimagined Mondrian’s masterpiece Broadway Boogie-Woogie (1943) in his artwork ComplexCity (2000). This screen-based work explored Simon’s experience of Manhattan, an experience mediated through Mondrian’s original 1940s response to the city. The pixel-like construction of Mondrian’s city grid and the primary color system of red, yellow, and blue are endlessly reformulated in Simon’s animated abstraction. Both Noll and Simon trap Mondrian’s practice at a moment of time, utilizing the power of permutation to create endless variations of Mondrian’s visual system. In Endless Victory (2005), Simon re-imagines the visual schema of Mondrian’s Victory Boogie-Woogie (1944), the modernist’s last yet unfinished work.  In Simon’s algorithm, most impressively realized in his media installation HD Traffic (2009), Mondrian’s project remains unfinished, but this time constant deferment comes in the endless stream of stop and start motion found in flow of city traffic.
ComplexCity, 2000, John F. Simon Jr., Software, Macintosh Powerbook G3, acrylic, 19 x 16 x 3 1/2 inches, Edition of 12 © John F. Simon Jr.
HD Traffic, 2009, John F. Simon Jr., Streaming Museum, Cocor MediaChannel, Bucharest, Romania © Streaming Museum, NYC. YouTube link: http://youtu.be/etMSra7gW-k
Because Noll was a scientist –not a trained artist– his position as a creative force and visionary is not recognized in traditional histories of art. It should be. Long before today’s artists were able to sample and remix from the global traffic of images and sounds, Noll understood that digital modes were essential to future cultural production. He saw too that the infinite expressive potential of computer languages, the poetics of code that software artists describe today, would shift our understanding of what constitutes the art object. Using the same radical machines that produced the twentieth-century’s most extraordinary achievement –space exploration– he sought to probe the structural boundaries of art. While he was not the first to ask if creativity should only be associated with human production, Noll was the first to attempt to formulate and test an algorithmic model for creative visual behavior. In many ways, Noll fired the opening salvo in the field of computational creativity. Noll’s experiment examined new aspects of machine agency beyond intelligence, positing the computer as a possible creative subject, or at least a close collaborator with the artist. And through his Mondrian experiment, he brought together the fields of artificial intelligence, cognitive psychology, and philosophy, an interdisciplinary perspective that would eventually define the expanded field of new media art.
1. A. Michael Noll, “Patterns by 7090,” Bell Labs Technical Memorandum, August 28, 1962.
2. Noll, “Patterns by 7090,” 4.
3. A. Michael Noll, “The Digital Computer as a Creative Medium,” IEEE Spectrum, Vol. 4, No. 10, (1967): 89-95.
4. Stuart Preston, “Art Ex Machina,” New York Times, April 18 1965, X23.
5. Meyer Schapiro, Modern Art: 19th & 20th Centuries, Selected Papers, (New York: Braziller, 1978). 252-254.
6. See Mondrian’s Composition in Line (1916/1917) at the Kröller-Müller Museumhttp://kmm.nl/object/KM%20106.482/.
7. A. Michael Noll, “Human or Machine: A Subjective Comparison of Piet Mondrian’s ‘Composition with Lines’ (1917) and a Computer-Generated Picture.” The Psychological Record 16 (1966): 11.
8. Noll, “The Digital Computer,” 89.
9. Immanuel Kant, Critique of Judgement, trans. J. C Meredith (Oxford: Clarendon Press, 1952), 168.
10. Herbert W. Franke, Computer Graphics – Computer Art. Translated by G Metzger, (New York: Phaidon, 1971), 57.
11. Noll, “The Beginnings of Computer Art,” 41.
14. Arthur C. Clarke, Voice Across the Sea (New York: Harper & Row, 1974), p 152.
15. A. Michael Noll, “Bell Lab” http://noll.uscannenberg.org (accessed August 10, 2012)
16. Jon Gertner, The Idea Factory: Bell Labs and the Great Age of American Innovation, (New York: Penguin Press, 2012), 2.
17. Caroline Jones, Machine in the Studio: Constructing the Postwar American Artist, (Chicago: The University of Chicago Press, 1996).
18. Gertner, The Idea Factory, 3.
19. Lillian Schwartz, “Computer and Appropriation Art: The Transformation of a Work or Idea for a New Creation,” Leonardo, Vol 29, No. 1. (1996): 43-49.
20. See Mondrian’s Victory Boogie-Woogie (1944) at the Gemeentemuseum Den Haaghttp://www.gemeentemuseum.nl/en/collection/item/4444.
Grant Taylor is an art historian who specializes in early digital art. He completed his graduate and post-graduate work at the University of Western Australia. Taylor currently teaches art history at Lebanon Valley College in Pennsylvania.