A. Michael Noll — Gaussian Quadratic

1963

Artists: A. Michael Noll · Year: 1963 · Medium: Computer-generated drawing, microfilm plot · Substrate: IBM 7090, microfilm plotter · Place: Murray Hill, New Jersey, USA · Institution: Bell Telephone Laboratories A. Michael Noll was an electrical engineer at Bell Labs who began making computer-generated images in 1962 more or less as a lunchtime habit. Gaussian Quadratic, from 1963, is an early plotter work — a sparse composition of straight line segments plotted according to two-dimensional Gaussian distributions. It predates Nees's Stuttgart exhibition by nearly two years and is among the earliest surviving computer-generated drawings intended as art. What Noll is best known for is a different piece of work. In 1965 he ran an experiment: he generated, algorithmically, a drawing in the style of Piet Mondrian's 1917 Composition with Lines. He then took his computer-generated image and the original Mondrian and showed both, side by side, to a hundred subjects — students and colleagues at Bell Labs. He asked them to identify which was the Mondrian. Fewer than half guessed correctly. A slight majority actually preferred the computer-generated version. His paper on the experiment, "Human or Machine: A Subjective Comparison of Piet Mondrian's Composition with Lines and a Computer-Generated Picture," was rejected by art journals and eventually published in The Psychological Record in 1966. Noll's own position on the experiment was more careful than it has been remembered. He was not claiming that the computer had made art equal to Mondrian's, and he was not claiming that viewers could not tell them apart in principle. He was claiming that the perceptual criteria people actually use to recognize a Mondrian are statistical rather than authorial — that a composition obeying Mondrian's rules at roughly the right density reads as Mondrian, whether Mondrian made it or not. The experiment was about perception, not replacement. It was widely misread in both directions. Unusually for a first-generation figure, Noll wrote his own history. He has published extensively on the early years of computer art at Bell Labs, and his own archive at USC's Annenberg School is one of the field's best primary sources. Born 1939, still active. He is the rare pioneer whose memory of the period can still be consulted. Lore Noll was a Bell Labs engineer who'd started making computer-generated images more or less as a lunchtime habit. In 1965 he ran the experiment for which he's best remembered: he generated an algorithmic near-copy of Mondrian's 'Composition with Lines' (1917), showed both images side by side to a hundred subjects, and asked which was which. Most could not reliably tell them apart — and a small majority actually preferred the computer-generated version. The paper he wrote on it was rejected by at least one art journal on the grounds that a computer could not, by definition, produce art. Noll later argued the entire project was about aesthetic perception, not replacement. He is still living and has published extensively on the early history himself — he is one of the few first-generation voices to write his own. Sources • Noll, A. Michael — 'Human or Machine: A Subjective Comparison...' (The Psychological Record, 1966) • Bell Labs Technical Memoranda • Noll's own writings at noll.uscannenberg.org