With the recent release of Magnus, an app that can recognize artworks using visual data, we turn back to the Summer 1977 issue of ARTnews, in which the editors wondered if it was possible for a computer to appreciate art. “Would there be a call for computer art critics?” the editors asked. No, Magnus does not meet that description, but it is being called “Shazam for the art world,” and it can tell you titles, artists, mediums, and prices when viewers take pictures of the work in question with their phones. The 1977 article follows in full below.
“Can computers be programmed to appreciate art?”
Summer 1977
Writing in a recent issue of the British contemporary arts journal Leonardo, Wales psychologist Michael J. Apter asked the fascinating question “Can computers be programmed to appreciate art?” Well, maybe, according to Apter, if we reevaluate our notion of what art is.
The idea of art as a profoundly spiritual—or, at least, emotional—undertaking is questioned by Apter, for starters: “It is possible to think of the whole artistic process mechanistically,” writes the psychologist. Apter’s strategy is arriving at such a mechanistic understanding of art involves dispensing with the question of “whether computers can be said to be creative in a human sense or feel conscious emotions,” since “these important and interesting questions have been discussed by philosophers and others.” Instead, he advises thinking of the human brain itself as a kind of computer. “It might seem that an art work should be considered as a collection of data produced by an artist’s ‘computer’ that is analyzed by an observer’s ‘computer’,” is Apter’s way of putting it.
He goes on to an interpretation of the “collection of data” that is an artwork as a program of the sort fed into a computer. The program, or artwork, operates on the “emotional subsystem of the observer’s ‘computer’ so that certain emotions are automatically aroused. It is as if an observer’s ‘computer’ was a musical instrument being played on by the artwork’s program.” New styles of art such as Impressionism or Cubism are not generally understood when first encountered, writes Apter, because one is not emotionally aroused. Yet soon enough the brain develops a “compiler” (a special program used by computers to decode “high level language”) responsive to these styles. For Apter, this suggests a function of art beyond emotional satisfaction: “…it exercises and develops one’s own ability to understand new and complex structures.”
Apter’s “structure” gets more complex by the page. A work of art, he states, is a model—”a structure that is simultaneously an exemplar of a theoretical construct and a simplification of reality.” Models are also used in science—an airplane model used for a wind test in a tunnel, for example. Works of art such as figurative paintings “may be said to exemplify some general idea, theme or argument in a particular way and at the same time to represent in a simplified and selective manner chosen aspects of the real world,” like the airplane in the wind tunnel. Apter cites Goya’s anti-war paintings as good examples of models of “abstract religious notions,” and Rembrandt’s drawings as “simplification…of an extreme form, a few lines representing a whole complex scene.”
Having cinched his arguments for the brain as a computer and the artwork as scientific model, Apter has to admit that “there are many difficulties at the present stage in converting such a mechanistic interpretation into a viable computer program.” Great strides have been taken in this direction, however, since, the author points out, computers are getting better and better at processing “natural language” and in incorporating “emotional subsystems.” In fact, as computers evolve and become more and more lifelike, they may evidence a need for an art on their own—programmed, of course. “Such programs would … be functionally equivalent to artworks; but they would be specially designed for computers… If computers themselves were used to produce such artwork, then we would have the intriguing development of computer art for computers themselves.” Intriguing indeed. Would there be a call for computer art critics, their insightful, critical printouts reading something like, “The works of Computer 352911 operated effectively on the emotional subsystem of this observer, exercising my ability to understand new and complex structures and setting my compiler whirling”?