Books I'm Reading: Aaron's Code

Dr. Harold Cohen and his creation, a computer program/drawing machine named Aaron, seem to be virtually unknown today, even among my artist friends. From my first personal encounter with Aaron (in early 1985), however, I thought him an extraordinary creation. Having just read Pamela McCorduck's well written, deeply researched book about Cohen and his machine, Aaron's Code: Meta-art, Artificial Intelligence, and the Work of Harold Cohen (Freeman, 1991), my admiration for Aaron and his creator has only grown.

Before reading the book, I knew little about Cohen's background. I had assumed he was a computer scientist. To the contrary, Cohen was a successful abstract painter in his native Britain before he ever began to dabble in computer programming in the late 1960s, a time when computers were primitive. In 1968, dissatisfied with the life of a painter, he left the UK to accept a position at UC San Diego, where he met Jeff Raskin (later of Apple Computer), who taught him how to write code.

Cohen's initial motivation appears to have been creating a machine that would help him make his own art, but he quickly became fascinated by fundamental questions about the process of artistic creation, the very definition of art, and how art generates meaning—about the cognitive processes that underly art-making and the experience of art. He started down a decades-long road that resulted at first in a program that created crude drawings, based on a complex set of general rules about line and spatial organization, but eventually evolved into a vastly more sophisticated program with "knowledge" not only about the process of filling a sheet of paper with marks that are meaningful to human beings but with knowledge about objects in the real world, including plants and the human figure. In its initial, mark-making phase, Aaron drew in black ink on white paper. Later Cohen gave Aaron color.

Aaron is infinite. Run the program and Aaron will create a unique drawing based on his knowledge about what makes an image. While Aaron has a distinct style, no two of his drawings are ever the same. Now that his creator, Dr. Cohen has died (in April, 2016, at the age of 87), Aaron will learn nothing new, but he will forever be capable of creating art, so long as the platform he runs on remains viable and there are people interested in seeing a new drawing.  

The existence of Aaron raises interesting questions about art, about intelligence, and about the meaning of art and the meaning of making art. Is Aaron intelligent? In some sense, yes. Is Aaron self-aware? No. Is Aaron expressing anything when he draws? No: Aaron is a computer program with no feelings, no motivations, no memory even of what he has drawn before. Yet, what Aaron creates generates meaning in the minds of those who see his work. And this is, perhaps, the most interesting idea McCorduck's book conveys: that Dr. Cohen saw art as generating meaning in the viewer rather than communicating meaning from the artist. Aaron is an excellent argument for the idea that the meaning in a work of art is almost wholly in the mind of the viewer. The book and its subjects are rich beyond my ability to express them. Suffice it to say, this is likely to be an exciting read for anyone with an interest in the human activity we call art-making. Aaron's Code is out of print, but I easily found a copy for a few dollars on line, and good libraries may have a copy*. Recommended.

*I just did an online search: In the Bay Area, there are copies in the libraries of UC Berkeley, UC Davis, Stanford, and in the San Francisco Public Library system.
SaveSave

Art I'm Looking At: A Drawing by Aaron

I was living in Japan in the spring of 1985. Mike, a former college roommate, was working at the American pavilion at Expo '85 in Tsukuba, just north of Tokyo (officially known as The International Exposition, Tsukuba, Japan, 1985), a World's Fair with a focus on science and technology. I took an opportunity to visit Mike, staying a night at his lodgings. He showed me his fat diplomatic passport. He told me about the vetting he had been through. Apparently, someone in the State Department thought it wise to interview some of his grade school classmates. What did they want to know? Maybe Mike ate paste as a child and was not to be trusted? He was the only member of the US delegation not a Mormon, I learned. Mormons don't get drunk, Mormons don't do drugs; they're a safe bet for international duty. With delegations from all over the world at hand, the atmosphere was somewhat party-like after hours. The Mormons seemed keenly focused on finding sex partners. Some of Mike's friends had made plans to go out for a beer with people from the Swedish delegation on the night of my visit. No sex was involved, but I can say "Jag förstår inte svenska" as a result of tagging along that evening ("I don't understand Swedish"). My pronunciation was praised.

Being a stamp collector, I was eager to get Expo cancellations for the stamps issued to commemorate the event. I stood in line. The Japanese have an interest in stamps far beyond anything I've ever seen here. The country still produces beautiful stamps. Collectors stand in line when new issues appear. Stamps in the US are now cheap stickers. US stamps once were gems of good design and often exquisite miniature engravings on fine paper, but the thrill is gone. Shown here are the souvenir sheet Japan issued to commemorate the 1985 event and a strip of five of the ¥50 Tsukuba Expo issue with special cancellations from the site.

At the American pavilion I overheard a Japanese visitor muttering about how the US had come down in the world. Visitors seemed unimpressed by the drab, technical displays. The State Department paid more attention to vetting delegates than to impressing the locals. The Japanese took a World's Fair seriously. The Reagan administration didn't. The Japanese wanted spectacle. They wanted talking robots and cars that drove themselves. They wanted skinny "companions" in slight bathing suits commenting on glitzy displays. What they got at the American pavilion was a rather dryly presented look at recent US developments in artificial intelligence. One corner featured an optical character reader scanning book pages. Another featured Aaron. The rest was not memorable. Aaron was mostly a horizontal plotter, about three feet wide and four feet long, that drew pictures. Aaron wasn't flashy. Toshiba had installed a vertically oriented billboard-sized multi-color plotter in front of its corporate pavilion. The Toshiba plotter drew crowds of people looking up, watching the moving plotter arms, impressed by the size and the color, but the Toshiba machine simply regurgitated pre-programmed images—mostly images promoting Toshiba.

I fell in love with Aaron almost immediately. He wasn't big and he drew in black and white. He may not have impressed most visitors to the American pavilion, but Aaron did something special. He didn't simply reproduce images fed from a computer. The real Aaron was a computer program that controlled the plotter using algorithms to create images—each one unique. Aaron could draw faces and crude figures and rocks and leaves. Aaron liked to put heads with crudely indicated facial features in rocky landscapes. Aaron knew that objects in the foreground seem to overlap picture elements behind them and further away. Aaron had a rudimentary sense of depth and a distinctive, angular, linear style. Aaron was an early—perhaps, the earliest—example of truly creative artificial intelligence. He was about 13 years old when I met him. His creator, Dr. Harold Cohen, later gave Aaron the ability to color his drawings, but, in 1985, Aaron had not yet discovered color.

I coveted one of Aaron's drawings, and I came home with one. Mike secured one for me signed by both Aaron and Dr. Cohen (above). It's dated March 13, 1985, four days before the Expo officially opened, on March 17. At this remove, I can't remember if Mike said Cohen had pre-signed paper sheets fed into Aaron or whether he signed the program's output later. I visited Tsukuba in August 1985, so the example I received had been produced a few months before my visit. I hadn't seen the drawing in many years when I came upon it today in the course of reorganizing my studio. Perhaps it's time to get Aaron's handiwork framed?

Since writing the above, I've found a copy of Aaron's Code: Meta-art, Artificial Intelligence, and the Work of Harold Cohen, by Pamela McCorduck (Freeman, 1991), a detailed account of Aaron's genesis and development. I look forward to reading it. The exhibition of Aaron at the Tsukuba Expo is mentioned. I can see from quickly flipping through the pages that Aaron may have received more attention there than my brief visit suggested. Perhaps the day I watched Aaron draw was just a slow day at the American pavilion, but the space was virtually empty and the visitors I saw were dismissive. One, as noted above, was contemptuous.

I will review Aaron's Code once I've read it*. In the meantime, here are author McCorduck's own words about her book, which, although out of print, is available used online. I found a copy on Amazon for a few dollars.

I've also found this piece about Cohen written following his recent death (in April 2016). He was 87. I had been wondering whether he was still alive. There is much more by and about Dr. Cohen and his work at "Aaron's Home."

*[Update: I have read it now and reviewed it. See the review.]
SaveSave