June 1996
Roger N. Shepard was always a prankster. As a boy, he delighted in playing visual tricks -- like the time he surreptitiously moved all the furniture out of his sister's room.
He went on to a distinguished career at Bell Labs, Harvard, and Stanford, all the while continuing to play visual tricks on people. By recording and analyzing their reactions, he discovered astonishing truths about mental processes. Shepard showed, for example, that when people compare two objects rotated at different angles, they first reorient the objects in their mind's eye. The astonishing part: Everyone does this at approximately 60° per second.
In the 1950s, Shepard began to discover universal laws that govern how people and animals perceive similarities among sensory input like colors, sounds, or smells. He worked out a computer method for measuring the perceived difference between stimuli.
When Shepard applied this method to colors, the computer yielded a circle. The physical wavelengths of visible light fall along a straight line, moving from red to violet; but the circle more accurately depicts the psychological (as opposed to physical) truth. This is that people perceive red and violet as more similar to each other than either color is to an intermediate color such as green. Shepard also discovered that:
- The distance between stimuli decreases along a particular curve -- a steep drop, followed by a slow tailing off (in mathematical terms, an exponential decay curve).
- All stimuli -- simple or complex -- produce the same curve.
- So do all experimental subjects -- whether humans, monkeys, rats, or pigeons.
Until Shepard began publishing his imagery work in the 1970s, no one had objectively measured mental imagery. It was thought to be impossible. The scientific community has since taken notice.
Today, his work influences fields as diverse as psychology, philosophy, computer science, linguistics, and neuroscience. Shepard says his image orientation work has been used in a pilot aptitude test. The test accurately identifies trainees who are most likely to make pilot errors. His imaging methods were also used in the design of a computer-based diagnostic system for breast and prostate cancers. Joseph L. Young, the NSF program director who has overseen Shepard's NSF grants for more than 19 years, anticipates other practical benefits such as better organization of control rooms and cockpit displays, as well as more effective educational programs.