A

Newell, Allen

Newell, Allen

American Scientist and Mathematician
1927–1992

A scientist and mathematician, Allen Newell is best remembered for his work and research on artificial intelligence (AI). Some of his most well known initiatives include the Logic Theorem Machine, a mechanical device that would be used to create new theorems, as well as the SOAR project, a research initiative that attempted to implement cognitive or rule-based computer simulations.

Newell was born in San Francisco, California, on March 19, 1927, the son of Dr. Robert R. Newell, a distinguished professor of radiology at Stanford Medical School, and Jeanette Le Valley Newell. He attended Lowell High School—the intellectual high school of San Francisco—where he was inspired academically and fell in love (at age sixteen) with fellow student Noel McKenna. Newell and McKenna married at age twenty and remained married for forty-five years.

Newell had no intention of following a scientific career upon graduation from high school. However, after working a summer in a shipyard, he enlisted in the U.S. Navy, and it was during his tenure in the navy that he became interested in scientific enterprise. He was serving on a ship that carried scientific observers to the Bikini atoll (island) to study the effect of nuclear tests, and Newell was assigned the task of mapping the radiation distribution over the atolls. Newell discovered how exciting science could be, and thereafter, he characterized himself simply as a scientist.

Newell received his bachelor of science degree in physics from Stanford University in 1949, spent a year at Princeton doing graduate work in mathematics, and obtained a Ph.D. from Carnegie Institute of Technology (now Carnegie Mellon University) in industrial administration in 1957.

Newell's primary interest, like that of his colleague Herbert A. Simon (1916–2001), was in understanding human intelligence and cognition. He developed the SOAR project with students and colleagues, including John E. Laird, a professor at the University of Michigan, and Paul S. Rosen-bloom, a professor at the University of Southern California, and others. Essentially, SOAR was a rule-based computer simulation or emulation of a cognitive system that was capable of learning and solving problems. The rules are defined by such structures as "If … then …," similar to structures that are thought by some to govern human behavior.

Like Simon, Newell was primarily interested in organizations and their behavior, but he soon moved toward individual cognition. He and Simon had met while Newell was working for the RAND Corporation in Santa Monica, California. It was Simon who influenced Newell to come to Carnegie Tech to obtain a doctorate. Simon and Newell collaborated, and Newell continued to work for RAND in Pittsburgh, as a one-man "office" until he became part of the Carnegie Institute faculty in 1961.

Newell's interest in human learning and thinking was also spurred by Oliver Selfridge, an artificial intelligence researcher, who created theories on pattern recognition, that is, the recognition of letters and other patterns. This led Newell to think of computing as a symbolic manipulation, rather than an arithmetic one, and led him to write a chess playing program c. 1955, which was then implemented by himself, Simon, and John Clifford Shaw in 1956.

Newell also collaborated with Simon and Shaw on the Logic Theorem Machine, a program to find or develop theorems. The theorems were discovered by working backward from the theorem to the axioms in an inductive method of discovery, looking for patterns or regularity in the data. This was an interest that would last his entire life. Newell was also a member of the initial Dartmouth conference, considered to be the first conference in artificial intelligence, along with Simon, Marvin Minsky, John McCarthy, and others.

Newell won the A. M. Turing Award with Simon in 1975. He was also the recipient of the first Award for Research Excellence from the International Joint Conference on Artificial Intelligence and was elected the first president of the American Association for Artificial Intelligence.

With Stuart Card and Thomas Moran, Newell also participated in some of the early research in Human Computer Interaction. This involved the GOMS system of Goals, Operators, Methods, and Selection, a structure for studying human behavior with the computer (and other machines, e.g., calculators) as well as the performance of any task. He also developed the mechanism, with Simon, of "talking out loud" to study the way people solve problems, and the pair was instrumental in developing means-end analysis, a way of explaining how people solve problems that is based on the theory that people notice a discrepancy between their current state and some goal state and employ some operator or operation to remove or overcome the difference.

With Shaw and Simon, Newell developed the information processing languages (IPL-I through IPL-V), which, although not as popular as LISP, were early languages for artificial intelligence. He later took a lead in the effort to develop OPS5 (and other OPS languages), a rule-based language for building artificial systems such as expert systems. His final project, ongoing after his death, was, however, the SOAR system. This was a system that purported to give an architecture of cognition, meant to explore the nature of a unified theory of cognition, a "mental" architecture.