ALLEN NEWELL 1927–1992

BY EDWARD A. FEIGENBAUM

ALLEN NEWELL, a pioneering computer scientist with broad ranging contributions to information science and technology, died on July 19, 1992, at the age of sixty-five. Newell is considered one of the founders of the field of artificial intelligence, and was a major scientific figure in the field of cognitive psychology.

Newell was elected to the National Academy of Engineering in 1980 and the National Academy of Sciences in 1972. His scientific career was distinguished not only by deep insights and remarkable innovation but also by his concern with creating and nurturing institutions suitable for furthering the growth of computer science. He was a founder of the Carnegie Mellon Computer Science Department—now one of the world's major departments; he was a founder of the American Association for Artificial Intelligence and was its first president. He was also president of the Cognitive Science Society. Over the years he served as adviser to the major government funding agencies for computer science, psychology, and health sciences research.

Newell received his B.S. in physics from Stanford University in 1949 and his Ph.D. in industrial administration from Carnegie Mellon University (CMU) in 1957. Newell was involved with computing almost from the beginning of the computer era. he began his career at the Rand Corporation, helping to start their Systems Research Laboratory for studying behavior of air defense teams. Newell was responsible for programming simulated air defense environments on the primitive computers of the early 1950s. He would later win a major award of the Human Factors Society for this pioneering work.

In the mid-1950s, Newell's scientific attention was captured by the idea of the digital computer as an adaptive mechanism—indeed the possibility of the computer as a cognitive mechanism. His paper on an adaptive chess playing program was quickly followed by a collaboration with J. C. Shaw of Rand and Herbert A. Simon of Carnegie Mellon University.

Rand transferred Newell to Pittsburgh to work with Simon, and in December of 1955 the two of them conceived, and then in a creative burst programmed, the first heuristic problem solving program, Logic Theorist (LT), which proved theorems in propositional calculus. This is the landmark program usually thought of as the birth of artificial intelligence, though the field did not pick up that term until the Dartmouth conference in the summer of 1956.

The method of programming LT was itself a landmark of programming technique: Information Processing Language (IPL). A series of IPLs from 1956 to 1959 introduced list processing, recursive programming, stacks, generators, and many other pieces of the art of programming that we take for granted today.

These language innovations were used to program the influential Newell- Shaw-Simon chess player, and the very influential General Problem Solver (GPS) that set the paradigm for work in artificial intelligence for almost a decade.

After several years as a Rand staff member in Pittsburgh, Newell joined the faculty of CMU in 1961, and later became the U. A. and Helen Whitaker University Professor of Computer Science. He published more than 250 works, including ten books. His most recent book, Unified Theories of Cognition, published by Harvard University in 1990, brings to the public Newell's renowned William James Lectures (1987).

In Newell's last lecture to his CMU students and colleagues in December 1991, he characterized his career as ''desires and diversions." His main desire was the quest for an understanding of mind—mechanistic models of mental processes so detailed that precise predictions of human behavior could be made and tested. A secondary desire was the engineering of computational artifacts that were capable of significant cognitive behaviors.

Newell's diversions would be considered important scientific career achievements to most. He collaborated with (NAE member) C. Gordon Bell on a series of works, including the first book to explain clearly and precisely the nature of a computer architecture. With two of his former students, later scientists at Xerox Corporation, he conceived and tested detailed models of human performance in certain skill tasks that were important in the use of computers.

However, the main line of his interest was to model the general "software architecture" of mind. GPS was his first attempt; a suite of programs called Soar was his last. Soar, begun by Newell at CMU, is now a major effort by dozens of researchers to build a general problem solving architecture that incorporates automatic learning and subgoaling.

Newell won all the major scientific awards of computer science, artificial intelligence, and cognitive psychology, including the A. M. Turning Award of the Association for Computing Machinery (1975), the Distinguished Scientific Contribution Award of the American Psychological Association (1985), the Emanuel R. Piore Award of the Institute of Electrical and Electronics Engineers, and the Louis E. Levy Medal of the Franklin Institute. In 1992, shortly before he died, President Bush awarded him the National Medal of Science.

His great contribution as a scientist is not the measure of Allen Newell. The words that best describe him are gentle, cheerful, smiling, magnanimous, generous, and of course large. There was always the calming influence of having Allen Newell working with you on a problem, whether a departmental problem, a thesis problem, a scientific problem, or a funding problem. His colleagues and his students came to love him because he gave so much to them—to their science and to their lives.