Allen Newell Papers

Allen Newell (b. San Francisco 1927-d. Pittsburgh 1992) received his B.S. in Physics from Stanford University in 1949. He pursued graduate work in Mathematics at Princeton into the next year, studying under mathematician George Polya, who introduced Newell to the study of heuristics, or use of selective search patterns in problem solving. Finding both mathematics and the new trend in game theory not to his taste in "experimental and theoretical research" aimed toward "understanding the nature of mind," Newell took leave from Princeton to work for the celebrated RAND Corporation "think tank" as a research scientist from 1950-1961.

At RAND, Newell became part of an environment that encouraged polymathic thinking and research interests that crossed disciplinary boundaries. There he met consultant Herbert A. Simon, the organizational theorist and future Nobel laureate in Economics, who was also interested in the study of information processing as a route to understanding human decision-making. Beginning under the RAND atmosphere, the Newell-Simon partnership grew into what Newell called a "highly personal and direct relationship" of scientific interests. RAND employed the pair on a project in U.S. Air Force defense logistics which also involved systems programmer J.C. "Cliff" Shaw; their collaborative simulations of radar screen information, mimicking enemy attack at diverse locations, studied the organizational processes of Air Force early warning teams and improved their capabilities. Done on pre-electronic computers, they were also the first efforts at creating a symbol-processing artificial intelligence (AI).

Newell’s long research association with Simon was furthered when Simon encouraged him to earn his doctorate in Industrial Administration (1957) through an AI-based thesis at the Carnegie Institute of Technology (CIT) where Simon taught. While working in Pittsburgh, Newell remained a research scientist on the payroll of California-headquartered RAND until his 1961 appointment as a CIT professor. Communicating with Cliff Shaw and RAND’s Johnniac computer in Santa Monica via an early, semiautomated phone-line computer network, the NSS (Newell-Simon-Shaw) team carried out RAND’s interest in computer simulation of complex thinking through a chess-playing program designed not to win games but to foster understanding of the analytical processes of human players. Equipped now with electronic computers, the NSS team seized the opportunity to use the computer as a general processor for symbols -- or thought simulator -- rather than as a mere arithmetic calculator.

The Carnegie-RAND group also invented the LT (or LTM), a logic theory machine that solved non-numerical problems by selective search. The LT depended on Newell and Shaw’s creation of a series of list processing languages developed by writing and testing programs that would solve logic tasks. To test the LT machine by human cognitive research, Newell ran subjects through problem-solving tasks designed by O.K. Moore at Yale University (1954) to see how closely the computer simulations paralleled human behavior. Finding that humans preferred a means-to-end form of analysis not factored into LT design, the NSS team devised a means-end analyzer, the General Problem Solver (GPS), which provided the bases for most of the AI programs of the next decade.

As a citizen of CIT and later Carnegie Mellon University (CMU) community, Newell was instrumental in encouraging and configuring -- with Simon and Alan J. Perlis -- the university’s first Computer Science degree program in 1961, and its Computer Science Department, established in 1965. Newell’s research and growing reputation were key in attracting and renewing National Institute of Mental Health grants for cognitive science research in the Psychology Department. His work also attracted Department of Defense grants (through its ARPA research arm) for core funding and project support to the university’s burgeoning Newell and Simon-supported cognitive science-artificial intelligence nexus. Newell worked on ARPA’s ZOG project -- a human-computer interface for the aircraft carrier USS Carl Vinson -- and on numerous micro-area projects. Spurred by his experiences with the ARPANET computer network -- the original Internet -- Newell served as president of CMU’s Task Force on the Future of Computing, a body which recommended and helped design the Andrew system, an IBM-CMU cooperation that was one of the country’s first campus-wide computer networks.

"Choose a final project to outlast you" was one of Newell’s maxims for scientists. For Newell that project was Soar (State, Operator and Result theory), an artificial intelligence system created with an eye toward application as a theory of human cognition. In what he described as "GPS done right," Newell synthesized 30 years’ research on the nature of AI and human cognition by himself and former students John Laird (University of Michigan) and Paul Rosenbloom (USC) into a program that integrated the critical symbolic processing stage of human decision-making into its operations. In his 1987 William James Lectures at Harvard University and the 1990 book, Unified Theories of Cognition, that emerged from them, Newell propounded Soar as an exemplar of the long-sought unified theory of human thought processes. His remarkably unbiased intelligence -- which welcomed from students, junior faculty, and esteemed colleagues alike opinions and research approaches opposed to his own – allowed him to take in stride initial misunderstandings that in Soar he had proposed the unified theory of cognition.

Newell authored or co-authored more than 250 publications, among them 10 books including Human Problem Solving with Herbert A. Simon (1972). A pure researcher unfazed by honors, Newell accepted for the advancement of the field the first presidency of the American Association for Artificial Intelligence in 1980 -- and later acknowledged the appointment as the honor he especially treasured. He was elected to the National Academy of Science and the National Academy of Engineering in 1972, and to the American Academy of Arts and Sciences in 1980. The University of Pennsylvania and Groningen University, Netherlands, awarded him honorary doctorates. Other honors included the Harry Goode Award of the American Federation of Information Processing Societies (1971); a Guggenheim fellowship (1967); the A.M. Turing Award of the Association for Computing Machinery (jointly with Herbert Simon, 1975); the Alexander C. Williams, Jr. Award of the Human Factors Society (co-recipient 1979); the Computer Pioneer Award of the IEEE Computer Society (charter recipient 1982); the Distinguished Research Contribution Award of the American Psychological Association (1985); the Research Excellence Award of the International Joint Conference on Artificial Intelligence (1989); the Emanuel R. Piore Award of the Institute for Electrical and Electronic Engineers (1990); and the Franklin Institute’s Louis E. Levy Medal (1992). On June 23, 1992 President Bush presented Newell the prestigious National Medal of Science at a White House ceremony.

Energetic in research and promotion, playful in his writing and theorizing, Allen Newell was an indefatigable promoter of artificial intelligence and cognitive science research, literally until days before his death of cancer in Pittsburgh on July 19, 1992. He was survived by his wife, Noël, and son, Paul Allen. At the time of his death Newell held the chair of U.A. and Helen Whitaker Professor of Computer Science at CMU. In October 1992 CMU’s School of Computer Science honored Newell through a symposium entitled Mind Matters, which focused on the state of AI and current computing technology. Professors from various universities attended in celebration of Newell’s interdisciplinary achievements.