JULY/AUGUST 2001 computer.org/intelligent 71

honors in his lifetime, including the Nobel Prize in Eco-nomic Science in 1978.

“What many of us remember is his grace, his humanity,and his sense of right and wrong. Herb’s enduring legacywill be his contributions to science,” says Raj Reddy, HerbertA. Simon university professor of computer science androbotics at Carnegie Mellon University. “Every field hetouched, from political science to economics, to cognitivescience and computer science, he has made deep,insightful, and seminal contributions that will endurebeyond the physical institutions he has helped to build.”

Contributions and predictionsSimon was a true renaissance man. He not only

contributed greatly to AI, he was an economist, apolitical scientist, and a psychologist. He researchedhuman decision-making and problem-solvingprocesses and the implications of those processes forsocial institutions.

“By far, his most important contribution was thecreation of the field of artificial intelligence. His viewwas that if intelligence can be demonstrated by largecollections of carbon-based molecules, then there is noreason, in principal, why the same would not be the casefor any other collection of molecules capable of manipu-lating physical symbols,” Reddy says. “Artificial intelli-gence was to lead him into an unending controversy. Thevery idea of artificial intelligence naturally upset manywho believe in the unique capabilities of human beings,such as creativity, emotion, and consciousness.”

That controversy led many to believe that Simon was argu-mentative. (According to a 1994 interview in Omni, his firstpublished piece was a letter to the editor in the MilwaukeeJournal defending atheism, written while in grade school.)Reddy disagrees. “He had very strong viewpoints. Herb pas-sionately defended his point of view against all comers. Moreimportantly, he systematically set out to prove that everyactivity that could be considered intelligent, when done by ahuman being, can also be done by a computer—through the-orems it can play chess, solve puzzles, solve problems,understand language, discover scientific laws, and so on.”

In 1957, Simon predicted that four things would happenin AI within 10 years: Computers would compose musicof aesthetic interest, most psychological theories wouldtake the form of computer programs, a computer wouldprove a significant mathematical theorem, and a computerwould be chess champion of the world. All but his chessprediction came true within the specified time—and thatcame to fruition 40 years later.

Fast forward to 2000. During the inauguration of Newell-Simon Hall, which houses Carnegie Mellon’s Departmentof Computer Science, Simon didn’t offer more predictionsfor the future of AI, but he did offer this. “I don’t believethat predicting the future is really what we’re about,” Simonsaid. “Our task is not really to predict it; our task is to

design it. Design a sustainable and acceptable world andthen devote our efforts to bringing that future about. We’renot observers of the future, we’re actors—who, whether wewish it or not, by our actions and very existence are going todetermine our future’s shape for better or ill.”

Significant workHerb Simon’s major awards in four areas—psychology,

economics, management science, and computer science—attest to his breadth of interest and expertise. His initialcontributions to computer science were in collaborationwith Allen Newell and Cliff Shaw—developing the firstheuristic programs, the first list-processing languages, andestablishing the field of AI, which they called “complexinformation processing.”

In the 1960s, Simon’s main research effort was extend-ing AI’s boundaries—with regards to the simulation ofhuman thought processes—from the well-structured tasksaddressed in early AI programs to wider ranges of tasks

When Herbert A. Simon passed away earlier this

year, the world of AI lost one of its founding

fathers. Simon, who, along with Allen Newell and J.C.

Shaw, wrote the first AI program in 1956, received many

Herbert A. Simon: AI PioneerBy Scott L. Andresensandresen@computer.org

H i s t o r i e s & F u t u r e s

72 computer.org/intelligent IEEE INTELLIGENT SYSTEMS

that call on substantial bodies of knowl-edge and that are relatively loosely struc-tured. Among the important early programswere the General Problem Solver, devel-oped jointly with Newell, and the Elemen-tary Perceiver and Memorizer, developedwith Edward A. Feigenbaum, Lee W.Gregg, and Howard B. Richman.

The work on the General Problem Solverled to a sustained program of research, withgraduate students and colleagues, onhuman problem solving. This involvedlaboratory experimentation paired withcomputer modeling of human performancein various problem environments. One lineof work, with William G. Chase, exploredthe role of knowledge-based recognitionprocesses in expert performance (usingchess as the principal domain of study).Recognition processes, modeled by pro-duction systems, have provided an impor-tant function for the expert systems nowcommon in AI applications.

Simon was concerned with the problemsof knowledge representation and its rela-tion to language understanding. He builtthe Understand program, which reads taskdescriptions in natural language and con-structs suitable representations of the taskto be used as inputs to the General ProblemSolver. The representation research led toinquiries into computational procedures forreasoning from diagrams.

This research has closely linked the con-struction of intelligent systems with thedevelopment of a computational theory ofhuman intelligence. The two chess programsSimon constructed with colleagues—theNSS program with Newell and Shaw (1958)and the Mater program with George W. Bay-lor—contrast with more powerful programssuch as DeepThought and HighTech. Hisprograms model the highly selective charac-teristics of human search instead of relyingon massive computing power to compensatefor lack of chess knowledge.

Simon’s impactSimon’s impact in the world is no more

apparent than at Carnegie Mellon. “It’sdifficult to overestimate his contributions,”Reddy says. “More than any one person inthe last 50 years, he’s responsible for thestatus of Carnegie Mellon. Arguably, hewas the most influential person.”

Robert Mehrabian, former CarnegieMellon president, agrees. “No one inCarnegie Mellon’s history more heavily

influenced the university’s intellectual andscholarly approach or its academic cultureas much as Herb Simon,” Mehrabian says.To further understand his impact on theuniversity, go to www.cs.cmu.edu/simon/remembrances.html. On this Web siteyou’ll find many stories, remembrances,and insights on Simon.

“His profound understanding of humancognition also led to some unusual insights,”Reddy said. “One day Herb said, ‘You can bea world-class expert in anything you choose,provided you’re willing to spend 15 years ofyour life, all the waking time, on that enter-prise and on that subject. This is true whetheryou want to be an Olympic athlete, a chesschampion, or a Pulitzer-Prize-winningauthor.’ Even now it is not clear to me howthis could be true, but nevertheless, I havelearned to trust Herb’s intuition.”

Simon’s legacy Simon left behind over 60 years of

research and theorems. His legacy to theworld of AI world simulations of humanbehavior over a wide range of perceptual,learning, and concept induction tasks; sys-tems for discovering lawful patterns inempirical data, initially in sequences of let-

ters or numbers (the Thurstone letter seriescompletion tasks); inquiries into computa-tional procedures for reasoning from dia-grams; an approach to causal modeling thathas broad applicability to research in quali-tative physics and to the automatic inductionof causal structures within AI; problems ofparallel architecture and programming interms of knowledge of hierarchical organi-zation; and other foundational issues incomputational AI.

He left the computer science world withthe heuristic compiler, in which he appliedtheories and techniques from other disci-plines. This was the first system with simplecapabilities for automatic programming.

“Herb leaves us, his disciples, with anunfinished revolution and a number ofunsolved problems,” Reddy says. “Can acomputer read a book in physics oraccounting and answer the questions at theend of the chapter? Can a robot look overthe shoulder of a person and learn toassemble a motor? Can computers makenew discoveries that will lead to thePulitzer Prize? Can a computer exhibitemotion, tell a joke, or display conscious-ness? Herb believed that the answer to allthese questions was ‘yes.’”

Herbert A. Simon was born in 1916 in Milwaukee, Wisconsin, and died 9 February2001 in Pittsburgh, Pennsylvania, from complications following surgery to remove acancerous tumor..

Simon earned his BS (1936) and PhD (1943) in political science from the Universityof Chicago. He received the Award for Distinguished Scientific Contributions from theAmerican Psychological Association in 1969. In 1975, Simon earned the A.M. TuringAward for his work in computer science. In 1978, he received the Alfred NobelMemorial Prize in Economic Sciences from the Royal Swedish Academy of Sciences“for his pioneering research into the decision-making process within economic orga-nizations.” His bounded-rationality theory said the abilities of decision-makers,rather than the desire to increase profits, limited business decisions.

In 1986, Simon was awarded the National Medal of Science, and in 1993 he wonthe American Psychological Association Award for Outstanding Lifetime Contribu-tions to Psychology. In 1994, he was one of only 14 foreign scientists ever to beinducted into the Chinese Academy of Sciences.

In 1995, Simon received the Award for Research Excellence from the InternationalJoint Conferences on Artificial Intelligence and the Dwight Waldo Award from theAmerican Society of Public Administration.

He also was inducted into the Automation Hall of Fame in Chicago because of hispioneering work in AI. The Hall of Fame, at the Chicago Museum of Science and Tech-nology, recognizes individuals who have made major contributions to the practice andphilosophy of manufacturing technology through advanced methods and research.

Simon was also active in science policy at the national level, primarily through thecouncil and committees of the National Academy of Sciences and its associatedNational Research Council, and the US President’s Science Advisory Committee.

Simon was an emeritus life trustee of Carnegie Mellon University, where he hadbeen a member of the board since 1973. Until his recent illness, Simon was activelyteaching and doing research at Carnegie Mellon. —SLA

Achievements