Herbert A. Simon

Overview

Herbert Alexander Simon was born on June 15, 1916, in Milwaukee, Wisconsin, and died on February 9, 2001, in Pittsburgh, Pennsylvania.^[1] His name is attached both to the first successful artificial-intelligence program and to a Nobel Prize in Economics.^[1] Both strands of that work circle one question: how a decision gets made when the information and the time available are limited.

Education

He received a BS in Political Science from the University of Chicago in 1937 and a PhD in Political Science from the University of Chicago in 1943.^[1] Both degrees were in political science, and the subject that occupied him afterwards was decision-making itself: not how a decision ought to be made by an ideal reasoner, but how one can be made by a person who has neither complete information nor unlimited time.

Bounded rationality

Simon argued that people intend to be rational but rarely have access to all the information or all the time they would need to make the optimally rational choice.^[1] An optimal choice, in this sense, is the best of all the options there are, found by weighing everything that could possibly matter. That standard is easy to state and very hard to meet: listing every option, gathering every relevant fact and comparing them all takes time and information that a person in a real situation does not have.

So, Simon argued, people must use simpler heuristics and accept satisfactory rather than optimal results.^[1] A heuristic is a rule of thumb — a shortcut that usually gives a good enough answer without an exhaustive search. Someone looking for a flat, for instance, visits a handful of places, compares rent and distance, and takes the first one that clears their own bar, rather than surveying every flat in the city. On Simon's account this is not reasoning that has gone wrong; it is what reasoning normally looks like.

Logic Theorist and the General Problem Solver

With Allen Newell and J. C. Shaw, Simon created the Logic Theorist of 1955–56, the first successful AI program, which was used to prove theorems of Russell and Whitehead's Principia Mathematica.^[1] A proof in this setting is a chain of statements in which every step either is an axiom or follows from earlier steps by a rule of inference, so proving a theorem means finding such a chain. The program searched for those chains, which is what makes it a reasoning machine rather than a calculator.

Simon famously announced to his undergraduate class that "Over the Christmas holiday, Allen Newell and I invented a machine that thinks."^[1]

Simon, Newell and Shaw then elaborated the Logic Theorist's principles in the General Problem Solver (GPS) of 1957–8.^[1] The name states the ambition of the project: to put the methods of problem solving into a form that would not be tied to one particular subject matter.

Career

He was Professor at the Graduate School of Industrial Administration of Carnegie-Mellon University from 1949, and Professor of Computer Science and Psychology there from 1965 to 2001.^[1] The later title joined two departments, and the pairing describes the research fairly well: computer programs on one side, theory and experiment about human thinking on the other.

Recognition and influence

He was elected to the National Academy of Sciences in 1972 and received the ACM Turing Award in 1975 jointly with Allen Newell.^[1] He received the Nobel Prize in Economics in 1978, the National Medal of Science in 1986 and the IJCAI Award for Research Excellence in 1995.^[1]

Set side by side, the two lines of Simon's work ask the same question from opposite ends: how a bounded agent finds a workable next step in a world too large to be searched exhaustively. The Logic Theorist handed that question to a machine and made it search; bounded rationality handed it back to people and described what they do instead of optimizing. Keeping programs and people inside one vocabulary is the common mark he left on artificial intelligence, psychology and economics.