Logic Theorist

1956

Herbert Simon (1916–2001), Allen Newell (1927–1992)

The development of computer programs that can perform complex, humanlike feats of problem solving has been the holy grail of cognitive scientists. In the 1950s, the partnership of scientist and researcher Allen Newell and political scientist Herbert Simon brought scientists closer to their goal. One of their breakthrough innovations was the computer program Logic Theorist, first presented at a conference in Dartmouth, New Hampshire, in 1956. This program sprang from Newell and Simon’s determination to describe thinking people as information processors. To do so, they developed what they called protocol analysis, in which they presented research subjects with logic problems and asked them to think out loud as they solved them. They then analyzed these protocols to discern the rules and patterns their subjects used, and these insights guided their computer programming. They described their general approach as one in which artificial intelligence would borrow ideas from psychology and psychology would borrow ideas from artificial intelligence.

Logic Theorist was designed to generate proofs for some of the basic theorems in Alfred North Whitehead and Bertrand Russell’s Principia Mathematica (1910, 1912, 1913), a landmark three-volume work on mathematical logic. When it actually generated a proof that was more elegant than the authors’ own, Newell and Simon attempted to have it published in a journal under the authorship of Logic Theorist (the editors did not allow it). Newell and Simon improved upon Logic Theorist with the development of General Problem Solver, which was based on means-end analysis, a heuristic strategy in which the current state and end goal of any problem are systematically compared. The option that reduces the distance between the two is chosen at every step until the distance between the current state and the goal is zero. At the time, Simon confidently proclaimed that a computer would be the world champion in chess within the next ten years. Although it would actually take forty years, in 1997 the program Deep Blue did beat reigning champion Garry Kasparov in a match that was televised around the world.

SEE ALSO Can Machines Think? (1843), Turing Machine (1937), Cybernetics (1943)

Mephisto Academy chess computer, 1989. In 1956, Simon confidently proclaimed that a computer would be the world champion in chess within the next ten years. In fact, it took forty years.