Shannon (1950a) has argued that the problem of providing a programme for a chess-playing computer is of theoretical interest, and its use might lead to a wide range of practical developments. The problem is also interesting psychologically. If the human and the mechanical players are to play the same game, they will each have to be directed by concepts which have a certain equivalence. But the concepts used by the skilled human chess-player are both subtle and complex, and for the purpose of programming a computer they will have to be reduced to their simplest form. Chess-masters are, as a class, men of considerable general intellectual ability, and come from the ranks of professional men, mathematicians, scientists, lawyers, etc. They have in addition a special ability. Very few chess-masters, who began the game early, did not show unusual excellence at it at a very early age. The specific chess ability begins to show itself, given the opportunity, at about the age of eleven. Furthermore, there are few, if any, chess-masters who cannot play blindfold, and play many games at once, achievements which are entirely beyond the powers of the ordinary player. The order of intellectual activity which we are required to reduce to simple terms is therefore of a superior kind.
Read full abstract