FINITE AUTOMATA AND HUMAN BEINGS

Arthur Burks is best known among philosophers for his views on in­ duction, probability and causation. He has elaborated and refined these views over many years, and his major work, Chance, Cause, Reason (1977), develops them in great detail and with much subtlety. No less interesting, however, is Burks's application of automata theory to such philosophical problems as the formalizability of intelligent be­ haviour, freedom of the will and the dispute between empiricists and rationalists concerning abstract ideas. While automata theory plays a role in Burks's investigations of causality, chance and inductive reason­ ing, it becomes especially prominent in his discussion of these other is­ sues. Unsurprisingly, Burks, who is as much a computer scientist as a philosopher, is exceptionally circumspect and scrupulous in bringing computer science to the aid of philosophy. In what follows, I consider whether Burks's strategy of reformulating philosophical problems as problems of computer science can render them more precise, intelligible, tractable. In agreement with Burks, I argue that certain philosophical problems - including the for­ malizability of intelligent behaviour, freedom of the will and the source of abstract ideas - can benefit from an infusion of computer science. But I also argue that application of Burks's strategy results in a sig­ nificant transformation of these problems as they are traditionally un­ derstood. Treating philosophical problems as problems of computer science may in the course of clarifying them reveal that they are im­ properly formulated or that they rest on controversial assumptions. A philosophical problem to which automata theory has direct applica­ tion is whether intelligent behaviour can be reproduced by machines. From a computer-theoretic standpoint, this is a problem, as Burks points out in "Logic, Computers and Men" (1972-73), about the