Purposeful and Non-Purposeful Behavior

In a recent essay (1) Professor Taylor criticizes the criteria used by Rosenblueth, Wiener and Bigelow in 1943 (2) to distinguish purposeful from nonpurposeful behavior. He also criticizes our definition of behavior, our concept of the vague as opposed to the general, our use of the word correlation, and our statement that a system may reach a final condition. Indeed, there seems to be little, if anything, in our paper to which he does not emphatically object. He maintains that the notions of purpose and teleology are not only useless for the understanding of mechanical behavior, but wholly incongruous when applied to this behavior. He further affirms that our use of the term purpose "bears no similarity whatever to the meaning which is ordinarily attached to it." He does not state, however, his own notions of purpose and teleology; and the meaning which he considers to be ordinarily attached to the term purpose. This omission weakens his criticism considerably. In the present paper we shall not deal with purely verbal issues, because we believe that they are trivial and barren. We shall discuss some of Professor Taylor's opinions and generalizations which we judge erroneous. We shall then complement our earlier distinction between purposeful and non-purposeful behavior, because, notwithstanding Professor Taylor's opinion to the contrary, we think that this distinction is meaningful and essential in science, and because Professor Taylor's misunderstanding of our earlier formulation suggests the possibility that others may also have deemed the problem to be largely a verbal one. In the interim between the publication of our first paper and the present time, we have devoted much effort to the clearing up of the set of categories necessary for the study of the statistical aspects of modern science. Some of this work has been presented by one of us (3). The present analysis is based on the ideas developed in that presentation. They go back to the work of Gibbs and of Heisenberg, which cannot be contained in the purely Newtonian frame of physics. Professor Taylor affirms that, in the study of the behavior which we have called purposeful, so long as distinctively human purposes are left out of account, there is no conceivable way of selecting some particular relationship between the behaving object and its surroundings as the goal toward which that object was directing itself. This statement is false. We can conceive many ways of selecting the goal or goals in purposeful behavior, and the only problem is to find out which of these several procedures is the most adequate for scientific analysis. The emphasis on human purposes is irrelevant. The purpose of the designer of a radar-controlled gun may have been to have the gun seek an enemy plane, but if the gun seeks the car of the commanding officer of the post, as this officer drives by, and destroys it, surely the purpose of the gun differs from that of the desig,ner. Indeed, this would be an excellent example of cross-purposes.