On an Example of Unpredictability in Human Behavior

Scriven’s example of essential unpredictability in human behavior goes like this. Assume a rational predictor P who wants to infer from information available to him the choice behavior of an individual C, where (i) C is choosing rationally and intelligently, (ii) C is a contrapredictive relative to P, i.e., C wishes to falsify any prediction made by P about his, C’s choice behavior, (iii) C knows the information available to P. It is easy to make these conditions more precise and then to prove that P cannot predict C’s behavior. Scriven fills out the bare bones of this example with many illustrations and comments, which I shall not attempt to sketch. His example of essential unpredictability is one I accept as in the main correct. The critical tenor of my remarks is to place the example in perspective from the standpoint of the theory of games and to challenge the significance of the example for the development of a quantitative and predictive behavioral science.