Despite the fact that the concepts of rationality and irrationality are central to the social sciences, social scientists have devoted remarkably little attention to them in their analyses. Instead, they have tended uncritically to assume models of human behavior that are either exclusively rationalistic or alternatively, strongly irrationalistic. Consider the case of political science. Influenced by scholars like Harold Lasswell, political science during the fifties took a strong interest in psychoanalytic models of political behavior. 1 More recently, the emphasis in political science has moved away from psychological models and towards rational choice analyses. 2 What is especially interesting about this is the completely either/or character of the two traditions of political science research -rational choice theorists tend to overlook the role of unconscious motives, while psychoanalytic approaches downplay the calculating aspects of that same behavior.
In this paper we propose a broader scheme for thinking about the rationality and irrationality of human action. This scheme, we shall argue in the first place, does better justice to the actual complexity of human action than either rational-choice or psychological models by themselves. And, in the second place, it explains why theorists (and laymen) so often differ in their assessments of the rationality or irrationality of the same action-they differ, we shall argue, both because they either ignore or else concentrate on the context in which the action has occurred.
We propose to describe human behavior in terms of a two-dimensional matrix -rational/irrational and conscious/unconscious -as in Figure 1. I. ACTION AND MERE BEHAVIOR Before we discuss the types of human action which this scheme yields, we must define action generally and distinguish action from what is mere behavior. An action is a goal-directed behavior. Thus a yawn when others Philosophical Studies 36 (1979) 333-343.