Dyadic interaction should be studied as a formal system, with the behaviors of each individual studied separately in such a way that the whole can be constructed from the parts. A mechanism for such interaction is interpersonal judgment, by which A selects a response to suit B's behavior. In two experiments, ss acting as teachers were given performance scores and asked to select goals for students. Their responses were a linear function of the performance score, independent of their assigned aim or the students' task. High norms and actual teaching experience lowered the goals chosen but did not affect the shape of the function. An illustration of the use of this data in a model of dyadic interaction leading to a selffulfilling prophecy is given.
w HESE sruDiEs are the first step in a pro-T gram of experimental research on formal models of the dyad as an interactional system. The goal is to understand some of the ways in which the behavior of A elicits responses from B which then affect A's subsequent behavior, which continues to influence B, etc. This paper will present (1) a method for the study of dyadic interaction as a system, (2) a mechanism of social process based on interpersonal judgments, and (3) data supporting this mechanism in one half of the interaction, that is, in -4's responses to B.
STUDY OF INTERACTION
The experimental problems raised by the dyad are familiar. If a pair is freely interacting, then each is setting conditions for the response of the other; the responses of each are stimuli for the other. So, whatever initial state is set by the experimenter as the independent variable, he will fairly quickly lose control of it, as each subject successively changes the stimulus situation for the other. The stimuli and responses which happen to emerge can then be correlated, but experimental control and the stronger inferences it permits have been lost.