ABSTRACT
Four key problems which have impeded research on nonverbal communication are examined: (a) sampling and labeling nonverbal behavior, (b) preserving the context in which the behavior occurs, (c) generating a source of nonverbal behavior, and (d) determining the criterion of accurate interpretation. These problems are particularly acute for research on the recognition of emotion.
Seven distinct varieties of context are identified, and the degree to which research has been faithful to these contexts is discussed. The criterion problem is described, and it is argued that this problem is insoluble within the emotion recognition tradition.
The paper presents a 9-cell typology of research designs, and the relative merits and frequencies of these nine design types are described. Finally, a new concept in nonverbal communication research is presented. This concept, embodied in the Social Interpretations Task (SIT), provides a solution to some of the problems of context and criterion which have beset the field of nonverbal communication.
