Abstract
We argue that recent work in connectionist modeling, in particular the parallel constraint satisfaction processes that are central to many of these models, has great importance for understanding issues of both historical and current concern for social psychologists. We first provide a brief description of connectionist modeling, with particular emphasis on parallel constraint satisfaction processes. Second, we examine the tremendous similarities between parallel constraint satisfaction processes and the Gestalt principles that were the foundation for much of modern social psychology. We propose that parallel constraint satisfaction processes provide a computational implementation of the principles of Gestalt psychology that were central to the work of such seminal social psychologists as Asch, Festinger, Heider and Lewin. Third, we then describe how parallel constraint satisfaction processes have been applied to three areas that were key to the beginnings of modern social psychology, and remain central today: impression formation and causal reasoning, cognitive consistency (balance and cognitive dissonance), and goal-directed behavior. We conclude by discussing implications of parallel constraint satisfaction principles for a number of broader issues in social psychology, such as the dynamics of social thought and the integration of social information within the narrow time frame of social interaction.