The historical emergence of a field devoted to the experimental investigation of effects identified as "social" required a radical break with traditional conceptions of the social. Psychological experimentation was limited to the investigation of effects that were proximal, local, short-term, and decomposable. A viable accommodation to these constraints occurred in the closely related programs of Moede's experimental crowd psychology and Floyd Allport's experimental social psychology. Later, Kurt Lewin attempted to provide a different conceptual foundation for the field by drawing on certain precepts of Gestalt psychology and the philosophy of scientific experimentation developed by Ernst Cassirer. These ideas were poorly understood and were soon replaced by a methodological regime in which a new generation of statistical procedures and experimental design shaped implicit conceptions of the social in social psychological experiments through such procedures as randomization and the additive combination of variables. ᭧ 2000 John Wiley & Sons, Inc.1. This use of "programmatic," like the closely related "research tradition," derives from the practice, frequently adopted in science studies, of taking, not specific experiments, but interlinked sets of experiments as the unit of historical analysis. The rationale for this usage was provided by Imre Lakatos' (1978) critique of falsificationism, which pointed out that the historical fate of a scientific theory depended, not on unique experimental outcomes, but on the general trend of "research programs." The existence of such programs is not a matter of rhetoric but of practice. Rhetoric may or may not play a role in establishing practice in specific cases. Thus the "explicit announcement of a new program" is not a criterion for establishing the existence of a research program, as Samelson's (2000) comments imply.