Beyond reductionism,with invidious comparisons between behavioral and paretian sociology

An acceptable reductive analysis does no more nor less than find the constitutive elements of a phenomenon. It should focus on observables, including indirectly given data but not including merely convenient assumptions. It should state propositions which refer to variables which are generalizable. Using these criteria, behaviorism is found to be wanting. In contrast, Pareto's model of five significant residues provides for a constitutive analysis which uses observables and generalizable variables. Revised version of a paper presented at the annual meeting of the American Sociological Association in New York, August 1976.