A Paradigm for Causal Analysis

LET ME begin with three assertions. First, causal analysis is an important and viable method for theory construction in sociology. Second, the Merton variety of functional analysis in sociology is a unique method of theory construction and is the major alternative to causal analysis. Third, inherent methodological difficulties in Mertonian functional ananalysis in sociology argument for its abandonment. The first of the above assertions constitutes my reason for this attempt to spell out a notion of causal analysis as a method for theory construction in sociology. The appreciation of the second and third assertions requires some discussion of the work of Robert K. Merton and Kingsley Davis. Merton's classic methodological paradigm for functional analysis was put forth in Social Theory and Social Structure in 1949. A number of assertions prior to the statement of the paradigm imply that Merton saw functional analysis as a unique method of theory construction. He claimed that much had been accomplished with its use and much more could be expected. The paradigm was offered as an explicit codification of the procedures used by several sociologists and anthropologists. In this light, then, it is interesting to find that ten years later the president of the American Sociological Association was claiming that the uniqueness of functional analysis in sociology was a myth. In the paper titled "The Myth of Functional Analysis as a Special Method in Sociology and Anthropology," Kingsley Davis (1959) argued that "the definitions most commonly agreed upon make functionalism synonymous with sociological analysis and make non-functionalism synonymous with either reductionist theories or pure description." Davis went on to claim that the ambiguities of the special terminology of functionalism "make the myth that it is a special method a liability," and that the myth should be abandoned. Davis' argument is essentially that of parsimony. If it can be shown that there are no differences between functional analysis and any other kind of analysis in sociology then it is clearly parsimonious to abandon