ABSTRACT
Robert Merton's study of the machine shows the advantages of rejecting society as the unit of analysis. Merton demonstrates the implausibility of assuming that functions go with entire societies by showing how functional analyses of religion ignore "commonplace facts regarding the role of religion in contemporary literate societies." One of the noteworthy features of Merton's 1949 essay is that it anticipates and answers many of the criticisms that gained common currency in the 1950s and 1960s. Merton refutes, for instance, the charge that adopting a functional mode of analysis commits one to an ideologically conservative perspective. Arthur Stinchcombe's valiant efforts to salvage functionalism have been lauded but nevertheless judged wanting by Jon Elster, who maintains that none of Stinchcombe's mechanisms meet the necessary conditions for a valid functional explanation. Elster is particularly critical of Stinchcombe's efforts to show that the environment exerts a selective force on behavior.
