The Questionable Logic of “Mistakes” in the Dynamics of Knowledge Growth in the Social Sciences

Neil J. Smelser The Questionable Logic of “Mistakes” in the Dynamics of Knowledge Growth in the Social Sciences ON RECEIVING AN INVITATION TO CONTRIBUTE TO THIS SPECIAL ISSUE on errors in science, I immediately responded to the editors that I did not wish to do so because it was a big mistake to think that the logic of “big mistakes” applies to the histoiy and dynamics of the behavioral and social sciences.* I gave them a few reasons why I thought that my reasoning was correct. The editors promptly came back to me and said what a great contribution it would be if I developed a case for my point ofview. The editors turned out to be cleverer than I, for in reflecting on their revised request, I came to believe that something useful could be said. And, as is frequently true, I ended up developing a more qualified and contingent view than I originally and somewhat flippantly shot back at the editors when they first approached me. THE EMULATION OF THE NATURAL SCIENCES IN THE BEHAVIORAL AND SOCIAL SCIENCES When the definitive histoiy of the social sciences is written, it will have to be acknowledged that perhaps the dominant and most persistent m otif is the effort to model those sciences after the physical and life sciences—or perhaps better, what the aspirants and practitioners ofthe social sciences believe the hard sciences to be. That effort is not the social research Vol 72 : No 1: Spring 2005 237 only or uniform story, because it has been challenged in so many ways, and so many alternative versions of the “human sciences” have been proposed over time. This emulative effort has not been a simple one, but has occurred on many different levels, among which the following can be identified: ►From an epistemological point ofview the main assumption has been that psychological and social phenomena should be studied as “natural” phenomena. The early political economists had the Newtonian model of physics before them (Halévy, 1949). Karl Marx announced that he was studying the “natural laws of capitalist production” like the physicist who “observes physical phenomena where they occur in their most typical form and most free from disturbing influence” (1949 [1868]: xvi-xvii). The most eloquent and thoroughgoing defense of sociology as a natural science was given by Émile Durkheim in his The Rules ofthe Sociological Method (1938 [1893]). There he staunchly laid out the subject m atter of sociology as objective “social facts,” which are found in nature and which resist distortion and are correspondingly “to be treated as things” (xliii), or data outside the mind of the investigator. This positivistic view pervaded early American sociology as well, and is the implicit informing logic of most empirical researchers in the social sciences to this day, despite the onslaughts of phenomenological and postmodernist critiques. Indeed, the persistence of the name “social sciences” testifies to the tenacity of this view. ►Many substantive models of the person and society have also been imported from the natural sciences as well—various mechanical causal models from physics, as well as models based on biological evolution, biological organism, and ecology. Both social evolution­ ary theory and classical functional sociology, for example, appealed to explicit anatomical, physiological, and evolutionaiy analogues (see Radcliffe-Brown, 1952). ►The early preoccupation with economic laws, psychological laws, and laws of society were all modeled on some perceived version of 238 social research natural laws that were thought to be characteristic of the physical and biological sciences. ►The logic of discovery, displacement of prior limited knowledge, and the systematic accumulation of knowledge also have been envisioned by practitioners and proponents of the social sciences, though this stress has receded over time. Later in the essay I will attempt to give an account ofwhat “accumulation” has meant in practice in the social sciences. ►The social sciences have consistently emulated the scientific methodology of the physical and life sciences. The most notable instance is the widespread and continuing use of the laboratory experiment in academic psychology (with more limited use in small-group sociology, economics, and political science). The aim of the experiment, we should remind ourselves, is to isolate and identify specific causes...