IS SOCIOLOGY A SCIENCE?

With the analysis of science in the previous section in mind, let us turn to sociology. Early sociologists tried to establish sociology as a science, and their arguments are mainly on the methodology of sociology. Comte claimed that sociology uses four different kinds of methodologies, namely observation, experiment, comparison and historical research as a special case of comparison. These are the methodology used in several other scientific fields, especially in biology. So if his sociology had really followed these methods, it would have been a strong case for sociology as a science. But actually, he never did empirical research so we cannot take his argument at the face value. But his argument influenced on other sociologists, especially Durkheim. For Durkheim, sociology is a study of social facts. A social fact is "a thing that is external to, and coercive of, the actor". Because they are external, social facts cannot be investigated by introspection. We should use empirical research. A typical use of this methodology is in his analysis of suicide. Durkheim used statistics on suicide rate to establish his argument that suicide is a social phenomenon. He refused alternative hypotheses because their predictions did not agree with the actual statistical data. This is an admirable attempt of empirical research of society, but there are several problems. Durkheim applied too strict criteria of falsification to rival accounts. Adoption of these strict criteria is suicidal for sociology because it is hard for a sociological theory to make a precise prediction, let alone to make a precise and correct prediction (and without this, the falsification criteria do not work). Another related problem is in his rejection of introspection as a sociological method. This restricts the scope of sociology too narrowly, and in fact, even Durkheim's own study becomes impossible. For example, Durkheim's definition of suicide is "any case of death 'resulting directly or indirectly from a positive or negative act of an individual against himself, which he knows must produce this result". But, without using introspection, how can we decide if 'he knows' the result or not, from external evidence only?

I think that Weber's methodology provides an answer to these problems. His keyword at this point is "Verstehen," a German word for "understanding" or "interpretation". According to him, we can "understand" other people's motivation through introspection of our own intentions, and this kind of knowledge is necessary for sociology. This is exactly what Durkheim denied as a method of sociology, but as we saw above even Durkheim himself used this "understanding" in his actual work. But, of course, the problem is if this is permissible as a scientific method. Strong falsification of a theory is almost impossible by such "interpreted" facts because if an interpreted fact runs counter to the theory we can just change the interpretation. But, as we saw in the last section, such strong falsification is given up by philosophers of science as too strict a criteria. Moreover, the arbitrariness of interpretation is not as great as one might worry. For example, Comte's three-stage theory (the detail of the theory does not matter here) has no follower today because there is no way we can reasonably interpret the evolution of society as obeying such a law. In this case, we can say that Comte's theory was falsified. As far as we have this minimal possibility of falsification, we can admit "Verstehen" as a scientific method of sociology, thus "interpretive" sociology as a science.

Before we proceed to next section, I would like to make a brief remark on the use of models in sociology. One of the reasons people may argue against sociology as a science is the lack of the sociological theory. We have Marx's theory, Durkheim's theory, Weber's theory and so on, but none of them are shared by all sociologists. This seems to make a strong contrast with other fields of science where scientists agree on the basic theories. But, as we saw in the last section, some philosophers think that even in other scientific fields what scientists are working on are piecemeal models, not a universal theory. And as for such models, we can find abundant models shared by many sociologists. Actually, this is what Weber called "ideal types". Ideal types are constructed by exaggerating some features of real cases. By comparing with ideal types we can find characteristics of each real case. These ideal types are useful conceptual tools for sociology just in the same sense as the planetary model of atoms is a useful conceptual tool for chemists. So, at this point, the difference between sociology and other scientific fields is not so great as it seems to be.