Methodological individualism and vertical integration in the social sciences

ABSTRACT: This paper argues against the false dichotomy between reductionism and holism in the social sciences. I make the points that reductionism is the mark of a mature science, that the social sciences will never progress until they drop their opposition to reductionism, that higher-level explanations, even when more appropriate and coherent than reductionist explanations, must not violate principles established at lower levels of explanation, and that reductionist explanations almost always absorb the explanatory efficiency of broad social categorizations and add incremental validity to them. I demonstrate the validity of these points by exploring them in the context of the four most frequently used variables in social science (gender, race, age, and social class). In each case it is demonstrated that such categories fail to capture the causes of the phenomena social scientists explore, and that by failing to consider more elemental explanations lead to reliance on hypothetical "social facts" which are at best incomplete, and are often demonstrably wrong. The history of science reveals that all disciplines at one time or another have resisted the incursions of the more fundamental sciences, then showed a grudging acceptance of them, and finally became fully integrated with them. The social sciences must do the same with regard to integration with the relevant biological sciences.

Professor Jones (1996) recently described in the pages of this journal what I consider to be a false dichotomy between methodological individualism (reductionism) and methodological collectivism (holism). He correctly points out that social scientists have steadfastly opposed the reduction of Durkheimian "social facts" to psychological or (heaven forbid!) biological principles. Jones wisely does not claim that social events cannot be reduced to propositions about individuals participating in them, but he does declare that nonindividualistic accounts are preferable and that "we have no good grounds for preferring individualistic theories." He also claimsquite wrongly-that methodological individualists "claim that only accounts using individualist accounts are legitimate" (1996, p. 125 [emphasis in original]). This dichotomy is unnecessarily divisive. I will attempt to demonstrate four points in response to Professor Jones:

1 ) That reductionist explanations are the mark of a mature science.

2) That social science will never develop into true science as long as it maintains its opposition to reductionism.

3) That even when holistic explanations are more appropriate and more coherent than reductionist explanations they must not violate principles established at lower levels of explanation, and that such violations frequently occur in social science.

4) That reductionist explanations generally absorb the explanatory efficiency of broad social categorizations and add incremental validity to them.

Greedy versus Good Reductionism and the Role of Methodological Collectivism

Professor Jones conflates the distinction that several philosophers have called "good reductionism" and "greedy reductionism" in his examples of English table manners and why square pegs don't fit into round holes. Daniel Dennett defines good reductionism as "the commitment to non-question-begging science," and greedy reductionism as that which skips over several layers of (higher level) complexity in the rush to fasten everything secure and neatly to a solid foundation (1995, p. 82). Using his now famous metaphors, Dennett describes antireductionists as folks who "yearn for skyhooks" (a sort of deus ex machina that miraculously lifts us out of scientific difficulty) and who "call those who settle for cranes [solidly grounded devices that also function to lift us out of scientific difficulty] `reductionists"' (1995, p. 80).

A greedy reductionist is a person who sees a chain of causal events running up the ladder from molecules to human behavior, and who may make claims such as: "'Our behavior is controlled by molecules-by nothing else" (Applewhite, 1981, p.1). The claim is probably based on the fact that we are composed of molecules and that there is no we" apart from the molecules that compose us; thus our behavior must be the result of some molecules moving around other molecules. Viewing behavior as the movement of molecules conceals higher-order explanations that have greater coherence, generality, and utility, and is a kind of reductionism rarely seriously advanced by anyone today. Few would argue with the point that many phenomena cannot be sensibly explained by reference to greedy reductionist propositions. For example, the meaning of a poem cannot be deduced from an examination of the individual words and letters from which it is composed. While there can be no poem without letters or the sounds they denote, the meaning of the poem lies in their distinctive relationship to one another. On another level, I know that when the water molecules gyrating in my kettle reach 100 degrees Celsius, I shall have my cup of tea, but I would lose coherence were I to try to translate this information about temperature (an emergent property of aggregate molecules) into the motion of individual molecules. The conflation of greedy reductionism with the kind of "cranebuilt" reductionism that has elevated the physical and life sciences haunts sociology, and if not exorcised could lead to its demise as intellectually respectable discipline (Crippen, 1994; Ellis,1996; Neilsen, 1994; van den Berghe, 1990; Walsh, 1995).