Liberty Matters

CLCA as a Positive Program in Social Science

George Smith in “Of What Use is Class Analysis?” and David Hart in “What Is a 'Class' and Can Class Analysis Be 'Wertfrei'?” both raise questions about the way in which scholars’ values enter into class analysis. Smith asks if CLCA requires us to assume that we understand too much about the interests of the individual members of the groups we term to be a class. Hart conjectures that 20th-century social scientists turned away from class analysis out of fear of appearing unscientific if they addressed questions with obvious normative implications.
These concerns strike me as related to F.A. Hayek’s discussion of social justice, in particular the claim that “only human conduct can be called just or unjust” and therefore “To apply the term ‘just’ to circumstances other than human actions or the rules governing them is a category mistake."[81] If we accept the idea that class matters, are we implicitly subjecting social structures to charges of injustice (even if the mode of our analysis is value-free)? In doing so do we violate the analytic principle that the actions of individuals, taken in service of subjectively defined values and with subjectively defined expectations, are the appropriate unit of analysis in social science?
I think it is useful here to distinguish between having a causal influence and being causally determinant. Understanding the class structures in effect in a society and the rules and restrictions they create, whether explicitly or implicitly, may allow the observer to identify patterns of behavior related to an individuals’ position in the class structure. For example, if a person lives in a society where they are restricted from highly skilled occupations because of their caste or gender, we can predict that the restriction will discourage them from investing in education. Similarly, if a person is in a group that grants them disproportionate access to the legislative process, we can predict that they will be more likely to invest in shaping laws in ways that suit their interests than they would be if influencing law was costlier for them.
Knowledge of these effects does not mean that studying class has enabled us to predict that a particular person will or will not lobby, or pursue higher education, or undertake any other specific action. This also does not mean that these people’s actions—whatever they are—have been dictated to them by their position in the class structure. As long as an individual’s valuation of the many ends they could choose to pursue and the many means they could choose for doing so is subjective, then purposive action can never be fully determined by external social factors.[82] The classed nature of the society has a knowable causal influence on behavior without defining the individuals’ interests or behavior, either to them or to the analyst.
The fact that class cannot and does not determine the interests or behavior of the individuals in a group is not at all a limit on the usefulness of the concept. This lack of causal determinacy and inability to generate specific predictions are true of all attempts to understand observable behavior patterns. Consider the prediction that an increase in the price of gas will lead to a decrease in gallons purchased. This simple prediction is always true in isolation, but only ceteris paribus. Other factors simultaneously impacting an individual’s purchasing decision can easily cancel out the observed effect of the price change. Any given person may or not not actually buy less gas than they did the day or year before, because their decision is going to be influenced by so many other factors in addition to that one single price change. The change in price has a predictable causal effect on the individual’s behavior without determining what they will choose. Few would suggest that this means the law of demand is not useful in understanding social behavior.
Returning to Smith’s concern that CLCA might require us to assume that we understand too much about individuals’ interests, I would argue that the distinction between causal influence and causal determinacy outlined above is significant to understanding how class analysis can be consistent with methodological individualism and subjectivism. Although it is possible to conduct class analysis in a way that presumes a set of interests—and indeed this has been done too often—CLCA need not understand classes as having well-specified or homogeneous interests in order to help the scholar understand a significant set of social structures and influences.
As a final note, if classical-liberal scholars have avoided CLCA out of fear of appearing inappropriately normative—a conjecture that is consistent with broader trends in 20th-century social science—then they have done so in error. All meaningful social science will have normative implications. This does not mean that the wertfrei ideal need be abandoned in the process of analysis, or that social structures need to be misidentified as engaging in purposive action. Instead, CLCA can generate useful a priori predictions about meaningful influences in the social world without presuming too much homogeneity of interests or knowledge on the part of the expert.
[81.] F. A. Hayek, Law, Legislation, and Liberty,vol. 2: The Mirage of Social Justice (Chicago: University of Chicago Press 1976), p. 31.
[82.] Paul Lewis, “Solving the ‘Lachmann Problem’: Orientation, Individualism, and the Causal Explanation of Socioeconomic Order,” American Journal of Economics and Sociology, 67, no. 5 (2008): 827-58.