Debates over the concept of class have been central in the study of social inequality. Indeed, the debate over how to define and study class has been central to the development of sociological theory. These debates have revolved around several polarities: Are classes merely categories of people in similar economic, educational, or occupational circumstances? Or, is class defined by boundaries of membership and shared interactions? Are classes best viewed as 'strata' -- cake-like layers of people aggregated and ranked by income, educational, or occupational status? Or, is the metaphor of strata completely inimical to the very nature of class as a concept designed to capture social inequality? If classes are distinctively unequal, are they so by virtue of an unequal distribution of material rewards? Or, are they unequal by virtue of exploitation and subordination?
A dominant approach to the study of class has conceived of it as part of a structural hierarchy of layers or strata. Here, individuals are ranked along a continuum divided into discrete categories. The criteria for operationalizing class has varied, but the general approach ranks individuals according to some indicator of economic rank, such as income, education, or occupational status. The most widely used approach is to calculate a single score derived from occupational ranking, average income level, and educational attainment. The result is an overall scale of socioeconomic status or SES (Blau and Duncan, 1967; Gilbert and Kahl 1987).
The operationalization of class along these lines has been generally associated with structural functionalism. According to structural functionalism, people are stratified along a continuum on the basis of status roles which they occupy in the workforce. Moreover, these rankings are determined by one's contribution to society and the key evaluative factors for defining, this role are "functional importance" and "differential scarcity" of talent and education (Davis 1945: 243-244; Davis and Moore 1949: 36; Parsons 1953: 403, 410).
Critics of this approach argue that the structural functionalist approach is inadequate for the purposes of understanding a class as a relationally defined group of people (Dahrendorf 1958; Stolzman and Gamberg 1974; Wrong 1963). Marxist inspired criticism has been most acute. On this view, classes are first defined by their relation to the means of production; that is, whether or not they own the means of production. Classes are, secondly, understood as existing or potentially existing groups. As such they are defined by their shared interactions and affiliations which make possible a sense of themselves as members of a class and who, accordingly, have shared interests. Marxist sociologists complain that the structural functionalist approach can only tell researchers about statistical aggregates of individuals defined precisely because they are independent of one another according to the protocols of survey sampling for example.
Perhaps the most sophisticated recent attempt to specify a Marxist approach to class is the work of Erik Olin Wright (197l; 1978; 1979; 1985). Wright brings the analytical theory construction, mathematical modeling, and statistical analysis to bear on the project of delineating or mapping out the contours of the contemporary class structure. As with earlier Marxist critiques of structural functionalist theories of stratification, Wright insists that class is a relational concept. Classes are always defined in relation to other classes, just as the concept of parent can only be defined in relation to the concept of child (1985: 34). The Marxist concept of class is antithetical to 'gradational' concepts of class which differ in the degree of some attribute such as status, income, or education. Thus, Wright argues that a relational concept of class must be defined primarily in terms of the processes of exploitation and subordination.
Where Wright has broken from some variants of Marxist sociology is in his insistence on utilizing the tools of mainstream social science: analytic theory construction, survey data and statistical analyses. He believes that a Marxist sociology can draw on survey sampling techniques and formal theory construction in an effort to revise Marxist theory an make it more adequate to the task of understanding and predicting class relations and the development of capitalism.
Hence, Wright has used large scale surveys in order to obtain indicators that will help build a theory of class structure. The concept of class was operationalized through questions designed to identify respondents' locations in the class structure. These questions distinguished between classes on the basis of three determinants of an individual's relationship to the means of production: ownership of economic surplus, control over the physical apparatus of production, and control over workers (1978: 73; 1979: 24). Thus, Wright operationlizes class in terms of relations to the ownership, control, and command of the means of production. Those who effectively possess all three attributes are defined as the modern bourgeoisie or capitalist class. Those who possess none of these attributes are defined as the proletariat or working class. Wright admits a third class, the petty bourgeoisie, who generate and control surplus, operate and manage their own business, but do not employ workers. The class structure of the contemporary United States, then, is composed of three main classes defined in terms of their relation to the means of production.
However, the impetus behind much of Wright's work has been to deal with the 'problem' or 'embarrassment' of the middle classes: those who do not fit neatly into the class categories of bourgeoisie, proletariat, or petty bourgeoisie. Wright argues that individuals who possess some but not all attributes which signify exploitative class relations do not, strictly speaking, form classes. Instead, they occupy 'contradictory' locations arrayed between the three main classes: executive managers, supervisors and foremen, small to medium capitalists, and semi-autonomous wage earners such as professionals. These contradictory class locations are differentiated on the basis of skill levels and Wright argues that these are best operationalized through questions that indicate educational attainment (Wright 1987: 24-29).
There are two problems entailed in Wright's attempt to operationalize class. First, despite his protestations otherwise and his attempt to build a relational model of class structure, Wright's analysis is driven by methodological individualism and a distributional analysis of class, both of which are the defining features of the structural functionalist approach to class. This is because Wright operationlizes class by using statistical survey and variable modeling and this approach cannot account for the Marxist conception of class as a relational process. Nor can it account for the concept of social change, as Stolzman and Gamberg (1974) argue. Wright operationalizes class as an independent variable that is derived from an aggregation of data about individuals. But, on the Marxist theory of class, this operation cannot account for social phenomena, social relations, and social change, history (Stolzman and Gamberg 1974: 121-122).
Wright's operationalization of class also conforms to a distributional model of inequality. For example, his attempt to map the dimensions of exploitation along the lines of skilled and unskilled labor reveals that, despite his insistence on a relational conception of class, he is willing to discard this principle: There is no necessary relationship between those who are skilled and those who are not. That is, one can be skilled but this does not mean, at least theoretically, that one's possession of a skill or skills translates into a relationship of exploitation with those who are unskilled (i.e., skilled blue collar workers do not exploit 'unskilled' domestic workers; professors do not exploit dental hygienists.) Thus, Wright's operationlization of this aspect of class depends on a distribution (and not a relational) model of skilled and unskilled labor. Wright recognizes the limitations in this aspect of his definition, conceding that different skill levels may well be a for of differentiation within classes and not between them (1985: 85, 95, 185).
As Burawoy (1989) and others influenced by labor process theories (Clegg 1994; Thompson 1989) have pointed out, a Marxist approach to the study of inequality and class structure seeks to understand how groups of workers are exploited and not merely how individuals are exploited. But focusing on distributional aspects of inequality, Wirght's theory cannot account for historical change (Carchedi 1987: 124-131). An historical analysis of class structure must conceptualize class in terms of production relations. A the center of Marx's theory is a concept of exploitation that is based on the alienation of workers from work and thus from their control over the process, product, and conditions of their labor. Marx's theory is not a theory of inequality 'per se.' Rather, inequality of income, education, or skill is symptomatic of the underlying structural relations and processes of historical change in the development of capitalist societies (Stolzman and Gamberg 1974: 108). Wright and other stratification models of class examine the symptoms of the disease, but not the disease.