All Marxists agree that manual workers directly engaged in the production of physical commodities for private capital fall into the working class. While there may be disagreement about the political and ideological significance of such workers in advanced capitalism, everyone acknowledges that they are in fact workers. There is no such agreement about any other category of wage-earners. Some Marxists have argued that only productive manual workers should be considered part of the proletariat. footnote1 Others have argued that the working class includes low-level, routinized white-collar employees as well. footnote2 Still others have argued that virtually all wage-labourers should be considered part of the working class. footnote3 If this disagreement were just a question of esoteric academic debates over how best to pigeon-hole different social positions, then it would matter little how these issues were eventually resolved. But classes are not merely analytical abstractions in Marxist theory; they are real social forces and they have real consequences. It matters a great deal for our understanding of class struggle and social change exactly how classes are conceptualized and which categories of social positions are placed in which classes. Above all, it matters for developing a viable socialist politics how narrow or broad the working class is seen to be and how its relationship to other classes is understood.

This essay will explore the problem of understanding class boundaries in advanced capitalist society. Rather than review the wide range of approaches Marxists have adopted in defining classes, I will focus primarily on the work of Nicos Poulantzas, in particular on his book Classes in Contemporary Capitalism. This work is, to my knowledge, the most systematic and thorough attempt to understand precisely the Marxist criteria for classes in capitalist society. While there are many points in Poulantzas’s argument with which I disagree, his work has the considerable merit of sharply posing the problem of defining classes in advanced capitalism and of providing some stimulating solutions. A critical discussion of Poulantzas’s work can, therefore, provide a very useful starting-point for the development of an explicit theory of classes in contemporary capitalism.

The first section below presents an outline exposition of Poulantzas’s theory of the structural determination of class. Poulantzas’s basic conclusion is that only manual, non-supervisory workers who produce surplus-value directly (productive labour) should be included in the proletariat. Other categories of wage-labourers (unproductive employees, mental labour, supervisory labour) must be placed in a separate class—either the ‘new’ petty bourgeoisie, or in the case of managers, the bourgeoisie itself. This exposition of Poulantzas will be followed in the second section by a general assessment and critique of his argument. The final section presents the preliminary outlines of an alternative conceptualization of class boundaries, that hinges on the concept of contradictory locations within class relations. I will argue that not all positions in the social structure can be seen as firmly rooted in a single class; some positions occupy objectively contradictory locations between classes. The analytical task is to give such positions a precise theoretical meaning and to relate them systematically to questions of class struggle.

The following presentation of Poulantzas’s ideas will necessarily be schematic and incomplete. I will discuss only the essential elements of his views on class boundaries and not deal with a variety of other important issues which he raises (such as class fractions, the relationship of classes to state apparatuses, etc.). While the exposition will lose many of the nuances of Poulantzas’s analysis, I hope that the basic contours of his argument will stand out. Critical comments will be kept to a minimum in this section.

Poulantzas’s analysis of social classes rests on three basic premises. 1. Classes cannot be defined outside of class struggle. This is a fundamental point. Classes are not ‘things’, nor are they pigeon-holes in a static social structure. ‘Classes’, Poulantzas writes, ‘involve in one and the same process both class contradictions and class struggle; social classes do not firstly exist as such and only then enter into class struggle. Social classes coincide with class practices, i.e. the class struggle, and are only defined in their mutual opposition.’ footnote4 Poulantzas does not mean by this proposition that classes can only be understood in terms of class consciousness. Class struggle, in Poulantzas’s analysis, does not refer to the conscious self-organization of a class as a social force, but rather to the antagonistic, contradictory quality of the social relations which comprise the social division of labour. Class struggle exists even when classes are disorganized. 2. Classes designate objective positions in the social division of labour. These objective positions, Poulantzas stresses, ‘are independent of the will of these agents’. footnote5 It is crucial not to confuse the analysis of the structure of these objective class positions with the analysis of the individuals (agents in Poulantzas’s terminology) who occupy those positions. While both analyses are important, Poulantzas insists that ‘the question of who occupies a given position, i.e. who is or becomes a bourgeois, proletarian, petty bourgeois, poor peasant, etc., and how and when he does, is subordinate to the first aspect—the reproduction of the actual positions occupied by the social classes’. footnote6 Poulantzas refers to the reproduction of these objective positions within the social division of labour as the ‘structural determination of class’. These first two propositions taken together imply that in order to define classes it is necessary to unravel the objective positions within the antagonistic social relations comprising the social division of labour. 3. Classes are structurally determined not only at the economic level, but at the political and ideological levels as well. This is perhaps the most distinctive (and problematic) part of Poulantzas’s analysis. While it is true that ‘the economic place of the social agents has a principal role in determining social classes’, footnote7 their position in ideological and political relations of domination and subordination may be equally important: ‘It must be emphasized that ideological and political relations, i.e. the places of political and ideological domination and subordination, are themselves part of the structural determination of class: there is no question of the objective place being the result only of economic place within the relations of production, while political and ideological elements belong only to [class struggle].’ footnote8 Political and ideological factors cannot be relegated to the transformation of a ‘class-in-itself’ into a ‘class-for-itself’, but lie at the heart of the very determination of class positions. footnote9 Given these premises, the basic theoretical strategy Poulantzas adopts for analysing class boundaries centres on elaborating the economic, political and ideological criteria which determine objective class positions within the social division of labour. We will first examine how Poulantzas does this for the working class and the new petty bourgeoisie, and then for the bourgeoisie.