Class politics, once the unquestioned centre of the socialist project, has became the object of intense controversy. There have been many reasons for this startling development—the appearance of the so-called new social movements and the continued failure of traditional Left parties to effect fundamental social change are just two. But at the heart of most critiques has been the notion that the working class is no longer a viable basis for socialism. Pointing to the contraction of the manual working class and the proliferation of ostensibly different strata, a number of socialists have argued that it is time to bid farewell to a social group that is anyway primarily turned to material preoccupations. For the advance of socialism, it appears, an alternative agency or agencies will have to be found.footnote1 What is striking, though perhaps not surprising, is that this abandonment of the first principle of Marxist political practice has not been rooted in a solid theorization of contemporary capitalist society. In fact, most recent contributions to the debate on class structure have rejected the older orthodoxy (as expressed in the writings of Poulantzas, Carchedi and Wright) that there exists a relatively large group of workers who cannot be regarded as either capital or wage-labour. Instead, there is now a majority view that much of the ‘new middle class’ is virtually indistinguishable from manual labour in its conditions of work and existence. The social structure would thus include a small capitalist class, a privileged middle class that is much narrower than previously thought, and a large and growing working class of manual and non-manual labourers. These theorists have, on balance, taken a step in the right direction. Their analyses seem far more consistent with contemporary reality—mass unemployment, declining real wages, erosion of the economic position of the traditional middle class, even the decline of the professions—than do those of the critics of class politics. They also recall Marx’s own basic insight that socialism is not a mere utopia, with no material basis in social reality, but the self-emancipation of an actually existing and exploited working class.footnote2 As will be argued below, however, what is still missing is a consistent Marxist definition of the nature of class and class conflict in contemporary capitalist society.

As Gavin MacKenzie has correctly remarked, discussions of class structure in contemporary Marxist and sociological theory have focused on the ‘boundary question’—i.e., on ‘where to draw the line’ among the various classes characteristic of capitalist social formations.footnote3 Virtually all the major contributors have agreed on a three-class model (Erik Olin Wright’s analysis of ‘contradictory class locations’ is a possible exception);footnote4 but there has been a shift in the general perception of the boundary between the working and middle class. Whereas many early commentators, including Poulantzas,footnote5 laid stress on the manual–non-manual divide, it is now widely agreed that many subordinate non-manual workers experience conditions of work, levels of pay and degrees of authority comparable to those of manual workers, and that they too should be regarded as forming part of the working class. One of the most influential attempts to locate the boundary between working class and middle class within the non-manual category was an article by John and Barabara Ehrenreich, ‘The Professional–Managerial Class’,footnote6 which posited the existence of an intermediate class of relatively privileged professional, intellectual, technical and administrative workers. This professional–managerial class (pmc) does not include many routinized, subordinate non-manual occupations, and thus constitutes a small proportion of the total workforce. As their criteria for designating this group a new class, distinct from the bourgeoisie above and the proletariat below, the Ehrenreichs point to (a) its non-ownership of the means of production and (b) its role in the reproduction of capitalist social relations. Whether in the labour process as such or in institutions such as schools, welfare agencies or the media, members of the pmc exercise social control over the working class and reproduce new generations of ‘docile’ wage-labourers.

Two more recent books, owing a great deal to Braverman’s Labour and Monopoly Capital, have taken up the pmc thesis in somewhat modified form. Nicholas Abercrombie and John Urry, in Capital, Labour and the Middle Classes, also contend that the ‘line’ should be drawn within the non-manual categoryfootnote7—between a de-skilled white-collar group and a ‘service class’footnote8 of privileged professionals and administrators, most of whom are in the middle layers of bureaucratic employment. The primary distinction between these two groups is in their market and work situations. De-skilled white-collar workers experience low pay, poor mobility, extreme subordination and routinized jobs, making them similar to the traditional manual working class. In contrast, the service class enjoys far greater material rewards, relative freedom from control, a degree of authority, and realistic prospects for career mobility. These sociological differences override the common functions performed by all non-manual labourers (control, reproduction and conceptualization), although Abercrombie and Urry seem to suggest that the important aspects of these functions may be increasingly concentrated in the ‘service class’.footnote9 The class map is thus similar to the Ehrenreichs’: a large heterogeneous working class, a small capitalist class, and a privileged middle class which, though distinct, shades off into the other two.

Finally, Martin Oppenheimer’s White Collar Politics footnote10 proposes a third analytic variant, splicing together the ‘service class’ thesis and Erik Olin Wright’s notion of ‘contradictory class locations’. White-collar work comprises a variety of class locations. Probably the largest group consists of low-level clerical and sales workers, many of them women, whose levels of skill, material reward and authority clearly place them in the working class. At least two white-collar groups, however, may not be designated as working class: the capitalist owners and top managers of large corporations; and a group in the middle (the professionals) whose position is ambiguous and even contradictory. Oppenheimer remarks of them: ‘The various fractions of the professional stratum, insofar as they can be identified, either belong to a vestigial class formation (the petty bourgeoisie) or are marginal to one of the major classes, or even both of them, depending on their function within the capitalist mode of production at any given moment—a moment that is constantly in motion, so that the situation of a given fraction of the professions, or of a profession, is “normally” contradictory, ambiguous, mystified.’footnote11 Oppenheimer locates the ambiguity of the professions in their ‘market and work situations’. Some of the lesser professions, and perhaps even some of the more prestigious, may be falling into the working class as their conditions of work and their material position deteriorate. Yet even the most ‘deprofessionalized’ enjoy substantial technical control over their work, and exclude working-class clients from the decision-making process—as a result, they form at best a marginal, ambiguous part of the working class. Top professionals, on the other hand, are directly involved in policy-making and blend off into the bourgeoisie, while middle professionals are the most ambiguous of all. Like subordinate workers, they are ‘involved in a set of oppressive and exploitative relations’ but are caught between capital and the labour force.footnote12 In sum, Oppenheimer too rejects the Ehrenreichs’ functionalism and hesitates to identify a clear-cut middle class;footnote13 but he does describe a middle stratum that does not really belong to either of the major classes in capitalist society.

These analyses represent significant advances in our understanding of contemporary social structure. It remains doubtful, however, whether any of them have solved ‘the boundary question’. In particular, if we examine the criteria they use in locating the barriers between the various classes (economic privilege, work conditions, function), many problems with their definitions of class remain. Consider, for example, an argument which enjoys wide acceptance in Marxist circles—that skill levels may be equated with class distinctions. This view is implicit in the willingness of many Marxists (including Oppenheimer and Urry) to range de-skilled non-manual workers in the working class, while debarring more highly skilled professional and technical workers. One might ask what is the basis for conflict between the skilled and de-skilled. Can one say that the skilled exploit the less skilled, as the bourgeoisie exploits the proletariat? Erik Olin Wright, in his most recent work, attempts to argue that they do, but even he is obliged to admit that ‘skill exploitation’ is a rather dubious concept.footnote14 Within capitalism, the skilled exploit the unskilled only in the very limited sense that they can temporarily mollify their own exploitation by maintaining their relative privilege. This is hardly the same thing as the irreconcilable, long-term conflict of interests implicit in the exploitation of the proletariat by the bourgeoisie. Moreover, quite apart from the increasingly evident problems in defining skill,footnote15 most students of the de-skilling process (including Braverman himself) have found the distinctions between skill levels to be exceedingly vague and difficult to draw. Thus Crompton and Jones, in their recent study of office de-skilling, found ‘no clear break between routine clerical work and administrative, professional and managerial work’ and questioned whether skill was the basis for class distinctions in bureaucratic structures.footnote16 We are thus entitled to ask just how de-skilled one has to be to qualify as working class. In addition historical studies indicate that occupations which were once highly skilled can and do experience a de-skilling process. This has certainly been the case for many manual occupations, and the process has been reproduced in some non-manual jobs—Margery Davies’s Woman’s Place Is at the Typewriter provides a detailed description of this for American clerical workers.footnote17 If skilled occupations are exposed to this process, why must we wait until it occurs before we accord them working-class status? In fact, Marxists do not do so in the case of highly skilled blue-collar occupations, even when they also have some supervisory responsibilities.

In reading Marxist studies that use skill as a criterion for drawing class distinctions, one sometimes detects a perverse kind of ‘optimism’, a feeling that this will solve the problems posed by contemporary capitalist social structure. The ostensible middle class of non-manual workers, so often held up as a refutation of Marxist class theory, is here seen as a shrinking group that de-skilling largely reduces to conditions like those of manual workers. Dale Johnson, for instance, suggests that the erstwhile middle classes of modern capitalism (an amorphous group including, amongst others, semi-autonomous employees, service professionals and middle-level administrators) are being ‘polarized’ by developments in the capitalist mode of production.footnote18 A small technocratic group, gaining in power and prestige, is increasingly engaged in the performance of control functions delegated by capital, and is thus being drawn closer and closer to the capitalist class. In contrast, a larger group finds that, with its labour-power devalued, it is being driven towards the working class. From this it is argued that the proletariat is a homogeneous class, all of whose members live and work in marginal economic circumstances.

Now, it is undeniable that de-skilling is a major force affecting the labour process and a constant threat to virtually all types of workers, both manual and non-manual. But it is not at all clear that the de-skilling process, under capitalism, will produce a permanent homogenization of labour. As Paul Thompson points out: ‘Homogenization is a weak link in the chain of argument on skills and the labour process. There is a great difference between all work being subject tendentially to the same trends with respect to skills, and saying that all work is the same.’ footnote19 Not only do workers respond differently to the experience of de-skilling— some seek to preserve lost skills, or to re-define new ones, while others capitulate—but the labour process continues to include a great diversity of types of work. Even ‘de-skilled’ jobs require different abilities (compare a typist with an assembly-line worker) and command different material rewards (this, after all, is what the movement for ‘comparable worth’ is trying to change). And while the situation of portions of the highly skilled ‘middle classes’ may have deteriorated in recent years, as Johnson suggests, the historical record should teach us that such trends are generally temporary, or overstated. The ‘middle class’ has been pronounced dead on many occasions. The reality is that skilled jobs persist: some, such as engineers, successfully resist de-skilling while other skilled jobs are created as the labour process evolves.footnote20 All in all, Marxists who hope that de-skilling will homogenize the labour force leave themselves open to the kind of criticism mounted by André Gorz when he denies that capitalism universalizes ‘general abstract labour’.footnote21 Skill, then, is not the same as class. If anything, de-skilling is more usefully understood as a symptom of class: that is, de-skilling—or what Marx referred to as the ‘real subordination of labour to capital’—can only occur once a worker has become a wage-labourer.