Chapter 3 Discourse Theories And Critical Cultural Analysis

Any critical cultural analysis must incorporate some sort of attention to linguistic analysis, some account of discourse and how it works. The general advantages of doing this have been explained in chapter 1 -- such a move is one way to bring marxism to bear upon modern cultural activities and processes that have been transformed by linguistic activity (in its broadest sense). To be very blunt about this project, using linguistics enables us to pin down the transformations away from the basic context provided by classic marxist analyses of the mode of production.


 In the traditions we are going to examine in this chapter, discourses can be defined as 'socially situated speech', according to Thompson (1984), and it is easy to tease out some pertinent implications from this basic definition. The intention is to rely largely on Thompson (1984) and Macdonell (1986) to guide the discussion. A social situation implies more than one participant, and therefore a social orientation of the speech, some relationship between speaker and listener(s) for example. Naturally, marxist analysts are especially interested here in the type of relations involved and how the conduct of the discourse influences it -- how a power relation is maintained, or how the real interests of the speaker are hidden or represented as universal. As we saw in chapter 1, other types of analyst might wish to pursue more 'functionalist', integrating, consensual or nomic mechanisms in discourses -- but it is easy to see how a critical, marxist agenda might be pursued instead.

 The social situation of discourses also helps us focus on the institutions in which speech takes place, and how their organisation distributes power to different participants to join in discourses, or be forced to be addressed by them. In the terms of this chapter, organisations like hospitals or even entire education systems embed acts and sequences of communication and the positions and viewpoints permitted in them: 'institutions...prompt people to speak...and store and distribute the things that are said' (Macdonell 1986: 4).

These concrete embodiments have a history, of course, and it becomes possible to analyse it -- how discourses replace each other in public life, for example, and how the institutions in which they are embedded develop and change. This provides another insight for Macdonell -- particular discourses bear traces of rival discourses in them, since they are often shaped by the need to oppose earlier ones, or address earlier agendas, or distinguish themselves from earlier or competing contemporaneous ones.

Focussing on discourses as units of study, therefore, seems to promise much progress. Speech (including writing, of course) is not being seen as simply determined by an economic structure or class position, but is granted some autonomy, creativity and effects of its own, including an ability to operate 'between classes' in Macdonell's phrase. On the other hand, linguistics is being combined with critical social investigations instead of being an abstract study of the structural relations between signs.

The latter development explains why early developments were seen as 'poststructuralist',  since the whole analysis shifts away from general structures of language and meaning in structuralist linguistics to specific concrete languages or 'sociolects', the actual words and expressions found in concrete texts, rather than some 'grammatical ideal', as Thompson puts it. We can now study 'concrete forms of differing social and institutional practices' rather than 'language', for Macdonell. We have progressed towards sociology or marxism, in other words, compared to the more general change toward the 'new semiology' in Barthes (which we examine in chapter 4).

Finally, although we have not discussed it at all here, there are the benefits of the much more general 'linguistic turn' in social theory: briefly, studying language in use is more promising in several ways than trying to study individual consciousness if we are interested in the effects of ideology or culture (see Bernstein (1985) on critical theory’s variant). However, there is no need to study such language in use formalistically, purely as language, as abstract linguistic rules. Instead, we need to rethink the relation between linguistics and sociology or marxism.

Discourses and social classes

The work of some discourse theorists can be connected to marxism without much additional work, as Macdonell explains. It is quite easy (still within a marxist problematic for the moment) to see that social class structures and limits the discourses that commonly occur, for example, and that the different social classes use words and expressions differently. Obviously 'political' discourses offer a chance to see this clearly, and two rather elementary empirical experiments are insightful. The first one is discussed in some detail by both of the writers we are using as guides.

Pêcheux and his associates asked a small sample of students to read and interpret a particular policy document to reveal how words like 'planning, political change, radical reform' could be (rather artificially I thought) incorporated into quite different political 'corpora'. A 'corpus' is a new linguistic term  referring to a cluster of language, so to speak. It is 'constituted by a series of discourses ...which are assumed to be "dominated" by stable and homogenous conditions of production' (Thompson 1984: 239. For the original, see Pêcheux 1982).

The terms used in this document permit this kind of ambiguity because they bear traces of past political struggles over them and, as Macdonnell explains, this led Pêcheux et al. to suggest that this residual ambiguity could be reactivated, so to speak. This leads to a more optimistic possibility than had been granted in Althusser’s famous work on ideological state apparatuses, which stressed the overwhelming power of those apparatuses. Now, those 'hailed' in particular ways by the apparatuses do not have to conform to the identities on offer but can 'counteridentify' (a kind of inversion in our terms -- see chapter 2) or even actively 'disidentify'(a more radical kind of break involving a rejection of the whole mechanism).

The other experiment was carried out using the work of a marxist linguist from a different tradition, but we can use it to make a similar point. According to Woolfson (1976), Volosinov's work also traces class struggle, and the effects of a partially separate 'sign community', in the 'multiaccentuality' of the sign. These persist, despite struggles by various ideologues to remove these traces and the contradictions they produce. Woolfson proposed to demonstrate these linguistic struggles and their effects on experience by asking Glasgow Corporation bus crews to read some press reports of political events and record their discussions. Woolfson claims to have found the effects of hegemonic ideology, but also some limited potential resistance, in the critical remarks made by the workers, the way they both used and challenged and reinterpreted the clichés of the reports, and performed a kind of 'counteridentification' by inverting and playing back terms like 'parasites' to refer to ministers instead of workers.

This kind of work, mixed rather curiously with a sociological study of political deviancy, led to some well-known early explorations of the television audience and its ability to resist or even oppose the 'codings' found in news or 'current affairs' programmes (see Harris 1992) ( and see Hall on coding here).

Other linguistic work can not be absorbed so easily into marxism, though, and has to be supplemented or reinterpreted. Both Macdonnell and Thompson perform work of this kind on different discourse theorists (including ethnomethodologists and some sociolinguists interested in educational discourses for Thompson), although Macdonnell favours Althusser and Thompson Habermas to provide an different integrating frameworks. One mystery arises for me in this integrative work, though, in that some obvious sociological studies of social class and language are omitted -- those of Bernstein for example (see Atkinson 1985 for some clues about how this work too might be subsumed more general critical concepts).

Critical linguistics

One study cited by Thompson does make progress in terms of detail, though, compared to the rather preliminary pieces in Pêcheux or Woolfson, which are designed to integrate marxism with linguistic concepts rather than to do much concrete analysis. Fowler and others (e.g. Fowler et al.. 1979) offer intriguing analyses of a number of actual strategic texts (to use the terms we introduced earlier), including materials deployed by 'middle management' on a management training course.

The language used orders the peculiar and contradictory social world of the middle manager, Fowler et al.. argue, preserving a certain tactical ambiguity about their precise roles, for example (by pursuing two kinds of terminology referring separately to 'who is called what', and 'who does what', or by preferring the abstract noun 'management' and using usually passive verbs instead of nominating precise agents and active verbs). You might find local examples, perhaps, and I have one here which reports that 'Change was welcomed by the Registrar's Department...'.

In another example, one very close to my own experience in scores of consultations with staff in various educational organisations, Fowler et al.. note that 'spatialisation of the ideological problem [of control and its legitimacy] pervades the syntax' (1979: 88) leading to the plethora of management diagrams and charts, or the widespread use of the 'line-management' metaphor to describe who reports to whom exactly. Such apparently harmless obsessions help conceal the real issues of where decisions are taken and by whom.

The chapter goes on to list various techniques found in writing and speech to represent 'management' as a benevolent group who happen to have emerged to fulfil vital abstract functions -- pronouns like 'we' are preferred to 'they'; 'modal verbs' like 'can' or 'think' (as in 'We think we can grow our college in the future', to take another local example) give a nice ambiguity to the issue of whose permission has been sought, and whether or not decisions have actually been taken yet. Various 'distancing devices' (like 'that would be difficult if...') or 'stalling devices' (including the very fashionable slight stutter at the start of spoken replies) both give the speaker time to think and create an impression that 'an extra amount of editing was applied by the speaker to the bit of speech that follows' (1979: 92), and, one might add, it alludes to an upper-class upbringing or participation in elite schooling, in Britain at least.

Detailed analysis follows of the substantial amounts of ambiguity that expresses itself in the avoidance of the present tense, for example (endlessly deferred futures or cautious subjunctives are common in my experience): 'linguistic equivocation mirrors the tension of the real situation' (Fowler et al.. 1979: 92). I also like the process referred to as the 'reclassification' of the 'syntactical problem' in management-speak, especially the way descriptions become activities, so that describing oneself as a manager becomes self-sufficient, a term requiring no further investigation or elaboration, one which takes on a sense of importance and status immediately.

Later chapters examine the ways in which various communication professionals deal with 'awkward facts' or anomalies when they construct press releases or television news programmes. What the authors call 'modal' activities can use indirectness and distance to code power relations, for example, while various 'transformations' can hide or displace agency, objectify events, classify actors in various disputes, and 'raise' or emphasise particular marginal aspects of a story. Very simple linguistic variations are used to perform these transformations: in one example, a story beginning 'South African police shot several black demonstrators today...' is transformed in to the much more passive and objective version 'Several black demonstrators were shot today in South Africa...'

Of course, considering television (and film) introduces other possible ways to analyse discourse in this sense, and we have discussed some of the issues about visual representations and cinematic narratives in earlier chapters (and see critical realism file) . We also know of newer work on the television audience as 'active', which raises new possibilities for research in Fowlerian critical linguistics. Do the actual readers of middle-management texts or organisational diagrams fall under the spell of these techniques, or do they too 'resist' in any of the ways we have described? Is there a growing cynicism among those ‘consulted at’? Can we see the emergence of charismatic management seminars styles or documents as a response?

There is also another highly detailed study in Thompson (1984), this time of political discourses associated with Nazism, as analysed by the French theorist Jean Pierre Faye. This one is of special interest methodologically speaking as well as substantively, and ,while re-reading it recently, I found many implications to pursue beyond the immediate ones for Nazism. Faye compares very favourably with the well-known studies of Thatcherism in British cultural studies (in Hall 1988, for example -see reading guide here), and offers a much more rigorous theoretical structure and one which demonstrates a clearer grasp of the connections between political narratives and actual events. One can also compare Faye’s approach with the similar-looking ‘topographical’ analysis of Disneyland in Marin (1977) (to which we will return in the last chapter).

Faye offers a complex analysis of the complex and developing narratives used to consolidate the power of the various groups in German politics in the emergence of Nazism. The narratives develop in response both to the raw materials available and to external events. The linguistic raw materials include various sets of terms used in political oppositions (like ‘conservative’ and ‘revolutionary’, or ‘völiksche’ and ‘Semite’). Linguistic practices available to the developers of narratives include transformations and weavings of these terms, so that Hitler is able to label himself as a ‘conservative revolutionary’ and to claim the middle ground, or Goebbels is able to identify the narrow interests of the far right with some sort of mass sentiment.

As narratives change and develop, they influence actual events. One example concerns the Nazi ‘economic miracle’ engineered by Schacht as President of the Reichsbank (Thompson 1984: 223--4). The spoken discourse embraced by Schacht apparently stressed the virtues of personal constraint and savings as a source of recovery, while his actual practice involved secretly channelling large sums of money to the armaments industry. This could be glossed as expenditure on ‘public welfare’ eventually. Such glosses permitted this expenditure, which was a real element in the economic recovery of Germany, of course, and one of the main causes of the subsequent War. In this way, ‘texts’ do not just exist in some ahistorical way, as a universal feature of human life, but are located in, and transmitted along ‘lines of force’, in Thompson’s phrase (1984: 231).

Thompson and ‘depth hermeneutics’

Thompson’s own position develops out of these detailed readings and critiques, and from his own earlier (1983) work which centred largely on Ricoeur and Habermas. His position is of particular interest to us since it offers a firm rejection of the notion that social life is best studied as a variant of linguistic behaviour, and it goes on to develop a thoroughly sociological or materialist account, based on an extension of Giddens with liberal helpings of Habermas.

The problems with linguistic approaches can be sketched by re-considering work like that of Fowler et al.. Although it is insightful (and even rather amusing or rather useful in helping sceptics perform exposés of current managerial practices), it is very abstract. The grammar of  the sentences is  the major focus, rather than the contents themselves. As we have seen before, content, a social context, a position of power and domination are smuggled in, so to speak, with the chosen example of management-speak. We know the context already, as indeed we do with the example developed by Faye (or, for that matter, those analyses of advertising developed by Barthes). The same combination of an implicit context, and a focus on a specific linguistic feature (grammar, syntax, narrative ,or semiotic for the ‘structuralist’ examples) affects much of the work we have reviewed so far.

What results is a kind of partial methodology -- the specific linguistic features somehow generate the effects of the examples without any further analysis (or with only limited analysis in the better examples). Thus ‘modalization’ or whatever is implicit in the power relations of management, narrative explains the acceptability and thus the rise of the Nazis, albeit in a very complex manner.

We can anticipate a little and invite readers to consider whether or not Barthes or the postmodernists operate in the same way, this time with largely (post)structuralist linguistic terms. As a clue, Thompson (1983) argues that the formal split between connotation and denotation in Barthes’ analyses of modern myths (discussed in the next chapter) is not sufficient on its own to explain the persistence of concrete ideologies, while our own later chapters consider whether the mere flourishing of terms like ‘intertextuality’ is sufficient as an instant analysis of actual films or videos. There is usually an implicit social theory lurking somewhere in these exercises, of course

Thompson (1984) suggests we need a far more explicit analysis both of discourse and of social structure if we are to undertake a critical analysis of ideology (whether postmodernist have abandoned this project altogether is something we shall discuss). His own model involves a further development (after a critique) of Giddens’s ‘structuration’ approach which we outlined briefly in chapter 1. It seems we now need to develop a three-fold model of everything to go beyond Giddens’s radical dualities, and we need much more concrete analytical tools to specify the effects of social structures in particular.

Thus Thompson offers us a threefold model of the social system with the levels of ‘action’ and ‘social structure’ in Giddens supplemented by a definite level of social institutions. Giddens’s split between ‘virtual’ and ‘structural’ dimensions is replaced by a more concrete division at each level. Thus we have actions as both reflexively-monitored general flows of activity and as concrete actual action-events, institutions as both ‘specific’ (e.g. the University of London) and ‘sedimented’ (‘the university system as such’), and social structures as both the elements necessary for any social life (a production system) and specific social formations (capitalist production) (1984: 129).

Power is still built in to the model, and takes different forms at each of the three main stages: an ability to act (at the level of action), a capacity to act (institutional level), the conditions which limit the range of institutional variations (structural level). These specific types of power replace the Giddens notion of rules and resources, which, we have already argued (chapter 1) are too general, and too dependent on Giddens’s own root metaphor (the use of a language -- see Thompson 1984 chapter 4 on Giddens). Here, we reinstate properly specific sociological categories, so to speak.

Having cleared the ground, we can now see how ideology works, mainly to obscure the relations between the concrete and the general levels of action, institution and structure we described in the penultimate paragraph. It is in the interests of domination (expressed in the ‘systematically asymmetrical’ conditions for action associated with the class system, but also with race and gender and with rather vaguely specified ‘other’ forms) to legitimise existing actions, institutions and structures. Dominant groups also dissimulate (conceal or block knowledge of processes or possibilities) and reify (by offering no concrete histories of the structures which exist).

We must pursue our enquires by analysing these forms of domination at the three levels, and here we can usefully incorporate people like Goffman on the situational specifics of action, or Bourdieu on the institutional contexts (of educational institutions, for example, to pick up one of my favourite topics). Marxism or Gouldner’s critical sociology seems to be useful at the structural level. Although this is impressive, of course, one cannot help thinking of the pursuit of the complex (which we have discussed in others) by ‘bolting on’ additional layers and analyses, however.

We can also analyse discourses. Here, analysis of narratives could be useful (not only as in Faye, but also in Barthes on myth -- and, presumably, in some of the examples of ideological narratives and representations we discussed in the last chapter). Thompson also wants to include critiques of grammatical structure as in Fowler et al.., and something that looks rather like what came to be known more fashionably as ‘deconstruction’ -- the interrogation of texts and the exposure of their ‘contradictions, ... inconsistencies, ... silences and lapsus’ (1984: 137)

Thompson wants to go beyond such ‘internal’ discursive efforts, though, and to revitalise Ricoeur’s ‘depth hermeneutics’ here (which we met, very briefly in chapter 1). This interprets existing discourses in a much less formal ‘structural’ linguistic manner than some of the alternatives discussed above, and keeps open the possibility of the explanation of discourses (by refusing to stay exclusively inside the text, by going out to the structural and institutional contexts outlined above).

As we saw, this draws upon the production contexts of texts, their socio-historical locations (134) and ‘lines of force’, and what might be termed the ‘social existence’ of texts after they have been written. They come to take on an ideological (e.g. legitimising) role as when different readings are quietly privileged by dominant groups. Thompson refers again here to the specific report discussed by Pêcheux which we have considered above, but we might think of the plagiarism example in the last chapter (where a student essay was rendered as ‘evidence’ for some judicial interpretation). To take another specific example, a struggle is underway in a local school to organise a debate about ‘opting-out’ of local government control: one party wants to take a statement of Government policy as some ‘factual’ or ‘naturalistic’ text apparently without a (party political) history, to structure the necessary meetings with parents as mere ‘information sessions’ where they can learn about this neutral text, and to manage rival policies as ‘outside disruptions’ to this impeccably non-political exchange of information!

Finally, Thompson is well aware that analyses of this kind are clearly ‘political’ ones themselves. There can be no attempt to pretend that we are doing social science here. Nor can we pretend that a neutral ‘philosophical’ or historical critique can be allowed to slide over into a political one in the hope that no-one notices (not an uncommon development in my experience). Thompson moves on to the ground staked out by Habermas to suggest that the only real way to ground such critical work is in some idea of genuine ‘generalisable’ or universal interests, (briefly) ones that would be generated if all the participants concerned were able to discuss matters freely and without any external constraint.

As with Habermas, Thompson realises that this is going to look rather an empty and abstract procedure, and that there are going to be real problems in using this sort of procedure to connect to complex actual political situations. Certainly, no theorist can simply expect any political  action to follow immediately from the analysis, and, in situations where there are asymmetrical power relations already, which would simply prevent any chance of unconstrained discussion, the only option seems to be a rather defensive ‘counterfactual assertion’, an insistence that there is another way.

However, these issues haunt any such endeavours, and at least we have an explicit discussion. An ‘aversion against the universal’ in Lyotard, say, ( to use Honneth's term) pours scorn on the possibility of unconstrained discussions to establish generalisable interests, but that leaves only an abstract and formalistic way to ground analysis -- some support for a general pluralism, some attempt to ‘fuse’ in some unclear way some activities with an indefinable ‘différance’ or to claim them as alluding to the realm of the sublime, as we shall see.



The problems for critical discourse theory are also focused nicely in the discussion of one of the best-known practitioners -- Foucault. Here, Thompson actually offers no sustained discussion (Macdonell does, however), but Foucault's work has been seen as rather uneasily compatible with marxist interests in the networks of power and domination dispersed throughout society.

Poster (1984) sees Foucault as offering an account of a much needed ‘mode of information’ to complement marxism’s ‘mode of production’ and to analyse the new ‘forms of language experiences that now inhabit our social landscape’ (167) (including new forms of electronic surveillance and interactions with computers). Foucault is exactly the person to pursue the project at the centre of the work in this chapter -- ‘to ‘make intelligible a level of analysis consonant with emergent [‘linguistic’] forms of social relations’(168). However, as we shall see, these enterprises have other consequences, including the need to abandon much of the usual material on power, ideology, domination by classes, the classical subject, and a good deal more.

We end this chapter with Foucault as a transitional figure, in other words, like Barthes, one who shows some of the less desirable consequences, for marxists, of attempting to develop new forms of ‘linguistic’ analysis. He will also serve as an exponent of ‘post-structuralist’ writing, which we have discussed briefly at the start of this chapter. To borrow from Poster’s discussion again, we can define this approach as offering readings that insist on the preservation of differences or discontinuities (between the past and the present, for example, or between different processes at work  at the same time).

This leads, in turn, to a reflexive awareness of the way in which conventional readings (including marxist or freudian ones) impose unities and smooth out differences. Post-structuralists are sceptical about what might be seen as the politics of such orthodox readings which claim to act in the name of some universal reason or subject but which find themselves complicit in new forms of domination. Referring to marxism, in Poster’s work, this leads to a number of problems, including intellectual elitism, the privileging of the claims of the proletariat over those of other equally oppressed groups, and an unfortunate alliance with the very forms of dominating reason (like liberalism) that were allegedly transcended (poster 1984, especially chapter 2).

Foucault’s own attempts to avoid these dominating readings are not always successful, argues Poster, especially when he refuses to account for the specific positions, contexts and politics of his own work -- a refusal of the function of author is no real substitute for such analysis. However, there is an attempt to develop a suitable style of analysis which keeps to post-structuralist tenets, and the development of suitable concepts (especially ‘discourse/practice’ as the unit of analysis rather than some essentialist category like ‘man’ or ‘mode of production’. This partly explains the strange and frustrating nature of Foucault’s descriptions too, as a number of commentators suggest (e.g. White in Sturrock 1979, to whom we have referred already).

With those warnings in mind, let us begin with a look at some attempts to borrow Foucauldian analysis for a more general marxist project. He has been seen quite frequently as providing the means to both concretise and 'elaborate' the basic gramscian concept of 'hegemony' in British 'cultural studies', for example, to extend and modernise such analysis, and to fight off the problems that came to attach themselves to marxism after the early attacks by postmodernism (see Harris 1992 for some examples).

In one of the clearest and most readable cases, Hargreaves (1986), uses Foucauldian themes to explore the ways in which discourses about sport, fitness and athletics and their organisation can be seen as 'disciplinary technologies' designed to enforce some suspect concept of a 'normal person' in a peculiarly contemporary manner -- by persuading people to regulate, manipulate and police their own bodies in a revisited version of the Protestant work ethic.

This example, apart from giving us all good reasons never to visit the Sports Centre again, gives some idea of the tremendous detail of description of the mechanisms in Foucauldian analysis to which we return below. Yet there are problems in such ‘elaborated gramscianism’, as a former exponent later came to realise. Bennett (1990: 246) argues that ‘the theoretical assumptions from which Foucault and Gramsci proceed are so sharply contrastive as to belie any significant points of contact’, and goes on to list the main differences -- Foucault abandons the ‘ideology/truth’ problematic of Gramsci (and, one might add, the purely negative prohibitory or ‘juridical’ concept of power), advocates a ‘grid’ of power relations as fundamental to social life, rather than the ‘State/civil society couplet’, denies the possibility of a ‘unified source of opposition’ as opposed to the gramscian ‘project of counter-hegemony’ and so on. For Bennett, Foucault does share with gramscianism the need to attend to concrete detail, however, although even here one might still entertain doubts about whether the point of the detail is to confirm a (very flexible) concept of hegemony, or to protect difference and deny totalisations.

No-one discusses what might be called ‘local domination’ in quite the same specific way as does Foucault, but then his objects of study provide wonderful levels of detail themselves, and perhaps the credit should really go to the designers of the disciplinary technologies and apparatuses in the first place, and the records they have left of the principles they have used. We can illustrate the range of the approach with reference to the example of prisons and punishment. The other main ‘applied’ example -- sexuality-- appears briefly in our discussions of sexual identity in later chapters.

Bentham might have used dubious abstractions in his general theory, but he displayed a wealth of detailed planning and  insight in his plans for a rational prison system. Panopticon was a system designed to keep every prisoner under individual surveillance and control from a central position (which accounts for the characteristic star shape of some of the older British prisons like Pentonville), and to offer detailed ways to modify and shape their behaviour, including maximum opportunities to repent in isolation, and a system of providing mattresses graded in thickness and comfort and awarded according to progress made towards repentance and reform. Embedding indeed!

In a way which inevitably evokes later work by Goffman on ‘total institutions’, all creature comforts were administered in this close 'payment by results' manner in order to shape the behaviour (and the intentions) of prisoners towards the desired end, and Bentham included detailed costs and procedures to inform the conduct of the administrators of the system. The whole system is a triumph of bourgeois rationality, as good as any in Weber’s work on the Protestant ethic, and, of course a triumphant demonstration of Bentham’s utilitarianism

Foucault argues that such technologies did not appear as an obvious alternative to the earlier system, nor is there a tight link to social evolution or to changes in the mode of production. Instead, we have the emergence of a more general discourse/practice in its own right, so to speak, one which appeared in other areas of social life too. This discourse/practice not only reflects conceptions of punishment it activates them, calls them into being, generates the actual prisoners and regimes it describes. Finally, we are to see these developments as perfectly rational in their own terms, with obvious critical implications for our own systems of discipline and punishment -- perhaps these are equally historically relative, equally dependent on a discourse/practice that appears normal to us.

Bentham was also an educational reformer, and he produced plans for a rational teaching system, based on a number of precedents and other people’s systems, but again developed to an amazingly detailed extent. In Chrestomathia Bentham proposes a thoroughgoing simplification and rationalisation of school subjects, for example, as well as detailed designs for a rational teaching scheme involving the older pupils teaching the younger ones, with periodic testing and seating in classes ranked according to merit. The scheme would be more fair and rational than the brutalised and scandalous teaching regimes in existence in Bentham's day, (which involved a good deal of flogging, for example) and far more cost-effective.

Nearly two hundred years after this sort of proposal, I have argued elsewhere (Harris 1987), Bentham's design is coming to fruition in the modern forms of 'individualised teaching', educational technology, and above all in features like modularisation, continuous assessment regimes, and the electronic monitoring of a range of student activities like attendance at lectures or library use (all of which will be coming to your university soon, if they are not already in place, I predict). These changes are already foreshadowed in discussions long ago in establishing early 'distance education' systems like the UK's Open University, I have argued, where again a well-meaning interest in the problems of educating effectively combines perfectly with an interest in cost-effective management.

Other writers have pursued the tracks of Panopticon into modern industrial practices, into attempts to manipulate intentions and behaviours in ‘just-in-time’ regimes, for example (see Sewell and Wilkinson 1992). Thus the labour process becomes more visible in such regimes partly because the drastic reduction of stocks ‘reduces the scope for workers to “hide” any defect’ (1992: 279), and because ‘management must erect a superstructure of surveillance and control which enhances visibility and facilitates the direct and immediate scrutiny of both individual and collective action’ (282). Sewell and Wilkinson include ‘the impact of teleconferencing on managerial self-discipline’ (286) as further examples of a ‘mechanism of Power/Knowledge which can bring out the minutest distinctions between individuals’ (287).

This sort of analysis works perfectly well within marxism (indeed, I drew upon 'critical theory' in my critique of educational technology rather than Foucault), but Foucault cannot be assimilated entirely, as we have seen. Macdonell rebukes him for his inconsistency on important issues (like whether the class system and class struggle does lie behind all the specific cases he examines). Without a consistent position, she argues, we have a rather arbitrary shift of positions from one analysis to the next, and, to pick up on a major problem noticed by other commentators, no real account of organised resistance either.

On the last point, the work of DeCerteau became popular, in cultural studies at least, for offering an account of activities like 'poaching, tricking, reading, speaking, strolling, shopping, desiring' (see Frow 1991 for a critical discussion) as techniques of resistance to the disciplinary technologies. We have already seen Fiske's development of this kind of argument. Poster (1984) cites additional studies of prison life which provide even more detail and which do include prisoner resistance (including the formation of prisoner subcultures, to make a link with an earlier chapter).

Foucault's innovations include an 'archaeological' method and a form of historical research he calls 'genealogical' (defined succinctly in Two Lectures in Foucault 1980 or in Poster 1984). This is a method of using the radical difference of past practices to undermine the rationality of the present, as we have suggested, and neither of these is the kind of rigorously marxist materialist analysis Macdonell would like, of course. Nor does Foucault's specific analysis of power (or rather power/knowledge couplets) lead to organised marxist politics, as we have seen, but points instead to more localised struggles to expose the workings of disciplinary technologies and to permit the excluded (prisoners, nurses, patients) to speak. This can look excessively 'pragmatic' or opportunistic to marxists.

Foucault suggests, as we have seen, that marxism (and all the other grand theories) are discourses themselves, however. Although intellectuals, in France at least, went along for years with the assumption that the categories of marxism simply did connect with the real state of affairs, that they represented the real class system, the real mode of production and so on, Foucault withdraws that belief. As we shall see in our discussions in chapter 5, marxism seemed much less coherent and plausible, and much less able to defend itself against this new doubt than many people thought. Local analyses of domination began to seem much more plausible than the attempts to link up with wider structures.

Secondly, French marxism in particular was every bit a disciplinary technology as any bourgeois discourse, partly because marxists were keen to embrace the title of 'science' in the bourgeois tradition (and to enjoy all the fruits of public recognition, university posts and so on). Foucault (1980: 85) wants to ask the usual questions:

'What types of knowledge do you want to disqualify ..[by this] demand? Which speaking discoursing subjects -- which subjects of experience and you then want to "diminish"...? Which theoretical-political avant-garde do you want to enthrone?'

You can see that this sort of questioning will lead to the familiar accusations that marxism ignores feminist struggle or non-class political movements. Foucault means to include victims of clinics or prisons too, however, and refers to the disinterest of his marxist colleagues. The French Communist Party dominated marxist thought at the time with its own suspect political agenda and its association with tainted Soviet practices (like using psychiatry to control political dissidents). This sort of work fits my own arguments that British gramscian marxism also ended with an uncritical adoption of the disciplinary conventions of the university (although it can be absolved from any involvement in the Gulag, of course).

However, the argument is not all simply critical of marxism, as ever -- Foucault himself seems to need to refer to the class system as the final embedding mechanism to explain the precise course of dominant discourse about sexuality, for example, according to Macdonnell, and, perhaps to overcome a certain acknowledged lack of integration or 'discontinuity' in his work. We have here the old problem, in other words of wanting to address the major struggles in society while refusing to provide any compelling theoretical reason to do so.

Why not study any discursive struggle, why not continue to amass further isolated studies of 'subjugated local knowledges' in institutions like local cricket clubs, churches or garden centres? Foucault knows why -- it would make his work look 'fragmentary, repetitive and discontinuous', a mere scholarly indulgence, ' a typical affliction of those enamoured of libraries, documents, reference works, dusty tomes, texts that are never read...[an addition to the] ...great warm and tender Freemasonry of useless erudition' (Foucault 1980: 79).


We have reviewed a number of approaches in this chapter which can help us to see the problems and prospects of extending marxist analysis into the area of modern culture. The attempts involved various linguistic theories designed to analyse ‘discourses’ in this rather special sense of ‘socially located speech’. Various options were discussed which attempted to deploy and control such analyses, to combine them with materialist (marxist or critical sociological) analyses of social structures. The attempts have generated work of great interest and some considerable ability to grasp the detail of the practices of cultural or ideological domination.

We have also hinted at a dissatisfaction with sort of work, though, most specifically via the work of Foucault. He exemplifies best of all in this collection of work the problem in that to fully grasp the specificity of modern cultural forms of domination, a new beginning seems necessary, the deployment of new concepts and new interests. We shall see another highly influential writer -- Roland Barthes-- treading the same trail away from marxist problematics altogether, in the next chapter. There is a sense in which we have tested marxist approaches to destruction in this chapter: it is not so much that we have refuted them decisively, but more that we have exhausted them, perhaps.

Forget Marx?

Poster’s (1984) commentary on the links between Foucault and Marx begins with a critique of Marx’s approach which has become familiar. I came across a similar critique first in the work of Habermas (1972 chapter 2 especially pp.33--40, and 1974 chapters 1 and 6 for example) in fact, so it might be possible to extend the discussion slightly into his work too. There are serious problems with the labour theory of value, for example, in explaining modern societies. We have outlined the basics in chapter 1, and many sociology students will be familiar with the usual criticism of Marx’s work -- that it is ‘economic reductionist’. This is not just a political flaw, though, which privileges the politics of class over, say, those of gender. It mean that Marx’s main explanatory devices and schemas no longer work.

The reasons are not too difficult to spot. Wealth these days is produced in a variety of ways apart from the classic mode of labour producing surplus value in manufacturing industry. There are problems arising from the widespread deployment of machinery which add value in their own right. There are also ways of creating wealth that do not seem to depend any longer on the direct exploitation of labour. A recent scandal in Britain uncovered the astonishing world of currency futures, for example. Banks got into currency markets initially, it seems, to facilitate international trade, as a service or supplement to industry, and, of course, bought currencies for the future in anticipation of demand from traders. That function would cause few problems for marxist analysis -- but banks then began to realise that they could make a great deal of money by speculating in foreign currencies as an activity in its own right, by buying currency futures at the current price, and expecting those prices to rise. We know from the Barings scandal that the sums of money involved ran into hundreds of millions, possibly even billions of pounds, far outweighing as a source of wealth the ‘normal’ service functions of the banks.

If the exploitation of labour is no longer central to capitalism, nor can be an analysis based upon it as a central motif. The effort needed to ground and conceal the exploitation system in legal and ideological systems becomes more marginal, possibly (although notions of private property and free trade were well to the fore in the Barings scandal!).The classes generated by exploitation need be central no longer, and other forms of division and inequality might emerge: apart from anything else, this leaves bankrupt the marxist account of the inevitable end of capitalism as the proletariat come to be aware of themselves as an active political agent., a ‘class for themselves’.

More technically, the model of domination based on alienation, commodity fetishism, or the exploitation of labour ceases to be the main organising concept for the analysis of ideology, of culture or of discourse. We need new models of communication as well. In Poster, Foucault supplies just such a new account of the ‘mode of information’ as we have seen.

Habermas, somewhat before, had already turned elsewhere, to bourgeois (American) sociology, to symbolic interactionism, and later to Parsons. He was able to insist that these writers had correctly identified a separate sphere of human life, an interest in interaction or communication in its own right.  This sphere could not be collapsed back into a marxist interest in ‘work’ or production, where communication was more instrumental, so to speak, more purposive rational (to borrow Weber), more ideological in classic marxist terms.

We have seen this tactic before, though, and we know that it seems to stave off one problem, only to raise others. We are left with rather uncomfortable -looking lists, or models of different levels. Above all, we are left with a suspicion that analysts have grasped complexity at the expense of incoherence covered by some of the devices we discussed at the end of chapter 1 (‘bolt ons’, for example, involving the quiet tactical switch into another discourse altogether).

In Habermas, this worry led to a massive effort to synthesis the different approaches. To be terribly brief, he was to argue, at a fairly early stage, for the existence of three separated universal ‘interests’ which defined the human condition -- ‘quasi-transcendental human interests’ in work, interaction and emancipation or critique (as in the famous Appendix to Habermas 1972). Each of the different approaches in social science could be fitted into the scheme as pursuing interests in work or interaction, as we have hinted, leaving critical theory uniquely qualified to pursue the third.

This rather evasive system was to give way to a scheme based on types of communication, however. Communication offers, after an evolutionary process at least, a universal capacity to reflect upon different types of knowledge in different spheres, to think in terms of different sorts of validity for that knowledge, and , above all to question the claims to validity of specific utterances and arguments (McCarthy 1984 or Dews 1992 offer much fuller summaries). This system too is highly controversial, of course, and we have hinted at Lyotard’s critique already (and we shall see his alternative in chapter 5).

Marxism, and its characteristic interest and critiques fits into this scheme but it is no longer the dominant approach. Marxism is not just plainly wrong or limited for Habermas, of course -- apart from certain philosophical errors, especially in the way Marx deals with Hegel (see Habermas 1974, chapter 4), it is partly that in Marx’s day, the other types of communication were ‘clamped’ to production and exploitation much more firmly than they were. Habermas leads us back to some of the discussions we have had on culture in (post) modernity here -- the spheres of culture have become much more autonomous since the early days of capitalism.

We can link these points to later work on ‘embedding’. This would be far too technical and neutral a term to describe the ways in which ideas are connected to social practices for serious marxists, of course, but it enjoys a certain current popularity. The sort of analysis we have pursued with Thompson’s ‘depth hermeneutics’ or with Foucault represents the high points of attempts to describe how social life is embedded in a general structure of domination.

Many writers now feel that is the wrong level upon which to operate, that general social constraints have become attenuated in complex and pluralistic societies. This is one sense in which culture is now thought to be 'disembedded'. However, there is still lots of work on what might be thought of as more local concrete embeddings, in institutions, organisations or in discourses, and I have tried to outline some of the more interesting examples. These show considerable variability, but  we are still far from the picture of complete drift and disembedding, the playful relativism that some see as characterising social life.

Forget Foucault?

As we saw with Foucault, though, general theory is not easily left behind, even if we wanted to run the risk of endless repetition and incoherence, and it tends to be smuggled in, especially at the level of choosing examples to analyse in the first place. It should also be said again that none of the alternatives are without problems. The work reviewed here offers some of the clearest recognitions of some of the major problems of analysis. In many ways, those problems have also not been transcended or resolved in the newer approaches, but have been just left aside in a flight from ‘seriousness’ and the model of the severe and responsible analyst operating with an awareness of a wider political context to their deliberations.

For Baudrillard, for example, all these serious endeavours are passé, and so is Foucault's whole corpus of work (and Marx's of course) for that matter. We are simply best off forgetting them, since the game has already moved on, away from all 'productivist' theories, designed to create meaning and system. Foucault described the mechanisms of power in tremendous detail, and was right to emphasise its positive side -- but his clinging to emancipatory politics, even in the form of a role for intellectuals in 'sapping' disciplinary technologies is politically and methodologically naive.

Just like Marx, he failed to realise that the critique of power only helps to validate it. Indeed, constant talk about power, and constant minute analysis of it is about the only thing left which is holding everything together for Baudrillard -- the real system has long ago declined into a mere simulation of power. The painstaking reconstructions of Foucault or Habermas (and others, like Deleuze), and their connections with the real remind Baudrillard of the relation between beautifully produced and detailed pornography and actual sex: pornography does not simply represent actual sex, but informs, initiates, or even constructs it.

Constant argument about the productive (as in 'world-producing') role of power/knowledge, discourse, desire, economic production, or universal pragmatics is at best a gesture -- an 'accumulation against death'. In political terms, fascists were the last to develop such a fascination with power, and to try and mobilise it on a total scale, to impose their political will on reality, according to Baudrillard. The implication, of course, is that Foucault and Habermas are conceptual fascists, so to speak, and like many of Baudrillard’s provocations, this is hardly new -- both writers have long recognised the dangerous connections between cognitive and real domination ( to phrase it in the terms of Adorno’s critical theory).

We end, therefore, with Baudrillard promising a philosophical position of great abstraction, able to fit grand theories together, to reconcile, for example, Marx and Nietzsche (see Kroker 1985) and to show the fundamental similarities between thinkers who had thought themselves opposed. Let us just anticipate some of the further discussions in chapter 5, though, if only to deny Baudrillard some triumphant ‘last word’.

From a participant's point of view, we have also reached absurdity. Here is a great French thinker who can see no differences of any importance between pornography, fascism and Habermas's social theory. Baudrillard's comments seem so unattached and remote as to be pointless, far too clever and innovative to be dispensed with, but irrelevant, simply to be left to one side while the rest of us get on with our lives.

In his interview with Lotringer, Baudrillard shows all his remoteness and brilliance. We discuss his evasiveness in connection with sociology in chapter 4, and he is apparently equally unconcerned about feminist analysis: 'I consider women to be the absence of desire. It is of little import whether or not that corresponds to real women. It is my conception of "femininity"' (Baudrillard and Lotringer 1987: 95). And what can and should one do with statements like ;'The drama of love is entirely in men, that of charm completely with women' (96). I know, of course, that one is meant to grasp these generalisations as poetic, but if one plays Baudrillard's game back and takes them literally we are left with something that really can be safely forgotten.

Baudrillard's interview also contains much insight into the dangers of taking his sort of radical 'coolness' to excess: 'The giddiness I'm talking about ended up taking hold of me...I felt I was going totally nuts...Dying doesn't do any good. You still have to disappear'. (Baudrillard and Lotringer 1987: 81--2).

Let Lotringer have the last word: 'I wonder if there isn't a kind of "skidding" endemic to theory. When theory manages to complete itself, following its own internal logic, that's when it disappears. Its accomplishment is its abolition' (127).

contents page