• Summary
  • Contents
  • Subject index

Organizational Culture and Identity discusses the literature concerned with culture in organizations and explains why the term has been invoked with such enthusiasm. Martin Parker presents further ways of thinking about organizations and culture which suggest that organizational cultures should be seen as `fragmented unities' in which members identify themselves as collective at some times and divided at others.

A Forgotten History of ‘Culture’
A forgotten history of ‘culture’

This chapter will develop my argument that a concern with culture in organizations is not a radically new development but is instead a synthesis of many central concerns in writing on organizations throughout the twentieth century. In general I want to argue that an understanding of the precursors of the culturalist literature can help to inform a rather different understanding of the relationship between organization and culture.

The chapter is therefore structured as a broadly chronological narrative beginning with Weber and Taylor and ending in the late 1970s before Peters et al. began to popularize their ideas. As I suggested in Chapter 1, it is also inevitably a catalogue of critiques of one-dimensional characterizations of bureaucracy and scientific management. By engaging in a debate with a caricature of Taylor and a narrow reading of Weber many writers enabled culture (or a similar term) to become both an explanatory variable and the basis for an anti-positivist epistemology. This is to say that much twentieth century writing on organization has rather a dualist character – a long running debate between what Gouldner has called the ‘rational’ and ‘natural system’ models (1965). Both ‘sides’ needed the other, and though the former was generally dominant (in the guise of managerial functionalism) there was a continuing attempt to articulate the latter prior to the emergence of modern culturalism. So, the general criteria for inclusion in the chapter is that the work should be in some way concerned with the non-structural features of organization – variously labelled ‘climate’, ‘atmosphere’, ‘personality’, ‘informal structure’ and so on. This strategy is intended to be suggestive and not in some way exhaustive. My aim is to chart what seems to me to be a forgotten history, and not worry too much about precise definitions.

Weber and Taylor

Organization theorists are often, whether they know it or not, debating with the ghost of Max Weber (1947; Gerth and Mills, 1948). In many ways he provides both ends of this chapter's speculative history of ‘culture’ within organizational analysis. Many managerial assessments of Weber reduce him to an organizational design consultant, but the retreat from positivism that acts as a direct precursor to organizational culturalism is also informed by his status as proponent of verstehen, of an attempt to understand the life-world of social actors (Silverman, 1970: 126–41). Beyond the narrow confines of an organizational behaviour textbook, the consensus now seems to be that Weber's formulation of bureaucracy should be properly seen as part of his historical thesis about the development of forms of rationality and legitimation. If it is not placed in this context then Weber is reduced to being no more than a precursor to Taylor, establishing organizational first principles which scientific management then distilled and applied to individual workers. As has been well established elsewhere (see forexample Mouzelis, 1967; Albrow, 1970; 1992) the bureaucratic ideal type is not a prescriptive category, but a heuristic device for gaining some purchase on some peculiarly modern forms of conduct. Many of the twentieth century debates about the supposed dysfunctions of bureaucracy stem from an interpretation of Weber as an example of grand theory in the ‘one best way’ tradition. Indeed, a diagnosis like March and Simon's is still all too common:

in general, Weber perceives bureaucracy as an adaptive device for using specialised skills, and he is not exceptionally attentive to the character of the human organism, (in Pugh, 1971: 30)

In a similar way, organizational behaviour texts often make little distinction between Weber and the ‘classical’ theory of Taylor, Urwick, Gantt, the Gilbreths and so on. All are framed as evidence which can be used to demonstrate the awfulness of anti-humanism and the amount of ‘progress’ that has been made in theories of work organization this century. This version of Weber is one in which roles should determine behaviour because any fluctuations could only interfere with the severe impartiality which is functional for the organization. Now, this is certainly a very selective reading of Weber, but it is one that has set much of the mythical backdrop for twentieth century theories of management. The other ‘straw man’ who is used to demonstrate something about how far organization theory has come since the turn of the century is Frederick W. Taylor.

Again, the accepted understanding of Taylor is that he espoused a rigid form of authoritarian behaviourism which relied on the assumption that all individuals were self-interested and motivated largely by economic reward. The only cultural elements of his work were the perturbations that he sought to eradicate – workers' ‘systematic soldiering’, management's reliance on rules of thumb and so on. Rose's stark comments on Taylor's attempt to reduce people to machines are well worth repeating.

The properly trained first class man is supposed to function as predictably as a piece of clockwork. He must be rested at appropriate moments. But to say this amounts to little more than conceding that a machine needs regular servicing.

Of course the speed at which a machine will run depends on the strength of the current or the octane of the fuel poured into it. Similarly, the worker can be encouraged to work harder by appealing to his materialism. The higher the money incentive, the greater his response.

Every other intluence upon his behaviour, social or psychological – the workgroup, a trade union, managerial ‘autocracy’ or whatever – is an unnatural interference that must be removed to allow optimal functioning. It is a bizarre conception. Taylor's worker is a monstrosity: a greedy machine indifferent to its own pain and loneliness once given the opportunity to maim and isolate itself. (1978: 62)

As with Weber, this understanding of Taylor's work provided a baseline for other authors to argue for a story of progress – the gradual humanization of the brutal face of industrial organization, partly through the heroic efforts of organizational theorists. Of course, as has been suggested by many other writers, neo-Marxists in particular, much of the supposed critique of scientific management was partial and many fundamental managerial assumptions remained unchallenged (Thompson and McHugh, 1995). I will return to this theme later in the chapter but for now I want to also briefly consider whether it is worth loosening the noose that so many have put around Taylor's neck.

In a very interesting essay, Lucy Taksa argues that aspects of the dominant characterizations of Taylor are actually rather flawed because structuralist (both functionalist and Marxist) organization theorists have simply not conceptualized Taylor carefully enough (1992; see also Jacques, 1996: 155). She suggests that he was centrally concerned with cultural issues – the harmonization of management and worker interests – and to argue otherwise is to mistake various scientific management techniques – like the Bedaux system for example – for his espoused rationale. Taylor's attempt to prevent the recurrent formation of shop-floor workgroups involved the replacement of orally transmitted counter-cultures by a single written organizational culture. Importantly, this was to be achieved through a division of agreed responsibilities and a consequent ‘mental revolution’ on the part of all employees – managers and workers – to ensure that the varied interests of all employees were reflected. Workers were to be encouraged to break with collective opposition from ‘dysfunctional’ counter-cultures, and become individual employees sharing common aims. This unitarist goal was to be ultimately achieved through selection, training and consensus management.

Reading Taylor in this way suggests that his attempt to engineer a sense of organizational harmony – even if it was based on a very narrow conception of interests – anticipates not only the Durkheimian strands of human relations theory but elements of the corporate culture literature too. Yet such a reading would profoundly disturb the chronology that is built into the narrative of the last 100 years. It seems that this dominant characterization of a Taylorist ‘other’ performs an important function for much of the supposedly humanized management of later in the century. I think this argument can be taken further if. for example, we also paid attention to Henri Fayol's interest in esprit de corps or Mary Parker Follett on participation in organizations. These were both contemporaries of Taylor whose work untidies the neat story of progress from ignorance to knowledge, from structure to culture. As Taksa suggests, pointing to these connections between Taylorism and culturalism begins to disturb the hidden assumptions of contemporary organizational culturalism and also illustrates just how difficult it is to disentangle the supposed newness of organizational culture from a long history of attempts at engineering organizational control strategies.

Human Relations

Nonetheless, the influence of these narrow readings of Weber and Taylor has been immense, and this is shown most clearly by the supposed contrast between scientific management and US ‘human relations’ theory which is now a canonical first move to enlightenment within the field known as organizational behaviour. There were, as Rose suggests, elements of a move towards quite a ‘social’ model of worker interests in some of the reports of the British industrial fatigue researchers (the ‘Myersians’) from the late 1920s onwards (1978: 75). Yet, despite the fact that the work of these early work psychologists was in many respects more rigorous than that of their American ‘human relations’ counterparts, the latter had far more influence on the shape of thought about organizations. As popularized by Elton Mayo, the human relations movement is usually said to have brought the social into the study of work – though perhaps it would be more accurate to say that these were really more sophisticated versions of how small groups might be managed. Though not internally coherent as a school, human relations work usually shares two general propositions which shape an understanding of the relation between people and their organizations. The first is that ‘informal’ patterns of interaction set up expectations and constraints that cannot be explained simply by reference to an organization chart or a desire for monetary reward. The second is that an employee's beliefs, attitudes and values are brought with them from non-work contexts and impinge upon the way they think about themselves and their organization. As Roethlisberger and Dickson put it:

Many of the actually existing patterns of human interaction have no representation in the formal organization at all, and others are inadequately represented by the formal organization. … Too often it is assumed that the organization of a company corresponds to a blueprint plan or organization chart. Actually, it never does. In the formal organization of most companies little explicit recognition is given to many social distinctions residing in the social organization, (in Merton et al., 1952:255)

Many of these ideas about the importance of the ‘informal’ appear to derive from Mayo's introduction to Pareto via his translator L. J. Henderson at Harvard. Both Talcott Parsons and Mayo were members of the ‘Pareto circle’ dining club at Harvard in the 1930s – as were Merton, Homans, Roethlisberger and Barnard (Burrell, 1996: 642) – and the influence of this group on twentieth century organization theory was substantial. Pareto's manifesto for the social engineering of ‘sentiments’ – the non-logical rationalizations for action – suggested that elites could manage better if they understood the irrationalities of ordinary human beings. As a result, for human relations theory, the concept of group values was almost exclusively restricted to the shop floor, not management, and attention was focused on small workgroups and not on larger collectivities within or without the organization. Of course this also echoes Taylor's highly normative conceptualization of the contented worker — any conflict with management being seen as a pathological deviation. The elitist paternalism which followed from this position essentially defines the ‘informal’ as being a property of ‘them’, and not of ‘us’ (Gouldner, 1965: 407). In terms of an early history of culturalism it is also interesting to note the influence of an imperial form of anthropology – another form of ‘us’ and ‘them’ – on the human relations movement at Harvard. One of its seminal figures was the Chicago structural anthropologist W. Lloyd Warner, a pupil of the British structural functionalist Alfred Radcliffe-Brown. By establishing close ties between the Chicago Anthropology Department and Harvard Business School he attempted to apply insights from his ethnographic studies with Australian Aborigines to an industrial setting, and the later phases of the Hawthorne studies were much influenced by this anthropological methodology (Trice and Beyer, 1993; Wright, 1994). Hence the bank wiring room study is conceptualized as if it were a society in miniature, and a functional society at that. The employee attempts to fit into the value system of this society and the task of the manager is to manipulate ‘them’ in order that they do not disrupt the logic of management.

One final point I want to make about the human relations movement is the central place that Elton Mayo's diagnosis of industrial civilization plays in the popularization of the analysis. He believed that traditional attachments to family and community were breaking down and giving way to a situation which again comes close to Durkheimian anomie. The potential disintegration of social ties and plunge into conflict could only be solved by a managerial elite who could give a new meaning to people's lives.1 To seek to restore to workers the satisfaction and meaning in their jobs was hence to regain a kind of lost solidarity. The social conscience of the enlightened entrepreneurial capitalist could hence be salved by realizing that their mission was to save society from collapse, and that this did not involve damaging their profits – a message later clearly echoed in culturalism. Whilst not all the writers who usually come under the human relations label shared Mayo's evangelism or skill at marketing, the search for community and social harmony is a theme which reappears in various guises later in the century. Pugh et al. describe human relations in general terms which could easily apply to Peters and Waterman:

the use of the insights of the social sciences to secure the commitment of individuals to the ends and activities of the organisation. (1971: 129)

In this sense managerial culturalism is centrally shaped by the elitist and romantic intentions of the human relations movement. Mayo, perhaps more than any other writer on organizations, can be seen as the precursor of In Search of Excellence. However, it does seem to me that reducing a concern with culture to a reinvention of human relations does miss at least one element of the distinctiveness of the former. As Thompson and McHugh point out, at least the culturalists recognize that management may be influenced by ‘non-logical’ factors too (1995: 50). Some of the later developments of human relations theory provide ample evidence of this, in both theory and practice.

Neo-Human Relations

First, it is worth mentioning two schools that further extended conceptualizations of the informal under the general banner of human relations – applied anthropology and Kurt Lewin's field theory. Conrad Arensberg and Eliot Chappie, again both at Harvard, founded the Society for Applied Anthropology in 1941. Through its journal Applied Anthropology they attempted to extend human relations beyond the factory gates. Importantly, in claiming this anthropological pedigree they also legitimated a lengthy immersion in the life of the organization or community under study – a move which began to break down the ‘them’ and ‘us’ divisions so constitutive of early human relations. The most well known result of this focus is contained in the work of W. F. Whyte, his classic Street Corner Society (1955) being followed by a series of studies of life in organizations (1948; 1961). Similar work was done within the ‘Yankee City’ studies supervised by W. Lloyd Warner. One of the series focused on The Social System of the Modern Factory (Warner and Low, 1947) and was a detailed description of a strike in a small community. These, and other works, helped to begin to make participant observation or ethnography respectable as a tool for the study of industrial societies, and many of the studies explicitly drew on the ideas of social anthropologists to justify their methodology (Warner and Low, 1947: 54; Whyte, 1955: 286).

Yet, despite this broad legitimation, these studies do not actually use many anthropological concepts. Terms like culture, myth, symbolism and so on are not applied with any frequency, consistency or reflexivity. When Warner and Low discuss their strike it is placed within a Durkheimian framework (1947: 54, 191) that stresses the fragmentation of consensus brought about by the division of labour. Similarly, Whyte's work, though extremely sensitively written, assumes in a touchingly naive way that careful observation will somehow reveal the reality of group and community systems. Many of his comments are illuminating and in places (1961: 34, 386, 433) he clearly anticipates the writings of Peters and Waterman as well as providing (I think) one of the earliest critiques of this kind of position (1961: 576). Nonetheless, his use of the term ‘symbol’, for example, is premised on the assumption that only certain organizational objects are symbols. The world is not symbolic, only parts of it, and the sensitive investigator must therefore attend carefully to those parts. That this is so stems largely from the prescriptive intent of many of these studies, an aspect of human relations that I have already mentioned, and which still tended to allow these researchers to see ‘culture’ as a problematic property of ‘them’, even if that group were now enlarged to include managers too.

A work which combines a similar attention to detail without using the term ‘culture’ comes from a student of Whyte's – Melville Dalton's Men Who Manage (1959). The book is an insider's description of four organizations, two of which Dalton worked in himself. Unlike Whyte, his study is primarily engaged with the post-war and post-Weberian writings on bureaucracy (see below). Like Whyte, it is work predicated on the assumption that ‘unofficial’ behaviour is an essential aspect of organization.

Organization is not seen as a chiseled entity, but as a shifting set of contained and ongoing counter phases of action. (1959: 4)

For Dalton the organization chart is only a ‘point of departure’ (1959: 17) and from there he goes on to investigate the power struggles and cliques that shape life within four organizations. The book contains a substantial amount of detail on social life, the informal code of dress, the use of the cafeteria, Masonic and ethnic ties, and the flexible use of formal and informal rules – such as the reward system (1959: 93, 94, 150, 180, 194). Dalton's formulation of the latter is worth quoting since it clearly presages elements of interactionist work which I will consider later in the chapter.

Those who regard this chapter as merely a set of episodes of theft have missed the point. Our study of unofficial rewards is not an attempt to justify internal plunder or to say that theft by membership is inevitable. Both ‘theft’ and ‘reward’ derive their meaning from the social context. (1959: 215)

Men Who Manage hence poses questions about the very use of the formal/ informal couplet. In doing so Dalton points towards both the theoretical inseparability of such a dualism and the practical ambiguity that managers face in their everyday lives. Indeed, he even suggests that conflict may be an endemic feature of bureaucracy, particularly in individualistic societies like the USA. Implicitly then he opens spaces in the very notion of organization itself as the where and the when of an organization's particularity become difficult to disentangle from its broader social context.

However, despite this laudable attempt at organizational ethnography, at points Dalton slips into making some rather predictable neo-human relations design prescriptions. For example, he suggests that some organizations rely too much on their informal or formal structures and do not effectively utilize their interrelation. It is difficult to see how he can argue for their inseparability and then go on to separate them when it comes to aparticular managerialist interest. In addition he occasionally seems to often do no more than rehearse a mass of local observations and anecdotes about cliques and compromises. Whilst many of his readers will certainly recognize the situations he writes about, he just places them in categories and leaves it at that. Dalton's neglect in the later culturalist literature is certainly surprising, but perhaps one explanation lies in his portrayal of managers as almost permanently rule breaking and self-interested. As Crozier perceptively observes:

He is so haunted by the fear of being misled by the formal structure and the formal definitions of roles, that, in his analysis of the way managers really behave, he reports only irregularities, back door deals and subtle blackmail. (1964: 149)

As Steven Feldman's essay on Dalton amply demonstrates (1996), such a jaundiced view of management is one that is unlikely to be approved of by consensus managerialists, so perhaps it is hardly surprising that this work is being increasingly ‘forgotten’.

The second development of human relations that is of importance here is more influenced by social psychology than anthropology. It is associated primarily with the work of Kurt Lewin and was largely an attempt to illustrate the importance of small group dynamics on leadership. The key significance of these studies for my purposes was the introduction of the term ‘climate’ to describe a characteristic of a group. As with most neo-human relations studies a managerial and prescriptive intent was central in determining their guiding rationale. The central tenet seemed to be that it should be possible to define a profile for the successful supervisor which will both give them coercive control and ensure employee co-operation – effectively a psycho-technology of management. The first mention of ‘social climate’ seems to be in a paper by Lewin, Lippitt and White from 1939 in which it is used interchangeably with the term ‘atmosphere’. The paper describes styles of leadership and draws conclusions about the utility of each with regard to aggressive behaviour. They conclude that aggression is highest in autocratic groups but acknowledge that culture or ‘style of living’ has an effect on these patterns. They appear to be using culture to refer to a society wide concept but there is sufficient ambiguity in their phrasing for the paragraph to be highly suggestive.

Whether or not a given amount of tension and given restraining forces will cause a person to become aggressive depends finally on the particular patterns of action which are customarily used in the culture in which he lives. The different styles of living can be viewed as ways in which a given problem is usually solved. A person living in a culture where a show of dominance is ‘the thing to do’ under certain conditions will hardly think of any other way in which the solution to the problem may be approached. Such social patterns are comparable to ‘habits’. (Lewin et al., reprinted in Pugh. 1971: 256)

The term ‘climate’ has since enjoyed some popularity in organizational psychology (Tagiuri and Litwin, 1968). Yet the development of this term during the 1960s and early 1970s was primarily shaped by Lewin's early work. Typically, responses to scaled Likert type questionnaires were used to categorize members' perceptions of autonomy, reward, warmth, constraint or whatever (Payne and Pugh, 1976: 1140). It then became possible to relate such measures to dimensions of structure and produce statistical correlation hypotheses about the functional ‘group personality’ for particular forms of organization. Whilst such work often looked highly sophisticated in attempting to link personality, climate and structure, its methodologies ran the usual danger of assuming that questionnaire response reflected practice. Change, conflict and ambiguity are simply not easily made visible on a questionnaire. The limitations of such an approach may be seen if we reflect upon Payne and Pugh's optimistic Comtean conclusion to their review of the area a quarter of a century ago.

The way forward is clear. We can benefit from the ability to design organization structures and climates which are appropriate to particular goals and needs.

Furthermore the planning and implementation of social changes will be more feasible with such knowledge. Thus, research which improves our understanding of organizational structure and climate will make an important contribution to our future. (1976: 1169)

Though ‘climate’ has since mutated into ‘culture’ as a factor in job design and job satisfaction (see for example Porter et al., 1975; Graves, 1986; Pheysey, 1993) the quantitative emphasis remains. This is much less true of other works within the rather more influential tradition of humanistic organizational psychology.

Leaders and Groups

A work which stands in a curious relation to writing on organizations before or since was written by another member of the Pareto circle – Chester Barnard's The Functions of the Executive (1966 [1938]). Its relevance for organizational culture is substantial, though it is rarely cited for this purpose, being usually viewed as a work on leadership. Like Dalton, Barnard wishes to look at ‘organization’ as both noun and verb, a synthesis of formal and informal. In defining organization as ‘a system of cooperative activities’ (1966: 75) he clearly stresses its processual and relational character. Management is a question of building common purpose or an ‘organization personality’ (1966: 88) which can be recognized by someone who has ‘observational feeling’. He disavows the reductionism of the organization chart in preference to ‘learning the organization ropes’ (1966: 121). Barnard's executive is a manipulator of co-operation and uses non-material means as much as economic incentives. By manipulating the ‘hierarchy of positions with gradation of honors and privileges’ (1966: 170) the able manager can ensure that the sentiments of organizational participants are favourable to their task at hand.

Barnard shows the influence of human relations in his writings but he also prefigures much that later characterized the debates on bureaucracy (see for example Barnard, 1952). For the purposes of this chapter however one of his major influences is in the status given to the executive. Management for Barnard is akin to statesmanship – ‘a matter of art rather than science, more aesthetic than logical’ (1966: 235). In passages that could easily come from a culturalist text he writes of the creativity and personal conviction that build strong organization codes. The quality of the leader makes myths that capture the hearts of employees. Barnard does not use the term ‘culture’, but he effectively formulates managers as culture heroes and shapers of destinies. His analysis swings between heroic myth making and Machiavellian practicality and both have been highly suggestive to writers who wish to privilege the managerial role in organization.

Another significant group of writers in terms of prefiguring ideas about the connections between culture and leadership is the humanistic psychologists who broadly built upon Abraham Maslow's now classic – or clichéd – formulation of a ‘hierarchy of needs’. McGregor, Likert, Bennis and Argyris all develop the notion that organizations need to be designed to satisfy human needs and not to thwart them. ‘Needs’ are here usually formulated as individual desires for esteem, involvement and so on but this potential essentialism is often tempered with an acknowledgement of the importance of organizational context in structuring such desires. Thus Douglas McGregor (1960) argues that managers can encourage either dependency (theory X) or autonomy (theory Y) depending on how they expect their employees to behave. Rensis Likert (1961) suggests that participation and communication ensure that the social systems of the workplace become tightly knit and productive. Warren Bennis (1966) argues that the ‘culture’ (by which he seems to mean group norms) of ‘target systems’ (groups to be changed) must be addressed as part of the ‘lab training’ or ‘T group’ change process. Finally, Chris Argyris (1957) proposes that many organizations treat their employees as if they were children, incapable of autonomous responsible action. If managers instead assume that direct supervision is not necessary, if they treat their employees as adults, then workers will respond to the increased responsibility. Clearly these are all again prescriptions for satisfying workers and managers simultaneously but they re frame elements of the early human relations studies by moving the focus of attention from the social structure of the workgroup to more interactive formulations of the relationship between social identities. The development of this work also seems to be substantially shaped by the idea that organizations could easily become theatres of oppression and cruelty – perhaps a disenchantment with modernity that had rather a lot to do with the extreme illustration that the Holocaust had provided only a few years before.

Whilst Likert's, Bennis's and McGregor's work is primarily within the managerial leadership tradition, Argyris most clearly exemplifies a more sophisticated conception of the dynamic relationship between the individual and the social. He sees them as mutually constitutive but also as potentially (and often in actuality) disruptive of each other. However, despite this avowed interactionism he still relies on a romantic vision of the potentially self-actualizing employee being repressed by an authoritarian bureaucracy. This effectively leads to the reduction of the social to a set of adaptive individual strategies – quite similar to the psychologistic accounts of climate I covered earlier. Hence, for example, his suggestion that child-like individuals may be suited to formalized structures – a thesis that comes too close to Taylor's assumptions about the ‘mentally sluggish’ (and culturally alien) Schmidt for comfort. What is valuable about Argyris for the purposes of this analysis is the space between ‘personality’ and organization that he opens and into which an analysis of culture could be inserted. Indeed, he makes some suggestive comments of his own to this end:

Perhaps social rankings and their constituent status symbols function to facilitate the self-actualization of the social organization. Why is it not possible to hypothesize that the agents of social organization create and then use such status symbols as desks, rugs, chairs, telephones, decorating and size of room to help the formal organization achieve its goals, maintain itself internally, and adapt to its external environment? … Why the agents of any particular system (e.g.: the formal or informal) decide to pick a particular set of symbols to denote status may be related to their personalities, to the culture within which they exist, and to the particular situation being observed. Research into these aspects would lead to insight into how the individual, organization and culture interact and transact to maintain themselves and each other. (1957: 244–5)

Developing in tandem with this psychological humanism was the organizational development work of Elliot Jaques which is still quite remarkable for its anthropological subtlety. In The Changing Culture of a Factory (1951) he reports on a period of action research when he was a consultant at Glacier Metals under its Chairperson and Managing Director Wilfred Brown. In this work Jaques applied his psychoanalytic training to a process of ‘working-out’ the stresses and tensions produced by the organization of factory life. The title for a 1948 pre-Glacier paper aptly summarizes his approach – ‘Interpretive Group Discussion as a Method of Facilitating Social Change’. Whilst this earlier work was firmly located within the Lewinian development of human relations he begins to go further with his participant observation methodology and a theorization of the relationship between culture, social structure and personality. ‘Social structure’ is defined as a fairly stable network of externalized roles and ‘personality’ as the (often unconscious) psychological makeup of the individual. The mediating term ‘culture’ is described in the following passage:

The culture of the factory is its customary and traditional way of thinking and of doing things, which is shared to a greater or lesser extent by all its members, and which new members must learn, and at least partially accept, in order to be accepted into service in the firm. Culture in this sense covers a wide range of behaviour: the methods of production; job skills and technical knowledge; attitudes towards discipline and punishment; the customs and habits of managerial behaviour; the objectives of the concern; its way of doing business; the methods of payment; the values placed on different types of work; beliefs in democratic living and joint consultation; and the less conscious conventions and taboos. Culture is part of second nature to those who have been with the firm for a long time. Ignorance of culture marks out the newcomers, while maladjusted members are recognised as those who reject or are otherwise unable to use the culture of the firm. … The culture of the factory consists of the means or techniques which lie at the disposal of the individual for handling his relationships, and upon which he depends for making his way among, and with, other members and groups. (1951:251)

Jaques's definition of culture is valuable and interesting but, like the industrial anthropology studies, it is sadly conspicuous by its neglect throughout most of the text. Despite providing the framework for an analysis, and also relying on an ethnographic methodology, the key dualism is still again the relation between individual and organization. Culture thus becomes a term that merely fills a gap. covering the ideational aspects of local social structure and the internalized group elements of personality. Jaques does occasionally hint that his factory as a social microcosm approach may need qualifying, such as when he suggests that management-worker splits may be related to a more general form of conflict (1951: 295), but his approach is broadly reflective of his adopted role as a ‘social analyst’, that is to say a psychoanalyst of collectives. In general then, most of this leadership and groups literature adopts a fairly local definition of the cultural which largely reflects a psychologistic neo-human relations agenda. A rather different treatment can be found in post-war work which located itself more within sociology, and hence began to deal more explicitly with the work of Max Weber.

The Retreat from Bureaucracy

Running in parallel with post-war humanistic work psychology is a large body of sociological work on informal structure within bureaucracy. The common thread that links much of this literature is an attempt to shift the level of analysis from the organization chart to its unanticipated consequences, dysfunctions, or informal structure. Many of these points had already been anticipated by Weber (for example in terms of his argument that bureaucrats tend to overextend their functions: Mouzelis, 1967: 21–3) but his ideal type formulation does certainly overstress the stability and formality of life in a bureaucracy if it is assumed to be an empirical description, which is an arguable point in itself. Albrow suggests that there are essentially two strands to this debate with Weber.

The first is an account of the empirical validity (both historical and predictive) of his account of the nature and development of modern administration. The second … is a rejection of his association of the ideal type of bureaucracy with the concepts of rationality and efficiency. (1970: 61–2)

This is certainly a useful way of approaching a wide body of writings but for the purposes of this text it is essential to note that both strands prefigure organizational culturalism. The culture thesis is an approach which claims both more explanatory purchase and to be a method of organizational design which is more efficient.

One of the writers who was influential in initiating this critique of bureaucracy was George Homans (1950). Though his work is perhaps best seen as a development of Chappie, Arensberg and Whyte, it does make a clear connection between those writers and the bureaucracy literature. Homans makes a Parsonian distinction between the external and internal aspects of the ‘organizational system’. The former refers to specified behavioural requirements that allow the group to survive in its environment, the latter to the expressive social relationships that reflect the sentiments – note the use of Pareto's term – that individuals have towards one another.

Remember that Homans was also a member of the Pareto circle, as was Robert Merton, the author of probably the most influential early contribution to this literature with his ‘Bureaucratic Structure and Personality’ (in Merton et al., 1952). With the ghost of the Holocaust again in mind, Merton suggested that individual bureaucrats may operate in ways that are dysfunctional for the goals of the organization – they may become obsessed with ‘red tape’. In a way that now defines the negative use of the term ‘bureaucracy’, instead of serving their clients bureaucrats turn means into ends and interpret criticism (from inside or outside the organization) as a need for more formalistic behaviour. Merton concludes that bureaucracy is hence not necessarily the most efficient mode of administration and, more importantly for the purposes of this chapter, that formal specifications do not determine actual behaviour. There was much literature elaborating on this point (see for example Merton et al., 1952), but I will restrict this review to four of the best known writers – Gouldner (1954), Blau (1962), Selznick (1957) and Crozier (1964). The main line of argument within all these texts is an empirical demonstration that bureaucracies do not (and should not) operate in the way that they felt Weber had outlined.

Alvin Gouldner was a student of Merton's, and though he uses functionalist language his treatment of bureaucracy is a highly processual and dynamic one which contains clear indications of his later move to become one of the major critics of the Parsonian orthodoxy (Warwick, 1974: 76–7). His description of the imposition of bureaucratic rules is particularly informative in this respect. In arguing that rules prescribe minimum standards of acceptable behaviour he notes that they may also encourage the bureaucrat to accomplish the minimum amount of work possible. Conformity can be reluctant since the rules can be viewed as alien and imposed. The consequent detrimental effect on productivity brings calls from management for tighter rules, and a spiral of counter-attempts to gain control over work is set in motion. In terms of an anthropological interest it is also worth mentioning Gouldner's frequent use of the term ‘myth’ and his interest in the belief systems, rituals and folklore of the workplace (1954: 79, 117 passim). Strip Gouldner's work of its Mertonian gloss and it becomes an excellent piece of industrial ethnography. His writing is certainly set within the notion of a system which is wavering around some kind of equilibrium point with uncertainty as the main destabilizing force, but there is an evident tension between this and his recognition of the often conflicting interests of particular groups and individuals within an organization. In a way that also explicitly challenges the functionalism of the time, he suggests that students of administration have often reified organizations but that only people can be said to have ‘ends’ in any meaningful sense (1954: 21).

In policy terms Gouldner does offer some kind of humanistic solution to bureaucratic dysfunctions in his formulation of ‘representative’ bureaucracy as opposed to the ‘punishment-centred’ bureaucracy. Individuals in the former structure see rules – such as safety guidelines – as collective agreements that are supported by all members. Of course here it needs to be noted that the use of the term ‘bureaucracy’ is effectively becoming synonymous with ‘organization’ itself. Gouldner's representative bureaucracy is not a bureaucracy at all in many theorists' understanding of the term – indeed it is perhaps closer to Peters and Waterman's strong culture. Yet, in terms of his analysis, if the same formal rules can have many different informal functions (explicate obligations, allow control at a distance, provide for impersonal authority, legitimate punishment or preserve apathy: 1954: 237) then they must be formulated as socially constructed. The possibility that a statement could be understood by an individual in a wide variety of ways according to (perceived) context opens a space for the prioritization of interpretation which prefigures the turn to the ‘action frame of reference’ in the 1970s. Indeed, Gouldner's later work on manifest and latent social roles expands these ideas by suggesting that organizational actors may have value and reference group orientations that shape action yet are not formally prescribed by the organization's explicit rules, and indeed might lie ‘outside’ the organization (1957). Echoing Dalton, people are not just members of an organization, they also have gendered, ethnic and religious identities (1957: 285) as well as professional allegiances that cross-cut each other. The usefulness of this recognition of multiple fractures and allegiances within and without organizations will become apparent later in the book. It was also rapidly recognized by Becker et al. (1960; 1961) who suggested translating the term ‘role’ into the term ‘culture’ – a move I will cover later in the chapter.

Peter Blau was another disciple of Merton whose participant-observer work in various organizations convinced him that they did not work as the organization chart suggested, but rather that rules were continually re-worked informally. In this light he suggested that the informal was needed to operationalize the formal and, more contentiously, that efficiency is actually often a result of turning a blind eye to official pronouncements. This becomes increasingly important when bureaucrats face tasks that are not easily routinized. In this regard he mentions language (1962: 106), ritual (166) and myth (194) but within a framework that is more directly influenced by Arensberg's version of structural functionalist anthropology. Blau's notion of structure is certainly flexible but he sees the system as prior to (and generative of) cultural elements. Compared to Gouldner the ‘cultural’ aspects of his analysis are less developed and more difficult to disentangle from his functionalism, partly because his work is simply less ethnographically textured.

In policy terms, Blau did believe that by initiating appropriate recruitment and training procedures members of organizations would be able to exercise appropriate autonomy and prevent the need for continual rule rewriting because of the existence of cultural consensus. The key factor is whether an individual identifies with the goals of the organization. Select the correct people, give them due responsibility and efficient administration will result. The alternative is to force individuals to ignore and adapt, in which case any rules become self-evidently useless. In this respect his suggestions are far more detailed than Gouldner's. He attempts to give prescriptive shape to his model of bureaucracy and, once again, some of his suggestions would not look out of place in a 1980s piece of organizational culturalism.

Philip Selznick took functionalism even more seriously than Blau, but he also conjoins this with many observations about the role of elites in shaping the ‘character’ of organizations. His central problem is the tension between organizational decentralization and the maintenance of agreement on goals – the difficulty that results in the frustration of both in his Tennessee Valley Authority case study (1966 [1949]). According to Selznick, when an organization is forced to decentralize in order to prevent the inertia of large bureaucratic structure, there is a tendency for the sub-units to focus on fulfilling their sub-goals at the expense of the major goals of the organization. Echoing Merton, the means become the ends, and echoing Gouldner, this leads to calls for more centralized control, which in turn leads to the need for decentralization and so on. This paradox is seen as part of organizational evolution and solving it requires the executive to become an institutional leader, not merely an administrator (1957:4).

Particularly important here is the term ‘institution’, which refers to the organization that has grown a character and hence become more than a rational bureaucracy.

In perhaps what is its most significant meaning, ‘to institutionalize’ is to infuse with value beyond the technical requirements of the task at hand. … From the standpoint of the committed person, the organization is changed from an expendable tool into a valued source of personal satisfaction. (1957: 17)

Through the idea of institutionalization Selznick's work produces observations about the ‘character’ of an organization which are undoubtedly indebted to Barnard's formulation of organizational personality and which have been highly influential in the US under the guise of ‘institutional theory’. Character is the ideational aspect of structure – in Parsonian terms the distinctive ‘central value system’ that the organism has adaptively produced. Yet ‘social evolution’ cannot always be relied on because organizational character formation can also be pathological if it is not given purpose or mission by good leadership which is capable of ‘transforming a neutral body of men into a committed polity’ (1957: 9). Selznick's leaders are capable of diagnosing the growth of their organization and assessing the strengths or inadequacies that their organization has produced (1957: 42). Within the organization there are thus roles, patterns of stratification, interest groups and beliefs that can be adjusted to engineer success. His discussion of myth is instructive in this respect.

To create an institution we rely on many techniques for infusing day-to-day behaviour with long-run meaning and purpose. One of the most important of these techniques is the elaboration of socially integrating myths. These are efforts to state, in the language of uplift and idealism, what is distinctive about the aims and methods of the enterprise. … The assignment of a high value to certain activities will itself help to create a myth, especially if buttressed by occasional explicit statements. (1957: 151)

Selznick's work can be seen as an exemplar of a high functionalist and managerialist approach to organizational culture. This text is essentially a guide for turning managers into statesmen (1957: 4) and allowing them to more effectively guide their enterprise. Despite claiming to be taking analysis ‘beyond efficiency’ and ‘beyond organization’ (1957: 134, 137) what he really achieves is to open a distinction between an organization (a bureaucratic structure) and an institution and hence yet again to critique the idea that bureaucracy necessarily equals efficiency. I will consider the US development of institutional theory in a while, but it is enough to say of Selznick that like many functionalists he becomes a victim of his own static metaphors. Envisaging organizations as personalities or organisms tends to marginalize conflict to an internal pathology which can be cured if the correct methods are applied. The organization is effectively reified and treated as a bounded system yet certain actors (the ‘leaders’) seem capable of escaping this constraint and manipulating the rules that others appear bound by. Nonetheless, despite these problems, Selznick does provide an explicit and elegant formulation of some of the US antecedents of organizational culturalism.

Finally, to turn to the work of Michel Crozier – the only non-US writer in this section and one who was much less influenced by ‘closed systems’ versions of functionalism. Crozier attempted to address the link between structural and psychological modes of explanation and hence to account for the problem of change (1964). He chooses to study organizations because he sees them as sites in which the central values of the wider culture are deployed. His particular interest is in discovering why bureaucracy, which he uses in the post-war sense of organizational dysfunction or maladaptation, is so prevalent in France. As with Gouldner, Blau and Selznick he relies heavily on participant observation and consequently provides substantial detail on two French organizations. Extrapolating from this evidence he then argues that the French desire for individuality and independence ensures that there is continual debate around uncertain situations. Organizational actors' power depends on their capacity to maintain a degree of uncertainty about exactly how they act. Bureaucratization is the process whereby this uncertainty is progressively restricted to higher levels in the organization whilst those lower down insulate themselves with rules and routine. Crozier regards this state of affairs as pathological and requiring a change in both French organizations and culture.

His diagnosis of the central problems of bureaucracy echoes much of the previous writing and I won't repeat it here. However, his use of the term ‘subculture’ (1964: 68, 80) and his recognition of chronic conflict within organizations show how thin functionalist arguments were beginning to wear by the time he was writing – even if they had to be dressed up with some rather elaborate language.

Instead of describing bureaucratic dysfunctions merely as the automatic consequence of the ordering of human and technical factors necessary for achieving a superior form of rationality, we have tried to understand them as the elements of a more complex equilibria affecting the patterns of action, the power relationships, and the basic personality traits characteristic of the cultural and the institutional systems of a given society. (1964: 294)

What is particularly important about Crozier is the articulation of his ethnography within a contextual description of the culture of a particular nation. If French organizations are to change then French society must change with them. A new form of rationality requires more than just organizational design; it needs a re-orientation of broader social assumptions too. Whilst Crozier's diagnosis is still open to debate, his widening of the boundaries of organizational analysis echoes Dalton, and I will be developing some of its implications later in the book.

So all these sociologists were empirically exploring the nature of bureaucratic rationality, though in a way that was still employing Weber rather narrowly. Albrow's comment that Merton was turning back to a pre-Weberian view which equated bureaucracy with inefficiency can be applied with some justification to most of the other authors I have covered in this section (1970: 55). However, placing Weber's writings on bureaucracy in the context of his views about rationality suggests some different lessons that were implicit in the retreat from bureaucracy writing. After all, if bureaucracy is not the highest point of modern administration then it might follow that scientific rationality may also be rather a one-dimensional instrument for conceptualizing organization. In other words, we might want to distinguish between different forms of rationality and hence acknowledge that conduct in organizations may not always be predicated on a dominant form of totalitarian instrumentality.2 It seems apposite to conclude this section with a quote from Weber himself. With characteristic rhetorical overstatement he suggested that the advantages of modern administration were such that it would be inconceivable

to think for a moment that continuous administration can be carried out in any field except by means of officials working in offices … The choice is between bureaucracy and dilettantism in the field of administration, (in Beetham, 1987:59)

Debates about the meaning of the term ‘bureaucracy’ left to one side, the key point of all the above writers is that organizations might find it difficult to operate at all if it were not for some form of dilettantism.

The ‘Eclipse’ of Culture: Systems and Contingencies

The influence of Parsonian functionalism in US organization studies from the 1950s onwards is evident in the formulations of social system assumed by many of the authors in the previous sections. However, a further development of this influence was the increasing attempt to theorize the links between ‘social systems’ and other supposed systems – most commonly the technological and the economic. Just as Parsons moved from the quasi-Weberian ‘unit act’ to develop more and more complex versions of systems and structure (1951), so did some theories of organization begin to subordinate the ‘social’ to other forms of suggested causation.

Robert Blauner's classic Alienation and Freedom (1964) is a good example of this kind of intellectual turn. Though it largely relies on survey data, it is also clearly influenced by the industrial anthropology tradition as were several other previous studies intended to gauge the ‘effects’ of technology on human relations (Whyte, 1948; Walker and Guest, 1952; Chinoy, 1955; Sayles, 1958). Blauner's general argument is that:

technology, more than any other factor, determines the nature of the job tasks performed by blue collar employees and has an important effect on a number of aspects of alienation. … Since technological considerations often determine the size of an industrial plant, they markedly influence the social atmosphere and degree of cohesion among the work force. (1964: 8)

As Eldridge observes, this was an assertion based on dubious evidence and suspect causality (1973: 188–93). Yet at various places in his text Blauner did recognize the importance of what he called the ‘social organization’ of a plant, the extent to which ‘custom and past practice’ and ‘folkways’ are related to work organization. Whilst this is clearly a neglected feature of his analysis, a comparison of his case studies of the textile and car industries indicates the extent to which community, kinship and religious factors allow the workforce to understand their technological experiences in a way that makes them more acceptable and less alienating. Blauner argues that because mill workers have low aspirations (and a large number of them are women) the ‘objective’ controls of the technology are less onerous because:

the interlocking of work and life in an integrated industrial community probably makes the job of a textile worker meaningful, if not necessarily highly gratifying and rewarding. (1964: 88)

Blauner goes on to argue that the car industry had ‘objectively’ similar technologies in terms of its potential for alienation, but no particular sense of community or group membership and hence a more completely alienated worker (1964: 113–14). It seems paradoxical then that he concludes that it is technology per se that substantially determines worker satisfaction since my re-reading would suggest that it is the cultural context of technology that might be formulated as the rather more convincing variable.

Yet even this form of industrial anthropology was more completely eclipsed with the development of the socio-technical or Tavistock approach (Trist et al., 1963; Rice, 1963; Miller and Rice, 1967; Lawrence and Lorsch, 1967) which firmly placed the social into an interlocking systems framework. Trist and Bamforth in particular had worked on Jaques's and Brown's Glacier project but in an early paper (1951) had begun to move towards a much more elaborate formulation of the possibilities for social engineering. This was an approach that treated the socio-cultural parts of the organization as being, at least, causally related to external pressures and, at most, determined by them. Since these pressures were variable but measurable a ‘best fit’ between social, technological and economic systems could be found for any particular context.

It seems that this is the point in my history where formulations of culture or similar concepts are most eclipsed. The Tavistock group's adoption of a systems metaphor and quasi-scientific methodology placed clear limitations on the nature of the social that could be operationalized within the research programme. Indeed, the major concern was with job design and patterns of supervision as methods for ensuring the success of the organization by sub-optimizing its various systems. Acceptance or rejection of technical change, for example, was seen as a behaviour pattern which could be managerially modified – or, as Miller and Rice put it, optimized by balancing the ‘task’ systems and the ‘sentient’ systems (1967). Such a formulation of the social left almost no space for considering how meaning could be constructed on an everyday basis, or for seeing conflict as anything but a temporary pathology. Instead various organic, mechanical or cybernetic analogies were adopted to claim a grand view of the extent to which an organization was achieving its ‘goals’ – as defined in a highly abstract sense. The UK Aston studies (Pugh and Hickson, 1976) were probably the clearest example of the flowering of this research programme. The informal organization all but disappeared underneath a deluge of contingently related variables intended to measure the relation between environment, technology, organization, group and individual. As Pugh et al. argued, their approach was concerned with

what is officially expected should be done and what in practice is allowed to be done; it does not include what actually is done, that is what really happens in the sense of behaviour beyond that instituted in formal organizational forms, (cited in Reed, 1992: 137)

Though the possible existence of climate or culture was hence acknowledged it was hypothesized as a dependent variable, or simply ignored. The systems and contingency theorists simply ruled certain ideas untestable or irrelevant – again rather like Pareto's sentiments – in favour of the analysis of what were defined as verifiable scientific facts.

That being said, the contingency approach did not hold with deterministic ‘one best way’ assumptions in favour of a multi-causal ‘best fit’. Further, some of the work influenced by these ideas also demonstrated a laudable attempt to define its terms of engagement with considerable clarity (Burns and Stalker, 1961; Woodward, 1965). This clarity was part of its attraction, and no doubt influenced its subsequent dissemination, but also caused this writing to be insensitive to organizational practices that did not easily fit with the methodological principles. For example, though the structural ramifications of the term ‘bureaucracy’ were being defined with precision it was largely being assumed that all organizational actors were behaving (and responding to questionnaires) in a narrowly rational fashion. Yet again, Weber's suggestion that formal rationality was one meaning framework amongst others was ignored, zweckrational becomes the inclusive term and wertrational action is either sidelined or presented as a pathology (Weber, 1947: 184–6). The dominance of the systems metaphor solidified organizations, and hence overstated their conceptual, political and temporal stability (Grint, 1991: 138–39).

A rather different, and largely US, formulation of fünctionalism was developed later as institutional theory. It owed more to the early Parsonian notions of action systems, combined with Merton's stress on the unintended consequences of organizational structure and Selznick's definition of an institution as a structure which is infused with meaning. Though it initially shared many of the assumptions of systems theory, it began to show how social constructionism and fünctionalism could converge in a way that was very influential on much later US writing on organizational culture. Institutionalist authors emphasized that bureaucratic structures were not efficient in some abstract sense but were ceremonial practices which legitimated certain ways of doing organization (Meyer and Rowan, 1981; Scott, 1987).3 Structures were hence not representative of what actually happened within the organization but were ‘isomorphic’ responses to pressures in the institution's environment to do things in particular ways (DiMaggio and Powell, 1983). In a way that finally began to re-evaluate a long neglected Weberian legacy, bureaucracy was formulated as a modernist rationalizing myth. In some of this later writing, particularly that influenced by Berger and Luckmann's chapter on ‘institutionalization’ in The Social Construction of Reality (1967: 65 passim), organizations are acknowledged to both reflect and create socially constructed realities and rationalities. Whilst this approach clearly begins to move away from managerialist definitions of efficiency and dualist models of organization and environment, little of the early research moved beyond abstractions about meaning systems. There were some valuable points made about sectoral similarities, which I will return to in later chapters, but little detail on everyday practices within organizations. Effectively a concept like culture is articulated as an outcome of structural relations, and not a local or everyday concept. In a sense, institutional theory is Mertonian middle range theory par excellence but it does, however, clearly contain the recognition that meaning must be taken to be central in the study of organizations (see Clegg, 1994), a theme I take up in more detail in the next section.

The ‘Return’ to Social Action

The extent to which ‘natural system’ models had been sidelined can be overstated but it is fair to say that the industrial anthropology tradition was marginal to much of the US sociology of organizations by the early 1960s. For example, Etzioni's textbook Modern Organizations (1964) contains a few references to the ‘problem’ of the informal organization but concentrates almost entirely upon developing a systems approach. However, by the late 1960s Parsonian grand theory was being attacked from a variety of directions and the impact of a turn away from Gouldner's ‘rational model’ in the study of organizations is clear (1965). In many of these writings Weber is finally reread as a theorist of verstehen and not the proponent of formal organizational prescriptions. As the institutionalists were pointing out, if bureaucracy is conceptualized as a socially legitimated order then many of the debates about ‘efficiency’ simply miss the point. Organizations have both formal and informal orders, both structures and cultures, in so far as organizational members believe that they do or act and talk in a way that suggests they do. Indeed Watson notes that Weber himself was involved in factory studies in the 1900s which, though unfirjished, show an interest in the ‘ethical, social and cultural background, the tradition and circumstances of the worker’ (Weber in Watson, 1995: 67). This reiterates the point I made at the start of this chapter – that Weber provides both the beginning and the end of this speculative history of ‘culture’ in twentieth century organizational analysis.

In the USA the writing and teaching of Everett Hughes (1958), Melville Dalton (1959) and Donald Roy (1960) form something of a bridge between the early work of Whyte, Warner et al. and the later development of this theme in the interactionism of Becker, Strauss and Goffman. Meanwhile in the UK, as Richard Brown points out, Parsonian notions of consensus never took as strong a hold as they did on the other side of the Atlantic. Brown suggests that this was partly due to the later development of sociology as a recognized discipline and the importance of trade unions in reflecting class and party divides (1992: 12). As a result post-war British industrial sociology had tended to be more oriented to Weberian ideas of interest group conflict (Scott et al., 1956; Dennis et al., 1956; Lockwood, 1958), with ‘unitarist’ models of a common industrial interest being treated with considerable scepticism (see Bate, 1994: 62). As noted above, even British contingency works like Burns and Stalker (1961), though adopting functionalist assumptions, did attempt to wed accounts of ‘action systems’ with notions of structural constraint. Nonetheless, most of this industrial sociology work had only local impact, partly perhaps because of the conceptual distinction between industrial and organizational sociology that existed in the UK. It was only when the debates about the relationship between action and structure reached the centre of social theory (Dawe, 1970) that some of these ideas began to provide legitimation for challenging the partial eclipse of the informal organization.

It was North American symbolic interactionists who first developed a more explicitly interpretive stance. Anselm Strauss's concept of ‘negotiated order’ (Strauss et al., 1963) gained wide circulation as a way to express how order, an apparently structural property of organizations, is continually remade in everyday interaction. Aspects of organization that were conventionally articulated as ‘real’ entities were hence reframed as complex and continually changing practices and local knowledges. In a sense this resulted in the industrial anthropology tradition being taken to its localizing limits as the language of everyday common sense and sense making began to challenge the Parsonian consensus. For example, Goffman's (1968 [1961]) and Becker et al.'s (1961) studies of organizations focused on the way in which understandings of self and other were constituted through organizational processes such as positive and negative labelling. Attention to everyday detail provides the methodological manifesto and there is a recognition that sense making is often contested, as organizational members defend alternative understandings of identity against the organization's attempt to reduce them to one-dimensional role players. Everyday life becomes a drama, a stage managed presentation with the continual danger of confusion lying beneath the surface. The dualism between individual psychologies and social structures is reframed as a set of flexible scripts that are played through by actors – and all, to a greater or lesser extent, are quite capable of modifying the rules for their own purposes but also being written by those rules at the same time. Indeed Becker and Geer (1960) could be claimed to be the first to explicitly use the term ‘organizational culture’ in their discussion of manifest and latent cultures which drew on Gouldner's initial suggestions about cross-cutting roles. Clothes, spaces, symbols, games, roles and rituals are seen to be deployed and arranged in complex constellations that can only be understand through participant observation, not armchair theorizing.

It … is my belief that any group of persons – prisoners, primitives, pilots, or patients – develop a life of their own that becomes meaningful, reasonable and normal once you get close to it, and that a good way to learn about any of these worlds is to submit oneself in the company of the members to the daily round of petty contingencies to which they are subject. (Goffman, 1968: 7)

Meanwhile, in the UK, David Silverman (1970; see Hassard and Parker, 1994) was pivotal in insisting on a theory of organizations that began with action, citing Weber, Schutz, and Berger and Luckmann (1967) as influences. Asserting the specificity of the human sciences he eschewed any approach that took social laws to exist independent of actors' understandings of the world. His approach echoed, and was influenced by, many themes in the US Carnegie school of decision making which had began to construct an interpretive social psychology of organization and provided an important bridge between US and UK writing (Clegg, 1994). Authors such as Simon, March, Cyert and later Weick asserted that assumptions about rationality and attempts to find ‘one best way’ simply overstated the extent to which actors were actually able to identify problems and make decisions. In parallel with the more structuralist writings of institutional theorists, the Carnegie school argued that actors justify their views of the world post hoc by using ideas about rationality and taking into account whatever they felt was important – the ‘bounded rationality’ or ‘enactment’ model of decision making (Simon, 1957; Cyert and March, 1963; Weick, 1979).

Though beginning with some rather different methodological assumptions, ethnomethodology also provided a rallying point for organizational sociologists who were sceptical about the structuralist assumptions of systems theory (Bittner, 1974; Clegg, 1975; Silverman and Jones, 1976; Garfinkel, 1986). For Bittner, for example, ‘organization’ becomes a common sense concept carried around in people's heads. To talk of organization as a noun hence ignores the extent to which organization is a verb, a practical accomplishment. Members use the concept as a resource to make reflexive and retrospective sense of their experiences. At the extreme, every conversational utterance needs to be carefully analysed as an example of accounting and typification practices, and hence generalizations about what exists ‘outside’ or ‘beyond’ the data are rigidly proscribed. For purist conversation analysts, Goffman's, Becker's or Strauss's attempt to empathize with the actors and describe their situation would be an improvement on Parsons, but not nearly radical enough in its humanist assumptions about the role of agency and reflexivity. However, like the interactionists, the ethnomethodologists were often subject to the accusation that they were overstressing the micro-social. They were failing to link the everyday life of the organization to wider contexts of, and contests over, power; and hence missing the woods for the trees (Reed, 1992: 150). The fact that Goffman and others tended to focus on total institutions or institutions intended to socialize members into detailed role compliance – prisons, hospitals, education organizations – was a further symptom of this intra-organizational focus, caused perhaps by a rather extreme reaction to grand theory. After all, Cicourel (1976) and Bittner (1973) attempted to describe the organization of the criminal justice system without much reference to structures of power and authority, such as the state. Of course, such criticisms can be neatly parried with the argument that terms like ‘power’ and ‘authority’ are only meaningful in the context of everyday practices (Clegg, 1975), a question I will return to later in this book.

Within more managerial versions of UK organization studies, some recognition of the importance of actor definitions began to drive later contingency theorists to question the degree of consensus within the systems they were attempting to describe. Hickson et al.'s (1971) work on ‘strategic contingency’ suggested that different departments within an organization might have more or less power to impose their view. This was rapidly followed by John Child's (1972) essay on ‘strategic choice’ which drew on both Silverman and the Carnegie school in suggesting that contingencies could mean different things to different actors and also that ‘dominant coalitions’ of organizational members had greater capacities to act upon supposed structural pressures. Hence the management faction were a coalition of actors who saw the world in a particular way and had the strategic resources to ensure that their perspective was dominant in the organization. The work of Andrew Pettigrew (1973; 1985) is a good example of such a position – a detailed account of an organization is provided to support an analysis of power-brokering and political manoeuvring in the management of meaning. Formulations of rationality, formal organization and environmental pressures are no longer foregrounded within a descriptive framework that stresses local specificity over normative or structural generalizability.

By the late 1960s and early 1970s a variety of other British writers were also publishing work which could be broadly characterized as industrial sociology. As noted above, the approach was neo-Weberian or conflict pluralist with an emphasis on investigating ‘shop floor’ respondents from the primary or secondary sector, and not the total institutions favoured by US interactionists. Anthropologists and sociologists at Manchester and Liverpool Universities respectively echoed the US community studies in combining elements of sociology with concepts derived from anthropology – focusing particularly on informal structures and traditions (Banks, 1960; Banks, 1963; Lupton, 1963; 1966).

A little later, the highly influential affluent worker studies (Goldthorpe et al., 1968; 1969) took factory case studies and attempted to relate them to theories of class formation; Ronald Fraser edited collections of biographical accounts of work lives (1968; 1969); and Barry Turner wrote about initiations into ‘occupational communities’ and invoked the notion of subculture – a concept largely derived from its development in the sociology of deviance and poverty (1971). In addition a number of industrial relations studies deepened the 1968 Donovan Commission report in analysing organizations from a standpoint in which internal and external influences are assessed for their influence on class perceptions, levels of industrial militancy, attitudes towards technology and so on (Beynon and Blackburn, 1972; Wedderburn and Crompton, 1972; Daniel, 1973). The concept of ‘work orientation’ is one that is shared across many of these studies – a concept that necessarily suggests that one person's or group's beliefs may be different to another's for a wide variety of reasons. Whilst this is not an epistemologically radical position, it is one that throws severe doubt on any assumptions of monolithic consensus within a complex organization.

At roughly the same time US and UK anthropologists were beginning to turn their focus towards the societies they lived and worked in. Abner Cohen's Two Dimensional Man (1974) is a good example of the beginnings of an ‘anthropology of power and symbolism in complex society’ in which the organization of culture is an ordering concept and organizations are the major focus. It is not surprising then that the term ‘culture’ crops up increasingly in industrial studies from the early 1970s onwards (Beynon and Blackburn, 1972: 156; Lane and Roberts, 1971: 232) and mention is also made of the discipline of ‘industrial anthropology’ (Mars, 1973: 200). Nonetheless, the term ‘culture’ is not usually applied to organizations themselves. Fox (1971) makes reference to working class, middle class and managerial cultures and subcultures but these are value systems that determine orientations to work in general, not to specific organizations.

Broad cultural values and ideologies may materially shape attitudes to work and to relationships of subordination and dependency and other variables … All values and attitudes capable of being expressed in the work situation, in fact, are potential influences on orientations to work and upon the social patterns that result. (1971: 15)

In general terms then, the 1960s and 1970s saw a general undermining of systems theory through an increasing emphasis on actors' meanings, orientations and definitions of the situation. Though the debts and lineages were varied, W. I. Thomas's aphorism ‘if men define situations as real, they are real in their consequences’ was being applied with increasing frequency. But most of these arguments were aimed against the epistemological assumptions of the rational model. In the final section I want to briefly look at some work which was aimed more at its politics, not because this was particularly influential on managerial culturalism, but because it is important to my argument later in the book.

Marxism and Culture

In 1974, Harry Braverman's highly influential Labor and Monopoly Capital reinvigorated a Marxian attempt to theorize organization as a process of domination. Though many of the subsequent debates were critical of Braverman's implicit structural determinism, this book coincided with and stimulated a substantial body of work that applied a broadly labour process analysis to practices in particular organizations and also began to invoke Gramscian notions of hegemony and the management of ideology. The influence of studies of labour and working class history provided the necessary comparative edge to tie in with an analysis of stages of industrial capitalism (Thompson, 1968; Marglin, 1980). Once again, it is difficult to draw precise divides here. The neo-Weberian and industrial relations studies referred to in the previous section often focused on class as a major explanatory variable and many broadly conformed with a model of Marxist political economy. Yet from the mid 1970s onwards there was an emerging body of work that investigated organizations primarily as instruments of power. The dialectic of control within the organization then becomes a story about how certain groups manage to achieve hegemonic status and the strategies and justifications they use to maintain that position (Clegg and Dunkerley, 1980).

Along these lines, Beynon (1974), Nichols and Armstrong (1976) and Nichols and Beynon (1977) describe subordination and resistance in everyday work in a way that explicitly relies on actor accounts of organizational life under capitalist systems of domination. In the US, Burawoy's (1979) work in the same plant that Donald Roy had worked in 30 years earlier pointed to the way in which ‘making out’ – or surviving – was not merely a local response to local difficulties but a response that could only be understood in the context of the changing nature of capital-labour relations. Later British studies by Pollert (1981), Cavendish (1982) and Westwood (1984) add a focus on gender inequalities to the shop floor ethnography. However, perhaps the most explicit, and historically aware, presentation of culture in these studies can be found in the work of Graeme Salaman. In exploring the practical activities that managers engaged in to ensure that control was achieved, and the ways in which these attempts at control were resisted, the role of contested meaning in organizations is central.

So far we have only concentrated on formal knowledge which is used and produced by organisations. Such knowledge constitutes only a fraction of the ideational and moral world of organisations. Distinctive cultures – both organisational and specialist – occur in all organisations, and play an important part in the distinctive and discrete character of different types of organisations. The difference between a military regiment and an electronics factory, or a university department and an insurance company, is not composed only of differences in structure, activity, technology and control mechanisms. It includes different ways of thinking and evaluating, different moralities and cultures. It is this difference we are referring to when we talk of the ‘feel’ of an organisation, or the ‘atmosphere’, or ‘climate’; the distinctive and habitual ways in which members of the organisation (or departments, or sections, or specialities) relate to each other, think, evaluate, know and conceptualise themselves, each other, their work, organisation and their objectives. (Salaman, 1979: 176–7)

Whilst Salaman tends to focus more on class than other dimensions of identification, his attempt to build an account of the processes of domination within organizations is clear, as is his rare understanding that ‘culture’ is related to several other (older) terms.

In most of these accounts, the term ‘ideology’ is of vital importance. If organizations are contested structures of domination then their endurance, their ‘naturalness’ has to be explained. Within a culturalist Marxist framework the ideas generated within a particular organization can been seen as representing, or misrepresenting, the interests of the various groups involved. But this is not simply a matter of the formal versus the informal since the actual practices of capitalists, professionals, proletarians and so on cannot be assumed from structural descriptions. Instead, ideologies have to be seen as justifications for particular states of affairs which are disseminated from various positions within capitalist organizations. Hence Nichols could argue that to understand management practice it is necessary to understand ‘business ideology’, a term which refers to ‘the social values and frames of reference of business men’ (1969: 12). Along similar lines Salaman (1979) suggested that official organizational cultures and ideologies were bodies of practice and belief that members had to be deliberately selected for and then socialized or indoctrinated into. These ideologies were counterposed to the unofficial, lower level, cultures that formed pockets of resistance against dominant assumptions and used sabotage, strikes and cynicism in response. The idea that cultural practices in organizations could be read as legitimations of, or resistance to, the power of elites was not in itself new since examining the conditions for the ‘legitimacy’ of authority is at the heart of the Weberian project too. However, when this is tied to a model of capitalist social, political and economic structure it becomes a powerful analytical and critical tool for questioning the taken-for-granted in everyday managerial practices.

In general terms, then, the development of culturalist Marxist accounts involved a critique of functionalism that paralleled that of action theorists. The main difference is that most of the criticisms are aimed at the managerialism of the various ‘one best ways’ and not the methodology of positivism – at the politics and not the epistemology. That being said, many 1970s neo-Marxists were also keen to disavow determinist readings of historical materialism. As Friedman (1977), Clegg (1979) and Edwards (1980) pointed out, the actual structuring of workplace relations depends as much on the relation between local resistance and history as it does on the pressures of a particular mode of production. In this sense most of this work presupposes a dialectic between the organization and its capitalist environment. Thus local specificity meets wider generality. Of course, from an interpretive point of view, the epistemology of much of the analysis is suspect since terms like ‘capitalism’, ‘ideology’ and so on are assumed to refer to enduring things and are not, as the interactionists or ethnomethodologists would have it, simply accounts of different forms of action.4 Yet within such an analysis the space for a formulation of organizational culture is clearly opened as the relation between the individual, an organization and a capitalist structure (see Eldridge and Crombie, 1974; Salaman, 1979). Capitalism must engineer actors' consent in order to operate, whether through economic coercion or ideological smokescreens; hence an understanding of organization necessarily involves an appreciation of the various methods that attempt to ensure co-operation is achieved.


As I have suggested, one of the themes that can be drawn from my speculative history is a persistent attempt to find different ways to think about and operate bureaucratic organizations. Many of these ideas have resonances with the organizational culturalism described in Chapter 1 in that they share the rhetoric of liberalizing and flattening the pyramid of bureaucracy, and attempting to insert accounts of meaning into the abstractions of organizational structure. So in general terms I hope that this chapter has demonstrated that many of the themes raised by the organizational culturalists in the early 1980s were far from new. This century-long history of writings on organization and management is rife with examples of approaches to organization that explicitly or implicitly call upon formulations of culture – manifested variously as climate, personality, atmosphere, institutionalization, informal organization and so on. That these strands found expression in Peters and Waterman et al. was hence the latest manifestation of a dialectic of ideas in which, broadly speaking, cultural accounts of organization have been framed as an ‘other’ to structuralist ones, or as Gouldner would have it, the ‘natural system’ to the ‘rational’ models. Certainly this ‘other’ was for a while written beneath a Parsonian hegemony, but it was never absent. Any historically sensitive account of organizational culture must, in that sense, develop this tradition and deal with this dualism. The claim that culture has been ignored is simply unsupportable, though it is possible to say that it has rarely been articulated as the disciplinary centre of organization studies.5 Gordon Marshall suggested that Deal and Kennedy were ‘rediscovering Burns and Stalker's sociological wheel’ (1990: 99). I would go further and suggest they were rediscovering another reading of Weber and Taylor. The epistemological tensions between verstehen and ideal type in Weber, and the political tensions between consen and coercion in Taylor, have never been resolved and probably never will be on their own terms. Any ‘new’ theory of organizational culture can only avoid these problems if it chooses one side and ignores the other, precisely as Peters et al. did from the early 1980s onwards.

So, the end of the 1970s saw a variety of theoretical approaches coalescing upon a mediation between action and structure. Interpretive sociologies. Gramscian versions of labour process theory, decision making theory, institutional theory and neo-Weberian pluralism were all attempting to accoun for the local and the general with various different levels of emphasis. Of course, there were still intellectual currents which did not correspond to this characterization (ethnomethodology or population ecology for example – Reed, 1992) but generally the attempt to find a mediating term was common. It could be argued that the formal-informal dichotomy was increasingly being perceived as misleading. As Mouzelis had suggested some time previously, they were ‘a pair of concepts not adequate to deal with the complexities of organizational behaviour and structure’ (1967: 149). Stressing either one is inadequate because it runs the risk of making one the repository of problems that must be solved by attending to the other. In other words, if the formal structure is a problem, use cultural means of analysis and control (human relations, humanistic psychology), but if the informal structure is a problem, use structural means of analysis and control (scientific management, systems approaches). These dualisms were certainly unhelpful, and did much to structure the debate with the ghost of Weber in ways that served to solidify the differences between various intellectual currents. Yet, as the next chapter will show, even when the agency and structure chasm had been identified as a problem, this did not prevent much research on organizational culture falling into some predictable patterns. So, the next chapter will look at how the term ‘culture’ has been theorized within more contemporary academic approaches to organization in an attempt to see what further resources might be found there.


1 Compare this to Casey's (1995) rather similar diagnosis of the ‘postmodern’ condition.

2 Rosabeth Moss Kanter's (1977) ethnography of Indesco is a good example of the later development of this line of US organizational sociology, containing as it does a recognition of the importance of gender in the reproduction and legitimation of organizational rationalities.

3 For a review and critique of this literature see Rowlinson (1997, particularly 82 passim).

4 Compare, for example, the fairly ethnomethodological writings of Stewart Clegg in 1975, much influenced by David Silverman, with his later more Marxist writings (1979; Clegg and Dunkerley, 1980).

5 The role that the ‘remembering’ and ‘forgetting’ of these various works has played in the construction of contemporary organization studies is an interesting one. It might be argued that sociological and psychological works on organization had to fall victim to amnesia before the ‘new’ discipline of organization studies could come into being.

  • culturalism
  • human relations
  • bureaucracy
  • neo-human relations
  • human relations theory
  • organizational culture
  • organizations
  • Loading...
locked icon

Sign in to access this content

Get a 30 day FREE TRIAL

  • Watch videos from a variety of sources bringing classroom topics to life
  • Read modern, diverse business cases
  • Explore hundreds of books and reference titles