Identifying and Explaining Organizational Cultures in the Public Sector: A Study of the Australian Public Service Using the Interaction Ritual Theory of Randall Collins

The Power Dimension: Order Giving and Order Taking

This section will show that people within the APS display different cultures because they occupy different locations in order giving hierarchies. Three such cultures can be identified: those of the order givers, the middle ranks, and the order takers. This section will also show that the combination of high social inequality, high social density, and low social diversity that typifies the APS generates, as Collins predicts, a culture of “coercive, highly visible deference rituals.” This culture is manifested in “rankism” and the “bureaucratic personality.” Collins argues that order givers and order takers within organizations acquire different cultural outlooks. Order givers are highly attached to official ideals as they must represent them when giving orders to others. They also tend to identify their role and the organization with their own ego. This is because they need to make themselves look impressive and in control when giving orders and because order giving provides them with ego rewards. Order takers by contrast are more cynical about official ideals and identify less strongly with their role, as they suffer ego loss when taking orders from others. They give perfunctory deference to official ideals but ignore or criticize them privately in their backstage group. Survey data from the APS support these claims. For example, hierarchical rank and organizational commitment among APS staff strongly correlate (Department of Finance [DOF], 1990; DITAC, 1992; Jans & McMahon, 1988). Similarly, SES officers have a sanguine view of the impact of the public sector reforms over which they preside, whereas the lower ranks are much more skeptical (Management Advisory Board/Management Improvement Advisory Committee [MAB-MIAC], 1992). Informants reported that their workmates were often skeptical about official statements. So too were my workmates. Survey data from the APS confirm Collins’s claim that hierarchical location powerfully shapes peoples’ cultural outlook. For example, one survey of APS staff discovered that hierarchical classification was the demographic variable which was most likely to yield varying responses (other variables included length of experience, type of agency, and gender). Indeed, variations in responses by staff were almost invariably linked with hierarchical classification (MAB-MIAC, 1992).

Collins (1988, p. 211) argues that order takers exhibit a ritualistic response of external conformity combined with private alienation. My workmates exhibited such behavior as they would complain about work to their workmates but conceal their disaffection from superiors. Pringle (1988) finds that Australian secretaries behaved similarly. Prandy et al. (1982) discover that lower level white-collar workers exhibited the highest levels of “self-estrangement.” The data from the APS confirm Collins’s claim that order givers identify with their organization whereas order takers are alienated from it. He argues that order giving is a ritual in which order givers seek to impose their view of reality on others. Ironically, however, because order takers are alienated from authority and because order givers must make a stronger commitment to official ideals if they are to represent them in a ritual, order givers end up indoctrinating themselves more than they indoctrinate others. For example, Jans, Frazer-Jans, and McMahon (1989) report that “Almost all the SES see that Tax is a well run organization, which cares for its staff and a place where morale is high; but only about two-thirds of all other Tax staff share this perception” (p. 36). Caiden (1967) likewise finds that

All senior officials consider morale to be high in their units . . . But lower grade officials, while outwardly respectful, hardworking and supporting the official view, tend to harbour grievances . . . one only has to listen to conversations between colleagues, to persistent criticism of supervisors out of their presence and to discussions away from the workplace to find out how deceptive appearances can be. (p. 226)

Drawing on the ideas of Goffman, Collins claims that order takers display an informal “backstage” culture, whereas order givers display a more formal “frontstage” culture. Within backstage cultures, people are less “image conscious” and are more forthright and uninhibited. Examples of backstage culture include rough informal dress, colloquial speech, open sexual remarks, elaborate griping, smoking, profanity, and reciprocal first naming. Junior staff within my department displayed such an informal backstage culture, whereas the higher ranks displayed a frontstage culture of middle-class respectability. For example, an all-female typing pool in my department was decorated with “beefcake” pinups, an adornment that I never encountered in the offices of senior female staff. Informants depicted the dress and demeanor of junior officers as being informal and uninhibited, whereas they saw senior officers as being more formal and restrained. Frontstage cultures tend to emerge where people’s work requires them to present a favorable public image. For example, a study of Department of Social Security (DSS) regional offices, Job Centers, and Unemployment Benefit Offices in Britain found that the more image conscious Job Centers encouraged a “professional” attitude toward work involving “civilized” behavior and dress. Conversations among staff in the Job Centers, for example, centered on “respectable” subjects such as families, holidays, and domestic issues. By contrast, staff in the DSS and Unemployment Benefits Offices displayed a backstage culture characterized by informal dress and ribald conversations in which “talk was often sexually explicit and language unrestrained.” Staff acknowledged the existence of a “benefits humor” which they described as “crude and vulgar” (Cullen, 1994). APS policy advisers exhibit a frontstage culture as they must present a “professional” image when dealing with high status figures such as ministers, senior officials, and lobbyists. By contrast, staff who perform technical and administrative work rarely encounter such high-status figures and therefore tend to exhibit a backstage culture. For example, an informant reported that his workmates in an information technology area openly discussed sex. When I mentioned this to a central agency informant, he was shocked. He said that sex was a “taboo subject” in his workplace, the ethos of which he described as “puritanical.”

The DFAT has a particularly marked frontstage culture. This is because the practice of diplomacy has historically involved ceremonies in which deference is paid to nations via their envoys and in which nations seek to impress others with a display of their power (Morgenthau, 1973). While working in Canberra, I met graduate recruits from a range of departments when attending various training courses. I found that graduate recruits from DFAT differed markedly from those of other departments. They had wider conversational interests, were better dressed, and exhibited an urbane demeanor. A non-DFAT informant who had worked in DFAT similarly considered that DFAT officers appeared to be “above themselves.” DFAT historically recruited many of its staff from the graduates of elite private schools (Encel, 1970). Diplomats in many European countries and in the United States likewise tend to have elite backgrounds. This is because diplomacy requires frontstage skills of the type that such elites tend to possess. Staff in my department (DITAC) did not exhibit the refined “patrician” culture of DFAT officers. Instead, they tended to exhibit a coarse “plebeian” backstage culture. For example, a workmate noted that staff in DFAT dressed more stylishly than those in DITAC. Indeed, senior DFAT officers would disparagingly refer to their DITAC counterparts as the “brown brigade” because of their alleged predilection for wearing brown suits. Jackall (1988) likewise notes the “mild taboo” against brown suits that is present among American corporate managers, as “brown is dull, a loser’s colour, winners choose blue” (p. 47).

Collins (1975, 1988) argues that the middle ranks exhibit a mixture of the order givers’ and the order takers’ cultures. Where promotion opportunities are present, he argues that order takers undergo anticipatory socialization into the order givers’ culture. This gives rise to what Collins calls the “conformist middle class” culture (see Figure 2). This culture is ascetic and moralistic, community oriented, respectable, and hard-working. Weber argued that the work situation of the lower middle class accounts for this style, for it is one in which success seems possible through self-discipline. The same is true of bureaucrats as they enjoy the possibility of career progression if they work hard. Senior and upwardly mobile APS officers accordingly tend to display this culture. For example, a social scientist who attained senior rank in the APS noted that its culture was marked by a

pervasive pressure to conform to the value systems, attitudes and behaviour which might broadly be described as belonging to Anglo-Saxon middle class morality of a generation ago . . . The civil servant, both in his official and unofficial capacity, is expected to observe what is regarded as conventional morality . . . (and) to conform to certain standards of dress . . . (quoted in Encel, 1970, p. 21).

As Encel (1970) observes, within government offices, “formality of dress and demeanour has always been essential” (p. 79). Others have likewise attributed to senior APS officials such qualities as conservatism, reserve, social conformity, Puritanism, and moralism (Caiden, 1967; Cumes, 1988; Pusey, 1991). American corporate managers also dress conservatively and are puritanical (Jackall, 1988; Kanter, 1977).

Collins (1988) argues that first line supervisors tend to exhibit the style of the “bureaucratic personality” or one that involves a stereotyped, rigid, and external identification with the organization’s rules and procedures (p. 213). He attributes this to the experience of taking orders from some people and giving them to others. In such instances, people identify with the organization officially but enact their role in a way that conforms externally without concern for its larger purposes. By contrast, top managers are able to create and develop plans for the organization and therefore think about its wider purposes. For example, whereas 81% of SES officers feel that they are “encouraged to come up with new and better ways of doing things,” only 44% of ASO 1- to 4- level staff feel similarly encouraged (MAB-MIAC, 1992). The inability of the middle ranks to create and develop organizational policy and to think about its wider purposes leads them to enforce their organizations rules and procedures in a rigid and perfunctory way. As an informant noted, public servants often tend to become “petty” people. Such pettifogging “bureaucratic personalities” are mostly encountered at the middle levels of APS agencies. For example, SES officers from spending departments reported that it was middle-level officials in the DOF (rather than senior officials) who imposed the most stringent controls on expenditure. One claimed that “My program managers got queried about the price of a lawnmower . . . They thought we spent too much money and that the machine was too powerful” (Campbell & Halligan, 1992, pp. 151, 156). These middle-level officials ignored the declared objective of the policy, which was to seek greater cost-effectiveness and to provide spending departments with greater discretion, and focused instead on the meafns, namely, cutting expenditure. A Royal Commission on Australian Government Administration (RCAGA; 1976) study of 100 APS middle managers similarly reported that “rule behaviour is strongly evident” (p. 309). Crozier (1964) also discovers strongly ritualistic behavior among junior order givers in a French government agency. Junior informants reported that their supervisors would often amend their written work simply to demonstrate their authority rather than to improve it. My own supervisors in DITAC behaved likewise.

The bureaucratic personality is common among people who enforce laws and represent the organization to the public, such as window clerks in government agencies (Collins, 1988, p. 214). Customs officers, for example, enforce the law and represent their organization to the public when controlling the importation of goods and the entry of visitors to Australia. A report found that they tend to have a ritualistic approach to policy and law enforcement with a focus on the letter of the law rather than its intention. Policies are treated as ends in themselves rather than as means to ends (Committee of Review Into the Australian Customs Service [CRACS], 1993). Collins (1988, p. 213) argues that the more coercion there is in order giving rituals, such as in the military, the more ritual obeisance there is between classes. In the U.S. military during World War II, this ritual obeisance was labeled “chickenshit” by soldiers. Examples include petty harassment of the weak by the strong, verbal public humiliation of subordinates, pleasure taken in rank for rank’s sake, and an insistence on the letter rather than the spirit of ordinances (Fussell, 1989). The Customs service possesses such a culture of ritual obeisance as its uniformed officers exercise police powers. I encountered this culture when working in a branch of Customs that had been relocated to DITAC. A workmate likened it to that of the army. Many first-line supervisors in DITAC were also petty authoritarians, as Collins’s theory predicts. For example, 18% of ASO 1- to 6-level staff reported that they had experienced bullying or harassment in the previous year compared with only 8% of staff at SES level (Australian Public Service Commission [APSC], 2005). Collins (2004, p. 349) argues that those who face such petty order givers from below are especially alienated by being an order taker. As many Customs officers are in this predicament and because many have working class origins, Customs has (like the military) a strong backstage order takers’ culture. This is manifested in acts of ritual profanation such as using obscene language and a “macho” culture of aggressive masculinity which enables male order takers to protect their self-esteem. For example, when my Director in a central personnel agency (a woman with a prim, upper-middle-class demeanor) observed during a meeting with two senior male Customs officers that Customs was “one of the more hierarchical departments,” she was visibly taken aback when one of them used an obscenity in his response.

Collins (1988) argues that when surveillance is used as a control technique, “the act of being observed is itself the focus of ritual attention and an enactment of authority and subordination. Hence attention is deflected from the task itself to the issue of compliance” (p. 455). Many of my supervisors behaved in exactly this way. For example, one would glance at his watch every time that I returned to my desk after an absence. Another would suddenly open the office door without knocking first and then peer inside. As a workmate observed, this practice had the apparent intent of catching off guard any subordinate who might be relaxing on the job. Other supervisors have behaved similarly. A woman recalls that in the 1950s, her supervisor would wait at the door for staff to return from lunch and write down the names of those who were a few minutes late (DSS, 1987). Encel (1970) reports that even an agency head had concerned himself with the “important matter” of checking on the punctuality of his subordinates in returning from lunch. Assigning people pointless work tasks to ensure that they are “kept busy” shows how the use of surveillance redirects the focus of ritual attention away from the task itself toward the issue of compliance. For example, when a supervisor discovered that I had no work to do, he gave me the task of rearranging columns containing hundreds of random numbers in ascending order. I later discovered that this task, which took me many weeks, could have been performed by a computer in seconds. My supervisor was not concerned with accomplishing our department’s objectives efficiently but with ensuring that I was visibly active. Accordingly, I found that the best way to avoid this fate was to “look busy” by shuffling papers on one’s desk. My conduct confirms Collins’s prediction that the use of surveillance to control order takers tends to alienate them and to yield minimal compliance, merely performing the physical motions demanded. For example, many of my workmates would conscientiously attend work and fill out their time sheets but exert only minimal levels of work effort. A number of my supervisors noted that the use of time sheets to record hours of attendance at work yielded this outcome, as it led staff to focus upon being physically present at their desks rather than upon performing their jobs.

Whereas the middle ranks implement policy, the senior ranks create it. They accordingly identify more strongly with their order giving roles and have a greater sense of the organization’s purpose. Collins (1975) argues that the culture of the upper class derives from the experience of being in command and is characterized by assuredness, arrogance, sociability, awesomeness, self-importance, and a deliberate and dignified manner. SES officers tend to display these attributes. For example, when asked to nominate the attributes of senior officers, informants invariably used words such as articulate, well presented in public, cool under pressure, self-confident, and assertive. Collins (1975) notes that the upper-class outlook emerges more sharply the more unmitigated deference people experience, so that insecure politicians who experience less of this than secure business tycoons possess the upper-class outlook to a lesser degree. Senior APS bureaucrats enjoyed a high level of job security until the mid-1980s. Caiden (1967) characterizes them as being a class of “secure, self-righteous, highly conservative and complacent official mandarins” (p. 15). He also identifies condescension and arrogance as distinctive traits of Australian administration (Caiden, 1990). Self-confidence verging on arrogance is also a characteristic of senior French and British civil servants (Wilks, 1990). The RCAGA (1976) notes that among many senior APS officials, “a sense of belonging to a distinctive, privileged and authoritative class of officers of the state” had emerged (Vol. 1, p. 23). This was especially true of the “permanent heads” of departments who occupied what one observer described as “an exalted status” (Spann, 1976). The greater level of respect enjoyed by the senior ranks is reflected in survey findings that SES officers enjoy better social relations with their superiors than the lower ranks do and are more often consulted by them (see DITAC, 1992; DOF, 1990; Task Force on Management Improvement [TFMI], 1992). As Butler (1973) observes, senior SES officers have a “heavy weight status” and “cannot easily be sat upon, even by the permanent secretary” (p. 29). By contrast, junior officers must learn to accept their “place” in the hierarchy. This is reflected in the informal convention that dress should be appropriate to rank. For example, it would be inappropriate for a junior male officer to wear a suit to work but entirely appropriate for a senior male officer to do so. Likewise, it would be inappropriate for a junior officer who held a doctorate to use her or his title in correspondence but perfectly acceptable for a senior officer who held a doctorate to do so.

The APS combines high social inequality with high ritual density. Collins (1988) argues that this combination of what Douglas calls “high grid” (inequality) and “high group” (solidarity) gives rise to rigidly stratified, rank conscious societies in which “higher ranking individuals are allowed to have individual selves, which may be worshipped by others in elaborate deference rituals . . . whereas lower-ranking persons have severely degraded selves” (p. 254). Individuals in higher social classes are given an exaggerated veneration for their individuality while lower class persons . . . are allowed very little individual self.” Fuller (2003) labels this phenomenon “rankism.” He argues that it occurs when people are divided into “somebodies” and “nobodies” on the basis of their organizational rank. The APS is pervaded by rankism. As an informant bluntly observed: “you are not taken seriously until you are a [class] nine, below a nine you are a nobody.” Senior officials by contrast are “somebodies.” For example, an agency head approvingly cited the following words from a government report:

A person who attains a permanent headship has reached the peak of the profession of public administration. His advancement to that status is an acknowledgment of excellence and ought to be seen by the community in that way. He is entitled to the respect and dignity that this achievement deserves . . . (cited in Tange, 1984, p. 12).

Other agency heads shared this view. For example, one head who was removed from his post by the government rejected the low-status ambassadorships that were offered to him in recompense on the grounds that they were “derisory.” Cumes (1988) observes that the offer by the government seemed designed to ensure that the head “anxious to hang on to such fragments of ‘dignity’ and ‘status’ as he still had, would turn them down” (pp. 113, 115). When Sir John Bunting, the head of the Department of Prime Minister and Cabinet, was transferred to the less-prestigious post of Head of the Cabinet Office by the Gorton government, many public servants and others considered that a slight had been inflicted on him and “were disturbed by the way Bunting had been treated” (Hawker, Smith, & Weller, 1979, pp. 113-114).

Whereas senior officers are entitled to and receive deference from others, junior officers have no such entitlement and receive little deference. Instead, they are at best ignored and at worst subjected to degrading treatment. One junior informant working in a file registry reported that

you’d talk to someone over the phone and you’re just treated like a dogsbody . . . I know that there was one lady that used to ring up and say ‘I want this file on my desk’ and hang up. We felt like not doing it, when you get commands like that.

Older or highly educated staff can undergo status degradation when they are ordered to perform menial tasks. An informant aged in his 30s reported that he had felt demeaned when an SES officer had got him to fetch a box of books from the trunk of his car. The use of status symbols and staff classifications (e.g., “APS 3”) ensures that peoples’ status is publicly visible and thereby serves as a constant reminder of their inferior rank. Status in my department was symbolized by office size and the proximity of one’s desk to a window. Higher ranking staff sat next to a window, those below them in rank sat at a slightly further remove while the lowest ranks sat in a windowless corridor. As Collins predicts, the combination of high ritual density and high social inequality that typifies the APS yields a culture of deference rituals as opposed to one of group inclusion rituals. Maconachie (1992), for example, reports that the culture of one APS agency was

male, paternalistic and autocratic. Standards of dress, speech, behaviour and mode of address were imposed, and it was expected that they would be rigidly followed. These rules imposed polite deferential behaviour between staff and superiors. (p. 226).

Staff in the DSS likewise report that until the 1960s, they were required to stand up whenever the Director-General entered the room. As one staff member recalls: “Nobody ever called the D-G by his Christian name” (DSS, 1987, p. 17).

We may conclude that cultural differences within the APS can be explained as the product of peoples’ location on the power dimension. Order givers exhibit a frontstage culture; the middle ranks exhibit a bureaucratic personality that blends the frontstage and backstage cultures, whereas the order takers exhibit a backstage culture. As Collins’s (1988) theory predicts, the combination of high social inequality, high social density, and low social diversity that typifies the APS generates a culture of “coercive, highly visible deference rituals.” Manifestations of this culture include “rankism” and bureaucratic ritualism (p. 224). Key features of the latter are the way in which the use of surveillance to control staff deflects the focus of ritual attention away from the work itself toward the act of being observed, rigid rule enforcement, and petty authoritarianism.

The Status Dimension: Localism Versus Cosmopolitanism

The second source of cultural differences is the amount of ritual density. Collins (2004) calls this the “Status” (as opposed to the “power”) dimension (pp. 111-118). Ritual density has four dimensions: ritual intensity, central-peripheral participation (“ritual centrality”), social density, and social diversity. These four dimensions are separate aspects of ritual that can vary independently of each other. Variations in ritual density generate differences in three aspects of culture (Collins, 1988, p. 200). These are the attitude to membership symbols, the attitude to non-members, and the reaction to symbolic violations. High ritual density generates a localistic culture in which people reify symbols by treating them as moral absolutes rather than as relative principles to be adopted under different conditions, in which they distrust non-members and in which they react to symbolic violations with righteous anger and punitive ritual. Low ritual density by contrast generates a cosmopolitan culture in which people view symbols in relative rather than in absolute terms, in which they trust non-members and in which they are lax in dealing with symbolic violations. This section will show that variations in ritual density within the APS underpin differences between localistic and cosmopolitan cultures. Specifically, it will show the following: (a) that each of the four dimensions of ritual density varies within the APS; (b) that social diversity exhibits the most variations and that these variations in social diversity stem from differences in people’s educational levels, the complexity of their work tasks, the diversity of their social contacts, agency recruitment and promotion practices, and the nature of their relations with clients; and (c) that variations in these factors within the APS are reflected in differences between localistic and cosmopolitan cultures. These cultural differences correlate with organizational rank, type of agency and regional as opposed to central office location.

The level of ritual intensity within the APS is generally moderate rather than high. This is the case because (a) bureaucratic norms discourage emotional expression, (b) workplace rituals are mostly “natural” rather than “intentional,” (c) the predominant emotion among junior staff is one of boredom, and (d) order takers lose emotional energy when engaging in power rituals. Ritual intensity is greatest among order givers because they experience high ritual centrality and the strongest effects of ritual membership: emotional energy, moral solidarity, and attachment to group symbols. Most APS staff experience high levels of social density as they work in open plan offices that entail a high level of physical co-presence with others. Although senior staff have private offices, their managerial and policy advising jobs involve high levels of face to face interaction. Accordingly, social density within the APS exhibits little rank-based variation. Social diversity by contrast does vary in accordance with rank and with such factors as educational levels, agency location, and work tasks. Collins argues that variations in social diversity stem from both the amount of diversity in the types of people that one encounters and in the content of one’s communications. Levels of education correlate with social diversity because the more educated tend to receive more written communication and to undertake greater reading. These factors increase the amount of diversity in their communications (Collins, 1988, p. 217).

The APS historically possessed a low level of social diversity because it was ethnically homogeneous, predominantly male, and staffed by ex-servicemen and school leavers. When I worked in DITAC in the mid- to late-1980s, its ethnic composition was predominantly Anglo-Celtic with very few non-Europeans. There was a shared culture that centered on following sports such as Australian Rules football, rugby league, and cricket. People who had different interests were somewhat marginalized. A workmate, for example, noted that a particular director in DITAC who had strong interests in the arts and in ideas was very different to other people in our department. In situations of high ritual density, we encounter strong conformity pressures and a suspicion of “outsiders” or those who are different. For example, 60% of Tax office staff agreed that there was a high level of conformity within the ATO (Jans et al., 1989). Crozier (1964) likewise finds that peer group social conformity was high within the French organizations that he studied due to the presence of strong horizontal as opposed to vertical ties. Although most staff in my department were university graduates, it had a markedly anti-intellectual culture. Those with intellectual proclivities were suspect. One of my supervisors expressed to me his relief in discovering that I was not an “ivory tower type” (he had been forewarned that this was the case by my previous supervisor). Anti-intellectualism is a long-standing feature of the APS. The public service union in the 1950s, for example, cautioned its members to be wary of “outside appointees, long-hairs, professors, authors, psychologists, advisers and economic witchdoctors” (Thompson, 1986, p. 42). Cumes (1988) reports that among his SES colleagues, “the capacity to reach out for new solutions . . . attracts suspicion to those who try to exercise it. They are usually categorized in pejorative terms—as ‘cranks,’ ‘airy-fairies,’ ‘troublemakers,’ ‘stirrers’ . . .” (pp. 58-59). Marshall (1992) finds that his SES interviewees believed that academics were “wankers.” Another academic (Wilson, 1990) similarly reports that many of the SES officers that he met were anti-intellectual. Anti-intellectualism reflects the preference of localists for viewing symbols in concrete rather than in abstract terms. Members of minority groups are also marginalized and discriminated against within localistic cultures. Surveys, for example, show that indigenous staff and those from non-English speaking backgrounds are more likely to report being victims of bullying and harassment than other APS staff (APSC, 2005).

Junior staff experience lower amounts of social diversity than senior staff because they are less educated, have less diverse social contacts, and perform less complex work. They also tend to place a greater value on in-group solidarity. This is the case for three reasons. First, such solidarity provides those with blocked mobility with an alternative source of recognition to promotion. Second, much of the work at junior levels provides few intrinsic rewards, and social rewards provide a substitute for this. Third, as Collins (1988, p. 212) observes, order takers are oriented away from the organization because this is the sphere in which they suffer ego loss. Instead, they are oriented either toward their fellow order takers or toward their life outside work. Goldthorpe, Lockwood, Bechhofer, and Platt (1968) refer to the former of these as the “solidaristic” orientation to work. Workers with this orientation strongly identify with their work group and against their employer. Given the presence of high levels of ritual density among junior staff and their preference for egalitarian rather than hierarchical interactions, they tend to display what Collins, following Douglas, calls a “low grid-low group” culture (also labeled by her as “enclavist” or “egalitarian”). In this culture, people exhibit strong in-group solidarity and out-group hostility. Among my workmates, for example, there was strong sense of in-group camaraderie. Tucker (1992) likewise reports that Data Processing Operators (DPOs) in one APS agency enjoyed the “happy social atmosphere” of the pool (p. 11). Many were fearful that the “depooling” associated with the process of workplace restructuring would lead to a loss of social support. She cites the case of “one tightly knit group of three who would go anywhere, but together” and reports that some DPOs even volunteered to “give back” their 4% pay rise so as to remain in the pool.

The enclavist culture emphasizes sharp external boundaries, involves ritual concern for enemies and pollution, and comprises group inclusion rituals and conformity to tradition. For example, junior officers surveyed by the RCAGA (1976) were more than twice as likely as senior officers to believe that it was “easy to make enemies” among their workmates. My workmates were suspicious of and hostile toward senior management and other areas in our department. Informants reported that their workmates felt likewise. Surveys show that APS staff distrust outsiders. For example, whereas 33% of DPIE staff have a “great deal” of trust and confidence in their immediate supervisor, only 4% regard top management similarly. Likewise, whereas 81% of DPIE staff believe that people cooperate to get the job done within their local work group, only 33% believe that this is true of the department as a whole (unpublished data). The division between insiders and outsiders is often based on organizational boundaries. A survey of one APS agency discovered that the establishment of differentiated service centers had produced a loss of cohesion with the emergence of “them and us” attitudes (Maconachie, 1993). Selth (1991) reports that “until recently few regional staff had regular (if any) formal contact with staff of other Commonwealth agencies . . . the other party was seen as being ‘not as good as us.’ The territorial imperative was very strong” (p. 53). Bate (1984) likewise finds a strong sense of territoriality in a number of British organizations that was manifested in a “them” and “us” mentality involving conflict and distrust between different organizational sub-sections. Kanter (1977) observes that workers with blocked mobility will often form closed close-knit groups that embrace an ethos of “anti-success peer solidarity” that stigmatizes the upwardly mobile. For example, an indigenous woman who was promoted within an ATO regional office reports that “It was as though it was okay when I was a lowly Aboriginal but when I gained any kind of success they just turned on me.” She had been called an “uppity Black bitch” and a “Black favorite,” and during a meeting, her work station had been referred to as the office “Black spot.” Staff had also refused to work with her or to answer her telephone and had excluded her from morning tea breaks (“Aboriginal Woman,” 1996). Those who violate norms of group solidarity within such cultures are viewed as enemies. For example, an informant reported that his workmates harbored strong animosity toward strike breakers or so-called “scabs.”

Because senior officers experience greater social diversity than junior officers, they are less inclined to reify symbols by seeing them as moral absolutes. For example, only 26% of junior officers agreed that it would be the “right thing” for a public servant to write a letter to a newspaper about his or her working conditions compared with 41% of senior officers. They were also more likely to believe that this act would be contrary to public service rules and regulations than were senior officers (47% as opposed to 23%). The disinclination of cosmopolitans to reify symbols and to punish symbolic violators is reflected in the fact that graduates were much more willing to condone protest action by public servants than were non-graduates (RCAGA, 1976). These findings are consistent with many studies that show that less-educated, lower status employees tend to hold more authoritarian values (Hofstede, 1980).

Academics experience less ritual density than public servants. This is not only because they are more educated but also because, as Collins (1988, p. 217) observes, the work of managers tends to consist of face to face meetings, whereas the more intellectual professions tend to spend more time working alone and more time reading. He argues that this factor explains why the higher social classes are split between liberals and conservatives on cultural issues. Academia exhibits what Collins, following Douglas, calls an “individualistic” culture. This culture is pluralistic, secular, and has low levels of ritualized interaction. Douglas and Collins maintain that high levels of social solidarity are associated with greater levels of ritual activity, because rituals both generate and symbolize social solidarity. Sociable rituals are more common in the APS than they are in academia. For example, it was common to celebrate staff birthdays within my APS workplaces, whereas this has never occurred in any of my academic workplaces. Such rites de passage symbolize social solidarity. Sporting contests also generate and express social solidarity. For example, there was an indoor cricket competition in my department that fostered interdivisional rivalry and team spirit. Informants reported that senior Treasury officials sponsored interdivisional football matches to build esprit de corps in their units and that they took a keen interest in them. Collins (1988, p. 194) argues that rituals generate pressures for social conformity by creating a common emotional mood among participants. For example, Treasury informants reported that they felt compelled to take part in football matches in spite of their fear of injury, as they did not want to be seen as “having let the side down.” By contrast sporting contests and their associated conformity pressures are rare in academic workplaces.

Although most APS agencies possess high levels of ritual density, levels of ritual density vary between agencies. Accordingly, APS agencies display different cultures. A major source of these variations in ritual density is the extent to which agencies enjoy adversarial as opposed to amicable relationships with their clients. Adversarial relationships with clients foster in-group cohesion among agency staff and hostility toward out-groups. Studies of police culture identify this as the “brotherhood syndrome” (Lewis, 1999). This term refers to the tendency of the police to close ranks against an outside world that they perceive as being hostile toward them and to adopt an “us” and “them” mentality. Customs officers likewise display the brotherhood syndrome. For example, an official report found that the Customs service had an insular culture that was characterized by intense loyalty to the organization, suspicion of clients, and a siege mentality (CRACS, 1993). Another reason why APS agencies possess different levels of ritual density is because they recruit different types of people. For example, Customs and Tax historically recruited working class Catholics with low educational levels and had high levels of internal promotion (Encel, 1970). These practices produced low levels of social diversity and localistic cultures. Encel (1970) notes that “the zeal of Catholic puritanism” was particularly strong in the Customs Service. An informant likewise reported that Tax officers in attendance at a training course described “current Tax culture” as being Roman Catholic, conservative, and “morally straight laced.” Collins argues that people in situations of high ritual density tend to reify symbols by treating them as absolute rather than as relative principles to be adopted under different conditions. Customs officers exhibit such an outlook as they have a ritualistic approach to policy and law enforcement with a focus on the letter of the law rather than on its intention. They tend to treat policies as ends in themselves rather than as means to ends (CRACS, 1993).

Another APS agency that has a strongly localistic culture is the Treasury. While living in Canberra, I befriended a number of Treasury officers. Through these key informants, I obtained an insight into the Treasury culture. Notwithstanding the fact that Treasury staff are highly educated (all “professional officers” have university degrees, often at honors level), the culture of the Treasury is strongly localistic. This is because the Treasury is a small, socially homogeneous agency (it has only 250 professional officers, 95% of whom have economics degrees) with a history of adversarial relationships with other departments and a high rate of internal promotion. This combination of factors generates high ritual density. For example, Pusey (1991) discovers that Treasury SES officers exhibited a higher level of social solidarity and esprit de corps than did SES officers from other departments. Treasury officers often describe their department as being a “family.” Even those staff who leave the Treasury to work elsewhere are said to remain lifelong members of the “Treasury family.” My informants depicted Treasury as being monocultural, insular, and conformist. Because Treasury staff experience high levels of ritual density, they tend to reify ideas and symbols and to punish symbolic violators. For example, my informants reported that the Treasury neo-classical economic paradigm was regarded by staff as being “holy writ.” They said that those Treasury staff members who questioned this paradigm were viewed as being “intellectual heretics.” As a department head once drily observed, a Treasurer has about as much chance of changing his officials’ views as a Protestant pope has of converting his cardinals (cited in Weller & Grattan, 1981).

In contrast to Customs, Tax, and the Treasury, some APS agencies possess low levels of ritual density and therefore display a cosmopolitan culture. This is because they have highly educated and/or socially diverse staff and amicable relationships with their clients. For example, the DURD recruited a large proportion of its staff from universities, state and local government, and the private sector. This unusually high level of social diversity was reflected in the culture of DURD. For example, Lloyd and Troy (1981) report that DURD exhibited more radicalism than was normal in Commonwealth departments. Painter and Carey (1979) likewise maintain that DURD officials often did not conform to the cultural norms of the public service due to the “hothouse experimental and often academic atmosphere that existed within DURD.” They report one non-DURD public servant as complaining that DURD staff conducted interdepartmental committee meetings as if they were “tutorials.” The Office of National Assessment (ONA), an intelligence agency that recruits international relations postgraduates, also has a cosmopolitan culture. An informant reported that the members of ONA saw themselves as being the “Bloomsbury group of Canberra.” The DFAT possesses a markedly cosmopolitan culture as it has a highly educated workforce and very diverse clients with whom it enjoys largely amicable relations. DFAT’s core task is the conduct of Australia’s diplomacy. This task involves seeking to understand the views of foreign governments and cultivating good relations with them. This combination of high social diversity and amicable relations with clients fosters a cosmopolitan outlook. For example, DFAT staff are often accused of exhibiting the “Lawrence of Arabia syndrome.” This is a tendency to act as advocates for foreign countries and to espouse vague internationalist notions ahead of a more hard headed appreciation of national interests (Johnstone, 1989). Because DFAT staff perform ceremonial diplomatic roles and experience low ritual density, they tend to exhibit the “sophisticated cosmopolitan, frontstage culture” depicted in the upper right of Figure 2. In contrast to DFAT, the Customs service seeks to protect Australia against exotic diseases, prohibited imports, terrorists, criminals, and illegal immigrants. Customs officers therefore tend to view their clients as being their adversaries. When combined with their low educational levels, this yields high ritual density and a localistic culture. Because they mostly enforce rules or take orders, Customs officers exhibit bureaucratic and backstage personalities (see the lower left of Figure 2) rather than the cosmopolitan, frontstage culture of DFAT.

The central and regional offices of departments also display different levels of social diversity. This is because central offices perform the more complex managerial and policy advising tasks while regional offices deliver services to clients. Regional offices also have flatter structures, lower educational levels, and less interdepartmental mobility than central offices. They therefore have lower levels of social diversity and a more localistic culture. This is manifested in a stronger tendency to reify symbols and to distrust outsiders. An informant who had worked in both the central and regional offices of the ATO summarized the differences thus:

Central office you’ve got higher qualified people, you’ve got more women across the spectrum and you’ve also got greater vision. In the branch offices, the hierarchy’s revered . . . purely men in the decision-making positions and quite a conservative culture.

Many regional office staff feel suspicion and resentment toward outsiders, especially those based in Canberra (Selth, 1991). Some perceive themselves as being the “victims” of central office (“Canberra does things to us”; TFMI, 1992). A sense of alienation from central authority also typifies street level bureaucrats in Britain (Cullen, 1994). Bate (1984) finds that the members of three British organizations tended to externalize their problems by blaming “head office,” “the workers,” “the unions,” and “the Government.” Collins (1988, p. 218) argues that in cultures with high levels of ritual density, people tend to personify larger social processes as forms of diabolical intervention by outsiders. This is because they fear such outsiders and tend to think in terms of concrete individuals rather than in terms of abstractions. Such personification is expressed in the practices of scapegoating and witch-hunting. For example, Tax officers report that “Staff suffer from low self-esteem . . . This causes them to look for a scapegoat and they make the managers the scapegoats”; “Managers are being singled out as the sole reasons for all the problems and quite simply this is wrong” (quoted in Jans & Frazer-Jans, 1990b, p. 6.18). Informants reported that their workmates often sought to find scapegoats for problems. A study of 100 APS middle managers found that they produced an “impressive list of scapegoats” when challenged over their performance (RCAGA, 1976).

Under conditions of high ritual density, the reaction to symbolic violations tends to involve righteous anger and punitive responses. Organizational whistleblowers, for example, are often ostracized, harassed, dismissed, or prosecuted as they are perceived as having betrayed their workmates. A widely reported instance of this occurred in Customs in the 2000s. Those who show disrespect for the sacred symbols of the group can elicit a similar response. Collins (2004) argues that people in situations of high ritual density tend to reify symbols by treating them as immutable and irreproachable realities, not to be questioned or criticized. Departmental policies and their underlying intellectual frameworks are often viewed in this way by APS staff. For example, Cumes (1988) reports that when he proposed a Keynesian solution to an economic policy problem to a senior official, the latter had replied, “That’s nonsense! . . . with some anger—and some venom” (p. 59). Cumes observes that his statement was so sharply at variance with the senior official’s beliefs that “it stood them upside down. He couldn’t have that—wouldn’t stand for it. He was psychologically—spiritually—revolted by my attempt to shatter his icons.”

Collins (1988, p. 196) maintains that righteous anger in such cases is not a calculated decision based on utilitarian calculations of what damage the violator has done but a spontaneous reaction which is “symbolic” inasmuch as it is designed to hurt the violator. For example, Smith (1978) reports that when he presented a group of senior APS officials with his survey findings that graduate recruits were disenchanted with the public service, “the initial reaction was outrage intermingled with threats to flog ‘those whingeing bastards’ upon returning to work” (p. 311). By expressing their displeasure with the APS, these recruits had shown disrespect toward an object that was held “sacred” by the senior officials and thereby provoked among them righteous anger and a punitive response. Order takers can also provoke such a response if they fail to display ritual deference toward order givers. This is because the latter are the foci of attention during order giving rituals and become sacred objects. For example, after a graduate recruit fell asleep during a long and boring speech by a senior official at an induction course which I attended, the senior official privately informed the course convener that were he to subsequently encounter that recruit while serving on a job selection committee, he would make sure that the said recruit did not obtain the job. Others in the audience had also closed their eyes during the speech and the convener lambasted them for having done so, claiming that they had shown gross disrespect toward the senior official. He then publicly interrogated these audience members demanding to know why they had shut their eyes during the speech. It is unlikely that a student falling asleep during a lecture would occasion such an angry and punitive response within the individualistic culture of academia. A senior APS official abandoned the practice of writing an anonymous newspaper column intended to expose inefficient and unethical practices after she or he was vilified by other senior officials. As this official noted, public servants who criticize “sacred cows” are treated in the same way as “the so called witches of Salem” ("Twixt the insults", 1991). Bate (1984) likewise finds that within three British organizations, sanctions were applied against deviants who challenged authority or those who took a personal moral stand which was at odds with the “party line.” As Martin (1986) observes, “a bureaucracy is analogous to an authoritarian state in its hierarchy, its imposed uniformity of perspective, and in its intolerance of dissent” (p. 189). We may conclude that variations in social diversity within the APS generate differences between localistic and cosmopolitan cultures. These differences are manifested in three aspects of culture: attitudes toward outsiders, attitudes to membership symbols, and the reaction to symbolic violations. These cultural differences can be observed when we compare senior and junior officers, different APS agencies, central and regional offices, and the APS and academia.

Discussion

This article has argued that Randall Collins’s IR theory can identify and explain the cultural differences that are present within the APS. Collins argues that there are two major sources of cultural variation within organizations: location in order giving hierarchies (the power dimension) and variations in ritual density (the Status dimension). The former of these generates three broad cultural outlooks: those of the order givers, the middle ranks, and the order takers. These three cultures are present within the APS. Senior APS officers are proud and self-assured and identify with their official order giving roles and their organization’s ideals. They exhibit frontstage personalities. The middle ranks identify with the organization officially but enact their role in a rigid, stereotyped way without concern for its larger purposes. They exhibit a bureaucratic personality that blends the frontstage and backstage cultures. The lower ranks outwardly obey commands on the frontstage of the order giving ritual but are inwardly alienated and cynical. They exhibit backstage personalities. Variations in ritual density on the Status dimension within the APS generate differences between localistic and cosmopolitan cultures. These differences are reflected in peoples’ attitudes toward membership symbols and non-members and in their reaction to symbolic violations. Most APS agencies possess high levels of ritual density and therefore display a localistic culture. A minority of agencies possess low levels of ritual density and therefore display a cosmopolitan culture. Variations in ritual density within the APS stem mainly from differences in social diversity. These in turn stem from the following sources: differences in educational levels and in the complexity of work tasks, differences in the extent to which relationships with clients are adversarial as opposed to amicable, recruitment practices, and the diversity of social contacts. Differences in social diversity are associated with differences in organizational rank, agency location, and central versus regional office location. The senior ranks, agencies with a high level of social diversity, and/or amicable relationships with their clients and central offices have a more cosmopolitan culture. By contrast, the junior ranks, agencies with a low level of social diversity, and/or adversarial relationships with their clients and regional offices have a more localistic culture.

When examined in terms of Collins’s four dimensions of ritual density, the APS exhibits moderate ritual intensity, high social density, high ritual centrality for both order givers and in-group members (but not for order takers and out-group members), and low social diversity. This combination of high social density, low social diversity, and high social inequality should, according to Collins (1988), generate a culture of “coercive, highly visible deference rituals” (p. 224). This is indeed the case. For example, the APS is characterized by a culture of “rankism” involving visible, asymmetric deference rituals. It also manifests a “bureaucratic personality” syndrome which involves the rigid and perfunctory enforcement of rules by the middle ranks.

The cultures that have been identified in this article represent points on a continuum rather than discrete entities. Survey data from the APS consistently show that the pattern of variation in responses from the different organizational ranks display a gentle gradient rather than sharp discontinuities. This is consistent with Collins’s (2004, p. 115) argument that power rituals are a continuum between the two extremes of order giving and order taking. There are several kinds of positions in the middle between the extremes such as order transmitters, those who take orders from someone else and give orders to others below. He notes that these persons tend to blend the order givers’ and order takers’ culture into a narrow and rigid “bureaucratic personality.” Collins (2004, pp. 349-350) argues that the Status dimension is likewise a continuum, so that we can derive multiple cultures or personality types from the intersection of the two dimensions. These cultures then do not correspond to particular groups in their entirety. They are relative tendencies. As Collins (1988) observes, “A given cultural trait is found in a greater percentage of one social class than another social class, yet is not found throughout a particular social class” (p. 218). For example, while most junior APS staff exhibit an order takers’ culture, a substantial minority display an order givers’ culture. This is because they have a good chance of being promoted to more senior levels. Collins argues that in such cases, order takers undergo anticipatory socialization into the order givers’ culture. Within the ATO, for example, nomination of advancement as a career goal strongly correlates with organizational commitment (Jans & McMahon, 1988). Differences in career goals among ATO staff are associated primarily with length of service rather than with hierarchical location, because those with greater length of service are less inclined to nominate career advancement as a goal (Jans & McMahon, 1988). Such longer serving workers are mostly found at the middle rather than the junior levels. This is reflected in the fact that the ASO 1s and 2s in the ATO in 1987 had slightly higher levels of organizational commitment and career satisfaction than did the ASO 3s and 4s (Jans & McMahon, 1988). Advancement as a career goal is nominated with least frequency by ATO staff at the ASO 5 and 6 levels (Jans & McMahon, 1988). An order takers’ culture can therefore emerge at these levels. As an informant observed,

The younger ones, a lot of them were quite keen, the lower ASO levels, were keen, either because they were young or they saw themselves going up, the ones which were the real walking dead . . . were the ASO 6’s. They’d died inside.

The APS possesses different sub-cultures that arise from variations on both the power and Status dimensions. How do these sub-cultures interact? Order givers and order takers interact when performing their assigned roles on the frontstage of the order giving ritual and thereby acquire different cultures. Interaction in such instances tend to reinforce rather than to undermine cultural differences. Order takers here comply outwardly with the commands of order givers but are privately alienated from them. I found that this was the case in my workplaces. For example, although I never witnessed a workmate directly disobey an order from a superior, my workmates rarely worked to their full capacity. They would assume the appearance of working hard when their superiors were present but that as soon as the latter were absent, they would abandon this pretence. As Collins (1988, pp. 205, 212) observes, when order givers are not present order takers can resume their informal backstage role in which they privately ridicule their superiors and complain to their workmates about them. Order takers generally prefer this role as taking orders is a source of ego loss for them. Order takers will therefore seek to avoid interacting with order givers. Order takers in any case rarely interact with order givers who are located more than one rank above them in the chain of command. This is yet another reason why order givers and order takers tend to inhabit different social milieu. For example, Jans et al. (1989) report that within the ATO “most managers are organizationally and socially remote from those whom they manage” (p. 22). At junior and middle levels, the cultures of the order givers and order takers often coexist side by side, because people with good career prospects tend to undergo anticipatory socialization into the order givers’ culture. Coexistence though does not entail cultural assimilation. Collins (1988, p. 218) argues that we may conceive of Weberian status groups as being a set of symbolic cults comprising localized groups and strata that are in varying degrees, mutually exclusive and even antagonistic. Upwardly mobile careerists and their immobile colleagues, for example, frequently view each other negatively (see Kanter, 1977).

The cultures on the Status dimension of social diversity also tend to be socially isolated. This is because localists and cosmopolitans tend to work at different organizational levels, in different agencies, and in regional as opposed to central offices. Another reason why cultures on the status dimension are isolated is that for an IR to succeed, people need to share symbols in common (Collins, 2004). The most basic of all rituals is sociable conversation. People with very different cultures will find it hard to sustain a ritually intense conversation as they have few interests in common. In such cases, people will gravitate toward those with whom they share a common culture. In so doing, they voluntarily segregate into different social groups. Given this tendency and the presence of the barriers to interaction between people located in different agencies and at different hierarchical levels that were noted earlier, we find that the different cultures on both the Status and power dimensions tend to be socially insulated.

How are the members of these cultures acculturated? In the case of the power dimension, it is the experience of giving and taking orders, of exercising autonomy at work and of perceiving, and experiencing upward mobility that is critical. Survey data from the APS show that these factors are very important in shaping people’s cultural outlook (Jans et al., 1989; Jans & McMahon, 1988). In the case of the Status dimension, it is the experience of social diversity measured both in terms of the diversity of people that one encounters and the diversity of communications that one receives, that is critical. People with more diverse social contacts and who receive more diverse ideas and information will tend to have a more cosmopolitan cultural outlook. Collins (1988, p. 218-219) argues that people’s cultural outlook is a product not just of their current location on the power and Status dimensions but also of their previous experience of these dimensions and of the class culture of the people that they associate with, such as parents and friends. For example, many SES officers exhibit a backstage culture of profane language and folksy informality. Campbell and Halligan (1992) note the tendency of their Australian SES respondents to “resort to more colourful language than that employed by their opposite numbers in other countries” (p. 151). We can attribute this to the fact that the Australian SES contains a higher proportion of people with working class origins than do the senior public services of the United States, the United Kingdom, and Canada (Campbell & Halligan, 1992, p. 103).

Collins argues that successful participation in rituals has certain ritual effects. It creates group solidarity, emotional energy in individuals, symbols of social relationships in the form of sacred symbols, and standards of morality. The fact that most APS staff experience only moderate levels of ritual intensity and occupy order taking roles means that their level of emotional energy is generally low. Collins (2004, p. 108) argues that high emotional energy is a feeling of confidence and enthusiasm for social interaction. Low emotional energy by contrast is a feeling of lack of confidence, lack of initiative, and depression. In between these extremes lies bland normalcy. He argues that order givers and those who are at the center of rituals tend to exhibit high emotional energy, whereas order takers and those who are at the periphery of rituals exhibit low emotional energy. Accordingly, he claims that there is a stratification of emotional energy within society, with an energized upper class lording it over a depressed lower class, with a moderately energized middle class in between. The APS exhibits precisely such a stratification of emotional energy. For example, hierarchical rank in the ATO strongly correlates with both organizational commitment and job involvement (Jans & McMahon, 1988). Jans et al. (1989) find that only 24% of ATO staff scored high on a measure of job involvement, whereas 48% of staff scored low. I thought that many of my workmates inhabited a “living death” because they were lethargic and dispirited. Jordan (1974) likewise entitles his account of working life at the junior levels of an APS agency “Living death in the social policy section.” When I asked a newly arrived workmate what he thought of the atmosphere in our division, he simply replied “comatose.” Another one of my workmates would ironically ask his coworkers, “Are you coping with the challenge?” or would simply announce deadpan while seated at his desk “I don’t think I can stand this excitement for much longer.” My workmates exhibited a marked lack of enthusiasm for their work and little emotional energy. Emotional energy by contrast is high among senior officers. Pusey (1991) in his study of the SES reports that “over and over again the interviews were suddenly charged with a certain intensity as the respondents warmed to their own phrases about ‘the challenge’ of the work and to their own feelings of ‘success’ and personal achievement” (pp. 87-88). Many SES officers work extremely long hours and are work obsessed. As one SES officer observes, “The APS is increasingly self-self-selecting in favour of task oriented driving ‘workaholics’ for its senior members (no-one else would want the job or could cope” (cited in Jans & Frazer-Jans, 1990a, p. 25). Collins (2004) argues that such work-obsessed individuals or “workaholics” obtain most of their emotional energy from working.

Conclusion

We may conclude that Collins’s IR theory can explain the cultural differences that are present within the APS in terms of peoples’ locations on the two dimensions of power and Status. The power dimension generates cultural differences between order givers and order takers while the Status dimension generates cultural differences between localists and cosmopolitans. The higher ranks in the APS exhibit a frontstage personality; the middle ranks exhibit a bureaucratic personality that blends the frontstage and backstage cultures while the lower ranks exhibit backstage personalities. The combination of high social inequality, high social density, and low social diversity that typifies the APS generates, as Collins predicts, a culture of “coercive, highly visible deference rituals.” Manifestations of this include “rankism” and the “bureaucratic personality.” Power rituals yield a stratification of emotional energy between an energized upper class and a depressed lower class. Variations in ritual density within the APS stem mainly from differences in the amount of social diversity. These generate a difference between localistic and cosmopolitan cultures that is reflected in peoples’ attitudes toward membership symbols and outsiders and in their responses to symbolic violations. These cultural differences are associated with variations in organizational rank, educational levels, agency tasks, relations with clients, and central versus regional office location. The APS is typified by high ritual density and high social inequality. The sub-cultures found within the APS therefore mostly fall into the “high ritual density” column in Figure 2. IR theory proposes that organizations contain not a single homogeneous culture, as is often assumed, but multiple and heterogeneous cultures, because people occupy different locations on the power and Status dimensions. The capacity of IR theory to identify and explain cultural differences within the APS suggests that it can make a significant contribution to the study of organizational culture. No previous studies have used IR theory for this purpose. The neo-Durkheimian theory of Mary Douglas has been widely used to study public sector culture. This article suggests that the ideas of Randall Collins, another prominent neo-Durkheimian theorist, can also contribute to this task.