Culture, Class, and Connections

by Bonnie H. Erickson
Citation
Title:
Culture, Class, and Connections
Author:
Bonnie H. Erickson
Year: 
1996
Publication: 
The American Journal of Sociology
Volume: 
102
Issue: 
1
Start Page: 
217
End Page: 
251
Publisher: 
Language: 
English
URL: 
Select license: 
Select License
DOI: 
PMID: 
ISSN: 
Abstract:

Culture, Class, and connections1

Bonnie H. Erickson

University of Toronto

Bourdieu's analysis of class and culture errs in neglecting two im- portant aspects of social structure: social networks and class rela- tions at work. He expects high-status culture to be useful in class because it is correlated with class, but culture used at work includes both genres related to class (used in domination) and genres unre- lated to class (used in coordination). High-status culture is corre- lated with class but excluded, not used, in the competitive private sector. The most widely useful cultural resource is cultural variety, and social network variety is a better source of cultural variety than is class itself.

Bourdieu's Distinction (1984) has done much to illuminate the profound connections between class and culture. However, Distinction neglects two important aspects of social structure: personal networks and work relationships. Personal networks are a major source of cultural resources and a more powerful source than class itself, as this article will show. Social relationships at work are the fundamental site of class processes in their most direct form. In this article I report research showing the importance of structure and of a more complex view of culture and class and the relationships between them.

Bourdieu offers both a model of class structure and a life course analy- sis of class reproduction. He argues that class and culture are both verti- cally ranked in mutually reinforcing ways. The culture of the highest classes becomes the most distinguished culture, apparently because it is innately superior but really because it is the culture of those who rule. In its turn, culture is a class signal that helps to maintain class domination and to shape individual life chances, much as economic capital does.

'I thank Scott Davies, Paul DiMaggio, Charles Kadushin, Mich&le Lamont, Keith Stewart, Lorne Tepperman, David Tindall, Barry Wellman, and an AJS reviewer for their helpful comments; the core research team (Jeff Cormier, Shannon Jackson, Don Lloyd, and David Tindall) for their effort and enterprise; the Social Sciences and Humanities Research Council of Canada for financial support; and the Centre for Urban and Community Studies, University of Toronto, for providing a supportive research home. Direct correspondence to Bonnie H. Erickson, Department of Sociol- ogy, University of Toronto, 203 College Street, Toronto M5T 1P9, Ontario, Canada.

O 1996 by The University of Chicago. All rights reserved.

0002-9602197110201-0006$01.50

AJS Volume 102 Number 1 (July 1996): 217-251 217 American Journal of Sociology

Indeed Bourdieu argues that culture is a form of capital as worthy of analytic attention as economic capital itself. "Cultural capital" varies with the prestige level of one's culture: cultural capital is smallest in volume for the culture typical at the bottom of the class structure and greatest in volume for the culture typical of elites.

Bourdieu uses both economic and cultural capital to describe a two- dimensional model of class (Bourdieu 1984, pp. 128-29). The vertical dimension concerns capital volume: the higher the class, the more eco- nomic and cultural capital people tend to have. The horizontal dimension concerns the composition of capital, or the strength of cultural capital versus economic capital. For example, professors and top business lead- ers both belong to the dominant class, but cultural capital matters more in academia and economic capital matters more in the business world. At the same time, professors have more cultural and economic capital than grade school teachers, and business elites have more capital of both kinds than those lower down in the business world. This model is an overview of the class structure; in more detailed analysis, Bourdieu em- phasizes "fields," or particular social settings in which class dynamics take place. Fields vary in the relative importance of economic versus cultural capital, but people always get ranked by some combination of these. Fields vary in their cultural level and range, but, within each field, the higher ranks have more prestigious culture, and this culture helps them both to dominate and to legitimate their domination. Fields vary in the particular forms of culture that pay off, but, in each field, the relevant forms are selected from the overall cultural hierarchy in ways that favor the dominant people in the field. Thus, each field should be analyzed separately, but each field has the same logic underlying its apparently unique details (e.g., Bourdieu 1984, pp. 113-14). Although Bourdieu stresses the theoretical importance of fields, he says little about what they are or how to identify them in research.

One theoretical problem with this two-capital model is neglect of a crucial third form of capital: "social capital, a capital of social connec- tions, honourability, and respectability" (Bourdieu 1984, p. 122). But social connections are too important to ignore. Earlier research related to the topic of social connections and class has shown that people with more advantaged origins and more prestigious jobs have better networks and that having richer networks can lead to better jobs (e.g., Granovetter 1995; Lin 1982). Here I focus on the more neglected topic of networks and culture and show that network variety is strongly linked to cultural variety; indeed, networks have more impact on culture than class does. I return to this theme below, after discussing how to conceptualize culture.

A second theoretical problem concerns cultural capital itself. Bourdieu argues that genres and their contents are ranked in prestige or "distinc-

Culture, Class, Connections

tion" and higher-class people have higher-class culture. But several North American studies (notably Lamont 1992; Peterson 1992; Peterson and Simkus 1992), as well as many of Bourdieu's own results for France, show a different picture. Peterson shows that higher-status people are more likely to consume highbrow culture than are lower-status people, but only a minority of high-status people consume any particular high- brow genre. There is no one kind of taste profile that advantaged people share. Moreover, higher-status people do not limit their tastes to the highbrow; they indulge more in many sorts of culture, not just the most elite forms. They are not cultural snobs but cultural "omnivores" (Peterson and Simkus 1992; Peterson 1992). Lamont (1992) shows that Amer- ican upper-middle-class men make less use of high culture than do their French counterparts when constructing social boundaries and that both sets of people use several valued forms of culture. Drawing on these contributions, I argue that cultural inequality is not so much a hierarchy of tastes (from soap opera to classical opera), as it is a hierarchy of knowledge (from those who know little about soap opera or opera to those who can take part in conversation about both). Those who have many cultural weapons can find one to suit the battle at hand, whether in the business company or in social company. Those who are culturally adroit will know when display of a particular cultural advantage will actually be an advantage or instead will show that one is not a competent player in the ongoing game. Thus the most widely useful form of cultural resource is cultural variety plus the (equally cultural) understanding of the rules of relevance.

In this article, I will only consider cultural resources in the sense of familiarity with various cultural genres. While Bourdieu includes many other aspects of culture such as taste and possessions (Lamont and Lareau 1988), and others too have discussed the complexity of cultural repertoires (Swidler 1986), my primary concern is forms of culture that can be used to advantage in seeking a better class position or conducting class rela- tionships. These uses are social: people make a better or worse impression in job interviews, in social relations on the job, or in building up social networks that can help in doing jobs or getting them (Granovetter 1995; Lin 1982). Sheer familiarity with culture is enough here. One need not like a genre or own any related artifacts in order to join in a conversation about it, and familiarity is the most portable and controllable form of culture, always available if needed or suppressible if inappropriate.

The argument in Distinction (Bourdieu 1984) also suffers from mis- matches between theory and the type of data available to test it. Bour- dieu's theoretical discussions emphasize the importance of fields and class struggles within fields, where fields are specific concrete social settings where class struggles take place. But, aside from anecdotes, Distinction American Journal of Sociology

is built on survey data relating occupations to leisure lifestyles in France. Respondents from the "same" occupation must come from many differ- ent fields, yet there is no way to distinguish the fields nor to observe the class relationship within fields. I argue that one important interpretation of a "field" is an industry in a specific local market. Industries tend to be distinctive locations in social structure, with their own labor markets and product markets, their own competitions between companies for business and between people for jobs, and their own characteristic forms of company organization and relationships between people in different positions. True, an industry study is a case study, and we must be cau- tious in making wider generalizations. Bourdieu himself, however, ar- gues that the dynamics of all fields are essentially the same and that we must examine specific fields to see these dynamics clearly. Here, I will report on one industry in one major metropolitan area. The industry is a highly competitive and little regulated one, so it offers a clear and stark example of private sector class relationships.

The mismatch between theory and data in Distinction includes both a failure to study fields and a focus on leisure lifestyles. In leisure life, class is just one of many influences on lifestyles (Halle 1984), and culture is just one of many ways to draw social boundaries (Lamont 1992). Class boundaries do exist in social ties away from work (Wright and Cho 1992), but they are much blurred by the effects of other influences on social relationships. Leisure life is partially related to class, but work life consti- tutes class. Thus class relationships at work should no longer be neglected in the study of class and culture.

What part does culture play in concrete class structure in work organi- zations? To answer this question, it is essential to remember that success- ful companies must have both domination and coordination. Bourdieu stresses domination: higher classes have power and authority over lower classes. In partial agreement with Bourdieu, I argue that the culture useful in domination is some kind of culture correlated with class. The more that those who run companies use culture that they know better than their inferiors, the more their superiority is emphasized and justified; the higher people rise, the more access they have to higher-level networks and their distinctive subcultures; and the more they adopt those subcul- tures the higher they can rise (as Useem [I9841 shows for the business elite). But what will this dominating culture be? I argue that it is defi- nitely not Bourdieu's "distinction," or the highest-status forms of cul- ture, in most private sector settings. Highbrow culture is defined as an irrelevant waste of time in the private sector and is actively excluded from the workplace (Erickson 1991). High-status culture has some value at the peak of the private sector, where support for the arts improves a company's standing and key players must maintain links to different Culture, Class, Connections

kinds of companies and to government (Useem 1984). But the business elite, though powerful, is a tiny fragment of the business world. Most establishments are modest in size (Spaeth and O'Rourke 1994, p. 883), like the ones reported in this article, and most business leaders' "concerns extend little beyond the immediate welfare of their own firms" (Useem 1984, p. 3).

A competitive, peripheral private sector industry is just the kind of field where Bourdieu expects cultural capital to be relatively limited and to be relatively unimportant compared to economic capital. But his argu- ment does not lead us to expect that cultural capital is excluded. He shows a correlation of class and high-status culture in all parts of his model of the class structure, though the peak levels of culture are lower in the more entrepreneurial regions. This pattern holds in the field studied here: owners and managers know more about high-status culture than supervisors and employees do. But owners and managers also know more about business culture, which has far more direct profitability for them. So they exclude high culture in favor of business culture, from the execu- tive lunch to the business journal, thus displaying their cultural superior- ity and their work ethic simultaneously. Even in the dynamics of domina- tion, correlation with class is not at all the same thing as usefulness in class.

Domination alone, however, cannot be enough. Company rulers must coordinate and motivate the efforts of all ranks in the company, and this calls for shared culture to smooth relationships across class boundaries. Culture that has little or no correlation with class is a necessary part of class relationships. Thus, relying on correlations is doubly wrong: first, some of the culture correlated with class in a field gets used in class relations and some does not, and, second, some culture gets used precisely because it is not correlated with class. I heartily agree with Bourdieu in the universal importance of culture of some kind, but there can be no single universally useful kind. There is no one type of culture (such as cultural capital) that fuels class dynamics in all fields to at least some degree, and every field needs at least two distinct types of culture, one for domination and one for integration, one class related and one not.

Cultural capital does matter some of the time, for example, as a basis for scholastic success and marriage choices (DiMaggio and Mohr 1985). Thus a person with a well-stocked, useful cultural repertoire ought to command quite a few types of culture in order to be distinguished or businesslike or popular as required. This returns us to my earlier claim that the most widely useful cultural resource is cultural variety and that cultural variety is closely linked to social network variety. Those who interact with a wider variety of people must respond to a wider variety of culture shown by others and, hence, develop a wider repertoire of American Journal of Sociology

culture themselves (DiMaggio 1987). Now people are more varied in their culture when they are in distinct structural locations since cultural differences and structural boundaries reinforce each other (DiMaggio 1987). I define "network variety" as the number of distinct classes repre- sented in a network, using a variation of Wright's (1985) class scheme described in detail below. Compared to other class typologies, including the occupational categories that Bourdieu uses, Wright's approach brings us closer to the real gaps in concrete social structure. The greater the network variety of classes, the more varied a person's culture, especially in terms of just those cultural variations that are most related to class differences. Thus, unlike Bourdieu, I give social networks a major rather than a peripheral role in the analysis of cultural resources.

Though Bourdieu underestimates the importance of class structure in workplaces and in social networks, he overestimates the importance of class in his life cycle analysis. He also oversimplifies the different stages of class trajectories, as though every step from one's parent's education to one's current job had the same kind of relationship to culture. I will argue that forms of class need as much unpacking as forms of culture and that forms of culture have markedly varying relationships with forms of class, with other inequalities, and with the class variety in one's net- works.

Like Mannheim (1952), Bourdieu argues that cultural orientations learned early in life are unconscious, taken for granted, hard to change, and powerful in shaping responses to later experiences. For Bourdieu it follows that children learn the class-based cultural orientation of their parents and this orientation shapes the child's class trajectory. Children with higher-class parents are socialized "naturally" to like and know just the kinds of higher-status culture that schools teach and reward. Hence such children tend to do well in long educational careers, which in turn add still more to their cultural capital. As people move from school to work, those equipped with better degrees and more cultivation get the better jobs.

This story overestimates the lifelong influence of parents' class. Bour- dieu thinks that children develop a deeply ingrained, largely unconscious orientation (habitus) that shapes all their outward manifestations of taste. Bourdieu acknowledges that people do pick up new cultural baggage as they move through life and that they experience this as the free choice of things they naturally like but that they like just those things that fit with their habitus. So they learn new things but the same kind of things, so their cultural capital stays much the same. If they try to increase their capital deliberately (e.g., by going to a wine seminar or taking up golf), they cannot do so naturally enough to make a socially convincing display Culture, Class, Connections

of taste more distinguished than that learned at home. Culturally speak-

ing, family is destiny.

Family is not destiny, however, in a rapidly changing society in which class structures and cultural possibilities both change considerably within one generation, so that the parents' cultural framework seems out of date, nor is it destiny in a society in which children gain massive cultural infusions from schooling that is longer and more important to life chances than their parents' educations (Hunter 1988). Neither is culture as im- mune to conscious manipulation as Bourdieu implies. Major life transi- tions, especially the transition to adulthood, can shake up old assump- tions and offer a "fresh encounter" (Mannheim 1952) with a range of new choices.

Instead of arguing that culture is a unidimensional matter deeply im- printed in early life, I argue that culture includes many genres learned at different times of life. For example, people often become sports buffs in childhood and adolescence, learn much of what they know about highbrow books or art in school but get their knowledge of current busi- ness trends or restaurants in the present day. It follows that one's family origins, schooling, past class locations, and current class have different relative importance for different genres.

As well as overestimating the formative influence of parental class, Bourdieu's life cycle analysis underplays the importance of forms of in- equality other than class. As Ha11 (1992) argues, gender and ethnicity are important independent bases of culture, including many varieties of cul- ture other than cultural capital. All major forms of inequality generate group boundaries that maintain cultural differences. All tend to generate cultural advantage of some kind for dominant groups, who will know more about some forms of culture that will be strategically useful just because more powerful people are more involved in them. All forms of advantage also tend to generate networks and hence greater cultural variety. Multiple forms of inequality are especially important given my emphasis on multiple cultural genres, since each genre can correlate with inequalities in different ways.

One ironic and important consequence is that genres that bridge gaps for some forms of inequality can simultaneously reinforce other forms. The research reported below shows that sports is a useful cross-class coordinating genre, popular in all class levels and widely seen as some- thing in common with others at work. Sports discussions help to build cooperative ties across class levels. But, at the same time, women and foreign-born people know much less about sports, so they are margin- alized in the informal networks that both keep companies integrated and help further individual careers.

American Journal of Sociology

A REVIEW OF THE ARGUMENT

The preceeding critical dialogue with Bourdieu leads us to a different view of social structure and culture and of individual trajectories and culture. Complex modern societies include many important forms of in- equality including gender and ethnicity as well as class. Each form goes with a different set of social boundaries and social networks, which sup- port a different set of relevant cultural differences. There is no single cultural hierarchy that correlates with all forms of inequality. Instead, genres that help to distinguish levels of one kind of inequality can cross boundaries and unite levels of another inequality. Sports, for example, divides men from women but unites men of all classes. Further, each form of inequality is itself a very complex structure including many differ- ent social settings. Though we speak of "the" class structure, the world of work includes multiple industries and markets with varying cultural distributions and rules of relevance. In some settings, such as networks of writers (Anheier, Gerhards, and Romo 1995), high-status culture is important in distinguishing social positions and rank. But in many oth- ers, including the vast mass of private sector companies below elite level, high-status culture is defined as an irrelevant waste of time; business culture prevails and distinguishes. Further still, within a single field, there will always be more than one kind of culture that matters in the dynamics of inequality. There must be relationships between unequal groups as well as boundaries between them, thus many actors draw on culture that crosses group boundaries as well as culture that differs be- tween groups.

From the individual's point of view, it follows that the most useful cultural resource is a little working knowledge of a lot of cultural genres combined with a good understanding of which culture to use in which context. Equipped with cultural variety and the rules of relevance, a person can navigate successfully in many settings; equipped with vast amounts of high culture alone, a person would be shipwrecked in many social seas. Cultural variety does not come primarily from class, or any other single kind of attribute or social location, since each of these is related to just some forms of culture. Instead, the most powerful single teacher of cultural variety is contact with people in many different loca- tions: network variety builds cultural variety. Advantaged people, in- cluding higher-class people, will certainly have better cultural resources, but this is not because of their class as such but because of the diverse networks that advantaged people have.

As people move through their life trajectories, they move through dif- ferent social settings in which different kinds of culture are salient, so there is no one life stage that dominates in learning all cultural genres.

As they move they continue to learn; the early influence of family of origin is just one influence among many and not so powerful overall as later effects of education and adult social networks.

I tested these arguments in a highly competitive private sector indus- try, the Toronto security industry. In this article, I first discuss measures of familiarity with several genres, how these genres relate to class, and the degree to which they enter the workplace. Second, I discuss my measure of network variety and its relationships with both class and culture. Finally I explore the possible roots of culture in early life (family background, other forms of inequality, and education), in past work career, and in current class and social network. I expected to find many useful forms of culture not one, many variations in how forms of culture and forms of inequality are related rather than one master trend of higher-status culture going with higher status throughout life, and a ma- jor connection between cultural diversity and network diversity.

METHODS

The research setting is the private contract security industry in Toronto, Canada. Following earlier research (Shearing, Farnell, and Stenning 1980), this study defined the industry as all companies offering security services on the open market. This does not include police personnel work- ing during off time for private employers, nor ordinary locksmiths, nor in-house security services that some companies provide for themselves. In the Toronto area, Canada's largest market, there are over 300 security companies offering a wide range of security guard, private investigation, and security hardware services. Guards work in construction sites, com- panies, malls, condominiums, housing projects, armored cars, and other sites, or respond to alarms. Private investigators do an even wider range of work including background checks on potential employees, checks on insurance claims, going undercover in warehouses with suspiciously high loss rates, tracing missing persons, and detecting shoplifters. Security hardware ranges from simple home alarms to elaborate systems of video surveillance, card access control, motion sensors, and so on. Most secu- rity companies are small (the median staff size is 10) but vary from one-person firms to firms with over 1,000 employees. Private security is a substantial business. Guards and investigators alone outnumber police by more than two to one (Campbell and Reingold 1994).

Although this is a case study of a single industry, the security industry is fairly typical of the intensely competitive peripheral sector of private enterprise. There is very little government regulation or self-regulation of private security companies in Ontario; there are no industrywide stan- dards nor means of enforcing them; levels of training are low; personal American Journal of Sociology

job security and prospects for advancement are weak; almost anyone can start a security company, and small firms spring up and die at high rates.

Fields in the competitive private sector may well differ from others in the way that culture works. In particular, such fields restrict the scope for high-status culture. First, the relentless pressure of the market forces company owners and managers to focus on business and exclude irrele- vant displays of cultivation. Second, the shortage of larger firms with internal labor markets restricts the use of education as a basis for hiring or promotion. If education has less value, there is less incentive to signal better education by displaying the higher culture that education helps to teach. Security in particular is a business service with considerable reve- nue from other organizations, and this also encourages industry owners and managers to stick to business as they work. Security workers may have more scope to display their culture away from work, but they have little contact with coworkers off the job; 53% reported that they had seen no fellow workers away from work in the previous month.

At the end of this article, I discuss how the game of culture may differ in other sorts of fields. Meanwhile, I note that this kind of field is important. The service sector is now the largest part of the economy, and the competitive periphery of the whole private sector employs substantial numbers of people. If high-status cultural capital is excluded from such fields, rather than simply playing a lesser role than elsewhere, this is a social fact with considerable empirical scope and theoretical importance. Further, low-status service industries represent capitalism in one of its most unfettered forms. Where once we studied factories to see class strug- gle in achetypal form, now perhaps we should study "McJobs" in the sweated services to see exploitation at its clearest.

My research team developed a list of security companies from the Yellow Pages, a listing of licensed guard and investigator firms, and preliminary telephone checks. Just over 50% of firms still in business at the time of our survey (161 companies) cooperated. In 150 companies, at least one owner or senior manager gave a personal interview, so our access to the upper classes of the industry was good. Access to others was more variable: some companies allowed no further access, some let us solicit volunteers through notices, some let us approach employees directly at work, and some let us take random samples from complete personnel lists. Thus the sample is probably more representative at higher-class levels, though the respondents are a diverse group drawn from all class levels and types of security work. The 393 respondents included 154 employees (70% men), 46 supervisors (69% men), 80 manag- ers (91% men), and 112 owners (94% men). The survey was conducted between May 1991 and January 1992.

Life course variables include parental education and class, respondent

education, past work career, and current class. Parental education and class variables have high levels of missing data but are important in Bourdieu's argument, so I included one variable with relatively good response rates for each parent. Mother's education is the number of years of formal schooling she completed. The rate of missing data for this item is 22.7%, but the rate is much higher for mother's class, since half of the mothers were housewives. Father's class is whether or not he was ever self-employed. The rate of missing data is 9.7%, while 23.9% have miss- ing data for father's education. I found very similar results whether these variables were included or omitted to increase the number of cases available. Turning to the respondent, education was the number of years of formal schooling completed. Past working life includes years of experi- ence in the security industry plus prior class, or whether the respondent was ever an owner or ever a manager before starting work at his or her current company. Current class is self-reported position as an owner, manager, supervisor, or employee. More complex measures (as in Wright 1985) are not necessary in this industry, where class position is very clear within companies and quite comparable between them. Other forms of inequality considered are race, nativity, and gender. Respondents were coded nonwhite if the ethnic group they felt most a part of was Chinese,

Japanese, Native Canadian, West Indian, African, Korean, Vietnamese, East Indian, or Pakistani. Nativity is the country of birth, coded as a dummy variable for Canada (native born) versus any other country. Gen- der was scored "1" for males, "0" for females. I also include age as part of early life in the sense that older people have had longer to build up cultural knowledge and social networks.

I developed a new measure of network variety for this study. Network variety is important here because contact with different types of people includes contact with different types of culture, so it is crucial to define "different types of people" in some way corresponding to cultural differ- ences. Since cultural differences go with distinct locations in social struc- ture (DiMaggio 1987) and one of the most consequential forms of social structure is class structure, I chose to focus on classes and class fractions. Respondents reported whether they knew people in each of 19 categories inspired by Wright's (1985) three major class dimensions: control of prop- erty, control of organizations, and control of skill. Interviewers instructed respondents "Now I am going to ask you whether you know anyone in a certain line of work at all in the Toronto area, for example, whether you know any lawyers. Please count anyone you know well enough to talk to even if you are not close to them." If the respondent reported knowing anyone in a category, the interviewer also asked whether he or she knew a close friend in that category and whether he or she knew a relative. This allows measurement of variety in both weak ties (knowing

anyone at all in each category) and in stronger ties (knowing a close

friend or a relative). The categories included one for control of property

(business owners outside your own company) and two for control of orga-

nizations (business managers who run an establishment other than your

own company and supervisors). Since people who control some kind of

skill form a large and varied part of the workforce, I used more categories

for this dimension than for the others, and I emphasized kinds of skill

with some relevance to the security industry. Categories included eight

professions and semiprofessions (lawyers, doctors, engineers, professors,

school teachers, bankers, insurance brokers, and accountants), four blue-

collar trades (carpenters, electricians, locksmiths, and plumbers), and

four police ranks (constable, sergeant, detective, and inspector).

This approach to network variety borrows Lin and Dumin's (1986) format but uses more structurally oriented content. Lin and Dumin asked about contacts with people in occupations with varying prestige levels. Their version is consistent with Lin's (1982) theory of networks and access to resources but is less appropriate for a study of structure and culture. Broad occupational titles combine many different particular jobs in distinct structural locations, and occupational prestige levels combine many different occupational titles, so prestige levels do not correspond to structural differences. Wright's class approach emphasizes location in structures of domination and exploitation and has closer correspondence with real cleavages in social structure, including both divisions at work and divisions in friendship (Wright and Cho 1992).

Quite independently of my research and using a somewhat different rationale, Peterson and Simkus (1992) developed another set of occupa- tional groups meant to correspond to cultural differences. They "grouped together occupations that involve the same job conditions and the same level of requirements for social and cultural skills" (Peterson and Simkus 1992, p. 155) and show that these groups are indeed quite varied in their tastes. Fortunately, there is considerable overlap between their occupa- tional status groups (Peterson and Simkus 1992, p. 156) and the compo- nents of my network variety measure. I include examples of at least six of their groups (higher and lower cultural, higher technical, higher and lower managerial, higher sales, skilled manual, and protective service), and these groups cover much of the cultural range while being somewhat biased toward the more culturally diversified groups (see Peterson and Simkus 1992, pp. 158, 162). This gives us some empirical, as well as theoretical, reason to believe that my measure of network variety is also a measure of contact with a variety of cultural repertoires.

To measure knowledge of genres, I drew upon results from an earlier study of class and culture in Toronto (Erickson 1991). That study ex- plored locally relevant measures of knowledge of four genres (books, art, movies, and restaurants) prominent in Bourdieu's work. Erickson found that these genres did vary with class but were excluded from private sector workplaces, where a businesslike culture prevailed. I decided to modify the earlier genre scales to include more items possibly related to business in general and the security industry in particular. Further, I wanted to go beyond the earlier study's focus on cultural capital, both to reduce the discomfort felt by some respondents (Erickson 1991, p. 264) and to probe the role of culture not related to class. I developed new items in consultation with a key informant who had many years of varied experience in the security industry.

For books, I kept two Canadian and two English classics but added more popular thrillers, science fiction, and action books, plus two books currently in vogue with business people. For restaurants, I created a largely new scale with more low-cost and large-scale restaurants than the original. I asked whether respondents had heard of or read each of 13 books (reading a book scored "2," hearing of it scored "1," neither scored "On). Details of all the items used will be discussed below in the results section. I asked whether people had heard of or been to each of 12 restaurants (going to a restaurant scored "2," just having heard of it scored "1," and neither scored "On). For art, I used 10 of the 20 artists from the earlier study, which found the shorter version reliable (having heard of an artist scored "1," not having heard scored "0"). I dropped movies since this scale had the weakest relationship to class and the worst reliability (Erickson 1991) to make room for two new genres. One was magazines, a selection of business and high-circulation magazines in- tended to probe both business and popular culture. I asked whether respondents had read each of 15 magazines in the past six months (yes scored "1," no scored "0"). The second new genre was sports, which I expected to be the topic most often discussed at work. I asked whether people had heard of 15 prominent athletes from a variety of sports (yes scored "1," no scored "On). The five genres chosen vary in their level of cultural capital, from high for arts and books to low for sports.

To probe the expression of culture in the workplace I coded responses to the open-ended question, "What interests do you have in common with people at work?" While respondents could have interpreted this question in many ways, their replies suggest they answered in terms of shared tastes and activities insofar as these were revealed through social interaction at work. Table 2 below gives a complete inventory of shared interests, which are all suitable topics for conversation. Some respondents clearly indicated that they were thinking of things they talked about at work. (The reported telling, discussing, communicating, talking, etc.)

Many of the variables were skewed and suitably transformed before computing correlations or multiple regressions, but tables showing mean American Journal of Sociology

class differences are left untransformed for easier interpretation (the transformed versions of these tables are not reported since they give the same results in a less approachable form). Numerous plausible possible interaction effects and curvilinear effects were explored but did not prove significant.

RESULTS

Genres and Class

Cultural classification systems vary, depending in part on social structure (DiMaggio 1987). Each field may have its own way of packaging cultural items into related sets or genres. Thus I began my analyses of culture by using factor analyses to identify related clusters of the survey items tap- ping knowledge of sports, art, books, restaurants, and magazines (details available on request). Then I summed the scores for items in each cluster to form scales of genre knowledge; most of these scales have good reliabil- ity, aside from one with a deplorable Cronbach's a of .44 (see table 1). I explain the content of these scales later in this section.

TABLE 1

Genre Employee Supervisor Manager Owner F qZ a

Sports: Popular .......... Esoteric ..........

Art: Mainstream ..... Esoteric ..........

Books: Books ............ Canadian ........ Action ............

Restaurants: Better ............ Chains ...........

Magazines: Business ......... Electronics ...... Popular ..........

NOTE.-Minimum row N = 380. Each row gives class means for one scale of knowledge of a genre, and F-test for the significance of one-way analysis of variance of class differences, qZto indicate the strength of class differences, and Cronbach's a as a measure of the scale reliability.

*P < .lo.
**P < .05.
***P< .01.

Table 1 shows that the security field includes many well-defined genres, including four that are related to class (q2 = .13-27) and many that are relatively weakly related to class (with q2 = .06 at most) or not related at all. So far as correlation with class goes, several genres could play a part in domination, and several could play a part in coordination, but only some do get used, as the next section indicates. To understand which get used, we must go beyond correlation alone.

Cultural candidates for use in domination include four class-related

scales. Two of these touch on core elements of cultural capital in Bour-

dieu's sense: art and books. "Mainstream art" includes Robert Bateman,

A. J. Casson, Salvador Dali, Cornelius Krieghoff, Claude Monet, Henry Moore, and Andy Warhol. These are relatively widely known, conven- tionally approved artists who appear often in the media and in museums. "Books" includes rather widely known books, whether classics (Pride and Prejudice, Oliver Twist) or recent best-sellers (Taipan, Burden of Proof) or a popular business success book (The Search for Excellence). Owners and managers know more about books and about art than do the supervisors and employees working below them, but, as I argued above, I do not expect the dominant classes to use this particular cultural advantage in the workplace, where high-status culture is out of place according to the dominant profit-oriented business culture that the own- ers and managers themselves insist upon.

More relevant, and even more strongly related to class, are two other scales. "Better restaurants" range from some chains that rise a bit above fast food (Fran's, The Keg, Licks, Old Spaghetti Factory), through a large place with conventional food and moderate prices (Ed's), to one formerly popular with the elite (Winston's). Higher-class people in the security industry have more discretionary time and money for frequenting such better restaurants and can use them for doing business. "Business magazines" include magazines aimed directly at the business leader in general (the Economist, Financial Times, Forbes, Harvard Business Re- view) or in the security industry specifically (Canadian Security, Security World). Like knowledge of better restaurants, familiarity with these mag- azines has direct business usefulness and can signal competence plus businesslike orientation simultaneously. These two genres are much more plausible candidates for use in domination than are books or art.

Turning to coordination, we find plenty of genres with little or no relation to class, the criterion I stressed above. Closer inspection suggests some other criteria that help us to understand which of these genres actually get used to smooth relations in the workplace. Many of these genres are useless because all classes are equally ignorant about them. "Esoteric sports" includes the athletes Ray Bourque, Rob Boyd, Matt Dunigan, Roddy Piper, Aryton Senna, Isaiah Thomas, and Tim Wal- American Journal of Sociology

lach. This is a scale of relatively esoteric knowledge typical of the true sports fan or the fan of more obscure sports. The average respondent has heard of three or four of these seven athletes, compared to six or seven out of the eight on the "popular sports" scale described below. Other genres are even less well known. "Esoteric art" includes Helen Franken- thaler, Michael Snow, and Mary Pratt. While this is a poor scale (a= .44), it seems to include critically acclaimed but poorly publicized artists. All classes had heard of a mean of about one-half of these three artists, compared to means from three to five on the "mainstream art" scale with a maximum of seven. Means (see table 1) are also low compared to scale maxima for "Canadian books" (Sunshine Sketches of a Little Town, The Stone Angel, Solomon Gursky Was Here; maximum score = 6); "action books" (The Art of War, Dune, the Foundation trilogy, Silence of the Lambs, Tommyknockers; maximum score = 10);"electronics magazines" (Popular Electronics, Radio Electronics; maximum score = 2); "popular magazines" (Macleans, Playboy, Reader's Digest, Scientific American, Sports Illustrated, Saturday Night, Time; maximum score =

7).

A popular, cross-class genre useful in coordination must be popular in two senses: little related to class and well-known within each class. This leaves us with two likely candidates: popular sports and chain restau- rants. "Popular sports" includes athletes very well known in the Toronto area because they were Canadian success stories (Kurt Browning, Eric Lindros, Elizabeth Manley, and Razor Ruddock) or played for major Toronto teams (Roberto Allomar, Kelly Gruber, and Rocket Ismail), or were famous everywhere (Martina Navratilova). The "chains" scale is based on Burger King, Country Style Doughnuts, Kentucky Fried Chicken, McDonald's, Tim Horton's Doughnuts. All classes score over 11 out of a possible maximum of 12. Cheap fast food restaurants are not only popular with most people but very popular indeed among security people, who work irregular hours for generally poor pay and who often want anonymous places for quick or covert meetings. However, chains lack one more thing useful in a coordinating genre: something to talk about. Chains are so standardized that their conversational possibilities end quickly, but popular sports offers new developments each day as well as a rich history.

I conclude by pointing out that the pattern of table 1 alone can be read as fully consistent with Bourdieu's expectations. Higher-class people do know more about some major forms of cultural capital. At the same time, being in a field where economic capital counts far more than cul- tural capital, the level of high-status cultural knowledge is somewhat limited. People know more about the more popular artists, books, or athletes than they know about the more esoteric items familiar to a real connoisseur of such genres. This parallels the findings of Peterson and Simkus (1992, pp. 162-63) on protective service occupations, which had below average rates of arts activities. Further, in the security industry class differences are more striking for business genres than for more cultivated ones. But the mere associations between class and culture do not suffice to tell us how culture will actually get used in class relations. For this, we must add other considerations, including some additions to my initial emphasis on the need for both culture related to class and culture unrelated to class.

Domination calls for genres that are both correlated with class and defined as relevant; in the security field, this means businesslike topics and not cultural capital, even though both are class correlated. Coordina- tion calls for genres that are well-known in all classes, nearly equal or equal between them, and rich in content; in the security world, sports is the most likely candidate we have seen.

Inclusion of these genres in the survey was no accident: their key role in the security world was strongly suggested by preliminary field observation and discussions with key informants. The next section of this article turns to more systematic data from the survey.

Genres at Work

One of our open-ended survey questions asked "What interests do you have in common with people at work?" People could only tell us of shared interests they knew about, that is, tastes and activities revealed through conversations at work, which is just what we would like to know about.

Many employees could not think of anything in common with cowork- ers (table 2). The alienating effect of poor working conditions and re- wards, combined with limited social contacts, keep some employees from engaging in much cultural expression of any kind on the job.

Work is the most common shared interest, consistent with the private sector business culture that regards anything but work as a waste of time (Erickson 1991), but classes vary dramatically in both level and type of shared work interest. Employees do the least rewarding work and are the least likely to see work as a common interest. The common interest is sometimes a negative one ("telling horror stories"). Their comments also tend to focus on their own jobs and the people they deal with directly ("providing good service," "watching each other's backs") with only two references to the company as a whole.

On the other hand, many owners and managers see the work itself as a major shared interest, described more positively and with more involvement. Both owners and managers often speak of teamwork American Journal of Sociology

TABLE 2

RESPONDENTSREPORTINGTOPICSAS THINGSIN COMMONWITH PEOPLEAT WORK. BY CLASS(%)

Topic Employees Supervisors Managers Owners N

None* .............................
Work ..........................
Sports .............................
Family ............................
Hobbies .........................
Politics ...........................
Life ................................
Social life .........................
Books .............................
Music ..............................
Policing ...........................
Outdoors .........................
Movies ............................
Ideas and values ................
Homes ............................
Humor ...........................
Dogs ..........................
Dining .............................
Shared background ............
Recreation ........................
Entertainment ...................
Computers .......................
People .............................
Television ........................
Travel ............................
Art .............................
Writing ..........................
Shopping .......................
Pets ................................
Self-defense ......................
N ...................................
No response .....................

*Twenty of the 23 employee nonresponses are blanks in self-administered questionnaires. and some may be equivalent to "none." Eight of the 14 owner nonresponses come from owners of one-person companies and are equivalent to "not applicable."

("We've got a good team here"). of sharing company goals and profes- sionalism. and of positive expectations of their futures in the industry . Their interests in business magazines and better restaurants are sub- sumed under their interests in work and do not get much attention as topics in their own right . Very few people at any class level see dining as a common interest. and none mention magazines. but both table 1 and

field observation show that owners and managers are active consumers of

business publications and better restaurants for business meetings.

Cultural capital makes only rare appearances in the security industry. Table 2 shows that only one person mentioned art as something in com- mon (this unusual person's mother once had an art gallery). Only 14 people mentioned books, and higher-class people were no more likely than others to mention books. Some of the other topics mentioned, like music, may reflect cultural capital, but these also are rarely mentioned, and there is no tendency for higher-class people to mention them more often. Of all the genres correlated with class, only business genres are allowed much expression because only these are seen as relevant in the eyes of the powerful.

Among genres found uncorrelated with class (table I), sports is strik- ingly popular. Sports are the most commonly mentioned shared interest after work itself and are mentioned far more often than other nonwork genres. The topics that come closest in popularity are well behind in frequency and most are not really genres ("family," "life," and "social life" concern the particulars of personal lives and "hobbies" includes a wide variety of different hobbies, with all but "cars" mentioned only once or twice). Sports is both remarkably popular and remarkably evenly spread across classes, with about one-fifth of the people in each class naming sports as something in common with coworkers. It is the lack of strong class effects and the general popularity of sports that are critical here: sports is so much a cross-class widespread interest that it is very useful in tending work relationships between or within classes.

Thus the correlation of culture with class is no predictor of the use of culture in class relations at work. Higher classes know more about some genres, like art, which they never mention at work. They know little if anything more about another genre, sports, which they and indeed all classes discuss often precisely because it is a classless topic in which all can participate equally. The distinction of a genre is also no guide to its use: books and art are distinguished genres but excluded from the work- place, sports is a popular genre but included in many workplace interac- tions.

Although the results reported in this section are striking, they are based on a single survey question; there is clearly a need for more research on how cultural resources are used in the working world, research including both more varied survey questions and field observation. Meanwhile, the survey results for this industry are consistent with my observations at three industry conferences, my discussions with several widely experi- enced key informants, and the project interviewers' observations through interviews, including many conducted at work sites.

American Journal of Sociology

Network Diversity and Class

Cultural variety is an important resource, since different types of culture have different uses: cultural capital pays off in education and in some kinds of work; business culture and cross-class sports culture pay off in a competitive private sector field. Cultural variety is related to class: higher-class people know as much or more than others do about all the genres examined here, but much of what seems to be individual class effects may be network effects in disguise, since class level and network variety are correlated. People in higher positions work more with people instead of things, span boundaries within and between work organiza- tions, and take part in more social settings away from work, so their contacts are more wide-ranging. Prior surveys consistently show that network size and diversity rise with social status (e.g., Marsden 1987; Lin and Dumin 1986).

Previous approaches to network diversity, however, are not structural enough for work on the connections among class, culture, and networks. Previous surveys measure network variety in terms of the variety of occupational prestige levels among those in one's network. Prestige levels combine many occupations, and occupations themselves combine many different sorts of jobs in distinct parts of the working world. Prestige levels do not identify distinct social settings that can maintain cultural differences, and prestige levels do not correspond well to class locations, so they can tell us little about interclass relationships. Thus I prefer to define network diversity in terms of the number of distinct class fractions to which a person has ties (see the methods section above for details).

Most prior surveys are also limited to a small number of close ties such as several intimates (Wellman 1982), several confidantes (Marsden 1987), or a few dozen potential supporters (Fischer 1982). Intimate ties are just a small fraction of a network, which averages about 1,400 people in North America (Killworth et al. 1990). Worse still, intimate ties are the ones least likely to provide a culturally stimulating flow of nonredundant information. The closer people are, the more similar they tend to be in their social locations and their resources (Granovetter 1973; Burt 1992), so it is the weaker ties that connect us to the greatest variety of classes and of culture.

The connection between class, strength of tie, and network diversity is shown in table 3. Note that weaker ties give substantially greater access to a variety of classes. People have relatives in only about two of the classes included in our survey item. They have friends in twice to three times more classes than they have relatives. And they have ties of any strength, including their weakest ties, in twice as many classes as they can reach through friends. Note also that class differences in net-

Culture, Class, Connections

TABLE 3

CLASSAND NETWORKDIVERSITY:MEAN NUMBER OF CLASSES IN WHICH RESPONDENT

KNOWSSOMEONE,BY RESPONDENT'S

CLASS

Current Class Knows a Relative Knows a Close Friend Knows Anyone

Employee ..................... 2.3
Supervisor .................... 2.1
Manager ...................... 2.3
Owner ......................... 2.5
F ............................... .50
N ............................... 384
q2 .............................. .oo

work variety also vary with the strength of the tie considered. People of all classes have about the same number of kin and all tend to have kin similar to themselves, so all classes are equally restricted in the range of their kin ties. The structural advantages of higher-class location show up more for friends and, most of all, for the whole network. Surveys confined to stronger ties greatly underestimate the extent to which class is correlated with larger, richer networks.

Network Diversity and Genres

If weaker ties include contacts with a wider range of classes, they must bring wider information about those genres that are related to class. Table 4 shows simple correlations between familiarity with each of the subgenres introduced in table 1 on the one hand, and network diversity of all three kinds (kin, friend, anyone) on the other. The top section shows those subgenres that we found more related to class. The results are powerful, simple, and consistent: network diversity is positively re- lated to familiarity with all five class-linked genres, and the relationship gets stronger as the tie considered gets weaker. If we consider the full range of networks, not just the intimate centers, then who you know and what you know are strikingly correlated.

The rows under "other genres" in table 4 show those subgenres that we found to be less related to class and, indeed, not related to class at all in most cases. Most of these subgenres are less related to network class diversity than are the corresponding class-linked subgenres. The pattern is particularly clear for the most important form of network diver- sity, class diversity in the whole network. The shared variance (R2)is .07 for popular sports but .03 for esoteric sports; .l3 for mainstream art but .04 for esoteric art; .14 for books but .02 for action books; .20 for American Journal of Sociology

TABLE 4 CORRELATIONS DIVERSITYAND KNOWLEDGE

BETWEEN NETWORK OF GENRES NO.OF CLASSES RESPONDENT

IN WHICH

KNOWSSOMEONE
KNOWLEDGEOF GENRES Knows a Relative Knows Close Friend Knows Anyone
Class-related genres:      
Popular sports .................. .14*** ,IS*** .27***
Mainstream art ................. .11** .12*** .36***
Books ............................. .06 .19*** .37***
Better restaurants ............. .21*** .31*** .45***
Business magazines ........... .lo** .31*** .41***
Other genres:      
Esoteric sports .................. .12*** .09** .18***
Esoteric art ...................... .05 ,14*** .21***
Canadian books ................ .08* -.01 .lo**
Action books .................... .11** .05 .IS***
Chain restaurants .............. .08* .07* ,18***
Electronics magazines ........ .lo** .09** .05
Popular magazines ............ .05 .18*** ,21***

NOTE.-These are pairwise correlations; minimum N = 376; all the signihcance levels are one-tailed.

*P< .lo. **P< .05. ***P< .01.

better restaurants but .03 for chains; .17 for business magazines but .04 for popular magazines.

Thus the class-linked genres are linked not just to personal class loca- tion but also to personal contacts ranging throughout the upper reaches of the class structure. The network of ties that links relatively advantaged people also nourishes the varieties of culture that advantaged people more often know, and this is true for all five of the more class-linked subgenres, even if they are not very strongly linked to class at the individ- ual level (as for sports) and even if they are not distinguished (as for sports, restaurants, and business magazines). Diverse networks encour- age cultural omnivores, not specialists in distinguished culture nor spe- cialists in culture specific to higher classes.

The apparent link between network diversity and cultural diversity, however, may be misleading since networks and culture both have roots in class trajectory, education, and other forms of inequality. A life course analysis of the roots of culture will give a clearer view.

Genres and the Life Course: Early Life

In criticizing Bourdieu earlier, I argued that he overestimates the effect of parental class on culture, that he treats different aspects of class from

Culture, Class, Connections

TABLE 5

Early Life,
Early Life and Past Work,
Predictors Early Life Past Work and Current Life
Born in Canada ...................      
Nonwhite ...........................    
Male ..............................    
Age ..................................    
Mother's education ...............    
Father self-employed .............    
Own education ....................    
Security experience ...............    
Former owner .....................    
Former manager ..................    
Owner .............................    
Manager .............................    
Supervisor ..........................    
Network variety ...................    
Adjusted R2 ........................    
N ......................................    

NOTE.-There is no significant difference between the R2 values for early life and for early life and past work, but other comparisons are significant at P < .01.

*P < .lo. **P< .05. ***P < .01.

different parts of the life course too uniformly as though all relate to culture in the same way, that he underemphasizes the role of other forms of inequality in generating cultural advantage, and that he neglects the role of network diversity, which is a very important source of cultural variety.

Since the role of culture in class is a major concern here, I will only consider the five subgenres that are more class-related: popular sports, mainstream art, books, better restaurants, and business magazines, shown genre by genre in tables 5-9. I analyze each of these genres with a series of multiple regressions: one for variables that occur or are set early in life (place of birth, age, race, gender, mother's education, whether the father was ever self-employed, and education), one for early life plus past working life (years of security industry experience, and experience as an owner or as a manager before entering the current company), and one for early life, past working life, and current life (cur- rent class, with employees as the omitted class, and network variety). Since network variety is greatest and most powerful for the network as a American Journal of Sociology

TABLE 6

KNOWLEDGE ART: STANDARDIZED MULTIPLE REGRESSIONS

OF MAINSTREAM USING PREDICTORSFROM EARLY LIFE, PAST WORK LIFE, AND CURRENTLIFE

Early Life, Early Life and Past Work, Predictors Early Life Past Work and Current Life

Born in Canada ...................
Nonwhite ........................... -
Male .................................
Age ...................................
Mother's education ...............
Father self-employed .............
Own education ....................
Security experience ...............
Former owner .....................
Former manager ..................
Owner ............................... ,038
Manager ............................. ,093
Supervisor .......................... ,002
Network variety ................... .204***
Adjusted R". ...................... .222*** .250*** .279***
N ...................................... 291 291 291

NOTE.-All pairs of R2values differ significantly (P< .01).

*P< .lo.
**P< .05.
***P< .01.

whole, I only consider this network variable. Obviously these regressions simplify a more complex life story in which networks build culture, both help people to get better jobs, better jobs add richness to networks and culture, and so on in a lifelong set of feedbacks.

Though simplified as much as possible, the regressions include enough variables to reduce the sample size considerably (from the original 393 to just under 300; see tables 5-9). The variables with the highest missing data rates are mother's education and whether father was ever self-employed (see methods section above). These theoretically important variables must be included in the tables and including them does not change major results for other variables. In particular, network diversity is strongly related to genre knowledge with or without controls for all the other variables in these regressions (cf. tables 5-9 with table 4).

Let us first consider the relative importance of early life as a whole, by comparing the R2for these variables (the first column of tables 5-9) to the R2for all the life course variables considered (the last column of tables 5-9). Where Bourdieu tends to emphasize a pervasive effect of

Culture, Class, Connections

TABLE 7

Early Life,
Early Life and Past Work,
Predictors Early Life Past Work and Current Life
Born in Canada ...................      
Nonwhite ...........................    
Male ................................    
Age .................................    
Mother's education ...............    
Father self-employed .............    
Own education ....................    
Security experience ...............    
Former owner .....................    
Former manager ..................    
Owner ...............................    
Manager ...........................    
Supervisor ..........................    
Network variety ...................    
Adjusted R2 ........................    
N ......................................    

NOTE.-A~~ pairs of RZvalues differ significantly (P< .01).

*P< .to. **Pi.05. ***P< .01.

early life on habitus, these data show considerable differences between genres. The early life variables "explain" 80% as much as the whole set of variables for art, 74% as much for sports, 50% as much for books or better restaurants, and 41% as much for business magazines. Differences between the R2 values for early life, early life plus past work, and the full set of variables are almost all significant (see tables 5-9).

Where Bourdieu stresses the role of cultural capital in class reproduc- tion, with higher-class parents giving children higher-class culture that helps them to gain higher-class status, these data show no consistent link between the utility of a genre and the genre's rootedness in early life. Sports, better restaurants, and business magazines are all good things to know about in the private sector workplace, but sports is far more tightly tied to early life than is knowledge of business magazines. Art and books are useless in security work, but art is far more tightly tied to early life. The life course timing of genre knowledge depends on the social script for what people do at various life stages: playing sports and talking sports in youth, studying art and books in schools but eating in currently available restaurants and reading current books. Timing also depends

American Journal of Sociology

TABLE 8

Early Life,
Early Life and Past Work,
Predictors Early Life Past Work and Current Life
-----------      
Born in Canada ...................      
Nonwhite ...........................      
Male ..................................      
Age ...................................      
...............Mother's education      
Father self-employed .............      
Own education ....................      
Security experience ...............      
Former owner .....................      
Former manager ..................      
...............................Owner      
Manager .............................      
Supervisor ..........................      
Network variety ...................      
Adjusted R? .......................      
N ......................................      
NoTE.-AII pairs of R2values differ signihcantly (P <  
"P< .lo.      
""P < .05.      
***P < .01.      

on the relative importance of classic, certified culture (as in mainstream art) that holds its value over time versus culture whose value lies in being up-to-date (as in business magazines).

Where Bourdieu stresses the importance of parental class, these tables show no difference between those with a self-employed father and those without one. (In results not shown, I found no effect for father's educa- tion nor for father's class-employee, supervisor, manager, selfemployed-when the respondent was 16 years old). Class effects from early life are limited to mother's education and one's own education. Whether or not we control for later life experiences, those with more educated mothers or more education of their own know more about the more distinguished genres (art and books, tables 6 and 7), which are just those genres embedded in school curricula. More educated people also know more about the other genres, though the effect is sometimes indi- rect; for example, more educated people have no advantage in knowledge of popular sports once we control for security experience (table 5). We often think of education as a proxy for class location or as part of class reproduction but should not forget that the world of education is quite a

Culture, Class, Connections

TABLE 9

KNOWLEDGEOF BUSINESSMAGAZINES:STANDARDIZED USINGMULTIPLE REGRESSIONS PREDICTORSFROM EARLY LIFE, PAST WORK LIFE, AND CURRENTLIFE

Early Life,
Early Life and Past Work,
Predictors Early Life Past Work and Current Life
Born in Canada ...................      
Nonwhite ...........................      
Male ..................................    
Age ...................................    
Mother's education ...............    
Father self-employed .............    
Own education ....................    
Security experience ...............    
Former owner .....................    
Former manager ..................    
Owner ...............................    
Manager .............................    
Supervisor ..........................    
Network variety ...................    
Adjusted R". ......................    
N ......................................    
NOTE.-A~~ pairs of R2values differ signihcantly (P<    
*P < .lo.    
**P< .05.    
***P< .01.    

different field from the world of work and has a different set of connec- tions to culture.

Moreover, these data suggest considerable importance for other forms of inequality that operate from early life onward: birthplace, race, and gender. People born in Canada have many more years than the foreign born to learn details of local culture, so they have a considerable advan- tage in knowledge of sports, art, and restaurants. This advantage persists even when we control for later life experiences (tables 5, 6, and 8). The foreign born can catch up most easily in the codified and accessible world of print, and they know as much as natives do about books (table 7) while they actually pay more attention to business magazines than natives in comparable current locations (table 9). Nonwhites know less than whites about all five genres, though many of the differences fade when we control for work careers and social networks (tables 5-9). This pattern suggests lifelong problems of racist exclusion from workplaces and net- works rich in mainstream culture (see also Breton et al. 1990).

Men and women have different cultural specialties (Collins 1992). Some differences seem rooted in gender gaps across the life course, since American Journal of Sociology

these differences persist despite all controls: men know more about sports (table 5) and women about books (table 7). But if men read more business magazines, that is because they are more likely to be owners or managers; gender differences vanish when we control for work history (table 9). On the other hand, controls reveal one later-blooming gender difference in table 8: women know more about restaurants, allowing for their lower- class position. Dining out for pleasure is a particular pleasure for those who often cook and clean up at home, and women often manage a fam- ily's leisure social events. Although my focus is on direct class relations in the workplace and the role of wide connections to classes through social networks, leisure life also plays a part in cultural repertoires.

Age brings greater knowledge of most genres. This is partly because age gives some people time to move up to higher-class positions where business culture is more useful (see tables 8 and 9, where the age advan- tage in knowledge of restaurants and magazines vanishes when work history is controlled). Age is also useful in giving time to build up knowl- edge of distinguished genres (books and art) with deep institutional roots that help old knowledge stay valued (see tables 6 and 7, where the age advantage persists despite all controls).

Overall, the results remind us that all forms of structural division contribute to cultural differences. Class alone can never tell the whole story of culture (see also Hall 1992), and different forms of inequality have different degrees of relevance for various genres. Advantaged peo- ple tend to get advantages in genre knowledge, too, but not across the board. The effects of various forms of inequality show different patterns for each of the five genres considered above, and looking at more genres would probably reveal more variations.

Looking at inequalities other than class helps us to understand the multiple roots of genre knowledge. It also leads to some important addi- tions to my earlier discussions of the role of popular sports in the work- place. If we look at class alone, sports is a relatively classless genre useful in coordinating ties between classes. Table 5 lends further support to this view. Sports knowledge is nearly classless (it is unrelated to mother's education and father's class and weakly related to one's own education before controls while unrelated after). Sports knowledge seems to grow with chances to use it at work (knowledge is greater for those who have been in the security field longer and who have become managers, the class most responsible for interclass relations). But, if we turn from class to other forms of inequality, we see that sports knowledge contributes to domination in these even while it contributes to coordination between classes. Men and the native born know a good deal more about sports, controls or no controls, and the security industry is male dominated at every class level (the percentage male runs from 70% of employees to 91% of owners). Sports talk can link the male majorities in all classes but excludes women, which may be one more reason for its popularity in a very macho industry. Returning briefly to the question, "What do you have in common with people at work?" it is interesting to note that men cite sports twice as often as women do (27% vs. 13%) while women more often cite books (8% vs. 3%). Two other differences suggest that women are marginalized in the workplace flow of talk and culture: women more often say they have nothing in common with people at work (19% vs. 10%) and they less often cite work itself as a shared interest (38% vs. 56%). The native born are also in the majority in every class, barely (57%) for employees but more clearly (70%) for each of the three more powerful classes, and sports talk also reinforces their advan- tage in the workplace. Companies need cross-class coordination, but this does not have to include everyone if it includes groups that dominate within all classes.

Sports culture handicaps women and the foreign born all the more because, as we have seen, sports knowledge has roots early in life. True, motivated people can learn a lot about sports by simply reading the newspapers at any time in life, but a long history of attention to a genre develops a cognitive framework that makes new details easier to assimi- late, so that "just" reading the sports news is more informative for those who already know more about sports. Still more, those who start early have a great advantage in personal historical experience. They not only know that Paul Henderson scored the big goal in the 1972 Canada-Soviet hockey series, they watched him do it and remember where they were at the time, and quite a few of them were hockey players themselves or played some other sport. Native-born Canadian men can share many memories of personal sports experiences. These shared memories fuel interest in sports as a topic relevant to the self and act as a kind of ticket of entry to sports conversations by showing authentic involvement.

Genres and the Life Course: Working Life

Though parental class has remarkably little to do with genre knowledge, one's own class trajectory and working experience are more important. The number of years of experience in the security industry goes with greater knowledge of all five genres (tables 5-9). This effect is probably not specific to security work: adult employment of all kinds tends to expand opportunities to consume culture and to learn more culture from one's expanding social network. Once we control for social networks and current class, the effect of sheer experience vanishes. Beyond sheer experience as passage of time, past experience of command positions goes with more knowledge of just those genres directly useful in business American Journal of Sociology

management. Thus past owners and managers read more business maga- zines (table 9), and past managers know more about restaurants (table 81, but past owners and managers do not know any more than others do about sports, art, or books (tables 5-7). Better familiarity with business culture may help people to get or keep higher-class position and to enrich their networks, so the effects of past ownership and management vanish when current class and social network variety are controlled (tables 8 and 9).

In terms of current class location, owners and managers know more about each of the five genres we are now examining (table I), but this does not necessarily mean that higher-class people know more because of their higher class itself. To the contrary, current high class in itself should only increase the kinds of genre knowledge that owners and man- agers encounter and practice in their work. Once we control for other sources of culture, including social network variety, we find that current owners and managers know substantially more about the specifically business-related genres (better restaurants and magazines, tables 8 and 9) but not about the distinguished genres that the dominant classes ex- clude from work (art and books, tables 6 and 7). Sports knowledge is a more subtle case. On the one hand, talking sports is useful in cross-class coordination and hence useful to owners and managers, and in particular to managers, who of all classes spend the most time talking to those at their own level and those above and below as well. This is probably why managers know more than others about sports, net of all controls (table S), but, on the other hand, the very use of sports to coordinate implies that people at all levels are included in sports conversations, so the use of this genre to bridge class gaps also maintains the very low correlation between class and this genre. Class has little to do with sports, while other factors like gender, native birth, and networks matter more (table

5). Thus the strongest of the class differences in genre knowledge, that is knowledge of restaurants and business magazines, is directly related to class differences in what people do at work. What a banal finding this is, compared to Bourdieu's more subtle and counterintuitive argument relating class and distinction. Class culture is to a considerable degree actually about class, with higher-class people consuming forms of culture that have direct application in their jobs. Now that we have reviewed several different components of class across the life course, note how differently they relate to culture. Parental class had no effect on the five genres. Mother's education and one's own education facilitated learning of all five, whether directly or through later life experiences. One's own class, past or present, is directly related to only some genres: the ones that higher-class people use more at work

than lower-class people do. If we want to understand how class and culture are related, we must decompose both culture and class into more specific components.

Genres and the Life Course: Network Variety

There is only one form of advantage that is an advantage for all five genres with all other possible sources held constant. This persistently powerful source of culture is not class, nor any other form of inequality considered here, but social network diversity. In this survey we asked people whether they knew anyone at all in each of 19 locations drawn from Wright's (1985) major class dimensions: control of property, organi- zations, or skills (see details above). Each of these locations is a somewhat separate social setting with its own cultural mix. Any contact with such a setting, even if it is just an acquaintance with one person in that setting, provides a channel of access to the distinctive cultural repertoire. The more diverse the set of such contacts one has, the more variety of culture one will encounter and hence maintain or learn. Social networks are the continuing adult education of culture, and diversified networks are the liberal arts programs teaching a little of almost everything. Thus network variety is one important source of cultural variety.

The role of networks helps us to understand why culturally varied people (or "omnivores") have such different kinds of cultural variety (Peterson 1992). Omnivores are alike only in the large number of tastes and practices that they have, not in having the same tastes and practices. This leaves us wondering why one omnivore may choose one cultural mix while another chooses something quite different. Part of the answer lies in the varied networks that help generate varied culture since two people with equally varied networks can have contacts in two different combinations of structural locations, hence access to distinct cultural inputs. As Simmel (1955) observed long ago, complex modern social structures generate great numbers of social circles, and each individual has a unique combination of social circle memberships. People with many social circles have great network variety and great cultural variety but do not share any one profile of contacts nor of culture. This reinforces the value of cultural variety since those who know more have a greater chance of finding common ground with those they deal with, despite the fact that these people have no predictable set of cultural interests.

DISCUSSION

This article has expanded and modified Bourdieu's argument by looking harder at two aspects of social structure: social networks and class struc- American Journal of Sociology

ture in the workplace. At work, the uses of culture are more complex than Bourdieu leads us to expect. He expects that culture correlated with class, especially high-status culture, helps higher classes to dominate lower ones, but, in fact, correlation with class is not the same as use- fulness in class relations. Some forms of culture correlated with class do not get used in the private sector because they are profitless irrelevancies that the upper classes themselves exclude through their intent focus on doing business. High-status culture in particular is just such an excluded waste of time in the business world. Business culture itself supports domi- nation. Meanwhile, profitable work organizations need coordination within and between classes as well as domination. Culture useful in coordination is uncorrelated or almost uncorrelated with class, popular in every class, and rich enough to provide enjoyable conversation, but the culture that integrates classes may reinforce other kinds of inequali- ties. The sports talk that links native-born men in all levels of security work simultaneously excludes the female and foreign-born minorities in each class.

How might culture work in other kinds of work organizations? No doubt high-status culture plays a greater role in many organizations out- side the competitive private sector. There is more room for talk without an immediate business use in organizations less exposed to market forces, for example, government bureaucracies and universities. High-status cul- ture has more value as a signal of education in organizations that use education as a basis of admission and promotion, for example, in larger private sector organizations with internal labor markets as well as much of the public sector. High-status culture is more work related in organiza- tions that produce or market it, especially in public sector organizations like schools rather than market-sensitive private sector organizations like publishing houses, and high-status culture is more observable, hence perhaps more important, wherever people meet each other away from work as well as just on the job.

All these kinds of organizations, however, still need both domination and coordination. Academia, a familiar field that provides large scope for high-status culture for all the reasons cited above, illustrates this point. Academics do spend a lot more work time talking about cultivated topics than security owners and managers do, but academics also spend considerable time on displays of their equivalent of "business culture," for example, displays of familiarity with academic journals (instead of business magazines). Academics also engage in coordinating talk across levels from distinguished professor to humble student. Sometimes the common culture used is even sports, and sometimes women and foreign-born students feel cut out of the loop by such talk. The weight placed on different cultural genres varies, but the general use of culture in class is similar. This is a very Bourdieu-like contention since he emphasizes the underlying similarity of cultural games in superficially different fields, but the common game I describe is a more complicated one, even at the level of the individual field.

Within a field, more than one kind of culture is useful. Since fields emphasize varying genres, a varied knowledge of culture is even more useful for those who cross between fields. Those who span fields tend to be higher-class people, who more often represent their home organization to the outside world and who more often are mobile among organizations. The more varied their culture, the more effectively they can bridge orga- nizational and field boundaries. Thus the most useful overall cultural resource is variety plus a well-honed understanding of which genre to use in which setting.

To the extent that anything in my analysis replaces high-status culture in Bourdieu's analysis, my "cultural variety" replaces his "distinction," but the changes I suggest are not quite so simple. I argue that we need not one master variable for culture but several variables: high-status culture, field-specific forms of dominating and coordinating culture, and cultural variety. We also need multiple class variables since different forms of culture have differing relationships to parental class, parental education, one's own education, and one's past and present class posi- tions.

Not only is the effect of class trajectory on culture more complex than Bourdieu implies, it is weaker as well. Many genres have stronger roots in other social cleavages, including the other forms of inequality studied here: gender, race, and nativity. Since each structurally distinct group can carry a somewhat different cultural repertoire, the most consistently powerful source of cultural knowledge is contact with people in a wide variety of social locations. The strongest single source of cultural variety is social network variety. Unlike Bourdieu, who gives mere glancing references to social capital, I argue that network variety is more impor- tant than class as a source of cultural variety.

My research is confined to a North American industry and Bourdieu's to a survey of France, so national differences may underlie some of the differences in our views. Lamont (1992) finds some differences in uses of culture by upper-class men in France and the United States and provides an insightful discussion of structural differences between the two coun- tries. But she too finds many reasons to modify Bourdieu's analysis, even for France itself. We need more comparative research that includes the two structural themes stressed in this article: social networks and class relations in specific fields.

Two lines of questioning are especially important. First, just how do networks contribute to cultural resources? I have argued that contact American Journal of Sociology

with culturally varied people naturally teaches varied culture as people talk about their interests, but we know little of the processes involved. Moreover, we need more work on network variety itself. I have inter- preted my network variety measure in terms of contact with people in widely varied class locations, but the locations measured could be read as class fractions in Wright's sense or in Bourdieu's, or they could be read as measures of exposure to varying occupational subcultures. My own argument suggests that we should take class or occupational variety as just a beginning; we need work on contacts with various ethnic groups, contacts through men and through women, and so forth. Second, how are cultural resources put to work in social fields of many kinds? There is an enormous amount of qualitative work needed just for fields in the working world: How are cultural resources deployed in interactions between equals and nonequals, in different kinds of industries and mar- kets? We also need parallel investigations of fields related to gender, ethnicity, and other forms of inequality and social cleavage.

REFERENCES

Anheier, Helmut K., Jurgen Gerhards, and Frank P. Romo. 1995. "Forms of Capital and Social Structure in Cultural Fields: Examining Bourdieu's Social Topogra- phy." American Journal of Sociology 100:859-903.

Bourdieu, Pierre. 1984. Distinction, translated by Richard Nice. London: Routledge & Kegan Paul.

Breton, Raymond, Wsevelod W. Isajiw, Warren E. Kalbach, and Jeffrey G. Reitz. 1990. Ethnic Identity and Equality: Varieties of Experience in a Canadian City. Toronto: University of Toronto Press.

Burt, Ronald S. 1992. Structural Holes: The Social Structure of Competition. Cambridge, Mass.: Harvard University Press. Campbell, Gayle, and Bryan Reingold. 1994. "Private Security and Public Policing in Canada." Juristat Service Bulletin 14 (10): 1-18.

Collins, Randall. 1992. "Women and the Production of Status Cultures." Pp. 213-31 in Cultivating Differences: Symbolic Boundaries and the Making of Inequality, edited by Michkle Lamont and Marcel Fournier. Chicago: University of Chicago Press.

DiMaggio, Paul. 1987. "Classification in Art." American Sociological Review 52: 440-55. DiMaggio, Paul, and John Mohr. 1985. "Cultural Capital, Educational Attainment, and Marital Selection." American Journal of Sociology 90:1231-61. Erickson, Bonnie H. 1991. "What Is Good Taste Good For?" Canadian Review of Sociology and Anthropology 28:255-78. Fischer, Claude. 1982. To Dwell among Friends: Personal Networks in Town and City. Chicago: University of Chicago Press. Granovetter, Mark. 1973. "The Strength of Weak Ties." American Journal of Sociol- ogy 78:1360-80. . 1995. Getting a Job: A Study of Contacts and Careers, 2d ed. Chicago: University of Chicago Press. Hall, John R. 1992. "The Capital(s) of Cultures: A Nonholistic Approach to Status Situations, Class, Gender, and Ethnicity." Pp. 257-85 in Cultivating Differences:

Symbolic Boundaries and the Making of Inequality, edited by Mich2le Lamont and Marcel Fournier. Chicago: University of Chicago Press. Halle, David. 1984. America's Working Man: Work, Home, and Politics among Blue- Collar Property Owners. Chicago: University of Chicago Press.

Hunter, Alfred A. 1988. "Formal Education and Initial Employment: Unravelling the Relationships between Schooling and Skills over Time." American Sociological Review 53:753-65.

Killworth, Peter, Eugene Johnson, H. Russell Bernard, Gene Ann Shelley, and Chris- topher McCarty. 1990. "Estimating the Size of Personal Networks." Social Net- works 12:289-312.

Lamont, Michhle. 1992. Money, Morals, and Manners: The Culture of the French and the American Upper-Middle Class. Chicago: University of Chicago Press. Lamont, Mich2le, and Annette Lareau. 1988. "Cultural Capital: Allusions, Gaps, and Glissandos in Recent Theoretical Developments." Sociological Theory 6:153-68.

Lin, Nan. 1982. "Social Resources and Instrumental Action." Pp. 131-45 in Social Structure and Network Analysis, edited by Peter V. Marsden and Nan Lin. Beverly Hills, Calif.: Sage.

Lin, Nan, and Mary Dumin. 1986. "Access to Occupations through Social Ties." Social Networks 8:365-86.

Mannheim, Karl. (1928) 1952. "The Problem of Generations." Pp. 276-32 1 in Essays on the Sociology of Knowledge, edited by D. Kecskemeti. London: Routledge & Kegan Paul.

Marsden, Peter V. 1987. "Core Discussion Networks of Americans." American Socio- logical Review 52:122-31. Peterson, Richard A. 1992. "Understanding Audience Segmentation: From Elite and Mass to Omnivore and Univore." Poetics 21:243-58.

Peterson, Richard A., and Albert Simkus. 1992. "How Musical Tastes Mark Occupa- tional Status Groups." Pp. 152-86 in Cultivating Dtyerences: Symbolic Boundaries and the Making of Inequality, edited by Michhle Lamont and Marcel Fournier. Chicago: University of Chicago Press.

Shearing, Clifford D., Margaret B. Farnell, and Philip C. Stenning. 1980. Contract Security in Ontario. Toronto: University of Toronto, Centre of Criminology. Simmel, Georg. 1955. "The Web of Group Affiliations." Pp. 125-95 in Conflict and the Web of Group Affiliations. Glencoe, Ill.: Free Press. Spaeth, Joe L., and Diane P. O'Rourke. 1994. "Designing and Implementing the

National Organizations Study." American Behavioral Scientist 372372-90.

Swidler, Ann. 1986. "Culture in Action." American Sociological Review 51:273-86.
Useem, Michael. 1984. The Inner Circle: Large Corporations and the Rise of Business
Political Activity in the US, and the U.K. New York: Oxford University Press.

Wellman, Barry. 1982. "Studying Personal Communities." Pp. 61-80 in Social Structure and Network Analysis, edited by Peter V. Marsden and Nan Lin. Beverly Hills, Calif.: Sage.

Wright, Eric Olin. 1985. Classes. London: Verso. Wright, Eric Olin, and Donmoon Cho. 1992. "The Relative Permeability of Class Boundaries to Cross-Class Friendships." American Sociological Review 57:85-102.

Comments
  • Recommend Us