The Blackwell Companion to Sociology (76 page)

people get more autonomous jobs and maintain better health. Erickson (1996)

shows that people who know others in a wider range of occupations also know

more about each of a variety of cultural genres: artists, books, sports, business magazines, restaurants. On the one hand, diverse networks act as the continuing

adult education school of culture: varied contacts have varied interests and talk to you about them. On the other, knowing at least about a lot of things helps you to find something in common with a wider range of the people you meet, and

hence to add more variety to your network. The cultural diversity that varied

networks help to teach is another valuable network benefit, since people in

higher level work need to be able to deploy different kinds of culture for different purposes, such as discussing the latest Harvard Business Review with fellow

managers to show command of business culture, but discussing sports with co-

workers of all levels to smooth cross-class relationships through topics popular from high to low (Erickson, 1996).

We have done enough to know that network diversity is important; but given

that, it is surprising how narrow is the research done so far. The study of

diversity is largely limited to the variety of different occupations in which a

person knows someone. For example, one may ask ``Do you know anyone at all

who is a lawyer? Anyone who is a plumber?'' and so forth. Pionered by Lin, this

is a valuable approach because it allows us to by-pass the huge problems of

studying people's weak ties. Instead, we ask about what matters in weak ties (at least in Lin's theoretical approach): to what kinds of resources do weak ties

324

Bonnie Erickson

connect a person? Since occupation is a master role in modern society, it is an

important indication of structural location, and of access to resources of many

kinds. But since modern society is complex and includes multiple forms of

important inequalities and differences, looking at occupation alone cannot

suffice. What about ties to people in different ethnic groups, or ties to men

and women? Almost everything we know about these is limited to close ties. Yet

the gender and ethnicity of weak ties must matter too. For example, men and

women tend to join voluntary associations of different kinds, and to work in

gender segregated jobs, and hence to meet people of their own gender. Women

tend to join smaller voluntary associations and to work in lower-level jobs, so

their networks become smaller as well as female dominated. These networks

then lead women into yet more jobs and other settings dominated by women.

Such feedbacks between networks and locations lead to growing occupational,

network, and other differences over the life course (Smith-Lovin and McPher-

son, 1993).

Analysis of this kind helps us to understand what it is about gender that makes

it important (gender differences in work, network, culture, and so forth), which is a great improvement over the all too common practice of using gender as a

variable in some statistical analysis and then speculating about why the observed effects occur. More generally, network analysis helps us to directly study the

kinds of things that actually matter in social life, such as concrete social

structures and the processes that go on within them, instead of relying on

individual attributes that are often just crude proxies for the things we should be studying.

A review of the network benefits described above will help to illustrate

another important benefit of network analysis: clarification of important but

fuzzy structural concepts, which are endemic in sociological discussions. Social capital is one very trendy example. Broadly, social capital is the good things

about social networks, the aspects of social networks that produce desirable

results for people or for groups. But we have just seen that different aspects of networks are good for different outcomes: stronger ties for help that takes

willingness but weaker ties for help that takes access to resources, kin for some kinds of help and neighbors for others, and so on. Moreover, every network

feature produces both desirable and undesirable results: our intimates may help

us or may grieve us by getting ill or betraying us; a diverse network may enrich our opportunities and our culture, but take time to maintain or create conflicting demands; density may help a politicized group to achieve unity of opinions, or

may encourage a slew of equally well placed people to fight for leadership, as for the Florentine oligarchs. Thus there is not such a thing as social capital singular, and we need to carefully work out which aspects of networks have which

outcomes (bad as well as good) under which conditions. Such conceptual refine-

ment is bad for our ability to create catchy sound bites such as ``social capital is declining,'' but good for getting us beyond a muddling of many different possible forms of social capital, as in Coleman's (1988) much cited but preliminary essay.

More generally, network analysts have already developed a great many useful

and precise structural concepts, while the network analysis orientation compels

Social Networks

325

us to get specific and clear about any new structural thought we may be trying to have.

Returning yet again to rival accounts of the strength of weak ties brings us to

the fascinating field of micro±macro integration in network terms. By any

account, the structural point of weak ties is that they are the part of a personal network that connects someone to a wider social structure in the most advantageous ways. Accounts differ in the kind of wider social structure they assume,

and hence in the ways that weak ties give advantage. Granovetter describes

separated pockets of resources linked only by weak ties, while Lin describes a

hierarchy of occupational prestige in which upward-leading ties are weak ties.

But we actually know rather little about overall social structure in network

terms, and about how personal networks plug into this wider picture.

Simmel (1955) provides the most ambitious and thought-provoking set of

speculations about how personal networks and social structure have connected

in changing ways in different societies and different times. The prime mover in

his story is the complexity of the overall social structure. If people live in small repetitive units like villages, all the members of each village have much the same network with the available variety of people: everyone in town. But in a complex modern society, each person can seek out a special social circle for every taste or necessity: fellow mechanics, fellow birdwatchers, fellow conservatives, fellow

neighbors, and so on and so on. Each social circle shares one limited thing in

common, and everyone in the circle is in that sense alike, more alike than the

village circle can be. Each person lives in the intersection of a unique combination of social circles, and thus in a unique social network of identities and

influences, the opposite of the shared networks of villagers. The modern (or

even postmodern) condition is thus one of individualism, but is not the indi-

vidual's doing, being instead the result of massive elaborations of the division of labor and societal complexity. Simmel's work is unusual in its attention to

history; too much of our work is very much limited to the present. Even our

studies of the present are usually static snapshots of social structure with no

attention to how networks change from year to year (though see Suitor et al.,

1997).

Our incomplete knowledge of macro±micro network linkages is all the more

important because we believe that networks are a major part of how overall

social divisions and forms of inequality are reproduced over time. If the rich get richer and the poor get poorer (as has unfortunately been true in North America

recently), and if in general those on top tend to stay on top, it is in part because advantaged social location provides every kind of network benefit needed to

increase one's chances of staying happy, healthy, and wealthy. Those with better work and education have larger networks of close ties to higher status yet more

varied people, and have larger overall networks with more variety of access to

many kinds of resources. If people of a certain ascribed kind (men, women,

members of ethnic groups, etc.) tend to keep getting into the same kinds of work, this is in part because a lot of hiring is done through networks and people know people like themselves. Tilly (1998) offers a fine analysis of such durable inequalities and the role of networks in them.

326

Bonnie Erickson

We have come full circle, from the study of whole networks to the study of

actor-centered networks to the need to integrate both kinds of work. There has

only been room, in this brief essay, to hint at the immense richness of network

analysis. To learn more, one could read a good introductory book such as Scott

(1991) or Degenne and Forse (1999). For network numbers, the definitive hand-

book and text is Wasserman and Faust (1994). The journal Social Networks is a

rich source of current work. But network analysis is not limited to such

obviously network-centered publications. It is everywhere now, in journals and

books of all substantive persuasions. Enjoy it as you find it, and consider adding it to your own repertoire.

23

Networks and Organizations

David Knoke

Organizational networks are increasingly pervasive social phenomena reaching

into a variety of social contexts. They span the gamut from multinational

corporate strategic alliances through community small enterprise systems to

mentoring bonds between employees and their supervisors. Unraveling the com-

plex economic, political, and interpersonal connections among organizations

and their participants poses serious theoretical and methodological challenges.

Over the past two decades, an explosive growth in scholarly research on organ-

izational networks has moved the core concepts from suggestive metaphor to

proto-theory, with powerful analytic tools at its disposal (for useful literature reviews see Alter and Hage, 1993; Knoke and Guilarte, 1994; Wasserman and

Galaskiewicz, 1994; Grandori, 1998). Three main research themes emerge from

this flourishing interdisciplinary field. First, many empirical researchers emphasize the development of rigorously defined concepts and precisely measured

structural dimensions of organizational networks. Second, some analysts seek

to specify the origins of network characteristics, in particular the formation and transformation of ties among actors. Third, other investigators examine the

impacts of organizational networks on their members' attitudes and behaviors

and try to explain the consequences for larger social systems within which

multilevel organizational networks are embedded. This chapter offers an inter-

pretive overview of these themes, emphasizing opportunities for deepening our

already considerable knowledge about organizational networks.

The organizational network perspective is one manifestation of more general

structural approaches to explaining social action. Structural action theorists

explicate behavior in terms of patterned social relationships among actors rather than by analyzing actors' statuses or such internal psychological states as perceptions, attitudes, and beliefs. A key structural action assumption is that the

totality of connections among multiple actors in a particular social system jointly 328

David Knoke

affects the behaviors of both its individual members and the system as a whole.

For example, small group theories of social action suggest that work team

productivity primarily depends on neither employees' education levels nor

their achievement drives, but on whether supportive or antagonistic interper-

sonal relations sustain or fragment group solidarity. Further, the actions of

individual team members will also vary according to their structural locations

within the network of interpersonal ties, with those employees who are more

socially integrated into the team outperforming the more isolated workers.

Similar emphases on patterned connections apply at more macrostructural

levels; for example, to explain the investment and borrowing practices of

banks and firms by analyzing their locations within networks of interlocking

corporate directorates. The primacy of structural ties among social actors infuses the organizational network perspective with a vision distinct from other prominent organizational theories, such as transaction cost economics, organizational

ecology, evolution, institutionalism, and resource dependence.

Both organization studies and network analysis are increasingly interdisciplin-

ary enterprises. Investigating their central problems draws concepts, theories,

data, and methods from the basic disciplines of sociology, economics, history,

political science, anthropology, and such applied fields as law, management, and public administration. Although comprehensive detailed coverage of such a vast

subject is impossible in this brief chapter, I sketch the broad contours of the

organizational network perspective. Following a brief presentation of core net-

work concepts and methods, I examine the application of network principles to

three important topics: interorganizational economic behavior, interorganiza-

tional political action, and intraorganizational networks. I conclude with some