You are on page 1of 23

A Network Theory of Social Capital

Nan Lin
Duke University
nanlin@duke.edu

January 2005

To appear in Handbook on Social Capital, edited by Dario Castiglione,


Jan van Deth and Guglielmo Wolleb, Oxford University Press

Lin

The concept of social capital has captured the imagination and attention of a
wide range of scholars and professionals in diverse disciplines and practical arenas. Since
the notion of social capital has generated multiple definitions, conceptualizations and
empirical measurements, the continued diversity in such usages without integration may
undermine and ultimately bring its downfall as a rigorous scientific concept and theory
and approach in social analysis. The purpose of this chapter is to describe a networkbased theory of social capital and how we may resolve several critical issues: linking
social network features to social capital, studying both access and mobilization of social
capital, measuring social capital, linking micro- and macro-level analysis, and
differentiating homophilous and heterophilous social capital.

Definition and Theory


To gain a better understanding of social capital, it is necessary to place it in the
context of different theoretical types of capital (Lin, 2001b). Capital, first of all, is
both a concept and a theory.i As a concept, it represents investment in certain types of
resources of value in a given society. As a theory, it describes the process by which
capital is captured and reproduced for some sort of return. For example, in the classical
theory of capital, Marx defines capital as part of surplus value created in a production
process. He also describes it as a process in which those controlling the means of
production capture the surplus value, including capital, through their taking for
themselves the difference in values generated in the production market -- where labor is
paid the lowest possible wage -- and those generated in the trade and consumption
markets (Lin 2001 : Chapter 1) where the produced commodity is priced for higher value.
Neo-capitalist theories offer a similar definition of capital but different theories. The
human capital theory, for example, postulates that investment in certain human resources
(skills and knowledge) may also generate economic returns, even for laborers
participating in the production market. Likewise, social capital theory conceptualizes
production as a process by which surplus value is generated through investment in
social relations (Lin 2001: 2).
Social capital is defined as resources embedded in ones social networks,
resources that can be accessed or mobilized through ties in the networks (Lin 2001 :

Lin

Chapter 2). Through such social relations or through social networks in general, an actor
may borrow or capture other actors resources. These social resources can then generate
a return for the actor. The general premise that social capital is network-based is
acknowledged by all scholars who have contributed to the discussion (Bourdieu, 1980;
1983/1986; Burt, 1992; Coleman, 1988; 1990; Cook, forthcoming ; Erickson, 1995;
1996; Flap, 1994; 1991; 1994; Lin, 1982; Portes, 1998; Putnam, 1993; 1995a; 2000).
However, various conceptual and operational analyses then diverge from this
point of agreement. For some analysts, social networks, and certain features of the
networks (e.g., closure), are social capital. For others, especially those working at the
more macro-level (e.g., civic engagement), social capital assumes a metamorphosed ,
more generalized, form along the lines of voluntary associations, non-governmental
organizations (NGO), or generalized trust in others. The conceptual linkage between
network-based social capital and these macro-forms has never been made clear.
To resolve these confusions and potential conflicts, I propose staying close to the original
conceptual understanding of social capital, as defined by Coleman, Bourdieu, Lin, and
others. As already indicated, they have defined social capital as resources embedded in
social networks , or resources that can be accessed or mobilized through ties in the
networks (Lin 2001 : Chapter 2). A theory of social capital describes the process by
which such resources are captured through investment in social relations for returns on
that investment.
Social capital thus defined allows us to formulate theoretical propositions for
identifying the sources of social capital and the returns to social capital. Elsewhere (Lin
2001) I have identified three principal sources (exogenous variables) for social capital:
(1) structural positions (actors position in the hierarchical structure of social
stratification -- the strength-of-position proposition), (2) network locations -- (actors
location in the networks that exhibit certain features, such as closure or openness, or
bridging, as illustrated in the strength-of-tie propositions), and (3) purposes of action
(instrumental or expressive motivations) (Lin 2001: Chapter 5). Two types of returns
have also been identified: instrumental (wealth, power, and reputation) and expressive
(cohesion, solidarity, and well-being). Propositions, then, link these sources and types of
return on social capital in causal sequences. Further elaboration of sources and returns

Lin

becomes possible and feasible (Lin, forthcoming-a; e.g., see Cook, forthcoming for
conceptualizing trust as a contingent condition for social capital).
In the remainder of the paper, I will address several issues that are essential for
furthering research on a network-based theory of social capital: (1) how social capital can
and needs to be differentiated from social networks, (2) whether social capital refers
to the accessible pool of embedded resources or just to the embedded resources actually
mobilized for a particular action, (3) how social capital should be measured, (4) what
conceptual basis can be used to elevate the network-based social capital into collective
social capital, and (5) how to resolve the differential utility of homophilous versus
heterophilous social capital.

Social Networks and Social Capital


While social capital is contingent on social networks, they are not equivalent or
interchangeable terms. Networks provide the necessary condition for access to and use of
embedded resources. Without networks, it would be impossible to capture the embedded
resources. Yet networks and network features by themselves are not identical with
resources. Rather, variations in networks or network features may increase or decrease
the likelihood of having a certain quantity or quality of resources embedded. Thus,
network features should be seen as important and necessary antecedents exogenous to
social capital. For example, for a given network, density or closure of networks may
increase the sharing of resources among participants as individuals and/or as a group
(Bourdieu, 1980; 1983/1986; Coleman, 1990: Chapter 12). On the other hand, sparse or
open networks may facilitate access to better or more varied resources or information,
control or influence (Burt, 2001; Lin, 1999a).
Thus, equating dense or closed networks with better or greater amount of social
capital is conceptually flawed. What is needed is to specify conditions under which
network features such as density or openness lead to the capturing of certain resources
that generate certain kinds of returns. Elsewhere (Lin, 2000), I have argued that once
network features (closed or open) are treated as exogenous variables, modeling of the
social capital process may proceed to specify how features of networks (e.g., closed or

Lin

open), social capital (e.g., diversity of embedded resources), and returns (instrumental or
expressive) form a sequential set of variables for analysis.
Likewise, the notions of bridging and associated concepts (e.g., strength of ties
and structural holes) are network concepts that should be seen as sources rather than
constitutive elements of social capital. A bridge links two clusters or blocks of actors and
allows possible flow of information or control and influence through the gate-keeping
actors. However, whether and to what extent a bridge brings added value (i.e., capital) to
the gate-keeping actor or a block of actors depend on the resources embedded in the other
linking actor or block of actors (Lin 2001: 69 71). As a proposition, we may argue that
bridging, as a network feature, increases the likely availability of such added value
social capital -- but bridging by and of itself cannot and should not be considered as
social capital, as some bridges may not in fact bring capital to the gate-keeping actor or
the block utilizing such bridging.

Access and Mobilization


While the definition of resources embedded in social networks provides a
framework, any theory of social capital needs further specification about what kind and
amount of investing entails. Extent of investment in any type of resources for expected
returns, as suggested in the various theories on capital, may be estimated in two ways.
(1) Degree of access to such resources is the extent to which a potential pool of resources
capable of generating returns is available to the actor. It indicates the capacity of capital.
For example, for human or cultural capital, educational experience may reflect such
access. Amount and type of education represents an individuals build-up of potential or
capacity of resources, whether skills and knowledge, or behavioral and ideological guides
provided by particular norms and values. Increase in education, therefore, represents
increased human or cultural capital capacity. (2) Social capital also may be measured
directly by the actual capacity of resources embedded in social networks. An assessment
or inventory of resources in the social networks of an actor -- accessible or embedded
resources -- reflects such capacity.
A possible distinction between access and embeddedness is therefore in order
here. Some scholars, including this writer, at times have followed the convention of

Lin

using the term embedded resources to represent the capacity or pool of resources
embedded in the social networks, while at other times, they have used the term access
(Lin, 1999a) instead. Embeddedness applies more appropriately to the description of
the pool of resources in a social network, from a structural or gestalt perspective. An
inventory of all or representative resources in a complete network reflects or measures
the embedded resources. Access more appropriately applies to an actors conscious
cognitive map or knowledge of such embedded resources. A network may embed certain
resources not present in the cognitive map of an actor. Such resources therefore cannot be
determined by asking the actor, even though they are embedded in the actors overall
network. So if the analysis concerns all the pooled resources of a network as a whole
(e.g., in an organization), embeddedness may be appropriate (Granovetter, 1985)
whereas if the analysis concerns actors (whether individuals or collectivities ) awareness
of resources embedded in their ties or networks, access would be more appropriate.
On the other hand, actual use of the accessed resources in production or
consumption in the marketplace represents a selection of specific ties and resources from
the resource pool for a particular action at hand. For example, presenting a diploma,
curriculum vitae, or degree certificate in applying for a specific job opportunity indicates
mobilized human capital. The particular choice of document does not reflect the entire
applicants human resources pool; in fact, it may not even be the optimal choice from the
resource pool for the specific action at hand. Nevertheless, it makes use of a particular
aspect or portion of the human resource pool for a particular action. Likewise, the use of
specific social ties to help in a job search represents the mobilization of particular
resources in the social resources pool. They too may or may not be the optimal choices
for the action at hand. Nevertheless, they are a choice from among the pool of social
connections and their resources accessible to the actor.
Both access and mobilization of resources have appeared in all theories on capital.
While mobilization and use of resources may seem a more direct way to assess the effects
of capital, it is nevertheless only a partial representation of the capital pool and its
capacity. Unless the capital pool is also independently estimated, reliance on only a
particular mobilization and/or use of resources may underestimate the utility of available
capital. For example, social capital may not be adequately reflected in a particular use of

Lin

specific overt social ties to help in a job search, since other accessed resources and social
ties may be involved in latent or invisible ways in the marketplace. Access to rich and
diverse social network resources may in fact routinely facilitate unsolicited job
information from others, which may eventually become critical in getting a better job,
without the actor actually searching for that or indeed any job (Lin, 2003a). Relying
solely on data on specific job contacts in the job-search process will likely miss a
significant portion of the invisible hand of and returns to social capital. Again, the
invisible return to invested resources is not unique to social capital; human, cultural, and
other types of capital also operate in both visible and invisible ways. Therefore, in
research, access to and use of social capital should be both measured and closely
examined, if possible.

Measurement
Access to social capital has traditionally been measured with a name-generating
methodology. Typically, a question is posed, such as, Whom do you usually discuss
work problems with? and a sampled respondent is asked to provide a list of names of
those who provide such services or exchanges. Further questions about the
characteristics of the named (name interpreters), as well as relationships among them and
between the respondent and each of them, provide data for reconstructing the density of
the network, and for estimating the quantity and/or quality of social resources (e.g.,
socioeconomic statuses) of those named.
However, this name-generating methodology has several limitations. First, the
content universe from which a particular question (e.g., work discussion) is drawn is
usually undefined or unknown to the researcher. Sometimes multiple questions are posed
to capture multiple content areas (Fischer, 1977; Wellman, 1979). Since the universe is
unknown, it is difficult to argue that such questions representatively sample a particular
universe. Second, the number of names generated is limited, typically ranging from only
three to five. Therefore the reconstructed network is of limited range and scope. Some
studies have tried to overcome these limitations by leaving the list open-ended (Wellman,
1982). However, such an approach is costly, time-consuming, and impractical for coding
in larger-scale surveys. Finally, since the names that come to the respondents mind

Lin

usually are those with stronger relationships to the respondent, the resources in the
captured pool tend to be homogeneous and relations homophilous relative to the
respondent. As research has demonstrated and argued, weaker and bridging ties to other
parts of the social structure may nevertheless be critical (Burt, 1992; Granovetter, 1974;
Lin, 1982). Missing data on such potential links to other levels of a social hierarchy may
underestimate, for example, the utility of an individuals social capital for instrumental
purposes, such as social mobility.
An alternative methodology has recently appeared (Lin & Dumin, 1986; Lin, et
al., 2001a). The position-generating methodology systematically samples a list of
positions in a social hierarchy (e.g., ranked occupations in a society). By using
systematic sampling (e.g., equal intervals) or stratified sampling (e.g., occupations
prevalent for different genders, ethnic/racial groups, or classes, each sampled occupation
is presented to a respondent, who is asked to indicate whether she/he knows anyone in
that sampled position. Since the rank distance is known between every pair of sampled
positions and among all the sampled positions, the responses to the set of positions can
then be used to estimate, with known measurement errors, the potential pool of resources
(i.e., in the occupational hierarchy) accessible to each respondent. Indexes (e.g., the total
number of accessed positions, the range or difference between rank scores of the highest
and lowest accessed positions, etc.) can be constructed to represent social capital, that is,
the capacity or pool of resources embedded in the respondents networks. Since such
access is not contingent on the strength of ties (which can be assessed relative to each
accessed position), it largely (but not completely) overcomes the tendency to evoke
homogenous or homophilous ties present in the social networks. Research over the past
decade in North America, Europe, and Asia has demonstrated that the position generator
methodology is reliable (across samples and over time) and valid (across different
communities and societies with respect to certain types of returns, for example, social
mobility) (Angelusz & Tardos, 1991; Erickson, 1996; Hsung, 1992; Volker & Flap, 1999;
for a summary , see Lin, 1999b; and for a collection of studies around the world see Lin
and Erickson, forthcoming-b), . This methodology has proven to be flexible,
allowing for variety in types of hierarchical positions (e.g., relative to social, political,

Lin

cultural, or economic resources) for different societies, populations, or returns, and for
incorporating additional dimensions for analysis (e.g., gendered or ethnic social capital).
It should be noted that the name generator and the position generator
methodologies also differ on another set of conceptual grounds. Name generating is
intended to create a list of individuals in egos networks, resulting in a sample of
respondents social ties and nodes in their networks: It is a person-focused methodology.
Position generating, on the other hand, canvasses the extent of access to structural
positions in a hierarchy: It is a structure-focused methodology. The name generator is
useful for identifying significant others in egos personal networks; whether they occupy
similar or different hierarchical positions is of secondary significance and interest. On the
other hand, the position generator is useful for assessing vertical reaches in the hierarchal
structure to which ego has access through social ties. How many persons there are or
how strong the relationship is at each accessed position is of secondary analytical
importance. In either case, further probing may yield additional information. For
example, the name generator may also reveal information about each named persons
socioeconomic characteristics and thus their structural positions. The position generator
may also reveal whether each accessed position has multiple occupants whom ego knows
and how close their relationship is. Nevertheless, in the case of the name generator this
additional information does not recover missing information about the range of
respondents contacts with various structural positions, in the case of the name generator,
the thickness of contacts with the full range of positions in the structure is probably
under-represented. Thus they represent alternative strategies, suited for different
conceptual purposes. The name generator is suitable for probing the depth of close ties,
whereas the position generator facilitates studying breadth of access to various levels of a
hierarchy.
Mobilization of embedded resources for a particular action is a complementary
rather than substitute measurement of access to embedded resources, as it inevitably
focuses on a particular and limited number of ties and their resources used in a particular
action. Research typically employs a critical-episode or - incidence approach to identify
the use of social capital. For example, a large body of research examines whether
personal contacts are used in job searches and whether the resources the contacts possess

Lin

(e.g., socioeconomic characteristics) make a difference in the likelihood of success or the


level of attained statuses. It is clear from the literature that mere use of any personal
contacts provides no relative advantage in the labor market. However, contact resources
that represent mobilized social capital do make a difference. That is, among those who
use contacts in a job search, those who mobilize contacts with better resources (better
socioeconomic characteristics or statuses) tend to obtain better jobs. This confirms the
significance of mobilizing embedded resources in the labor market.
Questions have been raised as to whether the lack of evidence for the use of social
contacts in many job searches suggests that social capital may be of little significance.
As mentioned earlier, however, absence of identified help may not reflect the lack of
utility of social capital. Current arguments and research show that job information can
flow in networks, especially networks rich with embedded resources, without any parties
actively seeking jobs or job information (Lin, 1999b; 2003a). Such flow and utility of
information and contacts may reflect the informal workings of social capital, or its
invisible hand.ii Thus, measuring the actual utility of social capital for returns in a
marketplace (be it instrumental or expressive) requires assessment of access and both
visible and invisible use of resources embedded in networks.
The measurement of contact resources as mobilized social capital has also been
criticized (Mouw, 2003) on the grounds that much of the effect (i.e., any association
between the contacts occupational status and respondents post-contact attained
occupational status) is due to the homogeneity effect (similarity between the contacts
occupation and the respondents attained occupation), rather than the contacts superior
status positively affecting respondents superior attained status. However, the theory of
social capital principally hypothesizes that it is the benefit of mobilized embedded
resources (contacts status) relative to the initial status of the jobseeker that should make
a difference the strength-of -position hypothesis. That is, it predicts that the contacts
relatively superior positions, in comparison to the job-seekers initial positions, should be
evidence of the utility of social capital. Indeed, from the same data set Mouw used to
demonstrate his argument (the Detroit study), even when those cases that showed
similarity between respondents initial occupations with contacts occupations were
removed from the sample (to eliminate the homogeneous ties), the positive association of

Lin

10

respondents original statuses with contacts statuses retains its significance. This means
that seeking, obtaining, and successfully utilizing contacts superior resources are
positively associated with obtaining better statuses.
The attained status represents improved status resulting from the utility of a
superior contact, thus closing the status distance between contacts status and
respondents initial status. This is not only not surprising, but even expected, as many of
the respondents ought to be now at a similar or approximate status level as compared to
that of the contacts themselves the general homogeneity principle applies to occupants
at comparable or horizontal level of positions.
Consider, for example, Fernandezs study of telemarketers who made referrals for
new hires (Fernandez & Weinberg, 1997). All successful referrals brought in new
telemarketers, thus achieving complete homogeneity between contact (referrers) status
and the job-seekers (referreds) newly attained status. This would reduce the remaining
observations for Mauws demonstration to zero. It is the status gap between the original
positions of successful applicants and their referrers (i.e., most of the referred probably
initiated with lower statuses than the telemarketer referrers) that attests to the utility of
social capital.
Thus, in measuring mobilized social capital for specific actions, it is important to
measure the initial and attained positions or statuses for ego as well as the positions or
statuses of contacts in order to reflect completely the process by which social capital
returns added value.

Inferring from Micro- to Macro-Level Analysis


One important concern about a network theory of social capital is that it is built
on micro-data and meso-relations, which raises questions about whether and how it can
be translated into macro-level analysis. Unless this is resolved, macro-level analysis
conducted for voluntary organizations and other collectivities at the societal level will
remain segregated conceptually from the micro-meso-level analysis. In fact,
conceptualizing this potential linkage is an opportunity for further understanding of the
micro-to-macro linkage, deemed so important in conceptualizing overall micro-macro
social processes (Coleman 1990: 8; Lin 2001: 186).

Lin

11

As defined here, social capital is conceived as resources embedded in social


networks, with the understanding that at the micro-meso level the richness and diversity
of such resources for individual actors depends on two meso-conditions: (1) the extent to
which resources are embedded among ties directly accessible to ego, and (2) the extent to
which additional resources can be accessed through ties of ties. The first condition
specifies resources ego can gain through direct ties within egos networks, and the second
condition specifies resources to which ego can have access through indirect ties, that is,
through ties as bridges to other networks. The first condition specifies how capital is
captured within egos immediate social circle, and the second condition suggests how
capital may be captured beyond egos immediate social circle. Egos social capital,
therefore, is the resources embedded in the immediate social circle and by those
embedded in egos extended circles. For example, in the context of an instrumental
action such as the job - search process, it has been found that the help chain may involve
a single helper (one intermediary in the chain) or multiple helpers (two more
intermediaries in the chain) (Bian, 1997; Lin, 2003b). In the former, we see the
mobilization of a tie in egos immediate social circle, whereas in the latter, tie chains in
extended social circles.
Such conceptual and empirical clarification provides the necessary foundation for
understanding how social capital operates at the macro-level. We may assume that
voluntary associations constitute social circles (Simmel, 1950) that bring various
members with certain shared interests together for one or more purposes. The association
has its own collective capital, including properties, endowments, liquid assets, equipment,
human resources, etc. Likewise, it has social capital. The amount of social capital
available to each collectivity can be estimated by two characteristics: (1) the extent to
which resources that could be brought to bear on the associations projects by its
members may in fact be offered by them as resources for the collectivity, and (2) the
extent to which the association, through its officials and members other social ties and
participation in other social circles and associations may also be connected to the
resources of such other social circles, including other informal voluntary associations as
well as more formal organizations. The first characteristic refers to resources embedded
in the association itself provided by its members, whereas the second involves resources

Lin

12

accessible through its own or its members connections to other social circles. Then it
follows that an associations richness and diversity of embedded resources, representing
social capital, can be assessed in terms of (1) resources directly available from members
in the association, and/or (2) resources accessible through the associations its members
connections to other associations and individuals.
This macro-level conceptualization of social capital resolves several crucial
issues. First, it dispels the general misconception that any or all associations are equally
endowed with capital. In fact, just like individual actors, collectivities also vary in terms
of their store of capital, including social capital. Second, the effectiveness of an
association theoretically depends on the amount of social capital represented by the
resources accessible from its members and by the resources of other associations and
organizations to which it connects through members and officials. Third, this micromacro conceptualization offers operational tools for examining and analyzing access and
mobilization of social capital by associations as themselves social actors, just as it does
for individual actors. Empirical research on social capital, therefore, entail the analysis of
an association s composition and the number of members and resources they can bring to
bear, as well as analysis of the members and the associations connections to other
associations and organizations , their embedded resources , and their connections
through their chains of relationships.
Some recent research lends initial empirical support to this micro-macro
conceptualization. Paxton (2002) found that nations with more interconnected
associations tend to show a greater likelihood of being a participatory democracies. Son
and Lin (2004) show that greater organizational diversity , as reflected in their members
diverse gender, education, and race/ethnicity , is associated with more network assets and
voluntary actions of the individual members.
From conceptualizing social capital for collectivities such as associations, it takes
only an additional step to similarly characterize social capital for a larger social unit, such
as a community, collection of associations and organizations, or whole society.
Analytically for each unit, collectivities within it become component members, and their
resources become embedded resources for the larger unit. The units connections to

Lin

13

other units (e.g., other communities and/or societies) and those units resources also
constitute part of the units social capital.
As conceptual and empirical work is extended to larger macro-units, fine-tuning
the theoretical processes must keep pace as well. As in micro-level analysis, social
capital -- the embedded resources in ones social networks -- depends on a number of
factors, such as the actors position in the hierarchical structure, his/her location in the
networks, and institutional and personal contingencies ; the macro-level analysis also
requires specification of such factors as the location of a units position in the more
global hierarchy, its location in the networks of units, and institutional and other unit
characteristics for these other implicated units. Concepts such as bridges, strength of ties,
structural holes, size of networks, density or openness of networks, and institutional
expectations for social relations all become important exogenous variables in the
modeling of social capital at multiple levels. The nature and elements of these variables
need to be operationalized in various ways, each dictated by the particular empirical
systems under analysis. But the underlying conceptual and theoretical formulations
should remain consistent and credible.
Such macro-level analysis, furthermore, extends beyond voluntary associations
and civic engagement. Social capital analysis of economic and other formal
organizations needs to be similarly conceptualized, as well, for which social capital
capacity and mobilization should likewise be theorized. Performance of an economic
organization, for example, can be hypothesized as linked to the recruitment, deployment,
and work assignment of agents with social connections and network resources to
positions that demand social skills, such as those toward the top of the authority
hierarchy, or at the edge of the organization where positions require contacts and
exchanges with external units (e.g., marketing, advertising, and public relations), as well
as positions that interact with humans rather than materials (Lin, 2004). Its performance
should also be contingent on the organizations connections to other organizations and
units with rich and diverse resources.

Heterophilous versus Homophilous Social Capital

Lin

14

Yet another conceptual issue deserves our attention: the differential utility of
social capital, or different types of social capital for instrumental or expressive purposes.
Two types of social capital can be identified relative to embedded resources in networks:
heterophilous or homophilous. It is generally assumed or hypothesized that homophilous
social capital enhances expressive actions and heterophily in social capital helps
instrumental actions. For Coleman (1990), Putnam (2000) and others, social capital is
conceived as a network with high density or closure because it promotes trust and
reciprocity, which, in turn, enhance collective solidarity, or expressive returns. Since
density and reciprocity are highly associated with homophily in characteristics and
resources among the participants (Burt, 1987; Homans, 1950; Marsden, 1988; Lazarsfeld
& Merton, 1954), this conception postulates an association between homophilous social
capital and expressive purposes.
On the other hand, in studying social mobility, economic performance, or social
movement, Burt (2001b), Lin (1999a) and others have argued that sparse or open
networks allow access to novel or new information and resources through bridges (e.g.,
weaker ties, structural holes). Such added resources help achieve instrumental goals (e.g.,
better jobs, higher salaries and bonuses, or changing institutional distribution of
resources). Thus, diverse resources embedded in networks, or heterophilous social
capital, is seen as useful for instrumental purposes.
The differential utility, instrumental or expressive, of social capital specified for
different types of social capital should not be seen as differential conceptualization of
social capital for different levels (i.e., micro-meso versus macro-) of analysis. Both types
of utility are meaningful for both micro-meso and macro-level analyses (Lin, 1982; Lin,
2001b) Chapter 4). For example, Putnam (2000) acknowledged both bonding and
bridging functions for civil engagement or a macro-level analysis of social capital.
Coleman (e.g., merchants in Cairos markets, 2000) and Lin (e.g., social support provided
by strong ties, 1982; Lin, Dean and Ensel 1986) have conceived expressive functions of
social capital for small-scale networks or individual actors. More importantly, it is false
to assume, as sometimes implied in the literature, that it is an either-or choice a network
can be either heterophilous, therefore good for instrumental purposes, or homophilous,
good for expressive purposes. A better conceptual articulation is needed.

Lin

15

The initial discussion on homophily (Fischer, 1982; Homans, 1950; Laumann,


1966; Lin, 1982; Lazarsfeld & Merton, 1954; Wellman, 1979) characterizes social ties:
that actors share certain characteristics, including lifestyles. Particular shared
characteristics or lifestyles can be identified as the focal elements in a specific analytic
context. For example, gender, education and race/ethnicity tend to be central elements in
describing ties as friends that is, friendships tend to form between actors of similar
gender, education, or racial/ethnic backgrounds. However, these ties may or may not
share other characteristics (e.g., age, marital status, religion, or occupations). Further,
homophily is a dynamic variable. Interacting ties may increase similarity with one
another in other characteristics and lifestyles (e.g., preferences for food, music and
movies) over time. Nevertheless, at any moment, ties may be homophilous in some
characteristics and heterophilous in other characteristics. The degrees of homophily and
heterophily between ties may be independent variables: some ties are homophilous across
many characteristics while other ties are homophilous on some characteristics, but not on
others. Further, some degree of homophily must be present for any tie by definition,
while the degree of heterophily may fluctuate widely across ties. That is, variance in
homophily should be smaller across ties than that in heterophily.
Similarly, a network may be homophilous over some characteristics, such as
gender, education, and/or race/ethnicity. Yet heterophily is likely also to be present in
the network. The degrees of homophily or heterophily in a network again are
independent. A network can be homophilous over most characteristics, and not
heterophilous; while another network may be homophilous over some characteristics, and
heterophilous over other characteristics, at the same time.
This discussion makes it clear that a simple dichotomized assertion that a tie or
network contains homophilous (or bonding) or heterophilous (or bridging) social capital
is unlikely to yield fruitful results. Two clarifications can improve our understanding of
these two types of social capital. First, we need to assess separately the extent of
homophilous and heterophilous social capital, for both individuals, at the micro-mesolevel, and for collectives, at the macro-level. We cannot assume that a high degree of
homophily over some characteristics implies a low degree of heterophily across all

Lin

16

characteristics. Homophily and heterophily vary across characteristics for specific ties
and networks.
A second clarification is the relative utility of the two types of social capital. We
may argue that each type of social capital has its own merit in generating either
expressive or instrumental returns. However, if the above discussion regarding (1) the
essential homophily over some characteristics for all ties and networks, and (2) the
relative greater variance in heterophily than in homophily has validity, then, we can make
the following as a working hypothesis: relatively, heterophily is more significant than
homophily in characterizing social capital. In this scenario, the assumption is that a
certain degree of homophily is present to any tie or network, which by definition requires
a sufficient amount of interaction and shared sentiment. Homophily takes precedence
over heterophily in ordinary and normative form of social relations. Actions to maintain
resources usually take precedent over actions to acquire additional resources (Lin 2001:
Chapters 3, 4, 8, and 9) because maintenance and stability of relations through homophily
is more important for resource utility than the heterophilous risks in trying to acquire new
resources. Any network must be homophilous to a certain extent, if it is to maintain a
sufficient degree of connectivity for access to and sharing and exchange of resources.
The expectation is that a network must contain certain shared characteristics or lifestyles
among its members. Even in a generally heterophilous network, there must be enough
homophily for meaningful contacts.
The reverse, however, may not be true. In fact, the extent of heterophily varies
across networks and can be the key to understanding how various networks generate
different levels and types of returns to members as well as to the collectivity. If ties or
networks, while maintaining interactions and sharing certain characteristics and resources,
nevertheless also encompass characteristics and resources that are different among the
actors, be they individuals or collectives, access to diverse information and resources is
enhanced for all ties or members in the networks. This is so because different
characteristics and resources imply that the actors may also engage in other ties and
networks, different from the focal ones. That is, ties and members become bridges by
bringing such new or different information and resources, through their other ties and

Lin

17

networks, to be shared. Therefore, what gives a tie or network a significant advantage, or


added value, is the extent of its members heterophily in characteristics and resources.
In this theoretical formulation, the crucial analysis is to ascertain the extent to
which heterophily is present in a network. Indeed, the presence of heterophily suggests a
greater likelihood that some of the members, engaged with other networks, bring more
diverse resources to the focal network, resources beyond those already shared by
members in this network, thus enriching the potential pool of resources available to the
network. There is, of course, a possible cost associated with wider heterophily in a
network, as this may weaken shared sentiment, interest, and thus interaction, the
foundation of solidarity and cohesion. It is nevertheless conceivable that if the network
consists of members who have more diverse characteristics and yet is able to maintain
sufficient homophily to allow contacts and networking with each other, then such a
network ought to benefit from its members having such mixed characteristics, by being
able to become both expressive and instrumental, stable and flexible, cohesive yet
dynamic.
This formulation has significant implications for understanding the utility of
social capital. Instead of looking for networks that carry homophilous or heterophilous
social capital, researchers should explore the extent of heterophily in social capital
(diversity in embedded resources) in ties and networks and ascertain their instrumental as
well as expressive utilities. That is, networks with heterophilous social capital should be
expected serve well both instrumental and expressive functions for the collectivity and its
members.
If we apply this hypothesis to voluntary associations conceived of as networks,
for example, then a crucial analytic focus is to examine the degree of diversity among the
members in a given association. The hypothesis suggests that the greater the diversity of
lifestyles and other characteristics among the members, the more effective the association
will be as means for access to or mobilizing of social capital (embedded resources) for
both instrumental and expressive purposes of the collectivity and its members (barring
any failure of exchange due to differences in characteristics, as outlined above). Thus, all
associations are expected to be cohesive to an extent, while some of them gain more
benefits due to the diversity of its participating members.

Lin

18

Summary
This essay introduces a network-based theory of social capital. Conceived as
embedded resources in social networks, social capital is seen as contingent on network
features such as closure or openness, bridging, etc., but not identical to these features. It
can and should be measured for both its capacity (accessed resources) and actual uses for
particular actions (mobilized resources). Recent development in the position-generator
methodology facilitates a research program that can now be based on precise theoretical
and measurement requirements. Further conceptual clarification also makes possible the
translating the network-based theory and methods from a micro-meso basis to the macrocontexts. The essay also postulates that it is the heterophily in social capital (diversity in
embedded resources) that may be the key feature in assessing the utility (both
instrumental and expressive) of social capital. These discussions, it is hoped, will set the
stage for further programmatic developments in social capital theory and research.

Notes: ------------------------------------------1

Applying both concept and theory to a term has been a common practice in the social

sciences as in the cases of classical Marxist theory, and of human and cultural capital, as
well as social capital theory.
2

For an economic formulation arguing for network effects, see

(Arrow & Borzekowski, 2004). In their formulation, the number of ties to different firms
-- an access indicator -- rather than particular mobilized ties is the estimator for network
resources.

Lin

19

REFERENCES CITED

Angelusz, Robert and Robert Tardos. 1991. "The Strength and Weakness of "Weak
Ties"." Pp. 7-23 in Values, Networks and Cultural Reproduction in Hungary,
edited by P. Somlai. Budapest: The Coordinating Council of Programs.
Arrow, Kenneth J. Arrow and Ron Borzekowski. 2004. "Limited Network Connections
and the Distribution of Wages," Finance and Economics Discussion Series, 200441, Board of Governors of the Federal Reserve System (U.S.)., August.
Bian, Yanjie. 1997. "Bringing Strong Ties Back In: Indirect Connection, Bridges, and Job
Search in China." American Sociological Review 62(3, June):366-85.
Bourdieu, Pierre. 1980. "Le Capital Social: Notes Provisoires." Actes de la Recherche en
Sciences Sociales 3:2-3.
_____. 1983/1986. "The Forms of Capital." Pp. 241-58 in Handbook of Theory and
Research for the Sociology of Education, edited by J. G. Richardson. Westport,
CT.: Greenwood Press.
Burt, Ronald S. Burt, Ronald S. 1987. "A Note on Strangers, Friends, and Happiness."
Social Networks 6:311-32.
_____. 1992. Structural Holes: The Social Structure of Competition. Cambridge, MA:
Harvard University Press.
_____. 2001. "Structural Holes Versus Network Closure as Social Capital." In Social
Capital: Theory and Research, edited by N. Lin, K. Cook and R. S. Burt.
Hawthorn, NY: Aldine de Gruyter.
Coleman, James S. 1988. "Social Capital in the Creation of Human Capital." American
Journal of Sociology 94:S95-S121.
_____. 1990. Foundations of Social Theory. Cambridge, MA: Harvard University Press.
Cook, Karen S. Forthcoming. "Networks, Norms and Trust: The Social Psychology of
Social Capital." Social Psychology Quarterly.
Erickson, Bonnie H. 1995. "Networks, Success, and Class Structure: A Total View."
Presented at the Sunbelt Social Networks Conference, February, Charleston, S.C.

Lin

20

_____. 1996. "Culture, Class and Connections." American Journal of Sociology 102(1,
July):217-51.
Fernandez, Roberto M. and Nancy Weinberg. 1997. "Sifting and Sorting: Personal
Contacts and Hiring in a Retail Bank." American Sociological Review 62
(December):883-902.
Fischer, Claude S. 1977. Networks and Places. New York: Free Press.
_____. 1982. To Dwell Among Friends: Personal Networks in Town and City. Chicago:
University of Chicago Press.
Flap, Henk D. 1991. "Social Capital in the Reproduction of Inequality." Comparative
Sociology of Family, Health and Education 20:6179-202.
_____. 1994. "No Man Is An Island: The Research Program of a Social Capital Theory."
Presented at the World Congress of Sociology, July, Bielefeld, Germany.
Granovetter, Mark. 1974. Getting a Job. Cambridge, MA: Harvard University Press.
_____. 1985. "Economic Action and Social Structure: The Problem of Embeddedness."
American Journal of Sociology 91:481-510.
Homans, George C. 1950. The Human Group. NY: Harcourt, Brace.
Hsung, Ray-may, Hwang. 1992. "Social Resources and Petit Bourgeois." Chinese
Sociological Quarterly 16:107-38.
Laumann, Edward O. 1966. Prestige and Association in an Urban Community.
Indianapolis: Bobbs-Merrill.
Lazarsfeld, Paul F. and Robert K. Merton. 1954. "Friendship as Social Process: A
Substantive and Methodological Analysis." Pp. 298-348 in The Varied Sociology
of Paul F. Lazarsfeld, edited by P. L. Kendall. NY: Columbia University Press.
Lin, Nan. 1982. "Social Resources and Instrumental Action." Pp. 131-45 in Social
Structure and Network Analysis, edited by P. V. Marsden and N. Lin. Beverly
Hills, CA: Sage.
_____. 1995. "Les Ressources Sociales: Une Theorie Du Capital Social." Revue
Francaise de Sociologie XXXVI(4, October-December):685-704.
_____. 1999a. "Building a Network Theory of Social Capital." Connections 22(1):28-51.
_____. 1999b. "Social Networks and Status Attainment." Annual Review of Sociology
25:467-87.

Lin

21

_____.

2000. "Social Capital: Social Networks, Civic Engagement or Trust?"


Introductory Presentation. Presented at the Workshop on Social Capital, 1920/October. Trento, Italy: University of Trento.

_____. 2001b. Social Capital: A Theory of Structure and Action. London and New York:
Cambridge University Press.
_____. 2003a. "The Invisible Hand of Social Capital." Presented at the Creation and
Returns of Social Capital, 10/29. Amsterdam: Royal Netherlenads Academy of
Arts and Sciences.
_____. 2003b. "Job Search in Urban China: Gender, Network Chains, and Embedded
Resources." Pp. 145-71 in Creation and Return to Social Capital, edited by H.
Flap and B. Volker. NY: Praeger.
_____. 2004. "Social Capital and Organizations." Presented at the Inaugural Conference,
International Association for Chinese Management Research, June, Beijing.
_____. Forthcoming-a. "Social Capital." In Encyclopedia of Economic Sociology, edited
by J. Beckert and M. Zagiroski. London: Rutledge.
Lin, Nan and Mary Dumin. 1986. "Access to Occupations Through Social Ties." Social
Networks 8:365-85.
Lin, Nan and Bonnie Erickson , editors. Forthcoming-b. Social Capital: Advances in
Research. NY: Transaction-de-Gruytal.
Lin, Nan, Yang-chih Fu, and Ray-may Hsung. 2001a. "Position Generator: A
Measurement for Social Capital." Pp. 51-87 in Social Capital: Theory and
Research, edited by N. Lin, K. Cook and R. S. Burt. New York: Cambridge
University Press.
Marsden, Peter V. 1988. "Homogeneity in Confiding Networks." Social Networks 10:5776.
Mouw, Ted. 2003. "Social Capital and Finding a Job: Do Contacts Matter?" American
Sociological Review 68 (December):868-98.
Paxton, Pam. 2002. "Social Capital and Democracy." American Sociological Review 67
(April):254-77.
Portes, Alexandro. 1998. "Social Capital: Its Origins and Applications in Modern
Sociology." Annual Review of Sociology 22:1-24.

Lin

22

Putnam, Robert D. 1993. Making Democracy Work: Civic Traditions in Modern Italy.
Princeton, N. J.: Princeton University Press.
_____. 1995a. "Bowling Alone: American's Declining Social Capital." Journal of
Democracy 6(1, January):65-78.
_____. 2000. Bowling Alone: The Collapse and Revival of American Community. NY:
Simon & Schuster.
Simmel, Georg (trans. and edited by Kurt H. Wolff). 1950. The Sociology of Georg
Simmel. Glencoe, Ill.: Free Press.
Son, Joon-mo and Nan Lin. 2004. "Diversity of Civic Organizations and Voluntary
Actions of Individuals." Presented at the Annual Meeting, The American
Sociological Association, August, San Francisco.
Volker, Beate and Henk Flap. 1999. "Getting Ahead in the GDR: Social Capital and
Status Attainment Under Communism." Acta Sociologica 41(1, April):17-34.
Wellman, Barry. 1979. "The Community Question: The Intimate Networks of East
Yorkers." American Journal of Sociology 84:1201-31.
_____ 1982. "Studying Personal Communities." Pp. 61-80 in Social Structure and
Network Analysis, edited by P. V. Marsden and N. Lin. Beverly Hills: Sage.
i

Applying both definition and theory to a term has been a common practice in social

sciences. It is also true in the cases of the classical Marxist theory, human capital,
cultural capital as well as social capital.
See a recent essay on how trust may be a contingent condition for social capital,
Cook, forthcoming.
ii

For an economic formulation in arguing for network effects, see (Arrow &

Borzekowski, 2004). In this formulation, number of ties to different firms, an


access indicator, rather than specific mobilized ties, is the estimator for network
resources.

Lin

23

You might also like